Author ORCID Identifier

0000-0002-5416-5789

Date of Award

12-16-2019

Degree Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Biology

First Advisor

Dr. Irene Weber

Second Advisor

Dr. Robert Harrison

Third Advisor

Dr. Giovanni Gadda

Abstract

Globally, 62% of 38 million HIV-infected people are receiving antiretroviral therapy. Inhibitors targeting the viral protease have been clinically successful as 9 protease inhibitors (PIs) have been approved by the FDA since 1995. However, drug resistance arising by mutations in the protease undermines effective treatment. Analysis of protease and its mutants by structural biology methods paired with enzymology has given insight into the molecular mechanisms for drug resistance and guided new strategies for inhibitor design. Recently, highly resistant variants of HIV-1 protease from clinical isolates have been identified with ~20 mutations and several orders of magnitude worse binding affinity for clinical PIs such as darunavir. Three such mutants, PR20, PRS17, and PRS5B, are the focus of this body of work and show 800-10,000-fold less susceptibility to darunavir than wild-type protease. Understanding the molecular mechanisms driving the extreme drug-resistance of these three protease mutants aid rational drug design efforts to fight the HIV/AIDS pandemic.

Structure-guided strategies for drug design have resulted in an investigational inhibitor, GRL-142, with modified ligands derived from the darunavir scaffold that shows 16-fold better inhibition than darunavir of resistant mutant PR20. The crystal structure of PR20 in complex with GRL-142 reveals how the expanded binding pocket, dynamic flaps, and faster dimer dissociation of PR20 are counteracted by the larger moieties of GRL-142.

Resistant variant PRS17, which was rationally selected from the HIVdb genotype-phenotype database by machine learning, shows ~3-fold better inhibition by peptide substrate analogs compared to wild-type protease. Crystal structures of PRS17 with substrate analogs show a major effect of drug-resistance mutations V82S and G48V improving interactions with substrates consistent with better inhibition, suggesting a novel mechanism for resistance.

Finally, structural studies of another mutant selected by machine learning, PRS5B, reveal coordinated structural changes leading to decreased intra-subunit interactions and intermediate levels of resistance to PIs. The sum of knowledge on resistant variants PR20, PRS17, and PRS5B illuminates the evolution of HIV-1 protease in the era of accessible PI treatments. These results illustrate the power of combining structural analysis of proteins with enzyme kinetics for combatting drug resistant HIV.

File Upload Confirmation

1

Available for download on Tuesday, December 01, 2020

Share

COinS