Inhibition of Norovirus 3CL Protease by Bisulfite Adducts of Transition State Inhibitors

Abstract

Noroviruses are the most common cause of acute viral gastroenteritis, accounting for >21 million cases annually in the U.S. alone. Norovirus infections constitute an important health problem for which there are no specific antiviral therapeutics or vaccines. In this study, a series of bisulfite adducts derived from representative transition state inhibitors (dipeptidyl aldehydes and α-ketoamides) was synthesized and shown to exhibit anti-norovirus activity in a cell-based replicon system. The ED₅₀ of the most effective inhibitor was 60 nM. This study demonstrates for the first time the utilization of bisulfite adducts of transition state inhibitors in the inhibition of norovirus 3C-like protease in vitro and in a cell-based replicon system. The approach described herein can be extended to the synthesis of the bisulfite adducts of other classes of transition state inhibitors of serine and cysteine proteases, such as α-ketoheterocycles and α-ketoesters.

Noroviruses are a major cause of waterborne and foodborne acute gastroenteritis.^1–4 Outbreaks of viral gastroenteritis are common because of the highly contagious nature of noroviruses. Noroviral gastroenteritis is the cause of significant morbidity and may lead to fatal infection in children, the elderly, and immuno-compromised individuals.⁵ There are currently no effective vaccines or antiviral agents for combating norovirus infection; consequently, there is an urgent need for the discovery of small molecule therapeutics for the management and treatment of norovirus infection.^6–7 Recently-reported small molecule norovirus inhibitors include cyclic and acyclic sulfamide derivatives,^8–10 piperazine derivatives,¹¹ pyranobenzopyrones,¹² nitazoxanide,¹³ and other chemotypes.¹⁴

The norovirus RNA genome encodes a polyprotein which is processed by a virus-encoded 3C-like cysteine protease (3CLpro) to generate mature non-structural proteins.¹⁵ Co- and post-translational processing of the polyprotein by norovirus 3CLpro is essential for virus replication (Figure 1); consequently, norovirus 3CLpro is an attractive target for the discovery of anti-norovirus small molecule therapeutics.

Figure 1.

Proteolytic cleavage of the nonstructural polyprotein of norovirus (Norwalk virus) encoded in open reading frame 1 (ORF1). The indicated cleavage sites at Q/G, E/A or E/G (corresponding to the P₁-P₁′ scissile bond) are mediated by the virus-encoded 3CL cysteine protease.

We have recently demonstrated that peptidyl aldehydes,¹⁶ α-ketoamides,¹⁷ and α-ketoheterocycles¹⁷ potently inhibit norovirus 3CLpro in vitro, as well as norovirus in a cell-based replicon system. In an attempt to identify suitably-functionalized dipeptidyl transition state inhibitors that possess potent pharmacological activity, as well as molecular properties that are important for oral bioavailability and favorable ADMET characteristics,^18–24 we describe herein the synthesis of bisulfite adducts of transition state inhibitors (I) (Table 1), and their subsequent utilization in the inhibition of norovirus 3CLpro in vitro, as well as viral replication in a cell-based replicon system (Figure 2). To the best of our knowledge, this is the first report on the use of bisulfite adducts of transition state inhibitors to inhibit a cysteine protease²⁵. We furthermore describe the results of preliminary structure-activity relationship studies related to the probing of the S₂ subsite²⁶ of norovirus 3CLpro, as well as the nature of the “cap” that projects toward the S₃ subsite and beyond.

Table 1.

Inhibitory activity of compounds 6, 7a–j, 9, 10

Compounda	R1	R2	R3	IC50 (μM)	ED50 (μM)
6a	benzyl	Isobutyl	CHO	0.6	0.2
7a	benzyl	Isobutyl	CH(OH) SO3Na	-b	0.3
6b	benzyl	n-propyl	CHO	6.1	2.3
7b	benzyl	n-propyl	CH(OH) SO3Na	-b	1.5
6c	benzyl	n-butyl	CHO	4.5	1.2
7c	benzyl	n-butyl	CH(OH) SO3 Na	-b	1.3
6d	benzyl	cyclohexylmethyl	CHO	0.5	0.05
7d	benzyl	cyclohexylmethyl	CH(OH) SO3 Na	-b	0.06
6e	benzyl	benzyl	CHO	5.1	1.8
7e	benzyl	benzyl	CH(OH) SO3Na	-b	1.1
6f	p-fluorobenzyl	Isobutyl	CHO	1.8	0.5
7f	p-fluorobenzyl	Isobutyl	CH(OH) SO3 Na	-b	0.4
6g	m-fluorobenzyl	Isobutyl	CHO	0.7	0.3
7g	m-fluorobenzyl	Isobutyl	CH(OH) SO3Na	-b	0.2
6h	2-phenylethyl	Isobutyl	CHO	1.9	1.1
7h	2-phenylethyl	Isobutyl	CH(OH) SO3Na	-b	1.0
6j	2-cyclohexylethyl	Isobutyl	CHO	0.6	0.3
7j	2-cyclohexylethyl	Isobutyl	CH(OH) SO3 Na	-b	0.3
917				3.4	1.1
10				5.3	0.8

^aCC₅₀: All compounds, except 6j and 7j, showed no toxicity up to 320 μM (CC₅₀: > 320μM). The CC₅₀ values for 6j and 7j were 210 μM, and 240 μM, respectively.

^bNot determined (see text).

Figure 2.

General representation of the interaction and interrelationships between 3CL protease (E-Cys-SH), transition state inhibitor, X(C=O)Z, and bisulfite adduct, XZ(OH)SO₃Na.

The synthesis of dipeptidyl inhibitors 6a–j, 7a–j, 9, and 10 is summarized in Scheme 1.

Scheme 1. Reagents.

(a) CCI₃0(C=0)CI/dioxane; (b)Triethylamine/R¹OH; (c) Li0H/THF/H₂0; (cl) EDCI/HOBt/DIEA/DMF; (e) LiBH₄/THF; (f) Dess-Martin periodinane/DCM; (g) NaHS03/EtOAc/EtOH/H₂0; (h) Cyclopropyl isonitrile/HOActhen K2CO3/CH3OH/H2O.

Reaction of an appropriate amino acid ester hydrochloride with trichloromethyl chloroformate yielded the corresponding isocyanate which was subsequently reacted with an appropriate alcohol in the presence of triethylamine to yield carbamate derivative 2. Hydrolysis with lithium hydroxide in aqueous THF followed by coupling with a glutamine surrogate²⁷ yielded ester 4 which was further elaborated to yield aldehydes 6a–j via sequential reduction to the alcohol with lithium borohydride, followed by Dess-Martin oxidation.²⁸ The reaction of aldehyde 6 (R¹ = benzyl, R² = isobutyl) with cyclopropyl isonitrile/HOAc followed by treatment with potassium carbonate in aqueous methanol yielded alcohol 8 which was then oxidized to the corresponding α-ketoamide 9 using Dess-Martin reagent. The generated aldehyde and α-ketoamide bisulfite adducts were readily obtained by stirring aldehydes 6a–j and α-ketoamide 9 with sodium bisulfite.²⁹ The interaction of the generated compounds with norovirus 3CLpro was investigated in vitro as previously described.^16–17 The activity of the compounds against norovirus was also investigated in a cell-based system^30–34 and the combined results are listed in Table 1.

The rationale underlying the studies described herein rested on the following considerations: (a) bisulfite adducts of amino acid-derived isocyanates are readily-accessible, stable, water-soluble solids which function as latent isocyanates. These adducts have been shown to be highly effective, time-dependent, irreversible inhibitors of mammalian serine proteases, such as neutrophil elastase, cathepsin G, and proteinase 3;³⁵ (b) bisulfite adducts of aldehydes, methyl or cyclic ketones, and α-ketoesters are readily-synthesized, stable solids having high aqueous solubility. Treatment of the addition products with acid or base yields the precursor carbonyl compounds;³⁶ (c) we hypothesized that the bisulfite adducts of transition state (TS) inhibitors of proteases (serine and cysteine), such as peptidyl aldehydes, α-ketoamides, and others could potentially function as a latent form of the precursor TS inhibitor (Figure 2), generating the active form of the inhibitor in the gastrointestinal tract and blood plasma. In principle, the bisulfite adducts could also function as transition state mimics³⁷ and, (d) the high aqueous solubility and pH-dependent equilibria between the precursor carbonyl compound and adduct were also envisaged to have a significant effect on potency and the ADMET and PK characteristics of the precursor TS inhibitors. It was envisioned that the bisulfite adducts might be suitable candidates for fulfilling such a role.

As shown in Table 1, the dipeptidyl aldehydes exhibited low to sub-micromolar inhibitory activity toward NV 3CLpro in vitro. The enzyme shows a strong preference for an R² = isobutyl, which is in agreement with the known substrate specificity of the enzyme. The strong preference of NV 3CLpro for a P2 Leu is supported by substrate specificity studies using peptidyl p-nitroanilide substrates, as well as X-ray crystallographic studies.³⁸ The results in Table 1 suggest that replacement of the isobutyl group by a cyclohexylmethyl group at R² yields an inhibitor that is equipotent to 6a (Table 1, compounds 6a and 6d). However, in sharp contrast to compound 6a, the ED₅₀ of compound 6d was found to be an order of magnitude lower than that of 6a, presumably because of its better cellular permeability. The nature of the “cap” (R¹) which projects toward the S₃ pocket and beyond was briefly explored by replacing the benzyl group with meta- and para-fluorobenzyl, 2-phenylethyl, and 2-cyclohexylethyl. The m-fluorobenzyl and 2-cyclohexylethyl groups were equipotent to the benzyl group, and about 2-fold better than other substitutions (Table 1, compounds 6a, 6g, 6j versus 6f and 6j).

The activity of the bisulfite adducts of the synthesized aldehydes was investigated in a cell-based system. The potency trends observed with the precursor aldehydes were generally reflected in the corresponding bisulfite salts, with bisulfite 7d (R² = cyclohexylmethyl) being the most potent. In order to determine the nature of the active species, the behavior of aldehyde 6a and its corresponding bisulfite salt 7a was examined by mass spectroscopy. In separate experiments, compounds 6a and 7a were dissolved in dimethyl sulfoxide and diluted 1 to 1000 in either acetonitrile or water and examined by MS and tandem MS-MS. In acetonitrile the expected peaks for aldehyde 6a were 404.4 M + H⁺ (dominant peak) and 426.3 M+Na. The mass spectra of bisulfite salt 7a using negative mode detection, showed a dominant peak at 484.5 for (M-1)⁻, a loss of H⁺ from the sulfonic acid moiety. Aldehyde 6a in aqueous solution showed peaks corresponding to the aldehyde (404.6), the aldehyde + sodium (426.4) and hydrated aldehyde + sodium (444.2) in positive mode. In water, bisulfite adduct 7a displayed a dominant peak at 484.5 in negative mode and the relative intensities of this parent ion and other ions remained unchanged over 24 h (a time course study was carried out). In the case of 6a, the hydrated form was the dominant species after only five minutes exposure to water, while 7a remains unchanged as the bisulfite form after 24 h. The results indicate that the bisulfite adduct of 7a is stable in aqueous solution; however, in buffer solution, pH 7.4, compound 7a gradually dissociates into the corresponding aldehyde 6a within an hour, rapidly becoming hydrated. These observations are in agreement with the results of X-ray crystallographic studies showing that incubation of bisulfite adduct 7a with norovirus 3CLpro in buffer solution results in the formation of an enzyme-aldehyde complex, with the active site cysteine residue covalently bonded to the carbonyl carbon of aldehyde 6a.³⁴ Lastly, the variable stability of the bisulfite values of the adducts in buffer solution has precluded accurate determination of the IC₅₀ inhibitors.

In summary, the utilization of bisulfite adducts of transition state inhibitors of cysteine and serine proteases in the in vitro and cell-based inhibition of norovirus 3CL protease has been described for the first time.