Disulfiram can inhibit MERS and SARS coronavirus papain-like proteases via different modes

Abstract

Severe acute respiratory syndrome coronavirus (SARS-CoV) emerged in southern China in late 2002 and caused a global outbreak with a fatality rate around 10% in 2003. Ten years later, a second highly pathogenic human CoV, MERS-CoV, emerged in the Middle East and has spread to other countries in Europe, North Africa, North America and Asia. As of November 2017, MERS-CoV had infected at least 2102 people with a fatality rate of about 35% globally, and hence there is an urgent need to identify antiviral drugs that are active against MERS-CoV. Here we show that a clinically available alcohol-aversive drug, disulfiram, can inhibit the papain-like proteases (PL^pros) of MERS-CoV and SARS-CoV. Our findings suggest that disulfiram acts as an allosteric inhibitor of MERS-CoV PL^pro but as a competitive (or mixed) inhibitor of SARS-CoV PL^pro. The phenomenon of slow-binding inhibition and the irrecoverability of enzyme activity after removing unbound disulfiram indicate covalent inactivation of SARS-CoV PL^pro by disulfiram, while synergistic inhibition of MERS-CoV PL^pro by disulfiram and 6-thioguanine or mycophenolic acid implies the potential for combination treatments using these three clinically available drugs.

Keywords: 6-Thioguanine; Disulfiram; MERS- and SARS-CoV; Mycophenolic acid; Papain-like protease; Synergistic inhibition.

Fig. 1

Inhibitory effects of disulfiram on coronaviral PL^pros. DUB activity of MERS-CoV (A) and SARS-CoV (B) PL^pro in the presence of disulfiram (6–50 μM) was measured. The concentration of fluorogenic substrate (Ub-AFC) was 0.25 μM, while the concentration of coronaviral PL^pro was 0.2 μM in both cases. The lines show best-fit results in accordance with the IC₅₀ equation (Eq. (1)).

Fig. 2

Inhibition of coronaviral PL^pros by disulfiram. The proteolytic activity of MERS-CoV (A) and SARS-CoV (B) PL^pro were measured in the presence of different peptide substrate concentrations (9–80 μM) and various concentrations of disulfiram (6–50 μM). The solid lines are best-fit results in accordance with noncompetitive (A) or competitive (B) inhibition models. The R_sqr values are 0.989 and 0.977, respectively. The experiments were repeated to ensure reproducibility. Kinetic parameters such as K_M, k_cat and K_is from the best-fit results are shown in Table 1.

Fig. 3

Mutual effects of coronaviral PL^proinhibitors. The activity of MERS-CoV PL^pro was measured without and with either 6TG (A) or MPA (B) in the presence of various concentrations of disulfiram, and that of SARS-CoV PL^pro was measured without and with either 6TG (C) or NEM (D) in the presence of various concentrations of disulfiram. The concentrations of peptidyl substrate and MERS-CoV PL^pro (A and B) were 20 and 0.6 μM, respectively, while those of peptidyl substrate and SARS-CoV PL^pro (C and D) were 15 and 0.05 μM, respectively. The points are the reciprocals of the initial velocities and the lines are the best fit of the data to Eq. (5). The results suggest that the α values for the four experiments (A–D) are 0.1, 0.17, 18.2 and 109.3, respectively.

Fig. 4

Effect of zinc ion ejection by disulfiram and its influence on PL^prostability. (A) MERS- and SARS-CoV PL^pro each was incubated without and with 5 μM disulfiram. The release of zinc ions from the enzyme was detected as the increase of the fluorescence signal of the zinc-specific fluorophore FluoZin-3. (B) and (C) Thermostability of MERS-CoV PL^pro, SARS-CoV PL^pro or SARS-CoV PL^pro C271A mutant in the absence or presence of 5 μM disulfiram was detected by circular dichroism spectrometry. The protein concentration was 0.2 mg/ml. The wavelength used was 222 nm and the cuvette pathlength was 1 mm. The right and left dotted lines show the melting temperature of SARS-CoV PL^pro without and with disulfiram, respectively. These results indicate that disulfiram destabilized the enzyme.

Fig. 5

Slow-binding inhibition of SARS-CoV PL^proby disulfiram. (A) DUB activity of disulfiram-treated MERS- and SARS-CoV PL^pro in the absence or presence of 5 mM β-ME. The enzyme was incubated without or with 200 μM disulfiram for 1 h and the mixture was then desalted using a Sephadex G-25 column. The concentrations of fluorogenic substrate (Ub-AFC) and enzyme were 0.25 and 0.2 μM, respectively. (B) 0.05 μM SARS-CoV PL^pro was incubated with different concentrations of disulfiram (0 μM, closed circles; 2–12 μM, open circles), after which its proteolytic activity was measured for 5 min using 15 μM peptidyl substrate. The solid lines are best-fit results in accordance with the slow-binding equation (Eq. (6)). (C) The observed inactivation rate constants (k_inact) from panel B were replotted against disulfiram concentration. The solid line is the best-fit result in accordance with the saturation equation (Eq. (7)). Kinetic parameters K_inact and k_max corresponding to the best-fit curve are shown in Table 1.

Fig. 6

Binding of disulfiram to SARS-CoV PL^pro. Overlay of model structure of SARS-CoV PL^pro in complex with DDC (magenta) (A) or disulfiram (orange) (B) with the crystal structure of SARS-CoV PL^pro in complex with ubiquitin (gray, PDB code: 4M0W). DDC and disulfiram are modeled based on the binding sites of βME and glycerol, respectively. The red dashed lines show putative polar interactions while the black dashed line shows the distance between residue Cys271 and disulfiram as 4.0 Å. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 7

Schemes of proposed kinetic mechanisms for the inhibition of SARS-CoV and MERS-CoV PL^proby disulfiram. The upper diagram denotes enzyme catalysis, mixed inhibition and inactivation of SARS-CoV PL^pro by disulfiram. The lower diagram shows noncompetitive inhibition of MERS-CoV PL^pro by disulfiram and triple inhibition with two other FDA-approved drugs, 6TG and MPA. SH symbolizes the thiolate of catalytic triad residue Cys.