Methods
Compound library and candidate compounds
TA was identified from the Food and Drug Administration (FDA) compound library (US Drug Collection), a unique collection of 1280 small molecules that have reached clinical trials. TA was purchased from the United States of America. TS-984 and TS-1276 from the Topscience compound library (Cat. No. L6000, Topscience, Shanghai, China), and contained 1584 natural compound products. All the compounds were provided in 10 mM of dimethyl sulfoxide (DMSO). The compounds of the libraries were diluted to 100 μM for HTRF assay.
Experimental cell lines and reagents
Cell lines: ACE2-overexpressing 293T cells (with mCherry labeled vector), 293T cells (only vector overexpressed), ACE2-overexpressing Capan2 cells (mCherry labeled vector), Capan2 cells (only vector overexpressed), were all donated from Chaonan Qian’s laboratory (SYSUCC, Guangzhou, China). SARS-CoV-2_S (D614G)-pseudotyped lentivirus (> 108 TU/ml, 10 × 100 μl, HBSS buffer solution) vector was VB900088-2229upx, with the polybrene (5 mg/ml) both of which were purchased from VectorBuilder China (Guangzhou, China).
ACE2 tagged with C-Fc and labeled with DRA36, and 2019-nCoV-S-protein tagged with C-6His and labeled with C05Y, were purchased from Novoprotein (Shanghai, China). Positive inhibitor control Nafamostat mesylate (T2392) and Emodin (T2869) were purchased from Topscience (Shanghai, China). The PAb Anti Human IgG-d2 (61HFCDAB) and MAb Anti-6HIS-Tb cryptate Gold (61HI2TLB) were purchased from CisBio Bioassays (Codolet, France). DMEM (Gibco, Cat#C11995500BT), FBS (Invitrogen, Cat# 10500064), 0.25% Trypsin (Invitrogen, Cat#25200056). D-PBS (Invitrogen, Cat#14190169).
Preparation of the working compound library
To prepare the working compound library, 2 μl of each compound from the stock library was dispensed into each well of the 384-well plate using a Voyager pipette (Integra, Zizers, Switzerland). A phosphate buffered saline (PBS) (1×) was used as a negative control, and emodin (100 μM) was used as a positive control [
2,
3]. The final reaction volume was 20 μl, and the final compound concentration was 100 μM in PBS.
Optimization of the HTRF-based spike protein/ACE2 inhibitor screening system
To optimize the interaction between the spike protein (S protein) and ACE2, we performed cross-titration experiments to determine the maximal effect. In brief, ACE2 and SARS-Cov-2 S-protein were prepared at multiple concentrations with a PBS containing 0.1% BSA, 5 μl ACE2 (Mammalian, C-Fc, DRA36, Novoprotein) and 5 μl SARS-Cov-2 S-protein RBD (Mammalian, C-6His, C05Y, Novoprotein). Each concentration was added into each well of the 384-microplate (ProxiPlate™ 384-shallow well Microplates, 66PLP96025, CISBIO) and the mixture incubated at 37 °C for 1 h. Next, 5 μl PAb Anti Human IgG-d2 (61HFCDAB) and 5 μl MAb Anti-6HIS-Tb cryptate Gold (61HI2TLB) were added to each well with the ACE2/S-protein RBD mixture (final reactive volume of 20 μl) following the supplier’s protocols. After 30 min final incubation at room temperature, HTRF signals were measured using a Multimode Reader (Spark 10M, Tecan) equipped with an excitation filter of 340 nm, and fluorescence detected at 620 and 665 nm with a lag time of 100 μs and an integration time of 200 μs. The results were analyzed using a two-wavelength signal ratio: [intensity (665 nm)/intensity (620 nm)] × 10
4 (HTRF Ratio). The Z factor was calculated using the following equation:
$${\text{Z}} = 1 - \frac{{3*{\text{SD}}\;{\text{negative}} + 3*{\text{SD}}\;{\text{positive}}}}{{{\text{MEAN}}\;{\text{negative}} - {\text{MEAN}}\;{\text{positive}}}}$$
Standard deviation (SD).
The initial screening assay was repeated twice and the hits confirmed by the determination of IC50 (HTRF) in quadruplicates. IC50 (HTRF) was defined as the compound concentration at which the combination of ACE2 and S-RBD decreased by 50%.
Pseudovirus neutralization assay on ACE2-overexpressing 293T cells
To further test the inhibiting effect of the candidate compounds on the binding of S protein and ACE2 at the cellular level, we performed pseudovirus neutralization assay [
4,
5]. One day before SARS-CoV-2 pseudovirus transduction (day 0), 293T cells were washed once with D-PBS and dissociated using 0.25% of Trypsin. Approximately 3 × 10
4 ACE2-overexpressed and vector-overexpressed 293T cells were seeded in each well of the 96-well plates at 37 °C with 5% CO
2 overnight. On the first day of SARS-CoV-2 pseudovirus transfection (day 1), the frozen SARS-CoV-2_S (D614G)-pseudotyped lentivirus was melted on ice and gently pipetted several times to mix the dissolved virus particles. Then, 50 μl of virus solution was added to 450 μl of fresh complete culture medium (DMEM + 10% FBS) containing 5 μg/ml of polybrene, and mixed gently. And the candidate compounds TS-984, TS-1276 and nafamostat mesylate were made into 100 mM of stock solutions with DMSO, meanwhile, TA was prepared in 100 mM of stock solutions with PBS. Then all the stock solutions were diluted to 50 and 100 μM with the mixture of fresh complete culture medium with virus. TA was diluted to 15 and 30 μM with the mixture of fresh complete culture medium with virus. The original medium was then changed with 70 µl of the above mixture with candidate compounds. Finally, the plate was shaken gently so that the virus solution covered every cell, and then placed into a carbon dioxide incubator at 37 °C and 5% CO
2 for culturing. After 24 h infection, the cultures were subjected to fluorescence measurement using a Nikon ECLIPSE Ti2.
Pseudovirus neutralization assay on ACE2-overexpressing pancreatic carcinoma cell line Capan2
One day before SARS-CoV-2 pseudovirus transduction (day 0), Capan2 cells were washed once with D-PBS and dissociated using 0.25% of Trypsin. Approximately 3 × 104 ACE2-overexpressed and vector-overexpressed Capan2 cells were seeded in each well of the 96-well plates at 37 °C with 5% CO2 overnight. On the first day of SARS-CoV-2 pseudovirus transfection (day 1), the frozen SARS-CoV-2_S (D614G)-pseudotyped lentivirus was melted on ice and gently pipetted several times to mix the dissolved virus particles. Then, 50 μl of virus solution was added to 450 μl of fresh complete culture medium (DMEM + 10% FBS) containing 5 μg/ml of polybrene, and mixed gently. And the candidate compounds TS-984, TS-1276 and nafamostat mesylate were made into 100 mM of stock solutions with DMSO, meanwhile, TA was prepared in 100 mM of stock solutions with PBS. Then TS-984, TS-1276, and nafamostat mesylate were diluted to 50 and 100 μM with the mixture of fresh complete culture medium with virus. TA was diluted to 15 and 30 μM with the mixture of fresh complete culture medium with virus. The original medium was then changed with 70 µl of the above mixture with candidate compounds. The original medium was then changed with 70 µl of the above mixture and a certain volume of concentration-graded candidate compounds. Finally, the plate was shaken gently so that the virus solution covered every cell, and then placed into a carbon dioxide incubator at 37 °C and 5% CO2 for culturing. After 24 h infection, the cultures were subjected to fluorescence measurement using a Nikon ECLIPSE Ti2.
Molecular docking
HTRF-based assay and pseudovirus neutralization assay suggested that TS-984 effectively inhibited the binding of the coronavirus S-protein and the human cell AEC2 receptor. To understand the structural basis of the inhibitory effects, we further investigated the binding mode of TS-984 to ACE2. The docking of TS-984 and the S protein/ACE2 complex was completed with the software Autodock 4.0 [
6]. Firstly, the 2D structures of TS-984 were constructed in chimera [
7] and optimized in autodock 4.0. There are 10 different conformations for the ligand TS-984. The crystal structure of the S protein/ACE2 complex was obtained from PDB database and the protein ID was 6m0j [
8]. The simple optimization to the S protein/ACE2 complex includes adding the side chain of amino acid residues, adding the missing loop part in the crystal structure, distributing the protonation state of amino acid residues, and optimizing the whole protein structure under the condition of OPLS2005 [
9] force field. In the docking process, 500 positions are generated in the initial stage, and the highest scoring-100 positions are minimized by conjugate gradient minimization. Q-site was used to find the possible binding pocket of TS-984 in the S protein/ACE2 complex structure.
Statistical analyses
Data are presented as means ± SD. Student’s t tests were performed for all the experiments, except where indicated differently in the figure legends.
Discussion
Currently, there are few effective anti-SARS-CoV-2 drugs available for clinically treating COVID-19. Given the urgency of the pandemic, agencies and companies worldwide are giving top priority to repurpose existing drugs. Up to now, several small-molecule compounds have been repurposed for treating COVID-19 [
11]. These compounds include Hydroxychloroquine (HCQ) developed by Sanofi, Remdesivir by Gilead, Favipiravir by Toyama, Lopinavir-ritonavir by Abbott, fluvoxamine by SolVay Pharmaceutics, molnupiravir by Merck and Ridgeback, and Paxlovid by Pfizer [
11‐
13]. These agents are considered to exert their antiviral effects through different mechanisms such as blocking viral entry into host cells, inhibiting an essential virally encoded enzyme, targeting a host enzyme required for viral replication, and obstructing coronavirus particle formation [
14]. However, while some early studies have reported that these agents appeared to inhibit SARS-CoV-2, large-scale clinical trials showed that most of them failed to provide significant benefits for hospitalized COVID-19 patients. For instance, Remdesivir, approved for use in patients with severe COVID-19, is believed to deter virus from replication by inhibiting a critical RNA polymerase, it was never demonstrated to effectively block SARS-CoV-2 infection and improve clinical outcomes [
15‐
18]. Other large-scale clinical data have further revealed that Remdesivir does not exert any significant anti-SARS-CoV-2 effects in patients with severe COVID-19 [
19,
20].
TMPRSS2, a serine protease, primes the S protein of human coronaviruses MERS-CoV and SARS-CoV, and promotes its entry into the host cell. Camostat mesilate, a protease inhibitor developed for the treatment of pancreatitis in Japan in the 1980s, was reported to inhibit TMPRSS2 and block the entry of SARS-CoV and SARS-CoV-2 into host cells [
21]. Previous studies have demonstrated that camostat mesilate suppresses virus-cell membrane fusion in vitro and in vivo, thus, viral replication [
22,
23]. Another study revealed that camostat mesilate blocks the entry of the SARS-CoV-2 virus into the lungs [
24]. Therefore, camostat mesilate has been repurposed for the treatment of COVID-19 in clinical trials [
25].
Nafamostat mesilate was also found to suppresses TMPRSS2 and virus-cell fusion, and thus block the entry of SARS-CoV-2 virus into host cells [
26,
27]. Notely, nafamostat mesilate displayed a much more potent inhibiting effect compared with camostat mesilate on TMPRSS2 [
23]. In present study, we further demonstrated that nafamostat mesilate can block the interaction of the coronavirus S protein and its host ACE2 receptor. Based on our results, nafamostat mesilate exerted a more effective inhibiting power on the binding of the S protein to ACE2 compared with the positive control emodin. Thus, nafamostat mesilate is a dual inhibitor of TMPRSS2 and the binding of the S protein to its ACE2 receptor.
Cellular entry of SARS-CoV-2 relies on two key procedures: the activation of the viral surface S protein by TMPRSS2 proteolytic processing, and the binding of the activated S protein (S1) to the cell surface receptor ACE2 for fusion of the virus-cell membrane. On the other side, maturation of the SARS-CoV-2 virions in host cells depends on a proteolysis of the viral precursor polyprotein by the main protease (M
pro/3CL
pro). A previous study uncovered that tannic acid is able to inhibit both TMPSS2 and the main protease (M
pro/3CL
pro). Thus, tannic acid is a potential dual inhibitor of both the SARS-CoV-2 main protease (M
pro) and the TMPRSS2 [
28]. Apparently, targeting the TMPRSS2 as well as M
pro is a better option for the treatment of COVID-19. However, no previous studies have investigated whether nafamostat mesilate or tanic acid can inhibit the binding of the S protein and ACE2 as well. To date, except for emodin which shows only a moderate inhibition effect [
2], no other compounds have been reported to exert an inhibition effect on the binding of the S protein and ACE2. In this study, we demonstrated that tannic acid inhibits the binding of the S protein to the ACE2 host cells, indicating that tannic acid is a triple inhibitor for TMPRSS2, M
pro, and S-ACE2 binding. At present, Tannic acid is the only identified drug that can inhibit TMPRSS2, M
pro, and the interaction between the coronavirus S protein and the ACE2 human cell receptor. However, our data shows that tannic acid has higher cytotoxicity that leads to cell death when the effective concentration is applied. For this reason, tannic acid might be an unsuitable drug for SARS-CoV-2 treatment, as it cannot be directly repurposed for the treatment of COVID-19 patients before the cytotoxicity is reduced.
In this present study, we identified a more potent inhibitor for blocking the binding of the S protein and ACE2. TS984 (9-Methoxycanthin-6-one), is an indole alkaloid and one of the main constituents in Eurycoma longifolia and Simaba multiflora. TS-984 has never been used as an antineoplastic and antiplasmodial agent [
29‐
31]. In this study, we identified for the first time that TS984 is able to block the binding of the coronavirus S protein to the host ACE2 receptor. Compared with tannic acid and nafamostat mesilate, TS984 (9-Methoxycanthin-6-one) presented a much stronger inhibiting effect with a lower cytotoxicity. We also demonstrated that TS-1276, anthraquinone, exhibited an inhibiting effect on the binding of the S protein to ACE2. But the effect of TS-1276 was much weaker compared with TS-984. Intriguingly, both TS-1276 and emodin belong to anthraquinone compounds. Emodin is a derivative of TS-1276.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.