Alexander D. MacKerell

CHARMM general force field: A force field for drug‐like molecules compatible with the CHARMM all‐atom additive biological force fields

The widely used CHARMM additive all‐atom force field includes parameters for proteins, nucleic acids, lipids, and carbohydrates. In the present article, an extension of the CHARMM force field to drug‐like molecules is presented. The resulting CHARMM General Force Field (CGenFF) covers a wide range of chemical groups present in biomolecules and drug‐like molecules, including a large number of heterocyclic scaffolds. The parametrization philosophy behind the force field focuses on quality at the expense of transferability, with the implementation concentrating on an extensible force field. Statistics related to the quality of the parametrization with a focus on experimental validation are presented. Additionally, the parametrization procedure, described fully in the present article in the context of the model systems, pyrrolidine, and 3‐phenoxymethylpyrrolidine will allow users to readily extend the force field to chemical groups that are not explicitly covered in the force field as well as add functional groups to and link together molecules already available in the force field. CGenFF thus makes it possible to perform “all‐CHARMM” simulations on drug‐target interactions thereby extending the utility of CHARMM force fields to medicinally relevant systems.

A simple polarizable model of water based on classical Drude oscillators

A simple polarizable water model is developed and optimized for molecular dynamics simulations of the liquid phase under ambient conditions. The permanent charge distribution of the water molecule is represented by three point charges: two hydrogen sites and one additional M site positioned along the HOH bisector. Electronic induction is represented by introducing a classical charged Drude particle attached to the oxygen by a harmonic spring. The oxygen site carries an equal and opposite charge, and is the center of an intermolecular Lennard-Jones interaction. The HOH gas-phase experimental geometry is maintained rigidly and the dipole of the isolated molecule is 1.85 D, in accord with experiment. The model is simulated by considering the dynamics of an extended Lagrangian in which a small mass is attributed to the Drude particles. It is parametrized to reproduce the salient properties of liquid water under ambient conditions. The optimal model, refered to as SWM4-DP for “simple water model with four site...

Automation of the CHARMM General Force Field (CGenFF) I: Bond Perception and Atom Typing

Molecular mechanics force fields are widely used in computer-aided drug design for the study of drug-like molecules alone or interacting with biological systems. In simulations involving biological macromolecules, the biological part is typically represented by a specialized biomolecular force field, while the drug is represented by a matching general (organic) force field. In order to apply these general force fields to an arbitrary drug-like molecule, functionality for assignment of atom types, parameters, and charges is required. In the present article, which is part I of a series of two, we present the algorithms for bond perception and atom typing for the CHARMM General Force Field (CGenFF). The CGenFF atom typer first associates attributes to the atoms and bonds in a molecule, such as valence, bond order, and ring membership among others. Of note are a number of features that are specifically required for CGenFF. This information is then used by the atom typing routine to assign CGenFF atom types based on a programmable decision tree. This allows for straightforward implementation of CGenFFs complicated atom typing rules and for equally straightforward updating of the atom typing scheme as the force field grows. The presented atom typer was validated by assigning correct atom types on 477 model compounds including in the training set as well as 126 test-set molecules that were constructed to specifically verify its different components. The program may be utilized via an online implementation at https://www.paramchem.org/ .

Computational Biochemistry and Biophysics

Computational methods: atomistic models and force fields dynamics methods conformational analysis treatment of long-range forces and potential internal co-ordinate simulation method implicit solvent models normal mode analysis of biological molecules free energy calculations reaction rates and transition pathways computer simulation of biochemical reactions with QM-MM methods. Experimental data analysis: X-ray and neutron scattering as probes of the dynamics of biological molecules applications of molecular modelling in NMR structure determination. Modelling and design: comparative protein structure modelling Bayesian statistics in molecular and structural biology. Computer-aided drug design. Advanced applications: protein folding simulations of electron transfer proteins the RISM-SCF/MCSCF approach for the chemical processes in solutions nucleic acids simulations membrane simulations. Appendix: useful Web sites.

Six-site polarizable model of water based on the classical Drude oscillator.

A polarizable water model, SWM6, was developed and optimized for liquid phase simulations under ambient conditions. Building upon the previously developed SWM4-NDP model, additional sites representing oxygen lone-pairs were introduced. The geometry of the sites is assumed to be rigid. Considering the large number of adjustable parameters, simulated annealing together with polynomial fitting was used to facilitate model optimization. The new water model was shown to yield the correct self-diffusion coefficient after taking the system size effect into account, and the dimer geometry is better reproduced than in the SWM4 models. Moreover, the experimental oxygen-oxygen radial distribution is better reproduced, indicating that the new model more accurately describes the local hydrogen bonding structure of bulk phase water. This was further validated by its ability to reproduce the experimental nuclear magnetic shielding and related chemical shift of the water hydrogen in the bulk phase, a property sensitive to the local hydrogen bonding structure. In addition, comparison of the liquid properties of the SWM6 model is made with those of a number of widely used additive and polarizable models. Overall, improved balance between the description of monomer, dimer, clustered, and bulk phase water is obtained with the new model compared to its SWM4-NDP polarizable predecessor, though application of the model requires an approximately twofold increase on computational resources.

Inclusion of multiple fragment types in the Site Identification by Ligand Competitive Saturation (SILCS) approach

The site identification by ligand competitive saturation (SILCS) method identifies the location and approximate affinities of small molecular fragments on a target macromolecular surface by performing molecular dynamics (MD) simulations of the target in an aqueous solution of small molecules representative of different chemical functional groups. In this study, we introduce a set of small molecules to map potential interactions made by neutral hydrogen bond donors and acceptors and charged donor and acceptor fragments in addition to nonpolar fragments. The affinity pattern is obtained in the form of discretized probability or, equivalently, free energy maps, called FragMaps, which can be visualized with the target surface. We performed SILCS simulations for four proteins for which structural and thermodynamic data is available for multiple diverse ligands. Good overlap is shown between high affinity regions identified by the FragMaps and the crystallographic positions of ligand functional groups with similar chemical functionality, thus demonstrating the validity of the qualitative information obtained from the simulations. To test the ability of FragMaps in providing quantitative predictions, we calculate the previously introduced ligand grid free energy (LGFE) metric and observe its correspondence with experimentally measured binding affinity. LGFE is computed for different conformational ensembles and improvement in prediction is shown with increasing ligand conformational sampling. Ensemble generation includes a Monte Carlo sampling approach that uses the GFE FragMaps directly as the energy function. The results show that some but not all experimental trends are predicted and warrant improvements in the scoring methodology. In addition, the potential utility of atom-based free energy contributions to the LGFE scores and the use of multiple ligands in SILCS to identify displaceable water molecules during ligand design are discussed.

Chapter 1 Considerations for Lipid Force Field Development

Abstract The underlying approach to development of the CHARMM lipid force field, and the current ab initio and molecular dynamics methods for optimization of each term are reviewed. Results from the recent revision of the alkane force field and new results for esters illustrate the dependence of torsional surfaces on level of theory and basis set, and how changes in the surface manifest themselves in alkanes and dipalmitoylphosphatidylcholine (DPPC) bilayers. The following properties from simulation and experiment on DPPC bilayers are compared: structure factors from X-ray diffraction; deuterium order parameters; NMR spin lattice relaxation times; lipid translational diffusion constants; elastic moduli; and the dipole potential. The importance of including long-range Lennard-Jones interactions and taking finite system size into account is stressed. Theoretical and practical aspects associated with surface tensions and surface areas of lipid bilayers and monolayers are discussed.

Kirkwood-Buff analysis of aqueous N-methylacetamide and acetamide solutions modeled by the CHARMM additive and Drude polarizable force fields.

Kirkwood-Buff analysis was performed on aqueous solutions of N-methylacetamide and acetamide using the Chemistry at HARvard Molecular Mechanics additive and Drude polarizable all-atom force fields. Comparison of a range of properties with experimental results, including Kirkwood-Buff integrals, excess coordination numbers, solution densities, partial molar values, molar enthalpy of mixing, showed both models to be well behaved at higher solute concentrations with the Drude model showing systematic improvement at lower solution concentrations. However, both models showed difficulties reproducing experimental activity derivatives and the excess Gibbs energy, with the Drude model performing slightly better. At the molecular level, the improved agreement of the Drude model at low solute concentrations is due to increased structure in the solute-solute and solute-solvent interactions. The present results indicate that the explicit inclusion of electronic polarization leads to improved modeling of dilute solutions even when those properties are not included as target data during force field optimization.

TIT for TAT: the properties of inosine and adenosine in TATA box DNA.

The sequence dependent conformation, flexibility and hydration properties of DNA molecules constitute selectivity determinants in the formation of protein-DNA complexes. TATA boxes in which AT basepairs (bp) have been substituted by IC bp (TITI box) allow for probing these selectivity determinants for the complexation with the TATA box-binding protein (TBP) with different sequences but identical chemical surfaces. The reference promoter Adenovirus 2 Major Late Promoter (mlp) is formed by the apposition of two sequences with very different dynamic properties: an alternating TATA sequence and an A-tract. For a comparative study, we carried out molecular dynamics simulations of two DNA oligomers, one containing the mlp sequence (2 ns), and the other an analog where AT basepairs were substituted by IC basepairs (1 ns). The simulations, carried out with explicit solvent and counterinons, yield straight purine tracts, the A-tract being stiffer than the I-tract, an alternating structure for the YRYR tracts, and hydration patterns that differ between the purine tracts and the alternating sequence tracts. A detailed analysis of the proposed interactions responsible for the stiffness of the purine tracts indicates that the stacking between the bases bears the strongest correlation to stiffness. The hydration properties of the minor groove in the two oligomers are distinctly different. Such differences are likely to be responsible for the stronger binding of TBP to mlp over the inosine-substituted variant. The calculations were made possible by the development, described here, of a new set of forcefield parameters for inosine that complement the published CHARMM all-hydrogen nucleic acid parametrization.

RUNX2 and TAZ-dependent signaling pathways regulate soluble E-Cadherin levels and tumorsphere formation in breast cancer cells

Intratumoral heterogeneity and treatment resistance drive breast cancer (BC) metastasis and recurrence. The RUNX2 transcription factor is upregulated in early stage luminal BC. However, the precise mechanism by which RUNX2 regulates an oncogenic phenotype in luminal BCs remains an enigma. We show that RUNX2 is predictive of poor overall survival in BC patients. RUNX2 associated with the TAZ transcriptional co-activator to promote a tumorigenic phenotype that was inhibited by knockdown of TAZ. RUNX2 increased endogenous TAZ translocation to the nucleus, which was prevented by inhibiting RUNX2. RUNX2/TAZ interaction was associated with ectodomain shedding of an oncogenic soluble E-Cadherin fragment (sE-Cad), which is known to cooperate with human epidermal growth factor receptor-2 (HER2/ErbB2) to increase BC growth. Neutralizing E-Cadherin antibodies or TAZ knockdown reduced the levels of sE-Cad in RUNX2-expressing BC cells and inhibited tumorsphere formation. RUNX2 expression also increased HER2-mediated tumorsphere size, which was reduced after treatment with the HER2-targeting agents Herceptin and lapatinib. These data support a novel role for RUNX2 in promoting an oncogenic phenotype in luminal BC in the context of TAZ, sE-Cad, and HER2. Using this signaling pathway to monitor BC cell oncogenic activity will accelerate the discovery of new therapeutic modalities to treat BC patients.