Inhibitors GC373, GC376, and their derivatives, as well as the FRET assay peptide substrate, Abz-SVTLQSG-Y(NO2)-R, were synthesized according to methods previously described (Vuong et al., 2020; Vuong et al., 2021).

The FIPV WSU-79/1146 M^pro gene was synthesized (Bio Basic, Canada) and cloned into pET SUMO expression vector (Invitrogen, United States), generating a fusion protein with a His-tagged SUMO domain at the N-terminus. The construct was transformed into BL21 (DE3) Escherichia coli, where protein expression was induced with 0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) once OD₆₀₀ reached 0.5–0.6 and then grown for an additional 5 h at 32°C. The cells were harvested by centrifugation (5,000 ×g for 20 min at 4°C), suspended in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 5 mM imidazole, pH 7.8), and lysed using the Emulsiflex C3 High Pressure Homogenizer. Cellular debris was removed by centrifugation at 20,000 ×g for 45 min at 4°C. The isolated supernatant was applied onto a Ni-NTA column (Qiagen, Canada), the resin was washed with 10 column volumes of lysis buffer containing 20 mM imidazole, and the protein was eluted with a step gradient of 100–1000 mM imidazole in lysis buffer. The eluted fractions were analyzed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis, pooled based on purity and dialyzed against 20 mM Tris-HCl, 150 mM NaCl, 1 mM TCEP, pH 7.8, for 2 h at 4°C. The SUMO tag was cleaved off using His-tagged SUMO protease (McLab, United States) and both the N-terminal SUMO tag and SUMO protease were removed by passing the protein sample through a Ni-NTA column. The flow-through containing FIPV M^pro was further purified using size exclusion chromatography (Superdex increase 10/300 GL, GE Healthcare), with buffer containing 20 mM Tris-HCl, 150 mM NaCl, 1 mM TCEP, pH 7.8. The fractions containing FIPV M^pro were pooled and concentrated using an Amicon Ultra-15 filter with a MWCO of 10 kDa.

Purified FIPV M^pro was dialyzed against 5 mM Tris-HCl, 5 mM NaCl, 1 mM TCEP, pH 7.8 buffer at 4°C overnight and concentrated to 10 mg/ml using an Amicon Ultra-15 filter with a MWCO of 10 kDa. FIPV M^pro was incubated with GC373 or GC376 (5× molar excess) at 4°C for 2 h prior to crystallization. The protein was subjected to the PACT and JCSG crystallization screens (Molecular Dimensions, United States). Crystals were observed with sitting drop trays at room temperature. The crystals of FIPV M^pro with GC376 were obtained using a protein:buffer ratio of 1:1 with 2.0 M ammonium sulfate, 0.1 M Bis-Tris, pH 5.5. The crystals of FIPV M^pro with GC373 were obtained using a 2:1, protein:buffer ratio with 0.2 M calcium chloride dihydrate, 0.1 M MES, 20% (w/v) PEG 6000, pH 6.0. The crystals were frozen in liquid nitrogen using 19% glycerol as a cryoprotectant.

The diffraction data were collected at Canadian Light Source using beamline CMCF-BM (08B1) and PILATUS3 S 6M detector, Saskatchewan, Canada. Several data sets were collected from different crystals and were processed using SCALA and XDS. The diffraction data set of the GC373 was processed to 2.05 Å, in a monoclinic C2 space group, while the GC376 were processed to 1.93 Å, in an orthorhombic P2₁2₁2₁ space group. The structures were determined by molecular replacement using the crystal structure of the apo-FIPV Mpro (PDB entry: 5EU8) as the search model. GC376 and GC373 were manually fit in the density using Coot. The structures were then refined by using the Phenix software. Data statistics, processing, and model refinement are given in Supplementary Table S1.

A fluorescence resonance energy transfer (FRET)–based cleavage assay was performed using a synthetic peptide fluorescent substrate containing the cleavage site of FIPV M^pro [Abz-SVTLQ↓SG-Tyr (NO2)-R] as described previously (Vuong et al., 2020; Vuong et al., 2021). For K_i determination, 50 nM FIPV M^pro was preincubated with GC376 in the concentration range of 0.01–0.4 µM for 10 min at 37°C. The enzymatic reactions using 1–500 μM of FRET substrate in activity buffer (25 mM Bis-Tris, 1 mM DTT, pH 7.0) were started with the addition of protease. For IC₅₀ determination, 100 nM of FIPV M^pro was incubated with an inhibitor concentration range of 0.25 nM–100 µM in activity buffer. The reaction was started with 40 µM of FRET substrate. The fluorescence signal of the FRET peptide cleavage product was monitored at an emission wavelength of 420 nm with excitation at 320 nm, using a Cytation 5 Imaging Multi-Mode Reader (BioTek) for 7 min at 37°C. The kinetic data were analyzed using computer-fit calculation (Prism 9.0, GraphPad Software). The slopes of the Lineweaver–Burk plots were plotted versus the concentration of GC376, and the K_i was determined from the x-axis intercept as −K_i. The experiments were performed in triplicate.

Crystal structures of FIPV M^pro in complex with the drug GC373 (PDB: 7SNA) and prodrug GC376 (PDB: 7SMV) were solved to 2.05 Å and 1.93 Å, respectively (Supplementary Table S1). GC376 being the dipeptidyl aldehyde bisulfite adduct form of the drug converts into the active-form aldehyde GC373, thus making both structures identical. In both structures, FIPV M^pro crystallized as a dimer, with each protomer being comprised of three domains (Figure 1A), similar to other viral M^pro. Domains I and II have a six-stranded antiparallel β-barrel structure, and domain III is a globular cluster of five antiparallel α-helices, connected to domain II by a long loop. The active site of FIPV M^pro contains a Cys144–His41 catalytic dyad located in a cleft between domain I and domain II. Domain III regulates the dimerization of the M^pro which is required for its catalytic activity. The N-terminal residues (N-finger) of protomer A fits between domains II and III of the protomer A and interacts with residues in domain II of protomer B helping shape the S1 substrate-binding subsite in the active site.

The GC373 inhibitor covalently binds FIPV M^pro and is stabilized by hydrogen bonding and hydrophobic interactions in a similar manner to SARS-CoV and SARS-CoV-2 M^pro. In both structures, a covalent bond between Cys144 and the aldehyde of the feline drug reveals that the bisulfite leaving group indeed was removed upon binding (Figure 1B). Weak H-bonding was observed between the oxyanion of the inhibitor and His41, the general base in the catalytic dyad, which is distinct from SARS-CoV-2 (Figure 2A) (Vuong et al., 2020). For the P1 position of the inhibitor, the Nγ of the lactam ring sits in the S1 pocket and forms a H-bond with the carbonyl oxygen of Phe139 (Supplementary Figure S1A), a conserved feature in other M^pro structures with GC373 (Arutyunova et al., 2021). The S2 pocket that supports hydrophobic interactions of a Leu moiety is formed with His41, Ile51, and Leu164 (Supplementary Figure S2A). This differs from the stabilization network found in SARS-CoV-2 M^pro for the same feline drug. Meanwhile, the P3 benzyl moiety interacts with the P1 lactam ring by pi stacking, similar to that observed in the SARS-CoV-2 M^pro structure (Vuong et al., 2020).

FIPV and SARS-CoV-2 M^pro share 60% sequence similarity, however, the active-site region exhibits even greater conservation (Supplementary Figure S3). Comparing the overall structures, both bound to GC373, the RMSD was calculated to be 1.16 Å. While the active site cavity of both FIPV and SARS-CoV-2 M^pro are composed of identical residues, the structures of M^pro in complex with GC373 or GC376, reveal some differences in inhibitor binding. In FIPV M^pro, GC373 is stabilized in the active site by H-bonding with His41 (Figure 2A, Supplementary Figure S1B). By contrast, SARS-CoV (not shown) and SARS-CoV-2 M^pro form a stable acyl-intermediate with the drug through a H-bonding network with the backbones of Cys145, Ser144, and Gly143 residues (Figure 2B, Supplementary Figure S6). Furthermore, the S2 pocket that supports hydrophobic interactions of the drug's Leu moiety is formed with His41, Ile51, and Leu164 in FIPV M^pro, but with His41, Met49, and Met165 in SARS-CoV-2 M^pro (Supplementary Figure S2B). In SARS-CoV-2 M^pro, Gln189 plays an integral role in stabilizing the dipeptide backbone of the inhibitor (Bai et al., 2021), however in FIPV M^pro, we observe an unstructured loop fit between the S3 and S4 pocket to form hydrophobic interactions, thus further supporting binding of the inhibitor (Supplementary Figures S4, S5). In FIPV M^pro, Ser1 of the N-terminal finger from protomer B forms a weak H-bond (3.8 Å) with the cyclic glutamine analog nitrogen of GC373 in the active site of protomer A, however, this is not seen in SARS-CoV-2 M^pro (Figure 3A). Furthermore, the side chain hydroxyl group and backbone amide of Ser1 in protomer B form H-bonds with Glu165 and Phe139 in the active site of protomer A. This is comparable to SARS-CoV and SARS-CoV-2 M^pro structures where Ser1 of the N-terminal finger (protomer B) forms H-bonds with Glu166 and Phe140 (protomer A) to shape the P1 position (Figure 3B). These structural changes led us to examine the inhibitory parameters of GC376 with FIPV M^pro for comparison with SARS-CoV and SARS-CoV-2 M^pro.

IC₅₀ and K_i values quantitatively reflect the potency and affinity of a drug and are therefore important parameters to consider when undergoing inhibitor design. First, we determined the catalytic parameters of FIPV M^pro using our synthetic peptide FRET-substrate (Supplementary Table S2) (Arutyunova et al., 2021). Interestingly, feline coronavirus protease exhibited 24 times slower catalytic turnover rate than M^pro of SARS-CoV-2 with the same substrate, and a lower K_m value. The K_i values for GC376 inhibition were determined to be 2.1 nM for FIPV M^pro (Figure 4), lower in comparison to previously determined K_i values of SARS-CoV and SARS-CoV-2 M^pro, which were 20 and 40 nM, respectively (Table 1) (Vuong et al., 2020; Arutyunova et al., 2021).

We recently demonstrated that derivatives of GC376 with singly or doubly modified constituents resulted in improved potency with SARS-CoV-2 M^pro, having lower IC₅₀ and EC₅₀ values (Vuong et al., 2021). The singly modified compounds contain derivatives that include a cyclopropyl group (1a) in the P2 position where the S2 pocket typically recognizes a Leu residues side chain, and a 3-fluorobenzyl (2c) or 3-chlorophenylethyl group (2d) in the P3 position recognized by the S4 pocket (Table 2). The doubly modified compounds all included a cyclopropyl group in the P2 position, as well as a 3-fluorobenzyl (2c), 3-chlorophenylethyl (2d), or 4-methoxyindole (2e) group at the P3 position (Table 3). In order to assess if these inhibitor derivatives also have improved potency with FIPV M^pro as they did with SARS-CoV-2 M^pro, IC₅₀ values were calculated and compared. In vitro analysis with purified FIPV M^pro revealed that the doubly modified inhibitor had stronger effects on IC₅₀ values than a singly modified inhibitor, bringing the IC₅₀ to the double-digit nanomolar range. This is a similar trend as seen with SARS-CoV-2 M^pro using the same doubly modified inhibitors. Overall, this suggests that inhibitors targeting FIPV M^pro can be improved and warrant further assessment in cellular and animal studies.