Structure-Based Rational Design of Constrained Peptides as TIM-3 Inhibitors.

Blocking the immunosuppressive function of T-cell immunoglobulin mucin-3 (TIM-3) is an established therapeutic strategy to maximize the efficacy of immune checkpoint inhibitors for cancer immunotherapy. Currently, effective inhibition of TIM-3 interactions relies on monoclonal antibodies (mAbs), which come with drawbacks such as immunogenicity risk, limited tumor penetration, and high manufacturing costs. Guided by the X-ray cocrystal structures of TIM-3 with mAbs, we report an in silico structure-based rational design of constrained peptides as potent TIM-3 inhibitors. The top cyclic peptide from our study (P2) binds TIM-3 with a K D value of 166.3 ± 12.1 nM as determined by surface plasmon resonance (SPR) screening. Remarkably, P2 efficiently inhibits key TIM-3 interactions with natural TIM-3 ligands at submicromolar concentrations in a panel of cell-free and cell-based assays. The capacity of P2 to reverse immunosuppression in T-cell/cancer cell cocultures, coupled with favorable in vitro pharmacokinetic properties, highlights the potential of P2 for further evaluation in preclinical models of immuno-oncology.

Free Energy Simulations of Receptor-Binding Domain Opening of the SARS-CoV-2 Spike Indicate a Barrierless Transition with Slow Conformational Motions.

Infection by sarbecoviruses begins with the attachment of the homotrimeric viral "spike" protein to the angiotensin-converting enzyme 2 receptor on the surface of mammalian cells. This requires one or more receptor-binding domains (RBDs) to be in the open (up) position. Here, we present the results of long molecular dynamics simulations with umbrella sampling (US) to compute a one-dimensional free energy profile of RBD opening/closing and the associated transition times. After ≃3.58µs of simulation time per US window (∼229 µs in total), which was required to approach trajectory decorrelation, the computed free energy profile was found to be without large barriers. This suggests that the RBD diffuses between the open and closed positions without significant energetic hindrance. This interpretation appears consistent with experiments but is at odds with some previous simulations. Modeling the RBD motion as diffusive dynamics along the computed free energy profile, we find that the overall time required for the transition is only about 2 µs, which is 5 orders of magnitude shorter than experimentally measured transition times. We speculate that the most likely reason for the transition time mismatch is our use of very short glycans, which was required to make the simulations performed here feasible. Despite the long simulation times, the final free energy profile is not fully converged with statistical errors of ≃1.16 kcal/mol, which were found to be consistent with the slow time decay in the autocorrelation of the conformational motions of the protein. The simulation lengths that would be required to obtain fully converged results remain unknown, but the present calculations would benefit from at least an order-of-magnitude extension.

Molecular Simulations of Conformational Transitions within the Insulin Receptor Kinase Reveal Consensus Features in a Multistep Activation Pathway.

Modulating the transitions between active and inactive conformations of protein kinases is the primary means of regulating their catalytic activity, achieved by phosphorylation of the activation loop (A-loop). To elucidate the mechanism of this conformational activation, we applied the string method to determine the conformational transition path of insulin receptor kinase between the active and inactive conformations and the corresponding free-energy profiles with and without A-loop phosphorylation. The conformational change was found to proceed in three sequential steps: first, the flipping of the DFG motif of the active site; second, rotation of the A-loop; finally, the inward movement of the αC helix. The main energetic bottleneck corresponds to the conformational change in the A-loop, while changes in the DFG motif and αC helix occur before and after A-loop conformational change, respectively. In accordance with this, two intermediate states are identified, the first state just after the DFG flipping and the second state after the A-loop rotation. These intermediates exhibit structural features characteristic of the corresponding inactive and active conformations of other protein kinases. To understand the impact of A-loop phosphorylation on kinase conformation, the free energies of A-loop phosphorylation were determined at several states along the conformational transition path using the free-energy perturbation simulations. The calculated free energies reveal that while the unphosphorylated kinase interconverts between the inactive and active conformations, A-loop phosphorylation restricts access to the inactive conformation, thereby increasing the active conformation population. Overall, this study suggests a consensus mechanism of conformational activation between different protein kinases.

ppdx: Automated modeling of protein-protein interaction descriptors for use with machine learning.

This paper describes ppdx, a python workflow tool that combines protein sequence alignment, homology modeling, and structural refinement, to compute a broad array of descriptors for characterizing protein-protein interactions. The descriptors can be used to predict various properties of interest, such as protein-protein binding affinities, or inhibitory concentrations (IC50 ), using approaches that range from simple regression to more complex machine learning models. The software is highly modular. It supports different protocols for generating structures, and 95 descriptors can be currently computed. More protocols and descriptors can be easily added. The implementation is highly parallel and can fully exploit the available cores in a single workstation, or multiple nodes on a supercomputer, allowing many systems to be analyzed simultaneously. As an illustrative application, ppdx is used to parametrize a model that predicts the IC50 of a set of antigens and a class of antibodies directed to the influenza hemagglutinin stalk.

On the Rapid Calculation of Binding Affinities for Antigen and Antibody Design and Affinity Maturation Simulations.

The accurate and efficient calculation of protein-protein binding affinities is an essential component in antibody and antigen design and optimization, and in computer modeling of antibody affinity maturation. Such calculations remain challenging despite advances in computer hardware and algorithms, primarily because proteins are flexible molecules, and thus, require explicit or implicit incorporation of multiple conformational states into the computational procedure. The astronomical size of the amino acid sequence space further compounds the challenge by requiring predictions to be computed within a short time so that many sequence variants can be tested. In this study, we compare three classes of methods for antibody/antigen (Ab/Ag) binding affinity calculations: (i) a method that relies on the physical separation of the Ab/Ag complex in equilibrium molecular dynamics (MD) simulations, (ii) a collection of 18 scoring functions that act on an ensemble of structures created using homology modeling software, and (iii) methods based on the molecular mechanics-generalized Born surface area (MM-GBSA) energy decomposition, in which the individual contributions of the energy terms are scaled to optimize agreement with the experiment. When applied to a set of 49 antibody mutations in two Ab/HIV gp120 complexes, all of the methods are found to have modest accuracy, with the highest Pearson correlations reaching about 0.6. In particular, the most computationally intensive method, i.e., MD simulation, did not outperform several scoring functions. The optimized energy decomposition methods provided marginally higher accuracy, but at the expense of requiring experimental data for parametrization. Within each method class, we examined the effect of the number of independent computational replicates, i.e., modeled structures or reinitialized MD simulations, on the prediction accuracy. We suggest using about ten modeled structures for scoring methods, and about five simulation replicates for MD simulations as a rule of thumb for obtaining reasonable convergence. We anticipate that our study will be a useful resource for practitioners working to incorporate binding affinity calculations within their protein design and optimization process.

Multiscale affinity maturation simulations to elicit broadly neutralizing antibodies against HIV.

The design of vaccines against highly mutable pathogens, such as HIV and influenza, requires a detailed understanding of how the adaptive immune system responds to encountering multiple variant antigens (Ags). Here, we describe a multiscale model of B cell receptor (BCR) affinity maturation that employs actual BCR nucleotide sequences and treats BCR/Ag interactions in atomistic detail. We apply the model to simulate the maturation of a broadly neutralizing Ab (bnAb) against HIV. Starting from a germline precursor sequence of the VRC01 anti-HIV Ab, we simulate BCR evolution in response to different vaccination protocols and different Ags, which were previously designed by us. The simulation results provide qualitative guidelines for future vaccine design and reveal unique insights into bnAb evolution against the CD4 binding site of HIV. Our model makes possible direct comparisons of simulated BCR populations with results of deep sequencing data, which will be explored in future applications.

A Coarse-Grained Model of Affinity Maturation Indicates the Importance of B-Cell Receptor Avidity in Epitope Subdominance.

The elicitation of broadly neutralizing antibodies (bnAbs) is a major goal in the design of vaccines against rapidly-mutating viruses. In the case of influenza, many bnAbs that target conserved epitopes on the stem of the hemagglutinin protein (HA) have been discovered. However, these antibodies are rare, are not boosted well upon reinfection, and often have low neutralization potency, compared to strain-specific antibodies directed to the HA head. Different hypotheses have been proposed to explain this phenomenon. We use a coarse-grained computational model of the germinal center reaction to investigate how B-cell receptor binding valency affects the growth and affinity maturation of competing B-cells. We find that receptors that are unable to bind antigen bivalently, and also those that do not bind antigen cooperatively, have significantly slower rates of growth, memory B-cell production, and, under certain conditions, rates of affinity maturation. The corresponding B-cells are predicted to be outcompeted by B-cells that bind bivalently and cooperatively. We use the model to explore strategies for a universal influenza vaccine, e.g., how to boost the concentrations of the slower growing cross-reactive antibodies directed to the stem. The results suggest that, upon natural reinfections subsequent to vaccination, the protectiveness of such vaccines would erode, possibly requiring regular boosts. Collectively, our results strongly support the importance of bivalent antibody binding in immunodominance, and suggest guidelines for developing a universal influenza vaccine.

Molecular Simulation of Stapled Peptides.

Constrained peptides represent a relatively new class of biologic therapeutics, which have the potential to overcome several limitations of small-molecule drugs, and of designed antibodies. Because of their modest size, the rational design of such peptides is becoming increasingly amenable to computer simulation; multi-microsecond molecular dynamic (MD) simulations are now routinely possible on consumer-grade graphical processors (GPUs). Here, we describe the procedures for performing and analyzing MD simulations of hydrocarbon-stapled peptides using the CHARMM energy function, in isolation and in complex with a binding partner, to investigate their conformational properties and to compute changes in their binding affinity upon mutation.

Dynamic Connection between Enzymatic Catalysis and Collective Protein Motions.

Enzymes employ a wide range of protein motions to achieve efficient catalysis of chemical reactions. While the role of collective protein motions in substrate binding, product release, and regulation of enzymatic activity is generally understood, their roles in catalytic steps per se remain uncertain. Here, molecular dynamics simulations, enzyme kinetics, X-ray crystallography, and nuclear magnetic resonance spectroscopy are combined to elucidate the catalytic mechanism of adenylate kinase and to delineate the roles of catalytic residues in catalysis and the conformational change in the enzyme. This study reveals that the motions in the active site, which occur on a time scale of picoseconds to nanoseconds, link the catalytic reaction to the slow conformational dynamics of the enzyme by modulating the free energy landscapes of subdomain motions. In particular, substantial conformational rearrangement occurs in the active site following the catalytic reaction. This rearrangement not only affects the reaction barrier but also promotes a more open conformation of the enzyme after the reaction, which then results in an accelerated opening of the enzyme compared to that of the reactant state. The results illustrate a linkage between enzymatic catalysis and collective protein motions, whereby the disparate time scales between the two processes are bridged by a cascade of intermediate-scale motion of catalytic residues modulating the free energy landscapes of the catalytic and conformational change processes.

A restrained locally enhanced sampling method (RLES) for finding free energy minima in complex systems.

We present an extension of the locally enhanced sampling method. A restraint potential is introduced to drive the many-replica system to the canonical ensemble corresponding to the physical, single-replica system. Convergence properties are demonstrated using a model rugged two-dimensional potential, for which sampling by conventional equilibrium molecular dynamics is inefficient. Restrained locally enhanced sampling (RLES) is found to explore the space of configurations with an efficiency comparable to that of temperature replica exchange. To demonstrate the potential of RLES for realistic applications, the method is used to fold the 12-residue tryptophan zipper miniprotein in explicit solvent. The RLES algorithm can be incorporated into existing LES implementations with minor code modifications.

The trajectory of intrahelical lesion recognition and extrusion by the human 8-oxoguanine DNA glycosylase.

Efficient search for DNA damage embedded in vast expanses of the DNA genome presents one of the greatest challenges to DNA repair enzymes. We report here crystal structures of human 8-oxoguanine (oxoG) DNA glycosylase, hOGG1, that interact with the DNA containing the damaged base oxoG and the normal base G while they are nested in the DNA helical stack. The structures reveal that hOGG1 engages the DNA using different protein-DNA contacts from those observed in the previously determined lesion recognition complex and other hOGG1-DNA complexes. By applying molecular dynamics simulations, we have determined the pathways taken by the lesion and normal bases when extruded from the DNA helix and their associated free energy profiles. These results reveal how the human oxoG DNA glycosylase hOGG1 locates the lesions inside the DNA helix and facilitates their extrusion for repair.

Microsecond Molecular Dynamics Simulations of Proteins Using a Quasi-Equilibrium Solvation Shell Model.

We describe the development and implementation of a quasi-equilibrium hydration shell model of biomolecular solvation with adaptive boundaries. Applying the model to microsecond-long molecular dynamics simulations of several protein systems of varying complexity, we find that the model simulation results are of comparable quality to those obtained from simulations of fully solvated systems, but at a reduced computational cost. We discuss the dominant sources of error in the model and outline directions for future improvements.

Structure of the EmrE multidrug transporter and its use for inhibitor peptide design.

Small multidrug resistance (SMR) pumps represent a minimal paradigm of proton-coupled membrane transport in bacteria, yet no high-resolution structure of an SMR protein is available. Here, atomic-resolution structures of the Escherichia coli efflux-multidrug resistance E (EmrE) multidrug transporter in ligand-bound form are refined using microsecond molecular dynamics simulations biased using low-resolution data from X-ray crystallography. The structures are compatible with existing mutagenesis data as well as NMR and biochemical experiments, including pKas of the catalytic glutamate residues and the dissociation constant ([Formula: see text]) of the tetraphenylphosphonium+ cation. The refined structures show the arrangement of residue side chains in the EmrE active site occupied by two different ligands and in the absence of a ligand, illustrating how EmrE can adopt structurally diverse active site configurations. The structures also show a stable, well-packed binding interface between the helices H4 of the two monomers, which is believed to be crucial for EmrE dimerization. Guided by the atomic details of this interface, we design proteolysis-resistant stapled peptides that bind to helix H4 of an EmrE monomer. The peptides are expected to interfere with the dimerization and thereby inhibit drug transport. Optimal positions of the peptide staple were determined using free-energy simulations of peptide binding to monomeric EmrE Three of the four top-scoring peptides selected for experimental testing resulted in significant inhibition of proton-driven ethidium efflux in live cells without nonspecific toxicity. The approach described here is expected to be of general use for the design of peptide therapeutics.

Cytosolic expression, solution structures, and molecular dynamics simulation of genetically encodable disulfide-rich de novo designed peptides.

Disulfide-rich peptides represent an important protein family with broad pharmacological potential. Recent advances in computational methods have made it possible to design new peptides which adopt a stable conformation de novo. Here, we describe a system to produce disulfide-rich de novo peptides using Escherichia coli as the expression host. The advantage of this system is that it enables production of uniformly 13 C- and 15 N-labeled peptides for solution nuclear magnetic resonance (NMR) studies. This expression system was used to isotopically label two previously reported de novo designed peptides, and to determine their solution structures using NMR. The ensemble of NMR structures calculated for both peptides agreed well with the design models, further confirming the accuracy of the design protocol. Collection of NMR data on the peptides under reducing conditions revealed a dependency on disulfide bonds to maintain stability. Furthermore, we performed long-time molecular dynamics (MD) simulations with tempering to assess the stability of two families of de novo designed peptides. Initial designs which exhibited a stable structure during simulations were more likely to adopt a stable structure in vitro, but attempts to utilize this method to redesign unstable peptides to fold into a stable state were unsuccessful. Further work is therefore needed to assess the utility of MD simulation techniques for de novo protein design.

Role of framework mutations and antibody flexibility in the evolution of broadly neutralizing antibodies.

Eliciting antibodies that are cross reactive with surface proteins of diverse strains of highly mutable pathogens (e.g., HIV, influenza) could be key for developing effective universal vaccines. Mutations in the framework regions of such broadly neutralizing antibodies (bnAbs) have been reported to play a role in determining their properties. We used molecular dynamics simulations and models of affinity maturation to study specific bnAbs against HIV. Our results suggest that there are different classes of evolutionary lineages for the bnAbs. If germline B cells that initiate affinity maturation have high affinity for the conserved residues of the targeted epitope, framework mutations increase antibody rigidity as affinity maturation progresses to evolve bnAbs. If the germline B cells exhibit weak/moderate affinity for conserved residues, an initial increase in flexibility via framework mutations may be required for the evolution of bnAbs. Subsequent mutations that increase rigidity result in highly potent bnAbs. Implications of our results for immunogen design are discussed.

Coexisting origins of subdiffusion in internal dynamics of proteins.

Subdiffusion in conformational dynamics of proteins is observed both experimentally and in simulations. Although its origin has been attributed to multiple mechanisms, including trapping on a rugged energy landscape, fractional Brownian noise, or a fractal topology of the energy landscape, it is unclear which of these, if any, is most relevant. To obtain insights into the actual mechanism, we introduce an analytically tractable hierarchical trapping model and apply it to molecular dynamics simulation trajectories of three proteins in solution. The analysis of the simulations introduces a subdiffusive exponent that varies with time and associates plateaus in the mean-squared displacement with traps on the energy landscape. This analysis permits us to separate the component of subdiffusion due to a trapping mechanism from that due to an underlying fluctuating process, such as fractional Brownian motion. The present results thus provide insights concerning the physical origin of subdiffusion in the dynamics of proteins.

Regulation and Plasticity of Catalysis in Enzymes: Insights from Analysis of Mechanochemical Coupling in Myosin.

The mechanism of ATP hydrolysis in the myosin motor domain is analyzed using a combination of DFTB3/CHARMM simulations and enhanced sampling techniques. The motor domain is modeled in the pre-powerstroke state, in the post-rigor state, and as a hybrid based on the post-rigor state with a closed nucleotide-binding pocket. The ATP hydrolysis activity is found to depend on the positioning of nearby water molecules, and a network of polar residues facilitates proton transfer and charge redistribution during hydrolysis. Comparison of the observed hydrolysis pathways and the corresponding free energy profiles leads to detailed models for the mechanism of ATP hydrolysis in the pre-powerstroke state and proposes factors that regulate the hydrolysis activity in different conformational states. In the pre-powerstroke state, the scissile Pγ-O3ß bond breaks early in the reaction. Proton transfer from the lytic water to the γ-phosphate through active site residues is an important part of the kinetic bottleneck; several hydrolysis pathways that feature distinct proton transfer routes are found to have similar free energy barriers, suggesting a significant degree of plasticity in the hydrolysis mechanism. Comparison of hydrolysis in the pre-powerstroke state and the closed post-rigor model suggests that optimization of residues beyond the active site for electrostatic stabilization and preorganization is likely important to enzyme design.

QM/MM free energy simulations: recent progress and challenges.

Due to the higher computational cost relative to pure molecular mechanical (MM) simulations, hybrid quantum mechanical/molecular mechanical (QM/MM) free energy simulations particularly require a careful consideration of balancing computational cost and accuracy. Here we review several recent developments in free energy methods most relevant to QM/MM simulations and discuss several topics motivated by these developments using simple but informative examples that involve processes in water. For chemical reactions, we highlight the value of invoking enhanced sampling technique (e.g., replica-exchange) in umbrella sampling calculations and the value of including collective environmental variables (e.g., hydration level) in metadynamics simulations; we also illustrate the sensitivity of string calculations, especially free energy along the path, to various parameters in the computation. Alchemical free energy simulations with a specific thermodynamic cycle are used to probe the effect of including the first solvation shell into the QM region when computing solvation free energies. For cases where high-level QM/MM potential functions are needed, we analyze two different approaches: the QM/MM-MFEP method of Yang and co-workers and perturbative correction to low-level QM/MM free energy results. For the examples analyzed here, both approaches seem productive although care needs to be exercised when analyzing the perturbative corrections.

A Simple and Accurate Method To Calculate Free Energy Profiles and Reaction Rates from Restrained Molecular Simulations of Diffusive Processes.

A method is developed to obtain simultaneously free energy profiles and diffusion constants from restrained molecular simulations in diffusive systems. The method is based on low-order expansions of the free energy and diffusivity as functions of the reaction coordinate. These expansions lead to simple analytical relationships between simulation statistics and model parameters. The method is tested on 1D and 2D model systems; its accuracy is found to be comparable to or better than that of the existing alternatives, which are briefly discussed. An important aspect of the method is that the free energy is constructed by integrating its derivatives, which can be computed without need for overlapping sampling windows. The implementation of the method in any molecular simulation program that supports external umbrella potentials (e.g., CHARMM) requires modification of only a few lines of code. As a demonstration of its applicability to realistic biomolecular systems, the method is applied to model the α-helix ↔ ß-sheet transition in a 16-residue peptide in implicit solvent, with the reaction coordinate provided by the string method. Possible modifications of the method are briefly discussed; they include generalization to multidimensional reaction coordinates [in the spirit of the model of Ermak and McCammon (Ermak, D. L.; McCammon, J. A. J. Chem. Phys. 1978, 69, 1352-1360)], a higher-order expansion of the free energy surface, applicability in nonequilibrium systems, and a simple test for Markovianity. In view of the small overhead of the method relative to standard umbrella sampling, we suggest its routine application in the cases where umbrella potential simulations are appropriate.

Role of Protein Dynamics in Allosteric Control of the Catalytic Phosphoryl Transfer of Insulin Receptor Kinase.

The catalytic and allosteric mechanisms of insulin receptor kinase (IRK) are investigated by a combination of ab initio and semiempirical quantum mechanical and molecular mechanical (QM/MM) methods and classical molecular dynamics (MD) simulations. The simulations reveal that the catalytic reaction proceeds in two steps, starting with the transfer of a proton from substrate Tyr to the catalytic Asp1132, followed by the phosphoryl transfer from ATP to substrate Tyr. The enhancement of the catalytic rate of IRK upon phosphorylations in the enzyme's activation loop is found to occur mainly via changes to the free energy landscape of the proton transfer step, favoring the proton transfer in the fully phosphorylated enzyme. In contrast, the effects of the phosphorylations on the phosphoryl transfer are smaller. Equilibrium MD simulations show that IRK phosphorylations affect the protein dynamics of the enzyme before the proton transfer to Asp1132 with only a minor effect after the proton transfer. This finding is consistent with the large change in the proton transfer free energy and the smaller change in the free energy barrier of phosphoryl transfer found by QM/MM simulations. Taken together, the present results provide details on how IRK phosphorylation exerts allosteric control of the catalytic activity via modifications of protein dynamics and free energy landscape of catalytic reaction. The results also highlight the importance of protein dynamics in connecting protein allostery and catalysis to control catalytic activity of enzymes.