A Perspective on the Prospective Use of AI in Protein Structure Prediction.

AlphaFold2 (AF2) and RoseTTaFold (RF) have revolutionized structural biology, serving as highly reliable and effective methods for predicting protein structures. This article explores their impact and limitations, focusing on their integration into experimental pipelines and their application in diverse protein classes, including membrane proteins, intrinsically disordered proteins (IDPs), and oligomers. In experimental pipelines, AF2 models help X-ray crystallography in resolving the phase problem, while complementarity with mass spectrometry and NMR data enhances structure determination and protein flexibility prediction. Predicting the structure of membrane proteins remains challenging for both AF2 and RF due to difficulties in capturing conformational ensembles and interactions with the membrane. Improvements in incorporating membrane-specific features and predicting the structural effect of mutations are crucial. For intrinsically disordered proteins, AF2's confidence score (pLDDT) serves as a competitive disorder predictor, but integrative approaches including molecular dynamics (MD) simulations or hydrophobic cluster analyses are advocated for accurate dynamics representation. AF2 and RF show promising results for oligomeric models, outperforming traditional docking methods, with AlphaFold-Multimer showing improved performance. However, some caveats remain in particular for membrane proteins. Real-life examples demonstrate AF2's predictive capabilities in unknown protein structures, but models should be evaluated for their agreement with experimental data. Furthermore, AF2 models can be used complementarily with MD simulations. In this Perspective, we propose a "wish list" for improving deep-learning-based protein folding prediction models, including using experimental data as constraints and modifying models with binding partners or post-translational modifications. Additionally, a meta-tool for ranking and suggesting composite models is suggested, driving future advancements in this rapidly evolving field.

Antioxidant and Membrane Binding Properties of Serotonin Protect Lipids from Oxidation.

Serotonin (5-hydroxytryptamine, 5-HT) is a well-known neurotransmitter that is involved in a growing number of functions in peripheral tissues. Recent studies have shown nonpharmacological functions of 5-HT linked to its chemical properties. Indeed, it was reported that 5-HT may, on the one hand, bind lipid membranes and, on the other hand, protect red blood cells through a mechanism independent of its specific receptors. To better understand these underevaluated properties of 5-HT, we combined biochemical, biophysical, and molecular dynamics simulations approaches to characterize, at the molecular level, the antioxidant capacity of 5-HT and its interaction with lipid membranes. To do so, 5-HT was added to red blood cells and lipid membranes bearing different degrees of unsaturation. Our results demonstrate that 5-HT acts as a potent antioxidant and binds with a superior affinity to lipids with unsaturation on both alkyl chains. We show that 5-HT locates at the hydrophobic-hydrophilic interface, below the glycerol group. This interfacial location is stabilized by hydrogen bonds between the 5-HT hydroxyl group and lipid headgroups and allows 5-HT to intercept reactive oxygen species, preventing membrane oxidation. Experimental and molecular dynamics simulations using membrane enriched with oxidized lipids converge to further reveal that 5-HT contributes to the termination of lipid peroxidation by direct interaction with active groups of these lipids and could also contribute to limit the production of new radicals. Taken together, our results identify 5-HT as a potent inhibitor of lipid peroxidation and offer a different perspective on the role of this pleiotropic molecule.

Recent advances on polyproline II.

About half of the globular proteins are composed of regular secondary structures, α-helices, and ß-sheets, while the rest are constituted of irregular secondary structures, such as turns or coil conformations. Other regular secondary structures are often ignored, despite their importance in biological processes. Among such structures, the polyproline II helix (PPII) has interesting behaviours. PPIIs are not usually associated with conventional stabilizing interactions, and recent studies have observed that PPIIs are more frequent than anticipated. In addition, it is suggested that they may have an important functional role, particularly in protein-protein or protein-nucleic acid interactions and recognition. Residues associated with PPII conformations represent nearly 5% of the total residues, but the lack of PPII assignment approaches prevents their systematic analysis. This short review will present current knowledge and recent research in PPII area. In a first step, the different methodologies able to assign PPII are presented. In the second step, recent studies that have shown new perspectives in PPII analysis in terms of structure and function are underlined with three cases: (1) PPII in protein structures. For instance, the first crystal structure of an oligoproline adopting an all-trans polyproline II (PPII) helix had been presented; (2) the involvement of PPII in different diseases and drug designs; and (3) an interesting extension of PPII study in the protein dynamics. For instance, PPIIs are often linked to disorder region analysis and the precise analysis of a potential PPII helix in hypogonadism shows unanticipated PPII formations in the patient mutation, while it is not observed in the wild-type form of KISSR1 protein.

Small Oligomers of Aß42 Protein in the Bulk Solution with AlphaFold2.

Aggregation of amyloid-ß (Aß42) protein is one hallmark of Alzheimer's disease, and the conformations of the smallest Aß42 oligomers are largely unknown. Here, we explore the application of the deep learning AlphaFold2 method to the structure determination of Aß42 monomers up to hexamers. The results shed light on the early Aß42 aggregation steps in the bulk solution.

Discrete analyses of protein dynamics.

Protein structures are highly dynamic macromolecules. This dynamics is often analysed through experimental and/or computational methods only for an isolated or a limited number of proteins. Here, we explore large-scale protein dynamics simulation to observe dynamics of local protein conformations using different perspectives. We analysed molecular dynamics to investigate protein flexibility locally, using classical approaches such as RMSf, solvent accessibility, but also innovative approaches such as local entropy. First, we focussed on classical secondary structures and analysed specifically how ß-strand, ß-turns, and bends evolve during molecular simulations. We underlined interesting specific bias between ß-turns and bends, which are considered as the same category, while their dynamics show differences. Second, we used a structural alphabet that is able to approximate every part of the protein structures conformations, namely protein blocks (PBs) to analyse (i) how each initial local protein conformations evolve during dynamics and (ii) if some exchange can exist among these PBs. Interestingly, the results are largely complex than simple regular/rigid and coil/flexible exchange. AbbreviationsNeqnumber of equivalentPBProtein BlocksPDBProtein DataBankRMSfroot mean square fluctuationsCommunicated by Ramaswamy H. Sarma.

PBxplore: a tool to analyze local protein structure and deformability with Protein Blocks.

This paper describes the development and application of a suite of tools, called PBxplore, to analyze the dynamics and deformability of protein structures using Protein Blocks (PBs). Proteins are highly dynamic macromolecules, and a classical way to analyze their inherent flexibility is to perform molecular dynamics simulations. The advantage of using small structural prototypes such as PBs is to give a good approximation of the local structure of the protein backbone. More importantly, by reducing the conformational complexity of protein structures, PBs allow analysis of local protein deformability which cannot be done with other methods and had been used efficiently in different applications. PBxplore is able to process large amounts of data such as those produced by molecular dynamics simulations. It produces frequencies, entropy and information logo outputs as text and graphics. PBxplore is available at https://github.com/pierrepo/PBxplore and is released under the open-source MIT license.

Toward Atomistic Resolution Structure of Phosphatidylcholine Headgroup and Glycerol Backbone at Different Ambient Conditions.

Phospholipids are essential building blocks of biological membranes. Despite a vast amount of very accurate experimental data, the atomistic resolution structures sampled by the glycerol backbone and choline headgroup in phoshatidylcholine bilayers are not known. Atomistic resolution molecular dynamics simulations have the potential to resolve the structures, and to give an arrestingly intuitive interpretation of the experimental data, but only if the simulations reproduce the data within experimental accuracy. In the present work, we simulated phosphatidylcholine (PC) lipid bilayers with 13 different atomistic models, and compared simulations with NMR experiments in terms of the highly structurally sensitive C-H bond vector order parameters. Focusing on the glycerol backbone and choline headgroups, we showed that the order parameter comparison can be used to judge the atomistic resolution structural accuracy of the models. Accurate models, in turn, allow molecular dynamics simulations to be used as an interpretation tool that translates these NMR data into a dynamic three-dimensional representation of biomolecules in biologically relevant conditions. In addition to lipid bilayers in fully hydrated conditions, we reviewed previous experimental data for dehydrated bilayers and cholesterol-containing bilayers, and interpreted them with simulations. Although none of the existing models reached experimental accuracy, by critically comparing them we were able to distill relevant chemical information: (1) increase of choline order parameters indicates the P-N vector tilting more parallel to the membrane, and (2) cholesterol induces only minor changes to the PC (glycerol backbone) structure. This work has been done as a fully open collaboration, using nmrlipids.blogspot.fi as a communication platform; all the scientific contributions were made publicly on this blog. During the open research process, the repository holding our simulation trajectories and files ( https://zenodo.org/collection/user-nmrlipids ) has become the most extensive publicly available collection of molecular dynamics simulation trajectories of lipid bilayers.