## GLOBAL HEALTH ISSUES OF AFLATOXINS IN FOOD AND AGRICULTURE: CHALLENGES AND OPPORTUNITIES

EDITED BY: Mehdi Razzaghi-Abyaneh, Perng-Kuang Chang, Masoomeh Shams-Ghahfarokhi and Mahendra Rai PUBLISHED IN: Frontiers in Microbiology

#### *Frontiers Copyright Statement*

*© Copyright 2007-2015 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 978-2-88919-544-2 DOI 10.3389/978-2-88919-544-2

#### 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

### **GLOBAL HEALTH ISSUES OF AFLATOXINS IN FOOD AND AGRICULTURE: CHALLENGES AND OPPORTUNITIES**

#### Topic Editors:

**Mehdi Razzaghi-Abyaneh,** Pasteur Institute of Iran, Iran **Perng-Kuang Chang,** Southern Regional Research Center, USA **Masoomeh Shams-Ghahfarokhi,** Tarbiat Modares University, Iran **Mahendra Rai,** Sant Gadge Baba Amravati University, India

Aflatoxins are a group of polyketide mycotoxins that are produced mainly by members of the genus Aspergillus. Production of these toxic secondary metabolites is closely related to fungal development (Keller et al., 2005; Jamali et al., 2012). Contamination of food, feed and agricultural commodities by aflatoxins poses enormous economic and serious health concerns because these chemicals are highly carcinogenic and can directly influence the structure of DNA. The resulting genetic defects can lead to fetal misdevelopment and miscarriages; aflatoxins are also known to suppress immune systems (Razzaghi-Abyaneh et al., 2013). In a global context, aflatoxin contamination is a constant concern between the 35N and 35S latitude where developing countries are mainly situated. With expanding boundaries of developing countries, aflatoxin contamination has become a persistent problem to those emerging areas (Shams-Ghahfarokhi et al., 2013). The continuing threat by aflatoxin contamination of food, feed and agricultural commodities to the world population has made aflatoxin research one of the most exciting and rapidly developing study areas of microbial toxins.

The present research topic includes six review articles, three mini reviews and four original research articles. Contributors highlight current global health issues arising from aflatoxins and aflatoxigenic fungi and cover important aspects of aflatoxin research including contamination of crops, epidemiology, molecular biology and management strategies. Special attention is given to fungus-plant host interactions, biodiversity and biocontrol, sexual recombination in aflatoxigenic aspergilli, potential biomarkers for aflatoxin exposure in humans and safe storage programs.

**Citation:** Razzaghi-Abyaneh, M., Chang, P. K., Shams-Ghahfarokhi, M., Rai, M., eds. (2015). Global Health Issues of Aflatoxins in Food and Agriculture: Challenges and Opportunities. Lausanne: Frontiers Media. doi: 10.3389/978-2-88919-544-2

# Table of Contents


#### *82 The microRNAs as potential biomarkers for predicting the onset of aflatoxin exposure in human beings: a review*

Rafael Valencia-Quintana, Juana Sánchez-Alarcón, María G. Tenorio-Arvide, Youjun Deng, José M. R. Montiel-González, Sandra Gómez-Arroyo, Rafael Villalobos-Pietrini, Josefina Cortés-Eslava, Ana R. Flores-Márquez and Francisco Arenas-Huertero

### *96 Use of benzo analogs to enhance antimycotic activity of kresoxim methyl for control of aflatoxigenic fungal pathogens*

Jong H. Kim, Noreen Mahoney, Kathleen L. Chan, Bruce C. Campbell, Ronald P. Haff and Larry H. Stanker

#### *108 Aflatoxins and safe storage*

Philippe Villers

### Global health issues of aflatoxins in food and agriculture: challenges and opportunities

#### *Mehdi Razzaghi-Abyaneh1 \*, Perng-Kuang Chang2, Masoomeh Shams-Ghahfarokhi <sup>3</sup> and Mahendra Rai <sup>4</sup>*

*<sup>1</sup> Research Laboratory of Fungal Metabolites, Department of Mycology, Pasteur Institute of Iran, Tehran, Iran*

*<sup>2</sup> Food and Feed Research Unit, Southern Regional Research Center, New Orleans, LA, USA*

*<sup>3</sup> Department of Mycology, Faculty of Medical Sciences, Tarbiat Modares University, Tehran, Iran*

*<sup>4</sup> Department of Biotechnology, Sant Gadge Baba Amravati University, Amravati, India*

*\*Correspondence: mrab442@yahoo.com*

#### *Edited and reviewed by:*

*Giovanna Suzzi, Università degli Studi di Teramo, Italy*

#### **Keywords: aflatoxin,** *Aspergillus***, food, agriculture, public health**

Aflatoxins are a group of polyketide mycotoxins that are produced mainly by members of the genus *Aspergillus*. Production of these toxic secondary metabolites is closely related to fungal development (Keller et al., 2005; Jamali et al., 2012). Contamination of food, feed and agricultural commodities by aflatoxins poses enormous economic and serious health concerns because these chemicals are highly carcinogenic and can directly influence the structure of DNA. The resulting genetic defects can lead to fetal misdevelopment and miscarriages; aflatoxins are also known to suppress immune systems (Razzaghi-Abyaneh et al., 2013). In a global context, aflatoxin contamination is a constant concern between the 35N and 35S latitude where developing countries are mainly situated. With expanding boundaries of developing countries, aflatoxin contamination has become a persistent problem to those emerging areas (Shams-Ghahfarokhi et al., 2013). The continuing threat by aflatoxin contamination of food, feed and agricultural commodities to the world population has made aflatoxin research one of the most exciting and rapidly developing study areas of microbial toxins.

The present research topic includes six review articles, three mini reviews and four original research articles. Contributors highlight current global health issues arising from aflatoxins and aflatoxigenic fungi and cover important aspects of aflatoxin research including contamination of crops, epidemiology, molecular biology and management strategies. Special attention is given to fungus-plant host interactions, biodiversity and biocontrol, sexual recombination in aflatoxigenic aspergilli, potential biomarkers for aflatoxin exposure in humans and safe storage programs.

Perrone et al. (2014) reported the expected risk of a shift in aflatoxin problems toward new territories particular in South East of Europe due to increasing average temperatures. Giving an overview on genetic diversity of *A. flavus* populations in Europe, the authors stressed the importance of selecting stable atoxigenic *A. flavus* strains as biocontrol agents. In the review of climate change on *A. flavus* growth and aflatoxin production, Medina et al. (2014)focused on the potential impact of key environmental factors, such as water activity (aw), temperature and atmospheric CO2, and their interactions on ecology, growth and aflatoxin production by the *A. flavus* both *in vitro* and on maize. The authors showed that while such interacting abiotic factors have little effect on fungal growth, they however have a significant impact on aflatoxin biosynthetic gene expression and can stimulate the production of aflatoxins. In the insightful mini-review on sexual recombination in aflatoxin-producing *Aspergillus* species, Moore (2014) concisely summarized the potential negative impact of sexual recombination on the feasibility of using biological controls to reduce aflatoxin contamination of field crops. The author discussed specifically the implication of sexual recombination on the fate of two commercially available biopesticides: AF36 and Afla-Guard®. In the excellent review on the characteristics of *A. flavus* as well as the biocontrol strategy using non-toxigenic *A. flavus* strains, Ehrlich (2014) described the current state and outlook of this application in agricultural field. The author concluded that understanding genetic variations among *A. flavus* strains is critical for developing a robust biocontrol strategy, and it is unlikely that a "one size fits all" strategy will work for preharvest aflatoxin reduction.

Host resistance is a very attractive area on aflatoxin research, and various aspects of *A. flavus*-plant host interaction were investigated with special focuses on mechanisms resistant to fungal growth and aflatoxin production (Dolezal et al., 2014; Fountain et al., 2014; Hruska et al., 2014; Scarpari et al., 2014; Shan and Williams, 2014). In the up-to-date review on environmental influences on aflatoxin production on maize, Fountain et al. (2014) detailed the history of research on this complex interaction and pointed out future directions for elucidating host resistance and susceptibility to *A. flavus* colonization in relation to abiotic stress such as drought and heat stresses, and oxidative stress in which aflatoxin may function as an antioxidant to the producing fungus. Utilizing an aflatoxigenic Green Fluorescence Protein (GFP) *A. flavu*s strain, Hruska et al. (2014) investigated invasion, colonization and competition in maize kernels by this engineered strain. The authors showed that a decrease in aflatoxin production is correlated with depression of the aflatoxigenic population by the biocontrol strain, AF36, supporting the theory of competitive exclusion. Using a lipidomic approach to investigate *A. flavus*maize interactions, Scarpari et al. (2014) suggested that *A. flavus* elicits the production of oxylipins in host plants, which function as signals for regulating aflatoxins biosynthesis, conidiogenesis and sclerotia formation. Their results highlighted the important role of maize oxylipins in driving secondary metabolism in *A. flavus*. In the microarray study to identify maize genes expressed during pathogen infection, Dolezal et al. (2014) found that metabolic processes are linked to defense responses, which include physical changes within the kernel as well as a disruption in kernel development. Shan and Williams (2014) provided a concise but clear overview of current knowledge about quantitative trait loci of corn related to aflatoxin contamination and ongoing efforts in the development of resistant corn lines. The authors concluded that the "phenotypic traits/data" established based on transcriptomics and proteomics approaches could be translated into the practices for improving corn resistance.

Human exposure to aflatoxins is another challenging but not well studied area. In the concise overview of aflatoxin contamination of foods and related biomarker research, Mohd-Redzwan et al. (2013) described the historical problems related to aflatoxins in Malaysia, and how these problems have influenced the Malaysian population by highlighting the aflatoxin concentrations in basic food products and their comparison with established aflatoxin limits. The authors emphasized the importance of the legislation of law for a more controlled food production, legal enforcement to meet the set regulatory standards, and the improvement of pre and post-harvest techniques to reduce aflatoxin amounts in food and hence to decrease diseases in Malaysian population. In the comprehensive review about using microRNAs as specific molecular biomarkers in populations exposed to aflatoxins and as early markers for evidence of presence of or damage by hepatocellular carcinoma (HCC), Valencia-Quintana et al. (2014) described differential expression of microRNAs under specific conditions related not only to chemical and environmental pollutants but also to biological pollutants such as the presence of aflatoxins in humans and animals, and consequently, their influence over HCC. The authors provided findings that are important to toxicological research because microRNAs can be used to predict the toxicity of some compounds and will help to explore new treatments.

In a well designed study aimed at enhancing antimycotic activities of known antifungal chemicals by natural compounds, Kim et al. (2014) successfully increased chemosensitization of a kresoxim methyl (Kre-Me), a natural fungicide from strobilurin class by chemically-synthesized benzo analogs. The authors found that among tested benzo analogs, octylgallate (OG) inhibits both growth and aflatoxin production by toxigenic aspergilli more efficiently. The study provided good evidence of remarkable synergism between OG and Kre-Me, which enhances the effectiveness of Kre-Me considerably. The efficient chemosensitizing capability of OG in increasing the efficacy of Kre-Me could reduce effective dosages of strobilurins and alleviate negative side effects associated with the current antifungal treatment.

Finally, as one of the most important and practical issues on aflatoxin research, aflatoxin prevention and elimination, Villers (2014) presented laboratory and field data on an Ultra Hermetic™ storage system, which creates an unbreatheable atmosphere through insect and microorganism respiration alone, in preventing the exponential production and accumulation of aflatoxins. This system is proven useful during multi-month post-harvest safe storage tests of maize, rice and peanuts in hot, humid countries. The author further stressed the need for research on post-harvest protection against aflatoxin contamination by determining the frequency at which excessive aflatoxin levels are reached in the field vs. after months of post-harvest storage using this system.

In conclusion, this research topic opens exciting perspectives on global health issues related to aflatoxins in the food chain and on the development of suitable strategies for preventing toxigenic fungal growth in field and storage, thereby reducing or eliminating subsequent aflatoxin contamination of our food supplies.

#### **ACKNOWLEDGMENT**

Authors gratefully thank all contributors of the research topic who made possible a successful collaborative endeavor on elucidating aflatoxins as potential public health threats.

#### **REFERENCES**


Shan, X., and Williams, W. P. (2014). Toward elucidation of genetic and functional genetic mechanisms in corn host resistance to *Aspergillus flavus* infection and aflatoxin contamination. *Front. Microbiol.* 5:364. doi: 10.3389/fmicb.2014.00364

Valencia-Quintana, R., Sánchez-Alarcón, J., Tenorio-Arvide, M. G., Deng, Y., Montiel-González, J. M. R., Gómez-Arroyo, S., et al. (2014). The microRNAs as potential biomarkers for predicting the onset of aflatoxin exposure in human beings: a review. *Front. Microbiol*. 5:102. doi: 10.3389/fmicb.2014.00102

Villers, P. (2014). Aflatoxins and safe storage. *Front. Microbiol*. 5:158. doi: 10.3389/fmicb.2014.00158

**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: 21 July 2014; accepted: 24 July 2014; published online: 12 August 2014. Citation: Razzaghi-Abyaneh M, Chang P-K, Shams-Ghahfarokhi M and Rai M (2014) Global health issues of aflatoxins in food and agriculture: challenges and opportunities. Front. Microbiol. 5:420. doi: 10.3389/fmicb.2014.00420*

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

*Copyright © 2014 Razzaghi-Abyaneh, Chang, Shams-Ghahfarokhi and Rai. 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.*

# Biodiversity of Aspergillus section Flavi in Europe in

### relation to the management of aflatoxin risk

#### *Giancarlo Perrone1, Antonia Gallo2 and Antonio F. Logrieco1\**

<sup>1</sup> Institute of Sciences of Food Production, National Research Council, Bari, Italy

<sup>2</sup> Unit of Lecce, Institute of Sciences of Food Production, National Research Council, Lecce, Italy

#### *Edited by:*

Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran

#### *Reviewed by:*

Massimo Reverberi, Università La Sapienza, Italy Sofia Noemi Chulze, Universidad Nacional de Rio Cuarto, Argentina

#### *\*Correspondence:*

Antonio F. Logrieco, Institute of Sciences of Food Production, National Research Council, via Amendola 122/O, 70126 Bari, Italy e-mail: antonio.logrieco@ispa.cnr.it

Aflatoxins and the producing fungi Aspergillus section Flavi are widely known as the most serious and dangerous mycotoxin issue in agricultural products. In Europe, before the outbreak of aflatoxins on maize (2003–2004) due to new climatic conditions, their contamination was confined to imported foods. Little information is available on molecular biodiversity and population structure of Aspergillus section Flavi in Europe. Preliminary reports evidenced the massive presence of Aspergillus flavus L -morphotype as the predominant species in maize field, no evidence of the highly toxigenic S-morphotype and of other aflatoxigenic species are reported. The risk of a shift in traditional occurrence areas for aflatoxins is expected in the world and in particular in South East of Europe due to the increasing average temperatures. Biological control of aflatoxin risk in the field by atoxigenic strains of A. flavus starts to be widely used in Africa and USA. Studies are necessary on the variation of aflatoxin production in populations of A. flavus to characterize stable atoxigenic A. flavus strains. The aim of present article is to give an overview on biodiversity and genetic variation of Aspergillus section Flavi in Europe in relation to the management of aflatoxins risk in the field.

**Keywords:** *Aspergillus***, aflatoxin, genetic variation, climate change, atoxigenic strains, biocontrol**

#### **INTRODUCTION**

The relatively recent outbreak of aflatoxins risk contamination in Europe has driven the EU researchers to investigate on this topic usually confined to tropical and sub-tropical areas. In this respect, aflatoxins and aflatoxigenic fungi are by far the most serious, dangerous and studied mycotoxin issue in agricultural products.

Aflatoxins are secondary metabolites synthesized by several *Aspergillus* species and are highly toxic to humans and animals when ingested at high concentrations. They may cause severe liver damage accompanied by jaundice, hepatitis and death, especially in developing countries (Probst et al., 2007). In addition, they are genotoxic, carcinogenic, and teratogenic for both humans and animals. Daily consumption of foods contaminated with low levels of aflatoxin B1 (AFB1) can result in chronic aflatoxicosis with stunting in children, immune suppression, cancer, and reduced life expectancy (Shephard, 2008).

Aflatoxins could extensively contaminate pre-harvest corn, cotton, soybean, peanuts, and tree nuts, and in addition residues from contaminated feed may appear in milk (Bennett and Klich, 2003).

In general, and in most of the published studies, *A. flavus* and *A. parasiticus* are the major prominent species associated in aflatoxin contamination of agricultural crops. However, recently additional species of Section *Flavi* have been reported to be responsible of aflatoxin contamination, i.e., *A. arachidicola* in peanuts and *A. nomius* in corn, nuts, and brazil nuts, especially in certain geographical area (Varga et al., 2011). New aflatoxigenic species *A. mottae*, *A. sergii* and *A. transmontanensis* have been described from maize and almonds in Portugal, they all belong to *A. parasiticus* group (Soares et al., 2012). *A. flavus* isolates produce AFB1 and AFB2 and often cyclopiazonic acid (CPA), while most *A. parasiticus* strains produce AFG1 and AFG2 in addition to AFB1 and AFB2 but never produce CPA (Horn and Dorner, 1999). Production in *A. flavus* is highly variable and depends on genotype, substrate and geographic origin, climate change and agronomic practice. Moreover, *A. flavus* is considered the predominant contaminating organism in soil and in agricultural areas and this species is more invasive and out competes *A. parasiticus* when both species are together in soil. So, most of the extensive researches have been conducted on this species and on its diverse assemblage of strains: different vegetative compatibility groups (VCGs), sclerotial type variants, toxigenic and atoxigenic strains (Ehrlich, 2014). In fact, *A. flavus* populations include isolates with two morphologically distinct sclerotial size variants, L strains with average sclerotial size >400 μm and S strains with sclerotial size <400 μm. Both these morphotypes are found in maize regions around the world; L strains are usually less toxigenic than S strains and produce only AFBs, while S strains are often high producer of aflatoxins and could be divided in two chemotypes: the SBG producing both AFBs and AFGs, and the SB producing only AFBs. The closely related new species *A. parvisclerotigenus* and *A. minisclerotigenes* were potentially associated to the SBG and SB taxons of *A. flavus*, respectively (Saito and Tsuruta, 1993; Pildain et al., 2008). Isolates that share morphological characters with SBG have been reported from Thailand, Argentina, and Australia, but the exact taxonomic designation of SBG remains unclear (Cotty and Cardwell, 1999; Donner et al., 2009). More recently, Probst et al. (2012) assigned four phylogenetic distinct groups to the S strains: one seems to belong to *A. minisclerotigenes* (SBG from Australia, Nigeria, Argentina, USA); the second is the highly toxic Kenyan SB

group (Probst et al., 2007), and a third is the SB group from elsewhere in the world, e.g., Thailand, United States, and Philippines. Finally, there is the SBG strain group from Nigeria. In general, members of the SBG group are found in locations where high levels of aflatoxin contamination occur.

It is important to underline that, based on actual surveys, the Europe population of *A. flavus* seems constituted only by L strains, none of the studies conducted have found *A. flavus* S strains. In Europe, the occurrence of *A. parasiticus* in maize seems to be very rare, while it resulted more distributed in almond especially in Portugal with an atypical chemotype producing only AFBs toxins (Rodrigues et al., 2011).

Incidence of aflatoxin outbreaks are most severe in tropical and subtropical areas around the world and also temperate regions, such as the United States Midwest are subject to occurrence of aflatoxin contamination. Until 2004, the European perspective regarding aflatoxin contamination was confined to imported foods such as peanut cake, palm kernel, copra, and corn gluten meal (depending of origin; European Food Safety Authority [EFSA], 2004). Several surveys have been conducted for detection of AFs in feed samples in Europe founding a small percentage of materials contaminated with AFB1 concentration above the regulatory limit. In fact, a survey of 110 maize samples in Northern Italy in 2003, initially planned to monitor the occurrence of fumonisins, showed 75% positive samples to AFBs with a mean of 4.4 and a maximum of 154.5 μg/kg (Piva et al., 2006). In 2006, aflatoxin contaminated rice meal used in dairy cattle feed production had been identified as the cause of elevated AFM1 levels in Swedish milk.

However, a big survey conducted by European Food Safety Authority [EFSA] (2007) evidenced the emerging issue of potential aflatoxin contamination of corn, almonds, pistachios grown in areas of Southern Europe, due to the subtropical climate occurring in some recent years. In the study of Ibáñez-Vea et al. (2012), detectable levels of AFB1 were reported for all the 123 Spanish barley samples from 2007 to 2008, and it was found to cooccur with other mycotoxins like ochratoxin A and zearalenone. In the recent work of Streit et al. (2013), samples of feed and feed raw materials from over the world were analyzed during an 8 year period for contamination with different mycotoxins. Regarding aflatoxin contamination, European samples originated primarily from Central Europe showed a high percentage of around 30% positive samples, albeit the pool of samples tested was made up by finished feeds or imported feedstuffs for more than half. A recent warning for maize contamination in Northern Italy was issued in 2012–2013 in consequence of drought conditions favorable to *A. flavus* infection (Andreotti, 2013; Causin, 2013).

In the following chapters of this mini-review we intend to analyze the most important critical points that should be studied and keep under audit to prevent and reduce the spreading of aflatoxin risk in Europe in the next decades.

#### **CLIMATE CHANGE AND RISK OF AFLATOXIN CONTAMINATION IN EUROPE**

A wide body of evidence demonstrates that the ability of fungi to grow, survive and interact with a large variety of crop species and to produce mycotoxins is greatly influenced by environmental factors, mainly temperature, relative humidity, insects. These factors are greatly related to climate change and to the variation of temperature and rainfalls. In this sense, food safety has become a very important issue worldwide and the potential effects of climate change on yields and quality of food crops, especially for mycotoxins, have received special attention in the last years, in particular from a risk analysis perspective (Miraglia et al., 2009; Magan et al., 2011). A shift in traditional occurrence areas for mycotoxins is to be expected due to the increasing average temperatures. In this respect, the Mediterranean zones have been identified as a climate change hotspot where extreme changes in temperature, CO2 levels and rainfall patterns are predicted. Regarding aflatoxins, their contamination events are more prevalent during times of high heat and drought, which may stress the host plant thereby facilitating *A. flavus* infection (Schmidt-Heydt et al., 2009; Mohale et al., 2013).

In 2003 and sporadically in the following years a hot and drying season led to severe *A. flavus* infection of maize in Northern Italy, as mentioned above (Piva et al., 2006; Battilani et al., 2008). The use of this maize as feedstuff for dairy cattle led to a widespread AFM1 contamination in milk and several thousands of tons of milk exceeding the EU legal limit of 0.05 μg/kg had to be discarded. In the 2 years following this incidence from the study of Decastelli et al. (2007) the presence of AFM1 in milk and AFB1 in feed was higher than the maximum allowable in 1.7% of raw milk samples and in 8.1% of feed samples. In 2005, the presence of these aflatoxins was below the limits of EU regulations. So, because of the very dry conditions in those years, *A. flavus* became a significant problem.

Under heat/drought stress also peanuts and pistachio can develop cracking in pods or hull splitting resulting in a significant increase in aflatoxin contamination (Cotty and Jaime-Garcia, 2007). Drought is also a major stress for the plants and undermines their natural immunity against pathogens like mycotoxinproducing fungi (Bircan et al., 2008; Kebede et al., 2012).

New strategies to monitor and predict mycotoxin contamination, either in specific foods or in geographical regions are of recent development, they could be useful in the next years for identify and predict environmental conditions present in regions that mayfavor mycotoxin proliferation. In this regard, some models have been created to predict aflatoxin for pistachio nuts (Marín et al., 2012), peanuts (Boken et al., 2008), and other crops (Masuoka et al., 2010). More recently, as in Northern Europe *A. flavus* became a dominant pathogen in maize,Battilani et al. (2013) have developed a mechanistic model "AFLA-maize" for the prediction of *A. flavus* infection and AFB1 contamination in this crop.

Another potential consequence of climate change is that the biocontrol strain could be an inadvertent cause of increased damage to the plant, especially if growing conditions are less favorable for cultivation. Changes in soil environment and its microbiome due to temperature increase, could also subject the crop to amplified damage. So, across Europe, it is relevant to improve harmonization of surveillance and monitoring of aflatoxins; improve database on the geographical distribution and prevention methods for aflatoxin; develop models for the prediction of aflatoxin contamination in the new biogeographical agricultural scenarios.

#### **BIODIVERSITY AND GENETIC DIVERSITY OF** *Aspergillus* **SECTION** *Flavi* **IN EUROPE**

Soil populations of *A. flavus* are typically composed of isolates from hundreds of different VCGs which reflect phenotypic differences (or similarity) among individuals (Leslie, 1993). Individuals (genotypes) of a fungal species having the same heterokaryon or vegetative incompatibility loci can fuse and undergo genetic exchange through parasexuality (Glass et al., 2000). Fungal isolates that form stable heterokaryons are considered to belong to same VCGs. In *A. flavus* populations, most of variations in morphology and mycotoxin production can be attributed to differences among VCGs. Vegetative compatibility group was believed to be a strong barrier to genetic exchange but recent studies found that VCGs are able to outcross, leading to new VCGs and thereby increased diversity (Olarte et al., 2012).

Recently, sexual reproduction was demonstrated in *A. flavus* which resulted to be an heterothallic fungus with two mating type loci, MAT1-1 and MAT1-2 maintained separately in homokaryotic isolates (Ramirez-Prado et al., 2008; Horn et al., 2009). Recombination can occur within conidia or sclerotia when they harbor multiple nuclei of different mating type and thereby capable of recombination.

Sexual recombination occurs in *A. flavus* through the meiotic process of independent assortment and crossing over that may influence the toxin phenotype of *A. flavus* strains, with a reduction or a complete loss of toxicity (Olarte et al., 2012). The majority of the genetic variation in mycotoxin production arises from mutations in the aflatoxin biosynthetic gene cluster (Chang et al., 2005, 2006), including gene loss, recombination, DNA inversions, partial deletions, translocations, and other genomic rearrangements of the cluster likely due to proximity of the cluster to the telomeric region of chromosome (Carbone et al., 2007).

Comparative analyses of the aflatoxin cluster in various *Aspergillus* species have underlined the complex evolutionary history of this cluster and its role in species adaptation and diversification (Ehrlich et al., 2003; Moore et al., 2009). From analyses of *Aspergillus* populations, several distinct deletions within aflatoxin cluster have been described that may each be responsible for atoxigenicity in various isolates. Either part or the entire biosynthetic cluster resulted deleted, or the non-aflatoxigenicity was associated to inability to amplify selected aflatoxin genes (Chang et al., 2005; Criseo et al., 2008; Donner et al., 2010).

Not many studies on the molecular diversity of *Aspergillus* populations isolated in Europe are available. Among these, the work of Gallo et al. (2012) was about an *A. flavus* population isolated from maize in 2003, during the first outbreak of aflatoxin contamination documented in Northern Italy (Piva et al., 2006; Giorni et al., 2007). The strains were analyzed for the presence of seven aflatoxin biosynthesis genes, including the regulatory genes *aflR* and *aflS*, in relation to their capability to produce AFB1. All aflatoxin producing isolates exhibited the complete set of amplification products, whereas non-producing isolates did not yield amplified products for three, four or all seven tested genes. The genetic diversity of *A. flavus* populations collected from maize kernels in Northern Italy from 2003 to 2010 was assessed by analysis of VCG and presence or absence of several aflatoxin genes by Mauro et al. (2013). Forty-eight VCGs were identified by means of complementation

between nitrate non-utilizing mutants. Twenty-five of these VCGs contained only atoxigenic isolates and the remaining 23 only aflatoxin producers. In addition six deletion patterns of genes in aflatoxin cluster were detected. Regarding the atoxigenic isolates, 12 of them had no deletion in the cluster, 10 had the entire cluster deleted and only one had a deletion pattern only seen once before in Nigeria, with only two genes amplified out of the thirteen tested.

The genetic variability of aflatoxin cluster in non-aflatoxigenic isolates appears diversified and complex but its understanding is important for the selection of safe and effective non-producing strains potentially usable in biocontrol for limiting aflatoxin contamination.

#### **ATOXIGENIC STRAIN AND GENETIC VARIATION IN AFLATOXIN CONTROL STRATEGY**

Interest in the variation of aflatoxin production by strains of *Aspergillus* section *Flavi* has increased recently because atoxigenic strains of *A. flavus* are being used as biological control agents to reduce the risk of aflatoxin contamination (Atehnkeng et al., 2008; Wu and Khlangwiset, 2010). Atoxigenic strains may displace wildtype aflatoxigenic strains in crop environments so only the nonaflatoxigenic population of fungi would be present in the field.

The effectiveness of pre-harvest biocontrol strategies using atoxigenic strains is based on competition for substrate, the potential production of inhibitory metabolites, and on their inability to recombine with native toxigenic strains, thus preventing the reacquisition of aflatoxigenicity (Ehrlich and Cotty, 2004; Abbas et al., 2011). Anyway, the choice of a candidate biocontrol atoxigenic *A. flavus* strain could not be based only on the phenotypic characteristic of atoxigenicity; it is necessary to investigate the genotypic condition. The long-term effect of atoxigenic biocontrol strains on native population depends on possibility of sexual recombination, in presence of which the high aflatoxin heritability will induce the regaining of ability to produce aflatoxins. There is never complete inheritance of the atoxigenic phenotype in the offspring of a biocontrol parent, so the use of atoxigenic strains biocontrol with lacking cluster genes would be preferable to one with intact biosynthetic cluster (Olarte et al., 2012; Moore, 2014).

Other factors should be considered for a successful application of biocontrol strategy such as the better understanding of natural diversity of *A. flavus* populations in agricultural soil, the ability of the introduced non-aflatoxigenic strains to recombine with the existing aflatoxigenic strains, the adaptation of *A. flavus* isolates for growth on the plant, the potential damage to the plant from the introduced strain, the potential effect on the soil microenvironment, the timing and the economical cost of application of biocontrol isolates, the potential production of other toxic metabolites, in addition to aflatoxins, which could affect animal health (Ehrlich, 2014).

This form of competitive exclusion of toxigenic strains by nonaflatoxigenic biocontrol strains has been demonstrated under field conditions in cotton (Cotty, 1994), peanuts (Dorner, 2005; Alaniz Zanon et al., 2013) and maize (Abbas et al., 2006).

The aflatoxigenic isolate AF36, which is unable to produce aflatoxin because of a point mutation in the polyketide synthase gene (*pksA*) necessary for aflatoxin biosynthesis (Ehrlich and Cotty, 2004), is registered to be used in USA for the management of aflatoxin contamination in cottonseed fields. The strain NRRL21882 is the active ingredient of Afla-Guard, a biocontrol formulation consisting of spore-coated barley seeds and used in peanut fields (Chang and Hua, 2007); the isolate is missing the entire aflatoxin and CPA gene clusters and therefore is unable to produce both mycotoxins (Chang et al., 2005, 2009). Recently the possibility to replace grain seeds in this formulation with bioplastic based granules has been explored in field experiments conducted in Northern Italy showing that bioplastic formulations are effective in reducing aflatoxin contamination in corn (Accinelli et al., 2014). In some cases the competing fungi are used as cocktails that include application of multiple strains of non-aflatoxigenic *A. flavus* (Wu et al., 2013).

#### **CONCLUSION**

The occurrence of AFB1 at high levels in Europe in the years 2003–2004 and 2012–2013 underlines the fact that the climate change will entail a change in the mycotoxin distribution patterns observed today. Global trade of plant products can also contribute to the spread of aflatoxigenic fungi and to the increase of diversity of local fungal populations. The study of diversity of aflatoxigenic fungi occurring in maize in Europe, under different points of view-morphological, molecular, metabolic, and plant pathological is essential for the development of strategies for the control of aflatoxin contamination. In this regard, the molecular characterization of native atoxigenic strains, acting through competitive exclusion of aflatoxin producers, with superior adaptation to a geographical region, should provide benefit of long-term displacement of toxigenic strains in maize environment. Additional information on the behavior of these atoxigenic isolates in the target agro-ecosystem will be needed to choose the best biological control agents. Finally, the development of predictive models for aflatoxins occurrence based on regional weather data would be a valuable tool to estimate the risk of contamination after a given growing season, together with using biopesticides in the frame of an integrated pest management (IPM).

#### **ACKNOWLEDGMENT**

This work was partially supported by EU-FPVII project MYCORED (KBBE-2007-222690).

#### **REFERENCES**


as potential biocontrol agents for aflatoxin in maize. *Food Addit. Contam.* 25, 1264–1271. doi: 10.1080/02652030802112635


gene. *Appl. Microbiol. Biotechnol.* 65, 473–478. doi: 10.1016/S1087-1845(02) 00509–501


**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 June 2014; paper pending published: 02 July 2014; accepted: 04 July 2014; published online: 21 July 2014.*

*Citation: Perrone G, Gallo A and Logrieco AF (2014) Biodiversity of Aspergillus section Flavi in Europe in relation to the management of aflatoxin risk. Front. Microbiol. 5:377. doi: 10.3389/fmicb.2014.00377*

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

*Copyright © 2014 Perrone, Gallo and Logrieco. 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.*

### Effect of climate change on *Aspergillus flavus* and aflatoxin B1 production

#### *Angel Medina , Alicia Rodriguez and Naresh Magan\**

*Applied Mycology Group, Cranfield Soil and AgriFood Institute, Cranfield University, Cranfield, Bedford, UK*

#### *Edited by:*

*Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran*

#### *Reviewed by:*

*Paula Cristina Azevedo Rodrigues, Polytechnic Institute of Braganca, Portugal Russell Paterson, University of Minho, Portugal*

#### *\*Correspondence:*

*Naresh Magan, Applied Mycology Group, Cranfield Soil and AgriFood Institute, School of Applied Science, Cranfield University, Vincent Building, College Road, Cranfield, Bedford MK43 0AL, UK e-mail: n.magan@cranfield.ac.uk*

This review considers the available information on the potential impact of key environmental factors and their interactions on the molecular ecology, growth and aflatoxin production by *Aspergillus flavus in vitro* and in maize grain. The recent studies which have been carried out to examine the impact of water activity × temperature on aflatoxin biosynthesis and phenotypic aflatoxin production are examined. These have shown that there is a direct relationship between the relative expression of key regulatory and structural genes under different environmental conditions which correlate directly with aflatoxin B1 production. A model has been developed to integrate the relative expression of 10 biosynthetic genes in the pathway, growth and aflatoxin B1 (AFB1) production which was validated under elevated temperature and water stress conditions. The effect of interacting conditions of aw × temperature × elevated CO2 (2 × and 3 × existing levels) are detailed for the first time. This suggests that while such interacting environmental conditions have little effect on growth they do have a significant impact on aflatoxin biosynthetic gene expression (structural *aflD* and regulatory *aflR* genes) and can significantly stimulate the production of AFB1. While the individual factors alone have an impact, it is the combined effect of these three abiotic factors which have an impact on mycotoxin production. This approach provides data which is necessary to help predict the real impacts of climate change on mycotoxigenic fungi.

**Keywords: climate change factors, water activity, temperature, elevated CO2, growth, gene expression, aflatoxin production, ecology**

#### **INTRODUCTION**

Food security has become a very important issue world-wide and the potential effects of climate change on yields and quality of food is now receiving significant attention by scientists, especially from a risk analysis perspective. The moldy contamination of staple foods such as cereals has received attention because of their acute and chronic effects in humans and animals. Indeed, the increasing use of staple crops, especially maize for biofuel production, has put further pressure on such key food crops. There is particular interest in maize because it is a key staple food in both developed and developing regions world-wide. Maize is prone to infection by *Aspergillus flavus* and *Aspergillus parasiticus*, especially via insect damage during silking and contamination with aflatoxins. Aflatoxins have been rated as class 1A carcinogens by the International Agency for Research of Cancer (IARC, 2012). They are heat stable and difficult to destroy during processing. Thus exposure, both acute and chronic, can have significant impacts on vulnerable groups, especially babies and children. This has resulted in strict legislative limits in many parts of the world for aflatoxins and mycotoxins in a wide range of foodstuffs (European Commission., 2006). However, in African countries where legislation is often applied to export crops only, consumption of mycotoxin contaminated staple foods is a significant risk, with rural populations exposed to aflatoxins throughout their lives, with serious impacts on their health (Wagacha and Muthomi, 2008). This is exemplified by the relatively recent acute outbreak of severe aflatoxicosis in Kenya (Lewis et al., 2005).

Climate change is expected to have a profound effect on our landscape world-wide. For some areas, climatic models have projected a marked decrease in summer precipitation and increases in temperature, which would result in concomitant drought stress episodes. The environment in which crops will be grown in the next 10–25 years may change markedly with atmospheric CO2 concentrations expected to double or triple (from 350 to 700 or 900–1000 ppm). Thus, there has been a lot of interest in the impact that climate change scenarios may have on economically important crops/mycotoxigenic fungal infection and contamination with mycotoxins (Paterson and Lima, 2010, 2011; Magan et al., 2011; Wu et al., 2011). Indeed, climate change conditions may impact on the interactions between different mycotoxigenic species and indeed other mycobiota and determine the relative mycotoxin composition contaminating staple foods/feeds (Magan et al., 2010; Paterson and Lima, 2012). Because of this increase and that of other greenhouse gases, the global temperature is expected to increase by between +2 and +5◦C.

The EU green paper on climate change in Europe also suggests that effects will be regional and be either detrimental or advantageous depending on geographical area. Thus, in Southern Europe, changes may equate to an increase of 4–5◦C with longer drought periods, resulting in increasing desertification, and a decrease in crop yields. In areas of Western and Atlantic Europe, changes of 2.5–3.5◦C with dryer and hotter summers are envisaged. In Central Europe, an increase of 3–4◦C, higher rainfall and floods are forecast, although longer growing periods may benefit crop yields. Northern Europe would expect a mean temperature increase of 3–4.5◦C, with a significant increase in precipitation of 30–40%. This may lead to increases in crop yields and perhaps new crop cultivation patterns (European Commission., 2007; Solomon et al., 2007). Similar impacts have been described in other areas of the world, especially parts of Asia and Central and South America which are important producers of staple crops (IPCC, 2007). A recent study has predicted that, on a global scale, pests, and diseases are moving to the poles at the rate of 3–5 km/year (Bebber et al., 2013). This could have further impacts on contamination of staple foods such as maize, as increases in pest reproduction rates will lead to more damage and facilitate more infection by *A. flavus* and contamination with aflatoxins. However, in the recent predictions by Bebber et al. (2013) no focus on spread of mycotoxigenic fungi or mycotoxins or interactions between pathogens and pests were considered in the context of climate change.

In developing countries drought stress may be particularly important in terms of food security. For example, marginal land where stress tolerant sorghum was previously grown has now been replaced with maize in both West and East Africa. Maize as well as ground nuts are particularly prone to infection when water stress periods occur. This leads to increased aflatoxin contamination of such crops pre-harvest and post-harvest and can significantly impact on the ability to export the crop and also on the nutritional quality when consumed in rural subsistence communities.

Magan et al. (2011) suggested that climate change factors may result in xerophilic fungi such as *Wallemia sebi*, *Xeromyces bisporus*, and *Chrysosporium* species becoming more important as colonizers of food commodities, as they can grow under very dry conditions [0.65–0.75 water activity (aw)] where there is much less competition from the majority of mesophilic fungi (Magan, 2006; Magan and Aldred, 2007). For example, *W. sebi* can produce metabolites such as walleminol and walleminone which can be toxic to animals and humans (Piecková and Kunová, 2002). Studies also suggest that there are competitive interactions between these xerophilic fungi in dry and hot conditions and that secondary metabolites may play a role (Leong et al., 2010). This will certainly have an impact on agricultural productivity, especially of essential/staple food crops such as maize and nuts and also influence the interface between plants, insect pests and fungal infection of staple foods (Miraglia et al., 2009). This could have a profound effect on pre- and post-harvest mycotoxin contamination, especially aflatoxins in developing countries, where food quality and security issues are critical.

Examples of modified weather regimes impacting on mycotoxins were demonstrated by the 2003/2004 and subsequently in 2012 summer seasons in the Mediterranean region such as Northern Italy where drought and elevated temperatures resulted in a switch from *Fusarium verticillioides* and contamination with fumonisins to significant contamination of maize grain with *A. flavus* and aflatoxins and entry of aflatoxin M1 into the dairy chain via the animal feed chain (Giorni et al., 2007). More recently, a survey of Serbian maize samples in 2009–2011 contained no aflatoxins. However, prolonged hot and dry weather in 2012 resulted in 69% of samples containing aflatoxins (Kos et al., 2013). Similarly in Hungary it has also been shown that an increase in aflatoxins may be due to climate change conditions (Dobolyi et al., 2013). However, previous to these examples there are only a few concrete examples of such incidences where climate change factors have been implicated (Magan et al., 2011).

#### **EFFECT OF WATER STRESS × TEMPERATURE STRESS EFFECTS ON AFLATOXIN CLUSTER GENE EXPRESSION, GROWTH, AND AFLATOXIN PRODUCTION**

Generally, the aflatoxin biosynthesis genes of *A. flavus* and *A. parasiticus* are highly homologous and the order of the genes (approx. 30) within the cluster has been shown to be the same (Yu et al., 1995, 2004). These include key regulatory genes (*aflR* and *aflS*) and a series of up and downstream structural genes. It has been shown that both water availability and temperature modifications affect the expression of these clusters of genes, relative growth rate and aflatoxin production in both *A. flavus* and *A. parasiticus* (Schmidt-Heydt et al., 2010, 2011). It was shown that there was a good correlation between the expression of an early structural gene (*aflD*) and aflatoxin B1 (AFB1) (Abdel-Hadi et al., 2010). It has also been shown that temperature × aw interactions were related to the ratio of the two key regulatory genes (*aflR/aflS*). The higher the ratio, the higher the relative AFB1 production (Schmidt-Heydt et al., 2009, 2010; **Figure 1**). This suggests that under certain interacting conditions of two environmental stress factors significantly influences on the relative amounts of AFB1 produced.

The study by Abdel-Hadi et al. (2010) also showed that when examining the relationship between temporal AFB1 production and the relative expression of the *aflD* structural gene involved early in the biosynthetic pathway then the relative expression could be mapped over time (**Figure 2**). This also suggests that the optimum aw for *aflD* expression was at 0.90 aw, which is different from that for growth (0.95 aw).

More recent detailed studies using a mycotoxin microarray (Schmidt-Heydt and Geisen, 2007) have been useful in elucidating the relationship between both key structural genes and the regulatory genes and interacting conditions of aw × temperature and integrated the data on relative expression of 10 genes with growth and AFB1 data (Abdel-Hadi et al., 2012). **Figure 3** shows the effect of these factors on growth and AFB1 production. They were able to model and validate this relationship under elevated temperature and drought stress conditions but elevated CO2 was not included in these studies. However, the relative relationship between the regulatory genes and key structural genes were examined using relative expression data under conditions of changing temperature and water stress to better understand the relationships between the regulatory and structural genes (**Figure 4**). This development of such ternary diagrams can help to evaluate the relationships between 3 key regulatory and structural genes at a time under different temperature and water stress conditions to help identify which are critical in the biosynthetic pathway as environmental factors change.

**FIGURE 1 | The key structural and regulatory genes involved in aflatoxin production and the effect of temperature × water activity conditions, ratio of the two regulatory genes and relative amounts of aflatoxin B1 production (adapted from Schmidt-Heydt et al., 2010).**

Schmidt-Heydt et al. (2008) also demonstrated that there was a stimulation of toxin biosynthetic gene expression in different mycotoxigenic fungi when exposed to interacting aw × temperature stresses including *A. flavus*. They suggested that there was two peaks of expression one under optimum abiotic interacting conditions and one when water and temperature stress was applied. Studies by Yu et al. (2011) examined the effect of elevated temperature on the relative expression of the whole genome of a type strain of *A. flavus* to identify groups of up and down regulated genes. However, these studies were carried out over short time periods and did not include interactions with environmental stresses or with elevated CO2.

A significant amount of data exists on the effect of interactions between water availability and temperature on the life cycle of mycotoxigenic fungi and mycotoxin production (Sanchis and

Magan, 2004; Magan and Aldred, 2007). This includes the ecological conditions of aw × temperature which will facilitate growth and AFB1 production. The aw × temperature boundary conditions for toxin production are slightly different from that for growth. Based on this information it is possible to predict the effect of increased temperature (e.g., 37◦C) and water stress

(0.95, 0.90 aw) on growth and aflatoxin production (**Table 1**). This shows that as you increase the temperature to 37◦C you get significantly less AFB1 produced although the *A. flavus* is able to grow. However, this excludes the interaction with CO2 which is necessary to examine in more detail the impact of predicted climate change scenarios.

#### **CLIMATE CHANGE IMPACTS (a<sup>w</sup> × TEMPERATURE × CO2) ON AFLATOXIN GENE CLUSTER EXPRESSION, GROWTH AND TOXIN PRODUCTION**

Recently, Medina et al. (in press) examined the effect of existing environmental conditions and when conditions were changed from 34 to 37◦C, with drought stress and CO2 was increased to 650 and 1000 ppm. They examined the effects on growth of *A. flavus* and on the relative expression of the structural *aflD* and the regulatory *aflR* genes, as well as AFB1 production for the first time. These studies have shown that for growth of *A. flavus* there was relatively little effect of these interacting climate change conditions (**Table 2**). However, there were significant differences between growth in



*Key:* <sup>μ</sup> *max, Maximum growth rate (mm day*−*1);* <sup>μ</sup>+*3, Growth rate increasing 3*◦*C;* μ+*5, Growth rate increasing 5*◦*C;* τ *max, Maximum toxin production (*μ*g <sup>g</sup>*−*1);* <sup>τ</sup>+*3*◦*C, Predicted toxin increasing* <sup>+</sup>*3*◦*C;* <sup>τ</sup>+*5*◦*C, Predicted toxin increasing* +*5*◦*C, NP, No toxin production.*

**Table 2 | Comparison of growth of** *Aspergillus flavus* **under different interacting conditions of elevated temperature, drought stress, and elevated CO2** *in vitro* **on a conductive yeast-glucose medium (adapted from Medina et al., in press).**


*This was compared with the effect of these three-way interactions on the relative expression of both a structural gene and a regulatory gene (aflD, aflR) in modified aw* × *temperature* × *650 and 1000 ppm CO2.*

relation to aw, but no effect of aw × temperature × CO2 on growth rate.

**Table 3** summarizes the results of this study and shows the effects of the three-way interacting conditions on relative gene expression (*aflD*, *aflR*) and AFB1 production. This clearly shows that under slightly elevated CO2 conditions there was a stimulation of AFB1 production, especially under drought stress at 37◦C and 650 and 1000 ppm CO2 exposure. It seems that the interactions between these three factors together are critical in the impact that slightly elevated CO2 has. This is clear from the results obtained at 0.92 and 0.95 aw × 37◦C and 650 or 1000 ppm CO2 where a statistically significant increase in AFB1 was observed.

#### **DISCUSSION AND CONCLUSIONS**

This review has considered the impact of different key environmental factors on the growth, gene expression and AFB1 production by *A. flavus*. This has shown that while there are some examples of the impact that changes in climatic weather conditions may have resulted in a switch with contamination from fumonisins to aflatoxins in maize, there have been few studies to examine the three-way interactions of the key environmental factors. Previous studies have examined water stress × temperature interactions on relative biosynthetic genes involved in aflatoxin production and that by other mycotoxigenic fungi

**Table 3 | Summary of the impact that interactions between the three climate change variables have on relative expression of the structural and regulatory genes (***aflD***,** *aflR***), and aflatoxin B1 production (from Medina et al., in press).**


=*, variation lower than 2-fold. Numbers between brackets refer to the foldvariation with respect to the control.*

(Abdel-Hadi et al., 2010; Schmidt-Heydt et al., 2010; Medina et al., 2013). However, there has been little detailed evidence of the impact that three-way interacting factors may impact on aflatoxin production.

The recent study by Medina et al. (in press) is the first to attempt to quantify the effects of interacting factors of water stress × temperature × elevated CO2 on growth, biosynthetic gene expression and AFB1 production. This new study suggests that when the three climate change factors are interacting there are responses which are not obtained when examining aw × temperature conditions only. Thus, while growth is relatively unaffected by the addition of 2× and 3× existing CO2 levels at 37◦C under the different water stress treatments used, this is not the case with mycotoxin production. The relative increased expression of both the structural *aflD* and the regulatory *aflR* genes in this recent study suggests that there is a significant impact on the biosynthetic genes involved in secondary metabolite production by strains of *A. flavus*. This was especially so at 37◦C and under water stress (0.95, 0.92 aw) where more changes were observed. This study showed that there is a strong stimulation of mycotoxin production (from ×15.1 to ×79.2 depending on the climate exposure conditions used).

Additional studies are now required to evaluate whether this is a general stress response or whether the presence of elevated CO2 results in its incorporation into the biosynthetic pathways for enzyme production and secondary metabolite production. Perhaps new studies need to be carried out with the cell wall integrity (CWI) and high-osmolarity glycerol (HOG) pathways to examine whether they are triggered by stimuli of the three interacting factors of water stress × temperature × elevated CO2 or if this is a general stress response *per se* (Hayes et al., 2014). Work is in progress with maize grain to compliment the data obtained by Medina et al. (in press). A recent study by Vaughan et al. (2014) showed that twice the existing CO2 concentrations (400 and 800µmol CO2mol−1) increased the susceptibility of maize to *Fusarium verticillioides* proliferation although fumonisin B1 mycotoxin production was not affected. They showed that inoculation at silking the accumulation of sugars, free fatty acids, lipoxygenase transcripts, phytohormones, and downstream phytoalexins were reduced in the maize grown at elevated CO2 conditions. Further studies using this approach are required where maize is grown under such conditions and then examining the host-pathogen interaction under the climate change scenarios described here.

Abdel-Hadi et al. (2012) with the aim of forecasting the AFB1 production by *Aspergillus flavus* examined the integration of growth, gene expression of multiple aflatoxin genes and AFB1 production by using a mixed secondary metabolite model. This model was validated at 37 and 40◦C and different water stress levels and predicted AFB1 production at 37◦C under water stress conditions, but none at 40◦C. However, CO2 was not included in this model. The results obtained suggest that this model could be extended to include CO2 as a parameter and that this could be a very interesting tool to help in predicting the impact of climate change scenarios with experimental data sets as opposed to being based on historical data sets. This would be beneficial in quantifying impacts of climate change scenarios on economically important staple food crops.

There are some examples of previous studies using data on drought stress × temperature effects on *A. flavus* to predict impacts of interacting environmental factors. Work by Chauhan et al. (2008, 2010) demonstrated that it is possible to utilize an Agricultural Production Systems Simulator to calculate an Aflatoxin Risk Index (ARI) in both maize and peanuts in Australia. For maize they related seasonal temperature and soil moisture during the critical silking period to determine the ARI. They showed that both dry and hot climates made maize prone to a much higher aflatoxin contamination risk. For peanuts, they used the fractional amounts of available soil water during the crucial pod-filling period to determine the ARI. This showed that historically there has been an increase in aflatoxin contamination of peanuts in Australia related to increases in ambient temperature and decreases in rainfall. This has been developed into a web-interface tool for practically real-time use of this model. This approach is very valuable to predict low and high risk years in relation to climatic fluxes and may have application in West Africa where maize is also an important staple crop. However, these models may need modification to provide accurate predictions under climate change scenarios. Recently, Battilani et al. (2013) developed a mechanistic weather-driven model based on the infection cycle of *A. flavus* on maize to predict the risk of aflatoxin contamination in field on a daily basis from silk emergence to harvest. This included a probability index to exceed the legal limit of 5µg/kg maize for aflatoxin. They suggested that this approach can be used for prediction of *A. flavus* infection and aflatoxin contamination during the growing season and at harvest. It may be possible to input the type of data from the present study to make this approach more accurate and improve the predictions of relative risk to take account of climate changes.

Many of the recent reviews which have examined aspects of the impact of climate change have focused on plant breeding, plant diseases and mycotoxins in Europe, Australia, Africa, and the USA (Boken et al., 2008; Chauhan et al., 2008, 2010; Wu et al., 2011). These have predominantly examined the existing or historical information and tools relevant to the impacts on crop yield, the impact of drought episodes and lack of water or elevated temperatures. Magan et al. (2011) examined the impacts of aw × temperature stress on potential changes in mycotoxin production when the temperature is changed by +3 and +5◦C and under different water stress regimes. Other reviews have used this same data (Paterson and Lima, 2010, 2011, 2012) to make their prediction of potential impacts. However, these previous studies did not include the three way interactions between aw × temperature × elevated CO2. The laboratory based studies now available and those being done in FACE based systems need to be combined to be able to obtain more accurate information which can be used to predict on a regional basis the real impact that climate change scenarios may have on exposure to aflatoxins. This is especially important in at risk regions such as parts of Africa and Asia where the risks of exposure may increase under these predicted climate change conditions and threaten food security.

#### **REFERENCES**


fungi to chemical fungicides and antifungal proteins. *Cell. Mol. Life Sci.* 71, 2651–2666. doi: 10.1007/s00018-014-1573-8


**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 May 2014; accepted: 23 June 2014; published online: 22 July 2014.*

*Citation: Medina A, Rodriguez A and Magan N (2014) Effect of climate change on Aspergillus flavus and aflatoxin B*<sup>1</sup> *production. Front. Microbiol. 5:348. doi: 10.3389/ fmicb.2014.00348*

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

*Copyright © 2014 Medina, Rodriguez and Magan. 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.*

### Sex and recombination in aflatoxigenic Aspergilli: global implications

#### *Geromy G. Moore\**

Southern Regional Research Center, Agricultural Research Service, United States Department of Agriculture, New Orleans, LA, USA

#### *Edited by:*

Perng-Kuang Chang, United States Department of Agriculture, USA

#### *Reviewed by:*

Massimo Reverberi, Sapienza – Università di Roma, Italy Edwin Rene Palencia, United States Department of Agriculture, USA

#### *\*Correspondence:*

Geromy G. Moore, Southern Regional Research Center, Agricultural Research Service, United States Department of Agriculture, 1100 Robert E. Lee Blvd, New Orleans, LA 70124, USA e-mail: geromy.moore@ars.usda.gov

For most of the half century that aflatoxigenic species have been intensively studied, these molds were known only to reproduce asexually, with parasexuality found only in the laboratory between certain mutant strains. Therefore, the fairly recent discovery of their sexual (teleomorphic) states creates a new wrinkle in our understanding of the field behavior of these agriculturally significant fungi. Sex within populations of these fungi, and attendant genetic recombination, eventually may create difficulties for their control; and subsequently for the protection of important human and animal food supplies. Moreover, if fungal sex is a form of response to ecological and environmental stressors, then perhaps human influence and climate change could accelerate this phenomenon. This article will explore scientific research into sexuality and recombination in aflatoxigenic Aspergillus species; the potential impacts these phenomena could have on a popular method of preharvest prevention of aflatoxin contamination (i.e., use of non-aflatoxigenic A. flavus for biocontrol); and the outlook for maintaining control of aflatoxin contamination in an era of changing global climate.

**Keywords:** *Aspergillus***, aflatoxin, sexuality, recombination, climate change, biocontrol**

#### **INTRODUCTION**

Within the expansive community of agriculturally significant fungi are species that pose health risks to animals and humans through the production of mycotoxins; consequently, these mycotoxigenic fungi are important to understand and to control (Bennett and Klich, 2003). Aflatoxins are the most serious agricultural mycotoxins. They are mostly produced by a group of fungal species within the genus *Aspergillus*' section *Flavi* (Bennett, 2010), although a few other species outside of this section produce aflatoxins, and carry homologs of the genes for aflatoxin (AF) synthesis (Bradshaw, 2006; Cary et al., 2009; Bradshaw et al., 2013). *Aspergillus flavus* and *A. parasiticus* are the most prominent aflatoxigenic (AF+) species with agricultural significance, contaminating cereal and oilseed crops as well as tree nuts; *A. nomius* has also been reported in agricultural fields (Feibelman et al., 1998; Ehrlich et al., 2007). The global economic losses due to contamination by these fungi are in the billions of dollars (Wu et al., 2008). AF contamination causes negative impacts on health and life across the globe, especially in many developing nations where inhabitants lack the education regarding the risks of consuming aflatoxin-contaminated foods, lack understanding of the importance for proper food storage, or perhaps would rather risk eating contaminated crops than go hungry (Williams et al., 2004; Shephard, 2008). The burden of illness and death associated with aflatoxin consumption is a constant reminder that efficient AF control measures are in need of globalization.

Researchers are working to develop a food supply that is free of AF contamination through implementation of pre-harvest and/or post-harvest strategies (Hell et al., 2008; Abbas et al., 2011). One of the most appealing control measures involves the use of naturally occurring, non-aflatoxigenic (AF−) *A. flavus* as pre-harvest biological control instead of chemical fungicides. This method involves the field dispersal of a high volume of inocula composed of a highly competitive AF− strain. The presence of the AF− strains interferes with the proliferation of indigenous AF+ fungi (Abbas et al., 2011). Studies on the genetic background of these AF− strains show that they result from either random mutations in, or absence/loss of, genes necessary for AF synthesis (Ehrlich and Cotty, 2004; Chang et al., 2005; Moore et al., 2009). Since their introduction a decade ago, two commercially available biopesticides have been in active use throughout the United States known as AF36 and Afla-Guard®. Currently, other strains are being studied and developed as potential biocontrol agents, not only in the U.S. but also in other parts of the world (Pitt and Hocking, 2006; Abbas et al., 2011; Probst et al., 2011). This method of control is gaining favor due to its proven success at reducing AF contamination in the field, but our confidence in using AF− *A. flavus* strains as biopesticides is based largely on the logic that they are predominantly asexual and genetically stable (Ehrlich and Cotty, 2004).

#### **RECOMBINATION AND SEXUALITY IN SECTION** *FLAVI*

Ken Papa explored parasexual recombination in *A. flavus* and *A. parasiticus* 40 years ago, but his findings were limited to laboratory experiments between mutant strains (Bennett, 1985). David Geiser was one of the first to report evidence for genetic recombination in *A. flavus,* due to a cryptic sexual state, and hence the potential risk for using *A. flavus* strains as biocontrol (Geiser et al., 1998). His findings inspired further studies by Ignazio Carbone's research group to uncover supportive evidence of recombination within the AF gene clusters of *A. parasiticus* and *A. flavus* populations sampled from within the same peanut field in the U.S.

(Carbone et al., 2007; Moore et al., 2009). Recombination breakpoints within the AF gene clusters of *A. parasiticus* and *A. flavus* were observed, even though the*A. parasiticus* population exhibited less historical recombination compared to the *A. flavus* population in the same field (Moore et al., 2009). Moore et al. performed a global scale study of recombination for *A. flavus* and *A. parasiticus* field populations, representing five separate continents, and observed evidence of recombination within each *A. flavus* population examined and for two of the *A. parasiticus* populations (Moore et al., 2013a). Similar patterns of linkage disequilibrium (LD) could be observed for global *A. flavus* L-strain populations, although the same could not be observed for all of the sampled *A. parasiticus* or *A. flavus* S-strain populations. Ramirez-Prado et al. (2008) characterized mating alleles in *A. flavus* and *A. parasiticus*, and determined that since only one mating-type idiomorph (*Mat1-1* or *Mat1-2*) could be amplified for each isolate it could be concluded that both species exhibit a heterothallic mating system. They developed a PCR diagnostic to quickly identify the idiomorphs, and using this diagnostic test they investigated the distribution of mating-types for the U.S. populations of *A. flavus* and *A. parasiticus*. Global recombination rates appear to correlate with the distribution of mating-type idiomorphs within certain field populations – an equal distribution of *Mat1-1* and *Mat1- 2* will yield higher incidences of recombination as well as yield greater diversity of aflatoxigenicity among individuals (Moore et al., 2009). Recombination generates more individual offspring with genomes that differ from both parents, meaning that sex in fungi may increase the number of vegetative compatibility groups (VCGs; Olarte et al., 2011; Moore et al., 2013a). When fungi like *A. flavus* out-cross, they overcome barriers such as vegetative incompatibility (likely to prevent cell death) and allow the exchange of novel genetic material between two fertile strains (Pál et al., 2007; Horn et al., 2009a). Supportive evidence for this circumvention of vegetative incompatibility exists through mating studies in *A. flavus* and *A. parasiticus* whereby the only successful pairings involved strains from different VCGs and opposite mating types (Horn et al., 2009a,b).

Investigations into the sexuality of three AF+ species (*A. parasiticus*, *A. flavus*, and *A. nomius*) eventually led to the discovery and characterization of their teleomorph states using the taxonomic nomenclature of genus *Petromyces* (Horn et al., 2009a,b, 2011). The teleomorph discoveries for the AF+Aspergilli occurred in the laboratory, and there are some who might argue that experimental crosses in a controlled environment are not representative of natural field conditions. However, experimental field studies by Horn et al. (2013) demonstrated *A. flavus* sexuality in the field. Ears of corn were inoculated with compatible parental strains and later found to contain sclerotia. The sclerotia harvested from the ears of corn were not cleistothecium-bearing stromata; however, sclerotia collected from the corn and further incubated in and on non-sterile soil eventually contained cleistothecia. Reportedly, when these sclerotia fall onto the soil out-crossing is stimulated (Horn et al., 2013). Another teleomorph discovery is the evidence of hybridization reported between *A. flavus* and *A. parasiticus* (Worthington et al., 2011). This may mean that the biological species recognition concept, which defines a species as reproductively isolated (Taylor et al., 2006; Samson and Varga, 2009), cannot be applied to these aflatoxigenic Aspergilli.

*A. flavus*,*A. nomius,* and*A. parasiticus* exhibit self-incompatible (heterothallic) mating systems (Horn et al., 2009a,b, 2011). Of particular interest for *A. nomius* is that both mating-type idiomorphs may exist within a single strain, whereby certain isolates PCRamplified as both *Mat1-1* and *Mat1-2*, and either idiomorph may be functional when out-crossing (Horn et al., 2011). A possible explanation for the presence of both mating-type idiomorphs may relate to the fact that *Aspergillus* species possess multi-nucleic conidiospores and hyphal cells, and are potentially heterokaryotic (Olarte et al., 2011; Runa et al., 2011). Olarte et al. (2011) examined progeny from out-crossing *A. flavus* strains that have either a full, partial, or absent AF gene cluster. In one generation of experimental crosses, most of the offspring exhibited recombinant genomes, having specific locus similarity to one or the other parent, while others were genetically distinct from both parents. They reported absence of cyclopiazonic acid (CPA) and AF loci in array Comparative Genomic Hybridization (aCGH) analysis for progeny, despite both parents having full and present CPA and AF clusters within their genomes, and yet the loci could be amplified using PCR. They posit that cryptic alleles influence genome variability, since some offspring were observed to amplify portions of the CPA and AF clusters when both parents exhibited the partial/absent AF cluster configuration (Olarte et al., 2011). Olarte's research suggests that copy numbers of examined loci may influence the genomic condition observed, whereby low copy, ancestral alleles are often masked, but still present among heterokaryotic nuclei (Olarte et al., 2011).

#### **POTENTIAL IMPACTS OF SEXUAL RECOMBINATION ON** *A. FLAVUS* **BIOCONTROL**

The effectiveness of pre-harvest biocontrol strategies using AF− *A. flavus* strains is based on their aggressiveness as competitors coupled with their inability to recombine with native AF+ strains, thereby preventing the re-acquisition of aflatoxigenicity (Ehrlich and Cotty, 2004; Abbas et al., 2011). Indeed, the biocontrol strains may be incapable of reinstated AF production, but not all of the offspring that result from their out-crossing will inherit the AF− phenotype. Through annual inundation of fields with biocontrol strains, the potential, over many generations, to increase the load of "super-competitors" with AF+ properties increases (Moore et al., 2013b). Given the possibility of obtaining multiple VCGs in a single generation, sexually promiscuous offspring could recombine with each other and further increase the population of AF+ individuals, thereby rendering the current biocontrol methodology ineffective in the future.

According to Olarte et al. (2011), the incidence of AF+ offspring observed was higher for crosses involving AF36 (58%) than for the Afla-Guard strain® (36%). For offspring with an AF36 parent, replacement with a functional *pksA* gene can promote AF synthesis, but with the component strain in Afla-Guard® (NRRL 21882) lacking the entire AF gene cluster, a simple replacement is less likely. Since the re-acquisition of AF cluster genes in offspring is less likely, use of a biocontrol strain lacking cluster genes such as NRRL 21882 would be preferable to one with an intact AF gene cluster. There is never complete inheritance of the AF− phenotype in the offspring of a biocontrol parent, unless both parents exhibit partial- or absent-cluster genotypes (Olarte et al., 2011). In field populations where AF+ strains are present, sex may yield toxigenic progeny. At this time, no studies have been reported that test colonization aggressiveness for progeny resulting from AF+ and biocontrol strain pairings, though there is evidence that AF levels are not significantly higher in the offspring from these out-cross events (Olarte et al., 2011). However, Moore et al. (2013a)reported that higher incidences of toxin diversity exist in actively recombining populations of *A. flavus* and *A. parasiticus*. For example, balancing selection in *A. flavus* seeks to maintain the AF− phenotype, but active recombination will alter the overall AF load of the population by reducing the numbers of AF− individuals (Moore et al., 2009, 2013a).

With regard to hybridization between *A. flavus* and *A. parasiticus*, perhaps the impact from their recombination may be important to agriculture beyond their mycotoxigenic potential. Firstly, *A. parasiticus* is predominately a soil inhabitant (Angle et al., 1982), whereas *A. flavus* is more ubiquitous (Zuluaga-Montero et al., 2010). Hybridization may allow *A. parasiticus* to alter or increase its niche through genetic modification. Hybrid offspring could exhibit far more diversity than recombinant offspring within each species (Olarte et al., 2011; Worthington et al., 2011). Perhaps, the most recently described AF+ "species" such as *A. parvisclerotigenus* and *A. minisclerotigenes* are actually hybrids, between *A. flavus* and B+G AF producers such as *A. parasiticus* or *A. nomius,* which have persisted due to selective advantages resulting from their hybridization events. Current molecular techniques will allow us to refute or support this hypothesis.

In the future, when seeking to use AF− *A. flavus* as pre-harvest biocontrol, there should be more diligence in researching the field ecology where biocontrol strain dispersal is intended for use or is currently in use. Specifically, the population density of native, potentially fertile AF+ species should be ascertained first by thorough field sampling; and additionally, the mating-type distribution of these native field strains should be determined since this will influence the stability of biocontrol (Moore et al., 2013a). If the field is predominately skewed to one mating type or the other, then using a biocontrol strain of the majority mating type could further restrict recombinant opportunities and slow the progression of both strain and toxin diversity (Moore et al., 2013a). In addition, any AF− strains sampled in a field should undergo extensive phenotypic and genomic investigations for consideration as biocontrol in the field where sampled (Moore et al., 2013a). Merely exhibiting the AF− phenotype is no longer a sufficient phenotypic criterion to warrant consideration of a strain as a candidate agent for biocontrol. Given its genotypic condition, the opportunity to generate AF+ offspring with a biocontrol strain such as NRRL 21882 is less likely than that with a strain such as AF36 (Olarte et al., 2011). Not only does NRRL 21882 exhibit the AF− phenotype because of it lacks AF cluster genes, but it also is incapable of producing CPA, another mycotoxin that is considered by some to have contributed to the severity of the Turkey X outbreak in the 1960s (Richard, 2008), while the AF36 biocontrol strain is AF− but produces CPA (Abbas et al., 2011), and CPA production has been observed in all offspring resulting from out-crossing with the AF36 parent (Olarte et al., 2011).

#### **FUNGAL SEX AND RECOMBINATION IN A CHANGING GLOBAL CLIMATE**

There is increasing evidence that climate change is causing more unpredictability in global weather patterns. High heat and drought conditions stress plants and facilitate infection by aflatoxigenic species such as *A. flavus* (Scheidegger and Payne, 2003). Agricultural areas experiencing drought often suffer outbreaks of AF contamination. Moreover, diminished water availability limits the ability to irrigate and thereby mitigate the effects of drought (Kebede et al., 2012). At this time, incidences of AF outbreaks are most severe in tropical and sub-tropical areas (between latitudes 40oN and 40oS) around the world (Williams et al., 2004), and even temperate regions such as the United States Midwest are subject to occurrences of AF contamination. However, if current scientific reports are accurate, the average global surface temperature has been increasing by 0.15 oF each year since 1901 (United States Environmental Protection Agency [U.S.E.P.A], 2013). If temperatures continue to increase then the ideal climate for outbreaks of AF contamination will encompass more of our "temperate" agricultural regions and also become more frequent in occurrence. Therefore, it is imperative that as the research establishments continue to seek ways to control AF that researchers be aware of the potential impacts of climate change on the pathogenicity of AF+ fungi and the basic biology of these fungi. Sexual recombination often results from environmental stressors these fungi must overcome in order to adapt and survive. There is an extensive history of recombination in *A. flavus* (Moore et al., 2009). If global climate events assert constant negative pressures on AF+ Aspergilli, then this may accelerate the frequency of recombination in natural populations and lead to unfavorable outcomes for crop protection.

#### **CONCLUSION**

The ultimate goal for using AF− *A. flavus* as biocontrol agents is long-term crop protection. Although biocontrol strains are reported to persist for years in inoculated fields (Cotty, 2013), current strategies require annual re-application of biocontrol strains. If the signature of the biocontrol strain is lost then perhaps recombination is to blame. Potentially, even a low rate of recombination between native AF+ fungi and introduced AF− fungi is significant when one considers future food safety. Generations from now, the aflatoxin problem may become more intractable because of the short-term method currently being used to prevent pre-harvest contamination. These fungi are – and have long been – sexually active. Their ability to evolve new phenotypes and genotypes via sexual recombination is a fact that cannot be ignored.

Sometimes, pathologists refer to plant: pathogen interactions as an evolutionary arms race (Anderson et al., 2010). Perhaps the same could be said for the use of non-aflatoxigenic strains for aflatoxin control. Hopefully, with continued research and understanding, we can maintain a consistent level of control without future risk of exacerbating the aflatoxin problem.

#### **REFERENCES**


**Conflict of Interest Statement:** The associate editor Perng-Kuang Chang declares that despite being in the same research area at the United States Department of Agriculture as the author Geromy Moore (Food and Feed Safety Research at Southern Regional Research Center), the review process was handled objectively and no conflict of interest exists.

*Received: 16 December 2013; accepted: 17 January 2014; published online: 05 February 2014.*

*Citation: Moore GG (2014) Sex and recombination in aflatoxigenic Aspergilli: global implications. Front. Microbiol. 5:32. doi: 10.3389/fmicb.2014.00032*

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

*Copyright © 2014 Moore. 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: 10 February 2014 doi: 10.3389/fmicb.2014.00050

### Non-aflatoxigenic Aspergillus flavus to prevent aflatoxin contamination in crops: advantages and limitations

#### *Kenneth C. Ehrlich\**

Southern Regional Research Center, United States Department of Agriculture – Agricultural Research Service, New Orleans, LA, USA

#### *Edited by:*

Mahendra Kumar Rai, Sant Gadge Baba Amravati University at Amravati, India

#### *Reviewed by:*

Ebrahim Hadavi, Islamic Azad University, Iran Zhao Chen, Clemson University, USA

#### *\*Correspondence:*

Kenneth C. Ehrlich, Southern Regional Research Center, United States Department of Agriculture – Agricultural Research Service, New Orleans, LA 70124, USA e-mail: ken.ehrlich@ars.usda.gov

Aspergillus flavus is a diverse assemblage of strains that include aflatoxin-producing and non-toxigenic strains with cosmopolitan distribution.The most promising strategy currently being used to reduce preharvest contamination of crops with aflatoxin is to introduce nonaflatoxin (biocontrol) A. flavus into the crop environment. Whether or not introduction of biocontrol strains into agricultural fields is enough to reduce aflatoxin contamination to levels required for acceptance of the contaminated food as fit for consumption is still unknown. There is no question that biocontrol strains are able to reduce the size of the populations of aflatoxin-producing strains but the available data suggests that at most only a four- to five-fold reduction in aflatoxin contamination is achieved.There are many challenges facing this strategy that are both short term and long term. First, the population biology of A. flavus is not well understood due in part to A. flavus's diversity, its ability to form heterokaryotic reproductive forms, and its unknown ability to survive for prolonged periods after application. Second, biocontrol strains must be selected that are suitable for the environment, the type of crop, and the soil into which they will be introduced. Third, there is a need to guard against inadvertent introduction of A. flavus strains that could impose an additional burden on food safety and food quality, and fourth, with global warming and resultant changes in the soil nutrients and concomitant microbiome populations, the biocontrol strategy must be sufficiently flexible to adapt to such changes. Understanding genetic variation within strains of A. flavus is important for developing a robust biocontrol strategy and it is unlikely that a "one size fits all" strategy will work for preharvest aflatoxin reduction.

**Keywords: aflatoxin,** *Aspergillus flavus***, biocontrol, food safety, recombination, maize, cottonseed, population diversity**

#### **DEVELOPMENT OF THE BIOCONTROL STRATEGY**

Fungal growth on agricultural commodities, with or without mycotoxin production, does not occur in pure culture. Early studies found that aflatoxin production by *Aspergillus flavus* is reduced when it is cultivated with certain other fungi and bacteria (Ashworth et al., 1965; Weckbach and Marth, 1977; Wicklow et al., 1980, 1988; Horn andWicklow, 1983; Ehrlich et al., 1985). The soil microbiome, mainly fungi and bacteria, affects the ability of the fungi to produce secondary metabolites (Ashworth et al., 1965). It has long been known that some plants are never contaminated with aflatoxin even though aflatoxin-producing species are present in the soil (Wogan, 1966). However, when tissues of these resistant plants are sterilized, *A. flavus* has no trouble producing aflatoxin on the tissue (Ehrlich and Ciegler, 1985). The ability of a fungus to compete for a host depends on many factors including pH, soil type, nitrogen and carbon availability, and water and mineral content (Eugenio et al.,1970). In 1975 an outbreak of *A. flavus* contamination of field corn in Iowa led to an unexpectedly low level of aflatoxin contamination of the crop (Ehrlich et al., 1985). Wicklow and others noticed that co-culture of *A. flavus* with *A. niger* caused a marked reduction in the formation of aflatoxin beyond a simple displacement of one fungus by the other (Wicklow et al., 1980). We found that co-cultivation of *A. flavus* with *P. oxalicum* had a similar effect on aflatoxin production but, in this case, not only were aflatoxin amounts reduced to a level not accountable by simple displacement, but a metabolite of *P. oxalicum*, secalonic acid, was also reduced (Ehrlich et al., 1985). We speculated that the co-cultivation of the two organisms caused a competition for ATP that is needed for one or more of the oxidation steps in secondary metabolite formation. The order of inoculation was also important for determining which of the two competing fungi was successful in reducing mycotoxin production. Inoculation with the *P. oxalicum* first greatly inhibited *A. flavus* production of aflatoxins even when eventual growth was similar. We also showed that *A. flavus* mutants reduced in their abilities to produce aflatoxins also showed similar competitor ability to that of *P. oxalicum* (Ehrlich, 1987). Recent work (see below) has offered a different interpretation of the mechanism behind these competition results (Sweany et al., 2011).

This early work supported the concept that competition with *A. flavus* isolates incapable of aflatoxin production could remediate aflatoxin contamination. This mode of biocontrol is currently the most widely used biocontrol method for reducing aflatoxin contamination of cereal crops in maize and cottonseed where aflatoxin contamination is a persistent problem for human and animal health (Wu and Khlangwiset, 2010a). Cotty and co-workers in the 1990s found that one particular non-aflatoxigenic *A. flavus* isolate (AF36) isolated from Arizona cottonseed, is an especially good competitor for reduction of aflatoxin content of cottonseed (Cotty and Bhatnagar, 1994). They determined that isolates from this VCG group were generally good as competitors against aflatoxin-producing isolates (Chang et al., 2012). We subsequently found that this isolate was unable to produce aflatoxin because of a point mutation in the polyketide synthase gene that is necessary for aflatoxin biosynthesis (Ehrlich and Cotty, 2004).

Cotty and co-workers developed a method to apply the nonaflatoxigenic strain to cotton-growing fields to prevent aflatoxin production by the wild-type aflatoxigenic populations present in the growing regions (Cotty, 2006; Cotty and Mellon, 2006). The method involves spreading non-aflatoxigenic *A. flavus* spores onto the field at particular times prior to harvest (Jaime-Garcia and Cotty, 2007). They assumed that addition of the non-aflatoxin producing strain would then allow it to out-compete the wildtype populations for access to the cottonseed and thereby displace the wild-type fungus. Therefore, based on this concept they called this strategy a "displacement" strategy for biocontrol of aflatoxin contamination (Jaime-Garcia and Cotty, 2009). Assuming this is true than the resulting treated field should never have to be treated again because, then, only the non-aflatoxigenic population of fungi would be present in the field. This method to prevent aflatoxin contamination is now in widespread use in Arizona for cotton fields and in other places in the southeast U.S for treatment of maize-growing areas. In other countries where aflatoxin contamination of maize is an endemic problem such as Kenya and parts of China, other strains have been discovered that are being used for aflatoxin remediation. In some cases the competing fungi are used as "cocktails" that include application of multiple strains of non-aflatoxigenic *A. flavus* (Wu et al., 2013).

There is a growing awareness that *A. flavus* also produces an indole tetramic acid mycotoxin, cyclopiazonic acid (CPA), under the same conditions that it produces aflatoxin (Chang and Ehrlich, 2011). CPA is a specific inhibitor of sarcoplasmic and endoplasmic reticulum calcium-dependent ATPase, an enzyme necessary for proper muscle contraction and relaxation. Therefore, nonaflatoxigenic competitor isolates incapable of production of this metabolite are being used instead of AF36 for reduction of aflatoxin levels in maize (Abbas et al., 2011). In a subsequent section we will discuss the possibility that other mycotoxins are produced by both the non-aflatoxigenic and aflatoxigenic populations that could contribute to mycotoxin contamination and cause toxic effects in humans and animals upon consumption.

Although the biocontrol strategy for aflatoxin remediation is increasingly being adopted world wide, there are several potential pitfalls that should be addressed These include the need to better understand the natural diversity of *A. flavus* populations in agricultural soil, the effects of climate change on both this diversity and on plant susceptibility, the ability of the introduced biocontrol strain to outcross with existing aflatoxin-producing *A. flavus*, the adaptation of certain *A. flavus* isolates for predominant growth on the plant rather than in the soil, the difficulty in timing the application or controlling the stability of the inoculum, how the introduction of the biocontrol strain affects the soil microenvironment, the potential damage to the plant from the introduced strain, and the need to better understand the entire *A. flavus* toxin burden that may result from *A. flavus* contamination beyond that of aflatoxin. In addition the cost of the biocontrol method and the potential need to continue reapplication seasonally must also be considered in weighing the benefits of biocontrol *A. flavus* as a means of reducing food and feed contamination with aflatoxin.

#### **EFFICACY OF BIOCONTROL BY NON-AFLATOXIGENIC ISOLATES OF** *A. flavus*

More than 100 countries have enforced or proposed regulations for levels of aflatoxin in feeds and foods (Wu et al., 2013). Because these levels are so low, the regulations place a strong burden on grain intended for export. Some of these requirements are listed in **Table 1**. Normally maize contains only low levels of aflatoxin and usually meets these requirements (Yu et al., 2008), but in years with severe outbreaks of *A. flavus*, contamination levels can exceed 100–200 ppb (Johansson et al., 2006). In cottonseed grown in Arizona levels of aflatoxin frequently exceed the levels permitted for commerce and remediation by either diluting the contaminated meal with less contaminated grain or by chemical treatment to destroy aflatoxins is often necessary. **Table 2** presents a summary of results from several laboratories showing aflatoxin concentrations after treatments with several different biocontrol *A. flavus* and of several different crops. In some studies the reported reduction in aflatoxin content in treated versus untreated fields is as much as 20-fold (Dorner et al., 2003; Dorner, 2009, 2010; Abbas et al., 2011; Zanon et al., 2013). These data are for experimental laboratory studies where treatments were presumably done under optimized and highly controlled conditions. Generally a 5- to 20 fold reduction in aflatoxin levels would be sufficient for allowing the crop to meet the standards for consumption, but if the starting levels are particularly high, even a 20-fold reduction may not be enough. Recently, in several maize-growing regions in Kenya there have been reports of aflatoxin poisoning in humans (Probst et al., 2012; Wagacha et al., 2013; Yard et al., 2013). In these cases, ingestion was of maize that was contaminated after harvest by improper storage. It is not clear that a pre-harvest biocontrol strategy would be able to prevent such exposure and there may be simpler and more cost-effective methods to prevent ingestion of post-harvest contaminated maize.

#### **LIFE CYCLE AND DIVERSITY OF** *A. flavus*

*Aspergillus flavus* is the most common species associated with aflatoxin contamination of agricultural crops (Cotty et al., 1994; Cotty, 1997). *A. flavus* is found in temperate and tropical regions in soil and, in agricultural areas, most commonly, on maize, cotton, tree, and ground nuts (Samuel et al., 2013) and less frequently on rice (Chen et al., 2013). *A. flavus* populations are highly diverse and their stability in the soil and on the plant is not well understood. An atoxigenic relative of *A. flavus*, *A. oryzae*, is widely used in soybean and rice fermentation (Chang and Ehrlich, 2010). It is now increasingly clear that *A. oryzae* is not a separate species but actually is only one of many examples of atoxigenic variants of *A. flavus* (Geiser et al., 2000; Chang et al., 2006). Other aflatoxin-producing

**Table 1 | Allowable levels of aflatoxins in foods and feeds.**


#### **Table 2 | Efficacy of biocontrol treatments.**


fungi have been implicated in contamination of agricultural commodities. *A. parasiticus* has been associated with contaminations of peanuts in the United States (Horn, 2005), Argentina (Vaamonde et al., 2003), and West Africa (Ismail, 2001), but generally, the predominant contaminating organism is *A. flavus* (Cotty et al., 1994). *A. flavus* appears to be more invasive and out-competes *A. parasiticus* when both species are together in the soil. *A. nomius* is more rarely found in the soil, and usually is not associated with agricultural contamination episodes (Cotty et al., 1994; Bhatnagar et al., 2001; Cardwell and Cotty, 2002). Mis-identification of the contaminating organism, in some cases is possible. For example, in Thailand, some aflatoxin B-and G-producing organisms, found to be common in the soil resemble *A. flavus*, but have been conclusively identified as a new clade of *A. nomius* (Ehrlich et al., 2007a).

*A. flavus* is a diverse assemblage of strains which include toxin-producing and non-toxigenic strains, sclerotial type variants, strains with variability in response to light, strains residing in multiple vegetative compatibility groups (VCGs), and strains with variable ability to colonize living plant tissue. A cladogram showing *A. flavus* diversity is shown in **Figure 1**. As a predominantly saprophytic fungus, *A. flavus* resides in the soil, but as an opportunist it is readily able to colonize most environments whenever there is a rich source of carbon and nitrogen. *A. flavus*'s diversity, therefore, appears to be an evolutionary response to its cosmopolitan distribution. Its main mode of replication is by asexual sporulation but under some conditions, *A. flavus* forms sclerotia, hardened masses of desiccated and melanized mycelia that are able to survive adverse environmental and nutritional conditions.

*Aspergillus flavus* soil populations contain isolates from two morphologically distinct sclerotial size variants, termed the L-strain [also called *A. flavus* Group IB (Geiser et al., 2000)] for isolates with average sclerotial size >400 μm and the S-strain (Group IA) for isolates with sclerotial size less that 400 μm (Cotty, 1997). Both S-and L-strains of *A. flavus* are found globally in maize-growing regions of the world. On typical laboratory growth media, when grown in the dark, S-strain isolates produce higher levels of aflatoxins, more abundant sclerotia, and fewer conidia. Atoxigenic S-strain isolates are very rarely found in natural environments (Orum et al., 1997). *A. flavus* lacks the ability to produce G-aflatoxins due to a gap in the cluster that includes a required cytochrome P450-encoding gene, *cypA*. The size of the deletion that causes loss of a portion of *cypA* is 1.5 kb for S-strain isolates and 0.8 kb for L-strain isolates. Differences in sclerotial morphology correlate with the differences between the S- and L-strain *A. flavus* in the size of the deletion in the *norB*-*cypA* gene (Ehrlich et al., 2004). Soil populations of *A. flavus* are typically composed of isolates from hundreds of different VCGs. Although frequent genetic exchange among these groups has not been observed, historical recombination in populations probably has occurred. Because the 0.8 kb deletion in S-strain isolates is identical to the deletion in those *A. oryzae* isolates that possess most of the aflatoxin cluster, such isolates may have descended from a common ancestor that had the S-strain-type *norB*-*cypA* gene deletion (Chang et al., 2005). On average 30% of the *A. flavus* soil isolates in Arizona were identified as belonging to the S-strain (Cotty, 1997; Orum et al., 1997). Because S-strain isolates consistently produce more aflatoxin than L-strain isolates and aflatoxin production in this strain is not as strongly affected by nitrogen source, the concentration of S-strain isolates in the soil appears to be better correlated with major outbreaks of aflatoxin contamination in cotton-growing areas in Arizona and Texas (Orum et al., 1997; Jaime-Garcia and Cotty, 2006). Furthermore, up to 40% of the L-strain soil isolates of *A. flavus* found in Arizona and other regions of the United States (Horn and Dorner, 1999) were incapable of producing aflatoxins while S-strain isolates rarely were atoxigenic (Cotty et al., 1994). Interestingly, Two of the biocontrol strains used in the United States, AF36 and K49, have the S-strain type *norB-cypA* deletion which may correlate with their competitor abilities (Chang et al., 2012).

#### **POPULATION DYNAMICS OF** *A. flavus* **IN AGRICULTURAL ENVIRONMENTS**

*Aspergillus flavus* was considered to be incapable of forming a sexual state and therefore was expected to maintain an entirely asexual life-style. Populations are divided into VCGs. Vegetative compatibility was believed to be a strong barrier to genetic exchange and in *A. flavus* was thought to be controlled by as many as 12 genetic loci (Bayman and Cotty, 1991; Ehrlich et al., 2007b). We found that

*A. flavus* isolates from different VCGs formed genetically distinct groups suggesting that recombination is at most an infrequent event (Ehrlich et al., 2007b). Genetic isolation has been suggested by an additional study thatfound no evidence of gene flow between VCGs, including VCGs of opposite mating-type. Their results suggest that the VCGs diverged before domestication of agricultural hosts (> 10,000 year before the present; Grubisha and Cotty, 2009).

Recently *A. flavus* and many other presumed asexual fungi have been found to be capable of sexual reproduction, when grown in the dark under nutrient deprived conditions. *A. flavus*, as a heterothallic fungus, has two mating type loci, Mat1-1 and Mat1-2, maintained separately in homokaryotic isolates (Ramirez-Prado et al., 2008). Early evidence from genetic analysis suggested that

*A. flavus* populations are able to undergo recombination (Geiser et al., 1998). Recent studies found that *A. flavus* in different VCG are able to outcross, and that VCG is not a strong barrier to sexual recombination (Olarte et al., 2012a). In fact such outcrossing among VCGs leads to new VCGs, and thereby, increased diversity (Olarte et al., 2012b). Most of these recombination studies have been done under laboratory conditions. Recombination can occur within conidia or sclerotia when they harbor multiple nuclei of different mating type. Fusion of nuclei containing different fluorescent markers revealed that, while conidial populations are predominantly homokaryotic, a small percentage can become heterokaryotic and, thereby, capable of recombination. The frequency of mating-type genes in the population was found to be correlated with recombination in the aflatoxin gene cluster.

Recombination has been detected between aflatoxigenic and non-aflatoxigenic *A. flavus* with some of the offspring regaining the ability to produce aflatoxins (Olarte et al., 2012a; Horn et al., 2013). Clearly, such recombination is a source of diversity within *A. flavus*. Because of this ability to recombine, it is critical to assess the frequency of such events in agricultural environments where atoxigenic biocontrol *A. flavus* have been introduced. A recent study found that, under an agricultural environment, a small percentage of the sclerotia that can form on contaminated maize can be heterokaryotic if the seed is contaminated with isolates of both mating types. Upon contact with non-sterile soil, these sclerotia can develop into ascocarps, the sexual reproductive developmental forms (Horn et al., 2013). A separate study found that soil populations in agricultural environments that were not treated with biocontrol *A. flavus* had approximately equal populations of fungi of both mating types. The population of fungi obtained from the plant (maize) was skewed to overrepresent isolates with Mat1-2 loci fungi (Sweany et al., 2011). These recent studies show that both asexual/sexual reproduction and ecological factors influence recombination.

It has been shown that the populations of *A. flavus* in an agricultural environment contain abundant amounts of nonaflatoxigenic *A. flavus* (Horn and Dorner, 1999). This suggests that loss of aflatoxin-producing ability in *A. flavus* could be a consequence of adaptation to a carbon-rich environment that makes the aflatoxin cluster less genetically stable. The ability to produce aflatoxins (and other mycotoxins) may give the fungi a long-term advantage over a non-aflatoxigenic biocontrol strain for survival in the soil, but in agricultural environments this adaptive pressure may be partially lost. Larger effective population sizes tend to increase mean population mutation and recombination rates (Hartl and Clark, 1997), further driving the evolution of new VCGs, some of which have lost aflatoxinproducing ability due to mutations within the biosynthetic cluster or due to large chromosomal deletions resulting in losses of entire telomeric regions (Chang et al., 2006). Since the aflatoxin and CPA clusters reside near the telomere of chromosome 3 in *A. flavus*, such mutations result in a high frequency of loss of aflatoxin and CPA-producing ability. The detection of linkage disequilibrium blocks in partial clusters indicates that recombination has played a large a role in cluster disassembly, and multilocus coalescent analyses of cluster and non-cluster regions indicate lineage-specific gene loss in*A. flavus* (Moore et al., 2007)*.*

The long-term fate of the non-aflatoxigenic biocontrol strain in the agricultural environment has not yet been fully addressed. A preliminary report found that in cotton fields treated with biocontrol *A. flavus* the introduced biocontrol isolate while the highly toxigenic strain S increased to reach an equilibrium in which the population of the biocontrol strain was about 10% that of the aflatoxin-producing isolate after 4 years. After only 1 year, the soil of treated fields had *A. flavus* populations with greater than 50% of the biocontrol isolate. This result suggests that long-term longevity of the biocontrol *A. flavus* could be an important consideration in establishing treatment protocols (Jaime-Garcia and Cotty, 2013). Survival of sclerotia in soil has been studied (Wicklow et al., 1993). After 36 months exposure to an agricultural soil, 68–100% of the sclerotia survived, with the main loss being due to nematode fungivory (Wicklow et al., 1993; McCormick, 2013). While soil sclerotia are largely stable in soil, the conidial inoculum is less stable and is subject to losses due to ingestion and degradation by soil bacteria and that formulations not using wheat or barley as a carrier are desirable (Accinelli et al., 2009).

#### **OTHER SECONDARY METABOLITE GENE CLUSTERS IN** *A. flavus*

*Aspergillus flavus* is able to produce toxic secondary metabolites in addition to aflatoxins. This suggests that caution is needed in considering what isolates should be used as non-aflatoxigenic biocontrol agents (Rank et al., 2012). Among these secondary metabolites are the indole-diterpenes, aflatrem, paxillenes, paspalicines, and aflavinines, as well as is the ergot-like alkaloids CPA and pseurotin (**Figure 2**). While none of these metabolites is currently regulated as a food or feed contaminant, toxicity studies indicate that they could have neurotoxic and nephrotoxic effects on animals. Other metabolites are also frequent metabolites of *A. flavus*, including the Substance P neurotransmitter antagonist, ditryptophenaline. We have recently determined that some of these metabolites are produced in greater quantities in S strain *A. flavus* than in L strain and are produced by some of the non-aflatoxigenic competitor strains. The reported toxic effects on humans of ingestion of *A. flavus*-contaminated maize was growth retardation, immune suppression, and liver damage, the latter being manifested the most in people with hepatitis C infection (Probst et al., 2007; Probst et al., 2010; Wu and Khlangwiset, 2010b). These toxic effects have usually been ascribed to ingestion of aflatoxins. We suggest that simultaneous ingestion of other toxic *A. flavus* metabolites may contribute to these observed toxicities in people who have eaten aflatoxin-contaminated maize. The S morphotype *A. flavus*, the *Aspergillus* strain is most associated with the recent outbreaks of toxicity to humans in Kenya and Nigeria and may be far more toxic than the L strain (Donner et al., 2009; Mehl and Cotty, 2010; Probst et al., 2010). *A. flavus* is also able to produce metabolites that are usually not considered to be particularly toxic but could affect animal health. These include metabolites such as orcellinic acid, aspergillic acid and kojic acid (Varga et al., 2012) as well as iron-chelating siderophores similar to ferricrocin (Wallner et al., 2009). What effect these additional metabolites might have on animal and plant health is unknown.

#### **GLOBAL WARMING AND BIOCONTROL**

Global warming has increased daily high temperatures in the mid west and northern maize-growing regions of the United States and Canada. The resulting temperatures are predicted to eventually resemble those in the southern United States where aflatoxin contamination of maize is a frequent problem. Aflatoxin contamination of mid west maize has not been recognized as a problem as yet. Besides temperature shifts, global warming can cause climate changes which result in more unpredictable weather problems for agricultural areas. Aflatoxin contamination events are more prevalent during times of high heat and drought,

which may stress the host plant thereby facilitating *A. flavus* infection (Schmidt-Heydt et al., 2009; Roze et al., 2012; Mohale et al., 2013; Reverberi et al., 2013). Fungal stress has been correlated with increased expression of genes involved in both secondary metabolism production and sexual recombination as discussed above. Agricultural areas experiencing drought often suffer aflatoxin contamination outbreaks, and unpredictable changes in climate that result in drought may occur with increased frequency. Currently, incidences of aflatoxin contamination of crops are limited to tropical and sub-tropical areas (between latitudes 40◦N and 40◦S) around the world (Samuel et al., 2013). Because the average global surface temperature has increased by 0.8◦C since 1901, with most of that increase occurring in the last 30 years, it

is possible that by the end of the 21st century the favorable climate for aflatoxin contamination may encompass more of the maize-growing regions of the U.S. and outbreaks will become more frequent in occurrence.

Another potential consequence of climate change is that the biocontrol strain could be an inadvertent cause of increased damage to the plant, especially if growing conditions are less favorable for cultivation. Concomitantly, changes in the soil environment and its microbiome due to temperature elevation, could also subject the crop to increased damage. Understanding genetic variation within strains of *A. flavus* is important for developing a robust biocontrol strategy and it is unlikely that a"one size fits all" strategy will work.

#### **CONCLUSION**

The ultimate goal for using non-aflatoxigenic *A. flavus* as a biocontrol agent should be the long-term protection of crops against aflatoxin contamination. Current strategies utilize a program of annual re-application of biocontrol strains, and the fate of the biocontrol strains after one growing season is still unknown. Even a low rate of recombination for aflatoxigenic fungi could be significant for future food safety. There exist other challenges to the biocontrol strategy for remediation of aflatoxin contamination. The inherent diversity of *A. flavus* populations makes a biocontrol strategy more difficult because *A. flavus* populations differ in their abilities to produce aflatoxins and other toxic secondary metabolites. Some of these other secondary metabolites could be important for assessing the full toxic burden when grains contaminated with *A. flavus* are ingested*.* Climate change could increase stress on the plant and the fungus and environmental stress could increase plant susceptibility to the fungus and is a known inducer of secondary metabolite production. Stress could also affect the ability of the fungus to outcross with native populations of *A. flavus*. Also, in use of the biocontrol *A. flavus* care must be taken to prevent undue crop damage or damage to the soil microflora that might result.

#### **ACKNOWLEDGMENTS**

I thank my co-workers who contributed to the increased understanding of the population biology of *Aspergillus flavus*, in particular, Peter Cotty, Perng-Kuang Chang, Jeffrey Cary, Deepak Bhatnagar, and Geromy Moore. I also thank Ken Damann (LSU Baton Rouge) and Ignazio Carbone North Carolina State University) for their contributions to understanding the fundamentals of population diversity of *A. flavus*.

#### **REFERENCES**


and propagule density in Yuma County, arizona, soils. *Plant Dis.* 81, 911–916. doi: 10.1094/PDIS.1997.81.8.911


Zanon,A., Chiotta,M., Giaj-Merlera, G., and Barros, G. (2013). Evaluation of potential biocontrol agent for aflatoxin in Argentinean peanuts. *Int. J. Food Microbiol.* 162, 220–225. doi: 10.1016/j.ijfoodmicro.2013.01.017

**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: 18 December 2013; paper pending published: 14 January 2014; accepted: 23 January 2014; published online: 10 February 2014.*

*Citation: Ehrlich KC (2014) Non-aflatoxigenic Aspergillus flavus to prevent aflatoxin contamination in crops: advantages and limitations. Front. Microbiol. 5:50. doi: 10.3389/fmicb.2014.00050*

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

*Copyright © 2014 Ehrlich. 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.*

### *Aspergillus flavus* infection induces transcriptional and physical changes in developing maize kernels

#### *Andrea L. Dolezal 1, Xiaomei Shu2, Gregory R. OBrian2, Dahlia M. Nielsen3, Charles P. Woloshuk4, Rebecca S. Boston5 and Gary A. Payne2 \**

*<sup>1</sup> Monsanto Company, Waterman, IL, USA*

*<sup>2</sup> Department of Plant Pathology, North Carolina State University, Raleigh, NC, USA*

*<sup>5</sup> Department of Plant and Microbiological Sciences, North Carolina State University, Raleigh, NC, USA*

#### *Edited by:*

*Perng-Kuang Chang, Southern Regional Research Center, USA*

#### *Reviewed by:*

*Ann E. Stapleton, University of North Carolina Wilmington, USA Jeffrey William Cary, United States Department of Agriculture, ARS, SRRC, USA*

#### *\*Correspondence:*

*Gary A. Payne, Department of Plant Pathology, North Carolina State University, 851 Main Campus Drive, Raleigh, NC 27695-7567, USA e-mail: gary\_payne@ncsu.edu*

Maize kernels are susceptible to infection by the opportunistic pathogen *Aspergillus flavus*. Infection results in reduction of grain quality and contamination of kernels with the highly carcinogenic mycotoxin, aflatoxin. To understanding host response to infection by the fungus, transcription of approximately 9000 maize genes were monitored during the host-pathogen interaction with a custom designed Affymetrix GeneChip® DNA array. More than 4000 maize genes were found differentially expressed at a FDR of 0.05. This included the up regulation of defense related genes and signaling pathways. Transcriptional changes also were observed in primary metabolism genes. Starch biosynthetic genes were down regulated during infection, while genes encoding maize hydrolytic enzymes, presumably involved in the degradation of host reserves, were up regulated. These data indicate that infection of the maize kernel by *A. flavus* induced metabolic changes in the kernel, including the production of a defense response, as well as a disruption in kernel development.

**Keywords:** *Aspergillus flavus***, maize, transcription, genetic, aflatoxins, pathogenesis**

#### **INTRODUCTION**

*Aspergillus flavus* is an opportunistic fungal pathogen that infects developing maize kernels, attacking plants that are weakened by environmental stresses such as drought and heat. Disease reduces grain quality and contaminates the kernel with the carcinogenic mycotoxin aflatoxin (Scheidegger and Payne, 2003; Payne and Yu, 2010; Dolezal et al., 2013; Hruska et al., 2013; Kew, 2013). The development of resistant maize lines has proven difficult although there is evidence for sources of resistance (Brown et al., 1999; Windham and Williams, 2002; Mylroie et al., 2013; Warburton et al., 2013; Mideros et al., 2014). The lack of reliable resistance phenotyping markers, the inconsistency of disease development each year, and an insufficient understanding of host resistance mechanisms, all have made the selection of resistance difficult.

Advances in technology, such as microarrays, have enabled researches the ability to monitor transcription on a genomewide level and provided a better understanding of how organisms respond to their environment on a cellular level. Studies investigating plant gene expression during pathogen attack have found the defense response goes beyond PR-proteins and involves transcription changes in both primary and secondary plant metabolic pathways and detoxification pathways (Boddu et al., 2007; Doehlemann et al., 2008; Alessandra et al., 2010). Phytohormones like salicyclic acid (SA), jasmonic acid (JA), ethylene (ET) have long been known to be an integral part of the defense response (Glazebrook, 2005; Jones and Dangl, 2006; Robert-Seilaniantz et al., 2011). Yet carbohydrate metabolism pathways, though not typically associated with resistance, may be an important component of the plant defense response including in maize (Berger et al., 2007; Bolton, 2009). Higher maize stalk carbohydrate levels have been associated with increased resistance to stalk infecting fungi, many of which are also capable of infecting the ear and kernel (Dodd, 1980).

Transcriptional changes of maize kernels during infection by *A. flavus* have been studied using microarrays (Luo et al., 2011; Kelley et al., 2012) and qPCR (Jiang et al., 2011). Kelley et al. (2012) compared maize varieties that were either susceptible or resistant to aflatoxin accumulation. They found 16 genes highly expressed in the resistant variety and 15 in the susceptible variety and concluded that multiple mechanisms are likely involved in resistance to aflatoxin accumulation. Jiang et al. (2011) reported higher levels of gene expression in stress related genes in resistant lines of maize. Luo et al. (2011) found that more maize genes were induced by *A. flavus* in susceptible kernels compared with resistant kernels. In all these studies, defense-related, and regulatory genes were associated with the response to *A. flavus*. To provide a clearer understanding of maize kernel resistance to *A. flavus* we monitored the transcriptional response of maize kernels during infection by *A. flavus* in the field using a custom DNA microarray. We report changes in expression of well-characterized defense signaling pathways and defense related genes as well as striking changes in expression of genes related to carbohydrate metabolism.

*<sup>3</sup> Department of Genetics, North Carolina State University, Raleigh, NC, USA*

*<sup>4</sup> Department of Botany and Plant Pathology, Purdue University, West Lafayette, IN, USA*

There are several stages in the infection process that host resistance could restrict fungal growth and aflatoxin contamination. Kernel infection with *A. flavus* begins through silk colonization. Conidia germinate and grow on senescing silks, moving down the silk channel to the developing kernels, which can take as little as 8 days (Marsh and Payne, 1984; Payne et al., 1988b). Subsequent steps in the infection process are less defined, but data suggest that *A. flavus* can attack kernels during their six stages (Ritchie et al., 1997) of their development: silking (R1), blister (R2), milk (R3), dough (R4), dent (R5), and physiological maturity (R6). Recently, Reese et al. (2011) inoculated detached kernels at stages R2–R5 in the lab and found that kernels at these four stages were susceptible to infection by *A. flavus*. Fungal infection has been observed in injured kernels as young as the milk (R3) stage (Taubenhaus, 1920; Anderson et al., 1975). These young kernels tend to accrue high concentrations of aflatoxin because of prolonged colonization by the pathogen (Lillehoj et al., 1980; Payne et al., 1988a). Infection in non-injured kernels in the field is thought to take place later, during the dent (R5) developmental stage just prior to physiological maturity (R6) (Koehler, 1942; Marsh and Payne, 1984; Payne et al., 1988a; Smart et al., 1990; Windham and Williams, 1998). Once inside, *A. flavus* preferentially colonizes the oil-rich germ tissue (Fennell et al., 1973; Jones et al., 1980; Smart et al., 1990; Keller et al., 1994). Fungal growth within endosperm tissue, more specifically the nutrient-rich starchy endosperm, has been observed, but there are discrepancies in the literature as to the extent of colonization (Lillehoj et al., 1976; Smart et al., 1990; Keller et al., 1994; Brown et al., 1995; Dolezal et al., 2013).

Our studies focused on the transcriptional response of developing kernels that were inoculated with *A. flavus* through a wound. We realize that this approach could overlook some resistance mechanisms, but it results in more consistent disease development. Resistance to infection of wounded kernels is also relevant as it mimics insect injury, which is important in the epidemiology of the disease. Furthermore, to capture the response in the different stages of kernel development, we evaluated *A. flavus* infection of four kernels stages, R2–R5. We also chose a specific time of 4 days after inoculation to examine gene expression based on previous histological studies by Dolezal et al. (2013) who showed that within 4 days after inoculation *A. flavus* mycelium reached the aleurone, endosperm, and germ tissue. Thus, sampling at 4 days allowed assessment of host response in several tissue types within the kernels.

#### **EXPERIMENTAL PROCEDURE**

#### **FUNGAL STRAIN AND CULTURE CONDITIONS** *ASPERGILLUS FLAVUS*

NRRL 3357 was grown on potato dextrose agar (PDA) at 28◦C for 7–10 days. Conidia were dislodge with 0.05% (v/v) Triton X-100 and diluted to a working solution of 1 <sup>×</sup> <sup>10</sup><sup>6</sup> spores mL−1.

#### **MAIZE KERNEL INOCULATION AND HARVESTING**

Inbred maize genotype B73 was grown at the Central Crops Research Station in Clayton, NC. Ears were hand pollinated and the date recorded on the bag. Ears at the blister (R2), milk (R3), late milk (R3)–early dough (R4), dough (R4), and dent (R5) stages of development were either mock-inoculated or inoculated with *A. flavus* as outline in Dolezal et al. (2013). Briefly, ears selected for inoculation had the husk pulled back to exposed the developing kernels below. The protruding portion of the pins of the pinbar was dipped into the *A. flavus* conidial suspension and inserted into the crown of the kernel. The husk was repositioned and secured around the ear with a rubber band, and a paper bag placed over the inoculated ear. Ears inoculated at the blister (R2), milk (R3), dough (R4), and dent (R5) stages of development were removed from the plant 4 days after inoculation (dai), and the kernels flash frozen immediately after removal. Harvested kernels were stored at −80◦C until RNA was extracted using the protocol outlined in Smith et al. (2008). Additional ears inoculated at the late milk (R3)–early dough (R4) stage of development were left in the field and picked at end of the growing season. Kernels adjacent to the pinbar-inoculated rows were harvested. Kernels on noninoculated ears, pollinated the same day as the inoculated ears, were also collected and used as controls. Adjacent diseased kernels and control kernels were cut-in-half and visually compared to assess for physical changes in kernel structure resulting from *A. flavus* infection.

#### **MICROARRAY PROCESSING AND ANALYSIS**

Custom-designed *A. flavus* Affymetrix GeneChip DNA microarrays were used to identify genes differentially expressed in maize during *A. flavus* kernel colonization. This multi-species array, in addition to being capable of monitoring genome-wide transcription of *A. flavus*, has close to 9000 probe sets representing maize genes. This pairing of *A. flavus* and maize genes onto a single array allowed for simultaneous detection of disease-associated transcript in the plant-pathogen interaction. The majority (83%) of maize genes selected for the array came from seed-specific cDNA libraries. The remaining genes were chosen based on recommendations from members of the maize community and prior association with disease resistance. The quality of RNA extracted from the mock-inoculated and *A. flavus*-inoculated kernels was assessed before processing. All array work was carried out at the Purdue Genomic Core Facility (http://www.genomics.purdue. edu) in West Lafayette, IN, and standard Affymetrix protocols were followed.

CEL files generated from the GeneChip DNA microarray scans were imported into JMP Genomics and log2 transformed. Mismatched probes were not used in the calculation of the expression values. The expression profiles of *A. flavus* and maize genes were examined for each array, and arrays for the mock-inoculated treatment that had moderate-to-strong *A. flavus* signal intensities were removed from further analysis. While these kernels did not visually appear infected, they were likely inadvertently contaminated with *A. flavus*. Data were then normalized using Loess Normalization. Normalized data from arrays generated from blister (R2), milk (R3), dough (R4), and dent (R5) inoculated kernels stages were grouped into either a mock-inoculated or *A. flavus* inoculated treatment group. The assemblage of the different developmental stages into a single treatment group allowed for the identification of maize genes that consistently responded to *A. flavus* infection regardless of what age infection initiated. An analysis of variance (ANOVA) was performed comparing the mock- and *A. flavus*-inoculated treatment groups. To account for multiple testing, a significance threshold based on a false discovery rate (FDR) of 0.05 was used (Benjamini and Hochberg, 1995). The data were deposited into Gene Expression Omnibus. The series record number is GSE57629 (http://www. ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE57629). A volcano plot was generated showing significance on the y-axis and fold change on the x-axis using JMP 11 (**Figure 1**). Gene names were assigned by using Tophat to align the affy probe sequences to the ZmB73\_RefGen\_v2 reference genome. AgriGO was used to perform Singular Enrichment Analysis (SEA) on differentially expressed genes (Du et al., 2010).

For SEA, the AGRIGO toolkit was used (http://bioinfo.cau. edu.cn/agriGO/analysis.php). Default values were used for the advanced options including the Yekutieli (FDR under dependency) multi-test adjustment method at a significance level of 0.05.

#### **VALIDATION OF MICROARRAY DATA BY qRT-PCR**

For each of the developmental stages used for the microarray study, a second set of RNA isolations was performed. RNA was treated with DNase (Promega) and cDNA was synthesized using a First Strand cDNA Synthesis Kit (Fermentas). Quantitative realtime RT-PCR (qRT-PCR) was performed using a SYBR® Green kit (Applied Biosystems) according to the manufacturer's instructions. The expression levels of a ribosome gene were used for normalization. Data were analyzed by the comparative CT method with the amount of target given by the calibrator 2−--CT. The primers used for qRT-PCR analysis are listed in **Table 1**.

#### **RESULTS**

#### **DIFFERENTIALLY EXPRESSED GENES**

Maize kernels were inoculated at the blister (R2), milk (R3), dough (R4), and dent (R5) stages, and harvested 4 days later.

Transcriptional changes for 8875 maize genes were monitored with an Affymetrix GeneChip® DNA array. Data were grouped into either mock-inoculated or *A. flavus-*inoculated treatment groups. An ANOVA comparing mock-inoculated with *A. flavus-* inoculated treatment groups (α ≤ 0.05 FDR) identified 912 and 3737 of the Affymetrix GeneChip probe-sets up- and down-regulated, respectively, (Table S1). Each probe-set represents a unique maize gene, except for those with a suffix attached to the probe-set name (e.g., ZM\_a\_at). These probesets may contain probes that represent more than one gene within a gene family or contains conserved sequence common to multiple genes. Some maize genes are represented by multiple probe-sets. Consult http://www.affymetrix.com/estore/support/ help/faqs/mouse430/faq8.jsp for more detail on the different suffixes. This analysis showed the differential expression of genes associated with host resistance and defense signaling pathways, and of genes associated with sugar metabolism.

#### **GENE ENRICHMENT ANALYSIS (SEA)**

To gain additional insight into the collective biological function of proteins whose genes showed differential expression during infection, we performed annotation enrichment using SEA (Du et al., 2010) on the genes listed in Table S1. The resulting list of enriched Gene Ontology terms is shown in Table S2. Notable are transcriptional changes in several genes associated with carbohydrate metabolism. Representative examples include GO:0005975, GO:0006006, GO:0034637, GO:0044262, and GO:0019318.

#### **CHANGES IN EXPRESSION OF GENES ASSOCIATED WITH CARBOHYDRATE METABOLISM**

Infection of maize kernels with *A. flavus* resulted in transcriptional changes of several maize genes involved in primary and secondary metabolism, particularly those associated with the synthesis and hydrolysis of starch, and the mobilization of hexoses (**Figure 2**; **Table 2**; Table S2). As an example, genes encoding




starch biosynthetic enzymes, including two starch branching enzymes, (GRMZM2G088753 and GRMZM2G032628), were down regulated during infection as was ADP-glucose pyrophosphorlase (GRMZM2G429899), which catalyzes a key metabolic step in the synthesis of starch in higher plants (Greene and Hannah, 1998). In addition to the apparent down regulation of starch synthesis, there was an increase in transcription of genes involved in starch hydrolysis. The transcription of a β-amylaselike genes, GRMZM2G025833, was upregulated during *A. flavus* pathogenesis.

Associated with changes in starch accumulation were changes in the mobilization of hexoses. Three maize invertases (GRMZM2G139300, GRMZM2G394450, GRMZM2G123633) were more highly expressed in the *A. flavus* infected kernels than in non-infected kernels (**Table 2**; **Figure 2**). Invertases are responsible for hydrolyzing sucrose into glucose and fructose (Cheng et al., 1996; Chourey et al., 2006), and are important in maize kernel development (Weber et al., 1997; Roitsch et al., 2003). Up regulation of invertases in the maize kernel is predicted to cause an increase in free hexose levels in the kernel and affect seed storage reserves such as starch.

The conversion of sucrose to hexoses also was associated with the down regulation of six genes involved with starch biosynthesis. Genes in the starch biosynthetic pathway [*wx1* (GRMZM2G024993), *su1* (GRMZM2G138060), *ss1* (GRMZM2 G129451), *sbe1* (GRMZM2G088753), *su2* (GRMZM2G348551), and *ae1* (GRMZM2G032628)] as well as over 20 zeinannotated genes (Table S1) and the gene encoding the transcription factor that regulates 22-kD zein expression, *opaque2* (GRMZM2G015534), were all down-regulated during infection.

It was not possible to determine the exact pathways of hexose remobilization in our studies, but gene expression in both the glycolytic pathway and the Pentose Phosphate Pathway (PPP) was altered by infection (**Figure 2**; **Table 2**). Before glucose can be utilized by either pathway, it must be phosphorylated by a hexokinase (**Figure 2**; Spielbauer et al., 2006). The fructokinase (GRMZM2G080375) was up regulated during pathogenesis. Hexose kinases have been associated with sugar sensing in plants and are potentially involved in the plant defense response (Granot et al., 2013). Glucose-6-phosphate also can be involved in starch synthesis, but because starch biosynthesis genes are down regulated it is likely used by other pathways for the production of energy or defense-related compounds during pathogenesis.

Up regulation of genes in the shikimate pathway supports the premise that hexoses are shunted away from starch synthesis in *A. flavus* infected kernels (**Figure 2**; **Table 2**). Several bioreactive compounds from this pathway are known to be involved in host defense (Daayf and Lattanzio, 2009). The shikimate pathway is an entry to aromatic secondary metabolism (Herrmann, 1995) and chorismate synthesized from this pathway is used to make the aromatic amino acids Phe, Tyr, and Trp. These amino acids are precursors for aromatic secondary metabolites including flavonoids and phytoalexins (Herrmann, 1995). Genes in the flavonoid pathway, *fht1* [(GRMZM2G062396), *c2* (GRMZM2G422750) (Bruce et al., 2000)] and genes from other phenypropanoid pathways (Table S1) increased in transcription after *A. flavus* inoculation.

The shikimic acid pathway also provides precursors for the biosynthesis of lignins (Herrmann, 1995), compounds associated with basal resistance to pathogens. Genes involved in lignin biosynthesis have been reported to be induced after *A. flavus* infection in both susceptible (VA35) (Kelley et al., 2012) and resistant varieties (Eyl25) (Luo et al., 2011) of maize. Liang et al. (2006) found lignin concentrations to increase in response to infection by *A. flavus*, and they found a negative correlation between lignin content of peanut cultivars and infection by *A. flavus*. Magbanua et al. (2013) following colonization by a GFP expressing strain of *A. flavus,* found less colonization of maize cob tissue in the resistant inbred Mp313e than in cobs of SC212m, a more susceptible genotype. They attributed the more restricted growth in Mp313e to the highly cross-linked lignin found in Mp313e. In our study, infection of maize kernels by *A. flavus* led to higher expression of three genes (GRMZM2G099420, GRMZM2G131205, GRMZ M2G090980) involved in lignin biosynthesis (Table S1).

The carbohydrate metabolic methylerythritol phosphate (MEP) pathway was also found differentially expressed during infection. The following genes from this pathway, which utilizes pyruvate from glycolysis to produces an assortment of isoprenoids including the hormone abscisic acid (ABA) were likewise up-regulated during *A. flavus* infection: (GRMZM2G056975, GRMZM2G493395, GRMZM2G172032, GRMZM2G027059, GRMZM5G859195). We found one gene, 9 cis-epoxycarotenoid dioxygenase (GRMZM2G014392), involved in ABA synthesis up-regulated. The MEP pathway expression has been found induced in maize root colonized by arbuscular mycorrhizal fungi (Lange et al., 2000; Walter et al., 2000).

#### **DEFENSE SIGNALING PATHWAYS**

Phytohormones are chemical compounds synthesized by the plant that regulate biochemical processes necessary for growth, reproduction, and survival. The plant defense response is hormonally regulated predominantly by the phytohoromones salicylic acid (SA), jasmonic acid (JA), and ethylene (ET) (Niu et al., 2011; Mengiste, 2012; Derksen et al., 2013). Each hormone likely activates different components of the defense response system that are effective against specific pathogens. The JA/ET pathways are often induced in resistance to necrotrophic pathogens, whereas the SA pathway is typically induced by biotrophic and hemibiotrophic pathogens (Glazebrook, 2005; Derksen et al., 2013).

In this study kernel infection by *A. flavus* resulted in increased expression of the 12-oxo-phytodienoic acid reductases (OPR) encoding *ZmOPR3* (GRMZM2G000236), and the alcohol dehydrogenase encoding *ts2* (GRMZM5G840653). Expression of these genes has been linked with JA biosynthesis in maize (Vick and Zimmerman, 1984; Browse, 2009), and *Ts2* has been associated with the hypersensitive response and resistance to Northern Leaf Blight in maize (Delong et al., 1993; Wisser et al., 2011). Our finding suggests that JA may be involved in the kernel-*A. flavus* interaction.

Other lipid-derived defense-related compounds besides JA are generated from enzymes generally associated with the JAbiosynthesis pathway. As an example, plants contain multiple LOX and OPR genes. In maize, the exact copy number of functional LOX genes varies between maize genotypes (De La Fuente et al., 2013). The function for most LOX and OPR isoenzymes is independent from JA biosynthesis. LOXs catalyze the formation of various oxidized lipids called oxylipins that can act as signaling molecules separate from JA and are thought to have antimicrobial properties (Blée, 2002; Prost et al., 2005). Oxylipins are known to effect fungal growth, including that of *A. flavus*, and mycotoxin production in seeds (Burow et al., 1997; Wilson et al., 2001; Brodhagen and Keller, 2006). The functional role for most OPR isoenzymes is unknown. *ZmLOX* and *ZmOPR* genes were previously found expressed in response to *A. flavus* and other maize fungal pathogen infections (Wilson et al., 2001; Zhang et al., 2005). In accordance with these findings, we observed *LOX* and *OPR* genes differentially expressed during *A. flavus* kernel colonization. *ZmLOX4*, *7*, and *9* (GRMZM2G109056, GRMZM2G070092, GRMZM2G017616) were up-regulated during *A. flavus* infection, whereas *ZmLOX11* (GRMZM2G009479) was down-regulated. *ZmOPR1*, *2*, *3*, and *5* (GRMZM2G106303, GRMZM2G000236, GRMZM2G156712, GRMZM2G087192) were up-regulated in the diseased kernel with *OPR1* and *OPR3* having a 18 and 16 fold-change, respectively, (Table S1).

Whether JA and the other lipid-derived compounds increase maize resistance against pathogen attack may depend on the pathogen and which isoenzyme is expressed. Disruption of the *ZmLOX3* results in enhanced resistance to *F. verticillioides* (Gao et al., 2009), *Colletotrichum graminicola* (Gao et al., 2007), *Cochliobolus heterostrophus* (Gao et al., 2007), and *Exserohilum pedicellatum* (Isakeit et al., 2007). However, the maize *lox3* mutant shows increased susceptibility to *A. flavus* and *A. nidulans*, indicating this gene regulates disease resistance in a pathogen-specific manner (Gao et al., 2009).

#### **DEFENSE-ASSOCIATED GENES IN** *A. FLAVUS* **INFECTED SEEDS**

Pathogenesis-related (PR) proteins are the hallmark of the induced defense response and their expression has been associated with resistance (van Loon et al., 2006; Luo et al., 2011). Several genes annotated as encoding for PR-proteins including those for chitinases [GRMZM2G112538, GRMZM2G477128, *PR-10* (GRMZM2G075283).

GRMZM2G051943, GRMZM2G129189, GRMZM2G133781, *chn2* (GRMZM2G145461), GRMZM2G145518, GRMZM2G 162359] were up-regulated in the *A. flavus* infected kernels (Table S1). The expression of *chitinase2* and *PR-10* genes has been reported to be induced in fungal infected maize seed (Cordero et al., 1994). Furthermore, studies by Chen et al. (2006) showed that PR-10 has antifungal activity against *A. flavus in vitro*, and its production is increased upon *A. flavus* infection in the resistance line GT-MAS: gk, but not in the susceptible line Mo17. They also showed that repression of maize *PR-10* by RNAi gene silencing resulted in increased susceptibility to *A. flavus* and aflatoxin production (Chen et al., 2010). A Bowman-Birk-like proteinase inhibitor (GRMZM2G156632), which encodes a PR-like protein showed a 2.5 fold increase in gene expression during infection (Rohrmeier and Lehle, 1993). This gene has been associated with the maize hypersensitive response (Simmons et al., 2002; Chintamanani et al., 2010).

The oxidative burst is an integral part of early plant immunity and is associated with reactive oxygen species, programmed cell death, and the hypersensitive response (Lamb and Dixon, 1997; Dickman and Fluhr, 2013). This defense cascade leads to the production of antimicrobial compounds. Associated with these defense responses are the production of peroxidases and gluathione-S-transferases (GST). Several genes encoding peroxidase-annotated genes (AC197758.3\_FG004GRMZM2G080183, GRMZM2G089959, GRMZM2G095404, GRMZM2G103342, GRMZM2G108207, GRMZM2G138918, GRMZM2G149273, GRMZM2G173195, GRMZM2G320269, GRMZM2G321839, GRMZM2G382379, GRMZM2G419953, GRMZM2G441541, GRMZM2G471357) were up-regulated in the diseased kernels.

Four additional peroxidase-encoding genes (GRMZM2G 034896, GRMZM2G089895, GRMZM2G103169, GRMZM2G 315176) were down-regulated, implying that only certain peroxidase-isozymes are needed during *A. flavus* infection. Gluathione-S-transferases (GST) reduce host cellular damage by detoxifying toxins and xenobiotics commonly encountered during periods of disease and abiotic stress. Wisser et al. (2011) recently correlated ZmGST23 (NP\_001104994.1) with moderate resistance to multiple maize pathogens. Though *ZmGST23* was not differentially expressed in this study, the expression of GRMZM2G01909, predicted to encode a GST, had increased expression during *A. flavus* infection.

#### **VALIDATION OF MICROARRAY DATA BY qRT-PCR**

In order to validate the results of the microarray study, the expression levels of five selected genes were monitored by qRT-PCR: (AC205274.3\_FG001, GRMZM2G139300, GRMZM2G088753, GRMZM2G109056, GRMZM2G017616). The fold changes of these genes as determined by qRT-PCR were highly correlated with the results obtained from the microarrays (**Table 3**).

#### **PHYSICAL CHANGES WITHIN KERNEL IN RESPONSE TO NATURAL INFECTION WITH** *A. FLAVUS*

The molecular analysis of maize gene expression during pathogenesis indicated major metabolic effects within kernels in response to infection by *A. flavus*. To determine if such effects could be manifest in the physical structure of kernels, we examined naturally infected kernels at the end of the growing season. Maize kernels adjacent to wound-inoculated kernels were harvested at maturity, dissected, and examined for growth of *A. flavus* and structural integrity. These kernels did not show any obvious wounds or cracks within their pericarp. **Figure 3** shows a comparison of three representative infected kernels and noninfected kernels. The most striking modification was the reduced size of the zein-filled hard [horny] endosperm in infected kernels (**Figure 3**e). In diseased kernels the hard endosperm had been replaced with starchy endosperm (**Figure 3**d), but the consistency of the entire starchy endosperm was different from that of the non-infected kernel. Instead of being firm and intact, the starchy endosperm of the diseased kernel was fragile, friable, and filled with tiny air pockets. *A. flavus* could be discerned in some of these pockets including the gap between endosperm and germ (**Figure 3**d, l). Mycelium was also observed in the embryo around the plumule (**Figure 3**g) and primary root (**Figure 3**h), and the germ was discolored and shriveled (**Figure 3**f).

**Table 3 | qRT-PCR results for select differentially expressed genes are consistent with microarray results.**


#### **DISCUSSION**

*A. flavus* kernel colonization is most aggressive on maize plants that have been subjected to heat or water stress (Tubajika and Damann, 2001; Scheidegger and Payne, 2003; Widstrom et al., 2003; Guo et al., 2008). Such conditions frequently occur throughout the world on rain-fed fields, and thus *A. flavus* colonization of kernels and subsequent contamination with aflatoxins is a concern internationally. Resistance to aflatoxin accumulation shows low heritability in the field, owing to the quantitative nature of resistance, the lack of reliable phenotyping, and strong genotype by environment (GxE) interactions. Thus, any approach that facilitates the identification of genes contributing to host resistance could accelerate the development of resistant maize genotypes.

The overall goal of this study was to better characterize host response to *A. flavus* opportunistic infection by identify maize genes differentially expressed during infection that could have applications as genetic markers in future breeding programs. Our previous research (Dolezal et al., 2013) showed that *A. flavus* follows a predictable pattern of colonization after inoculation of maize kernels. While all tissue types of the kernel can be colonized, growth of the fungus into scutellum tissue appears to follow the formation of a biofilm-like structure by *A. flavus* (Dolezal et al., 2013). The scutellum is metabolically active and known to synthesize numerous hydrolytic enzymes and defenseassociated compounds (Casacuberta et al., 1991, 1992). Based on these studies, we chose 4 days after inoculation as the time to evaluate host response to *A. flavus*. Data presented in this study indicate that infected seed at this time to be transcriptionally responsive and express genes known to be involved in host defense.

Our observations show that B73 kernels mount a multipronged defense response to *A. flavus* typical of that associated with plant basal resistance. Many of the maize genes induced by *A. flavus* are important in resistance against maize foliar pathogens, underscoring a possible commonality of the resistance response in seeds and leaves. These data further suggest that *A. flavus* infection invokes defense tactics used against more aggressive maize pathogens including increased transcription of defense signaling pathways and genes several genes known to be involved in the host's defense response.

We also observed striking changes in the transcription of genes associated with carbohydrate utilization (Table S2). An analysis of these transcriptional changes leads us to conclude that infection by *A. flavus* decreases starch synthesis, increases starch degradation, and mobilizes hexoses into pathways associated with plant defense (**Figure 2**; **Table 2**). A physical examination of kernels naturally infected with *A. flavus* (**Figure 3**) showed changes in the structure of the maize endosperm that could reflect the remobilization of hexoses in the seed in response to infection.

Physical changes in seeds infected with fungi have been observed before (Fennell et al., 1973; Koltun et al., 1974; Huff, 1980; Shetty and Bhat, 1999; Cardwell et al., 2000; Pearson and Wicklow, 2006). Most researchers have speculated that hydrolytic enzymes secreted by infecting fungi are responsible for this loss in grain quality. However, maize mutants with abnormal expression levels of carbohydrate and protein biosynthetic pathway genes can also develop atypical endosperm tissue (Neuffer et al., 1997; Black et al., 2006). Our findings showing changes in kernel primary metabolism during *A. flavus* infection challenges the assumption that fungal produced enzymes are solely responsible for changes in kernel structure, and suggests the plant may also contribute to these changes through starch degradation and hexose mobilization away from starch synthesis.

While these metabolic changes could represent a defense response by the kernel to infection by *A. flavus*, the changes could instead promote host susceptibility to the pathogen. Fungi, particularly fungal plant pathogens, are capable of manipulating the plant's metabolism to create an environment advantageous for fungal growth (Govrin and Levine, 2000; Doehlemann et al., 2008). Increased invertase transcription in *A. flavus* infected kernels could indicate a higher-than-normal accumulation of freehexoses within diseased tissue. Because glucose is the preferred carbon source of *A. flavus,* the up-regulation of sucrose hydrolyzing enzymes would presumable promote disease development by providing a steady supply of nutrients to the pathogen. IVR1 was previously found induced in sugar-poor environments, and its expression associated with tumor formation in *Ustilago maydis* infected maize (Xu et al., 1996; Doehlemann et al., 2008). Simple carbohydrates are also known to promote aflatoxin in maize kernels. Woloshuk et al. (1997) found an *A. flavus* α-amylase to play an important role in the production of aflatoxin by providing simple sugars conducive for aflatoxin production. Thus, sugar status in kernels could condition increased susceptibility as well as aflatoxin contamination.

In contrast, other studies have noted increased levels of hexoses in-and-around the site of pathogen infection and have hypothesized that these starch-derived sugars are an integral component of the host defense response (Berger et al., 2007; Bolton, 2009). Free-hexoses are thought to be used in the generation of reducing agents [NAD(P)H], energy [ATP], and pathway intermediates needed to synthesize secondary metabolite compounds. Their presence may also help trigger the synthesis of defense-related compounds.

Data from these studies, along with previous transcriptional studies (Luo et al., 2009, 2011; Jiang et al., 2011; Kelley et al., 2012), lay the groundwork for future studies investigating *A. flavus* resistance in maize. Under normal growth conditions, inducible defenses of B73 genotype may be adequate in inhibiting or at least slowing down *A. flavus* disease development. However, external and internal factors could affect this response. Abiotic stress, such as drought, can have a negative impact on the defense response (Wotton and Strange, 1987; Duke and Doehlert, 1996; Luo et al., 2010). Also, inherently low expression of defense and defense-associated genes may predispose the plant to greater infection (Chen et al., 2001; Alessandra et al., 2010). Genes expressed during infection may not necessarily be involved in resistance and could be causing increased susceptibility to fungal disease. Knowing which genes are typically expressed in response to pathogen attack is useful when examining how genotype and abiotic stress influence the infection process. Progress on more fully understanding disease development will ultimately leads to the development of genetically resistant cultivars.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb.2014. 00384/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: 07 April 2014; accepted: 09 July 2014; published online: 31 July 2014.*

*Citation: Dolezal AL, Shu X, OBrian GR, Nielsen DM, Woloshuk CP, Boston RS and Payne GA (2014) Aspergillus flavus infection induces transcriptional and physical changes in developing maize kernels. Front. Microbiol. 5:384. doi: 10.3389/fmicb. 2014.00384*

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

*Copyright © 2014 Dolezal, Shu, OBrian, Nielsen, Woloshuk, Boston and Payne. 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.*

### Environmental influences on maize-Aspergillus flavus interactions and aflatoxin production

#### *Jake C. Fountain1, Brian T. Scully <sup>2</sup> , Xinzhi Ni <sup>3</sup> , Robert C. Kemerait 1, Robert D. Lee4 , Zhi-Yuan Chen5 , and Baozhu Guo1,2 \**

<sup>1</sup> Department of Plant Pathology, University of Georgia, Tifton, GA, USA

<sup>2</sup> Crop Protection and Management Research Unit, Agricultural Research Service – United States Department of Agriculture, Tifton, GA, USA

<sup>3</sup> Crop Genetics and Breeding Research Unit, Agricultural Research Service – United States Department of Agriculture, Tifton, GA, USA

<sup>4</sup> Department of Crop and Soil Sciences, University of Georgia, Tifton, GA, USA

<sup>5</sup> Department of Plant Pathology and Crop Physiology, Louisiana State University, Baton Rouge, LA, USA

#### *Edited by:*

Perng-Kuang Chang, Southern Regional Research Center, USA

#### *Reviewed by:*

Chuck Niblett, Venganza, Inc., USA Ahmad Fakhoury, Southern Illinois University Carbondale, USA

#### *\*Correspondence:*

Baozhu Guo, Research Plant Pathologist, Crop Protection and Management Research Unit, Agricultural Research Service – United States Department of Agriculture, 2747 Davis Road, Tifton, GA 31793, USA e-mail: baozhu.guo@ars.usda.gov

Since the early 1960s, the fungal pathogen Aspergillus flavus (Link ex Fr.) has been the focus of intensive research due to the production of carcinogenic and highly toxic secondary metabolites collectively known as aflatoxins following pre-harvest colonization of crops. Given this recurrent problem and the occurrence of a severe aflatoxin outbreak in maize (Zea mays L.), particularly in the Southeast U.S. in the 1977 growing season, a significant research effort has been put forth to determine the nature of the interaction occurring between aflatoxin production, A. flavus, environment and its various hosts before harvest. Many studies have investigated this interaction at the genetic, transcript, and protein levels, and in terms of fungal biology at either pre- or post-harvest time points. Later experiments have indicated that the interaction and overall resistance phenotype of the host is a quantitative trait with a relatively low heritability. In addition, a high degree of environmental interaction has been noted, particularly with sources of abiotic stress for either the host or the fungus such as drought or heat stresses. Here, we review the history of research into this complex interaction and propose future directions for elucidating the relationship between resistance and susceptibility to A. flavus colonization, abiotic stress, and its relationship to oxidative stress in which aflatoxin production may function as a form of antioxidant protection to the producing fungus.

**Keywords: maize, aflatoxin contamination, environment effects, oxidative stress, host resistance**

#### **INTRODUCTION**

*Aspergillus flavus* (Link ex Fr.; Teleomorph: *Petromyces flavus*; Horn et al., 2009a,b) is a facultative, plant parasitic pathogen, which has the ability to colonize a number of common crop species including corn, cotton, peanuts, and many other crops (Diener et al., 1987). Economic losses due to the infection of grain crops such as maize (*Zea mays* L.) by *A. flavus* is not primarily due to the expression of symptoms known as *Aspergillus* ear rot but, rather, is due to the subsequent contamination of the grain with the fungal metabolite aflatoxin. Aflatoxins are a group of polyketide-derived furanocoumarin secondary metabolites produced by certain species of fungi, including the genus *Aspergillus* (Bennett and Klich, 2003; Chanda et al., 2009). Aflatoxins are highly carcinogenic and can be acutely toxic or fatal if ingested in sufficient quantities for both livestock and humans (Shephard, 2008).

Although this species was first described by Link in 1809 (Amaike and Keller, 2011), major research on the biology and pathogenicity of *A. flavus* did not commence until the mid 1960s with the incidence of Turkey X disease which killed over 100,000 turkey poults due to aflatoxin contaminated feed associated with *A. flavus* infected peanuts (Wogan, 1966). Shortly thereafter broad screening of feed and food was initiated, the chemical structures of the major aflatoxins (B1, B2, G1, and G2) were elucidated, and research was conducted to prevent post-harvest contamination of grain crops through the modulation of storage conditions (Asao et al., 1965; Trenk and Hartman, 1970). However, it was found during a particularly severe outbreak of aflatoxin contamination in maize in the late 1970s in the U.S. that it was possible for *A. flavus* to both colonize and produce aflatoxin on developing maize kernels prior to harvest (Diener et al., 1983, 1987).

Since the extensive losses from the 1977 growing season due to aflatoxin contamination (Diener et al., 1983), research efforts have focused on determining the source of host plant resistance to prevent *A. flavus* colonization and subsequent aflatoxin production pre-harvest and before transportation to storage. These efforts have employed numerous techniques and approaches including modern plant breeding and genetics tools such as proteomic, transcriptomic, and biochemical analyses in an effort to discover the underlying mechanism of host plant resistance and the interaction between the two organisms. To date, these efforts have revealed that resistance is quantitatively inherited with a strong genotype by environment component. It is a complex interaction with a high degree of environmentally induced variability with abiotic and biotic stress strongly influencing resistance or susceptibility. Here we review the results of research into the

mechanism of host resistance to both *A. flavus* colonization and abiotic stress, and propose future research directions for determining the relationship between oxidative stress and aflatoxin contamination.

#### **HOST RESISTANCE AGAINST** *A. flavus***: GENE-FOR-GENE VS. GENOTYPE × ENVIRONMENT INTERACTION?**

From a gene-for-gene perspective, host resistance against *A. flavus* colonization and subsequent aflatoxin contamination has been approached as a single virulence factor produced by the invading pathogen that would be countered by a single avirulence or resistance protein in the host. This results in compatible or incompatible reactions based on specific recognition (Keen, 1990). The particular virulence mechanisms of these plant-microbe interactions are utilized to classify plant pathogens into groups such as biotrophic, necrotrophic, and hemibiotrophic pathogens (Glazebrook, 2005).

The classification of *A. flavus* into a particular class of plant pathogens has yet to be determined. A microscopy study of the growth of *A. flavus* in maize kernel tissue by Smart et al. (1990) showed that cellular components such as cell walls were broken down in advance of mycelia. This has been interpreted as being indicative of necrotrophic pathogenicity in the literature (Mideros et al., 2009), however, not all research concurs with this conclusion. Magbanua et al. (2007)found that the colonization of kernel tissue from resistant maize lines exhibited increased levels of salicylic acid (SA) and unchanged levels of jasmonic acid (JA). Such patterns are commonly associated with resistance to biotrophic pathogens in various plant species (Glazebrook, 2005). As a facultative parasite, which naturally exists as a saprophyte, *A. flavus* may possess a unique pathogenicity mechanism that does not categorically fit into this classification scheme.

In keeping with the concept of gene-for-gene resistance, virulence factors produced by *A. flavus* during the colonization of maize tissues as well as maize kernel avirulence proteins have been the focus of multiple studies. These efforts have helped to better characterize the nature of this plant–pathogen interaction, and have sought to identify pathogen and host-derived proteins encoded by potential single-gene sources of host resistance or susceptibility. Proteomics-based techniques have identified several virulence proteins produced by *A. flavus*, most of which being hydrolytic enzymes. Examples of these enzymes are amylases, cellulases, chitinases, cutinases (e.g., phyto-cutinase), lipases, pectinases (P2c), proteases such as alkaline protease, and xylanases (Cleveland and Cotty, 1991; Guo et al., 1995; Chen et al., 1998, 1999; Fakhoury and Woloshuk, 1999; Mellon et al., 2000; Brown et al., 2001; Cleveland et al., 2004; Chen et al., 2009a; Pechanova et al., 2013). These enzymes are consistent with the biology and classification of *A. flavus* as a saprophyte since they typically catabolize decaying plant materials as a source of nutrition.

Several constitutively expressed and inducible proteins have been described in the literature, which have been shown to counter the function of other hydrolytic virulence proteins produced by *A. flavus*. For example, Chen et al. (1998) described a 14-kDa trypsin inhibitor (TI) which functions as an inhibitor of α-amylase, a protein utilized by *A. flavus* for the catabolism of complex carbohydrates. The same group also showed that silencing the expression of the TI gene in maize increases the susceptibility of maize kernel tissue to *A. flavus* infection and aflatoxin contamination (Chen et al., 2009b). In addition, β-1,3-glucanases, chitinases, pathogenesis-related proteins 10 and 10.1, ribosome inactivating proteins (RIPs), and zeamatin have also been shown to be involved in the resistance of maize against *A. flavus* (Mauch et al., 1988; Walsh et al., 1991; Huynh et al., 1992; Guo et al., 1997; Lozovaya et al., 1998; Chen et al., 2006; Chen et al., 2010; Xie et al., 2010). In peanut (*Arachis hypogaea*), Liang et al. (2005) reported that the increase of β-1,3-glucanases among resistant lines was higher than in the susceptible lines after infection with *A. flavus*.

Accumulation of such antifungal and avirulence proteins has been shown to contribute to the resistance observed in several maize lines along with morphological characteristics of resistant kernels such as thickened wax cuticles (Guo et al., 1995). Various breeding techniques have been employed to develop varieties with enhanced resistance to *A. flavus* and aflatoxin contamination on the basis of utilizing phenotypic screenings and molecular markers associated with known avirulence genes to identify quantitative trait loci (QTL) for selection (Brown et al., 2013). These efforts have met with some success such as in the case of Willcox et al. (2013) who identified 20 QTL explaining 22–43% of phenotypic variation within a F2 mapping population derived from Mp313E × Va35. This study, along with others (Paul et al., 2003; Brooks et al., 2005; Kelley et al., 2012), illustrate two important challenges in breeding resistant maize varieties, including: (1) that resistance to *A. flavus* is a quantitative trait involving multiple genes rather than single-gene forms of gene-for-gene resistance, and (2) a large genotype x environment (G × E) interaction conditions the expression of this trait. For example Willcox et al. (2013) determined that only 11 of the 20 identified QTL were consistently expressed across different environments with these accounting for 2.4–9.5% of phenotypic variance.

Environmental factors can have a significant effect on maize resistance to *A. flavus* and aflatoxin production, particularly abiotic stresses such as drought and heat stress. This was clearly demonstrated as early as 1977 when widespread and intense drought conditions in the Midwestern and Southeastern United States resulted in a high degree of aflatoxin contamination of maize kernels pre-harvest (Zuber and Lillehoj, 1979; Diener et al., 1987). However, maize genotypes possessing drought tolerance, such as Lo964 and Tex6, tend to be less susceptible to aflatoxin contamination (Guo et al., 2008; Luo et al., 2008; Jiang et al., 2012).

#### **RELATIONSHIP OF DROUGHT TOLERANCE, AFLATOXIN CONTAMINATION AND** *A. flavus* **RESISTANCE**

It has been hypothesized that there is an underlying relationship among various molecular mechanisms of drought stress adaptation and resistance to *A. flavus* infection and subsequent aflatoxin contamination. This connection has been demonstrated experimentally in a number of studies. Chen et al. (2007) found a number of drought stress-related proteins that were induced in response to *A. flavus* colonization in maize endosperm tissue including late embryogenesis abundant proteins LEA 3, 14, peroxiredoxin (PER1), and a 17.9-kDa heat shock protein (HSP17.9). Pechanova et al. (2011) found that these proteins along with several antioxidant proteins, such as ascorbate peroxidase and superoxide dismutase (SOP), were up-regulated in resistant maize rachis tissue earlier in development and formed the basis, when combined with increased expression of antifungal and pathogenesis-related proteins, for resistance to*A. flavus* colonization.

It has also been shown that global defense regulators have the potential to modulate maize resistance to drought stress, oxidative stress, and possibly *A. flavus* resistance. The WRKY transcription factors have been shown to regulate the responses of multiple plant species to both biotic and abiotic stresses (Rushton et al., 2010). The transcription factor *Zm*WRKY33 was recently shown to enhance abscisic acid (ABA) signaling and enhance osmotic stress tolerance in maize seedlings (Li et al., 2013). *Zm*WRKY33 along with *Zm*WRKY19, whose homolog in *Arabidopsis thaliana* (*At*WRKY53) is known to induce a response to oxidative stress and regulate the expression of antioxidant enzymes like catalase (Miao et al., 2004; Eulgem and Somssich, 2007; Miao et al., 2007; Rushton et al., 2010). These factors were found to be up-regulated earlier in resistant maize varieties in response to *A. flavus* inoculation in whole kernel tissues (Fountain et al., 2013).

#### **OXIDATIVE RESPONSES OF PLANTS TO HERBIVORY**

Oxidative responses of plants to feeding damage by both chewing and piercing-sucking insects and nematodes have been described in recent years. Walling (2000) and Kaloshian and Walling (2005) described the piercing-sucking hemipteran insect feeding on crop plants as resembling plant pathogen infections that they are often associated with chitosan (oligogalacturonides produced by pectinases) and reactive oxygen species (ROS)-activated wound or defense signaling pathways in host plants. Kessler and Baldwin (2002) reviewed the molecular mechanisms underlying plant responses to insect herbivory and how they differ from pathogen infections. They concluded that insect herbivores are physiologically independent from their host plants, whereas pathogens are physiologically dependent on their host plants for their growth and development. Oxidative enzyme-mediated wounding responses also play a critical role in understanding plant responses to insect herbivory, although the insect-specific elicitors frequently modify the responses of their host plants, and allow the host plants to optimize their defenses against a specific insect pest (Kessler and Baldwin, 2002). Das et al. (2008)studied the accumulation of ROS in cowpea–root-knot nematode interaction and confirmed that the induction of resistance is relatively late in this system. Typically, hypersensitive response is closely associated with an oxidative burst in infected tissue.

Bi and Felton (1995) reported that corn earworm (*Helicoverpa zea* Boddie), herbivory caused significant increases in lipid peroxidation and hydroxyl radical formation in the soybean leaves. The activities of several oxidative enzymes (i.e., lipoxygenases, peroxidase, diamine oxidase, ascorbate oxidase, and NADH oxidase I) were also increased following *H. zea* herbivory on soybean (*Glycine max*) leaves. They concluded that oxidative responses in the soybean plants may have led to a decrease in herbivory and an increase

in oxidative damage to the plant. Ni et al. (2000) described the salivary enzyme profiles of the leaf-chlorosis-eliciting Russian wheat aphid (*Diuraphis noxia* Mordvilko), and the non-leaf-chlorosiseliciting bird cherry-oat aphid (*Rhopalosiphum padi* L.), which differ in oxidative enzyme activities. While only peroxidase activity was detected in *R. padi*, catalase activity was only detected in *D. noxia.* The oxidative responses of four cereal plants (i.e., susceptible "Arapahoe" and resistant "Halt" wheat, susceptible "Morex" barley, and resistant "Border" oat) to the feeding of the two species of aphid differed (Ni et al., 2001a). The chlorosiseliciting *D. noxia* feeding caused a three-fold increase in peroxidase activity in the resistant Halt wheat, and nine-fold increase in the susceptible Morex barley 9 days after infestation when compared to the control leaves. In contrast, *R. padi* did not cause any changes in peroxidase activity in any of the cereal leaves. At the same time, *D. noxia* feeding did not elicit any change in either catalase or polyphenol oxidase activity in comparison with either the *R. padi*-infested or the control cereal leaves (Ni et al., 2001a). Furthermore, oxidative bleaching in leaf chlorosis elicited by *D. noxia* was not detected, but Mg-dechelatase activity was increased in the *D. noxia*-elicited chlorosis in wheat leaves (Ni et al., 2001b).

Zavala et al. (2013) also proposed a cellular mechanism to decipher the influence of elevated CO2 on insect herbivory. They highlighted that oxidative enzyme activities in the subcellular organelles, such as peroxisomes and chloroplasts, via the jasmonic signaling pathway are likely to be critical factors in dissecting the molecular mechanisms of plant defenses against both biotic and abiotic stresses. In general, the oxidative responses of plants to pathogen infection and insect herbivory under varying environmental conditions (e.g., drought and the elevated CO2) are critical for the management of pest outbreaks and the reduction of mycotoxin contamination in agricultural crops.

#### **THE ROLE OF OXIDATIVE STRESS IN AFLATOXIN BIOSYNTHESIS**

The presence of increased expression of antioxidant mechanisms in resistant maize tissues leads to the hypothesis that increased resistance to ROS-induced oxidative stress may correlate to resistance to *A. flavus* and aflatoxin contamination. Although this conclusion seems credible due to the high degree of correlative evidence present in the literature, the exact mechanism of how this phenomenon functions has yet to be elucidated completely. More recent studies into the biology of *A. flavus* and the mechanisms regulating the production of aflatoxin may illuminate this issue (Magbanua et al., 2007; Roze et al., 2013).

Aflatoxin biosynthesis is a complex process involving multiple gene products and regulatory mechanisms coded for by an approximately 70-kb cluster of 25 genes (Yu et al., 2004). This pathway is responsible for the biosynthesis of five major mycotoxins: sterigmatocystin, and aflatoxins B1, B2, G1, and G2 (Yu et al., 2004), and is the focus of intensive research into methods of negatively regulating its function. Although the structure and biochemical characteristics of aflatoxins have been known since the 1960s (Asao et al., 1965;Wogan, 1966), the specific purpose of their production by *A. flavus* or other aflatoxigenic fungi has remained a mystery. Prior research exploring the roles of oxidative stress in regulating aflatoxin biosynthesis as well as recent discoveries into the upstream regulation of major pathway regulatory factors have begun to elucidate the biological function of aflatoxins (Roze et al., 2013). It has been shown that aflatoxin production by *A. flavus* is higher in maize kernel tissues containing higher levels of lipids, such as embryo tissues (Earle et al., 1946; Fabbri et al., 1980; Brodhagen and Keller, 2006). The roles of lipids in regulating aflatoxin biosynthesis in *Aspergillus spp*. have been investigated, particularly oxylipins (Reviewed in Gao and Kolomiets, 2009). Earlier research by Fabbri et al. (1983) found that seeds of high-oil crops such as peanut support higher levels of aflatoxin production by *A. parasiticus* than seeds of graminaceous plants such as maize or wheat, which contain higher levels of starch. In addition, they found that culturing *A. parasiticus* amended with peroxidized lipids resulted in significantly elevated aflatoxin production with no significant effect on fungal biomass (Fabbri et al., 1983).

Lipid peroxidation is a byproduct of lipid metabolism in peroxisomes as well as the reaction of naturally produced free fatty acids with ROS (Reverberi et al., 2012). Therefore, it is possible that excessive peroxisome function in the fungal mycelia or oxidative stress may be a causative factor in the production of aflatoxin. *In vitro*, Jayashree and Subramanyam (2000) showed that toxigenic strains of *A. parasiticus* have increased oxygen requirements, which they postulate to be a potential source of ROS accumulation, in comparison to non-toxigenic strains. In addition, they showed that higher levels of glutathione and thiobarbituric acid-reactive substances (TBARS), as well as antioxidant enzyme activities, were present in toxigenic strains in comparison to non-toxigenic strains. This indicated that oxidative stress may be a pre-requisite for aflatoxin production (Jayashree and Subramanyam,2000). Also,Reverberi et al. (2012)found that bezafibrate and transformation of the *Cymbidium ringspot virus* P33 gene into *A. flavus* induced peroxisome proliferation resulting in an increase in aflatoxin production both *in vitro* and when cultured on maize kernel tissues in addition to increased levels of antioxidant enzyme gene expression, lipid metabolism, oxylipin biosynthesis, and ROS accumulation in the *A. flavus* mycelia. Recent studies have also found that cAMP and G-protein-mediated quorum sensing signaling pathways based on oxylipin perception can play a vital role in growth regulation and aflatoxin production in *A. nidulans* (Affeldt et al., 2012). Such G-protein mediated signaling has also been reported in other *Aspergillus spp*. including *A. fumigatus* (Grice et al., 2013). Therefore, it may be concluded that oxidative stress in *A. flavus* induced by ROS and/or oxylipins in the growth environment/medium will result in increased aflatoxin production from a biochemical perspective.

Recent studies have also shown that ROS can play a role in the transcriptional regulation of aflatoxin and sterigmatosystin biosynthesis pathway genes. Reverberi et al. (2008) found that a putative binding site for the Ap*yapA* gene, which regulates oxidative stress tolerance and conidiogenesis, was present in the promoter of the regulatory gene *aflR* in *A. parasiticus*, and that silencing Ap*yapA* results in an increase in aflatoxin biosynthesis. It has also been shown that the basic leucine zipper (bZIP) transcription factor AtfB regulates the aflatoxin biosynthesis genes *fas-1*, *ver-1*, and *omtA* as well as antioxidant genes encoding for catalase and SOP (Hong et al., 2013). They also found that the promoter regions associated with AtfB also contained cAMP-responsive elements implicating cAMP in the regulation of aflatoxin biosynthesis (Hong et al., 2013).

#### **POTENTIAL REACTIVE OXYGEN SPECIES-MEDIATED CROSSTALK BETWEEN MAIZE AND** *A. flavus*

Given the apparent role of oxidative stress in the promotion of aflatoxin biosynthesis, the hypothesis has been proposed that aflatoxin may function as a form of antioxidant protection to Aspergilli (Reverberi et al., 2010). This would provide an explanation to the long standing question, rather the mystery, as to the biological significance of aflatoxin, although the potential antioxidant mechanism of action of aflatoxin has yet to be fully elucidated. In addition, this explanation also provides for the potential role of ROS and oxylipins cross-kingdom communication between maize and *A. flavus*.

Cross-kingdom communication between plants and various fungi through the use of oxylipins has been previously documented (Eckardt, 2008; Christensen and Kolomiets, 2011). Specifically, the role of oxylipins has been clearly illustrated in the specific interaction between maize and *A. flavus* (Gao and Kolomiets, 2009). In a recent study, it was found that maize lipoxygenase-3 (LOX3) is required for resistance to *A. flavus* indicating that certain 9-oxylipins can play important roles in suppressing aflatoxin biosynthesis while other oxylipins may promote aflatoxin biosynthesis (Gao et al., 2009). The role of LOX-1 in resistance mechanisms against *A. flavus*, *A. nidulans,* and *A. parasiticus* has also been implied in soybean (Doehlert et al., 1993; Burow et al., 1997), although there are contradicting reports on the subject in the literature. Mellon and Cotty (2002) found that soybeans lacking LOX activity were just as resistant to *A. flavus* and aflatoxin contamination as those possessing LOX activity. This seems to imply some degree of specificity in the role of LOX enzymes or their products in resistance to certain *Aspergillus* species.

In addition to oxylipins, other host-derived compounds may influence oxidative stress including ROS and phytohormones. It was found that 2-chloroethyl phosphoric acid (CEPA), the metabolic precursor to ethylene, was capable of reducing the expression of *aflR* and *aflD* (two key genes in the aflatoxin biosynthetic pathway), reducing the accumulation of oxidative compounds, and regulating glutathione redox in *A. flavus* mycelia (Huang et al.,2009). Therefore, host-derived ethylene may result in the reduction of ROS accumulation in*A. flavus* mycelia and reduce aflatoxin biosynthesis. This hypothesis seems plausible since the expression of the maize Ethylene Responsive Factor 1 (*Zm*ERF1), a key transcription factor involved in ethylene and JA signaling, was found to be higher in the immature kernel tissues of the resistant maize inbred TZAR101 (Menkir et al., 2008) in comparison to the susceptible maize inbred B73 following *A. flavus* inoculation (Fountain et al., 2013).

Previous research has shown that maize varieties resistant to *A. flavus* tend to accumulate antioxidant enzymes, such as peroxidase and SOP, and tend to be more resistant to drought and heat stress than varieties susceptible to *A. flavus* (Guo et al., 2008; Pechanova et al., 2011). Given the reported role of oxidative stress in *Aspergillus spp*. biology, it may be possible for host-derived antioxidant proteins, phytohormones, and oxylipins to negatively regulate the production of aflatoxin in infecting *A. flavus* by reducing the level of oxidative stress endured by both the host and the fungus, particularly during drought or heat stress. Such an interaction may explain several observed phenomena in the literature. For example, Guo et al. (1996, 1997) observed that the pre-incubation of maize kernels in high-humidity conditions for 3 days prior to inoculation with *A. flavus* results in a significant reduction in aflatoxin contamination in comparison with kernels inoculated without pre-incubation. Given the fact that ROS such as hydrogen peroxide accumulates to maximum quantities two days postimbibition (DPI) followed by increased catalase activity beginning at three DPI in maize kernels (Hite et al., 1999), a combination of host-derived resistance and antioxidant proteins and reduced ROS production at the time of inoculation may have contributed to the reduction in aflatoxin production (Guo et al., 1996).

#### **CONCLUSION**

Determining the role of oxidative stress in the regulation of aflatoxin biosynthesis as well as the role of host defenses against both *A. flavus* infection and mycotoxin biosynthesis are critical areas of research for the mitigation of aflatoxin contamination in maize. Maize resistance to *A. flavus* is a complex, quantitative trait which is the culmination of the interaction of numerous resistanceassociated proteins and antioxidant enzymes which have been the subject of more than 50 years of rigorous research. Solutions to the problem of aflatoxin contamination of crops, particularly maize, have been elusive given the high level of environmental influence on the interaction and the lack of stable resistance in maize germplasm across multiple environments. By better understanding the role of oxidative stress and its remediation by the host and the pathogen, additional tools will be made available to counter the threat aflatoxin poses to food safety and security and further enhance the knowledge of cross-kingdom interactions which may be applied to other mycotoxin producing pathogens in various agricultural commodities.

#### **ACKNOWLEDGMENTS**

This work is partially supported by the U.S. Department of Agriculture Agricultural Research Service (USDA-ARS), the Georgia Agricultural Commodity Commission for Corn, and AMCOE (Aflatoxin Mitigation Center of Excellence). Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. The USDA is an equal opportunity provider and employer.

#### **REFERENCES**


cereal aphid feeding. *Entomol. Exp. Appl.* 101, 159–166. doi: 10.1046/j.1570- 7458.2001.00900.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: 26 November 2013; accepted: 21 January 2014; published online: 05 February 2014.*

*Citation: Fountain JC, Scully BT, Ni X, Kemerait RC, Lee RD, Chen Z-Y and Guo B (2014) Environmental influences on maize-Aspergillus flavus interactions and aflatoxin production. Front. Microbiol. 5:40. doi: 10.3389/fmicb.2014.00040*

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

*Copyright © 2014 Fountain, Scully, Ni, Kemerait, Lee, Chen and Guo. 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.*

### Co-inoculation of aflatoxigenic and non-aflatoxigenic strains of Aspergillus flavus to study fungal invasion, colonization, and competition in maize kernels

#### *Zuzana Hruska1, Kanniah Rajasekaran2 \*, HaiboYao1, Russell Kincaid1, Dawn Darlington1, Robert L. Brown2 , Deepak Bhatnagar <sup>2</sup> and Thomas E. Cleveland2*

<sup>1</sup> Geosystems Research Institute, Mississippi State University, Stennis Space Center, Starkville, MS, USA

<sup>2</sup> Southern Regional Research Center, Agricultural Research Service – United States Department of Agriculture, New Orleans, LA, USA

#### *Edited by:*

Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran

#### *Reviewed by:*

Geromy G. Moore, Agricultural Research Service – United States Department of Agriculture, USA Fuguo Xing, Institute of Agro-Products Processing Science and Technology – Chinese Academy of Agricultural Science, China Hamed Abbas, Agricultural Research Service – United States Department of Agriculture, USA

#### *\*Correspondence:*

Kanniah Rajasekaran, Southern Regional Research Center, Agricultural Research Service – United States Department of Agriculture, 1100 Robert E. Lee Boulevard, New Orleans, LA 70124, USA e-mail: rajah.rajasekaran@ ars.usda.gov

A currently utilized pre-harvest biocontrol method involves field inoculations with nonaflatoxigenic Aspergillus flavus strains, a tactic shown to strategically suppress native aflatoxin-producing strains and effectively decrease aflatoxin contamination in corn. The present in situ study focuses on tracking the invasion and colonization of an aflatoxigenic A. flavus strain (AF70), labeled with green fluorescent protein (GFP), in the presence of a non-aflatoxigenic A. flavus biocontrol strain (AF36), to better understand the competitive interaction between these two strains in seed tissue of corn (Zea mays). Corn kernels that had been co-inoculated with GFP-labeled AF70 and wild-type AF36 were crosssectioned and observed under UV and blue light to determine the outcome of competition between these strains. After imaging, all kernels were analyzed for aflatoxin levels. There appeared to be a population difference between the co-inoculated AF70-GFP+AF36 and the individual AF70-GFP tests, both visually and with pixel count analysis. The GFP allowed us to observe that AF70-GFP inside the kernels was suppressed up to 82% when coinoculated with AF36 indicating that AF36 inhibited progression of AF70-GFP. This was in agreement with images taken of whole kernels where AF36 exhibited a more robust external growth compared to AF70-GFP.The suppressed growth of AF70-GFP was reflected in a corresponding (upto 73%) suppression in aflatoxin levels. Our results indicate that the decrease in aflatoxin production correlated with population depression of the aflatoxigenic fungus by the biocontrol strain supporting the theory of competitive exclusion through robust propagation and fast colonization by the non-aflatoxigenic fungus.

**Keywords:** *Aspergillus flavus***, biocontrol, green fluorescent protein (GFP), fungal competition, aflatoxin, fluorescence microscopy, corn (***Zea mays***)**

#### **INTRODUCTION**

The opportunistic, soil-borne fungus *Aspergillus flavus* is the main producer of aflatoxin, a mycotoxin known for its carcinogenic properties and other deleterious effects on health in both human and animal populations. Aflatoxins have been found in many agricultural commodities including corn, cottonseed, groundnut, and tree nuts (CAST, 2003), and contamination can occur at any time point between pre- and post- harvest; specifically, plant growth through crop storage. Aflatoxin affects food production and international commerce, prompting enforcement of regulations regarding acceptable aflatoxin limits in many countries around the world [Food and Agriculture Organization (FAO), 2004].

Different pre- and post-harvest strategies have been tested for the prevention and control of aflatoxin in order to protect the integrity of corn. Pre-harvest strategies include traditional and molecular breeding approaches to develop resistant germplasm (Rajasekaran et al., 2006; Cary et al., 2009), and biological control by utilizing various microorganisms including bacteria, yeasts, and non-aflatoxigenic fungi (Yin et al., 2008). The most promising pre-harvest biocontrol method involves field inoculation with non-aflatoxigenic *A. flavus*, a strategy that has proven successful at reducing aflatoxin contamination in the field (Cotty, 1994, 2001; Dorner,2009a,b). Two bio-pesticides,AF36 andAfla-Guard®, have been approved by the Environmental Protection Agency (EPA) for pre-harvest application to control aflatoxin contamination in the U.S. (Moore et al., 2013). Neither AF36 (NRRL 18543) nor the component strain in Afla-Guard® (NRRL 21882) can produce aflatoxin due to genetic mutations in, or lack of, critical genes along the aflatoxin synthesizing pathway, respectively, (Ehrlich and Cotty, 2004; Chang et al., 2005; Moore et al., 2009). This method of control has been effectively tested on several commodities including cottonseed (Cotty, 1994, 2001), corn, (Brown et al., 1991; Dorner, 2009a), and peanuts (Dorner, 2009b). The suppression of measurable aflatoxin has been attributed to competitive exclusion of aflatoxigenic strains by non-aflatoxigenic strains; however, the exact mechanism of suppression has yet to be elucidated. Exclusion of the aflatoxin producing strain may be achieved by a more robust propagation and faster growth of the competing non-aflatoxigenic fungal strain (Chang and Hua, 2007). Alternatively, the inhibition of aflatoxin contamination/production could be due to a close physical proximity between the competing strains

which may initiate a signal to down-regulate or inhibit synthesis of aflatoxin in the aflatoxigenic strain by the non-aflatoxigenic strain (Huang et al., 2011).

Dispersion of conidial inocula in the corn field is presumed to occur through a direct transfer by insects, or via colonization of silks, and subsequent movement toward the developing kernels (Payne, 1998; Mehl and Cotty, 2011). Entry into individual kernels is thought to occur primarily through the pedicel (tip cap) of intact kernels (Lillehoj, 1983; Rajasekaran et al., 2013). However, it is difficult to determine the mode of action taken by the competing fungal strains, particularly during the infection and colonization process inside individual kernels.

The use of fungal strains that express fluorescent proteins in the cytoplasm greatly enhances the potential for visual tracking of biological processes without markedly affecting fungal growth or pathogenicity (Rajasekaran et al., 2008; Crespo-Sempere et al., 2011). Enhanced forms of the green fluorescent protein (eGFP) were recently developed to tag *A. flavus* for study of fungal growth, mode of entry and colonization of cottonseeds (Rajasekaran et al., 2008), to track the progress of fungal infection within developing corn ears (Magbanua et al., 2013), to study infection, fungal growth, colonization, and aflatoxin production in intact corn kernels (Rajasekaran et al., 2013), and also to tag *A. carbonarius* for monitoring fungal colonization in grapes (Crespo-Sempere et al., 2011).

To better understand the mechanisms involved during competition between co-inoculated *A. flavus* isolates in corn, the current *in situ* study focuses on tracking the invasion and colonization by aflatoxigenic AF70, labeled with GFP, in the presence of the AF36 biocontrol strain.

#### **MATERIALS AND METHODS**

#### **FUNGAL STRAINS AND PREPARATION OF INOCULA**

Two *A. flavus* strains were used as competitors. Toxin producing AF70 labeled with GFP (AF70-GFP; Rajasekaran et al., 2008), and a non-aflatoxigenic strain of *A. flavus* (AF36) fungal cultures were obtained from the SRRC fungal collection (ARS-USDA, New Orleans, LA, USA), and grown on potato dextrose agar (PDA) media for 7 days at 30◦C. Bright green fluorescence was observed in the AF70-GFP culture from both the conidiophores and the mycelia. Harvest of conidia was accomplished by flooding a single culture for each strain with 20 ml of 0.01% (v/v) sterile Triton X-100 solution and scraping the surface mycelia with a sterile scraper. Conidial suspensions were adjusted to 4 <sup>×</sup> 106 cells/ml using sterile distilled water. The inocula were kept in separate containers at 4◦C. Immediately before use, 50 ml of each inoculum, at 1:1 ratio, was combined to make up the AF70-GFP+AF36 co-inoculation mixture.

#### *IN SITU* **INOCULATION**

Corn kernels (N78B-GT, Syngenta NK Brand Seeds, Laurinburg, NC, USA), collected in 2010 from the ARS Field Station in Stoneville, MS, USA, were utilized in all experiments. Whole, undamaged kernels of roughly uniform size were randomly assigned into four treatment groups and processed according to a modified kernel screening assay (KSA; Brown et al., 1995). All kernels were surface sterilized in 70% ethanol and rinsed in dH2O.

Ten kernels per treatment per day were inoculated by immersion and stirred for 1 min. The treatments included:(1) kernels coinoculated with AF70-GFP+AF36, (2) kernels inoculated with AF70-GFP only, (3) kernels inoculated with AF36 only, and (4) kernels inoculated only with dH2O as control for non-specific fluorescence, and/or inherent fungal contamination. Each group of kernels was incubated in a humidity chamber using a plastic tray with individual compartments (**Figure 1**). Kernels were incubated at 30◦C and examined at several time points (3, 4, 5, 7, and 9 days) after inoculation.

#### *GFP imaging and analysis*

At each specified time interval, kernels were removed from the incubator, wiped free of visible exterior mold growth, sliced longitudinally, and mounted on a slide with double sided tape. A total of 20 kernel cross-sections per treatment per day were analyzed for GFP fluorescence. Images were taken in the dark with an Olympus SZH10 GFP stereomicroscope (Olympus, Center Valley, PA, USA) equipped with 480 nm excitation and 535 nm emission filters for GFP fluorescence, and also with UV filters. Digital images were acquired with a Nikon Digital Camera DXM1200 (Nikon Instruments, Melville, NY, USA). Sectioned samples were then processed for aflatoxin analysis.

In order to determine the extent of suppressed expression of the GFP signal from the AF70-GFP by AF36 in the co-inoculated corn kernels, as compared with those kernels inoculated with AF70- GFP alone, all GFP-expressing images were analyzed with pixel count analysis in ENVI (ENVI software v 4.7, ITT Exelis, Boulder,

CO, USA). Pixel count analysis has been employed previously as a useful tool for quantifying fluorescence signals in digital images (Waters, 2009). Each 8-bit image contained 3200 x 2560 pixels with brightness values ranging from 0–255. In order to quantify the number of pixels exhibiting fluorescence within each image, a region of interest (ROI) was created by utilizing a band thresholding process. A threshold brightness value of 20 was selected to separate the pixels exhibiting fluorescence (>20) from the background (<20) and minimize non-specific background fluorescence. Regions of interest were drawn around all pixels exhibiting values above the set threshold, and totals for each cross section of each sample were recorded for statistical analysis.

#### **AFLATOXIN ANALYSIS**

Imaged kernels were dried at 60◦C for 2 days and analyzed for aflatoxin with the VICAM AflaTest assay (VICAM, Milford, MA, USA). Dried kernels were processed following the single kernel assay protocol of Yao et al. (2010), modified from the VICAM AflaTest instruction manual for corn, milo, grains, and feeds. Briefly, kernels from each day were first uniformly, semi-coarsely ground in a coffee grinder (KitchenAid® BCG100). Next, aflatoxin B from each ground sample was extracted with sodium chloride (NaCl) in methanol:water (80:20 v/v) maintaining the proportion of 1 g of sample plus 0.1 g NaCl in 2 ml of methanol:water, and filtered through a fluted filter. Filtered extract was diluted 1:5 with ultrapure water; 2 ml of each sample were passed through a glass

filter and pushed through the high affinity AflaTest column. Samples were washed with ultrapure water, eluted with pure methanol (HPLC grade), and fluorescence was measured using the VICAM Series-4EX fluorometer (VICAM, Milford, MA, USA). Raw data values were expressed as ppb or μg/kg equivalents.

#### **STATISTICAL ANALYSES**

Pixel counts from the images, as well as the aflatoxin concentrations from the aflatoxin extractions, were analyzed in a two-way ANOVA with repeated measures using the Excel Data Analysis package (Excel 2007, Microsoft Office, Microsoft). Graphs were created in the R statistical program (R statistics software, http://www.r-project.org).

#### **RESULTS**

#### **GFP IMAGING AND POPULATION (PIXEL) COUNT**

On intact corn kernels the observed external fungal growth was predominantly AF36 for the co-inoculated groups; in some instances AF36 completely covered the pericarp (**Figure 1**). Internally, the spread of AF70-GFP was evident by the presence of fluorescence (**Figure 2**). In corn kernels inoculated solely with AF70-GFP, the fungus spread unimpeded, filling the whole kernel from the basal transfer layer through the embryo and into the endosperm, eventually compromising the kernel's integrity. Although the point of entry was predominantly at the pedicel, it was not unusual to find GFP fluorescence in other areas of a kernel

appearing first. The co-inoculated kernels exhibited a radically different fungal spread (**Figure 2**). The AF70-GFP growth appeared limited to the edges of the corn kernel, presumably outcompeted by AF36. For kernels inoculated with AF36 alone, or with the sterile water control, there was no GFP fluorescence observed; therefore, they were not included in the pixel count analysis. No GFP fluorescence could be observed externally on co-inoculated kernels which supports the premise that AF36 predominated.

Statistical analysis of the fluorescent pixels occupied by AF70- GFP is presented in **Figure 3**. Data was analyzed in a two way ANOVA using each treatment (AF70-GFP+AF36 and AF70-GFP) with repeated measures of day (post-inoculation). There was a significant main effect of treatment with *p* < 0.001, and a statistically significant interaction of treatment by day *p* < 0.001, α = 0.01. These results indicate that there was a significant difference between co-inoculated kernels and those treated with AF70-GFP alone, and this difference changed over time. Suppression of AF70 by AF36 is presented in terms of percentages in **Figure 4**. The peak difference between the two groups appeared to be on day nine whereby the highest suppression of AF70-GFP by AF36 was observed to be 82%. Average suppression over the duration of the study was 66%.

#### **AFLATOXIN CONTENT**

Similar to the population count, the aflatoxin data was analyzed in a two way factorial design of each treatment (AF70-GFP+AF36 and AF70-GFP) with repeated measures of day (**Figure 5**). A significant main effect of treatment with *p* < 0.001, and a statistically significant interaction of treatment by day *p* < 0.001 with α = 0.01 were revealed, indicating a significant difference in aflatoxin concentration between the co-inoculated kernels and those treated with AF70-GFP alone that changed over time. The peak difference between the two groups appeared to be on day five (73%), dipped by day seven (25%), and recovered to a lesser degree on day nine (47%). The observed suppression of AF70-GFP by AF36 presented in **Figure 4** is also observed in the aflatoxin content data presented

**FIGURE 4 | Percent (%) growth suppression of AF70-GFP by AF36 in co-inoculated corn kernels as determined by calculating the percent difference between treatment means on a particular day.** Over 5% difference in suppression was observed between consecutive time points (day 3–day 9).

in **Figure 5**. Maximum decrease in aflatoxin production was determined by calculating the percent difference between means on a particular day and was 73% on day five.

#### **DISCUSSION**

The current study used an *A. flavus* strain, expressing the GFP gene, to investigate its invasion and colonization potential in the presence of a non-aflatoxigenic biocontrol strain in single corn kernels. Although all co-inoculated corn kernels exhibited growth on day three, external fungal growth over the entire growth period was predominantly attributed to AF36, confirmed by lack of fluorescence combined with aggressive growth. Furthermore, the fluorescence was very specific to AF70-GFP and was not present externally or internally in the AF36 controls. It appears the competitive edge exhibited by AF36 may be in its propensity for rapid growth and aggressive colonization of the host compared to AF70- GFP (**Figure 1**). Aflatoxin production requires expending energy in the form of ATP (Ehrlich et al., 2011), and perhaps this affects a toxigenic strain's ability to rapidly colonize, since it may be diverting energy toward the secondary metabolism pathway. Because this does not appear to be an effective competitive strategy targeted against non-aflatoxigenic strains, it appears that aflatoxigenic and non-aflatoxigenic strains are not in competition with each other, under natural conditions, and may coexist without interfering with each other's life cycles unless their natural proportions are disturbed, for example, by inundating fields with excess biocontrol (Yin et al., 2008).

Although the initial entry of the fungus through the pedicel agrees with previous research (Rajasekaran et al., 2013), there were instances where the GFP fluorescence initially appeared in other areas of the corn kernel, particularly when AF70-GFP and AF36 were co-inoculated. The AF70-GFP fluorescence in the coinoculated kernels was significantly reduced and limited to the edges of the kernels, presumably because of the more aggressive growth of the AF36 biocontrol strain. Entry points were randomly distributed along the hull, possibly created by peripheral damage to the pericarp by the robust growth of AF36 where the extensive hyphal growth may have weakened the external kernel tissue. Alternately, invisible weak points in the external tissue may have been created during kernel harvest. Greatest population difference observed for GFP fluorescence between the AF70-GFP and AF70-GFP+AF36 treatments are in agreement with the literature, whereby population, aflatoxin production was decreased by aflatoxin non-producers between 70 and 90% (Brown et al., 1991). The average suppression of aflatoxin production in the current study was 50% compared to results reported in previous studies involving cotton (Cotty, 1990, 1994). The discrepancy may be attributed to the shorter duration and a smaller sample size of the laboratory experiment, compared to field experiments. The fact that AF36 was first isolated from cotton and may have a host preference could be another reason for the difference. The timing of the inoculation may have contributed to the reduced inhibition of aflatoxin production observed in this study. Greater reduction (80–90%) of aflatoxin concentration in corn (Brown et al., 1991), and cottonseed (Cotty, 1990), was shown when the non-aflatoxigenic fungal strain was introduced 24 h prior to the aflatoxin producing strain rather than simultaneously. However, this may not be a reflection of what is actually happening in the field. Studies that explore pre-inoculation with aflatoxigenic strains would better demonstrate the field environment.

The observed population differences between the co-inoculated and the sole AF70-GFP treatments support the premise that growth suppression of AF70-GFP is due to competitive exclusion by AF36 (Brown et al., 1991; Mehl and Cotty, 2010). The exclusion could be due to the initial acquisition of nutrients or tissue space by the biocontrol strain, or something unknown that the fungus requires for survival in a particular situation (Mehl and Cotty, 2011, 2013). Since spore concentrations for inocula in our competitive study were approximately the same, faster growth, and a more robust propagation of AF36 effectively excludes the aflatoxin producing strain and consequently decreases measurable aflatoxin levels. An additional possibility for decrease in aflatoxin production is presented by the presence of specific volatiles. Aflatoxigenic fungi reportedly produce volatile compounds that

differ from non-aflatoxigenic fungi (De Lucca et al., 2012). Certain terpenes found in essential oils, such as alpha-pinene have known antifungal properties (Moghtader et al., 2011) and have been discovered among the volatile compounds produced by some non-aflatoxigenic *A. flavus* isolates (De Lucca et al., 2012). Furans are another class of anti-microbial volatiles produced by *A. flavus*; however, in non-aflatoxigenic strains they are released several days earlier than in aflatoxin producing strains (De Lucca et al., 2012). A possible competitive strategy may involve the production of specific volatiles by the non-aflatoxigenic fungal strains to inhibit growth progression of the aflatoxin producers which would affect the population size of the aflatoxigenic fungi, and consequently reduce the amount of aflatoxin produced. Because the non-aflatoxigenic strains do not appear affected by close proximity to aflatoxin producers, it seems unlikely non-aflatoxigenic strains target aflatoxin production specifically. Therefore, the decrease in aflatoxin may be an indirect result of the suppressed population size of the aflatoxin-producing strain.

This strategy would help explain the apparent success shown by using non-aflatoxigenic fungi to effectively control aflatoxin contamination under field conditions. Unfortunately the long-term consequences of this form of treatment have not been studied. Our study has demonstrated that although AF36 decreased aflatoxin production, the kernels colonized by the biocontrol strain were severely damaged by day nine. There may not have been aflatoxin produced by the AF36 fungus; however, by the end of the study most of the kernels lost their integrity and were overwhelmed by the fungus which may not benefit the crop should the biocontrol strain be allowed to remain during post-harvest storage. Another potential caveat to use AF36 as biocontrol is that it produces cyclopiazonic acid (CPA) which is a mycotoxin that reportedly targets some internal organs as well as the skeletal muscles (Chang et al., 2009; Abbas et al., 2011; Moore et al., 2013). Although the toxic effects of CPA are not as well-documented as the effects of aflatoxin (Chang et al., 2009), overuse of the *A. flavus* isolates that produce CPA may lead to unintended effects on animal and human health. Additionally, when using nonaflatoxigenic fungal strains for applications in biocontrol, it is also important to take into account the potential course the suppressed aflatoxigenic strains may take to overcome inundation and suppression by the competing non-aflatoxigenic strains. For example, the constantly suppressed strain may call upon its latent ability for sexual reproduction demonstrated under laboratory conditions resulting in generations of aflatoxigenic offspring far more virulent than anticipated (Moore et al., 2013). All of these concerns regarding potential effects on crop integrity and/or unanticipated health risks need to be carefully evaluated when considering specific non-aflatoxigenic strains for aflatoxin biocontrol applications.

The present study has offered *in situ* evidence that the AF36 biocontrol strain is successful at suppressing tissue proliferation, and subsequently, aflatoxin contamination by an aflatoxigenic *A. flavus* isolate. Use of a GFP-labeled, aflatoxin-producing isolate allowed us to easily track invasion and colonization for a better understanding of the competitive relationship between the two strains in corn kernels. Our findings support the theory of competitive exclusion, in favor of the biocontrol strain, based on its robust growth and proliferation. However, this study also points to valid concerns regarding the long-term use of non-aflatoxigenic fungi for suppression of native toxigenic strains in biocontrol strategies.

#### **ACKNOWLEDGMENTS**

Funding for the current project was provided through the USDA-ARS Specific Cooperative Agreement No. 58-6435-3-0121. The authors thank Mr. D. Ambrogio and Ms. S. Beltz at USDA-ARS-SRRC, for providing the fungal strains used in this study, Ms. C. Sickler, SRRCfor assistance with GFP imaging andMs. Y. Lau, GRI, MSU, Stennis Space Center, for assistance with statistical analyses and graphics. Thanks also to Dr. H. Abbas, ARS, Stoneville, MS, for graciously providing access to corn field for our experiments.

#### **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: 15 January 2014; paper pending published: 26 January 2014; accepted: 10 March 2014; published online: 27 March 2014.*

*Citation: Hruska Z, Rajasekaran K, Yao H, Kincaid R, Darlington D, Brown RL, Bhatnagar D and Cleveland TE (2014) Co-inoculation of aflatoxigenic and nonaflatoxigenic strains of Aspergillus flavus to study fungal invasion, colonization, and competition in maize kernels. Front. Microbiol. 5:122. doi: 10.3389/fmicb.2014. 00122*

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

*Copyright © 2014 Hruska, Rajasekaran, Yao, Kincaid, Darlington, Brown, Bhatnagar and Cleveland. 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.*

### Lipids in *Aspergillus flavus*-maize interaction

#### *Marzia Scarpari 1, Marta Punelli 1, Valeria Scala1, Marco Zaccaria1, Chiara Nobili 2, Matteo Ludovici 3, Emanuela Camera3, Anna A. Fabbri 1, Massimo Reverberi <sup>1</sup> \* and Corrado Fanelli <sup>1</sup>*

*<sup>1</sup> Dipartimento di Biologia Ambientale, Università Sapienza – Roma, Roma, Italy*

*<sup>2</sup> Unità Tecnica Sviluppo Sostenibile ed Innovazione del Sistema Agro-industriale, Laboratorio Innovazione Agroindustriale, ENEA C.R. Casaccia, Roma, Italy <sup>3</sup> IFO-S.Gallicano, Roma, Italy*

#### *Edited by:*

*Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran*

#### *Reviewed by:*

*Xueyan Shan, Mississippi State University, USA Gary L. Windham, USDA Agricultural Research Service, USA*

#### *\*Correspondence:*

*Massimo Reverberi, Dipartimento di Biologia Ambientale, Università Sapienza, L.go Cristina di Svezia 24, 00165 Roma, Italy e-mail: massimo.reverberi@ uniroma1.it*

In some filamentous fungi, the pathways related to the oxidative stress and oxylipins production are involved both in the process of host-recognition and in the pathogenic phase. In fact, recent studies have shown that the production of oxylipins in filamentous fungi, yeasts and chromists is also related to the development of the organism itself and to mechanisms of communication with the host at the cellular level. The oxylipins, also produced by the host during defense reactions, are able to induce sporulation and to regulate the biosynthesis of mycotoxins in several pathogenic fungi. In *A. flavus*, the oxylipins play a crucial role as signals for regulating the biosynthesis of aflatoxins, the conidiogenesis and the formation of sclerotia. To investigate the involvement of an oxylipins based cross-talk into *Z. mays* and *A. flavus* interaction, we analyzed the oxylipins profile of the wild type strain and of three mutants of *A. flavus* that are deleted at the Af*lox1* gene level also during maize kernel invasion. A lipidomic approach has been addressed through the use of LC-ToF-MS, followed by a statistical analysis of the principal components (PCA). The results showed the existence of a difference between the oxylipins profile generated by the WT and the mutants onto challenged maize. In relation to this, aflatoxin synthesis which is largely hampered *in vitro*, is intriguingly restored. These results highlight the important role of maize oxylipin in driving secondary metabolism in *A. flavus.*

**Keywords: maize kernels, aflatoxins, lipoxygenase, lipidomic, reverse genetic**

#### **INTRODUCTION**

*Zea mays* is one of the most cultivated grain crop. The yield, quality and safety of its kernels are continuously challenged by pathogens. Amongst these, the mycotoxigenic fungi are actually the most health hazardous. *A. flavus* under suitable conditions infect maize kernels both during development in the field as well as during the storage (Sheidegger and Payne, 2003). In most cases *A. flavus* produces the harmful and carcinogenic aflatoxins, among which B1 is considered by IARC (International Agency for Research on Cancer) as belonging to group 1, i.e., carcinogenic to humans and animals. Seed composition, notably lipid composition, may affect susceptibility to fungal infection and mycotoxins production (Reddy et al., 1992; Dall'Asta et al., 2012). Unsaturated fatty acids (FAs) have been frequently described as modulators of plant resistance pathway upon pathogen attack, even though their profile is strongly influenced by the environmental conditions experienced by plants during the entire growing season (Kachroo and Kachroo, 2009; Dall'Asta et al., 2012). Polyunsaturated FAs, released from membranes by lipases in response to attacks by biotic agents, play a key role in plantpathogen interaction either directly as free FAs or as precursors of oxylipins (Walley et al., 2013). As an example, linoleic acid levels contribute to the regulation of development, seed colonization and mycotoxin production in *Aspergillus* spp. (Calvo et al., 1999). FAs may also serve as precursors in oxylipins synthesis. These oxidized FA-derived compounds are involved in plant-fungi interaction. It is possible to suggest the presence of an "oxylipins signature profile" typical for every pathogen (Tsitsigiannis and Keller, 2007; Reverberi et al., 2010). Similarly, plants contain phyto-oxylipin pools conferring an oxylipin signature on a given organelle, tissue or plant (Blée, 2002; Camera et al., 2004). Lipoxygenases (LOX) are amongst the main oxylipinproducing enzymes in all organisms. LOX are widespread in animals, plants and fungi, and their products have a variety of biological functions, involved in important physiological processes, for instance, the events of infection (in plants) or inflammatory events (in animals) and the structural and metabolic changes in animal cells. LOX enzymes have been found in plants, animals, fungi, and in all organisms that possess the required substrates, i.e., PUFAs (Polyunsaturated Fatty Acids) (Feussner and Wasternack, 2002).

Lipoperoxidation seem to play an important role in the induction of the production of certain mycotoxins including aflatoxins. *In vitro*, synthetic inducers of lipid peroxidation, such as cumene hydroperoxide, stimulate the production of aflatoxins, when added to culture media inoculated with toxigenic strains of *A. flavus* and *A. parasiticus* (Passi et al., 1985). *In vivo*, the amount of toxin produced on sunflower seeds differs according to the age of seeds and the amount of peroxides found in the oil obtained from them; older will be the seeds, the greater the number of peroxides, and will be higher the amount of aflatoxin produced by these fungi (Passi et al., 1984). Several studies have shown how oxylipins in plants are capable of inhibiting the mycelial growth and germination of spores (Prost et al., 2005). The similar structure between the oxylipins of fungi and plants suggest the hypothesis that these molecules are very important for the host-pathogen communication.

According to Tsitsigiannis and Keller (2007), oxylipins are able to modulate the sporulation through specific transcription factors related to cleistothecia (NsdD) and conidiospores (BrlA) formation; they may control homeostasis through the lipogenic transcription factors SREBP -1 and SREBP -2 and induce the development of the fungus and the production of secondary metabolites, such as penicillin, through the pathway of G proteins involving protein kinase A (PkaA). Studies on *A. nidulans* and *A. fumigatus* has led to hypothesize that oxylipins are able to escape from the cell through specific transport proteins, act as autocrine or paracrine ligand with membrane receptors, GPCRs, both of the fungal cell itself than that of the host, and activate a downstream signaling cascade (Tsitsigiannis and Keller, 2007).

Linoleic acid host-derived oxylipins may drive aflatoxin synthesis, as widely demonstrated (Burow et al., 1997; Brodhagen et al., 2008; Reverberi et al., 2012). The oxylipin-driven actions are further complicated by the evidence that even fungal pathogens produce oxylipin during the interaction with the host. A recent evidence of the central role of oxylipins in the regulation in the biosynthesis of mycotoxins and fungal sporulation was obtained in *A. ochraceus* interacting with viable seeds of wheat (Reverberi et al., 2010). In this study, a mutant defective for a *lox*-like gene produces minimal amounts of 13S-HPODE and show a strong decrease in the production of ochratoxin A, a delayed formation of conidia and an increase in the production of sclerotia. Furthermore, seeds infected with this mutant are not able to accumulate normal levels of 9S-HPODE nor to induce the expression of the gene of defense PR1, suggesting that the fungal oxylipins may modulate the response of the host defense. These data demonstrate that, the plants oxylipins influence the processes of development of *Aspergillus* because of their similarity with those of fungi. It has been observed that the PSIB α oxylipins of *A. nidulans*, derived from linoleic acid, are very similar to those produced from fatty acids of seeds and that these, after infection, are able to regulate the development of the fungus mimicking and/or interfering with the signals that regulate the processes sporogenesis (Prost et al., 2005).

In this paper we demonstrate that combining previous microarray data (Reverberi et al., 2013) with reverse genetic approach and with a lipidomic procedure may provide evidence of the key role played by lipids and oxylipins in the interaction between *A. flavus* and maize kernels, notably in influencing aflatoxin biosynthesis.

#### **MATERIALS AND METHODS**

#### **FUNGAL STRAINS AND CULTURE CONDITIONS**

*Aspergillus flavus* NRRL 3357, aflatoxin B1 producer and an argD/uracil double auxotroph mutant (AFC-1) (Zhu-Mei et al., 2007) were used in these studies. An Af*lox1* transformant was generated from AFC-1 as described below. The fungal strains were

#### **FUNGAL GROWTH, CONIDIOGENESIS, AND AFLATOXIN PRODUCTION** *In vitro assays*

*Fifty* ml of Potato Dextrose Broth (PDB) (Difco, USA) in 100 mL Erlenmeyer flasks were inoculated with the WT or the 3 strains of Af*lox1*- (1.1; 3.2; 5.1), using 0.1 mL of conidial suspension (∼10<sup>6</sup> conidia) for each flask (*<sup>n</sup>* <sup>=</sup> 3); the cultures were then incubated at 30◦C for different time periods (0–14 days post inoculation; total *n* = 5 time intervals). *N* = 2 biological replica were carried out. At each point in time, fungal growth was determined by weighting the mycelium after filtration (Millipore filters, 0.45μm) and drying it for 48 h at 80◦C (d.w.). Conidia formation was determined at each time interval by washing the mycelia with a solution of triton 0.01% w/v, taking 0.5 ml of this solution and calculating conidia's number using a hemacytometer. The filtered mycelia were lyophilized and weighed for determining the amount of oxylipins and performing molecular analyses.

#### *In vivo assays*

Each stock of maize seeds (30 g) was superficially sterilized (2% v/v NaClO solution), rinsed threefold by sterilized distilled water, moistened up to 0.90 aw and then inoculated with 250μL of *A. flavus* conidia (*n* = 3 Erlenmeyer flasks for each test) suspension (104 conidia/mL) at 30◦C. Each biological replica (*<sup>n</sup>* <sup>=</sup> 2) consisted in non-contaminated (control) and contaminated seeds of maize kernels, harvested at *n* = 5 different time intervals (from 0 up to 14 days post inoculation, dpi).

Aflatoxin production was analyzed from cultures of the WT and Af*lox1* strains grown both in PDB and onto viable maize seeds following extraction with chloroform:methanol (2:1 v/v) procedure. The extracts were collected, the volume was reduced under a stream of nitrogen and the quantitative analyses were carried out by HPLC, as previously reported (Reverberi et al., 2008).

#### **PLASMIDS AND TRANSFORMATION**

To obtain the disruption cassette for the Af*lox1* gene,5 - and 3 - flanking ends of the gene have been amplified, inserting in the primers (see below) a complementary tail to the 5 -end (*primer* #3) and 3 (*primer* #4) of the *argD* gene; *argD* either has been amplified by inserting into the primers tails complementary to the 3 -end of 5 -flanking region (*primer* #2) and to the *5* -end of the 3 -flanking region (*primer* #5)

#### #1 (*for*) 5 -AAAGGCTGGCACGTAGAAGA-3

#2 (*for*) 5 -CCTAGTCGAGACGGAAAACGTAATTGCGGAGC AAATCACA 3

#3 (*rev*) 5 -TGTGATTTGCTCCGCAATTACGTTTTCCGTCT CGACTAGG-3

#4 (*for*) 5 -GAATCCCTGCATCAGAGGAATCAATTCCATCA TTCCACGA-3

#5 (*rev*) 5 -TCGTGGAATGATGGAATTGATTCCTCTGATGC AGGGATTC-3

#### #6 (*rev*) 5 -CTACTGTGGCCTTTCCCAAA-3

DNA extracted (50 ng) from *A. flavus* NRRL 3357 was amplified in a thermal mastercycler gradient (Eppendorf, Germany) following amplification steps (95◦C × 2 min; 95◦C × 30 s, 55◦C × 45 s, 72◦C × 1 min × 35 times; 72◦C × 8 min). 5 -flanking region (1386 bp) has been amplified by *primers* #1 and #3; *argD* gene fragment (2217 bp) by *primers* #2 e #5; 3 -flanking region (1321 bp) by *primers* #4 and #6. After the obtainment of the 3 fragments, a PCR with external primers, #1 and #6, was carried on for obtaining gene disruption cassette; a DNA *mix* composed by 104 ng of 5 *-flank* + 168 ng of argD + 100 ng of 3 *-flank* was used as template for the amplification. Protoplast transformations of *A. flavus* were performed by the polyethyleneglycol method as described elsewhere (Woo et al., 1999).

#### **SELECTION OF TRANSFORMANTS**

The selection of Af*lox1* transformant strains was conducted at 30◦C on CDA containing 30 mM uracyle; putative transformants were selected, transferred to fresh selective medium, and allowed to sporulate. To obtain homokaryons, single spores were isolated from each selected heterokaryotic transformant and transferred to fresh selective medium. This monoconidial transfer was conducted three times. Finally, 20 monoconidial progenies were selected and further sub-cultured to determine the occurrence of abortive transformants. The stability of these transformants was also tested by two additional single-spore transfers on nonselective medium and then again on selective medium, and by several mycelial transfers on selective plates.

#### **SOUTHERN BLOT HYBRIDIZATION**

For Southern blot analysis, 10μg of genomic DNA from *A. flavus* NRRL 3357 and Af*lox1* was completely restricted with *Eco*RI (10 U) at 37◦C for 4 h in the manufacturer's buffer at the recommended concentrations (Fermentas, Germany). *EcoR*Idigested DNA fragments were separated by electrophoresis for 3 h and 30 min at 40 V on 0.8% w/v agarose gel in TAE buffer. DIG-labeled *Hin*dIII cut lambda (λ) (Roche, Swiss) was used as MW standard. Fluorescent DNA probes were prepared according to the PCR DIG-labeling mix method (Roche, Swiss). The membranes were pre-hybridized according to the instructions of the manufacturer of the DIG- detection kit, at 64◦C in DIG-easy buffer (Roche, Swiss); they were then hybridized for 12–16 h in the same buffer containing 250 ng of freshly denatured digoxigenin *argD* probe at 65◦C.

#### **RT-PCR ANALYSES**

Total RNA from 100 mg of freeze-dried mycelia was extracted using the Tri-Reagent protocol (Sigma-Aldrich, USA) and was quantified by spectrophotometry, determining the optical density at 260 nm. RNA was treated with RNAse-free DNAse I and then re-suspended in 20μL of DEPC-treated water. RNA was extracted at different points in time (from 24 to 168 hpi; 3 tubes for each point in time) from *A. flavus* WT and Af*lox1*-CDB cultures and was used to develop *lox1* SYBR green RT-PCR assay, as previously reported (Reverberi et al., 2010). Gene expression in the WT strain and the Af*lox1* transformants were also measured by comparing mRNA levels in the different time intervals with their own basal expressions at the baseline, i.e., after conidia germination (time 0). *A. flavus* β-tubulin RNA was used as the housekeeping gene to normalize the differences in total RNA target input and quality and in RT efficiency, using specific primers as Afβtub (Afβtub\_for GGAAGTCAGAAGCAGCCATC; Afβtub\_rev GTGACCACCTGTCTCCGTTT).

#### **LIPOXYGENASE ASSAY**

The LOX activity in the WT strain and in the Af*lox1* transformants grown in PDB medium was assayed 0 to 14 days after inoculation using a Beckman DU 530 spectrophotometer by following up the formation of conjugate dienes at 234 nm as previously reported (Reverberi et al., 2008). In order to exclude the possible interference of laccase activity, KCN 1 mM was added to the reaction mixture before the spectrophotometric assay (Gülçin et al., 2005).

#### **LIPIDOMIC UNTARGETED ANALYSIS**

Free, conjugated and modified fatty acids were extracted as described by Stumpe et al. (2005) with slight modifications. An amount of maize kernels (100 g) and/or mycelia (20 mg) were lyophilized and ground in liquid nitrogen. An aliquot of 20 mg was collected in a clean tube and added of 1 mL of the extraction medium (hexane: 2-propanol 3:2 v/v, containing 0.0025% butylated hydroxytoluene w/v) and 5μL heptadecanoic acid standard solution 2 mg/ml in EtOH added as the internal standard reference as reported in Scala et al. (2013). Separation and accurate mass measurements of lipid compounds was performed with a 1200 series rapid resolution HPLC coupled with a G6220A series time of flight mass spectrometer (ToF-MS, Agilent Technologies, CA, USA) equipped with an electrospray (ESI) interface operating in the negative ion mode. LC/MS-ESI ToF data were acquired and deconvoluted into individual chemical peaks using the Mass Hunter™ acquisition software. Untargeted and semi-targeted mining of the HPLC/MS-ESI ToF data were performed with the molecular feature extraction (MFE) algorithm in the Mass Hunter™ software. All the analyses were performed as previously described in Scala et al. (2013).

#### **STATISTICS**

All the experiments were carried out in three replicates of two biological replica. The values presented in figures and tables are the mean ± SE of 6 different results. The mean values were compared by using the Mann-Whitney test; *p*-values above 0.05 were considered not significant. Analysis of variance (ANOVA) was applied in the comparison of the treatments, and significance of differences were tested at 95% confidence by Fisher's LSD test which is a least significant difference (LSD) method consisting in a two-step testing procedure for pair wise comparisons of several treatment groups. Calculations were performed using XLSTAT Addinsoft software [45]. Concerning the lipidomic approach, the MFE algorithm was used to extract individual molecular species by their accurate mass detected with the ToF MS. The detected species were characterized by accurate mass, isotopic pattern and absolute abundance and lists of molecular features were generated from each analyzed sample and converted into compound exchange files (CEF), which were then processed with Mass Profiler Professional (MPP) 12 (see below). Molecular features detected in the HPLC-MS system were aligned by their retention time (RT) in the chromatographic runs, and accurate mass axis in order to compare their expression across the different maize hybrids in different growth stages. Compounds detected in the different samples and presenting consistent RT (shift below 6 % of the RT) and accurate mass (mass error below 6 ppm) were assigned as the same molecular species. Relative abundance of individual features was obtained by normalizing their peak area by the area of the ISTD. Among all the metabolites, only those features consistently detected throughout all the analyzed maize samples were selected for further statistical analysis. ANOVA with Tukey's *post-hoc* test was performed on the entities detected with 100% frequency in the different samples. Fold changes of filtered entities were compared between WT, mutant strains and maize alone and significance was determined by Student's *T*-test. Differential expression was evaluated at each sampling time, including harvest, and visualized by Volcano plots. Changes higher than 2-folds, with *p* < 0.05 after the Benjamini-Hochberg correction, were considered as significant. Principal components analysis (PCA) was then performed on entities filtered following grouping of samples according fumonisins amounts. Compound identification and annotation was performed using the METLIN Personal Metabolite Database by means of the ID browser tool, and the Molecular Formula Generator algorithm. The LIPIDMAPS (http://www.lipidmaps.org/) database was used to infer compound identity. The annotation of free fatty acids (Fas) reported the number of carbon atoms and of double bonds. Other lipids were annotated consistently with the names reported on LIPIDMAPS.

#### **RESULTS**

#### **A***flox1* **DELETION: MOLECULAR AND PHYSIOLOGICAL CHARACTERIZATION OF DELETED STRAINS**

Microarray analysis performed in previous study (Reverberi et al., 2013), indicated that a *lox* gene (corresponding to the affy probe 2911\_m00089), coding for a Mn-dependent lipoxygenase (BLASTX result), was up-regulated in *A. flavus* during the pathogenic exploitation of maize kernels. To uncover the relation of oxylipin production in the fungus and in the maize kernels with aflatoxin synthesis Af*lox1* strains have been generated by inserting a deletion cassette containing *argD* as selectable marker (AFC-1 strain is auxotroph for arginine—see Methods section) (**Figure 1**) and used for maize kernel infection.

Transformants selected as indicated in the Methods section, showed striking differences with the native strain AFC-1. These differences concern the growth (30–52% slower both in rich - PDB—and minimal—CD- media) and conidiogenesis (−90% compared to AFC-1) (**Figure 2**). This latter was completely recovered when Af*lox1* strains were inoculated onto viable maize kernels (data not shown).

To confirm and quantify deletion cassette insertion into Af*lox1*strains a southern blot analysis has been performed by using *EcoR*I-digested genomic DNA of the putative Af*lox1* deleted mutants, AFC-1 and the WT with which AFC-1 was previously originated (i.e., *A. flavus* NRRL3357). Hybridization was performed by using an *argD* probe. Hybridization pattern confirm a single insertion of the deletion cassette into Af*lox1*- 1.1 and 3.2 strains whereas, a putative double insertion event can be suggested for Af*lox1*-5.1 strain (**Figure 3**).

Af*lox1* mRNA expression was analyzed into AFC-1 and in the 3 mutant strains selected for further analysis. All the fungal strains were grown in PDB and harvested at different time post inoculation (0–14 dpi). Results shown how *lox1* mRNA expression is under the detection limit into the Af*lox1* strains also compared to gene expression into AFC-1 strain (**Figure 4**).

Since *lox1* should be present in multiple copies in *A. flavus* as well as in other toxigenic fungi (Reverberi et al., 2010) we quantify the amount of LOX activity, by quantifying diene conjugates at λ234, into AFC-1 and Af*lox1* mutant strains. It emerges that even if severely unpaired, a residual LOX activity is still present into deleted strains even if its amount is significantly (*P* < 0.001) to AFC-1 (**Figure 5A**). LC-TOF analysis of lipids extracted indicated that the synthesis of HPODEs (the main LOX-related oxylipins—Feussner and Wasternack, 2002) is consistently hampered in all Af*lox1* mutant strains compared to AFC-1 strain (**Figure 5B**).

Aflatoxin synthesis has been monitored at different time after inoculation (0-14 dpi) in PDB medium of the 4 fungal strains (AFC-1, Afl*ox1*-\_1.1, Afl*ox1*-\_3.2, Afl*ox1*-\_5.1). It emerged that Af*lox1* deleted strains are unable to synthesize AFB1 in this culture conditions whereas AFC-1 has an AFB1 production similar to the native wild type NRRL 3357 (102 ± 15.2 ppb at 14 dpi) (**Figure 6**).

#### **MAIZE KERNELS INOCULATION WITH AF***lox1* **DELETED STRAINS**

To study the behavior of Af*lox1* mutant strains *in vivo* compared to the AFC-1 strain, an inoculum was performed directly on viable corn seeds in environmental conditions that mimicked the natural condition of pathogenesis. Since the mutants do not produce conidia, *in vitro,* the inoculation of was performed by a small portion (200 mg) of fresh mycelium into 10 g of viable seeds of maize, to make the test more homogeneous as possible, even the mycelium of the AFC-1 strain was inoculated in the same way, i.e., by inoculating the same portion of mycelium that had not differentiated conidia into maize seeds. The three mutant strains, already at early time of analysis (3 dpi), intriguingly recover the ability to produce conidia although the overall growth appeared anyway to be slower compared to the AFC-1 strain (data not shown). Aflatoxins biosynthesis was then evaluated in times of analysis similar to the *in vitro* test (3, 7, 10, and 14 dpi). The results of this analysis have highlighted the recovery, and in some cases the stimulation (in strain Af*lox1*-\_3.2), of the capacity for synthesizing aflatoxin B1 in the three Af*lox1* deleted strains examined (**Figure 7**).

#### **OXYLIPIN PROFILE INTO MAIZE KERNELS CHALLENGED WITH** *A. flavus*

Lipidome analysis and the characterization of specific oxylipin has been performed into maize kernels challenged with the 4 strains—1 native and 3 Af*lox1* deleted strains—of *A. flavus*

**FIGURE 2 | Phenotypic comparison among Af***lox1* **strains with the native AFC-1 strain.**

**FIGURE 3 | Gel blot analysis of Af***lox1* **gene replacement mutants.** Genomic DNA was isolated from the WT strain, the arginine auxotroph AFC-1 strain and gene replacement transformants (Af*lox1*-). The blot was hybridized at 65◦C with a 0.65-Kb *argD* DIG-labeled probe. Lane 1 DIG-labeled λhindIII (Roche) used as molecular weight marker; lane 2 *EcoR*I-restricted genomic DNA of *A. flavus* WT strain; lane 3 *EcoR*I-restricted genomic DNA of *A. flavus* AFC-1; lane 4–6 *EcoR*I-restricted genomic DNA of Af*lox1* strains (clone number 3 out of a set of 20 transformants screened); lane 7 *argD* 0.65 Kb PCR fragment.

(**Figure S1**). Through *Mass Hunter* software (Agilent tech, USA) it has been possible to pinpoint 5 oxylipins amongst the *plethora* of lipid compounds highlighted by lipidomic approach (**Figures S2**, **S3**). Notably, HPODEs, as described elsewhere (Burow et al.,

1997), and di-HODEs result produced at a different amount by pathogen- challenged and unchallenged maize kernels. By simply comparing the chromatograms it is possible to highlight, concerning HPODEs, striking difference among the different samples, with particular regard to HPODEs produced by maize kernels inoculated with AFC-1 compared to those produced when the seeds were challenged with the 3 Af*lox1* deleted strains. Furthermore, the same analysis was carried out on maize seeds un-inoculated in the same experimental conditions. The HPODE profile of the not inoculated seeds is considerably lower compared to preceding, to indicate that the presence of the fungus can effectively elicit in somehow the production of oxylipin in maize too (**Figure 8**).

At this point, we evaluated the level of production of di-HODE and HPODE in inoculated relative to un-inoculated maize kernels. In both cases oxylipin production in the seeds infected by mutant strains results higher than that in the seeds infected with the WT (**Figures 9A,B**).

#### **DISCUSSION**

One of the principal concern related to the production of cereal seed is the contamination with mycotoxins produced by *Aspergillus* spp. and *Fusarium* spp. In particular, Aspergilli are

**FIGURE 5 | (A)** LOX activity (U/mg protein) measured as diene conjugates formation at 234 mn in AFC-1 and Af*lox1* deleted strains in PDB at 30◦C at different days after inoculation (3–14 dpi). **(B)** HPODE synthesis (expressed as

commonly associated with crops such as maize, wheat, cotton, peanut, characterized by seeds rich in lipids, suggesting an important role played by the host lipids in influencing the ability of the pathogen to synthesize mycotoxins.

The role of lipids in the recognition and plant-pathogen communication has recently been revised and re-evaluated. Lipids and oxylipins currently represent one of the most effective control signals for morphogenesis, development and virulence of pathogens (Gao and Kolomiets, 2009; Christensen and Kolomiets, 2010; Reverberi et al., 2010).

Several studies have shown a direct correlation between the lipoperoxidative processes that occur in the seeds of various plant species (maize, sunflower) and the production of aflatoxins by *A. flavus* and *A. parasiticus* (Fabbri et al., 1983; Burow et al., 1997; Gao et al., 2009). In particular, experiments on maize have shown that the processes of lipoperoxidation in the seeds induce an alteration of the balance between oxidants and antioxidants, in relative abundance) in AFC-1 and Af*lox1* deleted strains in PDB grown at 30◦C at different days after inoculation (3–14 dpi). Results are the mean (±SE) of a total of six replications deriving from two independent experiments.

favor of oxidants accumulation into the fungal cell thus stimulating biosynthesis of mycotoxins in *A. parasiticus* (Reverberi et al., 2007). The production of the toxins could be interpreted as the result of a fungal cell response to an incomplete scavenging of reactive oxygen species at the intracellular level (Reverberi et al., 2008; Hong et al., 2013).

independent experiments.

Several studies showed that plants oxylipins are able of inhibiting the mycelial growth and germination of fungal spores (Prost et al., 2005). The similar structure found between fungal oxylipins and those of plants has led to the hypothesis that the latter can "mimic" the corresponding oxylipins acting directly on the physiology of the fungal pathogen try to alter its pathogenic behavior. Indeed, *in vitro* experiments have shown that linoleic acid and two of its oxidation products, namely 9S- and 13S-HPODE, play a significant role in differentiation processes in *A. nidulans*, in *A. flavus* and *A. parasiticus* (Burow et al., 1997; Gardner, 1998; Calvo et al., 1999; Gao et al., 2009).

Af*lox1* gene, Mn-lipoxygenase (LOX) encoding, knock-out mutants showed a significant reduction of mycelial growth and a complete inability to differentiate conidia and produce aflatoxins, if grown *in vitro* conditions. As shown by recent studies (Tsitsigiannis and Keller, 2007) in *A. nidulans* and *A. fumigatus*, the change in lipid metabolism has a strong influence on the production of sexual spores and the ability to produce toxins (Tsitsigiannis and Keller, 2006). The data obtained herein are therefore in line with the others reported in the literature. Intriguingly, inoculating Af*lox1* mutants on viable and germinating maize seeds, it was possible to highlight the almost total recovery of conidiogenesis and production of aflatoxins. Oxylipins released from the seed, probably simulating/substituting fungal ones, induce the activation of secondary metabolism and changes in morphogenesis, as also suggested in Brodhagen et al. (2008). In these studies maize 9-LOX (Zm*LOX3*) was cloned in a mutant of *A. nidulans* unable to produce conidia. The oxylipins produced by the Zm*LOX3* gene are able to induce again the conidiogenesis, suggesting a strong correlation between host-oxylipins and production of conidia and providing a clear proof of the mutual cross-talk between plants and fungi mediated by oxylipins.

The LC-TOF and chemometric based lipidomic analysis of the maize kernels infected with the mutants and the wild-type strain of *A. flavus* showed a complete modification of the lipid and oxylipin profile. A higher production of HPODE and di-HODE in maize seeds inoculated with Af*lox1* strains compared to those inoculated with the wild-type strain may be evidenced. The comparison with the production of the same oxylipins in non-inoculated maize seeds makes it possible to demonstrate not only that production is higher in seeds inoculated with the mutants, but also that the production of these particular oxylipins, is greatly increased if the host is interacting with the pathogen. To confirm this, it was also evaluated the production of HPODE and di-HODE in mycelia of the strains grew on media containing maize seeds (data not shown). The production was significantly reduced, showing that synthesis of these oxylipins is specifically charged to the seed. Recent studies have, in fact, show that LOX host genes expression is driven by the pathogen in *Aspergillus* infections leading to changes in the profile of the plant oxylipins (Tsitsigiannis and Keller, 2007). These results suggest that some derivatives of 9 -LOX may act as signal molecules for the production of mycotoxins and conidia. Experiments based on a knock-out mutant of the 9 -LOX gene of maize (Zm*LOX*3) confirm this hypothesis, showing that the lack of derivatives in the seed of the 9-LOX affect the pathogenesis and production of conidia and mycotoxins (Gao et al., 2007) and therefore, the importance of the role played by the host oxylipins for its exploitation by some pathogenic fungi. It can be suggested that the presence of the fungus (PAMP?) elicits the production of such oxylipins in the plant, which, in turn may influence the development of the fungus itself by stimulating certain physiological processes such as the production of conidia and the biosynthesis of toxin. The way in which these process aid pathogenesis is not yet elucidated.

#### **SUPPLEMENTARY MATERIAL**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb.2014. 00074/abstract

**Figure S1 | Representative Total ion chromatogram of lipid compounds of AFC-1 grown on maize kernels after 7 dpi.** The different lipid compounds

were sorted for exact mass and RT. A total of 1217 different lipid compounds were separated.

**Figure S2 | PCA score plot of data generated by HPLC-ESI/TOF-MS analysis of of lipid extract of maize kernels non-inoculated or inoculated with AFC-1 (WT) and the 3 deleted Af***lox1* **strains (1.1/3.2/5.1) at 3, 7, 10 and 14 dpi.** The results of the PCA analysis referred to three separate experiments performed in duplicate.

#### **Figure S3 | Eulero–Venn's diagram showing compounds differences present in maize kernels inoculated with AFC-1 (WT) or the 3 deleted**

**Af***lox1* **strains.** The labels of each circle report the number of compounds selected on the basis of significant difference and 100% frequency in at least one group.

#### **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: 30 December 2013; paper pending published: 22 January 2014; accepted: 11 February 2014; published online: 27 February 2014.*

*Citation: Scarpari M, Punelli M, Scala V, Zaccaria M, Nobili C, Ludovici M, Camera E, Fabbri AA, Reverberi M and Fanelli C (2014) Lipids in Aspergillus flavus-maize interaction. Front. Microbiol. 5:74. doi: 10.3389/fmicb.2014.00074*

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

*Copyright © 2014 Scarpari, Punelli, Scala, Zaccaria, Nobili, Ludovici, Camera, Fabbri, Reverberi and Fanelli. 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: 21 July 2014 doi: 10.3389/fmicb.2014.00364

### Toward elucidation of genetic and functional genetic mechanisms in corn host resistance to Aspergillus flavus infection and aflatoxin contamination

#### *Xueyan Shan1\* and W. Paul Williams <sup>2</sup>*

<sup>1</sup> Department of Biochemistry, Molecular Biology, Entomology, and Plant Pathology, Mississippi State University, Mississippi, MS, USA <sup>2</sup> Agricultural Research Service, United States Department of Agriculture, Corn Host Plant Resistance Research Unit, Mississippi, MS, USA

#### *Edited by:*

Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran

#### *Reviewed by:*

Perng-Kuang Chang, Southern Regional Research Center, USA Jong H. Kim, United States Department of Agriculture – Agricultural Research Service, USA

#### *\*Correspondence:*

Xueyan Shan, Department of Biochemistry, Molecular Biology, Entomology, and Plant Pathology, Mississippi State University, Mississippi, MS 39762, USA e-mail: xshan@bch.msstate.edu

Aflatoxins are carcinogenic mycotoxins produced by some species in theAspergillus genus, such as A. flavus and A. parasiticus. Contamination of aflatoxins in corn profusely happens at pre-harvest stage when heat and drought field conditions favor A. flavus colonization. Commercial corn hybrids are generally susceptible to A. flavus infection. An ideal strategy for preventing aflatoxin contamination is through the enhancement of corn host resistance to Aspergillus infection and aflatoxin production. Constant efforts have been made by corn breeders to develop resistant corn genotypes. Significantly low levels of aflatoxin accumulation have been determined in certain resistant corn inbred lines. A number of reports of quantitative trait loci have provided compelling evidence supporting the quantitative trait genetic basis of corn host resistance to aflatoxin accumulation. Important findings have also been obtained from the investigation on candidate resistance genes through transcriptomics approach. Elucidation of molecular mechanisms will provide indepth understanding of the host–pathogen interactions and hence facilitate the breeding of corn with resistance to A. flavus infection and aflatoxin accumulation.

**Keywords:** *Aspergillus flavus***, aflatoxin, corn, host resistance, transcriptomics**

#### **INTRODUCTION**

Aflatoxins are produced by *Aspergillus flavus*, an ear rot fungus that infects corn. Aflatoxins are carcinogenic to humans and animals. Consequently, they are of great concern as food contaminants and are closely monitored in oily-seeded crops and dairy products (Payne, 1992; Gourama and Bullerman, 1995). Aflatoxins are designated as B1, B2, G1, G2, M1, and M2 types based on their molecular structures. Aflatoxin B1 is the most toxic form which is found in fungal-infected corn kernels (Bhatnagar et al., 2006). The production of aflatoxin B1 starts soon after the colonization of *A. flavus* in the developing corn kernels (Thompson et al., 1983). Accumulation of aflatoxin B1 continues during the maturation of corn. Commercial corn hybrids are generally susceptible to *A. flavus* infection. The Food and Drug Administration (FDA) has imposed a strict limit of 20 ppb on aflatoxin concentration present in corn (Aflatoxin Guidelines, http://www.admcrs.com/Aflatoxin.html). Aflatoxins remain stable under food processing conditions. Once corn is contaminated with aflatoxins, few detoxification options are available. Therefore, the prevention of aflatoxin accumulation should start in pre-harvest corn stage.

*Aspergillus flavus* infection and aflatoxin accumulation in preharvest corn largely depend on the local weather and field conditions. High temperature and drought conditions are in favor of aflatoxin production, thus posing the high risk of an aflatoxin outbreak. The colonization of *A. flavus* also frequently occurs at sites damaged by insects. In many locations of the southeastern United States, including Mississippi, *A. flavus* infection in corn is a chronic problem. In 1998, a severe aflatoxin outbreak in corn fields occurred in Mississippi and throughout the Southeast, resulting in considerable economic loss to corn farmers (Robens and Cardwell, 2005). In Mississippi, efforts to reduce aflatoxin contamination in corn at the pre-harvest stage have focused on multiple strategies including bio-control using atoxigenic *A*. *flavus* strains, optimal agronomic practices (irrigation, fungicides, planting dates), and breeding for host resistance (Larson, 1997; Cleveland et al., 2004). While notable improvement can be achieved through optimizing cultural practices for control of pre-harvest aflatoxin contamination, combination strategy by including breeding for corn host resistance would be the most promising and effective avenue.

#### **RESISTANT CORN GERMPLASM DEVELOPED BY USDA–ARS AT MISSISSIPPI STATE**

As early as in the 1970s, USDA–ARS scientists at Mississippi State University initiated breeding programs for screening and developing corn germplasm with resistance to *A. flavus* infection and aflatoxin accumulation. Reliable techniques (such as side-needle inoculation) were first developed to assure that all germplasm lines were treated with the same amount of fungal inoculums (Zummo and Scott, 1989; Windham et al., 2005, 2009; Buckley et al., 2006). Numerous germplasm and breeding lines were screened with such consistent methods. The results of these evaluations established the foundation of today's collection of Mississippi corn inbred lines that possess naturally occurring resistance to aflatoxin accumulation (Windham and Williams, 1998, 1999, 2002; Windham et al., 1999, 2010; Williams et al., 2002, 2005a,b, 2008; Williams, 2006). Resistant corn inbred lines were developed by consecutive

selfing and selection against aflatoxin accumulation. In 1988, the first inbred line Mp313E with resistance to *A. flavus* infection was released (Scott and Zummo, 1990). Mp313E was developed from germplasm Tuxpan. Since the release of Mp313E, additional resistant corn germplasm lines (Mp420, Mp715, Mp717, Mp718, and Mp719) have also been released; all exhibit significantly low levels of aflatoxin accumulation under artificial inoculation conditions (Scott and Zummo, 1992; Williams and Windham, 2001, 2006, 2012). Mp715, Mp718, and Mp719 also have lineages tracking back to germplasm Tuxpan. Tuxpan is a US derivative of Tuxpeño, one of the most productive and successful Mexican races of corn.

In the continual field studies conducted under artificial inoculation conditions, aflatoxin levels were evaluated for selected resistant corn inbred lines, single-cross hybrids from the selected resistant corn inbred lines, and a set of commercial corn hybrids. In 1999, aflatoxin level in the resistant corn inbred line Mp715 was 24 ppb which was contrasted to the level of 1622 ppb in SC212m, a susceptible corn inbred line. In 1996, aflatoxin level in the single-cross hybrid Mp715 × Mp313E was 18 ppb compared to 1532 ppb in the single cross of GA209 × SC212m. In 2001, aflatoxin level for Mp715 was 14 ppb compared to 9289 ppb for SC212M. In 2008, aflatoxin level was as low as 17 ppb for the single-cross hybrid of Mp313E × Mp717, whereas it was as high as 10800 ppb for a commercial hybrid. Similar

results were observed in 2009, where aflatoxin levels ranged from 5 ppb for the single-cross hybrid of Mp313E × NC388 to 1992 ppb for a commercial hybrid. From these studies, the lowest aflatoxin levels were found in the single-cross hybrids of the resistant inbred lines developed by the USDA–ARS Corn Host Plant Resistance Research Unit (CHPRRU) at Mississippi State. Four of the single-cross hybrids had aflatoxin levels lower than the FDA action level of 20 ppb in the 2008 and 2009 field studies. Commercial corn hybrids generally contained consistently high levels of aflatoxins in the studies conducted in 2008 and in 2009 at Mississippi State (Daves et al., 2010).

#### **THE IDENTIFICATION OF MAJOR RESISTANCE QUANTITATIVE TRAIT LOCI (QTLs)**

The resistant corn inbred lines exhibited significant general combining ability for reduced aflatoxin accumulation in a diallel cross, indicating good breeding values for the resistance trait (Williams et al., 2007). To incorporate the resistance into commercially acceptable lines for corn farmers, CHPRRU has made important progress in combining traditional and molecular breeding methods to develop resistant corn lines. Among the resistant corn germplasm lines developed by CHPRRU at Mississippi State, Mp313E and Mp715 are two of the most important parental lines used for breeding of *A. flavus* resistance and aflatoxin reduction.

2012).

separate genome-wide microarray study. Eight genes were located to

They have been used as resistance donors for many of the new experimental lines in the CHPRRU breeding programs. Two quantitative trait loci (QTL) mapping populations with Mp313E as a parental line were developed and investigated. The population from Mp313E × Va35 included 216 F2:3 families and was evaluated for aflatoxin accumulation over 3 years (Davis et al., 2000; Willcox et al., 2013). A total of 15 QTL regions were identified (**Figure 1**). Four QTLs on chromosomes 1, 4, and 9 were above a significance level of 23.58 in likelihood ratio. The chromosome 4 (bin 4.04–4.08) QTL was associated with Mp313E. The Mp313E × B73 population contains 210 F2:3 families and was also evaluated for 3 years (Brooks et al., 2005). A total of seven QTL regions were identified (**Figure 1**). The two QTLs on chromosomes 2 and 4 were most significant and were associated with Mp313E. From these studies, up to 48% of the genotypic effects of the resistance can be explained by the QTLs identified on chromosomes 2, 3, and 4 from Mp313E. Significant progress has been made to increase corn host plant resistance by DNA marker-assisted breeding. However, a major obstacle has been that the genomic regions containing the QTLs are large and indecipherable without the knowledge of the underlying genetic and molecular information. Closely linked molecular markers to the QTLs, ideally from resistance genes, are needed to expedite the breeding process and to reduce the breeding cycles.

#### **COMPARATIVE GENOME-WIDE GENE EXPRESSION STUDY ON SELECTED CORN INBRED LINES**

Corn host resistance to *A. flavus* infection and aflatoxin accumulation is a quantitative trait potentially controlled by many genes. Characterization of the corresponding genes and their effects is essential for the breeding of resistant corn lines. To characterize genes involved in the corn host plant resistance, a microarray study was conducted to investigate gene expression patterns in two resistant corn inbred lines (Mp313E and Mp04:86) and two susceptible corn inbred lines (Va35, B73) under artificial inoculation in field conditions (Kelley et al., 2012).

Developing kernels of corn were collected from field for RNA preparation. The field experimental plan was a randomized, complete block design with split plots and three replications. The average aflatoxin levels in mature corn kernels from the inoculated primary ears of each genotype were measured as 195 ppb for Mp04:86, 140 ppb for Mp313E, 1243 ppb for Va35, and 3791 ppb for B73. Mp04:86 was a resistant recombinant inbred line derived from the cross of Mp715 (resistant) × Va35 (susceptible), and it was selected by the phenotypic trait of low-level aflatoxin accumulation. The microarrays (from the NSF Maize Oligonucleotide Array Project) used in this experiment contained 57,452 maize gene probes (Pontius et al., 2003).

The log2 fold changes (inoculation/un-inoculation) for each gene probe were used as gene expression values, resulting in a high-dimensional (57,452 × 4) data matrix. Principal component analysis (PCA) was used to analyze the resulting gene expression values. PCA is a commonly used technique in finding patterns and relationships between variables in data analysis and the results can be displayed in biplots. The direction and length of the arrows in the biplots represented the larger variances in the gene expression values of each corn inbred line (**Figures 2A,B**). By using PCA analysis, the resistant corn inbred lines (Mp313E and Mp04:86) were separated from the susceptible

**FIGURE 2 | Biplots of PCA on the log2 ratios (***I/U***) of expressed genes in corn inbred lines.** The eigenvectors evidently separated into resistant and susceptible groups, respectively. Genes coordinately expressed in resistant lines located near the eigenvectors for Mp313E and Mp04:86. Genes

coordinately expressed in susceptible lines located near the eigenvectors for Va35 and B73. **(A)** PCA of 13,107 expressed genes and **(B)** PCA of a subset of 500 expressed genes. The arrows represent eigenvectors. The direction and length represent the larger variance of the fold changes (Kelley et al., 2012).

corn inbred lines (Va35 and B73). It had important implications. The patterns of gene expression appeared to be similar in Mp313E and Mp04:86 in contrast to those of Va35 and B73. The fact that Mp04:86 was clustered with Mp313E (resistant, not a parent) but not with the susceptible parent Va35 strongly indicated that Mp313E and Mp04:86 might have shared similar resistance mechanisms. It is worth mentioning that the PCA analysis was directly applied to all gene expression values without any pre-selection by statistical significance levels. These findings demonstrated that all levels of the gene effects collectively contributed to the grouping of resistant genotypes versus susceptible genotypes.

The microarray data were also analyzed statistically to determine the significance levels and the significant genes were selected for further validation by quantitative real-time PCR (qRT-PCR) experiments. Fifty genes that were statistically significant from microarray study were selected for a qRT-PCR analysis in a separate time course experiment using developing kernel samples of Mp313E and Va35. Thirty-one of the 50 genes were found to be significantly differentially expressed (*P* < 0.05) by analysis of qRT-PCR data. Among these genes, eight were mapped to seven previously identified QTL regions (**Figure 1**; Kelley et al., 2012).

Three of the mapped significant genes (**Figure 1**) that were highly expressed in Va35 were known plant stress responsive genes. AI664980 encodes a glycine-rich RNA-binding protein (GRBP2; Naqvi et al., 1998; Govrin and Levine, 2000; Singh et al., 2011). TC234808 encodes a protein that governs the large and small ribosomal subunits assembly. Both AI664980 and TC234808 genes have RNA-binding domains and are likely to be involved in the post-transcriptional regulation in plant defense systems. BG266083 encodes stress-induced small heat shock proteins (sHSPs). The development of ear rot caused by *A. flavus* infection in the susceptible corn inbred line Va35 showed hypersensitivity. AI664980 (GRBP2) protein was reported to be associated with plant–pathogen hypersensitivity interactions (Naqvi et al., 1998).

The resistant corn inbred line Mp313E appeared to have highly expressed genes encoding RNA transport regulators, molecular chaperones, and detoxification proteins. TC231674 is homologous to the human nucleoporin NUP85 which is a component of the nuclear pore complexes (NPCs). Studies have strongly suggested that components of NPCs regulate the transport of R proteins (Cheng et al., 2009; Garcia and Parker, 2009; Faris et al., 2010). TC231674 gene was found highly expressed in the resistant line Mp313E. BM498943 encodes ethylene responsive protein (ETHRP) which belongs to the universal stress responsive protein family. BM379345 encodes a metallothionein-like protein (MTLP) involved in the detoxification of heavy metal ions. It was found that increased aflatoxin production was associated with high levels of certain trace metal elements (Lillehoj et al., 1974).

#### **EXPLORATION OF COMPUTATIONAL AND STATISTICAL METHODS ON ANALYSIS OF RNA PATHWAY GENE EXPRESSION DATA**

Evidence from the microarray study showed that TC231674, which encodes a possible RNA transport regulator, was highly

expressed in the resistant corn inbred line Mp313E. A number of previous studies have demonstrated that RNA transport pathway genes play direct roles in plant defense (Piffanelli et al., 1999; Chisholm et al., 2006; Walley et al., 2007). RNA transport pathways are made up from a number of different protein complexes. For example, the cap binding complex (CBC), spliceosome, transcription-export complex (TREX), exon-junction complex (EJC), and translation initiation factors (eIFs) are involved in the transport of mRNA (Nakielny et al., 1997). Importins, exportins, Ran-GTP-related protein complex, and the survival motor neuron complex (SMN) are for the transport of rRNA, tRNA, and snRNA molecules (Kindler et al., 2005; Jambhekar and Derist, 2007; Vazquez-Pianzola and Suter, 2012). All RNAs are transported across the nuclear membrane through NPCs (Köhler and Hurt, 2007; Meier and Brkljacic, 2009; Deslandes and Rivas, 2011; Rivas, 2012). Several reports have shown that nucleoporins are directly involved in the regulation of R protein activities.

The expression patterns of corn RNA transport pathway genes and their relations were studied by a qRT-PCR experiment in the selected resistant and susceptible corn inbred lines (Asters et al., 2014). The advance of qRT-PCR technique makes it possible to precisely describe and compare the level of gene expression changes (Heid et al., 1996; Kubista et al., 2006; Rieu and Powers, 2009). In contrast to the comprehensive genome-wide microarray or RNA sequencing techniques, qRT-PCR provides the flexibility to measure gene expression levels on more samples across a wide range of experimental conditions as "phenotypic traits." This is an important and novel strategy since it presents a way to overcome the commonly seen"phenotyping bottleneck"due to the limitation of available measureable traits compared with the huge amount of genomic single nucleotide polymorphisms (SNPs) data. We aimed at the exploration on using gene expression values as the "phenotypic data" in addition to the levels of aflatoxin accumulation to conduct future genetic marker analysis.

Experiments were conducted to establish a computational pipeline for analysis of qRT-PCR gene expression data (Asters et al., 2014). This research aimed at the qRT-PCR gene expression analysis of more corn inbred lines, focused pathways, with an experimental design in parallel to a typical breeding project, so that statistical analysis over potential DNA markers, candidate gene expression levels, and aflatoxin levels could directly apply. Resistant (Mp718, Mp719, and Mp04:104) and susceptible (Va35, Mp04:85, and Mp04:89) corn inbred lines were used in this experiment. Experimental conditions consisted of two treatments (inoculated and un-inoculated with *A. flavus*), six corn inbred lines with three replications for each, and two sample collection time points at 2 and 7 days after inoculation (DAI). Since these inbred lines were offspring from a single cross of Mp715 and Va35, only up to two possible alleles for each gene were involved in all samples. The gene expression variations observed in these samples were likely related to gene regulating patterns associated with the resistance or susceptibility.

Forty RNA transport pathway genes and 16 candidate genes identified from the previous microarray experiment were investigated. Trends of gene expression patterns were observed among the test corn inbred lines (**Figure 3**). AI664980 showed significant variations among samples with down-regulation patterns in resistant lines. TC231674 was found highly expressed in the resistant corn line Mp718. Of the 56 genes analyzed, 17 were significantly differentially expressed between the resistant and the susceptible groups. Most of the significant genes were from the NPC and SMN protein complexes. Resistance candidate genes BE050050, TC231674, and BM498943 were found positively correlated. Whereas the susceptibilityrelated gene AI664980 was found negatively correlated with BE050050 (−0.97; Asters et al., 2014). The inclusion of previously identified candidate genes in this research provided another way of validation. The observation was consistent with the previous findings on TC231674 where it was found highly expressed in a different resistant corn inbred line Mp313E.

The genes from the RNA transport pathways were considered as being selected from a static gene network. However, in contrast to the methods excessively used in thefunctional clustering by matching with Gene Ontology (GO) terms, the analysis in this study aimed to explore a method on determination of gene relationships from the empirical gene expression data. Susceptibility-related gene AI664980 was found clustered with six test genes (Nup88, eIF2, CD443591, CA399536, SPN1, and MAGOH) by a network analysis method, suggesting a similar expression pattern of these genes in response to*A. flavus* infection. Resistance candidate genes, such as TC231674 and BE050050, appeared as isolates in the network. Further experiments with more genes included are needed to reveal positions of these genes in the empirical network regarding corn defense against *A. flavus* infection and toxin production (Asters et al., 2014).

#### **CONCLUSION**

Many studies have shown that significantly low levels of aflatoxin accumulation have been achieved in a number of resistant corn inbred lines. However, an efficient transfer of such resistance to

gene across all samples. The horizontal axis represents the 72 samples

vertical axis represents the relative gene expression values (Asters et al., 2014).

commercially available corn lines has been proven difficult due to the complex nature of this quantitative trait. Many genes were proposed to be involved in the corn host plant resistance to the aflatoxin reduction. The exploration for effective methods to identify and prioritize resistance genes will expedite the discovery of DNA markers and facilitate the breeding for corn host resistance.

For future work, we propose the use of multi-environmental gene expression data as "phenotypic data" for calculation of the DNA marker genetic effects in performing genomic selection for the breeding of corn host resistance. Parallel quantitative proteomics and metabolomics approaches will also be explored to identify proteins and metabolites related to *A. flavus* responsive pathways. Using transcriptomics, proteomics, and metabolomics data as "phenotypic data" will facilitate the identification of connections among genome regions, resistance-related genes and proteins, aflatoxin contents, and other metabolites for statistical analysis and modeling. The application of the resulting statistical models will be very helpful on estimating marker genetic effects and predicting the breeding values for aflatoxin resistance in corn.

#### **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: 21 May 2014; accepted: 29 June 2014; published online: 21 July 2014. Citation: Shan X andWilliamsWP (2014) Toward elucidation of genetic and functional genetic mechanisms in corn host resistance to Aspergillus flavus infection and aflatoxin contamination. Front. Microbiol. 5:364. doi: 10.3389/fmicb.2014.00364*

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

*Copyright © 2014 Shan and Williams. 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.*

### A mini review on aflatoxin exposure in Malaysia: past, present, and future

#### *Sabran Mohd-Redzwan1, Rosita Jamaluddin1\*, Mohd Sokhini Abd.-Mutalib1 and Zuraini Ahmad2*

<sup>1</sup> Department of Nutrition and Dietetics, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia

<sup>2</sup> Department of Biomedical Sciences, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, Serdang, Malaysia

#### *Edited by:*

Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran

#### *Reviewed by:*

Eva-Guadalupe Lizárraga-Paulín, Universidad Nacional Autónoma de México, Mexico Perng-Kuang Chang, Southern Regional Research Center, USA

#### *\*Correspondence:*

Rosita Jamaluddin, Department of Nutrition and Dietetics, Faculty of Medicine and Health Sciences, Universiti Putra Malaysia, 43400 Serdang, Selangor, Malaysia e-mail: rosita@medic.upm.edu.my

This mini review article described the exposure of aflatoxin in Malaysia, including its presence in the foodstuffs and the detection of aflatoxin biomarkers in human biological samples. Historically, the exposure of aflatoxin in Malaysia can be dated in 1960s where an outbreak of disease in pig farms caused severe liver damage to the animals. Later, an aflatoxicosis case in Perak in 1988 was reported and caused death to 13 children, as up to 3 mg of aflatoxin was present in a single serving of contaminated noodles. Since then, extensive research on aflatoxin has been conducted in Malaysia. The food commodities such as peanuts, cereals, spices, and their products are the main commodities commonly found to be contaminated with aflatoxin. Surprisingly, some of the contaminated foods had levels greater than the permissible limit adopted by the Malaysian Food Regulation 1985. Besides, exposure assessment through the measurement of aflatoxin biomarkers in human biological samples is still in its infancy stage. Nevertheless, some studies had reported the presence of these biomarkers. In fact, it is postulated that Malaysians are moderately exposed to aflatoxin compared to those high risk populations, where aflatoxin contamination in the diets is prevalent. Since the ingestion of aflatoxin could be the integral to the development of liver cancer, the incidence of cancer attributable by dietary aflatoxin exposure in Malaysia has also been reported and published in the literatures. Regardless of these findings, the more important task is to monitor and control humans from being exposed to aflatoxin. The enforcement of law is insufficient to minimize human exposure to aflatoxin. Preventive strategies include agricultural, dietary, and clinical measures should be implemented.With the current research on aflatoxin in Malaysia, a global networking for research collaboration is needed to expand the knowledge and disseminate the information to the global scientific community.

**Keywords: aflatoxin, aflatoxin B1, aflatoxin exposure, Malaysia, Southeast Asia**

"fmicb-04-00334" — 2013/11/11 — 21:56 — page 1 — #1

#### **INTRODUCTION**

There are many toxic compounds produced by fungi and one of them is mycotoxin. The occurrence of mycotoxin in the food commodities can be dated as early as in the tenth century. The infection of fungi in the diets caused the outbreak of disease known as St. Anthony's or Holy Fire in many European countries (Paterson and Lima, 2010). Paterson and Lima (2010) reported the outbreak was intensified with the contamination of rye by ergot alkaloid, produced by *Claviceps purpurea*. Since then, many cases have been reported and the discovery of "Turkey X" disease caused by aflatoxin, in 1950s and early 1960s had opened new prospectus on the scientific research on the etiology of mycotoxicosis and preventive strategies in foods, animals, and humans (Kensler et al., 2011). Kensler et al. (2011) described the epidemic disease as the major cause of death of numerous poultry animals including ducklings and chicks due to the consumption of diet containing contaminated peanuts. Further investigations at that time revealed the toxicity was associated with *Aspergillus flavus*, a pathogenic fungus and extracts from the culture of the fungus isolated from the meal were found to have the capability to induce the "Turkey X" syndromes (Kensler et al., 2011). Due to this, most of the reported cases on mycotoxicosis focused on specific species of fungi that are found in many contaminated foods and feeds. Indeed, four major species of fungi have been discovered belonging to the species of *Aspergillus*, *Fusarium*, *Penicillium*, and *Claviceps* that produced some major mycotoxins such as aflatoxin, ochratoxin A, fumonisim, and zearalenone (Paterson and Lima, 2010). Of these four mycotoxins, research on aflatoxin has been extensively conducted as aflatoxin represents the global main threat due to its toxicity (Paterson and Lima, 2010; Kensler et al., 2011)

Aflatoxin is produced by *A. flavus*, *A. paraciticus*, and *A. nomius* and ubiquitously found in foodstuffs (Reddy et al., 2011). Since aflatoxins are visible under ultraviolet (UV) light, the first attempt to detect aflatoxin is mainly based on this criteria (Groopman et al., 2005). Later the isolation of purified aflatoxin metabolites with identical physical and chemical properties formed the core to the scientific research on aflatoxin (Kensler et al., 2011). These findings have stimulated numerous research efforts, which transcend to the present, to assess possible hazards of aflatoxin contamination in the human food sources and finally to reduce the exposure through various preventive strategies. In one of many instances, the development of analytical methods to detect and quantify aflatoxin in foods and feeds is significance (Kensler et al., 2011) as more studies are able to be conducted to determine the association of aflatoxin ingestion with diseases in human population, especially with the incidence of hepatocellular carcinoma (HCC). Moreover, as the science is moving forward in parallel to the advancement of technology, the development of method for structural characterization, and synthesis of the major aflatoxins has led to better understanding of the mechanistic studies of their toxicology and metabolism (Groopman et al., 2005; Paterson and Lima, 2010; Kensler et al., 2011). For instance, the isolation of aflatoxin biomarkers in human biological samples such as serum AFB1-DNA adduct, AFB1-lysine adduct, and other metabolites of AFB1 in urine and feces (Wang et al., 1999; Mykkänen et al., 2005; Polychronaki et al., 2008) can give a better indication on the extent of human exposure to aflatoxin. These data are essential for the risk assessment as more concrete evidence can be used to evaluate human exposure to aflatoxin in a population.

#### **PAST-HISTORY OF AFLATOXIN IN MALAYSIA**

Malaysia is one of the Southeast Asia countries and geographically located in the equatorial area. Due to that, Malaysia experiences tropical climate with high temperature around 28–31◦C (Abdullah et al., 1998). On average, the relative humidity ranged from 70 to 80% during wet season and 50–60% during the dry season (Sulaiman et al., 2007). Furthermore, commodities stored under these conditions are easily deteriorated and susceptible to fungal infections (Abdullah et al., 1998; Sulaiman et al., 2007). With these conditions, the most prominent contamination is the infection of foods by mycotoxin-producing fungi. Aflatoxin, one of the mycotoxin is produced by *Aspergillus* species of fungi and this toxin has been classified by the International Agency for Research on Cancer (IARC) as Group 1 carcinogen and is linked to the development of liver cancer (International Agency for Research on Cancer [IARC], 1993).

In Malaysia, the first indicator of problem associated with aflatoxin occurred around 1960s due to the outbreak of disease in two pigfarms inMalacca (Hamid,1997). Lim andYeap (1966)reported the presence of aflatoxin in the feed ingredients, supported the evidences of severe liver damage to the animals. However, an episode in Perak, Malaysia around 1988 reminded the hazard of aflatoxin to human health, as the public was shocked by the death of 13 children as a result of consuming contaminated noodles called *loh shi fun*. The post-mortem was performed and the cause of death was associated with acute hepatic encephalopathy (Lye et al., 1995). The culprit was aflatoxin as Lye et al. (1995) reported up to 3 mg of aflatoxin was detected in a single serving of *loh shi fun*. Furthermore, the wheat flour used to make the noodles was reported to contain aflatoxin due to the poor storage and processing, which in turn promoted the growth of pathogenic *Aspergillus* fungi and subsequently the production of aflatoxin. Since then, extensive research on aflatoxin has been carried out and the first monitoring was performed by the Institute of Medical Research (IMR), followed by other agencies such as The Food Technology Center (FTC), MARDI, and local universities (Hamid, 1997). Up till know, various studies in Malaysia have reported the presence of aflatoxin in the foodstuffs. Moreover, the biomarkers of aflatoxin have also been reported recently in serum and urine samples,

#### **PRESENT – OCCURRENCE OF AFLATOXIN IN THE FOOD COMMODITIES, HUMAN EXPOSURE AND RISK ASSESSMENT**

The contamination of foodstuffs by aflatoxin has been documented in the literatures. Peanuts, cereals, spices, and their products are the commodities most susceptible to aflatoxin contamination. These food commodities are generally used in many Malaysian delicacies either as a main ingredient or as a base material (Reddy et al., 2011). In fact, they are available almost in all retail shops in Malaysia and are inexpensive. For example, various studies have detected aflatoxin in the peanuts and peanut-based products as shown **Table 1**. Sulaiman et al. (2007) indicated that raw shelled peanut and its products are the good substrate for the growth of aflatoxin by *A. flavus* and *A. paraciticus.* Even though peanuts are not extensively produced in Malaysia, the occurrence of aflatoxin in this food commodity is frequently reported. Moreover, Leong et al. (2011) postulated that high level of AFB1 in nuts can result in the high potential risk of AFB1 exposure in Malaysia. In fact, Malaysians probably consume higher amount of AFB1 than Europeans and Americans but lower than those in Africa and China (Leong et al., 2011).

Cereals such as rice are one of the main agricultural products in Malaysia. The presence of aflatoxin in the rice could pose serious health issues. For example, in a study by Soleimany et al. (2012a), 33.3% of analyzed rice from the retail market had detectable level of aflatoxin ranging from 0.19 to 3.96 ng/g. Even though the levels are not considered dangerous and harmful to humans, it was an unexpected finding as rice is the main staple food consumed by majority of Malaysian. Statistically, Malaysians consume rice about 289.68 g/day (Ministry of Health [MOH], 2006). Thus it is assumed that dietary aflatoxin exposure from rice is somewhere between 55.04 ng/day and 1.15 μg/day [288.68 g/day × 0.19 or 3.96 ng/g] for an adult of 60 kg. Given that the aflatoxin-contaminated rice is consumed on a regular basis, this small amount of aflatoxin at the end will be accumulated in the body and can be detrimental to the health as previously reported.

As highlighted in **Table 1**, varieties of food samples were found to be contaminated with aflatoxin. The detection rate ranged from 16% up to 92% showing the ubiquitous presence of aflatoxin in the human food resources. Although the findings do not represent the whole scenario of aflatoxin contamination in Malaysia, such information shows the pervasiveness of human exposure to this food contaminant. In an instance, Reddy et al. (2011) detected aflatoxin in the foods normally consumed by Malaysian, where they are used and consumed on a daily basis. The presence of aflatoxin in the food chain is a serious matter but not knowing its impact to the health is a big problem and should be a public concern. Indeed in Malaysia, a recent survey reported low awareness and knowledge among the public on the problems associated with fungal and aflatoxin contamination in the diets (Mohd Redzwan et al., 2012a). Furthermore, the similar observations were also reported by several studies in African countries (Jolly et al., 2006, 2009; Ilesanmi and Ilesanmi, 2011).

"fmicb-04-00334" — 2013/11/11 — 21:56 — page 2 — #2


"fmicb-04-00334" — 2013/11/11 — 21:56 — page 3 — #3

**1| Occurrence of aflatoxin in the foodstuffs in Malaysia.**

**Table** 


 Aflatoxins; aTotal aflatoxin; bAFB1.

**Table 1 |**

**Continued**

"fmicb-04-00334" — 2013/11/11 — 21:56 — page 4 — #4

Exposure assessment by measuring aflatoxin levels in the food samples and extrapolating to estimate average intake at the population level is of low reliability (Kensler et al., 2011). The reason is such data cannot provide information on the individual's exposure as different individuals may have different exposure to aflatoxin. Nevertheless, the expansion of metabolomic study such as the use of aflatoxin biomarkers could provide the tools for molecular epidemiology to evaluate aflatoxin exposure among individuals within human population. Besides, measuring aflatoxin biomarkers could improve estimation of aflatoxin exposure (Liu and Wu, 2010). The aflatoxin biomarkes such as AFB1-DNA adduct, AFB1-lysine adduct, and urinary AFM1 have been documented in the literatures especially in African countries (Gong et al., 2002; Jolly et al., 2006; Shuaib et al., 2010) and China (Mykkänen et al., 2005; Xu et al., 2010), where aflatoxin contaminations are ubiquitous. These biomarkers are useful for epidemiologist and public health workers to make a better postulation on the extent and severity of aflatoxin exposure in a population. For example, the detection of AFB1-DNA adduct such as AFB1-N7-guanine in the urine samples could be an indicator of mutagenic effect of aflatoxin.

In Malaysia, exposure assessment by measuring aflatoxin biomarkers in human biological samples is still in its infancy. Nevertheless, several studies have reported the presence of these biomarkers among the Malaysia population. Indeed, it is believed that study by Zulhabri et al. (2009) is the first one in Malaysia that reported AFB1-albumin adduct in the HCC patients. Zulhabri et al. (2009) indicated that the HCC patients had significantly higher level of AFB1-albumin adduct compared to the healthy subjects. Aflatoxin can be an integral to the development of liver cancer (International Agency for Research on Cancer [IARC], 1993), thus the finding could provide an interesting insight on the synergistic effect of aflatoxin exposure. On the other hand, a recent study by Leong et al. (2012) found AFB1-lysine adduct in 97% of 170 subjects in Penang, with the levels ranging from 0.2 to 23.16 pg/mg albumin. The study also reported that Chinese and Indian respectively were 3.05 and 2.35 times more likely to have high AFB1-lysine adduct compared to the Malay (Leong et al., 2012). As a matter of fact, it was corroborated

with a report from the Ministry of Health, Malaysia that showed high incidence of liver cancer among Chinese compared to other main ethnicities (Zainal Ariffin and Nor Saleha, 2011). Statistically in 2007, the report listed liver cancer as number 5 out of 10 most frequent cancers among male Chinese with the incidence rate of 5.9% (Zainal Ariffin and Nor Saleha, 2011). Besides, the other aflatoxin biomarkers have also been investigated. A preliminary study involving 22 adult subjects found the presence urinary AFM1 ranging from 0.029 to 0.15 ng/ml (Sabran et al., 2012). Moreover, a significant and positive association was detected between urinary AFM1 with the consumption of milk and dairy products among 98 subjects with detectable level of AFM1 in Malaysia (Mohd Redzwan et al., 2012b). Since AFM1 is a metabolite of AFB1, such findings indicated by Mohd Redzwan et al. (2012b) and Sabran et al. (2012) are useful to estimate human exposure to aflatoxin as a study has found good correlation between dietary intake of AFB1 and urinary excretion of metabolite AFM1 (Zhu et al., 1987).

Given that ingestion of aflatoxin can be linked to the development of liver cancer, a risk assessment can provide additional information on the extent of human exposure to this food contaminant. Risk assessment is the process of estimating the magnitude and the probability of a harmful effect to individuals or populations for certain agents or activities (Liu and Wu, 2010). Several studies have reported Malaysian population could be exposed as high as 140 ng/kg body weight/day (**Table 2**). In one of the examples given in **Table 2**, Chin et al. (2012) indicated that Malaysian exposed to AFB1 ranging from 24.3 to 34 ng/kg body weight/day and such exposure could contribute to 12.4–17.3% of liver cancer in Malaysia. It was noteworthy as Liu and Wu (2010) described that more than 55 billion people worldwide facing uncontrolled exposure of aflatoxin and the burden of aflatoxin-induced liver cancer still remained unclear. Although it seems impossible to remove and eliminate aflatoxin completely from the food chain, these data are useful to develop plans to control and monitor aflatoxin in the foodstuffs. In fact, interventions could be done to prevent human exposure to aflatoxin and reduce the incidence of aflatoxin-related diseases in Malaysia.


"fmicb-04-00334" — 2013/11/11 — 21:56 — page 5 — #5

**Table 2 | Risk assessment of aflatoxin exposure in Malaysia.**

aDietary aflatoxin exposure.

bThe value was calculated based on formula given by Chin et al. (2012) – (Dietary exposure <sup>×</sup> 0.025 cancers/100, 000 in Malaysia). <sup>c</sup>The value was computed from formula by Chin et al. (2012) – (Estimated liver cancer risk per 10,000 population/incidence rate of liver cancer in

Malaysia of 4.9 per 100,000 <sup>×</sup> 100%). <sup>d</sup>Data obtained from estimated annual HCC per 100,000 population for HBsAG-positive.

#### **FUTURE – MONITORING AND CONTROLLING THE OCCURRENCE OF AFLATOXIN AND INTERVENTION STRATEGIES**

It is a very important task to monitor and control the presence of aflatoxin in the human food resources as it involves many aspects. The legislation of law for instance by setting a certain limit can be considered as the "first defense" to prevent aflatoxin exposure to human. This rule was established following the massive outbreak of diseases associated with aflatoxin back in 1960s as mentioned earlier. Prior to that in Malaysia, aflatoxin was not considered to be a significant problem as it was not covered by the Sale of Food and Drug Ordinance 1952. In fact, the ruling on aflatoxinassociated cases was referred from countries such as in UK and USA (Hamid, 1997). However, around 1980s, Malaysia has promulgated stricter standards to prevent the "flow" of aflatoxin in the food chain by the implementation of Malaysia Food Regulation 1985, through the Food Act 1983. Initially, a maximum limit of 35 μg/kg for all mycological contaminants was implemented as reported by studies back in 1990s (Hamid, 1997; Abdullah et al., 1998; Ali et al., 1999). Later, it was revised and the new legislation limit was set to 5 μg/kg for all mycological contaminants including aflatoxin. Moreover, a limit of 15 and 0.5 μg/kg was set for groundnut for further processing and milk products respectively.

With the enforcement of law, it will be meaningless if no proper actions are taken by the public to remove the aflatoxincontaminated foodstuffs from the line of production. Although Malaysia has joined Codex Alimentarius since 1971 and strengthened the law on the occurrence of aflatoxin in the foodstuffs, it is still believed that some of the contaminated foods might "escape" during the surveillance process and persist in the food chain. A few studies explained that some contaminated foods may be perceived as safe and edible if there are no sign of defects and contaminations (Jolly et al., 2006; Mohd Redzwan et al., 2012b). Since the main route of human exposure to aflatoxin is through the diets, humans can be directly and/or indirectly exposed. Predicated upon that, interventions strategies have been carried out globally. Liu andWu (2010) classified the strategies into three categories namely agricultural, dietary, and clinical. The agricultural interventions can be applied during pre- or post-harvest. This strategy is considered as the main primary intervention whereas dietary and clinical interventions are regarded as the secondary intervention (Liu and Wu, 2010).

Since then, various research around the globe have come up with measures to limit human exposure to this food contaminant. The use of adsorbents such as activated carbon, hydrated sodium calcium alumino silicate, zeolite, bentonite, and certain clays are beneficial as they are proven to prevent aflatoxin absorption (Denli and Okan, 2006; Thieu and Pettersson, 2008; Gallo et al., 2010). Recently, the use of probiotic has been studied as one the potential adsorbents of aflatoxin in the gastrointetstinal tract. The *in vitro* and animal studies found the potential of certain probiotic lactic acid bacteria in reducing the bioavailability of aflatoxin (El-Nezami et al., 1998; Haskard et al., 2000; Lahtinen et al., 2004; Gratz et al., 2006; Hernandez-Mendoza et al., 2009; Nikbakht Nasrabadi et al., 2013). In fact, probiotics have many beneficial health effects (Oelschlaeger, 2010), thus its use to counteract the toxicity of alfatoxin could be further examined as one the preventive strategies.

#### **CONCLUSION**

Malaysia is a country that is shifting to become a fully developed nation by 2020 and agricultural sector is one of the driving forces for the economy. The occurrence of aflatoxin in the agricultural commodities can be a big loss to the economy and consequently affect the nation. Aflatoxin is dangerous and potent toxin as a number of expert groups have reviewed the health effect of aflatoxin (International Agency for Research on Cancer [IARC], 1993; Joint FAO/WHO Expert Committee on food additives [JECFA], 1998). In Malaysia, reports and publications on the level of aflatoxin in the foodstuffs are abundant. Nevertheless, from the scarce literatures of human exposure to aflatoxin, it is clear that more works need to be done on this very important matter. This is because the information is useful to estimate the magnitude of aflatoxin exposure in Malaysia. On the global scale, a worldwide networking with other nations is essential to gather data and information for harmonization in order to prevent this problem continuing to happen in the future.

#### **ACKNOWLEDGMENT**

Sabran Mohd-Redzwan would like to thank Ministry of Higher Education,Malaysia for providing scholarship for his postgraduate studies.

#### **REFERENCES**

"fmicb-04-00334" — 2013/11/11 — 21:56 — page 6 — #6


Mehan and C. L. L. Gowda (Hyderabad: International Crops Research Institute for the Semi-Arid Tropics), 32–35.


"fmicb-04-00334" — 2013/11/11 — 21:56 — page 7 — #7

Zulhabri, O., Jamal, R., Isa., M. R., Harjit, S., and Wang Ngah, W. Z. (2009). Hotspot Ser-249 TP53 mutations, aflatoxin B1-adducts and hepatitis B infection in human hepatocellular carcinoma from Malaysia. *Med. J. Mal.* 64 (Suppl B), 88.

**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 August 2013; paper pending published: 09 September 2013; accepted: 22 October 2013; published online: 13 November 2013.*

*Citation: Mohd-Redzwan S., Jamaluddin R., Abd.-Mutalib MS, Ahmad Z (2013) A mini review on aflatoxin exposure in Malaysia: past, present, and future. Front. Microbiol. 4:334. doi: 10.3389/fmicb.2013.00334*

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

*Copyright © 2013 Mohd-Redzwan, Jamaluddin, Abd.-Mutalib and Ahmad. 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-04-00334" — 2013/11/11 — 21:56 — page 8 — #8

**REVIEW ARTICLE** published: 18 March 2014 doi: 10.3389/fmicb.2014.00102

### The microRNAs as potential biomarkers for predicting the onset of aflatoxin exposure in human beings: a review

*Rafael Valencia-Quintana1 \*, Juana Sánchez-Alarcón1, María G. Tenorio-Arvide2, Youjun Deng3, José M. R. Montiel-González 1, Sandra Gómez-Arroyo4, Rafael Villalobos-Pietrini 4, Josefina Cortés-Eslava4, Ana R. Flores-Márquez <sup>4</sup> and Francisco Arenas-Huertero5 \**

*<sup>2</sup> Departamento de Investigación en Ciencias Agrícolas, Benemérita Universidad Autónoma de Puebla, Puebla, México*

*<sup>3</sup> Department of Soil and Crop Sciences, Texas AgriLife, Texas A&M University, College Station, TX, USA*

*<sup>4</sup> Departamento de Ciencias Ambientales, Centro de Ciencias de la Atmósfera, Universidad Nacional Autónoma de México, Distrito Federal, México*

*<sup>5</sup> Laboratorio de Patología Experimental, Hospital Infantil de México Federico Gómez, Distrito Federal, México*

#### *Edited by:*

*Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran*

#### *Reviewed by:*

*Eva-Guadalupe Lizárraga-Paulín, Universidad Nacional Autónoma de México, Mexico Jing Shen, Columbia University, USA*

#### *\*Correspondence:*

*Rafael Valencia-Quintana, Evaluación de Riesgos Ambientales, Facultad de Agrobiología, Universidad Autónoma de Tlaxcala, Av. Universidad No.1, Col. La Loma Xicohténcatl, 90070 Tlaxcala de Xicohténcatl, Tlaxcala, México e-mail: prvq2004@yahoo.com.mx; Francisco Arenas-Huertero, Laboratorio de Patología Experimental, Hospital Infantil de México Federico Gómez, Dr. Márquez 162, Col. Doctores, 06720 Ciudad de México, Distrito Federal, México e-mail: farenashuertero@ yahoo.com.mx*

The identification of aflatoxins as human carcinogens has stimulated extensive research efforts, which continue to the present, to assess potential health hazards resulting from contamination of the human food supply and to minimize exposure. The use of biomarkers that are mechanistically supported by toxicological studies will be important tools for identifying stages in the progression of development of the health effects of environmental agents. miRNAs are small non-coding mRNAs that regulate post-transcriptional gene expression. Also, they are molecular markers of cellular responses to various chemical agents. Growing evidence has demonstrated that environmental chemicals can induce changes in miRNA expression. miRNAs are good biomarkers because they are well defined, chemically uniform, restricted to a manageable number of species, and stable in cells and in the circulation. miRNAs have been used as serological markers of HCC and other tumors. The expression patterns of different miRNAs can distinguish among HCC-hepatitis viruses related, HCC cirrhosis-derivate, and HCC unrelated to either of them. The main objective of this review is to find unreported miRNAs in HCC related to other causes, so that they can be used as specific molecular biomarkers in populations exposed to aflatoxins and as early markers of exposure, damage/presence of HCC. Until today specific miRNAs as markers for aflatoxins-exposure and their reliability are currently lacking. Based on their elucidated mechanisms of action, potential miRNAs that could serve as possible markers of HCC by exposure to aflatoxins are miR-27a, miR-27b, miR-122, miR-148, miR-155, miR-192, miR-214, miR-221, miR-429, and miR-500. Future validation for all of these miRNAs will be needed to assess their prognostic significance and confirm their relationship with the induction of HCC due to aflatoxin exposure.

**Keywords: AFB1, aflatoxin exposure, microRNAs, HCC, potential biomarkers**

#### **INTRODUCTION**

The aflatoxins were structurally identified in the early 1960s and over the last 50 years have been extensively studied with respect to their mechanisms of action, including their mutagenic and carcinogenic activity. This work has been paralleled by developments in biomarkers of aflatoxin metabolism, DNA adducts, and mutations applied to exposed human populations. The improvements in exposure assessment in epidemiological studies and the demonstration of a specific mutation in the *TP53* gene have contributed significantly to the identification of aflatoxins as human carcinogens. In addition, the studies of animal and human aflatoxin metabolism have provided opportunities to develop chemoprevention approaches in human populations (Wild and Turner, 2002; Valencia-Quintana et al., 2012). These findings stimulated extensive research efforts, which continue to the present, to assess potential health hazards resulting from contamination of the human food supply and to minimize exposure (Kensler et al., 2011).

The use of biomarkers that are mechanistically supported by toxicological studies will be important tools for identifying stages in the progression of development of the health effects of environmental agents. Since the development of a general paradigm for molecular epidemiology and biomarkers nearly 20 years ago, progress has been made in applying these tools to specific environmental situations that may be hazardous to humans, as exemplified by AFB1 studies. The major goals of molecular epidemiology research are to develop and to validate biomarkers that reflect specific exposures, their interactions, and predictions of disease risk in individuals. Presumably, after an environmental exposure each person has a unique response to both dose

*<sup>1</sup> Evaluación de Riesgos Ambientales, Facultad de Agrobiología, Universidad Autónoma de Tlaxcala, Tlaxcala, México*

and time to disease onset. These responses will be affected both by genetic, host and environmental modifiers. It is assumed that biomarkers that reflect the mechanisms of action of the etiologic agents will be strong predictors of an individual's risk of disease. It is also expected that these biomarkers can more clearly classify the status of exposure of individuals and general populations (Groopman et al., 2005).

Biomarkers can be used as outcome measures in these and primary prevention studies. Overall, the integrated, multidisciplinary research on aflatoxins has provided the scientific platform on which to base decisions regarding acceptable exposures and priorities for interventions to reduce human risk in a public health context (Wild and Turner, 2002).

#### **AFLATOXIN BIOMARKERS**

AFB1 requires metabolic activation to its ultimate carcinogenic form, a reactive epoxide (aflatoxin-8,9-epoxide), primarily by the cytochrome P450 (CYP) monooxygenase system. Epoxidation is catalyzed by CYP1A2 and CYP3A4 in humans (Gallagher et al., 1994; Ueng et al., 1995). Many other oxidation products, including aflatoxin M1, are also formed. The epoxide can react further by interacting with DNA to produce a promutagenic aflatoxin-N7-guanine adduct. This adduct is unstable in DNA, rapidly undergoes depurination, and is excreted in urine (Bennett et al., 1981). The epoxide can also form products that react with serum albumin to form long-lived lysine adducts (Sabbioni et al., 1987). In addition, the epoxide can be conjugated by certain glutathione S-transferases (GSTs), which are further metabolized to form aflatoxin-mercapturic acid detoxification products that can be excreted in urine (Scholl et al., 1997). Urinary measures of aflatoxin M1, aflatoxin-mercapturic acid, and the aflatoxin-albumin adduct are used as biomarkers of internal dose. Aflatoxin-N7 guanine in urine serves as an elegant biomarker of biologically effective dose because it is clear that formation of this adduct lies on the causal pathway to aflatoxin-induced HCC (Kensler et al., 2011).

An objective in development of AFB1 biomarkers is to use them as predictors of past and future exposure status in people (Kensler et al., 2011). However, two key attributes, one biological (tracking) and the other chemical (stability), need to be confirmed to successfully use biomarkers for these purposes. miRNAs are good biomarkers because they are well defined, chemically uniform, restricted to a manageable number of species, and stable in cells and in the circulation (Wang et al., 2012a).

#### **microRNAs AND ENVIRONMENTAL POLLUTANTS**

Exposure to environmental chemicals is well known to increase risks for various diseases (Crinnion, 2010; Newbold, 2010), and gene expression can be changed as a response to these exogenous stressors (Ueda, 2009; Patel and Butte, 2010; Hou et al., 2012), like tobacco and polycyclic aromatic hydrocarbons in urban air of megacities (Arenas-Huertero et al., 2011). Such changes may be regulated by specific miRNAs and emerged as a gene expression regulatory factor that may link environmental chemicals and their related diseases.

Secreted miRNAs have many requisite features of good biomarkers. miRNAs are stable in various bodily fluids, the sequences of most miRNAs are conserved among different species, the expression of some miRNAs is specific to tissues or biological stages, and the level of miRNAs can be easily assessed by various methods, as polymerase chain reaction (PCR), which allows for signal amplification. The changes of several miRNA levels in plasma, serum, urine, and saliva have already been associated with different diseases (for review see Etheridge et al., 2011).

Growing evidence has demonstrated that environmental chemicals can induce changes in miRNA expression (Hou et al., 2011). Arsenite exposure induced significant decrease in miR-19a expression in human lymphoblast cells line TK-6, resulting in cell growth arrest and apoptosis (Marsit et al., 2006). Metal sulfates have been shown to generate reactive oxygen species (ROS) and trigger the expression of specific miRNAs (Lukiw and Pogue, 2007). Bollati et al. (2010) found an increased expression of miR-146a related to inhalation of Cd-rich air particles in steel workers, and induced rapid changes in the expression of two inflammation-related miRNAs, miR-21 and miR-222. Aluminum exposure may induce genotoxicity via miRNA-related regulatory elements, for example, miR-146a, miR-9, miR-125b, and miR-128 (Lukiw and Pogue, 2007; Pogue et al., 2009).

Jardim et al. (2009) have shown extensive alterations of miRNA expression profiles in human bronchial epithelial cells treated with diesel exhaust particles. Schembri et al. (2009) have identified 28 miRNAs that were differentially expressed in smokers when compared to non-smokers, changes in miRNA expression were suggested to contribute to altered regulation of oncogenes, tumor suppressor genes, oxidative stress, xenobiotic metabolism, and inflammation. Izzotti et al. (2009a,b) have monitored the expression of 484 miRNAs in the lungs of mice exposed to cigarette smoking, the most remarkably downregulated miR-NAs belonged to several miRNA families, such as let-7, miR-10, miR-26, miR-30, miR-34, miR-99, miR-122, miR-123, miR-124, miR-125, miR-140, miR-145, miR-146, miR-191, miR-192, miR-219, miR-222, and miR-223. These miRNAs regulate expression of genes involved in stress responses, apoptosis, proliferation, and angiogenesis.

Zhang and Pan (2009) have evaluated the effects of Hexahydro-1, 3, 5-trinitro-1, 3, 5-triazine (also known as hexogen or cyclonite) (RDX) on miRNA expression in mouse brain and liver, most of the miRNAs that showed altered expression, including let-7, miR-17-92, miR-10b, miR-15, miR-16, miR-26, and miR-181, were related to toxicant-metabolizing enzymes, and genes related to carcinogenesis, and neurotoxicity, in addition, consistent with the known neurotoxic effects of RDX, the authors documented significant changes in miRNA expression in the brains of RDX-treated animals, such as miR-206, miR-30a, miR-30c, miR-30d, and miR-195. STS (sodium thiosulfate) treatment also resulted in differential expression of miR-124a and miR-133a in the treated embryos (Choudhuri, 2010). Fukushima et al. (2007) have demonstrated that rat exposed to acetaminophen or carbon tetrachloride showed down-regulation of miR-298 and miR-370 in the liver that was accompanied by hepatocyte necrosis and inflammation. Wang et al. (2009a) found increase serum concentration of hepatocyte-specific miRNAs including miR-122 and miR-192 within 1 h after acetaminophen exposure. In mouse exposure to Wy-14,643, peroxisome proliferator-activated receptor alpha (PPARα) agonist, up-regulate let-7C (Shah et al., 2007). Ethanol exposure down-regulate miR-21, miR-335, miR-9, and miR-153 (Sathyan et al., 2007). In rats, tamoxifen upregulate miR-17-92 cluster, miR-106a, and miR-34 (Pogribny et al., 2007). 4-(methylnitrosamino)-1-(3-pyridyl)-1-butanone (NNK), a tobacco carcinogen, down-regulate miR-34, miR-101, miR-126, and miR-199 (Kalscheuer et al., 2008). In humans, 1 alpha, 25-dihydroxyvitamin D(3) [1,25(OH)(2)D(3)], major vitamin D metabolite, up-regulate miR-125b (Mohri et al., 2009), and 5-fluorouracil (5-FU) an antineoplasic drug, up-regulate miR-200b (Rossi et al., 2007). Bisphenol A (Avissar-Whiting et al., 2010), dioxin (Elyakim et al., 2010), and diethylstilbestrol (DES) (Hsu et al., 2009) desregulate expression of miR-146a, miR-191, and miR-9-3, respectively. Moffat et al. (2007) looked at the effects of dioxin treatment on miRNA in mice, dioxin-resistant rats (Han/Wistar; Kuopio) and dioxin-sensitive rats (Long–Evans; Turku/AB), it is interesting to note that the dioxin sensitive rats had more affected miRNAs. X-ray exposure resulted in down regulation of miR-7 (Ilnytskyy et al., 2008). The ROS induction resulted in up-regulation of a specific set of miRNAs, including miR-9, miR-125b, and miR-128 (Lukiw and Pogue, 2007). Lin et al. (2009), and Cheng et al. (2009), found that hydrogen peroxide induce up-regulation of miR-21.

The increasing evidence that the expression of miRNAs is affected by several known toxicants as well as oxidative and other forms of cellular stress certainly suggest an important role of miR-NAs in toxicology, which could provide a link between environmental influences and gene expression (Lema and Cunningham, 2010). Analysis of the resulting molecular signatures provides new tools to identify mechanisms of toxicity, as well as to classify compounds based on the biological response they elicit and to identify cluster of genes detective or predictive of certain type of toxic response, which are employed as biomarkers (Gatzidou et al., 2007). miRNAs are one of the main mechanisms of epigenetic regulation of gene expression. Conversely to stress-related miR-NAs, the toxicological research of miRNAs associated to specific toxicants has started few years ago. Therefore, available publications are focused in a broad spectra of toxicants and compiled research about a particular miRNA is yet very limited (Lema and Cunningham, 2010).

#### **microRNA IN HCC AS POTENTIAL BIOMARKERS OF AFLATOXIN EXPOSURE**

In the recent years, several studies revealed that the expression of miRNAs is deregulated in human HCC in comparison with matched non-neoplastic tissue (Lu et al., 2005; Gramantieri et al., 2008; Law and Wong, 2011; Borel et al., 2012a; Wang et al., 2012a,b; Sun and Karin, 2013; Wong et al., 2013). Some miRNAs identified in HCC are reported in **Table 1**.

Cellular miRNAs can be released into the circulation, and circulating miRNA levels are also affected in HCC. Circulating plasma miRNA signatures may provide a novel diagnosis method for early, pre-symptomatic HCC patients, and may prove useful as prognosis biomarkers (Borel et al., 2012a).

In HCC has been reported up-expression of miR-21, miR-221, miR-22, miR-15, miR-517a, and down-expression of miR-122, miR-29 family, miR-26a, miR-124, let-7 family members, and miR-199a/b-3p (Szabo et al., 2012). Other miR- reported as markers involved in HCC have been miR-15b, miR-16, miR-17- 5p, miR-18, miR-18a, miR-20, miR-23b, miR-26a, miR-29, miR-34a, miR-92, miR-101, miR-106a, miR-125b, miR-130b, miR-143, miR-146a, miR-195, miR-203, miR-223, miR-224, miR-338, miR-378, miR-422b, miR-500 (Murakami et al., 2006; Budhu et al., 2008; Jiang et al., 2008; Chen, 2009; Zhang et al., 2009; Kerr et al., 2011; Qu et al., 2011; Liu et al., 2012; Singhal et al., 2012; Wong et al., 2013).

In a review, Gramantieri et al. (2008) show miRNAs aberrantly expressed in HCC compared to non-tumorous liver tissue (upexpression of miR-33, miR-130, miR-135a, miR-210, miR-213, miR-222, miR-331, miR-373, miR-376a, and down-expression of miR-130a, miR-132, miR-136, miR-139, miR-143, miR-145, miR-150, miR-200a, miR-200b, miR-214). However, specific markers and their reliability are currently lacking as occur with aflatoxinexposure.

Singhal et al. (2012), review molecular and serum markers in HCC as predictive tools for prognosis and recurrence. Aberrant expression of miR-21 can contribute to HCC by modulating PTEN expression and PTEN-dependent pathways (Meng et al., 2007). A significantly high expression of miR-224 in HCC patients promotes proliferation and inhibits apoptosis inhibitor-5 (API-5) transcript expression (Wang et al., 2008). An inverse correlation between miR-221 and both CDKN1B/p27 and CDKN1 C/p57, suggested miR-221 oncogenic function in hepatocarcinogenesis (Fornari et al., 2008). Also miR-221 has been involved in the modulation of Bmf, a proapoptotic BH3-only protein, regulating the cell proliferation and apoptosis proteins (Gramantieri et al., 2009).

HCC cells showed highly deregulated miR-223 expression and a strong inverse relationship with its downstream target Stathmin 1 (Wong et al., 2008). Murakami et al. (2006) and Li et al. (2008), found that miR-125b that suppress the cell growth and phosphorylation of Akt to be a prognostic marker of HCC. miR-122 found up to 70% of total miRNA in the liver, modulates cyclin G1, thus influences p53 protein stability and transcriptional activity and reduces invasion capability of HCC-derived cell lines (Fornari et al., 2009). Bcl-w is a direct target of miR-122 that functions as an endogenous apoptosis regulator in these HCC-derived cell lines (Lin et al., 2008). miR-122 is under the transcriptional control of HNF1A, HNF3A and HNF3B and loss of miR-122 results in an increase of cell migration and invasion. From a clinical point of view, miR-122 can be used as a diagnostic and prognostic marker for HCC progression (Coulouarn et al., 2009).

In HCC cell line, miR-34a directly targeted c-Met and reduced both mRNA and protein levels of c-Met, thus blocking cell migration (Li et al., 2009a). miR-18a high expression in HCC tumors (Liu et al., 2009). miR-101 has a downstream target of v-fos oncogene and it is involved in cell invasion and migration in overexpressed HCC cell lines (Li et al., 2009b). Xu et al. (2009) show that miR-195 may block the G(1)/S transition by repressing Rb-E2F signaling through targeting multiple molecules, including cyclin D1, CDK6, and E2F3. Upregulation of miR-143 expression transcribed by NF-kappa B in HBV-HCC promotes cancer cell invasion/migration and tumor metastasis by repression of

#### **Table 1 | microRNA deregulated in HCC.**


*(Continued)*

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


*(Continued)*

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


*Potential biomarkers of aflatoxin exposure.*

fibronectin type III domain containing 3B (FNDC3B) expression (Zhang et al., 2009).

The level of miR-338 expression can be associated with clinical aggressiveness of HCC (Huang et al., 2009). miR-23b can recognize target sites in the 3-UTR of uPA and of c-met mRNAs and translationally repress the expression of uPA and c-met decreasing migration and proliferation abilities in HCC cells (Salvi et al., 2009).

MiR-126 down-regulation has been suggested to be directly linked to alcohol-induced hepatocarcinogenesis (Morgan et al., 2004). Microarray profiling studies showed reduction in miRNAs expression specific of HCV and HBV-associated cases: downregulation of miR-190, miR-134, and miR-151 occurs in HCV cases, and of miR-23a, miR-142-5p, miR-34c, in HBV cases (Ura et al., 2009). MiR-96 was reported to be distinctively upregulated in HBV-associated HCC (Ladeiro et al., 2008), whereas miR-193b upregulation has been found upon transfection of HCV genome (Braconi et al., 2010). As quoted above, up-regulation of miR-NAs, including miR-17-92 cluster, miR-106a, and miR-34, occurs during tamoxifen-induced hepatocarcinogenesis in female rats (Pogribny et al., 2007), also long-term-administration of 2-AAF resulted in disruption of regulatory miR-34a-p53 feed-back loop (Pogribny et al., 2009). In mice administered a choline-deficient and amino acid-defined diet, in which steatohepatitis precedes HCC development, microarray analysis identified that miR-155 was consistently up-regulated (Wang et al., 2009b).

Ross et al. (2010) analyzed the miRNA expression levels in control and conazole-treated mice. Conazol exposure induced many more changes in miRNA expression. All but one of the altered miRNAs were downregulated compared to controls. The authors suggest that this pattern of the altered miRNA expression represents a signature for tumorigenic conazole exposure in mouse liver.

This newly emerging area of research should unravel novel biomarkers of diagnostic as well as prognostic value in HCC.

#### **microRNA AND AFLATOXIN B1**

Exposure to environmental carcinogens may affect miRNAs expression in liver cells. While this concept is largely acceptable in principle, the specific miRNAs that are deregulated by various toxic and/or carcinogenic agents are yet to be fully documented. What we know at best today is the end-point of the process: the miRNAs whose expression is altered in HCCs (**Table 1**). However, results from some reports suggest that changes in expression of miRNAs may occur early in the process (Jiang et al., 2008), and these changes may be related to specific etiological factors, such as AFB1. These still preliminary evidences suggest the possibility of using miRNAs as early markers for aflatoxins exposure.

Exposure to natural or chemical environmental agents contributes to HCC development (Wild, 2009). In this context, uncovering relationships between exposure to environmental carcinogens and expression of miRNAs may reveal practical and sensitive biomarkers of toxic exposures and/or carcinogenicity testing (Wang et al., 2009a). A few reports addressed this hypothesis and revealed the existence of differential miRNAs expression patterns in HCCs in accordance with specific risk factors suggesting that exposure to specific risk factors could be responsible for the appearance of characteristic pathogenetic miRNA signatures (Elamin et al., 2011).

Although the precise roles of miRNA in the response to xenobiotics, drugs and chemical toxicants, remain to be established, there is little doubt that miRNAs are important in the cellular and *in vivo* responses to xenobiotics (Taylor and Gant, 2008). At this time, no specific studies on the effect of AFB1 on miRNA expression have been reported.

The field of miRNA and toxicology, particularly as it pertains to AFB1 toxicological outcome, is still in its beginnings. Nonetheless, there seems to be an increasing interest among toxicologists trying to understand the contribution of miRNA in regulating various toxicological outcomes through regulation of gene expression. A number of questions need to be addressed, such as a global role of miRNAs in cellular toxicity and disease; how miRNA biogenesis and expression affect susceptibility/resistance to xenobiotic-induced toxicity or disease (Taylor and Gant, 2008); whether cellular miRNAs form a regulatory networks and how perturbations of such network can cause toxicity/disease including developmental toxicity; how miRNAs may regulate transgenerational toxicological response through epigenetic regulation of gene and genome expression; as well as whether homologous miRNAs can be identified in an animal species based on known miRNA species and their action in other species or even in plant kingdom (Choudhuri, 2010).

Until today specific microRNAs as markers for aflatoxinexposure and their reliability are currently lacking. The following are some potential candidates based on their elucidated mechanisms of action.

The high expression of miR-122 in the liver appears to correlate with a central role in various functions of normal and diseased livers (Lewis and Jopling, 2010; Negrini et al., 2011). It provides a very attractive target for aflatoxins. Rather surprisingly, given the high intracellular levels and numerous targets of miR-122, inactivation of the miRNA does not have any apparent adverse effects on liver physiology. However, reduced miR-122 expression does show an association with hepatocellular carcinoma, and further work will be necessary. In HCC, miR-122 is downregulated in approximately 70% of cases, suggesting a tumor suppressor function for this miRNA (Bai et al., 2009; Fornari et al., 2009; Ma et al., 2010; Callegari et al., 2013). In addition, loss of miR-122 expression in patients with liver cancer is correlated with the presence of metastasis and a shorter time to recurrence (Coulouarn et al., 2009; Fornari et al., 2009; Tsai et al., 2009). The role of miR-122 in liver cancer has been demonstrated directly by the generation of miR-122 knockout mice (Hsu et al., 2012; Tsai et al., 2012) These mice were characterized by hepatic inflammation, fibrosis, and development of spontaneous tumors similar to HCC, demonstrating the tumor-suppressor function of this miRNA and its important role in liver metabolism and differentiation of hepatocytes (Jensen et al., 2003; Gramantieri et al., 2007; Lin et al., 2008; Bai et al., 2009; Fornari et al., 2009; Tsai et al., 2009; Callegari et al., 2013).

On the other hand, up-regulation of miR-221, may be involved from the very early stage of hepatocarcinogenesis, and expression of the miRNA may progressively increase during malignant transformation. Especially, high expression of miR-221 can be used to predict local recurrence of HCC, and fold changes in miR-221 less than 1 can be used as a predictive marker of metastasis after curative surgical resection in patients with HCC (Yoon et al., 2011). Thus, among the miRNAs that are upregulated in HCC, there is evidence in support of the tumor-promoting activity of miR-221. It is upregulated in 70–80% of HCC samples (Fornari et al., 2008). From a functional point of view, HCC cells overexpressing miR-221 show increased growth, proliferation, migration, and invasion capability (Fornari et al., 2008; Medina et al., 2008; Garofalo et al., 2009; Gramantieri et al., 2009; Pineau et al., 2010; Callegari et al., 2012). Additionally, high level of miR-221 positively correlated with cirrhosis, tumor size and tumor stage, and negatively correlated with overall survival. miR-221 serum level monitoring could be of clinical relevance as a potential diagnosis tool and biomarker of treatment efficacy. It remains to be established which miRNA can sensitively and reliably be correlated with the presence of HCC at early stages of disease development and prognosis (Borel et al., 2012a).

miR-429 expression increased AFB1-DNA adducts in the SMMC-7721 Cells. To explore the effects of miR-429 expression on AFB1-DNA formation, Huang et al. (2013), accomplished a toxin experiment of AFB1 in the SMMC-7721 cells transfected by different mimics. Results showed that group with overexpression of miR-429 had elevated levels of AFB1-DNA adducts compared with control group. MiR-429 is classified as a member of miR-200 family and may play an important role in tumor prognosis. Overexpression of miR-429 induces cell proliferation and inhibits cell apoptosis. On the contrary, the suppression of miR-429 expression hindered cell proliferation and promoted cell apoptosis. These data suggest that this microRNA plays an important role in liver tumorigenesis, and functionally acts as an oncogene in HCC. Increasing evidence has shown that the levels of AFB1-DNA adducts correlate with HCC risk and prognosis, whereas the formation process of AFB1-DNA adducts can be modified by some factors such as detoxifying enzymes and DNA repair enzymes (Long et al., 2006, 2011; Xia et al., 2013). Is possibly that miR-429 can target some detoxification enzyme genes and/or DNA repair genes and reduce their detoxification capacity or DNA repair capacity and subsequently increase DNA damage and promote AFB1-DNA adducts formation. These results provided new insights into the mechanism of HCC induced by AFB1 (Huang et al., 2013).

The maintenance of genomic integrity through efficient DNA repair is essential for propagation of cellular life (Natarajan and Palitti, 2008). Nucleotide excision repair (NER) is one of the most versatile DNA repair system for elimination of bulky DNA adducts caused by environmental agents (Nouspikel, 2009) as AFB1 and other carcinogens. A possibility is that AFB1 could interfere with cellular NER through the regulation of microR-NAs. Several miRNAs involved in DNA repair have been identified (Crosby et al., 2009; Yan et al., 2010; Hu and Gatti, 2011). A recent study showed that miR-192 directly targets a NER-associated protein (Georges et al., 2008). A bioinformatic analysis of miRNAs which potentially played a role in NER, show that miR-192, was the most differentially upregulated miRNA. The expression of ERCC3 and ERCC4 were reduced when miR-192 was overexpressed. Also has been observed that the relative repair capacity of damage by HepG2 and HeLa cells was reduced (Xie et al., 2011). Since of AFB1 is an important risk factor of HCC and AFB1-DNA adducts are known to be repaired by NER, dietary AFB1 exposure could impaired NER mediated by miRNAs like miR-192.

miR-500 is an oncofetal miRNA, which is highly expressed in fetal liver, more than in adult normal liver, and aberrantly expressed in HCC. This miRNA was associated with liver maturation in a mouse model of liver development. Levels tended to be higher in HCC lines and tumor samples when compared with matched normal tissue. Importantly, significant difference in miR-500 expression was found between normal liver and liver cirrhosis samples, suggesting that miR-500 expression was upregulated during cirrhosis development. An increased amount of miR-500 was found in the sera of 3 out of 10 HCC patients, which means that liver cancer-specific miRNA such as miR-500 is circulating in the peripheral blood and can be a novel diagnostic marker. These results show that the miR-500 abundance profile in serum of the HCC patients might reflect physiological and/or pathological conditions. However, although results are promising for miRNA-based HCC screening, further validation is suggested (Yamamoto et al., 2009).

miR-148, another candidate. There are also reports suggesting that drug-metabolizing enzymes such as CYP family genes are targeted by certain miRNAs. The expression of drug- and xenobiotic-metabolizing enzymes and nuclear receptors and their regulation by miRNA could be important factors for the outcomes of toxicity (Yokoi and Nakajima, 2011). Members of the CYP family are the most important enzymes catalyzing the metabolism of xenobiotics including drugs, environmental chemicals, and carcinogens. The different profiles of the expression of P450 isoenzymes determine the amount of reactive intermediates formed and the resulting toxic response. P450s are also known to bioactivate many procarcinogens to their ultimate carcinogens as in the case of AFB1. Recently, some P450s and nuclear receptors have been found to be post-transcriptionally regulated by miRNAs. Aflatoxin B1 and G1 are known to be oxidized efficiently to genotoxic metabolite(s) by CYP3A (Shimada et al., 1989; Forrester et al., 1990), epoxidation of AFB1 is catalyzed by CYP1A2 and CYP3A4 in humans (Gallagher et al., 1994; Ueng et al., 1995). The role of miRNA in the regulation of the expression of CYP3A4 has been reported, Takagi et al. (2008), found that miR-148 modulated inducible and/or constitutive levels of CYP3A4 in human liver cancer.

miRNAs are important regulators for CYP3A. Among these differentially regulated miRNAs, miR-155 appears to be the most prominent regulator as it was significantly associated with lower hepatic CYP3A activity (Vuppalanchi et al., 2013). CYP3A4 is the most abundant hepatic and intestinal CYP enzyme in humans, contributing to the metabolism of various drugs (Gonzalez and Yu, 2006), as AFB1. Pan et al. (2009) suggest that intervention of miRNA pathways may modify CYP3A4 expression and alter CYP3A4-catalyzed drug activation. Of particular note, miR-148a has been shown to control post-transcriptional regulation of PXR and, consequently, affect the expression of CYP3A4 (Takagi et al., 2008). Another study suggests that miR-27a and miR-27b may target RXR and regulated of CYP3A4 transcriptional expression (Ji et al., 2009b). The results indicate that intervention of miRNA pathways can be translated into an altered sensitivity of cells to xenobiotics. These findings may provide increased understanding of the complex regulation of CYP3A4 expression, as well as determine the role of miRNAs in drug metabolism and disposition (Pan et al., 2009).

Aflatoxin B1 (AFB1) is carcinogenic due its potential in inducing the oxidative stress and distortion of the most antioxidant enzymes (Abdel-Wahhab et al., 2007; El-Agamy, 2010; Alm-Eldeen et al., 2013). Recently, the role of miRNAs in oxidative stress-mediated etiology is emerging. Dong et al. (2013) found that miR-214 directly bound to 3- -UTR of the GSR and POR genes, and repressed their endogenous expressions and activities. These findings suggested miR-214 mediating down-regulation of glutathione reductase and CYP oxidoreductase genes might play an important role in oxidative stress in live cells. Wang et al. (2008, 2013) reported that miR-214 is one of the most significantly downregulated miRNAs in HCC patients. Extensive research has suggested that continued oxidative stress is a common pathologic pathway for most chronic diseases including cancer, and liver diseases. Therefore, Dong et al. (2013) postulated that miR-214 could be a key post-transcriptional regulator in oxidative stressmediated human diseases. This microRNA will be also an important molecule to study in oxidative stress induced by AFB1 in liver.

Future validation for all of these miRNAs will be needed to assess their prognostic significance. It is notable that only a few miRNA signatures could potentially be used for diagnosis and prognosis, and even for these there is still a long way to go before they can be used in clinics. To achieve this goal, the miRNA signatures need to be further validated with high accuracy in prospective studies (Ji and Wang, 2009).

#### **FUTURE PERSPECTIVES IN TOXICOLOGICAL RESEARCH**

There is a need for novel markers that would combine the less invasiveness of a blood test and serve as a reliable early detection method. miRNAs definitely have this potential because not only they can be detected in plasma, but their sensitivity and stability are suitable for a clinical setting. Depending on the method, as little as one copy can be detected. The discovery of circulating miRNAs offers interesting clinical perspectives but this field of research is quite recent and more work has to be done.

Recently, measurement of circulating miRNAs has shown promise in identification of new biomarkers of liver injury. Further studies are needed to evaluate the sensitivity and specificity as well as validate the omics biomarkers of hepatotoxicityxenobiotic exposure related.

It is difficult to establish the precise cause-effect relationships among environmental chemicals, miRNA alterations, and diseases. Future studies will need to demonstrate the contribution of environment-miRNA interaction to environmental human disease. The rapidly growing evidence linking miRNAs and environmental chemical, coupled with the unique regulatory role of miRNAs in gene expression, makes miRNAs potential biomarkers for elucidating the mechanisms and developing more effective prevention strategies for environmental diseases (Hou et al., 2011).

Currently, over five billion people worldwide experience uncontrolled exposure to aflatoxin (Strosnider et al., 2006). What remains unknown is how many liver cancer cases can be attributed to this aflatoxin exposure worldwide. Recently Liu and Wu (2010) have developed a risk assessment for the contribution of aflatoxin to the global burden of HCC. Of the 550,000– 600,000 new HCC cases worldwide per year, they estimate about 25,200–155,000 (4.6–28.2%) may be attributable to aflatoxin exposure alone. The broad range in the estimate reflects limitations in determining levels of aflatoxin exposures, uncertainties in the nature of the dose-response curve, uncertainties in the mode of interaction between aflatoxins and viruses, and incomplete data on the prevalence of HBV in different regions of the world. Data driven estimates of the noncarcinogenic health effects of aflatoxins in humans have not been undertaken (Kensler et al., 2011).

The understanding of miRNA biology has advanced greatly in recent years, and the continuous technological advances in accurate miRNA detection, prospect a very promising role for miRNAs as novel biomarkers of environmental chemical exposure-related diseases Identifying chemical-specific miRNAs will not only help our understanding of environmental disease, but may open the way to novel biomonitoring and preventive strategies. Therefore, it is critically important to be able to identify and validate miR-NAs that can be induced by specific environmental chemicals and regulate gene expression (Hou et al., 2011).

Understanding the miRNAs roles in toxicological processes requires overall a toxicogenomic approach. On the other hand, miRNA profiling data looks promising as a tool to predict the potential toxicity of unknown compounds. Thus, miRNA signatures of a known toxic compound may include miRNAs related to cellular response to stress, xenobiotic metabolism, and/or DNA repair. These signatures derived from supervised classification algorithms may effectively identify potential toxic compounds. Several examples of miRNAs active in cellular stress as well as in interactions of a number of toxicants. miRNA profiling may lead to the discovery of miRNA exposure biomarkers, which might work as sentinel molecules to better predict both efficacy and safety. The miRNA field in toxicology is still in its early stages. However, progress is occurring at a fast pace and the numbers of publications featuring miRNAs are increasing. As the roles of miRNAs in cellular response to xenobiotic stress and the development of physiological changes and other toxicological phenomenon such susceptibility and resistance are gradually uncovered, the coming years promise to be full of exciting avenues of miRNA research in toxicogenomics (Lema and Cunningham, 2010).

Potential microRNAs that could serve as possible markers of HCC by exposure to aflatoxins are miR-27a, miR-27b, miR-122, miR-148, miR-155, miR-192, miR-214, miR-221, miR-429, and miR-500. Future studies should include some of these microRNAs and confirm their relationship with the induction of HCC due to aflatoxin exposure.

#### **REFERENCES**


miR-21 in hepadnavirus-associated hepatocellular carcinoma contributes to the malignant phenotype. *Am. J. Pathol.* 173, 856–864. doi: 10.2353/ajpath.2008. 080096


expression of cytochrome P450 3A4. *J. Biol. Chem.* 283, 9674–9680. doi: 10.1074/jbc.M709382200


mitogen-activated protein kinase-heat shock protein 27 pathway. *Hepatology* 51, 1614–1623. doi: 10.1002/hep.23566


**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 December 2013; paper pending published: 08 January 2014; accepted: 26 February 2014; published online: 18 March 2014.*

*Citation: Valencia-Quintana R, Sánchez-Alarcón J, Tenorio-Arvide MG, Deng Y, Montiel-González JMR, Gómez-Arroyo S, Villalobos-Pietrini R, Cortés-Eslava J, Flores-Márquez AR and Arenas-Huertero F (2014) The microRNAs as potential biomarkers for predicting the onset of aflatoxin exposure in human beings: a review. Front. Microbiol. 5:102. doi: 10.3389/fmicb.2014.00102*

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

*Copyright © 2014 Valencia-Quintana, Sánchez-Alarcón, Tenorio-Arvide, Deng, Montiel-González, Gómez-Arroyo, Villalobos-Pietrini, Cortés-Eslava, Flores-Márquez and Arenas-Huertero. 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.*

### Use of benzo analogs to enhance antimycotic activity of kresoxim methyl for control of aflatoxigenic fungal pathogens

#### *Jong H. Kim\*, Noreen Mahoney , Kathleen L. Chan , Bruce C. Campbell , Ronald P. Haff and Larry H. Stanker*

*Foodborne Toxin Detection and Prevention Research Unit, Western Regional Research Center, USDA-ARS, Albany, CA, USA*

#### *Edited by:*

*Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran*

#### *Reviewed by:*

*Masoomeh Shams-Ghahfarokhi, Tarbiat Modares University, Iran Eva-Guadalupe Lizárraga-Paulín, Universidad Nacional Autónoma De México, Mexico*

#### *\*Correspondence:*

*Jong H. Kim, Foodborne Toxin Detection and Prevention Research Unit, Western Regional Research Center, USDA-ARS, 800 Buchanan Street, Albany, CA 94710, USA e-mail: jongheon.kim@ars.usda.gov* The aim of this study was to examine two benzo analogs, octylgallate (OG) and veratraldehyde (VT), as antifungal agents against strains of *Aspergillus parasiticus* and *A.flavus* (toxigenic or atoxigenic). Both toxigenic and atoxigenic strains used were capable of producing kojic acid, another cellular secondary product. *A. fumigatus* was used as a genetic model for this study. When applied independently, OG exhibits considerably higher antifungal activity compared to VT. The minimum inhibitory concentrations (MICs) of OG were 0.3–0.5 mM, while that of VT were 3.0–5.0 mM in agar plate-bioassays. OG or VT in concert with the fungicide kresoxim methyl (Kre-Me; strobilurin) greatly enhanced sensitivity of *Aspergillus* strains to Kre-Me. The combination with OG also overcame the tolerance of *A. fumigatus* mitogen-activated protein kinase (MAPK) mutants to Kre-Me. The degree of compound interaction resulting from chemosensitization of the fungi by OG was determined using checkerboard bioassays, where synergistic activity greatly lowered MICs or minimum fungicidal concentrations. However, the control chemosensitizer benzohydroxamic acid, an alternative oxidase inhibitor conventionally applied in concert with strobilurin, did not achieve synergism. The level of antifungal or chemosensitizing activity was also "compound—strain" specific, indicating differential susceptibility of tested strains to OG or VT, and/or heat stress. Besides targeting the antioxidant system, OG also negatively affected the cell wall-integrity pathway, as determined by the inhibition of *Saccharomyces cerevisiae* cell wall-integrity MAPK pathway mutants. We concluded that certain benzo analogs effectively inhibit fungal growth. They possess chemosensitizing capability to increase efficacy of Kre-Me and thus, could reduce effective dosages of strobilurins and alleviate negative side effects associated with current antifungal practices. OG also exhibits moderate antiaflatoxigenic activity.

#### **Keywords: aflatoxin, antioxidant system,** *Aspergillus***, cell wall integrity, chemosensitization, octylgallate, strobilurin, veratraldehyde**

#### **INTRODUCTION**

Controlling fungi that produce hepato-carcinogenic aflatoxins in crops, such as tree nuts, corn, peanuts, etc., is problematic as effective commercial fungicides for treating aflatoxigenic fungi are very limited (Roze et al., 2013). Aflatoxins are secondary metabolites produced mainly by the filamentous fungi *Aspergillus flavus* and *A. parasiticus*. Significant amounts of harvested crop can be made unsuitable for sale and consumption as a result of aflatoxin contamination. Very low level (parts per billion) of aflatoxin contamination can have a perniciously negative effect on food safety and economic value of a number of crops from year to year (Campbell et al., 2003). Therefore, effective methods are continually needed for control of aflatoxigenic fungal pathogens.

Strobilurins are widely used agricultural fungicides. Strobilurins were initially identified in the fungus *Strobilurus tenacellus*, and were synthetically developed into several subgroups, such as kresoxim methyl (Kre-Me), azoxystrobin, pyraclostrobin, etc. (Bartlett et al., 2002 and references therein). The molecular target for strobilurins is the *bc*<sup>1</sup> complex (complex III; ubiquinol-cytochrome *c* oxidoreductase, EC 1.10.2.2) in the mitochondrial respiratory chain (MRC). Strobilurins specifically bind to the QP (QO) center of cytochrome *b* (Bartlett et al., 2002 and references therein), resulting in the inhibition of MRC.

Development of fungal resistance to conventional antifungal agents is a global agricultural issue (Ghannoum and Rice, 1999; Wood and Hollomon, 2003; Possiede et al., 2009; Cools and Hammond-Kosack, 2013). For example, agricultural fields receiving continuous applications of the fungicide, strobilurin, resulted in the development of insensitivity of fungi to this fungicide (Keinath, 2009). Noteworthy is that if strobilurincontaining fungicides are applied at suboptimal time-points of fungal growth, these fungicides can actually potentiate mycotoxin production in fungi (Ellner, 2005).

Similar type of fungicide-potentiation of mycotoxin production was documented in *Penicillium verrucosum*, a filamentous fungal pathogen producing the mycotoxin, citrinin (Schmidt-Heydt et al., 2013). When *P. verrucosum* was treated with the fungicide "Rovral" (Iprodione), an inhibitor of DNA/RNA biosynthesis and cell division, the growth of fungi was decreased. However, a concomitant strong induction of citrinin biosynthesis also occurred in fungi with the same treatment (Schmidt-Heydt et al., 2013). Altogether, studies indicated that certain conventional fungicides could stimulate secondary metabolism, especially mycotoxin production, in fungal pathogens. Therefore, effective strategies are urgently needed to overcome counterproductive repercussions of fungicides currently in use.

Recent studies have shown that safe, natural phenolic compounds or their structural derivatives (e.g., benzo derivatives) could act as potent antifungal or antimycotoxigenic agents (Beekrum et al., 2003). For example, vanillic or caffeic acid not only inhibited the growth of *Fusarium verticillioides*, but also reduced its production of the mycotoxin, fumonisin (Beekrum et al., 2003). Natural benzo derivatives could also effectively inhibit the growth of *A. fumigatus*, *A. flavus*, *A. terreus,* and *Penicillium expansum* (Kim et al., 2011). These fungi are causative agents of human invasive aspergillosis or are producers of mycotoxins, including aflatoxin, patulin, gliotoxin, etc. The redoxactive natural phenolic agents can be potent redox cyclers that inhibit fungal growth by disrupting cellular redox homeostasis (and thus triggering cellular oxidative stress) (Guillen and Evans, 1994; Jacob, 2006). For defense, the fungal antioxidant system, such as oxidative signaling pathway, plays an important role for fungal tolerance to those phenolic agents (Kim et al., 2011).

Genes involved in stress-signaling pathways are also important for fungal virulence, pathogenesis and protection from oxidative burst exerted by the host (Washburn et al., 1987; Hamilton and Holdom, 1999; Clemons et al., 2002; de Dios et al., 2010). Oxidative stress signals sensed by a fungal cell are incorporated into the upstream mitogen-activated protein kinase (MAPK) pathway, which regulates the expression of the downstream response genes (such as antioxidant enzyme genes) detoxifying the stress (Miskei et al., 2009). In *A. fumigatus*, SakA and MpkC are orthologous MAPKs to *Saccharomyces cerevisiae* Hog1p, which plays a key role in countering oxidative stress (Toone and Jones, 1998; Lee et al., 2002; Xue et al., 2004; Reyes et al., 2006; Miskei et al., 2009).

Chemosensitization is a strategy where combined application of certain types of compounds along with a conventional fungicide/drug enhances the effectiveness of the conventional agents (Niimi et al., 2004; Agarwal et al., 2012; Campbell et al., 2012). Noteworthy is that certain benzo derivatives possessed antifungal chemosensitizing capability. In our prior study, co-application of antifungal agents with chemosensitizing benzo derivatives, such as 2-hydroxy-5-methoxybenzaldehyde, greatly enhanced the efficacy of antifungal agents (Kim et al., 2011). Thus, chemosensitization could lead to lowering dosages of conventional fungicides/drugs required for control of pathogens, especially drug-resistant strains. Collectively, these studies showed the potential for safe, natural phenolics to serve as effective antifungal and/or antimycotoxigenic agents.

In nature, in addition to oxidative stress, high temperature (heat) stress is another type of environmental challenge that many microbes face, which also triggers "signaling cascades" in fungal cells (Morano et al., 2012). Heat treatment is also a strategy to prevent contamination by food spoilage fungi in foods (Dagnas and Membré, 2013). In corn, artificial drying of maize kernels with high temperatures is one of the postharvest practices to prevent fungal growth and aflatoxin production (Hawkins et al., 2005). A prior study showed that heat treatment at 70◦C significantly reduced the maize kernel infection of *A. flavus*. However, heat treatment had no effect on aflatoxin concentration, which reflects the heat stability of the mycotoxin (Hawkins et al., 2005). Also, heat treatment can result in deterioration of the quality of the crop (seed breakage, viability, etc.). Therefore, development of new alternative strategies, which warrant early intervention of mycotoxin production/fungal growth as well as the quality of harvested crops, is necessary.

In this study, we investigated the role of two benzo derivatives, octylgallate (OG) and veratraldehyde (VT), currently used as food additives, as antifungal agents against strains of *A. flavus, A. parasiticus,* or *A. fumigatus*. OG and VT are generally regarded as safe (GRAS) reagents (FDA, 2011). We also evaluated antifungal chemosensitizing capacities of OG and VT, especially for overcoming strobilurin resistance of *Aspergillus* MAPK gene deletion mutants (*sakA*-, *mpkC*-). Kre-Me, containing (*E*)-methyl methoxyiminoacetate group in the structure, was tested as an exemplary strobilurin. Kre-Me is currently applied in the agricultural field for control of various fungal diseases caused by ascomycete, basidiomycete, oomycete, etc., while it exhibited high toxicity to the agriculturally important insects, bees (Bartlett et al., 2002). Our results showed that OG could serve as a potent antifungal chemosensitizer to Kre-Me for controlling *Aspergillus* strains.

#### **MATERIALS AND METHODS**

#### **MICROBIAL STRAINS AND CULTURE CONDITIONS**

Microbial strains used in this study are summarized in **Table 1**. *Aspergillus* strains were cultured on potato dextrose agar (PDA) at 30◦C, except otherwise noted in the text. Yeast strains, wild type (WT) and gene deletion mutants of *Saccharomyces cerevisiae* (**Table 1**), were cultured on Synthetic Glucose (SG; Yeast nitrogen base without amino acids 0.67%, glucose 2% with appropriate supplements: uracil 0.02 mg mL−1, amino acids 0.03 mg mL−1) or Yeast Peptone Dextrose (YPD; Bacto yeast extract 1%, Bacto peptone 2%, glucose 2%) medium at 30◦C.

#### **CHEMICALS**

Antifungal compounds, kresoxim methyl (Kre-Me; strobilurin), octylgallate (octyl 3,4,5-trihydroxybenzoic acid; OG), veratraldehyde (3,4-dimethoxybenzaldehyde; VT), benzohydroxamic acid (BHAM) (**Figure 1**), were procured from Sigma Co. (St. Louis, MO, USA). Each compound was dissolved in dimethylsulfoxide (DMSO; absolute DMSO amount: <2% in media) before incorporation into culture media (except for those plates used in aflatoxin assays; see below). Throughout this study, control plates (No treatment) contained DMSO at levels equivalent to that of cohorts receiving antifungal agents, within the same set of experiments.

#### **Table 1 | Microbial strains used in this study.**


*<sup>a</sup> NRRL, National Center for Agricultural Utilization and Research, USDA-ARS, Peoria, IL, USA.*

*<sup>b</sup> Saccharomyces Genome Database. Available online: http://www.yeastgenome.org (accessed on 10 January 2014).*

#### **SUSCEPTIBILITY TESTING**

#### *Agar plate-bioassay in Aspergillus*

In agar plate-bioassays, susceptibility of filamentous fungi to OG, VT, or Kre-Me, alone or in combination, were measured based on percent radial growth of treated compared to control fungal colonies (see Figures and Tables for test concentrations). The percent inhibition of growth was calculated using the Vincent equation (Vincent, 1947) [% inhibition = 100 (*C* − *T*)/*C*; where *C* = diameter of fungal colony on control plate (receiving only DMSO), and *T* = diameter of fungal colony on the treated plate]. Minimum inhibitory concentration (MIC) values on agar plates were based on triplicate assays and defined as the lowest concentration of agent where no fungal growth was visible on the plate. For the above assays, fungal conidia (5 <sup>×</sup> 103) were diluted in phosphate buffered saline and applied as a drop onto the center of PDA plates with or without antifungal compounds. Growth was observed for 5–7 days.

#### *CLSI liquid bioassay in Aspergillus*

To determine the precise level of chemosensitizing activities of OG or BHAM (0.00625, 0.0125, 0.025, 0.05, 0.1, 0.2, 0.4, 0.8 mM) to Kre-Me (0.5, 1, 2, 4, 8, 16, 32µg mL−1) in the strains of *Aspergillus*, checkerboard bioassays (triplicate) (0.<sup>4</sup> <sup>×</sup> 104–5 <sup>×</sup> 104 CFU mL−1) were performed in microtiter wells using a broth microdilution method (in RPMI 1640 medium; Sigma Co., St. Louis, MO, USA), according to protocols outlined by the Clinical and Laboratory Standards Institute (CLSI) M38- A (CLSI, 2008). OG was chosen since this compound showed much higher antifungal activity than VT (see Results). For comparison, BHAM, an alternative oxidase (AOX) inhibitor conventionally used in combination with strobilurin, was also included as a reference chemosensitizer in this test. RPMI 1640 medium was supplemented with 0.03% L-glutamine and buffered with 0.165 mM 3-[*N-*morpholino] propanesulfonic acid. MICs, lowest concentration of agents showing no visible fungal growth in microtiter wells (200 µL per well), were assessed after 48 and 72 h. Minimum fungicidal concentrations (MFCs), lowest concentration of agents showing ≥ 99.9% fungal death, were determined following completion of MIC assays by spreading entire volumes of microtiter wells (200µL) onto individual PDA plates, and culturing for another 48 and 72 h. Compound interactions, Fractional Inhibitory Concentration Indices (FICIs) and Fractional Fungicidal Concentration Indices (FFCI), were calculated as follows: FICI or FFCI = (MIC or MFC of compound A in combination with compound B/MIC or MFC of compound A, alone) + (MIC or MFC of compound B in combination with compound A/MIC or MFC of compound B, alone). Interactions were defined as: "synergistic" (FICI or FFCI ≤ 0.5) or "indifferent" (FICI or FFCI > 0.5–4) (Odds, 2003).

#### **GROWTH RECOVERY TEST IN** *ASPERGILLUS* **STRAINS TREATED WITH HIGH TEMPERATURES: AGAR PLATE-BIOASSAY**

Agar plate-based bioassay was performed to evaluate differential susceptibility of *Aspergillus* strains to high temperatures. First, fungal conidia (5 <sup>×</sup> 103) were spotted on PDA (see above for method), and were initially incubated at four different temperatures (moderate: 30, 35◦C; high: 45, 55◦C). Triplicate PDA plates were then removed from each temperature (30, 35, 45, or 55◦C) at day 1, 2, 3, and 4, and were transferred to 30◦C for additional 6, 5, 4, and 3 days of growth, respectively, resulting in a total of 7 days of incubation for each treatment (i.e., 1 day growth at 55◦C + 6 days growth at 30◦C = Total 7 days growth, 2 days growth at 55◦C + 5 days growth at 30◦C = Total 7 days growth, 3 days growth at 55◦C + 4 days growth at 30◦C = Total 7 days growth, 4 days growth at 55◦C + 3 days growth at 30◦C = Total 7 days growth). For controls, *Aspergillus* strains were grown solely at respective temperature (30, 35, 45, or 55◦C) for 7 days. The level of growth recovery at 30◦C was evaluated based on fungal radial growth as described above.

#### **GROWTH RECOVERY TEST IN CELL WALL INTEGRITY MUTANTS OF** *SACCHAROMYCES CEREVISIAE* **TREATED WITH OG: YEAST DILUTION BIOASSAY**

To determine the effect of OG on the cell wall-integrity system of fungi, sorbitol recovery tests were performed using Petri plate-based yeast dilution bioassays. Ten-fold diluted, i.e., 100 (1 <sup>×</sup> 106 cells), 10−1, 10−2, 10−3, 10−4, 10−5, strains of *S. cerevisiae* BY4741 (WT), *bck1* and *slt2* were spotted on: (1) SG only, (2) SG + caffeine (5 mM; a positive control for cell wall disruption) or SG + OG (0.03, 0.04, 0.05, 0.06 mM) (Testing sensitivity of *bck1* and *slt2* mutants to compounds), and (3) SG + sorbitol (0.5 M) + caffeine or OG (Testing growth recovery of *bck1* and *slt2* mutants by sorbitol). Cell growth was monitored for 5–7 days. If the growth score on the sorbitol-containing medium was higher than that on the "no sorbitol" medium, OG was considered to negatively affect cell wall integrity.

#### **AFLATOXIN ASSAYS**

For aflatoxin analysis, spore suspensions (200 spores) were inoculated onto the center of PDA plates. Cells were grown at 30◦C for 7 days with or without OG (OG concentration: 0.025, 0.05, 0.1 mM; added directly into PDA), with each treatment in triplicate. Aflatoxins were quantitated as described previously (Rodriguez and Mahoney, 1994). Aflatoxin standard solutions (AFB1, AFB2, AFG1, AFG2), used for the quantification of aflatoxins from OGtreated samples, were prepared as described in AOAC 971.22 (AOAC International, 2005).

#### **KOJIC ACID ASSAYS**

To examine whether the cellular secondary-product metabolism other than aflatoxin production is normal in the atoxigenic strains used in this study (namely, *A. flavus* 21882, 18543), we determined kojic acid (KA) production in these atoxigenic strains. *A. flavus* 3357 and 4212, toxigenic strains, were used as controls for KA production. Fungal spores (1 <sup>×</sup> 105 spores per 100µL, 0.05% Tween 80) were inoculated into 50 mL potato dextrose broth, and were incubated at 30◦C without shaking. KA production was measured from the culture media using an HPLC system consisting of a degasser, autosampler, quaternary pump, and photodiode array detector (Agilent 1100, Santa Clara, CA USA). Culture media aliquots of 1 mL were filtered through 0.45µm sterile nylon 25 mm syringe filters (Fisherbrand, Thermo Fisher Scientific, Waltham, MA, USA), with injection volumes of 20µL on a 4.6 × 250 mm Inertsil 5µm ODS-3 column (GL Sciences, Torrance, CA, USA) using an isocratic mobile phase consisting of MeOH/0.1% H3PO4 (25:75, v/v) at a flow rate of 1 mL per minute. KA quantitation was based on UV detection at 265 nm. HPLC quantification of KA was linear in the range of 0.02–4.00µg per 20µL, with a retention time of 5.4 min.

#### **STATISTICAL ANALYSIS**

Statistical analysis (student's *t*-test) was performed based on "Statistics to use" (Kirkman, 2014), where *p* < 0.05 was considered significant.

#### **RESULTS**

#### **SUSCEPTIBILITY OF** *ASPERGILLUS* **STRAINS TO OG OR VT: AGAR PLATE-BIOASSAY**

As shown in **Table 2**, OG possessed much higher antifungal activity compared to VT. The average MICs of OG was 0.4 mM, while that of VT was higher than 4.8 mM (viz. more than 10 times higher antifungal activity of OG).

#### **Table 2 | Differential sensitivity of** *Aspergillus* **strains to OG, VT, or high temperature (55◦C).**


*Concentrations (mM) are MICs examined on PDA plates.*

*aMaximum day(s) of incubation at 55*◦*C, which can result in growth recovery of fungi at 30*◦*C.*

*bVT was tested up to 5.0 mM. For statistical calculation (student's t-test) purpose, 10.0 mM (doubling of 5.0 mM) was used.*

*cStudent's t-test for paired data (MICOG) was vs. MICVT determined in nine strains.*

Results also indicated differential susceptibility (namely, different MIC level) of *Aspergillus* strains to the treatments, where: (1) All *A. fumigatus* strains (WT, *sakA*-, *mpkC*-) were more sensitive to either OG or VT compared to other *Aspergillus* strains (*p* < 0.05 for OG, *p* < 0.005 for VT), (2) *A. fumigatus* MAPK mutants (*sakA*-, *mpkC*-) were more sensitive to OG compared to the WT (*p* < 0.005). Although MICs of VT for all *A. fumigatus* strains were similar (MICVT: 4.5 mM), the level of growth of MAPK mutants was also lower than the WT (see below; **Table 3**). Thus, results indicated MAPK mutants lack the defense mechanism to protect fungal cells from the toxicity generated by redox-active benzo derivatives, (3) *A. flavus* 21882 and *A. parasiticus* 2999 were more tolerant to OG compared to other *Aspergillus* strains (*p* < 0.05), (4) *A. flavus* 3357, 4212 and *A. parasiticus* 5862 showed similar levels of susceptibility to OG, and (5) *A. flavus* 3357, 4212, 21882, 18543 and *A. parasiticus* 5862, 2999 also showed similar levels of susceptibility to VT (up to 5 mM).

#### **SUSCEPTIBILITY OF** *ASPERGILLUS* **TO HIGH TEMPERATURES**

Heat responses of *Aspergillus* strains were compared to their responses to benzo derivatives. *Aspergillus* strains were treated with moderate (30, 35◦C) to high (45, 55◦C) temperatures, and were then transferred to 30◦C for growth recovery. From the pathogenicity perspective, *A. flavus* shares agro-infectivity with *A. parasiticus*, while *A. flavus* shares human-infectivity with *A.* *fumigatus* (**Figure 2A**). As shown in **Figure 2B**, *A. fumigatus* strains, both WT and MAPK mutants, were much more tolerant to high temperature (55◦C), compared to other *Aspergillus* strains examined. For example, *A. fumigatus* strains did not germinate on PDA when they were maintained at 55◦C for 7 days. However, *A. fumigatus* treated with heat (55◦C) for 1–4 days could recover the growth after incubation at 30◦C, while *A. flavus* 3357, 4212 and 21882 could not recover their growth even after 1 day-heat treatment at 55◦C (see **Table 2** for summary). The remaining *Aspergillus* strains (18543, 5862, 2999) showed growth recovery only with 1 day-heat treatment at 55◦C. *A. fumigatus* MAPK mutants were marginally more sensitive to heat (55◦C) (i.e., 2– 10% less radial growth during recovery w/ 3–4 day-treatment at 55◦C) compared to the WT (**Figure 2B**).

The order of susceptibility of *Aspergillus* strains to high temperature (55◦C) (higher susceptibility to lower susceptibility) was: *A. flavus* 3357, 4212, 21882 > *A. flavus* 18543, *A. parasiticus* 5862, 2999 > *A. fumigatus* strains. Therefore, like in benzo treatments, *Aspergillus* strains showed differential susceptibility to high temperature (55◦C), where *A. fumigatus* strains were the most heattolerant species among *Aspergillus* strains tested. However, there was no direct correlation in fungal responses between benzo and heat treatments, indicating the mechanism of antifungal action of benzo analog(s) is considered to be different from that of heat treatment (see **Table 2**). [*Aspergillus* strains were also sensitive to



*aData shown in the table are % radial growth of fungi compared to control (no treatment). SD* < *5%. No visible cell growth (0%) is in bold. bFew: Only few colonies appeared.*

45◦C. However, all *Aspergillus* strains recovered their growth after transfer from 45 to 30◦C (see Figure S1 for exemplary bioassay with *A. flavus*). As expected, *Aspergillus* strains grew normally at the moderate temperatures (30 and 35◦C) (Figure S1)].

#### **SUSCEPTIBILITY OF** *A. FUMIGATUS* **TO CHEMOSENSITIZATION**

Chemosensitizing capability of OG or VT to Kre-Me was tested in *A. fumigatus* strains by using agar plate-bioassays. Results showed that: (1) When Kre-Me was co-applied with VT, the MIC of VT was lowered from 4.5 mM to 3.0 mMMAPK or 4.0 mMWT, respectively (**Table 3**), and (2) When Kre-Me was co-applied with OG, the MIC of OG was lowered from 0.3 mMMAPK or 0.35 mMWT to 0.05 mMMAPKandWT. Therefore, both OG and VT possessed chemosensitizing capability to Kre-Me in *A. fumigatus* strains, where OG possessed much higher chemosensitizing capacity (viz. needed ∼10 times lower concentration for chemosensitization; *p* < 0.005) than VT (**Table 3**).

Noteworthy is that the *A. fumigatus* MAPK mutants were more tolerant to Kre-Me (**Table 3**), where: (1) The level of radial growth of MAPK mutants with independent treatment of Kre-Me was marginally greater (4–9% more growth) than the WT, and (2) When VT was co-applied with Kre-Me (25µM), MICVT was 3.0 mM for the WT, while that of MAPK mutants was 4.0 mM (namely, higher concentration of VT is needed compared to the WT) during this chemosensitization (**Table 3**). Therefore, although VT-mediated chemosensitization was achieved in all *A. fumigatus* strains, results showed relatively higher tolerance of MAPK mutants to "Kre-Me + VT" compared to the WT [see below (**Table 4**) for relative tolerance of MAPK mutants to "Kre-Me + BHAM" compared to the WT]. However, Kre-Me tolerance of *A. fumigatus* MAPK mutants was completely abolished by OG (**Table 3**).

Co-application of OG or VT with Kre-Me also enhanced growth inhibition in toxigenic strains of *Aspergillus* (*A. flavus* 3357, *A. parasiticus* 5862) (viz. either complete growth inhibition or reduced radial growth, depending on types of strains/combination of compounds; **Figure 3**). For example, the growth of *A. parasiticus* 5862 was completely inhibited by Kre-Me + VT, while a similar level of growth inhibition in *A. flavus* 3357 could be achieved by Kre-Me + OG (**Figure 3**). Independent treatment of each compound alone at the same concentration did not achieve such a level of growth inhibition. Thus, results indicated differential susceptibility of each strain to different combinations of antifungal agents, where 3357 showed higher susceptibility (complete growth inhibition) to Kre-Me + OG, whereas 5862 was more susceptible to Kre-Me + VT (complete growth inhibition). As observed in *A. fumigatus*, OG possessed much higher antifungal potency in toxigenic *Aspergillus* strains


**Table 4 | Antifungal chemosensitization of OG (mM) or BHAM (mM) to Kre-Me (µg mL−1), tested against** *Aspergillus* **strains: summary of CLSI-based microdilution bioassays (48 h)a.**

*aMIC, Minimum inhibitory concentration; MFC, Minimum fungicidal concentration; FICI, Fractional Inhibitory Concentration Indices; FFCI, Fractional Fungicidal Concentration Indices. Synergistic FICIs and FFCIs are in bold.*

*bKre-Me was tested up to 32*μ*g mL*−*1. For calculation purpose, 64*μ*g mL*−*<sup>1</sup> (doubling of 32*μ*g mL*−*1) was used.*

*cStudent's t-test for paired data (combined, i.e., chemosensitization) was vs. mean MIC or MFC of each compound (alone, i.e., no chemosensitization) determined in strains.*

*dBHAM was tested up to 0.8 mM. For calculation purpose, 1.6 mM (doubling of 0.8 mM) was used.*

*eND, Not determined (neutral interaction).*

compared to VT (namely, 50µMOG vs. 3 mMVT, respectively, for achieving chemosensitization).

In CLSI liquid bioassay, co-application of Kre-Me with OG completely inhibited the growth of *Aspergillus* strains (100% killing in MFC testing), where independent treatment of each compound alone at the same concentrations allowed the survival of fungi (**Figures 4**, **5**). For MICs, "synergistic" FICIs (FICI ≤ 0.5; see Materials and Methods for calculations) were found between OG and Kre-Me for all *Aspergillus* strains (**Table 4**). For MFCs, "synergistic" FFCIs (FFCI ≤ 0.5) were also achieved by co-application of OG with Kre-Me for all *Aspergillus* strains tested (**Table 4**). Although there was increased antifungal activity of BHAM and Kre-Me when co-applied in *A. fumigatus* strains (WT, *sakA*-, *mpkC*-) (FICI = 0.6–0.8; **Table 4**), no calculated synergism was found in any of the *Aspergillus* strains tested with BHAM + Kre-Me (**Table 4**). As with VT + Kre-Me

(see above), *A. fumigatus* MAPK mutant also showed higher tolerance to "BHAM + Kre-Me" chemosensitization, where the value of MICCOMBINED of Kre-Me was 4µg mL−<sup>1</sup> for WT, while that for MAPK mutants was 16µg mL−1, respectively, (viz. four times higher tolerance to Kre-Me) (**Table 4**). Noteworthy is that certain fungi with mutation(s) in the MAPK signaling pathway can also escape toxicity of the commercial fungicide fludioxonil (Kojima et al., 2004).

#### **GROWTH RECOVERY TEST IN CELL WALL INTEGRITY MUTANTS OF** *S. CEREVISIAE* **TREATED WITH OG**

In sorbitol remediation bioassay, sensitivity of *slt2* and *bck1*- to OG was alleviated by sorbitol. The growth rate of *slt2* and *bck1* on sorbitol-containing media was 100–1000 times higher for caffeine or OG, respectively, compared to controls without sorbitol (**Figure 6**). Thus, the remediation by sorbitol indicates disruption of cell wall/membrane integrity in fungi is one contributing mechanism of how OG generates chemosensitization.

#### **EFFECT OF OG ON AFLATOXIN PRODUCTION**

When toxigenic strains of *A. flavus* and *A. parasiticus* were treated with OG (tested at 25, 50, 100µM), the highest reduction in aflatoxin production was observed at 25 or 50µM of OG, depending on types of strains. Aflatoxin production in OG-treated plates was reduced 4–30% compared to non-treated control plates (**Table 5**).

#### **DISCUSSION**

In this study, nine *Aspergillus* strains, namely, four toxigenic strains of *A. flavus* and *A. parasiticus*, two atoxigenic strains of *A. flavus*, and three *A. fumigatus* strains, WT and MAPK mutants (*sakA*-, *mpkC*-), were examined for their responses to different treatments. The *A. fumigatus sakA*- (*sakA* gene deletion) is an osmotic/oxidative stress sensitive mutant, while the *mpkC*- (*mpkC* gene deletion) is a mutant of the polyalcohol sugar utilization system (Xue et al., 2004; Reyes et al., 2006). In a prior study, both *sakA* and *mpkC* mutants showed higher sensitivity to certain benzo derivatives compared to the WT (Kim et al., 2011). The molecular biological/genetic resources (such as gene deletion mutants) for *A. flavus* or *A. parasiticus* available are few in number. Hence, *A. fumigatus sakA* and *mpkC* mutants could serve as model strains for investigating potential modes of antifungal responses in congeners, such as *A. flavus* or *A. parasiticus*. Except for the incapability to produce aflatoxins, the atoxigenic strains *A. flavus* 21882 and 18543 could produce KA (Table S1), a different type of secondary metabolite produced by *A. flavus* and *A. parasiticus*. Thus, the cellular secondary-product metabolism other than aflatoxin production is thought to be normal in the atoxigenic strains examined in this study.

Results showed that OG- or VT-based chemosensitization could enhance antifungal activity of Kre-Me in *Aspergillus* strains, where *Aspergillus* strains tested were sensitive to Kre-Me + OG or VT. OG was a more potent chemosensitizing agent than VT or BHAM to Kre-Me, where the concentration of OG necessary to achieve antifungal "synergism" was much lower (≥10 times lower) than the other compounds. When fungi are treated with Kre-Me, cellular AOX allows completion of electron transfer and ATP production via the MRC, thus resulting in overcoming the toxicity triggered by Kre-Me (or other MRC inhibitors) (Costade-Oliveira et al., 2012; Inoue et al., 2012). Therefore, AOX inhibitors, such as BHAM, have the effect to enhance the activity of Kre-Me (or other MRC inhibitors) when co-applied. The OGbased chemosensitization to Kre-Me, performed in this study, indicated that the AOX-inhibitory activity of OG (Sierra-Campos et al., 2009; Robles-Martinez et al., 2013) is much higher than that of the conventional AOX inhibitor, BHAM. Noteworthy is that the efficacy of OG-based chemosensitization to Kre-Me (which was "synergistic") was higher compared to the prior chemosensitization test with "2-hydroxy-5-methoxybenzaldehyde (phenolic) + antimycin A (MRC inhibitor)" (Kim et al., 2011), in which the level of compound interaction was "additive/indifferent" but not "synergistic."

The inhibition of MRC not only disrupts cellular ATP production, but also triggers oxidative stress, which results from abnormal leakage of electrons from the MRC (Fujita et al., 2004; Ruy et al., 2006). The escaped electrons cause oxidative damage to cellular components, such as cell membranes. Therefore, the enhanced oxidative stress generated by both MRC inhibitor (Kre-Me) and the redox-active phenolic derivative (OG) would result in increased growth inhibition of fungi.

Caffeine (a "control" reagent used in the cell wall/membrane integrity bioassay) tends to disorganize cell surface in fungi. Thus, fungi having abnormalities in cell surface integrity show increased sensitivity to caffeine treatment (Lussier et al., 1997). Caffeine also activates the protein kinase C (PKC) pathway, where the MAPK pathway genes *SLT2* and *BCK1* play key roles for maintaining cell wall integrity (Martin et al., 2000). Therefore, *SLT2* and *BCK1*

**FIGURE 5 | Chemosensitization test in aflatoxigenic** *Aspergillus***: Kre-Me <sup>+</sup> OG.** Exemplary MFC bioassay showing co-application of Kre-Me (0.5µg mL−1)

gene deletion mutants (*slt2* and *bck1*-) are hypersensitive to caffeine treatment, while caffeine sensitivity of *slt2* and *bck1*- could be alleviated by sorbitol (Martin et al., 2000; Ferreira et al., 2006). As shown in this study, OG also disrupted cell wall integrity, where *S. cerevisiae* MAPK pathway mutants (*slt2*-, *bck1*-) showed enhanced sensitivity to OG, while this sensitivity was remedied by sorbitol.

with OG (0.2 mM) completely inhibited the growth of aflatoxigenic *A. flavus* 3357 and 4212.

OG possessed some moderate level of antiaflatoxigenic activity, where 4–30% of reduction in aflatoxin production was achieved depending on types of toxigenic *Aspergillus* strains. In a prior study, more than 95% of inhibition in aflatoxin production could be achieved with caffeic acid, which is another type of phenolic compound (Kim et al., 2008). Modulation of the expression of antioxidant genes, such as alkyl hydroperoxide reductases (Ahp1) that detoxify organic peroxides, has been the mechanism of antiaflatoxigenic activity of caffeic acid (Kim et al., 2008). However, caffeic acid did not exhibit potent antifungal activity in the same study, indicating antimycotoxigenic activity



*Aflatoxin amount produced:* μ*g per cm2 of fungal mat. aND, Not detectable.*

of phenolic compounds is not always parallel to their antifungal activity or *vice versa*. Although OG possessed potent antifungal activity, as determined in this study, its antiaflatoxigenic capacity is not comparable to other types of phenolic agents, such as caffeic acid (e.g., reduction of aflatoxin production: 4–30% w/ OG vs. >95% w/ caffeic acid). Therefore, the intervention of either aflatoxin production or the growth of fungal pathogens could be achieved by treating or modulating different cellular targets, such as antioxidant enzymes (Ahp1) for mycotoxin control or MRC and cell wall/membrane integrity, etc., for fungal growth control, respectively.

"Compound or chemosensitization—strain specificity" exists with OG, reflecting differential susceptibility of *Aspergillus* strains to the treatments. Of note, cellular signaling system, such as histidine kinase receptors, MAPK, etc., was involved in differential susceptibility of fungi to fungicide or drugs (Ochiai et al., 2002; Kojima et al., 2004; Chapeland-Leclerc et al., 2007). For example, (1) Three histidine kinase receptors were differentially involved in drug sensitivity or stress adaptation of the opportunistic yeast *Candida lusitaniae* (Chapeland-Leclerc et al., 2007), (2) The pathogenic yeast *C. albicans* (containing multiple histidine kinase genes) were sensitive to the fungicides iprodione and fludioxonil, while the model yeast *S. cerevisiae* (having only one histidine kinase gene) showed insensitivity to these fungicides (Ochiai et al., 2002), and (3) Intact MAPK signaling system is required for the effective antifungal activity of fludioxonil, where MAPK mutants of the fungal pathogen *Colletotrichum lagenarium* exhibited increased resistance to the fungicide. Thus, it is thought that (1) different set-up of signaling components in various species of *Aspergillus*, and/or (2) modification or mutation in cellular signaling system during evolution/stress adaptation could be possible mechanisms of actions for differential susceptibility of *Aspergillus* strains to fungicides or environmental stresses (such as Kre-Me tolerance of *A. fumigatus* MAPK mutants or heat responses, as determined in this study). Determination of precise mechanism of differential susceptibility to the treatments warrants future study.

As with agricultural application, the MRC is recently considered as a new antifungal target for clinical antimycotic therapy. Examples include: (1) MRC inhibitors increased fluconazole susceptibility of both patient and laboratory isolates of *C. albicans* (Sun et al., 2013); (2) co-application of antimycin A (another MRC-inhibitory drug) and BHAM significantly enhanced antifungal potency of posaconazole and itraconazole (Shirazi and Kontoyiannis, 2013), where the antifungal activity of the drugs was determined as "fungicidal" in mucormycosiscausing *Rhizopus oryzae*; (3) inhibition of MRC of *C. parapsilosis*, causative agent for neonatal and device-related infections, enhances susceptibility of the pathogen to caspofungin, which is a cell wall-disrupting drug [(Chamilos et al., 2006) and references therein].

Application of chemosensitizing agents, such as OG, would lower the effective dosage of MRC-inhibitory drugs or fungicides (e.g., strobilurins). This approach thus, lowers negative side effects of MRC inhibitors and/or prevents fungal tolerance to strobilurins, as described in this study. Consequently, lowered dosage level of antifungal agents would render treatment or agricultural practice, such as control of aflatoxin-producing *Aspergillus* strains, less expensive and safer.

In conclusion, OG and/or VT exhibit potential to serve as safe antifungal chemosensitizers that in concert with Kre-Me greatly potentiate antifungal activity. This capacity was shown to be effective in most of the *Aspergillus* strains tested in this study. Moreover, OG was shown to also have moderate antiaflatoxigenic activity.

#### **ACKNOWLEDGMENTS**

This research was conducted under USDA-ARS CRIS Project 5325-42000-037-00D. We thank Dr. Gregory S. May at the University of Texas M.D. Anderson Cancer Center, Houston, TX for kindly providing us *A. fumigatus sakA* and *mpkC*- strains.

#### **SUPPLEMENTARY MATERIALS**

The Supplementary Material for this article can be found online at: http://www.frontiersin.org/journal/10.3389/fmicb.2014. 00087/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: 25 January 2014; paper pending published: 06 February 2014; accepted: 18 February 2014; published online: 07 March 2014.*

*Citation: Kim JH, Mahoney N, Chan KL, Campbell BC, Haff RP and Stanker LH (2014) Use of benzo analogs to enhance antimycotic activity of kresoxim methyl for control of aflatoxigenic fungal pathogens. Front. Microbiol. 5:87. doi: 10.3389/fmicb. 2014.00087*

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

*Copyright © 2014 Kim, Mahoney, Chan, Campbell, Haff and Stanker. 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: 10 April 2014 doi: 10.3389/fmicb.2014.00158

### Aflatoxins and safe storage

#### *Philippe Villers\**

*GrainPro, Inc., Concord, MA, USA*

#### *Edited by:*

*Mehdi Razzaghi-Abyaneh, Pasteur Institute of Iran, Iran*

#### *Reviewed by:*

*Theo Varzakas, Technological Educational Institute of Peloponnese, Greece David Okello Kalule, National Semi Arid Resources Research Institute, Serere, Uganda*

#### *\*Correspondence:*

*Philippe Villers, GrainPro, Inc., 200 Baker Avenue, Suite 309, Concord, MA 01742, USA e-mail: Pvillers@grainpro.com*

The paper examines both field experience and research on the prevention of the exponential growth of aflatoxins during multi-month post-harvest storage in hot, humid countries. The approach described is the application of modern safe storage methods using flexible, Ultra Hermetic™ structures that create an unbreatheable atmosphere through insect and microorganism respiration alone, without use of chemicals, fumigants, or pumps. Laboratory and field data are cited and specific examples are given describing the uses of Ultra Hermetic storage to prevent the growth of aflatoxins with their significant public health consequences. Also discussed is the presently limited quantitative information on the relative occurrence of excessive levels of aflatoxin (*>*20 ppb) before vs. after multi-month storage of such crops as maize, rice, and peanuts when under high humidity, high temperature conditions and, consequently, the need for further research to determine the frequency at which excessive aflatoxin levels are reached in the field vs. after months of post-harvest storage. The significant work being done to reduce aflatoxin levels in the field is mentioned, as well as its probable implications on postharvest storage. Also described is why, with some crops such as peanuts, using Ultra Hermetic storage may require injection of carbon dioxide, or use of an oxygen absorber as an accelerant. The case of peanuts is discussed and experimental data is described.

**Keywords: post harvest, aflatoxin, hermetic, safe storage, grain storage, Cocoons™, pesticide free, Ultra Hermetic**

#### **INTRODUCTION**

Aflatoxins (*Aspergillus flavus and Aspergillus parasiticus*) are widely recognized as a major health problem, especially in hot, humid countries. This is a particular serious problem in such crops as maize, rice, peanuts, tree nuts, and dried fruits. Aflatoxin production normally occurs in the field, particularly when stimulated by drought, stress, and high temperatures or during prolonged drying. Aflatoxin-producing molds grow exponentially in conventional multi-month storage as a result of a combination of heat and high humidity (Hell et al., 2010).

**Figure 1** (Villers et al., 2008) shows the graphic relationship between mold growth and relative humidity at equilibrium. It explains the mold growth during storage, when sustained relative humidity is beyond 65%. **Table 1** shows the optimum conditions and ranges for *Aspergillus* growth and therefore aflatoxin development. When temperatures are below 65 degrees F, and the moisture of the maize is below 12–13%, development of the fungus usually stops (Sumner and Lee, 2012).

In East Africa post-harvest losses in maize, as shown in **Figure 2**, often reach 25%, or more (World Bank Report, 2011).

#### **AFLATOXIN EFFECTS**

Dr. J. H. Williams of the University of Georgia, USA, cites a survey of local African markets showing that 40% of the commodities found there exceeded permissible aflatoxin levels (in excess of the international standard of 10–20 ppb) and that an estimated 4.5 billion people in developing countries are at risk of uncontrolled or poorly controlled exposure to aflatoxins, and up to 40% of commodities in local African markets exceed allowable levels of aflatoxins in foods (Williams, 2011).

It is known that high aflatoxin levels in the bloodstream depresses the immune system, thereby facilitating cancer, HIV, and stunting the growth of children. A cross-sectional study conducted in Ghana and cited by Dr. Williams shows that immune systems of recently HIV-infected people are significantly modified if they have above-median levels of natural exposures to aflatoxins (Williams et al., 2004). Referring to another study, Dr. Williams notes, "People with a high aflatoxin biomarker status in the Gambia and Ghana were more likely to have active malaria" (Williams, 2011).

In 2014, the Global Forum for Innovations in Agriculture (GFIA) convened a high level meeting in Abu Dhabi, UAE, on revolutionizing global agriculture through innovations. Frank Rijsberman, the CEO of the CGIAR Consortium, in his report based on a Benin study (Gong et al., 2004) on the post-weaning exposure to aflatoxin, concludes that aflatoxins have impaired growth in children and are costing African farmers over \$450 million USD per year in lost exports (Rijsberman, 2014). In 2010, 10% of the Kenyan maize crop was condemned because of excessive aflatoxin levels. One of the laboratories in Kenya that year tested 130 maize samples out of which only 47 samples had aflatoxin levels less than 10 ppb. The highest level of aflatoxin recorded in that year was 830 ppb (FAO, 2011).

Excessive aflatoxin levels also cause failure to thrive (or even death) in farm animals such as chickens and turkeys. According to Dr. Oladele Dotun, a Veterinarian at the Animal Care Laboratory

present. His evidence suggests that maximum aflatoxin contamination of maize in Guatemala occurs during the rainy season. Maize samples analyzed 20 days after harvest had levels of 130µg aflatoxin per kg of total maize. The same samples analyzed 60 days later showed rapid increase, up to 1 680µg per kg (FAO, 1992).

**FIGURE 2 | 2011 World Bank Report on post-harvest losses of maize in East Africa.**

**Table 1 | Conditions favoring** *Aspergillu***s** *flavus* **development.**


*Table shows the optimum conditions for Aspergillus growth and aflatoxin development. When temperatures are below 65*◦*F and the moisture of the corn is below 12–13%, development of the fungus usually stops.*

in Nigeria, research has shown that aflatoxins cause infertility, abortions, and delayed onset of egg production in birds as well as sudden losses in egg production in actively laying birds. Furthermore, loss of appetite, skin discoloration, or even yellowish pigmentation on skin can be observed in fish (Oladele, 2014).

An increasing amount of scientific research has been devoted to learning more about aflatoxin growth problems and possible solutions, including using genetically modified or hybridized seeds formulated for mold resistance or through use of products such as AflaSafe, now used in Africa. AflaSafe's "biological approach" uses a closely related, non-aflatoxin-producing mold to out-compete the aflatoxin-producing molds. In temperate climates, aflatoxin problems have been controlled largely with ventilation during cooler nights and through lower winter temperatures (CAES, 2014).

A neglected part of the aflatoxin problem has been the rapid growth of post-harvest aflatoxin levels when stored for prolonged periods in the conditions of high temperature and relative humidity above 65% found commonly in all tropical climates (See **Figure 1**). In the tropics, field studies by Icrisat in Mali show that conventional storage for more than 2 months causes rapid growth of aflatoxins (Waliyar et al., 2002) (**Figure 3**).

In Guatemala, a field survey of maize sold in rural markets by Martinez-Herrera in 1968 found that considerable contamination by several fungi, including *Aspergillus* species were frequently

This data, as well as data from several other studies, strongly indicate the need to dry maize before storage. Drying is an important step in ensuring good quality grain that is free of fungi and microorganisms and that has desirable quality characteristics for marketing and final use. Diverse drying systems and equipment are available that use various sources of energy, including solar energy. Choice of system requires consideration of a number of factors: temperature and air velocity, rate of drying, drying efficiencies, kernel quality, air power, fuel source, fixed costs, and management (FAO, 1992).

An interesting question is what happens to aflatoxin growth in Ultra Hermetic Storage™ following the use of biological controls such as AflaSafe, which introduces non-aflatoxin-producing molds to compete with aflatoxin-producing molds. However, biological competition between mold populations alone will not totally eliminate aflatoxin mold prior to storage. We believe that storage under tropical conditions in classical containers such as silos, bags, and tarps will continue to produce the exponential aflatoxin growth shown in **Figure 1**, probably from a lower starting population.

#### **CHEMICAL APPROACHES TO POST-HARVEST CONTROL**

Many chemical and additive methods for controlling aflatoxin growth have been proposed, but used only to a limited extent. For instance, as noted by Lunven at FAO, "Chakrabarti showed that aflatoxin levels could be reduced to less than 20 ppb using separate treatments with 3% hydrogen peroxide, 75% methanol, 5% dimethylamine hydrochloride, or 3% perchloric acid. These treatments, however, induced losses in weight and also in protein and lipids. Other methods include the use of carbon dioxide plus potassium sorbate and the use of sulphur oxide." Lunven also noted another process that had received some attention, namely the use of calcium hydroxide, a chemical used for lime cooking of maize. Regarding this process, Lunven stated, "Studies have shown a significant reduction in aflatoxin levels, although the extent of reduction is related to the initial levels. Feeding tests with moldy maize treated with calcium hydroxide have shown a partial restoration of its nutritional value" (FAO, 1992).

However, to date, no chemical or additive method has gained general acceptance. Concerns about using chemicals or additives is still growing, encouraging the expanded use of methods that have no contamination potential.

#### **SAFE STORAGE USING ULTRA HERMETIC AIRTIGHT CONTAINERS**

Ultra Hermetic Storage™ is in use now at a varying scale in 103 countries. It has been shown that WHEN sufficiently airtight, it drastically inhibit mold growth and hence aflatoxin expression. Since molds need oxygen and high humidity, the natural respiration of the contained insects and other microorganisms, and sometimes respiration of the commodity itself, will use up the available oxygen faster than any residue leakage. After 10 days to 2 weeks at room temperature or above, this results in an unbreatheable atmosphere–typically 3% oxygen and 15% carbon dioxide. This does not kill the molds, but as shown in **Table 2**, in the case of peanuts, the unbreatheable atmosphere arrests mold development of Colony Forming Units (CFUs) so that even after several months, the levels of mold products, and therefore aflatoxins do not rise significantly. The same table shows inhibition of Free Fatty Acids (FFAs), which cause rancidity (Navarro et al., 2012).

Martin Gummert, post-harvest rice specialist at the International Rice Research Institute (IRRI), wrote about storage of rice and rice seed, "Hermetic Storage of rice is becoming increasingly popular across Asia, and for good reason—as well as being transportable, it is better than air-conditioned storage and almost as good as a cold room, at a fraction of the cost" (Villers and Gummert, 2009).

#### **THE SPECIAL CASE OF PEANUTS**

Peanuts often are often contaminated with aflatoxins before storage, but (unlike most grains) may take as long as 30 days or so to reach a 3% oxygen level in Ultra Hermetic storage (Navarro et al., 2012). This is too long a period to prevent major increase of aflatoxin levels. For this reason, one of two forms of accelerant is used. For field operation and portable bags, an oxygen-absorbing sachet weighing 65 g per 69 kg capacity is sufficient. For large storage units or volume production, injecting carbon dioxide to drive out the air is adequate. **Table 2** shows mold growth densities measured in CFUs after 90 days (Navarro et al., 2012). The table shows two orders of magnitude difference in mold density for


#### **Table 2 | FFA and CFU mold levels in peanuts.**

crops stored in conventional (the Control) vs. hermetic storage, using carbon dioxide injection. Hermetic storage, even without injected carbon dioxide, still shows a five-fold improvement vs. the Control. Unfortunately, in this particular study, aflatoxinproducing molds turned out to be largely absent in the mold mix, and therefore the data of **Table 2** does not allow direct measurement of the growth of aflatoxin, as distinct from all molds (CFU). **Table 2** shows the significant improvement with hermetic storage alone and even greater improvement with the early addition of CO2 in controlling the growth of rancidity-producing FFAs.

#### **AN UNSOLVED QUESTION**

Needing further study is the question of exactly how frequently aflatoxin levels exceed the international standard of 10–20 ppb at initial storage time vs. after long-term storage. The unanswered question is how frequently initial tolerable levels of aflatoxins become intolerable after improper drying and or improper postharvest storage. We believe it is quite likely, that the levels of aflatoxin in grains registering above 10–20 ppb in peanuts before storage may well be relatively infrequent but only exceed the international safety standard after long-term storage. The absence of a definitive answer to this question may well contribute to scientific attention still being devoted almost entirely to the problem of controlling aflatoxin in the field. To date, the demonstrated use of Ultra Hermetic storage, although widespread and available, has not yet resulted in much scientific work being focused on existing ways of preventing aflatoxin growth post-harvest and in longer-term storage. We believe such study is overdue.

#### **AVAILABLE FORMS OF ULTRA HERMETIC AIRTIGHT STORAGE**

Ultra-Hermetic™ (airtight) enclosures without pesticides or fumigants use the respiration of infesting insects and other microflora as their "engines" to produce an unbreatheable atmosphere. Available sizes of such airtight plastic enclosures vary over a wide range depending on the application. They range from the patented man-portable SuperGrainbags™, made from thin (0.078 mm) coextruded, multi-layer plastic with capacities from 25 kg to 2.8 tons (**Figure 4**), to flood- and tsunami-resistant Cocoons™ made from 0.8 mm PVC with capacities of 5- to 1000-tonnes (**Figure 5**). Their massive air and watertight bulk and geometry help explain their survival when full in several Philippine typhoons and in one flood.

Cocoons are also highly rodent-resistant, as a result of intentionally stretching their tough exterior surface to prevent rodents from gaining a tooth hold. Cocoons can be erected without specialized tools or equipment and can be rapidly installed either directly on the ground or on pavement, either indoors or outdoors, and have a 10–15 year useful life.

The SuperGrainbag, man-portable Ultra Hermetic storage shown in **Figure 4** has permeability to oxygen of only 3 cc/m2/per day, 500 times more airtight than ordinary plastics such as polyethylene. This results from the use of a proprietary barrier layer inside its 3-layer, coextruded design.

**Figure 5** shows 100-tonne Cocoons enclosing bags of grains stacked on the bottom of the Cocoon. After the bags are stacked, the Cocoon is closed by mounting its upper section over the pile.

**FIGURE 4 | Nepalese woman with SuperGrainbags, Mulpani Village near Kathmandu, Nepal.**

Then the two sections are zipped together with an airtight zipper, originally designed for astronauts.

Other designs, such as the 1- to 2-tonne capacity GrainSafe™ shown in **Figure 6**, are used for storing bulk grain, and are designed for continuous in-and-out access while maintaining hermeticity and avoiding empty headspace when partly empty.

The largest existing Ultra Hermetic storage units are known as Hermetic Bunkers™ and store 10,000- to 20,000-tonnes (**Figure 7**) (Navarro et al., 1993). They are designed for multiyear strategic storage. Hermetic Bunkers are used for up to 5-year storage of barley, maize or wheat, such as in Cyprus, Jordan, and Israel (Navarro et al., 1993).

**FIGURE 6 | GrainSafe in Guatemala storing maize.**

**FIGURE 7 | Bunker in Jordan storing wheat.**

Rice and other grain seeds now are among the leading commodities stored in Ultra Hermetic enclosures instead of air-conditioned or refrigerated facilities (Villers and Gummert, 2009).

#### **SAFE SOLAR DRYING**

All forms of longer-term storage historically have required proper drying before storage. Various strategies have been used, but in most developing countries the capital and consumption costs of using non-renewable fuels are unaffordable, and so solar drying is used. However, much solar drying is done on paved patios, paved roads, or even dirt roads, with no protection from rewetting when it rains. The Collapsible Solar Dryer (CDC)™ (**Figure 8**) has the special property of being manually reclosed on itself when it rains, creating a rainproof cover like a suitcase lid that keeps the seeds or grains dry until the rain stops.

#### **CONCLUSION**

Programs to control aflatoxin-producing molds in the field through improved seeds or biological control have yielded encouraging results, but they do not protect commodities from exponential aflatoxin growth in post-harvest storage under hot, humid conditions, which in many cases may be the dominant growth factor. Ultra Hermetic storage is an available, important

answer to keeping aflatoxin levels at (or below) the International Standards of a maximum of 10- to 20-ppm, now so often exceeded in hot, humid climates. It is already widely proven in the field but not yet utilized on a sufficient scale. To date, the importance of post-harvest protection against aflatoxin growth remains not widely understood by the scientific community. Existing literature on aflatoxin control focuses on means of reducing aflatoxin levels during the growing period, and more work is needed to prove the full public health significance of aflatoxin growth during conventional drying and in non-hermetic, post-harvest storage vs. levels of aflatoxin already existing by harvest time.

#### **REFERENCES**


exposure, potential health consequences, and interventions. *Am. J. Clin. Nutr.* 80, 1106–1122.

**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: 03 February 2014; accepted: 25 March 2014; published online: 10 April 2014. Citation: Villers P (2014) Aflatoxins and safe storage. Front. Microbiol. 5:158. doi: 10.3389/fmicb.2014.00158*

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

*Copyright © 2014 Villers. 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.*

### ADVANTAGES OF PUBLISHING IN FRONTIERS

FAST PUBLICATION Average 90 days from submission to publication

COLLABORATIVE PEER-REVIEW

Designed to be rigorous – yet also collaborative, fair and constructive

RESEARCH NETWORK Our network increases readership for your article

#### OPEN ACCESS

Articles are free to read, for greatest visibility

#### TRANSPARENT

Editors and reviewers acknowledged by name on published articles

GLOBAL SPREAD Six million monthly page views worldwide

#### COPYRIGHT TO AUTHORS

No limit to article distribution and re-use

IMPACT METRICS Advanced metrics track your article's impact

SUPPORT By our Swiss-based editorial team

EPFL Innovation Park · Building I · 1015 Lausanne · Switzerland T +41 21 510 17 00 · info@frontiersin.org · frontiersin.org