Anthony K. Felts

Integrated Modeling Program, Applied Chemical Theory (IMPACT)

We provide an overview of the IMPACT molecular mechanics program with an emphasis on recent developments and a description of its current functionality. With respect to core molecular mechanics technologies we include a status report for the fixed charge and polarizable force fields that can be used with the program and illustrate how the force fields, when used together with new atom typing and parameter assignment modules, have greatly expanded the coverage of organic compounds and medicinally relevant ligands. As we discuss in this review, explicit solvent simulations have been used to guide our design of implicit solvent models based on the generalized Born framework and a novel nonpolar estimator that have recently been incorporated into the program. With IMPACT it is possible to use several different advanced conformational sampling algorithms based on combining features of molecular dynamics and Monte Carlo simulations. The program includes two specialized molecular mechanics modules: Glide, a high‐throughput docking program, and QSite, a mixed quantum mechanics/molecular mechanics module. These modules employ the IMPACT infrastructure as a starting point for the construction of the protein model and assignment of molecular mechanics parameters, but have then been developed to meet specialized objectives with respect to sampling and the energy function.

Comparative Performance of Several Flexible Docking Programs and Scoring Functions: Enrichment Studies for a Diverse Set of Pharmaceutically Relevant Targets

Virtual screening by molecular docking has become a widely used approach to lead discovery in the pharmaceutical industry when a high-resolution structure of the biological target of interest is available. The performance of three widely used docking programs (Glide, GOLD, and DOCK) for virtual database screening is studied when they are applied to the same protein target and ligand set. Comparisons of the docking programs and scoring functions using a large and diverse data set of pharmaceutically interesting targets and active compounds are carried out. We focus on the problem of docking and scoring flexible compounds which are sterically capable of docking into a rigid conformation of the receptor. The Glide XP methodology is shown to consistently yield enrichments superior to the two alternative methods, while GOLD outperforms DOCK on average. The study also shows that docking into multiple receptor structures can decrease the docking error in screening a diverse set of active compounds.

Distinguishing native conformations of proteins from decoys with an effective free energy estimator based on the OPLS all-atom force field and the surface generalized born solvent model

Protein decoy data sets provide a benchmark for testing scoring functions designed for fold recognition and protein homology modeling problems. It is commonly believed that statistical potentials based on reduced atomic models are better able to discriminate native‐like from misfolded decoys than scoring functions based on more detailed molecular mechanics models. Recent benchmark tests on small data sets, however, suggest otherwise. In this work, we report the results of extensive decoy detection tests using an effective free energy function based on the OPLS all‐atom (OPLS‐AA) force field and the Surface Generalized Born (SGB) model for the solvent electrostatic effects. The OPLS‐AA/SGB effective free energy is used as a scoring function to detect native protein folds among a total of 48,832 decoys for 32 different proteins from Park and Levitts 4‐state‐reduced, Levitts local‐minima, Bakers ROSETTA all‐atom, and Skolnicks decoy sets. Solvent electrostatic effects are included through the Surface Generalized Born (SGB) model. All structures are locally minimized without restraints. From an analysis of the individual energy components of the OPLS‐AA/SGB energy function for the native and the best‐ranked decoy, it is determined that a balance of the terms of the potential is responsible for the minimized energies that most successfully distinguish the native from the misfolded conformations. Different combinations of individual energy terms provide less discrimination than the total energy. The results are consistent with observations that all‐atom molecular potentials coupled with intermediate level solvent dielectric models are competitive with knowledge‐based potentials for decoy detection and protein modeling problems such as fold recognition and homology modeling. Proteins 2002;48:404–422.

Free energy surfaces of β-hairpin and α-helical peptides generated by replica exchange molecular dynamics with the AGBNP implicit solvent model

We have studied the potential of mean force of two peptides, one known to adopt a β‐hairpin and the other an α‐helical conformation in solution. These peptides are, respectively, residues 41–56 of the C‐terminus (GEWTYDDATKTFTVTE) of the B1 domain of protein G and the 13 residue C‐peptide (KETAAAKFERQHM) of ribonuclease A. Extensive canonical ensemble sampling has been performed using a parallel replica exchange method. The effective potential employed in this work consists of the OPLS all‐atom force field (OPLS‐AA) and an analytical generalized Born (AGB) implicit solvent model including a novel nonpolar solvation free energy estimator (NP). An additional dielectric screening parameter has been incorporated into the AGBNP model. In the case of the β‐hairpin, the nonpolar solvation free energy estimator provides the necessary effective interactions for the collapse of the hydrophobic core (W43, Y45, F52, and V54), which the more commonly used surface‐area‐dependent nonpolar model does not provide. For both the β‐hairpin and the α‐helix, increased dielectric screening reduces the stability of incorrectly formed salt bridges, which tend to disrupt the formation of the hairpin and helix, respectively. The fraction of β‐hairpin and α‐helix content we obtained using the AGBNP model agrees well with experimental results. The thermodynamic stability of the β‐hairpin from protein G and the α‐helical C‐peptide from ribonuclease A as modeled with the OPLS‐AA/AGBNP effective potential reflects the balance between the nonpolar effective potential terms, which drive compaction, and the polar and hydrogen bonding terms, which promote secondary structure formation. Proteins 2004.

Protein tertiary structure prediction using a branch and bound algorithm

We report a new method for predicting protein tertiary structure from sequence and secondary structure information. The predictions result from global optimization of a potential energy function, including van der Waals, hydrophobic, and excluded volume terms. The optimization algorithm, which is based on the αBB method developed by Floudas and coworkers (Costas and Floudas, J Chem Phys 1994;100:1247–1261), uses a reduced model of the protein and is implemented in both distance and dihedral angle space, enabling a side‐by‐side comparison of methodologies. For a set of eight small proteins, representing the three basic types—all α, all β, and mixed α/β—the algorithm locates low‐energy native‐like structures (less than 6Å root mean square deviation from the native coordinates) starting from an unfolded state. Serial and parallel implementations of this methodology are discussed. Proteins 1999;35:41–57.

Identification of alternative binding sites for inhibitors of HIV-1 ribonuclease H through comparative analysis of virtual enrichment studies.

The ribonuclease H (RNase H) domain on the p66 monomer of HIV-1 reverse transcriptase enzyme has become a target for inhibition. The active site is one potential binding site, but other RNase H sites can accommodate inhibitors. Using a combination of experimental and computational studies, potential new binding sites and binding modes have been identified. Libraries of compounds were screened with an experimental assay to identify actives without knowledge of the binding site. The compounds were computationally docked at putative binding sites. Based on positive enrichment of natural-product actives relative to the database of compounds, we propose that many inhibitors bind to an alternative, potentially allosteric, site centered on Q507 of p66. For a series of hydrazone compounds, a small amount of positive enrichment was obtained when active compounds were bound by induced-fit docking at the interface between the DNA:RNA substrate and the RNase H domain near residue Q500.

Conformational landscape of the human immunodeficiency virus type 1 reverse transcriptase non-nucleoside inhibitor binding pocket: lessons for inhibitor design from a cluster analysis of many crystal structures.

Clustering of 99 available X-ray crystal structures of HIV-1 reverse transcriptase (RT) at the flexible non-nucleoside inhibitor binding pocket (NNIBP) provides information about features of the conformational landscape for binding non-nucleoside inhibitors (NNRTIs), including effects of mutation and crystal forms. The ensemble of NNIBP conformations is separated into eight discrete clusters based primarily on the position of the functionally important primer grip, the displacement of which is believed to be one of the mechanisms of inhibition of RT. Two of these clusters are populated by structures in which the primer grip exhibits novel conformations that differ from the predominant cluster by over 4 A and are induced by the unique inhibitors capravirine and rilpivirine/TMC278. This work identifies a new conformation of the NNIBP that may be used to design NNRTIs. It can also be used to guide more complete exploration of the NNIBP free energy landscape using advanced sampling techniques.

Protein Folding and Binding: Effective Potentials, Replica Exchange Simulations, and Network Models

Advances in computational biophysics depend on the development of accurate effective potentials and powerful sampling methods to traverse rugged energy landscapes. We have developed an approach that makes use of the combined power of replica exchange simulations and a network model for kinetics. We carry out replica exchange simulations to generate a very large set of states using an all-atom effective potential function and construct a kinetic model for the folding, using an ansatz that allows kinetic transitions between states based on structural similarity. We are also using replica exchange simulations to study the binding of ligands to proteins such as cytochrome P450. A better understanding of the relationship between the physical kinetics of the systems being studied to their “kinetics” in the replica exchange ensemble is needed to use this new technology to maximum advantage. To illustrate some of the challenges, we will discuss the results using a network model to “simulate” replica exchange simulations of protein folding.

Fold Recognition using the OPLS All-Atom Potential and the Surface Generalized Born Solvent Model

Protein decoy data sets provide a benchmark for testing scoring functions designed for fold recognition and protein homology modeling problems. It is commonly believed that statistical potentials based on reduced atomic models are better able to discriminate native-like from misfolded decoys than scoring functions based on more detailed molecular mechanics models. Recent benchmark tests, however, suggest otherwise. Further analysis of the effectiveness of all atom molecular mechanics scoring functions for detecting misfolded decoys and direct comparison with results obtained using a statistical potential derived for a reduced atomic model are presented in this report. The OPLS all-atom force field is used as a scoring function to detect native protein folds among the Park & Levitt large decoy sets. Solvent electrostatic effects are included through the Surface Generalized Born (SGB) model. The OPLS potential with SGB solvation (OPLS-AA/SGB) provides good discrimination between native-like structures and non-native decoys. From an analysis of the individual energy components of the OPLS-AA/SGB potential for the native and the best-ranked decoy, it is determined that a roughly even balance of the terms of the potential is responsible for distinguishing the native from the misfolded conformations. Different combinations of individual energy terms provide less discrimination than the total energy. The effects of scoring decoys using several dielectric models are compared also. With the SGB solvation model, close to 100% of the structures with energies within 100 kcal/mol of the native state minimum are native-like. In contrast, only 20% of the low energy structures are found to be native-like when a distance dependent dielectric is used instead of SGB to model solvent electrostatic effects. The results are consistent with observations that all-atom molecular potentials coupled with intermediate level solvent dielectric models are competitive with knowledge-based potentials for decoy detection and protein modeling problems such as fold recognition.