Optimization of quenched fluorescent peptide substrates of SARS-CoV-2 3CL pro main protease (Mpro) from proteomic identification of P6—P6' active site specificity

Abstract

ABSTRACT SARS-CoV-2 3C-like main protease (3CL pro ) is essential for protein excision from the viral polyprotein. 3CL pro inhibitor drug development to block SARS-CoV-2 replication focuses on the catalytic non-prime (P) side for specificity and potency, but the importance of the prime (P') side in substrate specificity and for drug development remains underappreciated. We determined the P6–P6' specificity for 3CL pro from >800 cleavage sites that we identified using Proteomic Identification of Cleavage site Specificity (PICS). Cleavage occurred after the canonical P1-Gln and non-canonical P1-His and P1-Met residues. Moreover, P3 showed a preference for Arg/Lys and P3' for His. Essential H-bonds between the N-terminal Ser1 of protomer-B in 3CL pro dimers form with P1-His, but not with P1-Met. Nonetheless, cleavage occurs at P1-Met456 in native MAP4K5. Elevated reactive oxygen species in SARS-CoV-2 infection oxidize methionines. Molecular simulations revealed P1-Met OX forms an H-bond with Ser1 and notably, strong positive cooperativity between P1-Met with P3'-His was revealed, which enhanced peptide-cleavage rates. The highly plastic S3' subsite accommodates P3′-His that displays stabilizing backbone H-bonds with Thr25 lying central in a “‘threonine trio” (Thr24-Thr25-Thr26) in the P'-binding domain I. Molecular docking simulations unveiled structure-activity relationships impacting 3CL pro -substrate interactions, and the role of these structural determinants was confirmed by MALDI-TOF-MS cleavage assays of P1'- and P3'-positional scanning peptide libraries carrying a 2nd optimal cut-site as an internal positive control. These data informed the design of two new and highly soluble 3CL pro quenched-fluorescent peptide substrates for improved FRET monitoring of 3CL pro activity with 15× improved sensitivity over current assays. IMPORTANCE From global proteomics identification of >800 cleavage sites, we characterized the P6–P6' active site specificity of SARS-CoV-2 3CL pro using proteome-derived peptide library screens, molecular modeling simulations, and focussed positional peptide libraries. In P1', we show that alanine and serine are cleaved 3× faster than glycine and the hydrophobic small amino acids Leu, Ile, or Val prevent cleavage of otherwise optimal non-prime sequences. In characterizing non-canonical non-prime P1 specificity, we explored the unusual P1-Met specificity, discovering enhanced cleavage when in the oxidized state (P1-Met OX ). We unveiled unexpected amino acid cooperativity at P1-Met with P3′-His and noncanonical P1-His with P2-Phe, and the importance of the threonine trio (Thr24-Thr25-Thr26) in the prime side binding domain I in defining prime side binding in SARS-CoV-2 3CL pro . From these analyses, we rationally designed quenched-fluorescence natural amino acid peptide substrates with >15× improved sensitivity and high peptide solubility, facilitating handling and application for screening of new antiviral drugs.