Yeast based small molecule screen for inhibitors of SARS-CoV

Abstract

Severe acute respiratory coronavirus (SARS-CoV) emerged in 2002, resulting in roughly 8000 cases worldwide and 10% mortality. The animal reservoirs for SARS-CoV precursors still exist and the likelihood of future outbreaks in the human population is high. The SARS-CoV papain-like protease (PLP) is an attractive target for pharmaceutical development because it is essential for virus replication and is conserved among human coronaviruses. A yeast-based assay was established for PLP activity that relies on the ability of PLP to induce a pronounced slow-growth phenotype when expressed in S. cerevisiae. Induction of the slow-growth phenotype was shown to take place over a 60-hour time course, providing the basis for conducting a screen for small molecules that restore growth by inhibiting the function of PLP. Five chemical suppressors of the slow-growth phenotype were identified from the 2000 member NIH Diversity Set library. One of these, NSC158362, potently inhibited SARS-CoV replication in cell culture without toxic effects on cells, and it specifically inhibited SARS-CoV replication but not influenza virus replication. The effect of NSC158362 on PLP protease, deubiquitinase and anti-interferon activities was investigated but the compound did not alter these activities. Another suppressor, NSC158011, demonstrated the ability to inhibit PLP protease activity in a cell-based assay. The identification of these inhibitors demonstrated a strong functional connection between the PLP-based yeast assay, the inhibitory compounds, and SARS-CoV biology. Furthermore the data with NSC158362 suggest a novel mechanism for inhibition of SARS-CoV replication that may involve an unknown activity of PLP, or alternatively a direct effect on a cellular target that modifies or bypasses PLP function in yeast and mammalian cells.

Figure 1. SARS-CoV PLP produces a slow growth phenotype in yeast.

A. Galactose induction of SARS-CoV PLP. Strain containing HA tagged PLP under the control of the yeast GAL1 promoter was grown in the presence of 0 to 2% galactose. Protein was extracted and analyzed by anti-HA western blot. B. Growth curve of yeast expressing either empty vector or HA tagged PLP grown in 2% galactose media.

Figure 2. Compounds that reverse the slow growth phenotype.

Yeast grown in media containing 2% galactose with the addition of either 1% DMSO or 50 uM compounds dissolved in 1% DMSO. B. Structures of compounds shown in A. C. Effects of compounds on PLP expression. Western blots were performed with protein extracted from HA tagged PLP expressing yeast grown in the presence of 2% galactose and 50 uM of each compound and visualized with anti-HA antibody.

Figure 3. Toxicity assays.

VeroE6 (A) or 293T (B) cells were treated with 0, 10 uM or 100 uM of each compound in 1% DMSO for 24 hours. Cells were analyzed for viability with the CellTiter Glo assay (Promega).

Figure 4. Effects of compounds on virus growth and RNA production.

VeroE6 (A) or MA104 (B) cells were treated with 50 uM of each compound or 1% DMSO alone and infected with SARS-CoV(GFP) at an MOI of 3. Virus titer was assayed by plaque assay on VeroE6 cells after 12 and 24 hours of growth. C. Fluorescence images of SARS-CoV(GFP) infected Vero and MA104 cells at 24 hours post infection. D. Dose curve of NSC158362 on SARS-CoV growth. Various concentrations of drug were added to Vero cells 2 hours before SARS-CoV was added at an MOI of 3. Aliquots were removed at 12, 18 and 24 hours post infection and titered on Vero cells. E. Vero cells were treated with either DMSO or NSC 158263 at -2 hours, 0 hours, + 2 hours or +6 hours after infection with SARS-CoV. RNA was isolated at 12 hours post infection and analyzed by RT-PCR for SARS-CoV specific transcripts and GAPDH. F. MDCK cells were treated with 50 uM of each compound or 1% DMSO alone and infected with influenza A/PR/8 at an MOI of 0.1. Virus titer was determined by hemagglutination assay.

Figure 5. Inhibition of SARS-CoV replication in human airway epithelial cells by NSC158362.

HAE cells were treated with either 1% DMSO or NSC158362 at 50 uM and infected with SARS-CoV(GFP). The apical surface of each culture was rinsed with PBS at 24, 48 and 72 hr post infection and virus titered on VeroE6 cells.

Figure 6. Effect of compounds on PLP enzymatic function.

A. PLP protease activity was assayed using a nsp2/3/GFP reporter plasmid. 293T cells were transfected with either nsp2/3/GFP alone or with HA tagged PLP. Cells were treated with 50 uM of each compound or 1% DMSO alone and protease activity was assayed by reduction in size of the nsp2/3/GFP fusion protein by western blot with anti-GFP antibody. B. PLP deubiquitinase activity was assayed in cells transfected with Flag tagged ubiquitin and HA tagged PLP. 293T cells were transfected with both plasmids. Cells were treated with 50 uM of each compound or 1% DMSO alone and deubiquitinase activity was assayed by reduction in ubiquitinated protein by western blot with anti-Flag antibody. C and D. Effects of the compounds on PLP's IFN antagonism ability were analyzed by poly IC treatment of RIGI transfection of cells with and IFNβ/luciferase reporter plasmid with and without PLP and the compounds. Western blot of transfected HA/PLP shown below each graph. E. Effect of 158362 on IFN induction after infection with SARS-CoV. Vero cells were transfected with IFNβ/luciferase reporter plasmid. RIG-I was transfected in 1 set of wells as a positive control. Cells were then treated with either DMSO or 158362 for 2 hours prior to infection with SARS-CoV at an MOI of 3. No increase in IFNβ induction was seen after infection.