bims-strubi 2021-10-17 papers

J Biomol Struct Dyn. 2021 Oct 12. 1-10

Fragmented blind docking: a novel protein-ligand binding prediction protocol.

Gianvito Grassoa, Arianna Di Gregorio, Bojan Mavkov, Dario Piga, Giuseppe Falvo D'Urso Labate, Andrea Danani, Marco A Deriu.

In the present paper we propose a novel blind docking protocol based on Autodock-Vina. The developed docking protocol can provide binding site identification and binding pose prediction at the same time, by a systematical exploration of the protein volume performed with several preliminary docking calculations. In our opinion, this protocol can be successfully applied during the first steps of the virtual screening pipeline, because it provides binding site identification and binding pose prediction at the same time without visual evaluation of the binding site. After the binding pose prediction, MM/GBSA re-scoring rescoring procedures has been applied to improve the accuracy of the protein-ligand bound state. The FRAD protocol has been tested on 116 protein-ligand complexes of the Heat Shock Protein 90 - alpha, on 176 of Human Immunodeficiency virus protease 1, and on more than 100 protein-ligand system taken from the PDBbind dataset. Overall, the FRAD approach combined to MM/GBSA re-scoring can be considered as a powerful tool to increase the accuracy and efficiency with respect to other standard docking approaches when the ligand-binding site is unknown.Communicated by Ramaswamy H. Sarma.

Keywords: Blind docking; molecular docking; protein–ligand interactions; virtual screening

DOI: https://doi.org/10.1080/07391102.2021.1988709

Biophys J. 2021 Oct 07. pii: S0006-3495(21)00828-6. [Epub ahead of print]

Refining conformational ensembles of flexible proteins against small-angle X-ray scattering data.

Francesco Pesce, Kresten Lindorff-Larsen.

Intrinsically disordered proteins and flexible regions in multi-domain proteins display substantial conformational heterogeneity. Characterizing the conformational ensembles of these proteins in solution typically requires combining one or more biophysical techniques with computational modelling or simulations. Experimental data can either be used to assess the accuracy of a computational model or to refine the computational model to get a better agreement with the experimental data. In both cases, one generally needs a so-called forward model, i.e. an algorithm to calculate experimental observables from individual conformations or ensembles. In many cases, this involve one or more parameters that need to be set, and it is not always trivial to determine the optimal values or to understand the impact on the choice of parameters. For example, in the case of small-angle X-ray scattering (SAXS) experiments, many forward models include parameters that describe the contribution of the hydration layer and displaced solvent to the background-subtracted experimental data. Often, one also needs to fit a scale factor and a constant background for the SAXS data, but across the entire ensemble. Here, we present a protocol to dissect the effect of free-parameters on the calculated SAXS intensities, and to identify a reliable set of values. We have implemented this procedure in our Bayesian/Maximum Entropy framework for ensemble refinement, and demonstrate the results on four intrinsically disordered proteins and a three-domain protein connected by flexible linkers. Our results show that the resulting ensembles can depend on the parameters used for solvent effects, and suggests that these should be chosen carefully. We also find a set of parameters that work robustly across all proteins. SIGNIFICANCE The flexibility of a protein is often key to its biological function, yet understanding and characterizing its conformational heterogeneity is difficult. We here describe a robust protocol for combining small-angle X-ray scattering experiments with computational modelling to obtain a conformational ensemble. In particular, we focus on the contribution of protein hydration to the experiments and how this is included in modelling the data. Our resulting algorithm and software should make modelling intrinsically disordered proteins and multi-domain proteins more robust, thus aiding in understanding the relationship between protein dynamics and biological function.

DOI: https://doi.org/10.1016/j.bpj.2021.10.003

J Chem Inf Model. 2021 Oct 15.

Improving Structure-Based Virtual Screening with Ensemble Docking and Machine Learning.

Joel Ricci-Lopez, Sergio A Aguila, Michael K Gilson, Carlos A Brizuela.

One of the main challenges of structure-based virtual screening (SBVS) is the incorporation of the receptor's flexibility, as its explicit representation in every docking run implies a high computational cost. Therefore, a common alternative to include the receptor's flexibility is the approach known as ensemble docking. Ensemble docking consists of using a set of receptor conformations and performing the docking assays over each of them. However, there is still no agreement on how to combine the ensemble docking results to obtain the final ligand ranking. A common choice is to use consensus strategies to aggregate the ensemble docking scores, but these strategies exhibit slight improvement regarding the single-structure approach. Here, we claim that using machine learning (ML) methodologies over the ensemble docking results could improve the predictive power of SBVS. To test this hypothesis, four proteins were selected as study cases: CDK2, FXa, EGFR, and HSP90. Protein conformational ensembles were built from crystallographic structures, whereas the evaluated compound library comprised up to three benchmarking data sets (DUD, DEKOIS 2.0, and CSAR-2012) and cocrystallized molecules. Ensemble docking results were processed through 30 repetitions of 4-fold cross-validation to train and validate two ML classifiers: logistic regression and gradient boosting trees. Our results indicate that the ML classifiers significantly outperform traditional consensus strategies and even the best performance case achieved with single-structure docking. We provide statistical evidence that supports the effectiveness of ML to improve the ensemble docking performance.

DOI: https://doi.org/10.1021/acs.jcim.1c00511

Nature. 2021 Oct 13.

Structure of Venezuelan equine encephalitis virus with its receptor LDLRAD3.

Bingting Ma, Cuiqing Huang, Jun Ma, Ye Xiang, Xinzheng Zhang.

Venezuelan equine encephalitis virus (VEEV) is an enveloped RNA virus that causes encephalitis and potentially mortality in infected humans and equines1. At present, no vaccines or drugs are available that prevent or cure diseases caused by VEEV. Low-density lipoprotein receptor class A domain-containing 3 (LDLRAD3) was recently identified as a receptor for the entry of VEEV into host cells2. Here we present the cryo-electron microscopy structure of the LDLRAD3 extracellular domain 1 (LDLRAD3-D1) in complex with VEEV virus-like particles at a resolution of 3.0 Å. LDLRAD3-D1 has a cork-like structure and is inserted into clefts formed between adjacent VEEV E2-E1 heterodimers in the viral-surface trimer spikes through hydrophobic and polar contacts. Mutagenesis studies of LDLRAD3-D1 identified residues that are involved in the key interactions with VEEV. Of note, some of the LDLRAD3-D1 mutants showed a significantly increased binding affinity for VEEV, suggesting that LDLRAD3-D1 may serve as a potential scaffold for the development of inhibitors of VEEV entry. Our structures provide insights into alphavirus assembly and the binding of receptors to alphaviruses, which may guide the development of therapeutic countermeasures against alphaviruses.

DOI: https://doi.org/10.1038/s41586-021-03909-1

EMBO Rep. 2021 Oct 14. e52981

The human GID complex engages two independent modules for substrate recruitment.

Weaam I Mohamed, Sophia L Park, Julius Rabl, Alexander Leitner, Daniel Boehringer, Matthias Peter.

The human GID (hGID) complex is a conserved E3 ubiquitin ligase regulating diverse biological processes, including glucose metabolism and cell cycle progression. However, the biochemical function and substrate recognition of the multi-subunit complex remain poorly understood. Using biochemical assays, cross-linking mass spectrometry, and cryo-electron microscopy, we show that hGID engages two distinct modules for substrate recruitment, dependent on either WDR26 or GID4. WDR26 and RanBP9 cooperate to ubiquitinate HBP1 in vitro, while GID4 is dispensable for this reaction. In contrast, GID4 functions as an adaptor for the substrate ZMYND19, which surprisingly lacks a Pro/N-end degron. GID4 substrate binding and ligase activity is regulated by ARMC8α, while the shorter ARMC8β isoform assembles into a stable hGID complex that is unable to recruit GID4. Cryo-EM reconstructions of these hGID complexes reveal the localization of WDR26 within a ring-like, tetrameric architecture and suggest that GID4 and WDR26/Gid7 utilize different, non-overlapping binding sites. Together, these data advance our mechanistic understanding of how the hGID complex recruits cognate substrates and provides insights into the regulation of its E3 ligase activity.

Keywords: E3 ligase; hGID/CTLH; oligomerization; substrate receptors; ubiquitin

DOI: https://doi.org/10.15252/embr.202152981

Protein Sci. 2021 Oct 14.

Match_Motif: A rapid computational tool to assist in protein-protein interaction design.

Martin Zacharias.

In order to generate protein assemblies with a desired function the rational design of protein-protein binding interfaces is of significant interest. Approaches based on random mutagenesis or directed evolution may involve complex experimental selection procedures. Also, molecular modeling approaches to design entirely new proteins and interactions with partner molecules can involve large computational efforts and screening steps. In order to simplify at least the initial effort for designing a putative binding interface between two proteins the Match_Motif approach has been developed. It employs the large collection of known protein-protein complex structures to suggest interface modifications that may lead to improved binding for a desired input interaction geometry. The approach extracts interaction motifs based on the backbone structure of short (4 residue) segments and the relative arrangement with respect to short segments on the partner protein. The interaction geometry is used to search through a data base of such motifs in known stable bound complexes. All matches are rapidly identified (within a few seconds) and collected and can be used to guide changes in the interface that may lead to improved binding. In the output an alternative interface structure is also proposed based on the frequency of occurrence of side chains at a given interface position in all matches and based on sterical considerations. Applications of the procedure to known complex structures and alternative arrangements are presented and discussed. The program, data files and example applications can be downloaded from: https://www.groups.ph.tum.de/t38/downloads/. This article is protected by copyright. All rights reserved.

Keywords: binding interaction motifs; interaction design; protein design; protein-protein binding; protein-protein interface design

DOI: https://doi.org/10.1002/pro.4208

Curr Opin Struct Biol. 2021 Oct 11. pii: S0959-440X(21)00135-4. [Epub ahead of print]72 114-126

Toward better drug discovery with knowledge graph.

Xiangxiang Zeng, Xinqi Tu, Yuansheng Liu, Xiangzheng Fu, Yansen Su.

Drug discovery is the process of new drug identification. This process is driven by the increasing data from existing chemical libraries and data banks. The knowledge graph is introduced to the domain of drug discovery for imposing an explicit structure to integrate heterogeneous biomedical data. The graph can provide structured relations among multiple entities and unstructured semantic relations associated with entities. In this review, we summarize knowledge graph-based works that implement drug repurposing and adverse drug reaction prediction for drug discovery. As knowledge representation learning is a common way to explore knowledge graphs for prediction problems, we introduce several representative embedding models to provide a comprehensive understanding of knowledge representation learning.

DOI: https://doi.org/10.1016/j.sbi.2021.09.003