*2.2. Molecular Docking Studies*

To investigate the individual effects of the positively identified components of the tested *A. viridiflora* sample, we used molecular docking simulations. Starting with the most active compound, quercetin 3-(6"-ferulylglucoside), all compounds are listed in Table 2 in order of their binding affinity for the S-glycoprotein receptor (PDB ID: 7BZ5). This target's binding pocket is depicted in Figure S3, and its constituent residues are given in Table S1 (Supplementary Materials).

**Table 2.** Molecular docking simulation results of *A. viridiflora* constituents and positive controls against wild type S-glycoprotein target (PDB ID: 7BZ5).


\* In the interacting residues column residues involved in hydrogen bonding are denoted in bold font with the interaction distances enclosed in brackets. \*\* Positive control compounds are bordered with frame.

These findings demonstrated that the observed inhibitory activity was a result of contributions from both polyphenolic groups. Quercetin 3-(6"-ferulylglucoside) demonstrated

the highest binding affinity (−8.035 kcal/mol). The most favorable binding orientation of this compound is presented in Figure 2. Preliminary 12.50 ns molecular dynamic simulation results for the quercetin 3-(6"-ferulylglucoside)-S-glycoprotein complex presented in Figures S4–S7 confirm the stability of the observed system. Radius gyration trajectory (Figure S5) deviations between 18.40 Å and 18.85 Å indicate a stable secondary protein structure with a high complexing potential for the studied ligand. Observing the root mean square deviation (RMSD) trajectory (Figure S6) reveals that after 2 ns, the complex reaches a stable state. The complex exhibits simulation-based deviations after that point that do not compromise the system's stability because oscillations between the mean and maximum value do not exceed 2.5 Å.

**Figure 2.** Quercetin 3-(6"-ferulylglucoside) interactions with S-glycoprotein (wild); (**a**) the most favorable binding pose of the compound; (**b**) 2D illustration of interaction types between the compound and target residues.

Two compounds from the ellagitannin class, tellimagrandin I and II, showed only a somewhat lower affinity for interacting with S-glycoprotein, with binding affinity energies of −8.022 and −7.955 kcal/mol, respectively. Additionally, every compound that was tested produced complexes with the target protein that were stabilized by regular hydrogen bonds on distances lower than 2 Å. One of the key interacting residues, Gln160, was previously identified as one of 23 virus residues that participate in stable hydrogen bonds, which let the virus bind to the ACE2 receptor. Most of the tested *A. viridiflora* polyphenols showed interaction with this residue, and the second-most potent compound, tellimagrandin I, interacted with it via an H bond at a distance of 2.82 Å (Figure 3) [22].

When complexing with the target S-glycoprotein, the two positive controls utilized in this investigation showed a very slight energy difference, with umifenovir forming a more stable complex (−7.384 kcal/mol) than quercetin (−7.189 kcal/mol).

According to a recent study, umifenovir inhibits the internalization of SARS-CoV-2 and its variants by directly binding to the S-glycoprotein {Shuster, 2021 #41}. However, 12 polyphenolic *Alchemilla* constituents exhibited more affinity for S-glycoprotein as a target than umifenovir, indicating more effective infection prevention (Table 2).

**Figure 3.** Tellimagrandin I interactions with S-glycoprotein (wild): (**a**) the most favorable binding pose of the compound; (**b**) 2D illustration of interaction types between the compound and target residues.

In addition to docking against wild strain-specific S-glycoprotein more docking simulations were performed on the V483A, N501Y417NE484K, N501Y, N439K, L452RT478K, K417N, G476S, F456L, and E484K strains to evaluate the stability of the observed inhibitory action on other viral strains resulting from mutations. The binding energy fluctuations curve for S-glycoprotein revealed that different drugs had varying binding affinities. In particular, the mutated strains identified in South Africa lineage B.1.351 (also known as 501Y.V2 variant) and P.1 lineage (a descendant of B.1.1.28) identified in December 2020 (in Manaus, Amazonas State, North Brazil) showed increased affinity for quercetin and tellimagrandin II, compounds with binding affinity on first and third place for wild type virus S-glycoprotein (Figure 4) [23,24]. The increased binding affinity seen in Figure 4 for the positive control umifenovir is also consistent with the findings of Shuster et al. (2021) about its maintained activity against new virus strains [25].

Although variations in binding affinity were observed for all identified constituents of *A. viridiflora* extract overall conclusion is that the range of the complex energies between −6.0 and −9.0 kcal/mol for all compounds proves they maintained significant inhibitory potential regardless of mutation changes in S-glycoprotein. According to these results, the tested extract should retain its efficacy against other virus strains, which is necessary given the significant mutational potential identified for the SARS-CoV-2 virus.

Molecular docking simulation results of *A. viridiflora* constituents and positive control against the NRP1 target are presented in Table 3. Seven compounds displayed a greater affinity for NRP1 than the positive control brevifolin carboxylic acid, showing that other polyphenols also significantly contribute to the inhibitory activity. Figure 5 presents the binding position and active site of pentagalloylglucose (compound with the highest binding activity) interaction with NRP1.

**Figure 4.** Binding energy (kcal/mol) curves for *A. viridiflora* constituents and positive controls against all tested S-glycoprotein structural variants.

**Table 3.** Molecular docking simulation results of *A. viridiflora* constituents and positive control against NRP1 target (PDB ID: 2QQI).





\* In the interacting residues column residues involved in hydrogen bonding are denoted in bold font with the interaction distances enclosed in brackets. \*\* 3D structure atoms color legend; red-oxygen, grey-carbon, light grey-hydrogen. \*\*\* positive control compound is bordered with frame.

**Figure 5.** Pentagalloylglucose binding pose and site in complex with NRP1.

In Jin et al. study from 2022, pentagalloylglucose, the molecule listed first for its affinity to NRP1, was already recognized as a plant dietary polyphenol with substantial in vitro inhibitory effect against SARS-CoV-2 infection in Vero cells. This study confirmed that part of this inhibitory potential could be attributed to the inhibition of SARS-CoV-2 main- and RNA-dependent RNA-polymerase. Additionally, researchers found efficacy against the SARS-CoV and MERS-CoV viruses, indicating pentagalloylglucose has broad-spectrum anticoronaviral potential [26].

Other compounds with energy values below −6.5 kcal/mol, apart from pedunculagin, were from the flavonoid class, indicating that the flavonoid subclass of *A. viridiflora* polyphenols may contribute more significantly to inhibition activity when the virus primarily uses the NRP1 receptor for its internalization. Flavonoid potential for interaction with NRP1 has been already discussed in Yasmin et al. (2017) study where they proposed quercetin and diosmin as small ligands with promising potential for targeting NRP1 receptors with implications for therapeutic benefits in neurology and oncology [27]. Multiple interaction types contribute to the stabilization of ligand-target complexes, and it is noteworthy that all tested ligands were stabilized by at least one conventional hydrogen bond at a distance closer to 3 Å. The pharmacophore model highlighted ligand interactions with NRP1 residues: Tyr353, Thr349, Tyr297, Asn300, and Ser298 as critical for inhibitory activity in the Perez-Miller et al. (2020) investigation that found and confirmed inhibitors of the

interaction between NRP1 and SARS-CoV-2 S-glycoprotein [28]. All these significant NRP1 residue interactions are also identified in the most favorable binding poses of *A. viridiflora* polyphenol constituents. In addition to H bonds with Tyr353, pedunculagin interacts with the same type of interaction with Asp320, a crucial residue for interaction with vascular endothelial growth factor C-terminal arginine [29]. In addition to conventional H-bond interactions, ligands, particularly those with a flavonoid structure, were stabilized by hydrophobic interactions. These interactions included ring B from the flavonoid structure and the amino acid residues Tyr297, Thr316, Tyr353 and Trp411 from the binding pocket presented in Figure S7.
