Novel Inhibitors of Severe Acute Respiratory Syndrome Coronavirus Entry That Act by Three Distinct Mechanisms

Cell lines and plasmids.

pCAGGS SARS-CoV Spike, pNL-4.3-luc, and pCDNA 3.1-ACE2 were provided by Paul Bates (University of Pennsylvania) (43). The pCDNA 3.1 soluble ACE2 plasmid was provided by S. Pohlmann (Institute of Virology, Hannover, Germany). A plasmid encoding the SARS-CoV spike RBD with the signal sequence of CD5 and the Fc domain of human IgG1, pCDM8-Spike RBD-Fc, was provided by Michael Farzan (New England Primate Research Center) (29, 44). pMDG-VSV-G was provided by Marc Johnson (University of Missouri, Columbia, MO) (45). pSV2tat72 was obtained through the AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from Alan Frankel (46). TZM-bl cells were obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH, from John C. Kappes, Xiaoyun Wu, and Tranzyme Inc. (15, 47–50). 293T cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 10 μg streptomycin (43).

Infection assay with pseudotyped virus.

293T cells were cotransfected with pCAGGS SARS-S DNA or pMDG-VSV-G DNA together with pNL-4.3-Luc-E⁻R⁻ by using calcium phosphate transfection reagents (43). After 48 h of incubation at 37°C and 5% CO₂, the supernatants, containing virions pseudotyped with SARS-S or the vesicular stomatitis virus glycoprotein (VSV-G), were harvested and stored at −80°C. For infection experiments, 293T cells transiently transfected with pCDNA 3.1-ACE2 were seeded into 96-well plates at a density of 1 × 10⁴ cells/well. After 15 min, equal volumes of the viral supernatant were added to the wells with the cells, and the mixtures were incubated for another 48 h. Bright-Glo luciferase reagent (Promega, Madison, WI) was added to the cells, and the luciferase activity was determined using a Veritas microplate luminometer (Turner Veritas Biosystems). To determine the amount of pseudotyped virus to be used for screening and identification of SARS-CoV entry inhibitors, we titrated the viral supernatant (1 to 50 μl) and the ACE2-expressing 293T cells, and the amount of viral supernatant that gave ∼25,000 relative luciferase units was chosen for use in infection assays. The multiplicity of infection (MOI) in the pseudotype-based assays was estimated to be 10 U/cell by using an enzyme-linked immunosorbent assay (ELISA) for the detection of HIV p24. Experiments performed below this MOI gave inconsistent results with the pseudotype-based assay. To determine the MOI, six 5-fold serial dilutions of 50 μl of the pseudotyped viral stock were used to infect 1 × 10⁴ ACE2/293T cells in a 96-well plate in quadruplicate, with noninfected ACE2/293T cells as a control. After 48 h, the supernatant in each well was replaced with a lysis buffer (137 mM NaCl, 1.94 mM K₃PO₄, 8.06 mM Na₃PO₄, and 2.7 mM KCl, pH 7.4, or phosphate-buffered saline [PBS] and 10% Triton X-100), and the p24 antigen in each sample was subsequently measured using an ELISA kit (Organon Teknika). The 50% tissue culture infective dose (TCID₅₀) was calculated according to the Spearman-Karber formula, as follows: 1/TCID₅₀ = B^e, where e is the exponent for the reciprocal titer and B is the fold dilution used in the dilution series (http://www.europrise.org/documents/NEUTNET/SOPS/11_NHRBC_PBMC.pdf). After determination of the TCID₅₀ of the viral stock (TCID₅₀/ml), the TCID₅₀ titer was then converted to the estimated number of infectious units per volume of virus material (U/ml) (similar to PFU/ml in a plaque assay) by multiplying the titer by 0.7 (51). To obtain the MOI in U/cell, the number of infectious particles was divided by the number of cells to be infected. For the purpose of screening to identify inhibitors of SARS-CoV entry, the compounds were incubated with ACE2-expressing 293T cells for 45 min, followed by addition of the appropriate amount of viral supernatant containing 100 TCID₅₀ (MOI of 10 U/cell). The cells were further incubated for 48 h, followed by measurement of the luciferase activity using a Veritas microplate luminometer (Turner Veritas Biosystems).

Effects of inhibitors on cathepsin L and cathepsin B activity.

Purified recombinant cathepsin L (2 units) was incubated at 37°C with a 25 μM concentration of the fluorogenic substrate Z-Phe-Arg-7-amido-4-methylcoumarin and assay buffer (Calbiochem) in the presence or absence of various concentrations of the compounds (2 to 80 μM), in a total reaction volume of 200 μl. The amount of fluorescence produced as a result of the release of amido-4-methylcoumarin (AMC) from the substrate was measured using a fluorimeter at an excitation wavelength of 380 nm and an emission wavelength of 460 nm. The data obtained were analyzed using GraphPad Prism 5.0 software (GraphPad, Inc.). To monitor cathepsin B activity, purified recombinant cathepsin B (2 units) was incubated at 37°C with a 25 μM concentration of the fluorogenic substrate Z-Arg-Arg-7-amido-4-methylcoumarin and assay buffer (Calbiochem) in the presence or absence of CA074 (cathepsin B inhibitor) and SSAA09E1, in a total reaction volume of 200 μl. The amount of fluorescence produced as a result of the release of AMC from the substrate was measured using a fluorimeter with an excitation wavelength of 380 nm and an emission wavelength of 460 nm. The data obtained were analyzed using GraphPad Prism 5.0 software (GraphPad, Inc.).

Expression and purification of SARS-CoV spike receptor binding domain.

Expression and purification of the SARS-CoV spike receptor binding domain were carried out as described previously (29). Briefly, 293T cells were transfected with pCDM8-Spike RBD-Fc, a plasmid encoding the SARS-S RBD with the signal sequence of CD5 and the Fc domain of human IgG1. Twenty-four hours after transfection, cells were washed in PBS and then incubated in 293 serum-free medium II (Invitrogen). After 48 h, the medium was harvested and the protein was precipitated using protein A Sepharose beads at 4°C for 16 h. Beads were then washed in PBS with 0.5 M NaCl, and the bound protein was eluted with 50 mM sodium citrate-50 mM glycine, pH 2.0, and neutralized to pH 7.0 with 5 M sodium hydroxide. Purified proteins were concentrated with Millipore Amicon Ultra centrifugal filter units (nominal molecular weight limit [NMWL], 10,000) and then dialyzed against PBS.

Immunoprecipitation and Western blotting.

Soluble ACE2 protein was obtained from the supernatant of transiently transfected 293T cells, followed by dialysis against PBS and protein concentration using a Centricon Plus Ultrafilter with a 10-kDa molecular mass cutoff (Millipore, St. Louis, MO) (26). Soluble ACE2 protein and the purified receptor binding domain of the spike protein (RBD-Fc) were incubated in the presence of various concentrations of the inhibitors (0, 5, 10, and 20 μM) for 2 h at 4°C, followed by the addition of protein A Sepharose slurry to the mixture. The samples were incubated for an additional hour and then separated by 10% PAGE, followed by transfer to nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany). ACE2 protein and SARS-S RBD-Fc were detected using a goat anti-human polyclonal ACE2 antibody (LifeSpan Biosciences, Inc.) with rabbit anti-goat–horseradish peroxidase (HRP) as the secondary antibody (Millipore, St. Louis, MO) and anti-human IgG–HRP (Millipore, St. Louis, MO), respectively.

Binding inhibition and flow cytometry.

Binding inhibition and flow cytometry were performed as described previously (29). Approximately 1 × 10⁶ ACE2-transfected 293T cells were detached in PBS with 5 mM EDTA and washed with PBS with 0.5% bovine serum albumin (BSA). Various concentrations of the compounds (0, 5, 10, and 20 μM) were added to the cells and then allowed to incubate on ice for 30 min, followed by addition of the purified RBD-Fc, and then the mixture was incubated on ice for 1 h. The cells were washed 3 times with PBS and 0.5% BSA and then incubated with anti-human IgG–fluorescein isothiocyanate (FITC) conjugate (Sigma). Cells were washed again 3 times with PBS and 0.5% BSA. The cells were then fixed with 4% paraformaldehyde and analyzed by flow cytometry at the University of Missouri Cell and Immunobiology Core Facility to determine the percentage of cells bound to the spike receptor binding domain in the presence and absence of the inhibitors.

Quantitative cell-to-cell fusion assay.

TZM-bl cells that stably express luciferase under the control of the HIV-1 promoter when activated by Tat protein (15, 47–50) were transfected with pcDNA 3.1-ACE2 plasmid. Simultaneously, 293T cells were cotransfected with plasmids encoding wild-type SARS-S and Tat cDNA (46). All transfections were done with the Lipofectamine 2000 (Invitrogen) transfection reagent. Forty-eight hours after transfection, the TZM-bl cells were seeded into 96-well plates at a density of 1 × 10⁴ cells/well. The ACE2-expressing TZM-bl cells were then overlaid with the same amount of 293T cells expressing SARS-S envelope and Tat protein (1 × 10⁴ cells/well). Dimethyl sulfoxide (DMSO) or compounds were added to the TZM-bl cells either 45 min prior to the overlay with 293T cells or 2 h after the overlay with 293T cells. Three hours after the overlay, cells were treated with serum-free medium containing 2 μg/ml trypsin to induce fusion and replaced with complete DMEM after 30 min. Six hours after fusion induction, the cells were monitored or analyzed for the expression of luciferase by using the Bright-Glo luciferase reagent and a Veritas microplate luminometer.

SARS-CoV cytopathic effect assay.

The assay for the identification of potential inhibitors of SARS-CoV entry was performed essentially as described previously (52). Briefly, Vero E6 cells were dispensed into black, clear-bottom, 96-well plates at a density of 10,000 cells/well in 50 μl DMEM with phenol red, supplemented with 10% FBS and 1% l-glutamine, by use of a WellMate microplate dispenser (Matrix, Hudson, NH) and then were incubated for 24 h at 37°C and 5% CO₂, with high humidity. For dose-response and compound 50% effective concentration (EC₅₀) determinations, 25-μl aliquots of various concentrations of the compounds (0.052 μM to 120 μM) were added for final plate well concentrations ranging from 0.013 μM to 30 μM and a final DMSO concentration of 0.5%. Eight concentrations of each compound were added to 96-well plates in triplicate to measure the effective concentration at which each compound inhibited viral cytopathic effect (CPE) by 50% (EC₅₀) or in duplicate, in the absence of virus, to measure the inhibitory concentration at which growth was inhibited by 50% (50% cytotoxic concentration [CC₅₀]). The plates were transported to a biosafety level 3 (BSL3) facility (Southern Research Institute, AL), where they were infected with 25 μl of diluted Toronto-2 strain of SARS-CoV (a gift of Heinz Feldman) at a concentration of 100 TCID₅₀ by use of a WellMate microplate dispenser (MOI of 0.1 PFU/cell). Internal controls consisted of wells containing cells only (cell control), cells infected with virus (virus control), and virus-infected cells treated with calpain inhibitor IV. Plates were then allowed to incubate at 37°C and 5% CO₂ for 72 h. After incubation, 100 μl of Promega Cell Titer Glo (Promega, Madison, WI) was added to each well by use of a microplate dispenser. Plates were shaken for 2 min on a Lab-Line plate shaker. Luminescence was then measured using a PerkinElmer Envision plate reader.

The percent CPE inhibition was defined as follows: % CPE inhibition = [(test compound − virus control)/(cell control − virus control)] × 100, with “test compound” defined as the amount of luminescence obtained from wells with the compound, cells, and virus; “virus control” is defined as the amount of luminescence obtained from wells with the virus and cells only, and “cell control” is defined as the amount of luminescence obtained from wells with cells only. Percent cell viability was defined as follows: % cell viability = (test compound/cell control) × 100. An active compound, or “hit,” was defined as a compound that exhibited a % CPE inhibition of >50% without compromising cell viability. Eight concentrations of each compound were added to 96-well plates in triplicate to measure the EC₅₀ or in duplicate in the absence of virus to measure the CC₅₀ of the compound alone.

The Z factor values were calculated as follows: Z = [1 − (3σ_c + 3σ_v)/(μ_c − μ_v)], where σ_c is the standard deviation of the cell control, σ_v is the standard deviation of the virus control, μ_c is the mean cell control signal, and μ_v is the mean virus control signal (53).

Cytotoxicity studies on 293T cells were also performed by assessing the effects of the inhibitors on cellular viability, using a commercially available XTT cytotoxicity assay kit (Roche Diagnostics, Indianapolis, IN) that measures metabolism of XTT {2,3-bis(2-methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino) carbonyl]-2H-tetrazolium hydroxide)}. This assay was conducted as previously described (54), and the results were in agreement with those obtained for Vero cells by cytotoxicity tests using Promega Cell Titer Glo (Promega, Madison, WI). The latter kit quantitates the amount of ATP present, which signals the presence of metabolically active cells.

SARS-CoV replicon assay with RNA detection by RT-qPCR.

The SARS-CoV replicon and mutants were generated as previously described (41, 55). Briefly, 293T cells were grown to 95% confluence on 35-mm-diameter plates and transfected with 4 μg of SARS-CoV replicon, a SARS-CoV nonreplicative construct (NRC) (Rep1b deletion mutant), or mock plasmid by using Lipofectamine reagent (Invitrogen) as directed by the manufacturer. Compounds (20 μM) were added to the replicon-transfected cells and NRC-transfected cells. At 48 h posttransfection (hpt), the total intracellular RNA was extracted using TRIzol (Invitrogen), followed by treatment with DNase I to digest remaining DNA. The extracted RNA was used as a template for subsequent reverse transcription–quantitative real-time PCR (RT-qPCR) analysis of N gene mRNA synthesis (NC). The reverse primer URB-28630RS (5′-TGCTTCCCTCTGCGTAGAAGCC-3′), complementary to nucleotides 511 to 532 of the N gene, and the forward primer URB-29VS (5′-GCCAACCAACCTCGATCTCTTG-3′), containing nucleotides 29 to 50 of the Urbani leader sequence, were used for amplification using a SuperScript One-Step RT-qPCR system with Platinum Taq DNA polymerase (Invitrogen) as suggested by the manufacturer. The SuperScript system is a real-time qPCR system that uses Sybr green for detection and quantitation of amplified DNA. The sequences of the forward and reverse primers used for the amplification of U6 mRNA as an endogenous control were as follows: U6 forward primer, 5′-CTCGCTTCGGCAGCACA-3′; and U6 reverse primer, 5′-AACGCTTCACGAATTTGCGT-3′. Primer pair amplification efficiencies were determined using 1:10 cDNA dilutions; test and housekeeping gene primer pairs with similar efficiencies were used for the qPCRs. Samples were normalized internally using the cycle threshold (C_T) of the U6 housekeeping gene, as follows: ΔC_T = (C_T NC) − (C_T U6). This was followed by determination of the mean for each sample, since the reactions were performed in triplicate. The mean value for each sample was normalized to the mean value for the NRC cells by using the following equation: ΔΔC_T = ΔC_T(sample) − ΔC_T(NRC). The relative quantity (RQ) values were calculated as follows: RQ = (2^−ΔΔCT). The RQ value for each sample was then normalized to the RQ value for the NRC (which is 1) in order to obtain percent relative RQ values. The data were plotted as percentages of relative replicon activity against inhibitor concentrations, in μM, using GraphPad Prism 5.0 (GraphPad Inc.). Data presented represent assays performed in triplicate in 3 independent experiments.

Primary screening for SARS-CoV entry inhibitors.

To identify compounds that can inhibit SARS-CoV entry into susceptible cells, screening of a library of pharmacologically active small molecules was carried out using the SARS/HIV-luc pseudotyped virus infection assay. Compounds that reduced luciferase activity by 50% or less were selected as potential leads. Approximately 3,000 compounds were screened from the Maybridge Hitfinder chemical library, and 44 compounds were found to cause ≤50% reduction in luciferase activity at 10 μM.

Validation of initial hits with VSV/HIV-luc pseudotyped virus.

To exclude compounds that reduced luciferase activity due to cytotoxicity, luciferase inhibition, or HIV replication inhibition from the SARS/HIV pseudotype-based assay, we tested these compounds with a VSV/HIV pseudotyped virus infection assay. Of the 44 compounds tested, we selected for further characterization 3 compounds that did not reduce luminescence below 80% in the VSV/HIV pseudotyped virus infection assay, indicating that these 3 compounds are specific inhibitors of SARS-CoV entry (Fig. 1 and 2). The compounds are (i) SSAA09E1 {[(Z)-1-thiophen-2-ylethylideneamino]thiourea}, (ii) SSAA09E2 {N-[[4-(4-methylpiperazin-1-yl)phenyl]methyl]-1,2-oxazole-5-carboxamide}, and (iii) SSAA09E3 [N-(9,10-dioxo-9,10-dihydroanthracen-2-yl)benzamide]. The structures of these compounds are shown in Fig. 3. Additional compounds that reduced luciferase activity by 20% to 50% in the VSV/HIV infection assay were considered to be less specific SARS/HIV pseudotyped virus inhibitors and are the subject of a separate investigation.

Dose dependence assay.

To further validate the efficiencies of these compounds for SARS/HIV pseudotyped virus entry inhibition, we performed dose-response experiments with these 3 compounds, using SARS/HIV and VSV/HIV pseudotype-based infection assays. The results indicated that these 3 compounds decreased luciferase activity with increasing concentrations of the compounds (Fig. 2). The calculated EC₅₀s are shown in Table 1. It should be noted that inhibition was significantly less efficient (higher EC₅₀ values) at higher MOIs (more virus).

Inhibition of cathepsin L activity.

To understand the mechanism by which each of these compounds inhibits SARS-CoV entry, we tested their effects on cathepsin L proteolytic activity, since this enzyme has been implicated in the cleavage of SARS-S to render the S2 domain of S accessible for cell fusion during SARS-CoV entry (27). In this experiment, we did a dose dependence fluorescence inhibition assay of cathepsin L activity, and the results indicated that only SSAA09E1 inhibited cathepsin L activity, with a 50% inhibitory concentration (IC₅₀) of 5.33 ± 0.61 μM (Fig. 4A), which is similar to the potency with which the compound blocks SARS/HIV entry. However, several studies have implicated a related enzyme, cathepsin B, in the entry of some coronaviruses, such as feline enteric coronavirus (56) and, to a lesser extent, SARS-CoV (27, 57). Notably, cathepsin B has a similar but distinct cleavage specificity from that of cathepsin L. Hence, to determine the specificity of SSAA09E1 as a cathepsin L activity inhibitor, we performed cathepsin B activity assays in the presence and absence of SSAA09E1 as described in Materials and Methods. A known specific cathepsin B inhibitor, CA074, was used as a positive control for this assay. The results show that although SSAA09E1 is the only one of the three compounds that can inhibit cathepsin L activity, it does not block cathepsin B activity, further suggesting that SSAA09E1 may inhibit SARS-CoV entry by specifically suppressing cathepsin L activity (Fig. 4B).

Effects of compounds on interactions of SARS-S RBD with soluble ACE2.

To further investigate the inhibition mechanism of SARS-CoV entry by these compounds, we used an immunoprecipitation and immunoblot assay, as described in Materials and Methods, to examine their effects on the interaction of the SARS-S RBD with the soluble ACE2 receptor. The results indicate that only SSAA09E2 interferes with the interaction of the RBD with ACE2 (Fig. 5). To ensure that SSAA09E2 does not interfere with surface expression of ACE2, we incubated ACE2-expressing cells with DMSO or SSAA09E2 for 48 h, followed by surface labeling of the cells with anti-ACE2 antibody–FITC conjugate. The results (not shown) establish that SSAA09E2 does not prevent surface expression of ACE2 and likely affects initial recognition of the virus by directly interfering with SARS-S RBD–ACE2 interactions.

Effects of compounds on fusion of the SARS-S envelope with the host cellular membrane.

Following endocytosis and cathepsin L processing of the SARS-S envelope, the S2 domain of the spike fuses with the host cellular membrane by using the class I fusion mechanism. Fusion is followed by the release of the viral genome for viral replication (27, 30, 58, 59). To investigate whether any of the three compounds inhibited entry by interfering with fusion of the viral membrane with the host membrane, we performed a fusion assay as described in Materials and Methods. As shown in Fig. 6A and B, SSAA09E2 appeared to inhibit fusion only when added prior to overlay of SARS-S 293T cells with ACE2-expressing TZM-bl cells (Fig. 6B), consistent with the above data suggesting that it interferes directly with SARS-S RBD–ACE2 interactions. SSAA09E1 did not inhibit fusion when it was added prior to or after the overlay, consistent with the above data that predict that in an assay of this type, where fusion occurs at the surface of the cells and not at an endosomal compartment requiring function of cathepsin L, this inhibitor that blocks cathepsin L would not be likely to inhibit fusion. This is consistent with a previous report that showed that SARS-S-driven cell-cell fusion was independent of cathepsin L (60). Finally, SSAA09E3 inhibited fusion of the viral membrane with the host cellular membrane when it was added before or after overlay of the two cell types, consistent with a scenario where it blocks the process of membrane fusion regardless of the initial recognition between the SARS-S RBD and ACE2.

Time-of-addition experiments to establish whether the compounds target “early” or “late” steps of the entry process.

To further confirm the stages of SARS entry at which the compounds exhibit inhibitory activity, we performed time-of-addition experiments using the infection assay with the SARS/HIV pseudotyped virions and added the compounds at various time points after infection, as shown in Fig. 7. The results showed that entry inhibition by SSAA09E2 was effective only if the compound was added for up to 1 h postinfection. In contrast, the other two compounds, SSAA09E1 and SSAA09E3, were effective until at least 3 h postinfection (Fig. 7). As a control, we used the known cathepsin L inhibitor Z-FY(t-Bu)-DMK (Calbiochem), which also showed effective inhibition until 3 h postinfection and, to a lesser extent, at 6 h postinfection. Taken together, these data strongly suggest that SSAA09E2 targets an “early” step of SARS/HIV entry, such as initial recognition of the SARS-S RBD by ACE2, whereas SSAA09E1 and SSAA09E3 inhibit later stages of viral entry.

Inhibition of infectious SARS-CoV.

The antiviral activity of SSAA09E3 was determined by using a SARS-CoV cytopathic effect assay as described in Materials and Methods. Independent experiments were carried out at the Southern Research Institute and the University of Louisville. The results showed that SSAA09E3 inhibits SARS-CoV infection of Vero cells with a submicromolar EC₅₀ (Table 1), resulting in a promising selectivity index (>100).

Effects of compounds on postentry steps of SARS-CoV replication.

Infectious SARS-CoV was significantly more sensitive to the inhibitors than the SARS/HIV pseudotyped virions. This could have been the result of differences in conditions, such as the cell type or the 100-fold difference in MOI, but it was also possible that fully infectious virus was more sensitive to inhibitors because other steps of the viral life cycle were also blocked. To determine whether the 3 compounds also affected steps that are downstream from viral entry, we performed a SARS-CoV replicon assay as described in Materials and Methods. This assay was described in detail in a previous publication (41). We performed these experiments with the SARS-CoV replicon in the presence and absence of SSAA09E1, SSAA09E2, and SSAA09E3. As a control, we used SSYA10-001, a compound that we discovered previously in a search for SARS-CoV helicase (nsp13) inhibitors (41). As shown in Fig. 8, none of the entry inhibitors interfered with replication of the SARS-CoV replicon, whereas, as expected, our nsp13 helicase inhibitor, SSYA10-001 (41), suppressed SARS-CoV replicon function. These data suggest that the three compounds presented here are specific for inhibiting the entry process of the virus.