Molecular biology grade DMSO was from Sigma-Aldrich (St. Louis, MO, United States). Sterile PBS was purchased from ThermoFisher Scientific (Waltham, MA, United States). Recombinant full-length untagged PLpro with His-Tag (SARS-CoV-2), assay buffers, inhibitors, fluorescently labeled substrates, and deubiquitinase substrates were from BPS Biosciences (San Diego, CA, United States). The phytochemicals were purchased from Sigma-Aldrich (St. Louis, MO), MedChemExpress (Princeton, NJ), and Cayman Chemicals (Ann Arbor, MI). The names of the phytochemicals, manufacturer names, and catalog numbers are listed in Supplementary Table S1.

Phytochemicals stock (8 mM) was prepared using either DMSO or PBS. Each phytochemical working stock (250 and 500 µM) was prepared in PBS and used for in vitro enzymatic assay. PLpro proteolytic and DUB assay were performed as described previously (Lewis et al., 2022). Briefly, 0.4 ng/μL of PLpro in 30 µL of assay buffer was pre-incubated with the 10 µL of 250 µM phytochemicals for 1 hour. The enzymatic reaction was initiated by adding 10 µL of 250 µM fluorescently labeled substrate. The deubiquitinase assay was initiated using PLpro-specific ubiquitinated substrate incubated for 24 h s at room temperature under dark. Fluorescent reading was taken at 360/40 excitation and 460/40 nm emission using Synergy HT fluorescent plate reader. For dose-dependent studies, compounds were screened from concentrations of 0–100 µM. 10µL of 1% DMSO with enzyme and 50 µM of substrate served as the positive control. Wells with 50 µM of GRL0617 compounds (BPS Biosciences) served as specificity controls. Wells with only 1% DMSO and 50 µM of substrate served as blank. All the values were subtracted from blank values to calculate the percent activity of the enzymes.

The cytotoxic effect of selected phytochemicals were carried out with Vero-E6 cells using PrestoBlue™ Cell Viability kit as described earlier (Lewis et al., 2022). Briefly, Vero-E6 cells (20,000 cells) were seeded overnight in 96 well plates. Then, 100 µL of complete media (EMEM+10% FBS+1%Pen/Strep) was added to refresh the cells along with 50, 100, and 200 µM selected phytochemicals, and incubation was continued for 24, 48, and 72 hs at 37°C respectively. Cytotoxicity detection regent (PrestoBlue solution) was added and incubated for additional 1h at 37°C. The absorbance was taken at 530/25 excitation and 590/35 nm emission using Bio-Tek Synergy HT fluorescent plate reader.

Molecular Operating Environment (MOE) 2020.09 was used to conduct molecular docking in silico studies using the Amber10:EHT forcefield. The crystal structure of SARS-CoV-2 PLpro was retrieved from the protein data bank (www.rcsb.org) with PDB format (ID: 7CMD and was prepared with the MOE QuickPrep application under default settings. The selected phytochemicals were prepared with the program ChemDraw Professional, Version 10, Cambridge Soft.

Integrated Computer-Aided Molecular design computing method MOE was used to dock both phytochemicals with PLpro. All phytochemicals were examined individually and refined with the Triangle Matcher placement method and induced fit protocol. The docked molecules were scored with the GBVI/WSA dG scoring function.

Statistical analysis was carried out using one-way analysis of variance (ANOVA) with Bonferroni’s Multiple Comparison test with 99.9% confidence intervals and represented as the mean ± SEM. Two-way ANOVA was conducted with Dunnett’s post-test to compare the grouped data. p values p < .05 considered statistically significant. Non-linear regression (curve fit) with four variable dose vs. inhibition was performed to calculate the IC₅₀ values. GraphPad Prism (version 8; La Jolla, CA, United States) was used for statistical analysis. Four individual experiments were performed and triplicates are included in each experiment for reproducibility.

Since phytochemicals are known to inhibit viral replication or viral entry, we performed the high throughput screening of a series of phytochemicals, selected based on the literature review, towards SARS-CoV-2 PLpro proteolytic activity (Supplementary Table S2). Our data suggest that nine out of the 53 selected phytochemicals have the potential to inhibit SARS-CoV-2 PLpro proteolytic activity at least by 50% or lower when compared to untreated samples. Figure 1A reveals that only colchicine out of all the tested alkaloids such as chelidonine, evodiamine, lycorine, sophocarpine and tetrandrine, could inhibit PLpro proteolytic activity. Colchicine was able to inhibit around 50% of the proteolytic activity at 200 μM. The four xanthone compounds tested were α, β, and γ-mangostin and mangiferin. Only γ-mangostin and mangiferin exhibited more than 50% inhibitory activity at the higher concentration (200 μM) against SARS-CoV-2 PLpro proteolytic activity (Figure 1B). We observed that the terpenoids such as artemisinin, betulinic acid, glycyrrhizic acid, obacunone, and organosulfur compounds, including diallyl disulfide and diallyl trisulfide, did not show inhibitory activity towards SARS-CoV-2 PLpro proteolytic activity at all the tested concentrations (Figure 1D).

All flavanols (Figure 2A), including catechin, EGCG, and theaflavin, exhibited a significant decrease in proteolytic activity. The inhibitory effect of catechin was 25%, 44%, and 64%, EGCG was 65%, 81%, and 97%, and theaflavin had an average of 94%, 100% at 50, 100, and 200 μM. We observed that none of the tested isoflavone compounds such as daidzein, daidzin, genistein, genistin, and isoflorentin demonstrated an inhibitory effect against PLpro proteolytic activity at the tested doses (Figure 2B). As shown in Figure 2C, flavanones such as myricetin inhibited 96% at as low as 50 μM. Interestingly, quercetin inhibited 50% at 50μM, and its inhibitory effect did not alter at higher concentration of 200 μM. Rutin exhibited 70% inhibitory activity at 50 and 100 μM and 88% at 200 μM. Non-etheless, none of the miscellaneous agents (Figure 3) except oleuropein and rosmarinic acid exhibited significant inhibition of the proteolytic activity of SARS-CoV-2 PLpro (60% and 50% at 200 μM).

Previous data revealed that catechin, EGCG, theaflavin, myricetin, quercetin, rutin, oleuropein, γ-mangostin, and mangiferin inhibited proteolytic activity of PLpro enzyme, and therefore, we performed DUB activity in the presence of these inhibitory phytochemicals. As shown in Figure 4A, flavanols such as catechin inhibited the DUB activity by 30%, 40%, and 50%, whereas, EGCG by 70%, 90%, and 95%, at 50, 100, and 200 μM, respectively. Interestingly, theaflavin exhibited complete inhibition of DUB activity at concentration as low as 50 μM. Flavonones (myricetin, quercetin, and rutin) were selected to test for potential inhibitory activity against SARS-CoV-2 PLpro’s deubiquitinase activity. Figure 4B suggests that myricetin inhibited DUB activity by 60%, 90%, and 100% at 50, 100, and 200 μM, respectively, whereas quercetin and rutin exhibited minimal inhibitory (20%–40%) activity against DUB activity of PLpro at the tested doses. Oleuropein was the only polyphenol compound classified as miscelleanous agent that exhibited partial inhibitory activity (50% at 200 μM) against SARS-CoV-2 PLpro’s DUB activity (Figure 4C). Interestingly, xanthones such as γ -mangostin (50%) and mangiferin exhibited (70%) inhibitory activity towards the proteolytic activity of the PLpro enzyme (Figure 1B), but we observed very minimal DUB inhibitory activity (10%–30%) even at 200 μM concentration (Figure 4B). At present we do not know the reason for the difference in the inhibitory activity of γ-mangostin and mangiferin against proteolytic and DUB activity of SARS CoV-2 PLpro enzyme. The higher inhibitory effect of mangiferin towards proteolytic activity of PLpro might be due to the presence glycosidic unit compared to γ -mangostin, as we observed with Aloin-A and B (29).

Phytochemicals that exhibited more than 50% inhibitory activity against proteolytic and DUB activity of SARS-CoV-2 PLpro enzyme were tested for dose-dependent inhibitory effects to determine the concentration required to inhibit the 50% of enzymatic activity (IC₅₀). Here we represent the IC₅₀ values of the selected phytochemicals that exhibited inhibitory activity against the PLpro enzyme. As shown in the Figure 5, catechin, EGCG, myricetin, mangiferin, rutin, and theaflavin exhibited IC₅₀ values of 14.2, 128.4, 12.1, 43.4, 95.3, and 7.3 μM, respectively towards PLpro proteolytic activity. However, the IC₅₀ values γ-mangostin, oleuropein, and quercetin are ambiguous. We observed that EGCG, mangiferin, myricetin, oleuropein, rutin, and theaflavin have IC₅₀ values of 44.7, 104.3, 29.2, 131.5, 61.7, and 13.2 μM, respectively for DUB activity (Figure 6).

Furthermore, we investigated the cytotoxic effect of EGCG, mangiferin, myricetin, oleuropein, rutin, and theaflavin since they exhibited inhibitory activity for both proteolytic and DUB activity of PLpro. Out of the six phytochemicals tested, four of them did not have cytotoxic effect on Vero-E6 for 24, 48, and 72 h s at 50, 100, and 200 µM, respectively. However, we observed minimal cytotoxic effect (20%–30%) from myricetin and oleuropein after 48 hs (Supplementary Figure S1). We selected Vero-E6 cell line because of its sensitivity towards SARS-CoV-2 infection (Ren et al., 2006; Asrani et al., 2021). Although selected phytochemicals exhibited minimal cytotoxic effects, it is possible that the cytotoxic effects of phytochemicals might be altered in the virus infected cells, therefore it is warranted that the effective concentration of these phytochemicals may change during viral challenge studies.

The structure of SARS-CoV-2 PLpro is divided into four main sub-domains, the N-terminal Ubiquitin-like domain, α-helical thumb domain, β-stranded finger domain and the palm domain. The structure of ubiquitin-specific proteases deubiquitinating enzyme (DUB) shares a low homology (10%) (Osipiuk et al., 2021) with that of SARS-CoV-2 PLpro. The fingers subdomain is made of six β-strands and two α-helices, whereas, the thumb comprises six α-helices and a small β-hairpin. The palm subdomain is comprised of six β-strands. The proteolytic and DUB sites in SARS-CoV-2 PLpro are independent of each other implying two possible activities of PLpro. The interface of palm and thumb subdomains is the location for conventional catalytic triad Cys¹¹¹-His²⁷²-Asp²⁸⁶. In addition to the catalytic triad, three additional residues play an important role in the enzymatic activity of SARS-CoV-2 PLpro: β-turn/loop (Glu²⁶⁶ -Gly²⁷¹) which closes upon substrate and/or inhibitor binding is found adjacent to the active site. The Tyr²⁶⁸ part of the (Glu²⁶⁶ -Gly²⁷¹) plays a critical role in the proteolytic activity of SARS-COV-2 PLpro and the Glu¹⁶⁷ in SARS-COV-2 PLpro plays an important role in ubiquitin core recognition. The mutation of Tyr²⁶⁸ has shown to interfere with the proteolytic activity of SARS-COV-2 PLpro and the mutations of Glu¹⁶⁷causes a significant loss of DUB activity (Osipiuk et al., 2021). Any molecule which forms hydrogen bond with Tyr²⁶⁸ or Glu¹⁶⁷ will interfere with the proteolytic and the DUB activity of the SARS-CoV-2 PLpro, respectively and hence will show inhibitory activity.

The interaction of EGCG, mangiferin, myricetin, oleuropein, theaflavin, and rutin to the ligand site of GRL-0617, in the SARS-COV-2 PLpro (PDBID: 7cmd) was analyzed using MOE software. Molecular docking studies of EGCG, mangiferin, myricetin, oleuropein, theaflavin, and rutin with PLpro showed that myricetin and rutin did not interact with the catalytic site of PLpro enzyme and hence we were not able to obtain any docking studies for them and these molecules are not discussed further in molecular docking studies. The molecular docking studies resulted in 15–20 orientations each for the rest of the ligands used in the docking studies. The analysis of protein-ligand interaction between the SARS-CoV-2 PLpro enzyme and EGCG, and theaflavin showed a strong interaction with Try²⁶⁸ and Gln¹⁶⁷ whereas, oleuropein and mangiferin interacted with Try²⁶⁸ (Figures 7A–D). As mentioned earlier, Glu¹⁶⁷ plays an important role in the deubiquitination of the enzyme molecular modeling predicted that EGCG, and theaflavin significantly impair the DUB activity of SARS-CoV-2 PLpro. The ligand-protein fingerprint scan of mangiferin, EGCG, theaflavin, and oleuropein (Figures 8A–D) also shows that these four ligands show interaction with Glu¹⁶⁷ and Tyr²⁶⁸, however mangiferin and oleuropein showed very weak interaction with Glu¹⁶⁷ as compared to EGCG, and theaflavin which is reflected in their low IC₅₀ values for DUB activity.