Molecular recognition between protein and ligand is the foundation of a wide variety of biological processes. Non-bonded interactions are responsible for molecular recognition in various protein-ligand and protein-protein complexes. Theoretical investigations on non-bonded interactions in large biomolecular systems have always been the subject of great theoretical interests. But it represents a challenging undertaking due to the large system size and the necessity to include electron correlations in treating non-bonded interactions. We have systematically investigated the molecular determinants responsible for the molecular recognition in the formation of protein-ligand and protein-protein complexes in the following biologically important systems:
In the first project, a large scale data mining and high level quantum chemical analysis was performed to investigate the molecular recognition between Non-Nucleoside Reverse Transcriptase Inhibitors (NNRTIs) and the Human Immunodeficiency Virus type 1 (HIV-1) Reverse Transcriptase (RT). Thirty nine crystal structures of HIV-1 RT with bound NNRTI were systematically examined, resulting in the discovery of Trp-229 of the HIV-1 RT as a key determinant in binding all NNRTIs. This represents an important finding since mutation of HIV-1 RT is the leading cause of NNRTI drugs failure, and Trp-229 was known to be one of the residues that cannot be mutated for the survival of HIV-1. Thus, inhibitors targeting Trp-229 have the promise of overcoming drug resistance to NNRTI inhibitors. Intermolecular interactions between NNRTI and Trp229 are systematical analyzed, leading to the identification of six key binding modes. CH-π and T-shaped π-π stacking interactions were found to be the two most dominant intermolecular interactions responsible for the molecular recognition between NNRTIs and Trp-229, with 27 out of 39 NNRTIs utilizing CH-π interactions and 21 out of 39 NNRTIs utilizing T-shaped π-π interactions. There is a good correlation between our calculated interaction energies and the observed resilience to resistant mutations: that the more intensive the interaction with Trp-229 is, the less sensitive is the NNRTI to resistance mutations. This work not only provides us with a better understanding of the molecular recognition between NNRTIs and Trp-229 of HIV-1 RT, but also will guide us in the design of the new generations of NNRTI drugs that are less sensitive to resistance mutations.
The second project aims at designing enzyme inhibitors for Aspartate-β-semialdehyde dehydrogenase (ASADH) in the aspartate pathway, with the development of a novel integrated structure-based drug design (SBDD) protocol. The latter integrates the shape-based comparison method, the docking method, the hybrid quantum mechanics/molecular mechanics (QM/MM) simulation, and the quantum chemical analysis methods. Because the aspartate pathway is only present in plants, bacteria, and fungi, and is completely absent in humans, blocking the key enzyme in this pathway, ASADH, offers a viable approach to develop highly selective novel anti-bacterial drugs. Our newly developed SBDD protocol has been applied to designing ASADH inhibitors that target the binding pocket of the coenzyme NADP. The de novo design method encoded with the QM/MM simulation, and the quantum chemical analysis method has also been applied to designing the ASADH inhibitors targeting the substrate binding site.
The third project deals with molecular recognition of cell surface receptor by botulinum neurotoxin B. Botulinum neurotoxins (BoNTs) are among the most poisonous toxins known to mankind. Decades of biochemical investigations have led to the realization that the extreme toxicities of BoNTs have a lot to do with the high affinity and specificity of their binding to the neuronal membrane. High level quantum chemical calculations were carried out to quantify the strengths of nonbonded interactions responsible for the molecular recognition of cell surface receptor Syt-II by the botulinum neurotoxin BoNT/B, on the basis of the x-ray crystal structure of the BoNT/B – Syt-II. The work resulted in the discovery of multiple modes of molecular recognition in the formation of the BoNT/B – Syt-II complex. The amphipathic α-helix of the protein receptor Syt-II interacts with the neurotoxin BoNT/B by forming two interfaces of binding interactions, a hydrophobic interface and a hydrophilic interface, for molecular recognition. The hydrophobic interface consists of residues Phe47, Leu50, Phe54, Phe55 and Ile58 of Syt-II interacting with residues Trp1178, Tyr1181, Tyr1183, Phe1194 and Phe1204 of Hcc of BoNT/B. The hydrophilic interface is composed of residues Lys53 and Glu57 of Syt-II interacting with residues Lys1113, Asp1115, Ser1116 and Lys1192 of Hcc of BoNT/B. Intermolecular interaction energies between BoNT/B and its receptor protein Syt-II were calculated by means of the supermolecular approach at the MP2 level, with solvation energy correction by means of the SM5.42R Solvation Model of Cramer and Truhlar. The binding strength of the hydrophobic interface is quantified and compared with that at the hydrophilic interface of the BoNT/B - Syt-II complex. It was found that the energetic contribution of the hydrophobic interface toward binding of BoNT/B with Syt-II (-9.49 kcal/mol) is much larger than that of the hydrophilic interface (-2.58 kcal/mol). Pair-wise intermolecular interaction analysis is also performed to decipher the molecular determinants for recognition of BoNT/B by Syt-II. The interest lies in determining which types of interactions are used by the protein receptor Syt-II for recognition of BoNT/B, and what their relative importance is. It is concluded that π – π stacking interactions among aromatic residues function as the major molecular determinants for recognition of cell surface receptor Syt-II by the botulinum neurotoxin BoNT/B. Our findings are in good agreement with results of biophysical and site-directed mutagenesis experiments reported in the literature.