Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2016 Jun 8;25(8):1517–1524. doi: 10.1002/pro.2955

Analysis of the interactions of sulfur‐containing amino acids in membrane proteins

José C Gómez‐Tamayo 1,†, Arnau Cordomí 1,†, Mireia Olivella 2, Eduardo Mayol 1, Daniel Fourmy 3, Leonardo Pardo 1,✉
PMCID: PMC4972207  PMID: 27240306

Abstract

The interactions of Met and Cys with other amino acid side chains have received little attention, in contrast to aromatic–aromatic, aromatic–aliphatic or/and aliphatic–aliphatic interactions. Precisely, these are the only amino acids that contain a sulfur atom, which is highly polarizable and, thus, likely to participate in strong Van der Waals interactions. Analysis of the interactions present in membrane protein crystal structures, together with the characterization of their strength in small‐molecule model systems at the ab‐initio level, predicts that Met–Met interactions are stronger than Met–Cys ≈ Met–Phe ≈ Cys–Phe interactions, stronger than Phe–Phe ≈ Phe–Leu interactions, stronger than the Met–Leu interaction, and stronger than Leu–Leu ≈ Cys–Leu interactions. These results show that sulfur‐containing amino acids form stronger interactions than aromatic or aliphatic amino acids. Thus, these amino acids may provide additional driving forces for maintaining the 3D structure of membrane proteins and may provide functional specificity.

Keywords: van der Waals interactions, membrane proteins, sulfur‐containing amino acids, mining of crystal structures

Abbreviations

BNZ

benzene

DME

dimethyl ether

DMS

dimethyl sulfide

MT

methanethiol

PRP

propane

Introduction

Non‐bonded interactions are crucial for protein stability, function, and ligand binding. These comprise electrostatic (including hydrogen bonds) and van der Waals (dipole–dipole, dipole–induced dipole, and induced dipole–induced dipole) interactions.1 All these type of interactions have been extensively characterized in terms of strength, directionality, and physicochemical properties.2 However, their prevalence and importance vary depending on whether or not they occur in membrane or globular proteins due to their different environment. Both globular and membrane proteins position hydrophobic amino acid side chains toward the protein core, and maximize hydrogen bond interactions among backbone atoms. However, in contrast to soluble proteins, the hydrophobic nature of the lipid bilayer imposes that residues pointing toward the membrane are also hydrophobic. Thus, dispersion forces (mainly aromatic–aromatic, aromatic–aliphatic, or aliphatic–aliphatic) are involved in stabilizing the tertiary structure of the protein or in structural changes.2, 3, 4, 5, 6, 7

As polarizabilities of the two interacting partners become larger, van der Waals forces become stronger. For example, the aromatic ring of aromatic amino acids has a quadrupole π system that is highly polarizable and provides strong aromatic–aromatic dispersion interactions. Thus, aromatic side chains importantly contribute to the folding and thermodynamic stability of proteins.8 Similarly, sulfur‐containing amino acids are also highly polarizable, as sulfur has filled 3p and empty 3d orbitals and contain a permanent dipole.9 Surprisingly, non‐bonded interactions (dipole‐induced dipole or dispersion) involving sulfur‐containing amino acids (Met and Cys) have received little attention10, 11, 12 in contrast to interactions involving aromatic amino acids.2 More than 30 years ago, Morgan et al. observed a high frequency of contacts between sulfur‐containing residues and aromatic residues in proteins, and identified large stacked arrangements composed of aromatic and Met or Cys residues.13 Further studies also demonstrated that Cys‐ and Met‐aromatic interactions were fairly common in protein crystal structures.12, 14, 15

In the present work, we aim to evaluate the occurrence of interactions involving Met and Cys side‐chains in crystal structures of membrane proteins and to characterize their strength in small‐molecule model systems at the ab‐initio level. The employed level of theory improves previous calculations in analogous systems.16, 17, 18, 19 Our results show that Met–Met, Met–Phe, Met–Leu, and Cys–Phe interactions are stronger in magnitude than Phe–Phe interactions.

Results and Discussion

Structural bioinformatics analysis of the presence of Cys and Met in membrane proteins

Table 1 summarizes the occurrence of the most frequent amino acids, together with Cys and Met, in the transmembrane region (i.e., excluding water‐soluble domains or loops) of α‐helix bundles of membrane proteins with known crystal structure (see Methods section). Amino acids such as Leu, Ile, Val, and Phe are the most frequent at the membrane‐embedded region, without preference for being localized in the protein core or in the membrane‐exposed region. In contrast, sulfur‐containing amino acids (i.e., Met and Cys) show lower frequencies and are mostly found buried in the core of the protein. This might indicate a functional role in stabilizing the 3D structure. Analysis of inter‐residue interaction of Met and Cys reveals that a significant percentage of Met residues form interactions with aliphatic residues (Leu, Ala, Ile, Val) and Phe, while a smaller proportion interact with Met and Cys (Table 1). Thus, we aim to determine the orientation and strength of these interactions of sulfur‐containing amino acids.

Table 1.

Structural Bioinformatics Analysis of Membrane Proteins

Amino acid distribution Side chain–side chain interactions
Total Buried Exposed Cys Met Phe Val Ile Ala Leu
Leu 8,896 (17%) 3,787 (43%) 5,111 (57%) 474 1,129 2,415 3,427 2,491 3,764 2,319
Ala 6,198 (12%) 3,988 (64%) 2,209 (36%) 396 795 1,616 3,437 2,42 2,039
Ile 5,801 (11%) 2,493 (43%) 3,309 (57%) 401 806 1,671 2,126 723
Val 5,761 (11%) 2,826 (49%) 2,935 (51%) 357 700 1,276 1,363
Phe 4,651 (9%) 2,464 (53%) 2,188 (47%) 276 707 859
Met 1,933 (4%) 1,336 (69%) 597 (31%) 111 199
Cys 768 (1.4%) 557 (73%) 211 (27%) 37
Open in a new tab

Amino acid type distribution (absolute frequencies and relative frequencies in percentage) observed in the survey of transmembrane domains of α‐helix bundles (the five most frequent residues and Met and Cys) classified as buried or exposed to the membrane. The most significant side–side chain interactions of the five most frequent amino acids and Met and Cys.

The orientation and strength of Met–Phe interactions

The Met–Phe interactions identified in the crystal structures of membrane proteins (a total of 707 pairs, Table 1) were clustered based on relative distances and angles between the two amino acids (see Methods section). Figure 1 shows the 2D histograms with the distribution of the Met–Phe interactions projected on the conformational space defined by P and θ angles (see Methods section and Supporting Information Table S1). In order to evaluate the magnitude of the energy of interaction between both side chains, we performed high level ab initio calculations (see Methods section) in small‐molecule models systems [Met and Phe were mimicked by dimethyl sulfide (DMS) and the benzene ring (BNZ), respectively, Supporting Information Fig. S1]. Figure 2 shows that the interaction energy along the distance between DMS and BNZ exhibits a wide minimum located between 4 and 6 Å. Clusters I (containing 24% of the observed interactions) and II (37%) reproduce the most favorable arrangements of the side chains (−2.9 kcal/mol) as calculated in comparable model systems 1 and 2. Arabic numbers depict optimized ab initio models whereas roman numbers represent clusters observed in crystal structures. The two planes defined by DMS and BNZ molecules are almost parallel in Model 1 and perpendicular in Model 2, but in both cases a methyl group of DMS is located on top of the negative charge density at the center of BNZ ring (π electrons) and the sulfur atom on top of the positive charged density at the exterior of BNZ ring (–CH groups). In Cluster III (12%) the CH atoms of Met are pointing to the aromatic ring of Phe and the sulfur atom is pointing toward opposite direction, which results in an interaction of −2.4 kcal/mol in Model 3. Finally, Cluster IV (28%) accounts for interactions in which the sulfur atom of Met acts as hydrogen bond acceptor for a –CH group from the phenyl ring of Phe. The interaction energy was −2.0 kcal/mol in the comparable Model 4. Interestingly, these computed energies are of the same magnitude as the values experimentally obtained for peptides in water.20

Figure 1.

Figure 1

Open in a new tab

The orientation and strength of Met–Phe interactions. 2D histograms of the frequencies of occurrence of these interactions, clustered according to the conformational space defined by the distance d and the angles P and θ (see Supporting Information Table S1 for a detailed description). Roman and arabic numbers indicate the position in the 2D histogram of the most representative structure in the cluster and the energy‐minimized structure, respectively. Ab initio geometry optimization at the MP2/6‐31 + G(d,p) level and calculated energy of interaction at the CCSD(T)/6‐311 + G(3df,2p) level (see Methods) are shown inside dotted circles as solid sticks. Representative structures obtained in the cluster analysis are shown as transparent sticks.

Figure 2.

Figure 2

Open in a new tab

Influence of the distance on the interaction energy. Calculations were done on the DMS–DMS (Met–Met), DMS–BNZ (Met–Phe), DMS–PRP (Met–Leu), PRP–PRP (Leu–Leu), and PRP–BNZ (Leu–Phe) model systems with the lowest energy. d refers either to the sulfur–sulfur or the sulfur–benzene (centroid) distance.

In order to study the influence in the energy of interaction of the highly polarizable sulfur, compared to oxygen or the methylene group, we performed analogous ab initio calculations with model compounds that replace the sulfur atom (dimethyl sulfide, DMS, mimicking Met) by oxygen (dimethyl ether, DME) or –CH2– (propane, PRP, mimicking aliphatic amino acids) (Supporting Information Fig. S1). Comparison of these energies of interactions of Model compounds 1–4 in DMS–BNZ complexes with analogous conformations of DME–BNZ shows that in all cases the sulfur‐containing molecule (DMS) interacts stronger with the aromatic ring (BNZ) than in the oxygen‐containing one (DME) with the exception of Model 4. This suggests that the induced positive charge density on the methyl group, involved in the interaction with the π electrons of the ring in Models 1–3, is larger in the presence of the sulfur atom than in the presence of oxygen. Reasonably, because sulfur is a poorer hydrogen bond acceptor, in Model 4 the sulfur atom forms weaker S···HC hydrogen bond interaction with the CH group of the ring than oxygen (O···HC). Importantly, the energies of interactions of Model compounds 1–4 in DMS–BNZ complexes are always more stable than in PRP–BNZ complexes, indicating that the interaction of aromatic rings with sulfur‐containing groups is always stronger than with aliphatic groups.

Because aromatic–aromatic interactions are considered key in the stability of membrane proteins,8 we next compared the energies of interaction in DMS–BNZ complexes with those in BNZ–BNZ complexes. Comparison with the well‐characterized21, 22, 23, 24 lowest energy configurations of BNZ–BNZ (Supporting Information Fig. S2), the T‐shaped (−2.4 kcal/mol), and parallel displaced (−2.1 kcal/mol), indicates that Met forms more stable interactions with aromatics rings (DMS–BNZ) than aromatic–aromatic interactions (BNZ–BNZ).

The orientation and strength of Met–Met interactions

Clusters I–V in Figure 3, obtained from the 199 Met–Met interactions present in membrane proteins (Table 1), are calculated in a similar way to the clusters of Met–Phe (see above). Cluster I, containing 11% of the interactions, corresponds to an anti‐parallel orientation, exhibiting the lowest interaction energy (−3.5 kcal/mol) in the comparable Model system 1. The orientation of the Met side chains in Cluster II (46%) is in the T‐shaped configuration, being the interaction energy of −3.0 kcal/mol in the comparable Model system 2. In these configurations 1 and 2 each sulfur atom interacts respectively with four and three hydrogen atoms of the methyl groups (S···HC interactions) that have positive charge density. The Met side chains in Cluster III (5%) are in a parallel‐displaced orientation, in a head‐to‐head configuration with both sulfur atoms engaged in the interaction (−2.2 kcal/mol in Model 3). Clusters IV (15%) and V (25%) account for the least favored Met–Met interactions (−1.5 and −1.3 kcal/mol in Models 4 and 5, respectively). Models mimicking these clusters reproduce a structure with a single sulfur atom interacting with four and two CH hydrogen atoms (S···HC interactions), respectively. Cluster IV shows a parallel orientation in a head‐to‐tail configuration of the Met side‐chains, while Cluster V shows a T‐shaped orientation in which the interactions occur through the methyl groups.

Figure 3.

Figure 3

Open in a new tab

The orientation and strength of Met–Met interactions. 2D histograms of the frequencies of occurrence of these interactions, clustered according to the conformational space defined by the distance d and the angles P and θ (Supporting Information Table S2). See legend of Figure 1 for further details.

The influence of the highly polarizable sulfur atom in these interactions was evaluated by performing analogous ab initio calculations with model compounds that replace the sulfur atom (Supporting Information Fig. S3) (dimethyl sulfide, DMS) by oxygen (dimethyl ether, DME). The anti‐parallel orientation of Model 1 (−3.5 vs. −2.7 kcal/mol) and the T‐shaped configuration of Model 2 (−3.0 vs. −2.8 kcal/mol) are more stable in the sulfur‐containing DMS–DMS complex than in the oxygen‐containing DME–DME complex. The opposite is observed for Models 3 (−2.2 vs. −2.4 kcal/mol), 4 (−1.5 vs. −1.6 kcal/mol), and 5 (−1.3 vs. −1.5 kcal/mol).

The orientation and strength of Met–Leu interactions

We have selected Leu as a representative residue to study the interactions of Met with aliphatic amino acids. The 1129 Met–Leu interactions present in membrane proteins (Table 1) were clustered in a similar manner as in Met–Met interactions (see above). Because the Cγ, Sδ, and Cε atoms of Met are analogous to the Cδ1, Cγ, and Cδ2 atoms of Leu, the relative orientation of Met–Leu residues in Clusters I–V (Fig. 4) were taken in analogy with Clusters I–V of Met–Met (Fig. 3). The computed interaction energies in comparable Model systems 1–5 (Supporting Information Fig. S4) show that the anti‐parallel orientation in Model 1 exhibits the largest interaction energy (−2.1 kcal/mol). Comparison of the interaction energies in DMS–DMS (Met–Met clusters), DMS–PRP (Met–Leu clusters), and PRP–PRP (Leu–Leu clusters, not shown) models allow us to study the influence of the sulfur atom in the interaction energy. Clearly, the rank order of energies on interaction is DMS–DMS (2 sulfur atoms) < DMS–PRP (1 sulfur atom) < PRP–PRP (0 sulfur atom) in Models 1 (−3.5 < −2.1 < −1.7 kcal/mol, respectively), 2 (−3.0 < −1.5 < −1.4 kcal/mol), 3 (−2.2 < −1.4 < −1.0 kcal/mol), 4 (−1.5 < −1.3 < −1.1 kcal/mol), and 5 (−1.3 < −1.2 < −1.0 kcal/mol).

Figure 4.

Figure 4

Open in a new tab

The orientation and strength of Met–Leu interactions. 2D histograms of the frequencies of occurrence of these interactions, clustered according to the conformational space defined by the distance d and the angles P and θ (Supporting Information Table S3). See legend of Figure 1 for further details.

The interactions of Cys

Cys interacts with Phe (a total of 276 pairs), Met (111 pairs), and hydrophobic amino acids such as Leu (474 pairs), Ala (396 pairs), Ile (401 pairs), and Val (357 pairs) (Table 1). In addition, Cys can interact with other Cys through a covalent disulfide bridge. Excluding disulfide bridges (S—S distances < 3 Å) only 37 Cys–Cys pairs were observed in crystal structures (Table 1) in which Cys was acting as hydrogen bond donor and/or acceptor. These Cys–Cys interactions were not further analyzed, as they belong to the common hydrogen bond interaction. Analysis of the crystal structures of Cys–Phe interactions in membrane proteins revealed three main interaction modes (Fig. 5). In Cluster I (12%), the sulfur Sγ atom is located on top of the aromatic ring, in Cluster II (49%) the Cβ atom is located on top of the ring, and in Cluster III (39%) the sulfur Sγ atom is coplanar to the phenyl ring. Ab initio energy optimizations of model compounds [Cys and Phe were mimicked by methanethiol (MT) and the benzene ring (BNZ), respectively, Supporting Information Fig. S5] positioned the sulfhydryl hydrogen absent in the crystal structures. In Model 1 (−3.0 kcal/mol) MT forms a S—H··π hydrogen bond with the phenyl ring, whereas in Model 2 (−2.9 kcal/mol) the S atom of MT is located on top of the positive charged density at the exterior of the BNZ ring and the methyl group on top of the negative charge density at the center of the ring. Importantly, ab initio energy minimization of Model compound 3, mimicking Cluster III, led either to Models 1 or 2.

Figure 5.

Figure 5

Open in a new tab

The orientation and strength of Cys–Phe interactions. 2D histograms of the frequencies of occurrence of these interactions, clustered according to the conformational space defined by the distance d and the θ angle. See legend of Figure 1 for further details.

Clustering of the 111 Cys–Met and 474 Cys–Leu interactions (Table 1) was challenging due to the absence of the sulfhydryl hydrogen in the crystal structures. Thus, we performed ab initio energy optimizations of model compounds (MT–DMS or MT–PRP, Supporting Information Fig. S5) in which one of the methyl groups of DMS, in the reported Met–Met (DMS–DMS, Supporting Information Fig. S3) and Met–Leu (DMS–PRP, Supporting Information Fig. S4) interactions, was replaced by hydrogen. Comparison of these energies of interaction in MT–DMS complexes (−3.0, −2.9, −1.4, −1.7, −1.3 kcal/mol for Models 1–5, respectively) with analogous conformations of DMS–DMS (−3.5, −3.0, −2.2, −1.5, −1.3 kcal/mol for 1–5, respectively) shows that DMS–DMS interactions are stronger with the exception of Model 4.

Interaction energy comparison with the AMBER force field

We computed the interaction energies of the small‐molecule model compounds using the AMBER force field,25 with the aim of assessing the accuracy of this force field in reproducing the interactions of Met or Cys (see Methods section). The results shown in Supporting Information Figure S5 indicate a reasonable quantitative agreement in the interactions of Cys (mimicked by MT) with Met (DMS), Leu (PRP), and Phe (BNZ). The deviations are larger for Met (Supporting Information Figs. S1, S3–S4) with an average difference relative to CCSD(T) of ∼0.5 kcal/mol. The larger deviations correspond to conformations in which sulfur atoms are in close proximity or when the sulfur atom is located on top of the benzene ring. Moreover, the rank order of Met–Phe interactions is not fully reproduced: CCSD(T) predicts 1 = 2 < 3 < 4 in DMS–BNZ models, while AMBER predicts 3 < 1 < 2 < 4 (Supporting Information Fig. S1). Similarly, CCSD(T) predicts 1 < 2 < 3 < 4 < 5 in DMS–DMS models, while AMBER predicts 1 < 2 < 4 < 3 = 5 (Supporting Information Fig. S3). In contrast, the interactions of Met with Leu are highly consistent, both in magnitude and rank order (Supporting Information Fig. S4). Overall, these results are in line with a recent report on π–π, CH/π, and SH/π interactions.26

Conclusions

Others and we have previously outlined the structural and functional role of Met–aromatic and Met–Met interactions in the family of G protein‐coupled receptors.35, 36, 37 In the present report, we addressed a quantitative characterization of such interactions in membrane proteins. The analysis of the inter‐residue interactions in crystal structures of membrane protein revealed that Met and Cys often interact with Leu, Ile, Val, Phe, and other Met or Cys. The characterization of their strength using ab‐initio calculations in small‐molecule model systems, predicted that Met–Met, Met–Phe, Cys–Phe, Met–aliphatic, and Cys–aliphatic interactions are stronger in magnitude than aliphatic–aliphatic interactions. Remarkably, Met–Met, Met–Phe, and Cys–Phe interactions are stronger than aromatic–aromatic. Thus, we can conclude that these types of interactions, which have often been misled, need to be taken into account when considering the forces that stabilize the overall fold in membrane proteins. In addition to the stronger interactions of Met and Cys, their more flexible side‐chains may provide extra versatility and adaptation to conformational changes. We believe that these interactions are also likely to be important in the interior of globular proteins or in the formation of protein–ligand or protein–protein complexes. However, further studies would be necessary in these regard.

Material and Methods

Analysis of membrane protein structures

A non‐redundant dataset of 327 α‐helical transmembrane bundles was taken from TMalphaDB.27 This data set consists of crystallographic structures deposited in the Protein Data Bank28 with resolution <3.5 Å.

Residues were classified, based on their circular variance29 (CV) of vectors drawn from the Cα atom of a given residue to the Cα atoms of neighbor residues, as exposed (CV > 0.7) or buried (CV ≤ 0.7) to the membrane. Met–Met, Met–Phe, Met–Leu, Cys–Met, Cys–Phe, and Cys–Leu interactions were considered if the distance d between the two side‐chains (measured as the distance between the atoms Sδ of Met, Sγ of Cys or Cγ of Leu or the centroid of the aromatic ring of Phe) was < 6 Å. The relative orientation of the interacting side‐chains was defined by the distance d, the angle P between side‐chain planes (each plane defined by atoms Cγ, Sδ, and Cε of Met; Cα, Cβ, and Sγ of Cys; Cδ1, Cγ, and Cδ2 of Leu; and the aromatic ring of Phe), and the angle θ between the plane defined by side‐chain A and the vector connecting the central atoms of each side‐chain (A and B). Definitions of P and θ angles were those used by Chakrabarti and Bhattacharyya to describe benzene dimer geometries.30 These interactions were clustered according to the conformational space defined by the distance d and the angles P and θ (see Supporting Information Tables S1–S3 for a detailed description).

Quantum mechanical calculations

For the representative structure of each cluster, the energy of interaction between side chains was calculated using ab initio methods on small‐molecule models systems: Met was mimicked by dimethyl sulfide (DMS), Cys by methanethiol (MT), Leu by propane (PRP), and Phe by benzene (BNZ). All chosen model structures were optimized at the MP2/6‐31 + G(d,p) level of theory, which has been shown to provide reasonably good geometries.31, 32 Next, single point energy calculations were performed at the CCSD(T)/6‐311 + G(3df,2p) level. In order to minimize the basis set superposition error, counterpoise method by Boys and Bernardi33 was utilized. Moreover, we also computed the interaction energies using the AMBER force field 25 (see Supporting Information Figs. S1–S5) in order to evaluate the ability of protein force fields in reproducing these interactions. All calculations were performed using GAUSSIAN09 program.34

Supporting information

Supporting Information

Click here for additional data file. (985.5KB, pdf)

References

  • 1. Muller‐Dethlefs K, Hobza P (2000) Noncovalent interactions: a challenge for experiment and theory. Chem Rev 100:143–167. [DOI] [PubMed] [Google Scholar]
  • 2. Meyer EA, Castellano RK, Diederich F (2003) Interactions with aromatic rings in chemical and biological recognition. Angew Chem 42:1210–1250. [DOI] [PubMed] [Google Scholar]
  • 3. Nishio M (2004) CH/pi hydrogen bonds in crystals. CrystEngComm 6:130–158. [Google Scholar]
  • 4. Brandl M, Weiss MS, Jabs A, Suhnel J, Hilgenfeld R (2001) C‐H…pi‐interactions in proteins. J Mol Biol 307:357–377. [DOI] [PubMed] [Google Scholar]
  • 5. Kim K, Karthikeyan S, Singh J (2011) How different are aromatic π interactions from aliphatic π interactions and non‐π stacking interactions? J Chem Theory Comput 7:3471–3477. [DOI] [PubMed] [Google Scholar]
  • 6. Tsuzuki S, Honda K, Uchimaru T, Mikami M. (2006) Estimated MP2 and CCSD(T) interaction energies of n‐alkane dimers at the basis set limit: comparison of the methods of Helgaker and Feller. J Chem Phys 124:114304. [DOI] [PubMed] [Google Scholar]
  • 7. Ringer AL, Figgs MS, Sinnokrot MO, Sherrill CD (2006) Aliphatic C‐H/pi interactions: methane‐benzene, methane‐phenol, and methane‐indole complexes. J Phys Chem A 110:10822–10828. [DOI] [PubMed] [Google Scholar]
  • 8. Hong H, Park S, Jimenez RH, Rinehart D, Tamm LK (2007) Role of aromatic side chains in the folding and thermodynamic stability of integral membrane proteins. J Am Chem Soc 129:8320–8327. [DOI] [PubMed] [Google Scholar]
  • 9. Grossfield A, Woolf TB (2002) Interaction of tryptophan analogs with POPC lipid bilayers investigated by molecular dynamics calculations. Langmuir 18:198–210. [Google Scholar]
  • 10. Valley CC, Cembran A, Perlmutter JD, Lewis AK, Labello NP, Gao J, Sachs JN (2012) The methionine‐aromatic motif plays a unique role in stabilizing protein structure. J Biol Chem 287:34979–34991. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Pal D, Chakrabarti P (1998) Different types of interactions involving cysteine sulfhydryl group in proteins. J Biomol Struct Dyn 15:1059–1072. [DOI] [PubMed] [Google Scholar]
  • 12. Pal D, Chakrabarti P (2001) Non‐hydrogen bond interactions involving the methionine sulfur atom. J Biomol Struct Dyn 19:115–128. [DOI] [PubMed] [Google Scholar]
  • 13. Morgan RS, Tatsch CE, Gushard RH, McAdon J, Warme PK (1978) Chains of alternating sulfur and pi‐bonded atoms in eight small proteins. Int J Pept Protein Res 11:209–217. [DOI] [PubMed] [Google Scholar]
  • 14. Zauhar RJ, Colbert CL, Morgan RS, Welsh WJ (2000) Evidence for a strong sulfur‐aromatic interaction derived from crystallographic data. Biopolymers 53:233–248. [DOI] [PubMed] [Google Scholar]
  • 15. Samanta U, Pal D, Chakrabarti P (2000) Environment of tryptophan side chains in proteins. Proteins 38:288–300. [PubMed] [Google Scholar]
  • 16. Pranata J (1997) Sulfur aromatic interactions: a computational study of the dimethyl sulfide benzene complex. Bioorg Chem 25:213–219. [Google Scholar]
  • 17. Cabaleiro‐Lago EM, Hermida‐Ramon JM, Rodriguez‐Otero J (2004) Computational study of the interaction in (CH3)(2)X dimer and trimer (X = O, S). J Phys Chem A 108:4923–4929. [Google Scholar]
  • 18. Duan GL, Smith VH, Weaver DF (2001) Characterization of aromatic‐thiol pi‐type hydrogen bonding and phenylalanine‐cysteine side chain interactions through ab initio calculations and protein database analyses. Mol Phys 99:1689–1699. [Google Scholar]
  • 19. Ringer AL, Senenko A, Sherrill CD (2007) Models of S/pi interactions in protein structures: comparison of the H2S benzene complex with PDB data. Protein Sci 16:2216–2223. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Viguera AR, Serrano L (1995) Side‐chain interactions between sulfur‐containing amino acids and phenylalanine in alpha‐helices. Biochemistry 34:8771–8779. [DOI] [PubMed] [Google Scholar]
  • 21. Janowski T, Pulay P (2007) High accuracy benchmark calculations on the benzene dimer potential energy surface. Chem Phys Lett 447:27–32. [Google Scholar]
  • 22. DiStasio RA, Jr , von Helden G, Steele RP, Head‐Gordon M (2007) On the T‐shaped structures of the benzene dimer. Chem Phys Let 437:277–283. [Google Scholar]
  • 23. Sinnokrot MO, Sherrill CD (2006) High‐accuracy quantum mechanical studies of pi‐pi interactions in benzene dimers. J Phys Chem A 110:10656–10668. [DOI] [PubMed] [Google Scholar]
  • 24. Tsuzuki S, Honda K, Uchimaru T, Mikami M, Tanabe K (2002) Origin of attraction and directionality of the pi/pi interaction: model chemistry calculations of benzene dimer interaction. J Am Chem Soc 124:104–112. [DOI] [PubMed] [Google Scholar]
  • 25. Cornell WD, Cieplak P, Bayly CI, Gould IR, Merz KM, Ferguson DM, Spellmeyer DC, Fox T, Caldwell JW, Kollman PA (1995) A 2nd generation force‐field for the simulation of proteins, nucleic‐acids, and organic‐molecules. J Am Chem Soc 117:5179–5197. [Google Scholar]
  • 26. Sherrill CD, Sumpter BG, Sinnokrot MO, Marshall MS, Hohenstein EG, Walker RC, Gould IR (2009) Assessment of standard force field models against high‐quality ab initio potential curves for prototypes of pi‐pi, CH/pi, and SH/pi interactions. J Comput Chem 30:2187–2193. [DOI] [PubMed] [Google Scholar]
  • 27. Perea M, Lugtenburg I, Mayol E, Cordomi A, Deupi X, Pardo L, Olivella M (2015) TMalphaDB and TMbetaDB: web servers to study the structural role of sequence motifs in alpha‐helix and beta‐barrel domains of membrane proteins. BMC Bioinformatics 16:266. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28. Berman HM, Westbrook J, Feng Z, Gilliland G, Bhat TN, Weissig H, Shindyalov IN, Bourne PE (2000) The Protein Data Bank. Nucleic Acids Res 28:235–242. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Mezei M (2003) A new method for mapping macromolecular topography. J Mol Graph Model 21:463–472. [DOI] [PubMed] [Google Scholar]
  • 30. Chakrabarti P, Bhattacharyya R (2007) Geometry of nonbonded interactions involving planar groups in proteins. Prog Biophys Mol Biol 95:83–137. [DOI] [PubMed] [Google Scholar]
  • 31. Riley KE, Platts JA, Rezac J, Hobza P, Hill JG (2012) Assessment of the performance of MP2 and MP2 variants for the treatment of noncovalent interactions. J Phys Chem A 116:4159–4169. [DOI] [PubMed] [Google Scholar]
  • 32. Hobza P, Selzle HL, Schlag EW (1996) Potential energy surface for the benzene dimer. Results of ab initio CCSD(T) calculations show two nearly isoenergetic structures: T‐shaped and parallel‐displaced. J Phys Chem 100:18790–18794. [Google Scholar]
  • 33. Boys SF, Bernardi F (1970) The calculation of small molecular interactions by the differences of separate total energies. Some procedures with reduced errors. Mol Phys 19:553–566. [Google Scholar]
  • 34.Gaussian 09, Revision E.01. 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, Sonnenberg JL, Hada M, Ehara M, Toyota K, Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O, Nakai H, Vreven T, Montgomery JA, Peralta JE, Ogliaro F, Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN, Kobayashi R, Normand J, Raghavachari K, Rendell A, Burant JC, Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene M, Knox JE, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts R, Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C, Ochterski JW, Martin RL, Morokuma K, Zakrzewski VG, Voth GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD, Farkas, Foresman JB, Ortiz J. V, Cioslowski J, Fox DJ. Gaussian, Inc, Wallingford CT; 2009.
  • 35. Cordomi A, Gomez‐Tamayo JC, Gigoux V, Fourmy D (2013) Sulfur‐containing amino acids in 7TMRs: molecular gears for pharmacology and function. Trends Pharmacol Sci 34:320–331. [DOI] [PubMed] [Google Scholar]
  • 36. Nygaard R, Zou Y, Dror RO, Mildorf TJ, Arlow DH, Manglik A, Pan AC, Liu CW, Fung JJ, Bokoch MP, Thian FS, Kobilka TS, Shaw DE, Mueller L, Prosser RS, Kobilka BK (2013) The dynamic process of beta(2)‐adrenergic receptor activation. Cell 152:532–542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Magnan R, Escrieut C, Gigoux V, De K, Clerc P, Niu F, Azema J, Masri B, Cordomi A, Baltas M, Tikhonova IG, Fourmy D (2013) Distinct CCK‐2 receptor conformations associated with beta‐arrestin‐2 recruitment or phospholipase‐C activation revealed by a biased antagonist. J Am Chem Soc 135:2560–2573. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

Click here for additional data file. (985.5KB, pdf)

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES