Discovery of Di- and Trihaloacetamides as Covalent SARS-CoV-2 Main Protease Inhibitors with High Target Specificity.

The main protease (Mpro) is a validated antiviral drug target of SARS-CoV-2. A number of Mpro inhibitors have now advanced to animal model study and human clinical trials. However, one issue yet to be addressed is the target selectivity over host proteases such as cathepsin L. In this study we describe the rational design of covalent SARS-CoV-2 Mpro inhibitors with novel cysteine reactive warheads including dichloroacetamide, dibromoacetamide, tribromoacetamide, 2-bromo-2,2-dichloroacetamide, and 2-chloro-2,2-dibromoacetamide. The promising lead candidates Jun9-62-2R (dichloroacetamide) and Jun9-88-6R (tribromoacetamide) had not only potent enzymatic inhibition and antiviral activity but also significantly improved target specificity over caplain and cathepsins. Compared to GC-376, these new compounds did not inhibit the host cysteine proteases including calpain I, cathepsin B, cathepsin K, cathepsin L, and caspase-3. To the best of our knowledge, they are among the most selective covalent Mpro inhibitors reported thus far. The cocrystal structures of SARS-CoV-2 Mpro with Jun9-62-2R and Jun9-57-3R reaffirmed our design hypothesis, showing that both compounds form a covalent adduct with the catalytic C145. Overall, these novel compounds represent valuable chemical probes for target validation and drug candidates for further development as SARS-CoV-2 antivirals.

Introduction

The ongoing COVID-19 pandemic is a timely reminder that direct-acting antivirals are urgently needed. Despite the expeditious development of mRNA vaccines, SARS-CoV-2 is likely to remain a significant public health concern in the foreseeable future for several reasons. First, variant viruses with escape mutations continue to emerge, which compromise the efficacy of vaccines.[1] Second, a portion of the population opt out of vaccination based on their religious beliefs, concerns of long-term side effects, or other reasons. As such, it is unpredictable when or whether herd immunity can be achieved. Third, the durability of COVID vaccines is currently unknown. Therefore, antivirals are important complements of vaccines to combat both current COVID-19 pandemic and future coronavirus outbreaks.

In combating the COVID-19 pandemic, researchers from different disciplines work relentlessly to discover countermeasures. Drug repurposing led to the identification of remdesivir as the first FDA-approved SARS-CoV-2 antiviral. EIDD-2801, another viral polymerase inhibitor discovered through a similar approach, is in human clinical II/III trials.[2] Among the drug targets exploited, viral polymerases including the main protease (Mpro) and the papain-like protease (PLpro) are the most extensively studied.[3] The Mpro is a cysteine protease and digests the viral polyprotein at more than 11 sites during the viral replication. Mpro functions as a dimer and has a unique preference for glutamine at the substrate P1 position. Mpro is a validated high-profile antiviral drug target, and Mpro inhibitors have demonstrated potent antiviral activity in cell cultures and animal models (Figure ).[4−8] Two Pfizer Mpro inhibitors PF-07304814 and PF-07321332 are advanced to phase I clinical trial.[9,10] Additional promising leads are listed in Table , which are in different stages of translational development. The success of fast-track development of SARS-CoV-2 Mpro inhibitors is a result of accumulated expertise and knowledge in targeting SARS-CoV Mpro and similar picornavirus 3C-like (3CL) proteases over the years.[11] Despite the tremendous progress in developing Mpro inhibitors, the selectivity profiling has thus far been largely neglected. It is essential to address the target selectivity issue early on to avoid catastrophic failures in the later clinical studies. Cysteine protease inhibitor has yet received FDA approval, and the lack of target specificity might be the culprit.

Figure 1

AdvancedSARS-CoV-2 Mpro inhibitors with translational potential.

Table 1

Target Specificity of SARS-CoV-2 Mpro Inhibitors

compd	SARS-CoV-2 Mpro, IC50 (nM)	cathepsin L, IC50 (nM)	additional off targets	ref
GC-376	33	0.99	calpain I (IC50 = 74 nM)	(8, 17, 18, 20, 24)
			cathepsin K (IC50 = 0.56 nM)
MPI8	105	1.2	cathepsin B (IC50 = 230 nM)	(15, 16)
			cathepsin K (IC50 = 180 nM)
PF-00835231	5	146	cathepsin B (IC50 = 1.3 μM)	(19, 21)
6e	10	<0.5		(19)
6j	7	<0.5		(19)
11a	8	0.21		(19, 23)

The majority of current reported SARS-CoV-2 Mpro inhibitors are peptidomimetic covalent inhibitors with a reactive warhead such as ketone, aldehyde, or ketoamide.[11] Some of the promising examples include the Pfizer compounds PF-07304814 (the parent compound PF-00835231),[10]11a,[12]GC-376,[7,13] the deuterated GC-376 (D2-GC-376),[5]6e, 6j,[14]MI-09, MI-30,[4] and MPI8(15) (Figure ). Although the high reactivity of these reactive warheads, especially the aldehyde, confers potent activities in the enzymatic assay and antiviral assay, it inevitably leads to off-target side effects through reaction with some host proteins.[16−19] For example, we and others have shown that GC-376 is a potent inhibitor of cathepsin L (Table ).[17,20] A recent study revealed that MP18, an analogue of GC-376 with an aldehyde warhead, inhibits cathepsins B, L, and K with IC50 values of 1.2, 230, and 180 nM, respectively.[15] The off-target effect is also a potential concern for some of the most advanced Mpro inhibitors including the clinical candidate PF-07304814,[21] compounds 6j and 6e which showed in vivo antiviral efficacy against MERS-CoV-2 infection in mice, and compound 11a with potent in vitro antiviral activity (Table ).[23] All of these compounds are potent inhibitors of cathepsin L. The high reactivity of the aldehyde warhead might confer the lack of target specificity, and the design of covalent inhibitors with a high target specificity remains a daunting task.

We report herein the rational design of covalent Mpro inhibitors with novel cysteine reactive warheads and high target specificity. Specifically, guided by the X-ray crystal structure of SARS-CoV-2 Mpro with 23R (Jun8-76-3A) (PDB code 7KX5), which was one of the most potent noncovalent Mpro inhibitors developed from our earlier study,[24] we systematically explored a number of novel electrophiles in the replacement of the P1′ furyl substitution in 23R. The aim is to identify C145 reactive electrophiles with both potent Mpro inhibition and high target selectivity. This effort led to the discovery of several novel cysteine reactive warheads including dichloroacetamide, dibromoacetamide, tribromoacetamide, 2-bromo-2,2-dichloroacetamide, and 2-chloro-2,2-dibromoacetamide. One of the most potent lead compounds Jun9-62-2R (dichloroacetamide) inhibited SARS-CoV-2 Mpro with an IC50 of 0.43 μM and viral replication with an EC50 of 2.05 μM in Caco2-hACE2 cells. Significantly, unlike GC-376, Jun9-62-2R (dichloroacetamide) and Jun9-88-6R (tribromoacetamide) are highly selective toward Mpro and do not inhibit the host calpain I, cathepsins B, K, L, caspase-3, and trypsin. X-ray crystal structure of SARS-CoV-2 Mpro with Jun9-62-2R (dichloroacetamide) and Jun9-57-3R (chloroacetamide) revealed that the C145 forms a covalent adduct with the reactive warheads. Overall, the discovery of these di- and trihaloacetamides as novel cysteine reactive warheads sheds light on the feasibility of developing SARS-CoV-2 Mpro inhibitors with high target specificity over tested calpain and cathepsins and a high cellular selectivity index. These novel compounds represent valuable chemical probes for target validation and drug candidates for further development as SARS-CoV-2 antivirals.

Results and Discussion

Synthesis of Covalent Mpro Inhibitors

The covalent Mpro inhibitors were synthesized by the one-pot Ugi four-component reaction (Ugi-4CR) as shown for Jun9-62-2 (Figure ) with yields from 33% to 88%. For compounds with potent enzymatic inhibition, the diastereomers were subsequently separated by chiral HPLC. The absolute stereochemistry of Jun9-57-3R and Jun9-62-2R was determined by X-ray crystallography, and the stereochemistry for the diastereomers of Jun9-90-4, Jun9-89-2, Jun9-89-4, and Jun9-88-6 was tentatively assigned based on their relevant retention times in chiral HPLC.

Figure 2

Synthesis route for the covalent SARS-CoV-2 Mpro inhibitors through Ugi-4CR. The R and S chirality refers to the chiral center at the pyridine substitution.

Rational Design of Covalent Mpro Inhibitors

23R was designed based on the superimposed X-ray crystal structure of GC-376 with ML188 and UAWJ254.[24,25] The X-ray crystal structure showed that the furyl substitution at the P1′ position of 23R is in close proximity with the catalytic cysteine 145 (3.4 Å between C145 sulfur and the C-2 carbon of furyl, PDB code 7KX5) (Figure A), suggesting that replacement of furyl with a reactive warhead might lead to covalent inhibitors (Figure B). 23R is an ideal lead candidate for the design of covalent Mpro inhibitors for several reasons: (1) the P1, P2, and P3 substitutions have already been optimized; (2) the designed compounds can be expeditiously synthesized by the one-pot Ugi-4CR; (3) a diversity of cysteine reactive warheads are commercially available and can be promptly introduced at the P1′ position to react with the C145.

Figure 3

Rational design of covalent SARS-CoV-2 Mpro inhibitors based on 23R. (A) X-ray crystal structure of SARS-CoV-2 Mpro with 23R (PDB code 7KX5). The distance between the furyl ring and the catalytic cysteine 145 is 3.4 Å. (B) Representative cysteine reactive warheads for covalent labeling of C145. (C) FDA-approved covalent inhibitors. The reactive warheads are colored in magenta. Pfizer compound 12 is a preclinical candidate. (D) Designed covalent SARS-CoV-2 Mpro inhibitors. The results are the average ± standard deviation of three repeats.

Although a number of thiol-reactive warheads have been exploited in the development of covalent protease and kinase inhibitors,[26−28] we decided to focus on pharmacologically compliant reactive warheads from the FDA-approved drugs. The majority of FDA-approved thiol-reactive drugs are kinase inhibitors including ibrutinib, osimertinib, zanubrutinib, acalabrutinib, dacomitinib, neratinib, and afatinib (Figure C).[26] As such, acrylamide and 2-butynamide were chosen as reactive warheads in our initial design of covalent SARS-CoV-2 Mpro inhibitors (Figure B). Chloroacetamide was also chosen as it was previously explored by Pfizer for the development of SARS-CoV and SARS-CoV-2 Mpro inhibitors (Pfizer compound 12) (Figure C).[21] Chloroacetamide is frequently used as a reactive warhead for designing chemical probes for target pull down.[29] Finally, we included azidomethylene as it was previously shown to be a relatively unreactive cysteine warhed.[30,31] The fluoroacetamide was included as a control.

The designed covalent SARS-CoV-2 Mpro inhibitors were shown in Figure D. All compounds were first tested in the FRET-based Mpro enzymatic assay. Active hits were further tested for cellular cytotoxicity to select candidates for the following antiviral assay against SARS-CoV-2. It was found that the azidoacetamide Jun9-61-1 and the fluoracetamide Jun9-61-4 were not active (IC50 > 20 μM). Surprisingly, the acrylamides Jun10-15-2 and Jun9-51-3 were also not active (IC50 > 20 μM), suggesting the acrylamide might not be positioned at the right geometry for reacting with the C145. Gratifyingly, Jun9-62-1 with the 2-butynamide warhead showed potent inhibition with an IC50 of 1.15 μM. However, Jun9-62-1 also had moderate cytotoxicity in both Vero E6 (CC50 = 17.99 μM) and Calu-3 (CC50 = 47.77 μM) cells. Similarly, covalent inhibitors with the chloroacetamide reactive warhead had potent inhibition against SARS-CoV-2 Mpro. The most potent compound Jun9-57-3R inhibited SARS-CoV-2 Mpro with an IC50 of 0.05 μM, comparable to the potency of GC-376 (IC50 = 0.03 μM). Interestingly, the diastereomer Jun9-57-3S was also a potent Mpro inhibitor with an IC50 of 1.13 μM. However, covalent inhibitors with the chloroacetamide warhead Jun9-54-1, Jun9-59-1, Jun9-55-2, Jun9-57-3R, Jun9-57-3S, Jun9-57-2, and Jun9-55-1 were highly cytotoxic in Vero E6 (CC50 < 11 μM) and Calu-3 (CC50 < 2 μM) cells, possibly due to their off-target effects on host proteins/DNAs. The low cellular selectivity index precludes further development of these covalent Mpro inhibitors as SARS-CoV-2 antiviral drugs.

Exploring Acrylamides and Haloacetamides as Novel Warheads for SARS-CoV-2 Mpro C145

For the acrylamide series of compounds, Jun9-72-3 and Jun10-31-4, both containing a 2-substituted acrylamide warhead, were not active against Mpro (IC50 > 20 μM) (Figure ). However, compound Jun10-38-2 with the 2-chloroacrylamide had potent inhibition with an IC50 of 4.22 μM.

Figure 4

SARS-CoV-2 Mpro inhibitors with novel acrylamide and haloacetamide warheads. The results are the average ± standard deviation of three repeats.

For the haloacetamide series of compounds, the reference compound Jun9-54-1 with the classical chloroacetamide reactive warhead had potent inhibition against SARS-CoV-2 Mpro with an IC50 of 0.17 μM. However, it was cytotoxic in both Vero E6 cells and Calu-3 cells with CC50 values less than 3.5 μM. To increase the cellular selectivity index, we reasoned that substituted chloroacetamides or haloacetamides might have reduced cellular cytotoxicity while maintaining potent Mpro inhibition. It was found that Jun9-77-1 with the 2-chloropropanamide warhead was not active (IC50 > 20 μM). Encouragingly, compound Jun9-62-2R with the dichloroacetamide warhead had potent inhibition against Mpro with an IC50 of 0.43 μM while being noncytotoxic to Vero E6 cells (CC50 > 100 μM). In comparison, the corresponding diastereomer Jun9-62-2S was not active (IC50 > 20 μM), which is consistent with the predicted binding mode (Figure A). Given these promising results, we further designed two additional dichloroacetamide compounds Jun9-90-3 and Jun9-90-4 with variations at the P3/P4 substitutions. Similar to Jun9-62-2R, both Jun9-90-3R and Jun9-90-4R were potent inhibitors with IC50 values of 0.30 and 0.46 μM, respectively. Both compounds were also noncytotoxic to Vero E6 cells (CC50 > 100 μM). In contrast, the corresponding diastereomers Jun9-90-3S and Jun9-90-4S were not active (IC50 > 20 μM).

We further explored di- and trisubstituted haloacetamides as Mpro C145 reactive warheads (Figure ). Jun9-89-2R with the dibromoacetamide warhead is highly active with an IC50 of 0.08 μM; however, the cell cytotoxicity also increased (CC50 = 8.94 μM). The diastereomer Jun9-89-2S also had potent inhibition against Mpro with an IC50 of 2.44 μM and comparable cytotoxicity (CC50 = 4.57 μM). Jun9-76-4 with the 2,2-dichloropropanamide warhead, Jun9-72-4 with the trichloroacetamide, and Jun9-77-2 with the 2-chloro-2,2-difluoroacetamide were all inactive against Mpro (IC50 > 20 μM). Jun9-89-3 with the 2-bromo-2,2-dichloroacetamide showed potent inhibition with an IC50 of 1.20 μM. The cytotoxicity of Jun9-89-3 also improved (CC50 = 32.43 μM). Jun9-89-4R with the 2-chloro-2,2-dibromoacetamide warhead is highly potent with an IC50 of 0.05 μM, but it was cytotoxic in Vero E6 cells (CC50 = 8.41 μM). The diastereomer Jun9-89-4S was less active (IC50 = 9.04 μM). Jun9-88-6R with the tribromoacetamide warhead had high potency against Mpro with an IC50 of 0.08 μM, while the diastereomer Jun9-88-6S was less active (IC50 = 7.16 μM). Both Jun9-88-6R and Jun9-88-6S had comparable cytotoxicity as Jun9-54-1 with CC50 values of 5.48 and 5.99 μM, respectively.

Pharmacological Characterization of SARS-CoV-2 Mpro Inhibitors with Novel Reactive Warheads

On the basis of the Mpro inhibition and cell cytotoxicity, four compounds Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R were selected for mechanistic studies (Figure ). Enzymatic kinetic studies suggested that these four compounds bind to Mpro in a two-step process: the first step is reversible binding (KI), and the second step is irreversible binding (kinact). The calculated kinact/KI values for Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R were 819.7, 1543.6, 867.4, and 7074.3 M–1 s–1, respectively (Figure A). These results were in agreement with the expected mechanism of action in which all four compounds form a covalent bond with the catalytic C145. In the thermal shift-binding assay, all four compounds stabilized the SARS-CoV-2 Mpro upon binding as reflected by the Tm shift to higher temperatures (Figure B). As the tribromoacetamide is sterically hindered, the mechanism of action of Jun9-88-6R might involve the nucleophilic attack of the carbonyl by the C145 thiol to give a thiohemiketal intermediate, followed by a 1,2-shift of the sulfur to displace one bromide (Figure S2).

Figure 5

Pharmacological characterization of the SARS-CoV-2 Mpro inhibitors. (A) Curve fittings of the enzymatic kinetic studies of four compounds Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R against SARS-CoV-2 Mpro. (B) Binding of four compounds Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R to SARS-CoV-2 Mpro in the thermal shift assay. (C) Fast dilution experiment. 10 μM Mpro was preincubated with 10 μM testing compounds for 2 h at 30 °C; the preformed compound–enzyme complex was diluted 100-fold into reaction buffer before initiation of the enzymatic reaction. The recovered enzymatic activity was compared with DMSO control. 23R is a noncovalent Mpro inhibitor, and it was included as a control. (D) Time dependent inhibition of Mpro by Jun9-62-2R. 100 nM SARS CoV-2 Mpro was preincubated with Jun9-62-2R for various periods of time (0 min to 2 h) before the addition of 10 μM FRET substrate to initiate the enzymatic reaction. 23R was included as a control. (E–H) Native mass spectrometry assay of SARS-CoV-2 Mpro reveals binding of Jun9-62-2R with mass modifications of 482 Da (E), Jun9-89-2R with mass modifications of 526 Da (F), Jun9-88-6R with mass modifications of 526 Da (G), and Jun9-89-4R with mass modifications of (a) 481 and (b) 561 Da (H). Mpro functions as a dimer, and both one drug per dimer (protein + 1 Mod) and two drugs per dimer (protein + 2 Mods) were observed. (I) FlipGFP assay characterization of the inhibition of the cellular enzymatic activity of SARS-CoV-2 Mpro by the four compounds Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R. (J) Curve fittings of the FlipGFP Mpro assay. The results are the average ± standard deviation of three repeats.

To provide additional lines of evidence to support the proposed mode of action of covalent binding, we performed three additional experiments. First, to demonstrate the reversibility of the binding of Jun9-62-2R to Mpro, we incubated 10 μM SARS-CoV-2 Mpro with 10 μM Jun9-62-2R for 2 h and monitored the enzymatic activity of Mpro following 100-fold dilution of the mixture. It was found that no enzymatic activity was recovered (Figure C). In contrast, the mixture with our previously developed noncovalent inhibitor 23R showed nearly complete recovery of enzymatic activity after dilution (Figure C). These results suggest that the binding of Jun9-62-2R is irreversible while the binding of 23R is reversible. Second, we repeated the FRET assay of Jun9-62-2R with different preincubation times and found that longer preincubation time gave lower IC50 values (Figure D). These data are consistent with the mode of action of covalent inhibitors.[32] In contrast, preincubation of Mpro with the noncovalent inhibitor 23R did not lead to significant changes of the IC50 value (Figure D). Third, we used native mass spectrometry to detect the covalent adducts of Mpro with Jun9-62-2R, Jun9-89-2R, Jun9-88-6R, and Jun9-89-4R. The expected mass shifts of 482 Da and 526 Da were observed for Jun9-62-2R and Jun9-89-2R, respectively (Figure E,F). Interesting, the expected dibromoacetamide conjugate was not observed for Jun9-88-6R, suggesting this conjugate might not be stable. Instead, the mass shift corresponding to the monobromo thiol adduct was observed (Figure G). For Jun9-89-4R, the mass shifts for both the chlorobromo and chloro thiol adducts were observed (Figure H).

To further profile the cellular Mpro inhibition, we tested these four compounds in our recently developed FlipGFP assay.[18,33] Briefly, the GFP is split into two parts, the β1–9 template and the β10–11 strands. The β10 and β11 strands were engineered with K5-E5 linker such that they are restrained in the parallel form. When the linker is cleaved by Mpro, β10 and β11 adopt antiparallel conformation, which allows association with the β1–9 template, leading to the recovery of the GFP signal. In the FlipGFP assay, the GFP signal is proportional to the Mpro enzymatic activity. It was found that all four compounds led to dose-dependent inhibition of the GFP signal with EC50 values of 0.96 μM (Jun9-62-2R), 0.91 μM (Jun9-90-3R), 1.57 μM (Jun9-90-4R), and 0.92 μM (Jun9-88-6R) (Figure I,J). The EC50 value for the positive control GC-376 was 1.80 μM. This result suggests that these four compounds can potently inhibit the Mpro in the cellular content.

Antiviral Activity of SARS-CoV-2 Mpro Inhibitors with Novel Reactive Warheads

The antiviral activity of the four lead compounds was evaluated in both Vero E6 cells and Caco2-hACE2 cells to exclude cell type dependent effect. Caco2-hACE2 with endogenous TMPRSS2 expression is a validated cell line for SARS-CoV-2 antiviral assay.[34−36]Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R inhibited SARS-CoV-2 replication in Vero E6 cells with EC50 values of 0.90, 2.07, 1.10, and 0.58 μM, respectively (Figure A). All four compounds showed comparable antiviral activity in Caco2-hACE2 cells with EC50 values of 2.05, 3.24, 1.43, and 2.15 μM, respectively (Figure B). In comparison, GC-376 inhibited SARS-CoV-2 replication in Vero E6 and Caco2-hACE2 cells with EC50 values of 1.51 and 2.90 μM. When tested in Calu-3 cells, Jun9-90-3R showed comparable antiviral activity with an EC50 value of 2.00 μM (Figure C).

Figure 6

Antiviral activity of Jun9-62-2R, Jun9-90-3R, Jun9-90-4R, and Jun9-88-6R against SARS-CoV-2 in different cell lines. (A) Antiviral activity against SARS-CoV-2 in Vero E6 cells. (B) Antiviral activity against SARS-CoV-2 in Caco2-hACE2 cells. (C) Antiviral activity of Jun9-90-3R in Calu-3 cells. The results are the average ± standard deviation of three repeats.

Profiling the Target Selectivity against Host Proteases

Lack of target specificity is one of the major reasons that many cysteine protease inhibitors failed in the clinical trials. To profile the target specificity of these SARS-CoV-2 Mpro inhibitors with a novel reactive warhead, we selected Jun9-62-2R and Jun9-88-6R as representative examples and included the canonical GC-376 with an aldehyde reactive warhead for comparison. The results showed that GC-376 had potent inhibition of the host proteases including calpain I, cathepsin B, cathepsin K, and cathepsin L with IC50 values in the submicromolar and nanomolar range. GC-376 did not inhibit caspase-3 and trypsin (IC50 > 20 μM) (Figure ). In comparison, both Jun9-62-2R and Jun9-88-6R had a significantly improved target selectivity and did not show potent inhibition against the host calpain 1, cathepsin B, cathepsin K, cathepsin L, caspase-3, and trypsin. Jun9-88-6R had weak inhibition against cathepsin L with an IC50 of 7.37 μM, conferring a 94-fold higher selectivity for inhibiting the SARS-CoV-2 Mpro. Collectively, the covalent SARS-CoV-2 Mpro inhibitors Jun9-62-2R with the dichloroacetamide warhead and Jun9-88-6R with the tribromoacetamide warhead have high target specificity against Mpro over host proteases.

Figure 7

Target selectivity of SARS-CoV-2 Mpro inhibitors against host proteases. (A) Heat map of target selectivity. (B) IC50 values of Jun9-62-2R and Jun9-88-6R against host proteases in the FRET-based enzymatic assay. aThe result was from ref (20).

X-ray Crystal Structures of SARS-CoV-2 Mpro in Complex with Jun9-62-2R and Jun9-57-3R

Using X-ray crystallography, we solved the complex structures of SARS-CoV-2 Mpro with Jun9-57-3R (2.25 Å, PDB code 7RN0) and Jun9-62-2R (2.30 Å, PDB code 7RN1) (Figure , Table S1). Jun9-57-3R and Jun9-62-2R have nearly identical chemical features to their noncovalent progenitor 23R (Jun8-76-3A) (PDB code 7KX5). As such, the binding poses are very similar. The pyridyl ring binds to the S1 pocket of Mpro, where it forms a hydrogen bond with His163. This hydrogen bond is critical for coordinating the Gln side chain of its substrate, a residue it is uniquely selective for. Consequently, a hydrogen bond acceptor at this position confers tremendous potency to Mpro inhibitors. The phenylpyrrole (Jun9-57-3R) or biphenyl (Jun9-62-2R) moieties insert into the hydrophobic S2 pocket where they form nonpolar contacts and stack with the catalytic base, His41. An amide group linking the pyridyl ring to an α-methylbenzene group accepts a hydrogen bond from the main chain of Glu166. This α-methylbenzene group flips down toward the core of the substrate channel, where it forms additional π-stacking interactions with the biphenyl or phenylpyrrole moieties. The key distinction between Jun9-62-2R, Jun9-57-3R, and analogues Jun8-76-3A and ML188 is the presence of an electrophilic chloroacetamide warhead, which forms a covalent adduct with the catalytic cysteine Cys145 (Figure C,D). The short distance of this covalent bond (1.8 Å) allows the inhibitor to press further into the oxyanion hole, causing the P2 benzene to rotate inward by ∼40°. Likewise, the chloracetamide warhead is forced toward the catalytic core, causing the P1′ chloride of Jun9-57-3R to lie closer to Cys145 (2.8 Å) than the corresponding furyl oxygen of Jun8-76-3A (3.2 Å).

Figure 8

X-ray crystal structures of SARS-CoV-2 Mpro in complex with Jun9-62-2R (A) and Jun9-57-3R (B). 2Fo – Fc electron density map, shown in gray, is contoured at 1σ. Structural superimpositions of the noncovalent analogues Jun8-76-3A (white, PDB code 7KX5) and ML188 (yellow, PDB code 7L0D) with Jun9-62-2R (C) and Jun9-57-3R (D) reveal a different mode of interaction with the catalytic core.

Conclusion

The majority of the reported Mpro inhibitors contain the aldehyde reactive warhead, which is known to have nonspecific reactivity toward host proteins.[16−19] It should be noted that both of the Pfizer Mpro inhibitors that are currently in clinical trials do not contain the aldehyde warhead.[9,10] As such, we are interested in developing SARS-CoV-2 Mpro inhibitors with high target specificity. A highly specific Mpro inhibitor is also needed for target validation as it separates the effect of Mpro inhibition from host protease inhibition such as cathepsin L. It is known that host cathepsin L is important in SARS-CoV-2 replication in Vero E6 cells, which are TMPRSS2-negative, but not in Calu-3 cells, which are TMPRSS2-positive.[37] In this study, we report the discovery of dichloroacetamide, dibromoacetamide, 2-bromo-2,2-dichloroacetamide, 2-chloro-2,2-dibromoacetamide, and tribromoacetamide as novel cysteine reactive warheads. To the best of our knowledge, these warheads have not been explored in cysteine protease inhibitors. The most promising lead compounds Jun9-62-2R with the dichloroacetamide warhead and Jun9-88-6R with the tribromoacetamide inhibited SARS-CoV-2 Mpro with IC50 values of 0.43 μM and 0.08 μM, respectively. These two compounds also showed potent inhibition against SARS-CoV-2 in both Vero E6 and Caco2-hACE2 cells with EC50 values in the single-digit micromolar to submicromolar range. Significantly, both Jun9-62-2R and Jun9-88-6R had high target specificity toward Mpro and did not inhibit the host proteases including calpain I, cathepsin B, cathepsin K, cathepsin L, caspase-3, and trypsin. In comparison, GC-376 was not selective and inhibited calpain I, cathepsin B, cathepsin K, and cathepsin L with comparable potency as Mpro. Regarding the translational potential of the di- and trihaloacetamide-containing Mpro inhibitors, the widely used antibiotic chloramphenicol contains the dichloroacetamide, suggesting Jun9-62-2R might be tolerated in vivo. Follow-up studies will optimize the in vitro and in vivo pharmacokinetic properties and in vivo antiviral efficacy of these novel compounds in SARS-CoV-2 infection animal models. Other potential strategies include developing selective Mpro inhibitors including allosteric inhibitors[38,39] or targeting the more reactive Cys44 at the S2 binding pocket.[40,41] Overall, these novel compounds represent valuable chemical probes for target validation and promising drug candidates for translational development as SARS-CoV-2 antivirals.