Hostname: page-component-cd9895bd7-p9bg8 Total loading time: 0 Render date: 2024-12-23T02:43:07.868Z Has data issue: false hasContentIssue false

Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a see-saw effect in the conformational space

Published online by Cambridge University Press:  27 April 2023

Vishal Annasaheb Adhav
Affiliation:
Department of Biology, Indian Institute of Science Education and Research, Pune, India
Sanket Satish Shelke
Affiliation:
Department of Biology, Indian Institute of Science Education and Research, Pune, India
Pananghat Balanarayan
Affiliation:
Department of Chemical Sciences, Indian Institute of Science Education and Research, Mohali, India
Kayarat Saikrishnan*
Affiliation:
Department of Biology, Indian Institute of Science Education and Research, Pune, India
*
Corresponding author: Kayarat Saikrishnan; Email: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

Divalent sulfur (S) forms a chalcogen bond (Ch-bond) via its σ-holes and a hydrogen bond (H-bond) via its lone pairs. The relevance of these interactions and their interplay for protein structure and function is unclear. Based on the analyses of the crystal structures of small organic/organometallic molecules and proteins and their molecular electrostatic surface potential, we show that the reciprocity of the substituent-dependent strength of the σ-holes and lone pairs correlates with the formation of either Ch-bond or H-bond. In proteins, cystines preferentially form Ch-bonds, metal-chelated cysteines form H-bonds, while methionines form either of them with comparable frequencies. This has implications for the positioning of these residues and their role in protein structure and function. Computational analyses reveal that the S-mediated interactions stabilise protein secondary structures by mechanisms such as helix capping and protecting free β-sheet edges by negative design. The study highlights the importance of S-mediated Ch-bond and H-bond for understanding protein folding and function, the development of improved strategies for protein/peptide structure prediction and design and structure-based drug discovery.

Type
Research Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2023. Published by Cambridge University Press

Introduction

Non-covalent interactions are fundamental for protein folding, structural stability and function. Traditionally, hydrogen bonds (H-bonds), hydrophobic effects and electrostatic and van der Waals interactions are assumed to be the major drivers of protein folding and stability (Dill and MacCallum, Reference Dill and MacCallum2012; Pace et al., Reference Pace, Fu, Fryar, Landua, Trevino, Schell, Thurlkill, Imura, Scholtz, Gajiwala, Sevcik, Urbanikova, Myers, Takano, Hebert, Shirley and Grimsley2014). However, the essentiality of other weak interactions in sculpting protein structures is also being discovered. For example, the importance of weak H-bonds, cation/anion-π, n → π* and dihydrogen bonding (H···H) interactions for the stability of protein structures is well understood (Derewenda et al., Reference Derewenda, Lee and Derewenda1995; Gallivan and Dougherty, Reference Gallivan and Dougherty1999; Manikandan and Ramakumar, Reference Manikandan and Ramakumar2004; Matta et al., Reference Matta, Hernández-Trujillo, Tang and Bader2003; Bartlett et al., Reference Bartlett, Choudhary, Raines and Woolfson2010; Lucas et al., Reference Lucas, Bauzá, Frontera and Quiñonero2016; Forbes et al., Reference Forbes, Sinha, Ganguly, Bai, Yap, Patel and Zondlo2017; Newberry and Raines, Reference Newberry and Raines2019; Juanes et al., Reference Juanes, Saragi, Jin, Zingsheim, Schlemmer and Lesarri2020). Apart from C, O, N and H that form these non-covalent interactions, divalent sulfur (S) is present in methionines (Met-Sδ), cysteines (Cys-Sγ) and cystines of proteins. S has unique bonding properties that allow it to interact with both electrophiles and nucleophiles (Rosenfield et al., Reference Rosenfield, Parthasarathy and Dunitz1977; Guru Row and Parthasarathy, Reference Guru Row and Parthasarathy1981). Consequently, methionine, cystine and cysteine are expected to contribute to distinct polar interactions in proteins, making it imperative to study them to better understand their structural properties and folding.

Although the polar bonding properties of S have long been known (Fig. 1a) (Rosenfield et al., Reference Rosenfield, Parthasarathy and Dunitz1977; Guru Row and Parthasarathy, Reference Guru Row and Parthasarathy1981), it has gained prominence in the field of organic molecules over the last decade (Andersen et al., Reference Andersen, Jensen, Mackeprang, Du, Jørgensen and Kjaergaard2014; Beno et al., Reference Beno, Yeung, Bartberger, Pennington and Meanwell2015; Rao Mundlapati et al., Reference Rao Mundlapati, Ghosh, Bhattacherjee, Tiwari and Biswal2015; Pascoe et al., Reference Pascoe, Ling and Cockroft2017; Wang and Fujii, Reference Wang and Fujii2017; Motherwell et al., Reference Motherwell, Moreno, Pavlakos, Arendorf, Arif, Tizzard, Coles and Aliev2018; Riwar et al., Reference Riwar, Trapp, Root, Zenobi and Diederich2018; Scilabra et al., Reference Scilabra, Terraneo and Resnati2019; Haberhauer and Gleiter, Reference Haberhauer and Gleiter2020; Kolb et al., Reference Kolb, Oliver and Werz2020; Wang et al., Reference Wang, Zhu, Feng, Yu, Hao, Zhu and Wang2020). In contrast, these properties of S are often overlooked in the study of proteins. Many standard textbooks of biochemistry categorise methionine as a non-polar hydrophobic amino acid (Lehninger et al., Reference Lehninger, Nelson and Cox2000; Berg et al., Reference Berg, Tymoczko and Stryer2002). However, the lone pairs of S can interact with electrophiles, such as H atoms, to form H-bonds with donor O or N (Fig. 1a) (Zhou et al., Reference Zhou, Tian, Lv and Shang2009; Rao Mundlapati et al., Reference Rao Mundlapati, Ghosh, Bhattacherjee, Tiwari and Biswal2015; Chand et al., Reference Chand, Sahoo, Rana, Jena and Biswal2020). Although S is less electronegative than O/N, the strength of S-mediated H-bonds such as N–H···S is comparable to the conventional N–H···O bond mainly because of its larger size, higher polarisability and the diffuse electron cloud (Gregoret et al., Reference Gregoret, Rader, Fletterick and Cohen1991; Rao Mundlapati et al., Reference Rao Mundlapati, Ghosh, Bhattacherjee, Tiwari and Biswal2015). Additionally, the divalent S has two electropositive regions on the extension of its two covalent bonds, referred to as σ-holes, which can interact with various nucleophiles (Fig. 1a) (Murray et al., Reference Murray, Lane, Clark, Riley and Politzer2012; Politzer et al., Reference Politzer, Murray and Clark2014; Politzer et al., Reference Politzer, Murray, Clark and Resnati2017). The interaction made by a σ-hole of S with a nucleophile is categorised as a chalcogen bond (Ch-bond) (Aakeroy et al., Reference Aakeroy, Bryce, Desiraju, Frontera, Legon, Nicotra, Rissanen, Scheiner, Terraneo, Metrangolo and Resnati2019).

Figure 1. An approach of electrophiles and nucleophiles towards divalent S. (a) Approach of an electrophile or a nucleophile (upper panel) and a covalently linked electrophile–nucleophile fragment (lower panel) towards S. (b) Fragments analysed using the Cambridge Structural Database and the Protein Data Bank (PDB). *Fragments are from protein structures in the PDB.

Ch-bonds have been shown to play important roles in the self-assembly and catalysis of organic molecules (Iwaoka et al., Reference Iwaoka, Takemoto and Tomoda2002; Benz et al., Reference Benz, López-Andarias, Mareda, Sakai and Matile2017; Mahmudov et al., Reference Mahmudov, Kopylovich, Guedes Da Silva and Pombeiro2017; Chen et al., Reference Chen, Xiang, Zhao and Yan2018; Lim and Beer, Reference Lim and Beer2018; Vogel et al., Reference Vogel, Wonner and Huber2019; Carugo et al., Reference Carugo, Resnati and Metrangolo2021; Jena et al., Reference Jena, Dutta, Tulsiyan, Sahu, Choudhury and Biswal2022). Ch-bonds have a specific directionality, and a recent spectroscopic study on thiophenes has shown that the bond can be as strong as a conventional H-bond (Pascoe et al., Reference Pascoe, Ling and Cockroft2017). However, unlike the latter, the strength of a Ch-bond is independent of solvent polarity (Pascoe et al., Reference Pascoe, Ling and Cockroft2017). The occurrence of a divalent S-mediated Ch-bond in proteins has been documented previously and hypothesised to be functionally significant (Pal and Chakrabarti, Reference Pal and Chakrabarti2001; Iwaoka et al., Reference Iwaoka, Takemoto and Tomoda2002; Iwaoka and Isozumi, Reference Iwaoka and Isozumi2012, Reference Iwaoka and Isozumi2006; Iwaoka and Babe, Reference Iwaoka and Babe2015). Despite its potential importance, the precise role of the Ch-bond in protein structures and its effect on protein stability have remained unaddressed.

Many functional groups in proteins contain both electrophilic and nucleophilic centres with which S can interact to form H- or Ch-bonds (Fig. 1a). Hence, the direction of approach of the functional groups with respect to S and the nature of the bond formed are interlinked and could influence protein conformation. This prompted us also to ask whether S forms an H-bond or a Ch-bond with a functional group having both electrophilic and nucleophilic centres and what determines the choice between them (Fig. 1a). Here, we have addressed the above questions through extensive computational, cheminformatics and bioinformatics analyses. The study reveals the importance of Ch-bonds in the stability of protein structures. Using potential energy surface (PES) scan and atoms in molecules (AIM) analysis, we show that both H- and Ch-bonds can augment the stability of protein secondary structures, and through bioinformatics analyses, we identify some of the mechanisms of fold stabilisation. Analysis of the structures in the Cambridge Structural Database (CSD) and the Protein Data Bank (PDB), and molecular electrostatic surface potential (MESP) calculations reveal the factors that influence the choice between H-bond and Ch-bond by S. The study, therefore, unravels the underappreciated role of S-mediated interactions, particularly Ch-bonds, in protein structure, stability and function.

Methods

Computational methods

All monomer and dimeric complexes used in this study were optimised at the M06/6-311++G(3DF,3PD) level using the Gaussian09 programme (Zhao and Truhlar, Reference Zhao and Truhlar2008; Frisch et al., Reference Frisch, Trucks, Schlegel, Scuseria, Robb, Cheeseman, Scalmani, Barone, Mennucci, Petersson, Nakatsuji, Caricato, Li, Hratchian, Izmaylov, Bloino, Zheng and Sonnenber2009). The positive eigenvalues of the Hessian evaluation confirmed that these optimised complexes are minima on the PES. The (CH3)2S:OH2 and (CH3)2S:NH3 complexes were used to investigate energetically favourable regions for S···H–O/N interaction. Similar investigations were performed on Cl(CH3)S:OH2 for S···O and Cl(CH3)S:NH3 for S···N interactions. To carry out a spherical energy scan using these complexes, d (distance between S and H/O/N) was kept constant, whereas θ (the angle between the centroid (c) of a triangle defined by C–S–C/Cl, S and H/O/N) and δ (the torsion angle between C, c, S and H/O/N) were varied in steps of 2˚. δ was varied from 90˚ to −90˚ and θ from 60˚ to 178˚. Complexation energies (ΔEs) for different θ and δ values without basis set superposition error correction were calculated for all these complexes (Hobza and Řezáč, Reference Hobza and Řezáč2016). MESP topographical analyses characterising the strength of lone pairs (V min) of divalent S were carried out using the rapid topography mapping implemented in DAMQT (Yeole et al., Reference Yeole, López and Gadre2012; Kumar et al., Reference Kumar, Yeole, Gadre, Lõpez, Rico, Ramírez, Ema and Zorrilla2015). Texturing of MESP of molecules at defined density surface (V S,max) was carried out using Gaussview 5.0 (Dennington et al., Reference Dennington, Keith and Millam2009). Model systems used to investigate the role of S···O interactions in α-helices and a β-strands were from the PDB coordinates of 1PVH and 4KT1. All the side chain atoms in these fragments were deleted, and the Cβ atom was replaced by H. A partial energy minimisation was carried out where all atoms other than hydrogen were frozen. The torsion angles (χ) between N, Cα, Cβ and S were varied for the PES scan. AIM analysis was carried out using AIM2000 (König et al., Reference König, Schönbohm and Bayles2001).

CSD analysis

Fragments F1–F6 provided in Fig. 1b were queried in structural data retrieved from CSD 5.43 (Groom and Allen, Reference Groom and Allen2014) (version 5.43 with updates till March 2022) using ConQuest version 2022 1.0 (Bruno et al., Reference Bruno, Cole, Edgington, Kessler, Macrae, McCabe, Pearson and Taylor 2002a). The following criteria of ConQuest were used for the search: (1) only intermolecular contacts; (2) 3D coordinates determined for all the atoms; (3) structures with crystallographic R-factor ≤10%; (4) no disorder in crystallographic data; (5) no error in 3D atomic coordinates; (6) no polymeric structures; and (7) normalisation of terminal H position. Data thus obtained were further processed and analysed using Mercury 2022 1.0 (Bruno et al., Reference Bruno, Cole, Edgington, Kessler, Macrae, McCabe, Pearson and Taylor 2002b).

Searches were done for only intermolecular contacts using the criteria dS···O ≤ 3.32 Å and dS···N ≤ 3.35 Å. For the identification of H-bond, dS···H ≤ 2.8 Å, which is 0.2 Å shorter than the sum of the van der Waal radii of S and H (Bondi, Reference Bondi1964), was chosen for higher stringency. The number of bonded atoms to S in F1–F6 were 2. The number of bonded atoms to O and N in F3–F6 were 2 and 3, respectively. The number of such fragments in CSD is provided in Table 1. The approach of H or O/N towards S in space was investigated using the 3D parameters provided in ConQuest (Fig. 2a). To segregate H- and Ch-bonds based on their θ and δ values, we calculated the mean values for clusters in F1 and F3. The range of θ and δ defining H- and Ch-bonds about the respective mean values were obtained by taking their mean ± standard deviation at 1 sigma of the calculated values. The values for the limits were rounded off to the closest value that was a multiple of 5. The mean of the angular values of δ was calculated using their modulus. This angular range of θ and δ for H- and Ch-bonds thus obtained were used throughout the study. All the plots and figures were generated using OriginPro 9.0 (Seifert, Reference Seifert2014).

Table 1. A summary of CSD and PDB analysis

Note: NT is the total number of independent pairs of fragments found in the CSD structures. NC is the total number of S···O or S···H–O and S···N or S···H–N contacts found in the CSD and the PDB, having distance between S and the atom less than the sum of their van der Waals radii. The values in the parentheses are equal to (NC/NT) × 100 and represent the frequency of occurrence of the above-mentioned contacts in the CSD.

Abbreviations: CSD, Cambridge Structural Database; PDB, Protein Data Bank.

a Fragments are from protein structures in the PDB.

Figure 2. Nature of non-covalent interactions formed by S. (a) Definition of geometrical parameters d, θ and δ. (b) Mapping of the θ and δ values of S···H–O contacts (blue dots) in F1 with computationally calculated ΔEs in the background (greyscale). The (CH3)2S:OH2 complex was used as a model system to calculate ΔEs for F1; (c) S···O contacts (red dots) in F3. The Cl(CH3)S:O(CH3)2 complex was used as the model system to calculate ΔEs in the background (greyscale). (d) S···O/N contacts with d S···H ˃ 2.8 Å in red and those with d S···H ≤ 2.8 Å in blue. (e) S···O contacts formed by methionine and cystine in F7. The boxes shown in the figures represent the statistically obtained favourable range for θ and δ values for H- or Ch-bonds.

PDB analyses

Protein structures determined using X-ray crystallography in the PDB (Rose et al., Reference Rose, Prlić, Altunkaya, Bi, Bradley, Christie, Di Costanzo, Duarte, Dutta, Feng, Green, Goodsell, Hudson, Kalro, Lowe, Peisach, Randle, Rose, Shao, Tao, Valasatava, Voigt, Westbrook, Woo, Yang, Young, Zardecki, Berman and Burley2017) were downloaded in July 2022. A set of protein structures with resolution ≤2.0 Å, R-factor≤25% and pairwise sequence identity ≤90% was generated using the PISCES server (Wang and Dunbrack, Reference Wang and Dunbrack2005), resulting in 25,107 structures. The criterion for pairwise sequence identity was excluded for searching ligands containing aromatic S. These structures were analysed using an in-house script written in Python 3.7.1. Search for H-bonds and Ch-bonds was made using the criterion of d (Å) ≤ (sum of van der Waals radii of S and O/N). To minimise the effect of structural constraints on the direction parameters, we excluded the contacts where S and O/N were separated by less than seven covalent bonds or were intramolecular. The direction criteria obtained using the CSD analysis were used to distinguish between H- and Ch-bonds.

To understand the effects of H- and Ch-bonds on proteins’ secondary structures, we searched the PDB for S···H–N (dS···N ≤ 3.6 Å, relaxed from 3.35 Å because it usually peaks at 3.6 Å in proteins (Zhou et al., Reference Zhou, Tian, Lv and Shang2009)) and S···O (dS···O ≤ 3.32 Å) bonds made by the peptide backbone with S. An additional criterion of 120˚ ≤ ζ ≤ 240˚, where ζ is a torsion angle (see Supplementary Fig. 1 for more details), was applied to exclude structures where N–H was not pointing towards the lone pair regions of S. Secondary structure information was obtained from the header in the PDB files. For N-terminal capping, we searched for S···H–N bonds where the amino group was of N1, N2 or N3 residue at the N-terminus of the helix. While for C-terminal capping, we searched for S···O interactions where the carbonyl O was of C1, C2 or C3 residue at the C-terminus of the helix (Aurora and Rose, Reference Aurora and Rose1998). A similar analysis was performed to investigate the effects of S···H–N, S···O and S···N interactions on the stability of α-helices and β-strands. In this case, the last four residues at the N- and C-termini of the helix were excluded, whereas all the residues belonging to the strand were considered. Search for metal-chelating cysteine used a distance between S and the metal to be within 1.9–2.8 Å, which ensured that interactions with metals of different ionic radii were identified. This distance range was calculated using an in-house script for all interactions between S and metals in protein structures in the PDB. Figures of protein structures were made using Chimera 1.13.1 (Pettersen et al., Reference Pettersen, Goddard, Huang, Couch, Greenblatt, Meng and Ferrin2004).

Results and discussion

Geometrical features that distinguish S-mediated Ch-bond from H-bond

Structures in the PDB solved using X-ray crystallography often lack positional information of H atoms, making it difficult to identify if the non-covalent bond between S and O/N is H-bond or Ch-bond. We devised a methodology to distinguish between the H- and Ch-bonds even if positional information of the H atom was absent. Towards this, we first analysed fragments F1 and F2 in the CSD having positional information of H to identify the preferred direction of approach of H towards S to form H-bond (Fig. 1b). Fragments F3 and F4 were studied to characterise the preferred direction of approach of O/N towards S to form Ch-bond. θ and δ were used to characterise the angular distribution for H- and Ch-bonds (Fig. 2a). The parameters measured for all such contacts in the CSD were plotted to illustrate the preferred range of distances and angles of these interactions (Fig. 2b, c and Supplementary Fig. 2a,b). The angular distribution thus obtained was compared with the complexation energy ΔE obtained from PES scans at different values of θ and δ for the model systems (Fig. 2b, c).

Two distinct clusters were observed in the θδ plot for S···H–O contacts. The boundary values of the two clusters were (i) 95° ≤ θ ≤ 145° and − 90° ≤ δ < −50° and (ii) 95° ≤ θ ≤ 145° and 50° < δ ≤ 90°, respectively. This matched with the location of the PES scan minima (Fig. 2b and Supplementary Fig. 2a). The clusters represented the direction of approach of the electrophile towards the lone pairs of S (Supplementary Fig. 3a). In the case of S···O contacts, a single cluster was observed at a different region of the θδ plot (115° ≤ θ ≤ 155° and − 50° ≤ δ ≤ 50°), which overlapped with the PES scan minimum (Fig. 2c and Supplementary Fig. 2b). The direction corresponded to the approach of the nucleophile towards the σ-hole on S (Supplementary Fig. 3a; Politzer et al., Reference Politzer, Murray and Clark2013; Aakeroy et al., Reference Aakeroy, Bryce, Desiraju, Frontera, Legon, Nicotra, Rissanen, Scheiner, Terraneo, Metrangolo and Resnati2019). Outliers in the plots were due to other strong interactions within the molecules, such as other H-bonds and stacking interactions (Supplementary Fig. 3b–d). The number of S···N interactions was much less than S···O (Table 1) possibly because of N being conjugated in most of the structures resulting in the lack of lone pair electrons for the formation of Ch-bond.

Delineation of Ch-bond from H-bond in groups having electrophilic and nucleophilic centres

We next sought the nature of bonding between S and functional groups having both electrophilic and nucleophilic centres because they are common in proteins. For this, we studied fragments F5 and F6 in the CSD (Fig. 1a,b and Table 1). Using distance criteria, we ensured that these fragments formed either S···H–O/N or S···O/N interaction. Due to structural constraints, both interactions did not occur simultaneously. From this set of interactions, contacts satisfying dS···H ≤ 2.8 Å were assigned as H-bond and the rest as Ch-bond. Note that this filtering strategy excluded those H-bonds having a distance between S and O/N greater than 2.8 Å. The CSD analysis revealed three clusters in the θδ plot (Fig. 2d). Interactions in two of these clusters had θδ values expected for H-bond and most of them satisfied the criterion dS···H ≤ 2.8 Å (Fig. 2b). S···O/N interactions with dS···H > 2.8 Å primarily clustered with θδ distribution that matched the directionality of Ch-bond (Fig. 2c). A few interactions in this cluster had dS···H ≤ 2.8 Å. Note that H–O/N groups that formed Ch-bond with S could also form H-bond with a neighbouring acceptor atom (Supplementary Fig. 3a).

In proteins, Ch- or H-bond can be formed between methionine and cystine with side chains of serine, threonine or tyrosine, or the backbone amide or water (fragments F7–F9) is equivalent to fragments F5 and F6. As there were very few examples of X1–S–H in the CSD analysis discussed above, we excluded interactions made by free cysteine from the analysis of PDB structures. As in the case of fragments F5 and F6, the θδ plot obtained by analysing F1–F4 revealed the segregation of angular values into three clusters corresponding to either H-bond or Ch-bond (Fig. 2e and Supplementary Figs 2c,d and 4). In summary, the θδ plot allowed us to identify and distinguish Ch-bond from H-bond in protein structures without positional information of H atoms.

Electronic environment of S determines the choice between Ch- and H-bond

We next sought to find what dictated the choice between the formation of Ch-bond and H-bond. In general, the formation of H- or Ch-bond depends on the strength of lone pairs and σ-holes on S, respectively, which are in turn affected by the nature of substituent groups (Adhikari and Scheiner, Reference Adhikari and Scheiner2014; Kumar et al., Reference Kumar, Gadre, Mohan and Suresh2014). Using MESP, we found the same in model systems relevant to biomolecules (Fig. 3a, b). The MESP analysis revealed two V min (MESP minimum), corresponding to the lone pairs on S in all the model systems (Fig. 3a). The electrostatic potential maps also showed the presence of two σ-holes on the extension of the S–X bonds, except in the case of [Fe(SCH3)4] complex (Fig. 3b) because of the anionic nature of metal-chelated S (Hirano et al., Reference Hirano, Takeda and Miki2016). Interestingly, we noted the ability of substituents to modulate the strength of the lone pairs and σ-holes on S in a reciprocal manner, which consequently was expected to affect the nature of the bond formed.

Figure 3. Effect of substitution on the electronic environment of S. (a) Molecular electrostatic surface potential (MESP) minimum values, V min, in kcal mol−1 represents the lone pair regions of the S-containing monomers used in this study. (b) MESP map of the monomers with the colour-coding range from −6.28 (red) to 43.93 kcal mol−1 (blue) and textured on a 0.01 au density isosurface. The two σ-holes marked by arrows and their magnitude, V S,max, in kcal mol−1.

To see if this was true, we categorised all the contacts obtained from CSD based on the substituents linked to S (Fig. 4a and Supplementary Table 1). We analysed the corresponding structures to check if S formed Ch-bond or H-bond. Eighty-eight percent of S(Ar) and 73% of E–S–Y formed Ch-bond. In sharp contrast, more than 97% of M–S–Y formed H-bond. In comparison, saturated C/S/H substituents (R–S–R) appeared to have a lesser influence on the choice of the bond formed. The number of H-bonds (52%) was almost comparable to the number of Ch-bonds (48%) (Supplementary Table 1). The observations matched the expectations from the MESP analysis performed on the model systems. In summary, divalent S, when part of an aromatic ring or bonded to an electron-withdrawing group, is most likely to form a Ch-bond, whereas S coordinated with a metal can form an H- but not a Ch-bond.

Figure 4. Rules for the formation of H- and Ch-bonds. (a) Histogram showing the frequency formation of Ch-bond and H-bond in the Cambridge Structural Database and (b) the Protein Data Bank with different electronic environments of S. M = any metal; Y = any element except M; R = saturated C, H and S; E = any electron-withdrawing group and S(Ar) = Aromatic S. (c) Substituent-dependent see-saw change in the strength of the lone pairs and σ-holes on S.

Disulphide-linked S preferentially forms Ch-bonds, whereas metal-chelated cysteine form H-bonds in proteins

We studied interactions made by S of methionine or cystine with the hydroxyl, amino or carbonyl group of backbone amide, side chains of serine, threonine, tyrosine, aspartate, glutamate, arginine, lysine, histidine, asparagine, glutamine, tryptophan and bound water (Supplementary Table 2). Our analysis revealed that the disulphide-linked Cys-Sγ was more frequently involved in Ch-bonds (85%) than H-bonds (15%). In comparison, Met-Sδ appeared to form H-bonds (57%) only marginally more than Ch-bonds (43%) (Fig. 4b and Supplementary Table 3). The MESP analysis showed that S bonded to two methyl groups (C–S–C), as in methionine, had comparable values of V min and V S,max (Fig. 3a, b). In contrast, V S,max on a disulphide-linked S (C–S–S) was larger than V min, thus providing a rationale for cystine to preferentially form Ch-bonds.

Aromatic S preferentially formed Ch-bonds with groups containing O. This observation was consistent with previous reports of Ch-bonds between S in the aromatic rings of drugs containing thiophene, thiazole and thiadiazole groups and O in target proteins (Thomas et al., Reference Thomas, Jayatilaka and Guru Row2015; Zhang et al., Reference Zhang, Gong, Li and Lu2015; Koebel et al., Reference Koebel, Cooper, Schmadeke, Jeon, Narayan and Sirimulla2016; Kristian et al., Reference Kristian, Fanfrlík and Lepšík2018). S-mediated interactions with functional groups containing N did not show these features presumably because the delocalised lone pairs of N in backbone amide or the side chain precluded the formation of Ch-bonds. We next analysed the PDB for non-covalent interactions formed by metal-chelated cysteines, which occur in many metalloproteins. Consistent with the rules stated above and independent of the identity of the metal, the thiolate of cysteine preferentially formed H-bonds (M–S–Y in Fig. 4a). In summary, our analyses of the structures in the CSD and the PDB revealed that the nature of the interaction between functional groups containing electrophilic and nucleophilic centres and S was influenced by the substituent-dependent see-saw change in the strength of the lone pairs and σ-holes on S (Fig. 4b).

Role of S in helix capping

Capping satisfies the H-bond-forming abilities of the free backbone N–H or C=O of the terminal residues of an α-helix and is essential for the stability of α-helices in proteins and peptides (Aurora and Rose, Reference Aurora and Rose1998). The role of polar side chains of serine, threonine and asparagine, the acidic side chain of aspartate, the backbone amide of a neighbouring residue and metal-chelated S of cysteine in helix capping are well documented (Doig and Baldwin, Reference Doig and Baldwin1995; Aurora and Rose, Reference Aurora and Rose1998), but not those of methionine and cystine. As the N-terminus and C-terminus of α-helices have free backbone N–H (electrophile) and free backbone C=O (nucleophile), respectively, we asked if Met-Sδ or Cys-Sγ would interact and cap them.

We analysed protein structures in the PDB and found a number of examples of Met-Sδ or Cys-Sγ interacting with backbone amino at the N-terminus or backbone carbonyl at the C-terminus of α-helix (Fig. 5a and Supplementary Table 4). Consistent with previous observations (Doig and Baldwin, Reference Doig and Baldwin1995), we found that metal-chelated thiolates capped only the N-terminus of the helix by H-bonds. In contrast, 75% of Cys-Sγ capped the C-terminus by Ch-bonds, whereas the remaining 25% capped the N-terminus by H-bonds (Fig. 5b). Amongst the examples involving Met-Sδ, 37% of the interactions were H-bonds with backbone N–H of the N-terminal residues and 63% Ch-bonds with backbone C=O of the C-terminal residues (Fig. 5c).

Figure 5. Helix capping and edge strand stabilisation. (a) Representative examples of H-bond capping the N-terminus of α-helices and (b) Ch-bond capping the C-terminus of α-helices. (c) Histogram showing the frequency of H-bond and Ch-bond interactions capping the N- and C-termini of α-helices by metal-chelating cysteine, methionine and cystine. (d) Representative examples of negative design involving Ch-bond and H-bond.

Augmentation of the stability of regular secondary structures by S

In addition to backbone H-bonds, other non-covalent interactions such as C–H···O and n → π* are important for the structural stability of α-helices or β-sheets (Derewenda et al., Reference Derewenda, Lee and Derewenda1995; Manikandan and Ramakumar, Reference Manikandan and Ramakumar2004; Bartlett et al., Reference Bartlett, Choudhary, Raines and Woolfson2010). An earlier study reported methionine-forming intra-helical and inter-strand Ch-bonds with backbone O (Pal and Chakrabarti, Reference Pal and Chakrabarti2001). This prompted us to find if S could contribute to the stability of α-helices and β-strands through H- and Ch-bonds. We analysed structures in the PDB for Ch- and H-bonds between Met-Sδ or Cys-Sγ with backbone O or N–H of residues of α-helices or β-strands (Supplementary Fig. 5 and Table 5). Interestingly, we found an S-mediated H-bond involving backbone N–H of β-strands at the edge of β-sheets (Fig. 5d). We also found Ch-bond formed by S with free backbone C=O of edge strands.

Many elements of negative design, a mechanism that prevents the β-strand dimerisation and stabilises an edge strand of β-sheets, have been documented previously (Richardson and Richardson, Reference Richardson and Richardson2002; Koga et al., Reference Koga, Tatsumi-Koga, Liu, Xiao, Acton, Montelione and Baker2012). This includes the interaction of other regions of the protein with the edge β-strand, disruption of backbone H-bond formation by proline or a β-bulge or use of inward-pointing charged residues to prevent strand-mediated dimerisation. Our analysis revealed that H-bond or Ch-bond formed by backbone N–H or C=O of edge β-strand, respectively, with a neighbouring Met-Sδ or Cys-Sγ, is another element of negative-design that can stabilise β-sheets. Additionally, we found that in some proteins, the free backbone N–H of the insertion residue of a classical β-bulge formed an H-bond with Met-Sδ located two residues ahead (Supplementary Fig. 5).

To gauge the potential contribution of Ch-bonds to stabilise regular secondary structures, we performed a PES scan by varying the torsion angle χ about the covalent bond between Cα and Cβ of the cystines whose S formed a Ch-bond. One fragment each from an α-helix and a β-strand were chosen for the calculations (Fig. 6a). A plot of relative conformational energy versus χ showed that the minimum conformational energy corresponded to the χ of the respective crystal structures (Fig. 6b). The AIM analysis for minimum energy conformations showed the presence of a bond critical point (BCP) between S and O. ρ-values of these BCPs were 0.007 au for 1PVH and 0.011 au for 4KT1 (Fig. 6b), which were in the range suggested previously for favourable non-covalent interactions, that is, 0.002–0.035 au (Bader, Reference Bader1991). The analysis, thus, strongly suggested that Ch-bonds could provide extra stability to a particular conformation in protein molecules.

Figure 6. Stabilisation of α-helices and β-sheets by S. (a) Representative examples of Ch-bonds found in α-helical and β-sheet regions. (b) The plot for conformational energies as a function of χ (E at χ = 170˚ was assigned as 0.0 kcal mol−1). Values of ρ at bond critical points for the S···O interaction for the most energetically favourable structures are given in atomic units.

Conclusions

In this study, we have tried to understand the role of H- and Ch-bonds formed by divalent S in proteins. Computational analyses showed that the S-mediated interactions contributed to the stability of protein conformation and secondary structures. Hence, we conclude that S-mediated Ch- and H-bonds, like other weak interactions, are an essential aspect of the energy landscape in protein folding that compensates for unfavourable conformational entropy changes through favourable interactions (Grantcharova et al., Reference Grantcharova, Alm, Baker and Horwich2001; Dobson, Reference Dobson2003). Furthermore, we envisage that cooperativity among S-mediated and other weak interactions is likely to modulate their strengths with direct implications for protein function, which remains to be studied (Adhav et al., Reference Adhav, Pananghat and Saikrishnan2022). For example, we speculate that the propensity and strength of Ch-bonds would increase upon the delocalisation of lone-pair electron density of Cys-Sγ to form an n → π* interaction with a vicinal carbonyl group (Kilgore and Raines, Reference Kilgore and Raines2018). Also, the computational analyses reported here do not delineate the contribution of hydrophobic and van der Waals interactions from those of the polar Ch-bond and H-bond interactions towards structural stability.

S-mediated Ch- and H-bonds can contribute to the structural stability and substrate specificity of proteins, like other interactions formed by polar amino acids. However, S-mediated interactions can have properties different from other polar non-covalent interactions, for instance, the resistance of Ch-bond strength to solvent polarity (Pascoe et al., Reference Pascoe, Ling and Cockroft2017), thus bringing additional diversity to the repertoire of weak interactions essential for biomolecular functions. This could be a reason why, despite their high biosynthetic cost (Doig, Reference Doig2017), nature selected S-containing amino acids as part of the 20 building blocks of proteins. The wide variety of functionally relevant interactions made by S in proteins necessitates that these non-covalent interactions too are considered in the energy functions used for determining protein structures, folding pathways and binding properties. Also, the design and engineering of proteins and peptides would benefit from a better understanding of the distinct bonding properties of methionine and cysteine/cystine.

Open peer review

To view the open peer review materials for this article, please visit http://doi.org/10.1017/qrd.2023.3.

Supplementary material

The supplementary material for this article can be found at http://doi.org/10.1017/qrd.2023.3.

References

Aakeroy, CB, Bryce, DL, Desiraju, GR, Frontera, A, Legon, AC, Nicotra, F, Rissanen, K, Scheiner, S, Terraneo, G, Metrangolo, P and Resnati, G (2019) Definition of the chalcogen bond (IUPAC recommendations 2019). Pure and Applied Chemistry 91, 18891892.CrossRefGoogle Scholar
Adhav, VA, Pananghat, B, and Saikrishnan, K (2022). Probing the Directionality of S···O/N Chalcogen Bond and Its Interplay with Weak C–H···O/N/S Hydrogen Bond Using Molecular Electrostatic Potential. The Journal of Physical Chemistry B, 126(40), 78187832.CrossRefGoogle ScholarPubMed
Adhikari, U and Scheiner, S (2014) Effects of charge and substituent on the S···N chalcogen bond. Journal of Physical Chemistry A 118, 31833192.CrossRefGoogle ScholarPubMed
Andersen, CL, Jensen, CS, Mackeprang, K, Du, L, Jørgensen, S and Kjaergaard, HG (2014) Similar strength of the NH⋯O and NH⋯S hydrogen bonds in binary complexes. Journal of Physical Chemistry A 118, 1107411082.CrossRefGoogle Scholar
Aurora, R and Rose, GD (1998) Helix capping. Protein Science 7, 2138.CrossRefGoogle ScholarPubMed
Bader, RFW (1991) A quantum theory of molecular structure and its applications. Chemical Reviews 91, 893928.CrossRefGoogle Scholar
Bartlett, GJ, Choudhary, A, Raines, RT and Woolfson, DN (2010) N→π* interactions in proteins. Nature Chemical Biology 6(8), 615620.CrossRefGoogle Scholar
Beno, BR, Yeung, K, Bartberger, MD, Pennington, LD and Meanwell, NA (2015) A survey of the role of noncovalent sulfur interactions in drug design. Journal of Medicinal Chemistry 58, 43834438.CrossRefGoogle ScholarPubMed
Benz, S, López-Andarias, J, Mareda, J, Sakai, N and Matile, S (2017) Catalysis with chalcogen bonds. Angewandte Chemie International Edition 56(3), 812815.CrossRefGoogle ScholarPubMed
Berg, JM, Tymoczko, JL and Stryer, L (2002) Biochemistry, 6th Edn. New York: W. H. Freeman.Google Scholar
Bondi, A (1964) Van der Waals volumes and radii. The Journal of Physical Chemistry 68(3), 441451.CrossRefGoogle Scholar
Bruno, IJ, Cole, JC, Edgington, PR, Kessler, M, Macrae, CF, McCabe, P, Pearson, J and Taylor, R (2002a) ConQuest – New software for searching the CSD and visualizing crystal structures. Acta Crystallographica Section B 58, 389397.CrossRefGoogle ScholarPubMed
Bruno, IJ, Cole, JC, Edgington, PR, Kessler, M, Macrae, CF, McCabe, P, Pearson, J and Taylor, R (2002b) Mercury 3.8 software. Acta Crystallographica Section B Structural Science B58, 389397.CrossRefGoogle Scholar
Carugo, O, Resnati, G and Metrangolo, P (2021) Chalcogen bonds involving selenium in protein structures. ACS Chemical Biology 16(9), 16221627.CrossRefGoogle ScholarPubMed
Chand, A, Sahoo, DK, Rana, A, Jena, S and Biswal, HS (2020) The prodigious hydrogen bonds with sulfur and selenium in molecular assemblies, structural biology, and functional materials. Accounts of Chemical Research 53(8), 15801592.CrossRefGoogle ScholarPubMed
Chen, L, Xiang, J, Zhao, Y and Yan, Q (2018) Reversible self-assembly of supramolecular vesicles and nanofibers driven by chalcogen-bonding interactions. Journal of the American Chemical Society 140, 70797082.CrossRefGoogle ScholarPubMed
Dennington, R, Keith, T and Millam, J (2009) GaussView, Version 5. Shawnee Mission, KS: Semichem Inc. Google Scholar
Derewenda, ZS, Lee, L and Derewenda, U (1995) The occurence of C–H···O hydrogen bonds in proteins. Journal of Molecular Biology 252, 248262.CrossRefGoogle ScholarPubMed
Dill, KA and MacCallum, JL (2012) The protein-folding problem, 50 years on. Science 338, 10421046.CrossRefGoogle Scholar
Dobson, CM (2003) Protein folding and misfolding. Nature 426, 884890.CrossRefGoogle ScholarPubMed
Doig, AJ (2017) Frozen, but no accident – Why the 20 standard amino acids were selected. FEBS Journal 284, 12961305.CrossRefGoogle ScholarPubMed
Doig, AJ and Baldwin, RL (1995) N‐ and C‐capping preferences for all 20 amino acids in α‐helical peptides. Protein Science 4, 13251336.CrossRefGoogle ScholarPubMed
Forbes, CR, Sinha, SK, Ganguly, HK, Bai, S, Yap, GPA, Patel, S and Zondlo, NJ (2017) Insights into thiol-aromatic interactions: A stereoelectronic basis for S–H/π interactions. Journal of American Chemical Society 139(5), 18421855.CrossRefGoogle ScholarPubMed
Frisch, MJ, Trucks, GW, Schlegel, HB, Scuseria, GE, Robb, MA, Cheeseman, JR, Scalmani, G, Barone, V, Mennucci, B, Petersson, GA, Nakatsuji, H, Caricato, M, Li, X, Hratchian, HP, Izmaylov, AF, Bloino, J, Zheng, G and Sonnenber, DJ (2009) Gaussian 09. Wallingford, CT: Gaussian, Inc., pp. 2–3.Google Scholar
Gallivan, JP and Dougherty, DA (1999) Cation–pi interactions in structural biology. Proceedings of the National Academy of Sciences of the United States of America 96, 94599464.CrossRefGoogle ScholarPubMed
Grantcharova, V, Alm, EJ, Baker, D and Horwich, AL (2001) Mechanisms of protein folding. Current Opinion in Structural Biology 11, 7082.CrossRefGoogle ScholarPubMed
Gregoret, LM, Rader, SD, Fletterick, RJ and Cohen, FE (1991) Hydrogen bonds involving sulfur atoms in proteins. Proteins: Structure, Function and Bioinformatics 9(2), 99107.CrossRefGoogle ScholarPubMed
Groom, CR and Allen, FH (2014) The Cambridge Structural Database in retrospect and prospect. Angewandte Chemie International Edition. 53, 662671.CrossRefGoogle ScholarPubMed
Guru Row, TN and Parthasarathy, R (1981) Directional preferences of nonbonded atomic contacts with divalent sulfur in terms of its orbital orientations. 2. S···S interactions and nonspherical shape of sulfur in crystals. Journal of the American Chemical Society 103, 477479.Google Scholar
Haberhauer, G and Gleiter, R (2020) The nature of strong chalcogen bonds involving chalcogen-containing heterocycles. Angewandte Chemie International Edition 59, 2123621243.CrossRefGoogle ScholarPubMed
Hirano, Y, Takeda, K and Miki, K (2016) Charge-density analysis of an iron–sulfur protein at an ultra-high resolution of 0.48 Å. Nature 534, 281284.CrossRefGoogle ScholarPubMed
Hobza, P and Řezáč, J (2016) Benchmark calculations of interaction energies in noncovalent complexes and their applications. Chemical Reviews 116, 50385071.Google Scholar
Iwaoka, M and Babe, N (2015) Mining and structural characterization of S···X chalcogen bonds in protein database. Phosphorus, Sulfur, and Silicon and the Related Elements 190, 12571264.CrossRefGoogle Scholar
Iwaoka, M and Isozumi, N (2006) Possible roles of S···O and S···N interactions in the functions and evolution of phospholipase A2. Biophysics 2, 2334.CrossRefGoogle Scholar
Iwaoka, M and Isozumi, N (2012) Hypervalent nonbonded interactions of a divalent sulfur atom. Implications in protein architecture and the functions. Molecules 17, 72667283.CrossRefGoogle ScholarPubMed
Iwaoka, M, Takemoto, S and Tomoda, S (2002) Statistical and theoretical investigations on the directionality of nonbonded S···O interactions. Implications for molecular design and protein engineering. Journal of the American Chemical Society 124(35), 1061310620.CrossRefGoogle ScholarPubMed
Jena, S, Dutta, J, Tulsiyan, KD, Sahu, AK, Choudhury, SS and Biswal, HS (2022) Noncovalent interactions in proteins and nucleic acids: Beyond hydrogen bonding and π-stacking. Chemical Society Reviews 51, 42614286.CrossRefGoogle ScholarPubMed
Juanes, M, Saragi, RT, Jin, Y, Zingsheim, O, Schlemmer, S and Lesarri, A (2020) Rotational spectrum and intramolecular hydrogen bonding in 1,2-butanedithiol. Journal of Molecular Structure 1211, 128080.CrossRefGoogle Scholar
Kilgore, HR and Raines, RT (2018) N→π∗ interactions modulate the properties of cysteine residues and disulfide bonds in proteins. Journal of the American Chemical Society 140, 1760617611.CrossRefGoogle ScholarPubMed
Koebel, MR, Cooper, A, Schmadeke, G, Jeon, S, Narayan, M and Sirimulla, S (2016) S···O and S···N sulfur bonding interactions in protein–ligand complexes: Empirical considerations and scoring function. Journal of Chemical Information and Modeling 56, 22982309.CrossRefGoogle ScholarPubMed
Koga, N, Tatsumi-Koga, R, Liu, G, Xiao, R, Acton, TB, Montelione, GT and Baker, D (2012) Principles for designing ideal protein structures. Nature 491, 222227.CrossRefGoogle ScholarPubMed
Kolb, S, Oliver, GA and Werz, DB (2020) Chemistry evolves, terms evolve, but phenomena do not evolve: From chalcogen–chalcogen interactions to chalcogen bonding. Angewandte Chemie International Edition. 59, 27.CrossRefGoogle Scholar
König, FB, Schönbohm, J and Bayles, D (2001) AIM2000 – a program to analyze and visualize atoms in molecules. Journal of Computational Chemistry 22, 545559.Google Scholar
Kristian, K, Fanfrlík, J and Lepšík, M (2018) Chalcogen bonding in protein−ligand complexes: PDB survey and quantum mechanical calculations. ChemPhysChem 19, 25402548.Google Scholar
Kumar, A, Gadre, SR, Mohan, N and Suresh, CH (2014) Lone pairs: An electrostatic viewpoint. Journal of Physical Chemistry A 118, 526532.CrossRefGoogle ScholarPubMed
Kumar, A, Yeole, SD, Gadre, SR, Lõpez, R, Rico, JF, Ramírez, G, Ema, I and Zorrilla, D (2015) DAMQT 2.1.0: A new version of the DAMQT package enabled with the topographical analysis of electron density and electrostatic potential in molecules. Journal of Computational Chemistry 36(31), 23502359.CrossRefGoogle ScholarPubMed
Lehninger, AL, Nelson, DL and Cox, MM (2000) Principles of Biochemistry, 4th Edn. New York: Worth Publishers.Google Scholar
Lim, JYC and Beer, PD (2018) Sigma-hole interactions in anion recognition. Chem 4, 731783.CrossRefGoogle Scholar
Lucas, X, Bauzá, A, Frontera, A and Quiñonero, D (2016) A thorough anion-π interaction study in biomolecules: On the importance of cooperativity effects. Chemical Science 7, 10381050.CrossRefGoogle ScholarPubMed
Mahmudov, KT, Kopylovich, MN, Guedes Da Silva, MFC and Pombeiro, AJL (2017) Chalcogen bonding in synthesis, catalysis and design of materials. Dalton Transactions 46, 1012110138.CrossRefGoogle ScholarPubMed
Manikandan, K and Ramakumar, S (2004) The occurrence of C–H···O hydrogen bonds in α-helices and helix termini in globular proteins. Proteins: Structure, Function and Genetics 54, 768781.CrossRefGoogle Scholar
Matta, CF, Hernández-Trujillo, J, Tang, TH and Bader, RFW (2003) Hydrogen–hydrogen bonding: A stabilizing interaction in molecules and crystals. Chemistry – A European Journal 9, 19401951.CrossRefGoogle ScholarPubMed
Motherwell, WB, Moreno, RB, Pavlakos, I, Arendorf, JRT, Arif, T, Tizzard, GJ, Coles, SJ and Aliev, AE (2018) Non-covalent interactions of π systems with sulfur: The atomic chameleon of molecular recognition. Angewandte Chemie – International Edition 57, 11931198.CrossRefGoogle Scholar
Murray, JS, Lane, P, Clark, T, Riley, KE and Politzer, P (2012) σ-Holes, π-holes and electrostatically-driven interactions. Journal of Molecular Modeling 18, 541548.CrossRefGoogle ScholarPubMed
Newberry, RW and Raines, RT (2019) Secondary forces in protein folding. ACS Chemical Biology 14, 16771686.CrossRefGoogle ScholarPubMed
Pace, CN, Fu, H, Fryar, KL, Landua, J, Trevino, SR, Schell, D, Thurlkill, RL, Imura, S, Scholtz, JM, Gajiwala, K, Sevcik, J, Urbanikova, L, Myers, JK, Takano, K, Hebert, EJ, Shirley, BA and Grimsley, GR (2014) Forces stabilizing proteins. FEBS Letters 558, 21772184.CrossRefGoogle Scholar
Pal, D and Chakrabarti, P (2001) Non-hydrogen bond interactions involving the methionine sulfur atom. Journal of Biomolecular Structure and Dynamics 19(1), 115128.CrossRefGoogle ScholarPubMed
Pascoe, DJ, Ling, KB and Cockroft, SL (2017) The origin of chalcogen-bonding interactions. Journal of the American Chemical Society 139(42), 1516015167.CrossRefGoogle ScholarPubMed
Pettersen, EF, Goddard, TD, Huang, CC, Couch, GS, Greenblatt, DM, Meng, EC and Ferrin, TE (2004) UCSF chimera – A visualization system for exploratory research and analysis. Journal of Computational Chemistry 25, 16051612.CrossRefGoogle ScholarPubMed
Politzer, P, Murray, JS and Clark, T (2013) Halogen bonding and other σ-hole interactions: A perspective. Physical Chemistry Chemical Physics 15(27), 1117811189.CrossRefGoogle ScholarPubMed
Politzer, P, Murray, JS and Clark, T (2014) σ-Hole bonding: A physical interpretation. Topics in Current Chemistry 358, 1942.CrossRefGoogle Scholar
Politzer, P, Murray, JS, Clark, T and Resnati, G (2017) The σ-hole revisited. Physical Chemistry Chemical Physics. 19, 3216632178.CrossRefGoogle ScholarPubMed
Rao Mundlapati, V, Ghosh, S, Bhattacherjee, A, Tiwari, P and Biswal, HS (2015) Critical assessment of the strength of hydrogen bonds between the sulfur atom of methionine/cysteine and backbone amides in proteins. Journal of Physical Chemistry Letters 6, 13851389.CrossRefGoogle Scholar
Richardson, JS and Richardson, DC (2002) Natural beta-sheet proteins use negative design to avoid edge-to-edge aggregation. Proceedings of the National Academy of Sciences 99, 27542759.CrossRefGoogle ScholarPubMed
Riwar, LJ, Trapp, N, Root, K, Zenobi, R and Diederich, F (2018) Supramolecular capsules: Strong versus weak chalcogen bonding. Angewandte Chemie – International Edition 57, 17259172645.CrossRefGoogle ScholarPubMed
Rose, PW, Prlić, A, Altunkaya, A, Bi, C, Bradley, AR, Christie, CH, Di Costanzo, L, Duarte, JM, Dutta, S, Feng, Z, Green, RK, Goodsell, DS, Hudson, B, Kalro, T, Lowe, R, Peisach, E, Randle, C, Rose, AS, Shao, C, Tao, YP, Valasatava, Y, Voigt, M, Westbrook, JD, Woo, J, Yang, H, Young, JY, Zardecki, C, Berman, HM and Burley, SK (2017) The RCSB Protein Data Bank: Integrative view of protein, gene and 3D structural information, Nucleic Acids Research 45(D1), D271D281.Google ScholarPubMed
Rosenfield, RE, Parthasarathy, R and Dunitz, JD (1977) Directional preferences of nonbonded atomic contacts with divalent sulfur. 1. Electrophiles and nucleophiles. Journal of the American Chemical Society 99, 48604862.CrossRefGoogle Scholar
Scilabra, P, Terraneo, G and Resnati, G (2019) The chalcogen bond in crystalline solids: A world parallel to halogen bond. Accounts of Chemical Research 52(5), 13131324.CrossRefGoogle ScholarPubMed
Seifert, E (2014) OriginPro 9.1: Scientific data analysis and graphing software – Software review. Journal of Chemical Information and Modeling 54, 1552.CrossRefGoogle ScholarPubMed
Thomas, SP, Jayatilaka, D and Guru Row, TN (2015) S⋯O chalcogen bonding in sulfa drugs : Insights from multipole charge density and X-ray wavefunction of acetazolamide. Physical Chemistry Chemical Physics 17, 2541125420.CrossRefGoogle ScholarPubMed
Vogel, L, Wonner, P and Huber, SM (2019) Chalcogen bonding: An overview. Angewandte Chemie International Edition 58, 1880.CrossRefGoogle ScholarPubMed
Wang, G and Dunbrack, RL (2005) PISCES: Recent improvements to a PDB sequence culling server. Nucleic Acids Research 33, W94W98.CrossRefGoogle ScholarPubMed
Wang, D and Fujii, A (2017) Spectroscopic observation of two-center three-electron bonded (hemi-bonded) structures of (H2S) n + clusters in the gas phase. Chemical Science 8, 26672670.CrossRefGoogle ScholarPubMed
Wang, W, Zhu, H, Feng, L, Yu, Q, Hao, J, Zhu, R and Wang, Y (2020) Dual chalcogen–chalcogen bonding catalysis. Journal of the American Chemical Society 142(6), 31173124.CrossRefGoogle ScholarPubMed
Yeole, SD, López, R and Gadre, SR (2012) Rapid topography mapping of scalar fields: Large molecular clusters. Journal of Chemical Physics 137, 074116.CrossRefGoogle ScholarPubMed
Zhang, X, Gong, Z, Li, J and Lu, T (2015) Intermolecular sulfur⋯oxygen interactions: Theoretical and statistical investigations. Journal of Chemical Information and Modeling 55, 21382153.CrossRefGoogle Scholar
Zhao, Y and Truhlar, DG (2008) The M06 suite of density functionals for main group thermochemistry, thermochemical kinetics, non-covalent interactions, excited states, and transition elements: Two new functionals and systematic testing of four M06-class functionals and 12 other functions. Theoretical Chemistry Accounts 120(1–3), 215241.CrossRefGoogle Scholar
Zhou, P, Tian, F, Lv, F and Shang, Z (2009) Geometric characteristics of hydrogen bonds involving sulfur atoms in proteins. Proteins: Structure, Function and Bioinformatics 76(1), 151163.CrossRefGoogle ScholarPubMed
Figure 0

Figure 1. An approach of electrophiles and nucleophiles towards divalent S. (a) Approach of an electrophile or a nucleophile (upper panel) and a covalently linked electrophile–nucleophile fragment (lower panel) towards S. (b) Fragments analysed using the Cambridge Structural Database and the Protein Data Bank (PDB). *Fragments are from protein structures in the PDB.

Figure 1

Table 1. A summary of CSD and PDB analysis

Figure 2

Figure 2. Nature of non-covalent interactions formed by S. (a) Definition of geometrical parameters d, θ and δ. (b) Mapping of the θ and δ values of S···H–O contacts (blue dots) in F1 with computationally calculated ΔEs in the background (greyscale). The (CH3)2S:OH2 complex was used as a model system to calculate ΔEs for F1; (c) S···O contacts (red dots) in F3. The Cl(CH3)S:O(CH3)2 complex was used as the model system to calculate ΔEs in the background (greyscale). (d) S···O/N contacts with dS···H ˃ 2.8 Å in red and those with dS···H ≤ 2.8 Å in blue. (e) S···O contacts formed by methionine and cystine in F7. The boxes shown in the figures represent the statistically obtained favourable range for θ and δ values for H- or Ch-bonds.

Figure 3

Figure 3. Effect of substitution on the electronic environment of S. (a) Molecular electrostatic surface potential (MESP) minimum values, Vmin, in kcal mol−1 represents the lone pair regions of the S-containing monomers used in this study. (b) MESP map of the monomers with the colour-coding range from −6.28 (red) to 43.93 kcal mol−1 (blue) and textured on a 0.01 au density isosurface. The two σ-holes marked by arrows and their magnitude, VS,max, in kcal mol−1.

Figure 4

Figure 4. Rules for the formation of H- and Ch-bonds. (a) Histogram showing the frequency formation of Ch-bond and H-bond in the Cambridge Structural Database and (b) the Protein Data Bank with different electronic environments of S. M = any metal; Y = any element except M; R = saturated C, H and S; E = any electron-withdrawing group and S(Ar) = Aromatic S. (c) Substituent-dependent see-saw change in the strength of the lone pairs and σ-holes on S.

Figure 5

Figure 5. Helix capping and edge strand stabilisation. (a) Representative examples of H-bond capping the N-terminus of α-helices and (b) Ch-bond capping the C-terminus of α-helices. (c) Histogram showing the frequency of H-bond and Ch-bond interactions capping the N- and C-termini of α-helices by metal-chelating cysteine, methionine and cystine. (d) Representative examples of negative design involving Ch-bond and H-bond.

Figure 6

Figure 6. Stabilisation of α-helices and β-sheets by S. (a) Representative examples of Ch-bonds found in α-helical and β-sheet regions. (b) The plot for conformational energies as a function of χ (E at χ = 170˚ was assigned as 0.0 kcal mol−1). Values of ρ at bond critical points for the S···O interaction for the most energetically favourable structures are given in atomic units.

Supplementary material: File

Adhav et al. supplementary material

Adhav et al. supplementary material

Download Adhav et al. supplementary material(File)
File 8.2 MB

Review: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R0/PR1

Conflict of interest statement

No conflict of interest.

Comments

Comments to Author: The manuscript entitled “Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space” is nice piece of work. However, the following corrections/clarifications are needed before its publication.

1. Several important references are missing in the introduction part while discussing non-covalent interactions with sulfur (hydrogen bonding and chalcogen bonding). A few noticeable references are J. Phys. Chem. Lett. 2015, 8, 1385-1389, J. Am. Chem. Soc. 2017, 139, 1842−1855, Chem. Sci. 2017, 8, 2667−2670, Acc. Chem. Res. 2020, 53, 8, 1580-1592, J. Mol. Struct. 2020, 1211, 128080, Chem. Soc. Rev., 2022,51, 4261-4286. These references need to be cited and discussed in the revised manuscript.

2. Some of the important parameters such as lower electronegativity, charge and higher polarizability of sulfur should not be ignored while discussing hydrogen bonding with sulfur.

3. Why have different basis sets been used to optimize the monomer and dimer?

4. Why have the criteria for S---H bond distance taken less than 2.8 Å, whereas their sum of van der Walls radii is 3Å? Also, the S---N bond distance criteria were different for CSD (3.35 Å) and PDB (3.6 Å) search. The authors should provide a logical explanation for this.

5. A more detailed explanation should be given in the context of negligible sigma-hole formation on S when it is chelated to a metal center.

6. In the PDB analysis (page 8), the authors mentioned that coordinate files greater than 1 MB in size were excluded from both sets. What is the reason behind choosing based on file size?

7. Fluorescence-based studies:

Why are the experiments done at pH 7.6, slightly higher than the average body pH? Is there any particular reason for this?

The emission maximum for isolated tryptophan in a buffer medium and a protein environment is 350 nm±10 nm. However, in this report, it is 330 nm. Why such a blue shift in the tryptophan emission for MetRS?

For curiosity, can the author explain why fluorescence intensity changes upon ligand binding?

The Kd units in Figures 6C and G need to be correctly written for consistency.

Also, it would be more informative for the readers if the authors provided the emission spectra for titration of proteins and ligands in the supporting information.

8. There is no discussion on CD experiments in the manuscript’s main text. Instead of supplementary 10, the authors can write Figure S10 while discussing it. (page 23)

9. Did the authors perform ITC experiments for all the studied protein-ligand complexes like fluorescence experiments or only for MetRS:Met and MetRS:Nle ? For MetRS:Met complex, the enthalpy contribution is 40%, whereas the entropy contribution is 60% of the total free energy. However, the authors concluded this as prominently entropy-driven; how? Also, how the hydrophobic interactions are responsible for the entropy-driven process?

Review: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R0/PR2

Conflict of interest statement

Reviewer declares none.

Comments

Comments to Author: The contribution by Adhav et al. represents a very comprehensive and thorough investigation of chalcogen bonds in proteins well worth publishing.

As a general comment, the manuscript has obviously been in the making for quite some time, indicated for example by the CSD searches being executed on a rather old version of the database. One example, I searched for the F1 fragment in Figure 1 B and found 9435 fragment pairs rather than 7854 as indicated, with Nc going from 4.7 % as indicated to 6.1 %. I do not think more updated data would change anything important as far as the major results are concerned, but thought it worthwhile to mention.

Second, after reading the text, although the English language is fine, I sometimes get the feeling pieces of text occur more than once, so I encourage the authors the see if they can clean up the text in certain places and shorten it down. In particular, references to the illustrations are extremely abundant; this could potentially have been done in a more efficient way, i.e. by merging several references to the same illustration.

My only scientific concern deals with the CSD searches. Looking at the structures I retrieved searching for the F1 fragment (see above), I noted that a very large fraction has O-H…S contacts as part of five-membered rings (see attached). As indicated for proteins on page 8, this would have a profound effect on the geometry of the interaction. I thus calculated phi/delta scatterplots for five-membered rings and other rings as well as intermolecular contacts, and they are indeed different (see attached). This would suggest that the selected ranges for δ may lie somewhat to low. Since there is nevertheless a clear difference between hydrogen bonds and chalcogen bonds, it does not impact the way phi and (in particular) delta values are used to distinguish between the two groups, but I nevertheless think this should be addressed in a revised version of the manuscript.

Other than this, I have only a series of minor comments and suggestions.

Page 6, line 3-6. The equation defines ΔE*AB and the text then continues with a discussion of ΔEs, which is evidently corrected for BSSE. Is it really useful to operate with both ΔE*AB and ΔEs, or could BSSE just be included throughout (with this being mentioned, obviously).

Page 6, line 10. MESP is used here for the first time, consider spelling it out (like PES on previous page).

Page 6, line -6: The reference to Figure 4 D here makes no sense.

Page 6, line -4: “CXXXXC” appears here like a Jack-in-a-box with no details provided. A few lines in the Introduction describing the reason for considering this motif would be nice.

Page 7, line -6: “The values for the limits were rounded off to the closest value, which was a multiple of 5 (Supplementary Figure S3 for θ and S4 for δ)”. The reference to Figure S3 is wrong. I assume Figure S4 is correct and that the text refers to the dashed boxes, but this is not really explained.

Page 8, line 3: Not sure I found the division of protein structures into two sets particularly revealing. Consider if you want to keep it.

Page 8, line -3: Reference to the selected 1.9 to 2.8 Å range?

Page 8, line -2: I am not familiar with the use of δ rather than <, i.e. 90° δ θ δ 140° rather than just 90° < θ < 140° as used in Figure 2 A, please explain. The frequent use of the length range notations could be simplified. Also note use on page 13, line -3, which is not clear.

Page 9, line 11-13: As the definitions are given on the top of the page, there is no need to repeat them here.

Page 10, line 7: Here “MetRS” pops out. Those not familiar with this protein need to read to page 21 to learn what this is. Consider adding some information in the introduction, or find another way to handle this.

Page 12, line 8-10: Definitions repeated again.

Page 16, line 10-11: The superscripts on S need to be explained.

Page 19, line 7: A few lines explaining what “negative design” represents in this context?

Page 24, line 14: In my manuscript there is a box in the n-pi* interaction.

Reference list:

I missed a reference to the work by Iwaoka & Babe, Phosphorous, Sulfur, and Silicon, 190:1257-1264, 2015.

Figure 2: Explain boxes. See initial comments on F1.

Figure 3 A. Give the value of the contour surfaces shown?

Recommendation: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R0/PR3

Comments

Comments to Author: Reviewer #1: The manuscript entitled “Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space” is nice piece of work. However, the following corrections/clarifications are needed before its publication.

1. Several important references are missing in the introduction part while discussing non-covalent interactions with sulfur (hydrogen bonding and chalcogen bonding). A few noticeable references are J. Phys. Chem. Lett. 2015, 8, 1385-1389, J. Am. Chem. Soc. 2017, 139, 1842−1855, Chem. Sci. 2017, 8, 2667−2670, Acc. Chem. Res. 2020, 53, 8, 1580-1592, J. Mol. Struct. 2020, 1211, 128080, Chem. Soc. Rev., 2022,51, 4261-4286. These references need to be cited and discussed in the revised manuscript.

2. Some of the important parameters such as lower electronegativity, charge and higher polarizability of sulfur should not be ignored while discussing hydrogen bonding with sulfur.

3. Why have different basis sets been used to optimize the monomer and dimer?

4. Why have the criteria for S---H bond distance taken less than 2.8 Å, whereas their sum of van der Walls radii is 3Å? Also, the S---N bond distance criteria were different for CSD (3.35 Å) and PDB (3.6 Å) search. The authors should provide a logical explanation for this.

5. A more detailed explanation should be given in the context of negligible sigma-hole formation on S when it is chelated to a metal center.

6. In the PDB analysis (page 8), the authors mentioned that coordinate files greater than 1 MB in size were excluded from both sets. What is the reason behind choosing based on file size?

7. Fluorescence-based studies:

Why are the experiments done at pH 7.6, slightly higher than the average body pH? Is there any particular reason for this?

The emission maximum for isolated tryptophan in a buffer medium and a protein environment is 350 nm±10 nm. However, in this report, it is 330 nm. Why such a blue shift in the tryptophan emission for MetRS?

For curiosity, can the author explain why fluorescence intensity changes upon ligand binding?

The Kd units in Figures 6C and G need to be correctly written for consistency.

Also, it would be more informative for the readers if the authors provided the emission spectra for titration of proteins and ligands in the supporting information.

8. There is no discussion on CD experiments in the manuscript’s main text. Instead of supplementary 10, the authors can write Figure S10 while discussing it. (page 23)

9. Did the authors perform ITC experiments for all the studied protein-ligand complexes like fluorescence experiments or only for MetRS:Met and MetRS:Nle ? For MetRS:Met complex, the enthalpy contribution is 40%, whereas the entropy contribution is 60% of the total free energy. However, the authors concluded this as prominently entropy-driven; how? Also, how the hydrophobic interactions are responsible for the entropy-driven process?

Reviewer #3: The contribution by Adhav et al. represents a very comprehensive and thorough investigation of chalcogen bonds in proteins well worth publishing.

As a general comment, the manuscript has obviously been in the making for quite some time, indicated for example by the CSD searches being executed on a rather old version of the database. One example, I searched for the F1 fragment in Figure 1 B and found 9435 fragment pairs rather than 7854 as indicated, with Nc going from 4.7 % as indicated to 6.1 %. I do not think more updated data would change anything important as far as the major results are concerned, but thought it worthwhile to mention.

Second, after reading the text, although the English language is fine, I sometimes get the feeling pieces of text occur more than once, so I encourage the authors the see if they can clean up the text in certain places and shorten it down. In particular, references to the illustrations are extremely abundant; this could potentially have been done in a more efficient way, i.e. by merging several references to the same illustration.

My only scientific concern deals with the CSD searches. Looking at the structures I retrieved searching for the F1 fragment (see above), I noted that a very large fraction has O-H…S contacts as part of five-membered rings (see attached). As indicated for proteins on page 8, this would have a profound effect on the geometry of the interaction. I thus calculated phi/delta scatterplots for five-membered rings and other rings as well as intermolecular contacts, and they are indeed different (see attached). This would suggest that the selected ranges for δ may lie somewhat to low. Since there is nevertheless a clear difference between hydrogen bonds and chalcogen bonds, it does not impact the way phi and (in particular) delta values are used to distinguish between the two groups, but I nevertheless think this should be addressed in a revised version of the manuscript.

Other than this, I have only a series of minor comments and suggestions.

Page 6, line 3-6. The equation defines ΔE*AB and the text then continues with a discussion of ΔEs, which is evidently corrected for BSSE. Is it really useful to operate with both ΔE*AB and ΔEs, or could BSSE just be included throughout (with this being mentioned, obviously).

Page 6, line 10. MESP is used here for the first time, consider spelling it out (like PES on previous page).

Page 6, line -6: The reference to Figure 4 D here makes no sense.

Page 6, line -4: “CXXXXC” appears here like a Jack-in-a-box with no details provided. A few lines in the Introduction describing the reason for considering this motif would be nice.

Page 7, line -6: “The values for the limits were rounded off to the closest value, which was a multiple of 5 (Supplementary Figure S3 for θ and S4 for δ)”. The reference to Figure S3 is wrong. I assume Figure S4 is correct and that the text refers to the dashed boxes, but this is not really explained.

Page 8, line 3: Not sure I found the division of protein structures into two sets particularly revealing. Consider if you want to keep it.

Page 8, line -3: Reference to the selected 1.9 to 2.8 Å range?

Page 8, line -2: I am not familiar with the use of δ rather than <, i.e. 90° δ θ δ 140° rather than just 90° < θ < 140° as used in Figure 2 A, please explain. The frequent use of the length range notations could be simplified. Also note use on page 13, line -3, which is not clear.

Page 9, line 11-13: As the definitions are given on the top of the page, there is no need to repeat them here.

Page 10, line 7: Here “MetRS” pops out. Those not familiar with this protein need to read to page 21 to learn what this is. Consider adding some information in the introduction, or find another way to handle this.

Page 12, line 8-10: Definitions repeated again.

Page 16, line 10-11: The superscripts on S need to be explained.

Page 19, line 7: A few lines explaining what “negative design” represents in this context?

Page 24, line 14: In my manuscript there is a box in the n-pi* interaction.

Reference list:

I missed a reference to the work by Iwaoka & Babe, Phosphorous, Sulfur, and Silicon, 190:1257-1264, 2015.

Figure 2: Explain boxes. See initial comments on F1.

Figure 3 A. Give the value of the contour surfaces shown?

Review: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R1/PR4

Conflict of interest statement

No conflict of interest.

Comments

Comments to Author: The authors have addressed the concerns raised by the reviewers in this revised manuscript. The manuscript can be accepted for publication in its current form.

Review: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R1/PR5

Conflict of interest statement

Reviewer declares none.

Comments

Comments to Author: In the revised version of the contribution by Adhav et al. the authors in my opinion have handled all issues raised by the editor and the two reviewers, including myself, in a satisfactory manner. My comments at this stage are only very minor, with no need for further review.

The English language generally reads very well, but I have some problems with mixing of plural and singular forms in the Abstract in particular. Consider the following corrections:

---

Divalent sulfur (S) formschalcogen bonds (Ch-bonds) via its σ-holes and hydrogen bonds (H-bonds) via its lone pairs. The relevance of these interactions and their interplay for protein structure and function is unclear. Based on an analyses of the crystal structures of small organic/organometallic molecules and proteins and their Molecular Electrostatic Surface Potential (MESP), we show that the reciprocity of the substituent-dependent strengths of the σ-holes and lone pairs correlates with the formation of either Ch-bonds or H-bonds. In proteins, disulfide-linked cysteines preferentially form Ch-bonds, metalchelated cysteines form H-bonds, while methionines form either of them with comparable frequencies. This has implications for the positioning of these residues and their role in protein structure and function. Computational analyses reveal that the S-mediated interactions stabilize protein secondary structures by mechanisms such as helix capping and protecting free β-sheet edges by negative-design. The study highlights the importance of S-mediated Ch-bonds and H-bonds for understanding protein folding and function, development of improved strategies for protein/peptide structure prediction and design, and structure-based drug discovery.

---

If you exclude “(MESP)”, Molecular Electrostatic Surface Potential should not be capitalized, I think. Now this abbreviation is introduced on page 5.

Considering the expression “disulfide-linked cysteines”, I realize this was introduced in response to a comment made by the editor, but personally I don’t like it since “sulfide-linked cysteines” cease to be “cysteines”, they are cystines (a name used only once, in Figure 5, in the current version of the manuscript).

The Results part of the manuscript is where most of the discussion takes place, the Discussion part being used more for concluding remarks. The editor might have a view on this matter.

Minor things:

Page 3, line 10: “… cysteines (Cys-Sγ) and cystines of proteins.”(referring to the above comment)

Page 3, line 12: “Consequently, methionine, cystine and cystine are …”

Page 3, line -1 and -3: Plural form “H-bonds” works better.

Page 3, line -1: Remove final word “bond”.

Page 4, line 7: First part of sentence “Identified in many crystal structures of small, supra, and biomolecules, …” does not make sense.

Page 4, line 16: “…, the precise role of the Ch-bond in protein structures …”

Page 4 line -4: “a H-hond”

Page 4, line -1: Write out AIM on this first occurrence (now done on page 6).

Page 5, line -1: “… in α-helices and β-strands …”

Page 6, line 6: “Fragments F1 - F6 provided in …”

Page 11, line -4: Consider to help the reader a bit by expanding the parenthesis to “(M-S-Y in Figure 4 A)”.

Page 22, legend to Figure 1. Part of the text for (B) obviously belongs to Table 1. The meaning of the asterisk * for F7 - F9 is not explained.

Page 22, legend to Figure 2. For (D) the legend reads “S···O/N contacts”, while in the Figure only “S···O” has been retained. I guess it is wrong in the Figure, as F5 and F6 are shown.

Recommendation: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R1/PR6

Comments

Comments to Author: Editor: I would like to thank the authors for their careful revision. I concur with reviewer #2 that the part most in need of further improvement is the abstract, and am grateful to the great suggestion. Personally, I would leave out any abbreviations from the abstract, though (and in this case decapitalize Molecular Electrostatic Surface Potential.

Regarding the minor suggestions of reviewer #2, it may be helpful to know that negative line numbers correspond to lines counted from the bottom of the page.

Finally, I wish to clarify that I have the same opinion as reviewer #2 regarding cystines. This was a misunderstanding by the authors: I did not ask them to remove cystines from the manuscript, but to carefully check if the correct term was used, as I discovered some that were incorrect (cystines referred to as cysteines or vice versa).

Please implement the final changes, such that we can go ahead publishing the article as soon as possible.

Reviewer #1: The authors have addressed the concerns raised by the reviewers in this revised manuscript. The manuscript can be accepted for publication in its current form.

Reviewer #2: In the revised version of the contribution by Adhav et al. the authors in my opinion have handled all issues raised by the editor and the two reviewers, including myself, in a satisfactory manner. My comments at this stage are only very minor, with no need for further review.

The English language generally reads very well, but I have some problems with mixing of plural and singular forms in the Abstract in particular. Consider the following corrections:

Divalent sulfur (S) formschalcogen bonds (Ch-bonds) via its σ-holes and hydrogen bonds (H-bonds) via its lone pairs. The relevance of these interactions and their interplay for protein structure and function is unclear. Based on an analyses of the crystal structures of small organic/organometallic molecules and proteins and their Molecular Electrostatic Surface Potential (MESP), we show that the reciprocity of the substituent-dependent strengths of the σ-holes and lone pairs correlates with the formation of either Ch-bonds or H-bonds. In proteins, disulfide-linked cysteines preferentially form Ch-bonds, metalchelated cysteines form H-bonds, while methionines form either of them with comparable frequencies. This has implications for the positioning of these residues and their role in protein structure and function. Computational analyses reveal that the S-mediated interactions stabilize protein secondary structures by mechanisms such as helix capping and protecting free β-sheet edges by negative-design. The study highlights the importance of S-mediated Ch-bonds and H-bonds for understanding protein folding and function, development of improved strategies for protein/peptide structure prediction and design, and structure-based drug discovery.

If you exclude “(MESP)”, Molecular Electrostatic Surface Potential should not be capitalized, I think. Now this abbreviation is introduced on page 5.

Considering the expression “disulfide-linked cysteines”, I realize this was introduced in response to a comment made by the editor, but personally I don’t like it since “sulfide-linked cysteines” cease to be “cysteines”, they are cystines (a name used only once, in Figure 5, in the current version of the manuscript).

The Results part of the manuscript is where most of the discussion takes place, the Discussion part being used more for concluding remarks. The editor might have a view on this matter.

Minor things:

Page 3, line 10: “… cysteines (Cys-Sγ) and cystines of proteins.”(referring to the above comment)

Page 3, line 12: “Consequently, methionine, cystine and cystine are …”

Page 3, line -1 and -3: Plural form “H-bonds” works better.

Page 3, line -1: Remove final word “bond”.

Page 4, line 7: First part of sentence “Identified in many crystal structures of small, supra, and biomolecules, …” does not make sense.

Page 4, line 16: “…, the precise role of the Ch-bond in protein structures …”

Page 4, line -1: Write out AIM on this first occurrence (now done on page 6).

Page 5, line -1: “… in α-helices and β-strands …”

Page 6, line 6: “Fragments F1 - F6 provided in …”

Page 11, line -4: Consider to help the reader a bit by expanding the parenthesis to “(M-S-Y in Figure 4 A)”.

Page 22, legend to Figure 1. Part of the text for (B) obviously belongs to Table 1. The meaning of the asterisk * for F7 - F9 is not explained.

Page 22, legend to Figure 2. For (D) the legend reads “S···O/N contacts”, while in the Figure only “S···O” has been retained. I guess it is wrong in the Figure, as F5 and F6 are shown.

Recommendation: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R2/PR7

Comments

No accompanying comment.

Recommendation: Sulfur-mediated chalcogen versus hydrogen bonds in proteins: a seesaw effect in the conformational space — R3/PR8

Comments

No accompanying comment.