Structure-based design of antiviral drug candidates targeting the SARS-CoV-2 main protease

1In late December 2019, a cluster of pneumonia cases caused by a novel coronavirus (CoV) was reported in Wuhan, China. Genomic sequencing showed that this pathogenic coronavirus is 96.2% identical to a bat coronavirus and shares 79.5% sequence identify to SARS-CoV. This novel coronavirus was named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) by the International Committee on Taxonomy of Viruses, and the pneumonia was designated as COVID-19 by the World Health Organization (WHO) on February 11, 2020. The epidemic spread rapidly to more than 212 countries and was announced as a global health emergency by WHO. No clinically effective vaccines or specific antiviral drugs are currently available for the prevention and treatment of COVID-19 infections. The combination of α-interferon and the anti-HIV drugs Lopinavir/Ritonavir (Kaletra®) has been used, but the curative effect remains very limited and there can be toxic side effects. Remdesivir, a broad-spectrum antiviral drug developed by Gilead Sciences, Inc., is also being explored for the treatment of COVID-19, but more data are needed to prove its efficacy. Specific anti-SARS-CoV-2 drugs with efficiency and safety are urgently needed.

2A maximum likelihood tree based on the genomic sequence showed that the virus falls within the subgenus Sarbecovirus of the genus Betacoronavirus. Coronaviruses are enveloped, positive-sense, single-stranded RNA viruses. The genomic RNA of CoVs is approximately 30 k nt in length with a 5′-cap structure and 3′-poly-A tail, and contains at least 6 open reading frames (ORFs). The first ORF (ORF 1a/b), about two-third of genome length, directly translates two polyproteins: pp1a and pp1ab, because there is an a-1 frameshift between ORF1a and ORF1b. These polyproteins are processed by a main protease (Mpro), also known as the 3C-like protease (3CLpro), and one or two papain-like proteases (PLPs), into 16 non-structural proteins (nsps). These nsps engage in the production of subgenomic RNAs that encode four main structural proteins (envelope (E), membrane (M), spike (S), and nucleocapsid (N) proteins) and other accessory proteins. Therefore, these proteases, especially Mpro, play a vital role in the life cycle of coronavirus.

3Mpro is a three-domain (domains I to III) cysteine protease involved in most maturation cleavage events within the precursor polyprotein. Active Mpro is a homodimer containing two protomers. The CoV Mpro features a non-canonical Cys-His dyad located in the cleft between domains I and II. Mpro is conserved among CoVs and several common features are shared among the substrates of Mpro in different CoVs. The amino acids in substrates from the N terminus to C terminus are numbered as fellows (-P4-P3-P2-P1↓P1′-P2′-P3′-), and the cleavage site is between the P1 and P1′. In particular, a Gln residue is almost always required in the P1 position of the substrates. There is no human homolog of Mpro which makes it an ideal antiviral target.

4The active sites of Mpro are highly conserved among all CoV’s Mpros and are usually composed of four sites (S1′, S1, S2 and S4). By analyzing the substrate-binding pocket of SARS-CoV Mpro (PDB ID: 2H2Z), novel inhibitors targeting the SARS-CoV-2 Mpro were designed and synthesized (Fig. 1). The thiol of a cysteine residue in the S1′ sites anchors inhibitors by a covalent linkage that is important for the inhibitors to maintain antiviral activity. In our design of new inhibitors, an aldehyde was selected as a new warhead in P1 in order to form a covalent bond with cysteine. The reported SARS-CoV Mpro inhibitors often have an (S)-γ-lactam ring that occupies the S1 site of Mpro, and this ring was expected to be a good choice in P1. Furthermore, the S2 site of coronavirus Mpro is usually large enough to accommodate the bigger P2 fragment. To test the importance of different ring systems, a cyclohexyl or 3-fluorophenyl were introduced in P2, with the fluorine expected to enhance activity. An indole group was introduced into P3 in order to form new hydrogen bonds with S4 and improve drug-like properties.

5The synthetic route and chemical structures of the compounds (11a and 11b) are shown in scheme S1. The starting material (N-Boc-L-glutamic acid dimethyl ester 1) was obtained from commercial suppliers and used without further purification to synthesize the key intermediate 3 according to the literature. The intermediates 6a and 6b were synthesized from 4 and acids 5a, 5b. Removal of the t-butoxycarbonyl group from 6a and 6b yielded 7a and 7b. Coupling 7a and 7b with the acid 8 yielded the esters 9a and 9b. The peptidomimetic aldehydes 11a and 11b were approached through a two-step route in which the ester derivatives 9 were first reduced with NaBH4 to generate the primary alcohols 10a and 10b, which were subsequently oxidized into aldehydes 11a and 11b with Dess-Martin Periodinane (DMP).

6Recombinant SARS-CoV-2 Mpro was expressed and purified from Escherichia coli (E. coli). A fluorescently labeled substrate, MCA-AVLQ↓SGFR-Lys (Dnp)-Lys-NH2, derived from the N-terminal auto-cleavage sequence from the viral protease was designed and synthesized for the enzymatic assay.

7Both 11a and 11b exhibited high SARS-CoV-2 Mpro inhibition activity, which reached 100% for 11a and 96% for 11b at 1 μM, respectively. We used a fluorescence resonance energy transfer (FRET)-based cleavage assay to determine the IC50 values. The results revealed excellent inhibitory potency with IC50 values of 0.053 ± 0.005 μM and 0.040 ± 0.002 μM, for 11a and 11b respectively (Fig. 2).

8In order to elucidate the mechanism of inhibition of SARS-CoV-2 Mpro by 11a, we determined the high-resolution crystal structure of this complex at 1.5-Å resolution (table S1). The crystal of Mpro-11a belong to the space group C2 and an asymmetric unit contains only one molecule (table S1). Two molecules (designated protomer A and protomer B) associate into a homodimer around a crystallographic 2-fold symmetry axis (fig. S2). The structure of each protomer contains three domains with the substrate-binding site located in the cleft between domain I and II. At the active site of SARS-CoV-2 Mpro, Cys145 and His41 (Cys-His) form a catalytic dyad (fig. S2).

9The electron density map clearly showed compound 11a in the substrate binding pocket of SARS-CoV-2 Mpro in an extended conformation (Fig. 3A and fig. S3, A and B). Details of the interaction are shown in Fig. 3, B and C. The electron density shows that the C of the aldehyde group of 11a and the catalytic site Cys145 of SARS-CoV-2 Mpro form a standard 1.8-Å C–S covalent bond. The oxygen atom of the aldehyde group also plays a crucial role in stabilizing the conformations of the inhibitor by forming a 2.9-Å hydrogen bond with the backbone of residues Cys145 in the S1′ site. The (S)-γ-lactam ring of 11a at P1 fits well into the S1 site. The oxygen of the (S)-γ-lactam group forms a 2.7-Å hydrogen bond with the side chain of His163. The main chain of Phe140 and side chain of Glu166 also participate in stabilizing the (S)-γ-lactam ring by forming 3.2-Å and 3.0-Å hydrogen bonds with its NH group, respectively. In addition, the amide bonds on the chain of 11a are hydrogen-bonded with the main chains of His164 (3.2 Å) and Glu166 (2.8 Å), respectively. The cyclohexyl moiety of 11a at P2 deeply inserts into the S2 site, stacking with the imidazole ring of His41. The cyclohexyl group is also surrounded by the side chains of Met49, Tyr54, Met165, Asp187 and Arg188, producing extensive hydrophobic interactions. The indole group of 11a at P3 is exposed to solvent (S4 site) and is stabilized by Glu166 through a 2.6-Å hydrogen bond. The side chains of residues Pro168 and Gln189 interact with the indole group of 11a through hydrophobic interactions. Interestingly, multiple water molecules (named W1-W6) play an important role in binding 11a. W1 interacts with the amide bonds of 11a through a 2.9-Å hydrogen bond, whereas W2-6 form a number of hydrogen bonds with the aldehyde group of 11a and the residues of Asn142, Gly143, Thr26, Thr25, His41 and Cys44, which contributes to stabilizing 11a in the binding pocket.

10The crystal structure of SARS-CoV-2 Mpro in complex with 11b is very similar to that of the 11a complex and shows a similar inhibitor binding mode (Fig. 3D and figs. S3, C and D, and S4A). The difference in binding mode is most probably due to the 3-fluorophenyl group of 11b at P2. Compared with the cyclohexyl group in 11a, the 3-fluorophenyl group undergoes a significant downward rotation (Fig. 3D). The side chains of residues His41, Met49, Met165, Val186, Asp187 and Arg188 interact with this aryl group through hydrophobic interactions and the side chain of Gln189 stabilizes the 3-fluorophenyl group with an additional 3.0-Å hydrogen bond (Fig. 3, E and F). In short, these two crystal structures reveal a similar inhibitory mechanism in which both compounds occupy the substrate-binding pocket and block the enzyme activity of SARS-CoV-2 Mpro.

11Compared with those of N1, N3 and N9 in SARS-CoV Mpro complex structures reported previously, the binding modes of 11a and 11b in SARS-CoV-2 Mpro complex structures are similar and the differences among these overall structures are small (Fig. 4 and fig. S4, B to F). The differences mainly lie in the interactions at S1′, S2 and S4 subsites, possibly due to various sizes of functional groups at corresponding P1′, P2 and P4 sites in the inhibitors (Fig. 4, A and C).

12To further substantiate the enzyme inhibition results, we evaluated the ability of these compounds to inhibit SARS-CoV-2 in vitro (Fig. 5 and fig. S5). As shown in Fig. 5, compounds 11a and 11b exhibited good anti-SARS-CoV-2-infection activity in cell culture with EC50 values of 0.53 ± 0.01 μM and 0.72 ± 0.09 μM using plaque-reduction assay, respectively. Neither compound caused significant cytotoxicity, with half cytotoxic concentration (CC50) values of >100 μM, yielding selectivity indices (SI) for 11a and 11b of >189 and >139, respectively. Both immunofluorescence and quantitative real-time PCR were also employed to monitor the antiviral activity of 11a and 11b. The results show 11a and 11b exhibit a good antiviral effect on SARS-CoV-2 (Fig. 5 and fig. S5).

13To explore the further druggability of the compounds 11a and 11b, both of the compounds were evaluated for their pharmacokinetic (PK) properties. As shown in table S2, compound 11a given intraperitoneally (5 mg/kg) and intravenously (5 mg/kg) displayed a half-life (T1/2) of 4.27 hours and 4.41 hours, respectively, and a high maximal concentration (Cmax = 2394 ng/mL) and a good bioavailability of 87.8% were observed when the compound 11a was given intraperitoneally. Metabolic stability of 11a in mice was also good (Clearance (CL) = 17.4 mL/min/mg). When administered intraperitoneally (20 mg/kg), subcutaneously (5 mg/kg) and intravenously (5 mg/kg), compound 11b also showed good PK properties (the bioavailability of intraperitoneally and subcutaneously are more than 80%, and a longer T1/2 of 5.21 hours when 11b was given intraperitoneally). Considering the danger of COVID-19, we selected the intravenous drip administration to further study for the reason that value of the area under the curve (AUC) is high and the effect is rapid. Compared with 11a administrated intravenously, the T1/2 (1.65h) of 11b is shorter and the clearance rate is faster (CL = 20.6 mL/min/mg). Compound 11a was selected for further investigation with intravenous drip dosing in Sprague-Dawley (SD) rats and Beagle dogs. The results showed (table S3) that 11a exhibited long T1/2 (SD rat, 7.6 hours and Beagle dog, 5.5h), low clearance rate (rat, 4.01 mL/min/kg and dog, 5.8 mL/min/kg) and high AUC value (rat, 41500 hours*ng/mL and dog, 14900 hours*ng/mL)). Those above PK results indicate that compound 11a is worth to warrant further study.

14An in vivo toxicity study (table S4) of 11a has been carried out on SD rats and Beagle dogs. The acute toxicity of 11a was measured on SD rats. No SD rats died after receiving 40 mg/kg by intravenous drip administration. When the dosage was raised to 60 mg/kg, one of four SD rats died. The dose range toxicity study of 11a was conducted for seven days at dosing levels of 2, 6, and 18 mg/kg on SD rats and at 10-40 mg/kg on Beagle dogs. All animals received once daily dosing (QD), by intravenous drip, and all animals were clinically observed at least once a day. No obvious toxicity was observed in either group. These above data indicated that 11a is good candidate for further clinical studies.