The HADDOCK2.4 web server for integrative modeling of biomolecular complexes.

Interactions between macromolecules, such as proteins and nucleic acids, are essential for cellular functions. Experimental methods can fail to provide all the information required to fully model biomolecular complexes at atomic resolution, particularly for large and heterogeneous assemblies. Integrative computational approaches have, therefore, gained popularity, complementing traditional experimental methods in structural biology. Here, we introduce HADDOCK2.4, an integrative modeling platform, and its updated web interface ( https://wenmr.science.uu.nl/haddock2.4 ). The platform seamlessly integrates diverse experimental and theoretical data to generate high-quality models of macromolecular complexes. The user-friendly web server offers automated parameter settings, access to distributed computing resources, and pre- and post-processing steps that enhance the user experience. To present the web server's various interfaces and features, we demonstrate two different applications: (i) we predict the structure of an antibody-antigen complex by using NMR data for the antigen and knowledge of the hypervariable loops for the antibody, and (ii) we perform coarse-grained modeling of PRC1 with a nucleosome particle guided by mutagenesis and functional data. The described protocols require some basic familiarity with molecular modeling and the Linux command shell. This new version of our widely used HADDOCK web server allows structural biologists and non-experts to explore intricate macromolecular assemblies encompassing various molecule types.

Predicting binding events in very flexible, allosteric, multi-domain proteins.

Knowledge of the structures formed by proteins and small ligands is of fundamental importance for understanding molecular principles of chemotherapy and for designing new and more effective drugs. Due to the still high costs and to the several limitations of experimental techniques, it is most often desirable to predict these ligand-protein complexes in silico, particularly when screening for new putative drugs from databases of millions of compounds. While virtual screening based on molecular docking is widely used for this purpose, it generally fails in mimicking binding events associated with large conformational changes in the protein, particularly when the latter involve multiple domains. In this work, we describe a new methodology aimed at generating bound-like conformations of very flexible and allosteric proteins bearing multiple binding sites. Validation was performed on the enzyme adenylate kinase (ADK), a paradigmatic example of proteins that undergo very large conformational changes upon ligand binding. By only exploiting the unbound structure and the putative binding sites of the protein, we generated a significant fraction of bound-like structures, which employed in ensemble-docking calculations allowed to find native-like poses of substrates, inhibitors, and catalytically incompetent binders. Our protocol provides a general framework for the generation of bound-like conformations of flexible proteins that are suitable to host different ligands, demonstrating high sensitivity to the fine chemical details that regulate protein's activity. We foresee applications in virtual screening for difficult targets, prediction of the impact of amino acid mutations on structure and dynamics, and protein engineering.

ARCTIC-3D: automatic retrieval and clustering of interfaces in complexes from 3D structural information.

The formation of a stable complex between proteins lies at the core of a wide variety of biological processes and has been the focus of countless experiments. The huge amount of information contained in the protein structural interactome in the Protein Data Bank can now be used to characterise and classify the existing biological interfaces. We here introduce ARCTIC-3D, a fast and user-friendly data mining and clustering software to retrieve data and rationalise the interface information associated with the protein input data. We demonstrate its use by various examples ranging from showing the increased interaction complexity of eukaryotic proteins, 20% of which on average have more than 3 different interfaces compared to only 10% for prokaryotes, to associating different functions to different interfaces. In the context of modelling biomolecular assemblies, we introduce the concept of "recognition entropy", related to the number of possible interfaces of the components of a protein-protein complex, which we demonstrate to correlate with the modelling difficulty in classical docking approaches. The identified interface clusters can also be used to generate various combinations of interface-specific restraints for integrative modelling. The ARCTIC-3D software is freely available at github.com/haddocking/arctic3d and can be accessed as a web-service at wenmr.science.uu.nl/arctic3d.

DeepRank-GNN-esm: a graph neural network for scoring protein-protein models using protein language model.

Motivation: Protein-Protein interactions (PPIs) play critical roles in numerous cellular processes. By modelling the 3D structures of the correspond protein complexes valuable insights can be obtained, providing, e.g. starting points for drug and protein design. One challenge in the modelling process is however the identification of near-native models from the large pool of generated models. To this end we have previously developed DeepRank-GNN, a graph neural network that integrates structural and sequence information to enable effective pattern learning at PPI interfaces. Its main features are related to the Position Specific Scoring Matrices (PSSMs), which are computationally expensive to generate, significantly limits the algorithm's usability. Results: We introduce here DeepRank-GNN-esm that includes as additional features protein language model embeddings from the ESM-2 model. We show that the ESM-2 embeddings can actually replace the PSSM features at no cost in-, or even better performance on two PPI-related tasks: scoring docking poses and detecting crystal artifacts. This new DeepRank version bypasses thus the need of generating PSSM, greatly improving the usability of the software and opening new application opportunities for systems for which PSSM profiles cannot be obtained or are irrelevant (e.g. antibody-antigen complexes). Availability and implementation: DeepRank-GNN-esm is freely available from https://github.com/DeepRank/DeepRank-GNN-esm.

Impact of AlphaFold on structure prediction of protein complexes: The CASP15-CAPRI experiment.

We present the results for CAPRI Round 54, the 5th joint CASP-CAPRI protein assembly prediction challenge. The Round offered 37 targets, including 14 homodimers, 3 homo-trimers, 13 heterodimers including 3 antibody-antigen complexes, and 7 large assemblies. On average ~70 CASP and CAPRI predictor groups, including more than 20 automatics servers, submitted models for each target. A total of 21 941 models submitted by these groups and by 15 CAPRI scorer groups were evaluated using the CAPRI model quality measures and the DockQ score consolidating these measures. The prediction performance was quantified by a weighted score based on the number of models of acceptable quality or higher submitted by each group among their five best models. Results show substantial progress achieved across a significant fraction of the 60+ participating groups. High-quality models were produced for about 40% of the targets compared to 8% two years earlier. This remarkable improvement is due to the wide use of the AlphaFold2 and AlphaFold2-Multimer software and the confidence metrics they provide. Notably, expanded sampling of candidate solutions by manipulating these deep learning inference engines, enriching multiple sequence alignments, or integration of advanced modeling tools, enabled top performing groups to exceed the performance of a standard AlphaFold2-Multimer version used as a yard stick. This notwithstanding, performance remained poor for complexes with antibodies and nanobodies, where evolutionary relationships between the binding partners are lacking, and for complexes featuring conformational flexibility, clearly indicating that the prediction of protein complexes remains a challenging problem.

An antibiotic from an uncultured bacterium binds to an immutable target.

Antimicrobial resistance is a leading mortality factor worldwide. Here, we report the discovery of clovibactin, an antibiotic isolated from uncultured soil bacteria. Clovibactin efficiently kills drug-resistant Gram-positive bacterial pathogens without detectable resistance. Using biochemical assays, solid-state nuclear magnetic resonance, and atomic force microscopy, we dissect its mode of action. Clovibactin blocks cell wall synthesis by targeting pyrophosphate of multiple essential peptidoglycan precursors (C55PP, lipid II, and lipid IIIWTA). Clovibactin uses an unusual hydrophobic interface to tightly wrap around pyrophosphate but bypasses the variable structural elements of precursors, accounting for the lack of resistance. Selective and efficient target binding is achieved by the sequestration of precursors into supramolecular fibrils that only form on bacterial membranes that contain lipid-anchored pyrophosphate groups. This potent antibiotic holds the promise of enabling the design of improved therapeutics that kill bacterial pathogens without resistance development.

Discriminating physiological from non-physiological interfaces in structures of protein complexes: A community-wide study.

Reliably scoring and ranking candidate models of protein complexes and assigning their oligomeric state from the structure of the crystal lattice represent outstanding challenges. A community-wide effort was launched to tackle these challenges. The latest resources on protein complexes and interfaces were exploited to derive a benchmark dataset consisting of 1677 homodimer protein crystal structures, including a balanced mix of physiological and non-physiological complexes. The non-physiological complexes in the benchmark were selected to bury a similar or larger interface area than their physiological counterparts, making it more difficult for scoring functions to differentiate between them. Next, 252 functions for scoring protein-protein interfaces previously developed by 13 groups were collected and evaluated for their ability to discriminate between physiological and non-physiological complexes. A simple consensus score generated using the best performing score of each of the 13 groups, and a cross-validated Random Forest (RF) classifier were created. Both approaches showed excellent performance, with an area under the Receiver Operating Characteristic (ROC) curve of 0.93 and 0.94, respectively, outperforming individual scores developed by different groups. Additionally, AlphaFold2 engines recalled the physiological dimers with significantly higher accuracy than the non-physiological set, lending support to the reliability of our benchmark dataset annotations. Optimizing the combined power of interface scoring functions and evaluating it on challenging benchmark datasets appears to be a promising strategy.

Improving the quality of co-evolution intermolecular contact prediction with DisVis.

The steep rise in protein sequences and structures has paved the way for bioinformatics approaches to predict residue-residue interactions in protein complexes. Multiple sequence alignments are commonly used in contact predictions to identify co-evolving residues. These contacts, however, often include false positives (FPs), which may impair their use to predict three dimensional structures of biomolecular complexes and affect the accuracy of the generated models. Previously, we have developed DisVis to identify FP in mass spectrometry cross-linking data. DisVis allows to assess the accessible interaction space between two proteins consistent with a set of distance restraints. Here, we investigate if a similar approach could be applied to co-evolution predicted contacts in order to improve their precision prior to using them for modeling. We analyze co-evolution contact predictions with DisVis for a set of 26 protein-protein complexes. The DisVis-reranked and the original co-evolution contacts are then used to model the complexes with our integrative docking software HADDOCK using different filtering scenarios. Our results show that HADDOCK is robust with respect to the precision of the predicted contacts due to the 50% random contact removal during docking and can enhance the quality of docking predictions when combined with DisVis filtering for low precision contact data. DisVis can thus have a beneficial effect on low quality data, but overall HADDOCK can accommodate FP restraints without negatively impacting the quality of the resulting models. Other more precision-sensitive docking protocols might, however, benefit from the increased precision of the predicted contacts after DisVis filtering.

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.

DeepRank-GNN: a graph neural network framework to learn patterns in protein-protein interfaces.

MOTIVATION: Gaining structural insights into the protein-protein interactome is essential to understand biological phenomena and extract knowledge for rational drug design or protein engineering. We have previously developed DeepRank, a deep-learning framework to facilitate pattern learning from protein-protein interfaces using convolutional neural network (CNN) approaches. However, CNN is not rotation invariant and data augmentation is required to desensitize the network to the input data orientation which dramatically impairs the computation performance. Representing protein-protein complexes as atomic- or residue-scale rotation invariant graphs instead enables using graph neural networks (GNN) approaches, bypassing those limitations. RESULTS: We have developed DeepRank-GNN, a framework that converts protein-protein interfaces from PDB 3D coordinates files into graphs that are further provided to a pre-defined or user-defined GNN architecture to learn problem-specific interaction patterns. DeepRank-GNN is designed to be highly modularizable, easily customized and is wrapped into a user-friendly python3 package. Here, we showcase DeepRank-GNN's performance on two applications using a dedicated graph interaction neural network: (i) the scoring of docking poses and (ii) the discriminating of biological and crystal interfaces. In addition to the highly competitive performance obtained in those tasks as compared to state-of-the-art methods, we show a significant improvement in speed and storage requirement using DeepRank-GNN as compared to DeepRank. AVAILABILITY AND IMPLEMENTATION: DeepRank-GNN is freely available from https://github.com/DeepRank/DeepRank-GNN. SUPPLEMENTARY INFORMATION: Supplementary data are available at Bioinformatics online.

Information-Driven Antibody-Antigen Modelling with HADDOCK.

In the recent years, therapeutic use of antibodies has seen a huge growth, "due to their inherent proprieties and technological advances in the methods used to study and characterize them. Effective design and engineering of antibodies for therapeutic purposes are heavily dependent on knowledge of the structural principles that regulate antibody-antigen interactions. Several experimental techniques such as X-ray crystallography, cryo-electron microscopy, NMR, or mutagenesis analysis can be applied, but these are usually expensive and time-consuming. Therefore computational approaches like molecular docking may offer a valuable alternative for the characterization of antibody-antigen complexes.Here we describe a protocol for the prediction of the 3D structure of antibody-antigen complexes using the integrative modelling platform HADDOCK. The protocol consists of (1) the identification of the antibody residues belonging to the hypervariable loops which are known to be crucial for the binding and can be used to guide the docking and (2) the detailed steps to perform docking with the HADDOCK 2.4 webserver following different strategies depending on the availability of information about epitope residues.

PPARγ lipodystrophy mutants reveal intermolecular interactions required for enhancer activation.

Peroxisome proliferator-activated receptor γ (PPARγ) is the master regulator of adipocyte differentiation, and mutations that interfere with its function cause lipodystrophy. PPARγ is a highly modular protein, and structural studies indicate that PPARγ domains engage in several intra- and inter-molecular interactions. How these interactions modulate PPARγ's ability to activate target genes in a cellular context is currently poorly understood. Here we take advantage of two previously uncharacterized lipodystrophy mutations, R212Q and E379K, that are predicted to interfere with the interaction of the hinge of PPARγ with DNA and with the interaction of PPARγ ligand binding domain (LBD) with the DNA-binding domain (DBD) of the retinoid X receptor, respectively. Using biochemical and genome-wide approaches we show that these mutations impair PPARγ function on an overlapping subset of target enhancers. The hinge region-DNA interaction appears mostly important for binding and remodelling of target enhancers in inaccessible chromatin, whereas the PPARγ-LBD:RXR-DBD interface stabilizes the PPARγ:RXR:DNA ternary complex. Our data demonstrate how in-depth analyses of lipodystrophy mutants can unravel molecular mechanisms of PPARγ function.

Teixobactin kills bacteria by a two-pronged attack on the cell envelope.

Antibiotics that use novel mechanisms are needed to combat antimicrobial resistance1-3. Teixobactin4 represents a new class of antibiotics with a unique chemical scaffold and lack of detectable resistance. Teixobactin targets lipid II, a precursor of peptidoglycan5. Here we unravel the mechanism of teixobactin at the atomic level using a combination of solid-state NMR, microscopy, in vivo assays and molecular dynamics simulations. The unique enduracididine C-terminal headgroup of teixobactin specifically binds to the pyrophosphate-sugar moiety of lipid II, whereas the N terminus coordinates the pyrophosphate of another lipid II molecule. This configuration favours the formation of a ß-sheet of teixobactins bound to the target, creating a supramolecular fibrillar structure. Specific binding to the conserved pyrophosphate-sugar moiety accounts for the lack of resistance to teixobactin4. The supramolecular structure compromises membrane integrity. Atomic force microscopy and molecular dynamics simulations show that the supramolecular structure displaces phospholipids, thinning the membrane. The long hydrophobic tails of lipid II concentrated within the supramolecular structure apparently contribute to membrane disruption. Teixobactin hijacks lipid II to help destroy the membrane. Known membrane-acting antibiotics also damage human cells, producing undesirable side effects. Teixobactin damages only membranes that contain lipid II, which is absent in eukaryotes, elegantly resolving the toxicity problem. The two-pronged action against cell wall synthesis and cytoplasmic membrane produces a highly effective compound targeting the bacterial cell envelope. Structural knowledge of the mechanism of teixobactin will enable the rational design of improved drug candidates.

Cyclization and Docking Protocol for Cyclic Peptide-Protein Modeling Using HADDOCK2.4.

An emerging class of therapeutic molecules are cyclic peptides with over 40 cyclic peptide drugs currently in clinical use. Their mode of action is, however, not fully understood, impeding rational drug design. Computational techniques could positively impact their design, but modeling them and their interactions remains challenging due to their cyclic nature and their flexibility. This study presents a step-by-step protocol for generating cyclic peptide conformations and docking them to their protein target using HADDOCK2.4. A dataset of 30 cyclic peptide-protein complexes was used to optimize both cyclization and docking protocols. It supports peptides cyclized via an N- and C-terminus peptide bond and/or a disulfide bond. An ensemble of cyclic peptide conformations is then used in HADDOCK to dock them onto their target protein using knowledge of the binding site on the protein side to drive the modeling. The presented protocol predicts at least one acceptable model according to the critical assessment of prediction of interaction criteria for each complex of the dataset when the top 10 HADDOCK-ranked single structures are considered (100% success rate top 10) both in the bound and unbound docking scenarios. Moreover, its performance in both bound and fully unbound docking is similar to the state-of-the-art software in the field, Autodock CrankPep. The presented cyclization and docking protocol should make HADDOCK a valuable tool for rational cyclic peptide-based drug design and high-throughput screening.

Pathogen-sugar interactions revealed by universal saturation transfer analysis.

Many pathogens exploit host cell-surface glycans. However, precise analyses of glycan ligands binding with heavily modified pathogen proteins can be confounded by overlapping sugar signals and/or compounded with known experimental constraints. Universal saturation transfer analysis (uSTA) builds on existing nuclear magnetic resonance spectroscopy to provide an automated workflow for quantitating protein-ligand interactions. uSTA reveals that early-pandemic, B-origin-lineage severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) spike trimer binds sialoside sugars in an "end-on" manner. uSTA-guided modeling and a high-resolution cryo-electron microscopy structure implicate the spike N-terminal domain (NTD) and confirm end-on binding. This finding rationalizes the effect of NTD mutations that abolish sugar binding in SARS-CoV-2 variants of concern. Together with genetic variance analyses in early pandemic patient cohorts, this binding implicates a sialylated polylactosamine motif found on tetraantennary N-linked glycoproteins deep in the human lung as potentially relevant to virulence and/or zoonosis.

Molecular Insights Into Binding and Activation of the Human KCNQ2 Channel by Retigabine.

Voltage-gated potassium channels of the Kv7.x family are involved in a plethora of biological processes across many tissues in animals, and their misfunctioning could lead to several pathologies ranging from diseases caused by neuronal hyperexcitability, such as epilepsy, or traumatic injuries and painful diabetic neuropathy to autoimmune disorders. Among the members of this family, the Kv7.2 channel can form hetero-tetramers together with Kv7.3, forming the so-called M-channels, which are primary regulators of intrinsic electrical properties of neurons and of their responsiveness to synaptic inputs. Here, prompted by the similarity between the M-current and that in Kv7.2 alone, we perform a computational-based characterization of this channel in its different conformational states and in complex with the modulator retigabine. After validation of the structural models of the channel by comparison with experimental data, we investigate the effect of retigabine binding on the two extreme states of Kv7.2 (resting-closed and activated-open). Our results suggest that binding, so far structurally characterized only in the intermediate activated-closed state, is possible also in the other two functional states. Moreover, we show that some effects of this binding, such as increased flexibility of voltage sensing domains and propensity of the pore for open conformations, are virtually independent on the conformational state of the protein. Overall, our results provide new structural and dynamic insights into the functioning and the modulation of Kv7.2 and related channels.

Interface refinement of low- to medium-resolution Cryo-EM complexes using HADDOCK2.4.

A wide range of cellular processes requires the formation of multimeric protein complexes. The rise of cryo-electron microscopy (cryo-EM) has enabled the structural characterization of these protein assemblies. The density maps produced can, however, still suffer from limited resolution, impeding the process of resolving structures at atomic resolution. In order to solve this issue, monomers can be fitted into low- to medium-resolution maps. Unfortunately, the models produced frequently contain atomic clashes at the protein-protein interfaces (PPIs), as intermolecular interactions are typically not considered during monomer fitting. Here, we present a refinement approach based on HADDOCK2.4 to remove intermolecular clashes and optimize PPIs. A dataset of 14 cryo-EM complexes was used to test eight protocols. The best-performing protocol, consisting of a semi-flexible simulated annealing refinement with centroid restraints on the monomers, was able to decrease intermolecular atomic clashes by 98% without significantly deteriorating the quality of the cryo-EM density fit.

DeepRank: a deep learning framework for data mining 3D protein-protein interfaces.

Three-dimensional (3D) structures of protein complexes provide fundamental information to decipher biological processes at the molecular scale. The vast amount of experimentally and computationally resolved protein-protein interfaces (PPIs) offers the possibility of training deep learning models to aid the predictions of their biological relevance. We present here DeepRank, a general, configurable deep learning framework for data mining PPIs using 3D convolutional neural networks (CNNs). DeepRank maps features of PPIs onto 3D grids and trains a user-specified CNN on these 3D grids. DeepRank allows for efficient training of 3D CNNs with data sets containing millions of PPIs and supports both classification and regression. We demonstrate the performance of DeepRank on two distinct challenges: The classification of biological versus crystallographic PPIs, and the ranking of docking models. For both problems DeepRank is competitive with, or outperforms, state-of-the-art methods, demonstrating the versatility of the framework for research in structural biology.

Emergence and spread of SARS-CoV-2 lineage B.1.620 with variant of concern-like mutations and deletions.

Distinct SARS-CoV-2 lineages, discovered through various genomic surveillance initiatives, have emerged during the pandemic following unprecedented reductions in worldwide human mobility. We here describe a SARS-CoV-2 lineage - designated B.1.620 - discovered in Lithuania and carrying many mutations and deletions in the spike protein shared with widespread variants of concern (VOCs), including E484K, S477N and deletions HV69Δ, Y144Δ, and LLA241/243Δ. As well as documenting the suite of mutations this lineage carries, we also describe its potential to be resistant to neutralising antibodies, accompanying travel histories for a subset of European cases, evidence of local B.1.620 transmission in Europe with a focus on Lithuania, and significance of its prevalence in Central Africa owing to recent genome sequencing efforts there. We make a case for its likely Central African origin using advanced phylogeographic inference methodologies incorporating recorded travel histories of infected travellers.

Prediction of protein assemblies, the next frontier: The CASP14-CAPRI experiment.

We present the results for CAPRI Round 50, the fourth joint CASP-CAPRI protein assembly prediction challenge. The Round comprised a total of twelve targets, including six dimers, three trimers, and three higher-order oligomers. Four of these were easy targets, for which good structural templates were available either for the full assembly, or for the main interfaces (of the higher-order oligomers). Eight were difficult targets for which only distantly related templates were found for the individual subunits. Twenty-five CAPRI groups including eight automatic servers submitted ~1250 models per target. Twenty groups including six servers participated in the CAPRI scoring challenge submitted ~190 models per target. The accuracy of the predicted models was evaluated using the classical CAPRI criteria. The prediction performance was measured by a weighted scoring scheme that takes into account the number of models of acceptable quality or higher submitted by each group as part of their five top-ranking models. Compared to the previous CASP-CAPRI challenge, top performing groups submitted such models for a larger fraction (70-75%) of the targets in this Round, but fewer of these models were of high accuracy. Scorer groups achieved stronger performance with more groups submitting correct models for 70-80% of the targets or achieving high accuracy predictions. Servers performed less well in general, except for the MDOCKPP and LZERD servers, who performed on par with human groups. In addition to these results, major advances in methodology are discussed, providing an informative overview of where the prediction of protein assemblies currently stands.