Edited by: Ophry Pines, Hebrew University of Jerusalem, Israel
Reviewed by: Paolo De Los Rios, Ecole Polytechnique Fédérale de Lausanne, Switzerland; Eilika Weber-Ban, ETH Zurich, Switzerland
*Correspondence: Stefan G. D. Rüdiger, Cellular Protein Chemistry, Bijvoet Center for Biomolecular Research, Utrecht University, Padualaan 8, 3584 CH Utrecht, Netherlands
This article was submitted to Protein Folding, Misfolding and Degradation, a section of the journal Frontiers in Molecular Biosciences
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
The composition of protein surfaces determines both affinity and specificity of protein-protein interactions. Matching of hydrophobic contacts and charged groups on both sites of the interface are crucial to ensure specificity. Here, we propose a highlighting scheme, YRB, which highlights both hydrophobicity and charge in protein structures. YRB highlighting visualizes hydrophobicity by highlighting all carbon atoms that are not bound to nitrogen and oxygen atoms. The charged oxygens of glutamate and aspartate are highlighted red and the charged nitrogens of arginine and lysine are highlighted blue. For a set of representative examples, we demonstrate that YRB highlighting intuitively visualizes segments on protein surfaces that contribute to specificity in protein-protein interfaces, including Hsp90/co-chaperone complexes, the SNARE complex and a transmembrane domain. We provide YRB highlighting in form of a script that runs using the software PyMOL.
Protein-protein interactions underlie all processes in the cell. Specificity of protein-protein interactions is determined by matching of complementary functional groups with those of the opposite surface (Chothia and Janin,
Assessment of major determinants for protein-protein interactions typically requires a combination of the aforementioned color schemes. It would be helpful for assessing protein interaction surfaces, e.g., for planning mutations, to highlight the major determinants for protein interactions in a single image. Such a color scheme should also indicate protein properties at atomic level, to account for differences in chemical properties within side chains.
Here, we present a scheme to color proteins by charge and hydrophobicity at atomic level. The YRB scheme specifically highlights all carbon atoms that have high potential to form hydrophobic interactions in yellow. Simultaneously, nitrogen atoms in the side chain of arginine and lysine are colored blue and oxygen atoms in the side chains of glutamate and aspartate are red. This visualizes complementarity in functional groups, such as hydrophobic and charged groups, as we illustrate in interaction of the Hsp90 chaperone with its partner proteins and within the interface of the SNARE complex. The YRB scheme can be used universally for a protein of interest when loaded into the visualization platform PyMOL. The script facilitates a quick assessment of protein surfaces and visualizes specificity in protein-protein interfaces.
To apply YRB highlighting, we used PyMOL [The PyMOL Molecular Graphics System, Delano Scientific, San Carlos, CA, USA, version 1.4 (Mac OS X) and version 0.99 (Windows)]. PyMOL is a widely used open-source molecular visualization platform, which can be extended by plugins and scripts written in Python (Python Software Foundation. Python Language Reference, version 2.7, released on July 3rd 2010), available at
The YRB script colors structural models of proteins at atomic level. Carbon atoms not bound to nitrogen or oxygen atoms are colored yellow, oxygens carrying the negative charges in glutamate and aspartate red and nitrogens carrying the positive charges in lysine and arginine blue, while all remaining atoms are white (Figure
The YRB script is loaded into PyMOL by selecting “
When the script is run in PyMOL, hydrogen atoms are removed and the exact colors are defined with RGB color values. A hash table is used to map amino acids to a list of ordered pairs of atoms with their color according to their position in the side chain and which atoms they are bound to (Figure
The script considers all molecules present in PyMOL individually. If the designation of the structure matches the given command, the script proceeds to highlighting. When no specifications are made by the user then all molecules will be highlighted.
First, the backbone atoms N, C, CA, and O of all amino acids are colored white. CB atoms that are not bound to nitrogen or oxygen atoms are colored yellow, while the oxygen-bound CB of serine and threonine remain white. Then the remaining atoms of the amino acid side chains are colored. The side chains are considered one at time when the script runs through the mapping and every single atom of the side chains is colored accordingly. For example, the CG atom of arginine is bound to carbon atoms and becomes yellow. The nitrogen atoms, NE, NH1 and NH2 of arginine are colored blue. This results in the molecule being colored in yellow, red, blue, and white at atomic resolution (Figure
All alternative color schemes were applied using PyMOL. The electrostatic potential maps were applied on protein surface by the option “
Install both PyMOL and Python on your computer (
Save “The YRB script” at a desirable location, either with .py (Windows) or .pym (MacOS X) extension.
Find the pdb file of interest in the Protein Data Bank.
Open the pdb file into PyMOL or type “
The surface representation is created by either typing “hide everything; show surface” or clicking on “s: show as surface” in the right-hand panel.
The YRB scheme is loaded onto the surface by clicking in the menu: “File: Run.”
Open the YRB file from the location where it was saved before.
The YRB script is applied by typing: “yrb.”
Additional steps are needed to display the interaction interface in YRB colors. Amino acids that are not involved in the interaction can be colored gray.
(9) Make the amino acid sequence visible by clicking on “
(10) Select the amino acids that are not involved in the interaction in the sequence window above the protein.
Color this selection
(11) Next to the selection “
We set out to use a highlighting scheme that combines several features that facilitate assessment of protein interaction interfaces considering the following features: (i) It should highlight surface properties at atomic level (ii) It should visualize hydrocarbon groups (CHn) with non-polar substitutions but not those with polar ones. (iii) It should visualize the charged groups in interaction interfaces. (iv) It should be limited to using primary colors to effectively highlight surface properties in an intuitive way.
To visualize hydrophobic and charge contributions to protein interfaces we use the following highlighting scheme at atomic level using the colors yellow, red, and blue (YRB). In this scheme, all carbon atoms not bound to nitrogen and oxygen atoms are highlighted in yellow, nitrogen atoms in the side chains of lysine and arginine are blue, oxygen atoms in the side chains of glutamate and aspartate are red and all remaining atoms white (Figure
This color scheme reflects the key points that are relevant for intuitive assessment of protein-protein interactions: (i) Amino acids are composed of atoms with different properties. (ii) CHn groups that are not bound to electronegative oxygen or nitrogen have a high potential for hydrophobic interactions. (iii) Charged pairs play a key role in protein-protein interactions. (iv) The YRB scheme combines hydrophobic and charged functional groups, which reflects specificity in protein-protein interfaces.
The benefit of highlighting at atomic level is particularly evident for residues with polar or charged groups. The side chains of polar and charged amino acids contain non-polar segments, which may contribute to the hydrophobic interface. They are, however, unable to engage in coulomb and polar-driven interactions, as would be suggested if the protein is highlighted at residue level. Therefore, YRB colors e.g., the β-carbon of arginine yellow and the nitrogen atoms of its positively charged head group blue (Figure
To test the potential of YRB highlighting in the understanding of protein-protein interfaces, we analyzed protein complexes with known crystal structures. A particularly suitable system is the ATP-dependent molecular chaperone Hsp90, which interacts with a plethora of co-factors that regulate its functional cycle (Pearl and Prodromou,
Cdc37 interacts with the N-terminal domain of Hsp90 (Hsp90-N), preventing its dimerisation (Roe et al.,
Now, we set out to investigate the interface of Hsp90-p23 to analyze whether hydrophobic interactions in combination with charged pairs are also represented by YRB. The interaction of the co-chaperone p23 with Hsp90 has been revealed in a crystal structure of the yeast homologs of a complex consisting of the folded part of p23 and all three domains of Hsp90 (Supplementary Table
To test whether YRB also represents an interface that contains many charged pairs, we colored Hsp90-Aha1 interface according to the YRB scheme. The cochaperone Aha1 activates the ATPase activity by interacting with the middle domain of Hsp90 and partially overlaps with the p23 interface in Hsp90. The core of the Hsp90-Aha1 interface primarily consists of a hydrophobic patch supported by seven charged pairs (Supplementary Table
We now set out to test the YRB scheme on a highly hydrophobic interface. A system in which hydrophobic interactions play a key role in complex formation is the SNARE family (Sutton et al.,
Next, we analyzed how pockets for small molecules are represented by YRB highlighting. Therefore, we analyzed the nucleotide binding pocket in Hsp90-N (Prodromou et al.,
Some of the most extensive hydrophobic surfaces are within membrane proteins. As membrane proteins need to interact with the lipid bilayer, they contain extended hydrophobic membrane domains. Previously, we have shown that YRB highlights hydrophobic contributions to surfaces. Therefore, we highlighted the NADH:quinone oxidoreductase complex in YRB. Na+-NQR is a membrane complex and facilitates Na+ translocation across the membrane (Steuber et al.,
The YRB scheme combines three features that facilitate assessment of protein interaction interfaces: (i) YRB highlights surface properties at atomic level. (ii) YRB only colors the hydrocarbons with non-polar substitutions hydrophobic and distinguishes between hydrocarbons groups with polar and non-polar substitutions. (iii) YRB visualizes charged groups in interaction interfaces. As mentioned in the introduction there are alternative coloring schemes that represent properties of proteins. How does the YRB scheme compare to other highlighting schemes, in particular CPK coloring, the hydrophobicity scale and the electrostatic potential?
The CPK scheme colors per atom type, whereas the YRB scheme colors atoms based on functional properties (Corey and Pauling,
Although CPK and YRB both color at atomic level, they may differ in the potential to reveal specifics of protein interfaces. Therefore, we compared the Hsp90-p23 interface colored in YRB and CPK. CPK colors all carbons and therefore it also reveals the hydrophobic contributions of the Hsp90-p23 interface as in YRB. Only YRB distinguishes between hydrocarbons with polar and non-polar substitutions, in other words YRB does not color the carbons in the polar backbone as hydrophobic, among others (Figures
We compared YRB and hydrophobicity highlighting (Eisenberg et al.,
As mentioned above, YRB revealed the hydrophobic membrane domain of the NADH:quinone oxidoreductase complex and a charged ring at the cytoplasmic side of the transmembrane domain. Both highlighting schemes faithfully represent the hydrophobicity of the transmembrane segment (Figures
The electrostatic potential map is a widely used coloring scheme that colors protein surface based on their overall charge distribution. As YRB colors the charged atoms separately, we wanted to know how YRB compares to the electrostatic potential map of the Hsp90-p23 interface. Therefore, we highlighted the Hsp90-p23 interface in the electrostatic potential map (Figure
YRB highlighting revealed six charged pairs of the SNARE interface. Does the electrostatic potential maps also display these charged pairs of the SNARE interface? The electrostatic potential map of the SNARE complex displays the overall charge distribution at the edge of the interface between v-SNARE and t-SNAREs (Figure
YRB highlighting is based on properties of functional groups of amino acids. It simultaneously highlights two determinants of specificity in protein-protein interaction at atomic level, hydrophobicity and charge. For assessing hydrophobicity, we highlight all carbon atoms not bound to nitrogen or oxygen atoms. Simultaneously, we color all nitrogens and oxygens of the charged groups of arginine, lysine, aspartate, and glutamate. Here, we have shown that YRB reveals the two specificity determinants, hydrophobicity and charges, by the interfaces of Hsp90/co-chaperones and the SNARE complex.
Highlighting of hydrophobic and charged atoms reflect that those residues are particularly important in protein-protein interfaces. Binding hot spots in general are enriched in tryptophan, tyrosine and arginine (Bogan and Thorn,
We should stress that the YRB scheme has intrinsic limitations that need to be taken into account when assessing protein-protein interactions. YRB visualizes structural models, and the quality of the representation depends on eventual limitations of the structural model. In particular, proteins are dynamic and may undergo local and even global conformational changes. Surface-exposed amino acids have a certain degree of freedom in solution, which is not represented in a static YRB colored protein model. When it comes to charges, YRB identifies salt bridges in protein complexes, but for assessing long-range electrostatic interactions the overall electrostatic potential map will in many cases be more suitable. We encourage any interested users to highlight protein models using different schemes, depending on the problem to be addressed.
Large hydrophobic and aromatic side chains also play an important role for the recognition by molecular chaperones, together with positively charged residues (Knoblauch et al.,
We have shown that YRB visualizes complementarity of functional groups, such as hydrophobic and charged interactions, with a set of examples that represent different types of interaction including Hsp90/co-chaperone interfaces, the SNARE complex and Hsp90 nucleotide binding. The interfaces in YRB display matching of complementary functional groups between opposite surfaces, visualizing key specificity determinants (Figures
YRB highlighting may be useful to identify binding interfaces that are determined by hydrophobic interactions, such as those of molecular chaperones and SNARE proteins. It provides an intuitive assessment of the potential of protein surfaces to interact with binding partners. YRB runs as a script in PyMOL and does not require programming experience.
DH and IAEMvB wrote the YRB script. DH has analyzed the potential of the YRB script by analysing the interfaces mentioned in the article. SGDR conceived the work. DH, SGDR, and IAEMvB have contributed to writing the article. DH made all figures and tables. TML has contributed to both supervision of the project and discussions and has provided comments on the manuscript.
SGDR was supported by Marie-Curie Actions of the 7th Framework programme of the EU [Innovative Doctoral Programme “ManiFold” (No. 317371) and Initial Training Network “WntsApp” (No. 608180)] and by the Internationale Stichting Alzheimer Onderzoek (ISAO; project “Chaperoning Tau Aggregation”; No. 14542).
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.
We are grateful to Ineke Braakman for continuous support. We thank Magdalena Wawrzyniuk for initial work on highlighting protein models.
The Supplementary Material for this article can be found online at: