*3.3. EGCG Impairs the Attachment of PCV2 to Host Cells*

To investigate which stage of the EGCG antiviral effect could be active during the PCV2 infective process, EGCG was added post-infection and incubated for various time. The results showed that compared with of the EGCG whole process treatment PCV2 infection, a significant inhibitory effect could not be detected with treatment for 4 h or 12 h after PCV2 infection, only a slight effect was observed when the post treatment duration reached 24 h (Figure 3a), indicating that EGCG might work at the early stage of infection course. Thereby, the binding assay was performed, the virions were added to the cell surface at 4 ◦C to achieve the binding event for 1 h, and the results indicated that bound PCV2 particles would decrease significantly in cells with the EGCG treatment measured by qRT-PCR (Figure 3b). The cells attached by the virions could be detected by flow cytometry [17], and its

proportion also dropped from ~70% to ~55% in cells with the EGCG treatment for 1 h (Figure 3c). This tendency was also detectable when the binding time was extended to 4 h to maximize virions binding, and the proportion of PCV2-attached cells reduced from about 85% to about 67% (Figure 3b,c). Taken together, these findings demonstrated that EGCG inhibited the binding of PCV2 to host cells, but exerted no distinct effect on downstream infection.

**Figure 3.** EGCG prevents PCV2 virions binding to host cells. (**a**) Cells were inoculated with PCV2 at MOI = 1.0 with 100 μM EGCG treatment during all the phases, or followed by post-treatment of EGCG for 4 h, 12 h and 24 h, respectively, after which the cells were maintained without EGCG until determined with immunoblotting at 72 hpi. (**b**,**c**) Virus was added to PK-15 cells at MOI = 5.0 in the presence of 100 μM EGCG. After the binding for 1 h or 4 h at 4 ◦C. the genome copy numbers of the attached virions were determined using the qPCR method (**b**); PCV2-attached cells were measured by flow cytometry (**c**). Mean ± SD of three independent experiments were shown and Student's *t*-test was conducted for statistical analysis (\* *p* < 0.05; \*\* *p* < 0.01).

#### *3.4. EGCG Competitively Inhibits the Interaction between Capsid and Heparan Sulfate*

PCV2 virions attached to host cells via direct interaction between viral capsid protein and cell surface heparan sulfate [10]. Therefore, we hypothesized that EGCG might also bind to capsid, and competitively block the interaction between PCV2 capsid and HS. The microscale thermophoresis assay (MST) was utilized to evaluate the affinity between compounds and protein by measuring the dissociation constant (Kd). The Kd between dcapsid and EGCG was 98.03 ± 4.76 μM, indicating that EGCG could indeed bind to capsid (Figure 4a), and capsid's affinity to EGCG was comparable to the affinity for heparin (an HS analog similar in structure to the sulfated domain of HS), or even a little stronger than the latter (Kd = 120.67 ± 4.73 μM) (Figure 4b). However, the EC, which did not exhibit the inhibitory activity against PCV2 infection, could also bind to capsid with a 10-fold lower affinity than EGCG (Kd = 951.33 ± 6.81 μM) (Figure 4c). These findings implied that the interaction between PCV2 capsid and HS might be hindered by EGCG. This was confirmed by heparin column chromatography assay. Briefly, the dcapsid protein was pre-loaded onto a heparin-Sepharose column, then eluted by the various competitor reagents. The results showed that approximately 40% protein fraction (42.33 ± 3.52%) could be eluted by the EGCG at a concentration of 0.05 mg/mL, similar to the eluted level of heparin (36.00 ± 6.25%), while in case of EC only 10.33 ± 4.50% protein sample was eluted from the heparin column (Figure 4d), which was significantly reduced compared to the level of EGCG or heparin competitor reagents. The eluted fraction was further increased when the concentration of competitor reached 0.5 mg/mL. In that case, the heparin and EGCG could elute 89.33 ± 4.16% and 90.00 ± 5.29% of total protein, respectively (Figure 4e), and the difference between then was still not significant. In summary, EGCG could directly interact with PCV2 capsid, competitively inhibiting the binding of latter to HS.

#### *3.5. Identification of Key Amino Acids in PCV2 Capsid Contributing to the Interaction with EGCG*

To further investigate the detailed structural character of the interaction between EGCG and capsid, the flexible docking model of the complex was established with AutoDock software 4.2 (Olson Lab, http://autodock.scripps.edu). The crystal structure of PCV2 dcapsid was obtained as PDB accession ID:3R0R and the whole domain of protein was docked with ligand. Several docking model structures were obtained, and the one with the lowest interface energy was adopted for further analysis, which showed that each EGCG molecule could be attached to a capsid monomer at a distinct pocket, which was completely exposed to the surface of PCV2 particles according to the revealed structure of the PCV2 VLPs [18], and the pocket was formed by some critical amino acids (Figure 5a), among which ARG-51, GLY 54, ASP-70, ARG-73, ASN-75, and ASP-78 were predicted to engage in hydrogen bonds with EGCG (Figure 5b). We established an amino acid sequence alignment to assess the conservation of these amino acid. It was demonstrated that all of these 6 amino acids were highly conserved in PCV2 ranging from PCV2a to PCV2f except for ASN-75 (Figure 5c), suggesting their critical importance for the interaction between capsid and EGCG molecule in the various subtypes of PCV2.

**Figure 4.** EGCG competitively inhibited capsid binding to heparan sulfate. (**a**–**c**) EGCG could interact with capsid or with heparin. The dcapsid protein was labeled with the Cy5 fluorophore and incubated with two-fold serial dilutions of EGCG or heprarin (10–50 nM), and then the Kd values of the dcapsid were measured for EGCG (**a**), heparin (**b**), and EC (**c**) were measured by MST. (**d**,**e**) EGCG eluted dcapsid protein bound to a heparin column with similar efficiency as heparin. Heparin-Sepharose HP column was pre-loaded with dcapsid and eluted with 0.05 (**d**) or 0.5 (**e**) mg/mL soluble competitor reagents (heparin, EGCG and EC). The eluted samples were analyzed using SDS-PAGE and immunoblotting. Three independent experiments were conducted. The results are presented as Mean ± SD and compared using Student's *t*-test (n.s., not significant; \* *p* < 0.05; \*\* *p* < 0.01).

**Figure 5.** Molecular docking model of interaction between EGCG and PCV2 Capsid. (**a**) Structural diagram of EGCG-capsid complex based on the flexible docking model. The amino acids involved in the formation of the pocket accommodating EGCG molecule (green) were labeled blue in the dcapsid (light grey) in the zoom graph. (**b**) The putative amino acids (yellow) in dcapsid that could establish hydrogen bonds with EGCG (green) based on the docking model. The predicted hydrogen bonds are shown as a yellow dotted line. (**c**) The conservation of the key amino acids in the capsid to maintain the interaction among PCV2 subtypes. All the amino acid sequence data were obtained from GenBank. Among the predicted amino acids which might form hydrogen bonds with EGCG, the ones with high conservation (ARG-51, GLY 54, ASP-70, ARG-73 and ASP-78) are highlighted with a red frame, and the rest (ASN-75) are labeled with a green frame.

To test this hypothesis, and to identify the key amino acids for the interaction, we made a single-point mutation of these amino acids to alanine (A) residue and measured the affinity of each mutant to EGCG, the results of MST showed that the Kd of different mutants and EGCG could fluctuate significantly compared to the wild type (WT) capsid; among all the mutants, the D70A and D78A showed the loweest affinity to EGCG with Kd values of 863.67 ± 7.23 μM and 1061.33 ± 8.08 μM, respectively (Figure 6c,f), almost 10 times lower than the Kd of WT to EGCG (Kd = 98.03 ± 4.76 μM) (Figure 4a). Additionally, mutants aimed at arginine (R) residues also exhibited comparable reduction in affinity to EGCG, the Kd values of R51A and R73A were 591.00 ± 5.57 μM and 971.00 ± 9.54 μM, respectively (Figure 6a,d). On the contrary, the affinity of G54A and N75A mutants did not change significantly (Figure 6b,e). These results indicated that ARG-51, ARG-73, ASP-70 and ASP-78 were critical for the interaction of EGCG with the PCV2 capsid. To further assess the effect of these amino acids on the binding properties of capsid to HS, the heparin column chromatography assay of different mutants was performed and it showed that the eluted level of R51A, D70A, R73A and D78A mutants was remarkably reduced, with 0.5 mg/mL EGCG, especially in the case of D70A and D78A mutants, which is consistent with the results of Kd determination. Meanwhile, the mutation targeted to arginine residues (R51A and R73A) displayed the most significant eluted reduction with 0.5 mg/mL heparin (Figure 6g). Taken together, these results indicated that the ARG-51, ARG-73, ASP-70 and ASP-78 of PCV2 capsid protein were crucial for its interaction with EGCG, and the two arginine residues were also essential for binding with cell surface HS.

**Figure 6.** Verification of key amino acids involved in the interaction of EGCG with capsid. (**a**–**f**) The affinity impairment of capsid to EGCG induced by the key amino acid mutation. The Cy5 fluorophore labeled dcapsid mutants were incubated with two-fold serially dilutions of EGCG at room temperature for 15 min before being tested with MST, and the Kd values between EGCG and capsid mutants R51A (**a**), G54A (**b**), D70A (**c**), R73A (**d**), N75A (**e**) and D78A (**f**) were measured, respectively. (**g**) The eluted level of capsid mutants bound to the heparin column. Heparin column preloaded with dcapsid mutants was eluted with 0.5 mg/mL soluble heparin or EGCG. The eluted level of each mutants was analyzed using immunoblotting. Mean ± SD of three independent experiments were statistically analyzed using Student's *t*-test (\* *p* < 0.05; \*\* *p* < 0.01).
