Predicting and Experimentally Validating Hot-Spot Residues at Protein-Protein Interfaces

Abstract

Here we describe a comparative analysis of multiple CAS methods, which highlights effective approaches to improve the accuracy of predicting hot-spot residues. Alongside this, we introduce a new method, BUDE Alanine Scanning, which can be applied to single structures from crystallography, and to structural ensembles from NMR or molecular dynamics data. The comparative analyses facilitate accurate prediction of hot-spots that we validate experimentally with three diverse targets: NOXA-B/MCL-1 (an α helix-mediated PPI), SIMS/SUMO and GKAP/SHANK-PDZ (both β strand-mediated interactions). Finally, the approach is applied to the accurate prediction of hot-residues at a topographically novel Affimer/BCL-xL protein-protein interface.

Advances in Molecular Dynamics Simulations and Enhanced Sampling Methods for the Study of Protein Systems

“…Ibarra et al in introducing the BudeAlaScan method has also simultaneously compared the performance of several CAS methods, including flex ddG, in accurately predicting ‘hotspot’ residues. The comparable results achieved for flex ddG and BudeAlaScan emphasized the importance of considering conformational heterogeneity to obtain a more precise prediction of ΔΔ G [ 74 ]. This comparative study also observed quality ΔΔ G prediction from mutation Cutoff Scanning Matrix (mCSM), a machine learning approach that was trained based on several curated databases documenting changes in protein stability and protein-protein affinity upon mutation, such as SKEMPI and ProTherm [ 111 , 112 , 113 , 114 ].…”

“…The prediction of residues crucial for the formation of favorable protein-ligand or protein-protein interactions is an asset in the field of rational drug design. These decisive sets of amino acids are called ‘hotspots’ and numerous studies with the objective of quantifying the contribution of residues towards binding affinity have been conducted to facilitate precise identification of ‘hotspot’ residues [ 73 , 74 , 75 , 76 , 77 ]. Experimentally, the determination of ‘hotspots’ proceeds through traditional methods such as alanine scanning mutagenesis, structure-activity relationship by NMR and multiple solvent crystal structures (MSCS) [ 78 , 79 , 80 ].…”

“…This method involved the site-directed mutagenesis of specific residues to alanine to assess the contribution of the amino acid towards protein stability and function [ 103 , 104 , 105 ]. Computational alanine scanning (CAS), a robust alternative to its experimental counterpart, permits the rapid scanning of protein surface for amino acid clusters vital for protein-protein association [ 74 ]. Over the years, several CAS methods have been developed, centering around free energy calculations performed on single structures or structure ensembles from NMR or MD trajectories to predict ‘hotspots’ at protein-protein interface [ 106 , 107 ].…”

“…Another recently established CAS tool is the BUDE Alanine Scanning (BudeAlaScan) method, which was built upon the empirical free energy approach of a molecular docking algorithm, BUDE (Bristol University Docking Engine) [ 74 , 110 ]. Similar to flex ddG, this method considers conformational plasticity by considering a set of side chain rotamers for charged residues (Asp, Glu, Arg, Lys and His) to approximate the entropy loss resulting from the formation of inter-protein salt bridges between the side chains of the aforementioned charged residues and a fixed protein backbone.…”

Computational Hot-Spot Analysis of the SARS-CoV-2 Receptor Binding Domain / ACE2 Complex

“…We next used BAlaS 14,15 , a computational alanine scanning mutagenesis server, to predict the change in free energy (ΔΔG) associated with mutating interface residues to alanine. Our analysis focused on residues within 13 Å from the binding interface (set as default) and residues plotted in Figure 3 are limited to those with predicted ΔΔG > 4 kJ/mol.…”