Accurate calculation of mutational effects on the thermodynamics of inhibitor binding to p38α MAP kinase: a combined computational and experimental study

Combining QM/MM Calculations with Classical Mining Minima to Predict Protein-Ligand Binding Free Energy

We developed an effective binding free energy prediction protocol which incorporates quantum mechanical/molecular mechanical (QM/MM) calculations to substitute the specified atomic charges of force fields with quantum-mechanically recalculated ones at a proposed pose using a mining minima approach with the VeraChem mining minima engine. We tested this protocol using seven well-known targets with 147 different ligands and compared it with classical mining minima and the most popular binding free energy (BFE) methods using different metrics. Our new protocol, dubbed Qcharge-VM2, yielded an overall Pearson correlation of 0.86, which was better than all the methods examined. Qcharge-VM2 performed significantly better than implicit solvent-based methods, such as MM-GBSA and MM-PBSA, but not as good as explicit water-based free energy perturbation methods, such as FEP+, in terms of root-mean-square error, RMSE (1.75 kcal/mol) and mean unsigned error, MUE (1.39 kcal/mol) on a limited set of targets. However, our protocol is substantially less computationally demanding compared with FEP+. The combined accuracy and efficiency of our method can be valuable in drug discovery campaigns.

Exploitation of dihydroorotate dehydrogenase (DHODH) and p53 activation as therapeutic targets: A case study in polypharmacology

The tenovins are a frequently studied class of compounds capable of inhibiting sirtuin activity, which is thought to result in increased acetylation and protection of the tumor suppressor p53 from degradation. However, as we and other laboratories have shown previously, certain tenovins are also capable of inhibiting autophagic flux, demonstrating the ability of these compounds to engage with more than one target. In this study, we present two additional mechanisms by which tenovins are able to activate p53 and kill tumor cells in culture. These mechanisms are the inhibition of a key enzyme of the de novo pyrimidine synthesis pathway, dihydroorotate dehydrogenase (DHODH), and the blockage of uridine transport into cells. These findings hold a 3-fold significance: first, we demonstrate that tenovins, and perhaps other compounds that activate p53, may activate p53 by more than one mechanism; second, that work previously conducted with certain tenovins as SirT1 inhibitors should additionally be viewed through the lens of DHODH inhibition as this is a major contributor to the mechanism of action of the most widely used tenovins; and finally, that small changes in the structure of a small molecule can lead to a dramatic change in the target profile of the molecule even when the phenotypic readout remains static.

Implicit ligand theory for relative binding free energies: II. An estimator based on control variates

Implicit ligand theory describes the relationship between the noncovalent binding free energy and the binding free energy between a ligand and multiple rigid receptor conformations. We have previously shown that if the receptor conformations are sampled from or reweighed to a holo ensemble, the binding free energy relative to the ligand that defines the ensemble can be calculated. Here, we apply a variance reduction technique known as control variates to derive a new statistical estimator for the relative binding free energy. In applications to a data set of 6 reference ligands and 18 test ligands, statistically significant differences between the estimators are not observed for most systems. However, in cases where such differences are observed, the new estimator is more accurate, precise, and converges more quickly. Performance improvements are most consistent where there is a clear correlation, with a correlation coefficient greater than 0.3, between the control variate and the statistic being averaged.

On Restraints in End-Point Protein-Ligand Binding Free Energy Calculations

The impact of harmonic restraints on protein heavy atoms and ligand atoms on end-point free energy calculations is systematically characterized for 54 protein-ligand complexes. We observe that stronger restraints reduce the equilibration time and statistical inefficiency, suppress conformational sampling, influence correlation with experiment, and monotonically decrease the estimated loss of entropy upon binding, leading to stronger estimated binding free energies in most systems. A statistical estimator that reweights for the biasing potential and includes data prior to the estimated equilibration time has the highest correlation with experiment. A spring constant of 20 cal mol-1 Å-2 maintains a near-native energy landscape and suppresses artifactual energy minima while minimally limiting thermal fluctuations about the crystal structure. © 2019 Wiley Periodicals, Inc.

Exploring Gatekeeper Mutations in EGFR through Computer Simulations

The emergence of resistance against drugs that inhibit a particular protein is a major problem in targeted therapy. There is a clear need for rigorous methods to predict the likelihood of specific drug-resistance mutations arising in response to the binding of a drug. In this work we attempt to develop a robust computational protocol for predicting drug resistant mutations at the gatekeeper position (T790) in EGFR. We explore how mutations at this site affects interactions with ATP and three drugs that are currently used in clinics. We found, as expected, that certain mutations are not tolerated structurally, while some other mutations interfere with the natural substrate and hence are unlikely to be selected for. However, we found five possible mutations that are well tolerated structurally and energetically. Two of these mutations were predicted to have increased affinity for the drugs over ATP, as has been reported earlier. By reproducing the trends in the experimental binding affinities of the data, the methods chosen here are able to correctly predict the effects of these mutations on the binding affinities of the drugs. However, the increased affinity does not always translate into increased efficacy, because the efficacy is affected by several other factors such as binding kinetics, competition with ATP, and residence times. The computational methods used in the current study are able to reproduce or predict the effects of mutations on the binding affinities. However, a different set of methods is required to predict the kinetics of drug binding.

In silico identification of genetic mutations conferring resistance to acetohydroxyacid synthase inhibitors: A case study of Kochia scoparia

Mutations that confer herbicide resistance are a primary concern for herbicide-based chemical control of invasive plants and are often under-characterized structurally and functionally. As the outcome of selection pressure, resistance mutations usually result from repeated long-term applications of herbicides with the same mode of action and are discovered through extensive field trials. Here we used acetohydroxyacid synthase (AHAS) of Kochia scoparia (KsAHAS) as an example to demonstrate that, given the sequence of a target protein, the impact of genetic mutations on ligand binding could be evaluated and resistance mutations could be identified using a biophysics-based computational approach. Briefly, the 3D structures of wild-type (WT) and mutated KsAHAS-herbicide complexes were constructed by homology modeling, docking and molecular dynamics simulation. The resistance profile of two AHAS-inhibiting herbicides, tribenuron methyl and thifensulfuron methyl, was obtained by estimating their binding affinity with 29 KsAHAS (1 WT and 28 mutated) using 6 molecular mechanical (MM) and 18 hybrid quantum mechanical/molecular mechanical (QM/MM) methods in combination with three structure sampling strategies. By comparing predicted resistance with experimentally determined resistance in the 29 biotypes of K. scoparia field populations, we identified the best method (i.e., MM-PBSA with single structure) out of all tested methods for the herbicide-KsAHAS system, which exhibited the highest accuracy (up to 100%) in discerning mutations conferring resistance or susceptibility to the two AHAS inhibitors. Our results suggest that the in silico approach has the potential to be widely adopted for assessing mutation-endowed herbicide resistance on a case-by-case basis.

Simple Entropy Terms for End-Point Binding Free Energy Calculations

We introduce a number of computationally inexpensive modifications to the MM/PBSA and MM/GBSA estimators for binding free energies, which are based on average receptor-ligand interaction energies in simulations of a noncovalent complex, to improve the treatment of entropy: second- and higher-order terms in a cumulant expansion and a confining potential on ligand external degrees of freedom. We also consider a filter for snapshots where ligands have drifted from the initial binding pose. The variations were tested on six sets of systems for which binding modes and free energies have previously been experimentally determined. For some data sets, none of the tested estimators led to results significantly correlated with measured free energies. In data sets with nontrivial correlation, a ligand RMSD cutoff of 3 Å and a second-order truncation of the cumulant expansion was found to be comparable or better than the average interaction energy by several statistical metrics.

Using the fast fourier transform in binding free energy calculations

According to implicit ligand theory, the standard binding free energy is an exponential average of the binding potential of mean force (BPMF), an exponential average of the interaction energy between the unbound ligand ensemble and a rigid receptor. Here, we use the fast Fourier transform (FFT) to efficiently evaluate BPMFs by calculating interaction energies when rigid ligand configurations from the unbound ensemble are discretely translated across rigid receptor conformations. Results for standard binding free energies between T4 lysozyme and 141 small organic molecules are in good agreement with previous alchemical calculations based on (1) a flexible complex ( R≈0.9 for 24 systems) and (2) flexible ligand with multiple rigid receptor configurations ( R≈0.8 for 141 systems). While the FFT is routinely used for molecular docking, to our knowledge this is the first time that the algorithm has been used for rigorous binding free energy calculations. © 2017 Wiley Periodicals, Inc.

Implicit ligand theory for relative binding free energies

Implicit ligand theory enables noncovalent binding free energies to be calculated based on an exponential average of the binding potential of mean force (BPMF)-the binding free energy between a flexible ligand and rigid receptor-over a precomputed ensemble of receptor configurations. In the original formalism, receptor configurations were drawn from or reweighted to the apo ensemble. Here we show that BPMFs averaged over a holo ensemble yield binding free energies relative to the reference ligand that specifies the ensemble. When using receptor snapshots from an alchemical simulation with a single ligand, the new statistical estimator outperforms the original.

Absolute Binding Free Energies between T4 Lysozyme and 141 Small Molecules: Calculations Based on Multiple Rigid Receptor Configurations

We demonstrate the feasibility of estimating protein-ligand binding free energies using multiple rigid receptor configurations. On the basis of T4 lysozyme snapshots extracted from six alchemical binding free energy calculations with a flexible receptor, binding free energies were estimated for a total of 141 ligands. For 24 ligands, the calculations reproduced flexible-receptor estimates with a correlation coefficient of 0.90 and a root-mean-square error of 1.59 kcal/mol. The accuracy of calculations based on Poisson-Boltzmann/surface area implicit solvent was comparable to that of previously reported free energy calculations.

Reparametrization of Protein Force Field Nonbonded Interactions Guided by Osmotic Coefficient Measurements from Molecular Dynamics Simulations

There is a small, but growing, body of literature describing the use of osmotic coefficient measurements to validate and reparametrize simulation force fields. Here we have investigated the ability of five very commonly used force field and water model combinations to reproduce the osmotic coefficients of seven neutral amino acids and five small molecules. The force fields tested include AMBER ff99SB-ILDN, CHARMM36, GROMOS54a7, and OPLS-AA, with the first of these tested in conjunction with the TIP3P and TIP4P-Ew water models. In general, for both the amino acids and the small molecules, the tested force fields produce computed osmotic coefficients that are lower than experiment; this is indicative of excessively favorable solute-solute interactions. The sole exception to this general trend is provided by GROMOS54a7 when applied to amino acids: in this case, the computed osmotic coefficients are consistently too high. Importantly, we show that all of the force fields tested can be made to accurately reproduce the experimental osmotic coefficients of the amino acids when minor modifications-some previously reported by others and some that are new to this study-are made to the van der Waals interactions of the charged terminal groups. Special care is required, however, when simulating Proline with a number of the force fields, and a hydroxyl-group specific modification is required in order to correct Serine and Threonine when simulated with AMBER ff99SB-ILDN. Interestingly, an alternative parametrization of the van der Waals interactions in the latter force field, proposed by the Nerenberg and Head-Gordon groups, is shown to immediately produce osmotic coefficients that are in excellent agreement with experiment. Overall, this study reinforces the idea that osmotic coefficient measurements can be used to identify general shortcomings in commonly used force fields' descriptions of solute-solute interactions and further demonstrates that modifications to van der Waals parameters provide a simple route to optimizing agreement with experiment.

Molecular Dynamics Analysis of Binding of Kinase Inhibitors to WT EGFR and the T790M Mutant

Epidermal growth factor receptor (EGFR) inhibitors interrupt EGFR-dependent cellular signaling pathways that lead to accelerated tumor growth and proliferation. Mutation of a threonine in the inhibitor binding pocket, known as the "gatekeeper", to methionine (T790M) confers acquired resistance to several EGFR-selective inhibitors. We studied interactions between EGFR inhibitors and the gatekeeper residues of the target protein. Thermodynamic integration (TI) with Amber14 indicates that the binding energies of gefitinib and AEE788 to the active state of the T790M mutant EGFR is 3 kcal/mol higher than to the wild type (WT), whereas ATP binding energy to the mutant is similar to the WT. Using metadynamics MD simulations with NAMD v2.9, the conformational equilibrium between the inactive resting state and the catalytically competent activate state was determined for the WT EGFR. When combined with the results obtained by Sutto and Gervasio, our simulations showed that the T790M point mutation lowers the free energy of the active state by 5 kcal/mol relative to the inactive state of the enzyme. Relative to the WT, the T790M mutant binds gefitinib more strongly. The T790M mutation is nevertheless resistant due to its increased binding of ATP. By contrast, the binding of AEE788 to the active state causes a conformational change in the αC-helix adjacent to the inhibitor binding pocket, that results in a 2 kcal/mol energy penalty. The energy penalty explains why the binding of AEE788 to the T790M mutant is unfavorable relative to binding to WT EGFR. These results establish the role of the gatekeeper mutation on inhibitor selectivity. Additional molecular dynamics (MD) simulations, TI, and metadynamics MD simulations reveal the origins of the changes in binding energy of WT and mutants.

Assessing the performance of MM/PBSA and MM/GBSA methods. 4. Accuracies of MM/PBSA and MM/GBSA methodologies evaluated by various simulation protocols using PDBbind data set

By using different evaluation strategies, we systemically evaluated the performance of Molecular Mechanics/Generalized Born Surface Area (MM/GBSA) and Molecular Mechanics/Poisson-Boltzmann Surface Area (MM/PBSA) methodologies based on more than 1800 protein-ligand crystal structures in the PDBbind database. The results can be summarized as follows: (1) for the one-protein-family/one-binding-ligand case which represents the unbiased protein-ligand complex sampling, both MM/GBSA and MM/PBSA methodologies achieve approximately equal accuracies at the interior dielectric constant of 4 (with rp = 0.408 ± 0.006 of MM/GBSA and rp = 0.388 ± 0.006 of MM/PBSA based on the minimized structures); while for the total dataset (1864 crystal structures), the overall best Pearson correlation coefficient (rp = 0.579 ± 0.002) based on MM/GBSA is better than that of MM/PBSA (rp = 0.491 ± 0.003), indicating that biased sampling may significantly affect the accuracy of the predicted result (some protein families contain too many instances and can bias the overall predicted accuracy). Therefore, family based classification is needed to evaluate the two methodologies; (2) the prediction accuracies of MM/GBSA and MM/PBSA for different protein families are quite different with rp ranging from 0 to 0.9, whereas the correlation and ranking scores (an averaged rp/rs over a list of protein folds and also representing the unbiased sampling) given by MM/PBSA (rp-score = 0.506 ± 0.050 and rs-score = 0.481 ± 0.052) are comparable to those given by MM/GBSA (rp-score = 0.516 ± 0.047 and rs-score = 0.463 ± 0.047) at the fold family level; (3) for the overall prediction accuracies, molecular dynamics (MD) simulation may not be quite necessary for MM/GBSA (rp-minimized = 0.579 ± 0.002 and rp-1ns = 0.564 ± 0.002), but is needed for MM/PBSA (rp-minimized = 0.412 ± 0.003 and rp-1ns = 0.491 ± 0.003). However, for the individual systems, whether to use MD simulation is depended. (4) both MM/GBSA and MM/PBSA may be unable to give successful predictions for the ligands with high formal charges, with the Pearson correlation coefficient ranging from 0.621 ± 0.003 (neutral ligands) to 0.125 ± 0.142 (ligands with a formal charge of 5). Therefore, it can be summarized that, although MM/GBSA and MM/PBSA perform similarly in the unbiased dataset, for the currently available crystal structures in the PDBbind database, compared with MM/GBSA, which may be used in multi-target comparisons, MM/PBSA is more sensitive to the investigated systems, and may be more suitable for individual-target-level binding free energy ranking. This study may provide useful guidance for the post-processing of docking based studies.

An interaction map of small-molecule kinase inhibitors with anaplastic lymphoma kinase (ALK) mutants in ALK-positive non-small cell lung cancer

Human anaplastic lymphoma kinase (ALK) has become a well-established target for the treatment of ALK-positive non-small cell lung cancer (NSCLC). Here, we have profiled seven small-molecule inhibitors, including 2 that are approved drugs, against a panel of clinically relevant mutations in ALK tyrosine kinase (TK) domain, aiming at a comprehensive understanding of molecular mechanism and biological implication underlying inhibitor response to ALK TK mutation. We find that (i) the gatekeeper mutation L1196M causes crizotinib resistance by simultaneously increasing and decreasing the binding affinities of, respectively, ATP and inhibitor to ALK, whereas the secondary mutation C1156Y, which is located far away from the ATP-binding site of ALK TK domain, causes the resistance by inducing marked allosteric effect on the site, (ii) the 2nd and 3rd generation kinase inhibitors exhibit relatively high sensitivity towards ALK mutants as compared to 1st generation inhibitors, (iii) the pan-kinase inhibitor staurosporine is insensitive for most mutations due to its high structural compatibility, and (iv) ATP affinity to ALK is generally reduced upon most clinically relevant mutations. Furthermore, we also identify six novel mutation-inhibitor pairs that are potentially associated with drug resistance. In addition, the G1202R and C1156Y mutations are expected to generally cause resistance for many existing inhibitors, since they can address significant effect on the geometric shape and physicochemical property of ALK active pocket.

Protein design algorithms predict viable resistance to an experimental antifolate

Methods to accurately predict potential drug target mutations in response to early-stage leads could drive the design of more resilient first generation drug candidates. In this study, a structure-based protein design algorithm (K* in the OSPREY suite) was used to prospectively identify single-nucleotide polymorphisms that confer resistance to an experimental inhibitor effective against dihydrofolate reductase (DHFR) from Staphylococcus aureus. Four of the top-ranked mutations in DHFR were found to be catalytically competent and resistant to the inhibitor. Selection of resistant bacteria in vitro reveals that two of the predicted mutations arise in the background of a compensatory mutation. Using enzyme kinetics, microbiology, and crystal structures of the complexes, we determined the fitness of the mutant enzymes and strains, the structural basis of resistance, and the compensatory relationship of the mutations. To our knowledge, this work illustrates the first application of protein design algorithms to prospectively predict viable resistance mutations that arise in bacteria under antibiotic pressure.