Biology subjects

Murail, S.

Publications and source records attributed to Murail, S..

5 recordsLinked to original sources

bioRxiv2026

Deep-learning predictions of biomolecular structures : persistent limitations and new horizons extended by explicit ion addition

The advent of deep learning-driven tools such as AlphaFold has revolutionized the prediction of biomolecular structures, offering unprecedented accuracy and accessibility for proteins, RNA, and their complexes. While these tools have demonstrated remarkable success in benchmarking competitions and enabled experimentalists to generate models with ease, their widespread use has also highlighted persistent challenges. These include difficulties in assessing model confidence, limitations in predicting transmembrane domains, nucleic acids, conformational diversity, and interactions with ions or ligands, as well as the tendency to misfold intrinsically disordered regions (IDRs). In this perspective, we critically evaluate the strengths and limitations of current AI-based structure prediction tools through illustrative examples with a particular emphasis on the impact of explicit ion modelling. We notably report how the explicit addition of a few potassium cations to the prediction of IDRs or G-quadruplexes can trigger massive conformational switches compared to "dry" predictions. On this basis, we suggest modelling sequences both "dry" and in the presence of explicit potassium cations as a simple, practical way to sample alternative conformations and to expose disordered regions that current predictors tend to over-fold. We discuss the importance of reporting confidence metrics in publications to avoid overinterpretation. Furthermore, we address the unique challenges of RNA structure prediction, where data scarcity and structural complexity limit the performance of both classical and deep learning methods. Our analysis underscores the need for continued methodological advancements, integration of complementary computational tools, and expansion of high-quality experimental datasets. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=111 SRC="FIGDIR/small/740587v1_ufig1.gif" ALT="Figure 1"> View larger version (16K): org.highwire.dtl.DTLVardef@e950e9org.highwire.dtl.DTLVardef@1bf22b3org.highwire.dtl.DTLVardef@17f5229org.highwire.dtl.DTLVardef@1eb2e9f_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2025

Robust conformational space exploration of cyclic peptides by combining different MD protocols and force fields

Cyclic peptides are an important class of pharmaceutical drugs. We used here replica-exchange molecular dynamics (REMD) and simulated tempering (ST) simulations to explore the conformational landscape of a set of nine cyclic peptides. The N-ter to C-ter backbone cyclized peptides of seven to ten residues were previously designed for a high conformational stability with a mixture of L and D amino acids. Their experimental NMR structures were available in the protein data bank (PDB). For each peptide we tested several force fields, Amber96, Amber14, RSFF2C and Charmm36m in implicit and explicit solvent. We find that the variability of the free energy maps obtained from several protocols is larger than the variability obtained by just repeating the same protocol. Running multiple protocols is therefore important for the convergence assessment of REMD or ST simulations. The majority of the free energy maps showed clusters with a high RMSD to the native structures, revealing the residual flexibility of this set of cyclic peptides. The high RMSD clusters had in some cases the lowest free energy, rendering the prediction of the native structure more difficult with a single protocol. Fortunately, the combination of four implicit solvent REMD and ST simulations, mixing the Amber96 and Am-ber14 force fields, predicted robustly the native structure. As implicit solvent in the REMD or ST setup are up to one hundred times faster than explicit solvent simulations, so running four implicit solvent simulations is a good practical choice. We checked that the use of an explicit solvent REMD or ST simulation, taken alone or combined with implicit solvent simulations, did not significantly improved our results. It results that our combination of four implicit solvent simulations is tied in terms of success rate with much more expensive combinations that include explicit solvent simulations. This may be used as guideline for further studies of cyclic peptide conformations.

biophysics↗

bioRxiv2024

Simulated Solute Tempering 2

Molecular dynamics (MD) simulations are powerful tools for studying the movement and interactions of molecules, but they can be computationally expensive, especially for large biomolecules like proteins. This is problematic because accurately simulating the motions of these molecules is key to understanding their function. Enhanced sampling methods, such as Simulated Tempering (ST), temperature Replica Exchange Molecular Dynamics (REMD) and Replica Exchange with Solute Tempering (REST and REST2), have been developed to overcome this challenge by improving the efficiency of MD simulations. This article presents a new enhanced sampling method called Simulated Solute Tempering 2 (SST2) that builds upon the strengths of ST and REST2. SST2 selectively scales the interactions inside a biomolecule and with its surrounding environment, effectively accelerating the exploration of its different structural states and their stability at various temperatures. SST2 was tested on three different systems (chignolin CLN025, Trp-Cage, and a protein-peptide complex, p97/PNGase) and found to achieve comparable or superior sampling efficiency to ST, SST1 and REST2 while requiring fewer temperature rungs. Notably, SST2 is particularly well-suited for investigating large biomolecular systems, making it a valuable tool for studying a wide range of biomolecular processes, from protein folding to ligand binding. Significance StatementAccurately simulating the motions of biological molecules like proteins is key to understanding their function. However, this is computationally expensive, especially for large systems or slow processes. To overcome this, we developed Simulated Solute Tempering 2 (SST2), a novel algorithm that improves the efficiency of molecular dynamics (MD) simulations. SST2 selectively scales the interactions inside a biomolecule and with its surrounding environment, accelerating the exploration of its different structures and their stability at various temperatures. This allows SST2 to achieve remarkable performances with a clear improved ease of use compare to similar methods, particularly for large systems like protein-ligand complexes. This makes SST2 a valuable tool for studying a wide range of biomolecular processes, from protein folding to ligand binding. TOC Graphic O_FIG O_LINKSMALLFIG WIDTH=200 HEIGHT=111 SRC="FIGDIR/small/613476v3_ufig1.gif" ALT="Figure 1"> View larger version (41K): org.highwire.dtl.DTLVardef@761755org.highwire.dtl.DTLVardef@f5df1forg.highwire.dtl.DTLVardef@ba93dorg.highwire.dtl.DTLVardef@193e7eb_HPS_FORMAT_FIGEXP M_FIG C_FIG

biophysics↗

bioRxiv2022

A novel computational method for head-to-tail peptide cyclization: application to urotensin II

Backbone head-to-tail cyclization is one effective strategy to stabilize the conformation of bioactive peptides, preventing enzymatic degradation and improving their bioavailibility. However, very little is known about the requirements to rationally design linkers for the cyclization of linear peptides. Recently, we have shown that large scale data-mining of protein structures can lead to the precise identification of protein loop conformations. Here, we transpose this approach to head-to-tail peptide cyclization. We first show that given a linker sequence and the conformation of the linear peptide, it is possible to accurately predict the cyclized peptide conformation improving by over 1 [A] over pre-existing protocols. Secondly, and more importantly, we show that it is possible to elaborate on the information inferred from protein structures to propose effective candidate linker sequences constrained by length and amino acid composition. As experimental validation, we first apply our approach to design linkers for the head-to-tail cyclizations of a peptide derived from Nrf2. The designed cyclized peptide shows a 26-fold increase in binding affinity. We then consider urotensin II, a cyclic peptide already stabilized by a disulfide bond, that exerts a broad array of biological activities. The designed head-to-tail cyclized peptide, the first synthesized bicyclic 14-residue long urotensin II analogue shows an excellent retention of in vitro activity. Overall, we propose the first framework for the rational peptide head-to-tail cyclization and reveal its potential for cyclic peptide-based drug design.

bioinformatics↗

bioRxiv2019

Glutathionylation primes soluble GAPDH for late collapse into insoluble aggregates

Protein aggregation is a complex physiological process, primarily determined by stress-related factors revealing the hidden aggregation propensity of proteins that otherwise are fully soluble. Here we report a mechanism by which glycolytic glyceraldehyde-3-phosphate dehydrogenase of Arabidopsis thaliana (AtGAPC1) is primed to form insoluble aggregates by the glutathionylation of its catalytic cysteine (Cys149). Following a lag phase, glutathionylated AtGAPC1 initiates a self-aggregation process resulting in the formation of branched chains of globular particles made of partially misfolded and totally inactive proteins. GSH molecules within AtGAPC1 active sites are suggested to provide the initial destabilizing signal. The following removal of glutathione by the formation of an alternative disulfide bond between Cys149 and Cys153 reinforces the aggregation process. Besides acting as a protective mechanism against overoxidation, S-glutathionylation of AtGAPC1 triggers an unexpected aggregation pathway with completely different and still unexplored physiological implications.

plant biology↗