3.1.5  Ketoamide inhibitors
Liu et al. reported dipeptidic α‐ketoamides as broad‐spectrum antiviral agents against the main proteases of human α and β‐CoVs as well as the 3C protease of enterovirus. The α‐ketoamide warhead group was promising, as it provides two hydrogen bond acceptors—one from the keto and one from the amide oxygen—whereas other warhead groups, such as Michael acceptor esters and aldehydes, provide only one hydrogen bond acceptor. Compound 63 was identified as SARS‐CoV Mpro inhibitor with an IC50 value of 1.95 µM (Figure 19). 156
Figure 19  Ketoamide inhibitors targeting SARS‐CoV‐2 Mpro. Mpro, main protease; SARS‐CoV, severe acute respiratory syndrome coronavirus Taking 63 as a lead, aided by its X‐ray structure in complex with SARS‐CoV‐1, HCoV‐NL63, and coxsackievirus Mpros, systematic structural modifications were investigated, focusing on the P2‐moiety. As a result, the replacement of P2‐phenyl with P2‐cyclohexyl (64) was found to be the best substitution, while P2‐cyclopentyl (65) showed similar potency against the enzyme SARS‐CoV‐1 Mpro. In Huh7 cells, 64 also showed strong antiviral activity with an EC50 of 400 pM, but in Vero cells the antiviral activity of 64 was drastically reduced to 5 µM. This compound also exhibited antiviral activity against a range of enteroviruses in various cell lines.
Due to the high similarity between SARS‐CoV‐1 Mpro and SARS‐CoV‐2 Mpro authors speculated that 64 was likely to inhibit the new virus as well. Zhang et al. recently reported this molecule as a SARS‐CoV‐2 Mpro inhibitor with an IC50 value of 0.18 µM. They first resolved the unliganded crystal structure of SARS‐CoV‐2 Mpro (Figure 20), 157  which is largely identical to that of SARS‐CoV‐1 Mpro with a 96% sequence identity. Compound 64 was docked to SARS‐CoV‐2 Mpro, and a series of structural modifications were performed to improve its pharmacokinetic properties. Specifically, masking the P2‐P3 amide bond with the pyridone ring could improve plasma half‐life; and exchanging the lipophilic cinnamoyl residue for the less lipophilic Boc group, could increase plasma solubility and reduce its binding to plasma proteins.
Figure 20  Crystal structure of SARS‐CoV‐2 Mpro. 157  Mpro, main protease; SARS‐CoV, severe acute respiratory syndrome coronavirus [Color figure can be viewed at wileyonlinelibrary.com] Indeed, the resulting 66 had a ~3‐fold improved plasma half‐life in mice when compared to the lead 65 (from 18 min to 1 h). The in vitro kinetic plasma solubility has been increased by a factor of ~19 (from 6 µM for the lead to 112 µM for best derivative), and the thermodynamic solubility by a factor of ~13 (from 41 to 530 µM). Compound 66 also showed reduced binding to mouse plasma protein. However, compared to the lead (IC50, 0.18 µM), the structural modifications caused a reduction of activity against SARS‐CoV‐2 Mpro (IC50, 2.39 µM) and enteroviral 3 C proteases. Nevertheless, the introduction of a cyclopropyl group as in 67 instead of P2‐cyclohexyl enhanced the antiviral activity against β‐coronaviruses.
Compound 67 (Figure 19) inhibited purified SARS‐CoV‐2 Mpro with an IC50 of 0.67 µM. It also inhibited SARS‐CoV‐1 Mpro (IC50, 0.90 µM) and MERS‐CoV Mpro (IC50, 0.58 µM) with similar potency. It was effective against SARS‐CoV‐1 replication with an EC50 value of 1.75 µM. In SARS‐CoV‐2 infected human Calu3 cells, it inhibited the viral replication with an EC50 of 4–5 µM, when in fact the Boc‐unprotected 68 was inactive, suggesting a bulky hydrophobic group is necessary for cellular membrane penetration. On the other hand, increasing hydrophobicity of molecules should be pondered carefully, as it can increase plasma protein binding as it was described for 64. The pharmacokinetic properties of 67 revealed striking lung tropism and was suitable for inhalation in mice without any perceived adverse effects.
Compound 67 was cocrystallized with the enzyme in two different forms at 1.95 and 2.20 Å (Figure 21). The key feature observed from this crystal structure was that the inhibitor binds to the shallow substrate‐binding site at the surface of each protomer, between domains I and II. The thioketal that resulted from the nucleophilic Cys145 attacking the inhibitor, is stabilized by a H‐bond from His41, whereas the amide oxygen of 67 accepts a H‐bond from the main‐chain amides of Gly143, Cys145, and in part, Ser144 that make up the cysteine protease's canonical oxyanion hole. 157  The P1 lactam moiety is deeply embedded in the S1 pocket where the lactam nitrogen donates a three‐center H‐bond to the main chain oxygen of the Phe140 and the carboxylate of Glu166. The carbonyl oxygen forms a H‐bond to His163. The P2‐cyclopropyl moiety fits into the S2 subsite. The P3‐P2 pyridone moiety occupies the space normally filled by the substrate's main chain. The Boc group is not situated in the canonical S4 site, rather it is located near Pro168, which explains why the removal of the Boc group as in 68 weakened the inhibitory activity.
Figure 21  Crystal structure of 67 with SARS‐CoV‐2 Mpro. Mpro, main protease; SARS‐CoV, severe acute respiratory syndrome coronavirus [Color figure can be viewed at wileyonlinelibrary.com]