PhD Oral Exam - Ridwan Sulaimon, Chemistry

When studying for a doctoral degree (PhD), candidates submit a thesis that provides a critical review of the current state of knowledge of the thesis subject as well as the student’s own contributions to the subject. The distinguishing criterion of doctoral graduate research is a significant and original contribution to knowledge.

Once accepted, the candidate presents the thesis orally. This oral exam is open to the public.

Abstract

Computer-aided Design of Inhibitors for Botulinum Neurotoxin Metalloprotease and Insights into Membrane Permeation of Drug Candidates from Molecular Simulations Abstract Botulinum neurotoxins (BoNTs), one of the deadliest toxins known, are responsible for botulism, a neuroparalytic condition that can cause flaccid muscle paralysis and, in severe cases, death. Although rare, botulism remains a serious health concern. Despite their toxicity, BoNTs are widely used in cosmetic and therapeutic applications, with occasional adverse effects arising from off-target activity. Current treatments are limited because available antitoxins provide only passive immunity and cannot neutralize BoNT once it has been internalized into neurons. Effective therapeutic inhibitors must therefore combine strong and selective binding with the ability to permeate cellular membranes successfully. In this thesis, a computational framework is developed that integrates computer-aided drug design (CADD), machine learning (ML), and enhanced molecular dynamics (MD) simulations to accelerate the discovery of novel small-molecule BoNT inhibitors and to elucidate the permeation mechanisms of promising candidates across lipid bilayers. Structure-based virtual screening followed by MD refinement and binding free-energy calculations led to the identification of sixteen candidate inhibitors, including Dinoprost, an FDA-approved muscle contraction stimulant, all with predicted binding affinities exceeding those of known inhibitors. Analyses of protein–ligand interactions revealed that effective inhibition is not restricted to classical zinc-chelating hydroxamate groups. To further elucidate molecular determinants of inhibitory activity, machine learning models were developed using experimentally validated compounds. Among the evaluated models, PubChem fingerprint descriptors combined with the XGBoost algorithm achieved the highest predictive performance. Feature importance analysis identified key structural determinants of inhibitory activity. These features appear to promote favorable hydrogen-bonding and hydrophobic interactions within the BoNT/A active site, particularly in regions surrounding the catalytic zinc ion. Finally, the membrane permeation of selected BoNT/A inhibitors was investigated using umbrella sampling molecular dynamics simulations and the inhomogeneous solubility-diffusion model. The results showed that electrostatic charge, hydration retention, and membrane composition strongly influence permeability. Cholesterol was also found to reduce the permeability of polar solutes by increasing membrane order and reducing free volume. Together, this work establishes an integrated computational framework for BoNT inhibitor discovery that advances understanding of both binding and membrane permeation, identifies key structural determinants of inhibitory activity, and supports the use of machine learning and molecular simulation as complementary tools for accelerating therapeutic development against BoNT.