Community-wide assessment of protein-interface modeling suggests improvements to design methodology

The CAPRI (Critical Assessment of Predicted Interactions) and CASP (Critical Assessment of protein Structure Prediction) experiments have demonstrated the power of community-wide tests of methodology in assessing the current state of the art and spurring progress in the very challenging areas of protein docking and structure prediction. We sought to bring the power of community-wide experiments to bear on a very challenging protein design problem that provides a complementary but equally fundamental test of current understanding of protein-binding thermodynamics. We have generated a number of designed protein-protein interfaces with very favorable computed binding energies but which do not appear to be formed in experiments, suggesting that there may be important physical chemistry missing in the energy calculations. A total of 28 research groups took up the challenge of determining what is missing: we provided structures of 87 designed complexes and 120 naturally occurring complexes and asked participants to identify energetic contributions and/or structural features that distinguish between the two sets. The community found that electrostatics and solvation terms partially distinguish the designs from the natural complexes, largely due to the nonpolar character of the designed interactions. Beyond this polarity difference, the community found that the designed binding surfaces were, on average, structurally less embedded in the designed monomers, suggesting that backbone conformational rigidity at the designed surface is important for realization of the designed function. These results can be used to improve computational design strategies, but there is still much to be learned; for example, one designed complex, which does form in experiments, was classified by all metrics as a nonbinder.

Integrating atom-based and residue-based scoring functions for protein-protein docking

Most scoring functions for protein-protein docking algorithms are either atom-based or residue-based, with the former being able to produce higher quality structures and latter more tolerant to conformational changes upon binding. Earlier, we developed the ZRANK algorithm for reranking docking predictions, with a scoring function that contained only atom-based terms. Here we combine ZRANK's atom-based potentials with five residue-based potentials published by other labs, as well as an atom-based potential IFACE that we published after ZRANK. We simultaneously optimized the weights for selected combinations of terms in the scoring function, using decoys generated with the protein-protein docking algorithm ZDOCK. We performed rigorous cross validation of the combinations using 96 test cases from a docking benchmark. Judged by the integrative success rate of making 1000 predictions per complex, addition of IFACE and the best residue-based pair potential reduced the number of cases without a correct prediction by 38 and 27% relative to ZDOCK and ZRANK, respectively. Thus combination of residue-based and atom-based potentials into a scoring function can improve performance for protein-protein docking. The resulting scoring function is called IRAD (integration of residue- and atom-based potentials for docking) and is available at http://zlab.umassmed.edu.

A machine learning approach for the prediction of protein surface loop flexibility

Proteins often undergo conformational changes when binding to each other. A major fraction of backbone conformational changes involves motion on the protein surface, particularly in loops. Accounting for the motion of protein surface loops represents a challenge for protein-protein docking algorithms. A first step in addressing this challenge is to distinguish protein surface loops that are likely to undergo backbone conformational changes upon protein-protein binding (mobile loops) from those that are not (stationary loops). In this study, we developed a machine learning strategy based on support vector machines (SVMs). Our SVM uses three features of loop residues in the unbound protein structures-Ramachandran angles, crystallographic B-factors, and relative accessible surface area-to distinguish mobile loops from stationary ones. This method yields an average prediction accuracy of 75.3% compared with a random prediction accuracy of 50%, and an average of 0.79 area under the receiver operating characteristic (ROC) curve using cross-validation. Testing the method on an independent dataset, we obtained a prediction accuracy of 70.5%. Finally, we applied the method to 11 complexes that involve members from the Ras superfamily and achieved prediction accuracy of 92.8% for the Ras superfamily proteins and 74.4% for their binding partners.

Protein-protein docking benchmark version 4.0

We updated our protein-protein docking benchmark to include complexes that became available since our previous release. As before, we only considered high-resolution complex structures that are nonredundant at the family-family pair level, for which the X-ray or NMR unbound structures of the constituent proteins are also available. Benchmark 4.0 adds 52 new complexes to the 124 cases of Benchmark 3.0, representing an increase of 42%. Thus, benchmark 4.0 provides 176 unbound-unbound cases that can be used for protein-protein docking method development and assessment. Seventeen of the newly added cases are enzyme-inhibitor complexes, and we found no new antigen-antibody complexes. Classifying the new cases according to expected difficulty for protein-protein docking algorithms gives 33 rigid body cases, 11 cases of medium difficulty, and 8 cases that are difficult. Benchmark 4.0 listings and processed structure files are publicly accessible at http://zlab.umassmed.edu/benchmark/.

Performance of ZDOCK and ZRANK in CAPRI rounds 13-19

We report the performance of the ZDOCK and ZRANK algorithms in CAPRI rounds 13-19 and introduce a novel measure atom contact frequency (ACF). To compute ACF, we identify the residues that most often make contact with the binding partner in the complete set of ZDOCK predictions for each target. We used ACF to predict the interface of the proteins, which, in combination with the biological data available in the literature, is a valuable addition to our docking pipeline. Furthermore, we incorporated a straightforward and efficient clustering algorithm with two purposes: (1) to determine clusters of similar docking poses (corresponding to energy funnels) and (2) to remove redundancies from the final set of predictions. With these new developments, we achieved at least one acceptable prediction for targets 29 and 36, at least one medium-quality prediction for targets 41 and 42, and at least one high-quality prediction for targets 37 and 40; thus, we succeeded for six out of a total of 12 targets.

Matrix metalloproteinase 2 inhibition: combined quantum mechanics and molecular mechanics studies of the inhibition mechanism of (4-phenoxyphenylsulfonyl)methylthiirane and its oxirane analogue

The inhibition mechanism of matrix metalloproteinase 2 (MMP2) by the selective inhibitor (4-phenoxyphenylsulfonyl)methylthiirane (SB-3CT) and its oxirane analogue is investigated computationally. The inhibition mechanism involves C-H deprotonation with concomitant opening of the three-membered heterocycle. SB-3CT was docked into the active site of MMP2, followed by molecular dynamics simulation to prepare the complex for combined quantum mechanics and molecular mechanics (QM/MM) calculations. QM/MM calculations with B3LYP/6-311+G(d,p) for the QM part and the AMBER force field for the MM part were used to examine the reaction of these two inhibitors in the active site of MMP2. The calculations show that the reaction barrier for transformation of SB-3CT is 1.6 kcal/mol lower than its oxirane analogue, and the ring-opening reaction energy of SB-3CT is 8.0 kcal/mol more exothermic than that of its oxirane analogue. Calculations also show that protonation of the ring-opened product by water is thermodynamically much more favorable for the alkoxide obtained from the oxirane than for the thiolate obtained from the thiirane. A six-step partial charge fitting procedure is introduced for the QM/MM calculations to update atomic partial charges of the quantum mechanics region and to ensure consistent electrostatic energies for reactants, transition states, and products.

The application of the ONIOM hybrid method to the cycloaddition reactions of bromo-substituted 2(H)-pyran-2-ones

We investigate the performance of Our own N-layered Integrated molecular Orbital and molecular Mechanics (ONIOM) hybrid computational method applied to Diels–Alder reactions of bromo-2(H)-pyran-2-ones, combining the B3LYP/6–31G(d) method with a variety of low-level methods. We show that ONIOM is able to reproduce full B3LYP calculations, including the prediction of the stereoselectivity, which requires accurate potentials. We focus on the various ways in which the performance and potential errors of ONIOM can be analyzed, and show that the best method combination depends on the property one is interested in.

Computational investigation of irreversible inactivation of the zinc-dependent protease carboxypeptidase A

Zinc proteases are ubiquitous and the zinc ion plays a central function in the catalytic mechanism of these enzymes. A novel class of mechanism-based inhibitors takes advantage of the zinc ion chemistry in carboxypeptidase A (CPA) to promote covalent attachment of an inhibitor to the carboxylate of Glu-270, resulting in irreversible inhibition of the enzyme. The effect of the active site zinc ion on irreversible inactivation of CPA was probed by molecular orbital (MO) calculations on a series of active site models and the Cl(-) + CH(3)Cl S(N)2 reaction fragment. Point charge models representing the active site reproduced energetics from full MO calculations at 12.0 A separation between the zinc and the central carbon of the S(N)2 reaction, but at 5.0 A polarization played an important role in moderating barrier suppression. ONIOM MO/MO calculations that included the residues within 10 A of the active site zinc suggest that about 75% of the barrier suppression arises from the zinc ion and its ligands. A model of the pre-reactive complex of the 2-benzyl-3-iodopropanoate inactivator with CPA was constructed from the X-ray structure of l-phenyl lactate bound in the active site of the enzyme. The model was fully solvated and minimized by using the AMBER force field to generate the starting structure for the ONIOM QM/MM calculations. Optimization of this structure led to the barrierless S(N)2 displacement of the iodide of the inhibitor by Glu-270, assisted by interaction of the zinc ion with the leaving group. The resulting product is in good agreement with the X-ray structure of the covalently modified enzyme obtained by irreversible inhibition of CPA by 2-benzyl-3-iodopropanoate.

Transition States of Binap–Rhodium(I)-Catalyzed Asymmetric Hydrogenation: Theoretical Studies on the Origin of the Enantioselectivity

By using the hybrid IMOMM(B3LYP:MM3) method, we examined the binap-Rh(I)-catalyzed oxidative-addition and insertion steps of the asymmetric hydrogenation of the enamide 2-acetylamino-3-phenylacrylic acid. We report a path that is energetically more favorable for the major enantiomer than for the minor enantiomer. This path follows the "lock-and-key" motif and leads to the major enantiomeric product via an energetically favorable binap-dihydride-Rh(III)-enamide complex. Our theoretical results are consistent with the mechanism that takes place via Rh(III) dihydride formation, that is, oxidative addition of H2 followed by enamide insertion.

Is the Protein Surrounding the Active Site Critical for Hydrogen Peroxide Reduction by Selenoprotein Glutathione Peroxidase? An ONIOM Study

In this ONIOM(QM:MM) study, we evaluate the role of the protein surroundings in the mechanism of H2O2 reduction catalyzed by the glutathione peroxidase enzyme, using the whole monomer (3113 atoms in 196 amino acid residues) as a model. A new optimization scheme that allows the full optimization of transition states for large systems has been utilized. It was found that in the presence of the surrounding protein the optimized active site structure bears a closer resemblance to the one in the X-ray structure than that without the surrounding protein. H2O2 reduction occurs through a two-step mechanism. In the first step, the selenolate anion (E-Se(-)) formation occurs with a barrier of 16.4 kcal/mol and is endothermic by 12.0 kcal/mol. The Gln83 residue plays the key role of the proton abstractor, which is in line with the experimental suggestion. In the second step, the O-O bond is cleaved, and selenenic acid (R-Se-OH) and a water molecule are formed. The calculated barrier for this process is 6.0 kcal/mol, and it is exothermic by 80.9 kcal/mol. The overall barrier of 18.0 kcal/mol for H2O2 reduction is in reasonable agreement with the experimentally measured barrier of 14.9 kcal/mol. The protein surroundings has been calculated to exert a net effect of only 0.70 kcal/mol (in comparison to the "active site only" model including solvent effects) on the overall barrier, which is most likely due to the active site being located at the enzyme surface.

Lysine carboxylation in proteins: OXA-10 β-lactamase

An increasing number of proteins are being shown to have an N(zeta)-carboxylated lysine in their structures, a posttranslational modification of proteins that proceeds without the intervention of a specific enzyme. The role of the carboxylated lysine in these proteins is typically structural (hydrogen bonding or metal coordination). However, carboxylated lysines in the active sites of OXA-10 and OXA-1 beta-lactamases and the sensor domain of BlaR signal-transducer protein serve in proton transfer events required for the functions of these proteins. These examples demonstrate the utility of this unusual amino acid in acid-base chemistry, in expansion of function beyond those of the 20 standard amino acids. In this study, the ONIOM quantum-mechanical/molecular-mechanical (QM/MM) method is used to study the carboxylation of lysine in the OXA-10 beta-lactamase. Lys-70 and the active site of the OXA-10 beta-lactamase were treated with B3LYP/6-31G(d,p) density functional calculations and the remainder of the enzyme with the AMBER molecular mechanics force field. The barriers for unassisted carboxylation of neutral lysine by carbon dioxide or bicarbonate are high. However, when the reaction with CO2 is catalyzed by a molecule of water in the active site, it is exothermic by about 13 kcal/mol, with a barrier of approximately 14 kcal/mol. The calculations show that the carboxylation and decarboxylation of Lys-70 are likely to be accompanied by deprotonation and protonation of the carbamate, respectively. The analysis may also be relevant for other proteins with carboxylated lysines, a feature that may be more common in nature than previously appreciated.

Elucidation of the Mechanism of Selenoprotein Glutathione Peroxidase (GPx)-Catalyzed Hydrogen Peroxide Reduction by Two Glutathione Molecules: A Density Functional Study

The mechanism of the hydrogen peroxide reduction by two molecules of glutathione catalyzed by the selenoprotein glutatione peroxidase (GPx) has been computationally studied. It has been shown that the first elementary reaction of this process, (E-SeH) + H(2)O(2) --> (E-SeOH) + H(2)O (1), proceeds via a stepwise pathway with the overall barrier of 17.1 kcal/mol, which is in good agreement with the experimental barrier of 14.9 kcal/mol. During reaction 1, the Gln83 residue has been found to play a key role as a proton acceptor, which is consistent with experiments. The second elementary reaction, (E-SeOH) + GSH --> (E-Se-SG) + HOH (2), proceeds with the barrier of 17.9 kcal/mol. The last elementary reaction, (E-Se-SG) + GSH --> (E-SeH) + GS-SG (3), is initiated with the coordination of the second glutathione molecule. The calculations clearly suggest that the amide backbone of the Gly50 residue directly participates in this reaction and the presence of two water molecules is absolutely vital for the reaction to occur. This reaction proceeds with the barrier of 21.5 kcal/mol and is suggested to be a rate-determining step of the entire GPx-catalyzed reaction H(2)O(2) + 2GSH --> GS-SG + 2H(2)O. The results discussed in the present study provide intricate details of every step of the catalytic mechanism of the GPx enzyme and are in good general agreement with experimental findings and suggestions.

Density Functional Study of the Optical Rotation of Glucose in Aqueous Solution

Optical rotation values were calculated for the eight most abundant structures of glucose in aqueous solution, following the TD-DFT/GIAO approach for the property and the PCM description for the solvent. The results show that all alpha structures give a large positive contribution to the OR property, while the beta structures give both positive and negative contributions. The good agreement of the calculated OR, obtained as a Boltzmann average of the property of the eight conformers, with experimental data proves the validity of the quantum-mechanical approach and of the solvent modelization.