Inhibitors of SARS-CoV-2 PLpro

Abstract

The emergence of SARS-CoV-2 causing the COVID-19 pandemic, has highlighted how a combination of urgency, collaboration and building on existing research can enable rapid vaccine development to fight disease outbreaks. However, even countries with high vaccination rates still see surges in case numbers and high numbers of hospitalized patients. The development of antiviral treatments hence remains a top priority in preventing hospitalization and death of COVID-19 patients, and eventually bringing an end to the SARS-CoV-2 pandemic. The SARS-CoV-2 proteome contains several essential enzymatic activities embedded within its non-structural proteins (nsps). We here focus on nsp3, that harbours an essential papain-like protease (PLpro) domain responsible for cleaving the viral polyprotein as part of viral processing. Moreover, nsp3/PLpro also cleaves ubiquitin and ISG15 modifications within the host cell, derailing innate immune responses. Small molecule inhibition of the PLpro protease domain significantly reduces viral loads in SARS-CoV-2 infection models, suggesting that PLpro is an excellent drug target for next generation antivirals. In this review we discuss the conserved structure and function of PLpro and the ongoing efforts to design small molecule PLpro inhibitors that exploit this knowledge. We first discuss the many drug repurposing attempts, concluding that it is unlikely that PLpro-targeting drugs already exist. We next discuss the wealth of structural information on SARS-CoV-2 PLpro inhibition, for which there are now ∼30 distinct crystal structures with small molecule inhibitors bound in a surprising number of distinct crystallographic settings. We focus on optimisation of an existing compound class, based on SARS-CoV PLpro inhibitor GRL-0617, and recapitulate how new GRL-0617 derivatives exploit different features of PLpro, to overcome some compound liabilities.

Keywords: COVID-19; GRL-0617; Nsp3; SARS-CoV-2; antiviral drug discovery; medicinal chemistry; papain like protease (PLpro); structure-activity relationship (SAR).

FIGURE 1

Reported apo and compound structures of SARS-CoV-2 PLpro. Shown are unliganded and compound bound structures publicly released in the Protein Data Bank (PDB) since the beginning of the COVID-19 pandemic. The unit cell for each space group is shown (thin red lines) and the corresponding symmetry mates from the asymmetric unit are depicted with matching colours. The structures are grouped according to their bound ligands, the ligand is labelled above each unit cell, and the corresponding PDB accession numbers shown below. The obtained resolution or resolution range for each crystallographic setting is indicated.

FIGURE 2

Molecular basis for inhibition of SARS-CoV-2 PLpro by GRL-0617. (A) Structure of SARS-CoV-2 PLpro bound to GRL-0617 in teal (PDB 7JRN (Ma et al., 2021)), with inhibitor in wheat colour in ball-and stick representation representing the (R)-enantiomer. A close-up view of the ligand binding site for GRL-0617 with key residues indicated is also shown and hydrogen bonds are displayed as a dashed yellow line. A superimposed structure of apo PLpro [purple, PDB 6WZU (Osipiuk et al., 2021)] shows that the inhibitor does not induce global conformational changes. The catalytic Cys is shown in ball and stick representation, and a bound zinc ion in apo PLpro is shown as a grey sphere. (B) Close-up view of the GRL-0617 binding site overlaid with ubiquitin from the ubiquitin-PLpro complex in orange [PLpro ∼ Ub, PDB 6XAA (Klemm et al., 2020)] and with ISG15 from PLpro bound to the C-terminal ISG15 Ubl fold in pink [PLpro ∼ ISG15^CT, PDB 6XA9 (Klemm et al., 2020)]. The orthomethyl resides in a pocket formed by Leu162, Tyr264, and Tyr273 occupying the position of the Arg74 in Ub or Arg155 in ISG15. Upon ligand binding, Leu162 rotates its side chain to block the channel and the path of the Ubl tail to the catalytic Cys111. The catalytic Cys111 is shown in ball and stick representation. (C) Close up view of the ligand binding site for GRL-0617 in SARS-CoV-2 PLpro in teal overlaid with SARS-CoV PLpro in green [PDB 3E9S (Ratia et al., 2008)]. Key residues are fully conserved between SARS-CoV and SARS-CoV-2 which explains cross specificity of compounds. Hydrogen bonds are displayed as a dashed yellow line.

FIGURE 3

Overview of the reported attempts to target Glu167 in PLpro. (A,B) Close up view of the GRL-0617 binding pocket with PLpro in teal in cartoon representation (A) or as with the calculated surface charge overlaid (B). The inhibitor is shown in wheat colour in ball-and stick representation. Residues Lys157 to Glu167 in the PLpro Thumb domain form a shallow negatively charged pocket. Several compounds target the side chain of Glu167 to improve potency of GRL-0617 . (C) The GRL-0617 scaffold (boxed) was extended at the para-position of the benzene ring (at the R position). The red line indicates the handle used for the respective substituents - in GRL-0617 an amino group is present at this position. Given compound data refer to, IC₅₀ from in vitro activity assays, and K_D values from SPR where available (See Table 3 for the compound identifiers from their respective publications). Basic amines appear to be most tolerated at this position while replacement with alkyl groups are detrimental to activity of the compound. Co-crystal structures for 1–4 and 17 (Figure 4) are shown [PDB IDs 1 (7JIT), 2 (7JIW), 3 (7KOL), 4 (7KOJ), 17 (7LBS)]. Compounds 3 and 4 were published in the PDB but excluded from the final publication, so no IC₅₀ data is available (Osipiuk et al., 2021).

FIGURE 4

Overview of compounds which successfully replaced the naphthyl ring in GRL-0617 . (A,B) Close up view of the GRL-0617 binding pocket with PLpro represented in teal in cartoon representation overlaid with the surface representation (A) or calculated surface charge of the protein (B). GRL-0617 in wheat colour and compound 17 in green are depicted in ball and stick representation. In (A) the key residues are noted to highlight the BL2 groove formed by closure of the blocking loop and induced upon ligand binding. 17 shows that the replacement of the naphthalene ring with a 2-phenylthiophene appears to effectively replace the dependency of the naphthyl group. The BL2 groove is then engaged by a basic amine tail to improve potency. (C) Boxed (left), the parent scaffold for modifying the naphthyl group. Boxed (right) Second iteration of compound designs, starting at the 2-phenyl thiophene scaffold ( 13 ). The red line indicates the handle used for the respective substitution. Compound data refer to IC₅₀ from in vitro activity assays, and K_D values from SPR where available (see Table 3 for the compound identifiers from their respective publications). 13–16, the 2-phenyl thiophene appears to successfully replace the naphthyl ring while still maintaining potency. 17–24 , Aromatics containing basic amines appear to be the most potent at this position. Co-crystal structures for 17–21 are shown [PDB IDs, 17 (7LBS), 18 (7LOS), 19 (7LLF), 20 (7LBR), 21 (7LLZ)]. In the case of 17 , the basic nitrogen interacts with residues lining the BL2 groove.

FIGURE 5

Modification of the amide bond in GRL-0617. (A) Close up view of the binding pocket for GRL-0617, 31 (PDB 7SDR) and 36 (PDB 7RZC). PLpro is represented in teal, lime green or grey respectively in cartoon representation and the inhibitors are shown in wheat, pink, or orange respectively in ball-and-stick representation. In GRL-0617 the side chain of Asp164 forms a crucial H-bond interaction with the amide nitrogen and the orthomethyl group remains invariant in most derivatives, as its binding pocket restricts the orientation of the phenyl ring. In 31 and 36 the amide bond was successfully replaced with a tertiary amine and both compounds appear to have effectively removed the dependency on the orthomethyl group by increasing the bond strength with Asp164. (B) The amide bond and orthomethyl appear to be highly sensitive to variation ( 25–27 ). Conservative substitution with a chlorine group ( 27 ) reduces the IC₅₀ 5-fold, to 4.8 µM. Hit compounds 28 – 29 from (Ma et al., 2021) enabled the merging of compound properties with GRL-0617. (C) Compounds explored in (Ma et al., 2021). A tertiary amine enables more extensive variations to the phenyl group not achieved prior, while retaining compound potency.