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

A Survey of Deep Learning Methods for Estimating the Accuracy of Protein Quaternary Structure Models

The quality prediction of quaternary structure models of a protein complex, in the absence of its true structure, is known as the Estimation of Model Accuracy (EMA). EMA is useful for ranking predicted protein complex structures and using them appropriately in biomedical research, such as protein-protein interaction studies, protein design, and drug discovery. With the advent of more accurate protein complex (multimer) prediction tools, such as AlphaFold2-Multimer and ESMFold, the estimation of the accuracy of protein complex structures has attracted increasing attention. Many deep learning methods have been developed to tackle this problem; however, there is a noticeable absence of a comprehensive overview of these methods to facilitate future development. Addressing this gap, we present a review of deep learning EMA methods for protein complex structures developed in the past several years, analyzing their methodologies, data and feature construction. We also provide a prospective summary of some potential new developments for further improving the accuracy of the EMA methods.

Computational Mutagenesis of Antibody Fragments: Disentangling Side Chains from ΔΔG Predictions

The development of highly potent antibodies and antibody fragments as binding agents holds significant implications in fields such as biosensing and biotherapeutics. Their binding strength is intricately linked to the arrangement and composition of residues at the binding interface. Computational techniques offer a robust means to predict the three-dimensional structure of these complexes and to assess the affinity changes resulting from mutations. Given the interdependence of structure and affinity prediction, our objective here is to disentangle their roles. We aim to evaluate independently six side-chain reconstruction methods and ten binding affinity estimation techniques. This evaluation was pivotal in predicting affinity alterations due to single mutations, a key step in computational affinity maturation protocols. Our analysis focuses on a data set comprising 27 distinct antibody/hen egg white lysozyme complexes, each with crystal structures and experimentally determined binding affinities. Using six different side-chain reconstruction methods, we transformed each structure into its corresponding mutant via in silico single-point mutations. Subsequently, these structures undergo minimization and molecular dynamics simulation. We therefore estimate ΔΔG values based on the original crystal structure, its energy-minimized form, and the ensuing molecular dynamics trajectories. Our research underscores the critical importance of selecting reliable side-chain reconstruction methods and conducting thorough molecular dynamics simulations to accurately predict the impact of mutations. In summary, our study demonstrates that the integration of conformational sampling and scoring is a potent approach to precisely characterizing mutation processes in single-point mutagenesis protocols and crucial for computational antibody design.

Evaluation of AlphaFold Antibody-Antigen Modeling with Implications for Improving Predictive Accuracy

High resolution antibody-antigen structures provide critical insights into immune recognition and can inform therapeutic design. The challenges of experimental structural determination and the diversity of the immune repertoire underscore the necessity of accurate computational tools for modeling antibody-antigen complexes. Initial benchmarking showed that despite overall success in modeling protein-protein complexes, AlphaFold and AlphaFold-Multimer have limited success in modeling antibody-antigen interactions. In this study, we performed a thorough analysis of AlphaFold's antibody-antigen modeling performance on 429 nonredundant antibody-antigen complex structures, identifying useful confidence metrics for predicting model quality, and features of complexes associated with improved modeling success. We show the importance of bound-like component modeling in complex assembly accuracy, and that the current version of AlphaFold improves near-native modeling success to over 30%, versus approximately 20% for a previous version. With this improved success, AlphaFold can generate accurate antibody-antigen models in many cases, while additional training may further improve its performance.

A gated graph transformer for protein complex structure quality assessment and its performance in CASP15

MOTIVATION

Proteins interact to form complexes to carry out essential biological functions. Computational methods such as AlphaFold-multimer have been developed to predict the quaternary structures of protein complexes. An important yet largely unsolved challenge in protein complex structure prediction is to accurately estimate the quality of predicted protein complex structures without any knowledge of the corresponding native structures. Such estimations can then be used to select high-quality predicted complex structures to facilitate biomedical research such as protein function analysis and drug discovery.

RESULTS

In this work, we introduce a new gated neighborhood-modulating graph transformer to predict the quality of 3D protein complex structures. It incorporates node and edge gates within a graph transformer framework to control information flow during graph message passing. We trained, evaluated and tested the method (called DProQA) on newly-curated protein complex datasets before the 15th Critical Assessment of Techniques for Protein Structure Prediction (CASP15) and then blindly tested it in the 2022 CASP15 experiment. The method was ranked 3rd among the single-model quality assessment methods in CASP15 in terms of the ranking loss of TM-score on 36 complex targets. The rigorous internal and external experiments demonstrate that DProQA is effective in ranking protein complex structures.

AVAILABILITY AND IMPLEMENTATION

The source code, data, and pre-trained models are available at https://github.com/jianlin-cheng/DProQA.

Replica-exchange optimization of antibody fragments

In the framework of the rational design of macromolecules capable of binding to a specific target for biosensing applications, we here further develop an evolutionary protocol designed to optimize the binding affinity of protein binders. In particular we focus on the optimization of the binding portion of small antibody fragments known as nanobodies (or VHH) and choose the hen egg white lysozyme (HEWL) as our target. By implementing a replica exchange scheme for this optimization, we show that an initial hit is not needed and similar solutions can be found by either optimizing an already known anti-HEWL VHH or a randomly selected binder (here a VHH selective towards another macromolecule). While we believe that exhaustive searches of the mutation space are most appropriate when only few key residues have to be optimized, in case a lead binder is not available the proposed evolutionary algorithm should be instead the method of choice.

MetaScore: A Novel Machine-Learning-Based Approach to Improve Traditional Scoring Functions for Scoring Protein-Protein Docking Conformations

Protein-protein interactions play a ubiquitous role in biological function. Knowledge of the three-dimensional (3D) structures of the complexes they form is essential for understanding the structural basis of those interactions and how they orchestrate key cellular processes. Computational docking has become an indispensable alternative to the expensive and time-consuming experimental approaches for determining the 3D structures of protein complexes. Despite recent progress, identifying near-native models from a large set of conformations sampled by docking-the so-called scoring problem-still has considerable room for improvement. We present MetaScore, a new machine-learning-based approach to improve the scoring of docked conformations. MetaScore utilizes a random forest (RF) classifier trained to distinguish near-native from non-native conformations using their protein-protein interfacial features. The features include physicochemical properties, energy terms, interaction-propensity-based features, geometric properties, interface topology features, evolutionary conservation, and also scores produced by traditional scoring functions (SFs). MetaScore scores docked conformations by simply averaging the score produced by the RF classifier with that produced by any traditional SF. We demonstrate that (i) MetaScore consistently outperforms each of the nine traditional SFs included in this work in terms of success rate and hit rate evaluated over conformations ranked among the top 10; (ii) an ensemble method, MetaScore-Ensemble, that combines 10 variants of MetaScore obtained by combining the RF score with each of the traditional SFs outperforms each of the MetaScore variants. We conclude that the performance of traditional SFs can be improved upon by using machine learning to judiciously leverage protein-protein interfacial features and by using ensemble methods to combine multiple scoring functions.

Connection between MHC class II binding and aggregation propensity: The antigenic peptide 10 of Paracoccidioides brasiliensis as a benchmark study

The aggregation of epitopes that are also able to bind major histocompatibility complex (MHC) alleles raises questions around the potential connection between the formation of epitope aggregates and their affinities to MHC receptors. We first performed a general bioinformatic assessment over a public dataset of MHC class II epitopes, finding that higher experimental binding correlates with higher aggregation-propensity predictors. We then focused on the case of P10, an epitope used as a vaccine candidate against Paracoccidioides brasiliensis that aggregates into amyloid fibrils. We used a computational protocol to design variants of the P10 epitope to study the connection between the binding stabilities towards human MHC class II alleles and their aggregation propensities. The binding of the designed variants was tested experimentally, as well as their aggregation capacity. High-affinity MHC class II binders in vitro were more disposed to aggregate forming amyloid fibrils capable of binding Thioflavin T and congo red, while low affinity MHC class II binders remained soluble or formed rare amorphous aggregates. This study shows a possible connection between the aggregation propensity of an epitope and its affinity for the MHC class II cleft.

Modeling Protein Complexes and Molecular Assemblies Using Computational Methods

Many biological molecules are assembled into supramolecular complexes that are necessary to perform functions in the cell. Better understanding and characterization of these molecular assemblies are thus essential to further elucidate molecular mechanisms and key protein-protein interactions that could be targeted to modulate the protein binding affinity or develop new binders. Experimental access to structural information on these supramolecular assemblies is often hampered by the size of these systems that make their recombinant production and characterization rather difficult. Computational methods combining both structural data, molecular modeling techniques, and sequence coevolution information can thus offer a good alternative to gain access to the structural organization of protein complexes and assemblies. Herein, we present some computational methods to predict structural models of the protein partners, to search for interacting regions using coevolution information, and to build molecular assemblies. The approach is exemplified using a case study to model the succinate-quinone oxidoreductase heterocomplex.

Antibody Affinity Maturation Using Computational Methods: From an Initial Hit to Small-Scale Expression of Optimized Binders

Nanobodies (VHHs) are engineered fragments of the camelid single-chain immunoglobulins. The VHH domain contains the highly variable segments responsible for antigen recognition. VHHs can be easily produced as recombinant proteins. Their small size is a good advantage for in silico approaches. Computer methods represent a valuable strategy for the optimization and improvement of their binding affinity. They also allow for epitope selection offering the possibility to design new VHHs for regions of a target protein that are not naturally immunogenic. Here we present an in silico mutagenic protocol developed to improve the binding affinity of nanobodies together with the first step of their in vitro production. The method, already proven successful in improving the low Kd of a nanobody hit obtained by panning, can be employed for the ex novo design of antibody fragments against selected protein target epitopes.

Benchmarking AlphaFold for protein complex modeling reveals accuracy determinants

High‐resolution experimental structural determination of protein–protein interactions has led to valuable mechanistic insights, yet due to the massive number of interactions and experimental limitations there is a need for computational methods that can accurately model their structures. Here we explore the use of the recently developed deep learning method, AlphaFold, to predict structures of protein complexes from sequence. With a benchmark of 152 diverse heterodimeric protein complexes, multiple implementations and parameters of AlphaFold were tested for accuracy. Remarkably, many cases (43%) had near‐native models (medium or high critical assessment of predicted interactions accuracy) generated as top‐ranked predictions by AlphaFold, greatly surpassing the performance of unbound protein–protein docking (9% success rate for near‐native top‐ranked models), however AlphaFold modeling of antibody–antigen complexes within our set was unsuccessful. We identified sequence and structural features associated with lack of AlphaFold success, and we also investigated the impact of multiple sequence alignment input. Benchmarking of a multimer‐optimized version of AlphaFold (AlphaFold‐Multimer) with a set of recently released antibody–antigen structures confirmed a low rate of success for antibody–antigen complexes (11% success), and we found that T cell receptor–antigen complexes are likewise not accurately modeled by that algorithm, showing that adaptive immune recognition poses a challenge for the current AlphaFold algorithm and model. Overall, our study demonstrates that end‐to‐end deep learning can accurately model many transient protein complexes, and highlights areas of improvement for future developments to reliably model any protein–protein interaction of interest.

Multiple-Allele MHC Class II Epitope Engineering by a Molecular Dynamics-Based Evolution Protocol

Epitopes that bind simultaneously to all human alleles of Major Histocompatibility Complex class II (MHC II) are considered one of the key factors for the development of improved vaccines and cancer immunotherapies. To engineer MHC II multiple-allele binders, we developed a protocol called PanMHC-PARCE, based on the unsupervised optimization of the epitope sequence by single-point mutations, parallel explicit-solvent molecular dynamics simulations and scoring of the MHC II-epitope complexes. The key idea is accepting mutations that not only improve the affinity but also reduce the affinity gap between the alleles. We applied this methodology to enhance a Plasmodium vivax epitope for multiple-allele binding. In vitro rate-binding assays showed that four engineered peptides were able to bind with improved affinity toward multiple human MHC II alleles. Moreover, we demonstrated that mice immunized with the peptides exhibited interferon-gamma cellular immune response. Overall, the method enables the engineering of peptides with improved binding properties that can be used for the generation of new immunotherapies.

TRScore: a three-dimensional RepVGG-based scoring method for ranking protein docking models

MOTIVATION

Protein-protein interactions (PPI) play important roles in cellular activities. Due to the technical difficulty and high cost of experimental methods, there are considerable interests towards the development of computational approaches, such as protein docking, to decipher PPI patterns. One of the important and difficult aspects in protein docking is recognizing near-native conformations from a set of decoys, but unfortunately traditional scoring functions still suffer from limited accuracy. Therefore, new scoring methods are pressingly needed in methodological and/or practical implications.

RESULTS

We present a new deep learning-based scoring method for ranking protein-protein docking models based on a three-dimensional (3D) RepVGG network, named TRScore. To recognize near-native conformations from a set of decoys, TRScore voxelizes the protein-protein interface into a 3D grid labeled by the number of atoms in different physicochemical classes. Benefiting from the deep convolutional RepVGG architecture, TRScore can effectively capture the subtle differences between energetically favorable near-native models and unfavorable non-native decoys without needing extra information. TRScore was extensively evaluated on diverse test sets including protein-protein docking benchmark 5.0 update set, DockGround decoy set, as well as realistic CAPRI decoy set, and overall obtained a significant improvement over existing methods in cross validation and independent evaluations.

AVAILABILITY

Codes available at: https://github.com/BioinformaticsCSU/TRScore.

Protein Docking Model Evaluation by Graph Neural Networks

Physical interactions of proteins play key functional roles in many important cellular processes. To understand molecular mechanisms of such functions, it is crucial to determine the structure of protein complexes. To complement experimental approaches, which usually take a considerable amount of time and resources, various computational methods have been developed for predicting the structures of protein complexes. In computational modeling, one of the challenges is to identify near-native structures from a large pool of generated models. Here, we developed a deep learning-based approach named Graph Neural Network-based DOcking decoy eValuation scorE (GNN-DOVE). To evaluate a protein docking model, GNN-DOVE extracts the interface area and represents it as a graph. The chemical properties of atoms and the inter-atom distances are used as features of nodes and edges in the graph, respectively. GNN-DOVE was trained, validated, and tested on docking models in the Dockground database and further tested on a combined dataset of Dockground and ZDOCK benchmark as well as a CAPRI scoring dataset. GNN-DOVE performed better than existing methods, including DOVE, which is our previous development that uses a convolutional neural network on voxelized structure models.

Factual insights of the allosteric inhibition mechanism of SARS-CoV-2 main protease by quercetin: an in silico analysis

SARS-CoV-2 main protease (Mpro) cleaves the viral polypeptide 1a and 1ab in a site-specific ((L-Q|(S, A, G)) manner and produce functional enzymes for mediating viral replication. Numerous studies have reported synthetic competitive inhibitors against this target enzyme but increase in substrate concentration often reduces the effectiveness of such inhibitors. Allosteric inhibition by natural compound can provide safe and effective treatment by alleviating this limitation. Present study deals with in silico allosteric inhibition analysis of quercetin, against SARS-CoV-2-Mpro. Molecular docking of quercetin with Mpro revealed consistent binding of quercetin at a site other than active site in multiple runs, with the highest binding energy of - 8.31 kcal/mol, forming 6 H-bonds with residues Gln127, Cys128, Lys137, Asp289 and Glu290. Molecular dynamic simulation of 50 ns revealed high stability of Mpro-quercetin complex with RMSD values ranging from 0.1 to 0.25 nm. Moreover, native-Mpro and Mpro-quercetin complex conformations extracted at different time points from simulation trajectories were subjected to active site-specific docking with modelled substrate peptide (AVLQSGFR) by ZDOCK server. Results displayed site-specific cleavage of peptide when docked with native-Mpro. While substrate peptide remained intact when docked with Mpro-quercetin complex, also the binding energy of peptide reduced from 785 to 86 from 1 to 50 ns as quercetin induced alterations in the active site cavity reducing its affinity for the substrate. Further, no interactions were noticed between peptide and active site residues of Mpro-quercetin complex conformations at 40 and 50 ns. Hence, quercetin displayed effective allosteric inhibition potential against SARS-CoV-2 Mpro, and can be developed into an efficient treatment for COVID-19.

Evaluation of CONSRANK-Like Scoring Functions for Rescoring Ensembles of Protein–Protein Docking Poses

Scoring is a challenging step in protein–protein docking, where typically thousands of solutions are generated. In this study, we ought to investigate the contribution of consensus-rescoring, as introduced by Oliva et al. (2013) with the CONSRANK method, where the set of solutions is used to build statistics in order to identify recurrent solutions. We explore several ways to perform consensus-based rescoring on the ZDOCK decoy set for Benchmark 4. We show that the information of the interface size is critical for successful rescoring in this context, but that consensus rescoring in itself performs less well than traditional physics-based evaluation. The results of physics-based and consensus-based rescoring are partially overlapping, supporting the use of a combination of these approaches.

Integrating ab initio and template-based algorithms for protein-protein complex structure prediction

MOTIVATION

Template-based and template-free methods have both been widely used in predicting the structures of protein-protein complexes. Template-based modeling is effective when a reliable template is available, while template-free methods are required for predicting the binding modes or interfaces that have not been previously observed. Our goal is to combine the two methods to improve computational protein-protein complex structure prediction.

RESULTS

Here, we present a method to identify and combine high-confidence predictions of a template-based method (SPRING) with a template-free method (ZDOCK). Cross-validated using the protein-protein docking benchmark version 5.0, our method (ZING) achieved a success rate of 68.2%, outperforming SPRING and ZDOCK, with success rates of 52.1% and 35.9% respectively, when the top 10 predictions were considered per test case. In conclusion, a statistics-based method that evaluates and integrates predictions from template-based and template-free methods is more successful than either method independently.

AVAILABILITY AND IMPLEMENTATION

ZING is available for download as a Github repository (https://github.com/weng-lab/ZING.git).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Protein docking model evaluation by 3D deep convolutional neural networks

MOTIVATION

Many important cellular processes involve physical interactions of proteins. Therefore, determining protein quaternary structures provide critical insights for understanding molecular mechanisms of functions of the complexes. To complement experimental methods, many computational methods have been developed to predict structures of protein complexes. One of the challenges in computational protein complex structure prediction is to identify near-native models from a large pool of generated models.

RESULTS

We developed a convolutional deep neural network-based approach named DOcking decoy selection with Voxel-based deep neural nEtwork (DOVE) for evaluating protein docking models. To evaluate a protein docking model, DOVE scans the protein-protein interface of the model with a 3D voxel and considers atomic interaction types and their energetic contributions as input features applied to the neural network. The deep learning models were trained and validated on docking models available in the ZDock and DockGround databases. Among the different combinations of features tested, almost all outperformed existing scoring functions.

AVAILABILITY AND IMPLEMENTATION

Codes available at http://github.com/kiharalab/DOVE, http://kiharalab.org/dove/.

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Energy-based graph convolutional networks for scoring protein docking models

Structural information about protein-protein interactions, often missing at the interactome scale, is important for mechanistic understanding of cells and rational discovery of therapeutics. Protein docking provides a computational alternative for such information. However, ranking near-native docked models high among a large number of candidates, often known as the scoring problem, remains a critical challenge. Moreover, estimating model quality, also known as the quality assessment problem, is rarely addressed in protein docking. In this study, the two challenging problems in protein docking are regarded as relative and absolute scoring, respectively, and addressed in one physics-inspired deep learning framework. We represent protein and complex structures as intra- and inter-molecular residue contact graphs with atom-resolution node and edge features. And we propose a novel graph convolutional kernel that aggregates interacting nodes' features through edges so that generalized interaction energies can be learned directly from 3D data. The resulting energy-based graph convolutional networks (EGCN) with multihead attention are trained to predict intra- and inter-molecular energies, binding affinities, and quality measures (interface RMSD) for encounter complexes. Compared to a state-of-the-art scoring function for model ranking, EGCN significantly improves ranking for a critical assessment of predicted interactions (CAPRI) test set involving homology docking; and is comparable or slightly better for Score_set, a CAPRI benchmark set generated by diverse community-wide docking protocols not known to training data. For Score_set quality assessment, EGCN shows about 27% improvement to our previous efforts. Directly learning from 3D structure data in graph representation, EGCN represents the first successful development of graph convolutional networks for protein docking.

Performance of ZDOCK and IRAD in CAPRI rounds 39-45

We report docking performance on the six targets of Critical Assessment of PRedicted Interactions (CAPRI) rounds 39-45 that involved heteromeric protein-protein interactions and had the solved structures released since the rounds were held. Our general strategy involved protein-protein docking using ZDOCK, reranking using IRAD, and structural refinement using Rosetta. In addition, we made extensive use of experimental data to guide our docking runs. All the experimental information at the amino-acid level proved correct. However, for two targets, we also used protein-complex structures as templates for modeling interfaces. These resulted in incorrect predictions, presumably due to the low sequence identity between the targets and templates. Albeit a small number of targets, the performance described here compared somewhat less favorably with our previous CAPRI reports, which may be due to the CAPRI targets being increasingly challenging.

Identifying protein-protein interface via a novel multi-scale local sequence and structural representation

Background

Protein-protein interaction plays a key role in a multitude of biological processes, such as signal transduction, de novo drug design, immune responses, and enzymatic activities. Gaining insights of various binding abilities can deepen our understanding of the interaction. It is of great interest to understand how proteins in a complex interact with each other. Many efficient methods have been developed for identifying protein-protein interface.

Results

In this paper, we obtain the local information on protein-protein interface, through multi-scale local average block and hexagon structure construction. Given a pair of proteins, we use a trained support vector regression (SVR) model to select best configurations. On Benchmark v4.0, our method achieves average I_rmsd value of 3.28Å and overall F_nat value of 63%, which improves upon I_rmsd of 3.89Å and F_nat of 49% for ZRANK, and I_rmsd of 3.99Å and F_nat of 46% for ClusPro. On CAPRI targets, our method achieves average I_rmsd value of 3.45Å and overall F_nat value of 46%, which improves upon I_rmsd of 4.18Å and F_nat of 40% for ZRANK, and I_rmsd of 5.12Å and F_nat of 32% for ClusPro. The success rates by our method, FRODOCK 2.0, InterEvDock and SnapDock on Benchmark v4.0 are 41.5%, 29.0%, 29.4% and 37.0%, respectively.

Conclusion

Experiments show that our method performs better than some state-of-the-art methods, based on the prediction quality improved in terms of CAPRI evaluation criteria. All these results demonstrate that our method is a valuable technological tool for identifying protein-protein interface.

A consensus protocol for the in silico optimisation of antibody fragments

We present an in silico mutagenetic protocol for improving the binding affinity of single domain antibodies (or nanobodies, VHHs). The method iteratively attempts random mutations in the interacting region of the protein and evaluates the resulting binding affinity towards the target by scoring, with a collection of scoring functions, short explicit solvent molecular dynamics trajectories of the binder-target complexes. The acceptance/rejection of each attempted mutation is carried out by a consensus decision-making algorithm, which considers all individual assessments derived from each scoring function. The method was benchmarked by evolving a single complementary determining region (CDR) of an anti-HER2 VHH hit obtained by direct panning of a phage display library. The optimised VHH mutant showed significantly enhanced experimental affinity with respect to the original VHH it matured from. The protocol can be employed as it is for the optimization of peptides, antibody fragments, and (given enough computational power) larger antibodies.

Predicting the Affinity of Peptides to Major Histocompatibility Complex Class II by Scoring Molecular Dynamics Simulations

Predicting the binding affinity of peptides able to interact with major histocompatibility complex (MHC) molecules is a priority for researchers working in the identification of novel vaccines candidates. Most available approaches are based on the analysis of the sequence of peptides of known experimental affinity. However, for MHC class II receptors, these approaches are not very accurate, due to the intrinsic flexibility of the complex. To overcome these limitations, we propose to estimate the binding affinity of peptides bound to an MHC class II by averaging the score of the configurations from finite-temperature molecular dynamics simulations. The score is estimated for 18 different scoring functions, and we explored the optimal manner for combining them. To test the predictions, we considered eight peptides of known binding affinity. We found that six scoring functions correlate with the experimental ranking of the peptides significantly better than the others. We then assessed a set of techniques for combining the scoring functions by linear regression and logistic regression. We obtained a maximum accuracy of 82% for the predicted sign of the binding affinity using a logistic regression with optimized weights. These results are potentially useful to improve the reliability of in silico protocols to design high-affinity binding peptides for MHC class II receptors.

Determination of an effective scoring function for RNA-RNA interactions with a physics-based double-iterative method

RNA–RNA interactions play fundamental roles in gene and cell regulation. Therefore, accurate prediction of RNA–RNA interactions is critical to determine their complex structures and understand the molecular mechanism of the interactions. Here, we have developed a physics-based double-iterative strategy to determine the effective potentials for RNA–RNA interactions based on a training set of 97 diverse RNA–RNA complexes. The double-iterative strategy circumvented the reference state problem in knowledge-based scoring functions by updating the potentials through iteration and also overcame the decoy-dependent limitation in previous iterative methods by constructing the decoys iteratively. The derived scoring function, which is referred to as DITScoreRR, was evaluated on an RNA–RNA docking benchmark of 60 test cases and compared with three other scoring functions. It was shown that for bound docking, our scoring function DITScoreRR obtained the excellent success rates of 90% and 98.3% in binding mode predictions when the top 1 and 10 predictions were considered, compared to 63.3% and 71.7% for van der Waals interactions, 45.0% and 65.0% for ITScorePP, and 11.7% and 26.7% for ZDOCK 2.1, respectively. For unbound docking, DITScoreRR achieved the good success rates of 53.3% and 71.7% in binding mode predictions when the top 1 and 10 predictions were considered, compared to 13.3% and 28.3% for van der Waals interactions, 11.7% and 26.7% for our ITScorePP, and 3.3% and 6.7% for ZDOCK 2.1, respectively. DITScoreRR also performed significantly better in ranking decoys and obtained significantly higher score-RMSD correlations than the other three scoring functions. DITScoreRR will be of great value for the prediction and design of RNA structures and RNA–RNA complexes.

Alternative interaction sites in the influenza A virus nucleoprotein mediate viral escape from the importin-α7 mediated nuclear import pathway

Influenza A viruses are able to adapt to restrictive conditions due to their high mutation rates. Importin-α7 is a component of the nuclear import machinery required for avian-mammalian adaptation and replicative fitness in human cells. Here, we elucidate the mechanisms by which influenza A viruses may escape replicative restriction in the absence of importin-α7. To address this question, we assessed viral evolution in mice lacking the importin-α7 gene. We show that three mutations in particular occur with high frequency in the viral nucleoprotein (NP) protein (G102R, M105K and D375N) in a specific structural area upon in vivo adaptation. Moreover, our findings suggest that the adaptive NP mutations mediate viral escape from importin-α7 requirement likely due to the utilization of alternative interaction sites in NP beyond the classical nuclear localization signal. However, viral escape from importin-α7 by mutations in NP is, at least in part, associated with reduced viral replication highlighting the crucial contribution of importin-α7 to replicative fitness in human cells.

Distinguishing crystallographic from biological interfaces in protein complexes: role of intermolecular contacts and energetics for classification

BACKGROUND

Study of macromolecular assemblies is fundamental to understand functions in cells. X-ray crystallography is the most common technique to solve their 3D structure at atomic resolution. In a crystal, however, both biologically-relevant interfaces and non-specific interfaces resulting from crystallographic packing are observed. Due to the complexity of the biological assemblies currently tackled, classifying those interfaces, i.e. distinguishing biological from crystal lattice interfaces, is not trivial and often prone to errors. In this context, analyzing the physico-chemical characteristics of biological/crystal interfaces can help researchers identify possible features that distinguish them and gain a better understanding of the systems.

RESULTS

In this work, we are providing new insights into the differences between biological and crystallographic complexes by focusing on "pair-properties" of interfaces that have not yet been fully investigated. We investigated properties such intermolecular residue-residue contacts (already successfully applied to the prediction of binding affinities) and interaction energies (electrostatic, Van der Waals and desolvation). By using the XtalMany and BioMany interface datasets, we show that interfacial residue contacts, classified as a function of their physico-chemical properties, can distinguish between biological and crystallographic interfaces. The energetic terms show, on average, higher values for crystal interfaces, reflecting a less stable interface due to crystal packing compared to biological interfaces. By using a variety of machine learning approaches, we trained a new interface classification predictor based on contacts and interaction energetic features. Our predictor reaches an accuracy in classifying biological vs crystal interfaces of 0.92, compared to 0.88 for EPPIC (one of the main state-of-the-art classifiers reporting same performance as PISA).

CONCLUSION

In this work we have gained insights into the nature of intermolecular contacts and energetics terms distinguishing biological from crystallographic interfaces. Our findings might have a broader applicability in structural biology, for example for the identification of near native poses in docking. We implemented our classification approach into an easy-to-use and fast software, freely available to the scientific community from http://github.com/haddocking/interface-classifier .

Mutation of key amino acids in the polygalacturonase-inhibiting proteins CkPGIP1 and GhPGIP1 improves resistance to Verticillium wilt in cotton

Verticillium wilt, one of the most devastating diseases of cotton (Gossypium hirsutum), causes severe yield and quality losses. Given the effectiveness of plant polygalacturonase-inhibiting proteins (PGIPs) in reducing fungal polygalacturonase (PG) activity, it is necessary to uncover the key functional amino acids to enhance cotton resistance to Verticillium dahliae. To identify novel antifungal proteins, the selectivity of key amino acids was investigated by screening against a panel of relevant PG-binding residues. Based on the obtained results, homologous models of the mutants were established. The docking models showed that hydrogen bonds and structural changes in the convex face in the conserved portion of leucine-rich repeats (LRRs) may be essential for enhanced recognition of PG. Additionally, we successfully constructed Cynanchum komarovii PGIP1 (CkPGIP1) mutants Asp176Val, Pro249Gln, and Asp176Val/Pro249Gln and G. hirsutum PGIP1 (GhPGIP1) mutants Glu169Val, Phe242Gln, and Glu169Val/Phe242Gln with site-directed mutagenesis. The proteins of interest can effectively inhibit VdPG1 activity and V. dahliae mycelial growth in a dose-dependent manner. Importantly, mutants that overproduced PGIP in Arabidopsis and cotton showed enhanced resistance to V. dahliae, with reduced Verticillium-associated chlorosis and wilting. Furthermore, the lignin content was measured in mutant-overexpressing plants, and the results showed enhanced lignification of the xylem, which blocked the spread of V. dahliae. Thus, using site-directed mutagenesis assays, we showed that mutations in CkPGIP1 and GhPGIP1 give rise to PGIP versatility, which allows evolving recognition specificities for PG and is required to promote Verticillium resistance in cotton by restricting the growth of invasive fungal pathogens.

Contact Potential for Structure Prediction of Proteins and Protein Complexes from Potts Model

The energy function is the key component of protein modeling methodology. This work presents a semianalytical approach to the development of contact potentials for protein structure modeling. Residue-residue and atom-atom contact energies were derived by maximizing the probability of observing native sequences in a nonredundant set of protein structures. The optimization task was formulated as an inverse statistical mechanics problem applied to the Potts model. Its solution by pseudolikelihood maximization provides consistent estimates of coupling constants at atomic and residue levels. The best performance was achieved when interacting atoms were grouped according to their physicochemical properties. For individual protein structures, the performance of the contact potentials in distinguishing near-native structures from the decoys is similar to the top-performing scoring functions. The potentials also yielded significant improvement in the protein docking success rates. The potentials recapitulated experimentally determined protein stability changes upon point mutations and protein-protein binding affinities. The approach offers a different perspective on knowledge-based potentials and may serve as the basis for their further development.

Integrating Cross-Linking Experiments with Ab Initio Protein-Protein Docking

Ab initio protein-protein docking algorithms often rely on experimental data to identify the most likely complex structure. We integrated protein-protein docking with the experimental data of chemical cross-linking followed by mass spectrometry. We tested our approach using 19 cases that resulted from an exhaustive search of the Protein Data Bank for protein complexes with cross-links identified in our experiments. We implemented cross-links as constraints based on Euclidean distance or void-volume distance. For most test cases, the rank of the top-scoring near-native prediction was improved by at least twofold compared with docking without the cross-link information, and the success rate for the top 5 predictions nearly tripled. Our results demonstrate the delicate balance between retaining correct predictions and eliminating false positives. Several test cases had multiple components with distinct interfaces, and we present an approach for assigning cross-links to the interfaces. Employing the symmetry information for these cases further improved the performance of complex structure prediction.

Binding affinity prediction of nanobody-protein complexes by scoring of molecular dynamics trajectories

Nanobodies offer a viable alternative to antibodies for engineering high affinity binders. Their small size has an additional advantage: it allows exploiting computational protocols for optimizing their biophysical features, such as the binding affinity. The efficient prediction of this quantity is still considered a daunting task especially for modelled complexes. We show how molecular dynamics can successfully assist in the binding affinity prediction of modelled nanobody-protein complexes. The approximate initial configurations obtained by in silico design must undergo large rearrangements before achieving a stable conformation, in which the binding affinity can be meaningfully estimated. The scoring functions developed for the affinity evaluation of crystal structures will provide accurate estimates for modelled binding complexes if the scores are averaged over long finite temperature molecular dynamics simulations.

Determining Complex Structures using Docking Method with Single Particle Scattering Data

Protein complexes are critical for many molecular functions. Due to intrinsic flexibility and dynamics of complexes, their structures are more difficult to determine using conventional experimental methods, in contrast to individual subunits. One of the major challenges is the crystallization of protein complexes. Using X-ray free electron lasers (XFELs), it is possible to collect scattering signals from non-crystalline protein complexes, but data interpretation is more difficult because of unknown orientations. Here, we propose a hybrid approach to determine protein complex structures by combining XFEL single particle scattering data with computational docking methods. Using simulations data, we demonstrate that a small set of single particle scattering data collected at random orientations can be used to distinguish the native complex structure from the decoys generated using docking algorithms. The results also indicate that a small set of single particle scattering data is superior to spherically averaged intensity profile in distinguishing complex structures. Given the fact that XFEL experimental data are difficult to acquire and at low abundance, this hybrid approach should find wide applications in data interpretations.

Performance of ZDOCK and IRAD in CAPRI rounds 28-34

We report the performance of our protein-protein docking pipeline, including the ZDOCK rigid-body docking algorithm, on 19 targets in CAPRI rounds 28-34. Following the docking step, we reranked the ZDOCK predictions using the IRAD scoring function, pruned redundant predictions, performed energy landscape analysis, and utilized our interface prediction approach RCF. In addition, we applied constraints to the search space based on biological information that we culled from the literature, which increased the chance of making a correct prediction. For all but two targets we were able to find and apply biological information and we found the information to be highly accurate, indicating that effective incorporation of biological information is an important component for protein-protein docking. Proteins 2017; 85:408-416. © 2016 Wiley Periodicals, Inc.

Ranking protein-protein docking results using steered molecular dynamics and potential of mean force calculations

Crystallization of protein-protein complexes can often be problematic and therefore computational structural models are often relied on. Such models are often generated using protein-protein docking algorithms, where one of the main challenges is selecting which of several thousand potential predictions represents the most near-native complex. We have developed a novel technique that involves the use of steered molecular dynamics (sMD) and umbrella sampling to identify near-native complexes among protein-protein docking predictions. Using this technique, we have found a strong correlation between our predictions and the interface RMSD (iRMSD) in ten diverse test systems. On two of the systems, we investigated if the prediction results could be further improved using potential of mean force calculations. We demonstrated that a near-native (<2.0 Å iRMSD) structure could be identified in the top-1 ranked position for both systems. © 2016 Wiley Periodicals, Inc.

CLUB-MARTINI: Selecting Favourable Interactions amongst Available Candidates, a Coarse-Grained Simulation Approach to Scoring Docking Decoys

Large-scale identification of native binding orientations is crucial for understanding the role of protein-protein interactions in their biological context. Measuring binding free energy is the method of choice to estimate binding strength and reveal the relevance of particular conformations in which proteins interact. In a recent study, we successfully applied coarse-grained molecular dynamics simulations to measure binding free energy for two protein complexes with similar accuracy to full-atomistic simulation, but 500-fold less time consuming. Here, we investigate the efficacy of this approach as a scoring method to identify stable binding conformations from thousands of docking decoys produced by protein docking programs. To test our method, we first applied it to calculate binding free energies of all protein conformations in a CAPRI (Critical Assessment of PRedicted Interactions) benchmark dataset, which included over 19000 protein docking solutions for 15 benchmark targets. Based on the binding free energies, we ranked all docking solutions to select the near-native binding modes under the assumption that the native-solutions have lowest binding free energies. In our top 100 ranked structures, for the ‘easy’ targets that have many near-native conformations, we obtain a strong enrichment of acceptable or better quality structures; for the ‘hard’ targets without near-native decoys, our method is still able to retain structures which have native binding contacts. Moreover, in our top 10 selections, CLUB-MARTINI shows a comparable performance when compared with other state-of-the-art docking scoring functions. As a proof of concept, CLUB-MARTINI performs remarkably well for many targets and is able to pinpoint near-native binding modes in the top selections. To the best of our knowledge, this is the first time interaction free energy calculated from MD simulations have been used to rank docking solutions at a large scale.

Evaluation of the coarse-grained OPEP force field for protein-protein docking

Background

Knowing the binding site of protein–protein complexes helps understand their function and shows possible regulation sites. The ultimate goal of protein–protein docking is the prediction of the three-dimensional structure of a protein–protein complex. Docking itself only produces plausible candidate structures, which must be ranked using scoring functions to identify the structures that are most likely to occur in nature.

Methods

In this work, we rescore rigid body protein–protein predictions using the optimized potential for efficient structure prediction (OPEP), which is a coarse-grained force field. Using a force field based on continuous functions rather than a grid-based scoring function allows the introduction of protein flexibility during the docking procedure. First, we produce protein–protein predictions using ZDOCK, and after energy minimization via OPEP we rank them using an OPEP-based soft rescoring function. We also train the rescoring function for different complex classes and demonstrate its improved performance for an independent dataset.

Results

The trained rescoring function produces a better ranking than ZDOCK for more than 50 % of targets, rising to over 70 % when considering only enzyme/inhibitor complexes.

Conclusions

This study demonstrates for the first time that energy functions derived from the coarse-grained OPEP force field can be employed to rescore predictions for protein–protein complexes.

Performance of ZDOCK in CAPRI rounds 20-26

We report the performance of our approaches for protein-protein docking and interface analysis in CAPRI rounds 20-26. At the core of our pipeline was the ZDOCK program for rigid-body protein-protein docking. We then reranked the ZDOCK predictions using the ZRANK or IRAD scoring functions, pruned and analyzed energy landscapes using clustering, and analyzed the docking results using our interface prediction approach RCF. When possible, we used biological information from the literature to apply constraints to the search space during or after the ZDOCK runs. For approximately half of the standard docking challenges we made at least one prediction that was acceptable or better. For the scoring challenges we made acceptable or better predictions for all but one target. This indicates that our scoring functions are generally able to select the correct binding mode.

Structural neighboring property for identifying protein-protein binding sites

Background

The protein-protein interaction plays a key role in the control of many biological functions, such as drug design and functional analysis. Determination of binding sites is widely applied in molecular biology research. Therefore, many efficient methods have been developed for identifying binding sites. In this paper, we calculate structural neighboring property through Voronoi diagram. Using 6,438 complexes, we study local biases of structural neighboring property on interface.

Results

We propose a novel statistical method to extract interacting residues, and interacting patches can be clustered as predicted interface residues. In addition, structural neighboring property can be adopted to construct a new energy function, for evaluating docking solutions. It includes new statistical property as well as existing energy items. Comparing to existing methods, our approach improves overall F_nat value by at least 3%. On Benchmark v4.0, our method has average I_rmsd value of 3.31Å and overall F_nat value of 63%, which improves upon I_rmsd of 3.89 Å and F_nat of 49% for ZRANK, and I_rmsd of 3.99Å and F_nat of 46% for ClusPro. On the CAPRI targets, our method has average I_rmsd value of 3.46 Å and overall F_nat value of 45%, which improves upon I_rmsd of 4.18 Å and F_nat of 40% for ZRANK, and I_rmsd of 5.12 Å and F_nat of 32% for ClusPro.

Conclusions

Experiments show that our method achieves better results than some state-of-the-art methods for identifying protein-protein binding sites, with the prediction quality improved in terms of CAPRI evaluation criteria.

Surfing the Protein-Protein Interaction Surface Using Docking Methods: Application to the Design of PPI Inhibitors

Blocking protein-protein interactions (PPI) using small molecules or peptides modulates biochemical pathways and has therapeutic significance. PPI inhibition for designing drug-like molecules is a new area that has been explored extensively during the last decade. Considering the number of available PPI inhibitor databases and the limited number of 3D structures available for proteins, docking and scoring methods play a major role in designing PPI inhibitors as well as stabilizers. Docking methods are used in the design of PPI inhibitors at several stages of finding a lead compound, including modeling the protein complex, screening for hot spots on the protein-protein interaction interface and screening small molecules or peptides that bind to the PPI interface. There are three major challenges to the use of docking on the relatively flat surfaces of PPI. In this review we will provide some examples of the use of docking in PPI inhibitor design as well as its limitations. The combination of experimental and docking methods with improved scoring function has thus far resulted in few success stories of PPI inhibitors for therapeutic purposes. Docking algorithms used for PPI are in the early stages, however, and as more data are available docking will become a highly promising area in the design of PPI inhibitors or stabilizers.

Molecular modeling of the binding modes of the iron-sulfur protein to the Jac1 co-chaperone from Saccharomyces cerevisiae by all-atom and coarse-grained approaches

The iron-sulfur protein 1 (Isu1) and the J-type co-chaperone Jac1 from yeast are part of a huge ATP-dependent system, and both interact with Hsp70 chaperones. Interaction of Isu1 and Jac1 is a part of the iron-sulfur cluster biogenesis system in mitochondria. In this study, the structure and dynamics of the yeast Isu1-Jac1 complex has been modeled. First, the complete structure of Isu1 was obtained by homology modeling using the I-TASSER server and YASARA software and thereafter tested for stability in the all-atom force field AMBER. Then, the known experimental structure of Jac1 was adopted to obtain initial models of the Isu1-Jac1 complex by using the ZDOCK server for global and local docking and the AutoDock software for local docking. Three most probable models were subsequently subjected to the coarse-grained molecular dynamics simulations with the UNRES force field to obtain the final structures of the complex. In the most probable model, Isu1 binds to the left face of the Γ-shaped Jac1 molecule by the β-sheet section of Isu1. Residues L105 , L109 , and Y163 of Jac1 have been assessed by mutation studies to be essential for binding (Ciesielski et al., J Mol Biol 2012; 417:1-12). These residues were also found, by UNRES/molecular dynamics simulations, to be involved in strong interactions between Isu1 and Jac1 in the complex. Moreover, N(95), T(98), P(102), H(112), V(159), L(167), and A(170) of Jac1, not yet tested experimentally, were also found to be important in binding.

NPPD: A Protein-Protein Docking Scoring Function Based on Dyadic Differences in Networks of Hydrophobic and Hydrophilic Amino Acid Residues

Protein-protein docking (PPD) predictions usually rely on the use of a scoring function to rank docking models generated by exhaustive sampling. To rank good models higher than bad ones, a large number of scoring functions have been developed and evaluated, but the methods used for the computation of PPD predictions remain largely unsatisfactory. Here, we report a network-based PPD scoring function, the NPPD, in which the network consists of two types of network nodes, one for hydrophobic and the other for hydrophilic amino acid residues, and the nodes are connected when the residues they represent are within a certain contact distance. We showed that network parameters that compute dyadic interactions and those that compute heterophilic interactions of the amino acid networks thus constructed allowed NPPD to perform well in a benchmark evaluation of 115 PPD scoring functions, most of which, unlike NPPD, are based on some sort of protein-protein interaction energy. We also showed that NPPD was highly complementary to these energy-based scoring functions, suggesting that the combined use of conventional scoring functions and NPPD might significantly improve the accuracy of current PPD predictions.

IFACEwat: the interfacial water-implemented re-ranking algorithm to improve the discrimination of near native structures for protein rigid docking

Background

Protein-protein docking is an in silico method to predict the formation of protein complexes. Due to limited computational resources, the protein-protein docking approach has been developed under the assumption of rigid docking, in which one of the two protein partners remains rigid during the protein associations and water contribution is ignored or implicitly presented. Despite obtaining a number of acceptable complex predictions, it seems to-date that most initial rigid docking algorithms still find it difficult or even fail to discriminate successfully the correct predictions from the other incorrect or false positive ones. To improve the rigid docking results, re-ranking is one of the effective methods that help re-locate the correct predictions in top high ranks, discriminating them from the other incorrect ones.

In this paper, we propose a new re-ranking technique using a new energy-based scoring function, namely IFACEwat - a combined Interface Atomic Contact Energy (IFACE) and water effect. The IFACEwat aims to further improve the discrimination of the near-native structures of the initial rigid docking algorithm ZDOCK3.0.2. Unlike other re-ranking techniques, the IFACEwat explicitly implements interfacial water into the protein interfaces to account for the water-mediated contacts during the protein interactions.

Results

Our results showed that the IFACEwat increased both the numbers of the near-native structures and improved their ranks as compared to the initial rigid docking ZDOCK3.0.2. In fact, the IFACEwat achieved a success rate of 83.8% for Antigen/Antibody complexes, which is 10% better than ZDOCK3.0.2. As compared to another re-ranking technique ZRANK, the IFACEwat obtains success rates of 92.3% (8% better) and 90% (5% better) respectively for medium and difficult cases. When comparing with the latest published re-ranking method F²Dock, the IFACEwat performed equivalently well or even better for several Antigen/Antibody complexes.

Conclusions

With the inclusion of interfacial water, the IFACEwat improves mostly results of the initial rigid docking, especially for Antigen/Antibody complexes. The improvement is achieved by explicitly taking into account the contribution of water during the protein interactions, which was ignored or not fully presented by the initial rigid docking and other re-ranking techniques. In addition, the IFACEwat maintains sufficient computational efficiency of the initial docking algorithm, yet improves the ranks as well as the number of the near native structures found. As our implementation so far targeted to improve the results of ZDOCK3.0.2, and particularly for the Antigen/Antibody complexes, it is expected in the near future that more implementations will be conducted to be applicable for other initial rigid docking algorithms.

Large scale characterization of the LC13 TCR and HLA-B8 structural landscape in reaction to 172 altered peptide ligands: a molecular dynamics simulation study

The interplay between T cell receptors (TCRs) and peptides bound by major histocompatibility complexes (MHCs) is one of the most important interactions in the adaptive immune system. Several previous studies have computationally investigated their structural dynamics. On the basis of these simulations several structural and dynamical properties have been proposed as effectors of the immunogenicity. Here we present the results of a large scale Molecular Dynamics simulation study consisting of 100 ns simulations of 172 different complexes. These complexes consisted of all possible point mutations of the Epstein Barr Virus peptide FLRGRAYGL bound by HLA-B*08:01 and presented to the LC13 TCR. We compare the results of these 172 structural simulations with experimental immunogenicity data. We found that simulations with more immunogenic peptides and those with less immunogenic peptides are in fact highly similar and on average only minor differences in the hydrogen binding footprints, interface distances, and the relative orientation between the TCR chains are present. Thus our large scale data analysis shows that many previously suggested dynamical and structural properties of the TCR/peptide/MHC interface are unlikely to be conserved causal factors for peptide immunogenicity.