Investigation of SARS-CoV-2 Proteins as Targets for Novel Therapies Against COVID-19

The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), presented an unprecedented global health challenge. Despite the rapid development of vaccines and antivirals, the ongoing emergence of new variants and associated side effects highlight the need for novel antiviral strategies. This thesis explores the usage of molecular dynamics (MD) simulations in advancing our understanding of SARS-CoV-2 protein dynamics and facilitating drug discovery efforts. By focusing on non-structural proteins (nsps) of SARS-CoV-2, this thesis aims to identify novel antivirals and characterize the dynamic interactions between these proteins and potential inhibitors. A general introduction is given in Chapter 2 to provide a background on the COVID-19 pandemic and SARS-CoV-2 biology, outlining the proteome of the virus and discussing targets that have been used as antiviral therapies. The following chapters are self-contained studies exploring various nsps of SARS-CoV-2 as antiviral targets. To demonstrate the workflow of the investigation of compounds against viral targets, Chapter 3 presents a study targeting the exoribonuclease (ExoN) domain of nsp14. Using an in-house library of COVID-19 related compounds, molecular docking was performed. Trisjuglone, ararobinol, corilagin, and naphthofluorescein were identified as lead compounds. Protein-RNA docking was performed, showing that when ligands are bound to the ExoN site, the binding of RNA is disrupted. MD simulations were performed for the nsp10-nsp14 complex bound with lead compounds and protein dynamics was analysed. It was found that changes in protein structure were modest in response to ligand binding, with principal components analysis (PCA) showing that energetically favourable conformations were achieved. Ligand binding was stable and largely driven by hydrogen bond formation, with residues involved in RNA-binding frequently forming contacts. Encoded by nsp3, the SARS-CoV-2 papain-like protease (PLpro) was also studied in Chapter 3 using previously identified lead compounds: cyanidin-3-O-glucoside, hypericin, and rutin. As well as assessing stability within the naphthalene binding site, binding free energy calculations were performed using molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) methods. Chapter 4 extends the investigation of SARS-CoV-2 PLpro using a combination of different in silico methods to characterise the lead compound, hypericin, followed by in vitro validation of inhibitory behaviour. MD simulations were performed on an extended timescale, showing that PLpro adopts two distinct conformations characterised by the expansion and contraction of the ubiquitin-like domain. Ligands were found to bind in a stable manner to the naphthalene-binding site of PLpro, driven by interactions with blocking loop 2. Using adaptive protein energy landscape exploration (PELE) Monte Carlo (MC) simulations, unbiased binding sites were identified to reinforce the preferential binding of ligands to canonical sites associated with inhibition of PLpro. Using enzymatic assays, hypericin demonstrated inhibition of protease and deubiquitinase activities within biologically relevant ranges. The characterisation of novel inhibitors against SARS-CoV-2 Mpro, encoded by nsp5, is presented in Chapter 5 utilising a similar approach. Identified from previous docking studies, hypericin and cyanidin-3-O-glucoside were explored as inhibitors against Mpro. MD simulations revealed modest changes in overall protein structure in response to ligand binding, with hypericin showing stable binding to the catalytic site. Adaptive-PELE MC simulations showed that hypericin also demonstrated binding to potential allosteric sites located in domain III of the Mpro dimer. From in vitro fluorogenic Mpro assays, both compounds inhibited protease activities within biologically relevant concentrations. The SARS-CoV-2 nsp10-nsp16 methyltransferase complex was studied in Chapter 6. Small molecules were docked to the S-adenosyl-L-methionine (SAM)-binding site of nsp16, and potential lead compounds were assessed with MD simulations. Among the six lead compounds identified through the initial screening process, only oleuropein demonstrated comparable activity to the control. Investigation of nsp10-nsp16 was extended in Chapter 7, with additional MD simulations performed for nine compounds. From this, a subset of compounds was further evaluated using a longer timescale for production runs: sinefungin, verbascoside, and ligstroside. These extended simulations showed more stable binding, with ligands forming contacts with key residues of the SAM-binding pocket. Through the investigation of various SARS-CoV-2 proteins, a substantial collection of MD simulations was accumulated, culminating in the creation of a trajectory database presented in Chapter 7. The database is comprised of over 300 MD simulations, including simulations involving variant spike proteins, as well as the studies on nsp3, nsp5, nsp14, and nsp16 detailed in this thesis. The repository of COVID-19 related MD simulations has been made publicly accessible at https://epimedlab.org/trajectories/. This thesis represents a comprehensive investigation of SARS-CoV-2 nsps, resulting in the identification of potential antivirals. The trajectory database established in this study serves as a valuable resource for the scientific community, facilitating future research and development of novel antiviral strategies. Beyond its immediate application to COVID-19, this work offers valuable insights into drug discovery workflows, enhancing our preparedness for potential future pandemics.