My initial research into protein structure prediction and modeling at UCLA formed the basis for my current interest in the molecular properties of amyloid fibrils: the unbranched, ordered protein aggregates associated with several diseases, including Alzheimer's disease and type II diabetes. My work enabled the development of a novel, rational design approach for creating D-amino acid peptide inhibitors of amyloid fibril formation. I have demonstrated that these inhibitors are effective on a construct of the tau protein associated with Alzheimer's disease in vitro and we are currently studying their effects in vivo. We also developed the first structure-based algorithm for predicting short protein segments with a propensity to form amyloid-like fibrils. I was able to confirm that several of the protein segments predicted to fibrillize in fact do so, and in several cases, I was able to determine atomic level structures for the cross-beta fibril spines adopted by the segments. Further, I studied the interaction of full-length insulin with fibril-forming peptides in its core, demonstrating that small protein segments can provide a nucleus for fibril growth and provide targets for structure-based inhibitor design. My work defines a protocol for creating potential therapeutics of amyloid diseases: (1) fibril core prediction, (2) atomic level structure determination, (3) protease resistant inhibitor design, and (4) in vitro fibrillation inhibition assessment.

I investigated a number of biological systems in my preliminary work on protein structure and modeling, including (1) predicting the structures of several domains from the mammalian major vault protein, (2) modeling the molecular interaction between a human phosphorlyase and the TCL1 oncoprotein, and (3) building a model of an enzyme from the parasite Trichomas vaginalis. My work enabled my collaborators and me to further our understanding of the functionality of these systems at the molecular level.