# **VIBRIO ECOLOGY, PATHOGENESIS, AND EVOLUTION**

# **Topic Editors Daniela Ceccarelli and Rita R. Colwell**

#### *FRONTIERS COPYRIGHT STATEMENT*

© Copyright 2007-2014 Frontiers Media SA. All rights reserved.

All content included on this site, such as text, graphics, logos, button icons, images, video/audio clips, downloads, data compilations and software, is the property of or is licensed to Frontiers Media SA ("Frontiers") or its licensees and/or subcontractors. The copyright in the text of individual articles is the property of their respective authors, subject to a license granted to Frontiers.

The compilation of articles constituting this e-book, wherever published, as well as the compilation of all other content on this site, is the exclusive property of Frontiers. For the conditions for downloading and copying of e-books from Frontiers' website, please see the Terms for Website Use. If purchasing Frontiers e-books from other websites or sources, the conditions of the website concerned apply.

Images and graphics not forming part of user-contributed materials may not be downloaded or copied without permission.

Individual articles may be downloaded and reproduced in accordance with the principles of the CC-BY licence subject to any copyright or other notices. They may not be re-sold as an e-book.

As author or other contributor you grant a CC-BY licence to others to reproduce your articles, including any graphics and third-party materials supplied by you, in accordance with the Conditions for Website Use and subject to any copyright notices which you include in connection with your articles and materials.

All copyright, and all rights therein, are protected by national and international copyright laws.

The above represents a summary only. For the full conditions see the Conditions for Authors and the Conditions for Website Use.

**ISSN** 1664-8714 **ISBN** 978-2-88919-289-2 **DOI** 10.3389/978-2-88919-289-2

### *ABOUT FRONTIERS*

Frontiers is more than just an open-access publisher of scholarly articles: it is a pioneering approach to the world of academia, radically improving the way scholarly research is managed. The grand vision of Frontiers is a world where all people have an equal opportunity to seek, share and generate knowledge. Frontiers provides immediate and permanent online open access to all its publications, but this alone is not enough to realize our grand goals.

### *FRONTIERS JOURNAL SERIES*

The Frontiers Journal Series is a multi-tier and interdisciplinary set of open-access, online journals, promising a paradigm shift from the current review, selection and dissemination processes in academic publishing.

All Frontiers journals are driven by researchers for researchers; therefore, they constitute a service to the scholarly community. At the same time, the Frontiers Journal Series operates on a revolutionary invention, the tiered publishing system, initially addressing specific communities of scholars, and gradually climbing up to broader public understanding, thus serving the interests of the lay society, too.

### *DEDICATION TO QUALITY*

Each Frontiers article is a landmark of the highest quality, thanks to genuinely collaborative interactions between authors and review editors, who include some of the world's best academicians. Research must be certified by peers before entering a stream of knowledge that may eventually reach the public - and shape society; therefore, Frontiers only applies the most rigorous and unbiased reviews.

Frontiers revolutionizes research publishing by freely delivering the most outstanding research, evaluated with no bias from both the academic and social point of view.

By applying the most advanced information technologies, Frontiers is catapulting scholarly publishing into a new generation.

### *WHAT ARE FRONTIERS RESEARCH TOPICS?*

Frontiers Research Topics are very popular trademarks of the Frontiers Journals Series: they are collections of at least ten articles, all centered on a particular subject. With their unique mix of varied contributions from Original Research to Review Articles, Frontiers Research Topics unify the most influential researchers, the latest key findings and historical advances in a hot research area!

Find out more on how to host your own Frontiers Research Topic or contribute to one as an author by contacting the Frontiers Editorial Office: researchtopics@frontiersin.org

## **VIBRIO ECOLOGY, PATHOGENESIS, AND EVOLUTION**

Topic Editors:

**Daniela Ceccarelli,** University of Maryland, USA **Rita R. Colwell,** University of Maryland and Johns Hopkins University,USA

Epifluorescent micrograph of V. cholerae labelled with fluorescent monoclonal antibodies (DFA) specific for *V. cholerae* serovar O1 (New Horizon Diagnostic, MD, USA). Image prepared by Daniela Ceccarelli

Vibrios are Gram-negative bacilli that occur naturally in marine, estuarine, and freshwater systems. Some species include human and animal pathogens capable of causing gastroenteritis, wound infections, cholera, and fatal septicemia.

Over the past decades, cutting edge research on Vibrio genomics has promoted a tremendous advance in our knowledge of these pathogens. Significant developments include the discovery of emerging epidemic clones, tracking the spread of new strain variants, and an intensified appreciation of the role of mobile genetic elements in antibiotic resistance spread as well as pathogenesis. Furthermore, improved

understanding of the interaction of Vibrios with a variety of living organisms in the aquatic environment has documented the significant role of environmental reservoirs in their seasonal cycle favoring persistence of the pathogen during inter-epidemic periods and enhancing disease transmission.

This Research Topic is dedicated to our current understanding in these areas and will bring together leading experts in the field to provide a deep overview of Vibrios ecology and evolution, and will suggest the pathway of future research in this field.

# Table of Contents

### *05 Vibrio Ecology, Pathogenesis and Evolution* Daniela Ceccarelli and Rita R. Colwell


Nicolas Carraro, Maxime Sauvé, Dominick Matteau, Guillaume Lauzon, Sébastien Rodrigue and Vincent Burrus

*48 Vibrio Cholerae O1 Epidemic Variants in Angola: A Retrospective Study Between 1992 and 2006*

Romy Valia, Elisa Taviani, Matteo Spagnoletti, Daniela Ceccarelli, Piero Cappuccinelli and Mauro M. Colombo

*54 Photobacterium Damselae Subsp. Damselae, a Bacterium Pathogenic for Marine Animals and Humans*

Amable J. Rivas, Manuel L. Lemos and Carlos R. Osorio

*60 The Function of Integron-Associated Gene Cassettes in Vibrio Species: The Tip of the Iceberg*

Rita A. Rapa and Maurizio Labbate


Pimonsri Mittraparp-Arthorn, Janelle Renee Thompson, Varaporn Vuddhakul and Gary J. Vora


Kristi S. Shaw, John M. Jacobs and Byron C. Crump

*190 Vibrio Cholerae Interactions With Mytilus Galloprovincialis Hemocytes Mediated by Serum Components*

Laura Canesi, Elisabetta Pezzati, Monica Stauder, Chiara Grande, Margherita Bavestrello, Adele Papetti, Luigi Vezzulli and Carla Pruzzo

*196 Adaptation of a Simple Dipstick Test for Detection of Vibrio Cholerae O1 and O139 in Environmental Water*

Subhra Chakraborty, Munirul Alam, Heather M. Scobie and David Allen Sack

*199 Erratum: Adaptation of a Simple Dipstick Test for Detection of Vibrio Cholerae O1 and O139 in Environmental Water*

Subhra Chakraborty, Munirul Alam, Heather M. Scobie and David Allen Sack


Alison F. Takemura, Diana M. Chien and Martin F. Polz

*235 How Community Ecology Can Improve our Understanding of Cholera Dynamics* Guillaume Constantin de Magny, Nur A. Hasan and Benjamin Roche

### Vibrio ecology, pathogenesis, and evolution

#### *Daniela Ceccarelli <sup>1</sup> \* and Rita R. Colwell 1,2\**

*<sup>1</sup> Department of Cell Biology and Molecular Genetics, Maryland Pathogen Research Institute, University of Maryland, College Park, MD, USA*

*<sup>2</sup> Department of Environmental Health Sciences, Johns Hopkins Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA \*Correspondence: dcec@umd.edu; rcolwell@umiacs.umd.edu*

#### *Edited and reviewed by:*

*Jonathan P. Zehr, University of California, Santa Cruz, USA*

#### **Keywords: Vibrio, ecology, genome, evolution, pathogenesis**

This Research Topic brings together 24 articles that highlight the most recent research findings concerning the biology of the genus *Vibrio* and covers pathogenicity and host interaction, genome plasticity and evolution, and the dynamics of factors influencing the ecology of vibrios.

*Vibrio* comprises one of the most diverse marine bacterial genera (Gomez-Gil et al., 2014), and its diversity is emphasized in two of the articles opening this set of Research Topic papers. Sawabe et al. (2013) present a molecular phylogeny of 86 *Vibrio* species based on genome sequencing that provides insight into *Vibrio* biodiversity and evolutionary history. In a study of more than 300 *Vibrio* genome sequences, Lukjancenko and Ussery (2014) conclude that the *Vibrio* pan-genome comprises 17,000 gene families, differentially present and/or expressed in any given species.

A remarkable feature of all *Vibrio* species is an highly plastic genome, a feature examined in five papers. The two chromosomes are shaped by horizontal gene transfer involving, among others, antibiotic resistance, virulence, and niche adaptation (Rowe-Magnus et al., 2001; Kirkup et al., 2010). *V. vulnificus* biotype 3 is a notable example. Efimov et al. (2013) suggest biotype 3 evolved from biotype 1 by acquisition of unique genes from other bacterial species, such as *Shewanella,* sharing the same ecological niche. Carraro et al. (2014) employ molecular and functional characterization of pVCR94, to identify the role of IncA/C plasmids in antibiotic resistance in a Rwandan *V. cholerae* isolate. A retrospective analysis of epidemic *V. cholerae*in Angola is reported by Valia et al. (2013), showing unexpected genomic variability among variants, highlighting the role of genomic islands, phages, and integrative conjugative elements in the genetic diversity of *V. cholerae* in a single epidemic. Rivas et al. (2013) describe acquisition by *Photobacterium damselae* subsp. *damselae* of virulence plasmid pPHDD1 that encodes pore-forming toxins and hemolysins which play a key role in virulence for both fish and humans. A review by Rapa and Labbate (2013) describes the role of integrons in *Vibrio* species for which gene cassettes comprise approximately 1–3% of the entire genome and are very likely involved in bacterial adaptation and evolution.

Nine of the manuscripts analyze *Vibrio* pathogenicity, disease development, specificity, and adaptation in both human and animal hosts. Tan et al. (2014) deciphered the biosynthetic network of the siderophore vulnibactin, essential in iron uptake from host proteins, the importance of which in *V. vulnificus* pathogenicity has been clinically demonstrated. Inhibition/resistance mechanisms developed by *V. salmonicida*, the causative agent of hemorrhagic septicemia in Atlantic salmon, is described by Bjelland et al. (2013), who show that it overcomes the salmon innate immune system to a point where the infection overwhelms the host. The role in bacterial virulence of diverse extracellular proteolytic enzymes secreted by human pathogenic *Vibrio* species is the focus of a review by Miyoshi (2013). The engagement of type VI secretion systems by *V. cholerae* is suggested as a means of intra- and inter-species predation and nutrient acquisition, inducing rapid multiplication in the human host (Pukatzki and Provenzano, 2013). The bioluminescent marine bacterium *V. campbellii* is used by Wang et al. (2013) to analyze the pyomelanin-pigmented phenotype, known to provide *Vibrio* species with greater UV and oxidative stress resistance and enhanced intestine colonization. The relationship between pathogenicity and motility in *Vibrio* species and the contribution of flagella to adhesion and biofilm formation are discussed by Zhu et al. (2013). The largely unexplored *V. fluvialis* mechanisms of pathogenesis, survival and fitness are reviewed by Ramamurthy et al. (2014). Twenty new *Vibrio* species associated with molluscans are described and their pathogenic potential for molluscs elucidated by Romalde et al. (2014). The exquisite bacteria–host interaction between *V. fisheri* and its squid host, *Euprymna scolopes*, is described in detail, as are the molecular pathways of biofilm formation, motility, and chemotaxis (Norsworthy and Visick, 2013).

The capacity of *Vibrio* species to persist in the aquatic environment, their ecology and association with abiotic and biotic factors, as well as environmental surveillance for public health (Lipp et al., 2002; Grimes et al., 2009; Johnson, 2013) comprise a section in the Research Topic that opens with a review by Lutz et al. (2013) elucidating complex mechanisms enabling *V. cholerae* to withstand starvation, temperature fluctuation, salinity variation, and predation. Haley et al. (2014) report water temperature increase can be correlated with rise of a diverse population of *V. parahaemolyticus*, some of which carry pandemic markers, in water and plankton along the Georgian coast of the Black Sea. *V. parahaemolyticus* and *V. vulnificus* populations associated with oyster, sediment, and surface water related to a hurricane event in the Chesapeake Bay are concluded to be influenced by wave energy and sediment resuspension (Shaw et al., 2014). Canesi et al. (2013) show the serum of *Mytilus galloprovincialis* promotes phagocytosis and killing by hemocytes of both *V. cholerae* O1 and non-O1/non-O139 in edible bivalves. Chakraborty et al. (2013) evaluate a sensitive and specific dipstick test to detect toxigenic *V. cholerae* in water, validating a simple, inexpensive method for use in areas at risk of cholera.

Three articles addressing *Vibrio* environmental diversity and dynamics complete this Research Topic. Mansergh and Zehr (2014) suggest that the natural shift of *Vibrio* populations in Monterey Bay is affected by larger oceanographic conditions (flow velocities and wind patterns), rather than individual environmental factors. Meta-analysis of environmental variables and *Vibrio* association with plants, algae, zooplankton, and animals are reviewed by Takemura et al. (2014). As a final point concerning environmental distribution, Constantin De Magny et al. (2014) propose temporal shifts, zooplankton community variability, and occurrence of *V. cholerae* in the aquatic environment are related to cholera dynamics. These factors, analyzed by metagenomics, permit greater understanding of community structure, function, and competition.

In summary, the collection of manuscripts provided in this Research Topic offers a comprehensive exploration of *Vibrio* biology, from the single gene to the bacterial community, elucidating *Vibrio* molecular pathways and evolutionary history. This special issue shows the significant progress achieved in understanding the complex biology of the genus *Vibrio* and should both stimulate discussion and offer a challenge to researchers in microbial ecology and evolution.

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 01 May 2014; accepted: 10 May 2014; published online: 28 May 2014. Citation: Ceccarelli D and Colwell RR (2014) Vibrio ecology, pathogenesis, and evolution. Front. Microbiol. 5:256. doi: 10.3389/fmicb.2014.00256*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Ceccarelli and Colwell. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### Updating the Vibrio clades defined by multilocus sequence phylogeny: proposal of eight new clades, and the description of *Vibrio tritonius* sp. nov.

*Tomoo Sawabe1 \*, Yoshitoshi Ogura 2, Yuta Matsumura1, Gao Feng1, AKM Rohul Amin1, Sayaka Mino1, Satoshi Nakagawa1, Toko Sawabe3, Ramesh Kumar 4, Yohei Fukui 5, Masataka Satomi 5, Ryoji Matsushima5, Fabiano L. Thompson6, Bruno Gomez-Gil 7, Richard Christen8,9, Fumito Maruyama10, Ken Kurokawa11 and Tetsuya Hayashi <sup>2</sup>*

*<sup>1</sup> Laboratory of Microbiology, Faculty of Fisheries Sciences, Hokkaido University, Hakodate, Japan*


*<sup>7</sup> A.C. Unidad Mazatlán, CIAD, Mazatlán, México*


*<sup>11</sup> Earth-Life Science Institute, Tokyo Institute of Technology, Tokyo, Japan*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Xiu-Lan Chen, Shandong University, China Matteo Spagnoletti, University College London, UK*

#### *\*Correspondence:*

*Tomoo Sawabe, Laboratory of Microbiology, Faculty of Fisheries Sciences, Hokkaido University, 3-1-1 Minato-cho, Hakodate 041-8611, Japan e-mail: sawabe@fish.hokudai.ac.jp* To date 142 species have been described in the *Vibrionaceae* family of bacteria, classified into seven genera; *Aliivibrio*, *Echinimonas*, *Enterovibrio*, *Grimontia*, *Photobacterium*, *Salinivibrio* and *Vibrio*. As vibrios are widespread in marine environments and show versatile metabolisms and ecologies, these bacteria are recognized as one of the most diverse and important marine heterotrophic bacterial groups for elucidating the correlation between genome evolution and ecological adaptation. However, on the basis of 16S rRNA gene phylogeny, we could not find any robust monophyletic lineages in any of the known genera. We needed further attempts to reconstruct their evolutionary history based on multilocus sequence analysis (MLSA) and/or genome wide taxonomy of all the recognized species groups. In our previous report in 2007, we conducted the first broad multilocus sequence analysis (MLSA) to infer the evolutionary history of vibrios using nine housekeeping genes (the 16S rRNA gene, *gapA*, *gyrB*, *ftsZ*, *mreB, pyrH*, *recA*, *rpoA*, and *topA*), and we proposed 14 distinct clades in 58 species of *Vibrionaceae*. Due to the difficulty of designing universal primers that can amplify the genes for MLSA in every *Vibrionaceae* species, some clades had yet to be defined. In this study, we present a better picture of an updated molecular phylogeny for 86 described vibrio species and 10 genome sequenced *Vibrionaceae* strains, using 8 housekeeping gene sequences. This new study places special emphasis on (1) eight newly identified clades (*Damselae*, *Mediterranei*, *Pectenicida*, *Phosphoreum*, *Profundum*, *Porteresiae*, *Rosenbergii*, and *Rumoiensis*); (2) clades amended since the 2007 proposal with recently described new species; (3) orphan clades of genomospecies F6 and F10; (4) phylogenetic positions defined in 3 genome-sequenced strains (N418, EX25, and EJY3); and (5) description of *V. tritonius* sp. nov., which is a member of the "*Porteresiae*" clade.

**Keywords: vibrios,** *Vibrionaceae***, multilocus sequence analysis, evolution, housekeeping protein gene,** *Vibrio tritonius*

#### **INTRODUCTION**

Bacterial systematics has evolved alongside the development of innovative methodologies and techniques (Wayne et al., 1987;

**Abbreviations:** AFLP, amplified fragment length polymorphism; ANI, average nucleotide sequence identity; AAI, average amino acid identity; DDH, DNA-DNA hybridization; MLSA, multilocus sequence analysis; MLST, multilocus sequence typing; MP, maximum parsimony; NJ, neighbor joining; ML, maximum likelihood; PFGE, pulse field gel electrophoresis; RAPD, random amplified polymorphic DNA. Stackebrandt et al., 2002; Gevers et al., 2005). The first definition of bacterial species in "phylogenetic terms" was developed in 1987 using the DNA-DNA reassociation and DNA sequencing. These approaches to bacterial systematics provided us with a uniform definition of prokaryotic species (Wayne et al., 1987). In 2002, an *ad hoc* committee listed additional innovative methods that could be used for bacterial systematics, such as 16S rRNA gene sequence analysis, DNA typing methods (AFLP, RAPD, Rep-PCR, PFGE), MLSA, WGS analysis, and proteomics (Stackebrandt et al., 2002). The primary purpose of the committee's statement was to promote dialogue among systematists, population and evolutionary geneticists, ecologists and microbiologists for the benefit of bacterial systematics in general, and to create a more transparent species concept in particular (Stackebrandt et al., 2002). Among those innovative methodologies, MLSA and the WGS analysis have become the most important and successful methodologies; their strong impact on bacterial systematics is due to data reproducibility and portability (see, e.g., Maiden et al., 1998; Aanensen and Spratt, 2005; Gevers et al., 2005; Konstantinidis and Tiedje, 2005; Staley, 2006; Goris et al., 2007; Richer and Rosselló-Móra, 2009; Auch et al., 2010).

MLST, the prototype for MLSA-based methodology, was used for the first highly portable typing of *Neisseria meningitides* from invasive disease and healthy carriers, and it yielded the first understanding of the epidemiology and population structure of that infectious agent (Maiden et al., 1998). Its high levels of discriminatory power between those strains, which required half the loci typically required for a classical allozyme electrophoresis, and its superior application to evolutionary, phylogenetic, or population genetic studies, allowed researchers to develop MLST schemes for a number of bacteria taxa (Aanensen and Spratt, 2005) (also refer to the MLST website; http://www.mlst.net/). It also opened the use of MLSA for bacterial systematics (e.g., Sawabe et al., 2007; Thompson et al., 2007; Bishop et al., 2009), and it guided the reconsideration and re-evaluation of prokaryotic species concepts (Gevers et al., 2005; Staley, 2006; Preheim et al., 2011). Even now, MLSA provides a better understanding of taxonomically controversial bacterial taxa; for example, the human origins of the *Agrobacterium* (*Rhizobium*) *radiobacter* clustered as a wellseparated genovar (Aujoulat et al., 2011) and the highly versatile aeromonads consisting of 3 major clades (Roger et al., 2012).

In the genome era, genome sequencing has been used to characterize new bacterial species (Haley et al., 2010; Hoffmann et al., 2012), to reclassify bacterial taxons such as *Neisseria* (Bennett et al., 2012), *Acinetobacter* (Chan et al., 2012) and *Vibrio* (Lin et al., 2010), and to challenge defined prokaryotic species (Konstantinidis and Tiedje, 2005; Thompson et al., 2009; Chan et al., 2012). Using these WGSs, *in-silico* DDH calculations can also be emulated, mainly in two ways: high-scoring segment pairs (HSPs) (Konstantinidis and Tiedje, 2005; Goris et al., 2007) and the genome-to-genome distance calculation, called the Digital DDH measurement (Auch et al., 2010). The criterion of more than 95% ANI is currently a widely used similarity value for species delineation. The WGS analysis also supercedes the limitations of MLSA, which is only capable of including genes that are successfully amplified by designed primers (Gevers et al., 2005; Thompson et al., 2005; Sawabe et al., 2007).

*Vibrionaceae* are at the forefront of bacterial taxons being tested with new innovative methodologies and techniques for bacterial systematics (Thompson et al., 2001, 2005, 2009; Sawabe et al., 2007). The number of species described in *Vibrionaceae* has increased remarkably since the establishment of genome fingerprinting techniques (Thompson et al., 2001) and MLSA schemes (Thompson et al., 2005, 2007; Sawabe et al., 2007, 2009). Now, a total of 142 species are recognized in the family *Vibrionaceae* (Association of Vibrio Biologists website; http:// www.vibriobiology.net/). *Vibrionaceae* are defined as a group of strains with the following characteristics: they are Gram-negative rods with a polar flagellum enclosed in a sheath, have facultative anaerobic metabolisms, are capable of fermenting D-glucose, and grow at 20◦C. The bacteria are primarily aquatic, and most species are oxidase positive, can reduce nitrate to nitrite, require Na+ for growth, and ferment D-fructose, maltose, and glycerol (Gomez-Gil et al., 2014). In addition, most vibrio species ferment a variety of carbohydrates without gas production, and grow on TCBS medium (Farmer et al., 2005; Thompson et al., 2009). As we experience a rapid expansion in the number of known species in the family *Vibrionaceae*, we face a number of unique vibrio isolates that lack one or more of the above common properties. *Vibrionaceae* species are metabolically versatile, and the number of species showing gas production, nitrogen fixation, phototrophy, and non-motility is increasing (Gomez-Gil et al., 2014). Considerably more attention should be paid to the biological and genetic plasticity of vibrios to help understand the dynamics of vibrio evolution (Sawabe et al., 2007; Grimes et al., 2009; Thompson et al., 2009).

Due to the limitations of using 16S rRNA gene phylogeny (Gomez-Gil et al., 2014) to elucidate an "Integrated Vibrio Biology" that includes a biodiversity assessment, an inferred evolutionary history, the population biology, and genomics, Sawabe et al. (2007) developed an MLSA scheme for the family *Vibrionaceae* using nine gene sequences (*ftsZ*, *gapA*, *gyrB*, *mreB*, *pyrH*, *recA*, *rpoA*, *topA* and the 16S rRNA gene). The analysis involved the complete sequence sets of 9 genes from 58 vibrio taxa, and it revealed 14 monophyletic clades with a significant bootstrap support. The species within each clade shared >20% DDH, <5% G+C (mol%), >85% MLSA sequence similarity, and >89% AAI (Sawabe et al., 2007).

Recent extraordinary progress in biodiversity studies and WGS projects in vibrios has resulted a substantial leap in novel vibrio species and taxonomically unassigned strains. In fact, after the proposal of 14 robust clades in 2007, more than 60 new species have been described. The number of genome-sequenced strains has exceeded 1000. It is, therefore, obvious that the phylogenetic tree based on the multilocus gene sequences reported in 2007 is insufficient to show the most recent molecular phylogenetic structure of vibrios.

In our previous analysis of multilocus gene phylogeny, eight housekeeping protein coding genes and the 16S rRNA gene were included in the MLSA. However, the 16S rRNA gene has a rather low interspecies resolution (100, 99.5, and 95.3% of maximum, median, and minimum resolution, respectively) (see Figure 5S, in Sawabe et al., 2007). It was also difficult to include the 16S rRNA gene sequences for the calculation of the radiation time of each clade. We have been debating whether the inclusion of the 16S rRNA gene sequences is necessary to infer the evolutional history of vibrios, but there are no other proper and fast tools to check the tree topologies constructed using a rather large data set (58 species).

A variety of methods have been proposed to tackle the problem of gene tree reconciliation to reconstruct a species tree. When the taxa in all the trees are identical, the problem can be stated as a consensus tree problem (Guénoche, 2013). The comparison of gene trees and their assembly into a unique tree representing the species tree is a general problem in phylogeny. However, in this study, we faced a different problem: determining whether the inclusion or exclusion of a given gene in the analysis would substantially change the outcome. This was the method used in this analysis to investigate if the inclusion or exclusion of the 16S rRNA gene sequences in the MLSA analysis would, or would not, affect the final result.

The aims of this study were to re-evaluate how the 16S rRNA gene sequence affects the final phylogenetic tree, as based on multilocus gene sequencing analysis; to update our knowledge in vibrio biodiversity and evolution on the basis of an 8-gene MLSA; and to reconstruct a better vibrio phylogeny. The analysis provided a further opportunity to propose additional eight clades to the most up-to-date vibrio phylogeny.

#### **MATERIALS AND METHODS**

#### **SUBTREE INCONGRUENCE TEST OF MULTILOCUS GENE PHYLOGENY**

The usual approach to compare trees is to count how many subtrees they share; a given subtree often has a different topology according to the method used (NJ, ML, or MP) but the differences are often subtle and generally not well-supported. Accordingly, it is best to compare trees according to bipartitions (a tree is considered as a set of bipartitions, each one corresponding to an internal edge of the tree, the external ones connecting the leaves to the tree) (Guénoche, 2013). This was the method used in this analysis to investigate whether the inclusion or exclusion of the 16S rRNA gene sequences in the MLSA analysis would, or would not, affect the final result. TreeDyn (Chevenet et al., 2006) was used to compare trees to subtrees sharing the same topology. A dedicated python script (using libraries from Huerta-Cepas et al., 2010) was used to compare trees with shared sub-trees, independently of their topologies.

Our previous MLSA revealed that the 16S rRNA gene, in contrast to the other gene sequences used, has a rather low interspecies resolution (Sawabe et al., 2007). To increase the sensitivity and reduce the time taken by MLSA to update the vibrio phylogeny, we compared the subtree topologies obtained from a nine-gene data set (that included the 16S rRNA gene) and an eight-gene data set (that excluded the 16S rRNA gene). We used the gene sequence data set from the 58 species, for which we had the sequence of every gene, and which is identical to the data set used in Sawabe et al. (2007). The method selected, (1) shared subtrees only if they had the same topology, and (2) subtrees that shared the same species, independently of their topology. The results were visualized using TreeDyn (Chevenet et al., 2006), and congruent subtrees with the same topologies were indicated using the same color.

#### **SEQUENCING OF HOUSEKEEPING PROTEIN-CODING GENES**

An additional eight housekeeping genes of type strains of the genera *Vibrio* and *Photobacterium* were sequenced manually according to Sawabe et al. (2007) with newly designed primers (Vgap150f:ACTCAYGGYCGTTTCAACGGYAC, Vgap957r:RC CGATTTCGTTRTCGTACCAAG, VftsZf55f:GTKGGTGGCG GCGGCGGTAA, VftsZ782r:ACACCACGWGCACCAGCAA GATCG, VftsZ-9f:ACCGATGATGGAAATGTCTGACGATG C, VmreB225f:RATGAAA GACGGCGTWATYGC, VmreB102 5r:TCGCCRCCGTGCATRTCGATCA) (**Table S1**). All strains were maintained in ZoBell 2216E agar and stored with 20% glycerol at −80◦C. Whole genome sequencing was performed in nine type strains (*V. aerogenes* LMG 19650T, *V. gazogenes* ATCC 29988T, *V. halioticoli* IAM 14596T, *V. neonatus* HDD3-1T, *V. porteresiae* MSSRF 30T, *V. rhizosphaerae* MSSRF 3T, *V. ruber* LMG 23124T, *V. tritonius* sp. nov. AM2T, and *V. superstes* G3-29T) using the Roche 454 FLX titanium genome sequencer alone or in combination with the Illumina MiSeq sequencer. The sequence reads were assembled using the Newbler software version 2.3 or later. Illumina reads were used only for sequence error correction. After auto-annotation by Microbial Genome Annotation Pipeline (MiGAP, http://www.migap.org/; Sugawara et al., 2009), relevant housekeeping gene sequences were retrieved and used for the MLSA. The house keeping genes necessary for updating the vibrio phylogeny were also retrieved from the latest version of the NCBI microbial genome and GenBank database (Release 197.0, 15 August 2013), and used in the analysis. All sequence data used in this study are listed in **Table S1**.

#### **SEQUENCE ANALYSIS**

MLSA was performed the same way as in Sawabe et al. (2007). The sequences were aligned using the ClustalX program (Larkin et al., 2007). The domains used to construct the phylogenetic trees shown in **Figures 2**, **3** were regions of the *ftsZ*, *gapA*, *gyrB*, *mreB*, *pyrH*, *recA*, *rpoA*, and *topA* genes of *Vibrionaceae*: positions 195– 630, 225–861, 441–1026, 390–897, 171–543, 429–915, 87–873, and 570–990 (*V. cholerae* O1 Eltor N16961 (AE003852) numbering), respectively. The regions were within those used in the previous study (Sawabe et al., 2007). Sequence similarity and the number of nucleotide and amino acid mutation were determined using MEGA version 5 (Tamura et al., 2011).

Split Decomposition Analysis (SDA) was also performed as described in Sawabe et al. (2007) using SplitsTree version 4.12.8, with a neighbor net drawing and Jukes-Cantor correction (Bandelt and Dress, 1992; Huson and Bryant, 2005). The concatenated sequences of the eight housekeeping genes were also generated using the program and used for a phylogenetic analysis combined with NJ, MP, and ML analyses (Sawabe et al., 1998).

#### **PHYLOGENETIC, GENETIC, PHENOTYPIC, AND CHEMOTAXONOMIC CHARACTERIZATION OF** *Vibrio tritonius* **sp. nov**

Four isolates of *V. tritonius* sp. nov., JCM 16456<sup>T</sup> <sup>=</sup> LMG <sup>25401</sup><sup>T</sup> <sup>=</sup> AM2T, JCM 16457 <sup>=</sup> LMG 25402 <sup>=</sup> MA12, JCM 16458 = LMG 25403 = MA17, and JCM 16459 = LMG 25404 = MA35, isolated from gut of a sea hare, *Aplysia kurodai*, were used in this study. The strains were cultured on ZoBell 2216E agar (Oppenheimer and ZoBell, 1952) and stored at −80◦C in 10% glycerol-supplemented broth.

A total of 1400 bp 16S rRNA gene sequences of the four strains were determined according to Sawabe et al. (1998) using four sequence primers (24F, 1100F, 920R, and 1509R). The 16S rRNA gene sequences were blasted to the latest release ver. 197 of GenBank and related sequences were retrieved. Finally, 16S rRNA gene sequences of *V. aerogenes* X74705, *V. brasiliensis* AJ316172, *V. cholerae* X76337, *V. fluvialis* X74703, *V. furnissii* X76336, *V. gazogenes* X74705, *V. hepatarius* AJ345063, *V. nereis* X74716, *V. porteresiae* EF488079, *V. rhizosphaerae* DQ847123, *V. ruber* AF462458, *V. tubiashii* X74725, and *V. xuii* AJ316181 were included in the phylogenetic analysis (**Figure 4**). Phylogenetic trees were constructed using three different methods (NJ, ML, and MP). For NJ analysis, distance matrices were calculated using the Kimura two parameters correction and using MEGA version 5.0 (Tamura et al., 2011). ML and MP analysis was conducted using PHYLIP (Phylogeny Inference Package, version 3.573c, distributed by J. Felsenstein, Department of Genetics, UW, Seattle, WA, USA). Sequences corresponding to positions 86–1420 of the *E. coli* gene (NC\_000913) were used in this analysis. **Figure 4** represents a subset of the final tree obtained using the NJ method with 500 bootstrap replications. Nodes supported by ML and MP analyses are indicated by the bootstrap values in **Figure 4**.

DNAs of bacterial strains were prepared following the procedures of Marmur (1961), with minor modifications. The mol% G+C content of DNAs was determined by HPLC (Tamaoka and Komagata, 1984). DNA-DNA hybridization experiments were performed in microdilution wells using a fluorometric direct binding method described by Ezaki et al. (1988); Ezaki et al. (1989). DNA-DNA similarity data were shown as the average value of triplicate experiments. *V. brasiliensis*, *V. furnissii*, *V. fluvialis*, *V. tubiashii*, *V. hepatarius*, *V. mytili*, *V. nereis*, *V. porteresiae*, and *V. xuii* were selected as the reference species of these DNA-DNA hybridization experiments, based on the results from the MLSA molecular phylogenetic assessment and the 16S rRNA gene phylogeny of *V. tritonius* sp. nov.

A total of 62 phenotypic characteristics were determined using the standard manual characterization method established in our laboratory (Sawabe et al., 1998). The carbon assimilation test was conducted using a basal seawater medium, as previously described (Sawabe et al., 1998). These phenotypic characterizations were performed at 25◦C. O/129 sensitivity was determined using the sensitivity basal agar medium (Nissui Pharmaceutical, Tokyo, Japan) at 30◦C.

#### **RESULTS**

#### **THE SUBTREE INCONGRUENCE TEST BETWEEN THE 9-GENE AND 8-GENE PHYLOGENIES**

Subtrees obtained in the 8-gene phylogeny were compared to those in the 9-gene phylogeny and the 16S rRNA gene phylogeny (**Figure 1**). Among the 13 subtrees reconstructed in the 8-gene phylogeny, 12 were retained in the 9-gene phylogeny; the only difference observed was the inclusion of *V. proteolyticus* in the *V. cholera* subtree (**Figures 1A,B**). Most of the 13 subtrees in the

(−16S rRNA) **(A)** and 9 gene (+16S rRNA) phylogeny **(B)**. The topology was also compared between 8 gene and only 16S rRNA gene phylogeny **(C)**.

8-gene phylogeny corresponded to the clades that we previously proposed based on the 9-gene phylogeny (**Figure 1B**). The results of the subtree incongruence test using the 58 vibrio taxa data showed that the inclusion of the 16S rRNA gene sequence is not a critical factor in optimizing the vibrio phylogeny on the basis of MLSA.

#### **THE LATEST VIBRIO PHYLOGENY BASED ON MULTILOCUS HOUSEKEEPING PROTEIN-CODING GENE SEQUENCES**

WGSs of key vibrios species that were resistant to the gene amplification, e.g., *V. gazogenes*, *S. costicola*, *V. porteresiae*, *V. caribbenthicus*, are now available. This result indicated that we could use the complete set of 8 housekeeping proteincoding gene sequences currently available from 86 described vibrio species and 10 genome-sequenced *Vibrionaceae* strains for the MLSA updating of the 8-gene phylogeny, on the basis of Splits Decomposition Analysis (SDA) (Bandelt and Dress, 1992; Huson and Bryant, 2005) (**Figure 2**) and a supertree reconstruction (**Figure 3**). On the basis of SDA, we could retain the 14 distinct monophyletic clades that were previously defined, and we were able to further define 8 new clades: *Damselae*, *Mediterranei*, *Pectenicida*, *Phosphoreum*, *Profundum*, *Porteresiae*, *Rosenbergii*, and *Rumoiensis* (**Figures 2**, **3**, **Table 1**). The robustness of these clades was high enough to propose their monophyly in the supertree reconstruction using three different molecular phylogenetic analyses (**Figure 3**). Using an 8-housekeeping protein-coding gene analysis, most of clades shared >80.5% ANI and >92% AAI, and the highest ANI (98.3%) was observed in the sequence comparison between *V. anguillarum* and *V. ordalii* (**Table 1**).

#### **NEW CLADES**

*Mediterranei* consisted of three species: *V. mediterranei*, *V. maritimus*, and *V. variabilis*. The 8-gene ANI and AAI were 89.5– 96.3% and 98.0–99.1%, respectively. The mol% G+C range of the clade members was 42–46.3 mol%. The genome-sequenced strain *V. mediterranei* AK1 showed 98.7% ANI and 99.9% AAI against the *V. mediterranei* type strain. Their known habitats are warm seawater and coral mucus.

*Porteresiae* consisted of *V. porteresiae* and the newly described *V. tritonius* sp. nov. Detailed information for this new species is described in the section "The bacterial taxonomical remarks

**Table 1 | Newly proposed and emended clades by means of 8 gene MLSA for vibrios.**


*(Continued)*


#### **Table 1 | Continued**

*\*Bold indicated new and/or emended information.*

of *V. tritonius* sp. nov." below. These two species shared 91.4% ANI and 97.2% AAI. Two of the unique phenotypes in these species were an efficient H2 production and nitrogen fixation. While the genome sequences of these two species are highly conserved (unpublished data), they have distinct habitats (**Table 1**). The mol% G+C ranged from 44.2 to 45.5, and the DDH value of *V. tritonius* type strain against *V. porteresiae* type strain was 59% (**Table 2**).

*Pectenicida* consisted of two species, *V. caribbeanicus* and *V. pectenicida* showing 82.8% ANI and 96.1% AAI. The reported habitats were tidal flats and diseased larvae, respectively (**Table 1**).

*Rumoiensis* consisted of two species, *V. litoralis* and *V. rumoiensis* showing 91.5% ANI and 98.3% AAI. These species were isolated from a tidal flat and sewage from a fishery product factory, respectively (**Table 1**). The reported DDH value between *V. litoralis* and *V. rumoiensis* was below 7.4%.

*Damselae*, *Phosphoreum*, *Profundum*, and *Rosenbergii* were the newly proposed clades that are included in *Photobacterium* spp. These four new clades are based on ANI (87.5–96.2% in range), AAI (95.7–99.3% in range), and branch separation according to the supertree analysis in comparison with those ranges and branch separations of other *Vibrio* clades. The *Damselae* clade consisted of two subspecies of *P. damselae*. The *Rosenbergii* clade consisted of *P. lutimaris* and *P. rosenbergii*.

#### **DEFINING ORPHAN CLADES**

These are the clades that are formed by only one species. *V. tapetis* and *V. proteolyticus* were not grouped with any other species (**Figure 2** and **Table 1**). The recently proposed genomospecies F6 and F10 also did not belong to any of the clades proposed in this analysis. Previously reported singletons, *V. agarivorans*, and *V. pacinii* were not included in this analysis due to the lack of some gene sequences.

#### **EMENDED CLADES**

We can find emendations in most of the clades previously defined (**Figures 2**, **3**, and **Table 1**): (1) *Cholerae* (inclusion of *V. parilis*); (2) *Fischeri* (incl. *Ali. sifae*); (3) *Gazogenes* (incl. *V. rhizosphaerae*); (4) *Halioticoli* (incl. *V. breoganii*, and *V. inusitatus*); (5) *Harveyi*(incl. *V. azureus*, and *V. communis*); (6) *Orientalis* (incl. *V. sinaloensis*); (7) *Scophthalmi* (incl. *V. ponticus*); and (8) *Splendidus* (incl. *V. cyclitrophicus, V. gigantis,* and *V. crassostrea*).

We first included the complete set of the 8-housekeeping protein-coding gene sequences of *Salinivibrio costicola* subsp. *costicola* in the MLSA for vibrio phylogeny, because its WGS data (ASAI01000001) are available. However, for the current analysis of the *Salinivibrio*/*Grimontia*/*Enterovibrio* grouping, we could use only the data set including a single species of *Salinivibrio*, 2 species of *Enterovibrio*, and a single species of *Grimontia* for the 8-gene phylogeny. As both SDA and supertree analysis showed less robustness in these genera/species grouping, we decided to tentatively define them as the *Salinivibrio-Grimontia*-*Enterovibrio* (SGE) super-clade.

#### **DEFINING THE CLADE OF GENOME SEQUENCED STRAINS**

Vibrios are one of the most advanced groups in WGS analysis; currently more than 900 genomes are available in the public database (http://www.vibriobiology.net/). The MLSA of the 10 genome sequenced strains revealed: (1) LGP32 and EX25 formed a robust cluster with *V. tasmaniensis* and *V. alginolyticus*, respectively; (2) N418 and EJY3 (Roh et al., 2012) were related to *V. scopthalmi* and *V. natriegens*, respectively; and (3) The orphan positions of genomospecies F6 and F10, and AK16 (**Figures 2**, **3**, and **Table 1**).

#### **THE BACTERIAL TAXONOMICAL REMARKS OF** *Vibrio tritonius* **sp. nov.**

On the basis of the 8-gene MLSA, the sea hare (*Aplysia kurodai*) isolates were highly likely to represent a new species within the family *Vibrionaceae*, more precisely within a new clade "*Porteresiae.*" Four strains of *V. tritonius* formed a robust cluster within the *Porteresiae* clade on the basis of 4-gene sequence SDA (data not shown). To confirm the taxonomic status of the sea hare strains, a standard polyphasic taxonomy was conducted.

The results of our phylogenic analyses based on the 16S rRNA gene sequence clearly showed that these strains belong to class *Gammaproteobacteria*, and more precisely to the family *Vibrionaceae*. The closest phylogenic neighbor of the four sea hare isolates was the *V. furnissii*-*V. fluvialis* cluster (**Figure 4**). *V. porteresiae* was not closely related, as shown by the 16S rRNA gene sequence phylogeny. Intra-species sequence similarities of the 16S rRNA gene among *V. tritonius* sp. nov. were above 99.5%. Four strains of *V. tritonius* sp. nov. showed 98.0–98.1% similarity, and 98.1% similarity toward *V. furnissii* (X76336) and *V. fluvialis* (X74703), respectively. Sequence similarities of *V. tritonius* sp. nov. to the other phylogenetic neighbors and to gas-producing vibrios were below 98%. The 16S rRNA gene sequence similarity between *V. tritonius* sp. nov. and "*Allomonas enterica*" AJ550855 was 98.3%. The 16S rRNA gene sequences of *V. fluvialis* X74703 and "*A. enterica*" AJ550855 were identical.

**FIGURE 4 | Unrooted phylogenetic tree on the basis of 16S rRNA gene sequences.** Scale bar: 0.005 accumulated change per nucleotide. This figure combines the results of three analyses i.e., neighbor-joining, maximum parsimony, and maximum likelihood. The topology shown was obtained using neighbor-joining and 500 bootstrap replications. Percentages indicate the branches that were also obtained both in the maximum likelihood analysis (*P* < 0.01) and in the most parsimonious tree.

#### **Table 2 | DNA relatedness among** *Vibrio tritonius* **and the related vibiro species.**


*\*The reciprocal DDH value of V. tritonius JCM 16456T against V. porteresiae MSSRF 30T probe was 46%.*

Mutual DDH experiments showed that the four strains of *V. tritonius* sp. nov., JCM 16456T, JCM 16457, JCM 16458, and JCM 16459, were conspecific and clearly separated from their phylogenetic neighbors, e.g., *V. porteresiae*, *V. fluvialis V. furnissii, V. tubiashii, V. hepatarius,* and *V. nereis* (**Table 2**). The mol% G+C content was 44.8 ± 0.6, which was within the range of the genus *Vibrio*.

#### **DESCRIPTION OF** *Vibrio tritonius* **sp. nov**

Etymology of the newly describing *Vibrio* species was provided here: *Vibrio tritonius* (tri.to'ni.us. L. masc. adj. tritonius, named after Triton (a sea-god, son of Neptune and the nymph Salacia, referring to the habitat of the bacteria).

Major phenotypic features of *V. tritonius* sp. nov. are shown in **Table 3**. The four sea hare strains have the major phenotypic features of the genus *Vibrio* (except for no growth on TCBS and gas production). These strains required salt for their growth, and they were motile, fermentative and oxidase positive. Apparent catalase activity was not observed. The four strains of *V. tritonius* sp. nov. were phenotypically most similar to *V. porteresiae*, but they differed from *V. porteresiae* in four traits (catalase production, and the assimilation of D-mannose, γ-aminobutyrate and pyruvate), out of 62 tested traits (**Table 3**). The four *V. tritonius* strains were sensitive to the vibrio-static agent O/129 (150μg). Positive assimilation of glucose, mannitol, gluconate, glucuronate, and xylose indicated the presence of three major carbohydrate metabolic pathways, the Embden-Meyerhof, Entner-Doudoroff, and pentose-phosphate pathways, in *V. tritonius* sp. nov. Presence of the gene set for those three central metabolic pathways of carbohydrates was supported by our preliminary WGS analysis of *V. tritonius* JCM 16456T (data not shown). Phenotypic traits differentiating *V. tritonius* sp. nov. from *V. aerogenes*, which shows a gas production phenotype, included nitrate reduction, amylase production, and arginine dihydrolase activity. Inability to grow on TCBS was a common trait of *V. tritonius* sp. nov. and *V. porteresiae* (**Table 3**).

The other phenotypic traits were also described below. No swarming cells were observed. Gas production from glucose and mannitol occurred. Cells are curved rods, with rounded ends, are 0.7–0.9μm in diameter and 2.6–2.7μm in length when the organism is grown on ZoBell 2216E medium; the cells occur singly on the agar. No endospores or capsules are formed. Colonies on ZoBell 2216E agar medium are beige, circular, and smooth and convex with an entire edge. Sodium ions are essential for growth. The bacterium can grow in presence of 0.5% to 6% NaCl. The bacterium is a mesophilic chemoorganotroph which grows at temperatures between 15 and 40◦C. Optimal growth is observed from 25 to 30◦C. Growth occurs from pH 4.5 to pH 9, and optimal growth is at pH 7.5–8.0. No growth occurs at 45◦C. The bacterium is positive for acid production from glucose and mannitol; for nitrate reduction, acetoin production, and hydrolysis of gelatin, DNA and casein. The bacterium also can assimilate N-acetyl-D-glucosamine, cellobiose, D-fructose, maltose, D-mannitol, D-galactose, lactose, L-glutamate, L-proline, acetate, citrate, fumarate, DL-malate, pyruvate, and succinate. The bacterium is negative for catalase; indole production; arginine dihydrolase, lysine decarboxylase, ornithine decarboxylase, **Table 3 | Phenotypic characteristics for distinguishing** *Vibrio tritonius* **from the related** *Vibrio* **species.**


*Data were obtained under same culture condition in our laboratory.* +*, Positive;* −*, negative; w, weak reaction; d*+*, variable but type strain positive; d*−*, variable but type strain negative (numbers of positive strains are shown in parentheses). All species are motile, require Na for growth, fermentative of Dglucose, growth at 15–25* ◦*C, oxidase-positive, produce gelatinase, DNase, and caseinase, reduce nitrate, produce acid from D-glucose, assimilate N-acetyl-D-glucosamine, D-fructose, D-galactose, gluconate, D-glucose, D-marmitol, Lglutamate, L-proline, acetate, citrate, fumarate, DL-malate, pyruvate, succinate. All species are negative for pigmentation, swarming, luminescence, growth at 4 and 45* ◦*C, alginase production, agarase production, lysine decarboxylase, ornithine decarboxylase, requirement for organic growth factor, assimilation of D-sorbitol, aconitate,* α*-ketoglutarate, L-tyrosine, meso-erythritol.*

luminescence, and pigmentation; the requirement of organic growth factors; hydrolysis of agar, alginate, starch, and Tween 80; and assimilation of D-glucosamine, D-sorbitol, aconitate, αketoglutarate, L-tyrosine, meso-erythritol, trehalose, putrescine, propionate, and D-glucosamine. The G+C content of DNA is 44.2–45.5 mol%. The type strain is JCM 16459T <sup>=</sup> LMG <sup>25401</sup><sup>T</sup> <sup>=</sup> AM2T.

#### **DISCUSSION**

Considerable biodiversity can be found within the family *Vibrionaceae* (Gomez-Gil et al., 2014), even after the first proposal of vibrio phylogeny and evolution was inferred on the basis of MLSA in 2007 (Sawabe et al., 2007). More than 60 species of *Vibrionaceae*, with a surprising level of biodiversity, have been described since 2007. These include a marine invertebrate isolates such as coral associated vibrios (Chimetto et al., 2011; Gomez-Gil et al., 2014), introduction of nitrogen-fixing vibrios within an endophyte-like ecological niche (Rameshkumar et al., 2008), and an isolation of new vibrio species from the surface of cheese (Bleicher et al., 2010) have been reported. In addition to the increasing number of newly described vibrio species, many strains showing interesting ecophysiological features have been genome sequenced. However, taxonomic information appears to be insufficient to push the elucidation of vibrio biodiversity and evolution forward. Such a rapid progress in the study of vibrio biodiversity, genomics and evolution prompted us to update the vibrio phylogeny on the basis of MLSA. We have retrieved the complete sets of 8 house-keeping protein-coding gene sequences for 30 additional *Vibrionaceae* species including a newly described vibrio species, *V. tritonius*sp. nov., as well as for 10 as yet unnamed *Vibrio*/*Enterovibrio* spp. The MLSA led us to propose eight new clades (*Damselae*, *Mediterranei*, *Pectenicida*, *Phosphoreum*, *Profundum*, *Porteresiae*, *Rosenbergii*, and *Rumoiensis*) in the family *Vibrionaceae*, in addition to those previously proposed in the report of Sawabe et al. (2007). In 2007, *V. mediterranei*, *V. petenicida*, *V. rumoiensis*, and *P. rosenbergii* were affiliated as singlet species, but they have now been grouped. Four orphan clades (*Tapetis*, *Proteolyticus*, F6 and F10) were newly defined. More efforts are required to isolate the closest neighbors of these orphan species. Strains EJY3, EX25, N418, and LGP32 clustered robustly with *V. natriegens*, *V. alginolyticus*, *V. scopthalmi*, and *V. tasmaniensis*, but a further systematic survey is required to analyze the phylogenetic position of AK16 and 1F-230 (**Figure 3**).

We are still facing a lack of *Photobacterium* spp. sequences to infer their precise evolutionary history. Among the 23 described *Photobacterium* spp., we could include only half of them in this study. This situation has arisen mainly due to the "primer problems" in MLSA. Unfortunately, there are also limited numbers of WGS of *Photobacterium* spp. available in public databases. However, in this analysis, considering

the results of SDA, supertree analysis, and the ANI and AAI similarity ranges in comparion to the other *Vibrio* spp. clades, we proposed four new clades for the *Photobacterium* spp.; (1) *Damselae*, (2) *Phosphoreum*, (3) *Profundum*, and (4) *Rosenbergii*. A *Salinivibrio*/*Grimontia*/*Enterovibrio* super-clade is also proposed.

In molecular phylogenetics, the use of minimum gene set is crucial to reduce time and cost, as well as to improve the accuracy, of analyses. This is of particular importance when identifying species and elucidating population structure and evolution in a super bacterial taxon such as the family *Vibrionaceae*, which has more than 140 species. The previous MLSA of 58 vibrio taxa (Sawabe et al., 2007) showed that 16S rRNA gene sequences have an extremely low species/strain discriminating power compared to the other genes tested. Therefore, before conducting the current vibrio MLSA, we evaluated whether the 16S rRNA gene data set could be eliminated from the MLSA. For this analysis, we developed a subtree incongruence test algorithm. The algorithm is a fast and reliable method for selecting subtrees that share the same topology or those that have different topologies but share the same species. The results of this analysis indicate that inclusion of 16S rRNA gene sequences is not necessary for reconstructing the vibrio phylogeny on the basis of MLSA.

We have experienced the first case in which the molecular phylogenies resulting from 16S rRNA gene sequences and from housekeeping gene sequences were largely incongruent in the species descriptions of *V. porteresiae* and *V. tritonius* (**Figures 2**, **4**). For the affiliation of the clade of *V. porteresiae*, we used only four genes (the *pyrH*, *recA*, *rpoA*, and 16S rRNA genes), and we confirmed that *V. porteresiae* was affiliated with the *Cholerae* clade (Rameshkumar et al., 2008).The 16S rRNA gene sequence phylogeny revealed that *V. tritonius* sp. nov. was the most closely related to *V. furnissi* and *V. fulvialis* with ca. 98% sequence similarity, and *V. tritonius* sp. nov. and *V. porteresiae* were distantly related in their phylogenetic relationship (**Figure 4**). Less phylogenetic relatedness in the 16S rRNA gene sequence tree between both the species and the lack of housekeeping gene sequences of *V. porteresiae*


*\*1 Data from Rameshkumar et al. (2008). Values in C16:1*ω*7c and C14:0 3-OH are from the summed percentage of feature 2 and 3, respectively.*

prevented no direct comparison of *V. porteresiae* and *V. tritonius* sp. nov. Fortunately, the whole genome nucleotide sequences of *V. porteresieae* and *V. tritonius* sp. nov. were determined in this study, and the first direct comparison of both species by MLSA and whole genome comparison was achieved. Surprisingly, the MLSA with 8-housekeeping genes phylogeny led to the conclusion that both *Vibrio* species, *V. porteresiae* and *V. tritonius* sp. nov., share a common ancestry and that they can be proposed as a new vibrio clade, "*Porteresiae.*" Our preliminary genome comparison of both species also supported monophyly because a strong synteny was observed between both genomes (unpublished data). Incongruences between the 16S rRNA gene sequence tree and the MLSA tree were also observed in *Mediterranei* and *Pectenicida* clades (Lambert et al., 1998; Chimetto et al., 2011; Hoffmann et al., 2012) in this study. Therefore, we conclude that at present the "8-housekeeping-gene phylogeny" is the most powerful method for delineating vibrio species description/biodiversity/population study/evolution, until alternative genome-based approaches are proposed. This analysis reduces the misidentification of ancestry clades of vibrios.

The polyphasic taxonomic approach reveals that the *Aplysia* gut isolates, *V. tritonius* sp. nov., to be a novel *Vibrio* species showing gas producing ability (**Figures 2**–**4**, **Tables 2**, **3**) and forming a robust clade, *Porteresiae*, with *V. porteresiae*. Production of gas during the fermentation of carbohydrates is not a prevalent property in the *Vibrio* genus (Shieh et al., 2000, 2003; Farmer et al., 2005; Kumar and Nair, 2007; Rameshkumar et al., 2008). *V. aerogenes*, *V. furnissii*, *V. gazogenes*, *V. porteresiae*, *V. ruber*, and *V. rhizosphaerae* are the species in which the gas production phenotype is retained in a stable way among 98 *Vibrio* species. For this reason, gas production is an atypical property in the genus *Vibrio*. The current phylogenetic network analysis using the 8-gene MLSA confirmed that *V. aerogenes*, *V. gazogenes*, *V. rhizosphaerae*, and *V. ruber* form a robust clade, *Gazogenes*(**Figure 2**) (Kumar and Nair, 2007; Sawabe et al., 2007). *V. furnissii* belongs to the *Cholerae* clade. The newly described *V. tritonius* sp. nov. belongs to a new clade, *Porteresiae*. The gas compositions of *Porteresiae* clade species but also of the *Gazogenes* and *Cholerae* clades, were identical for H2 and CO2, but the H2 production efficiencies differ between these clades. The H2 production efficiency of *V. tritonius* sp. nov. and *V. porteresiae* was high and comparable to that of enterobacterial species, such as *Escherichia coli*, *Salmonella*, *Enterobacter*, and *Klebsiella* (Nakashimada et al., 2002). Our preliminary genome comparisons suggests that *V. tritonius* sp. nov. and *V. porteresiae* contain very similar gene clusters responsible for H2 production and nitrogen fixation machinery (unpublished data). It would be intriguing to understand how those vibrios acquire and/or lose gas producing abilities from the standpoint of both evolutionary dynamics and metabolic diversity.

In conclusion, 8-gene MLSA is a reliable tool for delineating a species and a monophyletic group or "clade." Using the current data set reported in this study, <98% of 8-gene-concatenated nucleotide sequence identity may allow us to define a species boundary. We also showed that WGSs overcome the limitations of gene-by-gene multilocus sequencing tasks found in the MLSA for *Vibrionaceae* (primer problem, Chromosome 2 gene inclusion). In fact, we could not include *V. aerogenes*, *V. gazogenes*, *V. rhizosphaerae*, and *V. superstes* for the previous 9-gene MLSA by the primer problems (Sawabe et al., 2007), but the successful WGS for these species and the inclusion of 8 housekeeping protein gene sequences retrieved from the genome sequences can provide the better picture of current vibrio molecular phylogeny. More efforts to sequence the individual housekeeping genes and WGS of all remaining species in the family *Vibrionaceae* not involved in this study allow the possibility of elucidating the ultimate clade structure of these Vibrios. The ultimate phylogenetic trees allow us to provide the ideal phylogenetic backbone to elucidate the evolutional history, genome dynamics, and plasticity in the family *Vibrionaceae*.

#### **ACKNOWLEDGMENTS**

This work was supported by Ministry of Agriculture, Forestry, and Fisheries, Japan, and KAKENHI (Grants-in-Aid for Scientific Research from Ministry of Education, Culture, Sports, Science, and Technology of Japan) (No. 21380129, 2365817201, 2529212203, and 2529212283). This work was also supported by the National Council for Science and Technology (CONACYT) of Mexico (Grant No. 132328 awarded to Bruno Gomez Gil), by Grant in Aid for Scientific Research on Innovative Area "Genome Science" from Ministry of Education, Culture, Sports, Science, and Technology of Japan (No. 221S0002), and by JST-CNPq Strategic Japanese-Brazilian Cooperative Program, Biomass and Biotechnology. We gratefully thank Professor Jean Euzéby at the École Nationale Vétérinaire, Toulouse for his advice and suggestions on naming *Vibrio tritonius*, and Professor Elena P. Ivanova, Swinburne University of Technology, for a critical reading of the manuscript. We also thank Fumito Testukawa, Takashi Wakabayashi, and Yoshiko Kawahara for technical contributions.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb. 2013.00414/abstract

The GenBank accession numbers for the *ftsZ*, *gapA*, *gyrB*, *mreB*, *recA*, *rpoA*, *pyrH*, and *topA* gene sequences used in this analysis, as well as those used in Sawabe et al. (2007), are listed in **Table S1**.

**Table S1 | List of strains and sequences accession number used for the MLSA.**

#### **REFERENCES**


reconciliation of approaches to bacterial systematics. *Int. J. Syst. Bacteriol*. 37, 463–464. doi: 10.1099/00207713-37-4-463

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 27 September 2013; paper pending published: 18 October 2013; accepted: 16 December 2013; published online: 27 December 2013.*

*Citation: Sawabe T, Ogura Y, Matsumura Y, Feng G, Amin AKMR, Mino S, Nakagawa S, Sawabe T, Kumar R, Fukui Y, Satomi M, Matsushima R, Thompson FL, Gomez-Gil B, Christen R, Maruyama F, Kurokawa K and Hayashi T (2013) Updating the Vibrio clades defined by multilocus sequence phylogeny: proposal of eight new clades, and the description of Vibrio tritonius sp. nov. Front. Microbiol. 4:414. doi: 10.3389/fmicb. 2013.00414*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Sawabe, Ogura, Matsumura, Feng, Amin, Mino, Nakagawa, Sawabe, Kumar, Fukui, Satomi, Matsushima, Thompson, Gomez-Gil, Christen, Maruyama, Kurokawa and Hayashi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### *Vibrio* chromosome-specific families

#### *Oksana Lukjancenko1 and David W. Ussery1,2\**

*<sup>1</sup> Department of Systems Biology, Center for Biological Sequence Analysis, Technical University of Denmark, Lyngby, Denmark <sup>2</sup> Comparative Genomics Group, Oak Ridge National Laboratory, Biosciences Division, Oak Ridge, TN, USA*

#### *Edited by:*

*Rita R. Colwell, University of Maryland, USA*

#### *Reviewed by:*

*Martin G. Klotz, University of North Carolina at Charlotte, USA Vincent Burrus, Université de Sherbrooke, Canada*

#### *\*Correspondence:*

*David W. Ussery, Department of Systems Biology, Center for Biological Sequence Analysis, Technical University of Denmark, Kemitorvet, Building 208, 2800 Kgs. Lyngby, Denmark e-mail: dave@cbs.dtu.dk*

We have compared chromosome-specific genes in a set of 18 finished *Vibrio* genomes, and, in addition, also calculated the pan- and core-genomes from a data set of more than 250 draft *Vibrio* genome sequences. These genomes come from 9 known species and 2 unknown species. Within the finished chromosomes, we find a core set of 1269 encoded protein families for chromosome 1, and a core of 252 encoded protein families for chromosome 2. Many of these core proteins are also found in the draft genomes (although which chromosome they are located on is unknown.) Of the chromosome specific core protein families, 1169 and 153 are uniquely found in chromosomes 1 and 2, respectively. Gene ontology (GO) terms for each of the protein families were determined, and the different sets for each chromosome were compared. A total of 363 different "Molecular Function" GO categories were found for chromosome 1 specific protein families, and these include several broad activities: pyridoxine 5' phosphate synthetase, glucosylceramidase, heme transport, DNA ligase, amino acid binding, and ribosomal components; in contrast, chromosome 2 specific protein families have only 66 Molecular Function GO terms and include many membrane-associated activities, such as ion channels, transmembrane transporters, and electron transport chain proteins. Thus, it appears that whilst there are many "housekeeping systems" encoded in chromosome 1, there are far fewer core functions found in chromosome 2. However, the presence of many membrane-associated encoded proteins in chromosome 2 is surprising.

**Keywords:** *Vibrio* **pan-genome, chromosome-specific genes,** *Vibrio* **comparative genomics,** *Vibrio* **core-genome, comparative genomics**

#### **INTRODUCTION**

The *Vibrio* genus represents a large subgroup of *Gamma* subdivision of *Proteobacteria*, which are abundant, fast growers that can be highly variable. These bacteria have the ability to form biofilm on biotic and abiotic surfaces and are ubiquitous in marine and estuarine environments at notably high densities in fish, corals, shrimps, plankton, and mammals (Thompson et al., 2004; Reen et al., 2006; Froelich et al., 2013). Currently, the *Vibrio* genus contains more than 60 different species, although complete genome sequences are available for only 10 of them. Several species are known to be pathogenic for humans, fishes, and marine invertebrates, and are well studied. *V. cholerae* can act as the causative agent of the severe and sometimes lethal disease, cholera, and is probably the most sequenced and clinically important member of *Vibrio* species (Heidelberg et al., 2000; Egan and Waldor, 2003). *V. vulnificus* causes septicemia in wound infections; however, despite its high fatality rate, human infections of *V. vulnificus* are rare (Matsuoka et al., 2013; Tsao et al., 2013). *V. parahaemolyticus* and *V. furnissii* infections may lead to gastroenteritis in humans via consumption of raw seafood (Tanabe et al., 2011; Xiang et al., 2013). Strains of *V. anguillarum* species are life threatening to many economically important fish, including Atlantic salmon, seabass, cod, and rainbow trout (Wiik et al., 1995). *V. fischeri* participates in beneficial symbioses with many marine organisms, especially squids (Verma and Miyashiro, 2013). *V. harveyi* causes luminous vibriosis, which infects prawns, oysters, and lobsters (Yu et al., 2013). Finally, *V. splendidus* is known as an extensive bivalve pathogen (Tanguy et al., 2013).

All known *Vibrios* have two chromosomes; the presence of two chromosomes in *V. cholerae* was first documented in 1998 (Trucksis et al., 1998). Chromosome 1 is usually larger, with a relatively constant size of about 3 million base pairs, encoding around 2700 proteins that represent many essential functions. In contrast, chromosome 2 is smaller, about 1 million base pairs encoding roughly a thousand proteins, and contains a highly variable "super-integron" (Rowe-Magnus et al., 1999). *Vibrio* genomes contain many genomic islands, which can contain functions allowing adaptation to specific environments and, perhaps, can even represent speciation events (Vesth et al., 2010).

The existence of two chromosomes in all *Vibrio* genomes, and variance of chromosome 2, has been the main point of many investigations worldwide and has been the subject of multiple discussions about the purpose and origin of smaller chromosomes. It has been proposed that chromosome 2 originated as a megaplasmid, although later Heidelberg et al. have suggested that it may play an important role in the organism and could help optimize the fast replication rate (Okada et al., 2005; Reen et al., 2006; Kirkup et al., 2010; Dikow and Smith, 2013).

The aim of this study is to compare *Vibrio* chromosome specific genes, as well as the conserved core-genome and pangenomes, across more than 300 strains of the *Vibrio* genus, both complete and available draft genomes, as well as to focus on distribution of functional proteins and available Gene Ontology information between two chromosomes.

#### **MATERIALS AND METHODS**

#### **SELECTION AND CHARACTERISTICS OF BACTERIAL STRAINS**

A set of all publically available *Vibrio* strains was selected for this study and downloaded from the NCBI web pages (July 2012). The initial set included 368 genomes, 18 of them were complete and 350 were retrieved as Illumina raw reads from the NCBI Sequence Read Archive (SRA). Of these, 188 genomes were sequenced using a HiSeq 2000 sequencer and the remaining 162 were sequenced with an Illumina Genome Analyzer II.

Protein encoding gene predictions were carried out using the gene-finding tool Prodigal (Hyatt et al., 2010). 16S ribosomal RNA sequences were extracted for both the complete and the draft *Vibrio* genomes using RNAmmer (Lagesen et al., 2007). For each assembled genome, the number of fragments (contiguous pieces), protein coding genes, and the mean gene length were calculated; strains with an average gene length below 700 bp were excluded from further analysis. The resulting set consisted of 18 complete genomes, (**Table 1**), and 284 draft sequences (**Table S1**). The distribution of these characteristics for each genome is shown

#### **Table 1 | List of species used in the study.**


*\*ATCC 17749.*

in **Figure 1**. Note that on average there are about 7 or 8 rRNA operons per complete *Vibrio* genome, although in most draft genomes only one copy is given.

#### **PROTEOME COMPARISON**

Proteome comparison was performed with the PanFunPro tool (Lukjancenko et al., 2013). Briefly, protein-encoding sequences from each genome were extracted and annotated as described by Lukjancenko et al. (2013) and grouped into protein families. Results of pan- and core-genome analysis for chromosomes 1 and 2 were visualized as an accumulative pan-/core-plot and a pairwise comparison matrix.

The distribution of unique functional profiles between the chromosomes 1 and 2 was examined, followed by a brief investigation of available GO functional categories, specific for each of the chromosomes.

One representative proteome for each species was chosen from the pool of complete genomes and interspecies analysis of specific-genomes was performed between each pair of species. The results were visualized as a specific-matrix.

#### **RESULTS AND DISCUSSION**

The *Vibrio* dataset consisted of 302 genomes, representing 9 known and 2 unknown *Vibrio* species. A list of the species and accession numbers for the complete genomes is shown in **Table 1**, and a similar list for all 302 genomes is given in **Table S1**. Only 18 of the strains were completely finished, and for those independent proteomes for both chromosomes 1 and 2 were extracted. However, most of the genomes (284) were draft and partially assembled into several large pieces of continuous chromosomal DNA, although information concerning which protein belongs to which chromosome was not available. Thus, it was decided to build analysis around 2 sets: the finished genomes (18 genomes) and the whole dataset, including the WGS draft genomes (302 genomes).

The calculated basic features for each analyzed genome is shown in **Figure 1**, including the number of contiguous pieces, predicted protein coding genes, average gene lengths, and predicted 16S rRNAs. A large fraction of the assembled genomes contain between 150 and 190 contiguous pieces (contigs) of chromosomal DNA, with a group of outlier strains showing more than 200 pieces per genome. An obvious correlation can be seen between the number of contigs and the amount of predicted rRNAs and genes, followed by a shorter than average gene length in assembled genomes with higher numbers of contiguous sequences.

#### *VIBRIO CHOLERAE* **CHROMOSOME 1 AND CHROMOSOME 2 COMPARISON**

The *Vibrio cholerae* chromosome 1 is larger (about 3 Mbp) and is more stable, carrying many essential protein coding genes, whereas chromosome 2 is smaller (about 1 Mbp), contains a large genomic island (the "superintegron"), is more variable, and has fewer essential genes. A pairwise comparison of set of 18 genomes for both chromosomes is shown in **Figure 2**. Chromosomes 1 and 2 share a bit more than 10% of their protein families. Within chromosome 1 the range is 55 to 96%, and for chromosome 2 it is 25 to 95%. Since there are multiple genome sequences for

**FIGURE 1 | Predicted genome characteristics (A).** Distribution of the number of contiguous pieces **(B)**. Distribution of the protein number per genome **(C)**. Distribution of the average protein coding gene length per genome **(D)**. Number of predicted 16S rRNA sequences.

several different strains available for the *V. cholerae* species, a high similarity within chromosomes can be found with confidence, although on average only 10% of the proteins are shared between chromosomes 1 and 2.

The core-genome of complete strains contains 1269 conserved protein families shared within chromosome 1, and 252 core families shared within chromosome 2; only 104 functional profiles are shared between the two chromosomes. When additional draft

**and 2.** The distribution is shared both as percentage on the axis and the absolute number above the bar. The absolute number reflects the amount of GO IDs that were connected to the pathway. The color code is as follows: red is the biological process, green is the cellular component, and blue is the molecular function.

**FIGURE 6 | Pairwise interspecies-specific genome comparison for chromosome 1 (A) and chromosome 2 (B).** Analysis included a single representation of 7 known and 2 unknown species. The resulting percentage shows the ratio between the amount of species-specific families and the size of the total proteome. On average, each species contained between 18 and 33% specific protein families. Color intensity indicates the level of specificity.

#### **Table 2 | List of species analyzed in this study.**


*For each species the number of available genomes and sequence status are provided. Species are listed alphabetically.*

genomes were included, the numbers for both chromosome 1 and 2 dropped to 673 core-genomes and 140 protein families, followed by a decrease of shared functional profiles for a total number of 96. The core- and pan-genome summary results are shown in **Table 2** and conserved profiles and their functions in **Table S2**.

The pan genome for chromosome 1 (∼5500 gene families) is about twice the number of genes encoded in a single copy of chromosome 1 (e.g., 2650 genes in *V. cholerae* strain M66-2), whilst the pan-genome for chromosome 2 (∼3740 gene families) is more than three times the size found encoded in a single copy of chromosome 2 (e.g., 1043 genes for *V. cholerae* strain M66-2). Many of these additional gene families are likely to be found in the super-integron, which is a known variable region of chromosome 2.

assignment source: PfamA, Superfamily, TIGRFAM, and CD-HIT clustering **(B)**. Protein coding gene length distribution by each profile type.

A closer look at the distribution of functions within the coregenomes of two chromosomes showed that all of the shared proteins are found in the PfamA database (**Figure S1**) and most of them are involved in biological processes or molecular function (**Figure 3**). The presence of proteins involved in essential metabolic and regulatory processes in the shared genomic pool of both chromosomes is consistent with the claim that the smaller chromosome is not a plasmid, but is fundamental for growth and biological activity.

In order to explore the overlap between the core genes in chromosomes 1 and 2, we extracted the core proteins for each chromosome and then examined the overlap with the core of the other chromosome (**Figures 4**, **5**). A total number of 639 GO IDs could be extracted for the chromosome 1 core-specific profiles (1169 profiles). 438 of these were involved in biological processes, 53 in cellular component functions, and 363 in molecular functions. Equivalent analysis of chromosome 2 corespecific profiles yielded, in total, 109 GO IDs (of 153 profiles). 57 of the IDs were involved in biological processes, 10 in cellular components, and 66 in molecular functions. It is not surprising that whilst the core of chromosome 1 carries more proteins that are essential to sustain life and to reproduce, the specific core of chromosome 2 contains proteins involved in metabolic processes and enzyme and membrane associated activity. The addition of 284 draft genomes slightly reduced the number of specific proteins and specific pathway groups in chromosome 1, leaving 265 GO terms involved in the biological process, 39 in cellular component functions, and 197 in molecular functions (**Figure S2**). In contrast, chromosome 2 contained 15 GO terms in biological processes, 4 in cellular components, and 14 in molecular functions (**Figure S3**).

#### **SPECIES COMPARISON**

The genus *Vibrio* is comprised of a diverse group of bacteria, which can be either pathogenic or symbiotic to mammals and organisms of marine environments. Species-specific genomes may contain proteins responsible for pathogenicity or they may be crucial for survival in a given environment. To demonstrate the level of specificity between species of the same chromosome, 9 strains representing 7 known and 2 unknown species, a pairwise comparison of specific-genomes, was performed. Within chromosome 1, the fraction of unique proteomes varies from 18 to 33% (**Figure 6A**), whereas genomes of chromosome 2 differ in a greater portion of proteins, ranging from 18 to 64% (**Figure 6B**).

*Vibrio cholerae* spp. are known pathogens in humans and were chosen to examine for genome specific differences in gene content. Representative strains of *V. cholerae* species were compared to other strains, as shown in **Figure S4**. Chromosomes 1 and 2 contained a similar amount of specific profiles, 190, and 192, respectively. Most of them were CD-HIT clustering-based, however, 79 and 44 were annotated by PfamA and TIGRFAM collections. A complete list of profiles and corresponding functions are listed in **Table S3**.

#### **PROTEOMES OF** *V. CHOLERAE* **DRAFT GENOMES**

*V. cholerae* is one of the most important, highly documented, and most sequenced species of *Vibrios*. Our dataset included 279 *V.* *cholerae* strains, 8 completely sequenced and 271 draft genomes. For the draft genomes, chromosome specific genes could not be calculated. However, starting with the known core genomes from the finished genomes, it is possible to look for the presence of the known chromosome core genes across the draft genomes. Thus, core-genome analysis of 279 *V. cholerae*strains yielded in 776, 250, and 182 protein families, in large, small, and both of the chromosomes, respectively. Further, we examined the pan-genomes of both chromosomes within a set of 18 genomes. The distribution of the total number of 8325 functional profiles is as follows: 2333, 341 and 73 families assigned to PfamA, Superfamily, and TIGRFAM databases, respectively (**Figure 7**). We estimate that the 271 newly sequenced *V. cholerae* strains brings at least 2000 possible profile combinations to the pool of previously known functions that represent more than 70 different GO functional categories (**Figure 8**).

In conclusion, the *Vibrio* pan-genome can be quite large, with more than 17,000 gene families, although, any one *Vibrio* genome will contain only about 3500 genes, or about one-fifth of the size of the pan-genome. There is considerably more variability in chromosome 2 than in chromosome 1.

#### **ACKNOWLEDGMENTS**

The authors are grateful to all research groups that have submitted their genome sequences to public databases, without which this analysis would not have been possible. The authors are in part supported by the Center for Genomic Epidemiology at the Technical University of Denmark; part of this work was funded by grant 09-067103/DSF from the Danish Council for Strategic Research.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb. 2014.00073/abstract

**Figure S1 | Annotation and length distribution of proteins within core-genome of small and large chromosomes (A).** Distribution of profiles by assignment source: PfamA, Superfamily, TIGRFAM, and CD-HIT clustering **(B)**. Protein coding gene length distribution by each profile type.

**Figure S2 | GO term analysis in proteins shared within chromosome 1 and missing in the core of chromosome 2 in set\_302 genomes.** Distribution is shared both as percentage on the axis and absolute number above the bar. Absolute number shows the amount of GO IDs that were connected to the pathway. Color code is as follows: red is biological process, green is cellular component, and blue is molecular function.

**Figure S3 | GO term analysis in proteins shared within chromosome 2 and missing in the core of chromosome 1 in set\_302 genomes.** Distribution is shared both as percentage on the axis and absolute number above the bar. Absolute number shows the amount of GO IDs that were connected to the pathway. Color code is as follows: red is biological process, green is cellular component, and blue is molecular function.

**Figure S4 | Annotation and length distribution of proteins within** *V. cholerae* **species-specific proteomes for chromosome 1 (A) and chromosome 2 (B).** Annotation of profiles and protein coding gene length

distribution are visualized by assignment source: PfamA, Superfamily, TIGRFAM, and CD-HIT clustering.

#### **Table S1 | List of Sequence Read Archive (SRA) genomes used in the study.**

#### **Table S2 | Conserved functional profiles within genomes of set\_18.**

Whether profile consists of more than 1 domain, function is shown for each domain.

#### **Table S3 | Profiles, specific for** *V. cholerae* **species., in chromosome I and**

**chromosome II.** Whether profile consists of more than 1 domain, function is shown for each domain.

#### **REFERENCES**


desferrioxamine B as a xenosiderophore in Vibrio furnissii. *Biol. Pharm. Bull*. 34, 570–574. doi: 10.1248/bpb.34.570


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 01 October 2013; accepted: 10 February 2014; published online: 18 March 2014.*

*Citation: Lukjancenko O and Ussery DW (2014) Vibrio chromosome-specific families. Front. Microbiol. 5:73. doi: 10.3389/fmicb.2014.00073*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Lukjancenko and Ussery. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## Insight into the evolution of *Vibrio vulnificus* biotype 3's genome

#### *Vera Efimov1†, Yael Danin-Poleg1†, Nili Raz 1, Sharona Elgavish2, Alex Linetsky1 and Yechezkel Kashi <sup>1</sup> \**

*<sup>1</sup> Laboratory of Food Microbiology and Applied Genomics, Faculty of Biotechnology and Food Engineering, Technion – Israel Institute of Technology, Haifa, Israel <sup>2</sup> Bioinformatics Knowledge Unit, Lorry I. Lokey Interdisciplinary Center for Life Sciences and Engineering, Technion – Israel Institute of Technology, Haifa, Israel*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Alison Buchan, University of Tenessee-Knoxville, USA Carmen Amaro, University of Valencia, Spain*

#### *\*Correspondence:*

*Yechezkel Kashi, Laboratory of Food Microbiology and Applied Genomics, Faculty of Biotechnology and Food Engineering, Technion – Israel Institute of Technology, Haifa 32000 ISRAEL e-mail: kashi@tx.technion.ac.il*

*†These authors have contributed equally to this work.*

*Vibrio vulnificus* is an aquatic bacterium and an important human pathogen. Strains of *V. vulnificus* are biochemically classified into three biotypes. The newly emerged biotype 3 appears to be rather clonal and geographically restricted to Israel, where it caused an outbreak of wound infections and bacteremia. To understand the evolution of the bacterium's genome, we sequenced and analyzed the genome of biotype 3 strain VVyb1(BT3), and then conducted a microbial environmental survey of the hypothesized niche from which it probably evolved. The genome of this environmental isolate revealed higher similarity to the published biotype 1 genomes of clinical strains (90%) than to the environmental strains (87%), supporting the virulence of the biotype 3 group. Moreover, 214 of the total 5361 genes were found to be unique to strain VVyb1(BT3), having no sequence similarity to any of the known genomes of *V. vulnificus*; 35 of them function in DNA mobility and rearrangement, supporting the role of horizontal gene transfer in genome evolution. Interestingly, 29 of the "unique" genes had homologies among *Shewanella* species. In a survey conducted in aquaculture ponds in Israel, we successfully co-isolated *Shewanella* and *V. vulnificus* from the same niche, further supporting the probable contribution of *Shewanella* to the genome evolution of biotype 3. Indeed, one gene was found in a *S. algae* isolate. Surprisingly, molecular analysis revealed that some of the considered unique genes are harbored by non-sequenced biotype 1 strains isolated from the same environment. Finally, analyses of the biotype 3 genome together with the environmental survey suggested that its genome originated from a biotype 1 Israeli strain that acquired a rather small number of genes from other bacterial species in the niche, such as *Shewanella*. Therefore, aquaculture is likely to play a major role as a man-made ecological niche in bacterial evolution, leading the emergence of new pathogenic groups in *V. vulnificus*.

**Keywords:** *Vibrio vulnificus***, evolution, biotype 3, genome, unique gene, gene transfer, environment**

#### **INTRODUCTION**

*Vibrio vulnificus* is a gram-negative halophilic bacterium which belongs to the family *Vibrionaceae*. It is a highly invasive human pathogen that is naturally found in marine and estuarine environments the world over (Strom and Paranjpye, 2000; Oliver, 2006; Jones and Oliver, 2009). Strains of *V. vulnificus* are classified into three biotypes based on biochemical and serological characteristics, and differences in host range. Biotype 1, isolated mostly from shellfish in coastal estuarine areas, is the most common group worldwide and is responsible for numerous clinical cases (Linkous and Oliver, 1999; Farmer, 2003). Biotype 2 was isolated from diseased eels and is rarely associated with human infections (Amaro and Biosca, 1996).

The newly emerged biotype 3 appears to be geographically restricted to Israel, where it caused an outbreak of wound infections and bacteremia among Israeli fish farmers and consumers of Tilapia fish (Bisharat et al., 1999). This novel group is responsible for nearly all clinical cases of *V. vulnificus* in Israel. Biotype 3 was found to be a clonal group that is clearly distinct from the other biotypes, possessing biochemical properties of both biotypes 1 and 2 (Bisharat et al., 2005; Broza et al., 2012). All biotype 3 isolates exhibited the same genotype in multilocus sequence typing (MLST) of 10 housekeeping and 5 conserved hypothetical genes (Broza et al., 2012). Analysis of 12 simple-sequence repeats (SSRs) loci, that are highly mutable regions, discriminated between its isolates (Broza et al., 2012), still demonstrating the low genetic diversity of this group in the tested loci. It has been suggested that the biotype 3 clonal hybrid evolved as a consequence of genome hybridization of two different and independent populations (Bisharat et al., 2005).

Even though *V. vulnificus* is a human pathogen, its evolution happens mostly in its natural environment—aquaculture fish ponds and Tilapia fish, which are the reservoir host of *V. vulnificus* biotype 3 in Israel (Bisharat et al., 1999). This emphasizes the importance of the local environment in shaping the genomic evolution of individual community members (Medini et al., 2008), including human pathogens. It is likely that the high organic biomass in aquaculture farms in Israel plays a major role as a niche for bacterial development, as well as providing selective pressure leading to new strains and groups, such as biotype 3 (Broza et al., 2012). Thus, we hypothesize that this aquaculture niche is the "melting pot" from which biotype 3 probably evolved. The main mechanism governing the development and emergence of new virulent strains in *V. vulnificus* is the high and frequent horizontal gene transfer in the *Vibrionaceae* (Quirke et al., 2006; Kim et al., 2011).

In the last decade, whole-genome sequencing has been used for the study of bacterial evolution (Jackson et al., 2011). This approach enables identifying large genomic rearrangements, including insertions, deletions, inversions, translocations and duplications (Bryant et al., 2012). Several comparative genomic studies have been conducted with *V. vulnificus* species to identify the specific genomic composition of different isolates (Quirke et al., 2006; Gulig et al., 2010; Morrison et al., 2012). The complete genome sequences of two *V. vulnificus* biotype 1 strains, YJ016 with CMCP6, were compared and showed large chromosomal regions that were unique to each. This suggested a role for DNA acquisition in increasing diversity and possible adaptability of the organisms to new and changing environments (Quirke et al., 2006). Sequencing of several biotype 1 and 2 genomes and subsequent comparative genomic analysis identified numerous genes that are common to the most virulent strains but are lacking from attenuated strains. These candidate virulence genes encode Flp pili, GGDEF proteins, and genomic island XII. Sialic acid catabolism was similarly identified as a potential contributory factor in molecular pathogenesis (Gulig et al., 2010). Recently, a pyrosequencing-based comparative study of six biotype 1 isolates identified 167 and 278 genes specifically associated with environmental and clinical genotypes, respectively (Morrison et al., 2012). The only available genome of biotype 3 was recently published by our laboratory, and provides a representation of this biotype due to the high clonality and very low diversity among strains revealed using molecular tools (Bisharat et al., 2005, 2007; Broza et al., 2007, 2012). This genome possesses a large number of genes that do not exist in the published biotype 1 genomes (Danin-Poleg et al., 2013). Analyzing this genome, and identifying the origin of new inserted elements and genes that are unique to biotype 3 in a bacterial population isolated from the aquaculture from which biotype 3 probably evolved, and tracing *V. vulnificus* in its natural ecological niche, should give an indication of horizontal gene transfer between the species. This, in turn, is expected to contribute to an understanding of the evolution of the human pathogen and provide a broad perspective of the emergence of new pathogenic strains.

#### **MATERIALS AND METHODS**

#### **BACTERIAL STRAINS AND ENVIRONMENTAL SAMPLING**

Environmental samples were collected from Tilapia fish originating directly from artificial fish ponds, the fish store, and pond sediment in western Galilee, Israel, between 2009 and 2013 (during May–October). Fish samples included skin, gills and fins, and were isolated by a previously described procedure (Broza et al., 2009). Briefly, all samples were selectively enriched in alkaline peptone water with 4% NaCl, pH 6.9, overnight in duplicate, then plated on thiosulfate–citrate–bile salts–sucrose (TCBS) agar (HiMedia Laboratories, Mumbai, India) in 10-fold dilutions. Suspected *V. vulnificus* and *Shewanella* colonies (green and black, respectively) were further grown on chromogenic agar (CHROMagar Microbiology, Paris, France) for verification. *V. vulnificus* colonies were PCR-amplified for detection of the *vvh* gene (Broza et al., 2007). *V. vulnificus* isolates were identified by biochemical tests as previously described (Broza et al., 2012). Rapid crude DNA extraction from *V. vulnificus*- and *Shewanella*suspected colonies was performed by ethanol-based technique as described previously (Buhnik-Rosenblau et al., 2013).

#### **GENOME-SEQUENCE COMPARISONS**

The biotype 3 draft genome of VVyb1(BT3) (Danin-Poleg et al., 2013) was annotated using the RAST annotation server (Aziz et al., 2008). The complete genome was compared to the available *V. vulnificus* biotype 1 genomes of three clinical and three environmental strains, and similarity calculation was carried out by alignment of the whole genomes using MUMmer 3.0 software (Kurtz et al., 2004). The genome of VVyb1(BT3) was not compared to biotype 2 strains as there was no available genome sequence of biotype 2. In addition, genes "unique" to biotype 3 were identified by comparison of the annotated VVyb1(BT3) genes to the annotated genes of the published genomes of strains YJ016 (Chen et al., 2003) and CMCP6 (Kim et al., 2003) using stand-alone BLAST-2.2.23 (Altschul et al., 1990) and in-house scripts. In the second step, 435 "unique biotype 3 genes" were compared by BLASTn against the GenBank bacterial database.

#### **DETECTION OF SPECIFIC GENES BY PCR**

Five genes with known function, "unique" to biotype 3 and present in *Shewanella* were selected for PCR amplification, and specific primers were designed based on available genomes of strains belonging to *V. vulnificus* biotypes 1, 3, and *Shewanella* targeting conserved regions (**Table 1**). The primers were used to generate ∼200-bp fragments. The reactions were carried out in a Veriti 96-well thermal cycler (Applied Biosystems, Foster city, CA) as follows: 95◦C for 3 min, 30 cycles of 30 s at 95◦C, 30 s at the annealing temperature (52◦C and 60◦C), 90 s at 72◦C, 10 min at 72◦C, cooling to 12◦C. PCR-amplification products were verified by 1.2% gel electrophoresis and observed by UV fluorescence. Both strands of the amplified products (**Table 3**) were also sequenced for verification, followed by multiple alignments (see further on).



*aLocus taq in VVyb1(BT3) genome.*

#### **IDENTIFICATION OF** *Shewanella*

Nucleotide sequence analyses of 16S rDNA and topoisomerase subunit B (*gyrB*) genes were performed to confirm *Shewanella* species identity (Yamamoto and Harayama, 1995; Nilsson et al., 2003). Fragments of the 16S rDNA gene and *gyrB* were PCRamplified and the products were purified using a QIAquick PCR purification kit (Qiagen, Hilden, Germany). Purified DNA (20– 50 ng) was sequenced on both strands using a BigDye terminator v1.1 cycle sequencing kit (Applied Biosystems) and loaded into an ABI 3130 genetic analyzer. Results were analyzed with SeqScape 2.5 software (Applied Biosystems) and DNA sequencing analysis 5.2 software (Applied Biosystems). Sequences of the isolates were compared with those of other *Shewanella* species in the GenBank database. Multiple sequence alignments were performed using CLUSTALW software (Thompson et al., 1994). The alignment files were used to evaluate genetic relationships among the strains by the unweighted pair group method with arithmetic mean (UPGMA) by MEGA 4.0 (Tamura et al., 2007). Bootstrap confidence values were based on 500 simulated dendrograms.

#### **RESULTS AND DISCUSSION**

We recently published the first genome sequence of biotype 3 strain VVyb1(BT3), which afforded the opportunity to learn about the evolutionary process leading to the emergence of this new clonal pathogenic group of *V. vulnificus* (Danin-Poleg et al., 2013). The genome of strain VVyb1(BT3) exhibits features similar to those of published biotype 1 genomes and consists of two chromosomes and a plasmid (5.74 Mbp; 46.7% G + C content), including a total of 5361 protein-encoding genes. Whole-genome comparisons of strain VVyb1(BT3) to the available *V. vulnificus* biotype 1 genomes of three clinical and three environmental strains (Chen et al., 2003; Kim et al., 2003; Park et al., 2011; Morrison et al., 2012) showed that although VVyb1(BT3) is an environmental strain, it has higher similarity to the published genomes of clinical, rather than environmental, biotype 1 strains (∼90% vs. ∼87%, **Table 2**). This result, together with the high clonality of biotype 3 (Bisharat et al., 2005, 2007; Broza et al., 2009, 2012) and the fact that many of its isolates have a clinical origin, support the virulence of this group. However, more analysis is required to further support this assumption, as the compared environmental genomes are incomplete (presented as scaffolds and contigs).

To better understand the special genomic features of biotype 3, a detailed analysis was performed. In the first stage, the genome of strain VVyb1(BT3) and the published *V. vulnificus* genomes of biotype 1 strains YJ016 (Chen et al., 2003) and CMCP6 (Kim et al., 2003) were compared. The analysis revealed a set of 435 genes that were absent in these biotype 1 genomes, suggesting that most of them are unique to biotype 3 (referring to *V. vulnificus* species) and may contribute to its virulence and environmental adaptation. Moreover, among the unique genes were those encoding proteins that might confer an advantage to biotype 3 against the microbial community in the environmental niche, such as the ParE-ParD toxin-antitoxin system, and against the host, such as hemoglobin-binding protein. In the second stage, the 435 genes were analyzed for sequence similarity to four more recently published *V. vulnificus* genomes (Morrison et al., 2012) using the **Table 2 | Genome similarity as revealed by the whole-genome comparisons between biotype 3 strain VVyb1(BT3) and each of the six listed biotype 1 strains.**


BLASTn algorithm. Only half of the genes (214) were found to be unique to VVyb1(BT3) and had little or no similarity (filtering below 85% identity and 80% query coverage) to sequences in the known *V. vulnificus* genomes. This suggested that as more *V. vulnificus* genomes are compared, including those of biotype 2 strains, fewer genes will be recognized as "unique" to biotype 3.

The finding of "unique" biotype 3 genes led us to hypothesize that most of them are acquired horizontally from other bacterial species sharing the same ecological niche, such as other *Vibrio* and *Shewanella* species (see further on). Indeed, 35 of the annotated "unique" genes have functions in genome organization and DNA transfer, supporting a role for gene transfer in genome evolution. Therefore, in the last stage of the bioinformatics analysis, the BLASTn algorithm was used against the NCBI gene database to find homologs of the "unique" biotype 3 genes and to identify bacteria that might serve as donors for the gene transfer. The analysis revealed sequence similarity (>70% homology) of 87 genes to other bacteria (**Figure 1**; Table S1)*.* As expected, due to the high horizontal gene transfer in the *Vibrionaceae* (Quirke et al., 2006; Kim et al., 2011), 37 genes showed homology to other *Vibrio* species: 10 genes were similar to *Vibrio parahaemolyticus*, 7 to *Vibrio harveyi*, 6 to *Vibrio fischeri*, and 14 to *Vibrio cholerae*. Interestingly, *Shewanella* shared a large number of genes (29) with the *V. vulnificus* biotype 3 genome. These two bacteria share similar environmental and clinical properties. The genus *Shewanella* also belongs to the family *Vibrionaceae* and is widely distributed in marine and freshwater environments (Hau and Gralnick, 2007). *Shewanella algae* and *Shewanella putrefaciens* are frequently found in non-human sources but are opportunistically pathogenic to humans (Tsai et al., 2008). Human infections include, among others, bacteremia, cellulitis (skin and soft tissue infection) and wound infection. The typical predisposing factor for infection with *S. algae* or *S. putrefaciens* is exposure to a marine environment with a skin lesion or skin trauma; other factors include the presence of a severe underlying debility, liver disease, or malignancy, and a compromised immune system (Oh et al., 2008). This all pointed to *Shewanella* as a possible gene donor that contributed to the formation of biotype 3, and called for an environmental survey.

Additional analysis revealed that 117 of the "unique" biotype 3 genes have no sequence similarity to any of the known sequences in the NCBI nr database (BLASTn: query coverage <30%, and *E* value >10−10), and most of them (86%) were annotated as hypothetical proteins. This suggested that these genes were acquired from new unknown or unsequenced species present in the marine niche, or underwent major changes during or after their integration into the genome. Results also emphasized that genome comparison in general is strongly dependent on the information available in the databases, and is limited and may not reflect the real picture, further calling for molecular analyses of the bacterial community in the habitat.

Environmental sampling was conducted to retrieve *V. vulnificus* isolates, as well as other bacteria that could serve as "candidate gene donors"—focusing on *Shewanella*—from the evolutionary niche from which biotype 3 probably emerged. Samples were taken from artificial fish ponds and fish stores in the western Galilee region of Israel, as most clinical cases in the last few years have been associated with fish aquaculture in this area (Broza et al., 2009). Hundreds of bacterial isolates were identified after

selective enrichment from skin, gills and fins of Tilapia fish and sediment from a few different sample collections. Fish and sediment samples were found to be contaminated with *V. vulnificus*. Suspected *V. vulnificus* colonies were green on TCBS agar and showed the expected turquoise or white-colony phenotype on CHROMagar, indicating the presence of both biotypes 1 and 3 in the sample (Broza et al., 2009), with lower levels of the latter strains. Successful identification of *V. vulnificus* was confirmed by DNA amplification of the *V. vulnificus*-specific gene *vvh*. In addition, we found colonies that exhibit a typical *Shewanella* species phenotype, i.e., they were black on TCBS agar. Suspected isolates were found in sediment samples and were subjected to 16S rDNA and *gyrB* sequence analyses in comparison to *Shewanella* strains available from the GenBank database, confirming their identity as *Shewanella* species. Four isolates (Sh-4, Sh-6, Sh-9, and RD-4) presented different but rather similar sequence types in both genes. In a phylogenetic analysis, these isolates clustered together with *S. algae* strains and were separate from other *Shewanella* species, suggesting that the isolated colonies belong to *S. algae* (data not shown). Therefore, results showed successful co-isolation of *Shewanella* and *V. vulnificus* and further support the idea that *V. vulnificus* and *Shewanella* share the same environmental niche—aquaculture ponds in Israel—also supporting the probable contribution of *Shewanella* to the evolution of the biotype 3 genome.

Based on these findings, specific primers were designed to trace the presence of "unique" biotype 3 genes in *Shewanella* strains isolated from *V. vulnificus*'s natural environment by PCR amplification. Five different genes representing different cell pathways were checked: the ParE-ParD toxin-antitoxin system, two transcriptional regulators and a site-specific recombinase (**Table 3**). Six *V. vulnificus* strains, representing biotypes 1, 2, and 3, were used as controls. As expected, the tested genes were present in biotype 3 isolates. The five genes were not amplified in the two tested biotype 2 strains. Surprisingly, only one "unique" gene (2770) was present in one *Shewanella* strain (Sh-4), whereas the other four "unique" genes were not amplified in the four *S. algae* isolates*,* but were present in at least one biotype 1 isolate (**Table 3**). Analysis of the PCR-amplified products of the four tested loci showed sequence identity between biotype 1 and 3 strains. Moreover, 100% similarity was observed between *Shewanella* strain Sh-4



*aLocus taq in VVyb1(BT3) genome.*

*bBT1, BT2, and BT3—biotypes 1, 2, and 3, respectively.*

*cAmplification products were checked for sequence similarity.*

and *V. vulnificus* in the amplified fragment (243 bp) of the 2770 gene, supporting these genes' common origin. These results indicate that some of the genes considered to be unique to biotype 3 are present in biotype 1 strains isolated from the same environmental niche, suggesting that they are the possible origin for biotype 3 genes rather than *Shewanella* directly. Nevertheless, the isolated *Shewanella algae* and other yet unidentified *Shewanella* species in the environment might have served as gene donors for biotype 1 strains. Thus, we concluded that biotype 3 probably harbors unique genes acquired from other bacterial species, but the number of these is significantly less than the estimated 214.

This environmental survey, followed by the amplification of selected genes, indicated that the evolution of biotype 1 is a stepwise process that depends mainly on gene transfer within species as well as from other bacterial species sharing the same environment, thus leading to the constant creation of new strains. This is supported by a high diversity among biotype 1 isolates, as revealed by the SNP genotyping array of hundreds of *V. vulnificus* strains (Raz et al., unpublished), and by previous phylogenetic studies (Gulig et al., 2005; Broza et al., 2009, 2012; Sanjuan et al., 2011). According to this hypothesis, the continuous evolution and high horizontal gene transfer in *V. vulnificus* (Quirke et al., 2006; Kim et al., 2011) led to the creation of biotype 1 strains that are most similar to biotype 3, such as the clinical biotype 1 strain v252 that carried all five tested "unique" biotype 3 genes (**Table 3**) or other closely related strains isolated from the environment.

The genome analysis of biotype 3 showing its high similarity to other biotype 1 genomes, together with the presence of some considered biotype 3 "unique" genes in biotype 1 isolates and the SNP haplotype analysis (Raz et al., unpublished), leads to the conclusion that biotype 3's genome was created as yetundiscovered event, based on the core genome of a biotype 1 strain that gained a rather small number of genes by horizontal gene transfer from its natural environment, leading to a change in biotype. Thus, we hypothesized that single episode of genome hybridization of two bacterial populations as suggested previously (Bisharat et al., 2005) may occur in *V. vulnificus*, however, it is less likely that this was the main event for biotype 3 creation. The new strain probably possessed better fitness under the selective pressure in the niche, as demonstrated by the clonal emergence of this biogroup. Similarly, serotype conversion was observed in *V. cholerae* when the specific DNA region responsible for this change was acquired from a non-*Vibrio* source (Stroeher et al., 1992). In addition, changes in gene expression might also cause the biochemical switch that led to the formation of new biotype. Data received from whole-transcriptome comparisons of the two biotypes by RNAseq performed by us (data not shown) and others (Bisharat et al., 2013) may provide an answer to this question.

To learn more about the creation of biotype 3, a full-genome comparison with a phylogenetically related biotype 1 strain should be performed, as this would enable focusing on a limited number of genes that separate the two biotypes. Furthermore, the high genetic diversity among *V. vulnificus* strains calls for an extensive multistrain full-genome comparison study together with high-throughput sequencing of whole bacterial communities in their natural habitat to fully understand the evolution of this human pathogen and the emergence of new virulent strains and biogroups.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb. 2013.00393/abstract

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 29 August 2013; paper pending published: 25 September 2013; accepted: 02 December 2013; published online: 18 December 2013.*

*Citation: Efimov V, Danin-Poleg Y, Raz N, Elgavish S, Linetsky A and Kashi Y (2013) Insight into the evolution of Vibrio vulnificus biotype 3's genome. Front. Microbiol. 4:393. doi: 10.3389/fmicb.2013.00393*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Efimov, Danin-Poleg, Raz, Elgavish, Linetsky and Kashi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### Development of pVCR94-X from *Vibrio cholerae*, a prototype for studying multidrug resistant IncA/C conjugative plasmids

#### *Nicolas Carraro , Maxime Sauvé , Dominick Matteau , Guillaume Lauzon , Sébastien Rodrigue and Vincent Burrus\**

*Département de Biologie, Université de Sherbrooke, Sherbrooke, QC, Canada*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Daniele Provenzano, University of Texas Brownsville, USA Elisa Taviani, Biotechnology Center University Eduardo Mondlane, Mozambique*

#### *\*Correspondence:*

*Vincent Burrus, Département de biologie, Université de Sherbrooke, 2500 Boulevard de l'Université, Sherbrooke, QC J1K 2R1, Canada e-mail: vincent.burrus@ usherbrooke.ca*

Antibiotic resistance has grown steadily in *Vibrio cholerae* over the last few decades to become a major threat in countries affected by cholera. Multi-drug resistance (MDR) spreads among clinical and environmental *V. cholerae* strains by lateral gene transfer often mediated by integrative and conjugative elements (ICEs) of the SXT/R391 family. However, in a few reported but seemingly isolated cases, MDR in *V. cholerae* was shown to be associated with other self-transmissible genetic elements such as conjugative plasmids. IncA/C conjugative plasmids are often found associated with MDR in isolates of *Enterobacteriaceae*. To date, IncA/C plasmids have not been commonly found in *V. cholerae* or other species of *Vibrio*. Here we present a detailed analysis of pVCR94-X derived from pVCR94, a novel IncA/C conjugative plasmid identified in a *V. cholerae* clinical strain isolated during the 1994 Rwandan cholera outbreak. pVCR94 was found to confer resistance to sulfamethoxazole, trimethoprim, ampicillin, streptomycin, tetracycline, and chloramphenicol and to transfer at very high frequency. Sequence analysis revealed its mosaic nature as well as high similarity of the core genes responsible for transfer and maintenance with other IncA/C plasmids and ICEs of the SXT/R391 family. Although IncA/C plasmids are considered a major threat in antibiotics resistance, their basic biology has received little attention, mostly because of the difficulty to genetically manipulate these MDR conferring elements. Therefore, we developed a convenient derivative from pVCR94, pVCR94-X, a 120.5-kb conjugative plasmid which only codes for sulfamethoxazole resistance. Using pVCR94-X, we identified the origin of transfer (*oriT* ) and discovered an essential gene for transfer, both located within the shared backbone, allowing for an annotation update of all IncA/C plasmids. pVCR94-X may be a useful model that will provide new insights on the basic biology of IncA/C conjugative plasmids.

**Keywords:** *Vibrio cholerae***, cholera, antibiotic resistance, IncA/C, conjugative plasmid, pVCR94,** *oriT***, SXT**

#### **INTRODUCTION**

Cholera is an infectious disease caused by the Gram-negative bacterium *Vibrio cholerae* that remains a major threat worldwide, especially for vulnerable territories where water supplies and sanitation are inadequate. The main symptoms of cholera are a profuse watery diarrhea and vomiting caused by the cholera toxin CtxAB encoded by CTXφ, a filamentous phage carried by toxicogenic *V. cholerae* strains (Waldor and Mekalanos, 1996). The resulting rapid loss of fluids and electrolytes leads to severe and often lethal dehydration of patients within a few hours of onset, making oral antibiotics ineffective. Significant measures have been developed to fight this scourge, from public health improvement to vaccines and antibiotic therapies (Desai and Clemens, 2012; Harris et al., 2012). Since cholera is a self-limiting disease, patients usually recover rapidly with prompt and proper rehydration, and electrolyte replacement. Antimicrobial therapies may considerably reduce the severity of diarrhea and the duration of vibrio excretion yet increase the emergence and development of antibiotic resistance (Harris et al., 2012).

Currently, major mediators of antibiotic resistance in epidemic strains of *V. cholerae*remain Integrative and Conjugative Elements (ICEs), and mobile and chromosomal integrons (Waldor et al., 1996; Mazel et al., 1998; Hochhut et al., 2001; Burrus et al., 2006a; Mazel, 2006; Pugliese et al., 2009; Wozniak et al., 2009; Stalder et al., 2012). In fact, most epidemic strains recovered in the past 20 years harbor an ICE belonging to the SXT/R391 family (Burrus et al., 2006a; Reimer et al., 2011; Garriss and Burrus, 2013). The characterization of the first multi-drug resistant O139 *V. cholerae* epidemic strain isolated in India in the early 1990s led to the identification of SXTMO10, which was responsible for the resistance to sulfamethoxazole and trimethoprim (co-trimoxazole), chloramphenicol, and streptomycin of this isolate (Waldor et al., 1996). Nowadays, SXT-like relatives are found in most recent *V. cholerae* isolates including O1 strains such as the ones that caused the 2010 cholera epidemic in Haiti (Ceccarelli et al., 2011; Chin et al., 2011; Reimer et al., 2011; Sjolund-Karlsson et al., 2011; Katz et al., 2013). SXT, as well as all related ICEs, spread into and between bacterial populations *via* conjugation, and ensure their vertical inheritance into daughter cells during cell division by integrating into the host chromosome. In addition to resistance to several antibiotics, some members of the SXT/R391 ICEs also encode diguanylate cyclases that may facilitate their dissemination through manipulating the intracellular level of the second messenger c-di-GMP (Bordeleau et al., 2010).

A few *V. cholerae* epidemic surveys report the implication of conjugative plasmids in multi-drug resistance (MDR) (Sundaram and Murthy, 1984; Garrigue et al., 1986; Kruse et al., 1995; Dalsgaard et al., 2000a; Ceccarelli et al., 2006; Pugliese et al., 2009). Recent characterization of MDR in pandemic *V. cholerae* in Eastern China has shown that resistance to co-trimoxazole, ampicillin, streptomycin, gentamicin, tetracycline, and chloramphenicol was conferred by the large conjugative plasmid pMRV150, which belongs to the A/C incompatibility group (IncA/C) (Pan et al., 2008). Strikingly, while O139 epidemic strains isolated before 1997 were devoid of pMRV150, the proportion of isolates harboring pMRV150 or a related plasmid gradually rose to 92% between 2000 and 2006. IncA/C elements are large (>100 kb), broad host-range and single-copy conjugative plasmids that are mainly known for their considerable contribution to MDR phenotypes in pathogenic *Enterobacteriaceae* infecting humans, food products and food-producing animals (Welch et al., 2007; Fernandez-Alarcon et al., 2011). Comparative genomics studies revealed that the majority of the fully or partially sequenced IncA/C plasmids share a highly conserved backbone of genes of nearly 110 kb, coding for the conjugative transfer and replication machinery (Welch et al., 2007; Fricke et al., 2009; Fernandez-Alarcon et al., 2011). Interestingly, every predicted transfer gene encoded by the IncA/C plasmids are found in synteny in SXT/R391 ICEs, and the identities of these predicted protein sequences vary from 34 to 78% (Wozniak et al., 2009). Despite their prevalence in a wide range of pathogens isolated from very diverse geographic areas and their importance in the spread of MDR, characterization of IncA/C plasmids has been mostly epidemiological and remained limited to typing (Giske et al., 2012), antibiotic resistance profiling (Glenn et al., 2013), sequencing and comparative genomics (Welch et al., 2007; Fricke et al., 2009; Fernandez-Alarcon et al., 2011). Transcriptome analysis of the self-transferrable IncA/C plasmid pAR060302 revealed that most genes remain silent in laboratory conditions with the exception of a toxin/antitoxin gene locus and resistance genes, which seem to be constitutively transcribed (Lang et al., 2012). Interestingly, while little is known regarding their basic biology, IncA/C plasmids have been shown to specifically drive the *trans*-mobilization of the MDR-conferring *Salmonella* genomic island 1 (SGI1) of *S. enterica* (Doublet et al., 2005; Douard et al., 2010).

Rwanda was free of the seventh pandemic of cholera until 1978, when the disease started to be endemic in the African Great Lakes region. In years of political stability, cholera outbreaks in this region were shown to be influenced by climatic conditions, rainfall and fluctuation in plankton populations in what is known as the "cholera paradigm" (Bompangue Nkoko et al., 2011). However, from the 6th of April to the 4th of July 1994, unrest in the population led to a nationwide genocide, and to the exile of an estimated 8 00 000 refugees to the north of Goma, Democratic Republic of the Congo. Lack of an adequate and timely response, size of refugee camps, promiscuity and poor treatment in the early days led to a devastating cholera outbreak in the refugee population, affecting an estimated 36 000 individuals. Treatment was limited to oral rehydration in the early days of the outbreak. Isolated strains were shown to belong to the *V. cholerae* O1 El Tor Ogawa serotype, and were resistant to tetracyclines, aminopenicillins, co-trimoxazole, and nifuroxazide (Bioforce, 1996). An increase in resistance to nalidixic acid over the course of the outbreak was also observed, leaving few options for treatment (Cavallo et al., 1995). At the time, lack of data did not allow precise identification of determinants involved in antibiotic resistance.

In this study, we report the characterization of the novel IncA/C plasmid pVCR94, which is responsible for MDR of a Rwandan epidemic *V. cholerae* isolate. Sequence analysis of pVCR94 revealed its mosaic nature as well as the high degree of similarity of the structural genes responsible for transfer and maintenance among other IncA/C plasmids such as pIP1202 from *Y. pestis*. Transfer assays showed that pVCR94 is capable of mobilizing resistance determinants between *V. cholerae* strains as well as to and from *E. coli*. Genetic engineering of pVCR94 allowed us to initiate a more mechanistic study of this prototype IncA/C plasmid, in which we experimentally identified the location of the origin of transfer (*oriT*) together with a novel gene required for conjugative transfer, both within their shared backbone, allowing for an annotation update of all IncA/C plasmids.

#### **MATERIALS AND METHODS**

#### **BACTERIAL STRAINS AND MEDIA**

The bacterial strains and plasmids used in this study are described in **Table 1**. The strains were routinely grown in Luria-Bertani (LB) broth at 37◦C in an orbital shaker/incubator and were preserved at −80◦C in LB broth containing 15% (vol/vol) glycerol. For *E. coli*, antibiotics were used at the following concentrations: ampicillin (Ap), 100μg/ml; chloramphenicol (Cm), 20μg/ml; erythromycin, 200μg/ml; gentamycin (Gn), 10μg/ml; kanamycin (Kn), 50μg/ml; rifampicin (Rf), 50μg/ml; spectinomycin (Sp), 50μg/ml; streptomycin, 200μg/ml; sulfamethoxazole (Su), 160μg/ml; tetracycline (Tc), 12μg/ml; and trimethoprim (Tm), 32μg/ml. For *V. cholerae*, antibiotics were used at the following concentrations: chloramphenicol, 2μg/ml; kanamycin, 30μg/ml; streptomycin, 10μg/ml; tetracycline, 10μg/ml. When required, bacterial cultures were supplemented with 0.3 mM DL-2, 6-diaminopimelic acid (DAP) or 0.02% L-arabinose. Antibiotics susceptibility profiling were done in three independent experiments using broth microdilution tests (Jorgensen and Ferraro, 2009).

#### **BACTERIAL CONJUGATION ASSAYS**

Conjugation assays were performed by mixing equal volumes of each donor and recipient strains that were grown overnight at 37◦C. The cells were harvested by centrifugation for 3 min at 1200 g, washed in 1 volume of LB broth and resuspended in 1/20 volume of LB broth. Mating mixtures were then deposited on LB agar plates and incubated at 37◦C for 6 h. The cells were recovered

#### **Table 1 | List of strains and plasmids used in this study.**


*ApR, ampicillin resistant; CmR, chloramphenicol resistant; EmR, erythromycin resistant; GnR, gentamycin resistant; KnR, kanamycin resistant; NxR, nalidixic acid resistant; RfR, rifampicin resistant; SuR, sulfamethoxazole resistant; SmR, streptomycin resistant; SpR, spectinomycin resistant; TcR, tetracycline resistant; TmR, trimethoprim resistant; Ts, thermosensitive.*

from the plates in 1 ml of LB broth and serially diluted before plating. Donors, recipients and exconjugants were selected on LB agar plates containing appropriate antibiotics.

#### **MOLECULAR BIOLOGY METHODS**

Plasmid DNA was prepared using the EZ-10 Spin Column Plasmid DNA Minipreps Kit (Biobasic) according to manufacturer's instructions. All the enzymes used in this study were purchased from New England BioLabs. PCR assays were performed with the primers described in **Table 2**. The PCR conditions were as follows: (i) 3 min at 94◦C; (ii) 30 cycles of 30 s at 94◦C, 30 s at the appropriate annealing temperature, and 1 min/kb at 68◦C; and (iii) 5 min at 68◦C. When necessary, PCR products were purified using a EZ-10 Spin Column PCR Products Purification Kit (Biobasic) according to manufacturer's instructions. *E. coli* was transformed by electroporation as described by Dower et al. (1988) in a BioRad GenePulser Xcell apparatus set at 25 μF, 200 V and 1.8 kV using 1-mm gap electroporation cuvettes. Sequencing reactions were performed by the Plateforme de Séquençage et de Génotypage du Centre de Recherche du CHUL (Québec, QC, Canada).

#### **PLASMIDS CONSTRUCTION**

Plasmids and oligonucleotides used in this study are listed in **Tables 1**, **2**, respectively. λRed recombination encoding plasmid pMS1 was obtained after amplification of the gentamycin

#### **Table 2 | DNA sequences of oligonucleotides used in this study.**


*aRestriction sites are underlined.*

resistance cassette of pAH153 (Haldimann and Wanner, 2001) using primer pair Gen153F/Gen153(2)R, digestion of the resulting fragment using EcoRI and NcoI, and subsequent cloning into pSIM5 (Datta et al., 2006) digested with the same enzymes.

Plasmids used for complementation and mobilization assays containing *oriT* (pMA1), *oriT*-*vcrx001* (pMA5), and *vcrx062* (pMA2) were constructed in pACYC177. Fragments to be cloned were amplified using primers oriT144\_F/oriT144\_R (*oriT*), oriT144\_F/oriTmobIPstI\_R (*oriT-vcrx001*), and oriT94oldBamHI\_F/oriT94oldPstI\_R (*vcrx062*), respectively. Both inserts and vector were digested with BamHI and PstI, and ligated together by T4 DNA ligase reaction. Resulting plasmids were verified using restriction profiles and sequencing.

#### **CONSTRUCTION OF CHROMOSOMAL DELETIONS AND TARGETED DELETIONS IN pVCR94**

Deletion mutants were constructed using the one-step chromosomal gene inactivation technique (Datsenko and Wanner, 2000). All deletions were designed to be non-polar. Primers used are listed in **Table 2**. The *recA*::*aad7* mutation of *E. coli* VB45 was introduced in *E. coli* BW25113 using primer pair recAWF/recAWR and pVI36 as a template. Deletion of the MDR region located between *vcrx028* and *vcrx029* of pVCR94 was constructed in *E. coli* VB261 using primer pair vcr94W1F/vcr94W1R and pVI36 as template. The λRed recombination system was expressed using pMS1 as described for the pSIM expression vectors (Datta et al., 2006). The SpR cassette was removed from the resulting construction (pVCR94-X::*aad7*) by Flp-catalyzed excision using the pCP20 vector (Cherepanov and Wackernagel, 1995). The resulting strain containing pVCR94-X (VB557) was used for subsequent deletions of *oriT*1, *oriT*2, *vcrx062*, *vcrx001*, and *oriT*2-*vcrx001* using primer pairs oriT94\_1WF/oriT94\_1WR, oriT94\_2WF/oriT94\_1WR, oriT94old\_WF/oriT94old\_WR, mobI94\_WF/mobI94\_WR, and oriT94\_2WF/mobI94\_2WR, respectively. pVI36 and pSIM5 were used for all these constructions. All deletions were verified by PCR and antibiotic resistance profiling.

#### **SEQUENCING AND SEQUENCE ASSEMBLY OF pVCR94***-***X**

Using genomic DNA of *E. coli* MG1655 RfRharboring pVCR94 or pVCR94-X, Illumina sequencing libraries were prepared as previously described (Rodrigue et al., 2010), and sequenced using 100-bp paired-end reads. The resulting sequences were assembled using version 2.6 of the *de novo* gsAssembler (Newbler) software by first removing sequences fully mapping to the *E. coli* MG1655 genome. The contigs were combined by using PCR and Sanger sequencing reactions. The resulting pVCR94-X sequence was annotated using the RAST automated pipeline (Aziz et al., 2008), manually curated and submitted to Genbank under accession KF551948.

#### **BLAST ATLAS REPRESENTATION OF SEQUENCES COMPARISON**

A circular BLAST Atlas was computed by GView (Petkau et al., 2010) on the GView server (https://server.gview.ca/) for each sequenced IncA/C plasmid using the BlastN algorithm and mapped against pVCR94-X. Sequences were aligned using raw sequence data, with an expect value cutoff of 1 <sup>×</sup> <sup>10</sup>−10, an alignment length cutoff of 100, and a percent identity cutoff of 75%. Accession number of the sequences are as follow: pVmi603 (ACYU01000017.1), pIP1202 (NC\_009141.1), pYR1 (CP000602.1), pRA1 (NC\_012885.1), pR148 (NC\_019380.1), pR55 (NC\_016976.1), pNDM-KN (NC\_019153.1), LS6 (JX442976.1), pNDM100469 (JN861072.1), pKPHS3 (CP003225.1), pSN254 (CP000604.1), pSD\_174 (JF267651.1), pAM04528 (NC\_012693.1), pPG010208 (NC\_019065.1), pAPEC1990\_61 (NC\_019066.1), pAR060302 (NC\_019064.1), pUMNK88 (NC\_017645.1), pNDM-1\_Dok01 (NC\_018994.1), pNDM10505 (JF503991.1), pNDM102337 (JF714412.2), pEH4H (NC\_012690.1), pMR2011 (JN687470.1), pTC2 (JQ824049.1), pP99-018 (AB277723.1), pP91278 (AB277724.1), pXNC1 (FN667743.1).

#### **RESULTS**

#### **THE 1994 RWANDAN CHOLERA OUTBREAK INVOLVED A MULTIRESISTANT** *V. cholerae* **ISOLATE HARBORING A CONJUGATIVE PLASMID**

*V. cholerae* F1939 is a co-trimoxazole-resistant O1 El Tor clinical isolate recovered from a refugee camp during the 1994 Rwanda cholera outbreak (O'shea et al., 2004a,b). F1939 was initially identified as a strain capable of transferring the co-trimoxazole resistance to *V. cholerae* E4 (**Table 1**). As such, we expected that the resulting exconjugant, *V. cholerae* BI144, would carry an ICE of the SXT/R391 family, which are major vectors of co-trimoxazole resistance in epidemic *V. cholerae* strains. All our attempts to PCR amplify markers typical of SXT/R391 ICEs (*int*SXT and *setR*) from genomic DNA of BI144 failed (data not shown). Instead, plasmid typing based on PCR amplification of *repA* revealed that in the exconjugant *V. cholerae* BI144 co-trimoxazole resistance was conferred by a plasmid of the IncA/C group that we named pVCR94 (plasmid *Vibrio cholerae* Rwanda 1994).

#### **pVCR94 DISSEMINATES MULTIDRUG RESISTANCE AT HIGH FREQUENCIES**

To further characterize pVCR94, mating experiments were carried out to transfer the plasmid from *V. cholerae* BI144 to *E. coli* VB111 and VB112 strains (MG1655 Nx<sup>R</sup> and MG1655 RfR, respectively) by co-trimoxazole selection. As expected, since IncA/C plasmids have a broad host-range, pVCR94 efficiently transferred to and was stably maintained in *E. coli*, giving rise to NC213 and NC207, respectively. Additional antibiotic resistance testing of *E. coli* exconjugants revealed that pVCR94 also confers resistance to ampicillin, streptomycin, tetracycline, and chloramphenicol, but not to gentamycin, kanamycin, rifampicin, nalidixic acid, and erythromycin (**Table 3**).

Preliminary tests showed that pVCR94 self-transfers at high frequency. To evaluate its potential to disseminate MDR, tetracycline and chloramphenicol resistance markers were used to test conjugative transfer of pVCR94 among and between *E. coli* and *V. cholerae*. Intraspecific transfer of pVCR94 was first tested from its primary host *V. cholerae*. Two El Tor variants were used, NC212 (N16961 containing pVCR94) as the donor and E4 as the recipient, revealing that pVCR94 transfers at very high frequency (∼<sup>3</sup> <sup>×</sup> <sup>10</sup>−<sup>1</sup> exconjugant per recipient) (**Figure 1**). For intraspecific transfer of pVCR94 in *E. coli*, otherwise isogenic Nx<sup>R</sup> and RfR MG1655 derivative strains VB111 and VB112 were used. Despite a ∼6-fold reduction of transfer, pVCR94 still transferred very efficiently under these conditions (**Figure 1**). Interspecific transfer of pVCR94 using *V. cholerae* NC212 as a donor and *E. coli* VB111 as a recipient and *vice versa*, *E. coli* NC208 as a donor and *V. cholerae* N16961 as a recipient, occurred at roughly the same frequency (∼10−<sup>2</sup> exconjugant per recipient cell) (**Figure 1**).

#### **pVCR94 SHARES A COMMON BACKBONE WITH IncA/C PLASMIDS**

Genomic DNA of *E. coli* NC207 (VB112 harboring pVCR94) was extracted to sequence the plasmid using the Illumina technology.

**Table 3 | Minimal inhibitory concentrations (MIC) of 12 antibiotics against** *E. coli* **carrying pVCR94 or its** *-***X mutant.**


*and, not determined, test was only done on solid agar plate. All assays were carried out in three independent replicates.*

*bR, resistant; S, susceptible.*

*cThese assays were carried out using the RfR derivatives VB112, NC207 (VB112 pVCR94) and NC222 (VB112 pVCR94*-*X), respectively.*

donor, *E. coli* MG1655 Nx<sup>R</sup> (VB111) as the recipient and exconjugants were selected as NxR Tc<sup>R</sup> CmR resistant colonies. On the other hand, *E. coli* NC208, an otherwise isogenic strain auxotrophic for diaminopimelic acid (DAP) and containing pVCR94, was used as the donor, the Sm<sup>R</sup> strain *V. cholerae* N16961 as the recipient and *V. cholerae* exconjugants were selected as SmR Tc<sup>R</sup> CmR colonies in the absence of DAP. The bars represent the mean and standard deviation values obtained from at least three independent experiments. pVCR94 and pVCR94-X transfer frequencies are expressed as a number exconjugant per recipient colonies.

A draft assembly generated a sequence of 134,484 bp. Initial sequence analyses revealed the presence of only three resistance genes, *sul1*, *sul2*, and *dfrA15,* conferring resistance to cotrimoxazole. *sul1* and *dfrA15* belong to a class 1 mobile integron devoid of any other integron cassette (**Figure 2**) Identical integron structures have already been described in clinical *V. cholerae* strains belonging to non-O1, non-O139 serogroups isolated in the mid-90s in Thailand and India (Dalsgaard et al., 2000b; Thungapathra et al., 2002). Further analysis confirmed that the sequence of pVCR94 was partial; a segment of unknown size overlapping most other antibiotic resistance genes could not be properly assembled. We constructed a deletion mutant, pVCR94-X, which lacks this region (see below), and submitted the resulting 120,572-bp sequence to Genbank database (accession number KF551948) (**Figure 3A**).

Comparative genomics confirmed the conservation in pVCR94-X of the large core set of genes that are common to all members of the IncA/C group (**Figure 3B**), including genes that may be involved in the regulation of IncA/C conjugative transfer (**Table 4**). Among them, *vcrx148* and *vcrx149* are particularly interesting. Pfam analyses (database v27.0) revealed that the Vcrx149 protein contains an FlhC signature (Pfam PF05280) whereas Vcrx148 has very weak homology with the FlhD domain (Pfam PF05247). Sequence comparisons also brought to light variable regions encoding hypothetical proteins of unknown function (VR2, VR3, VR4, VR5, and VR6) (**Figure 3**). The 4.5-kb

variable region VR1 contains DNA that is only found in pVCR94 and codes for a putative cadmium-bromine pump efflux and its cognate regulator (*vcrx049*, **Table 4**) along with a transposase gene. The 1.5-kb VR7 region codes for a putative transmembrane protein with an EndoU\_bacteria nuclease domain (Pfam PF14436) and is predicted to be a secreted bacterial toxin (Zhang et al., 2011).

Interestingly, comparison of the genes found in variable regions of pVCR94 with those of other sequenced IncA/C plasmids indicate that pVCR94 is more closely related to plasmids recovered from *P. damselae* (pP91278 and pP99-018), *K. pneumoniae* (pR55), *E. coli* (pPG010208), and *P. stuartii* (pTC2) than from the plasmid identified in *V. mimicus* (pVmi603) (**Figure 3**).

#### **pVCR94***-***X, A CONVENIENT PROTOTYPE FOR THE STUDY OF THE BASIC BIOLOGY OF IncA/C PLASMIDS**

To facilitate future studies of IncA/C plasmid biology without the challenges and limitations associated with the MDR phenotype usually conferred by these mobile elements, we decided to construct a mutant of pVCR94 coding for a reduced set of antibiotic resistance. Antibiotic resistance markers carried by vectors used for expression of λRed recombination system, ApR for pKD46 and ApR, CmR, KnR, or SpR for pSIM vectors (Datsenko and Wanner, 2000; Datta et al., 2006), are not compatible with pVCR94 and most known IncA/C plasmids. To circumvent this problem, we constructed a Gn<sup>R</sup> derivative of pSIM5 (pMS1) to allow expression of λRed recombination function in this multiple antibiotic resistance context.

Using pMS1 and a SpR cassette, we deleted the large fragment containing the MDR-conferring genes in pVCR94 and located between *vcrx028* and *vcrx029* (*sul2*), which encode a hypothetical protein and resistance to sulfamethoxazole respectively. After elimination of the Sp<sup>R</sup> cassette, antibiotic resistance examination confirmed the sensitivity of *E. coli* MG1655 containing pVCR94-X (NC367) to all tested antibiotics but sulfamethoxazole (**Table 3**). Despite the large deletion, pVCR94-X was able to stably maintain in *E. coli*. Finally, mating experiments showed that pVCR94-X remains self-transmissible at the same frequency as wild-type pVCR94 (**Figure 1**). Although the exact gene content and size of the region that was deleted remains to be established, our functional tests indicate that pVCR94-X now constitutes a convenient prototype for in-depth molecular study of pVCR94 and related IncA/C plasmids.

#### **FIGURE 3 | Sequence analysis of pVCR94***-***X. (A)** Schematic

representation of the genetic organization of pVCR94-X. The location and orientation of ORFs are indicated by arrowed boxes. The color of the arrowed boxes depicts the putative function or relationships of each ORF deduced from functional analyses and BLAST comparisons: white, unknown function; blue, conjugative transfer; orange, replication; yellow, antibiotic resistance; gray, regulation; red, homologous recombination. The origin of replication (*oriV*) and the origin of transfer (*oriT*) are symbolized by an orange and a blue star, respectively. The position of the scar resulting from the deletion of the multidrug resistance gene cluster is indicated (FRT site). **(B)** Genetic comparison of pVCR94-X and other sequenced IncA/C plasmids. A BLAST Atlas was constructed with the pVCR94-X sequence set as the reference (innermost circle). All completely sequenced IncA/C plasmids available in

Genbank were aligned according to their raw sequence data toward pVCR94-X using a BlastN algorithm. Coding sequences of pVCR94-X appear on the innermost circle in blue for the positive strand, and red for the negative strand. All other aligned plasmids sequences are represented only according to their sequence homology toward the reference. Full color saturation represents 100% sequence identity, and gaps indicate regions of divergence (<75% percentage of nucleic acid identity). The black arrow indicates the position of the deletion that generated pVCR94-X. Part of an IncA/C plasmid closely related to pVCR94 was detected among at least 3 different contigs of the unassembled *V. cholerae* RC9 genome (Genbank accession number ACHX00000000). Since the sequence of this plasmid is not assembled and probably not complete, it was not included in this analysis.


**Table 4 | Open reading frames (ORFs) of pVCR94***-***X coding for putative transcriptional regulators.**

*aSize in amino-acids of the predicted translation product.*

*bLevel of expression of ortholog genes in the IncA/C plasmid pAR060302 (Lang et al., 2012).*

*cThis domain was reported as an insignificant Pfam-A match.*

#### **IDENTIFICATION OF** *oriT* **IncA***/***C, THE ORIGIN OF TRANSFER OF IncA/C PLASMIDS**

Conjugative transfer is initiated at a specific *cis*-acting locus called the origin of transfer (*oriT*) by a DNA relaxase, which is typically called TraI. Fricke et al. (2009) and Welch et al. (2007) proposed to position the *oriT* locus of IncA/C plasmids (*oriT*IncA/C) between the genes *traD* and *traJ*. This annotation was based not on experimental data but rather on an analogy with the location of *oriT* of the ICE SXT proposed by Beaber et al. (2002). However, the sequence located between *traD* and *traJ* in SXT has been shown to be unable to support the mobilization of a non-mobilizable plasmid and *oriT* of the SXT/R391 ICEs was subsequently relocated upstream of a gene named *mobI* (Ceccarelli et al., 2008). Therefore, we hypothesized that the region located between *traD* and *traJ* in IncA/C plasmids, which contains *vcrx062* in pVCR94-X, is not *oriT*IncA/<sup>C</sup> and experimentally investigated the location of the *oriT*IncA/<sup>C</sup> in our model. Noteworthy in the following experiments, the use of alternative donor and recipient strains caused a tenfold reduction in transfer of pVCR94-X compared to the previous experiments (**Figures 1**, **4A**), thereby suggesting that the genetic background has a notable influence on the efficiency of IncA/C plasmids transfer.

First, we verified whether the region containing *vcrx062* was required for transfer and whether it was sufficient to support the mobilization of the non-mobilizable low-copy plasmid pACYC177. We used as a donor *E. coli* MS2, which harbors pVCR94-X *vcrx062* (SpR), to mobilize a pACYC177 derivative containing *vcrx062* (pMA2, KnR) to the TcR strain *E. coli* CAG18439. Exconjugants were independently selected for acquisition of pMA2 (Tc<sup>R</sup> KnR) or pVCR94-X *vcrx062* (Tc<sup>R</sup> SpR). We observed that deletion of *vcrx062* did not affect the transfer efficiency of pVCR94-X (**Figures 4AI,II**). Furthermore, the *vcrx062* locus was incapable to initiate transfer of pMA2 as no Tc<sup>R</sup> KnR exconjugant could be recovered (**Figure 4AII**). These results provide convincing evidence that the locus located between *traD* and *traJ* is not an *oriT* for IncA/C plasmids. Interestingly, in the presence of pMA2, transfer of pVCR94-X dropped significantly. This phenotype was not observed in the presence of pACY177.

In the ICEs of the SXT/R391 family, *oriT* is located in a large intergenic region between two divergent genes: *mobI*, which is crucial for conjugative transfer, and *s003*, a gene of unknown function (Ceccarelli et al., 2008). Comparison of the *s003*-*mobI* locus of SXT with the corresponding region of pVCR94 revealed striking similarities in gene organization (**Figure 4B**). In fact, in pVCR94 *vcrx001* encodes a protein sharing a weak identity with MobI (**Figure 4C**) and *vcrx152* encodes a protein sharing 58% identity with S003. To test whether *oriT*IncA/<sup>C</sup> was located between *vcrx152* and *vcr001*, plasmid pMA1 (KnR) was constructed by cloning this intergenic region (putative *oriT*) into pACYC177 (**Figure 4D**). pMA1 was subsequently introduced into *E. coli* MS1 and MS5, two Sp<sup>R</sup> strains containing pVCR94-X *oriT*<sup>1</sup> or *oriT*2, respectively (**Figure 4D**). In both of these mutants, deletions were designed to preserve the promoter region upstream of *vcrx001*. We observed that while neither deletion abolished the transfer of pVCR94, both led to a significant 10 fold reduction of transfer (**Figures 4AIII,IV**). In addition, while mobilization of pACYC177 by pVCR94-X *oriT*<sup>1</sup> or *oriT*<sup>2</sup> was undetectable, pMA1 was mobilized at a frequency of 1.<sup>3</sup> <sup>×</sup> <sup>10</sup>−<sup>5</sup> to 4.<sup>0</sup> <sup>×</sup> <sup>10</sup>−<sup>5</sup> exconjugant per donor cell (**Figures 4AIII,IV**). Altogether, these results suggest the presence of an *oriT* between *vcrx152* and *vcrx001* in pVCR94-X as this locus is a weak yet suitable substrate for transfer initiation. However, the low efficiency of mobilization conferred by this locus and the ability of the *oriT*<sup>2</sup> mutant of pVCR94-X to transfer efficiently suggest that other alternative *oriT* loci may exist in IncA/C plasmids. Alternatively, the actual *oriT* may also include the promoter region of *vcrx001* and perhaps the 5 end this open reading frame (see below).

#### **vcrx001 IS AN ESSENTIAL GENE FOR CONJUGATIVE TRANSFER OF IncA/C PLASMIDS**

Transfer of SXT is abolished in the absence of MobI, a protein that is likely an auxiliary component of the relaxosome (Ceccarelli

**FIGURE 4 | Identification of the** *oriT* **locus of pVCR94. (A)** Conjugation and mobilization assays were carried out to assess the impact of deletion, and the ability to *trans*-initiate transfer of the indicated regions cloned into pACYC177 (see panel **D**). In every experiment, pVCR94-X was used as a positive control for efficiency of transfer, and pACYC177 was used as a negative control for mobilization assays. Frequency of transfer of each deleted region was compared to the mobilization frequency of its cognate cloned sequence. Within each mating experiment, exconjugants were selected for their acquisition of either the pACYC177 derivatives, pVCR94-X derivatives, or for cotransfer of both, when applicable. All mating experiments that involved *recA*<sup>+</sup> strains were made from BW25113 Nx<sup>R</sup> as donor toward the Tc<sup>R</sup> strain CAG18439 as recipient. Transfers done in a *recA*<sup>−</sup> background involved BW25113 *recA*::*aad7* (VB45) and KB1 as donor and recipient strains, respectively. The bars represent the mean and standard deviation values obtained from three independent experiments. Asterisk indicates that frequency of exconjugant formation was below the limit of detection (<10−8). N.A. indicates that the selection was not applicable in the mating experiment. Statistical analyses were performed using the two-tailed Student's *t*-test. *P*-values are indicated above the bars when comparison referred to pVCR94-X (panel **AI**) or above the brackets comparing two bars. **(B)**

Comparison of the genetic context of *oriT* loci in SXT and pVCR94. Arrows of similar color represent genes predicted to have similar functions: dark blue, conjugative transfer; orange, replication; gray, H-NS-like DNA-binding protein; red, site-specific recombination; green, DNA repair; white, unknown function. Blue stars indicate the *oriT* loci. The orange star indicates the position of the origin of replication (*oriV*) of pVCR94 based on identity with pRA1. The red star indicates the position of the *attP* site for chromosomal integration of SXT by site-specific recombination. The percent of identity of orthologous proteins are indicated on dashed lines. **(C)** Amino acid sequence alignment of the translation products of *vcrx001* and *mobI* computed by Clustal Omega (Sievers et al., 2011). Similarities (gray) and identities (black) are visualized using the BLOSUM62 substitution matrix. **(D)** Schematic representation of the region of pVCR94 encompassing the end of *vcrx152* and the end of *vcrx002*. The inserts of the plasmids used in mobilization experiments, all of which were derived from the low-copy non-mobilizable vector pACYC177, are represented above the genetic map by overlapping segments delimited by arrows pointing outwards. Deletions within the region are depicted below the genetic map by overlapping segments delimited by arrows pointing inwards. The positions of the oligonucleotides used for amplification and cloning or construction of the deletions are indicated (**Table 2**).

et al., 2008). MobI of SXT/R391 ICEs exhibits 28% identity (40% similarity) with its IncA/C ortholog *vcrx001* (**Figures 4B,C**). Annotated in all available IncA/C plasmids sequences as a gene coding for a hypothetical protein, the importance of *vcrx001* in conjugative transfer of these plasmids has never been investigated. Furthermore, the low frequency of mobilization of pMA1 observed above could be due to the lack of an adjacent *mobI*-like gene as it has already been reported for SXT (Ceccarelli et al., 2008).

We constructed *E. coli* MS3, which harbors pVCR94-X *vcrx001* (SpR), to test the importance of the *mobI* ortholog for IncA/C transfer. As reported for SXT *mobI*, transfer of the *vcrx001* mutant of pVCR94-X was completely abolished (**Figure 4AV**), confirming the essential role of *vcrx001* in conjugative transfer. To further investigate its function, the 1063-bp fragment overlapping *vcrx001* and the upstream intergenic region was cloned into pACYC177 to generate pMA5 (**Figure 4D**). Mating experiments using pMA5 were carried out to evaluate the impact of the *cis*-expression of *vcrx001* on a mobilizable plasmid containing *oriT*IncA/C. Exconjugants colonies selected for transfer of pMA5 or pVCR94-X *vcrx001* were recovered at a frequency of 1 <sup>×</sup> <sup>10</sup>−3, which is not dramatically different from transfer of pVCR94-X (**Figure 4A**V). Thus, abolition of transfer observed in the *vcrx001* mutant is not due to a polar effect of the deletion on expression of the *repA* gene, as it can be complemented by expression from its endogenous promoter on a plasmid. Furthermore, the presence of *vcrx001 in cis* significantly improved the transfer of pMA5 over pMA1 (10-fold increase).

Since pVCR94-X *vcrx001* and pMA5 share an identical 514-bp fragment corresponding to the *vcrx152*-*vcrx001* intergenic region, we tested whether the mobilization observed for pMA5 could result from cointegrate formation between the two plasmids mediated by homologous recombination. We used two complementary approaches to test this hypothesis. First to rule out the RecA-mediated recombination, mobilization of pMA1 and pMA5 by pVCR94-X was tested using *recA* donor and recipients strains (*E. coli* VB45 and KB1, respectively). Results showed that transfer of pVCR94-X is RecA-independent and that pMA5, as well as pMA1, were still mobilized in this *recA* background, with the same 10-fold improvement due to the presence of *vcxr001* in pMA5 (**Figure 4AVII**). Second, since homologous recombination could also potentially be mediated by the activity of the putative λRed recombination system carried by IncA/C plasmids (*ssb*, *bet*, *exo*, see **Figure 3A**) we constructed a -(*oriT*2-*vcrx001*) mutant of pVCR94-X devoid of homologous sequence in pMA5, thereby resulting in *E. coli* MS6 (**Figure 4D**). pVCR94-X -(*oriT*2-*vcrx001*) was no longer able to transfer and this mutation was not complemented by expression of *vcrx001* from pMA5, thereby confirming that *oriT*IncA/<sup>C</sup> is located within the deleted fragment (**Figure 4AVI**). Furthermore pMA5 itself remained mobilizable at very high frequency by pVCR94-X -(*oriT*2-*vcrx001*), despite the absence of homologous sequences between the two plasmids confirming that the insert of this plasmid contains *oriT*IncA/C. Interestingly, mobilization of pMA5 was even improved in this context suggesting that different replicons carrying *oriT*IncA/<sup>C</sup> likely compete against with each other during transfer.

### **DISCUSSION**

Cholera remains one of the most devastating human diseases in the world mainly because of toxicigenicity, transmissibility, the rapid multiplication of *V. cholerae* in favorable conditions, and the MDR phenotype of pandemic and epidemic strains. In most modern cholera outbreaks, MDR has been shown to be conferred by SXT/R391 ICEs (for review, Garriss and Burrus, 2013). In the present study, we described the IncA/C plasmid pVCR94 recovered from the multidrug-resistant strain *V. cholerae* O1 El Tor F1939 as the element most likely responsible for the cotrimoxazole resistance during the severe 1994 cholera epidemics in Rwanda. This conclusion is further supported by the absence of the ICE SXT from two other co-trimoxazole resistant *V. cholerae* O1 El Tor isolates, F1873 and F1875, recovered from Rwandan refugees in the summer of 1994 in Goma, Democratic Republic of the Congo (Waldor et al., 1996). Mating assays showed that pVCR94 is a high-efficiency driver of antibiotic resistance dissemination between *V. cholerae* strains as well as to and from *E. coli*. Studies carried out on Asian pandemic isolates of *V. cholerae* underline the rapid emergence and dissemination of IncA/C plasmids as key drivers of antibiotic resistance between 1994 and 2000 (Pan et al., 2008). Although the rate of dissemination of SXT/R391 ICEs between vibrios is rather low in the laboratory as their transfer frequency rarely exceeds 1 <sup>×</sup> <sup>10</sup>−<sup>5</sup> in intra-species mating experiments (Waldor et al., 1996; Burrus et al., 2006b; Osorio et al., 2008), these elements have been extremely successful in invading environmental and clinical *V. cholerae* in the past two decades (Chin et al., 2011; Mutreja et al., 2011; Yu et al., 2012; Garriss and Burrus, 2013; Katz et al., 2013). On the contrary, IncA/C plasmids such as pVCR94 are much more efficient (1 <sup>×</sup> <sup>10</sup>−1) and seem to transfer even better between vibrios than in *E. coli*. It has recently been shown that pVCR94 transfers at high frequency between *E. coli* strains in sludge resulting from the coagulation/flocculation treatment of surface water, reaching the highest frequency after 72 h with about 1 <sup>×</sup> <sup>10</sup>−<sup>2</sup> exconjugants per recipient cell (Pariseau et al., 2013). These observations indicate that, in laboratory conditions as well as in a simulated environmental setting, pVCR94 is a very efficient conjugative plasmid able to invade a significant proportion of the surrounding compatible cells. Knowing the prevalence of IncA/C plasmids in pathogenic bacteria isolated from humans and food-producing animals, their circulation in clinical and environmental *V. cholerae* isolates is worrisome and their impact on the emergence of new pathogenic isolates needs to be surveyed.

Until now, relatively little has been done to characterize the basic biology of IncA/C plasmids despite the significant threat that they represent in the war against MDR pathogenic bacteria. In fact, one of the major challenges encountered to genetically manipulate IncA/C plasmids is their propensity to confer MDR to their host. As a consequence, the majority of antibiotic resistance phenotypes conferred by molecular engineering tools (plasmids, selection gene cassettes) are also conferred by IncA/C plasmids. A gentamycin-resistant version of a λRed recombination expression plasmid that is compatible with pVCR94 allowed us to construct pVCR94-X, a derivative lacking the MDR region. Sequence analyses and experimental evidences revealed that pVCR94-X carries the core sequences necessary for self-transfer and maintenance of IncA/C plasmids. Using this plasmid as a convenient prototype for the study of IncA/C biology, we have identified the locus containing *oriT*IncA/C. Extensive protein sequence conservation and gene synteny between IncA/C plasmids and SXT/R391 ICEs were crucial in this process. We also showed that *vcrx001*, the ortholog of *mobI* of SXT, plays a key role in transfer and enhances the mobilization of a plasmid containing *oriT*IncA/<sup>C</sup> when it is located on the same replicon. In SXT, *mobI* was reported to be a *cis*-acting sequence coding for a putative auxiliary component of the relaxosome required for SXT transfer (Ceccarelli et al., 2008). Surprisingly, deletion of the center (*oriT*1) and the left part (*oriT*2) of the intergenic region between *vcrx152* and *vcrx001* did not affect much the transfer of pVCR94. This suggests that *oriT*IncA/<sup>C</sup> overlaps the right part of this intergenic region and perhaps the 5 end of *vcrx001*. Experiments aimed at discovering the minimal *oriT* sequence required for initiation of transfer of IncA/C plasmids are ongoing. Furthermore, the mutation *oriT*<sup>2</sup> which removed the promoter upstream of *vcrx152* likely prevents expression of three genes, two of unknown function (*vcrx152* and *vcrx151*) and one coding for a predicted H-NS-like protein (*vcrx150*) (**Table 2**). H-NS-like proteins encoded by conjugative plasmids have been shown to provide stealth function helping the transmission of the plasmid into a naïve host (Doyle et al., 2007). While the ortholog of *vcrx150* was shown to be expressed in the IncA/C plasmid pAR060302 (Lang et al., 2012), our deletion did not seem to alter significantly the stability of pVCR94 or its ability to transfer.

Mobile genetic elements are characterized by a modular structure, each module containing the genes and sequences involved in a same biological function (Toussaint and Merlin, 2002). Clustering is an efficient way to exchange and transfer "en bloc" fully functional modules and thus confers new adaptive traits in one event (Lawrence and Roth, 1996). Previous sequence analyses of IncA/C plasmids highlighted their modular structure, with specific variable regions corresponding to mobile integrons and transposable elements conferring adaptive traits such as multiple antibiotics resistance (Welch et al., 2007; Fricke et al., 2009). As observed for pIP1202 and other IncA/C plasmids, pVCR94 carries a class 1 integron. While this integron carries only a *dfrA15* cassette conferring resistance to trimethoprim, its impact should not be underestimated. On one hand, Baharoglu et al. (2010) demonstrated that incoming single-stranded DNA during conjugative transfer triggers the SOS response in the recipient cell, in both *E. coli* and *V. cholerae*. On the other hand, SOS response enhances integron cassette rearrangements through excision/integration, providing opportunities for different integrons present in the same host to exchange cassettes (Guerin et al., 2009). Thus, conjugative transfer of pVCR94 could lead to integron cassettes trapping and drive their dissemination between bacterial communities. Sequence analysis of pVCR94 also revealed the specific region VR1 corresponding to a putative transposon that could confer resistance to heavy metals. This resistance cluster is exclusively carried by the chromosome of various Gram-negative bacteria. Thus, IncA/C plasmids drive horizontal transfer of chromosomal loci and other mobile genetic elements at high frequency by *cis*-mobilization. Beyond intra- and inter-molecular rearrangements among and between chromosome and plasmid, pVCR94 could also mediate *trans*-mobilization of genomic islands present in its host range, as demonstrated for the *S. enterica* pathogenicity island SGI1 (Douard et al., 2010). Finally, two core genes of IncA/C plasmids code for homologs of λBet and λExo proteins of the λRed recombination system. Many conjugative plasmids and ICEs code for such proteins, which have been shown to generate diversity of SXT/R391 ICEs (Garriss et al., 2009, 2013). Similarly, IncA/C plasmids could enhance their plasticity by recombining with a replicon present in the same cell that shares short identical sequences. Altogether, these observations give glimpses of the high dynamics of IncA/C plasmids and their impact on genome plasticity, which could have significant implications for pathogenic bacteria and forecast a bleak future for antibiotherapies.

Many questions remain regarding the coexistence in *V. cholerae* of IncA/C plasmids and SXT/R391 ICEs as two different yet related entities capable of conferring MDR and in particular, resistance to co-trimoxazole. For instance, no single isolate bearing both types of element has been described to date with the exception of isolates from Eastern China that were found to bear both pMRV150-like IncA/C plasmids and SXT-like elements (Pan et al., 2008). Although the IncA/C plasmids in these strains were shown to be able to transfer to recipient cells, no exconjugant bearing also a copy of the SXT-like elements was found, thereby suggesting that the latter were not functional or that negative interference between the two families of mobile elements occurs. Given these observations, we are wondering whether there is a mutual exclusion of IncA/C plasmids and SXT/R391 ICEs. If so, given the relative efficiency of transfer of both types of mobile genetic elements, can we expect to observe a displacement of SXT/R391 ICEs by IncA/C plasmids in clinical and environmental populations of *V. cholerae* in the near future? Given the prevalence of IncA/C plasmids in a plethora of bacteria, their broad host-range and their ability of mobilize MDR-conferring genomic islands, close attention needs to be paid concerning their circulation and evolution. Sequencing of pathogenic isolates bearing IncA/C plasmids and sequences analyses provide valuable information regarding the epidemiology of IncA/C plasmids but molecular characterization of their mechanism of transfer remain unavoidable to unravel the characteristics that make them so successful in modern pathogens. Using pVCR94-X as a prototype, it is now easier to explore the biology and regulation of IncA/C plasmids.

#### **ACKNOWLEDGMENTS**

We are grateful to M.K. Waldor for the kind gift of *Vibrio cholerae* BI144. We thank Alain Lavigueur and Eric Bordeleau for their insightful comments about the manuscript. This work was supported by a Discovery Grant (Vincent Burrus and Sébastien Rodrigue) and Discovery Acceleration Supplement from the Natural Sciences and Engineering Council of Canada (Vincent Burrus), and by a team research project from the Fonds de recherche du Québec – Nature et technologies awarded to Sébastien Rodrigue and Vincent Burrus. Vincent Burrus holds a Canada Research Chair in bacterial molecular genetics and is a member of the FRSQ-funded Centre de Recherche Clinique Étienne-Le Bel.

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 26 August 2013; paper pending published: 06 October 2013; accepted: 21 January 2014; published online: 06 February 2014.*

*Citation: Carraro N, Sauvé M, Matteau D, Lauzon G, Rodrigue S and Burrus V (2014) Development of pVCR94*-*X from Vibrio cholerae, a prototype for studying multidrug resistant IncA/C conjugative plasmids. Front. Microbiol. 5:44. doi: 10.3389/ fmicb.2014.00044*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Carraro, Sauvé, Matteau, Lauzon, Rodrigue and Burrus. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### *Vibrio cholerae* O1 epidemic variants in Angola: a retrospective study between 1992 and 2006

#### *Romy Valia1†, Elisa Taviani 1,2†, Matteo Spagnoletti 3, Daniela Ceccarelli 4, Piero Cappuccinelli <sup>5</sup> and Mauro M. Colombo1,2\**

*<sup>1</sup> Dipartimento di Biologia e Biotecnologie C. Darwin, Università di Roma Sapienza, Rome, Italy*

*<sup>2</sup> Centro de Biotecnologia, Universidade E. Mondlane Maputo, Mozambique*

*<sup>3</sup> Department of Genetics, University College London Genetics Institute, Evolution and Environment, University College London, London, UK*

*<sup>4</sup> Department of Cell Biology and Molecular Genetics, Maryland Pathogen Research Institute, University of Maryland, College Park, MD, USA*

*<sup>5</sup> Dipartimento di Scienze Biomediche, Università di Sassari, Sassari, Italy*

#### *Edited by:*

*Hongyue Dang, Xiamen University, China*

#### *Reviewed by:*

*Irma N. G. Rivera, University of São Paulo, Brazil Yael Danin-Poleg, Technion - Israel Institute of Technology, Israel*

#### *\*Correspondence:*

*Mauro M. Colombo, Dipartimento di Biologia e Biotecnologie Charles Darwin, Università di Roma Sapienza, Via dei Sardi 70, 00185 Roma, Italy e-mail: mauro.colombo@uniroma1.it †These authors have contributed equally to this work.*

Cholera is still a major public health concern in many African countries. In Angola, after a decade of absence, cholera reemerged in 1987, spreading throughout the country until 1996, with outbreaks recurring in a seasonal pattern. In 2006 Angola was hit by one of the most severe outbreaks of the last decade, with ca. 240,000 cases reported. We analyzed 21 clinical strains isolated between 1992 and 2006 from several provinces throughout the country: Benguela, Bengo, Luanda, Cuando Cubango, and Cabinda. We used two multiplex PCR assays to investigate discriminatory mobile genetic elements (MGE) [Integrative Conjugative Elements (ICEs), VSP-II, GI12, GI14, GI15, K, and TLC phages] and we compared the profiles obtained with those of different reference *V. cholerae* O1 variants (prototypical, altered, and hybrid), responsible for the ongoing 7th pandemic. We also tested the strains for the presence of specific VSP-II variants and for the presence of a genomic island (GI) (WASA-1), correlated with the transmission of seventh pandemic cholera from Africa to South America. Based on the presence/absence of the analyzed genetic elements, five novel profiles were detected in the epidemic strains circulating in the 1990s. The most frequent profiles, F and G, were characterized by the absence of ICEs and the three GIs tested, and the presence of GI WASA-1 and the WASA variant of the VSP-II island. Our results identified unexpected variability within the 1990s epidemic, showing different rearrangements in a dynamic part of the genome not present in the prototypical *V. cholerae* O1 N16961. Moreover the 2006 strains differed from the current pandemic *V. cholerae* O1 strain. Taken together, our results highlight the role of horizontal gene transfer (HGT) in diversifying the genetic background of *V. cholerae* within a single epidemic.

#### **Keywords:** *V. cholerae* **O1, Angola, mobilome, genomic islands, epidemic variants**

#### **INTRODUCTION**

*Vibrio cholerae*, the etiological agent of cholera, a gastrointestinal infection, has been responsible for seven known pandemics with the seventh pandemic currently occurring. To date, over 200 distinct serogroups have been described, with only serogroups O1 and O139 associated with epidemic and pandemic cholera. *V. cholerae* serogroup O1 strains can be further classified in two biotypes, El Tor and Classical, based on differences in their phenotypic and genotypic traits (Kaper et al., 1995).

During the seventh pandemic several epidemic lineages of *V. cholerae* O1 El Tor have emerged. Most notable was the emergence, in 1992, of a new epidemic serogroup in India and Bangladesh, named O139 or Bengal, which initially displaced the local existing O1 El Tor strains (Faruque et al., 2003). Molecular analyses demonstrated horizontal gene transfer (HGT) as the mechanism by which *V. cholerae* O1 strains acquired the O139 surface antigen resulting in a new epidemic serogroup (Bik et al., 1995).

Subsequently, new variants of *V. cholerae* O1 with features of both the Classical and the El Tor biotypes have been repeatedly isolated in Asia and Africa and are collectively called "Atypical El Tor" strains. This group includes the Matlab types I, II, and III (Nair et al., 2002), Altered El Tor (Nair et al., 2006), Mozambique El Tor (Ansaruzzaman et al., 2004) and Hybrid El Tor strains (Safa et al., 2008).

Recent comparative genomics studies demonstrated that *V. cholerae* O1 strains circulating during the current seventh pandemic, which include prototypical El Tor strains, Atypical El Tor variants, and the O139 serogroup, belong to a single phyletic lineage named 7th pandemic clade (Chun et al., 2009). These variants are a result of HGT and slight divergence in strains from a common recent ancestor. Seventh pandemic El Tor strains are in fact characterized by a highly conserved genome background and different combinations of mobile genetic elements (MGE), referred to as the mobilome, such as Integrative Conjugative Elements (ICE), genomic islands (GIs), and prophages (Cho et al., 2010).

Seventh pandemic *V. cholerae* El Tor clones appear to have disseminated globally from a single source (the Bay of Bengal) in at least three independent waves, all of which reached Africa (Mutreja et al., 2011). Comparative genomic analyses have also suggested that the African continent served as a bridge for the spread of seventh pandemic *V. cholerae* strains from Asia to the Americas. Mutreja et al. (2011) observed that an Angolan strain, isolated in 1989, clustered at the base of the South American clade with a difference of only ten SNPs in the genomic backbone and the presence of two newly observed GIs: a novel variant of VSP-II, the VSP-II WASA (West African-South America), and a new GI named WASA-1 (Mutreja et al., 2011).

Angola is emerging from a 40 years war that ended in 2002 and resulted in significant social and economic disorders. Less than 4% of the population has access to potable water and adequate health and sanitary services. These conditions resulted in an almost immediate initiation of a cholera epidemic after *V. cholerae* isolates associated with the 7th cholera pandemic entered the African continent in the early 1970s. This region is now considered an endemic area for cholera (Lam et al., 2010). The most recent outbreak in Angola occurred from February to April 2013, and resulted in 610 cases and 8 deaths (International Society for Infectious Diseases, 2013).

Along with the most recent outbreak, two other major cholera epidemics were reported in Angola in the past 30 years. In April 1987 an epidemic started in the north-east of the country and rapidly spread to all 18 provinces, recurring in a seasonal pattern (Colombo et al., 1993). The outbreak was followed by a 10 years absence of the disease until 2006 when cholera re-emerged in Luanda and rapidly spread throughout the country. The 2006 outbreak was one of the most severe recent cholera epidemics in Africa (World Health Organization, 2007).

We previously demonstrated that the *V. cholerae* O1 strain responsible for the 2006 Angolan outbreak was an atypical O1 El Tor variant earlier detected only in Asia, and that it was clonally and genetically different from El Tor strains circulating in the 1990s in the same area (Ceccarelli et al., 2011a). Angolan strains from 2006 carried the SXT-related ICE*Vch*Ang3, associated with a more narrow multidrug resistance profile compared to the one conferred by plasmid p3iANG harbored by strains from the 1990s (Ceccarelli et al., 2006, 2011a). In light of the new insights into the phylogeny of seventh pandemic *V. cholerae* O1, we intended to reanalyze the Angolan epidemics in order to better understand the epidemiology of cholera in Angola and to confirm its role in bridging Africa and South America during waves of the seventh pandemic. In this study, we screened a collection of Angolan *V. cholerae* O1 isolates for the presence/absence of MGEs in order to study their collective mobilome and determine possible correlations with the known seventh pandemic variants (Spagnoletti et al., 2012).

#### **MATERIALS AND METHODS**

#### *V. cholerae* **STRAINS**

Twenty-one clinical *V. cholerae* strains isolated during the two separate epidemics in Angola in 1996 and 2006 were analyzed (Ceccarelli et al., 2006, 2011a). Five completely sequenced reference strains were used as positive controls (**Table 1**): *V. cholerae* O1 N16961, isolated in India in 1975 (Kaper et al., 1995; Heidelberg et al., 2000); *V. cholerae* O139 MO10, isolated in India in 1992 (Ramamurthy et al., 1993); *V. cholerae* O1 CIRS101 (Altered El Tor), isolated in Bangladesh in 2002 (Nair et al., 2006); *V. cholerae* O1 B33 (Mozambique variant), isolated in Mozambique in 2004 (Ansaruzzaman et al., 2004); *V. cholerae* O1 MJ1236 (Matlab type I), isolated in Bangladesh in 1994 (Nair et al., 2002). *V. cholerae* strains analyzed in this work are listed in **Table 1**. Bacterial strains were routinely grown on LB Agar at 37◦C for 16–18 h with appropriate antibiotic selection and were maintained at −80◦C in LB broth containing 15% (vol/vol) glycerol.

#### **PRIMERS AND PCR CONDITIONS**

Genomic DNA was prepared with a Wizard® Genomic DNA Purification Kit (Promega), according to manufacturer's instructions. PCR was performed in 50µl reaction mix containing 1 U of GoTaq DNA polymerase (Promega) and 50 ng/µL DNA. Amplification was performed in an automated thermocycler (BioRad MJ-Mini Personal Thermal Cycler). Oligonucleotides and PCR conditions for the double multiplex PCR were optimized as previously described (Spagnoletti et al., 2012).

WASA-1 and VSP-2 WASA presence was investigated with new primer pairs designed using Angolan strain *V. cholerae* O1 El Tor A5 (Mutreja et al., 2011) as reference (Accession no ERS013245): (i) wasa1F (5 -CCAAAGCAGAGAGACGCA-3) , and wasa1R (5 - GTTCTCACCTTCTTCCGCA-3 ) giving an amplicon size of 464 bp; (ii) VSP2wasaF (5 -GTGCTGTATTTGGTTTGATGGGT-3 ), and VSP2wasaR (5 -GATAGTGGTTTCGCTGAGATTGT-3 ), resulting in an amplicon size of 438 bp. Oligonucleotides were obtained from PRIMM srl (Milano, Italy).

#### **RESULTS**

Of the 21 strains analyzed during this study, 19 were isolated between 1992 and 1996 from the provinces of Benguela, Bengo, Luanda, Cuando Cubango (South-east), and the Cabinda enclave (North-west) (**Figure 1**). Two clinical strains isolated in Luanda in 2006 were also included in this study. All the strains were subjected to the double multiplex PCR screening and their amplicon profiles compared with the reference strains (**Table 1**, profiles A–J). WASA-1 analysis was not included in the profile definition but was analyzed separately to be consistent with the original double Multiplex PCRs approach (Spagnoletti et al., 2012). It is important to note that primers included in Multiplex 1 (VSPIIintF/R and VSPIIcutF/R) and used to discriminate prototypical VSP-II from other variants (Spagnoletti et al., 2012) gave no amplification for both the deleted CIRS-like VSP-II island and the WASA VSP-II. Therefore, rearrangements in the non-prototypical VSP-II islands were further tested with the VSP-2wasaF/R primer pair. The five reference strains showed unique expected profiles (Spagnoletti et al., 2012), as listed in **Table 1**.

#### **ANALYSIS OF** *V. cholerae* **O1 EPIDEMIC STRAINS ISOLATED BETWEEN 1992 AND 1996**

Six clinical strains isolated in 1992 and 1993 in Luanda showed two different mobilome profiles. Profile F (four strains from 1992 and one strain from 1993) is characterized by the presence of a

#### **Table 1 |** *V. cholerae* **strains analyzed in this study.**


*aVSP-II rearrangements were determined by either Multiplex 1 (VSPIIintF/R, VSPIIcutF/R) or VSP-II WASA (VSP2wasaF/R) PCR analysis (see Materials and Methods for further details).*

*bSibling ICEs (Ceccarelli et al., 2011b).*

*<sup>c</sup> Information about this strain is derived from Mutreja et al. (2011).*

non-prototypical VSP-II, and phages K and TLC; and profile G (strain 582 from 1992) carries the non-prototypical VSP-II and the TLC phage. All non-prototypical VSP-II islands were confirmed as carrying WASA insertion. Additionally, all six strains showed the expected amplification for WASA-1 GI.

Seven clinical *V. cholerae* O1 strains isolated in five different provinces of the country in 1994 (**Figure 1**) gave three profiles (**Table 1**). Similar to the 1992 and 1993 strains isolated from Luanda, the two most common profiles were F and G. Profile F strains were isolated from Luanda (isolate 1350), Cuando Cubango (isolate 1356), and Benguela (isolate 1383). Profile G strains were isolated from Bengo province (isolate 1349), and Dondo (Cuanza Norte) (isolates 819 and 1354). A previously unobserved profile H (non-prototypical VSP-II and phage K) was observed in strain 1382 isolated in Cuando Cubango. All non-prototypical VSP-II islands were confirmed as carrying the WASA insertion. *V. cholerae* strains 1349 and 1354 gave a positive amplification for WASA-1 GI while the remaining five strains gave negative amplification for this cluster.

Five clinical *V. cholerae* O1 strains from 1995 and one from 1996 were isolated in the provinces of Luanda and Cabinda (**Figure 1**). Strains isolated in four different municipalities of the Luanda province in 1995 showed two profiles: profile F (strains 1357 and 1359), and profile I (strains 1358 and 1360) in which the sole non-prototypical variant of VSP-II was amplified. The two strains isolated in the Cabinda enclave showed profiles F and I in 1995 (strain 1361) and 1996 (strain 908), respectively. All non-prototypical VSP-II islands were confirmed as carrying the WASA insertion. *V. cholerae* strains 1357, 1358, and 1359 from Luanda gave a positive amplification for the WASA-1 GI while the remaining 3 strains (isolates 1360, 1361, and 908 from Luanda and the Cabinda enclave, respectively) were negative for this cluster.

#### **ANALYSIS OF** *V. cholerae* **O1 EPIDEMIC STRAINS ISOLATED IN 2006**

As anticipated by our previous study (Ceccarelli et al., 2011a), the two *V. cholerae* O1 strains isolated in Luanda during the 2006 epidemic, shared a common and previously unobserved mobilome profile J (**Table 1**). These strains harbored the prototypical intact variant of VSP-II, an ICE element (ICE*Vch*Ang3, sibling of ICE*Vch*Ind5), and TLC. Both were negative for the presence of WASA-1 and VSP-II WASA genetic clusters.

#### **DISCUSSION**

This retrospective study is based on a collection of *Vibrio cholerae* O1 clinical strains, isolated from 1992 to 2006 from different provinces in Angola (**Figure 1** and **Table 1**). Despite encouragement from local national health authorities, it was not possible to isolate strains over a broader temporal and geographic scale due to security and logistic constraints caused by socio-economic conditions. Thus, we are aware of the potential spatial and temporal biases inherent in this set of strains. Half of our isolates were recovered from cases in Luanda, and all but two isolates were recovered between 1992 and 1996. Nevertheless, we believe the strains presented in this study represent an important source of data depicting the presence of a variable population of *V. cholerae* O1 strains in the country.

The epidemiology of *V. cholerae* O1 in Africa can be explained by multiple introductions of cholera from endemic regions of Asia, with the first introduction following the early dissemination of seventh pandemic in the 1970s (Kaper et al., 1995). In our previous analysis we hypothesized that *V. cholerae* O1 strains responsible for the first epidemic (1987–1996) in Angola are progeny of the prototypical strain represented by *V. cholerae* O1 El Tor N16961, with the only difference being the presence of the p3iANG plasmid (Ceccarelli et al., 2011a). However, our current study revealed that all *V. cholerae* O1 strains from 1987 to 1996 harbored the WASA variant of the VSP-II GI, and 10 of these carried also the WASA-1 GI. Therefore, these strains differed in MGE content from *V. cholerae* N16961 (profile A). An even greater level of diversity was identified in these strains when tested for the presence of the K and TLC phages. These analyses resulted in the assignment of four distinct profiles (F–I) among strains isolated between 1987 and 1996 in Angola.

The two strains isolated in 2006 showed a mobilome profile highly similar to that of the altered El Tor *V. cholerae* O1 strain CIRS101, including the presence of ICE*Vch*Ang3, an ICE element of the SXT/R391 family sibling with ICE*Vch*Ind5, the most widespread ICE present in *V. cholerae* O1 CIRS101-like strains isolated in the Indian Subcontinent and worldwide (Ceccarelli et al., 2011a,b,c). Again, significant differences between the 2006 Angolan strains and *V. cholerae* CIRS101 were identified in the VSP-II island. The 2006 Angolan strains encoded the prototypical VSP-II, while many currently isolated pandemic *V. cholerae* O1 CIRS101-like strains isolated worldwide after 2001 encode a deletion in VSP-II (Taviani et al., 2010). These results suggest that epidemic strains circulating in Angola in 2006 harbored a genetic background with features similar to strains responsible for the first (profile A) and the third (profile B) pandemic waves.

All strains in this analysis were negative for GIs-12, −14, and −15, which are specific GIs of atypical El Tor strains similar to *V. cholerae* O1 MJ1236 and B33 (Grim et al., 2010). The latter was isolated in Beira, Mozambique, in 2004, but it was recently demonstrated to be present in Austral Africa since the 1990s (Spagnoletti et al., 2012). By testing these three GIs as molecular markers, our analysis suggests that the Angolan epidemics were not directly influenced by contemporary eastern Africa epidemics, which are thought to have been initiated by the second pandemic wave (Mutreja et al., 2011).

Our analysis revealed the presence of the WASA-1 GI in all *V. cholerae* O1 strains from 1992, whereas the same element was absent in recently recovered isolates. The VSP-II WASA variant was instead present in all *V. cholerae* O1 strains. The presence of the WASA clusters in Angolan *V. cholerae* O1 strains confirmed the Mutreja's hypothesis, based only on one isolate from 1989, that the transmission of the seventh pandemic to South America may have occurred via the African continent and, specifically, through Angola (Mutreja et al., 2011). Furthermore, it suggests that transmission of a strain which encoded both VSP-II-WASA and WASA-1 is likely to have occurred during the 1980s, as previously reported (Lam et al., 2010; Mutreja et al., 2011).

In conclusion, the history of the Angolan epidemic adds new and valuable information to evolutionary history of *V. cholerae* O1 El Tor within the seventh pandemic. At the time-scale presented here and elsewhere, this evolution occurs mainly via lateral gene transfer events driven by environmental factors. Interestingly, none of the Angolan strains showed a mobilome profile similar to those of the representative 7th pandemic strains. The variability within these strains was greater than expected, as strains isolated within the same epidemic were shown to harbor different constellations of mobile elements. A similar variability was previously observed in a collection of strains isolated in Mozambique in 1997–1998 (Spagnoletti et al., 2012), thus reflecting frequent HGT dynamics similar to the Angolan epidemic.

#### **ACKNOWLEDGMENTS**

This work was funded by Ministry of Foreign Affair—DGCS, Italy, Ministero dell'Istruzione, dell'Università e della Ricerca— Italy (PRIN). Matteo Spagnoletti is supported by a fellowship from Istituto Pasteur—Cenci Bolognetti Foundation, Italy. The English revision was carried out by Bradd J. Haley.

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 04 October 2013; paper pending published: 24 October 2013; accepted: 07 November 2013; published online: 28 November 2013.*

*Citation: Valia R, Taviani E, Spagnoletti M, Ceccarelli D, Cappuccinelli P and Colombo MM (2013) Vibrio cholerae O1 epidemic variants in Angola: a retrospective study between 1992 and 2006. Front. Microbiol. 4:354. doi: 10.3389/fmicb.2013.00354 This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Valia, Taviani, Spagnoletti, Ceccarelli, Cappuccinelli and Colombo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## *Photobacterium damselae* subsp. *damselae*, a bacterium pathogenic for marine animals and humans

### *Amable J. Rivas, Manuel L. Lemos and Carlos R. Osorio\**

*Institute of Aquaculture, University of Santiago de Compostela, Santiago de Compostela, Spain*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Devaki Bhaya, Carnegie Institution for Science, USA Ming-Lun Chen, National Penghu University of Science and Technology, Taiwan*

#### *\*Correspondence:*

*Carlos R. Osorio, Institute of Aquaculture, Campus Vida, University of Santiago de Compostela, 15782 Santiago de Compostela, Galicia, Spain e-mail: cr.osorio@usc.es*

*Photobacterium damselae* subsp. *damselae* (formerly *Vibrio damsela*) is a pathogen of a variety of marine animals including fish, crustaceans, molluscs, and cetaceans. In humans, it can cause opportunistic infections that may evolve into necrotizing fasciitis with fatal outcome. Although the genetic basis of virulence in this bacterium is not completely elucidated, recent findings demonstrate that the phospholipase-D Dly (damselysin) and the pore-forming toxins HlyApl and HlyAch play a main role in virulence for homeotherms and poikilotherms. The acquisition of the virulence plasmid pPHDD1 that encodes Dly and HlyApl has likely constituted a main driving force in the evolution of a highly hemolytic lineage within the subspecies. Interestingly, strains that naturally lack pPHDD1 show a strong pathogenic potential for a variety of fish species, indicating the existence of yet uncharacterized virulence factors. Future and deep analysis of the complete genome sequence of *Photobacterium damselae* subsp. *damselae* will surely provide a clearer picture of the virulence factors employed by this bacterium to cause disease in such a varied range of hosts.

**Keywords:** *Photobacterium damselae***, hemolysin, damselysin,** *hlyA***, pore-forming toxin**

"fmicb-04-00283" — 2013/9/23 — 21:15 — page 1 — #1

#### *PHOTOBACTERIUM DAMSELAE* **SUBSP.** *DAMSELAE*

*Photobacterium damselae* subsp. *damselae* is a marine bacterium of the family *Vibrionaceae* that causes infections in a variety of marine animals and also in humans. A bit of historic perspective is necessary in order to understand its current taxonomic placement as well as the changes in its nomenclature during the past decades. In 1971 an "unnamed marine *Vibrio*" was isolated as the causative agent of a human infectious case (Morris et al., 1982)*.* Later, this same organism was isolated from skin ulcers of damselfish (*Chromis punctipinnis*) and the name *Vibrio damsela* was first coined (Love et al., 1981). Further genetic and phenotypic studies indicated that the strains of *V. damsela* were closely related to species of the genus *Photobacterium*, and the name *Photobacterium damsela* was proposed (Smith et al., 1991). In 1995, DNA–DNA hybridization data and 16S rRNA sequence analysis demonstrated that *Photobacterium damsela* was closely related to a fish pathogen formerly named *Pasteurella piscicida*, the causative agent of pasteurellosis in fish. Hence, these two organisms were assigned to the same species epithet, *Photobacterium damselae*, with category of subspecies (Gauthier et al., 1995), *Photobacterium damselae*subsp. *damselae* and *Photobacterium damselae*subsp. *piscicida* respectively. Despite their similarity at the 16S gene sequence and the high percentage of DNA–DNA relatedness between them, these two subspecies are clearly distinguished by several phenotypical traits (Fouz et al., 1992; Magariños et al., 1992; Thyssen et al., 1998; Botella et al., 2002). Differential phenotypical tests of interest for subspecies discrimination that are positive only for subsp. *damselae* include motility, nitrate reduction and hemolysis on sheep blood agar. Of special relevance is the ability of most subsp. *damselae* strains to grow at 37◦C (a temperature inhibitory for subsp. *piscicida*), a trait that allows *Photobacterium damselae* subsp. *damselae* to potentially colonize and establish an infection in a homeotherm animal.

#### **A PATHOGEN OF MARINE ANIMALS**

*Photobacterium damselae* subsp. *damselae* is an autochthonous member of aquatic ecosystems. Strains of this pathogen have been isolated from sea and estuarine waters, from seaweeds, from apparently uninfected marine animals (Buck et al., 2006; Serracca et al., 2011) and from seafood (Lozano-León et al., 2003; Chiu et al., 2013), and it is considered a common member of the natural microbiota of healthy carcharhinid sharks (Grimes et al., 1985).

In addition, *Photobacterium damselae* subsp. *damselae* is considered a primary pathogen of several species of wild fish (damselfish, catfish, shark, stingray, etc.), as well as of fish species of economical importance in aquaculture, causing wound infections and hemorrhagic septicemia. Cultivated species reported to be affected by this pathogen include turbot (*Psetta maxima*; Fouz et al., 1992), rainbow trout (*Oncorhynchus mykiss*; Pedersen et al., 2009), ovate pompano (*Trachinotus ovatus*; Zhao et al., 2009), eel (*Anguilla reinhardtii*; Ketterer and Eaves, 1992), sea bream (*Sparus aurata*; Vera et al., 1991), sea bass (*Dicentrarchus labrax*), yellowtail (*Seriola quinqueradiata*), redbanded seabream (*Pagrus auriga*), white seabream (*Diplodus sargus*), and meagre (*Argyrosomus regius*; Labella et al., 2006, 2010a,b), among others. The recent first reports on isolation of this pathogen from diseased marine fish of new cultured species, suggest that *Photobacterium damselae* subsp. *damselae* can be considered as an emerging pathogen in marine aquaculture (Labella et al., 2011).

Moreover, *Photobacterium damselae* subsp. *damselae* has been isolated as a pathogen of brown shark (*Carcharhinus plumbeus*; Grimes et al., 1984), of reptiles as the leatherback sea turtle (*Dermochelys coriacea*; Obendorf et al., 1987), molluscs (*Octopus joubini*; Hanlon et al., 1984), crustaceans (Song et al., 1993; Vaseeharan et al., 2007), dolphins (*Tursiops truncatus* and *Delphinus delphis*; Fujioka et al., 1988; Buck et al., 1991) and Bryde's whale (*Balaenoptera edeni*; Buck et al., 1991).

Virulent isolates are capable of survival in seawater microcosms at 14–22◦C as culturable bacteria for long periods of time, maintaining their infectivity for fish (Fouz et al., 1998). Similarly, this pathogen can infect new fish hosts through water, and the spread of the disease depends largely on water temperature and salinity (Fouz et al., 2000). Typical signs of the disease in infected fish include hemorrhaged areas on the body surface and ulcerative lesions. In damselfish, ulcers typically occur in the region of the pectoral fin and caudal peduncle and may reach 5–20 mm in diameter (Love et al., 1981), while in turbot the most remarkable symptoms are extensive hemorrhages in eyes, mouth, and jaws (Fouz et al., 1995).

Experimental inoculation of *Photobacterium damselae* subsp. *damselae* extracellular products (ECPs) in a redbanded seabream model was reported to cause lethargy, increase in the respiratory frequency, mucus production, presence of ascitic liquid, hemorrhagic and enlarged liver, and hemorrhages in the abdominal cavity (Labella et al., 2010b). A histological analysis of internal organs in experimentally infected turbot indicated that the ECPs and cells of virulent strains cause similar tissue damage (Fouz et al., 1995). Structural changes included destruction and necrosis of cells, as well as accumulation of blood cells in interstitial tissue.

#### *Photobacterium damselae* **subsp***. damselae* **AS A HUMAN PATHOGEN**

Most of the reported infections caused by *Photobacterium damselae* subsp. *damselae* in humans have their primary origin in wounds exposed to salt or brackish water, inflicted during fish and tools handling (Morris et al., 1982; Dryden et al., 1989; Yuen et al., 1993; Shin et al., 1996; Tang and Wong, 1999; Barber and Swygert, 2000; Goodell et al., 2004; Aigbivhalu and Maraqa, 2009). Unusual cases of infection after ingestion of raw seafood (Kim et al., 2009) and through the urinary tract by exposure to sea water (Alvarez et al., 2006) were also reported. The majority of the cases occurred in coastal areas of the United States of America, Australia, and Japan.

*Photobacterium damselae* subsp. *damselae* can cause an extreme variant of a highly severe necrotizing fasciitis, and antibiotic administration proved unable to control the progression of fatal infections in some cases (Clarridge and Zighelboimdaum, 1985; Fraser et al., 1997; Yamane et al., 2004). It is interesting to note that some authors recommend to surgically debride and amputate without hesitation at a very early point of the infection by *Photobacterium damselae* subsp. *damselae*, to save the lives of patients (Goodell et al., 2004). Some patients infected by *Photobacterium damselae* subsp. *damselae* developed multiple organ failure within a few hours from the onset of initial symptoms, despite intensive chemotherapy and surgical treatments. As an example, in a fatal case reported in 1984 in which a patient injured his hand while handling a catfish, bulle formation occurred on the hand and a marked edema extended through the forearm in less than 24 h (Clarridge and Zighelboimdaum, 1985), and although the affected area was extensively debrided the patient died after a series of complications. The bacterium was recovered in high numbers from the tissue sample but only in very small numbers from the bulle fluid. In another fatal case reported in a patient injured while handling fish, *Photobacterium damselae* subsp. *damselae* was isolated in pure culture from wound specimens but failed to be isolated from blood samples (Fraser et al., 1997). These observations prompted these authors to suggest that a virulence factor or systemic toxin released by this bacterium contributed to the tissue damage and to the fatal outcome, rather than the septicemia itself. However, in other clinical cases this pathogen was recovered from blood (Perez-Tirse et al., 1993; Shin et al., 1996;Yamane et al., 2004).

Necrotizing fasciitis due to *Photobacterium damselae* subsp. *damselae* demonstrates more serious complications and a higher mortality rate than that caused by *Vibrio vulnificus*. While *V. vulnificus* usually affects persons with underlying diseases (as chronic liver disease and diabetes mellitus), necrotizing fasciitis by *Photobacterium damselae* subsp. *damselae* sometimes occurs in healthy hosts (Morris et al., 1982; Perez-Tirse et al., 1993; Yuen et al., 1993).

#### **VIRULENCE FACTORS**

#### *Iron uptake systems*

Early studies reported that *Photobacterium damselae*subsp. *damselae* can utilize heme, hemoglobin and ferric ammonium citrate as sole iron sources *in vitro* (Fouz et al., 1994). The complete sequence of 10 genes encoding a system for the utilization of heme as iron source was described in a human isolate of *Photobacterium damselae* subsp. *damselae*, and cloning of the complete system into *E. coli* conferred to this species the ability to use hemin and hemoglobin as iron sources (Rio et al., 2005). The presence of the heme receptor gene *hutA* was demonstrated in subsp. *damselae* isolates from fish and humans, and the identity at the DNA sequence level between the heme uptake clusters of subsp. *damselae* and subsp. *piscicida* strains was 97% (Rio et al., 2005). Although no functional studies were conducted with the heme uptake genes of subsp. *damselae*, it was recently demonstrated that this cluster is essential for heme utilization in subsp. *piscicida*, and two genes of a hemin ABC-transporter proved to be expressed during the infective process in a fish model (Osorio et al., 2010). Actually, an increase in the susceptibility of both fish and mice to infection by virulent *Photobacterium damselae* subsp. *damselae* strains in virulence assays conducted with iron-overloaded animals had been demonstrated in former studies (Fouz et al., 1994). It is also known that this bacterium produces a hydroxamate-type siderophore, and the synthesis of several high-molecular weight outer membrane proteins induced under iron limitation conditions was reported (Fouz et al., 1997). Although the precise chemical structure of the siderophore(s) is so far unknown, recent unpublished work from our laboratory demonstrated that vibrioferrin is being produced by some strains.

#### *Cytotoxins with hemolytic activity*

"fmicb-04-00283" — 2013/9/23 — 21:15 — page 2 — #2

Pioneering studies (Kreger, 1984) reported the existence of a correlation between the ability of *Photobacterium damselae* subsp. *damselae* isolates to cause disease in mice and the production of large amounts of a heat-labile cytolytic toxin *in vitro*. Later, the same authors purified a toxin that exhibited strong hemolytic activity against erythrocytes of a variety of animal species (Kothary and Kreger, 1985). In subsequent studies, this toxin named damselysin (Dly) was defined as a phospholipase-D active against sphingomyelin, with hemolytic activity (Kreger et al., 1987), and its gene (*dly*) was cloned and sequenced (Cutter and Kreger, 1990). Dly was thus considered to be the main virulence factor of *Photobacterium damselae* subsp. *damselae* for mice. Further studies reported the existence of hemolytic strains of *Photobacterium damselae* subsp. *damselae* that tested negative for *dly* gene, which suggested that Dly was not the only hemolysin in the subspecies (Osorio et al., 2000). It was also demonstrated by several authors that presence of *dly* is not a prerequisite for the hemolytic activity and for the pathogenicity for mice or fish, since *dly* negative strains bear virulence potential for animals and also toxicity for homeotherm and poikilotherm cell lines (Osorio et al., 2000; Labella et al., 2010b).

The genomic context of *dly* gene remained uncharacterized for decades, and it was initially proposed that it could be carried on a mobile or unstable DNA element (Cutter and Kreger, 1990). Recently, the authors' laboratory identified and characterized a 150 kb plasmid, which was dubbed pPHDD1, that contains the genes for Dly as well as for a HlyA toxin of the pore-forming toxin family (Rivas et al., 2011). Only a fraction of *Photobacterium damselae* subsp. *damselae* strains harbor pPHDD1, and these strains exhibit a much wider hemolytic halo on sheep blood agar plates than the plasmidless strains (**Figure 1**). Interestingly, pPHDD1 occurs in both fish and human isolates and it is not restricted to a unique animal host species (Rivas et al., 2011). In addition to being necessary to cause strong hemolytic haloes on blood agar plates, the two pPHDD1-encoded hemolysins play a crucial role in virulence for fish and mice in strains that naturally harbor the plasmid. Hence, mutation of both *dly* and *hlyA* genes in a pPHDD1-harboring strain renders the strain non-virulent for fish, and only slightly virulent for mice (**Table 1**), and the hemolytic phenotype on sheep blood agar of a *dly* and *hlyA* double mutant resembles that of naturally plasmidless strains (**Figure 1**; Rivas et al., 2011, 2013).

The hemolytic activity exhibited by plasmidless strains was recently demonstrated to be caused by a chromosome-encoded *hlyA* gene, which was dubbed *hlyAch* in order to differentiate it from the plasmid *hlyA* gene (hereafter *hlyApl*; Rivas et al., 2013). It was found that all the hemolytic *Photobacterium damselae* subsp. *damselae* strains harbor *hlyAch* gene, which is the only hemolytic determinant in plasmidless strains. Thus, pPHDD1-harboring isolates produce three different hemolysins. In hemolytic assays carried out with bacterial ECPs and with sheep erythrocytes, it was demonstrated that Dly acts in a synergistic manner with HlyApl and HlyAch, whereas the effect between HlyApl and HlyAch showed to be additive but not synergistic (Rivas et al., 2013).

Although each of the three hemolysins individually proved to be sufficient to cause death in mice, each one contributes to virulence in a different degree. The contribution of HlyAchto virulence for mice is the lowest among the three toxins. Altogether, albeit

**Table 1 | Role of the three** *Photobacterium damselae* **subsp.** *damselae* **hemolysins in virulence for mice and fish (turbot).**

hemolytic phenotype, similar to that of a naturally plasmidless strain

**(C)**, that only produces HlyAch.


*Parental and mutant strains were inoculated in groups of 15 animals, at doses of 2.1* <sup>×</sup> *<sup>10</sup><sup>6</sup> bacterial cells per mouse and 2.1* <sup>×</sup> *<sup>10</sup><sup>4</sup> bacterial cells per fish. The number of dead animals out of the total number of inoculated animals (15) is indicated. Asterisks denote that significant differences exist between a given mutant and the parental strain, using U-test (\*P* < *0.05).*

"fmicb-04-00283" — 2013/9/23 — 21:15 — page 3 — #3

the highest values of mortality for mice are achieved only when the three hemolysins are being produced, Dly and HlyApl demonstrated to be main contributors in the virulence of *Photobacterium damselae* subsp. *damselae* for mice (Rivas et al., 2013; **Table 1**).

Interestingly, the contribution of each hemolysin to virulence was found to vary depending on whether the host animal tested was mouse or turbot. When virulence experiments were conducted with turbot, it was found that among all the hemolysin gene mutants only the Dly-producing strains caused death in fish. This finding demonstrated that any of the two HlyA alone does not cause death in turbot, but rather one of the two HlyA needs the presence of either Dly or the other HlyA to cause death in fish. The production of Dly in combination with any of the two HlyA caused an increase in the number of dead fish with respect to the production of Dly alone, and this increase was found to be particularly evident when Dly was combined with HlyAch. This clearly suggests that, unlike what is observed in mice, the contribution of hemolysins to virulence for fish is not so much based on the individual effects of each hemolysin but rather on the combined (synergistic) effects between Dly and HlyA (Rivas et al., 2013; **Table 1**). These findings also state the importance of pPHDD1 plasmid in virulence for fish, since Dly is necessary for the synergistic effect.

#### *Other exoenzymes and exotoxins: toxicity of the extracellular products*

Early studies detected several enzymatic activities in the ECPs of *Photobacterium damselae* subsp. *damselae*, which included phospholipase and hemolysin activities (Fouz et al., 1993). More recent data confirmed that the ECPs of *Photobacterium damselae* subsp. *damselae* are strongly lethal for fish, and enzymatic activities such as amylase, lipase, phospholipase, alkaline phosphatase, esterase–lipase, acid phosphatase, and β-glucosaminidase were evidenced (Labella et al., 2010b). Moreover, treatment at 100◦C for 10 min of the ECPs abolished the ability to cause death in fish, suggesting that toxicity was not due to the thermorresistant lipopolysaccharide content. *Photobacterium damselae* subsp. *damselae* ECPs also displayed cytotoxic activity for different fish and mammalian cell lines (Wang et al., 1998; Labella et al., 2010b). Different studies found a correlation between virulence of the strain and toxicity of the ECPs, with toxicity being limited to ECPs from strains that were also virulent for fish (Fouz et al., 1995; Labella et al., 2010b). Of maximum interest is the observation that strains lacking pPHDD1 plasmid and thus being negative for *dly* and *hlyApl* genes, are virulent for fish and their ECPs are cytotoxic for cell lines. In addition, a comprehensive study reported that none of the enzymatic activities detected in the *Photobacterium damselae* subsp. *damselae* ECPs could be related with the degree of toxicity either *in vivo* or *in vitro* (Labella et al., 2010b). Most *Photobacterium damselae* subsp. *damselae*

#### **REFERENCES**

Aigbivhalu, L., and Maraqa, N. (2009). *Photobacterium damsela* wound infection in a 14-yearold surfer. *South. Med. J.* 102, 425–426. doi: 10.1097/SMJ.0b013e 31819b9491

Alvarez, J. R., Lamba, S., Dyer, K. Y., and Apuzzio, J. J. (2006). An unusual case of urinary tract infection in a pregnant woman with *Photobacterium damsela*. *Infect. Dis. Obstet. Gynecol.* 2006, 80682. doi: 10.1155/IDOG/2006/80682

strains test negative for protease activities as caseinase and gelatinase (Fouz et al., 1992; Labella et al., 2010b). This suggests that other, yet uncharacterized molecules produced by *Photobacterium damselae* subsp. *damselae* cells play a role in toxicity for animals and for cell lines. In this regard, previous studies detected the existence of an acetylcholinesterase activity (ictiotoxin) with neurotoxic activity in several species of *Vibrionaceae*, including *Photobacterium damselae* subsp. *damselae* (Perez et al., 1998), although the genetic basis for this neurotoxic activity remains unknown.

#### **FUTURE PERSPECTIVES**

An interesting observation that remains to be explained at the genetic level is the finding that plasmidless *Photobacterium damselae* subsp. *damselae* strains are virulent for fish and toxic for homeotherm and poikilotherm cell lines (Fouz et al., 1993; Osorio et al., 2000; Labella et al., 2010b, 2011). Since plasmidless strains lack *dly* and *hlyApl* genes, and since *dly hlyApl* double mutants are significantly reduced in its virulence for both mice and fish (Rivas et al., 2013), it is evident that plasmidless strains encode virulence factors that either are not encoded by pPHDD1-harboring strains or their expression is repressed in presence of pPHDD1-encoded genes.

The recent completion of the genome sequence of the type strain (ATCC 33539) of this subspecies (deposited in GenBank database in several separate contigs, under accession number ADBS00000000), allows an *in silico* analysis to search for candidate genes encoding potential toxins and other virulence factors. The type strain harbors pPHDD1 plasmid, and preliminary analyses also indicated the presence of genes encoding a type III hemolysin (open reading frame number: VDA003208), and a putative murine toxin (VDA000322) among others. The existence of yet uncharacterized plasmids is also evidenced in the complete genome of ATCC 33539. Studies to functionally characterize novel plasmid content and candidate virulence genes of *Photobacterium damselae* subsp. *damselae* strains are currently under way. It is expected that a deep analysis of the complete genome sequence of *Photobacterium damselae*subsp. *damselae*strains with different isolation origins and virulence properties will provide a clearer picture of the virulence factors employed by this bacterium to cause disease in such a varied range of hosts.

#### **ACKNOWLEDGMENTS**

The work in the authors' laboratory is supported by grant EM2012/043 from Xunta de Galicia, Spain; and by grants AGL2012-39274-C02-01, from the Ministry of Economy and Competitiveness (MINECO) of Spain and CSD2007-00002 (Consolider Aquagenomics) from the Ministry of Science and Innovation (MICINN) of Spain, both (cofunded by the FEDER Programe from the European Union.

Barber, G. R., and Swygert, J. S. (2000). Necrotizing fasciitis due to *Photobacterium damsela* in a man lashed by a stingray. *N. Engl. J. Med*. 342, 824. doi: 10.1056/NEJM200003163 421118

"fmicb-04-00283" — 2013/9/23 — 21:15 — page 4 — #4

Botella, S., Pujalte, M. J., Macian, M. C., Ferrus, M. A., Hernandez, J., and Garay, E. (2002). Amplified fragment length polymorphism (AFLP) and biochemical typing of *Photobacterium damselae* subsp. *damselae. J. Appl. Microbiol.* 93, 681–688. doi: 10.1046/j.1365-2672. 2002.01748.x


121, 181–188. doi: 10.1111/j.1574- 6968.1994.tb07097.x


*reinhardtii*) due to *Photobacterium (Vibrio) damsela*. *Aust. Vet. J.* 69, 203–204. doi: 10.1111/j.1751- 0813.1992.tb07528.x


"fmicb-04-00283" — 2013/9/23 — 21:15 — page 5 — #5

in bivalve molluscs from Northwest Spain. *Bull. Eur. Assoc. Fish Pathol.* 23, 40–44.


hemolysins damselysin and HlyA are encoded within a new virulence plasmid*. Infect. Immun.* 79, 4617–4627. doi: 10.1128/IAI.05436-11


nov. with an emended description. *Int. J. Syst. Bacteriol.* 41, 529–534. doi: 10.1099/00207713- 41-4-529


"fmicb-04-00283" — 2013/9/23 — 21:15 — page 6 — #6

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 23 July 2013; paper pending published: 12 August 2013; accepted: 02 September 2013; published online: 25 September 2013.*

*Citation: Rivas AJ, Lemos ML and Osorio CR (2013) Photobacterium damselae subsp. damselae, a bacterium pathogenic for marine animals and humans. Front. Microbiol. 4:283. doi: 10.3389/fmicb.2013.00283*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Rivas, Lemos and Osorio. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## The function of integron-associated gene cassettes in *Vibrio* species: the tip of the iceberg

#### *Rita A. Rapa1,2 and Maurizio Labbate 1,2 \**

*<sup>1</sup> ithree Institute, University of Technology, Sydney, NSW, Australia*

*<sup>2</sup> Department of Medical and Molecular Biosciences, University of Technology, Sydney, NSW, Australia*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Mauro M. Colombo, Biotechnology Center, E. Mondlane University, Mozambique Genevieve Garriss, Institut Pasteur, France*

#### *\*Correspondence:*

*Maurizio Labbate, ithree Institute, University of Technology, Sydney, NSW, Australia; Department of Medical and Molecular Biosciences, University of Technology, PO Box 123, Broadway 2007, Sydney, NSW, Australia*

*e-mail: maurizio.labbate@uts.edu.au*

The integron is a genetic element that incorporates mobile genes termed gene cassettes into a reserved genetic site via site-specific recombination. It is best known for its role in antibiotic resistance with one type of integron, the class 1 integron, a major player in the dissemination of antibiotic resistance genes across Gram negative pathogens and commensals. However, integrons are ancient structures with over 100 classes (including class 1) present in bacteria from the broader environment. While, the class 1 integron is only one example of an integron being mobilized into the clinical environment, it is by far the most successful. Unlike clinical class 1 integrons which are largely found on plasmids, other integron classes are found on the chromosomes of bacteria and carry diverse gene cassettes indicating a non-antibiotic resistance role(s). However, there is very limited knowledge on what these alternative roles are. This is particularly relevant to *Vibrio* species where gene cassettes make up approximately 1–3% of their entire genome. In this review, we discuss how emphasis on class 1 integron research has resulted in a limited understanding by the wider research community on the role of integrons in the broader environment. This has the capacity to be counterproductive in solving or improving the antibiotic resistance problem into the future. Furthermore, there is still a significant lack of knowledge on how gene cassettes in *Vibrio* species drive adaptation and evolution. From research in *Vibrio rotiferianus* DAT722, new insight into how gene cassettes affect cellular physiology offers new alternative roles for the gene cassette resource. At least a subset of gene cassettes are involved in host surface polysaccharide modification suggesting that gene cassettes may be important in processes such as bacteriophage resistance, adhesion/biofilm formation, protection from grazers and bacterial aggregation.

**Keywords: integron, gene cassette,***Vibrio***, mobile DNA,mobile genetic elements, mobile genes, lateral gene transfer**

#### **INTRODUCTION**

Members of the *Vibrio* genus are ubiquitous in marine environments and show a wide range of niche specialization (Thompson et al., 2004). The capability of vibrios to occupy diverse niches is a testament to their ability to adapt and evolve. An important driver of this in vibrios is lateral gene transfer (LGT). LGT is the mechanism of DNA transfer from one bacterial cell to another without the requirement for cell division. It is followed by subsequent incorporation of the DNA into the recipients' genome such that DNA can be stably inherited, a process assisted by mechanisms such as homologous recombination or via a range of mobile genetic elements (MGEs) such as transposons and genomic islands (Stokes and Gillings, 2011). This mini review will focus on one important MGE called the integron, an element commonly known for its role in antibiotic resistance. The focus on the integron and its role in antibiotic resistance has driven a lack of understanding (and perhaps lack of interest) for the role this element plays in the broader environment. In contrast, we argue that understanding integron contribution to the antibiotic resistance problem requires an understanding of the role of integrons in their broad evolutionary context. Since integrons are present in almost all *Vibrio* species and comprise a significant proportion of their genome, they are excellent candidates for studying alternative roles of integrons outside of the clinical environment. Using recent work from *Vibrio rotiferianus* DAT722, we discuss possible environmental roles for this MGE.

#### **WHAT ARE INTEGRONS?**

"fmicb-04-00385" — 2013/12/5 — 18:15 — page 1 — #1

An integron is a site-specific recombination system capable of integrating and expressing open reading frames (ORFs) contained in modular structures called gene cassettes (**Figure 1**; Mazel, 2006; Labbate et al., 2009). The integron is defined by three components, an integrase gene (*intI*) that encodes a site-specific recombinase, an attachment site (*attI*), and a promoter (P*c*). The mobile units that insert into integrons are gene cassettes. Gene cassettes commonly consist of a single promoterless ORF and an IntI-identifiable recombination site called *attC.* The integration of gene cassettes is facilitated by an integrase-mediated recombination reaction between *attI* × *attC* and less commonly *attC* × *attC*. Multiple insertion events produce a contiguous cassette array with cassettes downstream of the P*<sup>c</sup>* promoter being co-transcribed. Induction of *intI* can also cause excision and rearrangement of a gene cassette(s) into a different position.

**"fmicb-04-00385" — 2013/12/5 — 18:15 — page 2 — #2**

Integrons are a diverse family of elements and are catalogued into classes based on the nucleotide sequence of the integrase gene. Currently, there are over 100 different integron classes, *most present on bacterial chromosomes* and found in approximately 10% of sequenced bacterial genomes (Boucher et al., 2007). The class 1 integron was the first described integron (Stokes and Hall, 1989), found linked to antibiotic resistance genes in resistance plasmids from Gram negative pathogens. This is because class 1 integrons in clinical contexts largely contain antibiotic resistance gene cassettes with approximately 130 so far described (Partridge et al., 2008). The accumulation of multiple resistance gene cassettes (up to about six) has associated these elements with multi-drug resistance (Leverstein-van Hall et al., 2002, 2003). Unlike clinical class 1 integrons which carry an identical integrase gene sequence, diverse class 1 integrons are also found in the chromosomes of environmental *Betaproteobacteria,* containing divergent integrase sequences and functionally diverse gene cassettes. This indicates that *Betaproteobacteria* were the original source of the clinical class 1 integron and that its initial capture by a transposon disseminated it across diverse Gram negative pathogens and human commensals (Gillings et al., 2008).

Although the class 1 integron is by far the most abundant integron in clinical contexts, others have been described (approximately five). An environmental source for all these clinical integrons strongly suggests that integrons have a much broader role in adaptation than conferring resistance to antibiotics in clinical environments (Rowe-Magnus et al., 2001; Boucher et al., 2007; Gillings et al., 2008). Phylogeny shows integrons to be ancient structures (Rowe-Magnus et al., 2001; Boucher et al., 2007) therefore, the gene cassette pool has been contributing to adaptation and evolution of bacteria for several hundred million years and not just in the last 70 years during the antibiotic revolution. This point is sometimes not well understood by researchers studying clinically derived integrons.

#### **RESEARCHING THE BROADER ROLE OF INTEGRONS IN A FIELD FOCUSED ON THEIR CONTRIBUTION TO THE DISSEMINATION OF ANTIBIOTIC RESISTANCE**

Due to the ongoing issue of bacterial antibiotic resistance, research is still heavily focused on the role of class 1 integrons. A PubMed search with the term "integron" retrieves in excess of 2200 publications. A search with the following terms "integron and (antibiotic or resistance or class 1)" retrieves 1847 publications showing that 83% focus on antibiotic resistance and/or class 1 integrons. Antibiotic resistance is a significant issue and we are certainly not suggesting that the emphasis on the role of integrons in this area is not justified however, we believe that this is impacting adversely on our understanding of these elements including in relation to the antibiotic resistance problem. Firstly, the focus on antibiotic resistance is skewing understanding for the general role of integrons in the wider research community. Given the hundreds of integron classes that exist, much of our knowledge is based on a single class (i.e., the class 1 integron). This overshadows the likely important role that integrons play in the broader environment and sometimes results in an erroneous dogma that knowledge of class 1 integrons can be extrapolated to all integron classes. The overshadowing of research in integrons outside the clinical setting is easily observed via a cursory examination of the literature over the last 3 years. Although we have known for nearly a decade that there are over 100 classes of integrons (most from non-clinical bacteria including in *Vibrio* species), publications still claim the existence of "4" (Madiyarov et al., 2010; Rezaee et al., 2012), "5" (Li et al., 2013), "10" (Salimizand et al., 2013), or "several" (Peymani et al., 2012) classes of integrons based on their knowledge of clinically derived integrons.

We and other authors have also experienced erroneous dogma in comments provided by expert reviewers for submitted manuscripts, particularly with regard to whether integrons described from natural environments correspond with what is known about "integrons" (mostly class 1). An excellent example as to why research from class 1 integrons cannot be extrapolated to all integron classes is shown in the recombination reaction rate of the class 1 integron and that from *Vibrio cholerae*. The *V. cholerae* integron has a 2600-fold higher rate of recombination in a *V. cholerae* background compared to an *Escherichia coli* background indicating the involvement of host factors (Biskri et al., 2005). In contrast, the class 1 integron shows no difference in rates of recombination in both backgrounds. The class 1 integron's capacity to operate in different backgrounds is the likely reason for why this particular integron has been successful in its mobilization across different bacteria. This trait and possibly a greater capacity to integrate cassettes with diverse *attC* sites (Biskri et al., 2005) are likely to make this integron an exception rather the rule.

Secondly, it has been over 12 years since the discovery that integrons are diverse and found in different environments (Mazel et al., 1998; Nield et al., 2001). Knowledge on the function of integrons with regard to site-specific recombination, transcription of gene cassettes, and regulation of the integron-integrase has significantly advanced and has been excellently reviewed elsewhere (Cambray et al., 2010; Roy Chowdhury et al., 2011). However, little

"fmicb-04-00385" — 2013/12/5 — 18:15 — page 3 — #3

progress has been made in addressing the precise ways in which gene cassette products contribute to the adaptation and evolution of bacteria outside of antibiotic resistance. Based on our knowledge of integrons in antibiotic resistance, we are aware of the power of this system in providing rapid adaptation under strong selection pressure(s). In approximately 70 years, the integron has assisted in making antibiotic treatment problematic and most likely obsolete in the next 10 years (World Health Organisation, 2013). Class 1 integrons are now a common fixture on plasmids from commensal bacteria and Gram negative pathogens. Re-entry of commensal and Gram negative pathogens into the broader environment through routes such as wastewater treatment ensures that access to the environmental gene cassette metagenome will be easy and rapid. Thus, a lack of understanding or distribution of misinformation regarding this greater resource, particularly in the antibiotic/clinical field, has the potential to be counterproductive in the quest to solve or improve the antibiotic resistance problem into the future.

In addressing the knowledge gap for the broader adaptive role of integrons, *Vibrio* species make excellent candidates. As already stated the cassette arrays of vibrios tend to be quite large comprising up to 3% of their genome and consisting of diverse and unique gene cassettes. To date, the largest cassette array is in*Vibrio vulnificus* CMCP6 consisting of 219 cassettes and totaling approximately 150 kb (Kim et al., 2003). In order to highlight the necessity for research into the role integrons play in bacterial adaptation and evolution and to focus attention on the lack of understanding that exists about the function of this element in bacteria generally, we will review and discuss the phenotypic functions of cassettes in*Vibrio* species in the context of what has been discussed above.

#### **A BIG BLACK BOX IN OUR UNDERSTANDING OF GENE CASSETTE FUNCTION**

Cassette arrays in *Vibrio* species are large and mostly consist of unique and novel genes with no identifiable function. In 2007, a bioinformatics survey of gene cassettes from multiple genome sequenced *Vibrio* species found that 65% of cassette proteins had no known homologs and that 13% had homologs of unknown function (Boucher et al., 2007). The remainder showed a wide range of non-specific functions in metabolism, cellular processes, and information storage. Similar statistics have been observed through PCR amplification of gene cassettes from metagenomic DNA (Elsaied et al., 2007; Koenig et al., 2008, 2009). Putting aside this massive knowledge gap in cassette function for the moment, large cassettes arrays provide an extra level of complexity. While some have argued that P*<sup>c</sup>* is the only driver of cassette transcription in large arrays (Guerin et al., 2009; Cambray et al., 2010), other studies have shown otherwise (Yildiz et al., 2004; Michael and Labbate, 2010). A study of the 116-gene cassette array of *V. rotiferianus* DAT722 showed that the majority of gene cassettes were transcribed and that numerous diverse promoters across the array were present that responded to different growth conditions (Michael and Labbate, 2010). The presence of these diverse promoters provides integrated cassettes with multiple regulatory options. This gives the capacity for cassettes to re-arrange with different promoters potentially building operon-like structures that express complimentary cassette proteins. Such an idea has been demonstrated in principal using artificial gene cassettes containing genes for tryptophan biosynthesis (Bikard et al., 2010). This complexity can be elevated when we consider that *Vibrio* species live in populations where gene cassettes might be considered a community resource not just a singular cell resource. For example, integrons might provide a way for the community to break down and/or extract energy from complex substrates without the entire pathway (and genetic burden) being owned by just one cell. Amusingly, the complexity of integrons has been used as proof for God/intelligent design (Hunter, 2010).

Even with the limited understanding of gene cassette function, a number of studies have sampled the gene cassette metagenome from different environments and attempted to determine how cassettes might influence adaptation and evolution (Elsaied et al., 2007, 2011, 2013; Koenig et al., 2008, 2009, 2011). Although correlations are observed such as homologs of genes in cassettes encoding pollution degrading enzymes from contaminated environments (Nemergut et al., 2004; Koenig et al., 2009) or environments showing a "gene cassette ecotype" (Koenig et al., 2008), it is still the case that ∼80% of the gene cassettes are of unknown function. In a study looking at gene cassettes from *Vibrio* species found in coral mucus, 12.5% of gene cassettes were implicated in biochemical processes also associated with antibiotic resistance (Koenig et al., 2011). The authors argued that gene cassettes provide a competitive advantage by delivering protection from, or by synthesizing, antimicrobials in the coral environment. While sound, the fact that this conclusion could be drawn clearly reflects the amount of research that has been done in the integron/resistance field. No other conclusions on the other cassette-assisted bacterial interactions present in the coral mucus could be made. So we are still left with a gaping hole in our understanding of how gene cassettes contribute to adaptation and/or evolution in this environment.

A handful of gene cassette products have been functionally characterized and these have been summarized in **Table 1**. In many instances, characterization of these gene cassettes was selected based on some homology to a known protein such that a phenotype could be tested which does not really address the bulk of unknown and hypothetical gene cassette products. In other instances, they were selected based on their capacity to be crystallized or were identified as part of mutant library or other screens. In the instances where gene cassettes have been removed from their natural bacterial host and expressed in *E. coli* or where *in vitro* techniques have been used to study protein activity, caution must be taken in how their function is interpreted. Interactions of cassette proteins with host pathways may modify how these gene cassettes affect cell or community behavior. This was observed in a study in *V. rotiferianus* DAT722 where deletion of a gene cassette encoding a putative topoisomerase I-like protein affected porin regulation. This phenotype could not have been predicted if characterized outside the host (Labbate et al., 2011). This is also true of the bioinformatic studies described above where in the small proportion of gene cassette products that could be identified were often proteins such acetyltransferases, methylases, or transcriptional regulators. Without knowing the primary substrate that is being modified by the acetyltransferase or methylase or the gene(s) controlled by the transcriptional regulator, the biological importance of the cassette(s) is still unclear. Therefore, an approach where gene cassettes are deleted or expressed in their natural host is arguably the best way to identify their physiological role.

#### **NEW INSIGHT INTO GENE CASSETTE FUNCTION IN THE VIBRIOS**

*Vibrio rotiferianus* DAT722 is the only microorganism where extensive physiological analysis has been conducted on isogenic mutants with gene cassettes deleted. This has revealed new insights into how gene cassettes affect adaptation and evolution. In one study, a cassette could not be deleted without a compensatory mutation (Labbate et al., 2011). The resulting mutants had abnormal regulation of their porins and impaired growth in minimal medium. The gene cassette in question was the 11th cassette from *attI* and appears to be strain specific by lacking close relatives elsewhere. The cassette 11 protein contains two domains, one with weak homology to nucleases and the other a C4-zinc finger domain commonly found in topoisomerase 1 proteins. These domains indicate a DNA binding/processing protein that is likely to have a regulatory role potentially through controlling the coiling state of gene promoters. Irrespective of the exact role, this study is important in demonstrating that cellular networks can rapidly integrate a mobile gene cassette such that it becomes advantageous for survival. It also shows that benefit need not necessarily come from acquisition of a novel functional gene(s) but through modification of existing host cellular networks (Labbate et al., 2012).

In a follow up to this study, the impact of deletions on the cassette array of *V. rotiferianus* DAT722 was addressed (Rapa et al., 2013). Indels are regularly observed in *V. cholerae* arrays and are likely in all large arrays however, their impact on bacterial physiology were unknown (Labbate et al., 2007; Szekeres et al., 2007; Yan et al., 2011). Three deletion mutants were subjected to physiological growth, stress, proteomic, and chemistry-based techniques to determine the effect of cassette deletions on vibrio physiology. The total deleted cassettes encompassed 58% of the array. Surprisingly, growth and stress assays of these mutants showed little change compared to the wild-type. Furthermore, proteomic analysis of one deletion mutant in different media and growth stages showed only 0.5–1% change in the proteome. This indicates that unlike deletion of cassette 11, the majority of cassettes are not integrated into host pathways and do not affect the major metabolic pathways of the cell, at least in the conditions observed.

Importantly, analyses did identify changes to host surface polysaccharide in the deletion mutants with proton nuclear magnetic resonance on whole cell polysaccharide indicating that gene cassette products decorate host cell polysaccharide via the addition or removal of functional groups. Consistent with this result, one mutant had modified biofilm-forming capabilities in a simple batch biofilm assay (Rapa et al., 2013). This is a significant result as it focuses future researchers who are addressing gene cassette function in vibrios to surface polysaccharide. We propose that at least a subset of cassettes are involved in modifying host surface polysaccharide and that deletion (and most likely rearrangements and acquisition) of cassettes is a mechanism for creating surface property diversity. There is significant biological

"fmicb-04-00385" — 2013/12/5 — 18:15 — page 4 — #4



importance to surface-associated polysaccharide and its modification as evidenced in the literature. This includes biofilm formation (Lee et al.,2013), bacterial cell co-aggregation (Vu et al.,2009), bacteriophage resistance (Scholl et al., 2005), evasion of immune cells (Pier et al., 2001) as well as resistance to antimicrobial peptides (Westman et al., 2008; **Figure 1**).

#### **CONCLUSION**

In the last 12 years, little progress has been made in the precise ways that cassette gene products contribute to adaptation and evolution of *Vibrio* species. One reason is the emphasis that is placed on studying integrons from clinical contexts. Another is that characterization of unknown genes is difficult and thus not considered a fruitful endeavor by researchers, especially in the current competitive research environment. However, if we are to learn more about the broader role of integrons, some of our focus needs to shift to identifying functions for gene cassettes. This will not only improve our understanding of this important genetic resource in a broader sense but give improved context for these elements clinically.

#### **ACKNOWLEDGMENTS**

We thank Dr. H. W. Stokes, who has since retired, for helpful comments and guidance in the preparation of this document. We also like to acknowledge his support as a mentor and for his contributions to the integron field throughout his career.

#### **AUTHOR CONTRIBUTIONS**

Rita A. Rapa and Maurizio Labbate both contributed to the writing of this manuscript.

#### **REFERENCES**

"fmicb-04-00385" — 2013/12/5 — 18:15 — page 5 — #5


*Vibrio cholerae* identifies a cationic drug-binding module. *PLoS ONE* 6:e16934. doi:10.1371/journal.pone.0016934


"fmicb-04-00385" — 2013/12/5 — 18:15 — page 6 — #6


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 04 September 2013; accepted: 25 November 2013; published online: 09 December 2013.*

*Citation: Rapa RA and Labbate M (2013) The function of integron-associated gene cassettes in Vibrio species: the tip of the iceberg. Front. Microbiol. 4:385. doi: 10.3389/ fmicb.2013.00385*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Rapa and Labbate. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00385" — 2013/12/5 — 18:15 — page 7 — #7

### Molecular characterization of vulnibactin biosynthesis in *Vibrio vulnificus* indicates the existence of an alternative siderophore

#### *Wenzhi Tan1, Vivek Verma1, Kwangjoon Jeong1, Soo Young Kim1, Che-Hun Jung2, Shee Eun Lee1,3 and Joon Haeng Rhee1 \**

*<sup>1</sup> Department of Microbiology, Clinical Vaccine R&D Center, Chonnam National University Medical School, Gwangju, South Korea*

*<sup>2</sup> Department of Chemistry, Chonnam National University College of Natural Science, Gwangju, South Korea*

*<sup>3</sup> Department of Pharmacology and Dental Therapeutics, School of Dentistry, Chonnam National University, Gwangju, South Korea*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Jesus L. Romalde, Universidad de Santiago de Compostela, Spain Daniele Provenzano, University of Texas Brownsville, USA*

#### *\*Correspondence:*

*Joon Haeng Rhee, Department of Microbiology, Clinical Vaccine R&D Center, Chonnam National University Medical School, Gwangju 501-746, South Korea e-mail: jhrhee@chonnam.ac.kr*

*Vibrio vulnificus* is a halophilic estuarine bacterium that causes fatal septicemia and necrotizing wound infections in humans. Virulent *V. vulnificus* isolates produce a catechol siderophore called vulnibactin, made up of one residue of 2, 3-dihydroxybenzoic acid (2, 3-DHBA) and two residues of salicylic acid (SA). Vulnibactin biosynthetic genes (VV2\_0828 to VV2\_0844) are clustered at one locus of chromosome 2, expression of which is significantly up-regulated *in vivo*. In the present study, we decipher the biosynthetic network of vulnibactin, focusing specifically on genes around SA and 2, 3-DHBA biosynthetic steps. Deletion mutant of isochorismate pyruvate lyase (VV2\_0839) or 2, 3-dihydroxybenzoate-2, 3-dehydrogenase (VV2\_0834) showed retarded growth under iron-limited conditions though the latter showed more significant growth defect than the former, suggesting a dominant role of 2, 3-DHBA in the vulnibactin biosynthesis. A double deletion mutant of VV2\_0839 and VV2\_0834 manifested additional growth defect under iron limitation. Though the growth defect of respective single deletion mutants could be restored by exogenous SA or 2, 3-DHBA, only 2, 3-DHBA could rescue the double mutant when supplied alone. However, double mutant could be rescued with SA only when hydrogen peroxide was supplied exogenously, suggesting a chemical conversion of SA to 2, 3-DHBA. Assembly of two SA and one 2, 3-DHBA into vulnibactin was mediated by two AMP ligase genes (VV2\_0836 and VV2\_0840). VV2\_0836 deletion mutant showed more significant growth defect under iron limitation, suggesting its dominant function. In conclusion, using molecular genetic analytical tools, we confirm that vulnibactin is assembled of both 2, 3-DHBA and SA. However, conversion of SA to 2, 3-DHBA in presence of hydrogen peroxide and growth profile of AMP ligase mutants suggest a plausible existence of yet unidentified alternative siderophore that may be composed solely of 2, 3-DHBA.

#### **Keywords:** *V. vulnifiucus,* **siderophore, salicylate, 2, 3-DHBA, hydroxyl radical, AMP ligase**

#### **INTRODUCTION**

*Vibrio vulnificus* is an opportunistic Gram-negative bacterial pathogen that causes fatal septicemia and necrotizing wound infections in susceptible individuals with high serum iron levels (Strom and Paranjpye, 2000). Due to its vital role as a redox cofactor of proteins, iron is an essential micronutrient for most life forms. Since iron availability is limited in biological systems, pathogenic bacteria have evolved an array of intricate mechanisms to scavenge iron from the host (Skaar, 2010). The low molecular weight compound called siderophore binds iron with high affinity (Braun and Killmann, 1999) and is an important virulence factor produced by *V. vulnificus* (Morris et al., 1987; Litwin et al., 1996). Virulent *V. vulnificus* produces a phenolate (catechol) siderophore called vulnibactin that enables it to acquire iron from highly iron saturated host proteins such as transferrin (Stelma et al., 1992).

which are involved in the formation of oxazoline rings with Lthreonine, which is bound to a norspermidine backbone (Okujo et al., 1994). The chemical structure of vulnibactin is closely related to the *Vibrio cholerae* catechol siderophore, vibriobactin. Vibriobactin was reported to be composed of three residues of 2, 3-DHBA, and like vulnibactin its biosynthesis was also found to be dependent on non-ribosomal peptide synthatases (NRPS) (Griffiths et al., 1984; Keating et al., 2000). Hence, it is highly likely that synthesis of 2, 3-DHBA in *V. vulnificus* is similar to that of *Vibrio cholerae.* The most important precursor in NRPS-dependent siderophore biosynthetic pathway is chorismate. 2, 3-DHBA could be formed from chorismate by a three step reaction catalyzed by isochorismate synthase (Liu et al., 1990), isochorismatase (Rusnak et al., 1990; Litwin et al.,

Vulnibactin was reported to be composed of one residue of 2, 3-DHBA and two residues of salicylic acid (SA), both of 1996), and 2, 3-dihydroxybenzoate-2, 3-dehydrogenase (Liu et al., 1989; Nahlik et al., 1989). Salicylate is a powerful scavenger of highly reactive hydroxyl radicals, resulting in its non-enzymatic conversion to 2, 3-DHBA, 2, 5-DHBA and catechol, with the ratio of these products dependent upon iron concentrations and pH (Hiller et al., 1983; Halliwell et al., 1988; Maskos et al., 1990; Chang et al., 2008). In bacteria, like 2, 3-DHBA, salicylate could also be derived from chorismate. In *Pseudomonas aeruginosa* and *Pseudomonas fluorescens*, isochorismate synthase and an isochorismate pyruvate lyase were identified as being responsible for salicylate biosynthesis (Gaille et al., 2002, 2003). 2, 3-DHBA and salicylate are subsequently activated by AMP ligase as precursors for siderophore assembly (Keating et al., 2000; Khalil and Pawelek, 2011). In *V. vulnificus*, the genes encoding above-mentioned enzymes are located at a single gene cluster in the chromosome 2 (**Figure 1A**). Previously, by an *in vivo* transcriptome analysis we observed a significantly up-regulated expression of the genes in this cluster (Rhee, unpublished data). In the present study, using molecular genetics tools, we characterize the vulnibactin biosynthetic pathway highlighting the contribution of 2, 3-DHBA and SA to vulnibactin biosynthesis in *V. vulnificus.* We focus on four genes in the mentioned gene cluster (**Figure 1A**) i.e., VV2\_0834 (2, 3-dihydroxybenzoate-2, 3-dehydrogenase), VV2\_0839 (isochorismate pyruvate lyase), and two genes encoding putative AMP ligases (VV2\_0836 and VV2\_0840) for activation of 2, 3-DHBA or SA. By gene deletion and respective in-trans complementation studies we establish the essentiality of these genes in bacterial virulence. Further, by exogenously supplying the products of deleted genes, we found that 2, 3-DHBA is more important than SA for growth under iron limited conditions, and endogenously synthesized SA could serve as a scavenger of hydroxyl radicals, supplying 2, 3-DHBA for siderophore biosynthesis. Furthermore, we cloned and purified VV2\_0836 and VV2\_0840 encoded AMP ligases and found that both of these enzymes are capable of activating 2, 3-DHBA and SA, and their essentiality for siderophore biosynthesis is dependent on iron levels. By way of molecular dissection of various pathways of siderophore synthesis in *V. vulnificus* we envisage the existence of an alternative siderophore composed solely of 2, 3-DHBA.

#### **MATERIALS AND METHODS**

#### **BACTERIAL STRAINS, PLASMIDS, AND GROWTH CONDITIONS**

Bacterial strains and plasmids used in the present study are enlisted in **Table 1**. *V. vulnificus* CMCP6 is a clinical isolate from a male patient, isolated at the Chonnam National University Hospital, South Korea. *V. vulnificus* CMCP6 was grown in 2.5% NaCl heart infusion (HI) medium while *Escherichia coli* strains were grown in Luria-Bertani (LB) medium supplemented appropriately with antibiotics. Bacteria were grown at 37◦C under shaking conditions (200 rpm). Thiosulfate citrate bile salt sucrose agar (TCBS) (Merck, Darmstadt, Germany) was used as the selective medium for *V. vulnificus*. For *E. coli,* antibiotics were used at the following concentrations: ampicilin (Amp) 100μg/ml, kanamycin (Km) 100μg/ml, chloramphenicol (Cm) 30μg/ml, and tetracycline (Tc) 12.5 μg/ml. For *V. vulnificus*, Amp (20μg/ml), Tc (2μg/ml) and Cm (2μg/ml) were used. To assess the growth of bacteria under iron-limited conditions, overnight grown bacterial cultures in HI medium were washed twice with phosphate buffered saline (PBS, pH 7.2) and inoculated into fresh HI broth supplemented with various concentrations of 2, 2-dipyridyl (DP) (Sigma) to a final concentration of 5 × 10<sup>5</sup> CFU/ml. The optical density at 600 nm (OD600) was measured spectrophotometerically (Ultrospec 6300 Pro, Amersham Biosciences) at selected time points.

#### **CONSTRUCTION OF IN-FRAME DELETION MUTANTS BY HOMOLOGOUS RECOMBINATION METHOD**

The chromosomal in-frame deletion mutants of VV2\_0834, VV2\_0839, VV2\_0836, and VV2\_0840 were constructed in *V. vulnificus* by allelic exchange method (Miller and Mekalanos, 1988). Primers used for PCR reactions are enlisted in **Table 2**. As per requirement, PCR primers were synthesized with overhangs recognized by specific restriction enzymes (REs) for

#### **Table 1 | Strains and plasmids used in this study.**


*Cmr , Cm resistance; Tc<sup>r</sup> , Tc resistance; Ap<sup>r</sup> , Ap resistance; Km<sup>r</sup> , Km resistance.*

ligation into appropriate vectors. Upstream and downstream 1000 base pair (bp) fragments of target genes were amplified separately and converted to 2 kbp fragments by cross over PCR (Horton et al., 1989). Fusion fragments were digested

#### **Table 2 | Primers used in the present study.**


*Underlined sequences indicate the RE recognition sites with their respective names written in parenthesis. \*Up forward, \*\*Up reverse,* #*Down forward,* ##*Down reverse.*

with appropriate REs and subcloned into pDM4 suicide vector. The resulting recombinant vector was transformed into *E. coli* SM10 λ *pir* and subsequently mated into *V. vulnificus* CMCP6 by conjugation. Stable Cm*<sup>R</sup>* transconjugants were selected on TCBS agar plate containing Cm. Plating of the transconjugants on 2.5% NaCl-HI agar plate containing 10% sucrose was performed to select clones that experienced the second homologous recombination events forcing excision of the vector sequence and leaving only mutated or wild type allele of the genes. Each in-frame deletion mutation was confirmed by PCR with the chromosomal DNA from the respective mutant as template. The resulting mutant strains are enlisted in **Table 1**.

#### **COMPLEMENTATION OF MUTANTS**

For complementation of the mutants, DNA fragments containing wild type genes with respective promoters were amplified using primers listed in **Table 2**. Amplified DNA fragments were cloned into pCR2.1 TOPO vector (Invitrogen). Fragments containing genes were cut out by appropriate REs and subcloned into the broad host range vector pLAFR3II (Kim et al., 2003). The resulting plasmids were transferred into the mutant strains by the triparental mating using a conjugative helper plasmid pRK2013 (Ditta et al., 1980). The transconjugants were screened on TCBS agar plates containing appropriate antibiotics and confirmed by PCR.

#### **RNA EXTRACTION AND REVERSE TRANSCRIPTION (RT-PCR)**

Total RNA from log phase *V. vulnificus* was extracted using RNeasy mini kit (QIAGEN, Germany) in accordance with the manufacturers' protocol. The quality of RNA was assessed using the NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA). The cDNA (from total RNA; 1 μg/reaction mixture) was synthesized with the QuantiTect® Reverse Transcription Kit (Qiagen) according to manufacturer's instructions. 0.5 μL of cDNA was used in each PCR reaction. Primers used for RT-PCR are listed in **Table 2**. Fragments were resolved by electrophoresis on 1% agarose gel.

#### *In vitro* **SALICYLATE (SA) MEASUREMENT**

SA produced in culture supernatants was determined as previously described (Meyer et al., 1992; Leeman et al., 1996). In brief, *V. vulnificus* strains were grown in Tris-HCl-buffered minimal medium (100 mM pH 7.5, 1.1 g NH4Cl, 0.27 g KH2PO4, 25 g NaCl, 4 g succinate, 2 g casamino acids/L) at 37◦C for 24 h. Cells were removed by centrifugation and culture supernatants were acidified with 1 N HCl to pH 2.0 and SA was extracted into CHCl3 by vigorous shaking (culture supernatant: CHCl3; 3:1). For quantitative measurements, 1 volume of 2.5 mM FeCl3 was added to the CHCl3 phase. The absorbance of purple Fe-SA complex developed in the aqueous phase was measured at 527 nm and quantified against a standard of SA dissolved in the same growth medium.

#### **DETECTION OF CATECHOL SIDEROPHORE PRODUCTION**

Siderophore production was detected in bacterial cultures grown in 2.5% NaCl-HI broth or CM9 minimal medium (0.4% glucose, 0.2% sodium succinate, 10 mg/L glutamate, 0.1μM FeCl3). Overnight grown bacterial cultures were washed in PBS and subcultured into 2.5% NaCl-HI supplemented with DP, or CM9. Siderophore production was quantified by the Arnow test (Arnow, 1937). Briefly, 0.2 ml of culture supernatant, 0.2 ml of 0.5 N HCl and 0.2 ml NaNO2-Na2MoO4.2H2O was mixed. After the formation of yellow color, 0.2 ml of 1 N NaOH was added resulting in the generation of red color. Total volume was brought to 1 ml with distilled water and the absorbance was measured at 510 nm using 2, 3-DHBA dissolved in the growth medium as standard. DP was assessed not to interfere with the Arnow test. Siderophore concentration was normalized to bacterial cell density (μM/OD600).

#### **PRODUCTION, PURIFICATION AND CHARACTERIZATION OF AMP LIGASE RECOMBINANT PROTEINS**

VV2\_0836 and VV2\_0840 encoding putative AMP ligases were amplified from *V. vulnificus* CMCP6 genomic DNA and cloned into the plasmid pTYB12 (New England Biolabs, Inc.). Amplified DNA fragments were sequenced by the dideoxy-chain termination method (Macrogen Inc., South Korea). Resulted plasmids were transformed into electrocompetent *E. coli* ER2566 (New England Biolabs, Beverly, MA) using the *E.coli* Pulsar (BioRad Inc.), for protein expression. Proteins were purified from IPTG induced (0.4 mM) transformed *E. coli* ER2566 after sonication (Sonics & Materials Inc. UK, Ltd.), by affinity chromatography using IMPACT™-CN Protein Purification System (New England Biolabs). The purity of recombinant proteins was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The concentration of the obtained proteins was determined by Quick Start™ 1 × Bradford dye (Bio-Rad Laboratories) and purified proteins were stored at −80◦C in stock buffer (25 mM Tris, pH 8.0, 10 mM MgCl2, 5 mM DTT, and 10% glycerol).

AMP ligase activity of purified recombinant protein was determined by measuring the pyrophosphate release (Rusnak et al., 1989). This reaction was coupled with the pyrophosphatase reaction. Reaction mixtures (total volume: 100μL) containing 1μM of VV2\_0836 or VV2\_0840 recombinant proteins, 0.2 unit of *E. coli* inorganic pyrophosphatase (Sigma), 75 mM Tris (pH 7.5), 10 mM MgCl2, 1.5 mM ATP, 5 mM DTT, and 0.6 mM salicylate (Junsei) or 2, 3-DHBA (Sigma) were incubated at 37◦C. The amount of inorganic phosphate (Pi) produced in the reaction mixture was assayed by measuring the chromophore generated after mixing with an ammonium molybdate-malachite green solution at 620 nm (Sousa et al., 2006; Khalil and Pawelek, 2011).

#### **DETERMINATION OF LD<sup>50</sup> IN MICE**

The LD50 of mutants to mice was determined by our previously described method (Kim et al., 2008). Briefly, 7-weekold specific pathogen-free (SPF) randomly bred ICR female mice (5 mice/group) were inoculated intraperitoneally with 10 fold serial dilutions of test strains (109–105 cfu/mouse). Deaths were observed for 48 h and LD50 values were calculated by the method of Reed and Muench (1938). All animal experiments were approved by the Animal Care and Use Committee of Chonnam National University, South Korea.

#### **STATISTICAL ANALYSIS**

Statistical significance of the growth differences at each time point and differences in siderophore production between strains were compared using the Student's *t*-test. *P*-values less than 0.05 were considered statistically significant. Statistical values were calculated using GraphPad Prism 5 or Microsoft Excel, wherever appropriate. All experiments were repeated three times in triplicates, and results from representative experiments are shown.

#### **RESULTS**

#### **EXPRESSION OF VV2\_0834, VV2\_0839, VV2\_0836, AND VV2\_0840 IS INDUCED BY IRON-LIMITATIONS AND IS UNDER THE CONTROL OF FUR**

To investigate the effect of iron levels on the expression of selected target genes, total RNA was extracted from wild type CMCP6 cultured in 2.5% NaCl-HI supplemented with various amounts of DP. Expression of target genes could not be detected in iron sufficient conditions (**Figure 1B**). However, expression of VV2\_0839 could be detected when DP was added to a final concentration of 50μM, while that of other three genes could be detected at 75μM DP concentration (**Figure 1B**). Fur (ferric uptake regulator) is the global regulator of iron acquisition system (Escolar et al., 1999; Panina et al., 2001). To confirm the regulatory effect of Fur, we developed a Fur deletion mutant of *V. vulnificus* and checked the expression of the four selected genes in this mutant grown in 2.5% NaCl-HI without any DP (**Figure 1B**). The expression of these four genes was found to be upregulated in the mutant suggesting that Fur acted as a repressor for expression of these four genes under iron-rich conditions.

#### **SA IS DISPENSABLE WHILE 2, 3-DHBA IS ESSENTIAL FOR** *V. vulnificus* **GROWTH UNDER IRON-LIMITED CONDITIONS**

The siderophore vulnibactin had been reported to contain one residue of 2, 3-DHBA and two residues of SA (Okujo et al., 1994). To investigate the importance of these two substrates for siderophore assembly, mutants with in-frame deletions of respective synthesis genes (VV2\_0834 and VV2\_0839) were constructed. The growth of mutants was estimated in iron-limited conditions. Against our expectations, -20839 mutant showed only a slight growth defect (*P* > 0.05) (**Figures 2A,B**), despite SA being a major building block of vulnibactin. To confirm the function of the gene VV2\_0839, SA production in -20839 mutant was measured. No SA production was observed in -20839 mutant (**Figure 2E**).

Under 140μM DP condition, -20834 exhibited initial significantly retarded growth (*P* < 0.01, 9 h time point) followed by an increased growth rate, ultimately catching up with the wild type strain (**Figure 2A**). However, the same mutant could not catch up the growth if the DP concentration was increased to 150μM (**Figure 2B**). The double mutant (-20839/20834) manifested more profound growth defect than single gene mutants under both 140 and 150μM DP conditions. These results clearly suggest that though VV2\_0839 is essential for SA synthesis but SA itself is dispensable, while 2, 3-DHBA is essential for the vulnibactin-dependent growth of bacterium under iron-limited conditions. More importantly, these results suggest that vulnibactin, composed of two SA and one 2, 3-DHBA may not to be the only type of siderophore secreted by *V. vulnificus* in response to iron-limitations.

#### **CATECHOL SIDEROPHORE PRODUCTION IS IMPEDED BY THE MUTATION OF VV2\_0834 AND VV2\_0839**

To further ascertain the importance of genes responsible for biosynthesis of the catechol siderophore in *V. vulnificus*, levels of siderophore production in mutants were measured by the Arnow test (Arnow, 1937). In 2.5% NaCl-HI broth supplemented with 140 or 150μM DP, more catechol siderophore was secreted

supplemented with either 140μM DP **(A,C)** or 150μM DP **(B,D)**. In some growth assays bacterial deletion mutants were complemented with plasmids carrying respective genes. Depending upon experiments mannitol (1 mM), SA (5μM), DHBA (5μM), or SA (5μM) with H2O2 (10μM) was added along with DP. SA produced by wild type and mutants cultured in minimal media was also assessed **(E)**. OD600 values at each time point are the means of two independent experiments done in triplicate. The error bars represent the standard errors. \*\**P* < 0.01.

by the -20839 mutant than the wild type strain (**Figure 3A**). But in minimal medium, no siderophore was detected in the culture supernatant of -20839 (**Figure 3C**) even though this mutant showed same growth levels as wild type (**Figure 3C,** satellite) suggesting the presence of an alternate siderophore composed solely of 2, 3-DHBA, produced in iron limited HI broth but not in minimal medium. However, results indicated that the siderophore synthesized without SA was less efficient in acquiring iron since the -20839 mutant was slightly impeded compared with wild type despite a higher level of siderophore produced in the culture. Interestingly, we found that -20834 mutant was capable of secreting catechol siderophore in 2.5% NaCl-HI supplemented with 140μM DP, and that siderophore was accumulated in the late growth stage (**Figure 3A**). The accumulated siderophore in the later growth phase seems to be responsible for the catch-up growth of the mutant (**Figure 2A**). Though the siderophore amount produced by the single mutant

was less than that secreted by wild type, it was significantly higher than that secreted by the double mutant (-20839/20834) (**Figure 3A**). However, the siderophore production in 2.5% NaCl-HI with 150μM DP and minimal medium was nearly abolished in both single and double mutants (-20839/20834) (**Figures 3A,C**) indicating that somehow SA contributed to the later growth stage production of catechol siderophore in -20834 mutant.

#### **HYDROXYL RADICALS FACILITATE THE PRODUCTION OF 2, 3-DHBA FROM SA IN VV2\_0834 MUTANT (***-***20834)**

During stationary growth phase, bacterial cells experience stressful conditions such as decreased pH or increased ROS in culture medium (Storz and Imlay, 1999; Poole, 2012). It has been previously reported that in the presence of free hydroxyl radicals SA is chemically converted to catechol acid, 2, 3-DHBA and 2, 5-DHBA (Grootveld and Halliwell, 1986). Thus, we presumed that SA produced by the -20834 mutant could have been attacked by ROS-derived hydroxyl radicals, which consequently lead to the formation of 2, 3-DHBA used for the ultimate production of vulnibactin during the observed late stage catch-up growth (**Figure 2A**). Mannitol has been shown to be an effective quencher of hydroxyl radicals, inhibiting the production of 2, 3-DHBA from SA (Wendel, 1987). To confirm our hypothesis that SA was being converted to 2, 3- DHBA, we grew the -20834 mutant in the presence of various concentrations of mannitol and found that 1 mM mannitol inhibited the late stage growth of -20834 mutant (**Figure 2A**). These results were further corroborated by measuring catechol siderophore production in -20834 mutant after mannitol supplementation. Compared to no mannitol condition, siderophore production in the -20834 mutant was significantly impeded in the presence of mannitol (**Figure 3A**). These results conclusively point toward the possibility of SA hydroxylation by free hydroxyl radicals.

#### **SA OR 2, 3-DHBA CAN RESTORE THE DEFECT OF** *-***20839/20834 DOUBLE MUTANT**

To further confirm the role SA and 2, 3-DHBA in the vulnibactin biosynthesis, we supplied the double mutant -20839/20834 with either SA or 2, 3-DHBA and observed the growth profile in iron-limited 2.5% NaCl-HI. As expected, 2, 3-DHBA fully restored the growth of double mutant in iron-limited conditions regardless of DP concentration (**Figures 2C,D**). However, SA could rescue the growth of double mutant in iron-limited condition till 140μM DP concentration showing a growth profile similar to the -20834 mutant under 140μM DP condition (**Figures 2A,C**). If the concentration of DP was increased to 150μM, SA alone could not rescue the growth defect of double mutant that could however be rescued in presence of H2O2 (**Figure 2D**). Considering that the generation of hydroxyl radical would be dependent upon iron availability and bacterial growth, it is likely that self-generated hydroxyl radicals from the double mutant were not sufficient to hydroxylate SA to 2, 3-DHBA for supporting mutant growth in 150μM DP condition. Growth patterns similar to that obtained after SA or 2, 3- DHBA supplementation of double mutant (-20839/20834) were obtained after complementation with either gene i.e., VV2\_0839 or VV2\_0834 (**Figures 2C,D**). Moreover the decrease pattern of siderophore in the presence of DP (**Figure 3A**) was similar to that observed with respective single deletion mutants (**Figure 3B**).

#### **ACTIVITY OF AMP LIGASES IS AFFECTED BY IRON LEVEL**

In the siderophore biosynthetic gene cluster, two genes (VV2\_0836 and VV2\_0840) encode putative AMP ligases for activation of 2, 3-DHBA and initiating siderophore assembly (Khalil and Pawelek, 2011). Their amino acid sequences showed 42% similarity. To characterize the roles played by these two genes in the siderophore synthesis, in-frame deletion mutants of these genes were constructed and growth was tested under iron-limited conditions. The single gene mutant -20840 showed more severe growth defect at only higher DP concentrations while growth of the -20836 strain was significantly retarded even at lower concentrations of DP (**Figures 4A–C**) indicating a more dominant role played by VV2\_0836 in vulnibactin assembly. As expected, the growth of double mutant -20836/20840 was significantly retarded by iron limitation that nonetheless could be restored fully by complementation with VV2\_0836 till at least 150μM DP condition, while only partially when complemented with VV2\_0840 (**Figure 4B**). In 160μM DP condition, the growth of double mutant could not be restored by either of the genes while growth defect of single gene mutants could be restored completely by respective gene complementation (**Figure 4C**). These results clearly indicate the essentiality of two AMP ligases for the siderophore biosynthesis under extremely iron-limited conditions.

#### **TWO AMP LIGASES CAN ACTIVATE 2, 3-DHBA AND SA**

To investigate the substrate specificity of two AMP ligases, we cloned the two genes (VV2\_0836 and VV2\_0840) into the expression vector pTYB12. Recombinant proteins were purified (**Figure 5A**) and tested for enzymatic activities with SA or 2, 3-DHBA as substrate. The AMP ligase reaction was assayed spectrophotometrically by coupling the formation of PPi to pyrophosphatase reactions. The generation velocity of Pi indicated the efficiency of AMP ligase. Results indicated that though both AMP ligases could act on 2, 3-DHBA and SA (**Figure 5B**) but AMP ligase encoded by VV2\_0836 had higher enzymatic activity than that by VV2\_0840. These results further indicate that SA might be the optimal substrate for AMP ligase encoded by VV2\_0836 since its action on SA was much prompt (*P* < 0.01) and significantly higher (*P* < 0.0001) than on 2, 3-DHBA.

#### **VV2\_0839 IS THE MAJOR CONTRIBUTOR TO MOUSE LETHALITY**

As shown in **Table 3**, mutation of either of the two AMP ligases (VV2\_0836 or VV2\_0840) resulted in a marginal 2-fold increase in LD50 compared to that of wild type strain. However, as expected, LD50 of double mutant -20836/20840 increased by 10-fold indicating an important synergistic contribution made by both AMP ligases toward virulence to mice, at least through intraperitoneal route. However, deletion of VV2\_0839 or VV2\_0834 gene resulted in an increase in LD50 by 137 and 2-fold respectively suggesting that SA played a more important role in virulence than 2, 3-DHBA. Considering 10-fold increase in LD50 by mutation of both AMP ligases, we could speculate that the significantly impaired virulence of -20839 mutant should not solely be due to the defect in the siderophore biosynthesis and SA might play physiological role through other yet unidentified mechanisms warranting the estimation of pathophysiological significance of SA synthesis in future studies.

**Table 3 | Effect of the mutation on the lethality to mice.**


#### **DISCUSSION**

The importance of iron for pathogenicity of *V. vulnificus* has been demonstrated both clinically and experimentally (Wright et al., 1981; Gulig et al., 2005). Siderophores mediate efficient iron uptake in most bacteria. It was reported that *V. vulnificus* produces a catechol-like siderophore called vulnibactin (Simpson and Oliver, 1983). Genes supposed to be involved in vulnibactin biosynthesis are clustered at a locus of chromosome 2 of *V. vulnificus*. We analyzed the transcriptome of this bacterium using rat peritoneal cavity infection model (Rhee, unpublished study) and found that genes in the aforementioned cluster were highly up-regulated *in vivo*. In the present study, the RT-PCR results showed that iron concentration tightly regulated the expression of selected target genes (VV2\_0834, VV2\_0839, VV2\_0836, and VV2\_0840) in this gene cluster. Expression of these genes was not detected in iron-sufficient conditions and could only be found after iron depletion by DP. Expression of each gene was initiated at different DP concentration. Fur is the global regulator of iron acquisition systems (Escolar et al., 1999). Up regulation of these four genes upon deletion of Fur suggested the repressor function of the regulator. However, we also observed expression of these genes in iron rich 0.9% NaCl-HI (data not shown); suggesting that Fur might not be the only regulator of these four genes as these might also be influenced by osmolarity sensing system. In this regard, siderophore genes may also be under the regulation of other global regulators *in vivo*.

In this study, we showed that deletion of VV2\_0839 abolished SA production and by extension siderophore production in minimal medium, but in iron-limited nutritious conditions higher amount of siderophore was produced suggesting that SA is dispensable for at least growth of bacterium in nutrition rich conditions as the mutant -20839 exhibited insignificant growth defect in HI medium. The chemical structure of vulnibactin was identified to contain one residue of 2, 3-DHBA and two residues of SA and was found to be closely related to that of vibriobactin, a type of catechol siderophore secreted by *V. cholerae* (Griffiths et al., 1984). It is likely that there is an alternative type of siderophore produced by *V. vulnificus,* composed solely of 2, 3-DHBA. However, the iron-sequestering activity of the alternative siderophore seemed to be weaker than vulnibactin (**Figures 2A,B**). Though for better conclusion of these observations future investigations regarding exact extraction and chemical characterization of siderophores would be required.

We found that the growth characteristics of -20834 mutant were tightly associated with the concentration of supplemented DP, hence the available iron levels. In HI with 140μM or lower DP concentrations, growth of -20834 mutant exhibited a stepwise pattern (**Figure 2A**) having a retarded early growth stage, short stable stage and the late growth stage catching up ultimately with the wild type. The late growth stage of -20834 mutant could be inhibited by increasing the DP concentration to 150μM, by addition of hydroxyl radical scavenger mannitol, or by additional deletion of VV2\_0839. SA itself is capable of scavenging reactive hydroxyl radicals (**·**OH) (Wendel, 1987) leading to the generation of catechol acid, 2, 3-DHBA and 2, 5-DHBA (Grootveld and Halliwell, 1986). In bacterial cultures, hydroxyl radicals could be generated from bacterial respiration or from Fenton reaction. We observed that the pH of culture medium of the -20834 mutant fell from 7 to 6 after 9 h of growth in 140μM DP condition. Hydroxylation of SA is favored in acidic pH resulting in higher conversion rates of SA to 2, 3-DHBA (Chakinala et al., 2007), explaining the stepwise growth pattern of -20834 mutant and ultimately reaching the growth profile of the wild type. From these data, we could conclude that SA was used as a scavenger of hydroxyl radicals protecting the bacteria from reactive oxygen species, and also supplying 2, 3-DHBA for siderophore biosynthesis. The complementation results and growth restoration by 2, 3-DHBA, SA, or SA with H2O2 further substantiated the hypothesis that SA was being converted to 2, 3-DHBA leading to the formation of vulnibactin even in the absence of VV2\_0834. These results also emphasized the essentiality of 2, 3-DHBA for bacterial growth in iron-limited conditions.

In the siderophore biosynthetic gene cluster, we found two genes encoding putative AMP ligases. Previously it has been predicted that the residues in the carboxyl acid binding pocket of AMP ligase help in discriminating between 2, 3-DHBA and SA (May et al., 2002). By alignment, the AMP ligase encoded by VV2\_0836 was more likely to act on 2, 3-DHBA, while the AMP ligase encoded by VV2\_0840 was more likely to act on SA. However, our *in vitro* enzymatic activity test results showed that both of AMP ligases were capable of acting on 2, 3-DHBA and SA. Moreover, AMP ligase encoded by VV2\_0836 exhibited much higher activity on SA than AMP ligase encoded by VV2\_0840. Both AMP ligases showed relatively low activity on 2, 3-DHBA compared to SA. By measuring the amount of secreted SA and 2, 3-DHBA by the double AMP ligase mutant cultured in the minimal media, we found more SA was accumulated than 2, 3-DHBA (data not shown). It is likely that *V. vulnificus* divide more chorismate for SA production than 2, 3-DHBA. For the growth under iron-limited conditions, VV2\_0836 encoded AMP ligase appeared to be more important than VV2\_0840 AMP ligase, except that both are required for growth in extremely ironlimited conditions. However, the mechanism through which iron influences the essentiality of AMP ligases remains unclear.

To assess the role of the four genes in the virulence of *V. vulnificus*, we determined the LD50 of mutants to ICR female mice. LD50 of -20839 was significantly higher than -20834. Our transcriptome analysis showed that *in vivo* expression of gene VV2\_0839 was up-regulated 187-fold compared with *in vitro* culture, while 96-fold for VV2\_0834, suggesting a more important role of SA in pathogenicity of *V. vulnificus* than 2, 3-DHBA. In this study, we observed that SA was scavenging the hydroxyl ions *in vitro*, and was capable of supplying 2, 3-DHBA for siderophore biosynthesis. Thus, it seems possible that *V. vulnificus* secrets SA for not only siderophore synthesis but also for protecting the bacterium from damage caused by hydroxyl radicals generated *in vivo*. Another interesting possibility for *in vivo* SA synthesis comes from the fact that SA possesses anti-inflammatory properties. There are a number of reports showing that SA interferes with intracellular signaling pathways such as protein-kinases (MAPK) cascade (Wang and Brecher, 1999) and NF-kB pathway (Kopp and Ghosh, 1994) and may diminish the host inflammatory response to the pathogen. However, this possibility remains to be verified by in appropriate animal model studies.

Taken together, in this study we addressed the important role of 2, 3-DHBA, SA and two AMP ligases for siderophore biosynthesis. Based upon the data obtained from the present study, we constructed a vulnibactin biosynthesis pathway map that is slightly different from that has been predicted in the KEGG database (http://www.genome.jp/kegg-bin/showpathway? vvu01053+VV20834). According to our proposed vulnibactin

biosynthetic pathway (**Figure 6**) isochorismate may be routed to 2, 3-DHBA or SA synthesis, mediated by specific enzymes (VV\_20834 and VV\_20839 respectively). Synthesized 2, 3-DHBA and SA are assembled into vulnibactin by two AMP ligases (VV2\_0836 and VV2\_0840). Moreover, SA in presence of hydroxyl radicals may be non-enzymatically converted to 2, 3-DHBA that may be utilized for the production of yet unidentified alternative siderophore that, like vibriobactin, may be composed solely of 2, 3-DHBA. Hence, it is intriguing to suggest that the type and quantity of siderophore synthesized would be determined by the environmental cues sensed by this opportunistic pathogen. We further suggest that SA in *V. vulnificus* serves as a scavenger of hydroxyl radicals and plays an importance role in the virulence of *V. vulnificus*. However, there are several unanswered questions such as exact contribution of SA and vulnibactin, not only as a virulence factor but also as a bacterial defense mechanism to evade host innate inflammatory responses, besides contribution of *in vivo* activated global regulators in siderophore synthesis that needs further exploration.

#### **ACKNOWLEDGMENTS**

Joon Haeng Rhee was supported by a grant from the Regional Technology Innovation Program of the Ministry of Knowledge Economy of the Republic of Korea (No. RTI 05-01-01). Shee Eun Lee was supported by an NRF grant from the MSIP (2013R1A2A2A01005011). Wenzhi Tan was supported by the BK21 Plus (Brain Korea 21) Program for Leading Universities and Students, Ministry of Education, Republic of Korea.

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 20 November 2013; paper pending published: 16 December 2013; accepted: 03 January 2014; published online: 24 January 2014.*

*Citation: Tan W, Verma V, Jeong K, Kim SY, Jung C-H, Lee SE and Rhee JH (2014) Molecular characterization of vulnibactin biosynthesis in Vibrio vulnificus indicates the existence of an alternative siderophore. Front. Microbiol. 5:1. doi: 10.3389/fmicb. 2014.00001*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Tan, Verma, Jeong, Kim, Jung, Lee and Rhee. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## Expression of *Vibrio salmonicida* virulence genes and immune response parameters in experimentally challenged Atlantic salmon (*Salmo salar* L.)

#### *Ane M. Bjelland1 \*, Aud K. Fauske1, Anh Nguyen2, Ingvild E. Orlien2, Ingrid M. Østgaard2 and Henning Sørum1*

*<sup>1</sup> Section for Microbiology, Immunology and Parasitology, Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, Oslo, Norway <sup>2</sup> Department of Pharmacy and Biomedical Laboratory Sciences, Faculty of Health Sciences, Oslo and Akershus University College of Applied Sciences, Oslo, Norway*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*David G. Weissbrodt, ETH Zürich and Eawag, Switzerland Luigi L. Fazio, Sapienza-Università di Roma, Italy*

#### *\*Correspondence:*

*Ane M. Bjelland, Section for Microbiology, Immunology and Parasitology, Department of Food Safety and Infection Biology, Norwegian School of Veterinary Science, PO Box 8146 Dep., 0033 Oslo, Norway e-mail: anemb@nvh.no*

The Gram-negative bacterium *Vibrio salmonicida* is the causative agent of cold-water vibriosis (CV), a hemorrhagic septicemia that primarily affects farmed Atlantic salmon (*Salmo salar* L.). The mechanisms of disease development, host specificity and adaptation, as well as the immunogenic properties of *V. salmonicida* are largely unknown. Therefore, to gain more knowledge on the pathogenesis of CV, 90 Atlantic salmon parr were injected intraperitoneally with 6 <sup>×</sup> <sup>10</sup><sup>6</sup> CFU of *V. salmonicida* LFI1238. Samples from blood and spleen tissue were taken at different time points throughout the challenge for gene expression analysis by two-step reverse transcription (RT) quantitative real-time polymerase chain reaction. Out of a panel of six housekeeping genes, *accD, gapA*, and 16S rDNA were found to be the most suitable references for expression analysis in *Vibrio salmonicida.* The bacterial proliferation during challenge was monitored based on the expression of the 16S rRNA encoding gene. Before day 4, the concentrations of *V. salmonicida* in blood and spleen tissue demonstrated a lag phase. From day 4, the bacterial proliferation was exponential. The expression profiles of eight genes encoding potential virulence factors of *V. salmonicida* were studied. Surprisingly, all tested virulence genes were generally highest expressed in broth cultures compared to the *in vivo* samples. We hypothesize that this general muting of gene expression *in vivo* may be a strategy for *V. salmonicida* to hide from the host immune system. To further investigate this hypothesis, the expression profiles of eight genes encoding innate immune factors were analyzed. The results demonstrated a strong and rapid, but short-lasting innate immune response against *V. salmonicida.* These results suggest that the bacterium possesses mechanisms that inhibit and/or resist the salmon innate immune system until the host becomes exhausted of fighting the on-going and eventually overwhelming infection.

**Keywords:** *Vibrio salmonicida***, cold-water vibriosis, Atlantic salmon, gene expression studies, virulence factors, innate immune response, two-step RT-qPCR**

#### **INTRODUCTION**

The motile Gram-negative rod *Vibrio salmonicida* is the causative agent of cold-water vibriosis (CV) in farmed Atlantic salmon (*Salmo salar* L.), rainbow trout (*Oncorhynchus mykiss*), and Atlantic cod (*Gadus morhua*) (Egidius et al., 1981, 1986; Holm et al., 1985; Jørgensen, 1987). The disease occurs mainly in late autumn to early spring and is a generalized septicemia characterized by anemia and extended internal and external hemorrhages (Holm et al., 1985; Poppe et al., 1985; Egidius et al., 1986). Although *V. salmonicida* has been known in Norwegian aquaculture for more than 25 years, only a few studies have so far identified components with possible roles in virulence. These include a surface antigen VS-P1, temperature-sensitive iron sequestration, possible production of hydrogen peroxide, quorum sensing and motility (Hjelmeland et al., 1988; Fidopiastis et al., 1999; Colquhoun and Sørum, 2001; Karlsen et al., 2008; Bjelland et al., 2012a,b). In addition, genomic analysis has identified three putative hemolysins, proteases and several protein secretion systems (Hjerde et al., 2008). The bacterium has, however, been described to be a poor producer of proteases and hemolysins, and a capacity of producing extracellular toxins has never been identified (Holm et al., 1985; Hjelmeland et al., 1988; Toranzo and Barja, 1993; Bjelland et al., 2012b).

Several studies have tried to uncover the pathogenicity of *V. salmonicida* (Totland et al., 1987; Espelid et al., 1988; Bøgwald et al., 1990; Evensen et al., 1991; Brattgjerd and Evensen, 1996). After challenge, *V. salmonicida* has been described to rapidly establish a bacteremia. Before the fish shows clinical signs of disease, bacterial cells have only been detected in the blood stream (Totland et al., 1987; Bjelland et al., 2012a). This latency period can persist up to 5–10 days in artificially infected fish. During this time it has been suggested that *V. salmonicida* uses the blood stream to proliferate and ensure a successful infection (Bjelland et al., 2012a). The first targets of *V. salmonicida* are reported to be the endothelial cells of capillaries and leukocytes of the blood. In the later stages of infection endothelial cells are completely disintegrated and actively proliferating bacteria can be detected in the extravascular space and in the surrounding tissue (Totland et al., 1987).

Little is known about the immune response toward *V. salmonicida* infections in salmonid fish. Previous investigations have mainly targeted the humoral immune response in a variety of vaccination studies. The dominant antigen VS-P1 has been described to specifically stimulate B lymphocytes and antibody production (Espelid et al., 1987; Espelid and Jørgensen, 1992). Although a strong humoral immune response against *V. salmonicida* is demonstrated and suggested to be of protective nature, a poor correlation between protection and antibody production in immune responses against *V. salmonicida* has been obtained (Lillehaug et al., 1993; Eggset et al., 1997). Good efficacy of fish vaccines in the absence of detectable antibodies has been postulated to be T-cell mediated (Eggset et al., 1997).

The threat of *V. salmonicida* to the fish farming industry has been mitigated by vaccination. However, a significant increase of CV outbreaks has recently been reported suggesting a reemerging pathogen (Johansen, 2013). If the pathogen were to reemerge, our lack of knowledge on the virulence mechanisms and host immune response would inhibit the development of new countermeasures. Thus, this study was conducted to further elucidate the pathogenesis of *V. salmonicida*. Due to the lack of identified virulence properties of *V. salmonicida*, we hypothesized that the bacterium may require specific host factors to express important virulence features such as extracellular toxins and adhesins. Therefore, we isolated total RNA from the blood of artificially challenged Atlantic salmon and analyzed the *in vivo* expression of potential virulence genes by two-step RT quantitative realtime (two-step RT-qPCR). To increase our knowledge about the immune responses in Atlantic salmon during CV, the transcription levels of eight innate immune parameters in spleen were evaluated by two-step RT-qPCR. Several hypotheses on the immune response against *V. salmonicida* have previously been formulated. For instance, the bacterium might hide from the host immune system during the latency period of infection, and proliferate. In contrast, in the absence of identified extracellular proteins it has also been suggested that a strong inflammatory response that eventually damage the host's own cells and tissue is responsible for the pathological signs of CV diseased fish (Bjelland et al., 2012a). Thus, this work aimed to verify this hypothesis and to date this is the first report on the innate immune response against a *V. salmonicida* infection.

#### **MATERIALS AND METHODS**

#### **FISH AND HOLDING CONDITIONS**

Ninety unvaccinated Atlantic salmon parr of approximately 50 g and 8 months old were obtained from Sørsmolt AS (Sannidal, Norway). The fish were transported to the aquarium at the Norwegian School of Veterinary Science (Oslo, Norway) in a tank containing 800 l of oxygenized freshwater with a dissolved oxygen concentration between 8 and 13 ppm. The salmon were kept in a 600 l tank supplied with well aerated freshwater purified on a carbon filter medium (Pentair Water, Minneapolis, MN, USA) at a temperature of ≈7◦C and with an oxygen saturation between 10 and 11 ppm. The fish were fed *ad libitum* and acclimatized to the experimental conditions 2 weeks prior to challenge.

#### **BACTERIA AND CHALLENGE PROCEDURE**

*V. salmonicida* wild type strains LFI1238 genome sequenced strain were taken from freeze stocks at −70◦C and cultivated in Luria Bertani broth with 1% NaCl (LB1) with agitation at 200 rpm at 8◦C for 2 days. To verify and prepare virulence of the bacterial strain used in challenge, three Atlantic salmon parr were anaesthetised in a water bath containing 0.0035% benzocaine (*Benzoak*® VET, Euro-Pharma, Chemainus, Canada) and injected intraperitoneally (i.p.) with 0.1 ml of the bacterial culture. The fish showed severe symptoms of CV 4 days post infection and were euthanized and post mortem examined. The bacterium was recovered by taking bacteriological samples from the head kidney using a sterile metal loop, plated on Blood Agar Base No. 2 (Oxoid, Cambridge, UK) supplemented with 5% ox blood and 2.5% NaCl (BA2.5) and incubated at 8◦C for 4 days. *V. salmonicida* LFI1238 strain directly isolated from the diseased fish was then cultivated for the gene expression experiment as described above. The challenge experiment was carried out with a lethal dose of *V. salmonicida* that was expected to kill approximately 80–90% of the fish. Under these conditions, the fish usually develop symptoms of disease after 4–5 days followed by mortalities from day 5–6 post infection (Nordmo et al., 1997; Nelson et al., 2007; Bjelland et al., 2012a,b). Ninety Atlantic salmon parr were anaesthetised as described above and divided in one test group of 70 fish and one control group of 20 fish. The fish fins were differentially clipped to distinguish between groups. After sedation, the fish were injected i.p. with 0.1 ml of cultures of strain LFI1238 grown on LB1 with absorbance at 600 nm (A600) of 0.3 representing 6 <sup>×</sup> <sup>10</sup><sup>6</sup> CFU. The final bacterial suspension was controlled by colony counting following serial dilution in LB1, plating of aliquots of 100μl in duplicate on BA2.5 and incubation at 8◦C for 4 days. The control fish were injected i.p. with 0.1 ml phosphate buffered saline (PBS). The challenge experiment was approved by The Norwegian Animal Research Authority (approval no. ID4877).

#### **SAMPLING**

Before challenge, 200μl of the final bacterial culture was transferred to 400μl of RNAlater (Ambion, Applied Biosystems, Foster City, CA, USA) in triplicate, vortexed and incubated for 5 min at room temperature followed by pelleting at 8600 rpm for 10 min using a Himac CT15RE tabletop centrifuge (Hitachi Koki Co., Ltd., Tokyo, Japan). The bacterial pellet was then stored at −20◦C until RNA isolation. At different time points starting from 2 h to day 6 after challenge, five fish were removed from the test tank and euthanized in a water bath containing 0.01% benzocaine. Blood samples were taken from the caudal vein using vacutainer blood collection tubes with EDTA anticoagulants. The procedure was followed by gently inverting the collection tubes several times to allow for anticoagulation. A volume of 200μl anticoagulated blood was then thoroughly mixed with 1 ml of RNAlater in a microfuge tube, incubated in room temperature for 1 h before pelleting (1 min, 13.000 rpm min) using a Himac CT15RE and stored at −20◦C until RNA isolation. In addition to blood sampling, the spleen was dissected and transferred to 1 ml RNAlater. The tissue samples were incubated for 24 h at 4◦C before stored at −20◦C until RNA extraction. Further, bacteriological samples from the blood and head kidney were plated on BA2.5 and incubated at 8◦C for 4 days. The presence of *V. salmonicida* and other bacterial species from each sample were graded based on the four-quadrant semi-quantitative scoring method (Cappuccino and Sherman, 2007). The first quadrant of the plate was streaked using a sterile metal loop and each successive quadrant was streaked using a new bacteriologic plastic loop in order to dilute the number of bacteria in each quadrant. Quantification was expressed as 1+, 2+, 3+, or 4+ based on the number of quadrants that demonstrated bacterial growth. Growth of *V. salmonicida* limited to quadrant 1 was categorized as 1+ (sparse amounts), bacterial growth limited to quadrants 1 and 2 was categorized as 2+ (moderate amounts), bacterial growth limited to quadrants 1, 2, and 3 was categorized as 3+ (rich amounts), and bacterial growth that extended to all four quadrants was categorized as 4+ (very rich amounts).

Identical sampling procedure was performed to 10 control fish; five fish before start and five fish at the end of the challenge. No moribund fish were sampled for gene expression experiments.

#### **RNA EXTRACTION**

Total RNA from bacterial pellets was extracted using the RNAeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions (Protocol 4 and 7) including an on-column DNA wipeout treatment (Appendix B1-4). To extract total RNA from blood, samples were incubated in 1 ml TRIzol reagent (Invitrogen, Carlsbad, CA, USA) for 10 min with regularly vortexing. Spleen samples were transferred to tubes with 5 mm stainless steel beads (Qiagen). Then, 1 ml TRIzol reagent was added followed by homogenization using TissueLyser II (Qiagen) at 25 Hz for 4 min. Further, 200μl chloroform was added to the blood and spleen samples. The samples were shaken by hands for 15 s, incubated at room temperature for 5 min and centrifuged at 11.400 rpm for 15 min at 4◦C min using a Himac CT15RE. A volume of 400–500μl of the aqueous phase was then mixed 1:1 with 70% ethanol, transferred to an RNAeasy® spin column (Qiagen). Total RNA was then isolated according to the manufacturer's protocol for extraction of total RNA from animal tissues using an RNAeasy® mini kit. The RNA was eluted in 30μl DEPC-treated water (Invitrogen) and stored at −70◦C until reverse transcription (RT). Gel electrophoresis with 1% agarose gel was used to confirm that isolated RNA was intact while the concentration and purity of the RNA extracts were analyzed by measuring the absorbances at 260 (A260) and 280 nm (A280) using a NanoDrop™ ND-1000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). Only total RNA samples of high quality with A260/A280 ratios between 1.9 and 2.2 and with tight bands of 18S/28S ribosomal RNA (rRNA) were used for RT.

#### **TWO-STEP REVERSE TRANSCRIPTION QUANTITATIVE REAL-TIME POLYMERASE CHAIN REACTION (TWO-STEP RT-qPCR)**

RT was conducted with QuantiTect® RT kit (Qiagen) according to manufacturer's instructions for the synthesis of complementary DNA (cDNA) and included a DNase wipeout treatment. Amounts of 1μg of RNA were used in each RT reaction conducted in a BioRad T100 (Bio-Rad, Hercules, CA, USA). In addition, to confirm the absence of any contamination with genomic DNA (gDNA) contamination, one RNA sample per round of extraction was randomly chosen and not treated with reverse transcriptase. The cDNA samples were diluted in 180 μl of DEPC-treated water and stored at −70◦C until quantitative real-time PCR (qPCR). The qPCR was performed using EXPRESS SYBR® GreenER™ qPCR Supermixes (Invitrogen) according to the manufacturer's instructions. The reactions were performed in triplicate with a Stratagene Mx3005P (Stratagene, La Jolla, CA, USA) detection system at the following conditions: 50◦C for 2 min, 95◦C for 2 min, 40 cycles of 95◦C for 15 s and 60◦C for 60 s, followed by 95◦C for 1 min, 55◦C for 30 s, 95◦C for 30 s (dissociation curve). Each plate included a no-template control for non-specific amplification that comprised DEPC-treated water instead of cDNA. Each primer pair was shown to have no primer dimer product in the no-template control and a single peak in the dissociation curve. Randomly chosen samples that were not treated with reverse transcriptase were tested to confirm any absence of contamination by gDNA. Data were captured using the Stratagene MxPro Mx3005P QPCR software. The potential virulence genes surveyed in this work were chosen to elucidate the role of extracellular toxins, secretion systems and adhesion factors in the pathogenesis of *V. salmonicida*. In addition, a panel of eight key innate immune genes was chosen based on previous reports to investigate the host's early response to a CV infection. Primer sequences targeting for the genes coding for 16S rRNA, AccD, FstZ, RpoD, LitR, EFN1β, βactin, 18S rRNA, TLR5S, TNFα, IL-1β, IL-6, IL-8, IL-12, IFNα, and C3 were obtained from previous publications (Olsvik et al., 2005; Løvoll et al., 2007; Hynes et al., 2011; Bjelland et al., 2012b). Other primers were designed from sequences found in GenBank using the Primer3 online program (**Table 1**). One sequence was used for each primer designed. Accession numbers (NCBI) for the sequences are as follows: *polA*, gi: 6989219; *gapA*, gi: 6987419; *vah2*, gi: 6988754; *vah5*, gi: 6986473; *hlyIII*, gi: 6986281; *tadA*, gi: 6962598; *vasA*, gi: 6989384; *tolC*, gi: 6961545; *epsD*, gi: 6988447. All primers were purchased from Invitrogen with standard desalting.

#### **DATA ANALYSIS OF GENE EXPRESSION**

The stability of six house-keeping genes was examined; *accD, fstZ, gapA, polA*, *rpoD,* and the 16S rRNA encoding gene (hereafter referred to as *16S rDNA*). The *geNorm* VBA applet for Microsoft Excel was used to calculate the gene expression stability measure *M* that is defined as the average pair-wise variation of a particular gene with all other potential reference genes i.e., the least stable gene gets the highest *M*-value (Vandesompele et al., 2002). The threshold cycle (Ct) values were transformed to quantities using the comparative Ct (--Ct) method and the highest relative quantities for each gene was set to 1 in accordance to the *geNorm*

#### **Table 1 | Primers used in this study.**


*\*This study; \*\*Bjelland et al., 2012b; \*\*\*Olsvik et al., 2005; \*\*\*\*Hynes et al., 2011; \*\*\*\*\*Løvoll et al., 2007.*

manual (Livak and Schmittgen, 2001). From this, a gene expression normalization factor was calculated for each sample based on the geometric mean of the reference genes. Normalized relative gene expression level for each target gene was then calculated by transforming Ct values to quantities using the comparative Ct method and dividing the quantities for each sample by the appropriate normalization factor. Relative fold changes of potential virulence gene transcripts were calculated compared to the bacterial culture before injection into the fish. The relative expression of salmon immune gene transcripts was calculated compared to the control fish which received an injection of PBS. The average of fold change values ± standard error of the mean (s.e.m.) was estimated using the Student's *t*-test and a *p* ≤ 0.05 was considered statistically significant. Statistical analyses and figure construction were performed using the GraphPad Prism 6 software (GraphPad Software, San Diego, CA, USA).

#### **RESULTS**

#### **CHALLENGE EXPERIMENT**

The bacterial concentrations in blood and head kidney samples of Atlantic salmon parr was by semi-quantitative estimation categorized to 1+ already 2 h after challenge. One day after challenge the concentration were categorized to 2+, and from 4 days post challenge the bacterial growth was categorized to 4+ (data not shown). Mortality was observed at day 6 when the experiment was terminated. *V. salmonicida* was grown in pure culture from the blood and head kidney from all diseased fish. No control fish showed symptoms of infection or mortality and neither *V. salmonicida* nor any other bacterial species were identified after cultivation from this group.

#### **EVALUATION OF POTENTIAL REFERENCE GENES IN TWO-STEP RT-qPCR STUDIES OF** *V. salmonicida*

The *geNorm* VBA applet for Microsoft Excel was used to determine the most stable genes from tested reference genes (Vandesompele et al., 2002). Based on the *M*-values, the stability of the six genes was ranked in the following order: *accD* > *gapA* > *16S rDNA* > *rpoD* > *polA* > *ftsZ*. The geometric mean of the three most stable reference genes was further used to calculate a gene expression normalization factor for each sample. *M*-values for these genes were 0.882 (*accD*), 0.891 (*16S rDNA*), and 1.115 *(gapA*). Previous studies have evaluated the stability of potential reference genes in qPCR studies of Atlantic salmon (Olsvik et al., 2005; Løvoll et al., 2009; Zhang et al., 2011). Based on these results, *EFN1*β, β*-actin* and the *18S rDNA* was chosen to be used as Atlantic salmon reference genes. The *M*-values in our study were 0.805 (β*-actin*), 0.930 (*EFN1*β), and 1.047 (*18S rDNA*).

#### **QUANTIFICATION OF** *V. salmonicida* **GROWTH IN TISSUE SAMPLES USING TWO-STEP RT-qPCR**

The 16S rRNA encoding gene (*16S rDNA*) was by *geNorm* demonstrated to be a stable housekeeping gene. *16S rDNA* was also the highest expressed gene per bacterial cell of all tested housekeeping genes throughout the study. In contrast, the most stable housekeeping gene *accD* demonstrated high Ct values close to the no-template-control (NTC) initially in the experiment. To ensure that the NTC does not contribute to the fluorescence signal of the target sequence, it is recommended that the Ct value of the unknown target gene should have Ct values of 3.3 cycles (a log value) fewer than that of the NTC Ct value (Smith et al., 2006). Thus, the relative expression of *16S rDNA* was used to illustrate the bacterial growth in the fish blood during challenge (**Figure 1**). The lowest expression of *16S rDNA* was observed at the first sampling point 2 h after challenge (*x*¯ = 1.00 ± 0.93). The relative expression of *16S rDNA* transcripts were therefore calculated compared to this time point. At 8 h post infection the relative expression of *16S rDNA* remained low (*x*¯ = 1.34 ± 0.59). One day after challenge the relative expression of *16S rDNA* increased by 16-fold (*x*¯ = 16.30 ± 6.78). The relative expression continued to increase with time and at 2 and 4 days post infection the relative fold change were 223 (*x*¯ = 223.15 ± 103.36) and 861 (*x*¯ = 860.77 ± 294.82) higher, respectively. Overall, the highest increase in relative expression was seen on day 6 after challenge and was at this time point 2107-fold higher compared to the initial sampling point (*x*¯ = 2107.36 ± 724.90). Statistical significant results were observed on day 4 (*p* = 0.03) and 6 (*p* = 0.03). No transcription products of *16S rDNA* were identified in the control fish.

#### **EXPRESSION OF POTENTIAL** *V. salmonicida* **VIRULENCE GENES DURING A COLD-WATER VIBRIOSIS INFECTION**

To investigate the contribution of bacterial virulence factors to the pathogenesis of *V. salmonicida,* the transcription levels of eight potential virulence genes in blood were evaluated by two-step RT-qPCR (**Figure 2**). The analysis were performed on samples taken from bacterial culture before challenge (*n* = 2) and from fish 5 h (*n* = 3), 2 days (*n* = 4), and 4 days (*n* = 3) post infection. Average Ct values detected in the bacterial challenge culture were for *litR*: 20.9 ± 0.8, *vah2*: 23.5 ± 0.3, *vah5*: 24.4 ± 0.08, *hlyIII*: 24.9 ± 0.3, *epsD*: 21.2 ± 0.2, *tolC*: 22.1 ± 0.3, *vasA*: 28.2 ± 0.5, and *tadA*: 23.7 ± 0.4.

Compared to transcription values of the *in vitro* bacterial challenge culture, the relative expression of *litR* significantly decreased

approximately 4 ± 1.2 (*p* = 0.05), 12 ± 1.1 (*p* = 0.001), and 10 ± 1.5 (*p* = 0.01) fold inside the fish at 5 h and on 2 and 4 days after challenge, respectively (**Figure 2A**). The relative expression of the three investigated hemolysin genes also decreased after the bacteria were injected into the fish. The change in transcription values from *in vitro* to *in vivo* conditions was only moderate for *vah2* (Mean 4 h p.i.: −9.5 ± 0.96, *p* = 0.003) and *hlyIII* (*x*¯ 4 h p.i. = −9.6 ± 0.57, *p* = 0.0008) showing between 5 and 10-fold decrease for both genes (**Figures 2B,C**). For *vah5*, however, the relative expression level was more than 100-fold higher *in vitro* compared to 2 days post infection (*x*¯ = −1.8 ± 45.6, *p* = 0.13) (**Figure 2D**). This result was not statistically significant, however, likely due to high individual variation.

The *in vivo* transcription values of *tolC*, *epsD,* and *vasA* were also reduced compared to *in vitro.* The relative expression of *tolC* was more than 100-fold higher *in vitro* compared to 2 (−117 ± 14.6, *p* = 0.006) and 4 (−245 ± 122.3, *p* = 0.22) days post infection (**Figure 2E**). Similar to the transcription profile of *vah5,* transcription values of *epsD* showed a significant 48-fold decrease 2 days after challenge (*x*¯ = −48 ± 7.4, *p* = 0.01) (**Figure 2F**). No transcription of *vasA* was detected 5 h after challenge (*p* < 0.0001) before the relative expression increased and was observed to be approximately 12-fold lower compared to *in vitro* levels on day 2 (*x*¯ = −12 ± 1.2, *p* = 0.001) and 4 (*x*¯ = −11 ± 1.8, *p* = 0.02) (**Figure 2G**).

Similar to all the other tested potential virulence genes, *tadA* showed significantly higher transcription rate *in vitro* compared to inside the fish host. Overall, *tadA* was related to largest differences between *in vitro* and *in vivo* gene expression with no transcription products detected 5 h after challenge (*p* < 0.0001), and fold change values of −822 ± 219.5 (*p* = 0.33) and −450 ± 28.6 (*p* = 0.001) on days 2 and 4, respectively (**Figure 2H**).

#### **EXPRESSION OF ATLANTIC SALMON IMMUNE GENES IN SPLEEN AFTER CHALLENGE WITH** *V. salmonicida*

The transcription levels of eight immune genes in spleen of Atlantic salmon were evaluated by two-step RT-qPCR after exposure to *V. salmonicida* (**Figure 3**). Average Ct values detected in control fish before challenge were for TLR5S: 30.9 ± 0.7, TNFα: 29.3 ± 0.4, IL-1β: 30.9 ± 0.7, IL-6: 34.9 ± 0.3, IL-8: 29.0 ± 0.4, IL-12: 29.5 ± 0.2, IFNα: 29.8 ± 0.3, and C3: 31.9 ± 0.8.

Relative gene expression levels of TLR5S remained low in the initial phase of the experiment. The fold changes increased from 8 h post infection (*x*¯ = 56.54 ± 27.51) until day 2, with day 2 showing the overall highest fold change (*x*¯ = 87.55 ± 40.38). From day 4, the relative gene expression decreased with relative fold changes of 16.93 ± 5.04 and 5.69 ± 1.88 on days 4 and 6, respectively (**Figure 3A**). Significant differences were observed between the challenged and control fish 4 days after challenge (*p* = 0.02). The highest fold change of TNFα was seen 8 h after challenge and was 122-fold higher compared to the control fish (*x*¯ = 122.15 ± 18.03). From this time point, the relative fold changes decreased with time to 60.89 ± 45.06, 48.58 ± 19.88, 14.26 ± 5.02, and 9.57 ± 3.97 on days 1, 2, 4, and 6, respectively (**Figure 3B**). Significant differences were observed between the challenged and control fish at 8 h (*p* < 0.001) and 4 days (*p* = 0.04) after challenge. The earliest response of IL-1β was observed 8 h after challenge with a fold change of 162 (±51.66, *p* = 0.02). The overall highest relative gene expression level was detected 1 day post infection with 742-fold change higher compared to the control fish (*x*¯ = 742.46 ± 489.99). This result was not statistic significant, however, likely due to high variance. From 2 days post challenge onwards, the relative fold changes decreased significantly with time to 388.75 ± 105.38 (*p* = 0.01), 119.15 ± 36.96 (*p* = 0.02), and 27.16 ± 2.56 (*p* < 0.0001), on days 2, 4, and 6, respectively (**Figure 3C**).

Relative fold changes of C3 increased with time from low levels 2 and 8 h post infection (2 h *x*¯ = 1.9 ± 0.76, 8 h *x*¯ = 3.21 ± 1.11) to the overall highest level on day 6 (*x*¯ = 73.64 ± 29.92). Significant differences were observed between the challenged and

control fish on days 1 (*p* < 0.001), 4 (*p* = 0.03) and 6 (*p* = 0.05) after challenge (**Figure 3D**). The transcription pattern of IL-6 was similar to IL-1β with the earliest response observed 8 h after challenge (*x*¯ = 29.80 ± 10.15). The highest fold change of IL-6 was the overall highest increase of all immune parameters that were measured and was detected 1 day post infection (*x*¯ = 816.75 ± 353.14). This result was not statistically significant, however, likely due to high variance. From 2 days post challenge, the relative fold changes decreased with time to 448.83 ± 197.55, 299.29 ± 44.27, and 127.91 ± 33.96 on days 2, 4, and 6, respectively (**Figure 3E**). Significant differences were observed between the challenged and control fish at 8 h (*p* = 0.03), 4 (*p* < 0.001), and 6 days (*p* = 0.01) after challenge. IL-8 transcripts were detected 8 h post infection (*x*¯ = 71.50 ± 30.92) and significantly increased on days 2 (*x*¯ = 112.17 ± 25.33, *p* = 0.005), 4 (*x*¯ = 203.09 ± 58.05, *p* = 0.01), and 6 (*x*¯ = 94.22 ± 27.89, *p* = 0.02) after challenge with the overall highest fold change on day 4 (**Figure 3F**).

Only small, but significant differences in IL-12 expression were observed in challenged fish compared to the control fish (**Figure 3G**). At 8 h post infection, transcription levels of IL-12 displayed a 2.5-fold increase (±0.55, *p* = 0.04). The highest level was observed on day 1 and was 10 times (±1.93, *p* = 0.003) higher compared to the control fish. For IFNα, a slight down-regulation was observed 2 h after challenge (*x*¯ = 0.289 ± 0.04) (**Figure 3H**). After this time point, the relative expression increased to levels above control fish levels, however, similar to the relative expression of IL-12 the fold changes remained low. The highest fold change was observed on day 2 (*x*¯ = 32.01 ± 26.75). Significant differences were observed between the challenged and control fish at 2 h (*p* = 0.0002) and 4 days (*p* = 0.02) after challenge.

#### **DISCUSSION**

#### **CHALLENGE**

To gain more insight into the pathogenesis of CV, an i.p. challenge experiment was performed with *V. salmonicida* in Atlantic salmon. Holding conditions should mimic environmental conditions to limit misinterpretation of the results (Eggset et al., 1997; Johansen et al., 2006). The experiment was therefore performed as closely as possible to a natural CV infection with respect to temperature and host specie. Previous studies have shown that a relative large number of *V. salmonicida* cells are necessary to develop CV in experimentally challenged Atlantic salmon. Bacterial inoculums between 10<sup>5</sup> and 107 CFU should give a total mortality between 70 and 90% where the lower challenge doses are observed to give a more prolonged disease development compared to the higher doses (Nordmo et al., 1997; Nelson et al., 2007; Bjelland et al., 2012a,b). Injection of a large volume of bacteria to the peritoneal cavity is, however, an artificial way of challenging fish which could affect the result of the study. A previous study has demonstrated that *V. salmonicida* rapidly invades the host's blood stream after bath as well as i.p. challenge (Bjelland et al., 2012a). Based on these findings, it was assumed that the expression of bacterial virulence and host immune genes in the fish blood is irrespective to challenge model and comparable with a natural infection of *V. salmonicida*.

#### **EVALUATION OF POTENTIAL REFERENCE GENES IN qPCR STUDIES**

To date, no studies have been performed to identify suitable candidate genes of *V. salmonicida* for normalization in qPCR based gene expression analysis. The present study therefore evaluated the suitability of six housekeeping genes as potential reference genes for expression *in vitro* as well as in blood and spleen tissue. The housekeeping genes *accD*, *gapA*, and *16S rDNA* were found to be the most stable expressed reference genes and further used for normalization of gene expression in *V. salmonicida*. A previous study has evaluated the stability of potential reference genes in qPCR studies of Atlantic salmon and ranked the stability of examined housekeeping genes in spleen as *EFN1*β > *S20* > *18S rDNA* > β*-actin* > *EFN1*α > *GAPDH* (Olsvik et al., 2005). Initially, the three most stable genes *EFN1*β, *S20*, and *18S rDNA* were chosen as reference genes in our study. However, the *S20* primer pair showed a double peak in the dissociation curve and subsequent attempts to optimize the reaction conditions did not solve the problem. Thus, β*-actin* was included in the reference gene panel. In contrast to the study of Olsvik et al. (2005), the most stable housekeeping genes in spleen were in our study ranked as β*-actin* > *EFN1*β > *18S rDNA.* The differences observed between the two studies could be related to factors as fish size, numbers, origin and health status since the study of Olsvik et al. included tissues from 15 healthy individuals with a weight range from 254 to 1898 g, while the present study included tissues from 28 fish, both healthy and infected, of approximately the same weight (50 g). Nevertheless, this demonstrates the importance of a careful selection of appropriate housekeeping genes for accurate and reliable normalization of gene expression data.

#### **QUANTIFICATION OF** *V. salmonicida* **GROWTH IN TISSUE SAMPLES USING TWO-STEP RT-qPCR**

The ability of *V. salmonicida* to rapidly establish a bacteremia has previously been demonstrated. The time span from the establishment of bacteremia until symptoms of CV are registered implies a latency period. It is suggested that this latency period is needed for bacterial proliferation to overcome the host's immune defense (Bjelland et al., 2012a). Although a large number of bacteria are reported in internal organs of moribund fish, no previous studies have been performed to evaluate the growth of *V. salmonicida in vivo* (Egidius et al., 1986; Totland et al., 1987). Thus, the relative expression of the stable bacterial reference gene *16S rDNA* was used to illustrate the bacterial growth in the fish blood during challenge. Semi-quantification of bacteria from the blood and head kidney by traditional culturing supported the two-step RT-qPCR results. This is in accordance with previous studies and demonstrates that traditional methods produce reliable results with high time- and cost-effectiveness (Løvoll et al., 2009; Bjelland et al., 2012a). The bacterial concentration remained low during the first 2 days of experiment. From day 4, a significant increase in bacterial transcripts was observed and the overall highest concentration of *V. salmonicida* was detected on day 6 when the experiment was terminated. These results are in accordance with the previous hypothesis that *V. salmonicida* requires a latency period in the blood stream for sufficient bacterial proliferation. The results illustrate a typical exponential bacterial growth curve starting with a prolonged lag phase where the bacterium adapts to the growth conditions and the host immune response. The latency period is followed by an exponential growth phase characterized by cell doubling at regular intervals. During this period, the bacterial growth is not limited by nutrition and host defense molecule, thus the host seems to be overwhelmed by the bacterial intruder. The present study was terminated on day 6 when the bacterial population still was in the exponential growth phase indicating a generalized septicemia.

#### **EXPRESSION OF POTENTIAL** *V. salmonicida* **VIRULENCE GENES DURING A CV INFECTION**

To compare the expression of potential virulence factors during *in vitro* and *in vivo* conditions and thus further elucidate the pathogenesis of *V. salmonicida,* the transcription levels of eight potential virulence genes were evaluated by two-step RTqPCR. Surprisingly, the potential virulence genes showed in general highest transcription levels *in vitro* compared to inside the fish host.

It has previously been demonstrated that the quorum sensing master regulator LitR negatively regulates adhesion, aggregation and biofilm production in addition to impacting virulence in *V. salmonicida*. Thus, *V. salmonicida* LitR is suggested to be an important factor for adapting the bacterium from a sea water living "biofilm mode" to a "planktonic mode" more suitable for infection (Bjelland et al., 2012b). Our study aimed to further investigate these hypotheses by *in vivo* expression analysis of *litR*. During an infection in a nutrition rich environment such as the fish tissue, increasing cell densities of *V. salmonicida* and expression of LitR should down-regulate this "biofilm mode." Surprisingly and in contrast to the previous *in vitro* study, the *litR* expression per cell *in vivo* was found to be cell-density independent. However, the increasing numbers of LitR-expressing *V. salmonicida* cells throughout the trial could likely maintain the planktonic mode.

It has been hypothesized that bacterial extracellular toxins are responsible for the extended petechial hemorrhages observed during a CV infection (Holm et al., 1985; Totland et al., 1987). With a few exceptions (on blood from mice and rabbit), production of hemolysins *in vitro* has never been described in *V. salmonicida*. However, genomic analysis have identified three putative hemolysins *vah2, vah5*, and *hlyIII* that shows 77, 52, and 70% identity to hemolysins of *V. anguillarum (vah2* and *vah5)* and *V. vulnificus (hlyIII),* respectively (Hjerde et al., 2008). The hemolytic activity of the *V. anguillarum vah* genes (*vah1–5*) has been considered to be the virulence factor responsible for hemorrhagic septicemia during vibriosis (Hirono et al., 1996; Rodkhum et al., 2005). Similarly, the *hlyIII* hemolysin of *V. vulnificus* is also demonstrated to play a role in virulence (Chen et al., 2004). By taking these previous reports into consideration, it was expected to observe an increase in production of hemolysins in this study. Most surprisingly the results came out to be the complete opposite. The results indicate that the expression of *vah2* and *hlyIII* is only moderately changed after entering the fish host. Thus, any significance to the pathogenesis of CV of this down-regulation is highly speculative. For *vah5*, however, the decrease in gene expression *in vivo* was significant and this result may indicate that the Vah5 protein in some way plays a role outside the fish host. The extensive hemolysis occurring in CV may then be a result of the activity of the immune system of the salmon.

Different types of secretion systems are described among *Vibrio* species and some of these are showed to be related to virulence e.g., rtx toxin and cholera toxin are secreted through type I and type II secretion systems, respectively. In *V. salmonicida*, the genes for six different secretion systems are identified and includes three type I (T1SSI, T1SSII, and T1SSIII), one type II (T2SS) and two type VI (T6SSI and T6SSII) secretion systems (Hjerde et al., 2008). Thus, the presence of these systems in *V. salmonicida* demonstrates that the bacterium has the tools required for the secretion of extracellular toxins and/or enzymes. To investigate the impact of secretion systems inside the fish host, our study included gene expression analysis of the three genes *tolC*, *epsD*, and *vasA* that code for essential proteins of *V. salmonicida* T1SSII, T2SS, and T6SSI, respectively. Equal to the other potential virulence genes investigated, *tolC*, *epsD*, and *vasA* showed highest transcription levels *in vitro* compared to inside the fish host. The similar transcription profiles of *epsD* and *vah5* might suggest that the hemolysin is secreted through T2SS. This hypothesis is, however, highly speculative and requires further investigations. Only three of six secretion systems identified in *V. salmonicida* were included in this study. Hence, there is a possibility that the remaining three systems are expressed in different ways compared to T1SSII, T2SS, and T6SSI. Nevertheless, the unanimous down-regulation of the investigated secretion system genes suggests that extracellular toxins and enzymes is of minor importance during the CV disease development.

Adhesion of bacteria to the host surface is one of the initial steps in microbial pathogenesis (Taylor, 1991; Kline et al., 2009). In *V. salmonicida*, genomic analysis has identified coding sequences for a flp-type pilus system. Flp-type pilus system is described in several bacteria to impact auto-aggregation and unspecific adherence, and in some species e.g., *Actinobacillus pleuropneumoniae* the system also impacts virulence (Li et al., 2012). In contrast, the flp-type pilus system of the fish pathogen *Aeromonas salmonicida* subsp. *salmonicida* is shown to make little or no contribution to the development of furunculosis (Boyd et al., 2008). Similar to the expression of the *V. salmonicida* T6SS *vasA* gene, *tadA* expression was totally absent 5 h after challenge. The muting of *tadA* after entering the fish host is in accordance with the previous presumption that down-regulation of adhesion properties is mandatory for adapting *V. salmonicida* to become a virulent fish pathogen. Thus, these results indicate that *tadA* mainly plays a role in bacterial survival during environmental stages of life. T6SS is reported to be related to highly variable phenotypes among different bacteria including biofilm formation (Aschtgen et al., 2010; Records, 2011; Liu et al., 2012). Thus, similar to *tadA,* the complete down-regulation of *vasA* could also be related to bacterial adaptation from the environment to the fish host.

The virulence gene expression study elicited the question about why *V. salmonicida* down-regulate many important and potential virulence properties after infecting the host. Obviously, the genes analyzed in this study may not be important virulence factors in *V. salmonicida* and more potential virulence genes should be included in future studies. qPCR is a precise and reproducible technology for assaying the expression of a small number of genes. However, with a large number of samples to test against each target gene, the qPCR technology has limitations when it comes to time and cost effectiveness. To identify large numbers of important pathogen and host genes methods such as microarrays and transcriptomics could be better choices (Moreira et al., 2012; Rader and Nyholm, 2012; Montanchez et al., 2013). The RT-qPCR method, however, is demonstrated to have a greater dynamic range and therefore allow for a more precise quantitation of expression levels than the high-throughput methods (Dallas et al., 2005). Nevertheless, these results might indicate that a general muting of gene expression is the bacterium's strategy to hide from the host immune defense system.

#### **EXPRESSION OF ATLANTIC SALMON IMMUNE GENES DURING A CV INFECTION**

To date, no studies have been conducted to investigate the innate immune response in Atlantic salmon during a CV infection. Therefore, to evaluate the results from the immune parameter analysis, previous reports from bacterial challenge experiment on salmonid fish were used for comparison (Løvoll et al., 2009; Raida and Buchmann, 2009; Ching et al., 2010; Hynes et al., 2011; Zhang et al., 2011; Aykanat et al., 2012; Kvamme et al., 2013). Toll-like receptor 5 (TLR5) plays an essential role in the innate immune defense against bacterial invasion through the recognition of the bacterial protein flagellin followed by the activation of various proinflammatory cytokines (Takeda et al., 2003). In salmonids, TLR5 exists both in a membrane bound (TLR5M) and a soluble (TLR5S) form. Hynes et al. (2011) have reported a significantly increase of TLR5S on day 4 with a fold change of 20–35 when Atlantic salmon was injected with purified flagellin. Another study that performed cohabitant challenge of Atlantic salmon with *Aeromonas salmonicida* has shown an approximately 2.5 and 10-fold increase of TLR5S after 3 and 14 days, respectively (Zhang et al., 2011). Furthermore, Raida and Buchmann (2009) have challenged rainbow trout fry with *Yersinia ruckeri* i.p. and showed that the expression of TLR5S in liver was 3.6-fold higher after 8 h followed by a steady decrease down to control fish expression levels.

The recognition of pathogen associated molecules such as LPS and flagella result in a proinflammatory cytokine cascade whereby TNFα is released followed by IL-1β and then IL-6 (Secombes et al., 2001). In two separate vaccine experiments using *V. anguillarum* antigens, the transcript levels of TNF1α was detected to be low (4-fold increases) or absent (Aykanat et al., 2012; Kvamme et al., 2013). In contrast, experiments with live *Vibrio anguillarum* demonstrated an approximately 15, 25, and 5-fold increases of TNFα expression 12, 24, and 72 h after i.p. challenge, respectively (Ching et al., 2010). More elevated levels of IL-1β have also been demonstrated after challenge with live bacteria than seen after vaccination. Ching et al. have reported an increasing expression of IL-1β from 20-fold changes after 6 h to approximately 100-fold changes 3 days after i.p.-challenge. In similar, an immersion challenge with *Moritella viscosa* in Atlantic salmon reported the first response of IL-1β after 2 days. The transcription level increased during the next days with the peak expression on day 7 (50–220-fold increase) (Løvoll et al., 2009). For IL-6, Hynes et al. (2011) have reported that the highest increase in expression was seen on day 4 (fold change of 78) followed by a significantly decrease. A similar profile has been demonstrated by Raida and Buchmann (2009) where IL-6 showed a significantly increase in fold changes at 8 h (4-fold), on days 1 (46-fold) and 3 (1100-fold) post infection, followed by a decrease in transcription levels.

In the present study, a minority of the results from the immune parameter experiment were statistically significant most likely due to variation between the fish's individual immune response (Lohm et al., 2002). To increase the probability for obtaining more significant results, more fish could have been included in the trial. However, the number of experimental animals is a balance between uncertainty in the expected results and the need to keep the number of animals at a low level. Another methodological consideration is that although the overall pattern of protein expression is similar to that of mRNA expression, the correlation between mRNA and protein levels may vary (Tian et al., 2004; Shebl et al., 2010). This could be explained by posttranscriptional and posttranslational regulation and by misclassification due to measurement errors including the total RNA content of the sample, the number of cells in the starting material, the RNA extraction efficiency and differential enzymatic efficiencies (Vandesompele et al., 2002). Thus, the use of proteomic techniques might help improve our understanding of the relationship between mRNA expression and protein production.

Nevertheless, the pattern of gene activation and expression suggests that the innate response during cold-water vibriosis is based on the pathogen's ability to engage TLRs signaling pathways to trigger the modulation of cytokines, in addition to a subsequent production of the complement component 3 (C3). *V. salmonicida* seems to induce a rapid and strong response of TLR5S in Atlantic salmon, with a relative fold change of more than 50 from 8 h to 2 days after challenge. However, as the bacterial growth became exponential after day 2, the transcription level of TLR5S decreased to lower levels. The gene expression of the three cytokines TNFα, IL-1β, and IL-6 showed similar profiles. The highest fold changes were observed after 8 h (TNFα) and 1 day (IL-1β and IL-6) and were 122, 742, and 817, respectively. These high peaks were all followed by a step-by-step decrease to low levels on day 6. It should be noted that the former studies used to compare our results include different fish species, infective agents and challenge methods. Nevertheless, our results demonstrate that *V. salmonicida* is recognized by the host and induces a rapid and strong, but short-lasting immune response. This indicates that *V. salmonicida* is resistant against important salmon cytokines and may also possess mechanisms that inhibit the host immune response during a CV infection.

Downstream of TNFα, IL-1β, and IL-6 other cytokines are released such as IL-8 that induce migration of neutrophils and phagocytes to the site of infection (Secombes et al., 2001; Zhu et al., 2013). Two previous studies have reported a rapid and small to moderate (5–15-folds) increase in IL-8 gene expression during challenge (Ching et al., 2010; Aykanat et al., 2012). In contrast, a later induction of IL-8 has also been reported (Hynes et al., 2011). In the present study, transcription products of IL-8 were identified 8 h after infection. However, in contrast to the other early detected cytokines TNFα, IL-1β, and IL-6, the gene expression of IL-8 demonstrated a more stable profile with fold values about 100 throughout the experiment. This chemokine attracts immune cells to the site of infection and our results may suggest that the increasing amount of bacteria during the infection maintain the strong stimulation of IL-8.

The cleavage of C3 gives e.g., the opsonin C3b that binds microbial cells and enhances phagocytosis (Nakao et al., 2011). The gene expression of C3 in Atlantic salmon has by Løvoll et al. (2009) been described to be induced 2 days after bath challenge with *M. viscosa* and followed by an increase throughout the study with a peak expression of 100–250-fold changes on day 7. Our study demonstrates similar results with increasing C3 gene expression levels from approximately three-fold changes 8 h after challenge to 74-fold changes on day 6. These results demonstrate that *V. salmonicida* has developed resistant mechanisms against the bactericidal effect of complement. Previously, it has been hypothesized that the host initiate a strong inflammatory response against *V. salmonicida* that eventually damage its own cells and tissue. The complement system has the potential to be extremely damaging to self-tissues, meaning its activation must be tightly regulated. Perhaps C3 is a factor that could address the pathological signs observed during CV.

The binding of microbial antigens to host recognition receptors also stimulate the transcription and secretion of the cytokines IL-12 and IFNα. IL-12 is the predominant cytokine driving the differentiation of naive T helper cells into TH1 cells (Ho and Glimcher, 2002). IFNα is mainly involved in antiviral defense (Robertsen, 2006). Similar to the report of Hynes et al. (2011), the present gene expression results of IL-12 and IFNα was mostly low or absent. IFNα was used as a marker for intracellular bacterial growth. Thus, the low or absent expression of IFNα indicates that the *V. salmonicida* pathogenesis mostly takes place extracellular. The low expression of IL-12 could inhibit the differentiation of T helper cells followed by a negative effect on the production of cytokines.

#### **CONCLUDING REMARKS**

In this work we have demonstrated that *V. salmonicida* proliferates in the fish blood after an initial latency period. To do this, *V. salmonicida* must have developed a proper strategy to resist the bactericidal effect of serum and avoid the immune cells of the blood stream.

It has earlier been suggested that the bacterium is hiding from the host immune system during the latency period and in this way increase in population size to ensure a successful infection. The muting of the potential virulence genes *in vivo* supports this hypothesis. In contrast, the immune parameter results demonstrate an initial rapid and strong immune response indicating that the pathogen is recognized by the salmon host. The immune response against *V. salmonicida* seems, however, to be shortlasting. This indicates that the bacterium possesses mechanisms that inhibits the development of a proper defense against CV or just simply uses resistance mechanisms to avoid the fish immune system. Finally, the host will become completely exhausted of fighting the overwhelming infection.

#### **ACKNOWLEDGMENTS**

The authors would like to acknowledge Stein Helge Skjelde (SørSmolt AS) for providing Atlantic salmon parr free of charge for the challenge experiment. Espen Brudal, Stine Braaen, and Elisabeth Furuseth Hansen (Institute of Food Safety and Infection biology, Norwegian School of Veterinary Science, Oslo) are also acknowledged for valuable advices and discussions. This work was supported by the Norwegian School of Veterinary Science.

#### **REFERENCES**

Aschtgen, M. S., Gavioli, M., Dessen, A., Lloubes, R., and Cascales, E. (2010). The SciZ protein anchors the enteroaggregative *Escherichia coli* Type VI secretion system to the cell wall. *Mol. Microbiol.* 75, 886–899. doi: 10.1111/j.1365- 2958.2009.07028.x


in cold seawater. *FEMS Microbiol. Ecol*. doi: 10.1111/1574-6941.12216. [Epub ahead of print].


gene copy and gene transcript numbers in environmental samples. *Environ. Microbiol.* 8, 804–815. doi: 10.1111/j.1462-2920.2005.00963.x


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 11 September 2013; paper pending published: 11 October 2013; accepted: 05 December 2013; published online: 20 December 2013.*

*Citation: Bjelland AM, Fauske AK, Nguyen A, Orlien IE, Østgaard IM and Sørum H (2013) Expression of Vibrio salmonicida virulence genes and immune response parameters in experimentally challenged Atlantic salmon (Salmo salar L.). Front. Microbiol. 4:401. doi: 10.3389/fmicb.2013.00401*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Bjelland, Fauske, Nguyen, Orlien, Østgaard and Sørum. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## Extracellular proteolytic enzymes produced by human pathogenic *Vibrio* species

#### *Shin-ichi Miyoshi\**

*Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, Okayama, Japan*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Li Sun, Institute of Oceanology, Chinese Academy of Sciences, China Jung Sup Lee, Chosun University, South Korea*

#### *\*Correspondence:*

*Shin-ichi Miyoshi, Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, 1-1-1, Tsushima-Naka, Kita-Ku, Okayama 700-8530, Japan e-mail: miyos-s@cc.okayama-u.ac.jp*

Bacteria in the genus *Vibrio* produce extracellular proteolytic enzymes to obtain nutrients via digestion of various protein substrates. However, the enzymes secreted by human pathogenic species have been documented to modulate the bacterial virulence. Several species including *Vibrio cholerae* and *V. vulnificus* are known to produce thermolysinlike metalloproteases termed vibriolysin. The vibriolysin from *V. vulnificus*, a causative agent of serious systemic infection, is a major toxic factor eliciting the secondary skin damage characterized by formation of the hemorrhagic brae. The vibriolysin from intestinal pathogens may play indirect roles in pathogenicity because it can activate protein toxins and hemagglutinin by the limited proteolysis and can affect the bacterial attachment to or detachment from the intestinal surface by degradation of the mucus layer. Two species causing wound infections, *V. alginolyticus* and *V. parahaemolyticus*, produce another metalloproteases so-called collagenases. Although the detailed pathological roles have not been studied, the collagenase is potent to accelerate the bacterial dissemination through digestion of the protein components of the extracellular matrix. Some species produce cymotrypsin-like serine proteases, which may also affect the bacterial virulence potential. The intestinal pathogens produce sufficient amounts of the metalloprotease at the small intestinal temperature; however, the metalloprotease production by extraintestinal pathogens is much higher around the body surface temperature. On the other hand, the serine protease is expressed only in the absence of the metalloprotease.

**Keywords:***Vibrio***, vibriolysin, thermolysin, collagenase, serine protease**

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 1 — #1

#### **HUMAN PATHOGENIC** *VIBRIO* **SPECIES**

Bacteria in the genus *Vibrio* are normal habitants of aquatic environments and play important roles in maintaining the aquatic ecosystem. Although more than 100 species are currently in this genus, at least, 11 species listed in **Table 1** are human pathogens (Janda et al., 1988; Chakraborty et al., 1997). Amongst, *Vibrio hollisae* is recommended to place into a novel genus *Grimontia* (Thompson et al., 2003). Human pathogenic species can be classified into two groups according to types of infectious diseases: the group causing gastrointestinal diseases and that causing extraintestinal diseases. Representative species in the former group are *V. cholerae* and *V. parahaemolyticus*, whereas *V. vulnificus* is the most important species in the latter group. Two species,*V. damsela* and *V. vulnificus*, are also fish pathogens.

Of human pathogenic vibrios,*V. cholerae*is the most extensively studied species because it is a causative agent of severe watery diarrheal disease, cholera. Cholera is characterized by uncontrolled purging of copious rice water stools leading to serious electrolyte depletion, dehydration, acidosis, shock, and, if left untreated, to death. Although *V. cholerae* is divided serologically into more than 200 groups, only two serogroups, O1 and O139, are etiologic agents of epidemic cholera. The non-O1/non-O139 serogroups are etiologic agents causing sporadic diarrheal cases and occasionally extra-intestinal infections, including skin, ear, sputum, urine, and cerebrospinal fluid infections. Cholera toxin (CT) produced by the O1 and O139 serogroups is a major toxic factor evoking severe watery diarrhea, whereas, hemolysin (Shinoda and Miyoshi, 2006), Zonula occludens toxin (Fasano et al., 1991), and accessory CT (Trucksis et al., 1993) have also been demonstrated to be additional enterotoxic factors.

*Vibrio parahaemolyticus* inhabits commonly coastal and estuary areas in the tropical and temperate regions, but this species has been recognized to cause gastroenteritis following consumption of seafood. The outstanding features of gastroenteritis are severe abdominal pain, diarrhea (frequently bloody stools), nausea, vomiting, mild fever, and headache. The mean incubation period is 6 to 12 h and diarrhea or soft stools persist for 4–7 days. Although the food poisoning caused by *V. parahaemolyticus* had been frequent in Japan, the number of outbreaks has been decreased drastically in the last 10 years. Production of the hemolysin designated as thermostable direct hemolysin is closely related to the bacterial pathogenicity (Shinoda and Miyoshi, 2006). *V. parahaemolyticus* also causes wound-infection through exposure of a new wound to contaminated seawater or estuarine water, whereas this type of diseases is independent of production of the hemolysin.

The first clinical isolation of *V. vulnificus* was from a human leg ulcer (Roland, 1970). However, because of similar bacteriological characteristics of the bacterium isolated, this case was reported as an extra-intestinal infectious disease by *V. parahaemolyticus* (Hollis et al., 1976). *V. vulnificus* causes two types of infections, the primary septicemia and the wound-infection,



"fmicb-04-00339" — 2013/11/14 — 21:12 — page 2 — #2

++*, Major disease;* +*, Minor disease;* ±*, Rare disease.*

*\*Include otitis media, cholesystitis, meningitis*

to human (Blake et al., 1980; Klontz et al., 1988). The primary septicemia is associated with consumption of contaminated raw seafood, especially shellfish such as oysters. However, this type is a typical opportunistic infection. Namely, most patients have an underlying disease(s) of liver dysfunction, alcoholic cirrhosis or hemochromatosis, which leads to increase in the plasma ferric ion level and to decrease in the activity of the innate immunity (Strom and Paranjpye, 2000; Inoue et al., 2008). In two-thirds of the patients, the edematous and/or hemorrhagic secondary skin lesions appear on the extremities and the trunk (Blake et al., 1980; Inoue et al., 2008). The wound-infection is characterized by the development of edema, erythema or necrosis around a new wound exposed to seawater contaminated with *V. vulnificus* (Blake et al., 1980; Klontz et al., 1988). This type of infection can occur in healthy persons, as well as in the immuno-compromised hosts, and may occasionally progress to septicemia.

#### **BACTERIAL EXTRACELLULAR PROTEOLYTIC ENZYMES**

Human pathogenic vibrios produce various extracellular factors including enterotoxin, hemolysin, cytotoxin, protease, collagenase, phospholipase, siderophore, and hemagglutinin (Janda et al., 1988). Of these factors, enterotoxin, hemolysin, and cytotoxin are directly related to the clinical symptoms; however, siderophore and hemagglutinin may play roles in the establishment of the infection.

Proteolytic enzymes hydrolyzing a peptide bond in proteins and peptides are essential for the homeostatic control in both eukaryotes and prokaryotes, and thus, the bacterial enzymes also have various physiological roles in the life cycle of the microorganisms. However, the enzymes produced by pathogenic species occasionally act as toxic factors to the infected host (Hase and Finkelstein, 1993; Harrington, 1996). Proteolytic enzymes are classified into several groups, such as aspartic, cysteine, and serine protease, and metalloprotease; however, many of the bacterial toxic proteases are in the metalloprotease group, which often contains a zinc (II) ion in the catalytic center (Hase and Finkelstein, 1993; Hooper, 1994). As shown in **Table 2**, human pathogenic *Vibrio* species also produce and secrete proteolytic enzymes, and several enzymes have been extensively characterized as direct toxic factors causing skin damage or indirect virulence factors processing other protein toxins. The enzymes produced by vibrios are in two metalloprotease groups (vibriolysin and collagenases) or one serine protease group (chymotrypsin-like proteases).

The progress in the molecular biology has provided much information on the DNA-derived amino acid sequences of metalloproteases and has shown the presence of the consensus sequence HEXXH as the zinc-binding motif. This motif was also found in some bacterial protein toxins including clostridial neurotoxins,



*Bacteroides fragilis* enterotoxin, and *Bacillus anthracis* lethal factor. Indeed, these bacterial toxins were verified to show the remarkably specific proteolytic action toward a target host protein (Miyoshi and Shinoda, 2000). For instance, clostridial neurotoxins can cleave the restricted protein components of the neuroexocytosis machinery, which leads to the blockade of neurotransmitter release and consequent muscle paralysis (Schiavo et al., 1993). In addition, a novel cytotoxin that consists of one A subunit and five B subunits was isolated from some enterohemorrhagic *Escherichia coli* strains, and the A subunit was indicated to be a subtilase-like serine protease (Paton and Paton, 2010). The RTX (repeated-intoxin) toxins are large multifunctional cytotoxins and are possible to modulate the virulence of a number of gram-negative bacterial pathogens including *V. cholerae* and *V. vulnificus* (Satchell, 2007; Prochazkova et al., 2009; Roig et al., 2011). In *V. cholerae* RTX toxin, the cysteine protease domain was reported to mediate autoprocessing of the toxin (Sheahan et al., 2007; Shen et al., 2009).

#### **VIBRIOLYSIN**

#### **BIOCHEMICAL PROPERTIES**

Zinc-containing metalloproteaes consist of four superfamilies based on the amino acid residues in the zinc-binding site, and the zincins superfamily is characterized by the possessing of the HEXXH motif (Hooper, 1994). The thermolysin family, in which prototype enzyme is thermolysin from *Bacillus thermoproteolyticus*, is a one of major members of this superfamily. The enzymes in this family are commonly characterized by the presence of Glu at the 25th position from the first His of the above motif. *V. proteolyticus* is a marine microorganism that was first isolated from the intestine of a small, wood-boring isopod crustacean (Merkel et al., 1964). Griffin and Prescott (1970) purified a highly active metalloprotease from the bacterial culture supernatant. Durham (1989) first disclosed the designation of this enzyme as vibriolysin in the patent literature. Thereafter, highly homologous metalloproteases have been isolated from other *Vibrio* species including human pathogens *V. cholerae*, *V. fluvialis*, *V. mimicus*, and *V. vulnificus*. Therefore, the name of vibriolysin is currently applicable to all of these proteolytic enzymes (Miyoshi et al., 2012a).

Vibriolysin hydrolyzes specifically the peptide bond at the amino group side of the P1 amino acid residue, which is usually a hydrophobic amino acid reside (e.g., Phe, Tyr, or Leu; Narukawa et al., 1993). Synthetic oligopeptides, such as carbobenzoxy (Z)-Gly-Phe-NH2 and Z-Gly-Leu-NH2, are thus commonly used as the suitable substrate. On the other hand, phosphoramidon [*N*-(α-rhamnopyranosyloxy-hydroxyphosphanyl)Leu-Trp] and zincov [2-(*N*-hydroxycarboxamido)-4-methylpentanoyl-Ala-Gly amide] are well-known competitive inhibitors. In addition, phenylazobenzyloxycarbonyl-Pro-Leu-Gly-Pro-Arg, the substrate developed for bacterial collagenases, is significantly hydrolyzed by the enzyme (Miyoshi et al., 1987b). Vibriolysin is also highly active on a wide variety of protein substrates. Namely, the enzyme exhibits significant proteolysis of casein, albumin, hemoglobin, type I and IV collagen, gelatin, elastin, fibrin, and fibrinogen (Miyoshi et al., 1995; Miyoshi and Shinoda, 2000).

Like thermolysin, vibriolysin is synthesized as an inactive precursor, and maturation of the precursor is achieved by several processing stages (Chang et al., 2007). In the case of the enzyme from *V. vulnificus* (**Figure 1**), it is initially synthesized as the preproenzyme (609 aa, 65,964 Da) with a typical signal peptide (Cheng et al.,1996). The signal peptide is cleaved during its passage through the inner membrane in the signal peptidase-dependent manner. In the periplasm, the propeptide that may function as an intramolecular chaperone mediating maturation of the enzyme and/or a specific inhibitor to protect autodigestion of the enzyme (Chang et al., 2007) is then cleaved by an autocatalytic mechanism, and finally, the mature vibriolysin (413 aa, 44,648 Da) is generated. The vibriolysin maturated consists of two functional domains: the N-terminal domain (314 aa, 34,049 Da) mediating the catalytic action, and the C-terminal domain (99 aa, 10,656 Da) essential for efficient attachment to protein substrates (Yun et al., 2012). The N-terminal domain is easily obtained by autocatalytic limiteddigestion of the C-terminal domain (Miyoshi et al., 1997). The N-terminal domain alone possesses sufficient proteolytic activity toward oligopeptides or soluble proteins, while it shows markedly reduced activity toward insoluble proteins such as type I collagen and elastin.

The N-terminal domain of vibriolysin is significantly related to other enzymes in the thermolysin family; however, its C-terminal domain may be unique. The feature of the proteolytic action, which is mediated by the N-terminal domain, is considerably similar to other enzymes. However, neither suitable specific peptide substrate nor competitive inhibitor for vibriolysin has been developed. On the other hand, the hemagglutinating action on rabbit erythrocytes, which is due to association of both N-terminal and C-terminal domains with the erythrocyte membrane, is a distinctive feature of vibriolysin so far as reported (Alam et al., 1995).

#### **PATHOLOGICAL ROLES**

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 3 — #3

As summarized in **Figure 2**, a variety of pathological roles of vibriolysin have been documented (Shinoda and Miyoshi, 2011). In the local infections such as the wound-infection, vibriolysin is thought to be a direct toxic factor that causes hemorrhagic tissue

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 4 — #4

damage through digestion of the basement membrane around vascular endothelial cells, and that forms edematous lesions through generation of inflammatory mediators (**Figure 1**). The enzyme from *V. vulnificus* can enhance the vascular permeability when injected into the mammalian dorsal skin. In rat skin, this reaction is due to the release of histamine from mast cells, because the vascular enhancement was abolished by simultaneous injection of diphenhydramine, an anti-histaminic agent (Miyoshi et al., 1987c). In guinea pig skin, however, the permeability enhancement is most likely due to activation of the factor XII-plasma kallikrein-kinin cascade (Miyoshi et al., 1987a). Namely, the skin reaction was not blocked by diphenhydramine, but it was modulated by *in situ* administration of the specific inhibitors affecting the cascade activation. Bradykinin, a well-known *in vivo* mediator of inflammation, is finally generated through activation of the cascade. Further studies to clarify the activation mechanism of the human cascade have been carried out, and the results demonstrated that vibriolysin could generate the active forms via limited proteolysis of the inactive zymogens (Miyoshi et al., 2004). Plasma prekallikrein was converted to the active kallikrein, which can liberate bradykinin from kininogen, by cleavage of the Arg371-Ile372 bond. On the other hand, factor XII was activated by hydrolysis of the Arg353-Val354 or Gly357-Leu358 bond, and activated factor XII could convert plasma prekallikrein to kallikrein. Vibriolysin also induces the hemorrhagic reaction in the mammalian dorsal skin. The enzyme from *V. vulnificus* showed the greatest hemorrhagic activity compared with some bacterial metalloproteases, thermolysin from *B. thermoproteolyticus*, serralysin from *Serratia* species, and collagenase from *Clostridium histolyticum* (Miyoshi et al., 1998). The levels of the *in vivo* hemorrhagic activities of

these proteases were correlated with those of the *in vitro* proteolytic activities toward the reconstituted basement membrane gel. Of two major basement components, laminin and type IV collagen, only the latter was easily digested by vibriolysin. This indicates that type IV collagen forming the framework of the basement membrane is the target protein. Therefore, specific degradation of type IV collagen causes destruction of the basement membrane, breakdown of capillary vessels, and finally the leakage of blood components including erythrocytes.

In the systemic infections including septicemia, vibriolysin may act as a synergistic pathogenic factor through disordered proteolysis of various plasma proteins, which in turn disturbs the physiological homeostasis, and eventually elicits an immunocompromised state in the infected host. For instances, vibriolysin has been documented to facilitate the bacterial infection by disturbance of the plasma protease-protease inhibitor systems (Miyoshi et al., 1995), and to interfere with the blood homeostasis through prothrombin activation and fibrinolysis (Chang et al., 2005; Kwon et al., 2007). Acceleration of the heme utilization was also reported as the pathogenic role of vibriolysin (Nishina et al., 1992). Ferric/ferrous ion is essential for *in vivo* growth of pathogenic microorganisms. However, the concentration of free ferric/ferrous ion in human body is very low (10−<sup>15</sup> to 10−<sup>18</sup> M) because of the presence of heme-proteins including hemoglobin and iron-binding proteins. Therefore, pathogenic microorganisms invaded into human body must operate the systems to acquire ferric/ferrous ion. Heme, a complex of porphyrin with ferric/ferrous ion and a prosthetic group in hemoglobin or other heme-proteins, is often used as an iron source by human pathogenic bacterial species (Stojiljkovic and Perkins-Balding, 2002; Tong and Guo, 2009). Nishina et al. (1992) documented that the vibriolysin-deficient mutants of *V. vulnificus* could not grow in the iron-restricted broth because of the inability to utilize hemoglobin as an iron source, but the bacterial growth was apparently restored by the addition of the purified vibriolysin. On the other hand, hemoglobin or ferric ion is known to be required for efficient transcription of the gene encoding the vibriolysin (Kawase et al., 2004; Sun et al., 2006). Incidentally, it should be noted that human plasma contains a broad-range protease inhibitor α2-macroglobulin (α<sup>2</sup> M) as a primary inhibitor for exogenous proteolytic enzymes including bacterial enzymes. Vibriolysin was also inactivated with α<sup>2</sup> M at a molar ratio of 1:1 by means of physical entrapment (Miyoshi and Shinoda, 1989). Therefore, the pathological actions of vibriolysin documented herein are possible to support the systemic infection but may be highly restricted *in vivo*.

Vibriolysin from enteric pathogens may increase the bacterial attachment to the intestinal surface through digestion of the mucosal substances or the bacterial outer membrane proteins (Alam et al., 2006). However, the inverse effect of the enzyme was also known. The vibriolysin from *V. cholerae* O1 was reported to accelerate the bacterial detachment from cultured cells by digestion of *V. cholerae* adhesins (Finkelstein et al., 1992). Vibriolysin also modulate the enterotoxicity of the bacterial toxins by limited proteolysis. For instance, the enzyme from *V. cholerae* O1 can activate CT through nicking of the A subunit of CT (Booth et al., 1984). Vibriolysin also converts the precursor of the enterotoxic hemolysin produced by *V. cholerae* to the mature toxin through removal of the 15 kDa N-terminal propeptide (Nagamune et al., 1996). Although the results shown above suggest indirect pathogenic roles of vibriolysin, the possibility of the direct roles has also been reported. Ghosh et al. (2006) purified vibriolysin from a CT gene-negative strain of *V. cholerae* non-O1/non-O139 and measured the enterotoxic activity. The purified enzyme caused accumulation of the hemorrhagic fluid in the rabbit ileal loop assay and increase in the intestinal short-circuit current in the Using chamber assay. Additionally, through the analysis with several mutants genetically constructed, Silva et al. (2006) showed that vibriolysin was necessary for full expression of enterotoxicity of *V. cholerae* O1.

#### **COLLAGENASES**

Yu and Lee (1999) and Kim et al. (2002) carried out cloning of the *V. parahaemolyticus* gene encoding a collagenase, which was designated as PrtV/PrtVp (562 aa, 63,156 Da) and VppC (814 aa, 89,833 Da), respectively. These enzymes were revealed to be metalloproteases in the zincins superfamily having the consensus zinc-binding HEXXH motif, but neither of the collagenases was in the thermolysin family. Miyoshi et al. (2008) showed that, only when *V. parahaemolyticus* was cultivated at 26◦C, the *vppC* gene was sufficiently expressed, and VppC was secreted from the bacterial cell after removal of the N-terminal 72 amino acid residues. In contrast, expression of the *prtV/prtVp* gene was negligible in the wild type strain. The gene was significantly expressed by disruption of the *vppC* gene; however, the product PrtV/PrtVp was not secreted into the cultivation broth (Miyoshi et al., 2008), suggesting PrtV/PrtVp is a cell-associated enzyme. VppC purified showed the steady activity to hydrolyze Z-Gly-Pro-Gly-Gly-Pro-Ala, the specific substrate for bacterial collagenases, and to digest gelatin. This indicates that VppC may contribute to the woundinfection by *V. parahaemolyticus* because putative digestion of the components of the extracellular matrix by VppC may accelerate the bacterial dissemination and may form cellulitic skin damage. *V. alginolyticus*, another species causing the woundinfection, is known to produce a VppC homolog (Takeuchi et al., 1992).

#### **CHYMOTRYPSIN-LIKE PROTEASES**

In 2002, two research groups reported individually purification of a serine protease, which was termed protease A and VPP1 respectively, from the culture supernatant of *V. parahaemolyticus* (Ishihara et al., 2002; Lee et al., 2002). These proteases were identical and corresponded to the VPA0227 protein (677 aa, 71,038 Da) of strain RIMD 2210633 (GenBank accession number: BA000032). However, the amino acid sequencing of the purified enzyme indicated that the N-terminal 121 amino acid residues were removed during the maturating process. This serine protease designated herein as protease A/VPP1 showed the immunological cross-reactivity with serine proteases from *V. metschnikovii* and *V. alginolyticus* (Ishihara et al., 2002). Indeed, protease A/VPP1 revealed highly similarity of the amino acid sequence to the enzyme from *V. metschnikovii* (Kwon et al., 1995). Production of protease A/VPP1 was much higher at 25 than 37◦C and was induced by the addition of gelatin to the cultivation broth. Miyoshi et al. (2008) reported that production of this serine protease was remarkably increased by disruption of the *vppC* gene, suggesting that protease A/VPP1 is a substitutive enzyme of VppC. The purified protease was found to be sensitive to chymostatin, the well-known competitive inhibitor for chymotrypsin, and to hydrolyze the specific peptidyl-4-methyl-coumaryl-7-amide (MCA) substrates for chymotripsin, such as Glutaryl-Ala-Ala-Phe-MCA and Succinyl-Ala-Ala-Pro-Phe-MCA. Therefore, it may be concluded that protease A/VPP1 is a chymotrypsin-like serine protease. The cytotoxicity against CHO, HeLa, or Vero cells and the mouse lethal toxicity of purified protease A/VPP1 was demonstrated by Lee et al. (2002). Our preliminary study showed the proteolytic activity of the purified enzyme toward extracellular matrix components, laminin and type I collagen. These results suggest that protease A/VPP1 also modulate the bacterial pathogenicity.

*Vibrio vulnificus* sometime causes severe hemorrhagic septicemia called vibriosis in eels of the culture farms (Tison et al., 1982; Biosca et al., 1991). A few strains isolated from diseases eels were recently found to have lost the 80 kb genomic region including the gene encoding vibriolysin, but instead of vibriolysin, these strains secrete a serine protease termed VvsA, which is an ortholog of protease A/VPP1 from *V. parahaemolyticus* (Miyoshi et al., 2012b). As protease A/VPP1, production of VvsA was extremely increased in the absence of the functional gene encoding vibriolysin (Wang et al., 2008). The *vvsA* gene constitutes an operon with a downstream gene *vvsB*, of which product VvsB may act as the chaperon supporting the maturation process of VvsA. The database analysis showed that several *Vbrio* species including *V. parahaemolyticus* have the homologus genes to *vvsAB*, indicating

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 5 — #5

widely distribution of the chymotrypsin-like serine protease in *Vibrio* species.

#### **REGULATION OF PRODUCTION OF PROTEOLYTIC ENZYMES**

*Vibrio* species are ubiquitous microorganisms in aquatic environments, but 11 species cause intestinal or extra-intestinal infections to humans. During the infection process, the bacterial cells must sense the change of environmental factors, such as temperature, pH, salinity, and osmolarity, and then, the bacterium must transmit the signals into the cells through the specific signaltransduction systems, which result in the change of expression of the genes. Especially, the genes encoding the toxic or virulence factors, which may be required for *in vivo* survival and growth, must be expressed at an appropriate place and time in a tightly regulated fashion (Heithoff et al., 1997; Lee et al., 1999). Amongst the environmental factors, temperature is thought to be the most important. The vibriolysin genes of the bacterium causing intestinal infections are expressed sufficiently at 37◦C, while the genes of the bacterium causing extra-intestinal infections are expressed more effectively around the body surface temperature (Watanabe et al., 2004; Miyoshi et al., 2006). The gene expression is also often affected by the salinity. For instance, the expression level of the *vppC* gene in *V. parahaemolyticus* is higher at 3% NaCl than 0.9% NaCl (Miyoshi et al., 2008). Although the intracellular second messengers, such as cyclic AMP and cyclic di-GMP, and the global general regulators including RpoS and histone-like nucleotide structuring protein are also involved in the gene regulation (Benitez et al., 2001; Wang et al., 2011, 2012), the molecular mechanisms how the bacterium senses the environmental signals and transmits the signals into the cells are not clarified.

Production of proteolytic enzymes is tightly dependent on the growth phase and reaches to the maximum level at the early stationary phase. Many pathogenic bacteria coordinate expression of the virulence genes in response to the bacterial cell density. This regulation system is termed the quorum sensing (QS) and is controlled by the small diffusible signal molecule called autoinducer (AI; Whitehead et al., 2001; Henke and Bassler, 2004). At the low cell density, the QS system cannot modulate the gene expression because the concentration of AI is too low. However, at the high cell density, the AI concentration around the bacterial cell reaches the threshold level, the AI molecule is sensed by the sensor protein, the signals are transmitted into the cell, and finally, the expression of a set of genes is started or stopped.

In *Vibrio* species including *V. cholerae*, *V. mimicus*, and *V. vulnificus* (Miller et al., 2002; Milton, 2006; Sultan et al., 2006), the AI molecule is detected by the specific membrane-bound sensor protein, which causes conversion of the sensor protein from kinase to phosphatase. Subsequently, the sensor protein/phosphatase mediates dephosphorylation of LuxU-LuxO, the response regulator proteins. The dephosphorylated LuxO has no activity to inhibit LuxR or its homolog, the master transcriptional regulator for the genes under the control of the QS system. Therefore, at the high cell density, the transcriptional status of the target genes is changed by the function of LuxR or its homologue. Production of the proteolytic enzymes by pathogenic vibrios is closely related with the extracellular AI level (Kim et al., 2003; Kawase et al., 2004; Raychaudhuri et al., 2006). For instance, the mutant of the AI synthetase showed apparently reduced production of vibriolysin. Therefore, the QS system markedly controls the expression of the proteolytic genes in *Vibrio* species. However, it should be noted that the QS system of *V. cholerae* or *V. mimicus* is operated sufficiently at 37◦C (Sultan et al., 2006), whereas, the system of *V. vulnificus* is operated effectively at 26◦C but not at 37◦C (Miyoshi et al., 2006).

#### **CONCLUSION**

The proteolytic enzymes produced by human pathogenic *Vibrio* species may play a variety of pathological roles: direct roles by digesting many kinds of host proteins or indirect roles by processing other toxic protein factors. Especially, vibriolysin from *V. vulnificus* is thought to be a major virulence factor. However, some contradictions of the pathogenic roles were also reported (Shao and Hor, 2000; Sun et al., 2006). It must be mentioned that the purified enzymes from *V. vulnificus* and *V. proteolyticus*, a non-pathogenic species, are difficult to distinguish in the *in vivo* actions, because both enzymes are members of vibriolysin and have comparative biochemical and toxic activities. However, it has been demonstrated that the high growing ability of *V. vulnificus* in the mammal host is important for the pathogenicity of the bacterium (Watanabe et al., 2004). In addition, production of the toxic or virulence factors including proteolytic enzymes is tightly regulated by environmental factors, the bacterial cell density and so on. Therefore, the overall experiments from various approaches are necessary for evaluation of the extracellular proteolytic enzymes as the virulence factors.

#### **REFERENCES**

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 6 — #6


"fmicb-04-00339" — 2013/11/14 — 21:12 — page 7 — #7


**Conflict of Interest Statement:** The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 28 August 2013; paper pending published: 03 October 2013; accepted: 26 October 2013; published online: 18 November 2013.*

*Citation: Miyoshi S (2013) Extracellular proteolytic enzymes produced by human pathogenic Vibrio species. Front. Microbiol. 4:339. doi: 10.3389/fmicb.2013.00339*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Miyoshi. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00339" — 2013/11/14 — 21:12 — page 8 — #8

## *Vibrio cholerae* as a predator: lessons from evolutionary principles

### *Stefan Pukatzki 1\* and Daniele Provenzano 2 ,3 \**

*<sup>1</sup> Department of Medical Microbiology and Immunology, University of Alberta, Edmonton, AB, Canada*

*<sup>2</sup> Department of Biomedical Sciences, University of Texas Brownsville, Brownsville, TX, USA*

*<sup>3</sup> Department of Biological Sciences, University of Texas Brownsville, Brownsville, TX, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Daniela Ceccarelli, University of Maryland, USA Aoife Boyd, National University of Ireland Galway, Ireland*

#### *\*Correspondence:*

*Stefan Pukatzki, Department of Medical Microbiology and Immunology, University of Alberta, 6 – 22 Heritage Medical Research Center, T6G 2S2 Edmonton, AB, Canada e-mail: spukatzki@ualberta.ca; Daniele Provenzano, Departments of Biomedical Sciences and Biological Sciences, University of Texas Brownsville, One West University Boulevard, Brownsville, TX 78520, USA*

*e-mail: Daniele.Provenzano@utb.edu*

Diarrheal diseases are the second-most common cause of death among children under the age of five worldwide. Cholera alone, caused by the marine bacterium *Vibrio cholerae*, is responsible for several million cases and over 120,000 deaths annually.When contaminated water is ingested, *V. cholerae* passes through the gastric acid barrier, penetrates the mucin layer of the small intestine, and adheres to the underlying epithelial lining. *V. cholerae* multiplies rapidly, secretes cholera toxin, and exits the human host in vast numbers during diarrheal purges. How *V. cholerae* rapidly reaches such high numbers during each purge is not clearly understood. We propose that *V. cholerae* employs its bactericidal type VI secretion system to engage in intraspecies and intraguild predation for nutrient acquisition to support rapid growth and multiplication.

**Keywords:***Vibrio cholerae***, intraguild predation, cholera, nutrient acquisition, typeVI secretion system, microbiome modification**

"fmicb-04-00384" — 2013/12/9 — 20:05 — page 1 — #1

*Vibrio cholerae* – the causative agent of the diarrheal disease cholera – is an ancient companion of human civilization. Reports of cholera symptoms date back to ancient Greek civilization and Sanskrit writings (Barua, 1992). Records of the disease describing seven cholera pandemics have been maintained since the 19th century (Blake, 1994). Although the discovery is often attributed to Koch (1884), the Italian anatomist Filippo Pacini first identified *V. cholerae* in 1854 as the causative agent of cholera (Bentivoglio and Pacini, 1995). Advances in molecular biology during the late 20th century greatly improved our understanding of *V. cholerae* pathogenesis. The discovery of the ToxR regulon (Miller and Mekalanos, 1984) and characterization of the principal virulence factors cholera toxin (CT) and toxin co-regulated pilus (TCP) shed light on molecular mechanisms that mediate pandemic cholera spread (Herrington et al., 1988). TCP biosynthesis genes are encoded within a horizontally transferred mobile genetic element known as Virulence Pathogenicity Island I (VPI-I) that in addition to TCP and other accessory genes encodes several transcriptional activators and virulence factors required for pandemic spread of *V. cholerae* (Karaolis et al., 1998). TCP functions as a receptor for the CTX-ϕ filamentous bacteriophage that transduces CT genes into the chromosomes of pandemic *V. cholerae* strains. The contribution of CTX-ϕ to bacterial pathogenesis brought to light a remarkable example of mutualism between a bacteriophage and a bacterial pathogen, because only*V. cholerae* strains encoding CT are capable of pandemic spread (Waldor andMekalanos,1996).

Cholera toxin and additional diarrheagenic factors instruct the epithelial lining of the gut to secrete electrolytes and water that constitute the diarrhea characteristic of cholera (Sharp and Hynie, 1971). However, what nutrient source permits *V. cholerae* to reach bacterial counts reaching up to 10<sup>9</sup> CFUs/mL during acute cholera is unclear. Digested food is likely not a main source of nutrients for the bacteria, because cholera patients frequently empty their stomachs through vomiting (Chatterjee, 1953) and disease symptoms are often more severe in patients suffering from malnutrition (Dutta et al., 1991). A probable source of energy for *V. cholerae*, especially during the early phases of colonization, is the mucus layer that coats the entire length of the human gastrointestinal tract. In support of this hypothesis, *V. cholerae* secretes the metallo-protease TagA to specifically cleave mucin glycoproteins and select cell-surface glycans (Szabady et al., 2011). The ToxR regulon induces expression of VPI-encoded *tagA in vivo* (Bina et al., 2003; Withey and Dirita, 2005). Sialic acids of mucous membranes are another likely nutrient source for *V. cholerae*. The Virulence Pathogenicity Island 2 (VPI-2), a horizontally acquired DNA element, contains several genes for sialic acid transport and catabolism (Chowdhury et al., 2012). Furthermore, HA-protease, a broad-spectrum protease with mucinase, neuraminidase, and additional enzymatic activities, is important for *V. cholerae* dissemination (Häse and Finkelstein, 1991; Finkelstein et al., 1992). Taken together, these findings support the hypothesis that saccharides of mucin and the glycocalyx on the surface of

epithelial cells provide carbon and nitrogen as an energy source for *V. cholerae* growth and multiplication during early stages of infection. Thus, mucin-supported growth may be important particularly in the small intestine – the host environment colonized by *V. cholerae*. However, whether host mucins are sufficient to sustain a rapidly multiplying biomass of *V. cholerae* that quickly replenishes between rice-water purges consisting in part of bacteria associated with mucus plugs (Nelson et al., 2007) remains to be demonstrated. We propose that *V. cholerae* engages in predatory behavior to supplement a mucin diet with nutrients acquired from lysed bacteria.

In the environment, *V. cholerae* extracts energy from chitin through what has been collectively coined as the chitin utilization program (Meibom et al., 2004). Remarkably, chitin induces cell density-dependent natural competence in *V. cholerae*. Cholera bacteria switch from degrading exogenous nucleic acids for acquisition into their own pool of nucleotides at low cell density to taking up intact DNA at higher cell densities leading to transformation (Blokesch and Schoolnik, 2008). Chitin therefore plays a critical role in the environmental life style of *V. cholerae* as a source of nutrients and as a signal for uptake of either nucleotides or whole DNA sequences. *V. cholerae*'s ability to bind chitin is mediated by a secreted *N*-acetylglucosamine binding protein (GbpA) that also functions as the adhesin for attachment to mucin in the host intestines (Kirn et al., 2005; Wong et al., 2012). Considering the overlapping roles of mucin and chitin in *V. cholerae* adherence, colonization and nutrient acquisition, this analogy appears to become an inescapable theme. Perhaps *V. cholerae's* ancestral role as a chitin degrader in marine environments is the reason for the ability of the cholera bacterium to bind mucin in the human intestine.

Chitin binding likely preceded cholera bacteria's ability to bind mucin in the small intestine because all strains are capable of persisting in the environment, but not all *V. cholerae* are host-colonization competent. Analogously, bacterial predation likely evolved before *V. cholerae* became intestinal colonizers because single cells competed against each other in environmental reservoirs prior to emergence of multicellularity. Therefore, acquisition of DNA coding instructions for the biosynthesis and delivery of bactericidal effector molecules followed by natural selection may have provided predatory cells a formidable competitive advantage in any given niche. However, effective predation requires mechanisms to distinguish prey from kin as is the case for *Escherichia coli*, a bacterial species adapted to reside in vertebrate intestines. Coligenic *Escherichia coli* secrete episomally encoded bactericidal proteins called colicins (Jacob et al., 1952) that kill related bacteria, but exclude kin bacteria protected by means of cognate immunity proteins (Chak and James, 1984). Lysed neighboring prey, which may include non-coligenic *Escherichia coli* as well as coligenic bacteria expressing different immunity proteins, can become a nutrient source for predatory colicin-producing *Escherichia coli* (Leisner and Haaber, 2012). However, the best characterized examples of nutrient acquisition as a direct result of predation are found among the deltaproteobacteria: *Bdellovibrionaceae* are diminutive aquatic dwellers that penetrate the periplasm of larger Gramnegative prey where they utilize macromolecules as a source of nutrients and divide until they exit the host to repeat the cycle (Stolp and Starr, 1963). Another example is Myxobacteria found in most terrestrial soils and aquatic environments that subsist from nutrients acquired from both bacterial and fungal prey. In contrast to *Bdellovibrionaceae*, Myxobacteria display cooperative behavior through group swarming and predation (Berleman and Kirby, 2009) which culminates in the biosynthesis of antibiotics and enzymes (Rosenberg et al., 1977) that kill and convert a wide range of prey cells into growth substrates utilized by the predators for growth and multiplication (Morgan et al., 2010). Perhaps not coincidentally *Myxococcus xanthus* was found to harbor an intact type VI secretion system (T6SS; Bingle et al., 2008).

Interactions between different species can broadly be divided into the following categories: neutralism, competition, predation/parasitism, mutualism, commensalism, and amensalism. Intraguild predation is a combination of competition and predation that occurs when neighboring species competing for limited resources within the same niche assume predatory behavior toward each other to acquire nutrients (Polis et al., 1989). Intraguild predation differs from classical predation because it reduces competition and also differs from traditional competition because the predator directly benefits energetically from the demise of organisms that compete for the same resource pool (Polis et al., 1989). The potential benefits of intraguild predation in intestinal colonization are evidenced by the observation that several bacterial species (Cheng et al., 1973) including *Escherichia coli* (Holme and Cedergren, 1961) and *V. cholerae* (Bourassa and Camilli, 2009) store glycogen in cytoplasmic inclusions during their tenure in the host. In addition to removing competitors from the colonization niche, predation would provide immediate energetic benefits from lysed bacteria to predators like *V. cholerae* upon entry into and before exiting the human host.

*Vibrio cholerae* likely engages in predatory behavior in the marine environment where it spends considerable time of its lifecycle. This would require a genetically conserved, ancestral bactericidal mechanism capable of efficiently discriminating kin cells from prey bacteria. Yet, *V. cholerae* is a highly diverse species as each genome sequenced appears to consist of a patchwork of horizontally acquired genetic elements. Any given *V. cholerae* strain harbors approximately 3,700 genes, while the *V. cholerae* pan-genome consists of ∼6,500 genes (Vesth et al., 2010). In spite of the occasional encounter and explosive population expansion in human hosts, *V. cholerae* is primarily a marine inhabitant; its genetic diversity may reflect strain-specific adaptation to specific niches. All pandemic strains harbor CTX-ϕ and VPI-I (Dziejman et al., 2002); however, several non-pandemic strains are also capable of causing cholera. For example, some environmental strains have been found to code a type III secretion system (T3SS) shown to be a potent diarrheagenic factor in the infant mouse model (Dziejman et al., 2005). Regardless, all strains sequenced thus far harbor the genes for a highly conserved T6SS. The presence of T6SS genes in ∼25% of all sequenced Proteobacteria (Bingle et al., 2008) suggests an ancestral origin of the secretion system that may have preceded the evolutionary divergence of *Vibrio* lineages.

"fmicb-04-00384" — 2013/12/9 — 20:05 — page 2 — #2

To engage in predation (MacIntyre et al., 2010; Unterweger et al., 2012), *V. cholerae* employs a T6SS consisting of a polymeric protein nanotube bearing remarkable structural and functional similarity to the bacteriophage tail complex (Leiman et al., 2009). This highly dynamic tube is rapidly assembled in the cytosol and acts as a scaffold for an outer sheath that upon contraction ejects the inner tube and delivers a puncturing device on the tip into adjacent cells (Basler and Mekalanos, 2012). The discovery of T6SS in *V. cholerae* was originally linked to pathogenesis because an effector that crosslinks actin causes toxicity in *Dictyostelium discoideum* (Pukatzki et al., 2007). Recently, three T6SS effectors with homology to lipases (Dong et al., 2013), pore-forming proteins (Miyata et al., 2013), and peptidoglycan-degrading enzymes (Brooks et al., 2013; Dong et al., 2013) were found to exclusively target bacteria. Only a fraction of *V. cholerae* strains examined to date express the T6SS under laboratory conditions; yet, the secretion system has been linked to intestinal inflammation in the infant mouse (Ma and Mekalanos, 2010) and infant rabbit (Zheng et al., 2010) cholera models. Furthermore, an *in vivo* expression technology (IVET) screen employing non-toxigenic *V. cholerae* strains carried out in human volunteers suggested that T6SS genes are transcribed in the human host (Lombardo et al., 2007). These findings are supported by RNA-Seq experiments that show T6SS gene transcription in mouse intestines employing a toxigenic *V. cholerae* strain that does not express the secretion apparatus under laboratory conditions (Mandlik et al., 2011). Collectively these reports provide abundant support for a role of *V. cholerae* T6SS expression *in vivo*.

Evidence that T6SS genes are also likely expressed in the environment emerged from a report showing that salt concentrations and temperatures analogous to those encountered by *V. cholerae* in the marine environment induce T6SS biosynthesis (Ishikawa et al., 2012). These experiments employed toxigenic *V. cholerae* strains that otherwise do not express the secretion apparatus under laboratory conditions. In addition, we recently showed that nontoxigenic *V. cholerae* strains isolated from the Rio Grande Delta express their T6SS constitutively (Unterweger et al., 2012) supporting a role for the secretion system in the marine life style of this species. Perhaps more interestingly, this strain collection allowed us to test the hypothesis that *V. cholerae*'s T6SS immunity system (Dong et al., 2013; Miyata et al., 2013) mediates resistance to strain-specific T6SS-mediated toxicity. In summary, our results showed that the extent of killing between unrelated strains differs leading to speculations that distinct toxin/antitoxin sets are harbored by different *V. cholerae* and that T6SS-mediated selective intraspecies killing permits strains to distinguish self from non-self (Unterweger et al., 2012).

*Vibrio cholerae* does not utilize the T6SS exclusively to kill bacteria of its own species; several additional prokaryotes are susceptible to T6SS-dependent killing. Some of these susceptible bacteria include*V. communis*,*V. harveyi*,*V. mimicus*,*V. fischeri*, and *Pseudoalteromonas phenolica* isolated from the same environmental niches as *V. cholerae* (Unterweger et al., 2012). Other bacteria susceptible to T6SS-mediated killing are less likely to encounter *V. cholerae* in the marine environment; these include *Escherichia coli*, *Salmonella typhimurium,* and *Citrobacter rodentium*. Thus far, only Gram-negative bacteria have demonstrated susceptibility to *V. cholerae* T6SS killing, while *Pseudomonas aeruginosa* was recently shown to be capable of counter attacking cholera bacteria utilizing its own T6SS (Basler et al., 2013). Gram-positive bacteria examined thus far such as *Enterococcus faecalis, Bacillus subtilis, Listeria monocytogenes*, and *Staphylococcus aureus* are resistant to *V. cholerae* T6SS bactericidal activity (MacIntyre et al., 2010). One possible explanation is that Gram-positive bacteria are naturally refractory to *V. cholerae*'s T6SS's bactericidal effect; alternatively we cannot rule out the possibility that additional effectors targeting Gram-positive species remain to be discovered and characterized. Either way, susceptibility of the host microbiome to T6SS-mediated killing may be one of numerous factors contributing to *V. cholerae*'s ability to successfully colonize the host.

The host microbiome experiences profound changes during cholera infections. Commensal bacteria belonging to the *Bacteroides*, *Firmicutes,* and *Actinobacteria* phyla are displaced while harmful *Proteobacteria* prosper (Monira et al., 2013). Benign *Escherichia coli* K12 capable of colonizing the human small intestine implanted at high inocula failed to maintain a foothold in the intestines of volunteers suffering from acute cholera (Gorbach et al., 1970). This outcome may be, in part, due to K-12's high susceptibility to *V. cholerae* T6SS killing (MacIntyre et al., 2010). While shifts in bacterial populations and modification to the host microbiome during cholera infections are likely attributed to their displacement by water and electrolyte efflux from host cells in the intestine as a result of fulminant diarrhea, we propose that T6SSmediated predation contributes to these shifts. We hypothesize that *V. cholerae* engage their T6SS in intraspecies competition by killing cholera strains harboring different effector/immunity proteins that compete for the same niche in the small intestine upon entry into the host, leading to kin selection and clonal expansion. Subsequently, killing of T6SS-sensitive bacteria competing for the same niche in the large intestine may allow *V. cholerae* to engage in intraguild predation thereby shifting bacterial populations to its advantage and acquire nutrients from prey. For either scenario, we suggest that *V. cholerae*'s T6SS-mediated bacterial killing has contributed to its success as an enteric pathogen.

#### **ACKNOWLEDGMENTS**

The authors thank present and past laboratory personnel for contributions leading to this manuscript as well as colleagues whose work is referenced. Stefan Pukatzki is supported by the Canadian Institute of Health Research Operating Grant MOP-84473 and by Alberta Innovates Health Solutions and Daniele Provenzano was supported by NIH grants MD001091-01 and GM068855-02.

#### **AUTHOR CONTRIBUTIONS**

Daniele Provenzano and Stefan Pukatzki wrote the manuscript and conceived the ideas therein.

#### **REFERENCES**

"fmicb-04-00384" — 2013/12/9 — 20:05 — page 3 — #3

Barua, D. (1992). *"*History of cholera,"in*Cholera*, eds D. Barua andW. B. Greenough (New York, NY: Plenum Publishing), 1–36.

Basler, M., Ho, B. T., and Mekalanos, J. (2013). Tit-for-tat: type VI secretion system counterattack during bacterial cell–cell interactions. *Cell* 152, 884–894. doi: 10.1016/j.cell.2013.01.042


"fmicb-04-00384" — 2013/12/9 — 20:05 — page 4 — #4


Zheng, J., Shin, O. S., Cameron, D. E., and Mekalanos, J. J. (2010). Quorum sensing and a global regulator TsrA control expression of type VI secretion and virulence in *Vibrio cholerae*. *Proc. Natl. Acad. Sci. U.S.A.* 107, 21128–21133 doi: 10.1073/pnas.1014998107

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 25 October 2013; paper pending published: 04 November 2013; accepted: 25 November 2013; published online: 10 December 2013.*

*Citation: Pukatzki S and Provenzano D (2013) Vibrio cholerae as a predator: lessons from evolutionary principles. Front. Microbiol. 4:384. doi: 10.3389/fmicb.2013.00384 This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Pukatzki and Provenzano. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00384" — 2013/12/9 — 20:05 — page 5 — #5

## *Vibrio campbellii hmgA*-mediated pyomelanization impairs quorum sensing, virulence, and cellular fitness

#### *Zheng Wang1 \*, Baochuan Lin1, Anahita Mostaghim1,2, Robert A. Rubin3, Evan R. Glaser 4, Pimonsri Mittraparp-arthorn5, Janelle R. Thompson6, Varaporn Vuddhakul <sup>5</sup> and Gary J. Vora1*

*<sup>1</sup> Center for Bio/Molecular Science & Engineering, Naval Research Laboratory, Washington, DC, USA*

*<sup>2</sup> School of Systems Biology, College of Science, George Mason University, Fairfax, VA, USA*

*<sup>6</sup> Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Zongze Shao, State Oceanic Administration, China Xiu-Lan Chen, Shandong university, China*

#### *\*Correspondence:*

*Zheng Wang, Center for Bio/Molecular Science & Engineering, Naval Research Laboratory, Code 6900, 4555 Overlook Avenue, SW, Washington, DC 20375, USA e-mail: zheng.wang@nrl.navy.mil*

Melanization due to the inactivation of the homogentisate-1,2-dioxygenase gene (*hmgA)* has been demonstrated to increase stress resistance, persistence, and virulence in some bacterial species but such pigmented mutants have not been observed in pathogenic members of the *Vibrio* Harveyi clade. In this study, we used *Vibrio campbellii* ATCC BAA-1116 as model organism to understand how melanization affected cellular phenotype, metabolism, and virulence. An in-frame deletion of the *hmgA* gene resulted in the overproduction of a pigment in cell culture supernatants and cellular membranes that was identified as pyomelanin. Unlike previous demonstrations in *Vibrio cholerae, Burkholderia cepacia,* and *Pseudomonas aeruginosa,* the pigmented *V. campbellii* mutant did not show increased UV resistance and was found to be ∼2.7 times less virulent than the wild type strain in *Penaeus monodon* shrimp virulence assays. However, the extracted pyomelanin pigment did confer a higher resistance to oxidative stress when incubated with wild type cells. Microarray-based transcriptomic analyses revealed that the *hmgA* gene deletion and subsequent pyomelanin production negatively effected the expression of 129 genes primarily involved in energy production, amino acid, and lipid metabolism, and protein translation and turnover. This transcriptional response was mediated in part by an impairment of the quorum sensing regulon as transcripts of the quorum sensing high cell density master regulator LuxR and other operonic members of this regulon were significantly less abundant in the *hmgA* mutant. Taken together, the results suggest that the pyomelanization of *V. campbellii* sufficiently impairs the metabolic activities of this organism and renders it less fit and virulent than its isogenic wild type strain.

**Keywords:** *Vibrio***, melanin, bioluminescence, quorum sensing, tyrosine catabolism**

#### **INTRODUCTION**

As a member of the L-tyrosine catabolism pathway in bacterial and eukaryotic organisms, the enzyme homogentisate 1,2 dioxygenase (HmgA) catalyzes the intermediate homogentisic acid into 4-maleylacetoacetate which is further catabolized to yield fumarate and acetoacetate. In some bacterial species, it has been demonstrated that the inactivation of the *hmgA* gene results in the accumulation of homogentisic acid which when autooxidized leads to the formation of the water-soluble brown pigment pyomelanin (Rodriguez-Rojas et al., 2009; Schmaler-Ripcke et al., 2009; Turick et al., 2009; Valeru et al., 2009; Wang et al., 2011). This phenotype has been observed in naturally pigmented environmental and clinical strains of *Vibrio cholerae* and has been shown to be due to mutations in the *hmgA* gene (Wang et al., 2011). Interestingly, pyomelanin pigmented *V. cholerae* demonstrate greater UV and oxidative stress resistance, virulence factor expression and infant mouse intestine colonization rates than their non-pigmented counterparts (Valeru et al., 2009). The ability of pyomelanin to confer increased resistance to oxidative stress appears to contribute to virulence by reducing the susceptibility of pigmented bacteria to host defense mechanisms. Because of these particular characteristics, it is not surprising that pyomelaninproducing *Pseudomonas aeruginosa* and *Burkholderia cepacia* are frequently isolated from cystic fibrosis patients (Zughaier et al., 1999; Rodriguez-Rojas et al., 2009). Furthermore, the production of pyomelanin has also been shown to provide greater protection from other environmental stresses such as hyperosmotic shock and elevated temperatures (Kotob et al., 1995) and act as a sole terminal electron acceptor and soluble electron shuttle to iron which may provide an additional fitness advantage to pyomelanin-producing mutants in anaerobic environments (Turick et al., 2002).

Despite these seemingly advantageous phenotypes, such pigmented mutants have not been reported from pathogenic members of the *Vibrio* Harveyi clade. Two of the most economically important Harveyi clade species, *V. campbellii* and *V. harveyi*, are common inhabitants of tropical marine environments and are among the most important bacterial pathogens of many

*<sup>3</sup> Mathematics Department, Whittier College, Whittier, CA, USA*

*<sup>4</sup> Division of Electronics Science and Technology, Naval Research Laboratory, Washington, DC, USA*

*<sup>5</sup> Department of Microbiology, Faculty of Science, Prince of Songkla University, Hat Yai, Thailand*

commercially farmed marine invertebrate and vertebrate species (Thompson et al., 2004; Austin and Zhang, 2006). As certain pathogenic members of both species are capable of producing quorum sensing induced bioluminescence, the disease caused by them is often referred to as luminescent vibriosis (Defoirdt et al., 2008) and is a disease manifestation that is frequently implicated in outbreaks within penaeid shrimp larval culture facilities worldwide (Austin and Zhang, 2006). Given the importance of shrimp hemocyte-mediated oxidative defense mechanisms in combatting *Vibrio* infections (Ji et al., 2011), it is not unreasonable to posit that pyomelanization may benefit the survival and perhaps exacerbate the virulence of vibrios in this host environment. However, the production of pyomelanin comes at the cost of impairing the tyrosine catabolism pathway and the effect of the inactivation of *hmgA* and/or pyomelanin production on global cellular metabolism is not known. In this study, we used *V. campbellii* ATCC BAA-1116, a bioluminescent marine bacterium that is best known as a model organism for quorum sensing studies (Bassler, 1999), to begin to determine the generality of pyomelaninmediated phenotypes and how the deletion of the *hmgA* gene and resulting pyomelanin production may affect cellular phenotypes, virulence, and transcription.

#### **MATERIALS AND METHODS**

#### **BACTERIAL STRAINS AND GROWTH CONDITIONS**

*V. campbellii* (ATCC strain BAA-1116; previously known as *V. harveyi* BAA-1116 or BB120; Lin et al., 2010) and the *hmgA* mutant were grown in Luria Marine (LM) medium (20 g NaCl, 10 tryptone, 5 g yeast extract per L, pH 7.8) or Tryptic Soy Broth containing 1% NaCl. *Escherichia coli* DH5α and SM10λpir used for standard DNA manipulation and conjugation were grown in Luria Broth (LB) medium.

#### **CONSTRUCTION OF THE hmgA IN-FRAME DELETION MUTANT**

The in-frame deletion of the *V. campbellii hmgA* gene (*hmgA*) was generated by overlap PCR (Warrens et al., 1997). Briefly, ∼500 bp DNA fragments upstream and downstream of the *hmgA* open reading frame were amplified from *V. campbellii* BAA-1116 genomic DNA using the primer pairs *hmgA*-a (5 - TAggatccTGTACGAAATCGACCATCTGAC)/*hmgA*-b (5 -c) and *hmgA*-c (5 -GAGGAGTACTAAGCGGGGGCAAGGATGAAA)/ *hmgA*-d (5 -CActcgagACTTCACCTTCGAAGTCAATCC), respectively. The two PCR products were annealed using their overlapping region and amplified using primers *hmgA*-a and *hmgA*-d. The resulting 1 Kb PCR fragment was cloned into the pCR4-TOPO vector using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA, USA). This assembled fragment was then digested from the TOPO vector with *Bam*HI and *Xho*I and cloned into the plasmid pZW125, which was constructed by inserting a chloramphenicol resistance gene into the oriR*R*6*Kg* plasmid pWM91 containing the *sacB* gene (Metcalf et al., 1996). The resulting plasmid (pZW025) was transformed into *E. coli* strain Sm10λpir and transferred into a *V. campbellii* spontaneous streptomycin resistant mutant (*V. campbellii*-str1) by conjugation. The conjugants were grown on LM agar plates containing 3μg/mL chloramphenicol and 1μg/mL streptomycin. The *hmgA* was selected on LM plates supplemented with 6% sucrose and verified by PCR.

#### **GROWTH CURVE ANALYSES**

Bacterial replication was measured using a Bioscreen C analyzer (Growth Curves USA, Piscataway, NJ, USA). Briefly, overnight cultures were diluted 1:5000 (∼10<sup>5</sup> cells/mL) in pre-warmed LM and five 200μL aliquots of the wild type (WT) and *hmgA* strains were transferred into a 100-well honeycomb plate. The plate was incubated at 30◦C for 48 h with continuous shaking and wide band OD450−<sup>580</sup> *nm* measurements taken every 30 min. Three independent experiments were performed in this manner.

#### **MEASUREMENT OF PIGMENTATION, BIOLUMINESCENCE AND CELLULAR SUSCEPTIBILITY TO H2O<sup>2</sup>**

Matched diluted overnight WT and *hmgA* cultures were used to inoculate 50 mL LM media in 250 mL flasks and incubated at 30◦C and 200 rpm. Every 24 h, three 100μL aliquots of culture were collected and bacterial cells were pelleted via centrifugation at 10,000× *g* for 5 min. Supernatant pigments were measured using a NanoDrop ND-2000c spectrophotometer (Thermo Scientific, Pittsburg, PA, USA) at OD400. Another three 100μL aliquots of culture were placed in a black U96 Nunc MicroWell™ plate (Thermo Scientific) and measured for bioluminescence using a Luminoskan Ascent Microplate Luminometer (Thermo Scientific). Three independent experiments were performed in this manner.

At the 48 h time point, WT, and *hmgA* cells were harvested, washed and resuspended in fresh LM media. They were then incubated with 2 mM H2O2 at room temperature for 15 min. The percentage survival was calculated by counting colony forming units (CFU) immediately before and after the H2O2 treatment on LM agar plates. The potentially protective effect of WT and *hmgA* culture supernatants against H2O2 treatment was also tested using WT cells. Briefly, mid-log phase WT cells were harvested, washed and resuspended in 0.2μm filter-sterilized supernatants from WT and *hmgA* 48 h LM media cultures*.* The cell suspensions were then incubated at room temperature in the presence of 2 mM H2O2 for 15 min. The percentage survival was calculated by counting CFUs immediately before and after the H2O2 treatment on LM agar plates. The data for each of these experiments was generated from three independent assays.

#### **PIGMENT PREPARATION AND ELECTRON SPIN RESONANCE SPECTROSCOPY**

Partial purification of the pigment from the *hmgA* strain was modified from the method previously described by Turick et al. (2002). Briefly, a 50 mL *hmgA* culture was grown in LM at 30◦C for 96 h with shaking at 200 rpm. The cells were harvested via centrifugation at 5000× *g* for 10 min, and the supernatant was removed and acidified with 6 N HCl to a final solution concentration of 0.4 N and was then allowed to precipitate for 12 h at room temperature. The concentrated pigment was collected by centrifugation at 8000× *g* for 20 min, washed twice with dH2O and then dried using a SpeedVac Concentrator (Thermo Scientific). A pure synthetic melanin that was chemically prepared from the oxidation of tyrosine was purchased and used as a control (M8631, Sigma-Aldrich, St. Louis, MO, USA). A second control, DHN-melanin, was prepared from the conidia of the fungus *Aspergillus niger* using the method of Youngchim et al. (2004). The pigment powder samples were characterized by electron spin resonance (ESR) at 300K in a Bruker 9.5 GHz spectrometer. Typical microwave powers of 5–20 mW with 1G modulation amplitude and 100 kHz field modulation were employed for these experiments.

#### **SHRIMP VIRULENCE ASSAYS**

The LD50 of the WT and *hmgA* strains were evaluated on the black tiger shrimp *Penaeus monodon*. Both strains were grown in Tryptic Soy Broth containing 1% NaCl at 30◦C with shaking at 150 rpm, harvested by centrifugation at 2000× *g* for 10 min and washed twice with sterile Marinum® artificial seawater (ASW) (Mariscience International Co. Ltd., Bangkok, Thailand). Bacterial cell suspensions in ASW were adjusted to <sup>2</sup>.<sup>6</sup> <sup>×</sup> <sup>10</sup><sup>8</sup> CFU/mL using a turbidimeter (Oxoid Ltd., United Kingdom) and twofold dilutions were performed to obtain the required concentrations of bacteria prior to injecting the shrimp. The juvenile shrimp used in this study were 10–13 g in weight and 4–5 inches in length. Each shrimp received an intramuscular injection of 100μL diluted *V. campbellii* (with batches of seven shrimp/dose) between the third and fourth abdominal segments. Control shrimp were injected with ASW. The experiments were performed in quadruplicate. The animals were maintained in a 70 L ASW glass tank at a temperature of 29 ± 1◦C and salinity of 17 ppt. Shrimp mortalities were observed within 48 h of injection and were confirmed by detecting bioluminescence in the organs of the dead shrimp. The LD50 was calculated using the method of Reed and Muench (1938).

#### **MICROARRAY-BASED TRANSCRIPTOME ANALYSES**

Aliquots of three cultures (3.<sup>0</sup> <sup>×</sup> <sup>10</sup><sup>8</sup> cells/mL) of the WT and *hmgA* strains grown in LM at 30◦C for 48 h with constant shaking at 200 rpm were harvested for total RNA extraction. RNA was isolated using the RiboPure™-Bacteria Kit (Life Technologies, Grand Island, NY, USA), treated with DNase and 10μg of total RNA from each culture was further purified using the MICROB*Express*™ Bacterial mRNA Enrichment Kit (Life Technologies) according to the manufacturer's specifications. All RNA preparations were quantified and analyzed using the Agilent 2100 Bioanalyzer (Agilent Technologies, Inc., Santa Clara, CA, USA) and normalized to 1μg. The normalized RNA was labeled, purified, fragmented and hybridized to a custom Affymetrix microarray (520694F) according to standard protocols (Affymetrix, Santa Clara, CA, USA). All hybridizations incubated for 16 h at 49◦C in the GeneChip® Hybridization Oven 640 at 60 rpm and the microarrays were then washed and stained with the GeneChip® Fluidics Station 450 and scanned using the GeneChip® Scanner 7G (Affymetrix). Hybridization signal intensities were analyzed with the GeneChip® Operating Software (GCOS) to generate raw image files (.DAT) and simplified image files (.CEL) with intensities assigned to each of the corresponding probe positions. The data collected was used to profile the expression levels of 4831 open reading frames. The data containing the distribution of the probe amplitudes were calculated and a classical analysis of variance (ANOVA; applying the CRAN R's aov function) was performed across conditions for each probe site. The median values revealing the gene-level measurement of differential expression was determined (Rubin, 2009). The gene designations and annotations utilized are from the Naval Research Laboratory's *V. campbellii* ATCC BAA-1116 genome sequencing effort (GenBank accession numbers CP006605, CP006606, CP006607) and the expression profiling data can be found in the GenBank Gene Expression Omnibus repository (accession number GSE46223).

#### **QUANTITATIVE REVERSE TRANSCRIPTION PCR**

Real-time reverse transcription PCR assays were conducted using the iScript™ One-Step RT-PCR Kit with SYBR Green (Bio-Rad Laboratories, Hercules, CA, USA). One nanogram of mRNA from two biological replicates were tested in triplicate on an iCycler (BioRad). The PCR primers were designed using Primer3 online software (v. 0.4.0) (http://frodo.wi.mit.edu/). Relative quantities of the transcripts were determined using the 2−--Ct formula where -Ct is the difference in Ct of the selected genes and Ct of the normalizer gene, and --Ct is the difference in -Ct from *hmgA* and -Ct from the WT. The *rpoS1* gene was used to normalize the expression levels of the selected genes as its transcription level was found to be constant in both the WT and *hmgA* strains.

#### **RESULTS AND DISCUSSION**

#### **CHARACTERIZATION OF** **hmgA PIGMENT AND PHENOTYPES**

An examination of the *V. campbellii* ATCC BAA-1116 genome revealed that 4 genes in the catabolic pathway of tyrosine metabolism, *hmgA* (M892\_02450), *hppd* (4 hydroxyphenylpyruvate dioxygenase, M892\_02455), *fahA* (fumarylacetoacetase, M892\_02445), and *maiA* (maleylacetoacetate isomerase, M892\_02440), appeared to form an operon and shared the same genetic synteny as other sequenced *Vibrio* species. An in-frame deletion of the *hmgA* gene was generated (*hmgA*) to investigate the role of this gene in pigment production and cellular physiology. When cultured on LM agar plates, in baffled glass Erlenmeyer flasks or polystyrene round-bottomed tubes for ≥ 48 h, the *hmgA* strain produced a brown pigment that was expected to be the result of the auto-oxidation and polymerization of homogentisate (**Figures 1A,B**). The pigment was found to be present in *hmgA* cell-free supernatants as well as washed cell pellets suggesting that the pigment was not only released into the microenvironment but could also be found associated with the bacterial cell membrane (data not shown). Interestingly, the production of this pigmentation was not observed when *hmgA* cells were cultured in 50 mL polypropylene conical tubes (**Figure 1C**). Under the experimental conditions utilized, the 50 mL conical tubes provided the least aeration. In the absence of sufficient aeration, which is required for the oxidation of homogentisate, the production of the pigment was not observed. In order to identify the oxidized homogentisate polymer, the extracted brown pigment from *hmgA* cell culture supernatants was examined using ESR. The ESR spectrum of the pigment revealed a distinct stable free radical signal that was characteristic of melanin and highly similar to synthetic eumelanin and DHN melanin of *Aspergillus niger* (**Figure 1D**) thus, confirming that the *hmgA* pigment was pyomelanin. The results demonstrate that like *V. cholerae*

*hmgA*, *V. campbellii hmgA* also produces pyomelanin but does not appear to do so in the same abundance or nearly as rapidly (Valeru et al., 2009).

Growth curve analyses in nutrient rich LM medium revealed that the WT and *hmgA* strains grew equally well during the lag, log, and early stationary phases of growth. However, *hmgA* displayed a lesser ability to survive during late stationary phase (post 48 h) (**Figure 2A**): a time point that coincided with the measurable production of pyomelanin (**Figure 2B**).

In other bacterial and fungal organisms, melanin has been demonstrated to have a role in protecting against certain environmental stressors as mutations in the *hmgA* gene and the resulting melanization have been shown to provide an increased resistance to UV irradiation and H2O2-mediated oxidative stress (Rodriguez-Rojas et al., 2009; Schmaler-Ripcke et al., 2009; Valeru et al., 2009). However, when we performed similar experiments comparing the response of *V. campbellii* WT and *hmgA* washed cells to different doses of UV irradiation and concentrations of H2O2, no significant differences in viability were observed (data not shown). Interestingly, however, we did observe the protective

**FIGURE 2 | Growth, pigment production and bioluminescence in** *V. campbellii* **WT and** *hmgA***. (A)** Grow curves of the WT and *hmgA* strains in LM at 30◦C. WT (closed circle), *hmgA* (open circle); **(B)** Absorbance of pigment-containing supernatants. WT (white bars), *hmgA* (gray bars); (**C**) Survival of WT *V. campbellii* cells in 2 mM H2O2 when resuspended in supernatants from WT or *hmgA* cultures; **(D)** Bioluminescence measurements. WT (white bars), *hmgA* (gray bars). Error bars represent the standard deviation from three experiments.

properties of pyomelanin when cell-free pyomelanin-containing supernatants from *hmgA* cultures were incubated with WT cells and then exposed to 2 mM H2O2 (**Figure 2C**). Comparatively, cell-free supernatants from WT cultures did not demonstrate this protective function. The combined H2O2 challenge results demonstrated the protective material property of *V. campbellii* pyomelanin and suggest that there is not sufficient membraneassociated pyomelanin in *V. campbellii* to afford this same level of protection in the absence of the accumulated extracellular pyomelanin.

Unlike any other pyomelanin-producing bacterium investigated to date, *V. campbellii* ATCC BAA-1116 is capable of generating quorum sensing-induced bioluminescence. Interestingly, previous studies have suggested that the luciferase enzyme and bacterial bioluminescence, like pyomelanin, also play a physiological role in protecting cells against UV and oxidative stress. For example, UV irradiation has been found to stimulate bioluminescence (Czyz et al., 2002) which in turn photoreactivates DNA repair processes (Kozakiewicz et al., 2005) and bioluminescent cells have been shown to be significantly more resistant to UV irradiation than their non-bioluminescent counterparts (Czyz et al., 2000; Kozakiewicz et al., 2005). In addition, various oxidants, such as H2O2, have been shown to severely impair the growth of vibrios lacking the luciferase enzyme (Szpilewska et al., 2003). This is due to the fact that in addition to the production of light, this enzyme is also capable of increasing cellular resistance to oxidative stress by detoxifying reactive oxygen species (Katsev et al., 2004). Therefore, in addition to the more common bacterial mechanisms of UV and oxidative stress protection, bioluminescent *V. campbellii* also contain luciferasebased protective mechanisms. This combination of luciferase and pyomelanin in the same organism introduced the possibility of additive or synergistic stress protection. However, when the bioluminescence output of *hmgA* was examined during measurable pyomelanin production (48 h), it was found to be significantly attenuated in comparison to the WT (**Figure 2D**). While this result demonstrated a diminution of light production in pyomelanin-producing *hmgA*, it was incapable of determining whether this was due to a decrease in luciferase abundance or activity, reduced intracellular O2 or necessary substrates.

#### **SHRIMP VIRULENCE MODEL**

In *V. cholerae hmgA* and *P. aeruginosa hmgA*, pyomelanization was shown to play a role in increasing virulence factor expression and adaptation to chronic infections in vertebrate animal models (Rodriguez-Rojas et al., 2009; Valeru et al., 2009). As *V. campbellii* BAA-1116 is known to pathogenize shrimp, we sought to determine whether *V. campbellii* pyomelanization would have a similar effect on invertebrate animal infections. *V. campbellii* WT and *hmgA* were used to infect juvenile black tiger shrimp (*Penaeus monodon*) and the LD50 of both were evaluated. Surprisingly, the LD50 of *hmgA* was ∼2.7 times higher (less lethal) than that of the WT indicating that the production of pyomelanin in *hmgA* was associated with decreased virulence in this model infection system (**Table 1**).

In invertebrates, one of two major immune responses against invading pathogens is the prophenoloxidase activating system



*Four independent LD50 experiments were performed. Mean values of LD50 between the WT and hmgA were significantly different (P* < *0.05, student's t-test).*

(proPO) (Cerenius and Soderhall, 2004). Upon infection, nonself molecules such as lipopolysaccharide, peptidoglycan and β-glucan can activate the proPO I cascade and result in the formation of melanin around the invading microorganisms. In this circumstance, the host formed melanin is thought to physically shield the pathogens to prevent or retard their growth. Host derived quinones, which are intermediates of melanin production, may also be involved in the production of cytotoxic molecules (e.g., superoxides, hydroxyl radicals) that could help inactivate the invading pathogens. The protective efficacy of the proPO system is further highlighted by the demonstration that gene silencing of PO activating enzymes in *Penaeus monodon* increases the susceptibility of the host to *V. harveyi* infection (Amparyup et al., 2009; Charoensapsri et al., 2009). Given these facts and the demonstrated phenotypes of *V. campbellii hmgA*, we suggest that pyomelanization may reduce this bacterium's virulence potential in two ways. First, pyomelanin production by *hmgA* cells may effectively add another layer to the host-assembled melanin around the sites of infection so as to further limit bacterial growth. The comparatively poor survival of *hmgA* in late stationary phase (the time of melanization) (**Figure 2A**) may allude to this possibility. Second, the pyomelanin produced by *hmgA* may be recognized by the host as another foreign moiety that could further stimulate the proPO system and enhance the clearance of these bacteria.

#### **TRANSCRIPTOME ANALYSES**

As the differences in pyomelanin production, bioluminescence and survival were observed during late stationary phase (48 h), we chose this time point to perform comparative microarray-based expression profiling analyses to understand how the deletion of the *hmgA* gene and subsequent production of pyomelanin gave rise to the observed phenotypes. Whole genome expression profiling revealed that 129 genes (2.7% of the interrogated genome) were significantly modulated in *hmgA* when compared to the WT (adjusted *p*-value < 0.001). Overall, inactivation of the *hmgA* gene appeared to affect the expression of genes involved in energy production and conversion, amino acid metabolism, lipid metabolism, and quorum sensing/bioluminescence (**Figure 3**). Interestingly, the transcript levels of all 129 genes were found to be less abundant in *hmgA* with approximately 70% of them

demonstrating a ≥ twofold reduction in transcript abundance (**Table 2**). The transcriptional modulation of 10 of these genes was also verified using quantitative RT-PCR (**Table 2**).

#### **ENERGY PRODUCTION AND ELECTRON TRANSFER**

Two of 3 genes (M892\_08460, M892\_08470) in an operon encoding subunits of ubiquinol-cytochrome c reductase were down-regulated in *hmgA*. This cytochrome complex catalyzes the oxidoreduction of mobile redox components generating an electrochemical potential (Kurowski and Ludwig, 1987). Two other modulated genes, NrfB (M892\_17240) and NrfC (M892\_17245), encode cytochrome-type components of the electron transfer chain of respiratory nitrite ammonification in γproteobacteria (Simon, 2002). This electron transfer reaction usually occurs during anaerobic growth and the electron donor formate is readily formed from pyruvate by pyruvate formate lyase. Coincidently, the gene encoding pyruvate formate lyase (M892\_05385) was also repressed in the mutant. In addition, a number of genes in the glycolysis pathway, such as those encoding fructose-bisphosphate aldolase (M892\_13305), glyceraldehyde 3-phosphate dehydrogenase (M892\_15785) and enolase (M892\_19270), were down-regulated as well, and likely result in the reduced production of pyruvate. Since HmgA is not a DNA regulatory factor, these results indicated that the pyomelanin produced in *hmgA* may decrease redox activity by serving as an electron sink which in turn would interfere with electron transfer and further weaken cellular respiration in late stationary phase.

#### **AMINO ACID AND LIPID METABOLISM**

The transcripts from three well characterized gene clusters involved in amino acid and lipid metabolism (*etfABD*, *ivd-ABCDEFG, liuABCDE*) were also significantly down-regulated in *hmgA*. These operons encode enzymes that take part in the branched-chain amino acid (i.e., isoleucine, leucine, valine) degradation pathway that is used for energy production in many proteobacteria (Kazakov et al., 2009). Proteins encoded by *etfABD* catalyze electron transfer from quinones to flavoproteins not only in the leucine degradation pathway but also in the fatty acid degradation pathway. The master regulator of the branched-chain amino acid degradation pathway, LiuR, is encoded by the *liuR* gene which resides between the *liuABCDE* and *ivdABCDEFG* operons. The *liuR* gene was also significantly down-regulated in *hmgA* (*p*-value = 0.0003) suggesting a regulatory mechanism for the pyomelanin-induced retardation of branched-chain amino acid degradation.

It was interesting to note that another amino acid degradation pathway was also affected by the production of pyomelanin. The proline utilization operon *putBCP*, encoding proline dehydrogenase, δ-1-pyroline-5-carboxylate dehydrogenase and permease, respectively, was also down-regulated in *hmgA*. However, the expression of *pruR*, which is adjacent to and has been reported to regulate the *putBCP* operon in *P. aeruginosa* PAO1 (Nakada et al., 2002), did not change. Interestingly, transcription of the *putBCP* operon has been reported to respond to osmotic stress by producing more of the final product glutamate in *V. vulnificus* (Lee et al., 2003) and coincidently, pyomelanin biosynthesis was also demonstrated to be induced by osmotic stress in *V. cholerae* (Coyne and Al-Harthi, 1992). Thus, it is possible that *hmgA* pyomelanin may function as a solute glutamate to counter osmotic stress during late stationary phase thus, alleviating the physiological signal for the increased expression of *putBCP* and resulting in the decreased transcription observed in this study.

#### **VIRULENCE, QUORUM SENSING, AND BIOLUMINESCENCE**

Although the virulence mechanisms that contribute to bioluminescent vibriosis are not completely understood, several biomolecules are known to be contributing factors (Austin and Zhang, 2006). The expression of many of these virulence factors, including the type III secretion system (Henke and Bassler, 2004a), extracellular toxin (Manefield et al., 2000), metalloprotease (Mok et al., 2003), siderophore (Lilley and Bassler, 2000), chitinase (Defoirdt et al., 2010), phospholipase, caseinase, and gelatinase (Natrah et al., 2011) are known to be regulated by the quorum sensing master regulator LuxR. Our findings from the shrimp virulence model demonstrated that *hmgA* was less virulent than the WT and led to the suggestion that this phenotype may have been due to the down-regulation of

#### **Table 2 | Differentially expressed genes in** *hmgA* **compared with WT** *V. campbellii***.**


*(Continued)*

#### **Table 2 | Continued**


*(Continued)*

#### **Table 2 | Continued**


*aAverage signal intensity log2 values from three experiments.*

*bTranscript ratio between WT and hmgA.*

*cTranscript ratio between WT and hmgA from qRT-PCR data. qRT-PCR Ct values were derived from the average of duplicates from two independent biological samples.*

*dUndivisible as no transcript was detected from hmgA.*

*eCOG functional categories: C, energy production and conversion. D, cell cycle control, cell division and chromosome partitioning; E, amino acid transport ad metabolism; G, carbohydrate transport and metabolism; H, coenzyme transport and metabolism; I, lipid transport and metabolism; J, translation, ribosomal structure and biogenesis; K, transcription; L, replication, recombination and repair; M, cell wall/membrane/envelop biogenesis; O, post-translational modification, protein turnover and chaperones; P, inorganic ion transport and metabolism; Q, secondary metabolites biosynthesis, transport and catabolism; R, general function prediction only; S, function unknown; T, signal transduction mechanisms; U, intracellular trafficking, secretion and vesicular transport; V, defense mechanisms.*

several virulence factors in *hmgA*. However, with the exception of two [hemolysin (M892\_26640) and azurin (M892\_25430)], the genes encoding these factors were not differentially expressed suggesting that the decreased virulence potential of *hmgA* was not caused by an overt down-regulation of accepted virulence factors.

The expression of bioluminescence has also been associated with virulence in shrimp (Manefield et al., 2000; Phuoc et al., 2009) and indirect evidence has suggested that bioluminescence and a toxic extracellular protein are co-regulated (Manefield et al., 2000). The molecular mechanisms of *Vibrio* quorum sensing and its resulting bioluminescence have been most extensively studied in *V. campbellii* BAA-1116 and the regulation of *luxCD-ABEGH* gene cluster responsible for bioluminescence is known to be positively regulated by the quorum sensing master regulator LuxR (Bassler et al., 1993; Henke and Bassler, 2004b; Lenz et al., 2004; Waters and Bassler, 2006; Tu and Bassler, 2007). In this study, the bioluminescence output of *hmgA* was found to be significantly attenuated and this phenotype could be attributed to the pronounced down-regulation of *luxCDABEGH* in *hmgA* (**Table 2**). Furthermore, the gene encoding the high cell density state quorum sensing master regulator LuxR and another quorum sensing regulator (LuxT) (Lin et al., 2000) were also significantly modulated in *hmgA*. It is estimated that this singular effect on *luxR* transcript levels accounted for 40% of the transcriptional modulation seen in *hmgA* as a comparison with the LuxR regulon (manuscript in preparation) revealed 52 modulated genes in common.

In addition to the *lux* bioluminescence genes, LuxR has also been shown to regulate the synthesis of storage polyhydroxybutyrates in bioluminescent vibrios (Miyamoto et al., 1998). Polyhydroxybutyrates, the most characterized member of the polyhydroxyalkanoates, are storage polyesters that are produced and accumulate in the bacterial cytosol in carbon-rich environments when other nutrients are limited (Reddy et al., 2003). When available carbon has been exhausted, these storage polymers can be catabolized for carbon and energy. In *V. campbellii hmgA*, the expression of the polyhydroxybutyrate synthesis gene cluster (*phaCDAB*) was markedly reduced in comparison to the WT. Thus, a blockage of the tyrosine catabolism pathway and production of pyomelanin not only retarded amino acid degradation but also kept *V. campbellii* from accumulating carbon storage polymers when in a carbon-rich environment. Evidence of these nutrient management deficiencies could be seen in the late stationary phase *hmgA* cultures (**Figure 2A**) where depleted carbon supplies may have contributed to diminished survival.

The transcriptional modulation of members of the LuxR quorum sensing regulon in *hmgA*, presumably due to reduced *luxR* transcript levels, was unexpected as HmgA is not known to be a regulatory protein. One possible explanation for this observation may be that extracellular pyomelanin is somehow interfering with the binding of autoinducer molecules to their cognate quorum sensing histidine kinase receptors. Alternatively, membrane-embedded pyomelanin may sufficiently alter membrane structure and impair the binding of autoinducer molecules or subsequent phosphorylation cascade. In either case, the quorum sensing signaling cascade would mimic a low cell density state resulting in the phosphorylation of the response regulator LuxO and activated transcription of regulatory small RNAs (Qrr sRNAs) (Lilley and Bassler, 2000). As the base-pairing of the Qrr sRNAs to *luxR* transcripts results in the degradation of *luxR* mRNA (Lenz et al., 2004; Tu and Bassler, 2007), pyomelanization may lock the cells in a low cell density state thus, explaining the transcriptome profiling data.

#### **CONCLUSIONS**

In this study, we used the model quorum sensing bacterium *V. campbellii* BAA-1116 to determine how the deletion of the *hmgA* gene and resulting pyomelanization affected cellular phenotypes, virulence, and transcription. While the material properties of *V. campbellii* pyomelanin were similar to previous descriptions, there did not appear to be a generality of pyomelanin-mediated phenotypes as pyomelanization had either a neutral or deleterious effect on cell survival. Despite the fact that pyomelanin production in other pathogenic bacteria had been shown to increase their virulence and adaptation to chronicity in mammalian infection models, our results using a shrimp infection model indicated that pyomelanized *V. campbellii* were actually less virulent than the isogenic WT strain. These observations may be due in part to the comparatively lesser amount of pyomelanin produced and retained in *V. campbellii* and the effect this production had on cellular metabolism. This was supported by the first transcriptome-level analysis comparing a pyomelaninproducing mutant with its isogenic WT strain. It is also worth noting that the immune responses from different animal models (invertebrate vs. vertebrate) may also play a large role in the differences seen in virulence. *V. campbellii* is primarily an invertebrate animal pathogen and this is the first time that the effect of pyomelanization has been tested in an invertebrate model system (i.e., a natural host organism). The transcriptional profiles demonstrated that the deletion of the *hmgA* gene led to significantly lower transcript abundance levels of several important metabolic processes that disrupted cellular homeostasis and fitness in stressful environments (e.g., stationary phase). Taken together, these findings may explain why naturally pyomelanized *V. campbellii* or *V. harveyi* have not been identified from the marine environment or infected eukaryotic host organisms.

#### **ACKNOWLEDGMENTS**

This work was funded by the Office of Naval Research via U.S. Naval Research Laboratory core funds. The opinions and assertions contained herein are those of the authors and are not to be construed as those of the U.S. Navy, military service at large, or U.S. Government.

#### **REFERENCES**


persistence in chronic lung infection. *Microbiology* 155, 1050–1057. doi: 10.1099/mic.0.024745-0


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 18 September 2013; paper pending published: 21 October 2013; accepted: 22 November 2013; published online: 11 December 2013.*

*Citation: Wang Z, Lin B, Mostaghim A, Rubin RA, Glaser ER, Mittraparp-arthorn P, Thompson JR, Vuddhakul V and Vora GJ (2013) Vibrio campbellii hmgA-mediated pyomelanization impairs quorum sensing, virulence, and cellular fitness. Front. Microbiol. 4:379. doi: 10.3389/fmicb.2013.00379*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Wang, Lin, Mostaghim, Rubin, Glaser, Mittraparp-arthorn, Thompson, Vuddhakul and Vora. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### Structure, gene regulation and environmental response of flagella in *Vibrio*

#### *Shiwei Zhu , Seiji Kojima and Michio Homma\**

*Division of Biological Science, Graduate School of Science, Nagoya University, Nagoya, Japan*

#### *Edited by:*

*Rita R. Colwell, University of Maryland, USA*

#### *Reviewed by:*

*Gary J. Vora, Naval Research Laboratory, USA Karen L. Visick, Loyola University Chicago, USA*

#### *\*Correspondence:*

*Michio Homma, Division of Biological Science, Graduate School of Science, Nagoya University, Chikusa-Ku, Nagoya 464-8602, Japan*

*e-mail: g44416a@cc.nagoya-u.ac.jp*

*Vibrio* species are Gram-negative, rod-shaped bacteria that live in aqueous environments. Several species, such as *V. harveyi*, *V. alginotyticus*, and *V. splendidus*, are associated with diseases in fish or shellfish. In addition, a few species, such as *V. cholerae* and *V. parahaemolyticus*, are risky for humans due to infections from eating raw shellfish infected with these bacteria or from exposure of wounds to the marine environment. Bacterial flagella are not essential to live in a culture medium. However, most *Vibrio* species are motile and have rotating flagella which allow them to move into favorable environments or to escape from unfavorable environments. This review summarizes recent studies about the flagellar structure, function, and regulation of *Vibrio* species, especially focused on the Na+-driven polar flagella that are principally responsible for motility and sensing the surrounding environment, and discusses the relationship between flagella and pathogenicity.

**Keywords: bacterial flagellum, ion-driven motor, motility, chemotaxis, pathogenicity**

#### **INTRODUCTION**

*Vibrio* species are Gram-negative, rod-shaped bacteria that live in all types of aqueous environments, including marine, freshwater, and estuary (Blake et al., 1980; Joseph et al., 1982; Johnson et al., 2012). All *Vibrio* species can move using flagella, which are cell surface organelles that can propel them. Unlike eukaryotic flagella, each bacterial flagellum is driven by a rotary motor embedded in the cell envelope, and the flagellar rotation is harnessed by the ion-motive force across the cell membrane (Berg, 2003; Terashima et al., 2008). The basic structure and function of the flagellum produced by *Vibrio* species are the same as those in other species (**Figure 1**) (Aizawa, 1996; Chen et al., 2011). Each flagellum consists of a filament acting as a helical propeller, a hook functioning as a universal joint and a basal body working as a rotary motor (**Figure 1**) (Sowa and Berry, 2008; Li et al., 2011). More than 50 gene products are involved in flagellar synthesis (MacNab, 2003). Since flagella are relatively large motility organelles for the cell, the formation of the flagella and the expression of their components are tightly regulated. Their assembly process has been described in many reviews (Chilcott and Hughes, 2000; Kim and McCarter, 2000; Aldridge, 2002; Chevance and Hughes, 2008).

All *Vibrio* species have single or multiple flagella at the cell pole (called "polar flagellum") and can swim freely in a liquid environment. With respect to the flagellum, *V. cholerae* has a single polar flagellum (monotrichous). However, some species, such as *V. parahaemolyticus* and *V. alginolyticus,* produce enormous numbers of flagella at lateral or peritrichous positions (called "lateral flagella") in addition to the single polar flagellum when they are exposed to viscous environments (**Figure 1A**) (McCarter et al., 1988; Kawagishi et al., 1996). The polar and lateral flagella are structurally and functionally distinct from each other (McCarter, 2004). Each polar flagellum is covered with a sheath that is contiguous with the outer membrane, so that it is thick and can be observed using dark-field microscopy with a mercury lamp. Lateral flagella are not covered with a sheath, so they are difficult to observe using light microscopy. Although both types of flagella are driven by rotary motors embedded at their bases, the power source is different. Polar flagella are driven by the Na+ motive force, whereas lateral flagella are driven by the H+-motive force, as seen in the lateral flagella of *Escherichia coli* or *Salmonella enterica* (**Figure 1**) (Atsumi et al., 1992; Asai et al., 2000; Blair, 2003).

In terms of the pathogenicity of *Vibrio*, some species, such as *V. cholerae*, *V. vulnificus*, and *V. parahaemolyticus*, have been described extensively as human pathogens (Daniels and Shafaie, 2000; Yildiz and Visick, 2009), and *Vibrios* are also pathogenic to fishes or the other animals. In this review, we focus on flagellar function and assembly and on the relationship between flagellar motility and pathogenicity.

#### **FLAGELLAR BASAL BODY STRUCTURE AND MOTOR**

The overall structure of the flagellar base is shown schematically in **Figure 1B**, based on electron microsopic images of the purified hook-basal body from a peritrichous flagellum of *Salmonella enterica* (right side) and from a polar flagellum of *Vibrio alginolyticus* (left side) (Francis et al., 1994; Thomas et al., 2006; Terashima et al., 2008, 2013). Both kinds of basal bodies share common features even though they originate from different species of Gram-negative bacteria: the hook and basal body with several rings embedded in the cell envelope. The flagellar basal body functions as a rotary motor, and consists of two parts: the rotary part (rotor) and the stationary part (stator). The stator complex is composed of two proteins, MotA/MotB (LafT/LafU), for the H+-driven motor of lateral flagella from *E. coli*, *Salmonella* and *Vibrio*, and PomA/PomB, for the Na+-driven polar flagellar

motor of *Vibrio* (**Figure 1**). The ion flux through the stator couples to the rotor-stator interaction that generates torque. The rotor contains several rings: from the cytoplasmic face, there is the C ring (also called the "switch complex") composed of FliG, FliM, and FliN, and the MS ring embedded in the cytoplasmic membrane (made of at most 26 copies of FliF). These rings are connected by a rod whose tip connects to the hook. The basal body contains two other rings, the P ring (FlgI), which is associated with the peptidoglycan layer, and the L ring (FlgH), which is located in the outer membrane (Aizawa, 1996; Terashima et al., 2008). Thus, the LP ring does not rotate but functions as a bushing for the central rod.

Although high resolution ultrastructural images have been reported for the basal body, intact images of the entire flagellar motor have remained ambiguous until recently due to the complexity of the stator units, which dynamically assemble around the rotor (Leake et al., 2006). However, the stator always dissociates from the detergent-solubilized flagellar basal body and no one has been able to isolate the basal body intact with the stator. In 2006, Murphy and co-workers (Murphy et al., 2006) first showed the structure of the complete flagellar motor *in situ* using the whole cell electron cryotomography method from the spirochete *Treponema primitia.* Their images captured electron densities corresponding to the stator units surrounding around the rotor. The resolution was still not sufficient to identify the detailed structure of the stator, but those images revealed the relative locations of the C-ring and the stator units for the first time. The structure of the polar flagellum of *Vibrio* contains two additional ring structures, named the T ring, which surrounds the periplasmic side of the P ring (Terashima et al., 2006), and the H ring, which is located at the outer rim of the L ring. The T ring is made of two proteins, MotX and MotY, that are essential components for the polar flagellar motor of *Vibrio*, and it is probably involved in stabilizing the stator surrounding the rotor (McCarter, 1994a,b; Okabe et al., 2001; Terashima et al., 2006). The crystal structure of MotY from *V. alginolyticus* in conjunction with biochemical analysis revealed that the C-terminal domain of MotY stabilizes the association with the peptidoglycan layer and that the N-terminal domain of MotY is involved in the association with the basal body (Kojima et al., 2008). The H ring is located at the outer rim of the LP ring, so that we initially thought that the LP ring was somehow bigger in this basal body (**Figure 2**). The loss of *flgT* function results in an almost non-motile phenotype and there is no formation of the T ring in addition to the lack of an outer rim structure of the LP ring. Therefore, FlgT is likely to be a component of the H ring, and is also important for assembly of the T ring. Our crystal structure and functional analyses of FlgT support this idea (Terashima et al., 2013). Perhaps these two unique extra ring structures in the polar flagellum are required to fix the stator units around the rotor (**Figure 2**). The *Vibrio* polar flagella can rotate extraordinarily up to 1700 Hz (Magariyama et al., 1994), while the rotational speed of the H+-driven flagellar motor is just around 300 Hz (Chen and Berg, 2000).

It is noteworthy that in the *E. coli* H+-driven motor, unexpectedly dynamic properties of the stator units were observed by fluorescent photobleaching experiments: the stator units are not really static components of the motor, but can dynamically associate with and dissociate from the rotor (Leake et al., 2006). Such a dynamic characteristic can be observed also for the PomA/PomB stators of the *Vibrio* polar flagellum. Their assembly around the rotor is dependent on the external Na+ concentration (Fukuoka et al., 2009). Removal of Na+ from the medium caused the dissociation of the stators from the rotor, and the subsequent addition of Na+ to the medium restored the stator incorporation to the motor. The physiological meaning of the dynamic property of the stator is not known, but the rapid and stable rotation of the motor may be compensated by the turnover of the units to components pooled on the membrane.

#### **FLAGELLAR GENE REGULATION AND ASSEMBLY**

About 50 gene products are involved in the construction of a functional flagellum. Since the flagellum is such a big organelle and its production and assembly requires a large commitment of resources, especially for filament formation, bacteria have developed a precise regulation system that controls flagellar construction (MacNab, 2003). The control mechanism has been well-characterized in the system of *Salmonella* flagella (Chilcott and Hughes, 2000). In brief, the flagellum assembles from the inner structure base to the outer ones, beginning with basal body construction followed by hook assembly, and finally by filament formation. Therefore, any defects in gene products that disrupt the basal body or the hook formation inhibit the filament assembly. The hook assembly step is the critical check point: if the hook

subunits are polymerized to the proper length (about 55 nm), then genes required for filament formation are induced. This assembly-coupled flagellar gene regulation is achieved by a cascade of flagellar gene operons (called the flagellar regulon). In *Salmonella*, there are three classes of operons: early, middle, and late (Kutsukake et al., 1990). The master regulator for the flagellar regulon (FlhDC) belongs to the early operon that induces the expression of the middle operon. The middle operon contains genes encoding for basal body and hook proteins. It also contains regulator σ<sup>28</sup> (FliA) that controls the expression of genes belonging to the late operon, which encodes filament and motor proteins (Ohnishi et al., 1990). After completion of the hook assembly, σ<sup>28</sup> is released from anti-sigma factor FlgM binding, and so is able to induce the expression of genes in the late operon (Ohnishi et al., 1992).

and P ring. Next, MotX and MotY assemble around the basal body

In *Vibrio*, the morphogenetic pathway for the polar and lateral flagella is quite likely to be the same as the one identified for *Salmonella*. However, the gene regulation for polar flagellar synthesis is more complex (**Figure 3**). Detailed analyses have been carried out for *V. cholerae* and *V. parahaemolyticus* and have found that the flagellar regulon is composed of a combination of both a RpoN (σ54) and a FlrA/FlaK (master regulator) dependent transcriptional hierarchy, organized into four classes (Klose and Mekalanos, 1998; Klose et al., 1998). Distinct from the *Salmonella* cascade, the middle operons are divided into two classes (class 2 and 3), and two sigma factors are involved in the regulation (Kim and McCarter, 2000). σ<sup>54</sup> controls class 2 and 3 operons, which encode basal body and hook components, and σ<sup>28</sup> (FliA) controls class 4 operons, which encode filament, chemotaxis, and motor components (Stewart and McCarter, 2003). Since *V. alginolyticus* is so closely related to *V. parahaemolyticus*, it is likely to have a similar mechanism that regulates the polar flagellar synthesis, and the putative organization of its polar flagellar regulon is shown in **Figure 3** (McCarter, 2001; Correa et al., 2005; Kojima et al., 2011). It is noteworthy that *Vibrio* species have two chromosomes: the larger one contains all genes involved in the polar flagellar systems, whereas the smaller one contains all genes for the lateral flagellar systems (Makino et al., 2003). Recent work that the flagellar regulatory hierarchy facilitates the correct spatiotemporal expression patterns for optimal *V. cholerae* colonization and disease progression, is consistent with the idea that motility and the expression of specific virulence factors are inversely regulated (Syed et al., 2009).

previous report (Terashima et al., 2013).

Another interesting topic regarding polar flagellar assembly is the regulation of their number and placement: *V. alginolyticus* cells have only a single polar flagellum at the cell pole. Genetic analyses identified two proteins, FlhF and FlhG, which are involved in this regulation (Kusumoto et al., 2008). FlhF is a GTPase and signal-recognition particle (SRP) homolog that positively regulates the number of polar flagella, and also determines the polar positioning (Salvetti et al., 2007; Balaban et al., 2009). FlhG is a MinD homolog and a putative ATPase that negatively regulates the flagellar number (Kusumoto et al., 2006). Biochemical and protein localization studies using GFP fusion derivatives revealed that FlhG binding to FlhF prevents the polar localization of FlhF, so that the proper number of FlhF is localized at the pole to initiate a single polar flagellum (**Figure 4**) (Correa et al., 2005; Kusumoto et al., 2008). From the *flhF flhG* double mutant, which mostly lacks flagella, a suppressor strain was isolated that forms peritrichous flagella in the majority of cells (Kojima et al., 2011). A mutation was recently found in a *dnaJ* family gene, named *sflA* (suppressor of *flhFG*), which changed the monotrichous flagellar *Vibrio* cells into peritrichous flagellated cells (Kitaoka et al., 2013). SflA protein is proposed to prevent the initiation of flagellar construction.

**Kojima et al., 2011).** The arrows indicate the transcription unit and the number at the beginning of arrow indicates the class of transcriptional hierarchy.

flagellum is limited to one. The schemes are drawn based on a previous

report (Kusumoto et al., 2008).

**SENSING THE ENVIRONMENTAL CONDITIONS** As seen in other bacteria like *E. coli*, *Vibrio* species also show chemotaxis, moving toward favorable conditions and avoiding unfavorable environments (**Figure 5**) (Szurmant and Ordal, 2004). The chemotaxis-related genes have been wellcharacterized for *V. cholerae.* Whole genome sequence analysis predicted three *che* gene clusters: cluster I and II on chromosome I and cluster III on chromosome II (Heidelberg et al., 2000). Although many homologs of various chemotaxis components found in *E. coli* have been identified, only one of three chemotaxis

operons, cluster II, is required for chemotaxis in *V. cholerae* (Gosink et al., 2002). Current understanding of chemotactic behavior using a polar flagellar system is similar to that of *E. coli* (**Figure 5**). Once an attractant chemical binds to the periplasmic region of the methyl-accepting chemotaxis proteins (MCPs), a conformational change is generated that triggers autophosphorylation of the cytoplasmic kinase CheA, which is assumed to be associated with MCPs via adaptor protein CheW. Consequently, CheA donates the phosphate group to a response regulator CheY that triggers a rotational switch from counter-clockwise to clockwise by binding to FliM, a component of the switch complex. Meanwhile, some specific residues in the cytoplasmic domain of the MCP will be in a dynamic state of methylation with the help of CheR, a constitutively active methyltransferase, and of CheB, a methylesterase activated by phosphorylation. Methylated MCPs cease the signal output, so that cells are back to the original state (Boin et al., 2004).

The number of chemotaxis genes in *V. cholerae* is much greater than in *E. coli*, but only some of them are functional under experimental conditions (Banerjee et al., 2002). In *V. cholerae*, only deletion of *cheA-2* in cluster II on chromosome I impairs chemotaxis and the *che* genes in this cluster are the most homologous to the *che* genes in other *Vibrio* species, suggesting that *che* cluster II is involved in chemotaxis. Moreover, only the Che-Y3 protein among the 5 *cheY* paralogs showed a relationship to flagellar rotation (Boin et al., 2004; Hyakutake et al., 2005). Furthermore, there are 45 proteins predicted to be potential chemoreceptors. However, only some chemoreceptors such as Mlp24 (*m*ethylaccepting chemotaxis *l*ike *p*rotein, MCP homologs) recently reported were implicated in pathogenicity (Nishiyama et al., 2012). Thus, multiple chemotaxis-related genes and a large number of predicted MCP-like proteins imply a distinct chemotaxis mechanism that allows *V. cholerae* to adapt to various environmental conditions. However, Hyakutake et al. suggested that *che* clusters I and III are perhaps involved in controlling function(s) other than chemotaxis (Hyakutake et al., 2005). With regard to chemotaxis and pathogenicity, a non-chemotactic strain with counter-clockwise-biased flagellar rotation displays increased colonization and infectivity whereas reduced competition of a clockwise-biased strain was found (Butler and Camilli, 2004, 2005). A global transcription profile approach from stool samples showed that expression of many chemotaxis genes is at a low level, but some of them (*cheA-2*, *cheY-3*, *cheW-1*, and *cheR-2*) are not altered (Merrell et al., 2002). CheA-1 is highly expressed during human infection (Hang et al., 2003). Furthermore, it has been reported that human-shed *V. cholerae* have a 10-fold lower oral infectious, and have a transiently reduced chemotactic state in which the protein level of CheW was reduced (Butler et al., 2006). This shows that a pathogen alters its chemotactic state in response to human infection.

In addition to the chemotaxis, which mediates the direction of flagellar rotation, the polar flagellum seems to be a mechanosensor. As described above, *V. alginolyticus* and *V. parahaemolyticus* induce numerous lateral flagella in response to increases in external viscosity. The details of this sensing mechanism are still not understood. Besides an increase in viscosity, lateral flagellar expression is also induced by the addition of an anti-polar flagellum antiserum (McCarter et al., 1988). Since polar flagellar rotation is restricted in these two conditions, the polar flagellum also functions as a mechanosensor to induce lateral flagellar systems. Moreover, it was reported that inhibition of polar flagellar rotation by the specific inhibitor phenamil could induce lateral flagellar expression in media devoid of viscous agents, indicating the possibility that *Vibrio* cells can sense a decrease in the rotation rate of (or sodium influx through) the polar flagellar motor as a trigger for lateral gene expression (Kawagishi et al., 1996). Recently, it was reported that the mechanosensing mechanism by flagella is due to the dynamic assembly of the stator in *E. coli* (Lele et al., 2013; Tipping et al., 2013). Tipping and colleagues showed that the number of the stators bound to the flagellar motor is dependent on the external mechanical load, with more stators at higher viscosity and fewer stators at lower loads (Tipping et al., 2013). The flagellar motor of bacteria senses external load and regulates the strength of stator binding to the rest of the motor. Through this process the flagella can sense the external viscosity, suggesting that they function as a mechanosenseor as well as a locomotion component, and impact cellular functions (Lele et al., 2013). In this regard, it is suggested that the polar flagellum in *Vibrio* perhaps applies the same mechanosensing mechanism to sense and respond to its environment.

#### **PATHOGENESIS AND MOTILITY**

Of particular interest to the pathogenesis of *Vibrio* is to determine the relationships between virulence factors, motility, and biofilm formation. *V. cholerae*, *V. parahaemolyticus*, and *V. vulnificus*, in addition to fish pathogens like *V. alginolyticus* have been established as important pathogens (DePaola et al., 2003; Xie et al., 2005; Jones and Oliver, 2009; Rasmussen et al., 2011; Kim et al., 2013; Ren et al., 2013; Sreelatha et al., 2013). In *V. cholerae*, the major virulence factors are cholera toxin (CT) and toxin co-regulated pilus (TCP), which are mediated in response to environmental stimuli through a hierarchical regulatory cascade called the ToxR regulon (Childers and Klose, 2007). ToxR and TcpP, two membrane-localized transcription factors, activate the expression of ToxT, another important transcription factor in the cytoplasm, which results in the activation of CT and TCP (DiRita et al., 1991; Krukonis and DiRita, 2003). CT can cause severe diarrhea in humans, and is encoded in the CTX prophage (Hassan et al., 2010). A large *Vibrio* pathogenicity island (VPI), encoded in the VPI prophage (Karaolis et al., 1998), is required for the TCP gene cluster, functioning both as an essential colonization factor and as a CTX receptor (Taylor et al., 1987; Herrington et al., 1988; Lowden et al., 2010). However, recent research has shown that the ToxR cascade is also involved in repressing the production of virulence factors, suggesting that the ToxR regulon in *V. cholerae* may play a broader role in pathogenesis (Bina and Bina, 2010; Bina et al., 2013).

With regard to the pathogenesis of *V. parahaemolyticus*, a wide variety of virulence factors, including adhesins, thermostable direct hemolysin (TDH), TDH related hemolysin and Type 3 Secretion Systems, have been well-described by Zhang and Orth (2013). Interestingly, although hemolysin is thought to be a major virulence factor (Makino et al., 2003; Zhang and Austin, 2005), a deletion of the gene encoding both hemolysins does not drastically affect its cytotoxic effects on cultured cells (Zhang and Orth, 2013), suggesting that other virulence factors play important roles in the pathogenesis of *V. parahaemolyticus*. With regard to the pathogenesis of *V. vulnificus*, a recent review described that it possesses a variety of virulence factors, including capsular polysaccharide, acid neutralization, iron acquisition, acid neutralization and expression of proteins involving in motility, attachment and adhesion (Jones and Oliver, 2009).

Several studies have probed the relationship between the motility and the pathogenicity of *V. cholerae in vivo* (Gardel and Mekalanos, 1996; Krukonis and DiRita, 2003; Silva et al., 2006; Syed et al., 2009). It was initially thought that the expression of motility genes may repress the virulence genes (Gardel and Mekalanos, 1996). Mutational analyses have shown that motility is a virulence determinant, since non-motile mutants do not adhere to isolated rabbit brush borders and non-motile strains seem to have attenuated virulence for humans (Richardson, 1991; Silva et al., 2006). However, non-motile mutants have no significant defect in the ability to colonize suckling mice, implying that motility does not affect colonization dependent on TCP (Silva et al., 2006). The expression of motility genes is likely to repress virulence genes and vice-versa (Gardel and Mekalanos, 1996). Subsequently, Häse and coworkers showed that reducing the motility of *V. cholerae* by increasing the medium viscosity or disrupting the Na+-motive force results in an increase in the expression of ToxT (Häse and Mekalanos, 1999). It is noteworthy that the flagellar stator as a mechanosensor responds to assembly around the flagellar rotor in the case of high viscosity and consequently conducts massive Na+ ions influx. Meanwhile, a large number of Na+ ions are inside of cell in the case of disrupting the Na+-motive force. Thus, we can speculate that the influx of Na+ ions through the flagellar stator regulates the expression of ToxT.

With regard to the motility and the pathogenicity, Silva et al. found a remarkable increase in CT and TCP major subunit in a non-motile strain (*motY*), but a decrease in CT production in both wild-type and mutant strains when flagellar motility was inhibited (Silva et al., 2006). Recent work revealed a clear contribution of the flagellar regulatory hierarchy to the virulence of *V. cholerae* (Syed et al., 2009). Virulence factors such as toxin

and hemolysin, which are up-regulated in flagellar regulatory mutants, were confirmed by quantitative reverse transcription PCR. The flagellar regulatory system positively mediates the transcription of diguanylate cyclase, named CdgD which results in the transcription of a hemagglutinin that enhances intestinal colonization (Syed et al., 2009). Thus, this whole-genome expression analysis supports the concept that motility and virulence gene expression are inversely regulated. Meanwhile, it has been found that TCP is expressed before the generation of CT during infection (Lee et al., 1999). These findings specify another transitional phase from motility to pathogenicity: colonization prior to release of CT (**Figure 6A**).

A similar relationship between motility and pathogenicity can be found between motility and biofilm formation in the extra-intestinal environment (**Figure 6B**). Once *Vibrio* cells are colonized, they form a biofilm and are encapsulated and thereby protected (Costerton et al., 1999). Biofilm formation is an important lifestyle of pathogenic *Vibrio* and renders bacteria resistant to environmental stresses, such as antimicrobial compounds or drugs (Yildiz and Visick, 2009; Lasarre and Federle, 2013). Biofilm formation in *V. cholerae* is controlled by quorum sensing (QS) through the modulation of cyclic di-guanosine monophosphate (c-di-GMP) that is used by bacterial pathogens to regulate the expression of genes involved in defense and invasion (Hammer and Bassler, 2003; Zhu and Mekalanos, 2003; Waters et al., 2008). A role of c-di-GMP in biofilm and motility is becoming clear. A *flaA motX* double mutant formed smooth colonies, indicating that flagella are involved in the initial stages of biofilm formation (Lauriano et al., 2004). For biofilm formation in *V. cholerae*, VpsT, a transcriptional regulator, has been reported to control *Vibrio* polysaccharide gene expression and to inversely regulate biofilm formation and motility, via c-di-GMP (Krasteva et al., 2010). Polysaccharide is a very important factor regarding the pathogenicity of *Vibrio* in the case of immune evasion and biofilm formation (**Figure 6B**) (Waldor et al., 1994; Naka et al., 2011; Johnson et al., 2012).

Regarding motility, several groups have shown that in *E. coli* and *Salmonella*, c-di-GMP directly bound to YcgR interacts with components of the flagellar motor to disrupt flagellar rotation, thereby leading to decreased motility (Boehm et al., 2010; Fang and Gomelsky, 2010; Paul et al., 2010). Since homologs of the YcgR protein that contains a binding domain of c-di-GMP extensively exist in *Vibrio*, the same impacting mechanism for the transition from the motile to the sessile phenotype by regulating c-di-GMP also exists in *Vibrio* (Pratt et al., 2007). Interestingly, on one hand, flagellar motility is involved in biofilm formation in the initial stage, but on the other hand, it is repressed when the biofilm is formed. This perhaps implies much broader functions for flagella in addition to motility and mechanosensing. Further, QS also influences motility and polar flagellar biogenesis via a master regulator for QS, SmcR that down-regulates *flhF* expression at the transcriptional level (Kim et al., 2012). Following the loss of motility and the formation of bacterial community, the responses induced by QS signals are usually involved as virulence determinants, such as toxins, proteases, polysaccharides and other relative fitness factors (Rutherford and Bassler, 2012).

#### **SUMMARY**

Overall, this Review has focused on the structure, gene regulation and sensing of the polar flagellum in *Vibrio* spp., and emphasizes the inverse relationship between motility and pathogenicity. However, it is undeniable that motility induced by the polar flagellum in *Vibrio* spp. contributes to the virulence of pathogenic *Vibrio* through adhesion or biofilm formation regardless of the environment. We are interested in understanding how the *Vibrio* flagellar structure and function are involved in the pathogenicity. We also want to know why and how this very complicated structure has evolved and been maintained in *Vibrio*.

#### **REFERENCES**


the crystal structure of MotY. *Proc. Natl. Acad. Sci. U.S.A.* 105, 7696–7701. doi: 10.1073/pnas.0800308105


with other virulent strains of *V. anguillarum and V. ordalii. Infect. Immun.* 79, 2889–2900. doi: 10.1128/IAI.05138-11


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 30 September 2013; paper pending published: 25 October 2013; accepted: 12 December 2013; published online: 25 December 2013.*

*Citation: Zhu S, Kojima S and Homma M (2013) Structure, gene regulation and environmental response of flagella in Vibrio. Front. Microbiol. 4:410. doi: 10.3389/fmicb. 2013.00410*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Zhu, Kojima and Homma. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

**REVIEW ARTICLE** published: 07 March 2014 doi: 10.3389/fmicb.2014.00091

### *Vibrio fluvialis*: an emerging human pathogen

*Thandavarayan Ramamurthy1\*, Goutam Chowdhury1, Gururaja P. Pazhani <sup>1</sup> and Sumio Shinoda2*

*<sup>1</sup> National Institute of Cholera and Enteric Diseases, Kolkata, India*

*<sup>2</sup> National Institute of Cholera and Enteric Diseases, Collaborative Research Center of Okayama University for Infectious Diseases in India, Kolkata, India*

#### *Edited by:*

*Rita R. Colwell, University of Maryland, USA*

#### *Reviewed by:*

*Carlos R. Osorio, University of Santiago de Compostela, Spain Brian Austin, University of Stirling, UK*

#### *\*Correspondence:*

*Thandavarayan Ramamurthy, National Institute of Cholera and Enteric Diseases, P-33, CIT Road, Scheme-XM, Beliaghata, Kolkata-700010, India e-mail: tramu@vsnl.net*

*Vibrio fluvialis* is a pathogen commonly found in coastal environs. Considering recent increase in numbers of diarrheal outbreaks and sporadic extraintestinal cases,*V. fluvialis* has been considered as an emerging pathogen. Though this pathogen can be easily isolated by existing culture methods, its identification is still a challenging problem due to close phenotypic resemblance either with *Vibrio cholerae* or *Aeromonas* spp. However, using molecular tools, it is easy to identify *V. fluvialis* from clinical and different environmental samples. Many putative virulence factors have been reported, but its mechanisms of pathogenesis and survival fitness in the environment are yet to be explored. This chapter covers some of the major discoveries that have been made to understand the importance of *V. fluvialis*.

**Keywords:***V. fluvialis***, diarrhea, virulence factors, antimicrobial resistance, molecular typing**

#### **INTRODUCTION**

*Vibrio fluvialis* is a halophilic Gram-negative bacterium, which has a curved cell morphology and polar flagella for motility. The important biochemical features of this organism include conversion of nitrate to nitrite, do not cleave L-lysine or ornithine, activate arginine dihydrolase, produce indole but not acetoin, ferment sucrose, D-mannitol, L-arabinose, maltose, trehalose, D-galactose, and D-galacturonate. Most of the vibrios, including *V. fluvialis* occur widely in the aquatic milieu, mostly in the seas, estuaries and brackish waters. Even though more than 100 spices have been reported in the Genus *Vibrio* (http://www.bacterio.net/uw/vibrio.html), about 13 of them have been reported to cause several human diseases. Among the pathogenic vibrios, *V. alginolyticus, V. cholerae, V. costicola, V. mimicus, V. cincinnatiensis, V. hollisae, V. furnissii, V. parahaemolyticus, V. vulnificus, V. carchariae* (a junior synonym of *V. harveyi*) and *V. metschnikovii* are clinically important as they cause different types of vibriosis. One of the *Vibrio* spp., *V. damselae* has now been renamed as "*Photobacterium damselae* subsp. *damselae*." The toxigenic *V. cholerae*, *V. parahaemolyticus* and *V. vulnificus* are associated with well-known cholera and diarrhea and extraintestinal infections, respectively. Prevalence of *V. cholerae* in developing countries is mostly related to the breakdown of sanitary conditions and/or due to scarcity of drinking water. On the other hand, infections caused by *V. parahaemolyticus* and other vibrios denote contamination of seafood in many countries, irrespective of their economic conditions.

*V. fluvialis* is one of the emerging foodborne pathogens all over the world. The distribution of virulence factors and molecular epidemiological features of this pathogen remain mostly unknown. Among the foodborne infections in the United States, there has been a considerable increase (43%) in the *Vibrio*-mediated infections till 2012 compared with the rates reported during 2006–2008 (Centers for Disease Control and Prevention (CDC), 2013). Several recent publications indicate the epidemiological importance of *V. fluvialis* (Chowdhury et al., 2012; Liang et al., 2013).

#### **IDENTIFICATION AND TAXONOMY**

Thiosulfate-citrate-bile salts-sucrose agar (TCBS) has been conventionally used as a selective medium for the isolation of clinically important vibrios. The colony morphology of *V. fluvialis* in this medium remains indistinguishable from *V. cholerae,* i.e., it grows as sucrose fermenting yellow color colonies after direct plating of clinical specimens or after enrichment in alkaline peptone water (pH 8.0). After preliminary screening in the TCBS, a battery of biochemical testes is essential for the species-specific identification of *V. fluvialis*. Minimal biochemical tests such as lysine decarboxylase, ornithine decarboxylase, arginine dihydrolase, and L-arabinose are mandatory for the identification of *V. fluvialis*. Without these minimal tests, the identification is incomplete and the isolate will be improperly classified as*V. cholerae* or *Aeromonas* spp. In most resource-poor countries, these tests are not methodically performed, which may lead to labeling of *V. fluvialis* as *V. cholerae*. Considering such situation, there is a high possibility that the *V. fluvialis* could be reported as *V. cholerae* non-O1, non-O139 or non-agglutinable vibrios (NAGs). It is worth to mention here that *V. cholerae* O1 and O139 serogroups can be easily confirmed by slide agglutination with corresponding antiserum.

For the identification of *V. fluvialis* and other vibrios, rapid identification kits must be used with caution as they need additional tests for the final confirmation. While testing the commercially available identification kits,*V. fluvialis* remain as a major challenge with API 20E and Vitek GNI+ systems (Israil et al., 2003; O'Hara et al., 2003). Biochemically, *V. furnissii* expresses fibrin and mucin hydrolysis but no phosphate or esculin hydrolysis, for which *V. fluvialis* varied. *V. fluvialis*,*V. furnissii,* and *V. mimicus* are distinctive from *V. cholerae*, as the later exhibit strong mannosesensitive hemagglutination. These test results may have a strong influence in the confirmation of strains.

Molecular tools such as PCR are useful in the identification of many uncommon vibrios and most of these assays are comparable to the conventional identification methods. The sequence of amplified 16S–23S intergenic spacers (IGSs) has demonstrated 37 ribosomal RNA (*rrn*) operons representing seven different IGS types in different *Vibrio* spp. with IGS(0), IGS(IA), and IGS(Glu) as major ones. The sequence difference in these IGS types was used to design species-specific primers for PCR for *V. fluvialis* and other vibrios (Lee et al., 2002). In some of the reports, a universal primer PCR that covers conserved regions of bacterial 16S rRNA genes followed by denaturing gradient gel electrophoresis (DGGE) was found to be useful in the identification of *V. fluvialis* either as axenic bacteria or mixed with other pathogens (Ji et al., 2004).

Initially,*V. furnissii* was taxonomically assigned with *V. fluvialis* and named as aerogenic biogroup of *V. fluvialis*. Based on DNA relatedness and several biochemical tests,*V. furnissii* has been separated as a new species (Lee et al., 1981; Brenner et al., 1983). In the phylogenetic analysis with several housekeeping genes, *V. furnissii* and *V. fluvialis* have been linked as close species. The nucleotide comparison of 16S-rRNA, *recA,* and *toxR* sequences showed that *V. furnissii* and*V. fluvialis* had 100% similarity. The gene *toxR* of *V. fluvialis* had 84% similarity with*V. harveyi* (Franco and Hedreyda, 2006). With the *gyrB*, *V. cholerae, V. mimicus, V. furnissii*, and *V. fluvialis* shared 93% sequence similarity.

Toxigenic vibrios have a homolog of the *toxRS* operon, which regulates the virulence expression. The gene *toxR* encodes a transcriptional activation domain (TAD), a transmembrane domain (TMD), and a periplasmic domain (PD). Among the vibrios, there is essentially no homology within the region between TAD and TMD. Hence, this region has been used in designing of primers for the species-specific identification of many vibrios. Chakraborty et al. (2006) described a species-specific identification of *V. fluvialis* by PCR targeted to the conserved transcriptional activation and variable membrane tether regions of the *toxR* gene. The functional virulence genes encoding hemolysin (*vfh*), heme-utilization (*hupO*), and central regulation (*vfpA*) have been used as targets in an multiplex PCR for the identification of *V. cholerae*, *V. parahaemolyticus*, and *V. fluvialis*, respectively (Vinothkumar et al., 2013). For the detection of clinical vibrios in seafood samples, a multiplex primer-extension reaction (PER) assay targeting the *rpoA* gene has also been reported (Dalmasso et al., 2009).

Pyrolysis-mass spectrometry with metastable atom bombardment and pattern recognition seemed to be suitable for the identification of *V. fluvialis* and other vibrios (Wilkes et al., 2005). The mass spectra have been generated via an alternative ionization method, metastable atom bombardment followed by component-discriminant analysis. Since the outer membrane protein K (OmpK) of *V. fluvialis*, *V. alginolyticus*, *V. mimicus*, *V. parahaemolyticus*, and *V. vulnificus* is highly similar, the antibodies against these proteins have been proposed in the diagnosis (Li et al., 2010). The whole cell protein profile using SDS-PAGE was also considered in the identification of clinically important vibrios including *V. fluvialis* (Lee et al., 2012).

Since simple phenotypic diagnostic tests are not available, Chen et al. (1995) used species-specific bacteriophages as a tool

for the identification of *V. fluvialis* and with a set of phages, the diagnostic probability of human isolates was more than 84%. At least in one study, the importance of phage-typing of *V. fluvialis* has been demonstrated using six specific bacteriophages with 73% typability (Suthienkul, 1993). However, availability of these bacteriophages makes this assay technique less popular.

#### **PHENOTYPIC AND GENETIC CHARACTERISTICS OF** *V. fluvialis*

Based on the somatic antigen variation, several serotypes of *V. fluvialis* have been identified. Though Shimada et al. (1999) identified more than 50 somatic antigens, the serological based typing of *V. fluvialis* remains non-customary. *V. fluvialis* strains belonging to serogroup O19 possessed the C (Inaba) antigen of *V. cholerae* O1, but not the B (Ogawa) or A (common) antigens (Shimada et al., 1987; Kondo et al., 2000). In the crossed immuno-electrophoresis, antibodies against the oral cholera vaccines containing killed whole cells (WC) of *V. cholerae* O1 Inaba El Tor reacted with a few strains of *V. fluvialis* (Ciznãr et al., 1989). Presence of shared WC antigens indicates that the oral cholera vaccine could stimulate immunity effectively against other vibrios also. It is known that the antigenic nature of flagella of vibrios is highly homologous. Tassin et al. (1983) and Shinoda et al. (1984) demonstrated independently that anti-L-flagella antisera of *V. fluvialis* did not agglutinate other *Vibrio* species in the H-agglutination tests. Further studies placed*V. fluvialis* and*V. furnissii* in the same lateral flagellar serogroup-HL8 (Shinoda et al., 1992). However, in practice, serotyping based on H-flagella is also not in use.

A chemotaxonomic study based on sugar composition of the polysaccharide portion of their lipopolysaccharide (LPS) has divided 35 O-antigen groups of *V. fluvialis* into 21 chemotypes (Iguchi et al., 1993). This seems to be a unique finding since the D-glycero-D-manno-heptose, and two kinds of uronic acids, i.e., galacturonic and glucuronic acids are rare in Gram-negative bacteria. In addition, 2-keto-3-deoxyoctonate, which is a typical sugar component of Gram-negative bacterial LPS was not detectable in any of the chemotypes.

Of all the molecular typing methods, the pulsed-field gel electrophoresis (PFGE) has proven to be highly useful in tying the bacterial isolates. Unlike *V. cholerae* O1 and pandemic *V. parahaemolyticu*s, the isolates of *V. fluvialis* from acute diarrheal patients exhibited large genetic diversity (Chowdhury et al., 2012, 2013).

#### **PREVALENCE OF** *V. fluvialis* **IN THE AQUATIC REALM**

Even though the presence of vibrios is mostly documented from coastal environs, the domination of a particular species depends on many physico-chemical and biological factors. In warmer regions like Florida, USA, *V. fluvialis* was predominantly detected in sediments during winter months (Williams and Larock, 1985). Due to rise in seawater temperature, the identification rate of *V. fluvialis* has increased considerably (29%) in several niches at the Toulon harbor, France (Martin and Bonnefont, 1990). However, in Chesapeake Bay, *V. fluvialis* infections are always less during winter months, indirectly reflecting its minimal occurrence in

this season (Hoge et al., 1989). *V. fluvialis* along with *V. vulnificus* and *V. cholerae* non-O1 unusually existed in the Seto Inland Sea of Japan, which is a eutrophic zone with riverine influence (Venkateswaran et al., 1989a). In South East Queensland, Australia, next to *V. cholerae* (10.2%), *V. fluvialis* (8.2%) has been isolated more frequently from river waters, sediments, and plants (Myatt and Davis, 1989).

Due to high load of pollution in the upstream of the river Ganges, presence of *V. fluvialis* (0.74%) with other potential pathogens have been detected in several points of Varanasi, India (De et al., 1993). *V. fluvialis* has also been isolated from natural waters in Myanmar (Oo et al., 1993) and in a wide range of coastal environments of Japan (Uchiyama, 2000). Compared to other vibrios, the recovery of *V. fluvialis* has been high (41.4%) from suburban community effluents in South Africa. However, their occurrence was not associated with any season or plankton blooms, but positively correlated with temperature, salinity, and dissolved oxygen (Igbinosa et al., 2011a).

In many investigations, the detection frequency of *V. fluvialis* was very high in marine mollusks, mostly in bivalves, as they accumulate large number of pathogens during the process of filter-feeding. Findings of Kelly and Stroh (1988) from Pacific Northwest showed that oysters are the main source of *V. fluvialis* and other vibrios especially during warmer seasons. In Hong Kong, *V. fluvialis* was one of the important pathogenic vibrios identified in coastal waters and seafood sold in the markets (Chan et al., 1986, 1989). *V. fluvialis* has been isolated from mussels from Senegal (Schandevyl et al., 1984), Brazil (Matté et al., 1994), bivalves and mud from Costa Rica (García and Antillón, 1990) and cultured fishes from Denmark (Pedersen et al., 1999), copepods from Southern Italy (Dumontet et al., 2000) and cockles of Malaysia (Elhadi et al., 2004). In Turkey, next to *V. alginolyticus* (>30%), *V. fluvialis* was the most common *Vibrio* in blue crabs and retail fishes (>10%; Yalcinkaya et al., 2003; Yücel and Balci, 2010).

Generally, fecal pollution has been monitored in aquaculture areas to forecast human pathogens in the products. In Italy, about 11–27% of the mollusks and shrimps contained *V. fluvialis* without any association between presence of this pathogen and conventional fecal pollution indicators (Ripabelli et al., 2004). The micro fauna and flora occasionally support the occurrence human pathogens. *V. fluvialis* (36.5%) was significantly associated with plankton in the effluents of a rural wastewater treatment facility in the Eastern Cape Province of South Africa (Igbinosa et al., 2009). In the Atlantic coast of France, Deter et al. (2010) showed that chlorophyll-A had a significant influence on pathogenic vibrios including *V. fluvialis* in mussels.

There are few reports about identification of *V. fluvialis* from wound infections that took place in recreational areas. Since *V. fluvialis* has been cultured from the teeth of a great white shark (*Carcharodon carcharias*), there may be an association of this pathogen with wound infections caused by sharks in humans (Buck et al., 1984). Fibropapillomatosis (FP) is a mutilating disease among turtles that cause tumors on the skin and other internal organs. In a study conducted by Aguirre et al. (1994) showed the presence of *V. fluvialis* (47%) in green turtles with FP.

In the marine environment, *V. fluvialis* plays a major role in the production of hydrogen from starch acquired from the algal mass in the presence of *Rhodobium marinum*. In co-culture experiments,*V. fluvialis* degrade starch leading to the formation of acetic acid and ethanol, which are subsequently utilized for hydrogen production by *R. marinum* (Ike et al., 1999).

#### **SPORADIC CASES AND OUTBREAKS OF DIARRHEA DUE TO** *V. fluvialis*

Early reports from the US indicated involvement of *V. fluvialis* with gastroenteritis among infants (Hickman-Brenner et al., 1984; Bellet et al., 1989; Kolb et al., 1997). Since 1979, *V. fluvialis* was isolated as one of the important pathogens in Tenri Hospital, Japan (Aihara et al., 1991). Prevalence of *V. fluvialis* among children with diarrhea was very less during 1988 (0.6%) in Calcutta (now, Kolkata), India (Chatterjee et al., 1989). In the same region, progressive increase in the prevalence of *V. fluvialis* (>2%) among hospitalized acute diarrheal patients has been reported in the following years (Chowdhury et al., 2012). During 1996–1998, prevalence of *V. fluvialis* was 9.4% among hospitalized diarrheal patients in North Jakarta (Lesmana et al., 2002). In Zhejiang Province, China, *V. fluvialis* was identified as the second most pathogen (12%) among acute diarrheal cases but next to *V. parahaemolyticus* (64%; Jiang, 1991). Investigations carried out after the 1998 floods in Bangladesh showed involvement of *V. fluvialis* in a diarrhea outbreak (Tanabe et al., 1999). However, the number of cases was less compared to *V. cholerae* O1 and O139 infections.

*Vibrio-*mediated infections frequently occur in countries where the raw seafood is largely consumed. In many instances, *V. fluvialis* was found to be associated with cholera-like diarrhea (Allton et al., 2006). Between 1982 and 1988, 10 gastroenteritis cases of *V. fluvialis* have been reported in Florida due to consumption of contaminated seafood (Klontz and Desenclos, 1990). In the Gulf coast, the majority of the *Vibrio*-mediated gastroenteritis has been associated with intake of raw oysters and in about 6% of the cases *V. fluvialis* was the causative pathogen (Levine and Griffin, 1993). Foodborne outbreaks were reported in several communities implicating *V. fluvialis* alone or with either *V. parahaemolyticus/Salmonella* spp. (Tokoro et al., 1984; Chowdhury et al., 2013).

Foodborne diarrheal outbreaks caused by *V. fluvialis* have been reported during 1981 in Maharashtra (Thekdi et al., 1990) and 2012 in Kolkata (Chowdhury et al., 2013). In Brazil, and USSR, the first report on the association of *V. fluvialis* with diarrhea was reported during 1990 and 1991, respectively (Magalhães et al., 1990; Libinzon et al., 1991). Though the incidence of cholera among high socioeconomic population in Brazil was very low (0.07%), but the other vibrios including *V. fluvialis* comparatively prevailed more (1.2%; Magalhães et al., 1993). In Volga delta, Russia, acute enteric infections caused by *V. fluvialis* reaches about 30% during the summer months, mainly due to consumption of water than sea/fresh water fishes (Bo˘iko, 2000). Among travelers with diarrheal symptoms, the incidence of *V. fluvialis* seems to be low compared to other enteric pathogens. Early studies conducted with US Peace Corps volunteers in Thailand identified *V. fluvialis* in about 3% of the cases (Taylor et al., 1985).

#### **OTHER INFECTIONS**

*Vibrio fluvialis* causes a variety of infections in immunecompetent/HIV patients, including bacteremia, biliary tract infection and acute diarrhea (Albert et al., 1991; Usó et al., 2010; Liu et al., 2011). The other rarely reported infections caused by this pathogen include suppurative cholangitis (Yoshii et al.,1987), peritonitis (Lee et al., 2008), acute otitis (Cabrera et al., 2005; Chen et al., 2012) and endophthalmitis (Penland et al., 2000). Large numbers of (29%) endophthalmitis patients were reported to have mixed infection with *V. fluvialis* (Hassan et al., 1992). A report from Cuba showed that *V. fluvialis* was one of the predominantly identified pathogens from different extraintestinal samples (Cabrera et al., 2007). Cases of bacteremia with diarrhea (Lai et al., 2006) hemorrhagic cellulitis and cerebritis (Huang and Hsu,2005), peritonitis (Ratnaraja et al., 2005) have also been reported.

#### **QUORUM SENSING**

Quorum sensing (QS) is a process in which bacterial cells in a population are able to crosstalk with one another, thereby supporting them as a unit to synchronize gene regulation and consequent phenotypic changes. The importance of QS in pathogenic *V. cholerae* has been well established. Wang et al. (2013) have shown that QS in *V. fluvialis* regulates two potential virulence factors, including an extracellular protease and hemolysin. In addition, QS also regulates *in vitro* cytotoxic activity against epithelial cell lines.

#### **VIRULENCE FACTORS**

The clinical as well as environmental *V. fluvialis* strains express many putative virulence factors. The common virulence factor in *V. fluvialis* reported in several investigations is the expression of hemolysin that can be easily identified in sheep-blood agar plates. In majority of the toxin detection assays, eukaryotic cell lines are being used *in vitro*. In cell-free extracts, *V. fluvialis* has expressed Chinese hamster ovary (CHO) cell elongation factor, CHO cellkilling factors, cytolysins against erythrocytes and proteases active against azocasein (Lockwood et al., 1982). Various putative virulence factors of *V. fluvialis* are presented in **Table 1**. However, the ability to produce these factors is not uniform in all the isolates (Liang et al., 2013).

Purification of cytotoxin produced by *V. fluvialis* showed that the protein was heat-labile, and deactivated by proteases. The



culture supernatant retained hemolytic and phospholipase A2 activities and were coeluted in the gel filtration (Wall et al., 1984). The purified extracellular hemolysin produced by *V. fluvialis* showed virulence features including lyses of erythrocytes of different animal origin and activation of fluid accumulation in suckling mice (Han et al., 2002; Kothary et al., 2003).

The transmembrane regulatory protein (ToxR) is essential for the expression of virulence factors in pathogenic vibrios. Similar to *V. cholerae*, the ToxR plays a major role in bile resistance of *V. fluvialis*, which is an initial phase in the progression of vibrios as potential intestinal pathogens (Provenzano et al., 2000). Adaptability of vibrios to the intestinal environment, especially the bile salts favors colonization and expression of virulence factors. After initial adaptation to the bile salts under *in vitro* conditions, the *V. fluvialis* exhibited swarming mobility, biofilm formation and adherence (Di Pietro et al., 2004). In the animal models, *V. fluvialis* and the cholera toxin (CT) produced by *V. cholerae* O1 strains confirmed skin permeability factor (SPF). However, the antibodies against CT did not neutralize the SPF of *V. fluvialis* (Rodrigues et al., 1993; Ahsan et al., 1988).

The exocellular metalloprotease produced by *V. fluvialis* (VFP) was found to be similar to the one produced by*V. vulnificus*, which has also been used for the hemagglutination activity (Miyoshi et al., 2002). In addition, the amino acid sequence of VFP was found to be a member of the thermolysin family. It is interesting to note that most of the *V. fluvialis* isolated from the diarrheal patients harbored genes encoding hemolysin and metalloprotease (Chowdhury et al., 2012).

#### **SURVIVAL**

*Vibrio fluvialis* has the capacity to survive in the seawater microcosm for more than 15 days at ambient temperature regardless of carbonated substrate uptake (Munro et al., 1994). In microcosms, *V. fluvialis* has been shown culturally viable for a year without losing its virulence and in sediments this organism was recovered from viable but non-culturable stage, even after 6 years (Amel et al., 2008).

#### **ANTIMICROBIAL RESISTANCE**

Compared to other clinical vibrios, antimicrobial resistance (AMR) is largely reported in *V. fluvialis*. In Mediterranean fish farms, many of the vibrios including *V. fluvialis* were resistant to ampicillin, carbenicillin, kanamycin, cefalotin, and sulfadiazine-trimethoprim (Laganà et al., 2011). In South Africa, treated effluent system was found to be the reservoir for *V. fluvialis* strains, which are resistant to ampicillin, penicillin-G, streptomycin, sulfamethoxazole, trimethoprim, chloramphenicol, erythromycin, ciprofloxacin, and polymyxin B (Igbinosa et al., 2011b). In China, majority of the *V. fluvialis* strains were resistant for β-lactams, azithromycin, and sulfamethoxazole (Liang et al., 2013).

Several mobile genetic elements carrying AMR have been found in *V. fluvialis.* The integrative and conjugative element (ICE) is a conjugative transposon commonly detected in *V. cholerae*, which carries resistance genes for sulfamethoxazole-trimethoprim (SXT), chloramphenicol and streptomycin (Srinivasan et al., 2006; Taviani et al., 2008). This SXT element has also been reported

in *V. fluvialis* that has integrase gene similar to that of *V. cholerae* (Ahmed et al., 2005). The aminoglycoside acetyltransferase encoding gene *aac(3)-Id* was identified in class 1 integron from a clinical *V. fluvialis* strains (Ahmed et al., 2004).

Transfer of large plasmids carrying AMR genes is rarely detected in *V. fluvialis* (Rajpara et al., 2009). Efflux systems responsible for nalidixic acid and ciprofloxacin resistance have been reported in several clinical *V. fluvialis* strains (Srinivasan et al., 2006). Two putative multi antimicrobial extrusion (MATE) protein family efflux pumps viz., H- and D-type were found to be responsible for fluoroquinolones resistance in *V. fluvialis*. The sequences of these MATE encoding genes were found to be ∼99% identical to *V. cholerae* (Mohanty et al., 2012). In addition, many *V. fluvialis* strains had mutation (serine to isoleucine) at position 83 of the quinolone resistance-determining region (QRDR) of *gyrA*. Apart from this mutation, presence of plasmid-borne *qnrVC*-like genes have been reported for quinolone resistance in some of the *V. fluvialis* strains (Singh et al., 2012). *V. fluvialis* isolated from diarrheal patients in Kolkata were resistant to fluoroquinolones and β-lactam antimicrobials had mutations in the QRDR of GyrA at position 83 and of ParC at position 85 (Chowdhury et al., 2011). In addition, these strains carried a transferrable 150-kb plasmid that harbored the quinolone resistance *qnrA1* in a complex *sul1*-type integron, the ciprofloxacin-modifying enzyme-encoding gene *aac(6 )-Ib-cr* and genes encoding for extended-spectrum β-lactamases such as *blaSHV* and *blaCTX*−*M*−3.

#### **CONCLUSION**

Though the pathogen *V. fluvialis* has known for quite some time, its clinical importance is realized now, as the prevalence of diarrhea cases is reportedly increasing. In depth studies on the pathogenesis of *V. fluvialis* has to be established as there are many descriptions about the putative virulence factor.

#### **ACKNOWLEDGMENTS**

This work was supported by the Japan Initiative for Global Research Network for Infectious Diseases, Ministry of Education, Culture, Sports, Science and Technology, Japan.

#### **REFERENCES**


caused by dual pathogens, *Salmonella enterica* serovar Weltevreden and *Vibrio fluvialis* in Kolkata, India. *Foodborne Pathog. Dis.* 10, 904–906. doi: 10.1089/fpd. 2013.1491


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 17 December 2013; accepted: 19 February 2014; published online: 07 March 2014.*

*Citation: Ramamurthy T, Chowdhury G, Pazhani GP and Shinoda S (2014) Vibrio fluvialis: an emerging human pathogen. Front. Microbiol. 5:91. doi: 10.3389/fmicb.2014.00091*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Ramamurthy, Chowdhury, Pazhani and Shinoda. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## New *Vibrio* species associated to molluscan microbiota: a review

#### *Jesús L. Romalde\*, Ana L. Diéguez, Aide Lasa and Sabela Balboa*

*Departamento de Microbiología y Parasitología, CIBUS-Facultad de Biología, Universidad de Santiago de Compostela, Santiago de Compostela, Spain*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Patrick Monfort, Centre National de la Recherche Scientifique, France Eva Benediktsdóttir, University of Iceland, Iceland*

#### *\*Correspondence:*

*Jesús L. Romalde, Departamento de Microbiología y Parasitología, CIBUS-Facultad de Biología, Universidad de Santiago de Compostela, Campus Vida s/n, Santiago de Compostela 15782, Spain e-mail: jesus.romalde@usc.es*

The genus *Vibrio* consists of more than 100 species grouped in 14 clades that are widely distributed in aquatic environments such as estuarine, coastal waters, and sediments. A large number of species of this genus are associated with marine organisms like fish, molluscs and crustaceans, in commensal or pathogenic relations. In the last decade, more than 50 new species have been described in the genus *Vibrio*, due to the introduction of new molecular techniques in bacterial taxonomy, such as multilocus sequence analysis or fluorescent amplified fragment length polymorphism. On the other hand, the increasing number of environmental studies has contributed to improve the knowledge about the family*Vibrionaceae* and its phylogeny. *Vibrio crassostreae*, *V. breoganii*, *V. celticus* are some of the new *Vibrio* species described as forming part of the molluscan microbiota. Some of them have been associated with mortalities of different molluscan species, seriously affecting their culture and causing high losses in hatcheries as well as in natural beds. For other species, ecological importance has been demonstrated being highly abundant in different marine habitats and geographical regions. The present work provides an updated overview of the recently characterized *Vibrio* species isolated from molluscs. In addition, their pathogenic potential and/or environmental importance is discussed.

**Keywords:***Vibrionaceae***, genus** *Vibrio,* **molluscan microbiota, new species, pathogenicity, ecology**

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 1 — #1

#### **INTRODUCTION**

Coastal and estuaries environments are growing areas of bivalve molluscs which become an important industry in many countries, due to the increasing importance of these animal as protein for human consumption. Due to their filter-feeding habit, bivalves accumulate a rich and diverse bacterial microbiota, composed of various species belonging to different genera like *Vibrio*, *Pseudomonas*, *Acinetobacter*, *Photobacterium*, *Moraxella*, *Aeromonas*, *Micrococcus*, and *Bacillus* (Murchelano and Brown, 1970; Kueh and Chan, 1985; Prieur et al., 1990).

Vibrios are Gram-negative curved rods that occur naturally in marine, estuarine, and freshwater systems worldwide. They occupy habitats ranging from the deep sea to shallow aquatic environments (Reen et al., 2006), being some species important for natural systems, including the carbon cycle and osmorregulation (Johnson, 2013), as free-living inhabitants in the water column or associated to particulate material. On the other hand, vibrios are also associated with a wide variety of poikilotherm and homoiotherm animals, including humans, for some of which are pathogens. Paillard et al. (2004) recognized that the emergence of vibrios as etiological agents of diseases is likely to increase over the coming years due to ocean warming.

The repeated episodes of mortality due to bacterial infections constitute one of the main problems in the culture of bivalve molluscs, since they reduce the production and cause high economical losses. Some members of the genus *Vibrio* have been described as the main aetiological agents of diseases affecting all life stages of molluscan shellfish (Liu et al., 2000; Allam et al., 2002; Waechter et al., 2002; Lee et al., 2003; Anguiano-Beltrán et al., 2004; Estes et al., 2004; Gay et al., 2004a,b; Paillard et al., 2004; Gómez-León et al., 2005; Prado et al., 2005; Labreuche et al., 2006a,b; Garnier et al., 2007, 2008).

The aim of the present work is to provide an overview on the diversity of *Vibrionaceae* associated with bivalve molluscs. Special emphasis is made on the species described in the last years in an attemp to clarify, not only their taxonomy, but also their pathogenic or ecological importance.

#### **VIBRIOS AS MICROBIOTA OF BIVALVE MOLLUSCS**

In the literature, studies analyzing the diversity, distribution, and density of marine bacteria associated with bivalve molluscs are scarce. These studies date back to the 1960s and in general, results agree in the dominance of Gram negative over Gram positive bacteria in molluscs, as well as in the high abundance of bacteria belonging to the genus *Vibrio* (Colwell and Liston, 1960; Beenson and Johnson, 1967; Kueh and Chan, 1985).

From 1990s the diversity of *Vibrio* species associated with bivalves in different geographical areas has been the subject of various studies (Montilla et al., 1994; Hariharan et al., 1995;Arias et al., 1999; Pujalte et al., 1999; Maugeri et al., 2000; Caballo and Stabili, 2002; Castro et al., 2002; Guisande et al., 2004; Beaz-Hidalgo et al., 2008, 2010a; Lafisca et al., 2008). The main common conclusions obtained from these studies were that environmental parameters, such as variations in the water temperature and salinity, can influence the diversity of *Vibrio* spp. in the environment, as well as the physiological state of the bivalve and its susceptibility to bacterial infections (Arias et al., 1999; Pujalte et al., 1999; Maugeri et al., 2000; Paillard et al., 2004; Garnier et al., 2007).

In most studies, the predominating species associated with bivalves from different geographical locations (Spain, Canada, Italy, or Brazil), all from temperate climates, were either *V. splendidus*, *V. alginolyticus*, *V. harveyi*, or any combination of these species (Montilla et al., 1994; Arias et al., 1999; Pujalte et al., 1999). More recently, Beaz-Hidalgo et al. (2008) analyzed the diversity of *Vibrio* spp. in cultured Manila clams (*Venerupis philippinarum*) and carpet-shell clams (*Venerupis decusata*) by means of phenotypic and genotypic methods. The predominant species that accounted for 66.6% of the total identified strains were *Vibrio cyclitrophicus*,*V. splendidus*, and*V. alginolyticus*. Other species such as*V. fluvialis*,*V. vulnificus,* and*V. mimicus* have also been associated with molluscs (Maugeri et al., 2000; Caballo and Stabili, 2002). In those studies, the identification of *Vibrio* species was established using only phenotypic methods and, therefore, the real diversity present in bivalves could be underestimate. In fact, a high phenotypic variability was described within the *V. splendidus*-like and the *V. harveyi*-like groups which makes impossible to discriminate among several species (Thompson et al., 2005; Le Roux and Austin, 2006; Pascual et al., 2010).

As mentioned, most of the studies described the influence of environmental parameters (i.e., salinity and water temperature) on the diversity and alternance of *Vibrio* species (Kaspar and Tamplin, 1993; Motes et al., 1998; Arias et al., 1999; Pujalte et al., 1999). For instance, in bivalves from the Mediterranean Sea, *V. splendidus* has been found to be dominant during winter and spring and *V. harveyi* during the warmer months (Arias et al., 1999; Pujalte et al., 1999). Another example, in shellfishgrowing areas of the US Northern Gulf Coast, the densities of *V. vulnificus* were high and almost constant at temperatures above 26◦C and/or at salinity below 25 ppt, but decreased drastically below this temperature and/or above this salinity (Motes et al., 1998). The latter species together with *Vibrio parahaemolyticus* and *V. cholerae* are considered important human pathogens, producing important outbreaks after the consumption of contaminated shellfish (mainly oysters), and therefore have been the subject of many studies (Morris, 2003; Su and Liu, 2007; Jones and Oliver, 2009).

#### **NEW** *Vibrio* **SPECIES ASSOCIATED TO MOLLUSCS**

The introduction of molecular techniques such as the fluorescent amplifiedfragment length polymorphism (FAFLP) and multilocus sequence analysis (MLSA) has allowed a more precise identification of *Vibrio* species which were previously masked under other taxa (Thompson et al., 2001, 2005; Beaz-Hidalgo et al., 2008, 2010a; Pascual et al., 2010).

In this sense, molecular studies have demonstrated the genetic diversity and the polyphyletic nature of *V. splendidus* (Thompson et al., 2001, 2005; Le Roux et al., 2002) and have enabled many new species to be described, such as *Vibrio kanaloae*, *V. pomeroyi*, *V. chagasii,* or *V. gallaecicus* (Thompson et al., 2003c; Beaz-Hidalgo et al., 2009b). Furthermore, phenotypically identified *V. harveyi* strains were re-classified as *V. campbellii* by FAFLP, DNA–DNA hybridization (DDH), and MLSA (Gomez-Gil et al., 2004a; Thompson et al., 2007). Pascual et al. (2010) investigated the usefulness of an MLSA approach with six housekeeping genes to discriminate six tightly related species with DDH values close to 70%, namely*V. harveyi*,*V. campbellii*,*V. rotiferianus*,*V. parahaemolyticus*,*V. alginolyticus*, and*V. natriegens*. They recognized the genes*tox*R (cholera toxin transcriptional activator) and *rpo*D (Sigma factor σ70) as the most reliable for species identification, and proposed a scheme for species definition on the basis of the similarities of the concatenated sequences of the most resolving genes.

In the last decade, more than 50 new species have been described in the genus *Vibrio*, many of them associated to marine environments and aquatic eukaryotic organisms. To mention some examples, among the new species described as free-living seawater bacteria are *V. agarivorans* (Macián et al., 2001b),*V. ruber* (Shieh et al., 2003),*V. aestivus* and*V. quintilis* (Lucena et al., 2012), *V. azureus* (Yoshizawa et al., 2009), or *V. sagamiensis* (Yoshizawa et al., 2010). Associated to different marine organisms have been described, among others, *V. caribbeanicus* from sponges (Hoffmann et al., 2012), *V. hemicentroti* from sea urchin (Kim et al., 2013), *V. corallilyticus*, *V. maritimus*, *V. shiloi*, *V. stylophorae*, and *V. variabilis* from corals (Kushmaro et al., 2001; Ben-Haim et al., 2003; Chimetto et al., 2011; Sheu et al., 2011), *V. rotiferianus* from rotifers (Gomez-Gil et al., 2003a), *V. comitans*, *V. gallicus*, *V. inusitatus*, *V. neonates*, *V. rarus*, and *V. superstes* from abalones (Hayashi et al., 2003; Sawabe et al., 2004a,b, 2007a), *V. atypicus*, *V. hispanicus*, *V. jasicida*, *V. owensii*, *V. pacinii*, *V. zhanjiangensis*, and *V. zhuhaiensis* from crustaceans (Gomez-Gil et al., 2003b, 2004b; Cano-Gómez et al., 2010; Wang et al., 2010; Jin et al., 2012, 2013; Yoshizawa et al., 2012), *V. hippocampi* from sea horses (Bálcazar et al., 2010), and *V. alfacsensis*, *V. sinaloensis*, and *V. tasmaniensis* from fish (Thompson et al., 2003d; Gomez-Gil et al., 2008, 2012).

Regarding the vibrios described as associated with bivalve molluscs, and beside the well known species *V. alginolyticus*, *V. harveyi*, *V. mytili*, *V. parahaemolyticus*, *V. pectenicida,* or *V. vulnificus* (Pujalte et al., 1993; Lambert et al., 1998; Arias et al., 1999; Pujalte et al., 1999; Maugeri et al., 2000; Caballo and Stabili, 2002;Paillard et al., 2004; Beaz-Hidalgo et al., 2010a; Romalde et al., 2013), since the turn of the century 19 new species and 2 new subspecies have been described within the genus*Vibrio* (**Figure 1**). These new species and subspecies are listed below, in alphabetical order, including their key features, as well as their pathogenic potential and/or ecological relevance. It is noteworthythat about 50% of the new species belong to only one clade, Splendidus, comprising taxa phylogenetically closely related.

#### *Vibrio aestuarianus* **SUBSP.** *Francensis* **(Garnier et al., 2008)**

The species *Vibrio aestuarianus* was described back by Tison and Seidler (1983) on the basis of a group of isolates obtained from oysters, clams and seawater in Oregon and Washington coasts (USA). Few years later, it was described as sharing close similarity with *Vibrio anguillarum*, and*V. pelagius*(Pillidge et al., 1987). This finding was confirmed 20 years later when Sawabe et al. (2007b) on the basis of MLSA of all known*Vibrio* species included*V. aestuarianus* within the Anguillarum clade of the genus. It has been described as a ubiquitous species in different geographic areas from the Baltic Sea (Eiler et al., 2006) to Hong Kong (Wang et al., 2006).

In the last years, it was associated with the syndrome known as "summer mortality" of the oyster (*Crassostrea gigas*) in the

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 2 — #2

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 3 — #3

French coasts (Gay et al., 2004b; Garnier et al., 2007). The syndrome is thought to be multifactorial involving physiological and environmental factors as well as pathogens. Labreuche et al. (2006a,b) demonstrated the pathogenic potential of *V. aestuarianus* in experimental oyster challenges. The characterization of a group of isolates obtained from diseased oysters in France led to the description of a new subspecies, named *V. aestuarianus* subsp. *francensis*, on the basis of DDH values close to the boundary limit for species definition (70%) and several phenotypic differences with the American isolates.

#### *Vibrio artabrorum* **(Diéguez et al., 2011)**

In a previous study (Beaz-Hidalgo et al., 2008), a collection of isolates obtained from Manila and carpet-shell clams and tentatively allocated to the genus *Vibrio* based on their phenotypic features were analyzed by FAFLP. One of the groups established, cluster 70, consisted of 8 isolates that could not be assigned to any of the known species of the genus *Vibrio*. Analysis of the 16S rRNA gene sequences allocated the isolates within the Splendidus clade forming a tight group. MLSA of five housekeeping genes, *atp*A (α-subunit of bacterial ATP synthase), *pyr*H [uridine monophosphate (UMP) kinase], *re*cA (RecA protein), *rpo*A (RNA polymerase α-chain), and *rpo*D, supported their inclusion in that clade forming a well differentiated group with respect to the rest of species, being its closest neighbors *V. pomeroyi* and *V. kanaloae*. DNA–DNA reassociation values confirmed its status of new species within the genus *Vibrio*. It is considered as an environmental species without pathogenic potential for clams.

#### *Vibrio atlanticus* **(Diéguez et al., 2011)**

From the same study of Beaz-Hidalgo et al. (2008) another group of five strains, designated as cluster 5, was likely to be also a new *Vibrio* species, being further characterized using the same approach employed for the description of *V. artabrorum*. The phenotypical characterization, chemotaxonomy, MLSA, and DDH techniqeus confirmed the hypothesis that the clam isolates constituted a new species, related with *V. tasmaniensis*, *V. kanaloae*, and *V. cyclitrophicus* within the Splendidus clade. As *V. artabrorum*, and since until now no pathogenic activity can be proved for *V. atlanticus*, it seems to be part of the normal environmental and clam microbiota.

#### *Vibrio brasiliensis* **(Thompson et al., 2003a)**

Six isolates obtained from lion's paw scallop (*Nodipecten nodosus*) larvae were identified as a tight group during a wide study on vibrios by FAFLP (Thompson et al., 2001). Further characterization of those isolates on the basis of phenotypic features, 16S rRNA gene sequencing, G + C content and DDH, allowed the description of the new species *V. brasiliensis* within the Orientalis clade. Its pathogenic potential was demonstrated in experimental challenges using rainbow trout (*Oncorhynchus mykiss*) and *Artemia nauplii* as animal models (Austin et al., 2005). The extracellular products (ECP) of the strain tested were also harmful to the animals.

#### *Vibrio breoganii* **(Beaz-Hidalgo et al., 2009a)**

A group of seven non-motile, facultative anaerobic alginolytic marine bacteria isolated from cultured Manila and carpet-shell clams in Galicia (NW Spain) were characterized employing a polyphasic approach, including the sequence analysis of the 16S rRNA gene and three housekeeping (atpA, recA, and rpoA) genes, FAFLP fingerprinting, G + C content, DDH, and phenotypic features. Phylogenetic analysis based on 16s rRNA gene sequences showed that the isolates were closely related to the species *V. comitans*,*V. rarus,* and *V. inusitatus*, with sequence similarities of approximately 99%. However, phylogenies based on the sequences of the housekeeping genes grouped the isolates together and allocated them within the Halioticoli clade, although they can be differentiated from the other species in the clade by their FAFLP profiles. DDH experiments confirmed that they represented a new *Vibrio* species, that was named *V. breoganii*.

Some years later, this species was consistently detected in a meta-analysis of three separated studies aimed to identify the ecological population structure of *Vibrionaceae* in the Plum Island Sound Estuary (Ipswich, MA, USA), mainly associated to large particles and zooplankton (Preheim et al., 2011). Interestingly, population of *V. breoganii* showed essentially identical results with respect to genetic breadth in all studies, regardless the season or the sampling method. This species constitutes a good example of how genotypic clusters established by MLSA can serve as a reasonable initial definition of cohesive unit from an ecological point of view, as well as of the ubiquity of *Vibrio* species in the marine environment.

In 2012, a strain of this species was included in one of the first studies examining the complete methylation pattern of a bacterial genome (Murray et al., 2012). The knowledge of

the methylome could be of great interest due to the recognized importance of methylation for understanding fundamental microbiological processes, microbe adaptability, and disease pathogenicity.

#### *Vibrio celticus* **(Beaz-Hidalgo et al., 2010b)**

A group of four motile facultative anaerobic marine isolates obtained from cultured pullet carpet-shell (*Venerupis pullastra*) and Manila clams during 2004 and 2005 in Galicia (NW Spain) were studied using a polyphasic approach. It was found that they formed a tight phylogenetic group based on sequences of the 16S rRNA gene and four housekeeping (*atp*A, *rec*A, *rpo*A, and *rpo*D) genes, indicating that the four isolates represented a novel species in the Splendidus clade of the genus *Vibrio*, for which the name *V. celticus* was proposed. In addition, the strains showed potential pathogenic activity for adult clams in virulence assays.

Recently, a study on the diversity of *Vibrio* spp. in the Eastern English Channel by means of sequencing of the housekeeping gene *pyr*H (Tall et al., 2013), revealed that *V. celticus* was the predominant species among other 20 *Vibrio* species isolated at ambient environmental temperature.

#### *Vibrio cortegadensis* **(Lasa et al., 2013b)**

It was described as a results of the polyphasic characterization of a group of four marine strains isolated from carpet-shell and Manila clams in Galicia (NW Spain). The study of the phenotypic characteristics, the analysis of chemotaxonomic features, the sequencing of the 16S rRNA and five housekeeping (*atp*A, *pyr*H, *rec*A, *rpo*A, and *rpo*D) genes, as well as DDH, allowed the identification of the isolates within the genus *Vibrio*, being their closest neighbors *V. tapetis*, *V. pomeroyi,* and *V. crassostreae* (97.9%). The phylogenetic analysis of the five concatenated genes indicated the allocation of these strains in between the Splendidus and Anguillarum clades.

#### *Vibrio crassostreae* **(Faury et al., 2004)**

Described in 2004 on the basis of five strains obtained from oyster haemolimph, and originally identified as *V. splendidus*-like isolates. The authors employed a polyphasic approach including besides biochemical tests, fatty-acid methyl ester (FAME) analysis, 16S rRNA, and *gyr*B (DNA gyrase subunit B) genes sequencing, FAFLP fingerprinting, and DDH. Although all the genetic studies supported that the five strains constituted a novel *Vibrio* species within the Splendidus clade, their differentiation of the closest relatives was not possible on the basis of 17 phenotypic characters. However, the presence of fatty acids 16:0 iso and 14:0 iso allowed the differentiation of the new species from other *V. splendidus*-like species. It was described as a species with pathogenic potential for the oyster *C. gigas* (Gay et al., 2004a).

#### *Vibrio crosaei* **(González-Castillo et al., 2014)**

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 4 — #4

The description of this new *Vibrio* species was based on the study and characterization of one isolate obtained from cultured oyster (*C. gigas*) in Sonora (Mexico). The phenotypic characteristics and the 16S rRNA gene sequence of the isolate clearly placed it within the genus *Vibrio*, with *V. orientalis* and *V. rotiferianus* as closest relatives. Curiously, these both species belong to different clades, as proposed by Sawabe et al. (2007b), the Orientalis clade and the Harveyi clade, respectively. MLSA technique clarified the definitive allocation of the isolate within the Orientalis clade, and the DNA relatedness measures by DDH experiments confirmed that it constituted a new *Vibrio* species. The proposed name, *Vibrio crosaei*, was chosen to honor Prof. Dr. Jorge Crosa, microbiologist and specialist in vibrios.

#### *Vibrio fortis* **(Thompson et al., 2003b)**

The species was defined on the basis of 10 isolates obtained between 1994 and 1999 from different hosts, including healthy and diseased lion's paw scallop larvae, diseased *C. gigas* larvae, shrimp (*Litopenaeus vannamei*) larvae, Atlantic salmon (*Salmo salar*), as well as sea water. The geographical origins included different North and South American Countries, Tasmania, and UK. Genotypic analysis such as 16S rRNA gene sequencing and DDH confirmed the deliniation of this new species, differentiating it from the closets neighbors *V. pelagius* or *V. mytili*. Austin et al. (2005) confirmed the pathogenic potential of the species using rainbow trout and *A. nauplii* as animal models.

*Vibrio fortis* was further isolated from spotted rose snapper (*Lutjanus guttatus*) in Mexico (Gomez-Gil et al., 2007) and from crown-of-thorns starfish (*Acanthasther planci*) in Australia and Guam (Rivera-Posada et al., 2011). It was also identified as one of the predominant *Vibrio* species in the Cariaco Basin, Venezuela (García-Amado et al., 2011).

The extracellular polymeric substances from this biofilm forming *Vibrio* species were characterized (Kavita et al., 2013), showing potential for industrial applications.

#### *Vibrio gallaecicus* **(Beaz-Hidalgo et al., 2009b)**

This species within the Splendidus clade was described on the basis of the characterization of three strains isolated from Manila clams in Galicia (NW Spain). Phylogenetic analysis of the 16S rRNA gene and four housekeeping (*atp*A, *pyr*H, *rec*A, and *rpo*A) genes, indicated that these strains were closely related to the Splendidus clade, being its closest relatives *V. splendidus*,*V. gigantis,* and *V. pomeroyi*. The FAFLP fingerprints and DDH values supported the MLSA results. It is considered as an environmental species without proved pathogenic potential.

#### *Vibrio gigantis* **(Le Roux et al., 2005)**

The polyphasic characterization of four isolates obtained from *C. gigas* haemolymph allowed the description of this new *Vibrio* species within the Splendidus clade. Although 16S rRNA gene sequence analysis did not permit a clear differentiation of *V. gigantis* from other phenotypically related species, other techniques including FAFLP, DDH, and sequencing of four housekeeping [*gyr*B, *rct*B (replication origin-binding protein), *rpo*D, and *tox*R] genes demonstrated that the isolates formed a tight genomic group, clearly differentiated from the neighboring species. The authors suggested that, as other *Vibrio* species present in the shellfish haemolymph, *V. gigantis* may play a role in the health of the host.

#### *Vibrio kanaloae* **(Thompson et al., 2003a)**

It was described on the basis of five isolates with different origins, including diseased oyster (*Ostrea edulis*) larvae from France, shrimp (*Penaeus chinensis*) from China and sea water from Hawaii (USA). Therefore, it has been described as an ubiquitous species in the aquatic environment. The five strains were originally detected in a wide FAFLP study (Thompson et al., 2001) as a separate cluster, showing a pattern clearly different from other *Vibrio* species, with which share the main phenotypic traits of the genus. Further DDH experiments confirmed that they were in fact a new species within the Splendidus clade.

Later studies on the virulence of other related strains were performed on the basis of experimental infections of *C. gigas*. After injection of strains, bacteria were localized at the periphery of the muscle and induced extensive lesions of the translucent part of the adductor muscle. Unfortunately, although using a polyphasic approach these strains were confirmed to be *V. splendidus*-related, no clear discrimination between *V. kanaloae* and *V. pomeroyi* was possible with the techniques employed. (Gay et al., 2004b). Austin et al. (2005) confirmed its pathogenic potential for aquatic animals including fish and crustaceans.

#### *Vibrio lentus* **(Macián et al., 2001a)**

The study of 12 marine bacteria by means of cultural and physiological characterization, ribotyping, G+C content, DDH, and phylogenetic analysis on the 16S and 23S rRNA genes allowed the description of this new *Vibrio* species in 2001. All the strains had been isolated from Mediterranean oysters in Spain, and were phenotypically similar to *V. splendidus*. The name *V. lentus* was proposed since the strains showed a slow growth on Marine Agar. Thus, colonies of some of the isolates were not larger than 0.2 mm diameter after 3 days of incubation. Some years later *V. lentus* was isolated from diseased wild octopus (*Octopus vulgaris*) and from turbot (*Scophthalmus maximus*) also in Spain (Farto et al., 2003; Montes et al., 2003, 2006). In the case of octopus, experimental infections by bath challenge demonstrated that *V. lentus* was able to reproduce the skin lesions, colonize the internal organs, and induce mortality in healthy octopuses (Farto et al., 2003).

The presence of a lethal extracellular 39-kDa protease, similar to that of *Vibrio pelagius*, was detected in 15% of the ECP assayed belonging to strains of the*Vibrio splendidus*–*V. lentus*related group by Farto et al. (2006), which suggested their potential risk for the health of reared aquatic organisms.

#### *Vibrio neptunius* **(Thompson et al., 2003c)**

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 5 — #5

Described during a polyphasic study of 21 isolates with diverse origins, like healthy and diseased lion's paw scallop larvae, rotifers, and turbot larvae. The results clearly indicated that this group od strains constituted a new species within the Corallilyticus clade of the genus. *Vibrio neptunius* was further identified as aetiological agent of a mortality episode of oyster (*O. edulis*) larvae occurred in a Galician hatchery (Prado et al., 2005). Pathogenicity was confirmed in experimental tests where it shown to cause high mortalities (ranging from 98.5 to 100%) in 72 to 96 h after inoculation of larval cultures. The work of Prado and co-workers constituted the first description of *V. neptunius* as a molluscan pathogen. Later studies with New Zealand Greenshell mussel (*Perna canaliculus*) larvae confirmed the pathogenic potential for other molluscan species (Kesarcodi-Watson et al., 2009a,b).

*Vibrio neptunius* was also found in environmental studies as a predominant bacteria in the anoxic zone (García-Amado et al., 2011). In addition, during a study searching for novel antimicrobials in marine *Vibrionaceae* (Wietz et al., 2010), *V. neptunius* has been identified as a potential resource of antibacterial compounds with future applicability.

#### *Vibrio ostreicida* **(Prado et al., 2014)**

The species description relies on three strains isolated from a flat oyster (*O. edulis*) hatchery in Spain after episodes of high mortality (Prado et al., 2005). Pathogenicity was confirmed in experimental tests where the strains were able to cause high larval mortalities. The results of the phenotypic and genotipic analysis revealed that this group of strains constituted a new *Vibrio* species, closely related to *V. pectenicida*.

#### *Vibrio pomeroyi* **(Thompson et al., 2003a)**

As in the case of *V. kanaloae*,*V. pomeroyi* was originally detected as a group of four isolates showing a characteristic FFAFLP pattern during a study on the genomic diversity amongst *Vibrio* isolates from different sources (Thompson et al., 2001). Two strains had been isolated from healthy bivalve larvae (*N. nodosus*) in Brasil and two from turbot in Spain. They were confirmed as a new *Vibrio* species within the Splendidus clade by means of DDH, phenotypic characterization, and FAME analysis. The studies of Gay et al. (2004a) and Austin et al. (2005) mentioned before demonstrated either non- or low virulence of *V. pomeroyi* in animal models.

#### *Vibrio ponticus* **(Macián et al., 2004)**

It has been described to accommodate four marine bacteria isolated from mussels, fish, and seawater at the Mediterranean coast of Spain. Phylogenetic analysis locate these strains in the vicinity of the Fluvialis–Furnissii clade, sharing with these species similarities slightly higher tan 97% in their 16S rRNA gene sequences. Since one of the isolates were isolated after direct plating of a kidney sample from a diseased gilthead seabream (*Sparus aurata*), the pathogenic potential of the species cannot be discarded.

#### *Vibrio tapetis* **subsp.** *britannicus* **(Balboa and Romalde, 2013)**

*Vibrio tapetis*, described by Borrego et al. (1996), is the causative agent of an epizootic infection described in adult clams called Brown Ring Disease (BRD) constituting a major limiting factor for the culture of Manila clams. This pathogen was considered for years as a highly homogeneous taxon on the basis of its phenotypical features, but the isolation of new strains from different hosts revealed some variability both at serological and genetic level, allowing the description of three major groups related to the host origin of the isolates (Castro et al., 1997; Romalde et al., 2002; Rodríguez et al., 2006). Balboa and Romalde (2013) performed for the first time a phylogenetic study for this pathogen, where *V. tapetis* strains appeared clearly separated in two main robust clusters, one containing the isolates from the British Isles and other one containing the isolates from all other geographic origins. The two clusters, that showed values of DDH between 65.05 and 79.8%, were easily distinguishable for their capacity to produce acid from mannitol and arabinose and for the use of citrate. On the basis of these results, not only an emended description was provided for *V. tapetis*, but also the new subspecies *V. tapetis* subsp. *britannicus* was proposed.

#### *Vibrio toranzoniae* **(Lasa et al., 2013a)**

It was recently described in a polyphasic study of four strains isolated from cultured carpet-shell and Manila clams in the Northwest of Spain. The techniques utilized included phylogenetic analysis based on sequences of 16S rRNA and MLSA of five housekeeping genes (*atp*A, *rec*A, *pyr*H, *rpo*A, and *rpo*D), DDH, FAME analysis and more than 100 phenotypic traits. All the closest relatives were *Vibrio* species included in the Splendidus clade, such as *V. kanaloae*, *V. artabrorum*, *V. gigantis,* or *V. celticus*, from which it can be easily differentiated by several phenotypic characteristics. Current studies of some Chilean *Vibrio* strains isolated from fish seem to indicate that the geographical and host distribution of this species could be wider than expected.

#### *Vibrio xuii* **(Thompson et al., 2003c)**

Three isolates obtained from bivalve and shrimp systems were identified as a tight group during a wide study on vibrios by FAFLP (Thompson et al., 2001). Further characterization of those isolates on the basis of phenotypic features, 16S rRNA gene sequencing, G + C content, and DDH, allowed the description of the new species *V. xuii* within the Nereis clade. Considered as an environmental species, *V. xuii* demonstrated either non- or low virulence in the animal models (Austin et al., 2005).

#### **MOLLUSC AND OTHER** *Vibrionaceae*

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 6 — #6

Other *Vibrionaceae* described in the last years also associated to molluscan shellfish are representatives of the genera *Aliivibrio* and *Photobacterium*.

The genus *Aliivibrio* was established by Urbanczyk et al. (2007) to accommodate the species *V. fischeri*, *V. logei*, *V. salmonicida*, and *V. wodanis*, on the basis of a study based on 16S rRNA gene sequencing and MLSA which results indicated that the four species represented a lineage within the *Vibrionaceae* distinct from other genera. Therefore, the authors proposed the reclassification of the species as *Aliivibrio fischeri* (the type species), *A. logei*, *A. salmonicida* and *A. wodanis*, respectively. The genus includes symbiotic (*A. fischeri*) and pathogenic (*A. salmonicida*) species for marine organisms (Urbanczyk et al., 2007).

*Photobacterium* was one of the oldest established genus in the family *Vibrionaceae.* The type species is *Photobacterium phosphoreum*, which had been described by Cohn in 1878 as "*Micrococcus phosphoreus*" (Gomez-Gil et al., 2011). At the time of writing, the genus *Photobacterium* contained 23 species (http://www.vibriobiology.net). Although most species have no described pathogenic activity and are common inhabitants of marine environment, some species, i.e., both subspecies of *Photobacterium damselae*, are pathogenic for aquatic animals, mainly for fish.

#### *Aliivibrio finisterrensis* **(Beaz-Hidalgo et al., 2010c)**

This species was described after the phenotypic and genotypic characterization of four strains isolated from cultured Manila clam in the north-western coast of Spain. Phylogenetic analyses based on the 16S rRNA gene sequences indicated that these bacteria were closely related to *A. wodanis*, *A. salmonicida*, *A. fischeri*, and *A. logei* with sequence similarities between 98.1 and 96.0%. Phylogenetic analysis based on MLSA of four housekeeping genes and FAFLP experiments clearly showed that these novel isolates form a tight genomic group different from any currently known *Aliivibrio* species.

#### *Photobacterium swingsii* **(Gomez-Gil et al., 2011)**

The characterization of six Gram-negative coccobacilli, isolated from Pacific oysters (*C. gigas*) from Mexico and haemolymph of spider crabs (*Maja brachydactyla*) from Spain, allowed the description of this species within one of the oldest established genera in the family *Vibrionaceae*. Repetitive palindromic PCR (REP-PCR) analysis revealed a high degree of genomic homogeneity among the isolates. Several phenotypic traits differentiated the isolates from the type strains of species of the genus *Photobacterium*, including its closest relatives *P. aplysiae* and *P. frigidiphilum*.

#### **FUTURE PERSPECTIVES**

Although classification of bacteria into a natural system has been hampered by the lack of a generally applicable species concept, the introduction of MLSA has provided much higher resolution for microbial identification and taxonomy (Gevers et al., 2005). The groups or species defined by means of MLSA are of particular interest for microbial ecology, since some theories predict that they correspond to ecologically cohesive populations (Fraser et al., 2009; Preheim et al., 2011). Some examples have been mentioned in this review, such as *V. breoganii* or *V. celticus* which, some years after their description as species, have been identified in ecological studies in different geographical areas as predominant populations. It is likely expected that in the near future more efforts will be made to identify ecological populations using these or other approaches, including single-cell amplification of multilocus genes or single-cell genomics (Stepanauskas and Sieracki, 2007; Rodrigue et al., 2009). As indicated by Preheim et al. (2011), to establish reproducible associations between bacterial species and environmental categories may be helpful to predict their occurrence and to get a deeper knowledge on the ecological factors driving their evolution.

It has been indicated that a phylogenetic hypothesis based on complete genomes is desired for *Vibrionaceae* (Dikow and Smith, 2013), and the new pyrosequencing and bioinformatic tools available would be very helpful to obtain such goal. Comparative genome analyses have already revealed a variety of genomic events, including mutations, chromosomal rearrangements, loss of genes by decay or deletion, and gene acquisitions through duplication or horizontal transfer (e.g., in the acquisition of bacteriophages, pathogenicity islands, and super-integrons), that are probably important driving forces in the evolution and speciation of vibrios (Hazen et al., 2010; Morrison et al., 2012; Daccord et al., 2013; Dikow and Smith, 2013).

On the other hand, a better knowledge of the in situ or real-time function of vibrios is needed, both in the environment or within the microbiota of aquatic animals (Frias-Lopez et al., 2008; Jones et al., 2008). Metatranscriptomics would be a valuable method, not only to reveal "near instantaneous" responses to environmental changes, but also to determine the real role of vibrios in different habitats or hosts.

#### **CONCLUDING REMARKS**

The present study overviewed the diversity of *Vibrio* species associated with bivalve molluscs. Ongoing studies on the disease and pathogenicity of bivalves primarily relies on the use of phenotypic and molecular methods for an exact species identification. It remains to be investigated to what extent some of the recently discovered species are commensal, opportunistic or pathogenic organisms. Knowledge of the infection mechanisms used by classical and emerging*Vibrio* spp. to develop disease in bivalve molluscs will help to establish adequate preventive measures to control the transmission of these pathogens in hatcheries and in coastal growing areas.

Finally, and as pointed out by Grimes et al. (2009), future information on completed genomes, metagenomics, and metatranscriptomics will increase the understanding on the biology and ecology of vibrios, providing new insights and solutions to problems with disease, nutrient cycling in the ocean, and opportunities in marine biotechnology.

#### **ACKNOWLEDGMENTS**

The studies of the University of Santiago reviewed here were supported in part by grants AGL2003-09307-C02-01, AGL2006- 13208-C02-01, and AGL2010-18438 from the Ministerio de Ciencia y Tecnología (Spain).

#### **REFERENCES**

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 7 — #7


within the *Vibrio halioticoli* clade. *Int. J. Syst. Evol. Microbiol.* 59, 1589–1594. doi: 10.1099/ijs.0.003434-0


"fmicb-04-00413" — 2014/1/2 — 10:54 — page 8 — #8


"fmicb-04-00413" — 2014/1/2 — 10:54 — page 9 — #9


"fmicb-04-00413" — 2014/1/2 — 10:54 — page 10 — #10


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 08 October 2013; paper pending published: 29 October 2013; accepted: 16 December 2013; published online: 02 January 2014.*

*Citation: Romalde JL, Diéguez AL, Lasa A and Balboa S (2014) New Vibrio species associated to molluscan microbiota: a review. Front. Microbiol. 4:413. doi: 10.3389/fmicb.2013.00413*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Romalde, Diéguez, Lasa and Balboa. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00413" — 2014/1/2 — 10:54 — page 11 — #11

## Gimme shelter: how *Vibrio fischeri* successfully navigates an animal's multiple environments

### *Allison N. Norsworthy and Karen L. Visick\**

*Department of Microbiology and Immunology, Loyola University Medical Center, Maywood, IL, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Spencer V. Nyholm, University of Connecticut, USA Nicole Webster, Australian Institute of Marine Science, Australia*

#### *\*Correspondence:*

*Karen L. Visick, Department of Microbiology and Immunology, Loyola University Medical Center, 2160 S 1st Avenue, Building 105 Room 3936, Maywood, IL 60153, USA e-mail: kvisick@lumc.edu*

Bacteria successfully colonize distinct niches because they can sense and appropriately respond to a variety of environmental signals. Of particular interest is how a bacterium negotiates the multiple, complex environments posed during successful infection of an animal host. One tractable model system to study how a bacterium manages a host's multiple environments is the symbiotic relationship between the marine bacterium, *Vibrio fischeri,* and its squid host, *Euprymna scolopes*. *V. fischeri* encounters many different host surroundings ranging from initial contact with the squid to ultimate colonization of a specialized organ known as the light organ. For example, upon recognition of the squid, *V. fischeri* forms a biofilm aggregate outside the light organ that is required for efficient colonization.The bacteria then disperse from this biofilm to enter the organ, where they are exposed to nitric oxide, a molecule that can act as both a signal and an antimicrobial. After successfully managing this potentially hostile environment, *V. fischeri* cells finally establish their niche in the deep crypts of the light organ where the bacteria bioluminesce in a pheromone-dependent fashion, a phenotype that *E. scolopes* utilizes for anti-predation purposes.The mechanism by which*V. fischeri* manages these environments to outcompete all other bacterial species for colonization of *E. scolopes* is an important and intriguing question that will permit valuable insights into how a bacterium successfully associates with a host. This review focuses on specific molecular pathways that allow *V. fischeri* to establish this exquisite bacteria–host interaction.

**Keywords:** *Vibrio fischeri, Euprymna scolopes***, symbiosis, biofilm, chemotaxis, antimicrobials, bioluminescence**

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 1 — #1

#### **INTRODUCTION**

Bacteria are remarkably successful organisms because they can effectively sense and acclimatize to a wide variety of environments. This domain of life can flourish in habitats ranging from deep sea hydrothermal vents, to scum growing on a lakebed, and to the gastrointestinal tracts of humans (Orcutt et al., 2011; Salzman, 2011). To thrive in particular environments, bacteria use molecular signaling cascades that recognize extracellular signals and activate intracellular pathways, often leading to a change in gene expression. These changes in gene regulation allow a cell to manage the array of extracellular signals and adapt accordingly.

Of particular interest are the signaling cascades that permit a microbe to cope with the multiple environments found within a eukaryotic host. These pathways presumably sense changing environmental factors such as osmolarity, fluctuating nutrient sources, other microorganisms, antimicrobials, and components of the immune system. Furthermore, a bacterium must integrate the multiple inputs to identify which location, if any, is an appropriate niche. To ask in-depth questions about signaling pathways involved in host colonization, researchers often study "simplified" model systems, in which only one or a few bacterial species successfully infect a host (McFall-Ngai et al., 2013). One model system used for this purpose is the symbiosis between the luminescent marine bacterium, *Vibrio fischeri*, and its nocturnal squid host, *Euprymna scolopes*. In this symbiosis, *V. fischeri* is the only bacterium capable of colonizing a specialized symbiotic organ, the light organ. This monospecific association permits researchers to ask deeply reductionist questions about bacteria/host interactions, and has provided insights into how a single bacterial species controls its gene expression to cope with different host environments.

There are a number of experimentally tractable steps involved in colonization of *E. scolopes*, many of which are facilitated by known signaling pathways in *V. fischeri*. Newly hatched squid are aposymbiotic and must acquire*V. fischeri* cells from the surrounding seawater (Wei andYoung,1989). Ventilation by the squid brings seawater and any bacterial cells into the mantle cavity where the light organ is located (**Figure 1**). To aid in the recruitment of bacteria, the surface of the light organ has epithelial fields with cilia that circulate the seawater (McFall-Ngai and Ruby, 1991). This motion draws cells toward six pores leading into the light organ. In as little as 1 h,*V. fischeri* and other Gram-negative bacteria make contact with cilia and then form biofilm-like aggregates around the cilia and within mucous shed by the host in response to bacterial peptidoglycan (Nyholm et al., 2000; Altura et al., 2013). During these early processes, *V. fischeri* cells secrete molecules, known as microbe-associated molecular patterns (MAMPs), that induce morphological changes and alterations in gene expression in the squid, thereby resulting in a host environment actively shaped by the symbiont (for reviews, seeNyholm andMcFall-Ngai, 2004; Visick and Ruby, 2006; McFall-Ngai et al., 2012) Ultimately, *V. fischeri* cells dominate over other bacteria within the aggregate

McFall-Ngai (2004).

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 2 — #2

through unknown mechanisms (Nyholm and McFall-Ngai, 2003; Altura et al., 2013). After these initial interactions, *V. fischeri* cells then leave the aggregate, enter into the ducts of the light organ, travel through antechambers (spaces not permissive for colonization), and arrive within the crypts, the sites of colonization. Within the location of these different host tissues, *V. fischeri* cells are subjected to host-derived stresses such as reactive oxygen species (ROS) and reactive nitrogen species (RNS), that they must sense and resist (Tomarev et al.,1993;Weis et al.,1996; Small andMcFall-Ngai, 1999; Davidson et al., 2004). When the bacteria finally reach the crypt spaces, they grow to high cell density and begin to bioluminesce. Bioluminescence is a key component of the symbiosis: in exchange for a nutrient-rich niche, the bacteria provide light that the squid can use to avoid predation (Ruby, 1996; Jones and Nishiguchi, 2004). Every day at dawn, the squid expel ∼95% of the *V. fischeri* cells back into the seawater environment, leaving the remaining *V. fischeri* cells to repopulate the light organ (Lee and Ruby, 1994). It has been suggested that this process allows the squid to prevent bacterial overgrowth, thus relieving the burden of carrying a dense growth of bacterial cells (Ruby and Asato, 1993).

exposure to the bacteria. Initiation of colonization requires that *V. fischeri*

Research in the *V. fischeri/E. scolopes* symbiosis field has identified a number of molecular signaling pathways that facilitate the various steps of colonization. A few of these pathways within *V. fischeri* include controlling biofilm formation during the aggregation step, motility and chemotaxis to propel and direct the bacteria toward the crypts, ROS and RNS management during all steps of colonization, and bioluminescence within the light organ. This review will focus on these well-known signaling cascades, although it should be noted that other important pathways exist within *V. fischeri* to promote the symbiosis (reviewed in Nyholm and McFall-Ngai, 2004; Visick and Ruby, 2006; Dunn, 2012; McFall-Ngai et al., 2012; Stabb and Visick, 2013).

#### **INITIATING THE SYMBIOSIS: BIOFILM FORMATION**

The first step of colonization requires that *V. fischeri* cells come into the vicinity of and sense the presence of the squid. This seemingly simple task, however, can be considered a limiting factor in colonization. For example, within the Hawaiian water where *E. scolopes* reside, *V. fischeri* constitute ∼100 to 1500 cells per ml of seawater representing as little as 0.01% of the total bacterial population (Lee and Ruby, 1994; Nyholm and McFall-Ngai, 2004). Furthermore, the light organ is not openly exposed to the seawater; instead *E. scolopes* vents seawater through its mantle cavity and across the entrance to the light organ. It has been estimated that a miniscule volume of seawater (1.3 μl) and thus only a few*V. fischeri* cells enter the mantle during each half-second ventilation (Nyholm et al., 2000). Additionally, one *V. fischeri* cell constitutes only one-millionth the volume of the mantle cavity (McFall-Ngai and Montgomery, 1990; Nyholm et al., 2000). Theoretically,*V. fischeri* cells would have to locate all six pores in a brief amount of time before they are expelled from this cavity (Nyholm et al., 2000; Nyholm and McFall-Ngai, 2004). So how does this microbe manage the transition from seawater to squid?

*V. fischeri* cells do not immediately enter the light organ during ventilation; they first interact with mucous and cilia on the host's epithelial cells, and then they begin to coalesce into a bacterial aggregate (Nyholm et al., 2000; Altura et al., 2013; **Figure 1**). *V. fischeri* strains that fail to form this aggregate or form an enhanced aggregate either fail to colonize the squid or exhibit an enhancement of colonization, respectively (Nyholm et al., 2000; Millikan and Ruby, 2002; Yip et al., 2006; Morris and Visick, 2013). Because this stage of colonization is a critical step in establishing the symbiosis, much research has focused on the mechanisms by which*V. fischeri* cells form these squid-specific aggregates. Of note is the discovery of the 18 gene *syp* (*sy*mbiosis *p*olysaccharide) locus, which was found to be important for the formation of a biofilm, or a community of cells encased in a protective matrix often composed of polysaccharides and other macromolecules (Yip et al., 2005, 2006). The *syp* locus encodes proteins predicted to regulate, produce, or transport the biofilm polysaccharide, and most of the *syp* genes are critical for both *in vitro* biofilm formation colonization (Yip et al., 2005; Shibata et al., 2012). Perhaps not surprisingly, given its importance for initiating the symbiosis, there are layers of controls in place that regulate the formation of this biofilm (Yip et al., 2006; Morris and Visick, 2013).

Production of the Syp biofilm is controlled by a two-component signaling (TCS) cascade, a ubiquitous class of signaling pathways consisting of two types of proteins: a sensor kinase (SK) that receives input signals from the environment, causing it to autophosphorylate, and a response regulator (RR), a protein that receives the phosphoryl group from the SK (reviewed in Stock et al., 2000). This phospho-transfer often changes the activity of the effector domain on the RR, thus leading to a cellular response. The particular TCS pathway that controls production of the Syp biofilm is more complicated than canonical TCS cascades; it contains at least two SKs and two RRs (**Figure 2**; reviewed in Visick, 2009). Overexpression of the SK predicted to be at the top of the hierarchy, RscS, is sufficient to induce biofilm formation *in vitro* and *in vivo* by affecting the activity of two downstream RRs, SypG and SypE (Yip et al., 2006; Hussa et al., 2008; Morris et al., 2011). Phospho-SypG promotes transcription from the four *syp* promoters (Yip et al., 2005; Ray et al., 2013). Unphosphorylated SypE inhibits biofilms at a level below *syp* transcription; however, when SypE is phosphorylated, it functions as a positive regulator (Morris and Visick, 2013). SypE controls biofilm formation by changing the phosphorylation state of a small STAS domain protein, SypA, but the exact function of SypA is unknown (Morris and Visick, 2010, 2013). The *sypA, sypE,* and *sypG* genes are located within the *syp* locus whereas *rscS* is an orphan SK gene hypothesized to be acquired through horizontal gene transfer (Visick and Skoufos, 2001; Mandel et al., 2009). An additional putative SK gene, *sypF*, is located between *sypE* and *sypG*. This location suggests that SypF is yet another SK involved in regulating biofilms. In support of this, an "active" allele of *sypF, sypF\*,* was sufficient to promote biofilms in a *sypG-*dependent manner (Darnell et al., 2008). Surprisingly, SypF\*-induced biofilms also required *vpsR,* a putative RR that is predicted to be involved in cellulose biosynthesis (Darnell et al., 2008).

Although much is known about the Syp signaling pathway, there are still outstanding questions that have yet to be fully answered. For example, what are the signals that RscS and SypF recognize? Do these SKs function as separate inputs into downstream regulators? What, if any, is the connection between Syp biofilms and cellulose biosynthesis? Furthermore, the sole function of SypE appears to be controlling the activity of SypA (Morris and Visick, 2013), yet what does SypA do? Lastly, although

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 3 — #3

#### **FIGURE 2 | Regulation of biofilm formation in** *V. fischeri***.**

Two-component regulators control the production of the Syp biofilm. RscS and SypF are proposed to function as sensor kinases, resulting in the phosphorylation of the two downstream response regulators, SypE and SypG. SypG functions as a transcription factor to control expression of the *syp* locus at its four promoters, while SypE functions downstream of

*syp* transcription to control the phosphorylation state of the small STAS domain protein, SypA. SypF is also predicted to control the activity of a RR VpsR putatively involved in cellulose biosynthesis. Biofilms can be assessed *in vitro* as a wrinkled colony on an agar plate, or *in vivo* as a bacterial aggregate that forms on the surface of the light organ. Adapted from (Visick, 2009).

the ability to form the Syp biofilm is required for aggregate formation, it is not required for outcompeting colonizationincompetent species of bacteria that can also aggregate outside the light organ (Nyholm et al., 2000; Nyholm and McFall-Ngai, 2003; Altura et al., 2013). Hence, what Syp-independent mechanisms establish early specificity in the symbiosis by promoting the dominance of *V. fischeri* cells within this aggregate? The answers to these questions will permit a detailed and mechanistic understanding of a critical, early stage of host colonization.

#### **TRAVERSING THE TERRAIN OF THE SQUID: MOTILITY AND CHEMOTAXIS**

Once *V. fischeri* cells aggregate outside the squid's light organ, they must leave this matrix-encased biofilm, migrate through the pores against outward water currents produced by beating cilia, and traverse across the antechamber and into the crypt spaces in the organ (McFall-Ngai and Montgomery, 1990; McFall-Ngai and Ruby, 1998; Nyholm and McFall-Ngai, 2004). This migration process requires that *V. fischeri* cells have the capability to move through fluids or across surfaces and to direct this movement toward their final destination. For these processes to occur, *V. fischeri* cells utilize flagella for locomotion and chemotaxis proteins to alter the direction of movement.

#### **MOTILITY**

Flagella are large macromolecular appendages with a membraneembedded motor. This motor rotates the long flagellar filament using energy from ion gradients across the membrane (Berg, 2003). The number and location of the flagella (polar or peritrichous) vary among bacterial species; *V. fischeri* in particular has a tuft of 1–5 sheathed flagella at one pole (Ruby and Asato, 1993; McCarter, 2001). Studies have demonstrated that flagellar-dependent motility is required for early stages of host colonization; non-motile or hypermotile strains fail to efficiently colonize *E. scolopes* (Graf et al., 1994; Millikan and Ruby, 2002, 2004; Wolfe et al., 2004; Brennan et al., 2013b). Interestingly, although cells begin the colonization process flagellated, they lose these appendages within the light organ, suggesting that motility is not important within this environment (Ruby and Asato, 1993). Before dawn, *V. fischeri* begin to express flagellar genes and, once released from the light organ during venting from the squid at dawn,*V. fischeri* cells again have fully formed flagella (Ruby and Asato, 1993; Wier et al., 2010). These data suggest that *V. fischeri* cells change their flagellation status based on a particular environmental cue. In support of this idea, flagellation and thus motility of *V. fischeri* depends on magnesium, a divalent cation common in seawater; thus, the seawater environment might promote flagellar synthesis (O'Shea et al., 2005). The observation that *V. fischeri* cells are not flagellated within the light organ signifies that this region of the squid might constitute a low Mg2<sup>+</sup> environment; however, the abundance and/or role of Mg2<sup>+</sup> *in vivo* has not been assessed.

What is the mechanism by which *V. fischeri* cells control flagellation? In well-studied species of bacteria, such as *Escherichia coli*, *Salmonella enterica* Typhimurium and *Vibrio cholerae,* flagellar biosynthesis is regulated in a hierarchical, temporal fashion such that the most proximal structural proteins are expressed and assembled before the more distal ones (Chilcott and Hughes, 2000; Prouty et al., 2001). In *V. cholerae,* there are four classes of genes

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 4 — #4

#### **FIGURE 3 | Predicted flagellar synthesis pathway in** *V. fischeri***.** The *V. fischeri* model of flagella gene regulation is based on the pathway elucidated in the related microbe, *V. cholerae* (Prouty et al., 2001). Class I consists solely of the regulator, FlrA, which, together with σ54,controls expression of Class II genes. Class II proteins include both those necessary for building the base of the flagellum and also the regulators FlrB, FlrC and

σ28(FliA). FlrB and FlrC control transcription of Class III genes necessary for synthesis of the distal basal body, hook, and filament, while σ<sup>28</sup> regulates transcription of Class IV genes involved in the production of motor proteins and other miscellaneous factors. Regulators in red indicate they are important for *V. fischeri* to colonize the squid (Millikan and Ruby, 2003, 2004; Hussa et al., 2007; Brennan et al., 2013b).

that code for either regulatory or structural proteins (**Figure 3**). The sole Class I gene encodes FlrA, a transcriptional activator that controls expression of Class II genes in a manner that depends upon the alternative sigma factor, σ54. Two Class II regulatory proteins, FlrC and the alternative sigma factor σ<sup>28</sup> (FliA), control expression of Class III and Class IV genes, respectively. Classes II, III, and IV also encode different subunits of the flagellar apparatus. This temporal regulation of gene expression ensures proper, step-wise assembly of the flagellum (**Figure 3**).

Bioinformatic studies suggest that*V. fischeri* cells use regulators similar to*V. cholerae* to control flagellar assembly, and mutagenesis studies thus far have supported this hypothesis (for an extensive list, see Brennan et al., 2013b). Mutations in a few of these genes also cause pleiotropic effects. For example, mutations in the motility regulators *rpoN* (σ54) and *flrC* affected bioluminescence, biofilms, and growth in various media (Millikan and Ruby, 2003; Wolfe et al., 2004; Hussa et al., 2007). Additionally, a deletion of the master regulator of flagellar synthesis, *flrA*, affected the expression of a number of genes and proteins unrelated to motility, including a predicted topoisomerase, an ADP-ribosyltransferase similar to the CTX toxin in *V. cholerae* (halovibrin A), phosphoglycerate kinase, a potassium efflux protein, and genes involved in chromosome partitioning (Millikan and Ruby, 2004; Brennan et al., 2013b). These data suggest that motility regulators in *V. fischeri* might be involved in flagellar-independent pathways and predict that these other pathways might also impact colonization. However, the contribution of these motility-independent pathways in host association has yet to be assessed (Millikan and Ruby, 2003; Wolfe et al., 2004; Hussa et al., 2007).

Motility in *V. fischeri* requires the expression of many putative flagellar structural proteins (Brennan et al., 2013b). A few of the flagellin proteins, which polymerize to form the long, external flagellar filament, have been studied in some depth (Millikan and Ruby, 2004). An insertional mutation in *flaA*, which encodes the major flagellin protein (FlaA), resulted in fewer flagella and caused partial defects in motility and colonization (Millikan and Ruby, 2004). The partial defects could be attributed to the presence of at least 5 other flagellin genes in the *V. fischeri* genome. In support of this hypothesis, an insertional mutation in another flagellin gene, *flaD*, also caused a motility defect; conversely, a mutation in the *flaC* flagellin gene had no observable effect on motility (Millikan and Ruby, 2004; Brennan et al., 2013b). To date, no other flagellin genes have been studied in detail. It is not clear whether these "alternative" flagellin proteins are (i) only minor constituents of the flagella, (ii) only utilized in a subpopulation of cells, (iii) specific for the squid association, or (iv) perform yet unknown functions.

Although *V. fischeri* must be motile to colonize the squid, many intriguing questions about this phenotype remain unanswered. For example, many flagellar proteins can be found within light organ exudates, but it is thought that *V. fischeri* is largely aflagellate in the light organ (Ruby and Asato, 1993; Schleicher and Nyholm, 2011). Thus, what is the functional significance, if any, of the presence of these proteins within the light organ? Could these proteins serve as signaling molecules to other *V. fischeri* cells or to the squid? Furthermore, what environmental signals control the loss and/or regeneration of flagella? What are the levels of

magnesium associated with different squid tissues, and do the levels impact flagellation in symbiosis? If so, what is the mechanism? If not, are there molecules released by *E. scolopes* or specific environmental cues that dictate the flagellation state of *V. fischeri* cells? Further research into the control of flagellation should shed light on the mechanism by which this important phenotype is altered during the multiple stages of symbiosis.

#### **CHEMOTAXIS**

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 5 — #5

To identify and reach the colonization-permissive locations within *E. scolopes*, *V. fischeri* cells appear to use chemotaxis, a mechanism bacteria utilize to sense and move toward attractants and away from repellants (reviewed in Manson et al., 1998; Wadhams and Armitage, 2004; Sourjik and Wingreen, 2012). Chemotaxis, a process well-studied in other bacteria, consists of a series of "runs" (smooth swimming) and "tumbles" (for re-orientation). These events depend upon a complex TCS pathway in which receptors, or methyl-accepting chemotaxis proteins (MCPs), are coupled to a SK, CheA, and two downstream RRs, CheY and CheB (Hess et al., 1988; Bourret and Stock, 2002; **Figure 4**). Phospho-CheY directly interacts with the base of the flagellar motor, causing the flagellum to switch its rotation, leading to tumbling and reorientation of the cell (Wadhams and Armitage, 2004). Binding of attractants has the effect of reducing phospho-CheY levels, thereby decreasing tumbling (and increasing smooth swimming). Similarly, deleting *cheY* generates a strain that cannot tumble; therefore, it exhibits a "smooth" run. For a bacterium to continually respond to signals within a chemogradient, it must desensitize and reset the chemotaxis system. To do this, bacteria often control MCP activity by methylating or demethylating specific residues, which activates or deactivates the MCP, respectively (Borkovich et al., 1992; Manson et al., 1998; **Figure 4**). The constitutive methyltransferase, CheR, and the inducible methylesterase, CheB, reversibly control the methylation state of MCPs (Springer and Koshland, 1977; Kehry et al., 1983). Mutation of *cheR* or *cheB* causes cells to exhibit smooth runs or to tumble, respectively, because they cannot adapt to chemogradients (Springer and Koshland, 1977; Stock and Koshland, 1978; Borkovich et al., 1992).

The *V. fischeri* genome contains many genes predicted to be involved in chemotaxis, including the RR, *cheY,* and the methyltransferase, *cheR*. Strains with mutations in *cheY* or *cheR* exhibit "smooth" swimming, similar to *cheY* and *cheR* mutants of *E. coli* (Hussa et al., 2007; Deloney-Marino and Visick, 2012). Importantly, *cheY* and *cheR* mutants fail to compete with wild-type cells for colonization of the squid (Hussa et al., 2007; Deloney-Marino and Visick, 2012). These results suggest that *V. fischeri* cells respond to chemogradients, and that this promotes efficient host colonization (Hussa et al., 2007; Deloney-Marino and Visick, 2012).

Chemotaxis studies performed with*V. fischeri* cells revealed that these bacteria can chemotax to serine, nucleosides, and a variety of sugars, including N-acetylneuraminic acid (NANA) and two chitin components; the monosaccharide, GlcNAc, and the disaccharide, (GlcNAc)2 (DeLoney-Marino et al., 2003; Mandel et al., 2012). Interestingly, these three sugars are associated with the squid environment; NANA is found in squid mucous, while GlcNAc and (GlcNAc)2 are found within the light organ (Nyholm et al., 2000;

Heath-Heckman and McFall-Ngai, 2011). These data suggest that *V. fischeri* cells could use these sugar molecules to chemotax toward the mucous outside the light organ and then into the crypt spaces within the light organ (Nyholm et al.,2000; DeLoney-Marino et al., 2003; Mandel et al., 2012). In support of this, upon exposure to *V. fischeri*, *E. scolopes* expresses a chitin-degrading enzyme in and around the ducts that is predicted to establish a gradient of chitin degradation products, such as (GlcNAc)2, that the bacteria can use for chemotaxis (Kremer et al., 2013). Furthermore, prior exposure of *V. fischeri* to (GlcNAc)2, such as might occur during symbiotic aggregation, induced a four-fold increase in chemotaxis to this molecule (Kremer et al., 2013). Finally, and perhaps most importantly, Mandel et al. (2012) determined that disruption of the (GlcNAc)2 gradient hindered *V. fischeri* from colonizing due to its inability to enter the ducts; the bacteria formed aggregates around the pore, but they rarely entered the light organ.

To guide their migration through the squid, *V. fischeri* cells presumably use MCPs to sense attractants, such as GlcNAc2. In other organisms, the number of MCPs can range from 4 in *E. coli* to over 45 in *V. cholerae*, and it is believed that the number of MCPs reflect the complexity of environments a bacterium experiences (Boin and Hase, 2007; Lacal et al., 2010). The *V. fischeri* genome contains 43 putative MCPs, suggesting it has the capability to navigate toward or away from a large repertoire of attractants and repellants, respectively (Ruby et al., 2005; Mandel et al., 2012; Brennan et al., 2013a). Research into these MCPs in *V. fischeri*, however, has proven more difficult than anticipated. Although 19 of the putative MCP genes have been mutated, only one MCP mutant exhibited abnormal chemotaxis toward amino acids *in vitro*, and none exhibited a colonization defect (Mandel et al., 2012; Brennan et al., 2013a). These data suggest that some MCPs may sense the same signal, and/or perhaps MCPs important for *in vitro* and *in vivo* motility have yet to be studied. Identifying the functions of MCPs will surely provide insights into how *V. fischeri* cells direct their movement toward colonization-permissive sites. Furthermore, studying these MCPs may also shed light on host mechanisms and molecules used to promote colonization.

#### **ON THE DEFENSE: COMBATING ANTIMICROBIALS**

Every environment has the potential to be hostile toward a bacterium. This is especially true when a bacterium is exposed to the animal environment, where host immune pathways are implemented to eradicate an unsolicited microbe. As a result, a bacterial symbiont must be able to effectively respond to antimicrobials to promote an interaction with its host. The symbiosis between *E. scolopes* and *V. fischeri* is proving to be a useful model for understanding how beneficial microbes manage antimicrobial challenges received from a host. From initial aggregation outside the light organ to persistent colonization within, *V. fischeri* cells continually interact with antimicrobials and have evolved mechanisms to manage these molecules and thus maintain specificity within the symbiosis (Ruby and McFall-Ngai, 1999; McFall-Ngai et al., 2010; Nyholm and Graf, 2012).

#### **NITRIC OXIDE**

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 6 — #6

One host molecule experienced by *V. fischeri* during all steps of colonization is nitric oxide (NO), a gaseous, readily diffusible molecule with a wide range of functions (Davidson et al., 2004; Bowman et al., 2011). NO has been studied in mammals, where it controls cellular signaling and, importantly, participates in secondary reactions that produce antimicrobial RNS (Crane et al., 2010; Bowman et al., 2011; **Figure 5**). Presumably, all domains of life produce NO as phylogenetically-distinct organisms contain at least one gene that encodes NO synthase, or NOS (Crane et al., 2010). *E. scolopes*, too, produces an NOS enzyme, as NOS protein and NO molecules were detected within mucous outside the light organ, within the ciliated epithelial fields on the light organ surface, and in the ducts and antechamber of the light organ (Davidson et al., 2004). Surprisingly, after 18 h of colonization by *V. fischeri*, these tissues exhibited a decrease in NOS and NO levels relative to those of aposymbiotic squid (Davidson et al., 2004). This decrease was due to the symbiont's release of two MAMPs, lipopolysaccharide (LPS) and TCT, a component of the peptidoglycan (Altura et al., 2011). Together, these data suggest that NO plays a key role in the cross-talk between *E. scolopes* and *V. fischeri*.

What is the function of NO within this symbiosis? Within many host environments, NO can participate in reactions that generate antimicrobials; however, *V. fischeri* cells exposed to NO do not exhibit a growth defect, at least not when grown aerobically (Wang et al., 2010a,b). In contrast, squid treated with an NO scavenger allowed*V. fischeri* and even the non-symbiotic relative,*Vibrio parahaemolyticus,* to hyper-aggregate around the light organ (Davidson et al., 2004). Combined, these data suggest that NO might be toxic for *V. fischeri* under particular conditions, but this organism may have developed pathways to sense and resist free NO.

Indeed, *V. fischeri* encodes H-NOX, a protein that, in other bacteria, was predicted to be an NO sensing protein because it binds to NO but, until recently, had few known physiological roles (Boon and Marletta, 2005; Price et al., 2007; Carlson et al., 2010). As expected, *V. fischeri*'s H-NOX did bind NO (Wang et al., 2010a). The novel discovery was that an *hnoX* mutation disrupted*V. fischeri*'s normal transcriptional response to NO exposure, leading to the hypothesis that *hnoX* might sense NO and lead to the detoxification of this molecule during colonization (Wang

et al., 2010a). This was not found to be true, as an *hnoX* mutant substantially outcompeted wild-type cells for initiation of colonization, although the difference was diminished after 48 h. Upon inspection of the expression differences in an *hnoX* mutant, it was revealed that, instead of the expected NO defense genes, a set of 10 iron acquisition genes, including hemin receptors, was up-regulated. This suggests that H-NOX usually inhibits iron uptake. In support of this idea, the *hnoX* mutant grew better in hemin-supplemented minimal media than wild-type cells (Wang et al., 2010a).

Why would NO sensing by HNOX lead to a downregulation in the seemingly unrelated iron uptake pathways? The answer to this question remains murky. One possibility is that a high concentration of accumulated intracellular iron in *V. fischeri* within the light organ may be detrimental (Halliwell and Gutteridge, 1984). In fact, it is known that an increase in iron concentrations within a bacterium can lead to the production of harmful hydroxyl radicals through the Fenton reaction, in which H2O2 is converted to ROS (Halliwell and Gutteridge, 1984; Touati, 2000; **Figure 5**). It was proposed that early NO sensing through H-NOX "primes" *V. fischeri* for the crypt environment, where survival may depend upon the ability of the bacterium to combat the formation of these hydroxyl radicals by controlling the levels of free iron (McCormick et al., 1998; Semsey et al., 2006; Wang et al., 2010a; **Figure 5**). In contrast to this hypothesis, haem uptake genes in *V. fischeri* were upregulated 28 h post inoculation and were required for persistence within the light organ (Septer et al., 2011). Furthermore, a mutation in *glnD*, which led to a growth defect in low iron conditions, caused a defect in squid colonization (Graf and Ruby, 2000). Clearly, iron uptake is a complex process that seems to be partly regulated by NO and H-NOX, although the exact role of these regulators in this pathway remain unclear.

H-NOX responds to NO, but it does not induce expression of enzymes that neutralize NO. How, then, do *V. fischeri* cells defend against antimicrobials produced from NO? The genome of *V. fischeri* encodes additional regulators known to affect the expression of NO detoxification pathways in other bacteria (Rodionov et al., 2005; Spiro, 2007). One of the best characterized regulators is NsrR, a transcriptional repressor that inhibits NO mediators including globins and reductases (Bowman et al., 2011). The most conserved gene within the NsrR regulon is *hmp*, which codes for flavohemoglobin, a NO dioxygenase that eliminates NO by converting it to nitrate (Tucker et al., 2010). Similar to other bacteria, exposure of *V. fischeri* to NO and/or deletion of the repressor, *nsrR*, promoted *hmp* expression (Wang et al., 2010a,b). In addition, an *hmp* mutant exposed to NO exhibited a growth defect and a deficiency in oxygen consumption, consistent with its putative function in NO elimination and NO functioning as an antimicrobial. Furthermore, complementation of *hmp* on a high copy plasmid made the cells hyper-resistant to NO (Wang et al., 2010b). NO detoxification via Hmp was also important for colonization: not only was *hmp* promoter activity induced in response to host-derived NO, but cells deleted for *hmp* exhibited a colonization defect at the aggregation stage. Additionally, treating squid with an NOS inhibitor increased the competitiveness of the *hmp* mutant for colonization (Wang et al., 2010b). Finally, *in vitro* experiments demonstrated that pre-treatment of *V. fischeri* cells

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 7 — #7

with NO reduced the severity of the growth arrest upon a second NO challenge. This result indicates that NO exposure can prime *V. fischeri* for subsequent NO challenges, and suggests that perhaps NO sensed by *V. fischeri* cells at the beginning of colonization may serve as a signal to prepare them for subsequent exposure to NO within the light organ (Wang et al., 2010b).

Although much research has demonstrated the importance of NO in the *V. fischeri/E. scolopes* symbiosis, many intriguing questions remain. For example, what proteins directly sense NO and activate mediators of the NO detoxification response? Are other enzymes besides Hmp involved in detoxifying NO and are they involved in the symbiosis? What exactly is the functional link between NO sensing and iron acquisition in *V. fischeri*? Pursuing these questions is of interest not only to the *E. scolopes*/*V. fischeri* field, but also to the many areas of research that study how host NO production affects colonization by symbiotic and/or pathogenic bacteria (Richardson et al., 2006; Baudouin et al., 2007; Bouchard and Yamasaki, 2008).

#### **REACTIVE OXYGEN SPECIES**

In addition to RNS, the squid produces enzymes predicted to generate ROS (Tomarev et al., 1993; Schleicher and Nyholm, 2011). The light organ and the gills of the squid, tissues known to be exposed to bacteria, produce halide peroxidase (HPO; Tomarev et al., 1993; Weis et al., 1996; Small and McFall-Ngai, 1999; Schleicher and Nyholm, 2011). HPO converts H2O2 into HOCl, a chemical that is toxic to *V fischeri* and other bacteria (**Figure 5**). Upon colonization by *V. fischeri*, the levels of HPO in host tissues decrease, although the mechanisms behind this remain unknown (Small and McFall-Ngai, 1999). One potential mechanism to manage HOCl levels is through production of a catalase enzyme; this enzyme converts H2O2 to water and oxygen, thereby lowering the amount of H2O2 available for conversion to HOCl (Mishra and Imlay, 2012). In *V. fischeri*, mutation of the putative catalase gene, *katA*, caused increased sensitivity to H2O2 (Visick and Ruby,1998). Furthermore, the addition of H2O2 to cells induced *katA* expression, indicating that the bacteria recognize and respond to this antimicrobial. Finally, a *katA* mutant exhibited a defect in competing with the wild-type for colonization, suggesting that reducing H2O2 levels and/or preventing HOCl formation is important for the symbiosis (Visick and Ruby, 1998).

*E. scolopes* and *V. fischeri* cells produce many other enzymes involved in generating or attenuating ROS, respectively; however, the signals that induce their expression, and the mechanisms by which they might function remain unknown (Stabb et al., 2001; Schleicher and Nyholm, 2011). One unusual pathway hypothesized to reduce ROS is the bioluminescence pathway, due to the requirement for O2 by the light-producing enzyme, luciferase. Thus, this pathway could lower levels of oxygen molecules and reduce the potential for their conversion into superoxide (O− 2 ; Ruby and McFall-Ngai, 1999; **Figure 5**). Whether bioluminescence, in fact, reduces the levels of ROS within the symbiosis remains to be determined.

#### **OTHER ANTIMICROBIALS**

*E. scolopes* expresses a variety of enzymes that are potentially antimicrobial such as Cathepsin L, chymotrypsin protease,

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 8 — #8

lysozyme, and five peptidoglycan-recognition proteins (PGRPs; Wier et al., 2010; Schleicher and Nyholm, 2011; Collins et al., 2012; Kremer et al., 2013). One PGPR protein, PGPR2, can bind and degrade components of peptidoglycan; however, it seems to play a role in maintenance of the symbiosis rather than as an antimicrobial (Troll et al., 2010). Whereas normally peptidoglycan can serve as an immune stimulant for the host, the PGRP2 enzyme degraded the peptidoglycan components, thereby preventing an inflammatory response, perhaps protecting the host and/or preventing the host from clearing *V. fischeri* cells in the light organ. Whether PGPR2 or any other putative antimicrobial enzyme expressed by the squid is toxic toward *V. fischeri* or other microorganisms, or whether *V. fischeri* has mechanisms to combat these potential antimicrobials remains unknown.

#### **LIGHT IN A DARK PLACE: BIOLUMINESCENCE Lux AND QUORUM SENSING**

One of the first characterized and perhaps most striking phenotype exhibited by*V. fischeri* is its ability to bioluminesce, a phenomenon requiredfor a productive symbiosis with the squid (Wei andYoung, 1989; Visick et al., 2000). In exchange for nutrients, *V. fischeri* supplies light to *E. scolopes* so that the squid can mask its silhouette cast by moonlight (Jones and Nishiguchi, 2004). This process, known as counterillumination, is hypothesized to protect the squid from predation while it hunts for food at night (Jones and Nishiguchi, 2004). The importance of this phenotype to the symbiosis was established when it was determined that mutants unable to produce light failed to persist in symbiosis (Visick et al., 2000).

The structural proteins necessary for light generation (LuxCD-ABEG) are encoded by the *lux* operon (Gray and Greenberg, 1992; **Figure 6**). Numerous other proteins control light production, including the first gene in the *lux* operon, *luxI*, which encodes an autoinducer synthase, and the divergently transcribed gene, *luxR*, which encodes a transcription factor that regulates the *lux* operon. LuxI synthesizes a pheromone, 3-oxo-C6-HSL, that promotes *lux* transcription by binding to and activating LuxR (**Figure 6**; for reviews, see Stabb et al., 2008; Ng and Bassler, 2009). As a result, these regulators participate in a positive feedback loop, such that the LuxR-3-oxo-C6-HSL complex promotes synthesis not only of the Lux enzymes, but also of more 3-oxo-C6-HSL, thus amplifying light production.

LuxR itself is controlled at the level of transcription via input from a complex phosphorelay pathway (see reviews Stabb et al., 2008; Ng and Bassler, 2009; **Figure 6**). This pathway is comprised of two SKs, AinR and LuxP/Q, and additional downstream regulators, including the histidine phosphotransferase, LuxU, and the σ54-dependent RR, LuxO. At low cell density, the SKs function as kinases to autophosphorylate and serve as phosphodonors to a single phosphotransferase, LuxU, which, when phosphorylated, can donate a phosphoryl group to LuxO. At high cell density, the phospho-transfer pathway is reversed, with the SKs functioning as phosphatases to remove the phosphoryl group from LuxU (and LuxO). When LuxO is phosphorylated (at low cell density), it activates expression of the sRNA, *qrr1,* which inhibits the translation of the *litR* mRNA (Miyashiro et al., 2010). LitR is the direct transcriptional activator of *luxR*; thus, inhibition of *litR* leads to an

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 9 — #9

inhibition of bioluminescence (Fidopiastis et al., 2002; Miyashiro et al., 2010). When LuxO is dephosphorylated (at high cell density), *qrr1* levels decrease, LitR is translated, *luxR* is transcribed, and the *lux* operon is expressed.

The SKs in the phosphorelay, AinR and LuxPQ, are predicted to sense and respond to the presence of specific pheromones, C8-HSL and AI-2, produced by the autoinducer synthases AinS and LuxS, respectively (reviewed in Stabb et al., 2008). It is predicted that these pheromones accumulate at high cell density, causing the SKs to function as phosphatases and thus promote light production. However, the two inputs do not equally control bioluminescence. For example, an *ainS* mutation caused a severe bioluminescence defect both *in vitro* and in the squid, while a *luxS* mutation did not dramatically alter any bioluminescence phenotype (Lupp et al., 2003; Lupp and Ruby, 2004, 2005). Similarly, colonization experiments found that the *ainS* mutant exhibited defects in both initiation and persistence, while the *luxS* mutant did not (Lupp and Ruby, 2004, 2005). This suggests that the AinS/R branch is more important for controlling light production and colonization than the LuxS-LuxP/Q branch. One possible explanation for this differential importance is suggested by the finding that the AinSproduced C8-HSL can interact directly with LuxR, albeit at a lower affinity than occurs with 3-oxo-C6-HSL, produced by LuxI (Schaefer et al., 1996; Egland and Greenberg, 2000); thus, C8-HSL can exerts its impact both directly and indirectly to control *lux* expression and therefore light levels (Lupp et al., 2003). These and other data support a model in which the C8-HSL-LuxR complex initiates transcription of the *lux* operon, while 3-oxo-C6-HSL-LuxR ultimately takes over as the critical player that promotes positive feedback of *lux* transcription.

Although the above model fits with what is known about the Lux pathway in other bacteria, the roles of the upstream players in *V. fischeri* remain poorly understood. For example, a mutation in the C8-HSL synthase gene, *ainS,* severely impacted bioluminescence, yet a deletion in *ainR,* the SK predicted to respond to C8-HSL, appeared to exert only a minor effect on luminescence (Lupp et al., 2003; Lupp and Ruby, 2004; Ray and Visick, 2012). These results may indicate that the role of C8-HSL in activating LuxR may be more important than its role in controlling the AinR-mediated phosphorelay.

Surprisingly, the bioluminescence phenotypes of some regulators do not correlate with the predicted colonization phenotypes. For example, a *litR,* mutant exhibited a bioluminescence defect but was not impaired for colonization; in fact, this mutant colonized *E. scolopes* better than wild-type in competition experiments after 48 h (Fidopiastis et al., 2002; Miyashiro et al., 2010). However, in a different study, the same *litR* mutant exhibited a colonization disadvantage at an earlier, 12 h time point (Lupp and Ruby, 2005). Additionally, although mutants deleted for the negative regulators of bioluminescence, *luxO* or *qrr1,* exhibited increased bioluminescence *in vitro,* as expected, they exhibited a defect in colonization when competed with wild-type cells (Hussa et al., 2007; Miyashiro et al., 2010). These results suggest that bioluminescence regulators may have bioluminescence-independent functions. In support of this hypothesis, it was found that LuxO and LitR have complex regulons: LuxO controls many genes outside of the canonical Lux pathway, and LitR contributes to controlling whether the cells secrete or import acetate for metabolism, known as the acetate switch (Fidopiastis et al., 2002; Lupp and Ruby, 2005; Hussa et al., 2007; Studer et al., 2008).

Lastly, an interesting observation made about light production in *V. fischeri* is that most environmental strains of this bacterium are visibly bioluminescent when grown in culture, but strains isolated from *E. scolopes* are "dim," meaning they are bioluminescent but do not emit visible light outside of the squid (Boettcher and Ruby, 1990; Stabb et al., 2008). Furthermore, an experimental evolution model has revealed a correlation between increased colonization and decreased light production (Schuster et al., 2010). These findings indicate that some aspect of the light organ environment enriches for dim strains of *V. fischeri*. Perhaps this is not surprising, as producing bioluminescence is energetically taxing (Bourgois et al., 2001). The luciferase enzyme, consisting of the LuxA and LuxB heterodimer, can constitute up to 5% of total protein in visibly luminescent cells (Hastings et al., 1965). Furthermore, under particular growth conditions, expression of the *lux* operon causes a growth defect (Bose et al., 2008). Therefore, light production must in some way benefit *V. fischeri* both inside and outside a host. Numerous possibilities have been proposed, including the removal of oxygen from the environment, which could decrease the production of ROS (e.g., see Stabb et al., 2008). Additionally, because the squid can detect the bacterial bioluminescence, it is hypothesized that the squid may play an active part in maintaining a population of *V. fischeri* cells with a particular bioluminescence phenotype (Tong et al., 2009; Heath-Heckman et al., 2013). However, no single explanation has yet been established for how strains with particular luminescence levels are enriched within the squid, or how brightly luminescent bacteria survive in other marine environments.

#### **ADDITIONAL REGULATORS OF BIOLUMINESCENCE**

It has long been known that squid symbionts produce a level of light during symbiosis that is about ∼1000X brighter than that produced in culture (Boettcher and Ruby, 1990). These data suggest that there are squid specific-signals that affect luciferase production, and that *V. fischeri* might sense these signals using regulators that are outside of the canonical Lux pathway. In fact, a variety of environmental conditions affect the ability of *V. fischeri* to produce light, all of which might have the potential to be sensed by non-Lux regulators. These putative signals that affect bioluminescence include changes in oxygen levels, osmolarity, Mg2<sup>+</sup>

levels, and iron levels (Stabb et al., 2004; Bose et al., 2007; Lyell et al., 2010; Lyell and Stabb, 2013; Septer et al., 2013).

One key non-Lux regulator is ArcA, a RR predicted to function as a transcription factor when phosphorylated by its cognate SK, ArcB (Iuchi et al., 1990; Iuchi and Lin, 1992; Georgellis et al., 1997, 2001; Pena-Sandoval et al., 2005). In *V. fischeri,* a deletion of *arcA* caused a dramatic ∼500 fold increase in bioluminescence in culture, resulting in light levels that were similar to the levels found during symbiosis (Bose et al., 2007). ArcA appears to exert a direct effect on *lux* transcription by binding to a site upstream of the *lux* operon (Bose et al., 2007). These results suggest that ArcA functions as a transcriptional inhibitor of luminescence genes under culture conditions, and that this inhibition is relieved once *V. fischeri* is inside the light organ. The effect of the *arcA* mutation on bioluminescence, however, was dependent on the presence of an intact *luxI* gene (Septer and Stabb, 2012). This result indicates ArcA primarily functions to inhibit the positive feedback loop that relies on the LuxI-synthesized molecule, 3-oxo-C6-HSL (Septer and Stabb, 2012).

In other systems, the ArcA/B pathway is predicted to sense and respond to the redox state of the cell (reviewed in Malpica et al., 2006). In *E. coli*, reducing conditions sensed by ArcB lead to the phosphorylation and thus activation of ArcA, while oxidizing conditions generate unphosphorylated ArcA. Thus, it has been proposed that this two-component pathway in *V. fischeri* senses the oxidized state in the light organ, leading to unphosphorylated ArcA and a de-repression of bioluminescence (Bose et al., 2007); however, experimental evidence of this has yet to be found. Furthermore, although an *arcA* mutation amplified bioluminescence levels, ArcA might not be the only non-Lux regulator of bioluminescence. For example, a recent transposon screen identified mutations in additional genes that led to an increase in light production (Lyell et al., 2010; Lyell and Stabb, 2013). Whether these genes are directly involved in controlling light production, and/or whether they sense particular environments to regulate bioluminescence remain an active area of research.

#### **CONCLUSION**

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 10 — #10

For such a seemingly simple symbiosis, the interaction between *V. fischeri* and *E. scolopes* requires numerous, complicated regulatory pathways to promote host specificity and colonization. It should be noted that this review focuses only on regulators within *V. fischeri* known to be important for colonization, yet a considerable repertoire of information exists that details host specific pathways that are integral for the symbiosis to occur (see reviews Nyholm and McFall-Ngai, 2004; McFall-Ngai et al., 2012; Stabb and Visick, 2013). Although much information exists about the *V. fischeri* pathways described in the above sections, including biofilms, chemotaxis, responses to antimicrobials, and light production, many discoveries within these signaling cascades have uncovered yet more questions that can still be addressed. For example, what molecular pathways allow *V. fischeri* to outcompete other non-symbiont bacteria during aggregate formation? How is flagellation and deflagellation controlled during the various steps of colonization? What regulators sense antimicrobials and lead to their detoxification? And lastly, what

environmental signal within the light organ promotes bioluminescence? All of these subjects are active areas of research, and they will surely uncover new mechanisms that will expand the knowledge of how bacteria and a host establish a life-long, beneficial relationship.

#### **ACKNOWLEDGMENTS**

We thank members of the lab for helpful suggestions on this manuscript. Work on the *V. fischeri-E. scolopes* symbiosis in the Visick lab is funded by NIH grant GM59690 awarded to Karen L. Visick.

#### **REFERENCES**


"fmicb-04-00356" — 2013/11/28 — 21:02 — page 11 — #11


"fmicb-04-00356" — 2013/11/28 — 21:02 — page 12 — #12


"fmicb-04-00356" — 2013/11/28 — 21:02 — page 13 — #13


Yip, E. S., Grublesky, B. T., Hussa, E. A., and Visick, K. L. (2005). A novel, conserved cluster of genes promotes symbiotic colonization and sigma54-dependent biofilm formation by *Vibrio fischeri*. *Mol. Microbiol.* 57, 1485–1498. doi: 10.1111/j.1365- 2958.2005.04784.x

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 04 September 2013; paper pending published: 20 September 2013; accepted: 09 November 2013; published online: 29 November 2013.*

*Citation: Norsworthy AN and Visick KL (2013) Gimme shelter: how Vibrio fischeri successfully navigates an animal's multiple environments. Front. Microbiol. 4:356. doi: 10.3389/fmicb.2013.00356*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Norsworthy and Visick. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00356" — 2013/11/28 — 21:02 — page 14 — #14

## Environmental reservoirs and mechanisms of persistence of *Vibrio cholerae*

### *Carla Lutz 1, Martina Erken1,2 , Parisa Noorian1, Shuyang Sun3 and Diane McDougald1,2 \**

*<sup>1</sup> Centre for Marine Bio-Innovation, School of Biotechnology and Biomolecular Science, University of New South Wales, Sydney, NSW, Australia <sup>2</sup> Advanced Environmental Biotechnology Centre, Nanyang Environment and Water Research Institute, School of Biological Sciences,*

*Nanyang Technological University, Singapore, Singapore*

*<sup>3</sup> The Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, Singapore, Singapore*

#### *Edited by:*

*Rita R. Colwell, University of Maryland, USA*

#### *Reviewed by:*

*Barbara J. Campbell, Clemson University, USA Yechezkel Kashi, Technion, Israel Crystal N. Johnson, Louisiana State University, USA*

#### *\*Correspondence:*

*Diane McDougald, Centre for Marine Bio-Innovation, School of Biotechnology and Biomolecular Science, University of New South Wales, Sydney, NSW 2052, Australia e-mail: d.mcdougald@unsw.edu.au*

It is now well accepted that *Vibrio cholerae,* the causative agent of the water-borne disease cholera, is acquired from environmental sources where it persists between outbreaks of the disease. Recent advances in molecular technology have demonstrated that this bacterium can be detected in areas where it has not previously been isolated, indicating a much broader, global distribution of this bacterium outside of endemic regions. The environmental persistence of *V. cholerae* in the aquatic environment can be attributed to multiple intra- and interspecific strategies such as responsive gene regulation and biofilm formation on biotic and abiotic surfaces, as well as interactions with a multitude of other organisms. This review will discuss some of the mechanisms that enable the persistence of this bacterium in the environment. In particular, we will discuss how *V. cholerae* can survive stressors such as starvation, temperature, and salinity fluctuations as well as how the organism persists under constant predation by heterotrophic protists.

**Keywords: starvation adaptation, biofilms, chitin, zooplankton, protozoa, predation, stress, viable but nonculturable**

#### **INTRODUCTION**

While it is likely to have been responsible for human infections and mortality throughout human history, cholera outbreaks have only been formally known to science since 1817 (Pollitzer, 1954). Sir John Snow was credited in 1849 as being the first person to connect contaminated water with cholera outbreaks and to use that information as an infection control strategy (Snow, 1855). In addition to being the genesis of modern epidemiology, his observation may also be the first study on the ecology of *Vibrio cholerae*. However, it took another 120 years for *V. cholerae* to be recognized as an autochthonous aquatic bacterium rather than a human pathogen that is a transient resident of the aquatic environment (Colwell et al., 1977). *V. cholerae* has over 200 serogroups, with O1 and O139 being the causative agents of cholera, due to their carriage of the genes encoding cholera toxin (CT) and the toxin co-regulated pilus (TCP; Chatterjee et al., 2007). Surveys performed in non-endemic areas have shown that the majority of *V. cholerae* strains isolated are non-toxigenic (Faruque et al., 2004; Haley et al., 2012; Islam et al., 2013), which suggests that associations with the human host is only one small aspect of the *V. cholerae* life cycle and is not necessary for environmental persistence.

*Vibrio cholerae* inhabits a vast geographical range from the tropics (e.g., the Bay of Bengal where pandemics still occur, e.g., Albert et al., 1993; Huq et al., 2005; de Magny et al., 2011) to temperate waters world-wide (e.g., USA, South America, Australia, Sweden, and Italy, e.g., Vezzulli et al., 2009; Collin and Rehnstam-Holm, 2011; Schuster et al., 2011; Islam et al., 2013; Tall et al., 2013; **Figure 1**). An increasing understanding of the ecology of *V. cholerae*, along with advances in molecular detection has demonstrated that this bacterium is a cosmopolitan aquatic species that is capable of causing illness in humans (Sack et al., 2004).

The capability to survive in many different environmental niches is largely due to the evolution of a range of adaptive responses that allow*V. cholerae*to survive stressors such as nutrient deprivation, fluctuations in salinity and temperature and to resist predation by heterotrophic protists and bacteriophage. One such strategy is the conversion into a viable but non-culturable (VBNC) state during unfavorable conditions (Colwell, 2000; Thomas et al., 2006). Additionally, *V. cholerae* attaches to abiotic and biotic surfaces (chitinous as well as gelatinous zoo- and phytoplankton) as biofilms (e.g., Huq et al., 1996; Akselman et al., 2010; Shikuma and Hadfield, 2010). Biofilm formation is associated with increased stress resistance, increased access to nutrients and as a means of dispersal when attached to living, mobile hosts (Costerton et al., 1995; Hall-Stoodley et al., 2004). Here, the current understanding of how *V. cholerae* is able to adapt to, and persist in the aquatic environment is summarized.

#### **SURFACE COLONIZATION AND BIOFILM FORMATION ENHANCE** *V. cholerae* **PERSISTENCE**

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 1 — #1

For aquatic bacteria, surface attachment provides a selective advantage through access to nutrients that accumulate at the liquid–surface interface (Dawson et al., 1981). Therefore, surface adhesion may be a survival strategy that allows bacteria to persist in nutrient-limited natural environments (Dawson et al., 1981; **Figure 2**). Additionally, some biotic surfaces may provide nutrients for attached bacteria (e.g., chitin; Nalin et al., 1979). Thus, it is not surprising that *V. cholerae* has been detected on many abiotic

**FIGURE 1 | Global distribution of** *Vibrio cholerae***.** Triangles indicate where *V. cholerae* was detected by molecular and/or culture-based methods. Red indicates O1/O139 detection, light blue non-O1/non-O139 detection, and dark blue did not specify. Referenced studies here are only a small fraction of the studies published for certain areas and should guide as an example. North – and Middle America: (Colwell et al., 1981; Ogg et al., 1989; Blackwell and Oliver, 2008; Lizárraga-Partida et al., 2009; Hill et al., 2011; Dickinson et al., 2013), South America: (Franco et al., 1997;

Lipp et al., 2003; Leal et al., 2008; Martinelli Filho et al., 2010; Sá et al., 2012); Africa: (Taviani et al., 2008); Europe: (Andersson and Ekdahl, 2006; Covazzi Harriague et al., 2008; Kirschner et al., 2008; Vezzulli et al., 2009, Vezzulli et al., 2011; Böer et al., 2013; Cantet et al., 2013; Tall et al., 2013); Middle East: (Bakhshi et al., 2009; Grim et al., 2010; Gurbanov et al., 2011; Rashid et al., 2013); Asia Pacific: (Islam et al., 1994, 2013; Desmarchelier et al., 1995; Miyagi et al., 2003; Alam et al., 2006; Vimala et al., 2010; de Magny et al., 2011; Singh et al., 2012).

which may aid in survival.

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 2 — #2

and biotic surfaces, including ship hulls (Shikuma and Hadfield, 2010), zooplankton (Tamplin et al., 1990; Epstein, 1993; Huq et al., 2005; Turner et al., 2009), macroalgae (Hood and Winter, 1997), and as floating aggregates (Alam et al., 2006).

attached. Permanent attachment is mediated by pili (ChiRP and MSHA) and

*Vibrio cholerae* attachment is mediated by pili, which are surface expressed proteins, comprised of pilin subunits that promote surface attachment and subsequent biofilm formation. The ability of *V. cholerae* to attach to a range of surfaces is reflected in the variation in pilin subunits, and hence variation in pili, expressed by *V. cholerae* (Boyd and Waldor, 2002; Aagesen and Häse, 2012). One ecologically important substratum is chitin, and*V. cholerae,* as are most Vibrionaceae, is chitinolytic and possesses multiple conserved genes to attach to and degrade chitin (Meibom et al., 2004; Hunt et al., 2008). This organic polymer of *N*-acetylglucosamine (GlcNAc/NAG) is the second most abundant organic polymer in nature and is an excellent carbon source for bacteria (Rinaudo, 2006; Martínez et al., 2009). The binding of *V. cholerae* to chitin involves the GlcNAc binding protein, GbpA (Kirn et al., 2005; Stauder et al., 2010), as well as the mannose sensitive hemagglutinin (MSHA), which is a type IV pilus (Chiavelli et al., 2001; **Figure 2**). Furthermore, the TCP, which is a colonization factor of human intestinal epithelia, has a duel role in association with chitin. TCP is required for differentiation of attached biofilms, and undifferentiated biofilms lacking TCP have reduced ecological fitness, as they are less efficient at degrading chitin (Reguera and Kolter, 2005).

After initial surface attachment, *V. cholerae* forms "matrixenclosed, surface-associated communities" or biofilms (Yildiz and Visick, 2009). *V. cholerae* biofilm formation is enhanced through the actions of type IV pili, flagella and production of the biofilm matrix, Vibrio polysaccharide (VPS; Watnick and Kolter, 1999). VPS is involved in cell immobilization, microcolony formation, and biofilm maturation (Watnick and Kolter, 1999; Watnick et al., 2001). High and low VPS producing *V. cholerae* colony types are referred to as "rugose" and "smooth," respectively, with the rugose having a higher protective effect toward a variety of stresses, including chlorine (Rice et al., 1992; Morris et al., 1996; Yildiz and Schoolnik, 1999), low pH (Zhu and Mekalanos, 2003), osmotic and oxidative stress (Wai et al., 1998), anti-bacterial serum (Morris et al., 1996), SDS (Fong et al., 2006), phage (Nesper et al., 2001), and heterotrophic protists (Sun et al., 2013). The importance of VPS for protection in the environment is still unknown as there are few published reports on the occurrence of rugose *V. cholerae* in the environment (Ali et al., 2002; Jubair et al., 2012).

The structural genes for VPS production are encoded on two carbohydrate biosynthesis operons located on the large chromosome, which encodes many essential housekeeping genes (Yildiz and Schoolnik, 1999; Fong et al., 2010). The *vps*I operon contains the genes *vpsA* to *vpsK* and the *vps*II operon contains the genes *vpsL* to *vpsQ*. The six genes located between the two *vps* operons (*rbmA*–*F*) are also involved in biofilm formation (Fong and Yildiz, 2007; Absalon et al., 2011; Berk et al., 2012). The requirement for sugars in the synthesis of VPS is an important determinant for the control of biofilm formation (discussed in Section "*V. cholerae* Responses to Environmental Stresses *–* Bottom-up Control of *V. cholerae*"). In addition to sugars, multiple regulators control the expression of VPS. For example VPS biosynthesis is positively regulated byVpsR (Yildiz et al., 2001) and VpsT (Casper-Lindley and Yildiz, 2004) in a c-di-GMP-dependent manner (Krasteva et al., 2010; Srivastava et al., 2011). C-di-GMP is an intracellular secondary messenger that controls the surface association of bacteria in response to environmental conditions (Yildiz, 2008).

Bacterial cell–cell communication, or quorum sensing (QS), is critical for biofilm maturation and subsequent dispersal (Liu et al., 2007; Muller et al., 2007). At high *V. cholerae* cell densities, the QS response regulator, HapR, positively regulates the transcription of *hapA* encoding the hemagglutinin protease (HAP; Jobling and Holmes, 1997; Zhu et al., 2002), *cytR*, a repressor of biofilm development, flagellum biosynthesis genes (Yildiz et al., 2004), and represses VPS production and *toxR,* the regulator of virulence (Jobling and Holmes, 1997; Zhu et al., 2002; Hammer and Bassler, 2003; Zhu and Mekalanos, 2003; Yildiz et al., 2004). It is proposed that the coordination of QS and c-di-GMP controlled traits allows for survival through rapid adaptation to environmental conditions. For example, the switch from a free-swimming to an attached lifestyle (Yildiz and Visick, 2009; Srivastava and Waters, 2012) enables natural competency and horizontal gene transfer (HGT; Lo Scrudato and Blokesch, 2012) as well as provides enhanced predation resistance (Matz et al., 2005). Mechanisms such as biofilm formation enable the persistence of *V. cholerae* and are not limited to interactions with nutritive biotic factors. Indeed, as described in the following sections, many abiotic factors including temperature, salinity, and pH influence the expression of adaptive traits.

#### **"VIABLE BUT NON-CULTURABLE"** *V. cholerae* **IN PLANKTON**

In contrast to starved cells, VBNC cells fail to grow on culture media normally used to grow them, and are often reduced in size but remain metabolically active (Nilsson et al., 1991; McDougald et al., 1998; Oliver, 2010). Since the discovery that *V. cholerae* can enter the VBNC state (Xu et al., 1982), many bacteria, pathogens as well as non-pathogens, have been shown to enter the VBNC state under unfavorable conditions (McDougald et al., 1998, 1999; Oliver, 2005, 2010). Factors known to induce VBNC formation in *V. cholerae* include extremes in temperature and salinity as well as nutrient deprivation (Colwell et al., 1985; Ravel et al., 1995; Carroll et al., 2001; González-Escalona et al., 2006; Thomas et al., 2006; Mishra et al., 2012). VBNC cells of *V. cholerae* have been detected on the surface of higher organisms, such as crustaceans and algae in the plankton and benthos, attached to chironomid egg masses, as well as suspended in bacterioplankton (e.g., Louis et al., 2003; Binsztein et al., 2004; Alam et al., 2007; Halpern et al., 2007). Interestingly, *V. cholerae* appears predominately asVBNC cells within the bacterioplankton and as culturable cells in biofilm consortia, either as aggregates or attached to biotic and abiotic surfaces (Alam et al., 2006). The importance of the VBNC state in cholera epidemiology was demonstrated by Mishra et al. (2012), where virulence and colonization traits were actively expressed in VBNC *V. cholerae* incubated in freshwater microcosms.

For the VBNC response to impart protection allowing for persistence during unfavorable conditions, the cells must be able to resuscitate and divide when conditions become favorable (McDougald et al., 1998). For example, *Vibrio vulnificus* enters the VBNC state and can be resuscitated when incubated in environmental diffusion chambers in the marine environment (Oliver et al., 1995). Just as there are numerous conditions that induce VBNC formation in different species, there are numerous factors that induce resuscitation, including temperature upshift (Nilsson et al., 1991; Mishra et al., 2012) or an increase in nutrients (Binsztein et al., 2004; Senoh et al., 2010).

VBNC *V. cholerae* cells have also been shown to regain culturability by passage through animal digestive tracts (Colwell et al., 1985; Alam et al., 2007; Asakura et al., 2007). Furthermore, the ingestion by human volunteers of *V. cholerae* cells that had been VBNC for as long as 7 weeks resulted in colonization of the intestine and excretion of culturable cells (Colwell et al., 1996). Thus, VBNC cells represent an important environmental reservoir of *V. cholerae* as an agent of disease. However, VBNC cells remain capable of resuscitation for a limited time, and eventually,

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 3 — #3

these cells lose the ability to resuscitate (Weichart et al., 1997). For example, VBNC cells can be resuscitated after co-incubation with eukaryotic cell lines, but resuscitation does not occur for cells that have been VBNC for a prolonged time (more than 91 days; Senoh et al., 2010).

Recently, QS has been implicated in the regulation of the VBNC state. For example, transition of culturable*V. cholerae*to theVBNC state involves biofilm formation and was shown to be dependent on QS (Kamruzzaman et al., 2010). In accordance with these results, VBNC cells from surface waters in Bangladesh have been resuscitated by natural or chemically synthesized QS autoinducers, as high colony forming unit (CFU) counts were detected after 4–5 h of exposure to two different autoinducers (Bari et al., 2013).

One hypothesis for the non-culturability of viable cells on common agar plates is that accumulation of reactive oxygen species (ROS) in the non-growingVBNC cells is detrimental when growth is initiated after nutrient addition. Thus, increased nutrient could lead to an imbalance in metabolism resulting in the production of ROS and cell death (Bloomfield et al., 1998). In fact, treatment of VBNC*Escherichia coli*with catalase or peroxide-degrading compounds can restore culturability (Mizunoe et al., 1999) and elimination of hydrogen peroxide from starved cultures of *E. coli* can prevent VBNC formation (Arana et al., 1992). Furthermore, loss of culturability of *V. vulnificus* under low temperature incubation has been correlated with loss of catalase activity, making the cells ROS sensitive (Kong et al., 2004).

It was recently hypothesized that VBNC cells resuscitate in a stochastic manner rather than in response to environmental parameters (Epstein, 2009). The authors argue that some cells of a dormant community will randomly revive from dormancy and if conditions are favorable, they will grow. Thus these revived cells can be likened to "scouts" inspecting environmental conditions (Buerger et al., 2012a,b). If conditions are not permissive for growth, the scouts will die, resulting in the loss of only a small fraction of the population. However, if conditions are favorable, then the genetic pool is amplified and maintained. The authors demonstrated that sampled marine and soil bacteria randomly became culturable during long term incubation in the wells of microtiter plates containing single cells. Furthermore, strains that were slow growing initially, with a cultivation time of 3–4 weeks could be sub-cultured within 48–72 h (Buerger et al., 2012b). In this way, the VBNC state represents a low cost population-based strategy that allows bacteria to remain dormant in the environment for extended periods, and to potentially either revive when an appropriate cue is present, e.g., an inducing signal, or to randomly test their environment to subsequently grow when conditions are favorable. Although stochastic VBNC resuscitation was not tested with *V. cholerae*, it has implications for identifying resuscitation cues and for understanding triggers of *V. cholerae* growth and cholera outbreaks.

#### *Vibrio cholerae* **RESPONSES TO ENVIRONMENTAL STRESSES – BOTTOM-UP CONTROL OF** *V. cholerae*

The occurrence of*Vibrio* spp. in the environment is correlated with temperature, salinity, and phyto- as well as zooplankton (Turner et al., 2009, 2013; Johnson et al., 2010; Asplund et al., 2011). High

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 4 — #4

water temperature is a strong predictor for the presence of *Vibrio* spp. (Blackwell and Oliver, 2008; Lama et al., 2011; Johnson et al., 2012), as they are mainly detected in warmer waters (above 15◦C). Many studies have demonstrated that the abundance of *Vibrio* spp*.* follows a seasonal pattern that is dictated to a large extent by temperature (e.g., Louis et al., 2003; Binsztein et al., 2004). Increased temperature can influence the attachment of *V. cholerae* to chitinous zooplankton. At temperatures above 15◦C, attachment to chitin increases significantly due to an increase in the expression of the MSHA pilus and the colonization factor, GbpA (Castro-Rosas and Escartìn, 2005; Turner et al., 2009; Stauder et al., 2010). In contrast, despite the water temperatures in the Chesapeake Bay being above 19◦C, *V. cholerae* was found more often in the water column, as planktonic cells, than attached to plankton (Louis et al., 2003). Thus, in addition to elevated temperature, other factors must influence biofilm formation or dispersal, demonstrating the importance of environmental surveying, collecting, and interpreting metadata to determine those factors that influence cholera epidemics.

Temperature fluctuations due to seasonal changes, as well as freshwater influx can strongly influence the salinity of marine water bodies. Open ocean waters have an average salinity of 35 ppt. However, near coastal and estuarine areas the salinity can drop due to freshwater input from rivers or rain run-off (Jutla et al., 2011), and can increase in areas with higher solar evaporation, especially in the tropics. *Vibrio* spp. grow preferably at salinities <25 ppt (e.g., Jiang, 2001; Thomas et al., 2006; Baker-Austin et al., 2010). In high salinity environments *V. cholerae* increases the production of the protective pigment, melanin (Coyne and al-Harthi, 1992), which provides UV resistance (Valeru et al., 2009). The relationship between *V. cholerae* occurrence and salinity appears to be variable, with some studies reporting a significant correlation (Singleton et al., 1982; Johnson et al., 2010), while others demonstrate a lack of correlation between the occurrence of the organism and salinity (Johnson et al., 2012). For example, Stauder et al. (2010) showed that different salinities had no effect on attachment to surfaces, which is important for environmental persistence (as discussed in Section " Association with Other Organisms").

Seasonal fluctuations are often correlated with changing nutrient concentrations, as rain run-off is generally higher in spring/autumn and in coastal and estuarine areas. This can lead to higher phytoplankton abundance, followed by zooplankton blooms (e.g., Lobitz et al., 2000; Huq et al., 2005), which provide the chitinous surfaces that harbor bacteria such as *V. cholerae*. This may enable overall numbers of the organism to increase in the environment even though bacterivorous predators are also more abundant.

Nutrient sources in the environment are not uniformly distributed but occur as microscale patches, influenced by localized events such as cell lysis and waste excretion (Blackburn et al.,1998). Planktonic bacteria use motility and chemotaxis to take advantage of such nutrient patches (for a review of see, Stocker and Seymour, 2012). *V. cholerae* possesses a single sheathed polar flagella (Hranitzky et al., 1980) powered by sodium motive force (Kojima et al., 1999). The number of duplicated chemotaxis-related genes possessed by *V*. *cholerae* indicates the importance of this response for environmental survival (Heidelberg et al., 2000). *V. cholerae* have multiple chemotaxis genes, however not all are required for chemotaxis under standard laboratory conditions, suggesting that the other genes act as accessory chemotactic genes or have as yet undiscovered functions in the environment (Gosink et al., 2002). *V. cholerae* has been shown to be chemotactic toward all amino acids (Freter and O'Brien, 1981), suggesting that proteins, peptides, or amino acids may be important nutrient sources in the aquatic environment. In addition, *V. cholerae* upregulates chemotaxis genes in response to chitin oligosaccharides, facilitating detection and attachment to chitinous organisms (Meibom et al., 2004).

The ability of *V. cholerae* to persist in the environment is intrinsically linked to biofilm formation and VPS synthesis, both of which allow for the exploitation of nutrients available at the surface. Concentrations of sugars, phosphorus, and nitrogen influence attachment and biofilm formation *V. cholerae* cells. The presence of glucose and mannose induce VPS synthesis during biofilm development (Kierek and Watnick, 2003; Moorthy and Watnick, 2004). The phosphoenolpyruvate phosphotransferase system (PTS) is one of the major sugar transport systems in *V. cholerae* and activation of PTS results in derepression of VPS gene transcription and thus increased biofilm formation (Houot and Watnick, 2008; Houot et al., 2010). In addition, a *V. cholerae* PTS that responds to intracellular nitrogen concentrations, is believed to repressVPS production, however the receptor molecule and signaling pathway are still unknown (Houot et al., 2010).

Phosphorous also affects surface colonization. In phosphorus depleted environments, *V. cholerae* adopts a free-swimming planktonic lifestyle that is mediated by a two-component system, PhoBR. The histidine kinase, PhoR, phosphorylates the response regulator, PhoB, resulting in further repression of VPS production (Pratt et al., 2009; Sultan et al., 2010).

Planktonic *V. cholerae* cells have been shown to settle in response to extracellular DNA (eDNA), which is a component of the pre-established biofilm matrix (Haugo and Watnick, 2002). This occurs by repression of CytR,which in turn repressesVPS and biofilm formation (Haugo and Watnick, 2002). Since *V. cholerae* is rich in DNases (Focareta and Manning, 1991), the eDNA maybe utilized as a nutrient source (Seper et al., 2011).

Since nutrient availability fluctuates in the aquatic environment, the ability to store essential nutrients is an important traitfor bacteria that live a "feast-to-famine lifestyle." In bacteria, glycogen is stored in granules and can serve as a carbon source during periods of starvation (Preiss and Romeo, 1994). Under nutrient rich conditions *V. cholerae* increases glycogen storage precursors (Kan et al., 2004). In addition, glycogen granules are present in nutrient poor rice water stools shed by patients with cholera (Bourassa and Camilli, 2009), indicating that glycogen storage may provide nutrients to*V. cholerae* as it passages from the human host into the aquatic environment. In addition to glycogen storage, the ability to store inorganic phosphorus (Pi) facilitates protection against environmental stresses such as acidity, salinity fluctuations, and the damaging effects of hydrogen peroxide, as it is required for activity of the general stress response regulator, RpoS (Jahid et al., 2006). *V. cholerae* is also able to store Pi within membrane bound granules at 100 times the concentrations achieved by *E. coli*(Ogawa et al., 2000). *V. cholerae* mutants deficient in Pi storage displayed reduced attachment to abiotic surfaces, decreased motility and a delayed adaptation to high calcium media (200 mM) (Ogawa et al., 2000).

In addition to carbon and phosphorous, iron is also a growth limiting factor required for cellular metabolism as it is a component of many cofactors (Wackett et al., 1989) and has low solubility in aquatic environments (Martin, 1992). Iron concentrations in the aquatic environment are highly variable and are generally correlated with water depth (Martin and Michael Gordon, 1988). *V. cholerae* has evolved several iron transport systems and receptors that enable persistence in low iron environments (Heidelberg et al., 2000; Wyckoff et al., 2006, 2007). These iron acquisition systems include a catechol siderophore, vibriobactin (Griffiths et al., 1984), and a transport system, Feo, that can take up ferrous iron (Wyckoff et al., 2006). Most iron acquisition genes, such as *irgA* (Goldberg et al., 1991), are repressed by the ferric uptake regulator (Fur) in environments with sufficient iron, as Fe(II)-Fur binds to the promoter of iron-regulated genes, preventing their expression (Bagg and Neilands, 1987). *V. cholerae* can also make use of siderophores secreted by other microorganisms, such as fluvibactin synthesized by *Vibrio fluvialis*, as it possesses the required receptors and uptake systems (Yamamoto et al., 1993).

In nutrient limited environments, *V. cholerae* can enter a starvation state, in which the cells are non-growing but culturable. In a recent laboratory study, Jubair et al. (2012) described the longterm starvation survival of *V. cholerae* (700 days). The authors suggest the term "persister phenotype" to differentiate starved cells from the VBNC state. The growth of persister cells was enhanced in the presence of phosphate and chitin, both important nutrients, which further highlights their importance for *V. cholerae* survival. An earlier study on the behavior of *V. cholerae* starved for 40 days showed that chitin attachment ligands were maintained (Pruzzo et al., 2003). These findings demonstrate the importance of association with chitinous organisms with details of specific interactions discussed in the Section "Association with Other Organisms."

#### **TOP-DOWN CONTROL BY PREDATORY MICROGRAZERS**

While availability of nutrients exerts "bottom-up" control of *V. cholerae*, predation by heterotrophic protists is one of the major mortality factors faced by bacteria in the environment (Hahn and Höfle, 2001; Matz and Kjelleberg, 2005). As part of the bacterioplankton, *V. cholerae* is under constant predation pressure by phagotrophic protists and other bacterivorous members of the zooplankton community. The long-term persistence and seasonal accumulation of *V. cholerae* is dependent on how it responds to this stress. Microcosm studies of natural bacterioplankton communities from the Gulf of Mexico showed that ciliates and heterotrophic nanoflagellates (HNFs) efficiently eliminate *V. cholerae* from environmental water samples (Martínez Pérez et al., 2004). In addition, ciliates as well as flagellates can feed on *V. cholerae,* with grazing rates of up to 600–2,000 bacteria cell−<sup>1</sup> h−<sup>1</sup> (Macek et al., 1997). Control of *V. cholerae* numbers by heterotrophic protists was also demonstrated by Worden et al. (2006), where *V. cholerae* growth rates of up to 2.5 doublings per day were countered by heavy grazing mortality by HNFs. During intense phytoplankton blooms, these growth rates increased to more than four doublings per day, allowing

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 5 — #5

*V. cholerae* to overcome grazing pressure, potentially attaining sufficient numbers in the environment to reach an infectious dose.

*Vibrio cholerae* cells encased in a biofilm matrix are protected from predation by HNFs, while planktonic cells are rapidly eliminated (Matz et al., 2005). Predation induces biofilm formation and a smooth to rugose morphological shift, due to an increase in VPS production (Matz et al., 2005). VPS has subsequently been confirmed to be partly responsible for biofilm grazing resistance, where the *V. cholerae* cells encased in VPS were protected from predators (Sun et al., 2013). In addition to physical protection provided by biofilms, the high cell density in *V. cholerae* biofilms provides a sufficient quorum to promote expression of several QS-regulated anti-protozoal factors that cannot accumulate in the planktonic phase.

The importance of QS for protection from protozoal predation is supported by field tests demonstrating that QS-deficient *V. cholerae* was more susceptible to grazing than the wild type. However, the QS mutant strain did not lose all grazing resistance, suggesting that *V. cholerae* regulates anti-protozoal activities by a combination of QS and other regulatory systems (Erken et al., 2011). VPS mutants were less resistant than the wild type strain to surface grazing by the amoeba, *Acanthamoeba castellanii* and the HNF, *Rhynchomonas nasuta*, but were more resistant than the *hapR* mutant strain, indicating that QS is more protective thanVPS against predators (Sun et al., 2013). QS has been shown to regulate secreted compounds that provide resistance from functionally different predators such as *Tetrahymena pyriformis, Cafeteria roenbergensis,* and *Caenorhabditis elegans*, e.g., an uncharacterized anti-protozoal factor (Matz et al., 2005) and the PrtV protease (Vaitkevicius et al., 2006).

The type VI secretion system (T6SS) also functions as an anti-predation mechanism that is inhibitory against *Dictyostelium discoideum*, mammalian macrophages, and *E. coli* (Pukatzki et al., 2006; MacIntyre et al., 2010). Three proteins (VgrG-1, -2, and -3) secreted by the T6SS form syringe-like structures, puncturing the cell membrane and delivering a virulence factor, VasX, into *D. discoideum* (Pukatzki et al., 2007; Miyata et al., 2011). The expression of another major component of T6SS, Hcp, is positively regulated by QS in *V. cholerae* (Ishikawa et al., 2009). Although all *V. cholerae* strains have this system, expression differs between them (Unterweger et al., 2012). For example, pandemic El Tor strains do not express T6SS under laboratory conditions while in some non-O1/non-O139 strains T6SS is constitutively expressed (Miyata et al., 2010).

Another mechanism for surviving protozoan predation is the ability of the bacterium to survive digestion. Both clinical and environmental strains of *V. cholerae* can survive intracellularly in a range of amoeba (Abd et al., 2004, 2005; Jain et al., 2006). Several studies have demonstrated that*V. cholerae* growth is enhanced when associated with free-living amoeba (Thom et al., 1992; Sandström et al., 2010; Valeru et al., 2012), further demonstrating the role amoeba play as environmental reservoirs of *V. cholerae*. In addition to surviving within amoebic trophozoites,*V. cholerae* cells have been found in the stress resistant cysts formed by amoeba, providing protection from environmental stresses (Thom et al., 1992; Abd et al., 2004), as well as a vehicle for dispersal throughout

**"fmicb-04-00375" — 2013/12/12 — 20:31 — page 6 — #6**

the aquatic environment (Thom et al., 1992; Brown and Barker, 1999). Thus, amoeba cysts could potentially facilitate the spread of cholera (Winiecka-Krusnell and Linder, 2001). Although many reports have characterized the relationship between*V. cholerae* and amoeba (Thom et al., 1992;Abd et al., 2005, 2007; Sandström et al., 2010), very little is known about the mechanisms that facilitate intracellular survival, although survival of the acidic conditions encountered within the digestive vacuoles has been attributed to an inducible acid tolerance response (Merrell and Camilli, 1999). In addition, ToxR has been shown to be important for survival in amoeba (Valeru et al., 2012). The authors suggest that it may be the ToxR-regulated outer membrane proteins, OmpU and OmpT that are responsible for enhanced survival. Experimentally, attraction and attachment to a protozoan host cell has yet to be shown (Abd et al., 2009, 2011). However, adhesins such as MSHA or capsule/LPS O side chain are not involved (Lock et al., 1987; Abd et al., 2009).

There is a lack of knowledge regarding the type and function of other virulence factors in facilitating intracellular survival in protozoa, especially when compared with other bacteria such as *Legionella* spp. and *Salmonella* spp. (Rowbotham, 1980; Bozue and Johnson, 1996; Brandl et al., 2005). There are many important questions that need to be addressed regarding*V. cholerae*–protozoa interactions, including how prevalent these interactions are in the environment and whether they facilitate resuscitation from the VBNC state. In addition, it is important to explore the prevalence of survival and passage through predatory protists and whether the bacterium remains viable in fecal pellets. A further understanding of the roles higher organisms play in the enhancement of *V. cholerae* fitness traits is required to understand the persistence and spread of the pathogen in the environment.

In addition to aforementioned predation pressure by phagotrophic protists, phage, and predatory bacteria also affect the abundance and serogroup prevalence of *V. cholerae* in the environment. For example, the CTXϕ phage carries the genes encoding CT and is required for conversion of non-toxigenic to toxigenic strains (Miller and Mekalanos, 1988; Pearson et al., 1993; Waldor and Mekalanos, 1996). Phage predation has shaped cholera epidemics in Bangladesh, where high concentrations of phage are detected after an initial peak in cholera cases and numbers of *V. cholerae* cells in the aquatic environment (Faruque et al., 2005). Following the increase in phage numbers, the number of cholera cases decreases. An increase in phage numbers in the environment was also correlated with an increase in *V. cholerae* lytic bacteriophage in patient stool samples, with one of the predominant bacteriophage species belonging to the *Myoviridae* family (Seed et al., 2011). Environmental surveys have detected *Myoviridae* in regions where cholera outbreaks have occurred, such as Peru (Talledo et al., 2003), Kolkata (Sen and Ghosh, 2005), and Kenya (Maina et al., 2013). Control of *V. cholerae* by phage is supported by a continuous culture experiment, which suggests that *V. cholerae* populations may be influenced by phage to a larger extent than by nutrient limitation (Wei et al., 2011). Predatory bacteria such as *Bdellovibrio* sp. also prey on *V. cholerae* (Chen et al., 2012) and might also shape *V. cholerae* occurrence in the environment. However, there is limited information on the interactions between predatory bacteria and *V. cholerae*.

### **ASSOCIATION WITH OTHER ORGANISMS**

*Vibrio cholerae* is an integral part of the aquatic environment and in addition to heterotrophic protists, interacts with a wide range of organisms. It has been demonstrated to interact with water fowl (Halpern et al., 2008), fish (Kiiyukia et al., 1992; Senderovich et al.,2010), chironomids (Broza and Halpern,2001; Halpern et al., 2006), mussels (Deriu et al., 2002; Collin et al., 2012), cyanobacteria (Islam et al., 1999), diatoms (Binsztein et al., 2004; Seeligmann et al., 2008), and dinoflagellates (Binsztein et al., 2004; Akselman et al., 2010; **Figure 3**). The association of *V. cholerae* with zooplankton has been a topic of study since the discovery of cells attached to the surface of copepods in the early 1980s (Huq et al., 1983; Tamplin et al., 1990). Zooplankton are an important part of the aquatic food web, grazing on autotrophic and heterotrophic bacterio-, nano-, and micro-plankton and are in turn preyed upon by larger plankton, such as insect and crustacean larvae and fish. One well-studied interaction is that between *V. cholerae* and chitinous zooplankton, e.g., copepods and cladocerans (Nalin et al., 1979; Huq et al., 1983; Rawlings et al., 2007). For example, pivotal experiments link the transmission of cholera with zooplankton (Huq et al., 1996, 2005; Colwell et al., 2003). In a now classic experiment, the filtration of water through readily available sari cloth reduced *V. cholerae* numbers by 99% (Huq et al., 1996). This method proved to be effective in field trials in reducing the incidence of cholera cases and was continued by villagers as a treatment for drinking water (Colwell et al., 2003; Huq et al., 2010). de Magny et al. (2011) suggested the use of different zooplankter to predict cholera epidemics as they demonstrated that the cladocerans, *Monia* spp. and *Diphanosoma* spp. as well as the rotifer *Brachionus angularis*, were significantly correlated with the presence of *V. cholerae* and with cholera outbreaks. *V. cholerae* has repeatedly been found to be associated with the copepod *Acartia tonsa,* which appears to harbor higher numbers of *V. cholerae* than co-occurring copepods (e.g., Huq et al., 1983; Binsztein et al., 2004; Rawlings et al., 2007; Lizárraga-Partida et al., 2009, for further information, see Pruzzo et al., 2008).

The predominantly attached lifestyle of *V. cholerae* enables it to use many different biotic surfaces as nutrient sources. In addition to degrading chitin, *V. cholerae* has the ability to degrade the egg masses of chironomids (Broza and Halpern, 2001; Halpern et al., 2004). The production of the QS-regulated HAP is necessary for the degradation of the gelatinous matrix of the egg masses (Halpern et al., 2003). Although high numbers of *V. cholerae* were found attached to the egg masses (3.9 <sup>×</sup> 104 per egg mass; Halpern et al., 2007) 99.5% of the attached cells were other species, e.g., *Acinetobacter, Aeromonas, Klebsiella, Shewanella,* and *Pseudomonas*. These species may benefit from the nutrients that are released by *V. cholerae* as it degrades the egg mass. Alternatively, these species may have a negative impact on *V. cholerae* by expressing bacteriocins or competing for nutrients and space, which may in part explain why the majority of *V. cholerae* on the egg masses, 99.7%, were VBNC (Halpern, 2011). *V. cholerae* has been shown to be associated with chironomids in all four stages of development, from egg to adult (Broza and Halpern, 2001; Halpern et al., 2003; Broza et al., 2005), suggesting these insect eggs and larvae can serve as vectors for the transmission of cholera. Indeed, chironomids that were collected in air 3 km away from a water source

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 7 — #7

#### **FIGURE 3 |** *Vibrio cholerae* **interactions with other organisms and the environment.** *V. cholerae* is part of the bacterioplankton in aquatic environments. It is under predation pressure by protozoa and bacteriophage and is thus incorporated into the microbial loop. Low temperature and nutrient conditions can trigger the VBNC state, from which it resuscitates under more favorable conditions. *V. cholerae* can

also attach to autotrophic organisms such as phytoplankton or macroalgae, which can provide a carbon source. Attachment to chitinous zooplankton and gelatinous egg masses (e.g., chironomids) provide nutrients and also facilitate HGT. Fish and birds feed on plankton or mussels that might harbor *V. cholerae* and can potentially spread the bacterium across long distances.

were found to harbor *V. cholerae* and thus, these midges can carry the pathogen from one body of water to another (Broza et al., 2005). Although no toxigenic serogroups of *V. cholerae* have been detected in association with chironomids to date, it remains possible that these could also be associated with chironomids (Halpern, 2011).

The association between *V. cholerae* and phytoplankton has been well studied (e.g., Tamplin et al., 1990; Lobitz et al., 2000; Turner et al., 2009). Autotrophic protists, such as diatoms and dinoflagellates (Binsztein et al., 2004; Eiler et al., 2006), cyanobacteria (Islam et al., 1999; Eiler et al., 2007) as well as macroalgae (Vugia et al., 1997; Haley et al., 2012) support *V. cholerae* growth (e.g., Vezzulli et al., 2010). Various laboratory and environmental studies have shown that *V. cholerae* cells attach to microalgae cells. In a study off the coast of Argentina, Seeligmann et al. (2008) detected 1–10 VBNC *V. cholerae* cells attached to a single algal cell. It was suggested that attachment to phytoplankton might enable *V. cholerae* to survive prolonged exposure in freshwater environments due to the nutrients and salts excreted by the phytoplankton cells (Islam et al., 1989; Tamplin et al., 1990; Binsztein et al., 2004). Nutrients supplied by phytoplankton, e.g., due to a massive bloom, can also support explosive growth of *V. cholerae* (Mouriño-Pérez et al., 2003). In fact, remote sensing of chlorophyll *a* has been proposed as a method to predict cholera outbreaks (Lobitz et al., 2000).

Attachment of *V. cholerae* to macroalgae is induced by the plant-derived polyamine, norspermidine (Hamana and Matsuzaki, 1982). Norspermidine is bound by NspS, a periplasmic spermidine-binding protein that interacts with the periplasmic portion of the membrane protein, MbaA, resulting in induction of biofilm formation (Karatan et al., 2005). Mannitol, which is a product of photosynthesis by brown algae and secreted at high concentrations (Yamguchi et al., 1969; Ymele-Leki et al., 2013), induces *V. cholerae* settlement and VPS-dependent biofilm formation, resulting in transcription of *mtlA*, encoding a mannitol specific transport protein (Ymele-Leki et al., 2013). Once mannitol is transported into the *V. cholerae* cell it is possibly used as a carbon source or an osmoprotectant (Ymele-Leki et al., 2013).

*Vibrio cholerae* has also been detected in the gut of various species of fish that these feed on phyto- as well as zooplankton (Senderovich et al., 2010). *V. cholerae* numbers as high as <sup>5</sup> <sup>×</sup> <sup>10</sup><sup>3</sup> cells per gram of intestine content were detected in fish sampled from different marine and freshwater environments in Israel (Senderovich et al., 2010). Non-O1 *V. cholerae* has also been detected in the kidneys, livers, and spleens of diseased, or homogenates of healthy ayu fish sampled from different rivers in Japan (Kiiyukia et al., 1992). Although no isolate carried CT genes, the supernatant of the cultures produced fluid accumulation in suckling mice. In addition to fish, waterfowl have been connected to the spread of *V. cholerae* (Ogg et al., 1989). Sea birds feed on zooplankton as well as phytoplankton and come in contact with these organisms by swimming on the water. Here, these planktonic organisms can then attach to bird feathers and thus *V. cholerae* can potentially be spread by air (Halpern et al., 2008).

Most research on environmental *V. cholerae* has focused on the occurrence of the bacterium within the planktonic community or on the interactions of *V. cholerae* with planktonic organisms.

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 8 — #8

However, recent research has shown that benthic communities also harbor high numbers of Vibrios, including *V. cholerae* (e.g., Covazzi Harriague et al., 2008; Vezzulli et al., 2009; Collin and Rehnstam-Holm, 2011). For example, bivalves are benthic filter feeders connecting the plankton and the benthos by feeding on the plankton. Bivalves such as mussels and oysters can harbor high numbers of pathogenic *Vibrio* spp. in their tissue and are an important niche for these bacteria (e.g., Olafsen et al., 1993; Maugeri et al., 2001; Kirs et al., 2011). Food poisoning resulting from the ingestion of contaminated raw or undercooked seafood is a major threat to human health. While infection by *V. vulnificus* and *Vibrio parahaemolyticus* from ingestion of seafood are the most common (Wright and Harwood, 2013), mussels can also harbor high numbers of *V. cholerae* and thus are a potential health threat (e.g., Murphree and Tamplin, 1995; Bauer et al., 2006; Haley et al., 2012).

The bivalve immune system consists of hemocytes (phagocytic cells) and the hemolymph (i.e., lysosomal enzymes and antimicrobial peptides; Mitta et al., 2000; Pruzzo et al., 2005). In order to reside in bivalve tissues, bacteria need to survive the antimicrobial activity of the hemolymph and engulfment by the hemocytes. Vibrios are resistant to depuration treatments (Murphree and Tamplin, 1995) and show resistance toward the hemocytes of the blue mussel *Mytilus edulis* (Hernroth et al., 2010). Some Vibrios were able to inhibit filtration in adult *M. edulis*, which was not correlated with the binding of the bacteria to the gills of the mussels (Birkbeck et al., 1987), suggesting another mechanism is involved. Interestingly *V. cholerae* strains of different origin have different retention times in mussels (Collin et al., 2012). An environmental *V. cholerae* strain isolated from the blue mussel was both taken up and eliminated much faster than a clinical non-O1/O139 strain. The clinical strain had a much longer retention time, implying that pathogenic strains have better fitness in the mussel than environmental strains. This has implications for control measures such as depuration, as they will be less effective at removing clinical strains than environmental strains. In addition, Collin et al. (2012) identified a highly virulent El Tor strain that was not ingested at all, indicating that bivalves did not eliminate this pathogenic strain from the water column. These results highlight the importance of interaction of *V. cholerae* with other organisms in its environment and the evolution and selection for virulent strains.

In addition to being incorporated into the benthos by filter feeders,*V. cholerae* can be isolated from sediments in numbers that are much higher than in the planktonic phase (Covazzi Harriague et al., 2008; Vezzulli et al., 2009). Sediments may therefore also act as a reservoir for cholera, especially in colder months, seeding the water column when temperatures rise (Vezzulli et al., 2009). Interestingly, in this study nematodes accounted for the highest abundance of the meiofauna, and bacterivorous nematodes accounted for 50% of the total. This suggests that *Vibrio* spp. are under high grazing pressure and top-down control by these nematodes (Vezzulli et al., 2009). In a laboratory study with *C. elegans*, Vaitkevicius et al. (2006)showed that*V. cholerae* kills the nematode after ingestion by secreting the extracellular protease PrtV. Neither CT nor TCP were required for the killing. Interestingly, PrtV was also required to prevent grazing by the flagellate *C. roenbergensis* and the ciliate *T. pyriformis*. In a *hapR* strain, the ability to kill the nematode was strongly diminished. This is in accordance with the role of the QS response regulator, *hapR*, which is important for grazing resistance in the laboratory (Matz et al., 2005) as well as in the environment (Erken et al., 2011). Thus, *V. cholerae* has evolved or acquired a number of genetic systems that facilitate its ability to resist top-down control exerted by predatory eukaryotes.

#### **CONCLUSIONS**

*Vibrio cholerae* is a significant pathogen that has played an important role in human history. Its role in the spread of disease and in epidemics has been reported for more than 150 years and the organism has even played an important role in establishment of modern epidemiology. While the mechanisms leading to infection and epidemics have been well studied, the ecology and mechanisms that underpin environmental persistence have been less well documented. Interestingly, environmental *V. cholerae* strains are largely represented by non-toxigenic strains and indeed, environmental strains display a high degree of genetic variability which has been suggested to aid in *V. cholerae* environmental stress resistance and subsequent persistence. The bacterium has an array of genetic systems involved in stress resistance, when faced with nutrient starvation, iron limitation, or changes in salinity and temperature. One such adaptation is the ability to grow as a biofilm on a range of abiotic and biotic surfaces. This not only increases resistance to stress, but may also directly provide access to nutrients, such as when attached to chitinous surfaces. Biofilm formation has also been directly linked to avoidance of predation by microeukaryotes. Predation resistance can be provided either by physical protection offered by the biofilm, the production of antipredator compounds or defensive molecules or both. Perhaps not surprisingly, some of the gene systems involved in anti-predator defenses are the same as those associated with virulence during human infection. This may support the co-incidental virulence hypothesis that suggests that virulence factors evolve, at least in part, from the competition between predator and prey rather than against a human host. *V. cholerae* is a common inhabitant of many marine and freshwater habitats and this is most likely because it has evolved a range of strategies to enable its persistence in the natural environment. The identification and elucidation of these mechanisms, from ecological, evolutionary and molecular perspectives are likely to deliver exciting discoveries for the next 150 years.

#### **ACKNOWLEDGMENTS**

We thank Dr. Scott Rice and Dr. Sharon Longford as well as the anonymous reviewers for useful comments and suggestions providedfor earlier versions of this article.Work in the authors laboratory was supported by the Centre for Marine Bio-Innovation, University of New South Wales, the Advanced Environmental Biotechnology Centre, the Nanyang Environment and Water Research Institute, and the Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, and grants from the Australian Research Council.

#### **REFERENCES**

Aagesen, A., and Häse, C. (2012). Sequence analyses of type IV pili from *Vibrio cholerae*,*Vibrio parahaemolyticus*, and*Vibrio vulnificus*. *Microb. Ecol.* 64, 509–524. doi: 10.1007/s00248-012-0021-2


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 9 — #9

*Vibrio cholerae* in environmental water samples. *Proc. Natl. Acad. Sci. U.S.A.* 110, 9926–9931. doi: 10.1073/pnas.1307697110


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 10 — #10


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 11 — #11


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 12 — #12


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 13 — #13


*cholerae*. *FEMS Microbiol. Lett.* 302, 22–31. doi: 10.1111/j.1574-6968.2009. 01837.x


"fmicb-04-00375" — 2013/12/12 — 20:31 — page 14 — #14


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 01 October 2013; paper pending published: 29 October 2013; accepted: 21 November 2013; published online: 16 December 2013.*

*Citation: Lutz C, Erken M, Noorian P, Sun S and McDougald D (2013) Environmental reservoirs and mechanisms of persistence of Vibrio cholerae. Front. Microbiol. 4:375. doi: 10.3389/fmicb.2013.00375*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Lutz, Erken, Noorian, Sun and McDougald. This is an openaccess article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00375" — 2013/12/12 — 20:31 — page 15 — #15

### Molecular diversity and predictability of *Vibrio parahaemolyticus* along the Georgian coastal zone of the Black Sea

#### *Bradd J. Haley1, Tamar Kokashvili 2, Ana Tskshvediani 2, Nino Janelidze2, Nino Mitaishvili 2, Christopher J. Grim1,3†, Guillaume Constantin de Magny4, Arlene J. Chen1, Elisa Taviani 1†, Tamar Eliashvili 2, Marina Tediashvili 2, Chris A. Whitehouse5, Rita R. Colwell 1,3,6,7 and Anwar Huq1,8\**

*<sup>1</sup> Maryland Pathogen Research Institute, University of Maryland, College Park, MD, USA*

*<sup>2</sup> George Eliava Institute of Bacteriophages, Microbiology and Virology, Tbilisi, Georgia*

*<sup>3</sup> University of Maryland Institute for Advanced Computer Sciences, University of Maryland, College Park, MD, USA*

*<sup>4</sup> MIVEGEC UMR 5290 IRD-CNRS-UM1&2, IRD de Montpellier, Montpellier, France*

*<sup>5</sup> U.S. Army Medical Research Institute of Infectious Diseases, Fort Detrick, MD, USA*

*<sup>6</sup> Bloomberg School of Public Health, Johns Hopkins University, Baltimore, MD, USA*

*<sup>7</sup> CosmosID™, College Park, MD, USA*

*<sup>8</sup> School of Public Health, Maryland Institute for Applied Environmental Health, University of Maryland, College Park, MD, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Christopher Staley, University of Minnesota, USA Vaughn Cooper, University of New Hampshire, USA*

#### *\*Correspondence:*

*Anwar Huq, Maryland Pathogen Research Institute, University of Maryland, 3102 Biosciences Research Building, MD 20742, USA e-mail: huq@umd.edu*

#### *†Present address:*

*Christopher J. Grim, US Food and Drug Administration, Laurel, MD, USA*

*Elisa Taviani, Biotechnology Center University Eduardo Mondlane, Mozambique*

*Vibrio parahaemolyticus* is a leading cause of seafood-related gastroenteritis and is also an autochthonous member of marine and estuarine environments worldwide. One-hundred seventy strains of *V*. *parahaemolyticus* were isolated from water and plankton samples collected along the Georgian coast of the Black Sea during 28 months of sample collection. All isolated strains were tested for presence of *tlh*, *trh*, and *tdh*. A subset of strains were serotyped and tested for additional factors and markers of pandemicity. Twenty-six serotypes, five of which are clinically relevant, were identified. Although all 170 isolates were negative for *tdh*, *trh*, and the Kanagawa Phenomenon, 7 possessed the GS-PCR sequence and 27 the 850 bp sequence of *V. parahaemolyticus* pandemic strains. The *V. parahaemolyticus* population in the Black Sea was estimated to be genomically heterogeneous by rep-PCR and the serodiversity observed did not correlate with rep-PCR genomic diversity. Statistical modeling was used to predict presence of *V. parahaemolyticus* as a function of water temperature, with strongest concordance observed for Green Cape site samples (Percent of total variance = 70, *P* < 0.001). Results demonstrate a diverse population of *V*. *parahaemolyticus* in the Black Sea, some of which carry pandemic markers, with increased water temperature correlated to an increase in abundance of *V. parahaemolyticus*.

#### **Keywords:** *Vibrio parahaemolyticus***, predictive modeling,** *Vibrionaceae***, Black Sea, aquatic microbiology**

#### **INTRODUCTION**

*Vibrio parahaemolyticus*, a halophilic bacterium, is a causative agent of seafood-related gastroenteritis, wound infections, and septicemia and is known to occur in marine, estuarine, and brackish water environments globally with sporadic occurrence in fresh water (Sarkar et al., 1985; DePaola et al., 2000; Wong et al., 2000; Alam et al., 2009). In addition to notoriety as a causative agent of human infection, the organism is autochthonous to marine and brackish water ecosystems and, similar to other *Vibrio* spp., degrades chitin (Kaneko and Colwell, 1974; Kadokura et al., 2007). One of its main virulence factors, the type three secretion system-2 (TTSS2), plays an important role in preventing predation of its host by higher organisms, suggesting the virulence factors have evolved via environmental selection (Matz et al., 2011). Little work has been done on non-anthropocentric roles of this organism, but its ubiquity and association with animals demonstrate that its ecology extends beyond the human body.

The majority of clinical strains encode the thermostable direct hemolysin (TDH), within the *V. parahaemolyticus* pathogenicity island (Vp-PAI), one of the virulence factors responsible for enterotoxicity (Honda, 1993; Guang-Qing et al., 1995). However, some clinical isolates do not encode TDH, but other hemolysins instead, such as the TDH-related hemolysin (TRH), while all encode the thermolabile hemolysin (TLH). It has also been reported that two type three secretion systems (TTSS1 and TTSS2) are involved in *V. parahaemolyticus* pathogenicity (Bhattacharjee et al., 2006; Ono et al., 2006; Kodama et al., 2007; Matlawska-Wasowska et al., 2010). The TTSS1 found in all *V. parahaemolyticus* strains examined to date has been shown to translocate an effector protein (VP1686) into the cytosol of macrophages and induce DNA fragmentation and another effector protein (VP1680) has been shown to play a role in cytotoxicity in eukaryotic cells (Bhattacharjee et al., 2006; Ono et al., 2006). Interestingly, *V. parahaemolyticus* strains lacking TDH, TRH, and TTSS2 have frequently been isolated from patients not colonized by TDH-, TRH-, and TTSS2-positive strains, suggesting TTSS1 is also responsible for illness in humans (Suthienkul et al., 1995; Okuda et al., 1997; Vuddhakul et al., 2000; Laohaprertthisan et al., 2003; Cabanillas-Beltrán et al., 2006; Bhoopong et al., 2007; Meador et al., 2007; Serichantalergs et al., 2007; Chao et al., 2009, 2010; García et al., 2009; Harth et al., 2009).

*V*. *parahaemolyticus* has been frequently isolated from water samples collected from the Black Sea and sporadic cases of gastroenteritis caused by this bacterium and related vibrios have historically been reported in the Sea of Azov region (Libinzon et al., 1974, 1980, 1981; Shikulov et al., 1980; Clark et al., 1998; WHO, 2011). Further, human pathogenic vibrios are known to be endemic to the greater Caucasus (Narkevich et al., 1993; Gurbanov et al., 2011; Rashid et al., 2013) but the ecologies of these organisms are not well-elucidated in this region. The increasing global incidence of *V. parahaemolyticus* infections suggests it is important to fully understand the ecology of these regions in multiple locations so that public health assessments can be made more accurately (Baker-Austin et al., 2010). Members of the *Vibrionaceae* are known to have an intimate association with planktonic organisms and many studies have demonstrated the role of environmental conditions (namely water temperature and salinity) on the density of these organisms in water bodies. Generally, an increase in temperature of a water body is associated with an increase in *Vibrio* density (Turner et al., 2009; Oberbeckmann et al., 2012). To further understand the ecology of *V. parahaemolyticus* along the Georgian coast of the Black Sea we evaluated the presence of these organisms in water and plankton fractions over a 28 month period (June 2006 to October 2008) and modeled their presence in relation to environmental conditions (salinity, water temperature, pH, and dissolved oxygen). We further evaluated the molecular diversity and presence of virulence factors in a subset of *V. parahaemolyticus* isolates collected during this study.

#### **MATERIALS AND METHODS**

Water samples were collected monthly, except July to September when water was collected biweekly, from five stations on the coast of the Black Sea (**Figure 1**). One hundred liters of water were filtered through 200- and 64-μm plankton nets, to separate size fractions of plankton. Water temperature, salinity, pH, and dissolved oxygen were recorded at the time of sampling. The water fraction (100 ml) was filtered using a 0.45μm nitrocellulose membrane, which was incubated in alkaline peptone water (APW) at 37◦C for 24 h. An aliquot (1- to 5-ml) of each plankton fraction (64- and 200 μm) was also inoculated in APW and incubated at 37◦C for 24 h. A 10 microliter loop of the enrichment cultures were streaked onto thiosulfate citrate bile salts (TCBS) agar plates, which were incubated overnight at 37◦C. All colonies that appeared yellow to green at 24 h were considered presumptive *Vibrio* spp., picked with a sterile toothpick, and streaked to isolate colonies on Luria–Bertani (LB) agar. Presumptive *V*. *parahaemolyticus* colonies were confirmed by streaking onto CHROMagar™ *Vibrio* (mauve colonies) the latter were confirmed by PCR (presence of *tlh*, and *V. parahaemolyticus*-specific collagenase).

For molecular analyses, the following PCR primers were used; collagenase (Di Pinto et al., 2005), *tdh*, *trh*, and *tlh* (Bej et al., 1999), GS-PCR (Matsumoto et al., 2000), ORF8 (Nasu et al., 2000), Mtase (Wang et al., 2006), histone-like DNA-binding protein (HU-α ORF) (Williams et al., 2004), the 850 bp pandemic strain sequence (VPF2/VPR2) (Khan et al., 2002), VP1346 (*yop*) and VP1339 (*escC*) of TTSS2 (Chao et al., 2010), VP1680 (Whitaker et al., 2012) and VP1686 of TTSS1 (This study). Primer sequences for VP1686 were VP1686-F: TGCTTTTGTGATCGCTTTTG and VP1686-R: TGAAGGCAA ACTCAGCATTG (*Ta* = 56◦C; amplicon size = 169 bp) and were designed *in silico* using *V. parahaemolyticus* RIMD2210633 (NC\_004603.1/NC\_004605.1). DNA (25.0 ng) was mixed with 2.5 mM of dNTP, 15 mM of PCR buffer, and 5 U μL−<sup>1</sup> of Taq DNA polymerase, using 20μm of appropriate primer for each analysis. Amplicons were visualized on 1.5% agarose gel stained with ethidium bromide and examined under a UV transilluminator.

To approximate the molecular diversity of the *V*. *parahaemolyticus* isolates, rep-PCR was executed on a randomly selected subset of strains following the methods of Chokesajjawatee et al. (2008). PCR products were separated on a 1% agarose gel in TAE buffer. The resulting fingerprint patterns were documented using the GelDoc-It™ Imaging System (Ultra-Violet Products, Upland, CA). Banding patterns were identified by visual observation and dendrograms were calculated by the unweighted pair-group method using average linkages (UPGMA). Serotyping was performed as follows. Strains were streaked on LB agar with 3% NaCl and incubated overnight at 37◦C. One 10μl loopful of growth was homogenized in 1 mL of saline solution (0.9% NaCl). This solution was divided into two 500μl tubes, one of which was boiled for 2 h. Ten microliters of the boiled cell solution was then mixed with 10μl of each O-antisera and 10μl of the cell suspension that had not been boiled was mixed with 10μl of K-antisera on a glass slide and agglutination visually determined (Denka Seiken Co., Niigata-ken, Japan). Distilled water was used as a negative control for serotyping assays. *V. parahaemolyticus* strain RIMD2210633 (KP positive; serotype O3:K6) for assays.

Predictive models of *V. parahaemolyticus* detection were determined by examining the relationship between presence/absence (response variable) and recorded environmental parameters (explanatory variables) at the time of sample collection. Environmental parameters were also evaluated as explanatory variables by determining the distance from optimality for each data point. This was performed by subtracting the median values of all parameters for those samples in which *V. parahaemolyticus* had been detected (optimal parameters) from all data points following the methods of Jacobs et al. (2010) and Banakar et al. (2011). The absolute values of differences were used as explanatory variables in binary logistic regression analysis. For all measures of association, *p*-values ≤ 0.05 were considered significant. Statistical analyses were conducted on R (http://www.r-project. org/) and SAS softwares (Cary, NC, USA).

#### **RESULTS**

#### **DETECTION OF** *V. parahaemolyticus*

In total, 170 isolates of *V*. *parahaemolyticus* were recovered from Black Sea water and plankton samples collected along the Georgian coast, of which 101 were from water, 30 from the 64μm fraction, and 39 from the 200μm fraction of plankton (**Figure 2**).

*Vibrio parahaemolyticus* was isolated from 40 of a total of 106 water samples collected and 19 of 106 and 26 of 106 of 64- and 200-μm plankton fractions, respectively. Based on Cochran's *Q*test, water samples yielded *V. parahaemolyticus* significantly more frequently than either of the plankton fractions. The difference in *V. parahaemolyticus* isolation frequency was not significantly different between the two plankton fractions. When these distributions were binned to water temperature quartiles (11, 19.8, and 25.8◦C), water samples with temperature between 11 and 19.8◦C were significantly more likely to yield *V. parahaemolyticus* isolates than plankton.

Median water temperatures and salinities for all fractions positive for *V. parahaemolyticus* were higher than those that were negative for *V. parahaemolyticus*, while the opposite was observed for dissolved oxygen (**Table 1**). Median pH levels were slightly lower for all fractions positive for *V. parahaemolyticus* than those that were negative, excluding the P64 fraction (**Table 1**).

#### **SERODIVERSITY**

Twenty-seven serotypes of *V. parahaemolyticus* were detected the majority of which were O2:K28 (7 isolates), O3:K31 (7), O3:KUT (7), O4:KUT (7), and untypable (24) (**Table 2**). *Vibrio parahaemolyticus* O3 O-antigenic type was the most common, comprising 35% of the isolates. Untypable strains may represent strains with novel serology for which *V. parahaemolyticus* antisera has not yet been developed, or strains in which antigenic expression is altered or repressed.

#### **VIRULENCE FACTORS, MARKERS OF PANDEMIC CLONES, AND rep-PCR**

None of the *V. parahaemolyticus* isolates carried the genes for thermostable direct hemolysin (*tdh*) thermostable-related hemolysin (*trh*), TTSS-2, or MTase; all were both, Kanagawa phenomenon and urease negative (**Table 2**). Nineteen isolates resulted in PCR amplicons for the pandemic GS-PCR marker (*toxRS* sequence of pandemic strains), but only seven were 651 bp and 12 were ca. 750 bp. Twenty seven isolates carried the 850-bp pandemic sequence (VPF2/VPR2). Three of the 651 bp, GS-PCR-positive strains were positive for the 850 bp pandemic sequence, whereas six of the 750 bp, GS-PCR-positive isolates encoded this region. Each of the 651 bp, GS-PCR-positive isolates were different serotypes and were typed as O1:KUT, O3:KUT, O3:K31, O3:K33 O3:K65, OUT:K33, and UT, the most notable was the O1:KUT, related to pandemicity. This isolate was also positive for the 850 bp pandemic sequence but lacked all other markers of virulence except TTSS1. Rep-PCR was performed on 45 of the strains (**Figure 3**). A dendrogram of banding patterns revealed a high level of diversity suggesting a non-clonal population of *V. parahaemolyticus* in this environment.

#### **PREDICTIVE MODELING**

Among four explanatory variables in a logistic regression used to model presence/absence of *V. parahaemolyticus* as the response variable, water temperature was the only significant predictor



*A, statistic when V. parahaemolyticus was detected.*

*B, statistic when V. parahaemolyticus was not detected.*


**Table 2 | Molecular characteristics of serotyped strains.**

*aNumber of positive isolates.*

*bPercent of total isolates of that serotype.*

(**Table 3**). When data from all sites were combined, water temperature explained 37.3% of variance in isolation of *V. parahaemolyticus*, suggesting the dynamics of the population are driven by multiple factors. In the Chorokhi and Supsa estuaries, the proportion of variance in *V. parahaemolyticus* isolation explained by water temperature was 22 and 32.1%, respectively, but higher for Batumi Bulvard and Green Cape sites, 43.2 and 70.1%, respectively (**Table 3**).

#### **DISCUSSION**

Although commonly isolated from brackish waters, presence of *V. parahaemolyticus* suggests a public health concern to those utilizing these water sources or consuming products harvested from these waters. This risk is appreciable regardless of pathogenicity island presence in the genomes of circulating *V. parahaemolyticus*, since some infections are caused by isolates lacking *tdh*, *trh*, and TTSS2 (Suthienkul et al., 1995; Okuda et al., 1997; Vuddhakul et al., 2000; Laohaprertthisan et al., 2003; Cabanillas-Beltrán et al., 2006; Bhoopong et al., 2007; Meador et al., 2007; Serichantalergs et al., 2007; Chao et al., 2009, 2010; García et al., 2009; Harth et al., 2009). Isolates recovered in this study lacked the major virulence factors associated with the majority of clinical cases. However, these results are not surprising since typically <1% of environmental isolates encode these elements (McLaughlin et al., 2005). The historical reporting of *V. parahaemolyticus* infections in this region suggests that either infections have been caused by strains lacking major virulence factors, resident strains encoding these virulence factors were not detected using the methods employed by this study, or both.

Results of this study demonstrated a high level of diversity among isolates as measured by serotype distribution, presence/absence of pandemic markers, and rep-PCR banding patterns. Strains isolated in this study represented 9 O-antigens and 27 K-antigens, as well as untypable strains, a measure of antigenic diversity of natural isolates in this region. Mutations within antigen coding regions of the genome are common, as well as lateral transfer, allowing strains to adapt to microenvironments of the environment or evade predation by grazing protozoa (Lerouge et al., 2001; Woo et al., 2001; Wildschutte et al., 2004). Molecular divergence was noted by the heterogeneity observed among O3:K31 and O2:K28 strains by rep-PCR analysis suggesting that serology does not necessarily correlate with

**FIGURE 3 | Dendrogram showing relatedness of** *V. parahaemolyticus* **strains by rep-PCR.** Asterisks identify strains that are GS-PCR-positive. Numbers on branches indicate degree of divergence between isolates.

**Table 3 | Results of binary logistic regression analysis between** *V. parahaemolyticus* **water temperature.**


genome architecture. This genomic heterogeneity indicates the necessity of classifying strains by methods other than serology. The high degree of divergence among environmental *V*. *parahaemolyticus* strains in the Black Sea is corroborated by reports of similar findings in geographically distant regions (Wong et al., 1999; Matsumoto et al., 2000; Alam et al., 2009; Yu et al., 2011; Ellis et al., 2012; Paranjpye et al., 2012).

*V. parahaemolyticus* was detected across a broad range of salinities (3.4–20.8-) (**Table 2**). However, it was not significantly associated with *V. parahaemolyticus* presence in our model. This is most likely due to the relative stability of salinity readings at each site over the course of the study (data not shown). *V. parahaemolyticus* is a known member of estuarine and marine environments and salinity values detected during this study were typical of brackish waters (0.5 <sup>&</sup>gt; <sup>30</sup>-) suggesting a suitable salinity regime for *V. parahaemolyticus* presence at most sampling points. *V. parahaemolyticus* seasonality was observed at all sites, with a clear trend of increasing numbers as water temperatures increased from May to September. The organism was isolated from water samples at temperatures as low as 8◦C, but more frequently (ca. 93% of strains) at temperatures greater than 17◦C (**Table 1**). The highest percentage of total variance in detection, related to temperature, was at Green Cape (percent of total variance = 70, *P* < 0.05). At each site, the total variance in *V. parahaemolyticus* detection was significantly related to an increase in water temperature. However, these associations were not as strong for the Batumi Bulvard (43.18), Chorokhi estuary (22.01), and Supsa estuary (31.23) sites (**Table 3**). Interestingly, the associations between water temperature and *V. parahaemolyticus* detection were weaker for the two estuarine sites. Salinities at these two sites were much lower than the non-estuarine sites (Batumi and Green Cape) suggesting that either salinity played a role in *V. parahaemolyticus* presence, even though it did not show up as significant in our model, or that an unmonitored parameter common to both estuarine environments influenced *V. parahaemolyticus* presence. This trend is indicative of the patchiness of *V. parahaemolyticus* distribution in water bodies suggesting that environmental conditions are noticeably different at different locations within the same water body and that these differences contribute to *V. parahaemolyticus* presence.

In summary, an antigenically diverse population of *V*. *parahaemolyticus* inhabits the Georgian coast of the Black Sea. Although none of the strains collected during this study were Kanagawa phenomena-positive or *tdh* and *trh*-positive, the TTSS1 effector proteins and TLH were present in some isolates, which included a possible serovariant of the *V. parahaemolyticus* O3:K6 pandemic clone. These results, together with epidemiological data demonstrating strains lacking pathogenicity islands can cause disease, suggest there is a risk associated with occurrence of *V*. *parahaemolyticus* in Black Sea coastal waters. Warmer temperatures in the spring and summer lead to increased densities of *V. parahaemolyticus*. Recent clinical data on isolation of TDH-, TRH-, and TTSS2-negative *V*. *parahaemolyticus* suggests these strains represent underreported etiological agents of diarrhea, similar to *V*. *cholerae* non-O1/non-O139 strains lacking major virulence factors (Safrin et al., 1988; Ko et al., 1998; Lukinmaa et al., 2006; Shannon and Kimbrough, 2006; Chatterjee et al., 2009; Hasan et al., 2012; Marin et al., 2013). The high frequency of detection of *V*. *parahaemolyticus* lacking major virulence factors but associated with severe infection, suggests recreational water and shellfish harvesting areas in Georgia should be monitored, especially when water temperatures are seasonally high.

#### **ACKNOWLEDGMENTS**

The research described in this report was made possible by financial support provided by the Biological Threat Reduction Program of the U.S. Defense Threat Reduction Agency (DTRA) (Project # GG-13) through Bechtel National Inc., sponsor account number 24914416HC4W00000006. Christopher J. Grim was supported by an IC Postdoctoral Research Fellowship (NGA Grant #HM15820612010). Partial funding for this study was provided by NIH Grant No. 2RO1A1039129-11A2 and NSF Grant No. 0813066. Opinions, interpretations, conclusions, and recommendations are those of the authors and are not necessarily endorsed by the U.S. Army.

#### **REFERENCES**


*Vibrio parahaemolyticus* in New Hampshire shellfish waters as determined by multilocus sequence analysis. *Appl. Environ. Microbiol.* 78, 3778–3782. doi: 10.1128/AEM.07794-11


by *Vibrio cholerae* non-O1, non-O139, report of 3 hospitalized cases. *Diagn. Microbiol. Infect. Dis.* 54, 1–6. doi: 10.1016/j.diagmicrobio.2005.06.020


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The Associate Editor declares that despite being affiliated to the same institution as the authors Bradd J. Haley, Christopher J. Grim, Arlene J. Chen, Elisa Taviani, Rita R. Colwell and Anwar Huq, the review process was handled objectively and no conflict of interest exists.

*Received: 26 November 2013; paper pending published: 13 December 2013; accepted: 21 January 2014; published online: 10 February 2014.*

*Citation: Haley BJ, Kokashvili T, Tskshvediani A, Janelidze N, Mitaishvili N, Grim CJ, Constantin de Magny G, Chen AJ, Taviani E, Eliashvili T, Tediashvili M, Whitehouse CA, Colwell RR and Huq A (2014) Molecular diversity and predictability of Vibrio parahaemolyticus along the Georgian coastal zone of the Black Sea. Front. Microbiol. 5:45. doi: 10.3389/fmicb.2014.00045*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Haley, Kokashvili, Tskshvediani, Janelidze, Mitaishvili, Grim, Constantin de Magny, Chen, Taviani, Eliashvili, Tediashvili, Whitehouse, Colwell and Huq. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## Impact of Hurricane Irene on *Vibrio vulnificus* and *Vibrio parahaemolyticus* concentrations in surface water, sediment, and cultured oysters in the Chesapeake Bay, MD, USA

#### *Kristi S. Shaw1\*, John M. Jacobs <sup>2</sup> and Byron C. Crump1,3 †*

*<sup>1</sup> Horn Point Laboratory, Center for Environmental Science, University of Maryland, Cambridge, MD, USA 2Cooperative Oxford Laboratory, National Centers for Coastal Ocean Science, National Ocean Service, Oxford, MD, USA*

*3College of Earth, Ocean, and Atmospheric Science, Oregon State University, Corvallis, OR, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Bradd Haley, Environmental Microbial and Food Safety Laboratory – United States Department of Agriculture, USA*

*Crystal N. Johnson, Louisiana State University, USA*

#### *\*Correspondence:*

*Kristi S. Shaw, Horn Point Laboratory, Center for Environmental Science, University of Maryland, P.O. Box 775, Cambridge, MD 21613, USA e-mail: krististevensshaw@gmail.com*

#### *†Present address:*

*Byron C. Crump, College of Earth, Ocean, and Atmospheric Science, Oregon State University, Corvallis, OR, USA*

To determine if a storm event (i.e., high winds, large volumes of precipitation) could alter concentrations of *Vibrio vulnificus* and *V. parahaemolyticus* in aquacultured oysters (*Crassostrea virginica*) and associated surface water and sediment, this study followed a sampling timeline before and after Hurricane Irene impacted the Chesapeake Bay estuary in late August 2011. Aquacultured oysters were sampled from two levels in the water column: surface (0.3 m) and near-bottom (just above the sediment). Concentrations of each *Vibrio* spp. and associated virulence genes were measured in oysters with a combination of realtime PCR and most probable number (MPN) enrichment methods, and in sediment and surface water with real-time PCR. While concentration shifts of each *Vibrio* species were apparent post-storm, statistical tests indicated no significant change in concentration for either *Vibrio* species by location (surface or near bottom oysters) or date sampled (oyster tissue, surface water, and sediment concentrations). *V. vulnificus* in oyster tissue was correlated with total suspended solids (*r* = 0.41, *P* = 0.04), and *V. vulnificus* in sediment was correlated with secchi depth (*r* = −0.93, *P* < 0.01), salinity (*r* = −0.46, *P* = 0.02), tidal height (*r* = −0.45, *P* = 0.03)*,* and surface water *V. vulnificus* (*r* = 0.98, *P* < 0.01). *V. parahaemolyticus* in oyster tissue did not correlate with environmental measurements, but *V. parahaemolyticus* in sediment and surface water correlated with several measurements including secchi depth [*r* = −0.48, *P* = 0.02 (sediment); *r* = −0.97, *P* < 0.01 (surface water)] and tidal height [*r* = −0.96, *P* < 0.01 (sediment), *r* = −0.59, *P* < 0.01 (surface water)]. The concentrations of *Vibrio* spp. were higher in oysters relative to other studies (average *V. vulnificus* <sup>4</sup> <sup>×</sup> <sup>10</sup><sup>5</sup> MPN g−1, *V. parahaemolyticus* <sup>1</sup> <sup>×</sup> <sup>10</sup><sup>5</sup> MPN g−1), and virulence-associated genes were detected in most oyster samples. This study provides a first estimate of storm-related *Vibrio* density changes in oyster tissues, sediment, and surface water at an aquaculture facility in the Chesapeake Bay.

**Keywords: aquacultured oyster,** *Vibrio vulnificus***,** *Vibrio parahaemolyticus***, sediment resuspension, wind event, Chesapeake Bay, estuary, storm event**

#### **INTRODUCTION**

Storm events are thought to be important mechanisms for the distribution of benthic *Vibrio* populations into the water column via resuspension of sediments associated with high winds, and flushing due to large volumes of precipitation (Randa et al., 2004; Fries et al., 2008; Wetz et al., 2008; Johnson et al., 2010). Frequent storm events in the Chesapeake Bay are associated with the summer season, a time when *Vibrio vulnificus* and *V. parahaemolyticus*, autochthonous bacteria known to cause human illness, are at their highest densities in surface waters (Wright et al., 1996; Parveen et al., 2008; Jacobs et al., 2010; Johnson et al., 2012). The frequency and intensity of storm events are predicted to escalate in response to global climate change (Goldenberg et al., 2001), with increases in peak wind intensities and near-storm precipitation (Meehl et al., 2007) likely impacting mid-Atlantic

areas such as the Chesapeake Bay. In the Chesapeake Bay, a shallow, partially mixed estuary prone to tidal circulation (average depth 6.5 m), storm events may be expected to increase the overall *Vibrio* density in surface waters with relatively moderate wind speed and associated wave action. Increases in post-hurricane*Vibrio* infection has been documented (e.g., Hurricane Katrina), with a resultant need for heightened clinical awareness, particularly of wound infections, following exposure to flood waters (Centers for Disease Control and Prevention [CDC], 2005). Based on the reported increases in storm-related *Vibriosis* in other areas of the United States, it is conceivable that storm-induced increases in Chesapeake Bay *Vibrio* density may be linked to future *Vibriosis* outbreaks.

According to the U.S. Environmental Protection Agency, the Chesapeake Bay is home to 25% of the total approved shellfish harvesting waters in the United States (Environmental Protection Agency [EPA], 2011). Recently, the Chesapeake Bay has become a site of interest for oyster (*Crassostrea virginica*) aquaculture production to supplement the dwindling wild harvest, both through on-bottom (submerged land) and off-bottom (water column) leases (Maryland Department of Natural Resources, Shellfish Aquaculture Program). As of January 2013, 169 aquaculture operation permit applications (∼4000 acres) were submitted to Maryland Department of Natural Resources for water-column and submerged-land leases (Webster, University of Maryland Extension, personal communication), and a total of 300 submerged-land leases (∼3500 acres) and 23 water-column leases (∼94 acres) permitted. A small number of new aquaculture operations are in year-round production of retail oysters, with the supposition that many new operations will soon be joining their ranks.

Summer is generally considered to be a viable oyster harvest season in Maryland, but summer is also when *Vibrio* populations reach their peak in the Bay (Wright et al., 1996; Parveen et al., 2008; Jacobs et al., 2010; Johnson et al., 2012). Studies are currently being conducted to determine ways to reduce *Vibrio* concentrations in oysters (e.g., high salinity relay), but factors influencing the accumulation of high numbers or virulent strains of *Vibrio* in oysters are not completely understood (Warner and Oliver, 2008; Johnson et al., 2010; Froelich and Oliver, 2013). Thus, the harvest of oysters during seasons when surface water *Vibrio* populations are at high densities could become a pressing issue for seafood safety. If *Vibrio* density in oysters increases after storm events, shellfish managers may need to institute shellfish harvest closure periods to allow for oyster depuration or wait for suitable environmental conditions that favor a reduction in *Vibrio* concentrations, such as cooler water temperatures.

This study was conducted to test the hypothesis that a storm event, using Hurricane Irene as a proxy, generates enough wave energy to cause resuspension of sediment that would cause an increase in oyster-tissue density of *V. vulnificus* and *V. parahaemolyticus*. Oysters were tested in Taylor-style surface-water floats (Luckenbach et al., 1999) and in on-bottom cages, to determine if there was an accumulation difference based on water column position. Results from this study provide a first estimate of storm-related *Vibrio* density changes in oyster tissues, sediment and surface water at an aquaculture facility in the Chesapeake Bay.

### **MATERIALS AND METHODS**

#### **SAMPLING SITE**

The study was conducted at an oyster aquaculture facility in a mesohaline tributary of the Chesapeake Bay. The oyster farm was approximately 250,000 m2(6 acres) with a water depth of approximately 1.2 m (4ft) at low tide and 2.1 m (7ft) at high tide. Sediment types at the farm ranged from predominantly sand to predominantly silt. The sampling location within the oyster farm was chosen for the predominance of silty sediment (20.4% sand: 66.6% silt: 13.0% clay; Owens, Cornwell, University of Maryland Center for Environmental Science, personal communication), which is representative of the biodeposition typically produced by oysters (Haven and Morales-Alamo, 1972). Three sampling sub-locations were selected along the outermost matrix of oyster floats, which

covered approximately 1 acre, both for sediment composition and the likelihood of the area being unprotected from wind events and resultant resuspension activity. Estimates of wind speeds and resultant wave height were made using equations from Young and Verhagen (1996). Calculations of maximum bottom-sheer stress were made according to (Sanford, 1994) incorporating an approximate bottom depth of 1 m and sand grain roughness of 0.0005 m. Sand grain roughness is a measurement of characteristic bottom roughness height for use in hydrodynamic calculations. Erosion rate was calculated using the equation E (g m−<sup>2</sup> <sup>h</sup>−1) <sup>=</sup> Mo (kg m−<sup>2</sup> s <sup>−</sup><sup>1</sup> Pa−1) <sup>×</sup> 3600 s h−<sup>1</sup> <sup>×</sup> 1000 g kg−<sup>1</sup> <sup>×</sup> (τb–τc) (Pa), with site-specific estimates of τ<sup>c</sup> = 0.025 Pa and Mo = 0.000315 kg m−<sup>2</sup> s <sup>−</sup><sup>1</sup> Pa−<sup>1</sup> (τb: bottom-related sheer stress; τc: current-related shear stress; Pascal (Pa); Mo is erosion rate constant; Sanford, Kwon, University of Maryland Center for Environmental Science, personal communication). These calculations do not acknowledge the potential for a wave-dampening effect by the large array of oyster floats tied together at the aquaculture site, although a physical oceanographer conducting experiments at the same site shares that long period waves at the bottom of the water column are damped out by perhaps as much as 50% by the floats, but not so much that resuspension would be negated (Sanford, University of Maryland, personal communication).

#### **ENVIRONMENTAL SAMPLE COLLECTION**

Baseline surface water, oyster, and sediment samples were collected from the field location on August 26, 2011, the day before Hurricane Irene and associated storm impacts were forecast to be present along the Maryland coastline. Subsequent samples were taken at time points 1, 4, and 8 days after Hurricane Irene. All samples were collected at approximately 10:00 A.M. to approximate a uniform water and air temperature at the time of sampling due to solar irradiation.

Surface-water samples were collected at each sampling location in sterile wide mouth polypropylene 1 L bottles (Nalgene Thermo Scientific 2105-0032) following the methods described by Jacobs et al. (2009). Surface water (200 mL) was filtered through a 0.22 μm Sterivex-GP polyethersulfone filter (Millipore, Billerica, MA, USA) using a 60 mL BD luer lock syringe (BD, Franklin Lakes, NJ, USA), wrapped in Parafilm M laboratory wrapping film (Bemis Flexible Packaging, Oshkosh, WI, USA), and sealed in a labeled 7 oz Whirlpak bag (Nasco, Fort Atkinson, WI, USA). Filters were stored on ice until return to the laboratory (∼1 h), where they were stored at −20◦C until DNA extraction.

#### **PHYSICAL/CHEMICAL MEASUREMENTS**

Temperature, salinity, conductivity, and dissolved oxygen were measured using a YSI Model 85 (YSI, Yellow Springs, OH, USA) at 0.3 m depth and near-bottom (∼0.3 m off bottom). Secchi depth was recorded to the nearest 0.05 m. Total suspended solids (TSS) measurements were completed using 250–400 mL of surface water, filtered onto pre-weighed 47 mm glass fiber filters (Whatman GF/F, GE Healthcare Life Sciences, Piscataway, NJ, USA).

#### **SAMPLE SIZE**

Based on standard deviations reported in Johnson et al. (2010), sample size needed was calculated for a statistical power of 0.8, significance criterion of 0.05, and preferred detection difference of 500 CFU g−1. Based on this calculation, three samples were required for each depth (top and bottom), per sampling period.

#### **OYSTER SAMPLE COLLECTION**

Oyster samples (*C. virginica*) were collected from the top (*n* = 3) and bottom (*n* = 3) of the water column on each of the four sampling dates. Collected oysters [six per sample (Kaufman et al., 2003)] had shell heights (oyster hinge to opposite edge periphery) of ∼ 8 cm (3.1 in). Surface water oyster samples were collected from Taylor-style floats, which remained submerged in water continuously, and bottom-water oyster samples were enclosed in 1.3 cm mesh bags deployed inside of crab pots to keep the oysters at the bottom of the water column, but out of the sediment layer. Bottom oysters, collected from identical resident oyster stock as surface oyster samples, were deployed 1 month before the commencement of this study for acclimation purposes. Collected oysters were immediately placed on ice and processed within an hour.

Crab pots consistently had a coating of top layer sediment (∼1 cm) on the bottom of the pot from being deployed in the sediment. That sediment was collected at each of the three sites by filling a 50 mL Falcon sterile polypropylene conical centrifuge tube (BD Vacutainer Labware Medical 352070). Sediment samples were placed on ice, and stored frozen at −20◦C.

#### **OYSTER PROCESSING**

On each sampling date, a total of 36 oysters were examined, divided into six samples, for a total of 144 analyzed oysters over the four sampling periods. One sample of *n* = 6 oysters were collected from both the top and bottom layers at each of three sampling strata (Kaufman et al., 2003) and were homogenized following the three-tube MPN method described in the U.S. Food and Drug Administration Bacteriological Analytical Manual (BAM; DePaola and Kaysner, 2004) with slight modifications. Briefly, oysters were scrubbed, shucked with a sterile knife into a sterile blender, diluted with an equal weight of sterile phosphatebuffered-saline (Food and Drug Administration [FDA], 1998) and blended for 90 s to create a 1:1 (wt:wt) shellfish:diluent homogenate. A 1:20 dilution of oyster homogenate was made in triplicate by adding 1 mL of the 1:1 diluted homogenate to 9 mL alkaline peptone water (APW; 1% peptone, 1% NaCl, pH 8.5 <sup>±</sup> 0.2). Additional triplicate 10-fold dilutions to 5 <sup>×</sup> <sup>10</sup>−<sup>7</sup> were prepared volumetrically by transferring 1 mL portions into 9 mL APW. Following overnight incubation at 35 ± 2◦C, the top 1 mL of tubes showing growth was collected and frozen at −20◦C.

#### **DNA EXTRACTION, DETECTION, AND QUANTIFICATION**

DNA from surface water was extracted following a modified MO BIO Powersoil extraction protocol (Jacobs et al., 2009), and DNA from sediments was extracted using the standard MO BIO Powersoil extraction protocol. Extracted DNA was stored at −80◦C. Quantitative PCR was used to quantify CFU mL−<sup>1</sup> in water and CFU g−<sup>1</sup> in sediment. The reported extraction efficiency of surface water and sediment samples using their respective methods were comparable (Jacobs et al., 2010; Lloyd et al., 2010).

DNA template was obtained from MPN cultures by producing crude cell lysates by boiling 1 mL aliquots of APW cultures in 2 mL micro-centrifuge tubes for 10 min. Following boiling, tubes were plunged into ice until cool and then centrifuged at 14,000 × *g* for 2 min. Supernatant template was added to real-time PCR reactions (3–5 uL; see PCR methods) to determine presence or absence of *V. vulnificus* and *V. parahaemolyticus* in cultured samples. Biorad CFX96 TouchTM Real-Time PCR Detection System (Bio-rad, Hercules, CA, USA) was used to confirm the species with primers designed to detect*V. vulnificus* (Panicker and Bej, 2005) or*V. parahaemolyticus* (Nordstrom et al., 2007). Following initial detection, samples testing positive for either species were subjected to further PCR testing for virulence genes (*V. vulnificus*: virulence correlated gene, clinical variant (*vcgC*; Baker-Austin et al., 2010); *V. parahaemolyticus*: thermostable direct hemolysin (*tdh*), thermostable related hemolysin (*trh*) genes (Nordstrom et al., 2007).

Quantitative PCR was performed on surface water and sediment sample extracts by using 2.50 uL of 10X PCR Buffer (Qiagen, Valencia, CA, USA), 1.25 uL of 25 mM MgCl2 (Qiagen), 0.50 uL of 10 mM dNTP's solution (Qiagen), 5 uL Q solution (Qiagen), 0.45 uL of 5 U/uL TopTaq DNA polymerase (Qiagen), 0.188 uL of 10 uM internal control primers (each), 0.375 uL of 10 uM internal control probe, 2 uL internal control DNA, 0.50 uL of 10 uM primer (each), 0.188 uL of 10 uM probe, and 3 uL DNA template per reaction, with the exception of the *V. vulnificus vcgC* assay, in which 5 uL of DNA template was used. DNase–RNase free water was added in a quantity sufficient for a 25 uL total reaction volume. Two-stage qPCR cycling parameters were optimized to the conditions as described in Shaw et al. (2014). A unique internal control, including a primer set, probe and internal control DNA, was incorporated simultaneously into each assay, excluding *V. vulnificus vcgC*, to test for the presence and influence of inhibitors (Nordstrom et al., 2007). Positive controls used for each qPCR were *V. parahaemolyticus* USFDA TX2103 and *V. vulnificus* ATCC 27562. Standard curves were constructed as reported in Jacobs et al. (2010) from spiked environmental water and used during each qPCR analysis with appropriate parameters. Cycle threshold (Ct) value was plotted against the slope of the standard curve to determine PCR unit quantity.

#### **MOST PROBABLE NUMBER CALCULATION USING QPCR RESULTS**

Corresponding qPCR-MPN values were derived using the U.S. Food and Drug Administration MPN calculator, downloaded from the online publication "Bacteriological Analytical Manual, Appendix 2: Most Probable Number from Serial Dilutions."1

#### **STATISTICAL ANALYSIS**

Statistical analysis was completed using Intercooled Stata 9.1 for Macintosh statistical software (StataCorp LP, College Station, TX, USA). Oyster MPN g−1, sediment and surface water data (CFU mL−1) were log transformed (log10) to equalize variances. Each data set was analyzed for normality. Normally distributed oyster MPN g−1data were analyzed with multivariate analysis of variance

<sup>1</sup>http://www.fda.gov/Food/scienceResearch/LaboratoryMethods/Bacteriological AnalyticalManualBAM/ucm109656.htm

(MANOVA) to test for differences in sampling location (top vs. bottom oyster concentrations) and sampling date for each species of *Vibrio*. Surface water and sediment samples were tested with one-way analysis of variance (ANOVA). Data sets not meeting normality criteria were analyzed with Kruskal–Wallis non-parametric rank test for differences in sampling location and sampling date. Pearson pairwise correlation analysis was conducted for the experimental variables of oyster MPN g−1, surface water CFU mL−1, sediment CFU g−1, MPN g−1, salinity, temperature, TSS, dissolved oxygen, tidal height, and secchi depth. Spearman's rank correlation analysis was used for non-normally distributed data. Due to low sample numbers, virulence associated gene (*tdh* and *vcgC*) concentrations were not included in correlation analysis.

#### **RESULTS**

#### **HURRICANE DETAILS**

During the early morning hours of August 28, 2011, Hurricane Irene was just off the Delmarva coastline and the associated winds and rain impacted the Chesapeake Bay region. At the study site, there were ∼18.4 cm (7.23 in) of rainfall (NOAA, 2011). Wind gusts were recorded in excess of 26 m s−1(58 MPH). Highest sustained winds were measured at 19.5 ms−<sup>1</sup> (44 MPH) at 23:30 h on August 27, 2011 (Avila and Cangialosi, 2011; **Figure 1A**). Barometric pressure over the area reached a minimum of 976.2 mb at ∼18:40 h on August 28, 2011 (**Figure 1B**). Tidal height did not deviate from the predicted normal height on the first day of sampling, so there was no hurricane-related tidal forcing at the first sampling time point.

#### **PHYSICAL AND CHEMICAL CONDITIONS**

All physical and chemical measurements, whether taken at ∼0.3 m below the surface or ∼0.3 m from bottom, were found to be the same on each sampling date. As no water column stratification was detected, only one value per parameter is reported for each sampling date. Twenty-four hours after Hurricane Irene, salinity at the study site decreased from 10.6 to 8.0, and by day 8 returned to 9.9. Dissolved oxygen increased from 5.01 mg L−<sup>1</sup> to 6.37 mg L−<sup>1</sup> after the storm, and remained above 6 mg L−1. Water temperature decreased from 25.6◦C to 24.1◦C after the storm and by day 8 increased to 25.7◦C. Secchi depth increased from 0.4 to 0.45 m on the day after the storm, returned to 0.4 m on day 4, and increased to 0.55 m on day 8 (**Figure 2**). TSS started at 25.1 mg L−<sup>1</sup> and decreased over the course of the study to 19.5 mg L−<sup>1</sup> (day 1), 14.7 mg L−<sup>1</sup> (day 4), and 14.9 mg L−<sup>1</sup> (day 8). Tidal height ranged from low tide during initial sampling efforts [pre-storm: 0.20 m above mean lower low water (MLLW), Day 1: 0.15 m above MLLW] to high tide (day 4: 0.38 m above MLLW; day 8: 0.55 m above MLLW). While changes in temperature, salinity, dissolved oxygen, secchi depth, and TSS were small, tidal height was significantly correlated with temperature (*P* = 0.001, *r* = 0.6251), TSS (*P* < 0.001, *r* = −0.7512), and secchi depth (*P* < 0.001, *r* = 0.6621).

#### **RESUSPENSION CALCULATIONS**

Rates of erosion were calculated based on highest wind gusts (26.9 and 22.6 m s−1) and highest sustained wind speeds (9−9.8 m s−1).

Most winds during the storm were moving in a north-northeast or northeast direction. Erosion rates were predicted to range from 2,343 to 3,616 g m−<sup>2</sup> h−<sup>1</sup> during periods of wind gusts and 487 to 730 g m−<sup>2</sup> h−<sup>1</sup> during highest sustained winds. Given the lowest wind speed (m s−1) during the height of the storm, the oyster farm would have expected an erosion rate of <sup>∼</sup><sup>3</sup> <sup>×</sup> <sup>10</sup><sup>5</sup> g sediment h−1.

#### *Vibrio vulnificus*

#### *Oyster MPN*

*Vibrio vulnificus* oyster MPN g−<sup>1</sup> data were not normally distributed and Kruskal–Wallis non-parametric rank test determined no statistical difference in oyster *V. vulnificus* (MPN g−1) by location (top vs. bottom) or by date sampled. Spearman's rank correlation analysis of oyster*V. vulnificus* MPN g−<sup>1</sup> showed significant associations with TSS (*P* = 0.0455, *r* = 0.4119; **Table 1**).

Although non-significant statistically, a small concentration increase in average*V. vulnificus* in oysters (MPN g−1) was detected

between the first sampling pre-storm (August 26, 2011) and 1 day after the storm (August 29, 2011; **Table 2**). Average *V. vulnificus* decreased approximately between day 1 and day 4 post-storm, and then increased between day 4 and day 8. Despite these shifts, a very small change (1.6%) was measured in total *V. vulnificus* in oysters the entire study period.

#### *Surface water and sediment*

One-way ANOVA analysis of sediment and surface water CFU mL−<sup>1</sup> determined no statistically significant difference between dates for either sediment or surface water. Pearson's correlation analysis of sediment*V. vulnificus* revealed significant negative relationships with the environmental variables of salinity (*P* = 0.0224, *r* = −0.4641), secchi depth (*P* < 0.0001, *r* = −0.9343) and tidal height (*P* = 0.0256, *r* = −0.4548). Correlation analysis of surface water *V. vulnificus* found significant associations with sediment *V. vulnificus* concentrations (*P* < 0.0001, *r* = 0.9882) and secchi depth (*P* < 0.0001, *r* = −0.8917; **Table 1**).

While concentration changes detected were non-significant, average *V. vulnificus* decreased in surface waters and sediment on day 1 post-storm, increased on day 4, and decreased again to the lowest of this study's detected*V. vulnificus* concentrationsfor either substrate on day 8 (**Table 2**).

#### *Vibrio vulnificus virulence correlated gene*

The *V. vulnificus vcgC* was detected in oysters during each of the sampling dates, but concentrations were reduced during the day 1 and 4 sampling time points (393 and 105 MPN g−1, respectively) relative to concentrations pre-storm (789 MPN g−1) and on day 8 (622 MPN g−1; **Table 2**). The percentage *V. vulnificus vcgC* MPN g−<sup>1</sup> of overall *V. vulnificus* MPN g−<sup>1</sup> was appreciably the same on all sampled dates (0.2%). *V. vulnificus vcgC* was detected in both surface and bottom sampled oysters, but not in sediment or surface waters during this study.

### *Vibrio parahaemolyticus*

#### *Oyster MPN*

Multivariate analysis of variance found no statistical difference between the sampling locations or sampling dates for *V. parahaemolyticus* MPN g−1values of oysters. Oyster *V. parahaemolyticus* MPN g−<sup>1</sup> did not correlate significantly (Pearson's correlation) with any of the environmental variables tested (**Table 1**).

While not significant statistically, concentration changes of average overall *V. parahaemolyticus* MPN g−<sup>1</sup> increased 1 day post-storm from pre-storm concentrations and decreased 4 days post-storm, with a final increase on day 8 post-storm.

#### *Surface water and sediment*

One-way ANOVA analysis of difference among sampling dates for sediment and surface water CFU mL−<sup>1</sup> showed no statistically significant difference between dates for either sediment or surface water. Correlation analysis of sediment *V. parahaemolyticus* CFU g−<sup>1</sup> revealed significant associations with the environmental variables of temperature (*P* = 0.0124, *r* = −0.5019), TSS (*P* < 0.0001, *r* = 0.8569), dissolved oxygen (*P* = 0.0094, *r* = −0.5187), secchi depth (*P* = 0.0161, *r* = −0.4856), and tidal height (*P* < 0.0001, *r* = −0.9592). Correlation analysis of surface water *V. parahaemolyticus* CFU mL−<sup>1</sup> found a significant negative relationship with salinity (*P* = 0.0414, *r* = −0.4193), secchi depth (*P* < 0.0001, *r* = −0.9727), and tidal height (*P* = 0.0024, *r* = −0.5903). Conversely, a strong and statistically positive association was found between surface water *V. parahaemolyticus* and *V. vulnificus* CFU mL−<sup>1</sup> (*P* < 0.0001, *r* = 0.9595) and between surface water *V. parahaemolyticus* CFU mL−<sup>1</sup> and sediment*V. vulnificus* CFU g−<sup>1</sup> (*<sup>P</sup>* <sup>&</sup>lt; 0.0001,*<sup>r</sup>* <sup>=</sup> 0.9866; **Table 1**).

While not statistically significant, concentration changes of average *V. parahaemolyticus* were detected, with decreases in surface waters, but increases in sediment, 1 day after the storm. Surface water *V. parahaemolyticus* then increased on day 4 poststorm and decreased on day 8 post-storm. Conversely, sediment *V. parahaemolyticus* decreased on day 4 and decreased further on day 8 (**Table 2**).

#### *Vibrio parahaemolyticus tdh/trh*

The *trh* gene was not detected in any of the oyster MPN cultures, nor the sediment or surface water samples. The *tdh* gene was detected in oyster MPN cultures at all time points except on day 8. Two samples were positive for *tdh* during pre-storm sampling (average 658 MPN g−1), and three samples were positive post-storm (day 1, 1239 MPN g−1; day 8, 294 MPN g−1). Concentrations of *tdh* decreased over the sampling period, although overall percent*V. parahaemolyticus tdh* MPN g−1, when compared to total*V. parahaemolyticus* MPN g−1, was greatest at day 4 (2.9%). The percent of sampled oysters positive for *tdh* was lowest on day 8 [(2/6) = 33%].

#### **DISCUSSION**

Hurricane Irene produced a significant wind event for the Chesapeake Bay region and wave action was sufficient to cause sediment resuspension at the studied aquaculture facility, according to estimates of erosion based on wind speed


#### **Table 1 | Correlation table of environmental parameters and** *Vibrio* **concentrations in oysters, sediment and surface water.**

*Vp, Vibrio parahaemolyticus; Vv, Vibrio vulnificus; MPN, most probable number; CFU, colony forming units; ppt, parts per thousand;* ◦*C, Celsius; mg L*−*1, milligrams per liter; m, meter. \*Spearman's rank correlation used.*

and direction. Additionally, there was a large amount of precipitation (18 cm) during the storm event. Although our data lacks a sampling time point during the storm, *in situ* continuous monitoring data archives of turbidity (accessed at Maryland Department of Natural Resources "Eyes on the Bay;"2 depict sharp spikes in nephelometric turbidity units (NTU) during the peak of the storm winds and a rapid subsequent decrease of NTU, most likely due to the large amount of rainfall experienced during the storm and a resultant flushing effect (**Figure 3**). This flushing effect may be the cause of reduced turbidity and lowered surface

water CFU mL−<sup>1</sup> for both *Vibrio* species 1 day after the storm.

In general, many concentrations of *V. vulnificus* and *V. parahaemolyticus* detected during this study were greater than those found in similar studies documenting the detection of these species in the same sampled matrices in the Chesapeake Bay. Maximum concentrations of *Vibrio* detected in previous studies of oyster tissue were considerably lower [*V. parahaemolyticus*: 6.0 <sup>×</sup> <sup>10</sup><sup>2</sup> CFU g−<sup>1</sup> (Parveen et al., 2008), 1.0 <sup>×</sup> 104CFU g−1(Johnson et al., 2012); *V. vulnificus*: 1.2 <sup>×</sup> 104CFU g−<sup>1</sup> (Johnson et al., 2012)] than the findings of this study (*V. parahaemolyticus*: 4.1 <sup>×</sup> 105 MPN g−1; *V. vulnificus*: 1.14 <sup>×</sup> <sup>10</sup>6MPN g−1). In addition, Johnson et al. (2012) detected lower surface water and sediment

<sup>2</sup>http://mddnr.chesapeakebay.net/eyesonthebay/index.cfm


*V. vulnificus* concentrations [surface water: 150 CFU mL−<sup>1</sup> vs. 1.2 <sup>×</sup> 103 CFU mL−1(this study)*;* sediment: 3.5 <sup>×</sup> <sup>10</sup><sup>4</sup> CFU <sup>g</sup>−<sup>1</sup> vs. 3.6 <sup>×</sup> <sup>10</sup><sup>5</sup> MPN g−1(this study)], although *V. parahaemolyticus* concentrations found in Johnson et al. (2012) were approximately double the concentrations detected in this study [surface water: 60 CFU mL−<sup>1</sup> vs. 17.5 CFU mL−<sup>1</sup> (this study); sediment: 1.5 <sup>×</sup> <sup>10</sup><sup>4</sup> CFU g−1vs. 6.0 <sup>×</sup> <sup>10</sup><sup>3</sup> MPN g−<sup>1</sup> (this study)]. The lower oyster MPN g−<sup>1</sup> and surface water/sediment *V. vulnificus* values from previous studies may be due to a difference in sampling depth for oysters (i.e., natural oyster bar depth and open water versus near shore shallows) or a difference in recovery efficiencies of methodologies used in either study, such as under-detection (culture-based methods, previous studies) or detection of non-viable cells by qPCR (direct detection, this study) in sampled surface water and sediment matrixes.

While there was large variation in the average *V. vulnificus* and *V. parahaemolyticus* cell densities in oysters, surface water, and sediment, the values quantified in each of these substrates was not significantly different over the course of the study. There was a species difference in oyster tissue concentration immediately after the storm, with *V. parahaemolyticus* increasing substantially, but *V. vulnificus* increasing only slightly. A recent, similar study (i.e., sampling frequency, salinity, and temperature range) comparing oyster, sediment, and water concentrations of *V. vulnificus* and *V. parahaemolyticus* in the Gulf of Mexico reported comparable changes in oyster tissue *Vibrio* concentrations for both species over the course of the study (Givens et al., 2014). These findings contrast with the post-storm *Vibrio* concentration changes seen in this study, suggesting a species-specific dynamic post-storm during this study. Additionally, it has been shown that *V. vulnificus* outnumbers *V. parahaemolyticus* in sediment, oyster tissue and the water column (Johnson et al., 2010). During this study, *V. parahaemolyticus* cell g−<sup>1</sup> was approximately 5% of the total *V. vulnificus* cell g−<sup>1</sup> in sediment, which is consistent with the findings of Johnson et al. (2010). However, despite the relative dominance of *V. vulnificus* in sediments, post-storm increases in *Vibrio* were dominated by*V. parahaemolyticus,* suggesting speciesspecific variation during this study in the degree to which these bacteria were resuspended from sediments or were retained in oyster tissues, perhaps differing from *V. vulnificus* in properties of adhesion to marine aggregates, which may have been subsequently filtered by oysters.

Interestingly, on day 4 post-storm, oyster tissue *Vibrio* MPN g−1decreased precipitously from pre-storm concentrations (−74%, *V. vulnificus*; −56% *V. parahaemolyticus*), while surface water CFU mL−<sup>1</sup> and sediment CFU g−1increased substantially (+337 and +84%, respectively; **Table 2**). On day 8, oyster tissue *V. vulnificus* concentrations returned to pre-storm concentrations (−1.6%), while *V. parahaemolyticus* MPN g−<sup>1</sup> concentrations approximately quadrupled. Conversely, surface water and sediment concentrations decreased to a fraction of their original concentrations at day 8 post-storm (−92, −66% *V. vulnificus*, respectively; −100% for both sediment and surface water,*V. parahaemolyticus*). One possible explanation for these changes is a bacterial response to the flushing effect from the wind and rain at the study site, but more likely is storm-induced changes in

oyster filtration rates over the course of this study. In Givens et al. (2014), changes in *Vibrio* concentration were seen to be approximately replicated in surface waters and oyster tissues, suggesting that the opposing patterns of oyster and water *Vibrio* concentration detected in the days following Hurricane Irene were atypical.

Oysters have been shown to reduce or halt filtration during periods of high suspended solids, recommencing filtration at a normal or increased rate when water clarity returns to ambient conditions (Loosanoff and Tommers, 1948). If filtration stalled during the height of the storm and then resumed after sediment resuspension ceased, it may have explained the concomitant decrease in oyster *Vibrio* concentrations by 5–10 times (**Table 2**), while surface water *Vibrio* concentrations increased by 7–11 times on the fourth day post-Hurricane Irene (**Table 2**; **Figure 2**). However, filtration rates were not directly measured in this study and other factors, such as population turnover and physical transport, cannot be excluded as potentially important mechanisms for changes in *Vibrio* concentrations. Similar to Fries et al. (2008), who noted an increase in sediment concentrations of total *Vibrio* when Hurricane Ophelia impacted the Neuse River Estuary, NC, USA; there was also an increase in the sediment concentrations of both*Vibrio* species during the first four days post-storm (**Table 2**). However, this pattern then reversed with an overall decrease in sediment CFU g−1(−100%, *V. parahaemolyticus;* <sup>−</sup>66%, *V. vulnificus*). Whether this was due to a change in oyster filtration or a difference in how each *Vibrio* species was introduced into the water column as a function of resuspension, and associated particle adhesion, remains to be understood. In contrast to other studies (Fries et al.,2008; Hsieh et al.,2008;Wetz et al.,2008;Johnson et al., 2010), surface water CFU mL−<sup>1</sup> decreased following the storm (**Table 2**).

Notably, virulence-associated genes of *V. vulnificus and V. parahaemolyticus* were not detected in surface waters or sediment during the course of this study, possibly due to limitations of the direct extraction method (sediment, water) in relation to the MPN enrichment method (oyster samples). This is counter to other study findings, such as Johnson et al. (2010), which reported virulence-associated *V. parahaemolyticus* genes at similar frequencies in sediment, surface water and oysters. The *V. vulnificus vcgC* gene was found routinely in oyster tissues, but the percentage of *V. vulnificus* carrying *vcgC* was elevated at the beginning and end of the study (0.2%), and reduced one day after the storm and on day 4 (0.09%). Similarly, the percentage of *V. parahaemolyticus* carrying the *tdh* virulenceassociated gene was elevated before the storm and on day 4 (2%) and reduced one day after the storm (0.7%). Incidence and concentration of virulent *V. parahaemolyticus* was at its lowest point at day 8 (0%). These findings are in contrast to previous, laboratory-based studies, examining the relationship between *V. vulnificus'* virulence associated genes in oysters. These previous studies found no change in *V. vulnificus* virulence associated genes during the passage through the oyster (Groubert and Oliver, 1994; Staley et al., 2011). It is possible that the changes in virulence-associated genes percentages in this study are associated with population turnover within the oyster during the storm period.

Movement towards increased aquaculture production of oysters in the Chesapeake Bay, in combination with forecasted environmental responses to global climate change (e.g., warmer surface waters, increased frequency and/or intensity of storm events), may create a situation of higher *Vibrio* density in oysters, especially during the summer harvest season. An inventory of the last decade of tropical storms (2001–2011)<sup>3</sup> in the Chesapeake Bay elucidates that at least one tropical storm or depression is routinely seen in the region each year, and at least one hurricane within each decade, with an anticipated increase in tropical weather influenced by climate change conditions. Further research is needed to determine if patterns of adherence to oyster tissues is different between *V. parahaemolyticus* and *V. vulnificus*, as well as among virulent subsets of each species. As the storm event in this study consisted of both high winds and large amounts of precipitation, it would be useful to examine storm events with a range of wind speeds and precipitation to account for the individual response variables of resuspension and surface water flushing. Additionally, the role of nutrient introduction from terrestrial sources and the impact of plankton dynamics on *Vibrio* populations should be investigated in future studies to elucidate the impact of either variable on *Vibrio* concentration in the measured substrates. Such information would help managers of shellfish harvest decide if there should be a cessation or modification (e.g., post-harvest treatment) of harvest post-storm, what winds or rainfall would be significant for a given aquaculture site, and how long that suspension or modification of harvest should be recommended.

#### **ACKNOWLEDGMENTS**

Financial support for this research was graciously provided by National Oceanic and Atmospheric Administration award EA133C07CN0163.

We would like to thank Dr. Lawrence Harding and Mr. Steve Suttles (University of Maryland Center for Environmental Science, Horn Point Laboratory) for helpful conversations regarding the sampling site physical dynamics and how to appropriately calculate erosion rates for the site. We would also like to thank Dr. Jeffrey Cornwell and Mr. Michael Owens for information related to study site sediment properties and Dr. Jessica Jones (Food and Drug Administration, Gulf Coast Seafood Laboratory, Dauphin Island, AL, USA), Dr. Salina Parveen, Ms. Chanelle White (University of Maryland, Eastern Shore), and Mr. Matt Rhodes (NOAA Cooperative Oxford Laboratory) for assistance in determining appropriate microbiological, molecular, and quantitative techniques for *Vibrio* enumeration. We are appreciative of thoughtful edits provided by Ms. Carol McCollough (Maryland Department of Natural Resources). We are indebted to Dr. John Bowers (Food and Drug Administration, Center for Food Safety and Nutrition, College Park, MD, USA) for providing helpful assistance during statistical analyses of data. We thank Dr. Jessica Jones (Food and Drug Administration) and Dr. Carrie Givens (University of Georgia) for sharing data for comparison purposes. Lastly, we are extremely grateful to the aquaculture facility and its staff for all of their assistance and support.

#### **REFERENCES**

Avila, L. A., and Cangialosi, J. (2011). "*Tropical Cyclone Report: Hurricane Irene (AL092011), August 2011*." National Hurricane Center Summaries and Reports. Oxford, MD: NOAA.

<sup>3</sup>www.weather.gov/lwx/tropical


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

#### *Received: 15 November 2013; paper pending published: 16 December 2013; accepted: 17 April 2014; published online: 07 May 2014.*

*Citation: Shaw KS, Jacobs JM and Crump BC (2014) Impact of Hurricane Irene on Vibrio vulnificus and Vibrio parahaemolyticus concentrations in surface water, sediment and cultured oysters in the Chesapeake Bay, MD, USA. Front. Microbiol. 5:204. doi: 10.3389/fmicb.2014.00204*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Shaw, Jacobs and Crump. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## *Vibrio cholerae* interactions with *Mytilus galloprovincialis* hemocytes mediated by serum components

#### *Laura Canesi 1, Elisabetta Pezzati 1, Monica Stauder 1, Chiara Grande1, Margherita Bavestrello1, Adele Papetti <sup>2</sup> , Luigi Vezzulli <sup>1</sup> and Carla Pruzzo1\**

*<sup>1</sup> Dipartimento di Scienze della Terra, dell'Ambiente e della Vita, Università di Genova, Genova, Italy*

*<sup>2</sup> Dipartimento di Scienze del Farmaco, Università di Pavia, Pavia, Italy*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Spencer V. Nyholm, University of Connecticut, USA Gian Marco Luna, National Research Council – Institute of Marine Sciences, Italy*

#### *\*Correspondence:*

*Carla Pruzzo, Dipartimento di Scienze della Terra, dell'Ambiente e della Vita, Università di Genova, Corso Europa 26, 16132 Genova, Italy e-mail: carla.pruzzo@unige.it*

Edible bivalves (e.g., mussels, oysters) can accumulate large amount of bacteria in their tissues and act as passive carriers of pathogens to humans. Bacterial persistence inside bivalves depends, at least in part, on hemolymph anti-bacterial activity that is exerted by both serum soluble factors and phagocytic cells (i.e., the hemocytes). It was previously shown that *Mytilus galloprovincialis* hemolymph serum contains opsonins that mediate D-mannose-sensitive interactions between hemocytes and *Vibrio cholerae* O1 El Tor bacteria that carry the mannose-sensitive hemagglutinin (MSHA). These opsonins enhance phagocytosis and killing of vibrios by facilitating their binding to hemocytes. Since *V. cholerae* strains not carrying the MSHA ligand (O1 classical, non-O1/O139) are present in coastal water and can be entrapped by mussels, we studied whether in mussel serum, in addition to opsonins directed toward MSHA, other components can mediate opsonization of these bacteria. By comparing interactions of O1 classical and non-O1/O139 strains with hemocytes in artificial sea water and serum, it was found that *M. galloprovincialis* serum contains components that increase by at approximately twofold their adhesion to, association with, and killing by hemocytes. Experiments conducted with high and low molecular mass fractions obtained by serum ultrafiltration indicated that these compounds have molecular mass higher than 5000 Da. Serum exposure to high temperature (80◦C) abolished its opsonizing capability suggesting that the involved serum active components are of protein nature. Further studies are needed to define the chemical properties and specificity of both the involved bacterial ligands and hemolymph opsonins. This information will be central not only to better understand *V. cholerae* ecology, but also to improve current bivalve depuration practices and properly protect human health.

**Keywords:***Vibrio cholerae***, adherence, mussel, hemocytes, serum**

"fmicb-04-00371" — 2013/12/5 — 17:43 — page 1 — #1

#### **INTRODUCTION**

Microorganisms in seawater can be entrapped by filter feeding invertebrates that sieve suspended particles from the aquatic environment (Vazquez-Novelle et al., 2005; Beleneva et al., 2007). Accumulation of bacteria pathogenic to humans in the tissues of edible bivalves is of great concern to public health; in fact, consumption of raw or inadequately cooked bivalves has been implicated in numerous food poisoning outbreaks (Suzita et al., 2009; Scallan et al., 2011). Bivalve microbiological depuration (purification) in controlled waters is used worldwide to reduce the number of unwanted microorganisms to acceptable levels for human consumption (Teplitski et al., 2009). However, bacteria show different sensibility to depuration treatment; for instance, some *Vibrio* species have been reported to be resistant to the process and are able to persist and multiply within shellfish tissues (Tamplin and Capers, 1992; Murphree and Tamplin, 1995).

A relationship between microbial resistance to depuration and sensitivity to hemolymph bactericidal activity has been suggested (Harris-Young et al., 1995; Canesi et al., 2002, 2005). Shellfish hemolymph contains both hemocytes, which are responsible for cellular defense mechanisms (i.e., phagocytosis, production of reactive oxygen intermediates, and release of lysosomal enzymes), and soluble factors (e.g., opsonizing lectins and hydrolytic enzymes; Canesi et al., 2002; Pruzzo et al., 2005a). The capacities of different bacteria to survive hemolymph microbicidal activity depend on their sensitivities to combinations of these factors (Duperthuy et al., 2011; Balbi et al., 2013). Elucidation of underlying molecular mechanisms is crucial to improve current depuration practices and properly protect human health.

*Vibrio cholerae* is part of the endogenous bacterial component in estuarine areas (Kaper et al., 1995; Lipp et al., 2002). Only two serogroups, O1 (classical and El Tor biotypes) and O139 have been associated with cholera epidemic disease although nearly 200 serogroups of *V. cholerae* (named "non-O1/O139") have been described (Kaper et al., 1995; Nelson et al., 2009). The great majority of non-O1/O139 strains do not produce cholera toxin and are not associated with epidemic diarrhea. These strains are much more commonly isolated from the environment than are O1/O139 strains, and are occasionally isolated from cases of diarrhea usually due to consumption of raw or partially cooked shellfish (Iwamoto et al., 2010; Oliver et al., 2013).

Recent data on surface components involved in interactions of *V. cholerae* O1 with hemocytes of *Mytilus galloprovincialis,* an economically important, and appreciated seafood in the Mediterranean area, indicated that hemolymph serum contains opsonins specifically directed toward the"mannose-sensitive hemagglutinin (MSHA)" (Zampini et al., 2003), an adhesin expressed by El Tor and O139 strains (other *V. cholerae* strains either do not carry the *msha* gene or do not express it, i.e., strains of O1 serogroup, classical biotype; Chiavelli et al., 2001). It was suggested that Dmannose containing serum opsonins, capable to specifically react with MSHA, are involved in *V. cholerae* El Tor phagocytosis, and killing by mussel hemocytes (Zampini et al., 2003).

However, the above data did not rule out the possibility that in mussel serum, in addition to opsonins directed toward MSHA, other components are present that mediate interactions with hemocytes of vibrios different from O1 El Tor. To explore this possibility, we tested the capability of *M. galloprovincialis* serum to promote phagocytosis and killing by hemocytes of *V. cholerae* bacteria of O1 classical biotype and non-O1/O139 serogroups. The obtained results indicate that opsonizing molecules directed toward these bacteria are actually present in *M. galloprovincialis* hemolymph serum; a preliminary analysis of these components and the involved bacterial antigens is presented.

#### **MATERIALS AND METHODS**

#### **BACTERIA AND CULTURE CONDITIONS**

*Vibrio cholerae* CD81 (O1 classical ) was kindly provided by Dr. B. S. Srivastava (Microbiology Division, Central Drug Research Institute, Lucknow 226001, India); RC60, RC66, and RC69 (non-O1/O139) were kindly provided by Dr. A. Huq (Maryland Pathogen Research Institute, University of Maryland, College Park, MD 20742, USA). N16961 ATCC® 39315TM (O1 El Tor) was also used in some experiments. All cultures were grown in Luria-Bertani (LB) broth or agar under static conditions at 37◦C. To radiolabel bacteria, strains were grown overnight in LB broth containing 10 μCi of [methyl-3H]thymidine (25 Ci/mmol) ml−1, were then harvested by centrifugation (3,000 × *g* for 15 min at 4◦C), washed three times with phosphate-buffered saline (PBS; 0.1 M KH2PO4, 0.1 M Na2HPO4, 0.15 M NaCl, pH 7.2–7.4), and resuspended in PBS at an A650 of 1(2 <sup>×</sup> 108 to 4 <sup>×</sup> 108 bacteria ml−1). The number of counts per minute (cpm) per milliliter and the number of bacteria per milliliter were evaluated in triplicate samples to calculate the efficiency of cell labeling (number of bacteria per cpm) that varied in different bacterial preparations from 180 to 350. Artificial sea water [ASW; 35‰ (wt/vol) salinity, pH 7.9], filtered onto 0.22μm-pore-size Millipore filters (Bedford, MA, USA), was used throughout the experiments.

#### **PREPARATION OF MUSSELS HEMOCYTE MONOLAYERS**

Mussels (*M. galloprovincialis* Lam.) were obtained from the depuration plant Casa del Pescatore (Cattolica, Italy). Animals were transferred to the laboratory, cleaned of epibiota and kept in an aquarium at 16◦C in static tanks containing ASW (1 l/animal) for 1–3 days before use; sea water was changed daily. Hemolymph was extracted from the posterior adductor muscle of the mussels by using a sterile 1 ml syringe with a 18-gauge, 0.5-in. long needle. After the needle was removed, the hemolymph was filtered through sterile gauze and pooled. To prepare hemocyte monolayers, an approximately 0.3-ml portion of hemolymph (corresponding to about 2 <sup>×</sup> <sup>10</sup><sup>6</sup> to 3 <sup>×</sup> 106 cells) was seeded onto glass coverslips (20 by 22 mm) placed in plastic culture dishes. The coverslips were incubated at 18◦C for 30 min. Non-adherent hemocytes were removed by gently washing the preparations three times with 3 ml of ASW (Zampini et al., 2003).

To obtain hemolymph serum (i.e*.*, hemolymph free of cells), the whole hemolymph was centrifuged at 50 × *g* for 10 min; the supernatant was then passed through a filter (pore size, 0.22 μm).

#### **ADHESION TO AND ASSOCIATION WITH HEMOCYTE MONOLAYERS**

Evaluation of bacterial adherence to hemocytes was performed as follow: aliquots (1.5 ml) of either ASW or hemolymph serum containing radiolabeled bacteria at a final concentration of 2 <sup>×</sup> <sup>10</sup><sup>7</sup> to 3 <sup>×</sup> <sup>10</sup><sup>7</sup> bacteria ml−<sup>1</sup> were added to monolayers, and the dishes were incubated with gentle shaking at either 4◦C (to evaluate adhering bacteria) or 18◦C (to evaluate associated bacteria = adhering + internalized). Triplicate preparations were made for each sample. After 60 min incubation, the cover slips were rinsed with cold ASW, and transferred to PICO-FLUORTM15 scintillation fluid (Packard Instruments Company Inc., Meriden, CT, USA). For each sample, the number of bacteria per monolayer was calculated using the efficiency of cell labeling. Background counts due to bacterial attachment to coverslips were also evaluated (typically 50–250 cpm) and subtracted from the sample values. In experiments performed to define the chemical nature of the bacterial component(s) involved in interactions with hemocytes, bacteria were pretreated with pronase E and sodium meta-periodate. Pronase E was added to bacterial suspensions at a final concentration of 100 pgml−1. The suspensions were then incubated for 1 h at 37◦C in a shaking water bath and centrifuged. The pellets were resuspended in PBS to the original volume. Sodium meta-periodate pretreatment of bacteria was performed in PBS containing sodium meta-periodate at a final concentration of 1 mM. The suspensions were incubated at room temperature for 10 min, washed twice and resuspended in PBS.

#### **BACTERIAL KILLING BY HEMOCYTES**

"fmicb-04-00371" — 2013/12/5 — 17:43 — page 2 — #2

To evaluate bacterial sensitivity to killing by hemocytes,*V. cholerae* suspensions (about 10<sup>7</sup> bacteria ml−1) were added to hemocyte monolayers at 18◦C in the presence of hemolymph serum as described above. Triplicate preparations were made for each sampling time. Immediately after the inoculum (*T* = 0) and after 60 min of incubation at 18◦C supernatants were collected from monolayers, and hemocytes were lysed by adding 5 ml of cold distilled water and by 10 min of agitation. The collected monolayer supernatants and hemocyte lysates were pooled and 10-fold serially diluted in PBS; aliquots (100 μl) of the diluted samples were plated onto LB agar, and after overnight incubation at 37◦C the number of colony-forming units (CFU) per milliliter in the hemocyte monolayer (representing culturable bacteria survived to hemocyte bactericidal activity) was determined. Percentages of killing at 60 min were then determined relative to values obtained at *T* = 0. To evaluate the presence of endogenous bacteria in hemocytes, controls were performed with hemocyte monolayers without bacteria. The number of CFU in controls never exceeded 0.1% of those enumerated in experimental samples. To detect and correct for bacterial growth in hemolymph serum, separate samples were seeded with bacteria and 1.5 ml of sterile hemolymph serum. No appreciable bacterial growth was observed at the same time intervals used in killing experiments.

#### **SERUM FRACTIONATION**

Serum was fractionated into low and high molecular mass fractions (LMM and HMM) by ultrafiltration using the Vivaflow 200 complete system (VivascienceAG, Feodor-Lynen-Strasse 21, 30625 Hannover, Germany) comprising a pump (240 V), tubing, 500 ml sample/diafiltration reservoir, and a membrane 5,000 MWCO PES for ultradiafiltration (Vivascience). The LMM 5000 cutoff was chosen since electrophoretic separation of mussel soluble serum proteins (in both 1D and also 2D gels used for proteomic studies in different conditions) generally yields protein bands with MM generally ≥10,000. A diafiltrate, i.e., a LMM containing all the compounds with molecular masses less than 5000 Da, and a retentate, i.e., a HMM containing all the compounds with molecular masses greater than 5000 Da were obtained and, after restoring the initial volume, were tested.

#### **STATISTICS**

Experiments were repeated at least three times. Data shown in the Figures are the mean values ± standard deviation obtained in one representative experiment performed in triplicate. Data were analyzed for significance by the Mann–Whitney *U* test. Differences were considered significant at *P* < 0.05.

#### **RESULTS**

#### *IN VITRO* **INTERACTIONS OF** *Vibrio* **STRAINS WITH HEMOCYTES AND SENSITIVITY TO KILLING**

We studied interactions with hemocytes of *V. cholerae* strains that either do not carry the *msha* gene (RC60, RC66, and RC69 [non-O1/O139 serogroups]) or do not express it (CD81 [O1 serogroup, classical biotype]). Bacteria were added to hemocyte monolayers in both ASW and hemolymph serum at 18◦C. At this temperature, the number of associated (adhering + internalized) bacteria was evaluated. **Figure 1** shows that in hemolymph serum the association efficiency of the tested strains was 2.2–3.5-fold higher than in ASW (*P* ≤ 0.05).

To clarify to what extent the observed differences in association were due to different adhesion efficiencies, the same assays were performed at 4◦C, a condition that almost completely inhibits the internalization process (Zampini et al., 2003; **Figure 1**). The presence of hemolymph serum increased the number of vibrios adhering to hemocytes by 2.1–2.6-fold in comparison to ASW (*P* ≤ 0.05; **Figure 1**). Moreover, as expected, at 4◦C the number of bacteria interacting with hemocytes was lower than that at 18◦C. Interestingly, the increase in both association with and adherence to mussel hemocytes of the tested strains was about twofold lower than that observed with the MSHA-positive *V. cholera*e N16961 strain, used for comparison.

*Vibrio cholerae* strains were then tested for their ability to resist to killing by hemocytes, both in the presence and in the absence

of serum (**Figure 2**). It was found that, after 60 min incubation, the percentage of killed bacteria compared to that at *T* = 0 ranged from 9.3 to 14.7% in the experiments performed in serum and from 3.5 to 6.3% in experiments performed in ASW. Differences between the two experimental conditions were statistically significant (*P* ≤ 0.05). As for association and adhesion, the serummediated increase in killing efficiency of the tested strains by hemocytes was about twofold lower than that observed with *V. cholera*e N16961 strain carrying the MSHA.

As a whole, these results indicate that mussel serum, in addition to opsonins directed toward MSHA, contains factors that mediate binding between *Vibrio* cell wall component(s) and hemocyte surface receptor(s), and promote internalization and killing of these bacteria.

#### **PRELIMINARY ANALYSIS OF BACTERIAL SURFACE LIGANDS AND SERUM COMPONENTS INVOLVED IN INTERACTIONS OF** *Vibrio* **STRAINS WITH HEMOCYTES**

"fmicb-04-00371" — 2013/12/5 — 17:43 — page 3 — #3

In a first attempt to define the chemical nature of bacterial surface components involved in*V. cholerae* interactions with hemolymph, adhesion to hemocytes of the tested strains in the presence of

serum was evaluated after bacteria had been treated with either pronase E or sodium meta-periodate, which oxidizes polysaccharides. As shown in **Figure 3**, both treatments caused a decrease in serum-mediated attachment to hemocytes. In comparison to untreated controls, pronase E and sodium meta-periodate treatments reduced the interactions by 2.3–3.5 and 2.7–7.0-fold, respectively. Meta-periodate oxidation and pronase E digestion did not affect bacterial viability, as demonstrated by viable counts (data not shown).

To preliminarily assess the main chemical properties of serum factors capable to promote *Vibrio* strain interactions with hemocytes, hemolymph serum was first exposed (15 min) to different temperatures, ranging from 45 to 80◦C, then used in experiments to evaluate efficiency of bacterial adhesion to hemocytes. Serum progressively lost its capability to promote *V. cholerae* attachment to hemocytes with increasing temperature from 45 to 80◦C. In particular, the number of bacteria adhering to hemocytes in the presence of serum incubated at 80◦C was 2.2–3.9-fold lower than that observed in the presence of untreated serum (**Figure 3**), and similar to those obtained in experiments performed in ASW (**Figure 1**).

Serum was then fractionated into LMM and HMM fractions by ultrafiltration (see Methods). LMM fraction contained all the compounds with molecular masses lower than 5000 Da, and HMM contained all the compounds with molecular masses greater than 5000 Da. Capability of both fractions to promote adhesion to and association with hemocytes of the tested bacteria was analyzed. As shown in **Figure 4**, whereas LMM fraction did not cause any statistically significant increase in bacterial interactions with hemocytes, the HMM fraction caused a 1.6–2.4-fold increase of both adhesion to and association with monolayers of the tested strains. When the experiments were performed using a mixture of both LMM and HMM fractions, results similar to those obtained with HMM alone were obtained (not shown).

#### **DISCUSSION**

We previously showed that *M. galloprovincialis* hemolymph serum contains soluble factors that are involved in D-mannose-sensitive interactions between hemocytes and*V. cholerae* O1 El Tor carrying

"fmicb-04-00371" — 2013/12/5 — 17:43 — page 4 — #4

the MSHA adhesin (Zampini et al., 2003). These opsonins enhance phagocytosis and killing of these bacteria by facilitating their binding to hemocytes.

However, other *V. cholerae* strains, including human epidemic clones (O1 classical), not carrying the MSHA ligand are present in coastal waters (Pruzzo et al., 2005b; Oliver et al., 2013) and can be entrapped by mussels. These vibrios, which can persist and replicate in bivalves, were selected for the current study to evaluate if additional serum factors, different from those directed toward MSHA, can affect*V. cholerae* adhesion to hemocytes and sensitivity to killing, modulating its densities inside mussels.

By comparing interactions of O1 classical and non-O1/O139 strains with hemocytes in ASW and serum, it was found that *M. galloprovincialis* serum actually contains components that promote adhesion to, association with and killing by hemocytes of bacteria that do not carry the MSHA. Serum components different from those acting toward MSHA are likely to be involved as suggested by the fact that serum-mediated increase in bacterial interactions with hemocytes is higher toward *V. cholerae* O1 El Tor (carrying MSHA) than toward the other tested vibrios (without MSHA). Interestingly, when comparing efficiencies of interactions with hemocytes and killing sensitivity of the different tested strains, a variability among MSHA-negative isolates was observed. This might be due, at least in part, to differences in bacterial surface hydrophobicity and/or surface charge. Moreover, the most adhesive strains might carry surface molecules mediating non-opsonic interactions with hemocyte membrane.

The preliminary analysis of the chemical nature of the involved bacterial ligand(s) indicated that both cell wall protein(s) and carbohydrate(s) are responsible for the interactions of non-El Tor *V. cholerae* strains with serum. In fact, using vibrios pre-treated with proteinase E or sodium meta-periodate that disrupts cell surface polysaccharide, the observed serum-mediated increase of adhesion and association showed a significant reduction in comparison to untreated controls. On the other hand, the fact that serum exposure to high temperature (80◦C) abolished its capability to promote adhesion to and association with phagocytes suggests that heat-sensitive serum components are involved in the opsonizing activity. Experiments conducted with HMM and LMM fractions obtained by serum ultrafiltration indicate that these compounds have molecular mass higher than 5000 Da.

Further studies are needed to define the chemical nature and specificity of both the involved bacterial adhesins and hemolymph opsonins as well as physical and chemical conditions promoting or inhibiting such interactions inside bivalves *in vivo*. This information will be crucial to better understand the strategies used by mussels to control bacterial diffusion in their tissues and by *V. cholerae* to persist and spread in the aquatic environment.

Consumption of raw or partially cooked bivalves has been implicated in numerous food poisoning outbreaks. Thus, their microbial flora is of great concern to public health (Suzita et al., 2009; Scallan et al., 2011). The ability of bivalves to eliminate pathogens from their tissues is due, at least in part, to the ability of hemocytes to bind, phagocytize, and kill these bacteria (Canesi et al., 2002; Pruzzo et al., 2005a). Deciphering molecular mechanisms affecting pathogenic bacteria – hemocytes interactions represents the basis to improve depuration treatment and to prevent transmission of pathogens to humans through consumption of raw or partially cooked bivalves. As an example, to reduce the load of unwanted bacteria inside bivalves, depuration might be conducted in conditions that favor phagocytosis of pathogens and their clearance from hemolymph.

Unraveling the cellular and molecular mechanisms associated with hemolymph anti-bacterial activity is also central to understand the pathogenesis of bivalve diseases in cultured and wild populations of species susceptible to *Vibrio* spp. infection, and to set up new strategies to control summer mortalities affecting the bivalve production (e.g., *Crassostrea gigas* oysters) in aquaculture worldwide (Gay et al., 2004; Garnier et al., 2007).

#### **ACKNOWLEDGMENTS**

This work was supported by grants from Italian Ministry of Education and Research (MIUR, PRIN 2008) and Genova University (Italy).

#### **AUTHOR CONTRIBUTIONS**

Laura Canesi and Luigi Vezzulli organized and performed most experiments. Chiara Grande and Margherita Bavestrello took care of aspects related to mussel maintenance in aquarium and hemocyte monolayer preparations. Elisabetta Pezzati cultured and controlled bacterial strains. Monica Stauder supplied technical assistance. Adele Papetti carried out serum fractionation. Carla Pruzzo designed the experiments and supervised.

#### **REFERENCES**


"fmicb-04-00371" — 2013/12/5 — 17:43 — page 5 — #5

**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 09 October 2013; paper pending published: 29 October 2013; accepted: 20 November 2013; published online: 09 December 2013.*

*Citation: Canesi L, Pezzati E, Stauder M, Grande C, Bavestrello M, Papetti A, Vezzulli L and Pruzzo C (2013) Vibrio cholerae interactions with Mytilus galloprovincialis hemocytes mediated by serum components. Front. Microbiol. 4:371. doi: 10.3389/fmicb.2013.00371*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Canesi, Pezzati, Stauder, Grande, Bavestrello, Papetti, Vezzulli and Pruzzo. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00371" — 2013/12/5 — 17:43 — page 6 — #6

**METHODS ARTICLE** published: 29 October 2013 doi: 10.3389/fmicb.2013.00320

## Adaptation of a simple dipstick test for detection of *Vibrio cholerae* O1 and O139 in environmental water

### *Subhra Chakraborty1, Munirul Alam2 , Heather M. Scobie1 † and David A. Sack1\**

*<sup>1</sup> Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA*

*<sup>2</sup> International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Shawn R. Campagna, University of Tennessee, Knoxville, USA Marie-Laure Quilici, Institut Pasteur, France*

#### *\*Correspondence:*

*David A. Sack, Department of International Health, Johns Hopkins Bloomberg School of Public Health, 615 North Wolfe Street, E 5537, Baltimore, MD 21205, USA e-mail: dsack@jhsph.edu*

#### *†Present address:*

*Heather M. Scobie, Centers for Disease Control and Prevention, 1600 Clifton Road, MS A-04, Atlanta 30333, Georgia*

#### **INTRODUCTION**

Cholera is transmitted through water contaminated with *Vibrio cholerae*. Methods for identifying culturable*V. cholerae* from water and sewage include culture and PCR techniques (Barrett et al., 1980; Fields et al., 1992; Alam et al., 2010), but efforts for detecting *V. cholerae* in environmental samples have previously been limited by laboratory testing capacity in areas where cholera most often occurs and the difficulty of achieving reliable results in these areas.

Recently a dipstick rapid test, based on detection of LPS antigen was introduced for detecting*V. cholerae* directly from fecal samples (CrystalVC kit, Span Diagnostics Ltd., Udhna, Surat, India).When watery fecal samples were tested directly, using culture method as gold standard the test was found to be sensitive (>90%) and moderately specific (∼70%; Ingram et al., 1996; Rabbani et al., 2001; Bhuiyan et al., 2003; Nato et al., 2003; Wang et al., 2006; Harris et al., 2009; Mukherjee et al., 2010; Rosewell et al., 2012; Sinha et al., 2012; Boncy et al., 2013). The test can also be carried out with rectal swab specimens if these samples are first incubated in alkaline peptone water (APW) for 4 to 6 h, a vibrio-selective enrichment step, prior to testing with the dipstick (Bhuiyan et al., 2003). We felt that this dipstick test might be adapted for use in detecting *V. cholerae* O1 and O139 from environmental water samples in areas at risk for cholera if the specimen was incubated in APW first in order to amplify the concentration of vibrios in the specimen.

#### **MATERIALS AND METHODS**

First, we carried out dipstick experiments with water samples from the Chesapeake Bay that were spiked with 10-fold serial

The presence of *Vibrio cholerae* in the environment is key to understanding the epidemiology of cholera. The gold standard for laboratory confirmation of *V. cholerae* from water is a culture method, but this requires laboratory infrastructure. A rapid diagnostic test that is simple, inexpensive, and can be deployed widely would be useful for confirming *V. cholerae* in samples of environmental water. Here, we evaluated a dipstick test to detect *V. cholerae* O1 and O139 from environmental water samples in spiked samples and under field conditions. When environmental water samples were incubated in alkaline peptone water for 24 h at room temperature, samples spiked with <10 CFU could be detected using the dipstick test. When compared to culture, the test was 89% sensitive and 100% specific with environmental samples.

**Keywords: cholera, dipstick, Bangladesh, environmental water,***Vibrio cholerae*

**"fmicb-04-00320" — 2013/10/27 — 17:57 — page 1 — #1**

dilutions of *V. cholerae* serotype O1 or O139. We placed 1 mL of the spiked water samples into 9 mL of APW, followed by incubation at temperatures of 22, 30, and 37◦C (**Table 1**). The concentration of bacteria added to spike the water was determined by plating serial dilutions of the initial inoculum. We tested the APW after 2, 4, 6 and 24 h with the dipstick using methods as described by the manufacturer for fecal samples.

Next, we applied the dipstick testing method to environmental water samples from Bangladesh. The samples included specimens from both urban and rural sites that were being collected as part of cholera epidemiology studies at the International Centre for Diarrheal Disease Research, Bangladesh (icddr,b) during March 2011–June 2012. Use of these samples allowed for a direct comparison of the dipstick testing method to bacterial culture methods used at icddr,b.

A total of 550 environmental water samples (200 mL each) were filtered through individual 0.22 micron polycarbonate membrane filters. The filters were then placed in 10 mL of phosphate buffered saline (PBS). After resuspension, 2 mL of the PBS solution was added to 18 mL of APW, followed by incubation for 24 h at room temperature (∼22◦C). The APW broth was then tested using the dipstick. The same APW was streaked onto taurocholate tellurite gelatin agar (TTGA) and thiosulfate-citrate-bile salts-sucrose (TCBS) plates, and colonies suspected as being *V. cholerae* from either plate were identified as*V. cholerae* O1 or O139 using standard methods. The results from the dipstick and cultures were recorded without knowledge of results from the other test.


**Table 1 | Dipstick detection of** *Vibrio cholerae***-containing water samples incubated in alkaline peptone water (APW).**

*Crystal VC test was read as positive (*+*) if V. cholerae O1 or O139 and control bands were visible, or negative (*−*) if only the control band was visible, according to manufacturer's instructions.*

"fmicb-04-00320" — 2013/10/27 — 17:57 — page 2 — #2

#### **RESULTS**

All spiked specimens were negative at 2 and 4 h, and only the APW spiked with greater than ∼200 CFU were positive after 6 h of incubation at 37◦C. However, after 24 h, all specimens were positive, including those spiked with <10 CFU and incubated at 30◦C or at room temperature (22◦C). The APW inoculated with unspiked water was consistently negative at all time points and temperatures (**Table 1**).

A positive dipstick reading occurred when the concentration of *V. cholerae* exceeded <sup>∼</sup>10<sup>7</sup> per mL; thus, the small number of bacteria in the water that were added to the APW had to grow to this density to be detectable.

Among the 550 environmental samples tested in Bangladesh, 55 were positive for*V. cholerae* serotype O1 by the bacterial culture method, and 48 of these were also positive with the dipstick test (89% sensitivity). Of the 495 samples that were negative by culture, none were positive with the dipstick test (100% specificity). None of the samples were positive for *V. cholerae* O139 by dipstick or culture.

#### **DISCUSSION**

This dipstick method following incubation in APW provides a simple procedure for detecting *V. cholerae* O1 and O139 from environmental water. Although the bacterial culture method did detect more positive specimens than the dipstick test, the dipstick test was still very sensitive. It should be noted that the culture methods used at the icddr,b included streaking onto both TCBS and TTGA agars. Incubation on these two media will generally be more sensitive than methods which use only one type of agar, as is carried out in most laboratories (Alam et al., 2010). Thus, the sensitivity of the dipstick method may be as sensitive as using culture with TCBS alone.

While our preliminary study used samples spiked with O139 as well as O1 serotypes, there were no O139 strains detected with the environmental samples from Bangladesh. It seems plausible that this method can be used for the detection of serotype O139 from environmental water, but this has not yet been validated in the field. Recently, there have been no clinical cases of cholera nor environmental isolations related to *V. cholerae* O139 in Bangladesh. The finding that there were no positive dipstick tests for O139 was reassuring.

The high specificity of the dipstick test (100%) in this evaluation differs from the results observed in clinical studies of fecal samples in which many dip stick positive tests could not be confirmed by culture suggesting a high rate of false positives (Rabbani et al., 2001; Harris et al., 2009; Rosewell et al., 2012). We believe that this difference is because we carried out the test on water samples after incubation in APW rather than directly testing as done in fecal specimens. The reason that some stool samples yield false positive results is not known, but by using an APW selective enrichment step, cross reacting substances are diluted and vibrio antigens are amplified. Our results confirm that APW enrichment of samples helps to improve the specificity of the test, as previously demonstrated (Wang et al., 2006). For water samples, incubation in APW allows growth of vibrio species to concentrations detectable with the dipstick.

This study used the Crystal VC dipstick to detect *V cholerae* as this has been the test most widely used in developing country settings. Other rapid tests for *V. cholerae* are also available (Qadri et al., 1995) and might also be useful since this method primarily relies on the ability of vibrios to grow and for the LPS signal to be amplified in the APW broth prior to testing.

Some limitations to this dipstick method should be mentioned. *V. cholerae* can exist as viable but non-culturable (VBNC) bacteria (Alam et al., 2007). The dipstick test relies on the ability of the bacteria to grow in the APW medium; thus, it is not able to detect VBNC vibrios. Secondly, the Crystal VC dipstick test detects the LPS of *V. cholerae* but does not differentiate toxigenic from non-toxigenic strains. Non-toxigenic strains of serogroups O1 and O139 may also occur in environmental water but do not carry the same epidemic implications (Faruque et al., 2004). Thus, it would seem that some positive samples should be validated with follow up cultures to determine if they are toxigenic and to more fully characterize the strains. However, it should be noted that standard culture methods also do not determine if the vibrio strains isolated produce toxin. Depending on the epidemiological circumstances, the finding of *V. cholerae* O1 and O139 in water samples should stimulate a public health response and it is imperative that the finding of a positive dip stick test is rapidly confirmed.

Another limitation of this method, as it was carried out in Bangladesh, included filtration through a 0.22 micron filter. For some cholera endemic areas, this method of filtration may not be

possible. Thus, we are evaluating a less complex filtration method through cotton gauze, similar to the method used by Spira (Spira and Ahmed, 1981). Though filtration through gauze is likely to be somewhat less sensitive, it may still be useful in monitoring water samples. In fact the sensitivity of any assay also depends on the volume of water assayed, and no single test will be 100% sensitive.

We conclude that this dipstick method for testing environmental water samples is sensitive and specific for *V. cholerae* O1 and should be useful for water monitoring in cholera-endemic settings and in areas at risk for cholera. The method is technically simple and does not require use of an incubator. It is not "rapid" since an overnight incubation step is necessary but is "rapid" compared to traditional culture method.

#### **ACKNOWLEDGMENTS**

Support for the research was provided by National Institutes of Health grant no. R01AI039129-13 and by a grant from the Bill and Melinda Gates Foundation to Johns Hopkins University for the DOVE project (Delivering Oral Vaccine Effectively). In addition, the International Centre for Diarrhoeal Disease Research, Bangladesh acknowledges its donors for providing unrestricted support including the, governments of Bangladesh, Canada, the United Kingdom, Sweden, and Australia. Span Diagnostics, Ltd, Udhna, Surat, India donated the test kits.

#### **REFERENCES**


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 19 August 2013; paper pending published: 20 September 2013; accepted: 10 October 2013; published online: 29 October 2013.*

*Citation: Chakraborty S, Alam M, Scobie HM and Sack DA (2013) Adaptation of a simple dipstick test for detection of Vibrio cholerae O1 and O139 in environmental water. Front. Microbiol. 4:320. doi: 10.3389/fmicb.2013.00320*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Chakraborty, Alam, Scobie and Sack. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

"fmicb-04-00320" — 2013/10/27 — 17:57 — page 3 — #3

## Erratum: Adaptation of a simple dipstick test for detection of *Vibrio cholerae* O1 and O139 in environmental water

#### *Subhra Chakraborty1, Munirul Alam2, Heather M. Scobie1 and David A. Sack1 \**

*<sup>1</sup> Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA*

*<sup>2</sup> Centre for Communicable Diseases, International Centre for Diarrhoeal Disease Research, Dhaka, Bangladesh*

*\*Correspondence: dsack@jhsph.edu*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

**Keywords: cholera, dipstick, Bangladesh, environmental water,** *Vibrio cholerae*

#### **An erratum on:**

#### **Adaptation of a simple dipstick test for detection of** *Vibrio cholerae* **O1 and O139 in environmental water**

*by Chakraborty, S., Alam, M., Scobie, H. M., and Sack, D. A. (2013). Front. Microbiol. 4:320. doi: 10.3389/fmicb.2013.00320*

Two references cited in this article were incorrect. The reference to Ingram et al. (1996) should instead be Page et al. (2012) and the reference for Rabbani et al. (2001) should instead be Ley et al. (2012).

#### **REFERENCES**


*Received: 06 December 2013; accepted: 06 December 2013; published online: 24 December 2013.*

*Citation: Chakraborty S, Alam M, Scobie HM and Sack DA (2013) Erratum: Adaptation of a simple dipstick test for detection of Vibrio cholerae O1 and O139 in environmental water. Front. Microbiol. 4:405. doi: 10.3389/ fmicb.2013.00405*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2013 Chakraborty, Alam, Scobie and Sack. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## *Vibrio* diversity and dynamics in the Monterey Bay upwelling region

#### *Sarah Mansergh and Jonathan P. Zehr\**

*Ocean Sciences Department, University of California at Santa Cruz, Santa Cruz, CA, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Luigi Vezzulli, University of Genoa, Italy Rodrigo Costa, Centre of Marine Sciences, Portugal*

#### *\*Correspondence:*

*Jonathan P. Zehr, Ocean Sciences Department, University of California Santa Cruz, 1156 High Street, Santa Cruz, CA 95064, USA e-mail: zehrj@ucsc.edu*

The *Vibrionaceae* (*Vibrio*) are a ubiquitous group of metabolically flexible marine bacteria that play important roles in biogeochemical cycling in the ocean. Despite this versatility, little is known about *Vibrio* diversity and abundances in upwelling regions. The seasonal dynamics of *Vibrio* populations was examined by analysis of 16S rRNA genes in Monterey Bay (MB), California from April 2006–April 2008 at two long term monitoring stations, C1 and M2. *Vibrio* phylotypes within MB were diverse, with subpopulations clustering with several different cultured representatives including *Allivibrio* spp., *Vibrio penaecida*, and *Vibrio splendidus* as well as with many unidentified marine environmental bacterial 16S rRNA gene sequences. Total *Vibrio* population abundances, as well as abundances of a *Vibrio* sp. subpopulation (MBAY Vib7) and an *Allivibrio* sp. subpopulation (MBAY Vib4) were examined in the context of environmental parameters from mooring station and CTD cast data. Total *Vibrio* populations showed some seasonal variability but greater variability was observed within the two subpopulations. MBAY Vib4 was negatively associated with MB upwelling indices and positively correlated with oceanic season conditions, when upwelling winds relax and warmer surface waters are present in MB. MBAY Vib7 was also negatively associated with upwelling indices and represented a deeper *Vibrio* sp. population. Correlation patterns suggest that larger oceanographic conditions affect the dynamics of the populations in MB, rather than specific environmental factors. This study is the first to target and describe the diversity and dynamics of these natural populations in MB and demonstrates that these populations shift seasonally within the region.

**Keywords:** *Vibrio***, upwelling, Monterey Bay, seasonal variability, 16S rRNA gene diversity**

#### **INTRODUCTION**

The *Vibrionaceae* (*Vibrio*) are a group of physiologically-flexible marine bacteria that are ubiquitous in ocean waters and have been identified in most marine ecosystems (Wietz et al., 2010). *Vibrio* have the distinctive ability to break down and utilize many carbon, nitrogen, and phosphorus substrates (McDougald and Kjelleberg, 2006; Thompson and Polz, 2006; Dryselius et al., 2008; Lai et al., 2009; Salter et al., 2009) and their production of the external enzymes chitinase and laminarinase provide access to abundant nutrients that are unavailable to other organisms (Svitil et al., 1997; Riemann et al., 2000; Ansede et al., 2001; Alderkamp et al., 2007; Murray et al., 2007). In addition to their diverse metabolic capabilities, *Vibrio* species have developed adaptive responses to starvation and environmental stress which include conversion to an ultramicrobacterial morphology (<0.4μm diameter) (Denner et al., 2002) and retention of a high concentration of rRNA as a "stimulation ready" response that may contribute to this group's rapid growth potential after nutrient influxes (Eilers et al., 2000). Despite these capabilities, there have been few studies in oceanic upwelling regions, where metabolic flexibility, rapid nutrient response, and highly developed stress protection mechanisms should prove to be ecologically advantageous during changing environmental regimes.

Monterey Bay (MB) is an open embayment along the California coast that is distinguished by a near-shore deep-water canyon where periodic upwelling events sustain diverse sea life. Circulation within MB is variable and influenced by recently upwelled waters as well as offshore waters from the California and Davidson Currents that enter MB during relaxation events (Rosenfeld et al., 1994). Three hydrographic seasons have been defined within MB—an upwelling or "cold water phase" that usually occurs from mid-February through August, an oceanic period or "warm water phase" from mid-August through mid-October and a Davidson Current or "low thermal gradient" phase between mid-November and mid-February (Skogsberg and Phelps, 1946; Breaker and Broenkow, 1994). The periodic influx of nutrient-enriched waters into MB via upwelling results in a highly productive ecosystem presenting a unique environment for the study of *Vibrio* population dynamics.

*Vibrio* species are common isolates from MB waters but most studies have concentrated on pathogenic representatives (Kenyon et al., 1984; Kaysner et al., 1987; Miller et al., 2006). One phylogenetic screening study found that *Vibrio*-specific sequences made up 3.1% of the genetic information within a FOSMID library developed from a 100 m deep sample (Suzuki et al., 2004). *Vibrio* populations are estimated to average only about 1% of marine bacterial populations worldwide (Thompson and Polz, 2006) but it has also been suggested that differences in vertical distributions may be more significant than differences in geographic distribution for *Vibrio* populations, and thus, may make up a more significant portion of subsurface bacterial populations (Simidu and Tsukamoto, 1985).

This study was designed to determine the seasonal variability of *Vibrio* populations in MB by examining total *Vibrio* population dynamics as well as two subpopulations, a *Vibrio* sp. subpopulation (MBAY Vib7) and an *Allivibrio* sp. subpopulation (MBAY Vib4), and to analyze how population diversity changes with upwelling, season, and environmental factors. MB has a system of long term monitoring stations (Pennington and Chavez, 2000) and two of the stations within this network, C1 and M2 (**Figure 1**), were chosen as sample sites for this study. C1 is a coastal site that is influenced by aged upwelled waters and summer relaxation events. M2 is an outer Bay site that is influenced by upwelled waters that mix with California Current and Davidson Current waters. The examination of *Vibrio* population dynamics at these two different locations can provide insight into the effect of upwelling characteristics on *Vibrio* abundances.

#### **MATERIALS AND METHODS**

#### **SAMPLE COLLECTION**

Seawater samples were collected from MB at stations C1 (36.797 N; 121.847 W; **Figure 1**) and M2 (36.697 N; 122.378 W; **Figure 1**) during cruises of the MB Time Series project undertaken by the Biological Oceanography Group (BOG) at the Monterey Bay Aquarium Research Institute (MBARI). Samples were collected from April 2006 through April 2008 on a periodic basis spanning 19 sampling dates. Samples were collected using a SeaBird 911 CTD rosette equipped with physical sensors described by Pennington and Chavez (2000). A total of 82 samples were collected at station C1 at 5 m (15 samples), 10 m (17 samples), 20 m (18 samples), 30 m (17 samples), and 200 m (15 samples). At station M2 a total of 86 samples were collected at 5 m (12 samples), 10 m (14 samples), 20 m (14 samples), 40 m (16 samples), 100 m (16 samples), and 200 m (16 samples). For each sample 1–2 liters of seawater were filtered using gentle

peristaltic pumping through sequential in-line 25 mm diameter 10μm pore-size PE filters (GE Osmonics, Minnetonka, MN, USA) and 25 mm diameter 0.2μm pore-size Supor 200 membrane filters (Pall Corporation, Port Washington, NY, USA). The filters were transferred to 1.5 mL polypropylene microcentrifuge tubes containing 0.2 g of 0.1 mm and 0.5 mm diameter autoclaved glass beads (BioSpec Products, Bartlesville, OK, USA). Samples were flash frozen in liquid nitrogen before transfer to a −80◦C freezer onshore for storage until nucleic acids were extracted.

#### **ENVIRONMENTAL DATA**

Upwelling indices (UI) of both monthly and daily averaged upwelling conditions were obtained for 36◦N 122◦W from the Pacific Fisheries Environmental Laboratory (PFEL, Pacific Grove) for determination of upwelling patterns (http://www.pfeg.noaa. gov/products/PFEL/modeled/indices/PFELindices.html). Units are given as m<sup>3</sup> s <sup>−</sup><sup>1</sup> 100 m of coastline-<sup>1</sup> as the average amount of water upwelled through the bottom of the Ekman layer each second along each 100 m of coastline on a scale of about 200 miles. Surface data for nitrate, chlorophyll *a*, and temperature was obtained from the M2 mooring from MBARI LOBOviz (http://www.mbari.org/lobo/loboviz.htm). The BOG at MBARI provided environmental data from CTD measurements as well as surface (<10 m) phytoplankton concentrations, which were analyzed as described by Chavez et al. (1990). Specific phytoplankton groups (*Synechococcus,* total diatom, total dinoflagellate, and total phytoplankton) were chosen for analysis to assess the influence of different phytoplankton regimes on the variability of *Vibrio* groups in the water column.

#### **DNA EXTRACTION**

DNA was extracted from the 0.2μm pore-size filters using the Qiagen DNeasy Plant Kit (Hilden, Germany) as described by Moisander et al. (2008) with modifications to improve DNA recovery. Cells were lysed by a triplicate run of freeze-thaw steps in liquid nitrogen followed by a 65◦C water bath. Cells were further disrupted by a 2-min agitation of bead beating (Mini bead beater 96, BioSpec Products, Bartlesville, OK, USA) and DNA yield was increased with the addition of 0.45 μL of Proteinase K and incubation at 55◦C for 1 h. AE buffer was used as the elution medium. Two duplicate elutions of 25μL each were combined for a final elution volume of 50μL. Extracts were stored at −20◦C until use.

#### **AMPLIFICATION, CLONING AND SEQUENCING OF** *Vibrio***-SPECIFIC 16S rRNA GENE**

Samples were selected from 2007 to 2008 from 5 to 200 m depths at both station C1 and M2 to construct clone libraries of amplified 16S rRNA gene sequences from the MB. A total of 65 samples were processed for inclusion in the libraries. The PCR amplification utilized a universal forward primer (27F, **Table 1**) and a *Vibrio* specific reverse primer (680R, **Table 1**) and followed the two-phase PCR amplification technique outlined by Thompson et al. (2004b). Samples were amplified using Invitrogen *Taq* polymerase (Carlsbad, California), on a BioRad thermal cycler (Hercules, California). The PCR products were gel purified and cloned into pGEM-T vectors (Promega, Madison WI) using manufacturer's guidelines. Ten to fifteen clones were

#### **Table 1 | Primers and probes for PCR and qPCR (written 5- -***>***3- ).**



*The 16S rRNA gene and Total Vibrio primers are from Thompson et al. (2004a,b). The MBAY Vib4 and MBAY Vib7 primer and FAM/TAMRA labeled probe sets were developed based on sequences from the Monterey Bay 16S rRNA gene clone libraries.*

chosen from each sample, for a total of 911 clones, and prepared for sequencing with the Montage Plasmid Miniprep kit (Millipore, Billerica, MA) according to manufacturer's protocols. Cloned inserts were sequenced at the UC Berkeley DNA Sequencing Facility using T7 primers and analyzed on an Applied Biosystems 3730xl DNA Analyzer. Sequences were aligned and compared to published sequences using the Ribosomal Database Project (RDP) on-line interface (Cole et al., 2007, 2009) and were quality checked for chimeras using the RDP Chimera check program and Bellerophon (Huber et al., 2004). Phylogenetic analysis of 827 sequences was conducted in ARB (Ludwig et al., 2004) and neighbor-joining phylogenetic trees were constructed using the Jukes-Cantor correction. The distance matrix derived from the neighbor joining analysis was used in DOTUR for assignment of operational taxonomic units (OTUs) (Schloss and Handelsman, 2005). All sequences were submitted to the National Center for Biotechnology Information (NCBI) GenBank database as accession numbers KF941543–KF942369.

#### **QUANTITATIVE PCR (qPCR)**

Total *Vibrio* abundances were quantified using a SYBR Green qPCR method utilizing the *Vibrio* specific primers 567F and 680R (Table 1) from Thompson et al. (2004b) to ensure a broad spectrum of *Vibrio* species were included in the analysis. DNA template (2μl) was added to 12.5μl Applied Biosystems SYBR® Green PCR Master Mix, 8.25μl of 5 kD filtered water, 0.25μl of bovine serum albumin (BSA) and 1μl of each primer for a total volume of 25μl. The cycling conditions were 50◦C for 2 min, 95◦C for 10 min followed by 40 cycles of 95◦C for 15 s and 64◦C for 1 min. Each run was followed by a dissociation step (95◦C for 15 s and 60◦C for 1 min 95◦C for 15 s and 60◦C for 15 s) to determine a melt curve for analysis of specificity.

Two qPCR assays utilizing TaqMan® chemistry were designed to target two *Vibrio* subpopulations within MB based on sequences of interest identified from the MB 16S rRNA clone library. The MBAY Vib4 group was closely associated with *Alllivibrio* sp. and the MBAY Vib7 group was associated with *Vibrio penaecida.* Primers and probes were designed with Primer Express (**Table 1**) (ABI, Foster City, CA) and synthesized by Sigma-Genosys (Woodlands, TX). The reactions for the subpopulation assays contained 2μl of sample DNA, 12.5μl Applied Biosystems TaqMan® Universal PCR Master Mix, 8μl of 5 kD filtered water, 1μl each of the forward and reverse primer and 0.5μl of the probe. Samples were assessed using a TaqMan® qPCR method with cycling conditions of 50◦C for 2 min, 95◦C for 10 min followed by 45 cycles of 95◦C for 15 s and 60◦C for 1 min. The qPCR primer sets were used to analyze the full range of samples at both stations.

Both qPCR assays compared *C*<sup>T</sup> values to standard curves (equivalent to 101–108 gene copies per reaction) derived from group specific environmental isolate clones from the MB clone library linearized with Nad1. An internal control at a concentration of 10<sup>4</sup> gene copies of linearized standard assessed the runs for inhibition. If detected, DNA samples were diluted 1:10 and amplification was repeated. All qPCR runs were performed on a 7500 Applied Biosystems Real-Time PCR instrument. Final concentrations are reported as gene copies l−<sup>1</sup> of seawater sampled.

#### **STATISTICAL ANALYSIS**

Statistical analysis was performed in JMP version 9 (SAS Institute, North Carolina). Data was natural log transformed [ln (x + 1)] to scale variables for graphical interpretation and adjust for normalcy. A Principal Components Analysis (PCA) bi-plot was utilized to visualize relationships between *Vibrio* populations and physical and biological variables and Spearman's rank analysis (ρs) was used to define statistically significant relationships.

#### **RESULTS**

#### **ENVIRONMENTAL CONDITIONS IN MB**

Water temperatures from the sample sites ranged from 7.7 to 15.4◦C (**Table 2**). Salinity ranged from 32.7 to 34.4 ppt and chlorophyll *a* values ranged from 0 to 18.45μg liter−<sup>1</sup> (**Table 2**). The Monthly Upwelling Index (MUI) ranged from 46 to 237 m<sup>3</sup> s <sup>−</sup><sup>1</sup> 100 m of coastline−1. The daily Upwelling Index was analyzed against surface conditions at M2 (**Table 3A**). Results were consistent with expected trends for upwelling events with a negative correlation to temperature (ρ<sup>s</sup> = −0.47, *p* = 0.002) and a positive correlation to both nitrate (ρ<sup>s</sup> = 0.22, *p* = 0.035) and chlorophyll *a* (ρ<sup>s</sup> = 0.28, *p* = 0.009). Principal Component Analysis associated total diatoms, total dinoflagellates and all phytoplankton with the upwelling season and the monthly UI. *Synechococcus* was more closely associated with station M2 (**Figure 2**).

#### **16S rRNA GENE PHYLOGENETIC ANALYSIS**

Multiple distinct 16S rRNA-based phylotypes were recovered from the MB clone libraries (**Figure 3**). The primer sets utilized to develop the MB clone libraries targeted *Vibrio* sp. strains but 67 sequences within the library clustered with the closely related *Photobacterium* sp. There were also a significant number of sequences that clustered with unidentified *Vibrio* strains and uncultured marine bacteria. *Vibrio splendidus* was defined in several distinct clades within the phlogenetic tree. Other MB sequences clustered with *V. pomeroyi* and *V. lentus*;


**Table 2 | CTD measurement and** *Vibrio* **concentration ranges over the study period for all stations and depths.**

*nd, none detected.*

*V. pectenicida* and *V. rumoiensis*; and *V. aesturianus*. One clade included sequences that clustered with *V. agarivorans, V. hispanicus, V. haliticoli, V. ezurae, V. fortis, V. proteolyticus, V. sinaloensis, V. nigripulchritudo, V. harveyi, V. natriegens*, and *V. alginolyticus* (Assorted Vibrio in **Figure 3**). Quantitative PCR (qPCR) assays were developed for two specific phylotypes recovered from the MB clone libraries. The MBAY Vib4 group (**Figure 3B**) clustered with *Allivibrio* sp., which includes *A. salmonicida*, *A. fischeri*, *A. wodanis*, and *A. logei.* MBAY Vib4 included sequences obtained from samples collected from both stations C1 and M2. Within the clone library the MBAY Vib7 group included sequences that were only obtained at station M2 and clustered with *Vibrio penaecida* (**Figure 3C**).

Rarefaction curves were determined based on the neighborjoining distance matrix of sequences within the MB clone libraries. Of the 827 sequences included in the total sequences analysis, 158 OTUs were defined at 97% identity, 103 OTUs at 95% identity, 61 OTUs at 93% identity, and 30 OTUs at 90% identity. Only at 93 and 90% identity did the rarefaction curves begin to reach an asymptote. Analysis revealed that there was no significant difference in species richness between station C1 and M2 but that both sampling efforts did not attain full coverage to estimate total population OTUs. Rarefaction analysis of the 327 sequences from station C1 defined 97 OTUs at 97% identity, 69 OTUs at 95% identity, 45 OTUs at 93% identity, and 23 OTUs at 90% identity. Rarefaction analysis of the 500 sequences from station M2 defined 98 OTUs at 97% identity, 66 OTUs at 95% identity, 41 OTUs at 93% identity, and 21 OTUs at 90% identity.

#### **QUANTIFICATION OF** *Vibrio* **POPULATION ABUNDANCES**

There were observed differences in the *Vibrio* populations at both of the stations, with the MBAY Vib4 group having higher peak abundances at station C1 (**Figure 4**). The MBAY Vib4 group ranged in concentration from undetectable levels at all depths and stations up to 1.<sup>95</sup> <sup>×</sup> <sup>10</sup><sup>6</sup> gene copies liter−<sup>1</sup> seawater at 5 m at station C1 (**Table 2**). MBAY Vib4 was positively correlated to temperature (ρ<sup>s</sup> = 0.32, *p* = 0.002), oceanic season (ρ<sup>s</sup> = 0.41, *p* = 0.002), *Synechococcus* (ρ<sup>s</sup> = 0.20, *p* = 0.02), and total *Vibrio* abundance (ρ<sup>s</sup> = 0.55, *p* < 0.001) (**Table 3**). This group was negatively correlated to salinity (ρ<sup>s</sup> = −0.22, *p* = 0.02), chlorophyll *a* (ρ<sup>s</sup> = −0.16, *p* = 0.035), monthly UI (ρ<sup>s</sup> = −0.61, *p* < 0.001), surface phytoplankton (ρ<sup>s</sup> = −0.44, *p* < 0.001), and total diatoms (ρ<sup>s</sup> = −0.55, *p* < 0.001) (**Table 3**).

The MBAY Vib7 population was not as abundant as MBAY Vib4 (**Figure 4**), with concentrations ranging from undetectable to 5.<sup>18</sup> <sup>×</sup> <sup>10</sup><sup>5</sup> gene copies liter−<sup>1</sup> seawater (**Table 2**). MBAY Vib7 was positively correlated with depth (ρ<sup>s</sup> = 41, *p* < 0.001) (**Table 3**), with the highest detected concentration at 100 m at station M2 (**Table 2**). MBAY Vib7 populations were positively associated with station M2 (ρ<sup>s</sup> = 0.53, *p* < 0.001), surface *Synechococcus* populations (ρ<sup>s</sup> = 0.41, *p* < 0.001), and the Davidson season (ρ<sup>s</sup> = 0.15, *p* = 0.0486) (**Table 3**). Temperature (ρ<sup>s</sup> = −0.30, *p* = 0.0005), chlorophyll *a* (ρ<sup>s</sup> = −0.47, *p* < 0.001) and monthly UI (ρ<sup>s</sup> = −0.23, *p* = 0.012) were all negatively associated with MBAY Vib7 (**Table 3**). MBAY Vib7 was identified in samples from both station C1 and M2 but at lower concentrations and at deeper depths at station C1.

Total *Vibrio* populations showed some variability in abundance but were consistently present throughout the study at all depths and stations (**Figure 4**). Total *Vibrio* population abundances ranged from 1.<sup>80</sup> <sup>×</sup> 103 to 3.<sup>72</sup> <sup>×</sup> <sup>10</sup><sup>6</sup> gene copies liter−<sup>1</sup> seawater (**Table 2**). The abundance of total *Vibrio* populations was positively correlated to MBAY Vib4 (ρ<sup>s</sup> = 0.55, *p* < 0.001), and the oceanic season (ρ<sup>s</sup> = 0.31, *p* < 0.001) and negatively correlated to station number (ρ<sup>s</sup> = −0.32, *p* < 0.001) the monthly UI (ρ<sup>s</sup> = −0.26, *p* = 0.0039) and upwelling season (ρ<sup>s</sup> = −0.21, *p* = 0.0067) (**Table 3**). No other parameters examined in this study showed significant correlation with the abundance of the total *Vibrio* population.

#### **DISCUSSION**

Genetic diversity within *Vibrio* populations can be up to 7% when examining the entire 16S rRNA gene (Dorsch et al., 1992; Kitatsukamoto et al., 1993). This variability is mostly located **Table 3 | Spearman's rho (ρs) statistical analysis of physical and biological parameters.**



*Significant parameters (p* ≤ *0.05) are in bold type. UI, upwelling index. (A) Upwellling parameters verified against surface data from the M2 mooring station. (B) CTD data and environmental parameters. (C) Surface phytoplankton and M2 mooring station data.*

within 35 defined regions, 28 of which are located between the primers used to construct the clone library for this study (Wiik et al., 1995; Jensen et al., 2009). It is thus not surprising to have identified a significant number of OTUs and to observe deep branching within the phylogenetic tree of the MB clone library. Intraspecies sequence variability can also be high within *Vibrio* species (Jensen et al., 2009) and the identification of *Vibrio splendidus* within different branches of the tree is supported by observations of up to 2% difference within the 16S rRNA gene of this species (Jensen et al., 2009; Le Roux et al., 2009). Defining relationships within *Vibrio* populations is even more complex as their genomes contain multiple copies of the rRNA operon (Heidelberg et al., 2000). Thirteen copies have been identified in *V. natriegens* and 12 copies in both *Allivibrio fischeri* and *A. salmonicida* (Lee et al., 2009). Intragenomic heterogeneity within these copies may be high, as seen in *V. parahemolyticus* (Harth et al., 2007), or non-existent as is noted in other *Vibrio* species

(Coenye and Vandamme, 2003). Since this study was not designed to differentiate between heterogeneous intragenomic copies and individuals, our results may overestimate diversity within the MB clone library. With that being said, sequences did cluster with over 20 known *Vibrio* species suggesting that MB does in fact contain diverse *Vibrio* populations.

Multiple copies of the rRNA operon may also affect the qPCR estimates of abundance presented in this study, but is not expected to significantly change the interpretations of our findings. Even an estimate of 20 gene copies per cell results in estimates of abundances that fall within the range of concentrations observed in other coastal waters (Thompson et al., 2004a), and are higher than observations from samples in similar temperature ranges (Randa et al., 2004; Eiler et al., 2006). Temperature is often cited as a significant driver in *Vibrio* population abundance (Heidelberg et al., 2002; Maeda et al., 2003; Eiler et al., 2006, 2007; Hsieh et al., 2008), and while temperature variability was correlated to the two subpopulations examined in this study, it was the upwelling index (MBAY Vib4) and depth and station (MBAY Vib7) that were more significant factors in defining subpopulation abundance.

MBAY Vib4 and MBAY Vib7 are separate subpopulations of the *Vibrio* community with distinct niches in the MB. The MBAY Vib4 is defined by close association with *Allivibrio* sp. and represents an oceanic season subpopulation. This season is characterized by a lack of upwelling, lower sea surface salinities, warmer sea surface temperatures and waters that are influenced by wind relaxation events and the slow flowing California Current (Rosenfeld et al., 1994). MBAY Vib7 represents a deeper water subpopulation of *Vibrio* sp. with a greater association with the offshore waters at station M2. This station is influenced by upwelled waters mixed with California Current waters, the

#### **FIGURE 3 | Continued**

distance matrix method with a Jukes-Cantor correction. T, total number of sequences in clade, MB, MB sequences in clade. Sample number, station, depth, and date are listed for Monterey Bay sequences. **(A)** Major groups of environmental *Vibrio* sequences recovered from the MB. *Shewanella algae* included as the tree root. **(B)** *Allivibirio* and MBAY Vib4 group sequences, and **(C)** *Vibrio penaecida* and MBAY Vib7 sequences. Scales represent 10% or 0.1% nucleotide substitutions.

California Undercurrent and the Davidson Current-a northward flowing current that develops in winter months (Tisch et al., 1992; Breaker and Broenkow, 1994; Rosenfeld et al., 1994). The California Undercurrent usually flows below 100 m along the California shelf but often nears the surface during the Davidson season (Tisch et al., 1992; Pierce et al., 2000; Tseng et al., 2005). The largest hydrographic changes around MB occur during transitions between upwelling regimes (i.e., during oceanic phases) and in winter when the horizontal and vertical thermal gradients are reduced due to the northward flow of the Davidson Current (Bac et al., 2003; Storlazzi et al., 2003; Warn-Varnas et al., 2007). Dynamics of the MB *Vibrio* populations examined in this study seem to reflect these hydrographic processes, and may be highly influenced by them-MBAY Vib4 during the oceanic phases and MBAY Vib7 during the winter phases.

In 2006, the MB had experienced 4 years of anomalous oceanographic conditions. Delayed and unusually shallow upwelling affected the food web in MB up through higher trophic levels (Goericke et al., 2007). Changes included a shift in dominant toxin-producing algal species from diatoms to dinoflagellates, poor recruitment of krill and low zooplankton biomass as well as seabird reproductive failure (Peterson et al., 2006; Goericke et al., 2007; Jester et al., 2009). MBAY Vib4 was positively correlated to this year possibly due to delayed upwelling, reduced chlorophyll *a* concentrations and warmer surface temperatures (**Table 3**). Increased abundances of *Vibrio* populations could also have been influenced by decreased competition from phytoplankton for resources or reduced top down control from zooplankton. Following the 2002–2006 lull events, unusually strong upwelling was observed in 2007 (Kaplan et al., 2009), and in 2008 a strong development of upwelling was observed at the beginning of the season (**Figure 3**). MBAY Vib4 populations showed significantly reduced abundance during those upwelling seasons. MBAY Vib7 and the total *Vibrio* populations did not seem to be strongly affected by these events.

Overall, the *Vibrio* populations examined in MB seemed to be influenced by larger scale upwelling events and shifts in currents and oceanographic seasons rather than individual environmental factors. In the future, more extensive analysis of nutrient availability and oceanographic parameters (i.e., flow velocities and wind patterns) may provide better insight into how upwelling characteristics and water mass flows play a role in *Vibrio* population dynamics. Since *Vibrio* populations might have significant influence over nutrient availability through recycling chitin and other bound nutrient resources, understanding the role of these bacterial populations in both surface and deeper waters can better our understanding of the productivity potential within upwelling regions.

#### **ACKNOWLEDGMENTS**

The authors would like to thank the Biological Oceanography Group at the Monterey Bay Aquarium Research Institute, especially Francisco Chavez and Tim Pennington for providing us with the phytoplankton data and ship time. The authors would also like to thank Kendra Turk-Kubo, Fitnat Yildiz and Shellie Bench for helpful discussions. Partial funding was provided by a Gordon and Betty Moore Foundation Marine Investigator award (Jonathan P. Zehr).

#### **REFERENCES**

Alderkamp, A. C., Van Rijssel, M., and Bolhuis, H. (2007). Characterization of marine bacteria and the activity of their enzyme systems involved in degradation of the algal storage glucan laminarin. *FEMS Microbiol. Ecol.* 59, 108–117. doi: 10.1111/j.1574-6941.2006.00219.x


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 01 October 2013; paper pending published: 26 October 2013; accepted: 22 January 2014; published online: 12 February 2014.*

*Citation: Mansergh S and Zehr JP (2014) Vibrio diversity and dynamics in the Monterey Bay upwelling region. Front. Microbiol. 5:48. doi: 10.3389/fmicb. 2014.00048*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Mansergh and Zehr. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

## Associations and dynamics of Vibrionaceae in the environment, from the genus to the population level

#### *Alison F. Takemura†, Diana M. Chien† and Martin F. Polz\**

*Parsons Lab for Environmental Science and Engineering, Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Anwar Huq, University of Maryland, USA Darrell J. Grimes, The University of*

#### *\*Correspondence:*

*Southern Mississippi, USA*

*Martin F. Polz, Parsons Lab for Environmental Science and Engineering, Department of Civil and Environmental Engineering, Massachusetts Institute of Technology, Building 48, Room 417, 15 Vassar Street, Cambridge, MA 02139, USA e-mail: mpolz@mit.edu*

*†These authors have contributed equally to this work.*

the causative agent of cholera, and *V. vulnificus*, the deadliest seafood-borne pathogen, are a well-studied family of marine bacteria that thrive in diverse habitats. To elucidate the environmental conditions under which vibrios proliferate, numerous studies have examined correlations with bulk environmental variables—e.g., temperature, salinity, nitrogen, and phosphate—and association with potential host organisms. However, how meaningful these environmental associations are remains unclear because data are fragmented across studies with variable sampling and analysis methods. Here, we synthesize findings about *Vibrio* correlations and physical associations using a framework of increasingly fine environmental and taxonomic scales, to better understand their dynamics in the wild. We first conduct a meta-analysis to determine trends with respect to bulk water environmental variables, and find that while temperature and salinity are generally strongly predictive correlates, other parameters are inconsistent and overall patterns depend on taxonomic resolution. Based on the hypothesis that dynamics may better correlate with more narrowly defined niches, we review evidence for specific association with plants, algae, zooplankton, and animals. We find that *Vibrio* are attached to many organisms, though evidence for enrichment compared to the water column is often lacking. Additionally, contrary to the notion that they flourish predominantly while attached, *Vibrio* can have, at least temporarily, a free-living lifestyle and even engage in massive blooms. Fine-scale sampling from the water column has enabled identification of such lifestyle preferences for ecologically cohesive populations, and future efforts will benefit from similar analysis at fine genetic and environmental sampling scales to describe the conditions, habitats, and resources shaping *Vibrio* dynamics. The Vibrionaceae, which encompasses several potential pathogens, including *V. cholerae*,

**Keywords:** *Vibrio***, population, environmental correlation, ecology, niche, attachment, planktonic**

#### **INTRODUCTION**

The family *Vibrionaceae* (or vibrios for short) comprises a genetically and metabolically diverse group of heterotrophic bacteria that are routinely found in all ocean environments, ranging from coastal to open and surface to deep water (Thompson et al., 2004; Thompson and Polz, 2006). Moreover, a few *Vibrio* species have extended their range beyond the marine environment, occurring predominantly in brackish and even freshwater environments (Thompson et al., 2004). The study of the environmental distribution and dynamics of vibrios has a long history, largely because many species contain potential human and animal pathogens (Thompson et al., 2004, 2005). Hence there is considerable public health and economic interest in determining factors correlated to increased abundance of vibrios (Stewart et al., 2008). Moreover, vibrios are easily cultured on standard and selective media and thus were highly visible in the pre-molecular era of microbial ecology. In recent years, environmental dynamics have also been studied with culture-independent methods allowing for a more fine-scale assessment of environmental drivers of occurrence, and the vibrios have become a model for bacterial population biology and genomics. In fact, presently, the vibrios represent one of the best-studied models for the ecology and evolution of bacterial populations in the wild.

The early discovery that some fish species harbor high numbers of vibrios (e.g., Liston, 1954, 1957; Aiso et al., 1968; Sera et al., 1972) has led to the widespread notion that these bacteria are only transient members of microbial assemblages of the water column. Instead, vibrios were regarded as specifically associated with animals, and occurrence in water samples was thought to be primarily due to their excretion with fecal matter. This picture was enforced by the discovery that several luminescent *Vibrio* (*Allivibrio*) and related *Photobacterium* species form intimate symbioses with animals (e.g., fish, squid) (Ruby and Nealson, 1976; Stabb, 2006). More recent work has, however, revealed that the notion of vibrios being "enterics of the sea" (Liston, 1954) represents an oversimplification. Many *Vibrio* species grow actively in ocean water either in the free-living phase or associated with various types of organic particles, many of which are of non-animal origin (Lyons et al., 2007; Froelich et al., 2012). Thus although associationwith animals can be an important part of the life cycle of many *Vibrio* species, there are others that only loosely associate with animals or not at all, an aspect we explore in detail in this review.

Another widely held belief about vibrios is that they play a relatively minor role in chemical transformations in the ocean, despite the wide range of metabolisms [e.g., chitin degradation (Hunt et al., 2008a; Grimes et al., 2009)] of which they are capable. This belief is largely based on low to medium average relative abundance of *Vibrionaceae* in ocean water. Yet three considerations suggest that the role of vibrios has been underestimated. First, it has been pointed out that although vibrios' abundances are generally only around 10<sup>3</sup> to 10<sup>4</sup> cells per ml seawater (i.e., on the order of few percent of total bacteria), they have very high biomass (Yooseph et al., 2010). For example, an actively growing *Vibrio* can have 100× the biomass of *Pelagibacter*, which, at <sup>∼</sup>105 cells per ml, is typically the most abundant heterotrophic member of bacterial assemblages in the ocean (Yooseph et al., 2010). Second, new time-series analysis shows that vibrios are capable of blooms in the water column during which they can even become the predominant members of the total bacterial assemblage (Gilbert et al., 2012). These blooms had been missed previously because they are of relatively short duration, yet they confirm that vibrios, which are capable of very rapid growth in laboratory media, can reach high doubling rates in the environment. Finally, vibrios might be disproportionately subject to predation by protozoa and viruses (Worden et al., 2006; Suttle, 2007), likely due to their comparatively large size. For example, cells were found in one study to measure more than three times the community average in volume, and, along with other similarly large genera, suffered especially high grazing mortality (Beardsley et al., 2003). Taken together, these considerations suggest that vibrios should be re-evaluated for their role in biogeochemical processes in the ocean since they have disproportionately high biomass that is subject to high turnover by rapid growth in concert with high predation.

The purpose of this review is to provide an overview of known environmental factors and ecological associations affecting *Vibrio* abundance and dynamics. We note that although we look at the dynamics of potentially pathogenic species, we purposefully exclude data on pathogenesis itself since this is outside the scope of this review. We first focus on total *Vibrio* (i.e., the assessment of occurrence of members of the genus or family), which have often been measured as a proxy for potential pathogen occurrence, asking whether they can be treated as an environmentally cohesive unit. To what extent do total vibrios correlate to specific environmental variables, and do these measures have predictive power for individual species? To address this question, we present meta-analyses of the dynamics of *V. cholerae*, *V. parahaemolyticus*, and *V. vulnificus*, three species harboring genotypes potentially pathogenic to humans. The limitation to these three is necessary since public health interests have driven much of the research so that the literature is highly biased toward human pathogens. In this context, a further important question is to what extent easily measurable bulk parameters, such as temperature, salinity, nutrients, dissolved oxygen and/or chlorophyll a are good correlates for total vibrios or specific species, allowing easy and cost-effective risk assessment.

However, because our meta-analysis suggests poor or inconsistent performance of most bulk parameters, we researched alternative, frequently finer-scale environmental variables. These

Finally, we summarize recent research aimed at defining habitat characteristics and phylogenetic bounds of ecologically cohesive populations among co-existing vibrios, using the water column and macroinvertebrates as examples of adaptive landscapes. This research demonstrates that such populations, which may or may not correspond to named (taxonomic) species, represent eco-evolutionary units that allow testing of hypotheses of how populations are structured by environmental selection and gene flow.

#### **ENVIRONMENTAL CORRELATES OF** *Vibrio* **PRESENCE AND ABUNDANCE**

To better understand under what conditions vibrios occur and proliferate, most studies have investigated environmental variables that can be measured from bulk seawater such as temperature, salinity, dissolved oxygen, nitrogen, phosphorus, and chlorophyll a concentrations. These are attractive since they are easily measured and many are observable remotely by buoy or satellite (e.g., Lobitz et al., 2000) so that potential for presence of pathogenic vibrios might be easily assessed. In addition, several studies have extended measurements to more complex physicochemical and biotic variables, including dissolved organic carbon (DOC) and zoo- and phyto-plankton taxa.

In the following, we first ask how informative these variables are by conducting a meta-analysis to compare correlations across studies, for both total *Vibrio* as well as the potential pathogens *V. cholerae, V. parahaemolyticus*, and *V. vulnificus*, and, second, determine if the genus and species levels exhibit similar patterns. To determine the potential impact of environmental variables, we looked at how strong their correlations are by comparing coefficient of determination values, *R*2, reported in the literature. A goodness of fit parameter, *R*<sup>2</sup> varies from 0 (no explanation of variance in the dependent variable) to 1 (perfect explanation)**,** giving us a means of assessing, for example, whether temperature better predicts abundance of total *Vibrio*, than salinity does. Studies included have regression analyses with associated *R*2 values, or Spearman or Pearson correlations, whose rho values were squared to obtain *R*2. Additionally, we compare how their abundances trend along gradients in two particularly well-studied variables, salinity and temperature.

#### **TOTAL** *Vibrio*

When correlations across studies are compared, we see that the strongest environmental correlates to total *Vibrio* are temperature and salinity. These two variables most often explain the greatest amount of variance in total *Vibrio* abundance in the water column (**Figure 1**), whereas consideration of additional variables often makes only marginal improvements (e.g., in Heidelberg et al., 2002a,b; Oberbeckmann et al., 2012; Froelich et al., 2013). However, a minority of analyses has found temperature and

salinity to be non-significant toward explaining *Vibrio* abundance. This inconsistency might be a result of the ranges considered; for instance, temperature may be found non-significant due to a narrow range observed, such that *Vibrio* abundance varies little. In fact, evidence supports this hypothesis; the correlation strength of temperature to vibrios varies by season (Oberbeckmann et al., 2012; Froelich et al., 2013), suggesting the magnitude of the correlation may depend on the temperature range examined. For instance, Oberbeckmann et al. (2012) and Froelich et al. (2013) both observed the highest correlation of temperature and *Vibrio* during the seasons with the broadest temperature ranges, spring, and fall, respectively. Additionally, it is possible that at lower temperatures vibrios exhibit less variation in abundance; two studies assessing total vibrios in the cooler waters of the Baltic Sea and North Sea found non-significant correlations (Eiler et al., 2006; Oberbeckmann et al., 2012).

Compared to salinity and temperature, other environmental measures usually explain less variance in total *Vibrio*. Dissolved oxygen has had little explanatory power; for instance, in **Figure 1**, its largest *R*<sup>2</sup> was less than half that of temperature in the same analysis (Blackwell and Oliver, 2008). The same is true for nitrogen, whose highest *R*<sup>2</sup> was still less than temperature's (Blackwell and Oliver, 2008). In the environments examined, phosphate, pH, and turbidity explain little variance, and DOC explains none at all, albeit the number of studies used for DOC in this meta-analysis is limited. Of interest, though not depicted, potential host organisms, copepods, decapods, and cyanobacteria, have been found to explain relatively little variance in total vibrios when considered in a model that already incorporates temperature (Turner et al., 2009; Vezzulli et al., 2009), and similarly for dinoflagellates when salinity is first considered (Eiler et al., 2006). Turner et al. (2009) did observe that diatoms explained more variance than temperature. While this might imply a physical association, the correlation was negative, suggesting that total *Vibrio*, at least as a whole, do not associate with diatoms.

Chlorophyll a, on the other hand, has had noted importance in two datasets: the spring and summer of the study by Oberbeckmann et al. (2012), with *R*2-values of 60 and 26%, respectively. These were in fact higher than correlations to temperature or salinity in these seasons. Perhaps during this period, as temperature warms, growth conditions favor phytoplankton blooms that impact *Vibrio* abundance (Oberbeckmann et al., 2012). However, Froelich et al. (2013) did not make these same observations in their seasonal datasets. This inconsistency may be a product of the fact that different *Vibrio* species likely affiliate with or feed on exudates of specific algal taxa only, rather than algae in general, a subject further discussed in the section The Evidence for a Planktonic, Free-Living Lifestyle.

Given the frequent strength of temperature and salinity as correlates, we asked, how do total vibrios distribute with respect to these variables when their combined effect is considered? A few studies have modeled the bivariate relationship, finding that total *Vibrio* abundance increases as temperature and salinity increase (Hsieh et al., 2008; Turner et al., 2009; Froelich et al., 2013). The ranges investigated were also broad, lending confidence that these results are general; for example, Hsieh et al. (2008) modeled from 2.5 to 32.5◦C and 0 to 27 ppt, respectively.

#### *V. cholerae, V. parahaemolyticus***, AND** *V. vulnificus*

We compare environmental correlates and trends noted in total *Vibrio* to three species that have been well-sampled across locales: *V. cholerae, V. parahaemolyticus,* and *V. vulnificus*. While it would also be interesting to consider species beyond potential pathogens, their environmental data is much more limited.

In *V. cholerae*, we see an interesting shift from total *Vibrio* in the strength of correlating environmental variables: some biotic variables are as strong or, in fact, stronger than temperature or salinity (**Figure 2**). Total *Vibrio*, congenerics *V. vulnificus* and *V. parahaemolyticus*, as well as a dinoflagellate genus (*Prorocentrum*) and cladoceran species (*Diaphanosoma mongolianum*) have all significantly correlated to *V. cholerae* abundance (Eiler et al., 2006; Blackwell and Oliver, 2008; Kirschner et al., 2011; Prasanthan et al., 2011). Moreover, *V. parahaemolyticus* abundance has explained more *V. cholerae* abundance variance than nitrogen, temperature, or salinity in (Prasanthan et al., 2011), and dinoflagellate abundance has explained more variance than phosphorus, salinity, or temperature (Eiler et al., 2006). While correlations to plankton may represent direct associations, such high correlation of vibrios to each other is likely not indicative of causal interactions, but rather stems from overlap in environmental ranges and/or habitats (Blackwell and Oliver, 2008). *E. coli* and total coliforms have also correlated to *V. cholerae* abundance, though both groups may simply be responding to anthropogenic nutrient influxes favoring growth of heterotrophs (Blackwell and Oliver, 2008).

Long thought to be a reservoir of toxigenic *V. cholerae*, zooplankton, and particularly copepods, are hypothesized to correlate to *V. cholerae* abundance. Surprisingly, however, when de Magny et al. (2011) examined several zooplankton genera and species, including copepods *Cyclops* and *Diaptomus,* they did not find significant correlations to any zooplankter except the rotifer *Brachionus angularis* (not depicted in **Figure 2**, because Monte Carlo analysis did not yield *R*2-values)*.* While the association between *V. cholerae* O1/O139 and the copepod *Acartia tonsa* has also been studied (Huq et al., 2005; Lizárraga-Partida et al., 2009), quantitatively significant correlation in the environment has remained elusive. For instance, Lizárraga-Partida et al. (2009) demonstrated only a qualitative link between *V. cholerae* O1 presence coincident with an increase in *A. tonsa*, even though laboratory studies have shown ready attachment (e.g., Huq et al., 1984; Rawlings et al., 2007).

*V. cholerae* has also been hypothesized to correlate with chlorophyll a, a potential proxy of algal and zooplankton growth, and/or a eutrophic environment conducive to heterotroph growth, but chlorophyll a's general predictive value is unclear. While significant in Eiler et al. (2006), other studies have observed no correlation of chlorophyll a to *V. cholerae* abundance (Jiang and Fu, 2001; Kirschner et al., 2008; Mishra et al., 2012). Yet *V. cholerae* growth has been observed experimentally to depend on DOC, which could relate to phytoplankton abundance and thus chlorophyll a (Eiler et al., 2007). In microcosm experiments, Eiler et al. (2007) demonstrated that adding 2.1 mg carbon L−<sup>1</sup> of cyanobacterial-derived dissolved organic matter influenced bacterial growth more than a 12–25◦C change in temperature. The inconsistency of chlorophyll a, and, incidentally, bulk DOC (which showed no significant correlation) (Eiler et al., 2006; Blackwell and Oliver, 2008; Kirschner et al., 2008; Neogi et al., 2012) as correlates might be due to the quality of exudates; its composition of refractory humic substances (Kirschner et al., 2008) or derivation from different algal species, differentially stimulating *V. cholerae* growth [(Worden et al., 2006), see also section The Evidence for a Planktonic, Free-Living Lifestyle]. Interestingly, the lack of clear support for chlorophyll a's influence on *V. cholerae* environmental abundance is in contrast to the fact that chlorophyll a can correlate with cholera *disease* incidence (de Magny et al., 2008), and has been used in predictive models for cholera in Bangladesh (Bertuzzo et al., 2012; Jutla et al., 2013).

Like *V. cholerae*, *V. parahaemolyticus* abundance in water samples is also strongly correlated to temperature, and was found significant in all but one analysis reviewed here (DePaola et al., 1990; Zimmerman et al., 2007; Blackwell and Oliver, 2008; Caburlotto et al., 2010; Deter et al., 2010; Johnson et al., 2010, 2012; Böer et al., 2013), with maximal *<sup>R</sup>*<sup>2</sup> <sup>=</sup> <sup>50</sup>.6% (Deter et al., 2010) (**Figure 3**). Blackwell and Oliver (2008) found that *V. parahaemolyticus* correlates both to total *Vibrio* and congenerics, as well as coliforms and *E. coli*. These variables were only considered in a single study, however, so it is not known if the relationships hold across different sampling locations. The significance of salinity is variable for *V. parahaemolyticus* with only three of seven studies having non-zero *R*2-values (**Figure 3**) (Zimmerman et al., 2007; Caburlotto et al., 2010; Johnson et al., 2010), but this may be due to *V. parahaemolyticus* colonizing a large salinity range, as detailed below (**Figure 6**).

Correlation to environmental variables has also frequently been studied for *V. parahaemolyticus* occurring in sediment and shellfish, though trends remain unclear. In sediment, considered a potential reservoir (Vezzulli et al., 2009), individual regressions of *V. parahaemolyticus* abundance to temperature, salinity, and total organic carbon have yielded moderate *R*2-values, at times above 30% (Blackwell and Oliver, 2008; Deter et al., 2010; Johnson et al., 2012; Böer et al., 2013). However, some studies have found salinity or temperature to be a non-significant explanatory variable (Blackwell and Oliver, 2008; Deter et al., 2010; Johnson et al., 2010).

In shellfish, a common vehicle of virulent vibrios to humans, the incidence of temperature and salinity as correlates to *V. parahaemolyticus* is also inconsistent. Salinity has been found explanatory in some studies, with *R*<sup>2</sup> as high as 42% (DePaola et al., 2003; Johnson et al., 2010, 2012) and non-significant in others (Deepanjali et al., 2005; Deter et al., 2010; Sobrinho et al., 2010). Temperature can explain moderate amounts of variance in *V. parahaemolyticus* abundance (DePaola et al., 1990, 2003; Cook et al., 2002; Johnson et al., 2010, 2012; Sobrinho et al., 2010), with significant *R*<sup>2</sup> as high as 44% (Cook et al., 2002), though other studies have found little or no correlation (Deepanjali et al., 2005; Duan and Su, 2005; Deter et al., 2010). The absence of correlation is surprising, given that temperature's effect is amplified by influencing shellfish's ability to concentrate *V. parahaemolyticus* from surrounding water. Oysters can enrich *V. parahaemolyticus* over 100-fold (DePaola et al., 1990; Shen et al., 2009), and the magnitude of concentration is temperature-dependent, with effects greatest at

32◦C and less, but still evident, in cooler waters (Shen et al., 2009).

For *V. vulnificus* isolated from the water column, temperature is the strongest correlate among measured environmental variables, and often explains more variance in *V. vulnificus* than for other species or total *Vibrio*; several analyses found temperature explained over 50% of the variance in *V. vulnificus* sampled from water (Motes et al., 1998; Randa et al., 2004; Blackwell and Oliver, 2008; Nigro et al., 2011) (**Figure 4**). Moreover, temperature has been a stronger correlate than chlorophyll a (Randa et al., 2004; Johnson et al., 2010, 2012), dissolved oxygen (Pfeffer et al., 2003; Blackwell and Oliver, 2008; Ramirez et al., 2009), and nitrogen (Pfeffer et al., 2003; Blackwell and Oliver, 2008). While DOC is an inconsistent correlate, it has been more explanatory than temperature in at least one study (Jones and Summer-Brason, 1998). The variable pH, however, is not a significant correlate (Lipp et al., 2001; Pfeffer et al., 2003; Blackwell and Oliver, 2008; Ramirez et al., 2009; Franco et al., 2012), nor is phosphorus (Pfeffer et al., 2003; Blackwell and Oliver, 2008). Turbidity has been found nonsignificant in several studies (Lipp et al., 2001; Pfeffer et al., 2003; Wetz et al., 2008; Ramirez et al., 2009), or not as explanatory as temperature (Blackwell and Oliver, 2008). While salinity, when significant, has generally been less informative than temperature (Motes et al., 1998; Randa et al., 2004; Warner and Oliver, 2008; Johnson et al., 2010), it has, in one analysis, been more (Lipp et al., 2001).

Biotic correlates have also been identified for *V. vulnificus*. Total bacteria (Pfeffer et al., 2003; Randa et al., 2004; Blackwell and Oliver, 2008), enteroccous (Wetz et al., 2008; Ramirez et al., 2009), coliforms (Pfeffer et al., 2003; Blackwell and Oliver, 2008) and *E. coli* (Pfeffer et al., 2003; Blackwell and Oliver, 2008; Wetz et al., 2008) have been studied only sporadically, but their correlation strength to *V. vulnificus* has usually been less than temperature's; one exception, however, is enterococcus in (Ramirez et al., 2009), potentially indicative of a surge in nutrients overtaking temperature's effect on growth. Interestingly, total *Vibrio*

bar, *R*<sup>2</sup> was non-significant (i.e., *R*2= 0).

depicted grouped by variable, and then in rank order, with their associated

have explained substantial variance (*R*<sup>2</sup> <sup>=</sup> 43–54%) in *V. vulnificus* in more instances than for other *Vibrio* species (Pfeffer et al., 2003; Blackwell and Oliver, 2008; Wetz et al., 2008), suggesting they are responding similarly to their environments under the conditions studied. However, instances do occur where total *Vibrio* and *V. vulnificus* do not correlate (Høi et al., 1998; Wetz et al., 2008), underscoring that a species is not a constant component of a genus, and may respond to environmental conditions independently.

Isolations of the three potentially pathogenic species across salinity and temperature gradients were also looked at, and found to exhibit different patterns. *V. cholerae* has a wide temperature range (∼10–30◦C) in brackish water (1–10 ppt), and generally decreases with increasing salinity over the entire range examined (0–40 ppt) (**Figure 5**). Observed *V. cholerae* abundance is greatest around 20◦C and 0–10 ppt, on the order of 10<sup>3</sup> cells per mL. At less-favorable, higher salinities, *V. cholerae* has been found around this temperature, though in much lower abundances (on the order of 1 cell per mL). Interestingly, *V. cholerae*'s realized niche is much smaller than its fundamental one, as it has maximal temperature and salinity tolerances around 38◦C and 75 ppt (Materna et al., 2012), suggesting other controls on its abundance in the environment.

*V. parahaemolyticus* contrasts *V. cholerae* by having a more constant abundance that is broadly spread out over salinities of 3–35 ppt in a narrow, much warmer temperature range, centered roughly around 29◦C. (**Figure 6**). Consistent with this finding, it has been noted that this species prefers warmer waters (>20◦C) (Martinez-Urtaza et al., 2012), and has been observed to grow best at 25◦C *in vitro* (Nishina et al., 2004). However, isolations from shellfish can exhibit different trends from those observed in the water column; Martinez-Urtaza et al. (2008) detected *V. parahaemolyticus* in mussels gathered in much cooler, 15◦C water, consistent with the potential for shellfish to concentrate *V. parahaemolyticus*.

A previous literature-based analysis showed *V. vulnificus* to have a more complicated relationship to temperature and salinity than either *V. cholerae* or *V. parahaemolyticus*. It has a narrow temperature range at higher salinities (>10 ppt) while at low salinities (between 5 and 10 ppt) its temperature range more than doubles—from 22–30◦C to 10–32◦C (Randa et al., 2004). This suggests that, in temperate climates, this species is found yearround in estuarine, low salinity environments but can expand into full strength seawater during warmer months. In the tropics, this species should be endemic to the ocean.

#### **CONCLUSIONS FROM META-ANALYSIS**

From this meta-analysis, we find, first, that temperature and salinity often explain more variance than any other bulk water parameter, like phosphate, nitrogen, pH, or DOC. Yet some of the difficulty in making general statements regarding the relationship of vibrios to individual environmental variables likely stems from the fact that their strength can depend on the ranges examined, e.g., as for temperature, or in quality of the variable, such as DOC, which will encompass carbon derived from different sources that may impact *Vibrio* growth differentially. Second, we observe that trends that apply to the whole genus *Vibrio* do not necessarily reflect those of individual species. Total vibrios and the well-studied potential pathogens *V. cholerae*, *V. parahaemolyticus*, and *V. vulnificus* correlate with shared and distinct environmental variables. For *V. parahaemolyticus* and *V. vulnificus*, temperature often explains more variance than does salinity in the same analysis, and for *V. cholerae*, diverse biotic variables, including specific phyto- and zooplankton taxa, can be stronger correlates than abiotic variables. Unfortunately, biotic variables, particularly individual plankton taxa, have rarely been studied in more than one instance, making these observations difficult

studies are plotted against the temperature (◦C) and salinity values (ppt or psu) at which they were found. All studies report *V. cholerae*,

note the breaks are scaled for clearer visualization, and not linearly. (×) indicates no *V. cholerae* found in that sample.

to generalize. But the correlations reviewed above hint that there may be ecological relationships between *Vibrio* and plankton that merit deeper investigation.

Across salinity and temperature gradients, the pattern also differs between total *Vibrio* and individual species, and species' patterns differ from each other. Indeed, differences may occur even within taxonomic species; *V. parahaemolyticus* pathogenic genotypes have been observed to be a variable fraction of total *V. parahaemolyticus* (Zimmerman et al., 2007). For example, at their Alabama site, total *V. parahaemolyticus*—detected via thermolabile hemolysin marker (tlh)—remained at a more constant concentration of between 1 and 10 cells per mL, while toxigenic genotypes—thermolabile hemolysin+ and thermostable direct hemolysin+ cells—fluctuated in a much wider range: between 0.0001 and 10 cells per mL. This result argues against using the total species to infer the potential pathogens. Taken together with the results from the meta-analysis, these findings suggest that finer-scale sampling—of both the environmental parameters and the *Vibrio* population of interest—is necessary to link ecological parameters to cellular abundances.

#### **ASSOCIATIONS WITH COMPLEX AND PARTICULATE MARINE GROWTH SUBSTRATES**

The previous sections demonstrate that, with the exception of temperature and salinity, parameters measured in bulk seawater have shown limited power in explaining the environmental dynamics of *Vibrio* species. This may, in part, be due to the narrow focus on only a few (potentially) pathogenic species, and frequently limited comparability of measured parameters across studies. It is also likely, however, that bulk measurements, such as dissolved oxygen, nitrogen and phosphate concentration in seawater, only poorly capture the ecological parameters that *Vibrio* populations are associated with or respond to. Vibrios are often presumed to primarily attach to biological surfaces, yet may also subsist on dissolved resources of biological origin while free-living. Taking these resource associations into account, their environmental dynamics may be somewhat decoupled from parameters measurable in bulk seawater, and may depend more on the concentration and properties of relevant solid or dissolved resources. We review in the following sections the ample evidence for surface-associated niches, as well as more recent evidence for environmental dynamics including free-living states and formation of blooms.

From the perspective of bacteria attaching to surfaces, these are either metabolically inert or can be degraded as a source of growth substrates. Vibrios have the ability to attach to and degrade a considerable number of polymeric substrates (Johnson, 2013), suggesting that specific association with surfaces is an important growth strategy. For example, nearly all vibrios can metabolize the abundant biopolymer chitin (present in both crustacean and diatom shells in the marine environment) (Hunt et al., 2008a; Grimes et al., 2009), and various representatives can metabolize an array of plant/algal polysaccharides: agar, alginate, fucoidan, mannan, cellulose, pectin, and laminarin (Goecke et al., 2010). In addition, vibrios may metabolize plastic wastes, as suggested by a recent study documenting that vibrios make up the majority of bacteria attached to plastic wastes floating in the ocean, with electron microscopy showing individual cells residing at the bottom of pits (Zettler et al., 2013). Although this suggests that these plastics, which had been thought to be largely biologically inert, could be degraded by vibrios, such activity remains to be confirmed.

Evidence is also accumulating that vibrios may play a role in oil spill degradation: *Vibrio* representatives can metabolize oil-derived compounds (West et al., 1984; Moxley and Schmidt, 2010), and have been found to comprise a sizable fraction of oilassociated microbial communities from the Deepwater Horizon spill, both from sea-surface samples (>31% in the molecular study of Hamdan and Fulmer, 2011) and salt-marsh plants contaminated with oil mousse (57% in the study of Liu and Liu, 2013). While a clear positive effect of crude oil on *Vibrio* growth has yet to be demonstrated *in vitro*, it appears that many vibrios can at least persist in the presence of oil (Stephens et al., 2013). *Vibrio* representatives furthermore show resistance to inhibition by the oil dispersant Corexit (Hamdan and Fulmer, 2011), which was widely used following the Deepwater Horizon spill; this resistance may additionally support an ability to persist after oil spills.

Most associations with specific surfaces have, however, been described for plants, algae, and animals, and the following section explores these organisms as potential biological niches for vibrios.

#### **BIOLOGICAL NICHES FOR** *Vibrio*

*Vibrio* have been detected on a plethora of aquatic biological surfaces, but which of these associations represent more than transient, incidental attachments? In the following sections we consider which aquatic plants (**Table 1**) and animals (**Table 2**) may represent sustained *Vibrio* niches, on the basis of (1) numerical enrichment compared to the surrounding medium, and (2) knowledge of biological mechanisms, e.g., availability of nutrition and shelter, potentially supporting an association. In doing so, we also draw attention to the need for more quantitative and mechanistic approaches to understanding the ecological associations that allow vibrios to flourish—approaches that could underpin more powerful predictions of *Vibrio* dynamics arising from these diverse associations. We note also that many of the following observations are limited to *V. cholerae* because of its prominence as a pathogen, but the same niches may be available to other vibrios with similar biological activities.

#### **ASSOCIATIONS WITH PLANTS**

Vibrio survival is enhanced in association with certain freshwater and estuarine plants (**Table 1**). Plant hosts can provide nutrition (Andrews and Harris, 2000) and the opportunity to form predation-resistant biofilms (Matz et al., 2005), and have been postulated to modulate unfavorably cold temperatures as well (Criminger et al., 2007). Two freshwater aquatic plants have been observed to support both *in situ* enrichment (in freshwater bodies of Bangladesh) and *in vitro* survival advantage for *V. cholerae*: duckweed, *Lemna minor* (Islam et al., 1990b), and water hyacinth, *Eichhornia crassipes* (Spira et al., 1981), with preference for roots of the latter. Concentration on *E. crassipes* roots may indicate that root exudate is a particularly rich nutritional source, but may also be an artifact of the fact that the roots represent the greatest area exposed to water, and hence to inoculation by planktonic *Vibrio*. By contrast, duckweed's minimal structure, lacking stem or developed leaves, means that almost the entire plant is in contact with the water and thus available for inoculation.

Among estuarine plants, nitrogen-fixing representatives of several *Vibrio* taxa—including *V. diazotrophicus*, *V. natriegens*, *V. cininnatiensis* (Urdaci et al., 1988), and *V. parahaemolyticus* (Criminger et al., 2007)—appear to be noteworthy members of the rhizosphere, given that they represent more than half of the culturable diazotrophs associated with the dominant marsh grasses *Spartina* sp. and *Juncus roemerianus* (Bagwell et al., 1998; Larocque et al., 2004), and the herb *Salicornia virginica* (Bergholz et al., 2001; Criminger et al., 2007). While this numerical dominance may reflect culturing bias, later molecular studies of the *S. alterniflora* rhizosphere confirmed that vibrios (not taxonomically resolved below the level of the family) are stable constituents of the community (Lovell et al., 2008), with little seasonal fluctuation (Gamble et al., 2010). Nitrogen fixation thus appears to be an effective strategy supporting *Vibrio* survival in the anaerobic rhizosphere, demonstrating the ecological breadth granted by vibrios' facultatively anaerobic metabolism.

#### **ASSOCIATIONS WITH MICROALGAE AND FILAMENTOUS CYANOBACTERIA**

While early culture-based studies have demonstrated numerical dominance of vibrios on phytoplankton surfaces compared to surrounding water, e.g., Simidu et al. (1971), little is known about direct, physical associations with specific phytoplankton. Algal cells represent a nutritional opportunity in that they often excrete a high proportion of their photosynthetically fixed carbon, thereby creating a diffusive sphere (the phycosphere) around them, with elevated organic carbon concentration compared to the bulk (Paerl and Pinckney, 1996). However, *in vitro* survival advantage and persistence have been thus far been demonstrated only for *V. cholerae* in physical association with two microalgae: with the filamentous freshwater green alga *Rhizoclonium fontanum* (Islam et al., 1989), and inside the mucilaginous sheath of *Anabaena* sp. cyanobacteria under both freshwater (Islam et al., 1990a, 1999) and saline conditions (Ferdous, 2009) (**Table 1**).

Recent work has illuminated mechanistic details of the *V. cholerae* association with *Anabaena*, which may follow the canonical model of symbioses between heterotrophic bacteria and nitrogen-fixing freshwater cyanobacteria. In such associations, heterotrophs locate their hosts via chemotaxis and benefit from rich cyanobacterial exudate (Paerl and Gallucci, 1985). In return, their oxidative metabolism both relieves oxygen inhibition of nitrogen fixation (which would otherwise limit rapid algal growth), and generates carbon dioxide for photosynthetic assimilation (Paerl and Gallucci, 1985). For *V. cholerae*, chemotactic preference for components of the Anabaena mucilaginous sheath has been demonstrated (Mizanur et al., 2002). Furthermore, investigators have shown that both chemotaxis to and survival on *Anabaena* depend on *V. cholerae*'s expression of mucinase (Islam et al., 2002, 2006). The exact role of mucinase has yet to be defined, but activity of secreted mucinase might liberate from mucus the relevant chemotactic attractants, aid colonizing *Vibrio* in physical penetration of the mucilage, and/or convert mucilage to nutritive compounds supplementary to the cyanobacterial exudate.

#### **ASSOCIATIONS WITH MACROALGAE**

Numerous studies have shown that vibrios are one of the most abundant culturable constituents of macroalgal communities (**Table 1**): a recent meta-analysis of 161, predominantly culture-dependent macroalgal-bacterial studies determined that vibrios on average comprised 10% of these communities (Hollants et al., 2013), with 28, 28, and 44% of them found on brown, green, and red macroalgae, respectively. While no molecular studies have yet quantified *Vibrio* within macroalgal


**Table 1 | Plant and algae hosts for vibrio, as demonstrated by numerical enrichment and biological mechanisms supporting association.**

*(Continued)*

**Table 1 | Continued**


communities, numerical enrichment of culturable vibrios has been demonstrated for the brown algae *Ascophyllum nodosum* (Chan and McManus, 1969), and *Laminaria longicruris* (Laycock, 1974); the red algae *Hypnea* sp. (Lakshmanaperumalsamy and Purushothaman, 1982), *Polysiphonia lanosa* (Chan and McManus, 1969), and *Porphyra yezoensis* (Duan et al., 1995); and the green algae *Chaetomorpha* sp. (Lakshmanaperumalsamy and Purushothaman, 1982), *Enteromorpha* sp.


**Table 2 | Animal hosts for vibrio, as demonstrated by numerical enrichment and biological mechanisms supporting association.**

*(Continued)*

#### **Table 2 | Continued**


**Table 2 | Continued**


(Lakshmanaperumalsamy and Purushothaman, 1982), and *Ulva pertusa* (Duan et al., 1995). For *V. cholerae*, *in vitro* survival advantage has been shown on the green algae *Ulva lactuca* and *Enteromorpha intestinalis* and the red alga *Polysiphonia lanosa* (Islam et al., 1988).

As mentioned above, vibrios can metabolize many algal polysaccharides; they have furthermore been implicated in several other biological activities facilitating symbiosis with macroalgal hosts. These include antagonism directed toward potential bacterial or algal competitors for host surface area (Dobretsov and Qian, 2002; Kanagasabhapathy et al., 2008), developmental morphogenic effects on *Ulva pertusa* (Nakanishi et al., 1996), and stimulation of spore germination for *Ulva* sp. (Patel et al., 2003; Tait et al., 2005). Hence multiple lines of evidence point to significant *Vibrio* association with *Ulva* sp. (enrichment, survival, morphogenesis and spore modulation) and *Polysiphonia* sp. (enrichment, survival) in particular.

#### **ASSOCIATIONS WITH ANIMALS**

*Vibrio* interactions with animals include both specific, stable symbioses, and less well-defined associations (**Table 2**). Stable symbioses have been described for luminescent *V. fischeri* (*Aliivibrio*) with sepiolid squids (*Euprymna scolopes*) and loligonoid squids (Ruby and Lee, 1998), and for various luminescent *Vibrio* with flashlight fishes (Anamalopidae) and anglerfishes (Ceratioidei) (Haygood and Distel, 1993). The dynamics of the *V. fischeri*-*Euprymna symbiosis* have been particularly well-explicated: *V. fischeri* from surrounding waters colonize the developing squid light organ, successfully outcompeting non-symbionts in this process, which triggers a developmental program in the host. Once established, the symbionts undergo daily cycles of expulsion and regrowth (Ruby and Lee, 1998; Stabb, 2006). Thus the symbiosis regularly seeds the water column, such that luminous *V. fischeri* are enriched in the water surrounding *E. scolopes* (Ruby and Lee, 1998). This expedites continual recolonization of immature squid, which is likely further facilitated by *V. fischeri* chemotaxis toward squid mucus (DeLoney-Marino et al., 2003).

Some *Vibrio* have also been deemed facultative intracellular symbionts of *Acanthamoeba* protozoa: *Vibrio cholerae* O1 and O139, and *Vibrio mimicus* (Abd et al., 2005, 2007, 2010; Sandström et al., 2010). These vibrios can replicate intracellularly for at least 14 days without affecting host health, at least in nutrient-replete artificial medium, and have been observed in both cytoplasm and cysts of the protozoa. Like several other microbial taxa, then, most famously the pathogen *Legionella* (Rowbotham, 1980), vibrios appear capable of evading *Acanthamoeba* endocytosis to shelter intracellularly. Thus they gain protection from antibiotics (Abd et al., 2005, 2007, 2010), predation, and perhaps other adverse conditions, e.g., cold temperatures. Still to be investigated are the questions of why some *Acanthamoeba* cells encyst their *Vibrio* inhabitants while others do not; why the *Vibrio* do not appear to be detrimental to host survival; and how often *Vibrio* might be released following host lysis, or even actively ejected, thus returning to the water column. Moreover, all studies of the *Vibrio*-*Acanthamoeba* relationship have been experimental: *in situ* surveys are necessary to establish the environmental relevance of this potential symbiosis, and assess any effects on *Vibrio* population dynamics.

Vibrios may be neutral or benign inhabitants of coral hosts: they have been shown to comprise a significant portion of the mucus-dwelling bacterial community of healthy corals (e.g., Koren and Rosenberg, 2006; Kvennefors et al., 2010), being able to subsist on coral mucus as their sole carbon and nitrogen source (Sharon and Rosenberg, 2008). *V. splendidus*, for example, constituted 50–68% of clone libraries derived from *Oculina patagonica* coral mucus, but was scarce in the coral tissue itself (Koren and Rosenberg, 2006). Moreover, nitrogen-fixing *Vibrio* representatives, primarily *V. harveyi* and *V. alginolyticus*, have been found to dominate the culturable diazotrophs of the coral *Mussimilia hispida* (Chimetto et al., 2008), and likely share fixed nitrogen with either or both coral and zooxanthellae. Evidence also suggests immune interaction between *Vibrio* and coral hosts: adaptation of *Vibrio* commensals to coral antimicrobials has been suggested by significant antibiotic-resistance gene cassette content of their integrons (Koenig et al., 2011), while one *V. harveyi* coral isolate has been found to help defend its host by inhibiting colonization by a pathogen (Krediet et al., 2013).

In freshwater habitats, *V. cholerae* have been found to proliferate on egg masses of the abundant, widely distributed chironomid midges (Broza and Halpern, 2001; Halpern et al., 2008). These egg masses are embedded in thick, gelatinous material, which *V. cholerae* can use as a sole carbon source (Broza and Halpern, 2001); their degradation of the gelatinous matrix via secreted hemagglutinin/protease appears to be the primary cause of egg mass disintegration (Halpern et al., 2003). Accordingly, Halpern et al. (2006) were able to show correlations of chironomid egg mass with the abundance of attached *V. cholerae*, although they have not yet investigated any correlation of *V. cholerae* dynamics in the surrounding aquatic environment.

Zooplankton, primarily estuarine copepods such as *Acartia* and *Eurytemora*, have been investigated as a major reservoir of *V. cholerae* in particular, but while attachment has been demonstrated, it remains unclear whether the association is specific, and whether attached vibrios are consistently enriched compared to surrounding waters. Individual copepods have been shown to be able to host up to 10<sup>5</sup> *V. cholerae* cells (Colwell, 1996; Mueller et al., 2007), with preference often shown for attachment to the oral region and egg sac (next to the anal pore)—that is, regions offering close access to host exudates (Huq et al., 1983, 1984). Culture-based studies have detected enriched *Vibrio* occurrence on copepods compared to the surrounding water column (e.g., Simidu et al., 1971; Sochard et al., 1979), and one culture-based study showed *Vibrio* dominance of wild copepods' surface- and gut-attached bacterial communities (Sochard et al., 1979). However, other studies, both *in vitro* and *in situ*, have observed *V. cholerae* remaining predominantly free-living in the presence of copepods (Worden et al., 2006; Neogi et al., 2012) or attaching with greater preference to phytoplankton (Tamplin et al., 1990). Additionally, one culture-independent environmental study detected greater concentrations of Vibrio, including *V. cholerae*, in water compared to zooplankton (Heidelberg et al., 2002a,b). Perhaps such variability of association with copepods helps explain the difficulty in detecting correlated *Vibrio*-copepod dynamics, as mentioned above in the section Environmental Correlates of *Vibrio* Presence and Abundance.

Other uncertainties regarding *Vibrio* association with copepods exist. There is a lack of quantitative evidence demonstrating long-term proliferation of copepod-attached *Vibrio*: existing studies assessing survival advantage of *Vibrio* cultured with copepods have only demonstrated increased abundance of *Vibrio* in surrounding water, without monitoring attached abundance (Huq et al., 1983, 1984). Finally, it is not clear whether vibrios prefer colonizing live or dead copepods. While several *in vitro* studies have noted *V. cholerae* attachment preference for dead or detrital copepods (Huq et al., 1990; Tamplin et al., 1990; Mueller et al., 2007), one study instead observed survival advantage only upon association with live copepods, and found little attachment to dead copepods (Huq et al., 1983). Perhaps this question could be resolved by investigating from which part(s) exactly of the copepod vibrios derive nutrition: from oral/anal exudates or gut contents of actively feeding copepods, from degradation of the chitinaceous exoskeleton which for live copepods is protected by a waxy epicuticle that resists attachment (Tarsi and Pruzzo, 1999), or from degradation of other copepod detritus. In addition, variable host traits such as immune defenses, age, and time since molting or death (which likely affect epicuticle condition) should be taken into account. As of yet, evidence of association with live copepods as an ecological specialization has been demonstrated for only one *Vibrio* sp. nov. (F10) (Preheim et al., 2011a).

In addition, zooplankton other than copepods may represent potential *Vibrio* hosts as well. Kirschner et al. (2011) found cladoceran *Diaphanosoma mongolianum* to enhance growth more than the copepod *Arctodiaptomus spinosus* in microcosm experiments; when cladocerans were added, they enhanced the growth of *V. cholerae* strains in the surrounding medium relative to controls where cladocerans were excluded, while copepods did not. In addition, the number of cells attached to cladocerans per individual was on average 100 times higher than on copepods. When a back-of-the-envelope calculation is done to consider whether *V. cholerae* is enriched on zooplankton, however, we find that they are not, even on cladocerans; from six microcosms, 105–107 cells were estimated attached and 106–107 cells not attached, a result suggesting that cladocerans might enhance overall growth with frequent dispersal, rather than supporting exclusively attached growth.

For other animals in which *Vibrio* have been found to be abundant—fish, and shellfish—it has not yet been determined whether vibrios form specific, lasting associations as gut microbiota, or are merely transient occupants, temporarily proliferating on favorable nutrients until excreted or otherwise detached. In marine fish, numerous studies, both culture-dependent and independent, have demonstrated that *Vibrio* are major gut inhabitants, often dominating the community, and hence are substantially enriched compared to surrounding seawater. Surveyed fish include flatfish (Liston, 1957; Xing et al., 2013), jackmackerel (Aiso et al., 1968), bluefish (Newman et al., 1972), salmonids (Yoshimizu and Kimura, 1976), sea bream (Muroga et al., 1987), and various coral reef fishes (Sutton and Clements, 1988; Smriga et al., 2010). Notably, *Vibrio* abundances often appear comparable between culture-based and -independent studies: e.g., 35–74 and 83.4%, respectively, of flatfish inhabitants (Liston, 1957; Xing et al., 2013). The ability of *Vibrio* representatives to resist low pH and bile supports their survival within the fish gut (Yoshimizu and Kimura, 1976). Whether food or water intake is the greater source of inoculation is an open question: some studies have found a strong effect of food source on gut *Vibrio* composition (e.g., Grisez et al., 1997), whereas others found a stronger influence of *Vibrio* representation in the water column (e.g., Blanch et al., 2009). Conversely, *Vibrio* content of the fish gut has also been shown to be responsible for increasing *Vibrio* abundance in surrounding water when fish were introduced into a tank that did not otherwise support *Vibrio* growth, demonstrating significant excretion of viable cells from the fish gut (Sugita et al., 1985). Hence, regardless of length of association, the fish gut appears to represent a favorable refuge where *Vibrio* can rapidly proliferate, prior to being released again to the water column. Indeed, the bioluminescence of marine microbes, including many vibrios, has been suggested to be an adaptation encouraging fish ingestion: fish preferentially predate zooplankton that are glowing after having grazed bioluminescent *Photobacterium* (Zarubin et al., 2012).

Among shellfish, high *Vibrio* abundance has been reported on surfaces and in tissues of hosts including oysters (e.g., Murphree and Tamplin, 1995; Froelich and Oliver, 2013), abalone (Sawabe, 2006), and blue crabs (Davis and Sizemore, 1982), with uptake and population dynamics particularly well-documented for *V. vulnificus* in association with oysters (Froelich and Oliver, 2013). *V. haliotis* has been suggested to stably associate with gut of the herbivorous *Haliotis* abalone on the basis of reproducibly specific occurrence: it has never been isolated from other seaweedconsuming invertebrates (reviewed in Sawabe, 2006). Being alginolytic, *V. haliotis* has also been suggested to aid its host's digestion of algal polysaccharides (Sawabe, 2006). Otherwise, it is not clear whether copious *Vibrio* representation might solely be the result of non-specific uptake from food or water, particularly for filter-feeding shellfish, whose highly efficient filtration has been reported to increase *Vibrio* concentrations by up to 4 orders of magnitude in oysters compared to surrounding waters (Froelich and Oliver, 2013). Furthermore, filter feeders can produce copious amounts of mucus, which rapidly and efficiently removes associated microbes, so that their turnover may be high. Consequently, it is challenging to prove specific association on the basis of abundance. In the next section, we will review a metapopulation study that more explicitly addresses the problem of assessing *Vibrio* host specificity by analyzing population structure across and within macroinvertebrate hosts. Future application of the approach described could help to resolve the question of whether *Vibrio* colonization of animal hosts like fish and crabs is specific, or driven more by indiscriminate uptake from the water column.

#### **POPULATION DYNAMICS ASSOCIATED WITH MACROINVERTEBRATE HOSTS**

In a metapopulation study by Preheim et al. (2011a), relative abundances of *Vibrio* groups were compared across different shellfish and parts of shellfish. The study found that macroinvertebrates do not appear to be a strongly selective habitat for vibrios, when contrasted to preceding metapopulation studies of the water column, where differential associations of genotype clusters revealed ecologically distinct populations (described in detail in the section Using Ecology to Define Cohesive Populations). When different body parts of mussels and crabs were sampled by Preheim et al. (2011a), little host preference was evident, and the diversity and frequency of populations (identified by multi-locus sequence analysis) resembled that in water samples. For example, *V. splendidus* represented the dominant population in the water and on both animals. For mussels, which can retain particles when filter feeding (Vahl, 1972), the similarity between water column and animal-associated populations was particularly high, and there appeared to be relatively little difference when gills, stomach and gut walls and contents were compared. This was interpreted as population assembly being largely driven by filter-feeding activity, as was posited in the section above. In contrast to mussels' highly uniform population structure across individual hosts, crabs showed high variance in associated *Vibrio* populations, although composition across individuals' body parts was still similar to that in the water column. What causes the high variance among individual crabs is not known, although there was some evidence suggesting that they may be inoculated by food items, which could be of variable composition given their scavenging lifestyle.

The apparent lack of specificity for the animals was surprising considering that ecological theory predicts that habitats that are long-lived and stable compared to the colonizing species should be dominated by specialists (Kassen, 2002). Yet with regard to mussels and crabs as habitats, vibrios appear to be generalists whose population dynamics may be determined by direct inoculation from the water or via food items (Preheim et al., 2011a). A similar dynamic has recently been suggested to drive *V. vulnificus* accumulation in oysters (Froelich et al., 2010). These can only retain larger particles when filter feeding, and hence enrich pathogenic ecotypes of *V. vulnificus* that are particle-associated as compared to ecotypes that are predominantly free-living.

Overall, these studies demonstrate that colonization may be a complex process strongly influenced by dispersal. In contrast to water column populations, which showed varying degrees of specificity toward microhabitats (e.g., organic particles, zooplankton), *Vibrio* populations on larger invertebrates (mussels and crabs) showed little specificity either for host or host body parts. Whether similar patterns exist for other animals remains unknown; it will be valuable to test fish to determine whether their *Vibrio* inhabitants are true gut microflora. The above studies stress the importance of taking into account potential *Vibrio* sources, i.e., water and food, when assessing host association. For example, *V. splendidus* was the dominant population on both crabs and mussels, and on particles in the water column; had only mussels been sampled, *V. splendidus* may have appeared to have been a mussel specialist. Such erroneous conclusions can be avoided by "mass balancing" populations in a particular location by determining their frequency across different microhabitats or patches that are potentially connected by migration.

#### *VIBRIO* **PROLIFERATION IN THE WATER COLUMN**

Ocean water is a heterogeneous landscape of varying ecological opportunities on small scales, with a highly patchy distribution of resources that may represent microhabitats for vibrios. Some of these are hotspots of soluble organic material, which originates from exudates or excretions of larger organisms, while others are particulates of various origins. For example, as mentioned above, algal cells exude a zone of enriched organic material (Bell and Mitchell, 1972; Paerl and Pinckney, 1996). Several other processes can also generate ephemeral patches of dissolved nutrients, and it is likely that many bacteria, including vibrios, can chemotax toward these and take advantage of the elevated nutrient concentrations (e.g., for vibrios, Sjoblad and Mitchell, 1979; Mizanur et al., 2001, 2002). In addition, diverse processes are responsible for the formation of suspended particulate organic matter that can be colonized and degraded by bacteria. This includes dead biomass of small planktonic organisms, fecal pellets, and aggregates (marine snow) formed from polymers and other, smaller particles.

This section will address two main subjects, both seeking to situate *Vibrio* within the marine water column. Here, we will first review both experimental and environmental evidence that blooms of *Vibrio* can and do occur, despite their typically low representation in marine assemblages. Second, we will review the evidence for proliferation of *Vibrio* in the planktonic, freeliving phase, expanding the view of their niche range beyond the longstanding proposition that their lifestyle is predominantly attached.

#### *VIBRIO* **BLOOMS**

Thompson and Polz (2006) summed up three key *Vibrio* traits supporting the ability to bloom on sporadic nutrient pulses: *Vibrio* can (1) survive long-term under resource-limited conditions, as indicated by continued respiratory activity in mesocosms (Ramaiah et al., 2002; Armada et al., 2003); (2) recover from starvation and grow rapidly in response to substrate pulses, enabled by maintenance of high ribosome content (Hood et al., 1986; Flärdh et al., 1992; Kramer and Singleton, 1992; Eilers et al., 2000; Pernthaler et al., 2001); and (3) actively seek out nutrient patches via chemotaxis (Bassler et al., 1991; Yu et al., 1993), including under starvation conditions (Gosink et al., 2002; Larsen et al., 2004).

*Vibrio* proliferation on natural dissolved resources alone has been experimentally demonstrated by rapid growth of inocula in mesocosms or microcosms of filtered water from algal blooms. *V. cholerae* strain N19691 grew at a rate of up to 2.6 d-1 in dinoflagellate (*Lingulodinium polyedrum*) bloom water (Mouriño-Pérez et al., 2003), and up to 1.73 d-1 in water from a dense picophytoeukaryote and dinoflagellate bloom, surpassing the 0.76 d-1 average growth rate of the separately incubated native bacterial assemblage (Worden et al., 2006).

Experiments have furthermore demonstrated conditions where algal resources were sufficient for *Vibrio* to overcome competition and/or grazing pressure. Taking competition into account, but in the absence of predation, strains of both *V. cholerae* and *V. vulnificus* have been shown capable of increasing in relative abundance when in direct competition with the total bacterial community for filtered homogenate of a cyanobacteria bloom (dominated by *Nodularia spumigena*) (Eiler et al., 2007). Meanwhile, *V. cholerae* N19691 has been shown to overcome substantial protozoan grazing when proliferating on filtrate of a particularly dense algal bloom (Worden et al., 2006). Ample algal dissolved organic material may have permitted this *V. cholerae* growth by relieving resource competition, as the *V. cholerae* inocula grew at the same rate with or without the whole bacterial community filtered out from their bloom-water amendments. Similarly, an analysis of *Vibrio* dynamics sampled from the Arabian Sea suggested that algal resource supply can be a more significant control on *Vibrio* abundance than predation, enabling rapid turnover (Asplund et al., 2011).

Reinforcing these experimental findings, (Gilbert et al., 2012) observed an explosive *Vibrio* bloom in the environment, demonstrating that their potential for rapid growth is indeed relevant in the context of a full marine community. In 1 month, a single *Vibrio* sp., otherwise comprising only 0–2% of total rRNA genes, grew to constitute 54% of the community—the largest bloom of any bacterial group observed over the course of a 6-year time series. Furthermore, there was a correlated bloom of the diatom *Chaetoceros compressus*, itself typically rare within the phytoplankton community. Hence, nutrients exuded by the unusually proliferating diatom taxon may have sparked the *Vibrio* bloom, whether by specifically appealing to the species' metabolic palate, relieving resource competition, diluting protozoan grazing pressure by stimulating rapid growth of the surrounding bacterial community, or some combination of the three. Luminescent *Vibrio* blooming in association with algae have even been suggested to be responsible for the phenomenon dubbed "milky seas," where significant stretches of surface water are rendered white with bioluminescence (Lapota et al., 1988; Nealson and Hastings, 2006); one recent case was expansive enough (>17,700 km2) to be detectable by satellite. Whether such bloom events are rare remains unknown due to currently infrequent sampling and lack of time series; however, the observations cited above provide evidence that *Vibrio* are capable of rapid growth in the environment.

#### **THE EVIDENCE FOR A PLANKTONIC, FREE-LIVING LIFESTYLE**

The two mesocosm/microcosm studies discussed above (Mouriño-Pérez et al., 2003; Worden et al., 2006), both furnish evidence that vibrios can thrive while free-living. Mouriño-Pérez et al. (2003) demonstrate the ability of a *V. cholerae* strain to flourish purely on dissolved compounds derived from an algal bloom. Even more strikingly, Worden et al. (2006) observed *V. cholerae* N19691 remaining free-living in four out of their five seawater mesocosm experiments: one initiated from non-bloom seawater, and the other three initiated from seawater collected during distinctly different phytoplankton blooms. Notably, in two of these four experiments, *V. cholerae* attachment to cohabiting copepods was assessed and found to be insignificant (e.g., <1 *V. cholerae* cell found per copepod, averaged over a sampling of 10 copepods, in one of the experiments). This stands in contrast to the theory that *V. cholerae* preferentially attach to copepods, as discussed above in the section on animal associations. In the remaining experiment of (Worden et al., 2006), by contrast to the mesocosms in which *Vibrio* remained free-living, the *V. cholerae* inoculum was initially almost entirely free-living, but, as bloom decay progressed and algal detrital particles increased in size, the population became almost entirely particle-attached, presumably in response to nutrient limitation.

The factors determining whether *Vibrio* remain free-living versus particle-attached are still unknown, but both environmental and genetic determinants could come into play. Past studies have demonstrated effects of temperature, pH, ion concentration, and starvation state (Hood and Winter, 1997); salinity (Kumazawa et al., 1991; Hsieh et al., 2007); and growth-stage-dependent chitin content of diatom cell walls (Frischkorn et al., 2013) on *Vibrio* attachment. Perhaps encounters with relevant biological compounds, e.g., a specific algal cell wall component or polysaccharide, might also trigger lifestyle changes. Even less is known about the genetic mechanisms, diversity, and dynamics underlying vibrio lifestyle association; this remains a rich field of inquiry. For example, Shapiro et al. (2012) recently discovered genomic patterns underlying the ongoing ecological differentiation of two *V. cyclitrophicus* populations: the population with preference for association with larger particles possessed genes for attachment and biofilm formation that were absent from the preferentially free-living population. Such evidence of genetic bases for habitat specificity will provide invaluable insights into selective pressures exerted by different marine microhabitats.

The findings described above suggest great flexibility in *Vibrio* lifestyle, permitting many lines of attack on marine substrates, with different ecological implications for vibrios' dynamics in the water column. For example, biofilm attachment on particulate resources can decrease susceptibility to protozoan predation (Matz et al., 2005), while association with larger particles might increase probability of ingestion by macrofaunal predators, which could in turn facilitate rapid proliferation and dispersal, as discussed above in the section on fish associations. Given vibrios' possibilities for rapid growth and association with diverse marine niches and resources, their impacts on marine nutrient cycling and trophic structure might be much greater than previously believed. Understanding their dynamics will help to elucidate these fundamental marine processes, as well as *Vibrio*-specific models of pathogen persistence and transmission.

#### **USING ECOLOGY TO DEFINE COHESIVE POPULATIONS**

The studies summarized above suggest potential for association of vibrios with plants, algae, and animals as well as growth response to specific classes of particulate and dissolved organic matter; however, they have targeted primarily a single, taxonomically defined species, leaving several important questions unanswered. First, do such taxonomic species correspond to ecologically cohesive units, i.e., do they comprise several ecologically distinct populations or should they be merged with others to form one ecologically cohesive population? Second, if we can define such populations, do these partition resources or compete with each other? Finally, are vibrios primarily ecological generalists or specialists?

A series of studies explored to what extent ecologically coherent groups of vibrios could be distinguished by determining the distribution patterns of genotypes among different potential microhabitats in the coastal ocean (Hunt et al., 2008a; Preheim et al., 2011a,b; Szabo et al., 2013). Initially, this was done by isolation of vibrios from four consecutive size fractionations of ocean water, collected in the spring and fall, to distinguish free-living and attached genotypes (Hunt et al., 2008a). The rationale of this sampling scheme was that different types of microhabitats (e.g., organic particles of various origin, zoo- and phytoplankton) have characteristic size spectra and hence will be enriched in a specific size fraction. Consequently, bacteria specifically associated with a microhabitat should be enriched in the same specific size fraction. Further, because ecological associations may evolve on faster time scales than rRNA genes, isolates were also characterized at higher genotypic resolution using several protein coding genes in a multilocus sequence analysis (MLSA) scheme, to better capture the ecoevolutionary dynamics of environmental populations. Because of the complexity of the data, a statistical clustering algorithm (AdaptML) was developed that allows identification of groups of related genotypes with distinct and characteristic distributions among the sampled parameters (size fractions and seasons) (Hunt et al., 2008a).

The analysis of >1000 isolates identified a large number of genotypic clusters with clear microenvironmental preferences, consistent with the notion of an ecological population (Hunt et al., 2008a). Seasonal differentiation was particularly strong, with little overlap between spring and fall samples, supporting the observed significant correlation of some species to temperature discussed in above sections. The study also revealed that several populations appear free-living or predominantly freeliving, again supporting the notion that vibrios can pursue, at least temporarily (e.g., during a bloom), unattached lifestyles. Most populations, however, displayed various preferences for size fractions enriched in different types of organic particles or zooand phytoplankton. For example, *V. calviensis* appeared almost entirely free-living, while *V. alginolyticus* had significant representation in both the free-living and large-particle fractions, and *V. fischeri* occurred on small and large particle size fractions. Most strikingly, *V. splendidus* was broken up into several, very closely related populations with distinct distributions. Overall, 25 distinct populations could be identified in the two seasonal samplings. (Hunt et al., 2008a), demonstrating the fine-scale resource partitioning co-existing vibrios are engaged in.

To what extent does the commonly used rRNA marker gene resolve these populations? The *V. splendidus* example and several others demonstrate that at least some ecologically distinct genotypic clusters may not be resolved by rRNA analysis and do require high resolution protein-coding genes to identify genotypic clusters whose environmental distributions can be assessed (Preheim et al., 2011b; Shapiro et al., 2012). Most populations, however, were manifest as deeply divergent protein-coding gene clusters (Hunt et al., 2008a) that correspond to microdiverse rRNA gene clusters previously postulated to represent ecological populations (Acinas et al., 2004). Although overall reassuring for rRNA gene-based environmental surveys, variable performance of marker genes is expected since they are slowly evolving and may not capture populations at early stages of divergence (Shapiro et al., 2012).

Additional studies carried out at the same coastal site refined the habitat resolution for several populations, allowed identification of ecological generalists and specialists, and also demonstrated reproducible associations (Preheim et al., 2011a,b; Szabo et al., 2013). The actual microhabitat of several attached populations was identified by hand-picking under the microscope visually identifiable types of particles and zooplankton (Preheim et al., 2011a,b). This revealed high habitat specificity for several populations while others occurred more broadly, indicating different levels of ecological specialization. For example, *V. breoganii* occurred on algal derived detritus while a not yet formally described species (*Vibrio* F10) was highly specific for living zooplankton. On the other hand*, V. crassostreae* was associated with both zooplankton and algal detritus. Metabolic potential in these species, measured by growth assays and comparative genomics, reflects these associations. Both *V. breoganii* and *V. crassostreae* are able to exploit alginate, a brown algal cell wall component, as the sole carbon source, yet the algae-associated *V. breoganii* has acquired the ability to grow on the algal storage polysaccharide laminarin but has lost the ability to grow on chitin, a trait ancestral to vibrios (Hunt et al., 2008b). Moreover, such high specificity for algal derived material was unexpected for vibrios, which are reputed to be animal associated, and supports the evidence provided above that vibrios encompass algal specialists.

A recent study that attempted to reproduce the original size fractionation of ocean water collected at a similar time point, but 3 years after the initial sampling, showed that population structure was preserved for many of the originally detected populations, but also revealed populations as dynamic and environmentally responsive entities (Szabo et al., 2013). For example, *V. breoganii*, *V. crassostreae*, and *V. splendidus*, which range in ecological specialization from specialist to generalist, had highly reproducible distributions indicative of similar habitat associations. The study, however, also showed that several populations were nearly absent in either of the samplings, possibly due to the lower frequency of their habitat in the water samples. Moreover, some populations had shifted distributions among the size fractions. This was the case for a recently diverged population of *V. cyclitrophicus* that was associated with larger particles or organisms in the first study, but was highly represented in the free-living fraction in the second sampling. It was hypothesized that this shift represented a population expansion following a diatom bloom because the relative frequency of *V. cyclitrophicus* increased coincident with a shift from a copepod- to a diatom-dominated eukaryotic plankton community. Similarly, bloom dynamics, as have previously been observed for total vibrios in the water column, may cause the variable representation of several additional populations. Overall, the comparison of the two studies supports highly predictable population-habitat linkage but also provides additional support for the notion that vibrios may be subject to rapid population expansions or blooms in response to often overlooked or unknown environmental factors.

#### **POPULATIONS AS ECOLOGICAL, GENETIC, AND SOCIAL UNITS**

Populations as defined here are genotypic clusters (evident by MLSA) that act as ecologically cohesive units, i.e., their ecology is more similar within the cluster than between. Defining populations in this way has afforded the opportunity to test the hypothesis that, akin to sexual eukaryotes, gene flow boundaries across such clusters are strong enough for adaptive genes or alleles to spread in a population-specific manner. A population genomic analysis of two very recently diverged populations of *V. cyclitrophicus*, which are ecologically distinct but remain >99% similar in average nucleotide composition across their genomes, showed that specific genome regions have swept each of the two populations (recently reviewed by Polz et al., 2013). Moreover, annotation of these genome regions as well as behavioral and growth analysis suggest that these genome regions are adaptive for differential lifestyles (Shapiro et al., 2012; Yawata et al., 2014).

A second study showed that ecologically defined populations may also act as social units. This was evident in a test for potential of antagonistic interactions mediated by antibiotics between individuals from different ecological populations of vibrios (Cordero et al., 2012). Because of higher niche overlap among close relatives, it was expected that antagonism be more advantageous if directed against members of the same population. In stark contrast, however, antagonism was primarily directed against members of other populations while members of the same population were resistant to antibiotics produced within their own populations. This suggests synergism on the population level, especially since multiple antibiotics were produced within each population but each only by relatively few members.

Overall, this research shows agreement between ecological, genetic, and social population structure and suggests that, in many ways, populations can be regarded as species-like units in the wild. Importantly, these units are non-clonal, and their genetic exchange and social structure suggest that populations frequently coexist and re-assemble on small-scale habitats.

#### **CONCLUSION**

In this review, we examine what is known about *Vibrio* ecology at increasingly fine environmental and taxonomic scales, to reveal factors with potential for greater predictive and explanatory power for *Vibrio* dynamics.

We find that while bulk environmental variables are often inconsistent in their ability to explain variance in *Vibrio* abundances, at both the genus and species levels, temperature and salinity are usually the strongest abiotic correlates. Yet total *Vibrio* trends do not necessarily capture species-level trends, and thus it is necessary to monitor populations of interest directly to capture their dynamics. Correlations of species to specific plankton like those of *V. cholerae* to dinoflagellate (Eiler et al., 2006), cladoceran (Kirschner et al., 2011), and rotifer (de Magny et al., 2011) taxa—can provide the bases for hypotheses of biological associations—as was demonstrated by Kirschner et al. (2011) for cladoceran the *D. mongolianum*. Further investigation is necessary to confirm reproducibility and biological significance of such correlations.

Indeed, the breadth of vibrios' metabolic and attachment abilities mean that they can appear quite generalist in their ecological associations, making it difficult to discern which relationships with other organisms are specific and stable, rather than simply the product of promiscuous attachment followed by proliferation. Among the diverse biological associations that we review, some may be true mutualisms, on the basis of vibrios exchanging benefits with their hosts. The symbioses of luminescent vibrios with certain squid and fish are well-attested, while possible symbioses with other organisms are suggested by potentially mutual metabolic exchange (salt marsh plants, cyanobacteria, corals), or *Vibrio* modulation of host processes like development and reproduction (macroalgae), and response to infection (corals). Notably, diazotrophy may facilitate relationships with both marsh plants and corals. In numerous other cases, vibrios may simply be taking advantage of hosts as nutrient sources, and perhaps only temporarily and opportunistically be associated with microalgae, zooplankton, fish, shellfish, and chironomid egg masses, or as intracellular occupants of protozoa. Of these, we argue that evidence points toward a particularly significant ecological impact of *Vibrio* interactions with algae, given the abundant laboratory and environmental observations of vibrios' ability to live on algal exudates -including bloomsasfree-livingcells,ahistoricallyunderappreciated *Vibrio* lifestyle. Nonetheless, much work remains to be done in resolving more specific *Vibrio*-algae associations.

In light of these studies, we have several recommendations. Previous surveys of *Vibrio* abundance are predominantly culturedependent; going forward, molecular methods, such as fluorescent *in-situ* hybridization or quantitative PCR, can be used to gain less biased quantitative data. Such techniques also enable targeting of specific genotypic groupings, allowing better discrimination of pathogenic variants or ecologically meaningful populations than traditional taxonomic assays of species identity. Furthermore, to distinguish specialized association from incidental attachment, a "mass-balanced" approach is necessary: are vibrio enriched on a given microhabitat (e.g., a specific organic particle type or zooplankton) compared to the surrounding water? Or, is the habitat enriched in *Vibrio* compared to other habitats? This approach has provided support for many of the potential symbioses noted above, and enabled identification of specialist *Vibrio* populations, e.g., *V. breoganii* for macroalgaederived material and *V.* F10 for zooplankton (Hunt et al., 2008a; Preheim et al., 2011a,b; Szabo et al., 2013). It provides a strong basis from which to proceed to more detailed and, ideally, mechanistic elucidation of *Vibrio* associations: for example, identifying chemotactic preferences for or proliferation on host or host exudates, or taking advantage of vibrios' genetic tractability to demonstrate dependence of an association on particular metabolic pathways.

When considering the question of to what extent environmental affiliations may be shared among or within *Vibrio* taxa, we also explore the shifting perspective on the nature of microbial groupings: recent work has moved toward discerning ecologically cohesive *Vibrio* populations, rather than relying on named species as the unit of inquiry. Pursuing this approach, whereby habitat associations are mapped onto genotypic clusters, has been successful in identifying ecological, genetic and social units among vibrios in the wild. We stress, however, that the initial identification of environment-genotypic cluster associations by the "mass-balanced" approach outlined above must be treated as a hypothesis of population structure to be further explored by more mechanistic investigation of, for example, dynamic habitat associations, biological interactions and gene flow boundaries. As demonstrated above, this approach has already helped to resolve apparently generalist *Vibrio* taxa into specialized populations and to identify mechanisms of how adaptive genes spread amongst nascent, ecologically differentiated populations. By sampling the environment at fine scales and molecularly characterizing associated *Vibrio*, we will gain a deeper understanding of the ways in which vibrios live in the environment. Such a population-based framework serves as a means of understanding the ecology of microorganisms in general.

#### **AUTHOR CONTRIBUTIONS**

All authors conceived the structure and content, wrote the text as well as edited the manuscript in its entirety.

#### **ACKNOWLEDGMENTS**

This work was supported by grants from the National Science Foundation Evolutionary Ecology program (Evolutionary Processes), and the National Science Foundation and National Institutes of Health co-sponsored Woods Hole Center for Oceans and Human Health, the Moore Foundation, and the National Institutes of Health sponsored MIT Environmental Health Science Center (P30-ES002109).

#### **REFERENCES**


distribution of *Vibrio parahaemolyticus* in the Rias of Galicia, Spain. *Appl. Environ. Microbiol.* 74, 265–274. doi: 10.1128/AEM.01307-07


oysters (*Crassostrea plicatula*). *Int. J. Food Microbiol.* 136, 129–132. doi: 10.1016/j.ijfoodmicro.2009.09.012


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 16 November 2013; paper pending published: 27 December 2013; accepted: 20 January 2014; published online: 11 February 2014.*

*Citation: Takemura AF, Chien DM and Polz MF (2014) Associations and dynamics of Vibrionaceae in the environment, from the genus to the population level. Front. Microbiol. 5:38. doi: 10.3389/fmicb.2014.00038*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Takemura, Chien and Polz. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*

### How community ecology can improve our understanding of cholera dynamics

#### *Guillaume Constantin de Magny1\*, Nur A. Hasan2 and Benjamin Roche3*

*<sup>1</sup> Maladies Infectieuses et Vecteurs Ecologie, Génétique, Evolution et Contrôle, Institut de Recherche pour le Développement, UMR 224 IRD-5290 CNRS-UM1-UM2, Montpellier, France*

*<sup>2</sup> Maryland Pathogen Research Institute, University of Maryland, College Park, MD, USA*

*<sup>3</sup> Unité de Modélisation Mathématique et Informatique des Systèmes Complexes, Institut de Recherche pour le Developpement, UMI IRD/UPMC 209, Bondy, France*

#### *Edited by:*

*Daniela Ceccarelli, University of Maryland, USA*

#### *Reviewed by:*

*Ryan J. Newton, University of Wisconsin-Milwaukee, USA Yan Boucher, University of Alberta, Canada*

#### *\*Correspondence:*

*Guillaume Constantin de Magny, Maladies Infectieuses et Vecteurs Ecologie, Génétique, Evolution et Contrôle, Institut de Recherche pour le Développement, UMR 224 IRD-5290 CNRS-UM1-UM2, 911 avenue Agropolis, BP 64501 Montpellier, France e-mail: guillaume.demagny@ird.fr*

Understanding the seasonal emergence and reemergence of cholera is challenging due to the complex dynamics of different protagonists. The abundance of *Vibrio cholerae*, the causative agent of cholera and a natural inhabitant of aquatic environments, fluctuates according to abiotic, and biotic factors. Among the biotic factors, the zooplankton community dynamics has been suggested to play a pivotal role in the survival, persistence, and natural competence of *V. cholerae*. However, factors regulating *V. cholerae* population structure and seasonal dynamics are still not fully understood. Investigation of the temporal shifts and variability in aquatic community composition in relation to the occurrence or abundance of *V. cholerae* appears very promising yet remained underexplored. Recent advances in metagenomics, facilitated by high-throughput ultra deep sequencing, have greatly improved our ability for a broader and deeper exploration of microbial communities including an understanding of community structure, function, as well as inter- and intra-specific competitions. Here, we discuss possible areas of research focusing how combination of community ecology and metagenomic approaches could be applied to study the cholera system.

#### **Keywords: cholera, ecology, metagenomics, genetics, pathogen**

The diarrheal disease, cholera, is caused by the pathogenic bacteria *Vibrio cholerae*; affecting hundreds of thousands of people every year with thousands of deaths (World Health Organization, 2010). Depending on the health status, infection occurs due to the ingestion of 103–109 cells of pathogenic *V. cholerae* (either O1 or O139) through contaminated food and water. After the ingestion, the microbe has to pass through acidic conditions of the human stomach and the immune system defenses to reach the small intestine, where it attaches, and begins to produce cholera toxin (CT). Sufficient quantities of CT cause severe diarrhea and shedding of the pathogen, up to 107 per infected individual per stool (Zahid et al., 2008), into the environment facilitating spread of the disease.

Outside the human host, *V. cholerae* is an autochthonous member of natural aquatic environments such as lakes, rivers, estuaries, and the ocean, which serve as the principal natural reservoir for this organism. Ecologically, vibrios play an important role in the degradation of organic matter and act as a link that transfers dissolved organic carbon to higher trophic levels of the marine food web (Grossart et al., 2005). *V. cholerae* is a facultative anaerobic, asporogenous, gram-negative rod with capability of respiratory and fermentative metabolism. It is oxidase positive, reduces nitrate, and motile by a single polar, sheathed flagellum (Kaper et al., 1995). The pathogenic and non-pathogenic strains of *V. cholerae* are globally distributed in aquatic environments (Lipp et al., 2002). Historically, cholera is endemic in the Bengal basin with outbreaks occurring elsewhere due to lack of access to clean

water and sanitation facilities (World Health Organization, 2010). The role of climate and other abiotic factors on cholera has been well investigated (Lipp et al., 2002) but approaches taken to date to study *V. cholerae* have fallen short in separating the role of abiotic and biotic factors. Therefore, we are advocating the need for both genomics and community ecology approaches to elucidate such relationships.

How *V. cholerae* survives in two very distinct habitats, human hosts, and aquatic ecosystems, and mechanisms by which it periodically emerges as a human pathogen are compelling questions in the ecology of this species. Understanding the ecology of *V. cholerae* is, however, challenging because of the complex dynamics of different protagonists. In this paper, we will explore how knowledge from multiple different disciplines like ecology, genomics, and modeling, might contribute to establish an ecological framework of*V. cholerae*to provide critical insights in our understanding of this bacterium.

A widely accepted concept in theoretical ecology of *V. cholerae* is *the niche theory*. The niche theory is defined by three main factors that contribute to population growth rate: resources, the physical environment, and natural enemies (Vandermeer, 1972). How a species respond to these factors, spatially, and temporarily, determines its ability to invade, persist, and occasionally become dominant. This appears as a critical phenomenon in understanding the ecology of *V. cholerae*.

The firstfactor is resources. For*V. cholerae*, thisfactor is strongly linked to the availability of chitin in the natural environment. In the aquatic ecosystem, chitin is the most abundant polysaccharide and the principal component of zooplankton exoskeleton. *V. cholerae* is strongly associated with plankton, forming commensal relationships with chitinous organisms that are dominant among zooplankton populations (i.e., copepods, amphipods, and other crustaceans; Colwell and Huq, 1994). The copepod exoskeleton has been shown to support large population of vibrios, including the pathogenic *V. cholerae* (Colwell et al., 1981; Huq et al., 1983; Tamplin et al., 1990). An advantage of epibiotic organisms, such as *V. cholerae*, on biotic substrates, is their proximity to available nutrients. *V. cholerae* cells living on highly mobile zooplanktons are less likely to be nutrient limited than"free-living"or planktonic cells in the environment.

The commensal relationship between vibrios and chitinous organisms, like copepods, has important consequences (Pruzzo et al., 2008). This relationship represents a useful model to investigate the role of primary habitat selection in developing pathogenicity traits of the bacteria primarily inhabiting the aquatic environment. It has also been reported that *V. cholerae* becomes naturally competent and amenable to serotype conversion on chitin surfaces (Meibom et al., 2005). Hence, the ability of chitin to support both genetic evolutionary diversity and growth of cholera organisms in the environment strongly implicates copepods as a critical component in the environmental lifestyle of this human pathogen. Indeed, zooplanktons comprise a broad assortment of ecologically important heterotrophic groups. The composition of zooplankton community changes constantly during an annual cycle, with dramatic differences in species composition of freshwater and marine zooplankton communities. Thus, the plankton species composition plays a pivotal role in *Vibrio* seasonality, as seen, i.e., in waters off the Georgia (U.S.) coast, highlighting the complex relationship between seasonal shifts in plankton composition and number of vibrios in the aquatic environment (Turner et al., 2009).

Statistical model parameter estimates have indicated a strong direct relationship between *Vibrio* concentration and the relative abundance of copepods in the >200 μm fraction. Recently, the incidence of cholera and the occurrence of pathogenic *V. cholerae* with diverse zooplankton taxa were studied in rural areas of Bangladesh (Constantin de Magny et al., 2011). Chitinous zooplankton communities of several water bodies were analyzed in order to understand the interaction of the zooplankton population composition with the population dynamics of pathogenic *V. cholerae* and the incidence of cholera. Two dominant zooplankton groups, namely, rotifers and cladocerans, were consistently associated with detection of *V. cholerae* and/or occurrence of cholera cases. Specifically, the presence of rotifer, *Brachionus angularis,* was significantly associated with the presence of pathogenic *V. cholerae.* Local differences indicated some subtle ecological factors influencing the interactions between *V. cholerae*, its plankton hosts, and incidence of cholera. Like the determination of a potential keystone species (Mills and Doak, 1993), or assemblage of species associated with the occurrence of *V. cholerae*, understanding the community dynamics (diversity and abundance) of zooplanktons in the natural aquatic environment could be a key element in deciphering the complete picture of *V. cholerae* ecology.

The second factor of the niche theory for*V. cholerae* is the physical environment. For most*Vibrio* sp., temperature and salinity are two major abiotic factors strongly associated with their abundance (Lipp et al., 2002). Materna et al. (2012) remarkably investigated the shape and evolution of the fundamental niche of marine *Vibrio* by combining theoretical and experimental approaches on 2-dimensional niche shape associated with temperature and salinity, the two major factors in *Vibrio* ecology. It was found that for the studied marine *Vibrio* strains, ecological range, in reality, occupies a limited section of their potential niche (Materna et al., 2012). In addition to the physical drivers of the population growth, this bacterial species, like other gram-negative bacteria, has a selective advantage in its ability to enter a dormant stage when environmental conditions are unfavorable for active growth and cell division. This state is termed the viable but nonculturable (VBNC) state (Cottingham et al., 2003; Takeda, 2011). This represents another important aspect of the ecological niche of *V. cholerae,* namely its adaptability to fluctuations of the abiotic conditions.

The VBNC state of bacteria is defined as the state where cells remain viable but cease to grow or divide, on, or in, routinely used bacteriological media. Even if it is still in debate, evidences of the reversibility of this state back to active growth and cell division have accumulated (Senoh et al., 2010; Takeda, 2011). This nonculturability associated with the ability to aggregate on biofilm, even after having been non-culturable for more than one year (Alam et al., 2007), is another important aspect in the ecology of *V. cholerae.* Several studies have already demonstrated the importance of the VBNC state in the seasonal epidemics of cholera (Colwell et al., 1985, 1996; Alam et al., 2006a,b; Alam et al., 2007; Asakura et al., 2007). A recent study (Mishra et al., 2012), however, demonstrated a vital role of the VBNC state in cholera epidemiology manifesting an active expression of traits necessary for both viability, stress response, virulence, and colonization of the bacterium.

Yet another factor of the ecological niche of *V. cholerae* is the biotic antagonists. Among them, the lysogenic filamentous phage, CTX-, is found in toxigenic *V. cholerae* (Waldor and Mekalanos, 1996) and encodes the CT. The*V. cholerae* pathogenicity island, VPI, another antagonist, encodes a toxin co-regulated pilus that functions as a colonization factor, and a receptor for CTX-. It has been hypothesized that lytic bacterial viruses are responsible for the decline in the incidence of cholera during seasonal outbreaks (Faruque et al., 2005b), limiting *V. cholerae* population densities through predation. It has been suggested that the increased environmental phage abundance is the result of a host-mediated phage amplification during the cholera epidemic, decreasing the load of environmental *V. cholerae* resulting in the collapse of the epidemic (Faruque et al., 2005a). This predator/prey relationship have been debated due to the lack of evidences distinguishing the effect of growth-limiting resources and the lytic activity of the phage restricting the densities of *V. cholerae* population (Wei et al., 2010). The alternative hypothesis states that phage-resistant *V. cholerae* mutants are less fit than wild-type, possessing a number of characteristics that are likely to further reduce capacity to be maintained in natural populations and may be avirulent in human hosts (Wei et al., 2010). The role of the combined effect of growth resources and the lytic activity of phages on *V. cholerae* population abundances need to be clarified for better understanding of bacterial population dynamics driver.

In addition, it appears that*V. cholerae* does not give up without a fight. *V. cholerae*, like other Gram-negative bacteria, utilizes accessory virulence factors, e.g., the type VI secretion system (T6SS), to provide advantages in intraspecific and interspecific competition (Unterweger et al., 2012). It is hypothesized that the T6SS and toxin translocation play critical roles for *V. cholerae* to outcompete other bacteria and phagocytic cells. Thus, it persists in human hosts and in the environment (Miyata et al., 2010). Furthermore, it was shown that in order to successfully colonize organic matter in the marine environment, *V. cholerae* must compete against a high concentration of diverse bacteria that reside on the surface of particles. The interspecies antagonistic interactions involving allelochemicals can influence *V. cholerae* particle colonizations. These interactions can also be temperature sensitive (Long et al., 2005). However, the definite role of the microbial (viral and bacterial) community on *V. cholerae* abundance is still unknown. As microbes are interacting with each other, at an individual scale (Hajishengallis et al., 2012), or at a population level (Telfer et al., 2010), overlooking such interaction may lead to a partial understanding of the contributing factors. However, microbial communities are typically very diverse and complex*,* therefore, characterization of species diversity and the complex network of species interactions within these communities also pose a technical challenge to the scientific community.

The advent of next-generation sequencing technologies and bioinformatics opened up the door to a wide new era of research exploration and has offered an unprecedented opportunity to fill an existing knowledge gap in community ecology. Recent advances of such technologies and tools have already revolutionized the field of metagenomics, allowing a much more accurate, and detailed description of microbial communities. As these technologies and tools are becoming widely available and accessible, their application is becoming the cornerstone in addressing the outstanding hypotheses in various domains like biogeography as well as furthering our understanding of how ecological communities assemble, evolve, and function. Next-generation high-throughput sequencing (HTS) methods have already been proven as an outstanding way to describe microbial biodiversity and community ecology in a variety of systems and environments. By definition, an individual ecosystem is composed of populations of organisms, interacting within communities, and contributing to the cycling of nutrients, and the flow of energy. Conventional methods of sampling ecosystems are unquestionably biased toward the highly abundant and easy to characterize members of the community where HTS methods facilitate the opportunity to study every member of the community either individually, or as a whole, to reflect the entire microbial system. Additionally, decomposition and nutrient cycling are key processes in ecosystem functioning. Both are fundamental to ecosystem biomass production, and mostly operated by microbes, more precisely bacteria. Understanding microbial community dynamics in ecosystems is, therefore, one of the key elements deciphering ecosystem functioning. As *V. cholerae* survives in two very distinct habitats, human hosts, and aquatic ecosystems, it is, therefore, critical to investigate the influence of microbial population dynamics on seasonal emergence and reemergence of *V. cholerae*, especially to understand what happens in the community when there is an outbreak.

Historically, the first metagenomics research was focused on cultured soil microorganisms (Handelsman et al., 1998). Rapidly, it became a very promising source of knowledge in understanding the dynamics of microbial population, the interaction between pathogens, and the microbial community, and how they evolve in different habitats (e.g., host, vector, reservoir). Recently, Monira et al. (2013) explored gut microbiota of children suffering from cholera and during disease recovery*.* From the nine children included in the study, they reported *V. cholerae* accounted for 35% (minimum 5%, maximum 63%) of the total gut microbiota at day 0, with *V. cholerae* was the predominant species in six of the nine children (Monira et al., 2013). During the recovery, the effect of antibiotic therapy demonstrated a rapid shift in bacterial diversity affecting the top ten bacterial families, which accounted for more than 85% of the bacterial flora during cholera. Furthermore, cholera infection appears to induce expulsion of major commensal bacteria belonging to the phyla, e.g., Bacteroidetes, Firmicutes, and Actinobacteria, facilitating the colonization of the gut by harmful bacteria belonging to the phylum Proteobacteria (Monira et al., 2013).

Does this phenomenon of interspecific competition observed in the human gut during cholera infection also occur in the second distinct habitat, the aquatic ecosystem? Other compelling questions in microbial community ecology also arise from this. What is the structure and composition of the microbial communities in the aquatic environment directly sourced by the human populations for their daily needs? How did these microbial communities evolve over time? What space does *V. cholerae* occupy in the communities and how does it impact the cholera dynamics (emergence, transmission)? How similar microbial communities are in the aquatic environments where cholera is endemic, epidemic, or benign? Ongoing research in Bangladesh based on the limited scale application of 16S rRNA gene survey suggests that the abundance and diversity of the microbial community including *Vibrio* populations exhibited a complex seasonal distribution driven by temporal fluctuations of environmental parameters, such as sea surface temperature, and indicates that a specific consortium among the indigenous population, and their interaction might be crucial in responding to changes in environmental parameters that allow *V. cholerae* to overcome the natural competition and emerge in epidemic proportions.

Translating these capacities of a metagenomics approach to the study of *V. cholerae*, we suggest that such approach could provide crucial insights to infer competition processes within microbial communities.While studies thusfar havefocused mostly on the environmental dimension of *V. cholerae* niche (Preheim et al., 2011a,b; Materna et al., 2012), adopting metagenomics based community ecology approach could add a critical, perhaps keystone, dimension to the study of *V. cholerae*, and the robustness of predictive models could improve substantially by incorporating such a fundamental dimension.

Potentially, the integration of all the knowledge about the disease cholera, and the ecology of *V. cholerae*, should lead to an End-To-End model of cholera. End-to-End models are integrated models of an ecosystem, which encompasses climate, various ecosystem components, and in some cases human interaction with the ecosystem. These models are developed to describe, understand, and predict ecosystem dynamics. To this extent, integrating an ecosystem component in such a model, and the microbial community context involving natural enemies of bacteria that influence strongly the kinetics of *V. cholerae* in the environment and, consequently, in human populations, would be crucial in achieving the objective of such a predictive tool. This is in contrast to the traditional approach or the old paradigm of climate prediction for human health that seeks correlations between climatic variables, morbidity, mortality, exacerbation of chronic conditions, population-level outbreaks, or other indicators that are precursors to an outbreak. Additionally, there is a critical need to plug ecological niche modeling (ENM) into this approach. ENM is often used to inform conservation efforts (Peterson and Robins, 2003), to describe species potential range (Meentemeyer et al., 2004), and is increasingly being used to identify species of public health concern (Mullins et al., 2011). Advances so far in environmental predictions for human health can be illustrated by the excellent analogy for Earth System interactions provided by El Niño. Examples range from heat and cold wave related mortalities and morbidities to cholera, malaria, dengue, and meningitis. The impacts of global changes are already manifesting themselves in global indicators such as temperature and sea level rise, but the impacts on populations occur through local changes in weather, ecology, water resources, barometric pressure, etc.

The salience, reliability, and legitimacy of a successful prediction system will depend on filling the gaps in mechanistic linkages from changes in climate to effects on human health and well-being. Thus, simple empirical approaches of the past must be steered toward more mechanistic methodologies to identify the drivers from all fronts, e.g., environment, agriculture, and water. These include genetic, chemical, and biological factors, such as microbial contamination pathways, human behavior, exposure, social, and health systems infrastructure, demographics, and other measures of vulnerability that affect the onset, spread, and exacerbation of disease, and the resources needed (water, food, shelter) to maintain health. Networked phenomena including the flow and potency of pathogens will be a useful component. In addition to geographic measures, vulnerability, and risk clearly depend on various socio-economic and demographic factors. Regional Earth System Prediction can account for these factors in an adaptive, learningby-doing mode of model-data synthesis. More importantly, these disparate pieces have to be integrated into Earth System models, especially in the high resolution RESP framework (Murtugudde, 2009) with goal-oriented observational monitoring networks.

#### **REFERENCES**

Alam, M., Hasan, N. A., Sadique, A., Bhuiyan, N. A., Ahmed, K. U., Nusrin, S., et al. (2006a). Seasonal cholera caused by *Vibrio cholerae* serogroups O1 and O139 in the coastal aquatic environment of Bangladesh. *Appl. Environ. Microbiol.* 72, 4096–4104. doi: 10.1128/AEM.00066-06


**Conflict of Interest Statement:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

*Received: 15 November 2013; paper pending published: 06 December 2013; accepted: 17 March 2014; published online: 02 April 2014.*

*Citation: Constantin de Magny G, Hasan NA and Roche B (2014) How community ecology can improve our understanding of cholera dynamics. Front. Microbiol. 5:137. doi: 10.3389/fmicb.2014.00137*

*This article was submitted to Aquatic Microbiology, a section of the journal Frontiers in Microbiology.*

*Copyright © 2014 Constantin de Magny, Hasan and Roche. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited andthatthe original publication inthis journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.*