key: cord-0766996-f9th2j84
authors: Stille, Julia K.; Tjutrins, Jevgenijs; Wang, Guanyu; Venegas, Felipe A.; Hennecker, Christopher; Rueda, Andrés M.; Sharon, Itai; Blaine, Nicole; Miron, Caitlin E.; Pinus, Sharon; Labarre, Anne; Plescia, Jessica; Patrascu, Mihai Burai; Zhang, Xiaocong; Wahba, Alexander S.; Vlaho, Danielle; Huot, Mitchell J.; Schmeing, T. Martin; Mittermaier, Anthony K.; Moitessier, Nicolas
title: Design, synthesis and in vitro evaluation of novel SARS-CoV-2 3CL(pro) covalent inhibitors
date: 2021-12-11
journal: Eur J Med Chem
DOI: 10.1016/j.ejmech.2021.114046
sha: 4d6af1b2c773b0872d4b2fdfbcf7c4251adcad7c
doc_id: 766996
cord_uid: f9th2j84

Severe diseases such as the ongoing COVID-19 pandemic, as well as the previous SARS and MERS outbreaks, are the result of coronavirus infections and have demonstrated the urgent need for antiviral drugs to combat these deadly viruses. Due to its essential role in viral replication and function, 3CL(pro) (main coronaviruses cysteine-protease) has been identified as a promising target for the development of antiviral drugs. Previously reported SARS-CoV 3CL(pro) non-covalent inhibitors were used as a starting point for the development of covalent inhibitors of SARS-CoV-2 3CL(pro). We report herein our efforts in the design and synthesis of submicromolar covalent inhibitors when the enzymatic activity of the viral protease was used as a screening platform.

Coronaviruses. Coronaviruses (CoV) are a large family of viruses associated with some forms of common colds (together with rhinoviruses, respiratory syncytial virus, adenoviruses and others), as well as far more serious diseases including Severe Acute Respiratory Syndrome (SARS, caused by SARS-CoV infection), which made headlines worldwide in 2002-2003 with over 700 deaths including 43 in Canada [1] , and the Middle East Respiratory Syndrome (MERS, caused by MERS-CoV infection), which was reported in Saudi Arabia in 2012 and killed over 900 [2] . The current outbreak of novel coronavirus (COVID-19, caused by SARS-CoV-2 infection), its numerous variants, and the discovery of animal reservoirs provide significant motivation for the development of potent therapeutics against these viruses to prevent future outbreaks [3, 4] . SARS, MERS, and COVID-19 are respiratory illnesses characterized by fever, cough, and shortness of breath, posing significant danger to patients. The case fatality rates for those infected with SARS-CoV and MERS-CoV were estimated at about 10% and 35%, respectively [1, 2] .

Estimates for SARS-CoV-2 are ranging anywhere from of 0.1 to 25% depending on the age group, the country and the stage of the pandemic, although this number could change substantially as more accurate information on the numbers of infections and deaths becomes available [5, 6] . In contrast to SARS and MERS, COVID-19 has rapidly spread worldwide despite the severe restrictions imposed in many countries, and the official number of deaths now exceeds 5.0 million [7] (which is a well underestimated number as shown by excess mortality studies [8] ).

Vaccines and therapeutics. While vaccines are a central pillar of our efforts to end our current deadly phase of the COVID-19 pandemic, therapeutics offer a complementary approach with many distinct advantages. For example, oral therapeutics tend to be easy to store and administer and need only be given to the small minority of patients suffering more serious symptoms. In contrast, a large proportion of the population must be inoculated for vaccines to be effective and mRNAbased vaccines require complex logistics to maintain the cold chain, leading to enormous challenges in production, supply and administration. In addition, large vaccine campaigns require public compliance and amplifies the number of people suffering from adverse reactions to medication. To add to these difficulties, Pfizer recently announced that the immunity of their vaccine drops after about 6 months suggesting that regular injections would be needed, further amplifying the public compliance issue and burden to public health systems [9, 10] . Importantly, J o u r n a l P r e -p r o o f vaccines primarily induce an immune response against the spike protein [11] , while future variants of concern may have mutations in this protein that could allow them to evade immunity. In contrast, antiviral therapeutics can target a wide range of proteins including viral proteases (3CL pro , PL pro ), the RNA-dependent RNA polymerase (RdRp) and RNA helicase. Therefore, they can be equally effective against strains of the virus with mutations that escape spike-based vaccination or herd immunity. Overall, it is clear that effective therapeutics would be complementary to mass vaccination. Finally, some groups (pregnant and breastfeeding women, people with allergies, young children, immunocompromised patients or people with other conditions) may be at risk or not responsive to vaccines and alternative treatments (e.g., oral therapeutics) must be available [12] . Consequently, major efforts from a large number of research groups focused on the development of small molecules as antivirals against SARS-CoV-2 which culminated in the recent approval of Molnupiravir in the United Kingdom [13] .

3-chymotrypsin-like cysteine proteases (3CL pro ), also referred to as the main proteases (Mpro) or nsp5 (non-structural protein 5), which feature a Cys-His catalytic dyad (Cys 145 , His 41 ) and are required for viral replication and infection. 3CL pro enzymes were identified early on as attractive targets for antiviral development, resulting in several inhibitors and structures of SARS-3CL proinhibitor complexes (eg, PDB codes: 4TWY, 2ZU5, 2ALV [14] ). The 3CL pro enzymes from SARS-CoV and SARS-CoV-2 share nearly 80% sequence identity [15, 16] , suggesting that many of the lessons learned for developing SARS therapeutics can be applied to COVID-19. As a note, 3CL pro is not limited to coronaviruses but is also a drug target for the development of antivirals against noroviruses (such as the one involved in gastroenteritis [17] ) or antivirals against enteroviruses (e.g., antiviral drug 3CLpro-1 [18] targeting the hand, foot, and mouth disease enterovirus 71 and Rupintrivir - Figure 1 -originally developed to fight rhinoviruses [19] ).

The quest for novel antivirals against SARS-CoV and, more recently, SARS-CoV-2 has been intense, and several viral enzyme inhibitors and crystal structures of enzyme-inhibitor complexes have quickly been reported (e.g., PDB codes: 6LU7 [20] , 6M2N [21] , 6XQU [22] , 6WQF [23] ) [24] [25] [26] . The presence of a catalytic cysteine residue in the active site makes 3CL pro amenable to covalent inhibition, a strategy that was successfully employed following the SARS-CoV pandemic (SARS). In fact, many of the reported SARS-CoV inhibitors feature an electrophilic group, such as an α-ketoamide, epoxide, aziridine, α,β-unsaturated ester (Michael J o u r n a l P r e -p r o o f acceptor), or α-fluoroketone, which forms a covalent bond with the catalytic cysteine residue (Cys 145 ), as confirmed by X-ray crystallography (e.g., PDB code: 5N19) [24] . A crystal structure of the SARS-CoV-2 3CL pro with a covalent peptidic inhibitor bound to Cys 145 was quickly elucidated (PDB code: 6LU7). This pseudo-peptidic inhibitor, an analogue of Rupintrivir (tested on SARS [28] and COVID-19 [29] ), has been the starting point for a number of drug discovery campaigns [30] [31] [32] [33] [34] . 3CL pro inhibitor (PF-0730814 and PF-07321332 -Paxlovid -, Figure 1 ) entered clinical trials [32, 35, 36] and some encouraging phase 2/3 results were reported [37] .

Investigations from a group of Canadian researchers identified other warheads for this lead molecule with potential for further development [38, 39] . The identification of a potent warhead was also the focus of Hilgenfeld and co-workers [40] .

The structurally similar GC376 was originally identified as active against a feline coronavirus [41] and more recently confirmed as a SARS-CoV-2 3CL pro inhibitor [26] , and structure-activity relationship studies led to improved analogues [42] . Smaller, more drug-like inhibitors such as the isatine derivative 2 have been devised [43, 44] . More recently, Jorgensen and co-workers converted Perampanel, a known antiepileptic drug that is also a weak 3CL pro inhibitor, into potent inhibitors (3, Figure 1 ) using a combination of computational and experimental investigation [45] .

Reported covalent SARS-CoV-2 3CL pro inhibitors 1, GC376 [26] , PF-07321332, PF-0730814 [32, 35, 36] and an analogue [38] , and reported non-covalent inhibitors Masitinib [46] J o u r n a l P r e -p r o o f and Walcyrin B [47] . Reported inhibitors of SARS-CoV 3CL pro 2 [43] and 3 [45] . Orange spheres indicate the warheads for covalent binding.

As described in our recent review [48] , covalent drugs can be extremely effective and useful pharmaceuticals, yet they have been largely ignored in most drug design endeavours and particularly in those concerning structure-based drug design. Concerns about their potential offtarget reactivity and toxicity have often been raised [49] . Despite these concerns, there are many examples of covalent drugs on the market, including two of the ten most widely prescribed medications in the U.S., as well as several other common drugs like aspirin and penicillin [48] .

The advantages of covalent drugs are becoming increasingly recognized: they have extremely high potencies, long residence times, and high levels of specificity [50] . Although skepticism persists, many pharmaceutical companies are embracing covalent drugs as exemplified by Neratinib (Nerlynx®, Pfizer) and Afatinib (Gilotrif®, Boehringer-Ingelheim).

3CL pro inhibitor design. Many of the structure-based studies related to COVID-19 to date have employed virtual screening and machine learning techniques. Several potential 3CL pro inhibitors have been identified, however experimental verification has lagged [51] [52] [53] . As of today, much of the research has focused on peptidic substrate-like inhibitors ( Figure 1 ). There is currently a need for the development of drug-like inhibitors with synthetically accessible scaffolds that will allow for more thorough investigations of structure-activity relationships. We thought to benefit from our team's expertise in covalent inhibition and from our software that enables automated docking and virtual screening of covalent inhibitors, which is not possible with most commercial packages.

We present herein our efforts towards the development of novel potent covalent inhibitors of 3CL pro .

Inhibitor design through covalent docking. In the past years, we have successfully applied covalent docking to the design and discovery of prolyl oligopeptidase inhibitors [54, 55] and thought to apply a similar strategy to develop SARS-CoV-2 3CL pro inhibitors. An investigation of the crystal structure of a non-covalent inhibitor (X77, Figure 2 ) bound to 3CL pro of SARS-CoV-2 (PDB code: 6W63) suggested that it might be possible to modify this inhibitor by incorporating a covalent warhead in proximity to the catalytic cysteine residue. As shown in Figure 2 , the sulphur atom of Cys 145 is positioned at 3.2 Å from the imidazole moiety and at the same location as the J o u r n a l P r e -p r o o f covalent warhead of PF-00835231. Thus, replacement of the imidazole with a covalent warhead appeared to be a promising strategy to improve the inhibitory potency of this non-covalent inhibitor. Additionally, this scaffold could be prepared via a 4-component Ugi reaction [56] , enabling a combinatorial approach that would provide an efficient synthetic method for preparing diverse analogues. This would provide a significant advantage in exploring structure-activity relationships when compared to previously reported inhibitors, as a wide range of covalent warheads could be readily incorporated into the same inhibitor scaffold. As a note, a consortium of research groups including a group at the Weizmann Institute of Science in Rehovot (Israel) took a very similar strategy although focusing primarily on non-covalent inhibitors [57] . To validate the design strategy, a virtual library of modified inhibitors was prepared based on incorporation of covalent warheads that could be accessed via a traditional or modified Ugi 4 component coupling (4CC) reaction ( Figure 3 ). These compounds were docked to 3CL pro (PDB code: 6W63) using our docking program, FITTED [59] . The docked poses ( Figure 4 ) suggested that J o u r n a l P r e -p r o o f many of these modified inhibitors would be able to maintain the same non-covalent interactions as the original non-covalent inhibitor while also positioning the warhead close enough to Cys 145 to facilitate the formation of a covalent bond. Based on the promising docking results with multiple warheads, a small library of analogues was synthesized for experimental testing. Synthesis. Following a protocol reported by Jacobs et al. [56] , a 4-component Ugi reaction was used to prepare analogues bearing four different classes of covalent warheads (alkene Michael acceptor, α-halo ketone, alkyne Michael acceptor, and α-ketoamide, Scheme 1a). This synthetic strategy was later employed to probe some of the features of this class of inhibitors by modifying the R 2 , R 3 and R 4 groups.

J o u r n a l P r e -p r o o f Scheme 1. a) carboxylic acid (1.0 mmol, 1.0 eq.), 4-tert-butylaniline (1.0 mmol, 1.0 eq.), 3pyridinecarboxaldehyde (1.0 mmol, 1.0 eq.), cyclohexyl isocyanide (0.9 mmol, 0.9 eq.), MeOH (5 mL, 0.2 M), r.t., overnight, X77 (30%), ML188 (84%), 4a (76%), 6a (32%), 7a (94%), 8a (32%), 9a (73%), 10a (68%), 11a (80%), 12a (37%), 13a (92%), 15a (70%), 16a (93%), 17a (52%), 18a (36%), 20a (84%), 21a (68%), 22a (29%), 23a (43%), 8b (54%), 13b (82%), 6b (42%), 4b (84%), 13c (94%), 13d (86%), 16b (88%), 16c (91%), 11b (67%), 11c (79%), 11d (59%), 11e (26%), 11f (71%), 11g (29%), 11h (78%), 11i (37%), 8c (74%), 16d (87%), 16f (34%), 13e (90%), 13f (69%), 13g (91%), 13h (92%), 13i (22%), 13j (47%), 18b (37%), 13k (83%), 13l (57%), 13m (82%), 13n (89%), 13o (88%), 13p (89%), 13q (35%). b) 3-pyridinecarboxaldehyde (2.0 mmol, 2.0 eq.), glycinamide hydrochloride (2.0 mmol, 2.0 eq.), triethylamine (2.0 mmol, 2.0 eq.), acetic acid (2.0 mmol, 2.0 eq.), cyclohexyl isocyanide (2.0 mmol, 2.0 eq.), MeOH (5 mL, 0.2 M), r.t., overnight, 24a (79%). c) 3-pyridinecarboxaldehyde (1.0 mmol, 1.0 eq.), β-alanine (1.0 mmol, 1.0 eq.), cyclohexyl isocyanide (1.0 mmol, 1.0 eq.), MeOH (5 mL, 0.2 M), r.t., overnight, 25a (54%). D) phosphoric acid, MeOH, r.t., 68%.

A 3-component Ugi reaction was used to prepare additional analogues bearing nitrile [60] and β-lactam [61] covalent warheads following reported procedures (Scheme 1b, c and d).

In order to complete the synthesis of the analogues featuring warheads not accessible through J o u r n a l P r e -p r o o f Scheme 2. a) H3PO4, MeOH, rt, 5a (96%); b) Cl-CH2-CH2-SO2Cl, pyridine, DCM, 0ºC to rt, 14a (35%); c) Cl-CO2Et, pyridine, DCM, 0ºC to rt, 19a (76%).

In order to further probe the R 4 group of the potential inhibitors, a recently reported strategy to make isocyanides [62, 63] was employed (Scheme 3). The strategy illustrated in Scheme 2 was then used to prepare these analogues. Scheme 3. a) HCO2Et, 60 ºC, 26b (95%), 26c (98%), 26d (40%), 26e (99%), 26f (99%), 26g (99%); b) POCl3, Et3N, DCM, 27b (73%), 27c (83%), 27d (79%), 27e (71%), 27f (50%), 27g (74%); c) H3PO4, MeOH, rt, 28b (42%), 28c (45%), 28d (48%), 28e (26%), 28f (25%), 28g (40%); c) Cl-CH2-CH2-SO2Cl, pyridine, DCM, 0ºC, 14b (66%), 14c (45%), 14d (73%), 14e (80%), 14f (42%), 14g/h (75%);

3CL pro inhibitioncovalent warheads. Twenty potential warheads (compounds 6a-25a, Figure 3 ) and four non-covalent analogues (X77, ML188, 4a and 5a) were synthesized and evaluated for their inhibitory potency using a fluorescence inhibition assay. The compounds were initially screened at 50 μM and IC50 values were subsequently determined for compounds displaying greater than 80% inhibition (Table 1) . 

To evaluate the covalent inhibition hypothesis, the time dependence of our most potent inhibitors, 16a and 14a, were measured ( Figure S4 ). The level of inhibition increased with incubation time, an observation that is consistent with the formation of a covalent adduct. Additionally, the presence of 3CL pro -16a and -14a adducts were confirmed by LC-MS ( Figures S4a/b) . As a definitive proof of covalent inhibition and binding mode, crystal structures of 3CL pro co-crystallized with 16a and 14a were obtained ( Figure 5 ). J o u r n a l P r e -p r o o f Analysis of covalent warheads. We observed that GC376, X77 and ML188, previously reported SARS-CoV 3CL pro inhibitors (X77: IC50 = 3.4 μM [64] and ML188: IC50 = 4.8 μM [56] respectively) and SARS-CoV2 inhibitor (GC376: 139 nM), also inhibit SARS-CoV-2 3CL pro with similar potencies with our assay. (Table 1 , entries 1 to 3). Gratifyingly, low micromolar to submicromolar potencies were also observed for the covalent analogues containing acrylamide (6a), alkynylamide (13a), vinyl sulfonamide (14a), α-chloroamide (16a) and α-ketoamide (18a)

warheads. Interestingly, our two most potent inhibitors 16a and 14a (IC50 = 0.4 and 0.5 μM) were an order of magnitude more potent than the original non-covalent hit molecule (X77: IC50 = 4.1 μM).

An increase in potency was observed when increasing the electrophilicity of the warheadthe α-chloroamide (16a) was more active than the corresponding α-fluoroamide (15a), and the vinyl sulfonamide (14a) was more active than the corresponding acrylamide (6a). However, this trend was not observed when increasing the electrophilicity of the ketoamide (18a) with a CF3 group (22a), potentially due to an increase in the steric bulk and/or electrostatic properties of the warhead that are not tolerated in the active site. Similarly, while acrylamides are typically more reactive with cysteine than the corresponding alkynylamides (when tested in glutathione or cysteine binding assays [65] ), the aklynylamide warhead was more active against 3CL pro . A possible explanation could be that the sp geometry of the aklynylamide warhead positions the electrophile more favorably to the cysteine residue to facilitate covalent bond formation.

The binding pocket also seems to favor smaller warheadsany steric bulk around the acrylamide warhead resulted in a decrease in potency, regardless of electronics. As mentioned previously, a similar effect was observed when comparing the activity of inhibitor 22a and 18a.

The position of the covalent bond formation also appears to influence inhibitor activity. Minimal Another observation is the significant loss of potency when removing the heterocyclic ring of X77 or ML188 (X77/ML188 vs. 4a, Table 1 ). As illustrated in Figure 2 , the basic imidazole nitrogen of X77 forms a hydrogen bond interaction with the backbone of Gly 143 , an interaction that is also observed with the furan ring of ML188 (PDB code: 3V3M) or other heterocycles of the same chemical series [56, 66] . Gly 143 , together with the backbone amides of Ser 144 and Cys 145 , forms an oxyanion hole that contributes to the catalytic activity of this enzyme. Substitution of this heterocycle with a carbocycle of similar size but no hydrogen bonding groups (compounds 11a vs.

does not preserve the inhibitory potency even when this ring was converted to a warhead for covalent binding. Additionally, no difference in activity was observed when comparing the activity of compounds 4a and 5a, suggesting that the carbonyl group does not contribute significantly to the inhibitor's binding affinity.

As shown in Figure 2 , both X77 and PF-00835231 interact with the catalytic His 41 , either through a direct hydrogen bond in the case of PF-00835231, or though a water-mediated hydrogen bond (X77) via a conserved water molecule. In an attempt to reproduce this interaction, longer covalent groups were designed by incorporating an ethyl ester to an acrylamide warhead (17a) and

by incorporating a hydroxyl group to an alkynylamide warhead (23a). While 17a resulted in a loss of potency, 23a led to a nearly 10-fold improvement in potency over the alkynylamide analogue 13a. Although our initial hypothesis was that the increase in potency was due to the formation of a hydrogen-bond interaction with His 41 , docking suggests that the hydroxyl group of 23a instead occupies the oxyanion hole and interacts with the backbone amides of Gly 143 , Ser 144 and Cys 145 ( Figure S7 ).

ITC was employed to validate the initial fluorescence inhibition assay and to gain information of the kinetics of covalent bond formation.

ITC provides unique insights into the kinetics and thermodynamics of inhibitor binding that are not available in traditional enzyme assays [67] . ITC experiments measure the heat produced by We performed activity assays where 3CL pro was titrated into the sample cell, which contained either substrate alone or a mixture of substrate and inhibitor. We then fit the resulting kinetic traces to determine both the kinetic parameters of the enzyme and the rate and mechanism of inhibitor binding. Figure 6a shows the ITC trace obtained when 3CL pro was injected into substrate alone, with the first (0. In all cases, the total amounts of heat (areas of the peaks) were less than that obtained in the absence of inhibitor (panel a), indicating that the enzyme was inactivated before all the substrate was cleaved. To obtain quantitative information on inhibition, the traces were fit to three different models: (i) a reversible (rapid equilibrium) mechanism, E↔EI characterized by an equilibrium affinity constant, Ki; (ii) an irreversible mechanism, E→EI characterized by a second order rate constant kinact; and (iii) a two-step mechanism, E↔EI*→EI, consisting of a rapid pre-equilibrium described by Ki, followed by irreversible inhibition described by a first-order rate constant kinact.

In models (ii) and (iii), kinact can be tentatively assigned to the rate of covalent bond formation with the enzyme. The Michaelis Menten enzymatic parameters were held fixed at the values obtained in (a) and the parameters of the inhibition models were varied to minimize the residual-sum-ofsquared-deviations (RSS). In all cases, the pre-equilibrium irreversible model (iii) gave the best agreement with data. The improvements in RSS given by model (iii) compared to models (ii) and

(i) were calculated using F-test statistics [68] and found to be significant at levels of p≤10 -2 . The extracted values of Ki and kinact are listed in Table S1 . 14c had both the fastest rate of covalent bond formation (largest kinact) and tightest binding in the initial non-covalent step (smallest Ki) of the three inhibitors, tested. Both 14a and 14c have similar values of Ki to the parent non-covalent scaffold, which is consistent with the two-step binding model (iii), since the first step corresponds to non-covalent binding.

Structure-Activity Relationship. Following our search for an optimal warhead, several modifications were made to the core of the molecule (Tables 2, 3 and 4). Previously, it was shown that replacement of the tert-butylphenyl group (R 2 ) by smaller groups or differently substituted phenyl groups modulates the potency against SARS-CoV 3CL pro with slight improvements in some cases, while various hydrophobic groups were tolerated as R 3 [64] . We thought this information could help further improve these inhibitors for the highly homologous SARS-CoV-2 3CL pro . However, all of our attempts to replace the tert-butylphenyl and pyridyl groups have proven unsuccessful to date. More specifically, the use of a variety of aromatic heterocycles such as thiophene, benzothiazole, pyrimidine, pyrazine and benzothiophene led to loss of potency (Table 3) . Additionally, replacement of the heterocyclic ring with a hydantoin moiety also led to a J o u r n a l P r e -p r o o f loss in potency. Similarly, diversely functionalized phenyl groups in place of the tert-butylphenyl group were detrimental for the activity and replacement of the phenyl group with a benzyl moiety led to a complete loss of activity (Table 4) .

Replacement of the cyclohexyl group with groups of similar sizes did not improve the potency (Table 2 , entries 1-7 and 9). However, we have found that substitution of the cyclohexyl (R 4 group in Table 2 ) for longer groups (14b-l) improved the inhibitory potency over that of 16a and 14a.

The longer chains likely enable the inhibitor to form a hydrophobic interaction with Pro 168 , while the amine nitrogen of analogue 14f may facilitate a hydrogen-bond interaction with Gln 189 or the carbonyl of Glu 166 . Based on these optimizations, our current most potent compound (14c) has a potency similar to the previously reported inhibitor GC376. It is expected that one enantiomer is more potent that the other one. To confirm this hypothesis, we used a chiral isocyanide leading to easily separable diastereomers 14h and 14g. Diastereomer 14g was found to be approximately 20 times more potent that 14h in line with what has been observed previously [69] . 

cross-reactivity with the potential viral inhibitors, decreasing the therapeutical effect against SARS-CoV-2. [70] For example, the increase of the cathepsin L levels, a ubiquitous human protease, in plasma of patients with SARS-CoV-2 severe infections generates a target competence for the protease antiviral inhibitors [71] . To assess this potential undesired effect of our most potent 3CL pro inhibitors, we measured the inhibitor effect against cathepsin L. The low inhibition of cathepsin L activity for our inhibitors at 50 µM (25% ± 2, 19% ± 1 and 15% ± 2 for 16a, 14a and 14c, respectively, Figures S3a and S3b) indicates an excellent selectivity of these inhibitors against 3CL pro over cathepsin L.

Covalent inhibition of SARS-CoV-2 3CL pro is a promising strategy for the treatment of COVID-19. Our strategy relied on a previously reported imidazole-containing inhibitor of the similar coronavirus SARS-CoV responsible for the epidemic of SARS in the early 2000's. We first used our docking program FITTED, specifically modified to accommodate covalent inhibitors, and screened a set of covalent warheads. The docked poses confirmed that replacing the imidazole ring by a reactive group should lead to potent covalent inhibition. Gratifyingly, while the imidazole of X77 was known to be essential for the inhibitory potency, replacing it with many warheads maintained and even improved the potency, with our lead compounds 16a and 14a being an order of magnitude more potent. Both the inhibition pattern of enzymatic activity and the biophysical data first suggested that these inhibitors bind covalently to the viral protease, a binding mode later J o u r n a l P r e -p r o o f confirmed by crystallography; thus, the robustness of in silico rational-drug design was validated using in vitro detection of protein processing.

General Considerations. All other reagents were purchased from commercial suppliers and used without further purification. All 1 H, 13 Measured purities for all tested compounds are listed in Table S3 in the supporting information.

In a 6-dram vial equipped with a stir bar aldehyde (1.0 mmol, 1.0 eq.), aniline (1.0 mmol, 1.0 eq.) and carboxylic acid (1.0 mmol, 1.0 eq.) were combined in MeOH (4 mL). The obtained reaction mixture was stirred for 30 min at room temperature. Afterwards cyclohexyl isocyanide (0.9 mmol, 0.9 eq.) was added to the reaction mixture and the walls of the vial were washed with 1 mL of MeOH. The reaction mixture was continued to stir at room temperature overnight. The crude reaction mixture was evaporated in vacuo. Purification procedure A) The crude product was triturated with hexanes (5 mL) and

filtered. The obtained product was further washed with hexanes (3 x 3 mL). Purification procedure B) The crude product recrystallized from CHCl3/hexanes mixture, filtered and the obtained product was further washed with hexanes (3 x 3 mL). Purification procedure C) The crude product was redissolved in DCM. The obtained crude solution was deposited on silica. It was then purified using flash column chromatography using DCM/MeOH (gradient 0  5%) as eluent.

General Procedure D to Prepare Formamides. The synthesis of formamides was derived from known literature [62] . In a 1-dram vial equipped with a stir bar, 5 mmol (1 eq.) of amine was mixed with 15 mmol (3 eq.) of ethyl formate and stirred at 60 ºC until completion (monitored using J o u r n a l P r e -p r o o f TLC -1:1 EtOAc:hexanes or 2:1 EtOAc:hexanes). Once the amine was fully converted, ethyl formate was removed in vacuo and the product was used in the next step without further purification.

General procedure E to Prepare Isocyanides. The synthesis of isocyanides was derived from known literature [63] . In a 6-dram vial equipped with a stir bar, 2 mmol (1 eq.) of formamide was dissolved in 1 mL DCM with 10 mmol (5 eq.) of Et3N. The mixture was cooled to 0 ºC and 2 mmol

(1 eq.) of POCl3 was added dropwise. The reaction mixture was stirred at 0 ºC for 10 minutes then General Procedure G for Synthesis of Vinyl Sulfonamides. The synthesis of vinyl sulfonamides was derived from known literature [72] . In a 6-dram vial equipped with a stir bar, 0.2 to 0.3 mmol (1.0 eq.) of the previously made acetamide (see general procedure F) was dissolved in 5 mL of DCM with 0.7 to 3. 5 mmol (from 0.5 eq. to 3 eq.) of Et3N. The mixture was cooled to 0 ºC and 0.3 to 0.45 mmol (1.5 eq.) of 2-chloroethanesulfonyl chloride was added dropwise. The solution was stirred at 0 ºC for 2 h. The solution was then diluted with 5 mL DCM and washed with 10 mL sat. NaHCO3. The aqueous layer was extracted with 10 mL DCM and the combined organic layer was washed with 10 mL sat. NaCl solution and further dried with anhydrous Na2SO4.

The crude product was purified using flash column chromatography using DCM/ EtOAc (gradient 0%  50%) as eluent. 

carboxamide (ML188

Compound was made and purified using general procedure A, white solid 76 % yield, 280 mg. 1 

cinnamamide (7a). Compound was made and purified using general procedure A, pale white solid 94 % yield, 420 mg. 1 J o u r n a l P r e -p r o o f 

acetamide (16a It was then purified using flash column chromatography using DCM/MeOH (gradient 0  5%) as eluent. The product was obtained as colorless oil, 156 mg 54%. 1 

Compound was made and purified using general procedure A, white solid 84 % yield, 290 mg. 1 

Compound was made and purified using general found 132.0810. 

(28f). Compound was made and purified using general procedure F. Pale yellow solid with 25% yield, 346 mg (3 mmol 3 g used). Rf = 0.31 (5% MeOH in DCM). 1 

Compound was made and purified using general procedure F. Pale yellow solid with 40% yield, N-(4-(tert-butyl)phenyl)-N-(2-(cyclohexylamino)-2-oxo-1-(pyridin-4-yl) N-benzyl-N-(2-(cyclohexylamino)-2-oxo-1-(pyridin-3-yl)ethyl) propiolamide (13q).

Pyridine-3-carboxaldehyde (0.09 mL, 0.93 mmol) and benzylamine (0.10 mL, 0.93 mmol) were stirred in MeOH (3.7 mL) and for 30 min at rt. 2-Butynoic acid (78.5 mg, 0.93 mmol) and

cyclohexyl isocyanide (0.10 mL, 0.84 mmol) were added and the solution stirred at rt overnight.

The reaction was concentrated in vacuo and the crude product was dissolved in EtOAc washed with water and brine. The resulting aqueous phases were extracted three times for residual compound. The combined organic phases were washed with brine and dried over Na2SO4. The crude product was purified using column chromatography with a gradient of mixture of (MeOH + 1% NH4OH) and DCM (0 %  5 %). The reaction was further purified using column chromatography with a gradient of EtOAc in hexanes (0 %  100 % [73] , and Zhang et al. [25] , the protein was designed to end at the last glutamine comprising the natural C-terminus of the enzyme, followed by a glycine, a proline residue and six histidine residues. The protein this has a 6-His purification tag that can be cleaved using a modified-PreScission protease approach (SGVTFQ↓GP).

E. coli BL21 (DE3) cells were transformed using the plasmid described above and the CaCl2 method [74] and colonies were selected on LB-agar-ampicillin plates. A single colony was picked and grew over night in LB media and then used to inoculate a 400 mL LB culture. This was grown mM Tris pH8, 100mM NaCl, 1mM DTT and concentrated to 5mg/ml, and was then incubated with 450 µM of compound 16a for 1 hour at room temperature. Following incubation, the sample was filtered using a 0.22 µm filter and used for crystallization trials. Crystals were grown using the sitting drop method at 22 °C. 200 nL enzyme was mixed with 200 nL well solution (30% PEG2000 MME, 0.2M Potassium thiocyanate) and allowed to equilibrate against 50 µL well solution. The crystals were cryo-protected using well solution supplemented with 20% ethylene glycol and flash frozen in liquid nitrogen. Data was collected at the Canadian Light Source CMCF-BM beamline and processed in space group C 1 2 1. The structure was solved in PHASER [76] with a previously published structure of the enzyme (6WTK) [26] as a search model. Restraints for the covalently bonded inhibitor were generated using AceDRG [77] in CCP4i2 [78] , and the model was refined with REFMAC5 [79] and Coot [80] .

Addition information on LC-MS data, ITC measurement, dose-response curves and 1 H and 13 C NMR spectra.

tert-butyl)phenyl)amino)-N-(3-chlorophenethyl)-2-(pyridin-3-yl)acetamide (28c)

Compound was made and purified using general procedure F. Off-white solid with 45% yield

mg. Rf = 0.25 (1:1 EtOAc:hexanes). 1 H NMR (500 MHz, CDCl3): δ 8.64 (d, J = 20.0 Hz, 1H), 7.63 (dt, J = 7.9, 2.0 Hz, 1H), 7.32 -7.30 (m ,1H), 7.27 (d, J = 8.7 Hz, 2H), 7.21 -7.19 (m, 1H), 7.15 (t, J = 7.7 Hz, 1H), 7.09 (t, J = 1.9 Hz, 1H), 6.92 -6.87 (m, 2H)

EtOAc:hexanes). 1 H NMR (500 MHz, CDCl3): δ 8.64 (d, J = 3.3 Hz, 1H), 8.60 (dd

Hz, 1H), 6.87 (t, J = 5.9 Hz, 1H)

13 C NMR (126 MHz, CDCl3): δ 170

EtOAc:hexanes). 1 H NMR (500 MHz, CDCl3) δ 8.71 (d, J = 2.3 Hz, 1H), 8.62 (dd

Hz, 2H), 2.59 (td, J = 8.3, 7.6, 2.4 Hz, 2H), 1.88 -1.81 (m, 2H), 1.30 (s, 9H). 13 C NMR (126 MHz, CDCl3)

Compound was made and purified using general procedure B, off white solid 37 % yield

Hz, 1H), 7.28 (d, J = 8.6 Hz, 2H), 7.11 (s, 1H)

Hz, 2H), 1.73 -1.65 (m, 4H), 1.37 -1.22 (m, 15H). 13 C NMR (101 MHz, CDCl3) δ 169

enamide (8c). Compound was made and purified using general procedure A, white solid 74 % yield, 287 mg. 1 H NMR (500 MHz, CDCl3) δ 8.49 -8.45 (m, 2H), 7.28 (d, J = 8.9 Hz, 2H)

(m, 3H), 1.70 -1.64 (m, 2H), 1.58 (dt

ethyl)acetamide (16d). Compound was made and purified using general procedure A, white solid 87 % yield, 352 mg. 1 H NMR (500 MHz, CDCl3) δ 7.28 (s, 2H), 7.25 -7.21 (m, 1H)

88 (d, J = 8.0 Hz, 1H), 5.83 (s, 1H), 3.87 (d

Compound was made and purified using general procedure B, pale yellow solid 34 % yield, 140 mg. 1 H NMR (500 MHz, CDCl3) δ 8.62 (s, 1H), 8.39 (s, 1H), 7.37 (d, J = 8.2 Hz, 2H

Compound was made and purified using general procedure A, pale white solid 90

% yield, 392 mg. 1 H NMR (500 MHz, CDCl3) δ 9.11 (d, J = 0.7 Hz, 1H), 8.24 (s, 1H), 8.00 -7.94 (m, 2H), 7.82 -7.73 (m, 1H), 7.65 (ddd, J = 7.9, 6.9, 1.0 Hz, 1H)

Compound was made and purified using general procedure A, pale white solid 69

% yield, 298 mg. 1 H NMR (500 MHz, CDCl3)

Hz, 1H), 8.00 (d, J = 2.4 Hz, 1H)

Compound was made and purified using general procedure A, pale white solid 91 % yield

phenyl)but-2-ynamide (13h). Compound was made and purified using general procedure A, grey solid 92 % yield, 402 mg. 1 H NMR (500 MHz, CDCl3) δ 7.86 -7.82 (m, 1H), 7.77 -7.71 (m, 1H), 7.50 (s, 1H), 7.45 -7.34 (m, 2H), 7.14 -7.08 (m, 2H)

07 (s, 1H), 7.02 (s, 1H), 6.22 (d, J = 8.0 Hz, 1H), 6.14 (s, 1H), 4.06 -3.86 (m, 1H), 2.35 (s, 3H), 2.32 (s, 3H

Compound was made and purified using general procedure A, grey solid 83 % yield

Hz, 1H), 6.98 (s, 2H), 6.79 -6.70 (m, 2H), 6.10 (s, 1H), 6.00 (d, J = 8.0 Hz, 1H), 3.88 -3.81 (m, 1H), 3.79 (s, 3H), 1.94 -1.84 (m, 2H), 1.74 (s, 4H), 1.61 (s, 1H), 1.42 -1.11 (m, 6H). 13 C NMR (126 MHz, CDCl3)

ynamide (13m). Compound was made and purified using general procedure A, white solid 82 % yield, 320 mg. 1 H NMR (500 MHz, CDCl3) δ 8.42 (s, 2H), 7.96 (s, 1H), 7.83 -7.76 (m, 1H), 7.49 J o u r n a l P r e -p r o o f (dt, J = 7.9, 2.1 Hz, 2H)

Compound was made and purified using general procedure A, light brown solid 89 % yield, 345 mg. 1 H NMR (500 MHz, CDCl3)

cyclohexylamino)-2-oxo-1-(pyridin-3-yl)ethyl)-N-(4-(difluoromethoxy)phenyl)but-2-ynamide (13p

Hz, 1H), 7.50 (dt, J = 8.0, 2.0 Hz, 1H), 7.20 -7.10 (m, 2H), 7.03 -6.93 (m, 2H), 6.50 (t, J =

SARS-CoV-2 in fruit bats, ferrets, pigs, and chickens: an experimental transmission study

A review of studies on animal reservoirs of the SARS coronavirus

Temporal estimates of case-fatality rate for COVID-19 outbreaks in Canada and the United States

WHO Coronavirus Disease (COVID-19) Dashboard, in

Estimating COVID-19 mortality in Italy early in the COVID-19 pandemic

Pfizer's CEO says Covid vaccine effectiveness drops to 84% after six months

Six months of COVID vaccines: what 1.7 billion doses have taught scientists

Viral targets for vaccines against COVID-19

The Effectiveness of the Two-Dose BNT162b2 Vaccine: Analysis of Real-World Data

Merck and Ridgeback's Molnupiravir, an Oral COVID-19 Antiviral Medicine, Receives First Authorization in the World

Targeting SARS-CoV-2 Proteases and Polymerase for COVID-19 Treatment: State of the Art and Future Opportunities

Genome Composition and Divergence of the Novel Coronavirus (2019-nCoV) Originating in China

Structure-Guided Optimization of Dipeptidyl Inhibitors of Norovirus 3CL Protease

Design, synthesis, and evaluation of 3C protease inhibitors as anti-enterovirus 71 agents

Structure-Based Design, Synthesis, and Biological Evaluation of Irreversible Human Rhinovirus 3C Protease Inhibitors. 4. Incorporation of P1 Lactam Moieties as l-Glutamine Replacements

Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors

Anti-SARS-CoV-2 activities in vitro of Shuanghuanglian preparations and bioactive ingredients

Malleability of the SARS-CoV-2 3CL Mpro Active-Site Cavity Facilitates Binding of Clinical Antivirals

Structural plasticity of SARS-CoV-2 3CL Mpro active site cavity revealed by room temperature X-ray crystallography

An Overview of Severe Acute Respiratory Syndrome-Coronavirus (SARS-CoV) 3CL Protease Inhibitors: Peptidomimetics and Small Molecule Chemotherapy

Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved α-ketoamide inhibitors

Feline coronavirus drug inhibits the main protease of SARS-CoV-2 and blocks virus replication

An automatic pipeline for the design of irreversible derivatives identifies a potent SARS-CoV-2 Mpro inhibitor

Coronavirus Main Proteinase (3CLpro) Structure: Basis for Design of Anti-SARS Drugs

Research and Development on Therapeutic Agents and Vaccines for COVID-19 and Related Human Coronavirus Diseases

Rational design of a new class of protease inhibitors for the potential treatment of coronavirus diseases, bioRxiv

Lead compounds for the development of SARS-CoV-2 3CL protease inhibitors

The Discovery of Ketone-Based Covalent Inhibitors of Coronavirus 3CL Proteases for the Potential Therapeutic Treatment of COVID-19

Discovery of Ketone-Based Covalent Inhibitors of Coronavirus 3CL Proteases for the Potential Therapeutic Treatment of COVID-19

Study of PF-07321332 in healthy participants -NCT04756531

Pfizer's novel COVID-19 oral antiviral treatment candidate reduced risk of hospitalization o death by 89% in iterim analysis of phase 2/3 EPIC-HR study

Peptidomimetic α-Acyloxymethylketone Warheads with Six-Membered Lactam P1 Glutamine Mimic: SARS-CoV-2 3CL Protease Inhibition, Coronavirus Antiviral Activity, and in Vitro Biological Stability

Peptidomimetic nitrile warheads as SARS-CoV-2 3CL protease inhibitors

α-Ketoamides as Broad-Spectrum Inhibitors of Coronavirus and Enterovirus Replication: Structure-Based Design

Reversal of the Progression of Fatal Coronavirus Infection in Cats by a Broad-Spectrum Coronavirus Protease Inhibitor

Improved SARS-CoV-2 Mpro inhibitors based on feline antiviral drug GC376: Structural enhancements, increased solubility, and micellar studies

Synthesis and evaluation of isatin derivatives as effective SARS coronavirus 3CL protease inhibitors

Aryl methylene ketones and fluorinated methylene ketones as reversible inhibitors for severe acute respiratory syndrome (SARS) 3C-like proteinase

Potent Noncovalent Inhibitors of the Main Protease of SARS-CoV-2 from Molecular Sculpting of the Drug Perampanel Guided by Free Energy Perturbation Calculations

Drug repurposing screen identifies masitinib as a 3CLpro inhibitor that blocks replication of SARS-CoV-2

Identification of SARS-CoV-2 3CL Protease Inhibitors by a Quantitative High-Throughput Screening

Covalent inhibitors design and discovery

Drug bioactivation, covalent binding to target proteins and toxicity relevance

Covalent Drugs Form Long-Lived Ties

AI-aided design of novel targeted covalent inhibitors against SARS-CoV-2, bioRxiv

Potentially highly potent drugs for 2019-nCoV, bioRxiv

Main Protease Through Virtual Screening, ADMET, and Molecular Dynamic Simulation Studies

Moitessier, 3-Oxo-hexahydro-1H-isoindole-4-carboxylic acid as a drug chiral bicyclic scaffold: structure-based design and preparation of conformationally constrained covalent and noncovalent prolyl oligopeptidase inhibitors

Discovery of covalent prolyl oligopeptidase boronic ester inhibitors

Based Optimization of a Series of N-(tert-Butyl)-2-(Narylamido)-2-(pyridin-3-yl) Acetamides (ML188) as Potent Noncovalent Small Molecule Inhibitors of the Severe Acute Respiratory Syndrome Coronavirus (SARS-CoV) 3CL Protease

An automatic pipeline for the design of irreversible derivatives identifies a potent SARS-CoV-2 Mpro inhibitor, bioRxiv

Structure of COVID-19 main protease bound to potent broad-spectrum non-covalent inhibitor X77

Medicinal chemistry projects requiring imaginative structure-based drug design methods

Substituted 2-(Cyanomethyl-amino)-acetamides by a Novel Three-Component Reaction

Liquid-Phase Combinatorial Synthesis of Alicyclic β-Lactams via Ugi Four-Component Reaction

An efficient, catalyst-and solvent-freeNformylation of aromatic and aliphatic amines

Isocyanide 2.0, Green Chemistry

Broad spectrum non-covalent coronavirus protease inhibitors

An electrophilic warhead library for mapping the reactivity and accessibility of tractable cysteines in protein kinases

Discovery of N-(benzo[1,2,3]triazol-1-yl)-N-(benzyl)acetamido)phenyl) carboxamides as severe acute respiratory syndrome coronavirus (SARS-CoV) 3CLpro inhibitors: Identification of ML300 and noncovalent nanomolar inhibitors with an induced-fit binding

Rapid measurement of inhibitor binding kinetics by isothermal titration calorimetry

Data Reduction and ErrorAnalysis for the Physical Sciences

Expedited Approach toward the Rational Design of Noncovalent SARS-CoV-2 Main Protease Inhibitors

Challenges for Targeting SARS-CoV-2 Proteases as a Therapeutic Strategy for COVID-19

Cathepsin L plays a key role in SARS-CoV-2 infection in humans and humanized mice and is a promising target for new drug development

Thermal and high pressure intramolecular Diels-Alder reaction of vinylsulfonamides

Production of Authentic SARS-CoV Mpro with Enhanced Activity: Application as a Novel Tag-cleavage Endopeptidase for Protein Overproduction

Studies on transformation of Escherichia coli with plasmids

Enzymatic activity characterization of SARS coronavirus 3C-like protease by fluorescence resonance energy transfer technique

Phaser crystallographic software

Validation and extraction of molecular-geometry information from small-molecule databases

CCP4i2: the new graphical user interface to the CCP4 program suite

Direct incorporation of experimental phase information in model refinement

Coot: model-building tools for molecular graphics

Covalent docking was used to design SARS-CoV-2 3CLpro inhibitors

A micromolar non-covalent SARS-CoV 3CLpro inhibitor was converted into a submicromolar SARS-CoV-2 3CLpro covalent inhibitor

A set of warheads was evaluated and vinyl sulfonamide and alkynylamide were identified as promising warheads

We thank the Canadian Institutes of Health Research (CIHR, OV3-170644), the McGill Interdisciplinary Initiative in Infection and Immunity (MI4) and the Faculty of Science for funding.

J o u r n a l P r e -p r o o f