Theses Doctoral

High-throughput methods to investigate the function and pharmacological inhibition of viral proteases

Hong, Seo Jung

Viral pathogens have plagued human civilizations since ancient times and continue to pose a serious and constant global threat to not only human health but all facets of life. To date, more than 200 viruses capable of infecting humans have been identified, and the combined efforts of the academic and pharmaceutical sectors have yielded both extensive understanding of the biology and pathology of the viral infections as well as breakthrough interventions against a number of devastating diseases such as those caused by HIV (human immunodeficiency virus) and HCV (hepatitis C virus). In late 2019, SARS-CoV-2 (severe acute respiratory syndrome coronavirus 2), the etiological agent of COVID-19 (coronavirus disease 2019), rapidly spread worldwide, leading to detrimental public health and socioeconomic crises. While the immediate response of the scientific community to the pandemic, which involved investigation of the disease and discovery of several therapeutic options at an unprecedented pace, has been impressive, this recent experience exposes the serious need to continuously fortify our fundamental knowledge of virology and equip our antiviral arsenal in preparation for future outbreaks. Moreover, given the scale of the challenge at hand, it highlights the value in the development and application of experimental approaches that accelerate the rate at which this information is obtained.

In this dissertation, we utilize various techniques that allow high-throughput analysis of the SARS-CoV-2 3CL (3-chymotrypsin-like) protease to better understand its functional landscape as a favorable therapeutic target of the virus, and to investigate its response in developing resistance against the clinically used protease inhibitor, nirmatrelvir, at scale. We further expand our efforts to develop a platform for multiplexed drug screening that has the capacity to detect viral protease inhibitors for not only coronaviruses but also other targets across six additional virus families. Using this approach, we are able to rapidly identify broad-acting inhibitors, which are favorable for pandemic preparedness purposes where the exact nature of the future threat is difficult to predict a priori.

To perform our studies, we make use of a variety of model systems, from a simple yeast-based system for detecting viral protease activity to the passaging of live virus within cultured human cells. Utilizing our yeast-based reporters, we comprehensively profile the activity landscape of all possible single mutants of the SARS-CoV-2 3CL protease via deep mutational scanning (DMS), uncovering its general malleability while also identifying several immutable regions within the enzyme that can serve as targets for the design of the next generation of protease inhibitors. Among the sites that show tolerance to changes, we predict E166 to be a residue that may confer nirmatrelvir resistance upon mutation based on available structural data which reveal its critical role in the binding of the drug to the active site. We prove this to be true by demonstrating a 265-fold loss in EC50 for the E166V mutant relative to the wild type protease within the recombinant virus. Recognizing that the plasticity of the enzyme could translate to a lower genetic barrier to resistance, we extend our investigation to study the whole virus response to nirmatrelvir at scale via in vitro passaging of SARS-CoV-2 in increasing concentrations of the drug. Upon examining 53 independent viral lineages to explore the ways by which resistance can be acquired, we identify a total of 23 mutations that arise in often non-overlapping combinations, with T21I, P252L, and T304I being the most common precursor mutations within all analyzed mutational trajectories. Validation of select single, double, and triple mutants based on the frequency of their appearance reveals that most single mutations, including the aforementioned founder mutations, confer low-level resistance (~5 – 6 fold) while greater resistance is acquired with the accumulation of additional mutations.

Moreover, some mutations, such as T21I and L50F, appear to mediate, through a compensatory mechanism, the acquisition of secondary mutations such as E166V, which alone may confer much greater resistance but also cause significant loss in replicative fitness. Overall, the myriad of solutions that exist for the virus to escape the drug further corroborate the malleability of the SARS-CoV-2 3CL protease as established by our initial DMS study. These findings also establish a foundation for extended analysis of the mechanism of resistance and informed drug design. Lastly, by introducing additional viral proteases into our yeast cellular chassis and labelling each model with a set of unique DNA-barcodes, we develop a method of screening a pool of 40 unique protease targets simultaneously against small molecule libraries. Using this platform, we screen 2,480 structurally diverse compounds, and identify and orthogonally validate a series of broad-acting coronavirus 3CL protease inhibitors with a chromen-2-one structure. Together, the work described in this thesis underline the importance of innovative high-throughput approaches to investigating biology as demonstrated by their application to viral protease research.


This item is currently under embargo. It will be available starting 2028-03-07.

More About This Work

Academic Units
Pathobiology and Molecular Medicine
Thesis Advisors
Chavez, Alejandro
Ph.D., Columbia University
Published Here
March 29, 2023