We present a supercomputer-driven pipeline for in silico drug discovery using enhanced sampling molecular dynamics (MD) and ensemble docking. Ensemble docking makes use of MD results by docking compound databases into representative protein binding-site conformations, thus taking into account the dynamic properties of the binding sites. We also describe preliminary results obtained for 24 systems involving eight proteins of the proteome of SARS-CoV-2. The MD involves temperature replica exchange enhanced sampling, making use of massively parallel supercomputing to quickly sample the configurational space of protein drug targets. Using the Summit supercomputer at the Oak Ridge National Laboratory, more than 1 ms of enhanced sampling MD can be generated per day. We have ensemble docked repurposing databases to 10 configurations of each of the 24 SARS-CoV-2 systems using AutoDock Vina. Comparison to experiment demonstrates remarkably high hit rates for the top scoring tranches of compounds identified by our ensemble approach. We also demonstrate that, using Autodock-GPU on Summit, it is possible to perform exhaustive docking of one billion compounds in under 24 h. Finally, we discuss preliminary results and planned improvements to the pipeline, including the use of quantum mechanical (QM), machine learning, and artificial intelligence (AI) methods to cluster MD trajectories and rescore docking poses.

BackgroundThe conversion of plant biomass to ethanol via enzymatic cellulose hydrolysis offers a potentially sustainable route to biofuel production. However, the inhibition of enzymatic activity in pretreated biomass by lignin severely limits the efficiency of this process.ResultsBy performing atomic-detail molecular dynamics simulation of a biomass model containing cellulose, lignin, and cellulases (TrCel7A), we elucidate detailed lignin inhibition mechanisms. We find that lignin binds preferentially both to the elements of cellulose to which the cellulases also preferentially bind (the hydrophobic faces) and also to the specific residues on the cellulose-binding module of the cellulase that are critical for cellulose binding of TrCel7A (Y466, Y492, and Y493).Conclusions Lignin thus binds exactly where for industrial purposes it is least desired, providing a simple explanation of why hydrolysis yields increase with lignin removal.Electronic supplementary materialThe online version of this article (doi:10.1186/s13068-015-0379-8) contains supplementary material, which is available to authorized users.

Vaccines derived from chimpanzee adenovirus Y25 (ChAdOx1), human adenovirus type 26 (HAdV-D26), and human adenovirus type 5 (HAdV-C5) are critical in combatting the severe acute respiratory coronavirus 2 (SARS-CoV-2) pandemic. As part of the largest vaccination campaign in history, ultrarare side effects not seen in phase 3 trials, including thrombosis with thrombocytopenia syndrome (TTS), a rare condition resembling heparin-induced thrombocytopenia (HIT), have been observed. This study demonstrates that all three adenoviruses deployed as vaccination vectors versus SARS-CoV-2 bind to platelet factor 4 (PF4), a protein implicated in the pathogenesis of HIT. We have determined the structure of the ChAdOx1 viral vector and used it in state-of-the-art computational simulations to demonstrate an electrostatic interaction mechanism with PF4, which was confirmed experimentally by surface plasmon resonance. These data confirm that PF4 is capable of forming stable complexes with clinically relevant adenoviruses, an important step in unraveling the mechanisms underlying TTS.

Lignin is an abundant aromatic polymer found in plant secondary cell walls. In recent years, lignin has attracted renewed interest as a feedstock for bio-based chemicals via catalytic and biological approaches and has emerged as a target for genetic engineering to improve lignocellulose digestibility by altering its composition. In lignin biosynthesis and microbial conversion, small phenolic lignin precursors or degradation products cross membrane bilayers through an unidentified translocation mechanism prior to incorporation into lignin polymers (synthesis) or catabolism (bioconversion), with both passive and transporter-assisted mechanisms postulated. To test the passive permeation potential of these phenolics, we performed molecular dynamics simulations for 69 monomeric and dimeric lignin-related phenolics with 3 model membranes to determine the membrane partitioning and permeability coefficients for each compound. The results support an accessible passive permeation mechanism for most compounds, including monolignols, dimeric phenolics, and the flavonoid, tricin. Computed lignin partition coefficients are consistent with concentration enrichment near lipid carbonyl groups, and permeability coefficients are sufficient to keep pace with cellular metabolism. Interactions between methoxy and hydroxy groups are found to reduce membrane partitioning and improve permeability. Only carboxylate-modified or glycosylated lignin phenolics are predicted to require transporters for membrane translocation. Overall, the results suggest that most lignin-related compounds can passively traverse plant and microbial membranes on timescales commensurate with required biological activities, with any potential transport regulation mechanism in lignin synthesis, catabolism, or bioconversion requiring compound functionalization.

Slow diffusion of the lipids in conventional all-atom simulations of membrane systems makes it difficult to sample large rearrangements of lipids and protein-lipid interactions. Recently, Tajkhorshid and co-workers developed the highly mobile membrane-mimetic (HMMM) model with accelerated lipid motion by replacing the lipid tails with small organic molecules. The HMMM model provides accelerated lipid diffusion by one to two orders of magnitude, and is particularly useful in studying membrane-protein associations. However, building an HMMM simulation system is not easy, as it requires sophisticated treatment of the lipid tails. In this study, we have developed CHARMM-GUI HMMM Builder (http://www.charmm-gui.org/input/hmmm) to provide users with ready-to-go input files for simulating HMMM membrane systems with/without proteins. Various lipid-only and protein-lipid systems are simulated to validate the qualities of the systems generated by HMMM Builder with focus on the basic properties and advantages of the HMMM model. HMMM Builder supports all lipid types available in CHARMM-GUI and also provides a module to convert back and forth between an HMMM membrane and a full-length membrane. We expect HMMM Builder to be a useful tool in studying membrane systems with enhanced lipid diffusion.

Energetics of protein side chain partitioning between aqueous solution and cellular membranes is of fundamental importance for correctly capturing the membrane binding and specific protein–lipid interactions of peripheral membrane proteins. We recently reported a highly mobile membrane mimetic (HMMM) model that accelerates lipid dynamics by modeling the membrane interior partially as a fluid organic solvent while retaining a literal description of the lipid head groups and the beginning of the tails. While the HMMM has been successfully applied to study spontaneous insertion of a number of peripheral proteins into membranes, a quantitative characterization of the energetics of membrane–protein interactions in HMMM membranes has not been performed. We report here the free energy profiles for partitioning of 10 protein side chain analogues into a HMMM membrane. In the interfacial and headgroup regions of the membrane, the side chain free energy profiles show excellent agreement with profiles previously reported for conventional membranes with full-tail lipids. In regions where the organic solvent is prevalent, the increased dipole and fluidity of the solvent generally result in a less accurate description, most notably overstabilization of aromatic and polar amino acids. As an additional measure of the ability of the HMMM model to describe membrane–protein interactions, the water-to-membrane interface transfer energies were analyzed and found to be in agreement with the previously reported experimental and computational hydrophobicity scales. We discuss strengths and weaknesses of HMMM in describing protein–membrane interactions as well as further development of model membranes.

Lignin is an abundant aromatic heteropolymer found in secondary plant cell walls and is a potential feedstock for conversion into bioderived fuels and chemicals. Lignin chemical diversity complicates traditional structural studies, and so, relatively little experimental evidence exists for how lignin structure exists in aqueous solution or how lignin polymers respond to changes in their chemical environment. Molecular modeling can address these concerns; however, prior computational structural lignin models typically did not capture lignin heterogeneity, as only a few polymers were considered. LigninBuilder creates a framework for building structural libraries for lignin from existing topological libraries, permitting significantly greater diversity of lignin structures to be sampled at atomic detail. As a demonstration of its capabilities, LigninBuilder was applied to three libraries of lignin from hardwood, softwood, and grass, and the resulting polymer structures were simulated in an aqueous environment. The lignins adopted compact globular structures, as would be expected for polymers in poor solvents. The differences between the libraries were largest when quantifying the packing of nonadjacent aromatic residues, with greater branching within the polymer resulting in poorer aromatic packing. Individual lignin polymers were also found to undergo rapid conformational changes, with the dwell time within a state growing as the square of the molecular weight. This first application of LigninBuilder demonstrates the potential for atomic-level insight into lignin interactions. LigninBuilder is distributed as a plugin to the visualization software VMD, lowering the barrier for modeling lignin structure in diverse environments.

Lignin solvolysis from the plant cell wall is the critical first step in lignin depolymerization processes involving whole biomass feedstocks. However, little is known about the coupled reaction kinetics and transport phenomena that govern the effective rates of lignin extraction. Here, we report a validated simulation framework that determines intrinsic, transport‐independent kinetic parameters for the solvolysis of lignin, hemicellulose, and cellulose upon incorporation of feedstock characteristics for the methanol‐based extraction of poplar as an example fractionation process. Lignin fragment diffusion is predicted to compete on the same time and length scales as reactions of lignin within cell walls and longitudinal pores of typical milled particle sizes, and mass transfer resistances are predicted to dominate the solvolysis of poplar particles that exceed approximately 2 mm in length. Beyond the approximately 2 mm threshold, effectiveness factors are predicted to be below 0.25, which implies that pore diffusion resistances may attenuate observable kinetic rate measurements by at least 75 % in such cases. Thus, researchers are recommended to conduct kinetic evaluations of lignin‐first catalysts using biomass particles smaller than approximately 0.2 mm in length to avoid feedstock‐specific mass transfer limitations in lignin conversion studies. Overall, this work highlights opportunities to improve lignin solvolysis by genetic engineering and provides actionable kinetic information to guide the design and scale‐up of emerging biorefinery strategies.