We develop, implement, and apply computational methods to study the structure, dynamics, and functional mechanisms of biomolecules. Over the last year, we have worked in two main, complementary directions: i) refinement of a continuum model to represent molecular interactions in solutions, the predictive power of which was demonstrated in the structure prediction of peptides and mini-proteins from their primary sequence, and ii) development of a self-guided multiscale method for modeling many-protein systems in realistic crowded environments. We also utilize ab-initio quantum chemistry to investigate the geometry and energetics of bioactive compounds. This approach is particularly useful in elucidating the transition states of chemical reactions that cannot be probed experimentally. The mechanistic understanding provides a firm structural/theoretical basis for controlling the outcome of chemical reactions, e.g., the ratio of possible products (one patent granted, one patent application published). We have begun to develop and apply pattern-recognition techniques, including artificial neural networks (ANN) and genetic algorithms, to identify non-obvious behaviors in biological systems. We are currently using ANN in two ambitious projects: to predict the physiological effects of opioid analgesics from interactions at the receptor level, and to reverse engineer nanostructures that interact with membranes and proteins in specific ways. If successful, this approach will have implications in basic and applied research, with potential impact in translation medicine. Opioid analgesics used to treat chronic pain lead to addiction and have other serious side effects, including potential respiratory arrest. We are carrying out dynamics simulations to study the roles of agonists and antagonists of mu and delta opioid receptors in activating G-protein. We are combining computational modeling techniques developed for small molecules and for macromolecules to propose chemical substitutions that may lead to a deeper understanding of the effects of both opioid-like and non-opioid compounds. We are focusing our attention on a new series of compounds recently found to have G-protein biased properties. These properties are emerging as essential for avoiding side effects, such as respiratory depression. Using computer simulations, we have identified the residues that potentiate the beta-arrestin-2 bias to the cannabinoid receptor 1 (CB1R) antagonists. We have also provided a molecular rationale to design and synthesize biased antagonists that can have a better anti-diabetic and anti-obesity efficacy. A manuscript is in preparation. We have studied gold nanoparticles in serum and in cell media to optimize strategies for drug delivery and imaging. We have used a recently developed multiscaling technique to realistically represent biological media and are using these approaches to speed up both Monte Carlo and molecular dynamics simulations of multiprotein-multiparticle solutions. We have studied ultrasmall nanostructures of different designs and have demonstrated that they can be rationally designed to interact with biological matte in specific ways. We are currently using the same approach to predict binding modes, affinities, and kinetics in nanoparticle-protein binding, and to assess whether protein function can be modulated. We have carried out quantum chemical calculations to elucidate the fluorination mechanism of diaryliodonium salts at the atomic level. An understanding of this process is essential in the development of novel 18F-labeled PET probes for brain imaging. Ongoing studies of fluorination include the elucidation of the hydrocarbon fluorination catalyzed by CoF3 as well as the trifluoromethylation mechanism of aryliodonium salts with CuCF3. These studies will provide insight for the efficient synthesis of 11C-labeled fluoroform, e.g. (11C)CHF3, for PET imaging. In addition, we have investigated why some high-affinity full 5-HT1B receptor ligands penetrate the blood brain barrier (BBB) while others do not. Through quantum chemical calculations, we proposed that internal H-bonding makes some compounds more amenable to BBB penetration; this work was published. Male contraceptives have been pursued worldwide, but all clinical trials by the World Health Organization have failed at different stages because of long-term health concerns due to their hormonal basis. This project is centered on finding inhibitors of the Gonadotropin Regulated Testicular RNA Helicase (GRTH/DDX25) a novel protein discovered by our IRP collaborators, which may lead to the development of a non-hormonal contraceptive. Using a variety of molecular modeling techniques, we have developed a reliable model of DDX25 that we have used to develop a series of pharmacophores. We are now using the most promising ones to design a macro cyclic peptide inhibitor of DDX25 activity. This project is also the basis for new methodological developments for reverse-engineering of cyclic peptides. GATA2 deficiency is an inherited or sporadic genetic disorder characterized by distinct cellular deficiency, bone marrow failure, various infections, lymphedema, pulmonary alveolar proteinosis, and predisposition to myeloid malignancies. In a multidisciplinary study we have modeled the interactions between GATA2 and the corresponding DNA sequence to rationalize the role of mutations observed in patients with this blood disease. We proposed a set of novel, potentially deleterious mutations. We are investigating the structure and energetics of polymethylated 1,4,7,10-tetraazacyclododecane-1,4,7,10-tetraacetic acid (DOTA) ligand complexed with lanthanide ions. Quantum chemical calculations are being carried out to determine the structural requirements that modulate the thermochemical stability of one isomer over the other. These complexes may find application in magnetic resonance imaging and protein-structure studies. We have modeled a major ternary complex that contains a protein recently discovered by one of our IRP collaborators and shown to be a key player in itch sensation associated with eczema, psoriasis, hepatic cholestasis, uremia, and some neuropathies. Using a battery of experimental and computer modeling tools we were able to design an antibody to inhibit the protein and prevent itch in mice. We proposed a rather general experimental-modeling approach to find inhibitors of complexes with limited structural information based on cross-linking mass spectrometry (XL-MS). We are now applying the same technique to a series of chaperones/co-chaperones (multiprotein complexes and aggregates, where Hsp90 plays the central role) involved in the folding and regulation of protein function. We model proteins based on homology to interpret the structure-function effects of mutations observed in patients. In collaboration with NHGRI, we published a study linking HEPHL1 variants with an abnormal hair phenotype. We are working with NIDDK to study protein-RNA interfaces using computational analyses and experimental verification. We have expanded the capabilities of our MemExp software, which is used world-wide to recover distributions of effective lifetimes from kinetics data. The analysis of time-correlated single photon counting experiments has been improved. A second paper has been published with colleagues at the University of Illinois at Chicago. Another manuscript will be submitted soon with researchers at Florida State University.
Showing the most recent 10 out of 42 publications
