Systematic Studies on the Anti-SARS-CoV-2 Mechanisms of Tea Polyphenol-Related Natural Products

The causative pathogen of COVID-19, severe acute respiratory syndrome-coronavirus-2 (SARS-CoV-2), utilizes the receptor-binding domain (RBD) of the spike protein to bind to human receptor angiotensin-converting enzyme 2 (ACE2). Further cleavage of spike by human proteases furin, TMPRSS2, and/or cathepsin L facilitates viral entry into the host cells for replication, where the maturation of polyproteins by 3C-like protease (3CLpro) and papain-like protease (PLpro) yields functional nonstructural proteins (NSPs) such as RNA-dependent RNA polymerase (RdRp) to synthesize mRNA of structural proteins. By testing the tea polyphenol-related natural products through various assays, we found that the active antivirals prevented SARS-CoV-2 entry by blocking the RBD/ACE2 interaction and inhibiting the relevant human proteases, although some also inhibited the viral enzymes essential for replication. Due to their multitargeting properties, these compounds were often misinterpreted for their antiviral mechanisms. In this study, we provide a systematic protocol to check and clarify their anti-SARS-CoV-2 mechanisms, which should be applicable for all of the antivirals.


INTRODUCTION
Since its emergence in 2019 December, the coronavirus disease 2019 (COVID-19) has infected approximately 0.7 billion people and resulted in 6.9 million deaths.The pandemic continues with a case fatality rate of ∼1%. 1,2The causative agent of COVID-19 shares genomic homology with the severe acute respiratory syndrome-coronavirus (SARS-CoV) that caused an outbreak in 2002−2003 3,4 and was thus named SARS-CoV-2 by the World Health Organization (WHO).These human coronaviruses (CoVs) belong to the β genera of the CoV subfamily, which consists of single-stranded positivesense RNA viruses.The viral infection is initiated by binding of the receptor-binding domain (RBD) of spike on SARS-CoV and SARS-CoV-2 to human receptor angiotensin-converting enzyme 2 (ACE2).After virus attachment, spike is initially cleaved by the human transmembrane protease furin at the S1/ S2 site, dividing it into S1 and S2 fragments, followed by cleavage at the S2 site through human transmembrane serine protease 2 (TMPRSS2) to facilitate membrane fusion for viral RNA entry. 5Alternatively, the virus may undergo endocytosis mediated by ACE2 to form an endosome.Within the endosome, the spike is cleaved by cathepsin L at the S1/S2 site to facilitate membrane fusion and the subsequent release of the viral RNA into the cytosol. 6fter virus entry, the positive-sense viral RNA is translated into two viral polyproteins, pp1a and pp1ab, by the host cellular machinery.These polyproteins are then cleaved by the virus-encoded papain-like protease (PL pro ) and 3C-like protease (3CL pro ), which are embedded within the polyproteins and are self-cleaved for maturation, to yield 16 mature nonstructural proteins (NSPs). 7While 3CL pro primarily facilitates the maturation of NSPs, PL pro exhibits multifaceted activities, including polyprotein cleavage, removal of ubiquitin, and deISGylation of interferon (IFN)-stimulated gene product 15 (ISG15) to antagonize host immunity. 8,9Subsequently, NSP12 forms a complex with two accessories NSP7 and NSP8, possessing RNA-dependent RNA polymerase (RdRp) activity. 10,11Additionally, RdRp collaborates with helicase (NSP13) 12 as well as several RNA-processing enzymes such as NSP14, a bifunctional enzyme with 30-to-50 exoribonuclease (ExoN) and N7-methyltransferase activities, 13 to form the replication−transcription complex, which could further synthesize mRNA from RNA(+) and generate four structural proteins, spike (S), nucleocapsid (N), membrane (M), and envelope (E) proteins for assembly of new virus particles.
−16 Among the food and drug administration (FDA)-approved anti-COVID-19 drugs, antibodies such as bebtelovimab hinder viral attachment by binding to the RBD to inhibit its interaction with ACE2. 17emdesivir and molnupiravir inhibit RdRp, 18−20 while Paxlovid (active ingredient: nirmatrelvir) inhibits 3CL pro . 21ith the emergence of drug resistance, 22−25 more drug candidates are desired.−33 However, the antiviral mechanisms of certain effective natural products, such as tea polyphenols, remain incompletely understood due to their multiple-targeting properties.These natural products have been assessed through limited enzymatic assays and pseudovirus entry assays and/or predicted as inhibitors via computer modeling (refer to the Section 4) to be classified as anti-SARS-CoV-2 agents by inhibiting the targets assayed.It is now recognized that blocking the interaction between RBD and ACE2 as well as inhibiting enzyme activities of TMPRSS2, furin, and/or cathepsin L could block virus entry, whereas targeting viral enzymes like PL pro , 3CL pro , and/or RdRp could antagonize virus replication.Whether the antivirals inhibit virus entry or replication could be distinguished by assessing their effectiveness through pre-entry and postentry treatments.In this study, we systematically investigated the inhibitory effects of a series of tea polyphenol-related natural products (chemical structures shown in Figure 1) on various targets, including RBD/ACE2 interaction, TMPRSS2, furin, cathepsin L, 3CL pro , PL pro , and RdRp, and elucidated their antiviral mechanisms through entry or postentry treatment, to clarify their antiviral mechanisms.

Materials.
All of the natural products used herein were purchased from MedChemExpress (NJ), with the exception of epitheaflagallin-3-O-gallate obtained from GlpBio (CA), theaflavin-3-gallate from Cayman Chemical (MI), and TF3 from ChromaDex, Inc. (CA).All of the chemical reagents employed were of the highest grade.

Expression and Purification of Recombinant TMPRSS2.
The recombinant TMPRSS2 ectodomain (residues 109−492), excluding the transmembrane domain, was expressed using the ExpiSf Baculovirus Expression System (Thermo Fisher, catalog no.A38841, A39111, A39112) as reported previously. 34,35Briefly, the Escherichia coli DH10Bac cells were transformed with the TMPRSS2 plasmid to generate viral bacmid DNA, which was subsequently used to transfect ExpiSf cells for the production of recombinant baculovirus particles.These viral particles were then amplified from the P0 to P1 viral stocks.Recombinant P1 viruses were used to infect ExpiSf9 insect cells in ExpiSf CD medium.After 4 days of infection with cell viability dropping to 40−50%, the cell culture containing the secreted His-tagged TMPRSS2 was loaded onto a HisTrap column (Cytiva) to capture the target protein.The HisTrap column was washed with phosphatebuffered saline (PBS) buffer containing up to 25 mM imidazole and then eluted with the buffer containing up to 250 mM imidazole.The partially purified protein was then activated by the addition of enterokinase (NEB) and subjected to further purification using a Superdex 75 10/300 GL column (GE Healthcare).
2.3.TMPRSS2 Inhibition Assays.TMPRSS2 activity was assayed using a fluorogenic substrate Boc-Gln-Ala-Arg-AMC (Bachem, catalog no.4017019.0025,Switzerland) as reported previously. 34,35A microplate reader (BioTek Synergy H1) was used to measure the fluorescence increases at excitation and emission wavelengths of 355/460 nm.The assay was conducted in the presence of 1.3 nM TMPRSS2, 10 μM substrate, and various concentrations of inhibitors in a buffer of 20 mM tris−HCl (pH 7.4) containing 1% dimethyl sulfoxide (DMSO) from the stock solutions of inhibitors.The concentration-dependent inhibition curves of TMPRSS2 were fitted with the equation A(I) = A(0) x {1−[I/(I + IC 50 )]} using GraphPad Prism (v.9.4.0) software to determine the IC 50 values.Here, A(I) represents the enzyme activity at a given inhibitor concentration I, while A(0) denotes the enzyme activity in the absence of inhibitor, and I represents the inhibitor concentration.All of the measurements were triplicated to calculate the averaged IC 50 values and the standard deviations.
2.4.Furin Inhibition Measurements.Following the protocol as described in ref 35 a commercial Furin Protease Assay Kit (BPS Bioscience, catalog#78040) was employed to assay furin activities.It was conducted in 20 μL reaction mixtures on a 384-well plate (PerkinElmer) comprising 0.25 ng/μL furin, 2 μM substrate, and different concentrations of inhibitors in the assay buffer with 1% DMSO from the stock solutions of compounds.Fluorescence intensity was monitored at excitation/emission wavelengths of 380/460 nm.The concentration-dependent inhibition curves of furin were fitted with the equation A(I) = A(0) x {1−[I/(I + IC 50 )]} using GraphPad Prism (v.9.4.0).All of the measurements were performed in triplicate to determine the averaged IC 50 values and the standard deviations.
2.5.Cathepsin L IC 50 Measurements.Cathepsin L activity was determined using a commercial Cathepsin L Protease Assay Kit (Abcam, U.K.) in 20 μL reaction mixtures on a 384-well plate (PerkinElmer).Different concentrations of each inhibitor were added to the assay buffer supplied with dithiothreitol (DTT) in the kit, also containing 1% DMSO from the stock solutions of inhibitors.The fluorescence intensity was monitored for 30 min at excitation and emission wavelengths of 380/460 nm.The concentration-dependent inhibition curves of cathepsin L were fitted with the equation A(I) = A(0) x {1−[I/(I + IC 50 )]} using GraphPad Prism (v.9.4.0).The measurements were performed in triplicate to generate the averaged IC 50 values and the standard deviations.
2.6.RBD/ACE2 Cell-Based Binding Assay.The inhibition of RBD/ACE2 binding by the natural products was evaluated by using the NanoBiT technology commercial kit from Promega (WI).A stable cell line expressing SmBiTtagged human ACE2 on HeLa cells was established, and the recombinant RBD-LgBiT protein (amino acids 330−521 of spike) was produced. 35,36To monitor the interaction between RBD and ACE2, SmBiT-ACE2-expressing cells were seeded onto a white 96-well plate at a density of 1 × 10 4 cells per well (in triplicate).For each binding assay, cells were washed once with PBS and pretreated with 50 μL of the indicated compounds per well for 10 min.Next, a 50 μL reaction mixture containing 10 ng of RBD-LgBiT, 0.5 μL of Nano-Glo luciferase assay substrate, and 9.5 μL of luciferase assay diluent (Promega) was added into each well.The luminescence signal was recorded every 2 min continuously for 1 h using a microplate reader (BioTek Synergy HTX, VT) at 37 °C with a time-lapsed kinetics program.To calculate RBD inhibition by all agents, luminescence data from the time point showing the highest signal in the negative control sample was selected for downstream calculation.Inhibition (%) = [1-(luminescence signal of test sample)/(luminescence signal of negative control sample)] × 100.The compound concentration required to inhibit 50% of the interaction is defined as IC 50 .All of the measurements were performed in triplicate to determine the averaged IC 50 values and the standard deviations.
2.7.3CL pro Inhibition Measurements.Recombinant SARS-CoV-2 3CL pro was prepared as previously reported. 37he purified tag-free 3CL pro was dialyzed against a buffer containing tris−HCl (pH 7.5), 120 mM NaCl, 0.1 mM ethylenediaminetetraacetic acid (EDTA), and 2 mM DTT and stored at −70 °C until use.3CL pro activity was monitored using a fluorogenic peptide Dabcyl-KTSAVLQSGFRKME-Edans purchased from Yuan Yu Ltd. (Taiwan).This peptide contained a fluorescence quenching pair, and the fluorescence increased when the peptide was cleaved by the protease.Fluorescence resulting from the cleavage of the fluorogenic peptide by 3CL pro was measured over time at 538 nm with excitation at 355 nm using a fluorescence plate reader.IC 50 values of the active compounds were measured in reaction mixtures containing 35 nM 3CL pro and 6 μM fluorogenic substrate in a buffer of 20 mM Bis-tris (pH 7.0) in the absence and presence of various concentrations of inhibitors.No additional reducing agent was added since the 3CL pro storage buffer already contained 2 mM DTT, which should be present in the assay buffer as well.It is important to note that previous studies have demonstrated that in the absence of DTT, compounds like ebselen, disulfiram, carmofur, PX-12, tideglusib, and shikonin exhibit nonspecific inhibition not only against 3CL pro but also against a panel of viral cysteine proteases, including SARS-CoV-2 PL pro , as well as 2A pro and 3C pro from enteroviruses A71 and D68. 38All of the measurements were triplicated to yield the averaged IC 50 and standard deviations.
2.8.PL pro IC 50 Measurements.Recombinant SARS-CoV-2 PL pro was prepared as previously reported. 37The purified PL pro was dialyzed against a buffer containing tris−HCl (pH 7.5), 120 mM NaCl, 0.1 mM EDTA, and 2 mM DTT and stored at −70 °C until use.For the determination of IC 50 against proteolytic, deubiquitinating, and deISGylation activities of PL pro , the following fluorogenic substrates were utilized: Z-Arg-Leu-Arg-Gly-Gly-AMC (Cat.#79997, Bachem Bioscience, Switzerland) for proteolytic activity, Ub-AMC (Cat.#U550-050, Boston Biochem) for deubiquitinating activity, and ISG15-AMC (Cat.#U553-050, Boston Biochem) for deISGylation activity.The initial rates of the reactions involving 75 nM PL pro and 10 μM fluorogenic substrate were measured in a buffer of 20 mM HEPES (pH 7.5), in both the absence and presence of various concentrations of inhibitors.The measurements were taken at 460 nm upon excitation at 355 nm.No additional reducing agent was added as 2 mM DTT was already present in the PL pro storage buffer and should be present in the assay buffer as well.The initial rates of the inhibited reactions were plotted against different inhibitor concentrations to determine the IC 50 value by fitting with the following equation: A(I) = A(0) × {1-[I/(I + IC 50 )]}.All of the measurements were performed in triplicate to determine the averaged IC 50 values and the standard deviations.
2.9.Preparation and Assay of SARS-CoV-2 RdRp.The fusion protein nsp7−nsp8 (nsp7L8) was generated by inserting a GSGSGS linker sequence between the nsp7 and nsp8 coding sequences. 11The nsp7L8, nsp8, and nsp12 were produced and purified independently.The procedure involved using both E. coli for expressing nsp7L8 and baculovirus-infected insect cells for expressing nsp12. 39,40Subsequently, nsp12, nsp7L8, and nsp8 were combined at a molar ratio of 1:3:3 and preincubated on ice for 10 min to facilitate the formation of an active complex following a reported protocol. 41,42In vitro RdRp activity was monitored using a fluorescence plate reader (Synergy H1 Hybrid multimode Reader, BioTek) with a 2.10.Antiviral EC 50 Measurements.Vero E6 cells were seeded onto a 24-well culture plate in Dulbecco's modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS) and antibiotics 1 day before infection.For entry treatment, SARS-CoV-2 Delta virus (NTU92) or Omicron BA.5 (NTU280) at 50−100 plaque-forming units (pfu) was preincubated with test compounds for 1 h at 37 °C before being added to cells.Then, the mixture was incubated for an additional hour.After the removal of virus inoculum, the cells were washed once with phosphate-buffered saline (PBS) and overlaid with 1 mL of overlay medium containing 1% methylcellulose, with or without test compounds for postentry treatment or entry treatment, respectively.After a 5 day incubation at 37 °C, the cells were fixed with 10% formalin overnight.Following the removal of the overlay media, the cells were stained with 0.5% crystal violet, and the plaques were counted.The percentage of inhibition was calculated as [1−(V D /V C )] × 100%, where V D and V C represent the virus titer in the presence and absence of test compounds, respectively.The minimal concentrations of compounds required to reduce 50% of plaque numbers (EC 50 ) were calculated by regression analysis of the dose−response curves generated from plaque assays.For each compound, the measurements were repeated at least 3 times to obtain the averaged EC 50 values and standard deviation.To ascertain the stages at which the compounds exerted their antiviral effects, the compounds were added before infection, during infection, and/or after infection, as described in the Section 3.
2.11.Cytotoxicity CC 50 Measurements.Cytotoxicity of the compounds was determined using the acid phosphatase (ACP) assay.Vero E6 cells were seeded onto a 96-well culture plate at a concentration of 2 × 10 4 cells per well.Next day, the medium was removed, and each well was washed once with PBS before adding DMEM containing 2% FBS and different concentrations of test compounds.After 3 days of incubation at 37 °C, the medium was removed, and each well was washed once with PBS.Next, a buffer containing 0.1 M sodium acetate (pH = 5.0), 0.1% Triton X-100, and 5 mM p-nitrophenyl phosphate was added.After incubating at 37 °C for 30 min, 1 N NaOH was added to stop the reaction.The absorbance was measured using an ELISA reader (VERSAmax, Molecular Devices, Sunnyvale, CA) at a wavelength of 405 nm.The percentage of cell viability was calculated by using the following formula: Viability % = [ (A t /A s ) × 100]%, where At and As represent the absorbance of the test compound and the solvent control, respectively.The 50% cytotoxicity concentration (CC 50 ) was determined by nonlinear regression analysis.For each compound, the measurements were repeated at least 3 times to determine the averaged CC 50 values and standard deviation.
2.12.Molecular Docking.The molecular docking analysis was conducted utilizing the Discovery Studio (DS) 2022 software to forecast the interaction between natural products and TMPRSS2, furin, and cathepsin L. The three-dimensional (3D) structures of TMPRSS2 (PDB: 7MEQ), furin (PDB: 4RYD), and cathepsin L (PDB: 5MQY) were retrieved from the RCSB Protein Data Bank (PDB, https://www.rcsb.org/),with the elimination of all water molecules and bound ligands.The binding sites were prepared by extracting the binding cavity surrounding nafamostat, a recognized inhibitor of TMPRSS2, with 12 Å radius.Similarly, the binding cavity encompassing para-guanidinomethyl-Phac-R-Tle-R-Amba, a known furin inhibitor, was extracted with the same radius as that for furin.The binding cavity of cathepsin L was determined analogous to that of TMPRSS2 and furin.The natural products were prepared using the built-in function in the DS software for ligand preparation.Prior to docking, all proteins and ligands were optimized to ensure stable conformations.

Assay of Compounds for Inhibiting RdRp.
We proceeded to evaluate the inhibitory activities of all of the natural products against RdRp using a commercial assay kit and our prepared RdRp complex. 42We found that theaflavin, theaflavin-3′-gallate, TF3, and proanthocyanidin exhibited inhibition toward RdRp with IC 50 values of 40.2 ± 2.3, 23.2 ± 0.4, 2.3 ± 0.6, and 23.4 ± 1.3 μM, respectively (Figure 4k− n).Consistent with the computer prediction, 43 TF3 emerged as the most effective inhibitor among them.The data are summarized in Table 1.

Antiviral Activities and Cytotoxicity of Natural Products.
In the final step, the natural products were subjected to the antiviral plaque reduction assay using Delta (NTU92) and Omicron BA.5 (NTU280) strains of SARS-CoV-2.To discern the stage at which the active compounds exerted their antiviral activities, these compounds were preincubated with the virus or added during infection and removed after the infection (entry treatment) or added only after infection (postentry treatment).Notably, catechin did not inhibit the Delta SARS-CoV-2 at 10 μM (figure not shown).The EC 50 values for theaflavin, epitheaflagallin-3-O-gallate, theaflavin-3-gallate, theaflavin-3′-gallate, TF3, and proanthocyanidin against the entry of the Delta variant of SARS-CoV-2 into Vero E6 cells were measured to be 7.4 ± 0.1, 0.5 ± 0.1, 0.4 ± 0.0, 0.8 ± 0.1, 0.2 ± 0.0, and 0.5 ± 0.1 μM, respectively (Figure 5a−f), and the data are summarized in Table 1.The active antivirals exhibited inhibition not only against the Delta strain but also antagonized the Omicron BA.5 strain (see Figure S1 for the dose-dependent inhibition curves and Table 1   example, not only inhibited Delta and Omicron BA.5 but also inhibited the wild-type, α, and γ strains with similar EC 50 values of 0.3 ± 0.0, 0.5 ± 0.0, and 0.4 ± 0.2, respectively (see Figure S2).All of the antivirals were found to be active when added at the entry stage but not at the postentry stage, suggesting that they likely prevent virus entry to host cells rather than virus replication inside the cells.The CC 50 values of theaflavin, epitheaflagallin-3-O-gallate, theaflavin-3-gallate, theaflavin-3′-gallate, TF3, and proanthocyanidin derived from the plots were 66.4 ± 3.2, 109.2 ± 45.3, >100, 69.9 ± 12.9, >100, and >100 μM, respectively, as shown in Figure 5a−f.The derived selectivity index values for these compounds were 9.0, 232.3, >285.7,93.2, >476.2, and >188.7,respectively.

Computer Modeling of the Inhibitors for the Human Proteases Involved in the Virus Entry.
Given that most of the tea polyphenols inhibit viral entry, understanding their binding modes with key human proteases such as cathepsin L, furin, and TMPRSS2 as well as the RBD/ACE2 interaction is crucial.Based on the IC 50 data, it appears that the smaller theaflavin inhibited only cathepsin L, while the larger epitheaflagallin-3-O-gallate, theaflavin-3-gallate, and theaflavin-3′-gallate inhibited furin and TMPRSS2, with the exception of theaflavin-3′-gallate also weakly inhibiting cathepsin L. However, only theaflavin-3′-gallate and even larger TF3 inhibited TMPRSS2 potently.To rationalize the structure− activity relationship, computer docking studies were conducted to explore the binding interaction of these potent inhibitors with their respective targets.In the docking analysis, theaflavin, which exhibits an IC 50 of 11.1 μM forms several hydrogen bonds with residues C25, W26, N66, G68, and H163 of cathepsin L. Especially, the central 3,4,6-trihydroxy-5H-benzo- [7]annulen-5-one group, highlighted in pink, forms hydrogen bonds with both catalytic residues C25 and H163, and the residue W26 of cathepsin L (Figure 6a).The two chromane rings highlighted in cyan form hydrogen bonds with residues N66 and G68.As shown in Figure 6b, theaflavin-3-gallate, the most potent inhibitor of furin in this study with an IC 50 of 36.7 μM, forms hydrogen bonds with residues D154, D191, R193, P256, D258, S363, H364, and T365 of furin, surrounding the active site residues D153, H194, N295, and S368.Its central 3,4,6-trihydroxy-5H-benzo [7]annulen-5-one group, identical to that in theaflavin and depicted in pink, forms a hydrogen bond with the catalytic residue H194.An OH group on the chromane ring depicted in cyan forms hydrogen bonds with the backbone oxygen of P256 and the side-chain carboxylate oxygen of D258.Additionally, two OH groups on the gallate moiety shown in yellow form hydrogen bonds with the sidechain oxygens of S363 and T365, while one OH forms a hydrogen bond with the side-chain nitrogen of H364.On the other hand, theaflavin-3′-gallate (IC 50 = 6.2 μM) and TF3 (IC 50 = 5.5 μM) exhibited potent inhibition against TMPRSS2.The gallate group of theaflavin-3′-gallate, shown in yellow, occupies the catalytic pocket of TMPRSS2 and forms hydrogen bonds with residues C437, G464, and H296 around the catalytic residues H296, D435, and S441 (Figure 6c).Its chromane ring, shown in cyan, could also form hydrogen bonds with the G439 and S441 residues.TF3 also forms hydrogen bonds with residues D435, C437, and S441 of TMPRSS2 through its chromane group shown in cyan (Figure 6d).Due to the extra gallate in TF3, its binding orientation is different from that of theaflavin-3′-gallate.

DISCUSSION
While numerous herbs and natural products have demonstrated inhibition of CoVs, 26−33 their antiviral mechanisms remain incompletely understood.For instance, pseudotyped  virus assays have facilitated the screening of natural products as SARS-CoV-2 entry inhibitors, 44 but the precise targets, such as RBD/ACE2, TMPRSS2, furin, and/or cathepsin L, have yet to be fully identified.In this study, we chose a series of tea polyphenols with similar building blocks and structural features to investigate their exact antiviral targets.As illustrated in Figure 1 from catechin to proanthocyanidin, we identified theaflavin, epitheaflagallin-3-O-gallate, theaflavin-3-gallate, theaflavin-3′-gallate, TF3, and proanthocyanidin as active antivirals against SARS-CoV-2.These compounds exhibited antiviral activity only when they were preincubated with the virus and included during virus infection of cells (entry treatment) but not when added after virus infection (postentry treatment).While these compounds exhibit inhibitory activities on various targets (Table 1), their true antiviral targets are likely to be limited to those for virus entry, such as RBD/ACE2 interaction, TMPRSS2, furin, and/or cathepsin L. This inference is supported by the fact that their antiviral activities were observed specifically under entry treatment conditions.
Catechin is a flavanol characterized by a hydroxyl group at the C-3 position of the C ring and was initially regarded as a potential antiviral targeting various components of SARS-CoV-2, including proteases, RdRp, spike, and ACE2, as suggested by computer modeling. 45However, our current study reveals that catechin exhibits only weak inhibitory activity against PL pro (Table 1), consistent with its failure to inhibit SARS-CoV-2.In contrast, theaflavin that looks like containing two molecules of catechin but with a different 7-membered ring to fuse with the benzene ring of a catechin inhibits not only the RBD/ACE2 interaction (IC 50 = 12.3 μM) and cathepsin L activity (IC 50 = 11.1 μM) but also activities of 3CL pro (IC 50 = 16.1 μM), PL pro (IC 50 = 9.9, 10.6, and 11.2 μM), and RdRp (IC 50 = 40.2μM).However, its inhibitory effect on SARS-CoV-2 with an EC 50 of 7.4 μM against the Delta strain was only observed with entry treatment (Figure 5a), suggesting that its antiviral activity is primarily attributed to inhibiting RBD/ACE2 and cathepsin L.  entry.Therefore, inhibiting RBD/ACE2 and the human protease TMPRSS2 indeed contributes to their antiviral effect, although they also inhibit 3CL pro , PL pro , and RdRp.
It is intriguing that theaflavin, smaller than TF3 and lacking two 3,4,5-hydroxybenzoyl groups at the ends (see Figure 1 for the structures), inhibits cathepsin L rather than TMPRSS2, although both compounds inhibit RBD/ACE2 to block the virus entry.In contrast, theaflavin-3′-gallate inhibits TMPRSS2 more potently (IC 50 = 6.2 μM), although it also weakly inhibits furin and cathepsin L. Comparing the docking results of theaflavin and theaflavin-3′-gallate, it was found that a portion of the latter compound was located outside the active site cleft of cathepsin L, likely due to the presence of an additional gallate group (figure not shown).The largest TF3 with extra gallate groups inhibits only TMPRSS2, but not furin and cathepsin L. Therefore, the computer models shown in Figure 6 successfully explain the structure−activity relationship of these structurally related natural products, with the smaller theaflavin inhibiting cathepsin L and larger theaflavin-3′-gallate and TF3 inhibiting TMPRSS2.
Theaflavin, derived from black tea, has been reported to inhibit SARS-CoV-2 3CL pro . 46Additionally, it has been found to inhibit the RBD/ACE2 interaction and consequently block the pseudovirus entry. 47Here, we found that theaflavin also inhibits cathepsin L, further contributing to its ability to block SARS-CoV-2 entry by interfering with the RBD/ACE2 interaction.Moreover, TF3 was reported to effectively inhibiting the interaction between recombinant ACE2 and RBD at 60 μM, which is comparable to the concentration of this compound in tea beverages. 48TF3 is a black tea polyphenol produced by the polymerization and oxidation of green tea polyphenols epicatechin gallate and (−)-epigallocatechin-3-gallate, an antioxidant commonly found in green and black tea, during fermentation of fresh tea leaves.Previous studies have demonstrated TF3′s inhibitory effect on SARS-CoV 3CL pro in vitro 29 as well as blocking the RBD/ACE interaction. 47Moreover, TF3 has been shown to antagonize Zika virus's protease activity and virus replication. 49In this study, we demonstrate that TF3 not only antagonizes the RBD/ACE2 (IC 50 = 8.7 μM), but also inhibits TMPRSS2 (IC 50 = 5.5 μM), thereby effectively blocking SARS-CoV-2 entry.The dual-inhibitory action results in a potent antiviral EC 50 of 0.2 μM against the Delta SARS-CoV-2, without toxicity at a concentration of up to 100 μM (selectivity index SI >476.2).
Compared to the potentially repurposed drugs for COVID-19, certain herbs and natural products demonstrate antiviral properties that surpass some conventional drugs while exhibiting low toxicity. 50It appears that multiple targeting is a common feature of natural products with broad biological activities, including polyphenols described in this study.The antiviral ability of polyphenols may stem from the presence of multiple OH groups for forming H-bonding interactions and aromatic rings for hydrophobic and π−π interactions to inhibit the targets essential for SARS-CoV-2 entry, replication, and/or immune escaping.−54 Anionic polyphenols could act as ARBs, similar to anionic tetrazolates of antihypertensive drugs, such as elmisartan, candesartan, and losartan, which have been demonstrated to clinically protect hypertensive patients infected by COVID-19. 55With multiple targets, it is important to avoid misinterpretation of the antiviral mechanisms of polyphenols on the basis of only limited assays.We propose that for compounds acting as the SARS-CoV-2 entry inhibitors (effective via entry treatment), investigations should focus on accessing their impacts on targets such as RBD/ ACE2, TMPRSS2, furin, and cathepsin L. Conversely, for compounds inhibiting SARS-CoV-2 replication within cells (effective via postentry treatment), assessments should include targets such as 3CL pro , PL pro , and RdRp to gain a comprehensive understanding of their true antiviral mechanisms.

CONCLUSIONS
The systematic assays conducted in the current study have provided an insight into the anti-SARS-CoV-2 mechanisms of polyphenols.Through multiple targeting, natural products related to tea polypehenols inhibit various targets, albeit not always with high potency individually.However, the combination of these inhibitory activities results in sub-μM EC 50 , demonstrating their potential effectiveness against SARS-CoV-2.Extensive cohort studies and human intervention trials on COVID-19 patients, remarking on the possibility of decreasing virus multiplication and thus improving clinical signs, could be conducted following comprehensive in vivo investigations.In fact, a study demonstrated that epigallocatechin gallate, a green tea polyphenol, reduced SARS-CoV-2 replication in a mouse model. 56This suggests that the incorporation of polyphenols into food and clinical practice should be swiftly implemented, as they are natural mixtures derived from plants, already approved for use as herbs and food.Importantly, the use of polyphenols offers the advantage of a high safety profile with minimal risk of causing major side effects.Therefore, our studies presented here not only elucidate the antiviral mechanisms of certain natural products but also offer effective options for preventing and/or treating COVID-19 with these anti-SARS-CoV-2 natural products.

* sı Supporting Information
The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c02392.EC 50 profiles of natural products against Omicron BA.5 and that of TF3 against different strains of SARS-CoV-2 (PDF)

Figure 6 .
Figure 6.Modeled binding modes of selected natural products with cathepsin L, furin, and TMPRSS2 to rationalize their inhibition specificities.The binding modes of (a) theaflavin with cathepsin L (PDB: 5MQY), and (b) theaflavin-3-gallate with furin (PDB: 4RYD), and (c) theaflavin-3′gallate and (d) TF3 (bottom) with TMPRSS2 (PDB: 7MEQ), respectively.These compounds represent the best inhibitors of their targets in this study.The central 3,4,6-trihydroxy-5H-benzo[7]annulen-5-one moieties are colored pink, chromane rings are in cyan, and gallate groups are shown in yellow.Oxygen atoms are shown in red.In amino acids, carbon skeletons are colored green, nitrogen atoms in blue, and sulfur atoms in yellow.
for their measured EC 50 values).Specifically, TF3, as an