key: cord-0895839-tv5yyfak
authors: Zhang, Jianmin; Huitema, Carly; Niu, Chunying; Yin, Jiang; James, Michael N.G.; Eltis, Lindsay D.; Vederas, John C.
title: Aryl methylene ketones and fluorinated methylene ketones as reversible inhibitors for severe acute respiratory syndrome (SARS) 3C-like proteinase
date: 2008-03-04
journal: Bioorg Chem
DOI: 10.1016/j.bioorg.2008.01.001
sha: 83992aaca8041550749a8bb1a6607116135ada38
doc_id: 895839
cord_uid: tv5yyfak

The severe acute respiratory syndrome (SARS) virus depends on a chymotrypsin-like cysteine proteinase (3CL(pro)) to process the translated polyproteins to functional viral proteins. This enzyme is a target for the design of potential anti-SARS drugs. A series of ketones and corresponding mono- and di-fluoro ketones having two or three aromatic rings were synthesized as possible reversible inhibitors of SARS 3CL(pro). The design was based on previously established potent inhibition of the enzyme by oxa analogues (esters), which also act as substrates. Structure–activity relationships and modeling studies indicate that three aromatic rings, including a 5-bromopyridin-3-yl moiety, are key features for good inhibition of SARS 3CL(pro). Compound 11d, 2-(5-bromopyridin-3-yl)-1-(5-(4-chlorophenyl)furan-2-yl)ethanone and its α-monofluorinated analogue 12d, gave the best reversible inhibition with IC(50) values of 13 μM and 28 μM, respectively. In contrast to inhibitors having two aromatic rings, α-fluorination of compounds with three rings unexpectedly decreased the inhibitory activity.

Severe acute respiratory syndrome (SARS) is a lifethreatening form of atypical pneumonia that rapidly spread in Asia, North America and Europe in early 2003 [1, 2] . SARS is characterized by high fever, malaise, rigor, headache and non-productive cough or dyspnea and may progress to generalized interstitial infiltrates in the lung, requiring intubation and mechanical ventilation [3] . The causative agent of SARS is a coronavirus (SARS-CoV) [1, 2, 4] . Its 3C-like proteinase (3CL pro ), a chymotrypsin-like cysteine proteinase which is essential for viral replication, has been recognized as a key target for anti-SARS drug design [4] . The active site of SARS-CoV 3CL pro has a catalytic dyad with Cys145 as the nucleophile and His41 as the general base [4, 5] . This proteinase cleaves the initially translated polyproteins pp1a and pp1ab at 11 conserved interdomain sites, in which the P1 position has a well-conserved Gln residue, and the P2 position has a hydrophobic amino acid residue [6] .

Currently, there are no effective anti-viral agents for the treatment of SARS. However, rational design and librarybased screening have identified some small molecules that are potent inhibitors of 3CL pro , such as the HIV proteinase inhibitor TL-3 [7] , bifunctional aryl boronic acids [8] , thiophenylcarboxylate [9] , keto-glutamine analogues [10a], AG7088 analogues [11,12a] [10b] . In our previous studies, we demonstrated that a series of 5-halopyridin-3-yl aromatic esters act as very potent covalent inhibitors of 3CL pro with IC 50 values in the low nanomolar range (Fig. 1) [10b]. These compounds initially bind competitively and strongly to the active site, but are then hydrolyzed by the enzyme as substrates and released. Despite their potent inhibition of 3CL pro and relatively long half life in buffer at neutral pH, they are likely to be problematic as drug candidates because of the propensity of esters to be rapidly hydrolyzed by lipases, esterases and other enzymes in mammalian cells. These compounds and many of the other reported inhibitors of SARS-CoV 3CL pro can potentially also react non-specifically with other thiols or nucleophiles in mammalian cells, thereby leading to toxicity.

In order to develop stable and non-covalent inhibitors based on these pyridinyl esters, the ketone 5 was initially investigated as a potential SARS-CoV 3CL pro inhibitor (Fig. 2) . The carbonyl of this relatively unreactive compound could form a hemithioacetal with the enzyme's active site thiol, but this should be a reversible process.

Disappointingly, compared to its oxa analogue, the pyridinyl ester 1 (91% inhibition at 100 lM concentration, IC 50 = 7.9 lM, Fig. 1 ), the ketone analogue 5 displays no inhibition at 100 lM concentration [10b]. However, it seemed that corresponding fluorinated ketones might be better inhibitors, as they possess the combined features of a more electrophilic carbonyl group and a chemical structure that fits the enzyme active site. The fluorinated ketones should be more stable in vivo than the corresponding pyridinyl esters, and would also avoid the hydrolysis problem that leads to the reactivation of the enzyme [10b]. Furthermore, the difluoromethylene moiety has been proposed as a mimic of an oxygen atom in biological systems [13] . In the present work, we examine a group of methylene ketones and corresponding mono-and di-fluorinated methylene ketones as potential SARS-CoV 3CL pro inhibitors.

All reactions involving air or moisture sensitive reactants were done under argon. All reagents employed were of American Chemical Society (ACS) grade or finer and were used without further purification unless otherwise stated. Solvents were dried for anhydrous reactions. All reactions and fractions from column chromatography were monitored by thin layer chromatography (TLC) using glass plates with a UV fluorescent indicator (normal SiO 2 , Merck 60 F 254 ). Flash chromatography was performed using Merck type 60, 230-400 mesh silica gel. Nuclear magnetic resonance (NMR) spectra were obtained on Inova Varian 300, 400 and 500 MHz spectrometers. All the protons were assigned based on the structure numbering system indicated in the paper. The enzyme assay, mass spectrometry of enzyme-inhibitor complexes and molecular docking were performed as described in the literature [10a,10b]. Briefly, the assay was performed at 22°C in a solution containing 20 mM Bis-TRIS at pH 7.0, 2 mM DTT, 10 lM fluorogenic substrate (Abz-SVTLQSG-Tyr (NO 2 )R, 93% purity), 0.2 lM non-tagged 3CL pro and 1% DMSO without any preincubation. The increase in fluorescence (k ex 340 nm, k em 420 nm) was monitored using a spectrofluorimeter. Initial rates were calculated using the first 5 min of the progress curves.

2.2. 2-Fluoro-2-(pyridin-3-yl)-1-(thiophen-2-yl)ethanone (6) and 2,2-difluoro-2-(pyridin-3-yl)-1-(thiophen-2-yl)ethanone (7) To a solution of 5 [10b] (0.14 g, 0.70 mmol) in dry THF (25 mL) at À78°C was added LiHMDS (0.84 mL, 1.0 M solution in THF, 0.84 mmol) dropwise over 15 min. After 1.5 h of stirring at À 78°C, a solution of N-fluorobenzenesulfonimide (NFSi) (0.24 g, 0.77 mmol) in THF (5 mL) was added slowly. The mixture was stirred at À78°C for 6 h and then quenched with 1 M aqueous HCl (1 mL). Satu- rated NaHCO 3 was added to adjust the pH to 9 and the resulting solution was extracted with CHCl 3 (3 Â 50 mL). The combined organic layers were washed with brine (20 mL), dried over MgSO 4 and concentrated in vacuo. Purification of the product by flash chromatography on silica gel (EtOAc) afforded 6 (98 mg, 63%) and 7 (9.7 mg, 6 %) as solids.

Data 

To a solution of 19 (0.32 g, 2.10 mmol) in H 2 O (5 mL) was added conc. HCl (5 mL), and the mixture was heated to reflux at 100°C overnight. The solvent was removed in vacuo to afford (5-chloronicotin-3-yl)acetic acid, which was used for next reaction without further purification. A solution of this acid in MeOH (5 mL) was treated with a solution of MeOH (3 mL) and SOCl 2 (1 mL). The resulting mixture was heated to reflux overnight, and then the solvent was removed in vacuo. The residue was diluted with saturated NaHCO 3 , and the solution was extracted with EtOAc (3 Â 20 mL). The combined organic layers were washed with brine (15 mL), dried over MgSO 4 and concentrated in vacuo. The crude product was purified by column chromatography (EtOAc) to yield 9a as a liquid (0.28 g, 72% over two steps 

This was obtained from 3-chlorophenylacetic acid (1.71 g, 10 mmol) following the procedure described for the preparation of 9a. Purification by flash chromatography on silica gel (10:90 EtOAc/hexanes) afforded 9b as a liquid (1.63 g, 88%). Literature compound [14] . IR (CHCl 3 cast): 3000, 2953, 2843, 1741, 1599, 1576, 1477, 1434 cm À1 ; 1 H NMR (CDCl 3 , 400 MHz) d 7.31-7.14 (m, 4H, PhH), 3.71 (s, 2H, CH 2 CO 2 ), 3.61 (s, 3H, CO 2 CH 3 ); HRMS (EI) calcd for C 9 H 9 ClO 2 (M + ), 184.0291; found: 184.0290.

This was obtained from (5-bromopyridin-3-yl)acetic acid (8c) (1.04 g, 4.80 mmol) following the procedure for 9a. Purification by flash chromatography on silica gel (EtOAc) afforded 9c as a liquid (1.00 g, 91%). Literature compound [15] 

To a solution of 9a (175 mg, 0.94 mmol) in THF (5 mL) at À78°C was added LiHMDS (1.0 mL of 1.0 M solution in THF, 1.0 mmol) dropwise over 15 min. The solution was stirred for 1 h at À78°C. To this solution was added dropwise over 15 min a solution of 2-furoic acid (50 mg, 0.45 mmol) and carbonyl diimidazole (CDI) (80 mg, 0.49 mmol) in anhydrous THF (5 mL), which had been previously stirred for 1 h at 20°C. The mixture was stirred for 4 h at À78°C, and quenched with 1.0 M aqueous HCl (10 mL). The pH was adjusted to $8 by adding saturated aqueous NaHCO 3 , and the solution was extracted with EtOAc (3 Â 25 mL). The combined organic layers were washed with brine (15 mL), dried over anhydrous MgSO 4 and concentrated in vacuo. The crude product was purified by column chromatography (50:50 EtOAc/hexanes) to yield 10a as a liquid (90 mg, 72% 

This was obtained from 9c (1.81 g, 7.86 mmol) following the procedure for 10a. Purification by flash chromatography on silica gel (33:67 EtOAc/CHCl 3 ) afforded 10d as an oil (1.35 g, 83% 

A solution of 10a (70 mg, 0.25 mmol) in 50% H 2 SO 4 (5 mL) was heated to reflux at 100°C for 8 h. NaOH (6.25 M, 10 mL) and saturated NaHCO 3 (9 mL) were then added to neutralize the solution to pH 7. The solution was extracted with EtOAc (3 Â 20 mL). The combined organic layers were washed with brine (15 mL), dried over anhydrous MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (EtOAc) to yield 11a as a solid (47 mg, 85% 

This was obtained from 10b (160 mg, 0.58 mmol) following the procedure for 11a. Purification by flash chromatography on silica gel (25:75 EtOAc/hexanes) afforded 11b as an oil (75 mg, 59%), which solidified upon cooling to 4°C. Literature compound [16] . IR (CHCl 3 cast): 3135, 1732, 1673, 1598, 1570, 1466, 1432 cm À1 ; 1 H NMR (CDCl 3 , 500 MHz) d 7.59 (dd, 1H, J = 1.7, 0.7 Hz, H 5 ), 7.28-7.18 (m, 5H, H 3 and PhH), 6.53 (dd, 1H, J = 3.6, 1.7 Hz, H 4 ), 4.10 (s, 2H, COCH 2 ); HRMS (EI) calcd for C 12 H 9 ClO 2 (M + ), 220.0291; found: 220.0291.

This was obtained from 10c (0.52 g, 1.59 mmol) following the procedure for 11a. Purification by flash chromatography on silica gel (EtOAc) afforded 11c as a solid (0.30 g, 68%) that sublimes under vacuum. 

This was obtained from 10d (1.22 g, 2.81 mmol) following the procedure for 11a. Purification by flash chromatography on silica gel (EtOAc) afforded 11d as a solid (0.90 g, 85% 

To a solution of 11a (20 mg, 0.09 mmol) in dry THF (5 mL) was added LiHMDS (0.1 mL, 1.0 M solution in THF, 0.1 mmol) over 5 min. The mixture was stirred for 1 h at À78°C. A solution of NFSi (32 mg, 0.1 mmol) in dry THF (3 mL) was added dropwise to the mixture over 10 min. This was stirred for another 2 h at À78°C. Saturated NaHCO 3 (5 mL) was added to adjust the pH to 9, and the solution was extracted with EtOAc (3 Â 15 mL). The combined organic layers were washed with brine (10 mL), dried over MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (50:50 EtOAc/hexanes) to yield 12a as a solid (16 mg To a solution of 11a (10 mg, 0.045 mmol) in dry THF (5 mL) was added LiHMDS (0.1 mL, 1.0 M solution in THF, 0.1 mmol) over a period of 5 min. The mixture was stirred for 1 h at À78°C, and a solution of NFSi (32 mg, 0.1 mmol) in dry THF (3 mL) was added dropwise over 10 min. The mixture was stirred for another 2 h at À78°C. Saturated NaHCO 3 (5 mL) was added to adjust the pH to 9, and the solution was extracted with EtOAc (3 Â 15 mL). The combined organic layers were washed with brine (10 mL), dried over MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (50:50 EtOAc/hexanes) to yield 13a as a solid 

Acetyl chloride (30 mL) was slowly added to dry MeOH (30 mL) at 0°C to generate HCl and MeOAc. This solution was stirred for 10 min at 0°C, and then was added to a solution of 5-aminonicotinic acid (9.89 g, 71.6 mmol) in MeOH (120 mL) at 0°C. The mixture was heated to reflux for 18 h and then the solvent was removed in vacuo. The residue was treated with saturated NaHCO 3 until pH of the solution was $7, and the resulting solution was extracted with EtOAc (3 Â 80 mL). The combined organic layers were dried over anhydrous MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (EtOAc) to yield 15 as a solid (8.58 g, 79%). Literature compound [17] 

A solution of NaNO 2 (3.26 g, 47.2 mmol) in H 2 O (21 mL) was added over 30 min to a solution of 15 (5.91 g, 38.9 mmol) in conc. HCl (42 mL) and H 2 O (21 mL) at 0°C. The mixture was stirred for another 30 min at 0°C. To this was added HCl solution (10% w/ w, 50 mL), and then a solution of CuCl 2 (8.82 g) and CuCl (42 mg) in HCl solution (10% w/w, 30 mL). The mixture was stirred for 4 h at 0°C and then allowed to warm to 20°C. NaOH solution and saturated NaHCO 3 were added to neutralize the solution to pH $ 7. The aqueous layer (250 mL) was extracted with EtOAc (3 Â 80 mL). The combined organic layers were dried over anhydrous MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (50:50 EtOAc/hexanes) to yield 16 as a solid (4.74 g, 71%). Literature compound [18] . 

To a solution of 16 (0.51 g, 2.96 mmol) in MeOH (5 mL) and H 2 O (5 mL) was added KOH solution (0.24 g, 10% w/ w) to bring the pH between 10 and 11. The mixture was stirred for 24 h at 20°C. A precipitate appeared upon acidifying the mixture to pH 1 with 1 N HCl (5 mL). This solid was collected and then washed several times with H 2 O. The filtrate was concentrated in vacuo and then dissolved in dry MeOH. The insoluble impurities were removed by gravity filtration, and the filtrate was concentrated in vacuo to afford 17 as a solid (0.48 g, 84%). Literature compound [19] . IR 

To a solution of 17 (0.49 g, 2.54 mmol) in THF (20 mL) at 0°C was added dry Et 3 N (0.78 mL, 5.60 mmol), followed by ethyl chloroformate (0.29 mL, 3.05 mmol). The mixture was stirred for 1.5 h at 0°C, and then the precipitate, Et 3 N Á HCl, was removed by gravity filtration. To the filtrate at À78°C was added LiAlH 4 (3.05 mL of 1.0 M solution in THF, 3.05 mmol) over a period of 15 min. The mixture was stirred at À78°C for another 4 h, and then quenched with 5% NaOH (8 mL). The solvent was removed in vacuo and the residue was diluted with H 2 O (20 mL). Saturated NH 4 Cl was added to adjust pH of the solution to 8. The resulting mixture was stirred for 1 h and then EtOAc (30 mL) was added. The solution was filtered through celite, and then the two layers were separated. The aqueous layer was extracted with EtOAc (3 Â 50 mL). The combined organic layers were dried over anhydrous MgSO 4 and the solvent was removed to yield the product (5-chloropyridin-3-yl)methanol as an oil (0.38 g), which was used for next reaction without further purification. To a solution of the above alcohol (0.38 g, 2.65 mmol) in CH 2 Cl 2 (15 mL) at 20°C was added SOCl 2 (0.98 mL, 16.0 mmol). The mixture was stirred for 42 h. The solvent was removed in vacuo, and the residue was treated with saturated NaHCO 3 (25 mL) to give pH 8. The solution was then extracted with EtOAc (3 Â 40 mL). The combined organic layers were dried over anhydrous MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (50:50 EtOAc/hexanes) to yield 18 as a solid (0.32 g, 76% over three steps 

A solution of 18 (0.10 g, 0.62 mmol) and KCN (0.10 g, 1.54 mmol) in dry DMF (3 mL) was stirred for 48 h at 20°C. The solvent was removed in vacuo, and the residue was treated with K 2 CO 3 solution (15 mL, 10% w/w). The solution was then extracted with EtOAc (3 Â 15 mL). The combined organic layers were dried over MgSO 4 and concentrated in vacuo. The product was purified by column chromatography (EtOAc) 

This was prepared by modification of the literature procedure [20] . To a solution of ethyl chloro-oxoacetate (1.6 mL, 14 mmol) in THF (20 mL) was added dropwise TMSCHN 2 (21 mL, 2 M solution in hexane, 42 mmol). After 3 h of stirring at 20°C, the solvent was removed in vacuo and the product was purified by column chromatography on silica gel (25:75 EtOAc/hexanes) to yield 21 as a solid (1.35 g, 68%) 

To a stirred suspension of bis-copper acetylacetonate (5.2 mg) in benzene (5 mL) and 4-chlorobenzonitrile (1.8 g, 13.1 mmol) at reflux temperature was added ethyl diazopyruvate (21) (1.0 g, 6.06 mmol) in benzene (14 mL) during a period of 3 h. The mixture was heated until the complete consumption of the starting material was confirmed by TLC (ca 3 h). The solvent was removed in vacuo, and the residue was diluted with saturated NaHCO 3 solution (30 mL). The solution was then extracted with EtOAc (3 Â 30 mL). The combined organic layers were washed with brine (20 mL), dried over MgSO 4 and then concentrated in vacuo. The residue was purified by column chromatography on silica gel (25:75 EtOAc/hexanes) to afford 22 as a solid (240 mg, 14%). Literature compound [21] 

To a solution of 22 (180 mg, 0.72 mmol) in THF/H 2 O (8 mL/8 mL) at 0°C was added LiOH (39 mg, 0.93 mmol). The resulting solution was stirred for 2 h until complete consumption of the starting material was confirmed by TLC. The reaction was quenched with 1 N HCl to give pH 3, and solvent was removed in vacuo. The residue was diluted with H 2 O (20 mL) and then extracted with EtOAc (3 Â 20 mL). The combined organic layers were washed with brine (10 mL) and dried over MgSO 4 

Initially, the fluorinated methylene ketones 6 and 7 were prepared from the ketone 5 [10b] (Scheme 1). Deprotonation of ketone 5 with 1.05 equivalent of LiHMDS, followed by fluorination with 1.1 equivalent of N-fluorobenzenesulfonimide (NFSi), provides both the monofluoromethylene ketone 6 and the difluoromethylene ketone 7 in 63% and 6% yields, respectively. Compounds 6 and 7 were then tested as inhibitors of SARS 3CL pro using a continuous fluorometric assay in which enzymatic cleavage of an internally quenched fluorogenic peptide substrate, (Abz-SVTLQSG-Tyr(NO 2 )R) leads to enhanced fluorescence [10a,10b]. The monofluoro ketone 6 shows weak inhibition against 3CL pro (10% at 100 lM concentration, Table 1) , and no improved inhibition was observed after 2 h preincubation with the 3CL pro . However, the di-fluoro ketone 7 displays stronger inhibition against 3CL pro (38% at 100 lM concentration, Table 1 ). Similarly, preincubation of 7 with the 3CL pro for 2 h shows no improved inhibition. These results suggest that the fluorinated ketones 6, 7 are non-covalent and completely reversible inhibitors for SARS 3CL pro . As proposed, fluorination leads to the substantial improvement in the inhibitory activity against the 3CL pro . The di-fluoro ketone 7 appears to be a reasonable mimic of the corresponding pyridinyl ester 1 (91% inhibition at 100 lM), with only 2-to 3-fold less potent inhibition. As the pyridinyl ester 1 (IC 50 = 7.9 lM) is only a moderate inhibitor among the pyridinyl esters (2-4, IC 50 = 50-63 nM), it seemed that the difluoromethylene ketone mimics of the very potent esters (i.e. 2-4) could inhibit SARS 3CL pro very strongly in a non-covalent and reversible fashion. Based on this assumption, a series of methylene ketones and fluorinated methylene ketones were synthesized. Synthesis started from halogen-substituted phenyl or pyridinyl acetic acids, most of which are commercially available materials except the 5-chloropyridinyl acetic acid 8a (Scheme 2). Esterification of carboxylic acids 8a-c yields 9a-c in 88% to quantitative yields. Treatment of 9a-c with LiHMDS to generate the anions, followed by the addition of pre-activated CDI/acid solutions, provides the b-keto esters 10a-d in 72-96% yields. Hydrolytic decarboxylation of the b-keto esters 10a-d gives the corresponding ketones 11a-d in 59-85% yields. Fluorination of the ketones 11a-d with 1.1 equivalent of LiHMDS and NFSi affords the corresponding monofluoromethylene ketones 12a-d in 50-88% yields. Similarly, fluorination of the ketones 11a-d with 2.2 equivalent of LiHMDS and NFSi generates the corresponding difluoromethylene ketones 13a-d in 34-85% yields. 

Dash represents <10% inhibition. Compound 5-chloronicotinate 9a is not commercially available and was prepared as described in Scheme 3. Esterification of the commercially available 5-aminonicotinic acid (14) affords methyl 5-aminonicotinate (15) in 88% yield. Then treating the free amine of 15 with sodium nitrite under acidic conditions, followed by the addition of copper (I) chloride and copper (II) chloride yields the methyl 5chloronicotinate (16) in 79% yield. Hydrolysis of 16 with potassium hydroxide, and then acidic workup generates the acid 17 in 84% yield. Activation of 17 with ethyl chloroformate, followed by lithium aluminum hydride reduction produces the corresponding alcohol, which is readily converted by thionyl chloride to the chloride 18 in 76% yield over 3 steps. Nucleophilic attack of 18 by potassium cyanide provides compound 19 in 56% yield. Hydrolysis of 19, followed by esterification gives the desired ester 9a in quantitative yield.

Compounds 11a-d, 12a-d, 13a-d were tested as inhibitors of SARS 3CL pro using the continuous fluorometric assay described above. The testing results, which only examine the initial binding affinity, are listed in Table 1 . Surprisingly, most of the methylene ketones (11a-c) as well as their fluorinated methylene ketone analogues (12a-c, 13a-c) inhibit 3CL pro poorly. However, the methylene ketone 11d and its fluorinated methylene ketone analogues 12d, 13d are good inhibitors of 3CL pro with IC 50 values of 13-57 lM. Interestingly, introduction of a fluorine substituent to this group of inhibitors bearing the p-chlorophenyl ring (11d, 12d, 13d) decreases the inhibitory activity $2fold, which stands in contrast to results with inhibitors 11a-c, 12a-c, 13a-c that are lacking this ring. This suggests that the three-ringed inhibitors (11d, 12d, 13d) may have a different binding mode from the two-ringed inhibitors (11a-c, 12a-c, 13a-c). After 2 h preincubation of 25 lM methylene ketone (11d) or fluorinated methylene ketones (12d, 13d) with 3CL pro , no improved inhibitory activity is observed, indicating that these compounds act as noncovalent fully reversible inhibitors of SARS 3CL pro . The inhibition mechanism was also examined on a longer time scale using electrospray ionization-mass spectrometry (ESI-MS) studies. After mixing 10 equivalent of inhibitor 11d, 12d or 13d with 1 equivalent of 3CL pro , and incubating the solution for 24 h, no mass change is observed in the major mass peak of 3CL pro . This further demonstrates that compounds 11a-13d utilize a non-covalent reversible mechanism of inhibition. These compounds represent a new class of SARS-CoV 3CL pro inhibitors, and to our knowledge, are among the most potent non-covalent and reversible inhibitors of this enzyme.

As furan moieties have some potential to be metabolized rapidly in mammalian cells, two additional ketone analogues 25 and 28 were prepared and examined. Synthesis of 25 started from commercially available ethyl 2-chloro-2-oxoacetate (20) , as shown in Scheme 4. Nucleophilic reaction of 20 with TMSCHN 2 provides the diazo compound 21 in 68% yield. Treating the diazo compound 21 with copper salt generates the carbene intermediate, which readily reacts with 4-chlorobenzonitrile through [3+2] cyclization to form the ester 22 with the desired oxazole moiety in 14% yield. Hydrolysis of 22 with lithium hydroxide produces the carboxylic acid 23 in 94% yield. Activation of the acid 23 with CDI, and addition of this to the pre-generated anion formed by deprotection of 9c with LiHMDS, yields the b-keto ester 24 in 23% yield. Hydrolytic decarboxylation of the b-keto ester 24 gives the desired ketone 25 in 58% yield. The ketone 28 was also prepared in a similar method as described in Scheme 5.

Compounds 25 and 28 were tested against SARS 3CL pro . Compound 25 inhibits the proteinase with an IC 50 value of 75 lM, and compound 28 displays 35% inhi- bition at a concentration of 100 lM (Table 1 ). It is known that for non-covalent and reversible inhibitors, hydrogen bonds, ionic and van der Waal's interactions play crucial roles in binding affinity to the target enzyme [22] . For inhibitors 11d-13d, the oxygen atoms of their furan rings are suspected to have hydrogen bonds with the 3CL pro [10b,23] . In contrast, that position may be occupied by the oxygen atom of the oxazole ring of 25 and the carbon atom of the isoxazole ring of 28, both of which have lower electron density and may thus give weaker binding to the 3CL pro .

To gain further insight into the inhibition mechanism, modeling studies (Fig. 2 ) of 3CL pro with inhibitors 11a-13d were conducted [10b]. Since the S2 and S4 pockets in the active site of 3CL pro are relatively large, the shorter two-ringed compounds (e.g. 11c) cannot occupy the maximal volume in these S sites in any docked conformations. However, due to their extended end-to-end length, the three-ringed compounds can occupy more of the binding surface from the S2/S4 to the S1 pocket. Therefore, the three aromatic ring compounds (11d, 12d, 13d) are more effective in blocking the binding of substrates into the active site, and thus exhibit better inhibition against 3CL pro than the two-ringed ones, as revealed in the enzymatic assay. Based on our previous modeling studies [10b,23], the three-ringed esters utilize a non-covalent and reversible mechanism of inhibition in a S4-S1 binding mode, by blocking entry of substrates into the active site of SARS-CoV 3CL pro . This is supported by a recently reported crystal structure of a three-ringed thioester with the 3CL pro [24] . Docking results suggest that these ketone analogues (11d, 12d, 13d) adopt binding conformations similar to that of the corresponding esters (e.g. 4). Each of these three compounds is oriented in an extended conformation from S4 to S1 pocket, with the oxygen atom of their furan ring forming a hydrogen bond with the main chain NH of Glu166, an interaction that was also predicted for the three-ringed esters. The pyridinyl moiety preferentially binds inside the S1 specificity pocket, which limits the possible spatial orientations of the substituents at the a-position of the central ketone group. In the 3CL pro :12dand 3CL pro :13d complexes, the fluorine substituents of the inhibitor point towards the main chain carbonyl oxygen of His164 (Fig. 2) . This pushes 12d and 13d slightly out towards the solvent. Thus, the van der Waals interactions between the fluorinated compounds and the active site residues of the enzyme are likely to be weaker than those for 11d.

A series of aryl methylene ketones and fluorinated methylene ketones have been prepared and tested as potential SARS 3CL pro inhibitors. Derivatives 11d, 12d and 13d comprising three aromatic rings, including a 5-bromopyridin-3-yl moiety, show the best inhibition (IC 50 = 13-57 lM). Interestingly, in this series (as opposed to analogues with two aryl rings), fluorination decreases inhibition despite enhancing electrophilicity of the carbonyl carbon. Enzyme kinetics and ESI-MS studies indicate that these inhibitors utilize a non-covalent, reversible mechanism, providing a basis for development of more potent specific inhibitors.

The Organic Chemistry of Enzyme-Catalyzed Reactions

Financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC) to JCV and LDE, the Canada Research Chairs Program, the Alberta Heritage Foundation for Medical Research (AHFMR) and the Killam Foundation (Fellowship to J.Y.) is gratefully acknowledged. We thank Dr. Randy Whittal and Ms. Jing Zheng for help with mass spectra. We also thank Sean Ferland for assistance with synthesis of inhibitors.