Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2005 May;14(5):1328-39.
doi: 10.1110/ps.041222905. Epub 2005 Mar 31.

Improved side-chain modeling for protein-protein docking

Chu Wang  1 , Ora Schueler-Furman, David Baker
Affiliations

Improved side-chain modeling for protein-protein docking

Chu Wang et al. Protein Sci. 2005 May.

Abstract

Success in high-resolution protein-protein docking requires accurate modeling of side-chain conformations at the interface. Most current methods either leave side chains fixed in the conformations observed in the unbound protein structures or allow the side chains to sample a set of discrete rotamer conformations. Here we describe a rapid and efficient method for sampling off-rotamer side-chain conformations by torsion space minimization during protein-protein docking starting from discrete rotamer libraries supplemented with side-chain conformations taken from the unbound structures, and show that the new method improves side-chain modeling and increases the energetic discrimination between good and bad models. Analysis of the distribution of side-chain interaction energies within and between the two protein partners shows that the new method leads to more native-like distributions of interaction energies and that the neglect of side-chain entropy produces a small but measurable increase in the number of residues whose interaction energy cannot compensate for the entropic cost of side-chain freezing at the interface. The power of the method is highlighted by a number of predictions of unprecedented accuracy in the recent CAPRI (Critical Assessment of PRedicted Interactions) blind test of protein-protein docking methods.

PubMed Disclaimer

Figures

Figure 1.
Figure 1.
Rotamer approximation of side-chain conformations restricts accurate docking of 1CHO. (A) Low-RMSD model with a high energy due to side-chain clashes between Trp172 and Trp215. (B) High-RMSD model with a low energy with the clashes relieved. The predicted models are superimposed on the native complex structure based on the receptor backbone. The receptor backbone is colored gray. The ligand and the receptor Trp172 and Trp215 side chains in the native complex structure and in the predicted structure are colored black and gray, respectively.
Figure 2.
Figure 2.
RTMIN improves side-chain repacking and sequence redesign in monomeric proteins. (A) Side-chain repacking test; (B) sequence redesign test. The results are shown in the figure for the standard packing protocol (white) and the standard packing protocol plus RTMIN (gray). In both tests, positions which are ALA, GLY, PRO, and CYS in the native sequence are excluded from the calculation. (ALA and GLY do not have rotamers; PRO has such a restrained side chain that very limited torsion space is accessible to minimization; CYS is often involved in formation of disulfide bonds and may not be modeled properly without a more specialized treatment). In the side-chain repacking test, a side chain is considered to be correctly predicted if its angular deviations are <40° for both χ1 and χ2 angles from the native conformation. In the sequence redesign test, each sequence position is selected one at a time, with the rest of the protein fixed in its native conformation. Twenty amino acids with all their possible rotamers are considered at this position. The rotamer which yields the lowest energy determines the residue chosen for this position. If it matches the native amino acid, this sequence position is considered to be recovered. The first and last five residues in the protein are excluded from the sequence redesign test.
Figure 3.
Figure 3.
Comparison of results of side-chain repacking of interface residues in native complexes. The side-chain packing results are shown in the figure for the standard side-chain packing protocol (white), the standard side-chain packing protocol plus RTMIN (light gray), the standard side-chain packing protocol including native unbound rotamers (dark gray), and the standard side-chain packing protocol including native unbound rotamers and RTMIN (black).
Figure 4.
Figure 4.
Assessment of side-chain modeling based on residue interaction energy distributions. Distribution of interface residue energies within and between the protein partners (Eintra vs. Einter) for Arginine: (A) native bound protein complexes; (B) native bound protein complexes with repacked interfaces using the standard packing protocol; (C) native bound protein complexes with repacked interfaces using the improved side-chain modeling protocol (the standard packing protocol including unbound native rotamers and RTMIN); (D) true positive models (TP, low-score and low-RMSD models) of the bound docking perturbation runs using the improved side-chain modeling protocol (see Materials and Methods). When native interfaces are repacked, increased similarity to the native distribution is seen with the improved protocol, compared to the original protocol; note that in B, entropically unfavorable combinations (Eintra ~ 0, Einter ~ 0) are enriched, while entropically favorable conformations (Eintra ~ 0, Einter << 0 and Eintra << 0, Einter ~ 0 are less frequent, or even absent. There are 110 and 998 ARG interface residues included in the calculation in the native complexes (A–C) and the true positive models (D), respectively. The figure was created using Microsoft Excel.
Figure 5.
Figure 5.
Recapitulation of side-chain conformational changes in docking. Distributions of the number of residues that are conserved in rotamer conformation upon binding are shown for the interface GLU (A); ILE (B); ARG (C). White bars represent the counts for all interface residues. Gray bars represent the counts for those residues that do not change rotamer conformation upon binding. Symbol points represent distributions after repacking the interface of the native complex using different side-chain modeling protocols: standard packing protocol (open circle); standard packing protocol with RTMIN (cross sign); standard packing protocol including native unbound rotamers (plus sign); standard packing including native unbound rotamers and RTMIN (open square). The counts are distributed into bins of residue energy values in the native unbound structures. The white bars represent the extreme level of rotamer conservation assumed by classical rigid-body docking methods with all side chains fixed, while the gray bars show the distribution that a perfect flexible side-chain docking method which accounts for the right degree of side-chain flexibility would achieve. The symbol points indicate how well the different side-chain modeling methods handle the balance between rotamer conservation and side-chain flexibility.
Figure 6.
Figure 6.
The improved side-chain modeling method significantly improves the energy separation between correct and incorrect models. The normalized energy gap (Z-score) between near-native and nonnative models was computed as described in the Materials and Methods section for docked conformations generated for the 54 protein complexes in the benchmark of Chen et al. (2003). The Z-scores for the 54 protein complexes were binned into intervals of 0.5 Z-score units and the count for each bin is plotted in (A) bound docking perturbation studies with the standard side-chain modeling protocol (“−”) and with the improved side-chain modeling protocol (“+”). (B) Unbound docking perturbation studies with the standard side-chain modeling protocol (“−”) and with the improved side-chain modeling protocol (“+”). For both the bound and unbound cases, there are a significantly larger number of proteins with an energy gap between correct and incorrect docked arrangements >1 standard deviation (Z-score >1) when the new side-chain modeling method is applied.
Figure 7.
Figure 7.
Examples of accurate predictions in the CAPRI experiment. Overview of backbone orientation (left) and zoom-in view of side-chain prediction (right) of our predicted models in the CAPRI experiment are shown for (A) Target 12: Dockerin–Cohesin complex (Shimon et al. 1997; Lytle et al. 2001; Carvalho et al. 2003) (I_RMSD = 0.27 Å), (B) Target 14: MYPT–PP1 complex (Goldberg et al. 1995; Terrak et al. 2004) (I_RMSD = 0.38 Å), (C) Target 15: ImmD–ColD complex (Graille et al. 2004) (I_RMSD = 0.23 Å) (see http://capri.ebi.ac.uk for detailed description of CAPRI targets and evaluation reports). The predicted model is superimposed onto the native complex by the receptor backbone. The receptor and ligand of the native complex are colored red and orange. The ligand of the predicted model is colored blue. In Target 12, the side chain of LEU83 in the unbound conformation (green) clashes with the receptor in the native rigid-body position and it is moved to the native side-chain conformation after docking. This demonstrates the necessity of allowing side-chain flexibility in docking.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H., Shindyalov, I.N., and Bourne, P.E. 2000. The Protein Data Bank. Nucleic Acids Res. 28 235–242. - PMC - PubMed
    1. Camacho, C.J. and Vajda, S. 2002. Protein–protein association kinetics and protein docking. Curr. Opin. Struct. Biol. 12 36–40. - PubMed
    1. Carvalho, A.L., Dias, F.M., Prates, J.A., Nagy, T., Gilbert, H.J., Davies, G.J., Ferreira, L.M., Romao, M.J., and Fontes, C.M. 2003. Cellulosome assembly revealed by the crystal structure of the cohesin–dockerin complex. Proc. Natl. Acad. Sci. 100 13809–13814. - PMC - PubMed
    1. Chen, R. and Weng, Z. 2002. Docking unbound proteins using shape complementarity, desolvation, and electrostatics. Proteins 47 281–294. - PubMed
    1. Chen, R., Mintseris, J., Janin, J., and Weng, Z. 2003. A protein–protein docking benchmark. Proteins 52 88–91. - PubMed

Publication types