Automated docking of peptides and proteins by using a genetic algorithm combined with a tabu search.

A genetic algorithm (GA) combined with a tabu search (TA) has been applied as a minimization method to rake the appropriate associated sites for some biomolecular systems. In our docking procedure, surface complementarity and energetic complementarity of a ligand with its receptor have been considered separately in a two-stage docking method. The first stage was to find a set of potential associated sites mainly based on surface complementarity using a genetic algorithm combined with a tabu search. This step corresponds with the process of finding the potential binding sites where pharmacophores will bind. In the second stage, several hundreds of GA minimization steps were performed for each associated site derived from the first stage mainly based on the energetic complementarity. After calculations for both of the two stages, we can offer several solutions of associated sites for every complex. In this paper, seven biomolecular systems, including five bound complexes and two unbound complexes, were chosen from the Protein Data Bank (PDB) to test our method. The calculated results were very encouraging-the hybrid minimization algorithm successfully reaches the correct solutions near the best binded modes for these protein complexes. The docking results not only predict the bound complexes very well, but also get a relatively accurate complexed conformation for unbound systems. For the five bound complexes, the results show that surface complementarity is enough to find the precise binding modes, the top solution from the tabu list generally corresponds to the correct binding mode. For the two unbound complexes, due to the conformational changes upon binding, it seems more difficult to get their correct binding conformations. The predicted results show that the correct binding mode also corresponds to a relatively large surface complementarity score. In these two test cases, the correct solution can be found in the top several solutions from the tabu list. For unbound complexes, the interaction energy from energetic complementarity is very important, it can be used to filter these solutions from the surface complementarity. After the evaluation of the energetic complementarity, the conformations and orientations close to the crystallographically determined structures are resolved. In most cases, the smallest root mean square distance (r.m.s.d.) from the GA combined with TA solutions is in a relatively small region. Our program of automatic docking is really a universal one among the procedures used for the theoretical study of molecular recognition.     

Docking unbound proteins using shape complementarity, desolvation, and electrostatics

A comprehensive docking study was performed on 27 distinct protein-protein complexes. For 13 test systems, docking was performed with the unbound X-ray structures of both the receptor and the ligand. For the remaining systems, the unbound X-ray structure of only molecule was available; therefore the bound structure for the other molecule was used. Our method optimizes desolvation, shape complementarity, and electrostatics using a Fast Fourier Transform algorithm. A global search in the rotational and translational space without any knowledge of the binding sites was performed for all proteins except nine antibodies recognizing antigens. For these antibodies, we docked their well-characterized binding site-the complementarity-determining region defined without information of the antigen-to the entire surface of the antigen. For 24 systems, we were able to find near-native ligand orientations (interface C(alpha) root mean square deviation less than 2.5 A from the crystal complex) among the top 2,000 choices. For three systems, our algorithm could identify the correct complex structure unambiguously. For 13 other complexes, we either ranked a near-native structure in the top 20 or obtained 20 or more near-native structures in the top 2,000 or both. The key feature of our algorithm is the use of target functions that are highly tolerant to conformational changes upon binding. If combined with a post-processing method, our algorithm may provide a general solution to the unbound docking problem. Our program, called ZDOCK, is freely available to academic users (http://zlab.bu.edu/~rong/dock/).     

Protein docking using spherical polar Fourier correlations

We present a new computational method of docking pairs of proteins by using spherical polar Fourier correlations to accelerate the search for candidate low-energy conformations. Interaction energies are estimated using a hydrophobic excluded volume model derived from the notion of "overlapping surface skins," augmented by a rigorous but "soft" model of electrostatic complementarity. This approach has several advantages over former three-dimensional grid-based fast Fourier transform (FFT) docking correlation methods even though there is no analogue to the FFT in a spherical polar representation. For example, a complete search over all six rigid-body degrees of freedom can be performed by rotating and translating only the initial expansion coefficients, many unfeasible orientations may be eliminated rapidly using only low-resolution terms, and the correlations are easily localized around known binding epitopes when this knowledge is available. Typical execution times on a single processor workstation range from 2 hours for a global search (5 x 10(8) trial orientations) to a few minutes for a local search (over 6 x 10(7) orientations). The method is illustrated with several domain dimer and enzyme-inhibitor complexes and 20 large antibody-antigen complexes, using both the bound and (when available) unbound subunits. The correct conformation of the complex is frequently identified when docking bound subunits, and a good docking orientation is ranked within the top 20 in 11 out of 18 cases when starting from unbound subunits. Proteins 2000;39:178-194.     

Electrostatic contributions to protein-protein interactions: fast energetic filters for docking and their physical basis

The methods of continuum electrostatics are used to calculate the binding free energies of a set of protein-protein complexes including experimentally determined structures as well as other orientations generated by a fast docking algorithm. In the native structures, charged groups that are deeply buried were often found to favor complex formation (relative to isosteric nonpolar groups), whereas in nonnative complexes generated by a geometric docking algorithm, they were equally likely to be stabilizing as destabilizing. These observations were used to design a new filter for screening docked conformations that was applied, in conjunction with a number of geometric filters that assess shape complementarity, to 15 antibody-antigen complexes and 14 enzyme-inhibitor complexes. For the bound docking problem, which is the major focus of this paper, native and near-native solutions were ranked first or second in all but two enzyme-inhibitor complexes. Less success was encountered for antibody-antigen complexes, but in all cases studied, the more complete free energy evaluation was able to identify native and near-native structures. A filter based on the enrichment of tyrosines and tryptophans in antibody binding sites was applied to the antibody-antigen complexes and resulted in a native and near-native solution being ranked first and second in all cases. A clear improvement over previously reported results was obtained for the unbound antibody-antigen examples as well. The algorithm and various filters used in this work are quite efficient and are able to reduce the number of plausible docking orientations to a size small enough so that a final more complete free energy evaluation on the reduced set becomes computationally feasible.     

On the attribution of binding energy in antigen-antibody complexes McPC 603, D1.3, and HyHEL-5

Using X-ray coordinates of antigen-antibody complexes McPC 603, D1.3, and HyHEL-5, we made semiquantitative estimates of Gibbs free energy changes (delta G) accompanying noncovalent complex formation of the McPC 603 Fv fragment with phosphocholine and the D1.3 or HyHEL-5 Fv fragments with hen egg white lysozyme. Our empirical delta G function, which implicitly incorporates solvent effects, has the following components: hydrophobic force, solvent-modified electrostatics, changes in side-chain conformational entropy, translational/overall rotational entropy changes, and the dilutional (cratic) entropy term. The calculated delta G ranges matched the experimentally determined delta G of McPC 603 and D1.3 complexes and overestimated it (i.e., gave a more negative value) in the case of HyHEL-5. Relative delta G contributions of selected antibody residues, calculated for HyHEL-5 complexes, agreed with those determined independently in site-directed mutagenesis experiments. Analysis of delta G attribution in all three complexes indicated that only a small number of amino acids probably contribute actively to binding energetics. These form a subset of the total antigen-antibody contact surface. In the antibodies, the bottom part of the antigen binding cavity dominated the energetics of binding whereas in lysozyme, the energetically most important residues defined small (2.5-3 nm2) "energetic" epitopes. Thus, a concept of protein antigenicity emerges that involves the active, attractive contributions mediated by the energetic antigenic epitopes and the passive surface complementarity contributed by the surrounding contact area. The D1.3 energetic epitope of lysozyme involved Gly 22, Gly 117, and Gln 121; the HyHEL-5 epitope consisted of Arg 45 and Arg 68. These are also the essential antigenic residues determined experimentally. The above positions belong to the most protruding parts of the lysozyme surface, and their backbones are not exceptionally flexible. Least-squares analysis of six different antibody binding regions indicated that the geometry of the VH-VL interface beta-barrel is well conserved, giving no indication of significant changes in domain-domain contacts upon complex formation.     

Molecular modelling of protein-carbohydrate interactions. Docking of monosaccharides in the binding site of concanavalin A.

A general procedure is described for addressing the computer simulation of protein-carbohydrate interactions. First, a molecular mechanical force field capable of performing conformational analysis of oligosaccharides has been derived by the addition of new parameters to the Tripos force field; it is also compatible with the simulation of protein. Second, a docking procedure which allows for a systematic exploration of the orientations and positions of a ligand into a protein cavity has been designed. This so-called 'crankshaft' method uses rotations and variations about/of virtual bonds connecting, via dummy atoms, the ligand to the protein binding site. Third, calculation of the relative stability of protein ligand complexes is performed. This strategy has been applied to search for all favourable interactions occurring between a lectin [concanavalin A (ConA)] and methyl alpha-D-mannopyranoside or methyl alpha-D-glucopyranoside. For each monosaccharide, different stable orientations and positions within the binding site can be distinguished. Among them, one corresponds to very favourable interactions, not only in terms of hydrogen bonding, but also in terms of van der Waals interactions. It corresponds precisely to the binding mode of methyl alpha-D-mannopyranoside into ConA as revealed by the 2.9 A resolution of the crystalline complex (Derewenda et al., 1989). Some implications of the present modelling study with respect to the molecular basis of the specificity of the interaction of lectins with various monosaccharides are presented.     

ClusPro: an automated docking and discrimination method for the prediction of protein complexes

MOTIVATION: Predicting protein interactions is one of the most challenging problems in functional genomics. Given two proteins known to interact, current docking methods evaluate billions of docked conformations by simple scoring functions, and in addition to near-native structures yield many false positives, i.e. structures with good surface complementarity but far from the native. RESULTS: We have developed a fast algorithm for filtering docked conformations with good surface complementarity, and ranking them based on their clustering properties. The free energy filters select complexes with lowest desolvation and electrostatic energies. Clustering is then used to smooth the local minima and to select the ones with the broadest energy wells-a property associated with the free energy at the binding site. The robustness of the method was tested on sets of 2000 docked conformations generated for 48 pairs of interacting proteins. In 31 of these cases, the top 10 predictions include at least one near-native complex, with an average RMSD of 5 A from the native structure. The docking and discrimination method also provides good results for a number of complexes that were used as targets in the Critical Assessment of PRedictions of Interactions experiment. AVAILABILITY: The fully automated docking and discrimination server ClusPro can be found at http://structure.bu.edu     

A protein-protein docking algorithm dependent on the type of complexes

An efficient 'soft docking' algorithm is described to assist the prediction of protein-protein association using three-dimensional structures of molecules. The basic tools are the 'simplified protein' model and the docking algorithm of Wodak and Janin. The side chain flexibility of Arg, Lys, Asp, Glu and Met residues at the protein surface is taken into account. The complex type-dependent filtering technique on the basis of the geometric matching, hydrophobicity and electrostatic complementarity is used to select candidate binding modes. Subsequently, we calculate a scoring function which includes electrostatic and desolvation energy terms. In the 44 complexes tested including enzyme-inhibitor, antibody-antigen and other complexes, native-like structures were all found, of which 30 were ranked in the top 20. Thus, our soft docking algorithm has the potential to predict protein-protein recognition.     

F2Dock: fast Fourier protein-protein docking

The functions of proteins are often realized through their mutual interactions. Determining a relative transformation for a pair of proteins and their conformations which form a stable complex, reproducible in nature, is known as docking. It is an important step in drug design, structure determination, and understanding function and structure relationships. In this paper, we extend our nonuniform fast Fourier transform-based docking algorithm to include an adaptive search phase (both translational and rotational) and thereby speed up its execution. We have also implemented a multithreaded version of the adaptive docking algorithm for even faster execution on multicore machines. We call this protein-protein docking code F2Dock (F2 = Fast Fourier). We have calibrated F2Dock based on an extensive experimental study on a list of benchmark complexes and conclude that F2Dock works very well in practice. Though all docking results reported in this paper use shape complementarity and Coulombic-potential-based scores only, F2Dock is structured to incorporate Lennard-Jones potential and reranking docking solutions based on desolvation energy .     

PRL-Dock: protein-ligand docking based on hydrogen bond matching and probabilistic relaxation labeling

Protein-ligand docking is widely applied to structure-based virtual screening for drug discovery. This article presents a novel docking technique, PRL-Dock, based on hydrogen bond matching and probabilistic relaxation labeling. It deals with multiple hydrogen bonds and can match many acceptors and donors simultaneously. In the matching process, the initial probability of matching an acceptor with a donor is estimated by an efficient scoring function and the compatibility coefficients are assigned according to the coexisting condition of two hydrogen bonds. After hydrogen bond matching, the geometric complementarity of the interacting donor and acceptor sites is taken into account for displacement of the ligand. It is reduced to an optimization problem to calculate the optimal translation and rotation matrixes that minimize the root mean square deviation between two sets of points, which can be solved using the Kabsch algorithm. In addition to the van der Waals interaction, the contribution of intermolecular hydrogen bonds in a complex is included in the scoring function to evaluate the docking quality. A modified Lennard-Jones 12-6 dispersion-repulsion term is used to estimate the van der Waals interaction to make the scoring function fairly "soft" so that ligands are not heavily penalized for small errors in the binding geometry. The calculation of this scoring function is very convenient. The evaluation is carried out on 278 rigid complexes and 93 flexible ones where there is at least one intermolecular hydrogen bond. The experiment results of docking accuracy and prediction of binding affinity demonstrate that the proposed method is highly effective.     

BiGGER: a new (soft) docking algorithm for predicting protein interactions

A new computationally efficient and automated "soft docking" algorithm is described to assist the prediction of the mode of binding between two proteins, using the three-dimensional structures of the unbound molecules. The method is implemented in a software package called BiGGER (Bimolecular Complex Generation with Global Evaluation and Ranking) and works in two sequential steps: first, the complete 6-dimensional binding spaces of both molecules is systematically searched. A population of candidate protein-protein docked geometries is thus generated and selected on the basis of the geometric complementarity and amino acid pairwise affinities between the two molecular surfaces. Most of the conformational changes observed during protein association are treated in an implicit way and test results are equally satisfactory, regardless of starting from the bound or the unbound forms of known structures of the interacting proteins. In contrast to other methods, the entire molecular surfaces are searched during the simulation, using absolutely no additional information regarding the binding sites. In a second step, an interaction scoring function is used to rank the putative docked structures. The function incorporates interaction terms that are thought to be relevant to the stabilization of protein complexes. These include: geometric complementarity of the surfaces, explicit electrostatic interactions, desolvation energy, and pairwise propensities of the amino acid side chains to contact across the molecular interface. The relative functional contribution of each of these interaction terms to the global scoring function has been empirically adjusted through a neural network optimizer using a learning set of 25 protein-protein complexes of known crystallographic structures. In 22 out of 25 protein-protein complexes tested, near-native docked geometries were found with C(alpha) RMS deviations < or =4.0 A from the experimental structures, of which 14 were found within the 20 top ranking solutions. The program works on widely available personal computers and takes 2 to 8 hours of CPU time to run any of the docking tests herein presented. Finally, the value and limitations of the method for the study of macromolecular interactions, not yet revealed by experimental techniques, are discussed.     

A Fast Flexible Docking Method using an Incremental Construction Algorithm

We present an automatic method for docking organic ligands into protein binding sites. The method can be used in the design process of specific protein ligands. It combines an appropriate model of the physico-chemical properties of the docked molecules with efficient methods for sampling the conformational space of the ligand. If the ligand is flexible, it can adopt a large variety of different conformations. Each such minimum in conformational space presents a potential candidate for the conformation of the ligand in the complexed state. Our docking method samples the conformation space of the ligand on the basis of a discrete model and uses a tree-search technique for placing the ligand incrementally into the active site. For placing the first fragment of the ligand into the protein, we use hashing techniques adapted from computer vision. The incremental construction algorithm is based on a greedy strategy combined with efficient methods for overlap detection and for the search of new interactions. We present results on 19 complexes of which the binding geometry has been crystallographically determined. All considered ligands are docked in at most three minutes on a current workstation. The experimentally observed binding mode of the ligand is reproduced with 0.5 to 1.2 Å rms deviation. It is almost always found among the highest-ranking conformations computed.     

Escher: A new docking procedure applied to the reconstruction of protein tertiary structure

Evaluation of Surface Complementarity, Hydrogen bonding, and Electrostatic interaction in molecular Recognition (ESCHER) is a new docking procedure consisting of three modules that work in series. The first module evaluates the geometric complementarity and produces a set of rough solutions for the docking problem. The second module identifies molecular collisions within those solutions, and the third evaluates their electrostatic complementarity. We describe the algorithm and its application to the docking of cocrystallized protein domains and unbound components of protein-protein complexes. Furthermore, ESCHER has been applied to the reassociation of secondary and supersecondary structure elements. The possibility of applying a docking method to the problem of protein structure prediction is discussed. Proteins 28:556–567, 1997. © 1997 Wiley-Liss, Inc.     

Use of pair potentials across protein interfaces in screening predicted docked complexes.

Empirical residue-residue pair potentials are used to screen possible complexes for protein-protein dockings. A correct docking is defined as a complex with not more than 2.5 A root-mean-square distance from the known experimental structure. The complexes were generated by "ftdock" (Gabb et al. J Mol Biol 1997;272:106-120) that ranks using shape complementarity. The complexes studied were 5 enzyme-inhibitors and 2 antibody-antigens, starting from the unbound crystallographic coordinates, with a further 2 antibody-antigens where the antibody was from the bound crystallographic complex. The pair potential functions tested were derived both from observed intramolecular pairings in a database of nonhomologous protein domains, and from observed intermolecular pairings across the interfaces in sets of nonhomologous heterodimers and homodimers. Out of various alternate strategies, we found the optimal method used a mole-fraction calculated random model from the intramolecular pairings. For all the systems, a correct docking was placed within the top 12% of the pair potential score ranked complexes. A combined strategy was developed that incorporated "multidock," a side-chain refinement algorithm (Jackson et al. J Mol Biol 1998;276:265-285). This placed a correct docking within the top 5 complexes for enzyme-inhibitor systems, and within the top 40 complexes for antibody-antigen systems.     

Protein docking using surface matching and supervised machine learning

Computational prediction of protein complex structures through docking offers a means to gain a mechanistic understanding of protein interactions that mediate biological processes. This is particularly important as the number of experimentally determined structures of isolated proteins exceeds the number of structures of complexes. A comprehensive docking procedure is described in which efficient sampling of conformations is achieved by matching surface normal vectors, fast filtering for shape complementarity, clustering by RMSD, and scoring the docked conformations using a supervised machine learning approach. Contacting residue pair frequencies, residue propensities, evolutionary conservation, and shape complementarity score for each docking conformation are used as input data to a Random Forest classifier. The performance of the Random Forest approach for selecting correctly docked conformations was assessed by cross-validation using a nonredundant benchmark set of X-ray structures for 93 heterodimer and 733 homodimer complexes. The single highest rank docking solution was the correct (near-native) structure for slightly more than one third of the complexes. Furthermore, the fraction of highly ranked correct structures was significantly higher than the overall fraction of correct structures, for almost all complexes. A detailed analysis of the difficult to predict complexes revealed that the majority of the homodimer cases were explained by incorrect oligomeric state annotation. Evolutionary conservation and shape complementarity score as well as both underrepresented and overrepresented residue types and residue pairs were found to make the largest contributions to the overall prediction accuracy. Finally, the method was also applied to docking unbound subunit structures from a previously published benchmark set.     

Protein-DNA docking with a coarse-grained force field

ABSTRACT: BACKGROUND: Protein-DNA interactions are important for many cellular processes, however structural knowledge for a large fraction of known and putative complexes is still lacking. Computational docking methods aim at the prediction of complex architecture given detailed structures of its constituents. They are becoming an increasingly important tool in the field of macromolecular assemblies, complementing particularly demanding protein-nucleic acids X ray crystallography and providing means for the refinement and integration of low resolution data coming from rapidly advancing methods such as cryoelectron microscopy. RESULTS: We present a new coarse-grained force field suitable for protein-DNA docking. The force field is an extension of previously developed parameter sets for protein-RNA and protein-protein interactions. The docking is based on potential energy minimization in translational and orientational degrees of freedom of the binding partners. It allows for fast and efficient systematic search for native-like complex geometry without any prior knowledge regarding binding site location. CONCLUSIONS: We find that the force field gives very good results for bound docking. The quality of predictions in the case of unbound docking varies, depending on the level of structural deviation from bound geometries. We analyze the role of specific protein-DNA interactions on force field performance, both with respect to complex structure prediction, and the reproduction of experimental binding affinities. We find that such direct, specific interactions only partially contribute to protein-DNA recognition, indicating an important role of shape complementarity and sequence-dependent DNA internal energy, in line with the concept of indirect protein-DNA readout mechanism.     

Automated docking of ligands to antibodies: methods and applications

Many approaches to studying protein-ligand interactions by computational docking are currently available. Given the structures of a protein and a ligand, the ultimate goal of all docking methods is to predict the structure of the resulting complex. This requires a suitable representation of molecular structures and properties, search algorithms to efficiently scan the configuration space for favorable interaction geometries, and accurate scoring functions to evaluate and rank the generated orientations. For many of the available methods, tests on experimentally known antibody-antigen or antibody-hapten complexes have appeared in the literature. In addition, some of them have been used in predictive studies on antibody-ligand interactions to provide structural insights where adequate experimental information is missing. The AutoDock program is presented as example of a method for flexibly docking ligands to antibodies. Applying parameters of the second-generation AMBER force field, three antibody-hapten complexes (AN02, DB3, NC6.8) are used as new test cases to analyze the ability of the method to reproduce experimental findings. The X-ray structures could be reconstituted and the corresponding solutions were ranked with best energy score in all cases. Docking to the free instead of the complexed NC6.8 structure indicated the limits of the rigid protein treatment, although fairly good guesses about the location of the binding site and the contact residues could still be obtained if conformational flexibility was allowed at least in the ligand.     

Hydrophobic complementarity in protein-protein docking

Formation of hydrophobic contacts across a newly formed interface is energetically favorable. Based on this observation we developed a geometric-hydrophobic docking algorithm that estimates quantitatively the hydrophobic complementarity at protein-protein interfaces. Each molecule to be docked is represented as a grid of complex numbers, storing information regarding the shape of the molecule in the real part and information regarding the hydropathy of the surface in the imaginary part. The grid representations are correlated using fast Fourier transformations. The algorithm is used to compare the extent of hydrophobic complementarity in oligomers (represented by D2 tetramers) and in hetero-dimers of soluble proteins (complexes). We also test the implication of hydrophobic complementarity in distinguishing correct from false docking solutions. We find that hydrophobic complementarity at the interface exists in oligomers and in complexes, and in both groups the extent of such complementarity depends on the size of the interface. Thus, the non-polar portions of large interfaces are more often juxtaposed than non-polar portions of small interfaces. Next we find that hydrophobic complementarity helps to point out correct docking solutions. In oligomers it significantly improves the ranks of nearly correct reassembled and modeled tetramers. Combining geometric, electrostatic and hydrophobic complementarity for complexes gives excellent results, ranking a nearly correct solution < 10 for 5 of 23 tested systems, < 100 for 8 systems and < 1000 for 19 systems.     

Dissection of binding interactions in the complex between the anti-lysozyme antibody HyHEL-63 and its antigen

Alanine-scanning mutagenesis, X-ray crystallography, and double mutant cycles were used to characterize the interface between the anti-hen egg white lysozyme (HEL) antibody HyHEL-63 and HEL. Eleven HEL residues in contact with HyHEL-63 in the crystal structure of the antigen-antibody complex, and 10 HyHEL-63 residues in contact with HEL, were individually truncated to alanine in order to determine their relative contributions to complex stabilization. The residues of HEL (Tyr20, Lys96, and Lys97) most important for binding HyHEL-63 (Delta G(mutant) - Delta G(wild type) > 3.0 kcal/mol) form a contiguous patch at the center of the surface contacted by the antibody. Hot spot residues of the antibody (Delta Delta G > 2.0 kcal/mol) are organized in two clusters that juxtapose hot spot residues of HEL, resulting in energetic complementarity across the interface. All energetically critical residues are centrally located, shielded from solvent by peripheral residues that contribute significantly less to the binding free energy. Although HEL hot spot residues Lys96 and Lys97 make similar interactions with antibody in the HyHEL-63/HEL complex, alanine substitution of Lys96 results in a nearly 100-fold greater reduction in affinity than the corresponding mutation in Lys97. To understand the basis for this marked difference, we determined the crystal structures of the HyHEL-63/HEL Lys96Ala and HyHEL-63/HEL Lys97Ala complexes to 1.80 and 1.85 A resolution, respectively. Whereas conformational changes in the proteins and differences in the solvent networks at the mutation sites appear too small to explain the observed affinity difference, superposition of free HEL in different crystal forms onto bound HEL in the wild type and mutant HyHEL-63/HEL complexes reveals that the side-chain conformation of Lys96 is very similar in the various structures, but that the Lys97 side chain displays considerable flexibility. Accordingly, a greater entropic penalty may be associated with quenching the mobility of the Lys97 than the Lys96 side chain upon complex formation, reducing binding. To further dissect the energetics of specific interactions in the HyHEL-63/HEL interface, double mutant cycles were constructed to measure the coupling of 13 amino acid pairs, 11 of which are in direct contact in the crystal structure. A large coupling energy, 3.0 kcal/mol, was found between HEL residue Lys97 and HyHEL-63 residue V(H)Asp32, which form a buried salt bridge surrounded by polar residues of the antigen. Thus, in contrast to protein folding where buried salt bridges are generally destabilizing, salt bridges in protein-protein interfaces, whose residual composition is more hydrophilic than that of protein interiors, may contribute significantly to complex stabilization.