Shijun Zhong

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 Small-Molecule Inhibitor of BCL6 Kills DLBCL Cells In Vitro and In Vivo

The BCL6 transcriptional repressor is the most frequently involved oncogene in diffuse large B cell lymphoma (DLBCL). We combined computer-aided drug design with functional assays to identify low-molecular-weight compounds that bind to the corepressor binding groove of the BCL6 BTB domain. One such compound disrupted BCL6/corepressor complexes in vitro and in vivo, and was observed by X-ray crystallography and NMR to bind the critical site within the BTB groove. This compound could induce expression of BCL6 target genes and kill BCL6-positive DLBCL cell lines. In xenotransplantation experiments, the compound was nontoxic and potently suppressed DLBCL tumors in vivo. The compound also killed primary DLBCLs from human patients.

Rational Design of Human DNA Ligase Inhibitors that Target Cellular DNA Replication and Repair

Based on the crystal structure of human DNA ligase I complexed with nicked DNA, computer-aided drug design was used to identify compounds in a database of 1.5 million commercially available low molecular weight chemicals that were predicted to bind to a DNA-binding pocket within the DNA-binding domain of DNA ligase I, thereby inhibiting DNA joining. Ten of 192 candidates specifically inhibited purified human DNA ligase I. Notably, a subset of these compounds was also active against the other human DNA ligases. Three compounds that differed in their specificity for the three human DNA ligases were analyzed further. L82 inhibited DNA ligase I, L67 inhibited DNA ligases I and III, and L189 inhibited DNA ligases I, III, and IV in DNA joining assays with purified proteins and in cell extract assays of DNA replication, base excision repair, and nonhomologous end-joining. L67 and L189 are simple competitive inhibitors with respect to nicked DNA, whereas L82 is an uncompetitive inhibitor that stabilized complex formation between DNA ligase I and nicked DNA. In cell culture assays, L82 was cytostatic whereas L67 and L189 were cytotoxic. Concordant with their ability to inhibit DNA repair in vitro, subtoxic concentrations of L67 and L189 significantly increased the cytotoxicity of DNA-damaging agents. Interestingly, the ligase inhibitors specifically sensitized cancer cells to DNA damage. Thus, these novel human DNA ligase inhibitors will not only provide insights into the cellular function of these enzymes but also serve as lead compounds for the development of anticancer agents.

Computational Identification of Inhibitors of Protein-Protein Interactions

The ability to control protein-protein interactions (PPIs) for therapeutic purposes is attractive since many processes in cells involve such interactions. Recent successes in the discovery of small molecules that target protein-protein interactions for drug development have shown that targeting these interactions is indeed feasible. In the present review the use of computer-aided drug design (CADD) via database screening or docking algorithms for identifying inhibitors of protein-protein interactions is introduced. The principles of database screening and a practical protocol for targeting PPIs are described. The recent applications of these approaches to different systems involving protein-protein interactions, including BCL-2, S100B, ERK and p56lck, are presented and provide valuable examples of inhibitor discovery and design.

Uniformly convergent n-tuple-ζ augmented polarized (nZaP) basis sets for complete basis set extrapolations. I. Self-consistent field energies

We present a sequence of n-tuple-zeta augmented polarized (nZaP) basis sets designed for extrapolations of both self-consistent field (SCF) and correlation energies to the complete basis set (CBS) limit. These nZaP basis sets (n=2-6) are formulated to give consistent errors throughout the Periodic Table (e.g., a consistent of approximately 1 mhartree/electron error for the 2ZaP SCF energy and a consistent of approximately 1.4 muhartree/electron error for the 6ZaP SCF energy). The SCF energy exhibits systematic convergence to the CBS limit: E(SCF)(nZaP) approximately E(SCF)(CBS)+Ae(-an). A single parameter, a=6.30, describes the 2ZaP through 6ZaP errors of H through Xe within 10%. The SCF rms basis set truncation errors of H through Xe are 33.5mE(h), 4.58mE(h), 0.82mE(h), 0.18mE(h), and 0.047mE(h) for 2ZaP, 3ZaP, 4ZaP, 5ZaP, and 6ZaP, respectively. Linear extrapolations of the (2,3)ZaP, (3,4)ZaP, (4,5)ZaP, and (5,6)ZaP calculations (all with a=6.30) reduce these errors by an order of magnitude to 0.24mE(h), 0.056mE(h), 0.020mE(h), and 0.005mE(h), respectively. A test set of 34 atoms, ions, and molecules gives similar results, and the associated test set of 25 chemical energy differences also gives comparable absolute accuracy. However, the cancellation of errors between reactant and product is lost by extrapolation. As a result, these chemical energy differences show a more modest two-to-fourfold improvement with extrapolation.

Identification and validation of human DNA ligase inhibitors using computer-aided drug design.

Linking together of DNA strands by DNA ligases is essential for DNA replication and repair. Since many therapies used to treat cancer act by causing DNA damage, there is growing interest in the development of DNA repair inhibitors. Accordingly, virtual database screening and experimental evaluation were applied to identify inhibitors of human DNA ligase I (hLigI). When a DNA binding site within the DNA binding domain (DBD) of hLigI was targeted, more than 1 million compounds were screened from which 192 were chosen for experimental evaluation. In DNA joining assays, 10 compounds specifically inhibited hLigI, 5 of which also inhibited the proliferation of cultured human cell lines. Analysis of the 10 active compounds revealed the utility of including multiple protein conformations and chemical clustering in the virtual screening procedure. The identified ligase inhibitors are structurally diverse and have druglike physical and molecular characteristics making them ideal for further drug development studies.

Binding Response: A Descriptor for Selecting Ligand Binding Site on Protein Surfaces

The identification of ligand binding sites on a protein is an essential step in the selection of inhibitors of protein-ligand or protein-protein interactions via virtual database screening. To facilitate binding site identification, a novel descriptor, the binding response, is proposed in the present paper to quantitatively evaluate putative binding sites on the basis of their response to a test set of probe compounds. The binding response is determined on the basis of contributions from both the ligand-protein interaction energy and the geometry of binding poses for a database of test ligands. A favorable binding response is obtained for binding sites with favorable ligand binding energies and with ligand geometries within the putative site for the majority of compounds in the test set. The utility of this descriptor is illustrated by applying it to a number of known protein-ligand complexes, showing the approach to identify the experimental binding sites as the highest scoring site in 26 out of 29 cases; in the remaining three cases, it was among the top three scoring sites. This method is combined with sphere-based site identification and clustering methods to yield an automated approach for the identification of binding sites on proteins suitable for database screen or de novo drug design.

On the optimization of Gaussian basis sets

A new procedure for the optimization of the exponents, αj, of Gaussian basis functions, Ylm(ϑ,φ)rle−αjr2, is proposed and evaluated. The direct optimization of the exponents is hindered by the very strong coupling between these nonlinear variational parameters. However, expansion of the logarithms of the exponents in the orthonormal Legendre polynomials, Pk, of the index, j: ln αj=∑k=0kmaxAkPk((2j−2)/(Nprim−1)−1), yields a new set of well-conditioned parameters, Ak, and a complete sequence of well-conditioned exponent optimizations proceeding from the even-tempered basis set (kmax=1) to a fully optimized basis set (kmax=Nprim−1). The error relative to the exact numerical self-consistent field limit for a six-term expansion is consistently no more than 25% larger than the error for the completely optimized basis set. Thus, there is no need to optimize more than six well-conditioned variational parameters, even for the largest sets of Gaussian primitives.

Targeting NAD biosynthesis in bacterial pathogens: Structure-based development of inhibitors of nicotinate mononucleotide adenylyltransferase NadD.

The emergence of multidrug-resistant pathogens necessitates the search for new antibiotics acting on previously unexplored targets. Nicotinate mononucleotide adenylyltransferase of the NadD family, an essential enzyme of NAD biosynthesis in most bacteria, was selected as a target for structure-based inhibitor development. Using iterative in silico and in vitro screens, we identified small molecule compounds that efficiently inhibited target enzymes from Escherichia coli (ecNadD) and Bacillus anthracis (baNadD) but had no effect on functionally equivalent human enzymes. On-target antibacterial activity was demonstrated for some of the selected inhibitors. A 3D structure of baNadD was solved in complex with one of these inhibitors (3_02), providing mechanistic insights and guidelines for further improvement. Most importantly, the results of this study help validate NadD as a target for the development of antibacterial agents with potential broad-spectrum activity.

Possible binding modes for dinitrogen activation by the FeMo-cofactor in nitrogenase

Abstract Dinitrogen activation at the FeMo-cofactor site of nitrogenase was studied based upon the quantum chemical CNDO calculations performed on possible processes of dinitrogen approaching and binding to the FeMo-cofactor. The best dinitrogen binding mode suggested in this article is the dinitrogen-inserting end-on mode, in which one nitrogen atom is located at the geometrical center of the FeMo-cofactor, and another nitrogen atom keeps out of the Fe4 surface of the Fe6 prism of the FeMo-cofactor. The mechanism of protonation of dinitrogen at the FeMo-cofactor site is discussed based upon the suggested binding mode.