Collaborative Assessment of Molecular Geometries and Energies from the Open Force Field.

Force fields form the basis for classical molecular simulations, and their accuracy is crucial for the quality of, for instance, protein-ligand binding simulations in drug discovery. The huge diversity of small-molecule chemistry makes it a challenge to build and parameterize a suitable force field. The Open Force Field Initiative is a combined industry and academic consortium developing a state-of-the-art small-molecule force field. In this report, industry members of the consortium worked together to objectively evaluate the performance of the force fields (referred to here as OpenFF) produced by the initiative on a combined public and proprietary dataset of 19,653 relevant molecules selected from their internal research and compound collections. This evaluation was important because it was completely blind; at most partners, none of the molecules or data were used in force field development or testing prior to this work. We compare the Open Force Field "Sage" version 2.0.0 and "Parsley" version 1.3.0 with GAFF-2.11-AM1BCC, OPLS4, and SMIRNOFF99Frosst. We analyzed force-field-optimized geometries and conformer energies compared to reference quantum mechanical data. We show that OPLS4 performs best, and the latest Open Force Field release shows a clear improvement compared to its predecessors. The performance of established force fields such as GAFF-2.11 was generally worse. While OpenFF researchers were involved in building the benchmarking infrastructure used in this work, benchmarking was done entirely in-house within industrial organizations and the resulting assessment is reported here. This work assesses the force field performance using separate benchmarking steps, external datasets, and involving external research groups. This effort may also be unique in terms of the number of different industrial partners involved, with 10 different companies participating in the benchmark efforts.

Exploring protein-protein interactions using the site-identification by ligand competitive saturation methodology.

Protein docking methods are powerful computational tools to study protein-protein interactions (PPI). While a significant number of docking algorithms have been developed, they are usually based on rigid protein models or with limited considerations of protein flexibility and the desolvation effect is rarely considered in docking energy functions, which may lower the accuracy of the predictions. To address these issues, we introduce a PPI energy function based on the site-identification by ligand competitive saturation (SILCS) framework and utilize the fast Fourier transform (FFT) correlation approach. The free energy content of the SILCS FragMaps represent an alternative to traditional energy grids and they can be efficiently utilized to guide FFT-based protein docking. Application of the approach to eight diverse test cases, including seven from Protein Docking Benchmark 5.0, showed the PPI prediction using SILCS approach (SILCS-PPI) to be competitive with several commonly used protein docking methods indicating that the method has the ability to both qualitatively and quantitatively inform the prediction of PPI. Results show the utility of the SILCS-PPI docking approach for determination of probability distributions of PPI interactions over the surface of both partner proteins, allowing for identification of alternate binding poses. Such binding poses are confirmed by experimental crystal contacts in our test cases. While more computationally demanding than available PPI docking technologies, we anticipate that the SILCS-PPI docking approach will offer an alternative methodology for improved evaluation of PPIs that could be used in a variety of fields from systems biology to excipient design for biologics-based drugs.

Optimization and Evaluation of Site-Identification by Ligand Competitive Saturation (SILCS) as a Tool for Target-Based Ligand Optimization.

Chemical fragment cosolvent sampling techniques have become a versatile tool in ligand-protein binding prediction. Site-identification by ligand competitive saturation (SILCS) is one such method that maps the distribution of chemical fragments on a protein as free energy fields called FragMaps. Ligands are then simulated via Monte Carlo techniques in the field of the FragMaps (SILCS-MC) to predict their binding conformations and relative affinities for the target protein. Application of SILCS-MC using a number of different scoring schemes and MC sampling protocols against multiple protein targets was undertaken to evaluate and optimize the predictive capability of the method. Seven protein targets and 551 ligands with broad chemical variability were used to evaluate and optimize the model to maximize Pearson's correlation coefficient, Pearlman's predictive index, correct relative binding affinity, and root-mean-square error versus the absolute experimental binding affinities. Across the protein-ligand sets, the relative affinities of the ligands were predicted correctly an average of 69% of the time for the highest overall SILCS protocol. Training the FragMap weighting factors using a Bayesian machine learning (ML) algorithm led to an increase to an average 75% relative correct affinity predictions. Furthermore, once the optimal protocol is identified for a specific protein-ligand system average predictabilities of 76% are achieved. The ML algorithm is successful with small training sets of data (30 or more compounds) due to the use of physically correct FragMap weights as priors. Notably, the 76% correct relative prediction rate is similar to or better than free energy perturbation methods that are significantly computationally more expensive than SILCS. The results further support the utility of SILCS as a powerful and computationally accessible tool to support lead optimization and development in drug discovery.

Estimation of relative free energies of binding using pre-computed ensembles based on the single-step free energy perturbation and the site-identification by Ligand competitive saturation approaches.

Accurate and rapid estimation of relative binding affinities of ligand-protein complexes is a requirement of computational methods for their effective use in rational ligand design. Of the approaches commonly used, free energy perturbation (FEP) methods are considered one of the most accurate, although they require significant computational resources. Accordingly, it is desirable to have alternative methods of similar accuracy but greater computational efficiency to facilitate ligand design. In the present study relative free energies of binding are estimated for one or two non-hydrogen atom changes in compounds targeting the proteins ACK1 and p38 MAP kinase using three methods. The methods include standard FEP, single-step free energy perturbation (SSFEP) and the site-identification by ligand competitive saturation (SILCS) ligand grid free energy (LGFE) approach. Results show the SSFEP and SILCS LGFE methods to be competitive with or better than the FEP results for the studied systems, with SILCS LGFE giving the best agreement with experimental results. This is supported by additional comparisons with published FEP data on p38 MAP kinase inhibitors. While both the SSFEP and SILCS LGFE approaches require a significant upfront computational investment, they offer a 1000-fold computational savings over FEP for calculating the relative affinities of ligand modifications once those pre-computations are complete. An illustrative example of the potential application of these methods in the context of screening large numbers of transformations is presented. Thus, the SSFEP and SILCS LGFE approaches represent viable alternatives for actively driving ligand design during drug discovery and development. © 2016 Wiley Periodicals, Inc.

DIRECT-ID: An automated method to identify and quantify conformational variations--application to ß2 -adrenergic GPCR.

The conformational dynamics of a macromolecule can be modulated by a number of factors, including changes in environment, ligand binding, and interactions with other macromolecules, among others. We present a method that quantifies the differences in macromolecular conformational dynamics and automatically extracts the structural features responsible for these changes. Given a set of molecular dynamics (MD) simulations of a macromolecule, the norms of the differences in covariance matrices are calculated for each pair of trajectories. A matrix of these norms thus quantifies the differences in conformational dynamics across the set of simulations. For each pair of trajectories, covariance difference matrices are parsed to extract structural elements that undergo changes in conformational properties. As a demonstration of its applicability to biomacromolecular systems, the method, referred to as DIRECT-ID, was used to identify relevant ligand-modulated structural variations in the ß2 -adrenergic (ß2 AR) G-protein coupled receptor. Micro-second MD simulations of the ß2 AR in an explicit lipid bilayer were run in the apo state and complexed with the ligands: BI-167107 (agonist), epinephrine (agonist), salbutamol (long-acting partial agonist), or carazolol (inverse agonist). Each ligand modulated the conformational dynamics of ß2 AR differently and DIRECT-ID analysis of the inverse-agonist vs. agonist-modulated ß2 AR identified residues known through previous studies to selectively propagate deactivation/activation information, along with some previously unidentified ligand-specific microswitches across the GPCR. This study demonstrates the utility of DIRECT-ID to rapidly extract functionally relevant conformational dynamics information from extended MD simulations of large and complex macromolecular systems.

Pharmacophore modeling using site-identification by ligand competitive saturation (SILCS) with multiple probe molecules.

Receptor-based pharmacophore modeling is an efficient computer-aided drug design technique that uses the structure of the target protein to identify novel leads. However, most methods consider protein flexibility and desolvation effects in a very approximate way, which may limit their use in practice. The Site-Identification by Ligand Competitive Saturation (SILCS) assisted pharmacophore modeling protocol (SILCS-Pharm) was introduced recently to address these issues, as SILCS naturally takes both protein flexibility and desolvation effects into account by using full molecular dynamics simulations to determine 3D maps of the functional group-affinity patterns on a target receptor. In the present work, the SILCS-Pharm protocol is extended to use a wider range of probe molecules including benzene, propane, methanol, formamide, acetaldehyde, methylammonium, acetate and water. This approach removes the previous ambiguity brought by using water as both the hydrogen-bond donor and acceptor probe molecule. The new SILCS-Pharm protocol is shown to yield improved screening results, as compared to the previous approach based on three target proteins. Further validation of the new protocol using five additional protein targets showed improved screening compared to those using common docking methods, further indicating improvements brought by the explicit inclusion of additional feature types associated with the wider collection of probe molecules in the SILCS simulations. The advantage of using complementary features and volume constraints, based on exclusion maps of the protein defined from the SILCS simulations, is presented. In addition, reranking using SILCS-based ligand grid free energies is shown to enhance the diversity of identified ligands for the majority of targets. These results suggest that the SILCS-Pharm protocol will be of utility in rational drug design.

Mapping functional group free energy patterns at protein occluded sites: nuclear receptors and G-protein coupled receptors.

Occluded ligand-binding pockets (LBP) such as those found in nuclear receptors (NR) and G-protein coupled receptors (GPCR) represent a significant opportunity and challenge for computer-aided drug design. To determine free energies maps of functional groups of these LBPs, a Grand-Canonical Monte Carlo/Molecular Dynamics (GCMC/MD) strategy is combined with the Site Identification by Ligand Competitive Saturation (SILCS) methodology. SILCS-GCMC/MD is shown to map functional group affinity patterns that recapitulate locations of functional groups across diverse classes of ligands in the LBPs of the androgen (AR) and peroxisome proliferator-activated-γ (PPARγ) NRs and the metabotropic glutamate (mGluR) and ß2-adreneric (ß2AR) GPCRs. Inclusion of protein flexibility identifies regions of the binding pockets not accessible in crystal conformations and allows for better quantitative estimates of relative ligand binding affinities in all the proteins tested. Differences in functional group requirements of the active and inactive states of the ß2AR LBP were used in virtual screening to identify high efficacy agonists targeting ß2AR in Airway Smooth Muscle (ASM) cells. Seven of the 15 selected ligands were found to effect ASM relaxation representing a 46% hit rate. Hence, the method will be of use for the rational design of ligands in the context of chemical biology and the development of therapeutic agents.

Site-Identification by Ligand Competitive Saturation (SILCS) assisted pharmacophore modeling.

Database screening using receptor-based pharmacophores is a computer-aided drug design technique that uses the structure of the target molecule (i.e. protein) to identify novel ligands that may bind to the target. Typically receptor-based pharmacophore modeling methods only consider a single or limited number of receptor conformations and map out the favorable binding patterns in vacuum or with a limited representation of the aqueous solvent environment, such that they may suffer from neglect of protein flexibility and desolvation effects. Site-Identification by Ligand Competitive Saturation (SILCS) is an approach that takes into account these, as well as other, properties to determine 3-dimensional maps of the functional group-binding patterns on a target receptor (i.e. FragMaps). In this study, a method to use the FragMaps to automatically generate receptor-based pharmacophore models is presented. It converts the FragMaps into SILCS pharmacophore features including aromatic, aliphatic, hydrogen-bond donor and acceptor chemical functionalities. The method generates multiple pharmacophore hypotheses that are then quantitatively ranked using SILCS grid free energies. The pharmacophore model generation protocol is validated using three different protein targets, including using the resulting models in virtual screening. Improved performance and efficiency of the SILCS derived pharmacophore models as compared to published docking studies, as well as a recently developed receptor-based pharmacophore modeling method is shown, indicating the potential utility of the approach in rational drug design.

Estimation of ligand efficacies of metabotropic glutamate receptors from conformational forces obtained from molecular dynamics simulations.

Group 1 metabotropic glutamate receptors (mGluR) are G-protein coupled receptors with a large bilobate extracellular ligand binding region (LBR) that resembles a Venus fly trap. Closing of this LBR in the presence of a ligand is associated with the activation of the receptor. From conformational sampling of the LBR-ligand complexes using all-atom molecular dynamics (MD) simulations, we characterized the conformational minima related to the hinge like motion associated with the LBR closing/opening in the presence of known agonists and antagonists. By applying a harmonic restraint on the LBR, we also determined the conformational forces generated by the different ligands. The change in the location of the minima and the conformational forces were used to quantify the efficacies of the ligands. This analysis shows that efficacies can be estimated from the forces of a single conformation of the receptor, indicating the potential of MD simulations as an efficient and useful technique to quantify efficacies, thereby facilitating the rational design of mGluR agonists and antagonists.

Hysteresis-based mechanism for the directed motility of the Ncd motor.

Ncd is a Kinesin-14 family protein that walks to the microtubule's minus end. Although available structures show its α-helical neck in either pre- or post-stroke orientations, little is known about the transition between these two states. Using a combination of molecular dynamics simulations and structural analyses, we find that the neck sequentially makes intermediate contacts with the motor head along its mostly longitudinal path, and it develops a 24° twist in the post-stroke orientation. The forward (pre-stroke to post-stroke) motion has an ∼4.5 k(B)T (where k(B) is the Boltzmann constant, and T=300 K) free-energy barrier and is a diffusion guided by the intermediate contacts. The post-stroke free-energy minimum is higher and is formed ∼10° before reaching the orientation in the post-stroke crystal structure, consistent with previous structural data. The importance of intermediate contacts correlates with the existing motility data, including those for mutant Ncds. Unlike the forward motion, the recovery stroke goes nearly downhill in free energy, powered in part by torsional relaxation of the neck. The hysteresis in the energetics of the neck motion arises from the mechanical compliance of the protein, and together with guided diffusion, it may be key to the directed motility of Ncd.

Structure of the human marker of self 5-transmembrane receptor CD47.

CD47 is the only 5-transmembrane (5-TM) spanning receptor of the immune system. Its extracellular domain (ECD) is a cell surface marker of self that binds SIRPα and inhibits macrophage phagocytosis, and cancer immuno-therapy approaches in clinical trials are focused on blocking CD47/SIRPα interaction. We present the crystal structure of full length CD47 bound to the function-blocking antibody B6H12. CD47 ECD is tethered to the TM domain via a six-residue peptide linker (114RVVSWF119) that forms an extended loop (SWF loop), with the fundamental role of inserting the side chains of W118 and F119 into the core of CD47 extracellular loop region (ECLR). Using hydrogen-deuterium exchange and molecular dynamics simulations we show that CD47's ECLR architecture, comprised of two extracellular loops and the SWF loop, creates a molecular environment stabilizing the ECD for presentation on the cell surface. These findings provide insights into CD47 immune recognition, signaling and therapeutic intervention.

Identification and characterization of fragment binding sites for allosteric ligand design using the site identification by ligand competitive saturation hotspots approach (SILCS-Hotspots).

BACKGROUND: Fragment-based ligand design is used for the development of novel ligands that target macromolecules, most notably proteins. Central to its success is the identification of fragment binding sites that are spatially adjacent such that fragments occupying those sites may be linked to create drug-like ligands. Current experimental and computational approaches that address this problem typically identify only a limited number of sites as well as use a limited number of fragment types. METHODS: The site-identification by ligand competitive saturation (SILCS) approach is extended to the identification of fragment bindings sites, with the method termed SILCS-Hotspots. The approach involves precomputation of the SILCS FragMaps following which the identification of Hotspots, performed by identifying of all possible fragment binding sites on the full 3D structure of the protein followed by spatial clustering. RESULTS: The SILCS-Hotspots approach identifies a large number of sites on the target protein, including many sites not accessible in experimental structures due to low binding affinities and binding sites on the protein interior. The identified sites are shown to recapitulate the location of known drug-like molecules in both allosteric and orthosteric binding sites on seven proteins including the androgen receptor, the CDK2 and Erk5 kinases, PTP1B phosphatase and three GPCRs; the ß2-adrenergic, GPR40 fatty-acid binding and M2-muscarinic receptors. Analysis indicates the importance of considering all possible fragment binding sites, and not just those accessible to experimental methods, when identifying novel binding sites and performing ligand design versus just considering the most favorable sites. The approach is shown to identify a larger number of known binding sites of drug-like molecules versus the commonly used FTMap and Fpocket methods. GENERAL SIGNIFICANCE: The present results indicate the potential utility of the SILCS-Hotspots approach for fragment-based rational design of ligands, including allosteric modulators.

Determination of Ionic Hydration Free Energies with Grand Canonical Monte Carlo/Molecular Dynamics Simulations in Explicit Water.

Grand canonical Monte Carlo (GCMC) simulations of ionic solutions with explicit solvent models are known to be challenging. One challenge arises from the treatment of long-range electrostatics and finite-box size in Monte Carlo simulations when periodic boundary condition and Ewald summation methods are used. Another challenge is that constant excess chemical potential GCMC simulations for charged solutes suffer from inadequate insertion and deletion acceptance ratios. In this work, we address those problems by implementing an oscillating excess chemical potential GCMC algorithm with smooth particle mesh Ewald and finite-box-size corrections to treat the long-range electrostatics. The developed GCMC simulation program was combined with GROMACS to perform GCMC/MD simulations of ionic solutions individually containing Li+, Na+, K+, Rb+, Cs+, F-, Cl-, Br-, I-, Ca2+, and Mg2+, respectively. Our simulation results show that the combined GCMC/MD approach can approximate the ionic hydration free energies with proper treatment of long-range electrostatics. Our developed simulation approach can open up new avenues for simulating complex chemical and biomolecular systems and for drug discovery.

Characterization of Mg2+ Distributions around RNA in Solution.

Binding of metal ions is an important factor governing the folding and dynamics of RNA. Shielding of charges in the polyanionic backbone allows RNA to adopt a diverse range of folded structures that give rise to their many functions within the cell. Some RNA sequences fold only in the presence of Mg2+, which may be bound via direct interactions or occupy the more diffuse "ion atmosphere" around the RNA. To understand the driving forces for RNA folding, it is important to be able to fully characterize the distribution of metal ions around the RNA. In this work, a combined Grand Canonical Monte Carlo-Molecular Dynamics (GCMC-MD) method is applied to characterize Mg2+ distributions around folded RNA structures. The GCMC-MD approach identifies known inner- and outer-shell Mg2+ coordination, while also predicting new regions occupied by Mg2+ that are not observed in crystal structures but that may be relevant in solution, including the case of the Mg2+ riboswitch, for which alternate Mg2+ binding sites may have implications for its function. This work represents a significant step forward in establishing a structural and thermodynamic description of RNA-Mg2+ interactions and their role in RNA structure and function.

Conformational Heterogeneity of Intracellular Loop 3 of the µ-opioid G-protein Coupled Receptor.

G-protein coupled receptors (GPCRs), including the µ-opioid receptor, interact with G-proteins and other proteins via their intracellular face as required for signal transduction. However, characterization of the structure of the intracellular face of GPCRs is complicated by the experimental methods used for structural characterization. In the present study we undertook a series of long-time molecular dynamics (MD) simulations, ranging from 1 to 5 µs, on the µ-opioid receptor in both the dimeric and monomeric states. Results show intracellular loop 2 (ICL2) to sample an equilibrium between coiled and helical states. Intracellular loop 3 (ICL3) samples a wider range of conformations. Previously unobserved ß-sheet structures were primarily sampled in the simulations initiated from the inactive dimer conformation. In contrast, helical structures were sampled in simulations initiated from the active, monomer conformation. Notably, in the dimeric form of the receptor, both intramolecular and intermolecular ß-sheet structures were sampled, with the latter occurring between the two monomers. These results indicate that the sampling of ß-sheet structures can maintain the ICL3 in an inactive conformation that contributes to stabilization of the dimeric form of the receptor via interchain ß-sheet structures.

Sampling of Organic Solutes in Aqueous and Heterogeneous Environments Using Oscillating Excess Chemical Potentials in Grand Canonical-like Monte Carlo-Molecular Dynamics Simulations.

Solute sampling of explicit bulk-phase aqueous environments in grand canonical (GC) ensemble simulations suffer from poor convergence due to low insertion probabilities of the solutes. To address this, we developed an iterative procedure involving Grand Canonical-like Monte Carlo (GCMC) and molecular dynamics (MD) simulations. Each iteration involves GCMC of both the solutes and water followed by MD, with the excess chemical potential (µex) of both the solute and the water oscillated to attain their target concentrations in the simulation system. By periodically varying the µex of the water and solutes over the GCMC-MD iterations, solute exchange probabilities and the spatial distributions of the solutes improved. The utility of the oscillating-µex GCMC-MD method is indicated by its ability to approximate the hydration free energy (HFE) of the individual solutes in aqueous solution as well as in dilute aqueous mixtures of multiple solutes. For seven organic solutes: benzene, propane, acetaldehyde, methanol, formamide, acetate, and methylammonium, the average µex of the solutes and the water converged close to their respective HFEs in both 1 M standard state and dilute aqueous mixture systems. The oscillating-µex GCMC methodology is also able to drive solute sampling in proteins in aqueous environments as shown using the occluded binding pocket of the T4 lysozyme L99A mutant as a model system. The approach was shown to satisfactorily reproduce the free energy of binding of benzene as well as sample the functional group requirements of the occluded pocket consistent with the crystal structures of known ligands bound to the L99A mutant as well as their relative binding affinities.

Critical buckling length versus persistence length: what governs biofilament conformation?

We numerically study the length dependence in the bending stiffness of alpha helices, coiled coils, and a linear chain model. As the length approaches what we named the critical buckling length lc, the chain appears to become increasingly more flexible. This is due to weak nonbonded attractions that eventually lead to buckling instability and alter the chain's conformational ensemble. For alpha helices and coiled coils, lc is much less than their respective persistence lengths, so lc is the defining length scale for their conformations. These results elucidate the importance of weak nonspecific attractions that are inherent in many biofilaments in physiological conditions.

Modulation of elasticity in functionally distinct domains of the tropomyosin coiled-coil.

Alpha-helical coiled-coils are common protein structural motifs. Whereas vast information is available regarding their structure, folding, and stability, far less is known about their elastic properties, even though they play mechanical roles in many cases such as tropomyosin in muscle contraction or neck stalks of kinesin or myosin motor proteins. Using computer simulations, we characterized elastic properties of coiled-coils, either globally or locally. Global bending stiffness of standard leucine zipper coiled-coils was calculated using normal mode analysis. Mutations in hydrophobic residues involved in the knob-into-hole interface between the two alpha-helices affect elasticity significantly, whereas charged side chains forming inter-helical salt bridges do not. This suggests that coiled-coils with less regular heptad periodicity may have regional variations in flexibility. We show this by the flexibility map of tropomyosin, which was constructed by a local fluctuation analysis. Overall, flexibility varies by more than twofold and increases towards the C-terminal region of the molecule. Describing the coiled-coil as a twisted tape, it is generally more flexible in the splay bending than in the bending of the broad face. Actin binding sites in alpha zones show local rigidity minima. Broken core regions due to acidic residues at the hydrophobic face such as the Asp137 and the Glu218 are found to be the most labile with moduli for splay and broad face bending as 70 nm and 116 nm respectively. Such variation in flexibility could be relevant to the tropomyosin function, especially for moving across the non-uniform surface of F-actin to regulate myosin binding.