Martini 3: a general purpose force field for coarse-grained molecular dynamics

Specific interactions of peripheral membrane proteins with lipids: what can molecular simulations show us?

Peripheral membrane proteins (PMPs) can reversibly and specifically bind to biological membranes to carry out functions such as cell signalling, enzymatic activity, or membrane remodelling. Structures of these proteins and of their lipid-binding domains are typically solved in a soluble form, sometimes with a lipid or lipid headgroup at the binding site. To provide a detailed molecular view of PMP interactions with the membrane, computational methods such as molecular dynamics (MD) simulations can be applied. Here, we outline recent attempts to characterise these binding interactions, focusing on both intracellular proteins, such as phosphatidylinositol phosphate (PIP)-binding domains, and extracellular proteins such as glycolipid-binding bacterial exotoxins. We compare methods used to identify and analyse lipid-binding sites from simulation data and highlight recent work characterising the energetics of these interactions using free energy calculations. We describe how improvements in methodologies and computing power will help MD simulations to continue to contribute to this field in the future.

Lipid-Mediated Association of the Slg1 Transmembrane Domains in Yeast Plasma Membranes

Clustering of transmembrane proteins underlies a multitude of fundamental biological processes at the plasma membrane (PM) such as receptor activation, lateral domain formation, and mechanotransduction. The self-association of the respective transmembrane domains (TMDs) has also been suggested to be responsible for the micron-scaled patterns seen for integral membrane proteins in the budding yeast PM. However, the underlying interplay between the local lipid composition and the TMD identity is still not mechanistically understood. In this work, we combined coarse-grained molecular dynamics simulations of simplified bilayer systems with high-resolution live-cell microscopy to analyze the distribution of a representative helical yeast TMD from the PM sensor Slg1 within different lipid environments. In our simulations, we specifically evaluated the effects of acyl chain saturation and anionic lipid head groups on the association of two TMDs. We found that weak lipid-protein interactions significantly affect the configuration of TMD dimers and the free energy of association. Increased amounts of unsaturated phospholipids (PLs) strongly reduced the helix-helix interaction, while the presence of anionic phosphatidylserine (PS) hardly affected the dimer formation. We could experimentally confirm this surprising lack of effect of PS using the network factor, a mesoscopic measure of PM pattern formation in yeast cells. Simulations also showed that the formation of TMD dimers in turn increased the order parameter of the surrounding lipids and induced long-range perturbations in lipid organization. In summary, our results shed new light on the mechanisms of lipid-mediated dimerization of TMDs in complex lipid mixtures.

Multiscale molecular modelling: from electronic structure to dynamics of nanosystems and beyond

Important contemporary biological and materials problems often depend on interactions that span orders of magnitude differences in spatial and temporal dimensions. This Tutorial Review attempts to provide an introduction to such fascinating problems through a series of case studies, aimed at beginning researchers, graduate students, postdocs and more senior colleagues who are changing direction to focus on multiscale aspects of their research. The choice of specific examples is highly personal, with examples either chosen from our own work or outstanding multiscale efforts from the literature. I start with various embedding schemes, as exemplified by polarizable continuum models, 3-D RISM, molecular DFT and frozen-density embedding. Next, QM/MM (quantum mechanical/molecular mechanical) techniques are the workhorse of pm-to-nm/ps-to-ns simulations; examples are drawn from enzymes and from nanocatalysis for oil-sands upgrading. Using polarizable force-fields in the QM/MM framework represents a burgeoning subfield; with examples from ion channels and electron dynamics in molecules subject to strong external fields, probing the atto-second dynamics of the electrons with RT-TDDFT (real-time - time-dependent density functional theory) eventually coupled with nuclear motion through the Ehrenfest approximation. This is followed by a section on coarse graining, bridging dimensions from atoms to cells. The penultimate chapter gives a quick overview of multiscale approaches that extend into the meso- and macro-scales, building on atomistic and coarse-grained techniques to enter the world of materials engineering, on the one hand, and cell biology, on the other. A final chapter gives just a glimpse of the burgeoning impact of machine learning on the structure-dynamics front. I aim to capture the excitement of contemporary leading-edge breakthroughs in the description of physico-chemical systems and processes in complex environments, with only enough historical content to provide context and aid the next generation of methodological development. While I aim also for a clear description of the essence of methodological breakthroughs, equations are kept to a minimum and detailed formalism and implementation details are left to the references. My approach is very selective (case studies) rather than exhaustive. I think that these case studies should provide fodder to build as complete a reference tree on multiscale modelling as the reader may wish, through forward and backward citation analysis. I hope that my choices of cases will excite interest in newcomers and help to fuel the growth of multiscale modelling in general.

SPICA Force Field for Proteins and Peptides

A coarse-grained (CG) model for peptides and proteins was developed as an extension of the Surface Property fItting Coarse grAined (SPICA) force field (FF). The model was designed to examine membrane proteins that are fully compatible with the lipid membranes of the SPICA FF. A preliminary version of this protein model was created using thermodynamic properties, including the surface tension and density in the SPICA (formerly called SDK) FF. In this study, we improved the CG protein model to facilitate molecular dynamics (MD) simulations with a reproduction of multiple properties from both experiments and all-atom (AA) simulations. An elastic network model was adopted to maintain the secondary structure within a single chain. The side-chain analogues reproduced the transfer free energy profiles across the lipid membrane and demonstrated reasonable association free energy (potential of mean force) in water compared to those from AA MD. A series of peptides/proteins adsorbed onto or penetrated into the membrane simulated by the CG MD correctly predicted the penetration depths and tilt angles of peripheral and transmembrane peptides/proteins as comparable to those in the orientations of proteins in membranes (OPM) database. In addition, the dimerization free energies of several transmembrane helices within a lipid bilayer were comparable to those from experimental estimation. Application studies on a series of membrane protein assemblies, scramblases, and poliovirus capsids demonstrated the good performance of the SPICA FF.

Improving Martini 3 for Disordered and Multidomain Proteins

Coarse-grained molecular dynamics simulations are a useful tool to determine conformational ensembles of proteins. Here, we show that the coarse-grained force field Martini 3 underestimates the global dimensions of intrinsically disordered proteins (IDPs) and multidomain proteins when compared with small-angle X-ray scattering (SAXS) data and that increasing the strength of protein-water interactions favors more expanded conformations. We find that increasing the strength of interactions between protein and water by ca. 10% results in improved agreement with the SAXS data for IDPs and multidomain proteins. We also show that this correction results in a more accurate description of self-association of IDPs and folded proteins and better agreement with paramagnetic relaxation enhancement data for most IDPs. While simulations with this revised force field still show deviations to experiments for some systems, our results suggest that it is overall a substantial improvement for coarse-grained simulations of soluble proteins.

A unifying framework for amyloid-mediated membrane damage: The lipid-chaperone hypothesis

Over the past thirty years, researchers have highlighted the role played by a class of proteins or polypeptides that forms pathogenic amyloid aggregates in vivo, including i) the amyloid Aβ peptide, which is known to form senile plaques in Alzheimer's disease; ii) α-synuclein, responsible for Lewy body formation in Parkinson's disease and iii) IAPP, which is the protein component of type 2 diabetes-associated islet amyloids. These proteins, known as intrinsically disordered proteins (IDPs), are present as highly dynamic conformational ensembles. IDPs can partially (mis) fold into (dys) functional conformations and accumulate as amyloid aggregates upon interaction with other cytosolic partners such as proteins or lipid membranes. In addition, an increasing number of reports link the toxicity of amyloid proteins to their harmful effects on membrane integrity. Still, the molecular mechanism underlying the amyloidogenic proteins transfer from the aqueous environment to the hydrocarbon core of the membrane is poorly understood. This review starts with a historical overview of the toxicity models of amyloidogenic proteins to contextualize the more recent lipid-chaperone hypothesis. Then, we report the early molecular-level events in the aggregation and ion-channel pore formation of Aβ, IAPP, and α-synuclein interacting with model membranes, emphasizing the complexity of these processes due to their different spatial-temporal resolutions. Next, we underline the need for a combined experimental and computational approach, focusing on the strengths and weaknesses of the most commonly used techniques. Finally, the last two chapters highlight the crucial role of lipid-protein complexes as molecular switches among ion-channel-like formation, detergent-like, and fibril formation mechanisms and their implication in fighting amyloidogenic diseases.

Membrane thickness, lipid phase and sterol type are determining factors in the permeability of membranes to small solutes

Cell membranes provide a selective semi-permeable barrier to the passive transport of molecules. This property differs greatly between organisms. While the cytoplasmic membrane of bacterial cells is highly permeable for weak acids and glycerol, yeasts can maintain large concentration gradients. Here we show that such differences can arise from the physical state of the plasma membrane. By combining stopped-flow kinetic measurements with molecular dynamics simulations, we performed a systematic analysis of the permeability of a variety of small molecules through synthetic membranes of different lipid composition to obtain detailed molecular insight into the permeation mechanisms. While membrane thickness is an important parameter for the permeability through fluid membranes, the largest differences occur when the membranes transit from the liquid-disordered to liquid-ordered and/or to gel state, which is in agreement with previous work on passive diffusion of water. By comparing our results with in vivo measurements from yeast, we conclude that the yeast membrane exists in a highly ordered and rigid state, which is comparable to synthetic saturated DPPC-sterol membranes.

Membrane permeability of small molecules depends on the composition of the lipid bilayer. Here, authors compare permeability measured on membranes in different physical states and conclude that the yeast membrane exists in a highly ordered phase.

Correction Schemes for Absolute Binding Free Energies Involving Lipid Bilayers

Absolute binding free-energy (ABFE) calculations are playing an increasing role in drug design, especially as they can be performed on a range of disparate compounds and direct comparisons between them can be made. It is, however, especially important to ensure that they are as accurate as possible, as unlike relative binding free-energy (RBFE) calculations, one does not benefit as much from a cancellation of errors during the calculations. In most modern implementations of ABFE calculations, a particle mesh Ewald scheme is typically used to treat the electrostatic contribution to the free energy. A central requirement of such schemes is that the box preserves neutrality throughout the calculation. There are many ways to deal with this problem that have been discussed over the years ranging from a neutralizing plasma with a post hoc correction term through to a simple co-alchemical ion within the same box. The post hoc correction approach is the most widespread. However, the vast majority of these studies have been applied to a soluble protein in a homogeneous solvent (water or salt solution). In this work, we explore which of the more common approaches would be the most suitable for a simulation box with a lipid bilayer within it. We further develop the idea of the so-called Rocklin correction for lipid-bilayer systems and show how such a correction could work. However, we also show that it will be difficult to make this generalizable in a practical way and thus we conclude that the use of a “co-alchemical ion” is the most useful approach for simulations involving lipid membrane systems.

Interaction With the Lipid Membrane Influences Fentanyl Pharmacology

Overdose deaths from fentanyl have reached epidemic proportions in the USA and are increasing worldwide. Fentanyl is a potent opioid agonist that is less well reversed by naloxone than morphine. Due to fentanyl’s high lipophilicity and elongated structure we hypothesised that its unusual pharmacology may be explained by its interactions with the lipid membrane on route to binding to the μ-opioid receptor (MOPr). Through coarse-grained molecular dynamics simulations, electrophysiological recordings and cell signalling assays, we determined how fentanyl and morphine access the orthosteric pocket of MOPr. Morphine accesses MOPr via the aqueous pathway; first binding to an extracellular vestibule, then diffusing into the orthosteric pocket. In contrast, fentanyl may take a novel route; first partitioning into the membrane, before accessing the orthosteric site by diffusing through a ligand-induced gap between the transmembrane helices. In electrophysiological recordings fentanyl-induced currents returned after washout, suggesting fentanyl deposits in the lipid membrane. However, mutation of residues forming the potential MOPr transmembrane access site did not alter fentanyl’s pharmacological profile in vitro. A high local concentration of fentanyl in the lipid membrane, possibly in combination with a novel lipophilic binding route, may explain the high potency and lower susceptibility of fentanyl to reversal by naloxone.

What have molecular simulations contributed to understanding of Gram-negative bacterial cell envelopes?

Bacterial cell envelopes are compositionally complex and crowded and while highly dynamic in some areas, their molecular motion is very limited, to the point of being almost static in others. Therefore, it is no real surprise that studying them at high resolution across a range of temporal and spatial scales requires a number of different techniques. Details at atomistic to molecular scales for up to tens of microseconds are now within range for molecular dynamics simulations. Here we review how such simulations have contributed to our current understanding of the cell envelopes of Gram-negative bacteria.

Supramolecular Engineering of Alkylated, Fluorinated, and Mixed Amphiphiles

The rational design of perfluorinated amphiphiles to control the supramolecular aggregation in aqueous medium is still a key challenge for the engineering of supramolecular architectures. Here we present the synthesis and physical properties of six novel non-ionic amphiphiles. We also studied the effect of mixed alkylated and perfluorinated segments in a single amphiphile and compared it with only alkylated and perfluorinated units. To explore their morphological behavior in aqueous medium, we used dynamic light scattering (DLS) and cryo-TEM/EM measurements. We further confirmed their assembly mechanisms with theoretical investigations, using the Martini model to perform large-scale coarse-grained molecular dynamics simulations. These novel synthesized amphiphiles offer a greater and more systematic understanding of how perfluorinated systems assemble in aqueous medium and suggest new directions for rational designing of new amphiphilic systems and interpreting their assembly process. This article is protected by copyright. All rights reserved.

Membrane Mixer: A Toolkit for Efficient Shuffling of Lipids in Heterogeneous Biological Membranes

Molecular dynamics (MD) simulations of biological membranes have achieved such levels of sophistication that are commonly used to predict unresolved structures and various properties of lipids and to substantiate experimental data. While achieving sufficient sampling of lipid dynamics remains a major challenge, a commonly used method to improve lipid sampling, e.g., in terms of specific interactions with membrane-associated proteins, is to randomize the initial arrangement of lipid constituents in multiple replicas of simulations, without changing the overall lipid composition of the membrane of interest. Here, we introduce a method that can rapidly generate multiple replicas of lipid bilayers with different spatial and conformational configurations for any given lipid composition. The underlying algorithm, which allows one to shuffle lipids at any desired level, relies on the application of an external potential, here referred to as the "carving potential", that removes clashes/entanglements before lipid positions are exchanged (shuffled), thereby minimizing the energy penalty due to abrupt lipid repositioning. The method is implemented as "Membrane Mixer Plugin (MMP) 1.0" in VMD, with a convenient graphical user interface that guides the user in setting various options and parameters. The plugin is fully automated and generates new membrane replicas more rapidly and conveniently than other analogous tools. The plugin and its capabilities introduced here can be extended to include additional features in future versions.

Coarse-Grained Model of Thiol-Epoxy-Based Alternating Copolymers in Explicit Solvents

The cosolvent method has been widely used in the self-assembly of amphiphilic alternating copolymers (ACPs), but the role of good and selective solvents is rarely investigated. Here, we have developed a coarse-grained (CG) model for the widely studied thiol-epoxy-based amphiphilic ACPs and a three-bead CG model for tetrahydrofuran (THF) as the good solvent, which is compatible with the MARTINI water model. The accuracy of both the CG polymer and THF models was validated by reproducing the structural and thermodynamic properties obtained from experiments or atomistic simulation results. Density in bulk, the radius of gyration, and solvation free energy in water or THF showed a good agreement between CG and atomistic models. The CG models were further employed to explore the self-assembly of ACPs in THF/water mixtures with different compositions. Chain folding and liquid-liquid phase separation behaviors were found with increasing water fractions, which were the key steps of the self-assembly process. This work will provide a basic platform to explore the self-assembly of amphiphilic ACPs in solvent mixtures and to reveal the real role of different solvents in self-assembly.

The phase behaviour of cetyltrimethylammonium chloride surfactant aqueous solutions at high concentrations: an all-atom molecular dynamics simulation study

We explore the phase behaviour of aqueous solutions of the cetyltrimethyl ammonium chloride (CTAC) surfactant and in particular the transition from the micellar phase (L₁) to the hexagonal columnar phase (H₁) by employing all-atom (AA) molecular dynamics (MD) simulations for six CTAC concentrations in the range of 34.1 wt% to 70.5 wt%, at the temperature of 318 K and pressure of 1 atm. For the concentrations considered, we examine the spontaneous occurrence of the H₁ phase by testing a number of plausible values for the linear density (molecules per unit length) along the cylindrical columns. Using large simulation cells and starting from random initial configurations, the MD simulations demonstrate that the micellar phase occurs for concentrations up to 50.0 wt%, with CTAC molecules self-assembling into a mixture of spherical and rod-like micelles. At even higher concentrations, the system self-organizes into the H₁ phase in accordance with the available experimental data. For the analysis of the MD trajectories, we devise a clustering algorithm based on Voronoi tesselation which enables (a) the thorough characterization of the shape and structure of both molecules and assemblies, and (b) the investigation of the positional and orientational order in the system that are further scrutinised using radial pair correlation functions and X-ray diffraction patterns. Our work paves the way for the investigation of the phase behaviour at high concentrations of other surfactants.

Assembly and Analysis of Cell-Scale Membrane Envelopes

The march toward exascale computing will enable routine molecular simulation of larger and more complex systems, for example, simulation of entire viral particles, on the scale of approximately billions of atoms─a simulation size commensurate with a small bacterial cell. Anticipating the future hardware capabilities that will enable this type of research and paralleling advances in experimental structural biology, efforts are currently underway to develop software tools, procedures, and workflows for constructing cell-scale structures. Herein, we describe our efforts in developing and implementing an efficient and robust workflow for construction of cell-scale membrane envelopes and embedding membrane proteins into them. A new approach for construction of massive membrane structures that are stable during the simulations is built on implementing a subtractive assembly technique coupled with the development of a structure concatenation tool (fastmerge), which eliminates overlapping elements based on volumetric criteria rather than adding successive molecules to the simulation system. Using this approach, we have constructed two "protocells" consisting of MARTINI coarse-grained beads to represent cellular membranes, one the size of a cellular organelle and another the size of a small bacterial cell. The membrane envelopes constructed here remain whole during the molecular dynamics simulations performed and exhibit water flux only through specific proteins, demonstrating the success of our methodology in creating tight cell-like membrane compartments. Extended simulations of these cell-scale structures highlight the propensity for nonspecific interactions between adjacent membrane proteins leading to the formation of protein microclusters on the cell surface, an insight uniquely enabled by the scale of the simulations. We anticipate that the experiences and best practices presented here will form the basis for the next generation of cell-scale models, which will begin to address the addition of soluble proteins, nucleic acids, and small molecules essential to the function of a cell.

PyLipID: A Python Package for Analysis of Protein-Lipid Interactions from Molecular Dynamics Simulations

Lipids play important modulatory and structural roles for membrane proteins. Molecular dynamics simulations are frequently used to provide insights into the nature of these protein-lipid interactions. Systematic comparative analysis requires tools that provide algorithms for objective assessment of such interactions. We introduce PyLipID, a Python package for the identification and characterization of specific lipid interactions and binding sites on membrane proteins from molecular dynamics simulations. PyLipID uses a community analysis approach for binding site detection, calculating lipid residence times for both the individual protein residues and the detected binding sites. To assist structural analysis, PyLipID produces representative bound lipid poses from simulation data, using a density-based scoring function. To estimate residue contacts robustly, PyLipID uses a dual-cutoff scheme to differentiate between lipid conformational rearrangements while bound from full dissociation events. In addition to the characterization of protein-lipid interactions, PyLipID is applicable to analysis of the interactions of membrane proteins with other ligands. By combining automated analysis, efficient algorithms, and open-source distribution, PyLipID facilitates the systematic analysis of lipid interactions from large simulation data sets of multiple species of membrane proteins.

Conformational ensembles of intrinsically disordered proteins and flexible multidomain proteins

Intrinsically disordered proteins (IDPs) and multidomain proteins with flexible linkers show a high level of structural heterogeneity and are best described by ensembles consisting of multiple conformations with associated thermodynamic weights. Determining conformational ensembles usually involves the integration of biophysical experiments and computational models. In this review, we discuss current approaches to determine conformational ensembles of IDPs and multidomain proteins, including the choice of biophysical experiments, computational models used to sample protein conformations, models to calculate experimental observables from protein structure, and methods to refine ensembles against experimental data. We also provide examples of recent applications of integrative conformational ensemble determination to study IDPs and multidomain proteins and suggest future directions for research in the field.

Protein dynamics and lipid affinity of monomeric, zeaxanthin-binding LHCII in thylakoid membranes

The xanthophyll cycle in the antenna of photosynthetic organisms under light stress is one of the most well-known processes in photosynthesis, but its role is not well understood. In the xanthophyll cycle, violaxanthin (Vio) is reversibly transformed to zeaxanthin (Zea) that occupies Vio binding sites of light-harvesting antenna proteins. Higher monomer/trimer ratios of the most abundant light-harvesting protein, the light-harvesting complex II (LHCII), usually occur in Zea accumulating membranes and have been observed in plants after prolonged illumination and during high-light acclimation. We present a combined NMR and coarse-grained simulation study on monomeric LHCII from the npq2 mutant that constitutively binds Zea in the Vio binding pocket. LHCII was isolated from ¹³C-enriched npq2 Chlamydomonas reinhardtii (Cr) cells and reconstituted in thylakoid lipid membranes. NMR results reveal selective changes in the fold and dynamics of npq2 LHCII compared with the trimeric, wild-type and show that npq2 LHCII contains multiple mono- or digalactosyl diacylglycerol lipids (MGDG and DGDG) that are strongly protein bound. Coarse-grained simulations on npq2 LHCII embedded in a thylakoid lipid membrane agree with these observations. The simulations show that LHCII monomers have more extensive lipid contacts than LHCII trimers and that protein-lipid contacts are influenced by Zea. We propose that both monomerization and Zea binding could have a functional role in modulating membrane fluidity and influence the aggregation and conformational dynamics of LHCII with a likely impact on photoprotection ability.

The Bak core dimer focuses triacylglycerides in the membrane

Apoptosis, the intrinsic programmed cell death process, is mediated by the Bcl-2 family members Bak and Bax. Activation via formation of symmetric core dimers and oligomerization on the mitochondrial outer membrane (MOM) leads to permeabilization and cell death. Although this process is linked to the MOM, the role of the membrane in facilitating such pores is poorly understood. We recently described Bak core domain dimers, revealing lipid binding sites and an initial role of lipids in oligomerization. Here we describe simulations that identified localized clustering and interaction of triacylglycerides (TAGs) with a minimized Bak dimer construct. Coalescence of TAGs occurred beneath this Bak dimer, mitigating dimer-induced local membrane thinning and curvature in representative coarse-grain MOM and model membrane systems. Furthermore, the effects observed as a result of coarse-grain TAG cluster formation was concentration dependent, scaling from low physiological MOM concentrations to those found in other organelles. We find that increasing the TAG concentration in liposomes mimicking the MOM decreased the ability of activated Bak to permeabilize these liposomes. These results suggest that the presence of TAGs within a Bak-lipid membrane preserves membrane integrity and is associated with reduced membrane stress, suggesting a possible role of TAGs in Bak-mediated apoptosis.

Martini 3 Model of Cellulose Microfibrils: On the Route to Capture Large Conformational Changes of Polysaccharides

High resolution data from all-atom molecular simulations is used to parameterize a Martini 3 coarse-grained (CG) model of cellulose I allomorphs and cellulose type-II fibrils. In this case, elementary molecules are represented by four effective beads centred in the positions of O2, O3, C6, and O6 atoms in the D-glucose cellulose subunit. Non-bonded interactions between CG beads are tuned according to a low statistical criterion of structural deviation using the Martini 3 type of interactions and are capable of being indistinguishable for all studied cases. To maintain the crystalline structure of each single cellulose chain in the microfibrils, elastic potentials are employed to retain the ribbon-like structure in each chain. We find that our model is capable of describing different fibril-twist angles associated with each type of cellulose fibril in close agreement with atomistic simulation. Furthermore, our CG model poses a very small deviation from the native-like structure, making it appropriate to capture large conformational changes such as those that occur during the self-assembly process. We expect to provide a computational model suitable for several new applications such as cellulose self-assembly in different aqueous solutions and the thermal treatment of fibrils of great importance in bioindustrial applications.

Druggable Lipid Binding Sites in Pentameric Ligand-Gated Ion Channels and Transient Receptor Potential Channels

Lipids modulate the function of many ion channels, possibly through direct lipid-protein interactions. The recent outpouring of ion channel structures by cryo-EM has revealed many lipid binding sites. Whether these sites mediate lipid modulation of ion channel function is not firmly established in most cases. However, it is intriguing that many of these lipid binding sites are also known sites for other allosteric modulators or drugs, supporting the notion that lipids act as endogenous allosteric modulators through these sites. Here, we review such lipid-drug binding sites, focusing on pentameric ligand-gated ion channels and transient receptor potential channels. Notable examples include sites for phospholipids and sterols that are shared by anesthetics and vanilloids. We discuss some implications of lipid binding at these sites including the possibility that lipids can alter drug potency or that understanding protein-lipid interactions can guide drug design. Structures are only the first step toward understanding the mechanism of lipid modulation at these sites. Looking forward, we identify knowledge gaps in the field and approaches to address them. These include defining the effects of lipids on channel function in reconstituted systems using asymmetric membranes and measuring lipid binding affinities at specific sites using native mass spectrometry, fluorescence binding assays, and computational approaches.

Insights into lipid-protein interactions from computer simulations

Lipid-protein interactions play an important direct role in the function of many membrane proteins. We argue they are key players in membrane structure, modulate membrane proteins in more subtle ways than direct binding, and are important for understanding the mechanism of classes of hydrophobic drugs. By directly comparing membrane proteins from different families in the same, complex lipid mixture, we found a unique lipid environment for every protein. Extending this work, we identified both differences and similarities in the lipid environment of GPCRs, dependent on which family they belong to and in some cases their conformational state, with particular emphasis on the distribution of cholesterol. More recently, we have been studying modes of coupling between protein conformation and local membrane properties using model proteins. In more applied approaches, we have used similar methods to investigate specific hypotheses on interactions of lipid and lipid-like molecules with ion channels. We conclude this perspective with some considerations for future work, including a new more sophisticated coarse-grained force field (Martini 3), an interactive visual exploration framework, and opportunities to improve sampling.

Automatic multi-objective optimization of coarse-grained lipid force fields using SwarmCG

The development of coarse-grained (CG) molecular models typically requires a time-consuming iterative tuning of parameters in order to have the approximated CG models behave correctly and consistently with, e.g., available higher-resolution simulation data and/or experimental observables. Automatic data-driven approaches are increasingly used to develop accurate models for molecular dynamics simulations. However, the parameters obtained via such automatic methods often make use of specifically designed interaction potentials and are typically poorly transferable to molecular systems or conditions other than those used for training them. Using a multi-objective approach in combination with an automatic optimization engine (SwarmCG), here, we show that it is possible to optimize CG models that are also transferable, obtaining optimized CG force fields (FFs). As a proof of concept, here, we use lipids for which we can avail reference experimental data (area per lipid and bilayer thickness) and reliable atomistic simulations to guide the optimization. Once the resolution of the CG models (mapping) is set as an input, SwarmCG optimizes the parameters of the CG lipid models iteratively and simultaneously against higher-resolution simulations (bottom-up) and experimental data (top-down references). Including different types of lipid bilayers in the training set in a parallel optimization guarantees the transferability of the optimized lipid FF parameters. We demonstrate that SwarmCG can reach satisfactory agreement with experimental data for different resolution CG FFs. We also obtain stimulating insights into the precision-resolution balance of the FFs. The approach is general and can be effectively used to develop new FFs and to improve the existing ones.

Multiscale Simulation of Protein Corona Formation on Silver Nanoparticles: Study of Ovispirin-1 Peptide Adsorption

The exposure of nanoparticles (NPs) to biofluids leads to the rapid coverage of proteins, named protein corona, which alters the NPs' chemicophysical and biological properties. Fundamental studies of the protein corona are thus critical to the increasing applications of NPs in nanotechnology and nanomedicines. The present work utilizes multiscale simulations of a model biological system, small ovispirin-1 peptides, and bare silver nanoparticles (AgNPs) to examine the NPs' size and surface hydrophilicity effects on formation dynamics and the structure of the peptide corona. Our simulations revealed the different adsorption dynamics of ovispirin-1 peptides on the NPs, including the direct adsorption of a single peptide and peptide aggregates and multistep adsorption, as well as an intermediate cycle of desorption and readsorption. Notably, the whole process of peptide adsorption on hydrophilic AgNP surfaces can be generalized as three stages: diffusion to the surface, initial landing via hydrophilic residues, and the final attachment. The decrease in AgNP's size leads to faster adsorption with more heterogeneous peptide interfacial dynamics, a denser and inhomogeneous peptide packing structure, and a wider distribution of adsorption orientations. Subsequent atomistic molecular dynamics simulations demonstrated that on the hydrophilic AgNP surfaces, adsorbed peptides display moderate changes in their secondary structure, resulting in further changes of corona composition, i.e., amino acid residue distribution on the surface.

Temperature-Transferable Coarse-Grained Model for Poly(propylene oxide) to Study Thermo-Responsive Behavior of Triblock Copolymers

Thermo-responsive behavior of ethylene oxide (EO)-propylene oxide (PO) copolymers makes them suitable for many potential applications. Reproducing the origins of the tunable properties of EO-PO copolymers using coarse-grained (CG) models such as the MARTINI force field is critically important for building a better understanding of their behavior. In the present work, we have investigated the effects of coarse-graining on the water-polymer interaction across a temperature range. We compared the performance of different all-atom force fields to find the most appropriate one for the purpose of PO block parameterization in the MARTINI platform. We parameterized a CG temperature-dependent PO model based on the reproduction of the atomistic free energy of transfer of propylene oxide trimer from octane to water over a range of temperatures (20-60 °C) and compared the atomistic bond and angle distributions. Then, we used the model to study the effects of EO/PO ratio, molecular weight, and concentration on the thermo-responsive behavior of EO-PO copolymers in water. The results show an excellent agreement with experiments in different areas. Our temperature-dependent model reproduces (1) micellar phase above critical micelle temperature (CMT) and unimer phase below CMT for different Pluronics (a class of EO-PO triblock copolymers) spanning many EO/PO ratios and molecular weights; (2) spherical-to-rodlike micellar shape transition for Pluronics with 60 wt % of PO content or more; (3) diffusion coefficients for Pluronics with high PO content (P104 Pluronic with a PO mass of 3500 g mol^-1) across a broad range of temperatures; and (4) micelle core size and micelle diameter similar to experimental results. Overall, our model improves the temperature sensitivity of EO-PO copolymers of existing models significantly, particularly for copolymers that are dominated by PO agents.

Mapping Mechanostable Pulling Geometries of a Therapeutic Anticalin/CTLA-4 Protein Complex

We used single-molecule AFM force spectroscopy (AFM-SMFS) in combination with click chemistry to mechanically dissociate anticalin, a non-antibody protein binding scaffold, from its target (CTLA-4), by pulling from eight different anchor residues. We found that pulling on the anticalin from residue 60 or 87 resulted in significantly higher rupture forces and a decrease in koff by 2-3 orders of magnitude over a force range of 50-200 pN. Five of the six internal anchor points gave rise to complexes significantly more stable than N- or C-terminal anchor points, rupturing at up to 250 pN at loading rates of 0.1-10 nN s-1. Anisotropic network modeling and molecular dynamics simulations helped to explain the geometric dependency of mechanostability. These results demonstrate that optimization of attachment residue position on therapeutic binding scaffolds can provide large improvements in binding strength, allowing for mechanical affinity maturation under shear stress without mutation of binding interface residues.

Supramolecular Organization of SARS-CoV and SARS-CoV-2 Virions Revealed by Coarse-Grained Models of Intact Virus Envelopes

The coronavirus disease 19 (COVID-19) pandemic is causing a global health crisis and has already caused a devastating societal and economic burden. The pathogen, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has a high sequence and architecture identity with SARS-CoV, but far more people have been infected by SARS-CoV-2. Here, combining the structural data from cryo-electron microscopy and structure prediction, we constructed bottom-up Martini coarse-grained models of intact SARS-CoV and SARS-CoV-2 envelopes. Microsecond molecular dynamics simulations were performed, allowing us to explore their dynamics and supramolecular organization. Both SARS-CoV and SARS-CoV-2 envelopes present a spherical morphology, with structural proteins forming multiple string-like islands in the membrane and clusters between the heads of spike proteins. Critical differences between the SARS-CoV and SARS-CoV-2 envelopes are the interaction pattern between the spike proteins and the flexibility of the spike proteins. Our models provide structural and dynamic insights into the SARS virus envelopes and could be used for further investigation, such as drug design and membrane fusion and fission processes.

Polyply; a python suite for facilitating simulations of macromolecules and nanomaterials

Molecular dynamics simulations play an increasingly important role in the rational design of (nano)-materials and in the study of biomacromolecules. However, generating input files and realistic starting coordinates for these simulations is a major bottleneck, especially for high throughput protocols and for complex multi-component systems. To eliminate this bottleneck, we present the polyply software suite that provides 1) a multi-scale graph matching algorithm designed to generate parameters quickly and for arbitrarily complex polymeric topologies, and 2) a generic multi-scale random walk protocol capable of setting up complex systems efficiently and independent of the target force-field or model resolution. We benchmark quality and performance of the approach by creating realistic coordinates for polymer melt simulations, single-stranded as well as circular single-stranded DNA. We further demonstrate the power of our approach by setting up a microphase-separated block copolymer system, and by generating a liquid-liquid phase separated system inside a lipid vesicle.

Complex nanoemulsion for vitamin delivery: droplet organization and interaction with skin membranes

Lipid nanoemulsions are promising nanomaterials for drug delivery applications in food, pharmaceutical and cosmetic industries. Despite the noteworthy commercial interest, little is known about their supramolecular organization, especially about how such multicomponent formulations interact with cell membranes. In the present work, coarse-grained molecular dynamics simulations have been employed to study the self-assembly of a 15-component lipid nanoemulsion droplet containing vitamins A and E for skin delivery. Our results display aspects of the unique "onion-like" agglomeration between the chemical constituents in the different layers of the lipid nanodroplet. Vitamin E molecules are more concentrated in the center of the droplet together with other hydrophobic constituents such as the triglycerides with long tails. On the other hand, vitamin A occupies an intermediate layer between the core and the co-emulsifier surface of the nanodroplet, together with lecithin phospholipids. Coarse-grained molecular dynamics simulations were also performed to provide insight into the first steps involved in absorption and penetration of the nanodroplet through skin membrane models, representing an intracellular (hair follicle infundibulum) and intercellular pathway (stratum corneum) through the skin. Our data provide a first view on the complex organization of commercial nanoemulsion and its interaction with skin membranes. We expect our results to open the way towards the rational design of such nanomaterials.

Machine learning-driven multiscale modeling reveals lipid-dependent dynamics of RAS signaling proteins

RAS is a signaling protein associated with the cell membrane that is mutated in up to 30% of human cancers. RAS signaling has been proposed to be regulated by dynamic heterogeneity of the cell membrane. Investigating such a mechanism requires near-atomistic detail at macroscopic temporal and spatial scales, which is not possible with conventional computational or experimental techniques. We demonstrate here a multiscale simulation infrastructure that uses machine learning to create a scale-bridging ensemble of over 100,000 simulations of active wild-type KRAS on a complex, asymmetric membrane. Initialized and validated with experimental data (including a new structure of active wild-type KRAS), these simulations represent a substantial advance in the ability to characterize RAS-membrane biology. We report distinctive patterns of local lipid composition that correlate with interfacially promiscuous RAS multimerization. These lipid fingerprints are coupled to RAS dynamics, predicted to influence effector binding, and therefore may be a mechanism for regulating cell signaling cascades.

Challenges and frontiers of computational modelling of biomolecular recognition

Biomolecular recognition including binding of small molecules, peptides and proteins to their target receptors plays a key role in cellular function and has been targeted for therapeutic drug design. However, the high flexibility of biomolecules and slow binding and dissociation processes have presented challenges for computational modelling. Here, we review the challenges and computational approaches developed to characterise biomolecular binding, including molecular docking, molecular dynamics simulations (especially enhanced sampling) and machine learning. Further improvements are still needed in order to accurately and efficiently characterise binding structures, mechanisms, thermodynamics and kinetics of biomolecules in the future.

A conserved function of Human DLC3 and Drosophila Cv-c in testis development

The identification of genes affecting gonad development is essential to understand the mechanisms causing Variations/Differences in Sex Development (DSD). Recently, a DLC3 mutation was associated with male gonadal dysgenesis in 46,XY DSD patients. We have studied the requirement of Cv-c, the Drosophila ortholog of DLC3, in Drosophila gonad development, as well as the functional capacity of DLC3 human variants to rescue cv-c gonad defects. We show that Cv-c is required to maintain testis integrity during fly development. We find that Cv-c and human DLC3 can perform the same function in fly embryos, as flies carrying wild type but not patient DLC3 variations can rescue gonadal dysgenesis, suggesting functional conservation. We also demonstrate that the StART domain mediates Cv-c's function in the male gonad independently from the GAP domain's activity. This work demonstrates a role for DLC3/Cv-c in male gonadogenesis and highlights a novel StART domain mediated function required to organize the gonadal mesoderm and maintain its interaction with the germ cells during testis development.

Computational methods and theory for ion channel research

Ion channels are fundamental biological devices that act as gates in order to ensure selective ion transport across cellular membranes; their operation constitutes the molecular mechanism through which basic biological functions, such as nerve signal transmission and muscle contraction, are carried out. Here, we review recent results in the field of computational research on ion channels, covering theoretical advances, state-of-the-art simulation approaches, and frontline modeling techniques. We also report on few selected applications of continuum and atomistic methods to characterize the mechanisms of permeation, selectivity, and gating in biological and model channels.

Thermo- and Ion-responsive Silk-elastin-like Proteins and Their Multiscale Mechanisms

Silk-elastin-like protein (SELP) is an excellent biocompatible and biodegradable material for hydrogels with tunable properties that can respond to multiple external stimuli. By integrating fully atomistic, replica exchange molecular dynamics...

Sequence-complementarity dependent co-assembly of phosphodiester-linked aromatic donor–acceptor trimers

Sequence-defined trimers of phosphodiester-linked aromatic donor–acceptors self-assemble according to monomer order, and co-assemble into new structures with their complementary sequence.

Synaptotagmin-1 C2B domains cooperatively stabilize the fusion stalk via a master-servant mechanism

Synaptotagmin-1 is a low-affinity Ca2+ sensor that triggers synchronous vesicle fusion. It contains two similar C2 domains (C2A and C2B) that cooperate in membrane binding, being the C2B domain the...

Role of density and electrostatic interactions in the viscosity and non-newtonian behavior of ionic liquids – a molecular dynamics study

Both viscosity and the shear-thinning of ionic liquids are determined mainly by ionic interaction, with density having a secondary effect.

A Martini 3 coarse-grain model for the simulation of the photopolymerizable organic phase in dental composites

A coarse grain model of Bis-GMA and TEGDMA is presented and used to study aspects relevant to dental composites. It was found that the simulated polymerization shrinkage and double bond conversions are in good agreement with experimental data.

Improved Parameterization of Phosphatidylinositide Lipid Headgroups for the Martini 3 Coarse-Grain Force Field

Phosphoinositides are a family of membrane phospholipids that play crucial roles in membrane regulatory events. As such, these lipids are often a key part of molecular dynamics simulation studies of biological membranes, in particular of those employing coarse-grain models because of the potential long times and sizes of the involved membrane processes. Version 3 of the widely used Martini coarse-grain force field has been recently published, greatly refining many aspects of biomolecular interactions. In order to properly use it for lipid membrane simulations with phosphoinositides, we put forth the Martini 3-specific parameterization of inositol, phosphatidylinositol, and seven physiologically relevant phosphorylated derivatives of phosphatidylinositol. Compared to parameterizations for earlier Martini versions, focus was put on a more accurate reproduction of the behavior seen in both atomistic simulations and experimental studies, including the signaling-relevant phosphoinositide interaction with divalent cations. The models that we develop improve upon the conformational dynamics of phosphoinositides in the Martini force field and provide stable topologies at typical Martini time steps. They are able to reproduce experimentally known protein-binding poses as well as phosphoinositide aggregation tendencies. The latter was tested both in the presence and absence of calcium and included correct behavior of PI(4,5)P₂ calcium-induced clusters, which can be of relevance for regulation.

Explicit-pH Coarse-Grained Molecular Dynamics Simulations Enable Insights into Restructuring of Intestinal Colloidal Aggregates with Permeation Enhancers

Permeation enhancers (PEs) can increase the bioavailability of drugs. The mechanisms of action of these PEs are complex, but, typically, when used for oral administration, they can transiently induce the alteration of trans- and paracellular pathways, including increased solubilization and membrane fluidity, or the opening of the tight junctions. To elucidate these mechanistic details, it is important to understand the aggregation behavior of not only the PEs themselves but also other molecules already present in the intestine. Aggregation processes depend critically on, among other factors, the charge state of ionizable chemical groups, which is affected by the pH of the system. In this study, we used explicit-pH coarse-grained molecular dynamics simulations to investigate the aggregation behavior and pH dependence of two commonly used PEs—caprate and SNAC—together with other components of fasted- and fed-state simulated intestinal fluids. We also present and validate a coarse-grained molecular topology for the bile salt taurocholate suitable for the Martini3 force-field. Our results indicate an increase in the number of free molecules as a function of the system pH and for each combination of FaSSIF/FeSSIF and PEs. In addition, there are differences between caprate and SNAC, which are rationalized based on their different molecular structures and critical micelle concentrations.

Concluding remarks: peptide-membrane interactions

This article is based on the concluding remarks lecture given at the Faraday Discussion meeting on peptide-membrane interactions, held online, 8-10th September 2021.

Computer-aided design of molecular machines: techniques, paradigms and difficulties

With their development in the past decade, molecular machines, which achieve specific tasks by responding to external stimuli, have gradually come to be regarded as powerful tools for a wide range of applications, rather than interesting molecular toys. This conceptual change in turn motivates scientists to design molecular machines with complex architectures. Due to the lack of general principles bridging the functions and the chemical structures of molecular machines, experience-based design becomes difficult with the increase of size and complexity of the architectures. Computer-aided molecular-machine design, therefore, has attracted widespread attention on account of its ability to model and investigate complex molecular architectures without too much time and expense required for synthetic experiments. Using leading-edge numerical-simulation techniques, the mechanisms underlying achieving tasks through response to external stimuli of a large number of existing molecular machines have been successfully explored. Based on the experience of studying existing molecular machines, generalized methodologies of predicting the properties and working principles of molecular candidates have been established, paving the way for de novo computer-aided design of molecular machines. In this perspective, we introduce cutting-edge techniques that have been applied for investigating and designing molecular machines. We show paradigms of computer-aided design of molecular machines, which can serve as guidelines for the investigation of new supramolecular architectures. Moreover, we discuss the limitations and possible future developments of current techniques and methodologies in the field of computer-aided design of molecular machines.

Computational Approaches to Investigate and Design Lipid-binding Domains for Membrane Biosensing

Association of proteins with cellular membranes is critical for signaling and membrane trafficking processes. Many peripheral lipid-binding domains have been identified in the last few decades and have been investigated for their specific lipid-sensing properties using traditional in vivo and in vitro studies. However, several knowledge gaps remain owing to intrinsic limitations of these methodologies. Thus, novel approaches are necessary to further our understanding in lipid-protein biology. This review briefly discusses lipid-binding domains that act as specific lipid biosensors and provides a broad perspective on the computational approaches such as molecular dynamics (MD) simulations and machine learning (ML)-based techniques that can be used to study protein-membrane interactions. We also highlight the need for de novo design of proteins that elicit specific lipid-binding properties.

Multiscale simulations of protein and membrane systems

Classical multiscale simulations are perfectly suited to investigate biological soft matter systems. Owing to the bridging between microscopically realistic and lower-resolution models or the integration of a hierarchy of subsystems, one gets access to biologically relevant system sizes and timescales. In recent years, increasingly complex systems and processes have come into focus such as multidomain proteins, phase separation processes in biopolymer solutions, multicomponent biomembranes, or multiprotein complexes up to entire viruses. The review shows factors that have contributed to this progress - from improved models to machine-learning-based analysis and scale-bridging methods.