The aerial parts of R. retinorrhoea Steud, ex Olive, locally known as ‘Sumac/Heishar’ were collected from the southern region of Saudi Arabia in March 2009. The plant material was authenticated (voucher specimen no. 15371) by Dr. Mohammad Yusuf, a plant taxonomist at College of Pharmacy, King Saud University Riyadh.

The ethanolic extract of the aerial parts of the plant was further fractionated in dichloromethane following the isolation of several known or new compounds belonging to different classes of phytochemicals. The majority of these were either obtained in very low quantity or were unsuitable candidate for testing against HBV. Based on available literature on their structural similarity and non-cytotoxic flavonoids reported against other viruses, two compounds were finally selected, namely C251 and C253, as previously described (11,18). For structure elucidation, ¹H and ¹³C, and 2D nuclear magnetic resonance (NMR) spectroscopy of C251 and C253 were recorded at 700 and 175 MHz, respectively, on the Bruker Avance spectrometer (Bruker BioSpin GmbH)equipped with a 5-mm cryoprobe, in deuterated DMSO, using standard pulse programs. All organic solvents were purchased from Sigma-Aldrich (Merck KGaA).

HepG2.2.15 cells, which were established by stably transfecting human hepatoma HepG2 cells with the full genome of HBV, were generously provided by Dr S. Jameel (Virology Group, ICGEB, New Delhi, India). HepG2.2.15 cells were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific Inc.) supplemented with 10% bovine serum albumin (Gibco; Thermo Fisher Scientific, Inc.) and 1X penicillin-streptomycin solution (HyClone; Cytiva) at 37°Cin an incubator with 5% CO₂. Prior to treatment, cells at a density of 0.5×10⁵/100 µl/well were grown overnight in a 96-well plate (Corning, Inc.). LAM triphosphate (or 3TC) and quercetin (QRC; both from Sigma-Aldrich; Merck KGaA) served as positive controls, as previously described (26–30). For consistency and reproducibility, the assays were performed in duplicate.

Although the Rhus spp., including R. tripartite, are known to be non-toxic (2,3), the optimal non-cytotoxic doses of the isolated compounds were first assessed in HepG2.2.15 cells. Briefly, compounds dissolved in DMSO (Sigma-Aldrich; Merck KGaA) were prepared in RPMI-1640 to produce four doses with 6.25, 12.5, 25 and 50 µg/ml of each compound. Following HepG2.2.15 cell incubation overnight, the culture medium was replaced with treatment media (in triplicate), including negative control (0.1% DMSO) medium, followed by incubation at 37°C for 72 h. The cells were periodically monitored directly under an inverted microscope. Subsequently, cells were treated with MTT solution (TACS MTT Cell Proliferation Assay Kit; Sigma-Aldrich; Merck KGaA), according to the manufacturer's instructions. The optical density at a wavelength of 570 nm was measured using the Elx800 microplate reader (BioTek Instruments, Inc.). The results were analyzed in Excel 2010 (Microsoft Corp.) and presented in relation to the negative control.

Initially, HBsAgs were dose-dependently inhibited (6.25, 12.5 and 25 µg/ml) by the isolated compounds to determine the maximally active concentration. HepG2.2.15 cells were cultured overnight and the culture medium was then replaced with treatment medium, including negative and positive control media, and incubated for an additional two days (a single time-point). Following the determination of the maximal dose, the time-dependent inhibition of HBsAg by the compounds was then assessed. HepG2.2.15 cells were treated with 25 µg/ml SEK or VEL and the corresponding controls, and incubated for several days. The culture was directly monitored every day under microscope and it was replenished with treatment media every alternate day. The culture supernatants collected and clarified (150 × g; 5 min; 22°C) on day 1, 3 and 5 were quantitatively analyzed for HBsAgs using the diagnostic HBsAg ELISA kit (cat. no. 72348; Monolisa HBs Ag ULTRA assay; Bio-Rad Laboratories Inc.) in a microplate, according to the manufacturer's protocol. The optical density of the samples at a wavelength of 450 nm was measured and the results were then analyzed in relation to the negative control (Excel software 2010; Microsoft Corp.) and compared with the positive control.

The treated culture supernatants collected and clarified (150 × g; 5 min; 22°C) on day 1, 3 and day 5 were also quantitatively analyzed for HBeAg production using a HBeAg ELISA kit (cat. no. KAPG4BNE3; HBeAg/Anti-HBe Elisa Kit; DIAsource ImmunoAssays SA) according to the manufacturer's instructions. The recorded optical density (λ=450 nm) of the samples were analyzed in relation to the negative control (Excel software 2010; Microsoft Corp.), and compared with the positive control. All samples were tested in triplicate and the experiment was repeated for two times.

Based on their promising anti-HBV activities in cultured cells, VEL and SAK were further subjected to virtual structure-activity analysis to uncover the potential mechanisms underlying their inhibitory effects. The viral POL and CORE proteins served as target drugs, while their respective inhibitor molecules LAM and heteroaryldihydropyrimidine (HAP) acted as standard ligands (28,34). Notably, in the absence of crystallographic data or 3D model for HBV POL, an in-house constructed POL structure was used, as previously described (28). The available 3D structures of HBV CORE (PDB code, 5E0I; http://www.rcsb.org/) and the ligands LAM, VEL and SAK (https://pubchem.ncbi.nlm.nih.gov/) were retrieved. The target proteins were prepared by removing any solvent molecules or co-crystallized ligands and via adding hydrogen atoms and Kollman charges (28). For docking, the published catalytic or active residues of LAM (28) and CORE (34) were confirmed using the SEINA program (35). The two target proteins were prepared and energy-minimized in Maestro software (36). The ligand-target interactions were visualized using the 2D (Maestro) and 3D (UCSF ChimeraX) modes (37). The ligands were docked onto their corresponding target binding pocket or active site using AutoDock Vina 1.2.3 software (38,39).

All data were analyzed using SPSS 17.0 (SPSS Inc.). Data are expressed as the mean ± SEM of three independent experiments. The results were compared with the negative control group using one-way ANOVA followed by Dunnett's post hoc test. P<0.05 was considered to indicate a statistically significant difference.

The two isolated compounds, 251 and 253, which were subjected to ¹H and ¹³C, and 2D NMR analyses (Table I; Fig. S1, Fig. S2, Fig. S3, Fig. S4, Fig. S5, Fig. S6, Fig. S7, Fig. S8, Fig. S9, Fig. S10, Fig. S11, Fig. S12), were identified as the structurally-related flavonoids SEK (4′,5-dihydroxy-7-methoxyflavanone) and VEL (5,4′-dihydroxy-7,3′-dimethoxyflavone), respectively (Fig. 1; upper panel).

MTT assay demonstrated that the flavonoids SAK and VEL did not show any hepatotoxicity in cells cultured for 72 h even at the maximal dose tested (Fig. S13). This was consistent with the microscopic observation of cells, as treated cells exhibited intact morphology as with negative cells. Therefore, the 50% cytotoxicity concentration (CC₅₀) values could not be determined.

Optimal dose assessment revealed that SAK and VEL at a dose of 25 µg/ml showed the maximal inhibition of HBsAg on day 2 (Fig. 1; lower panel). However, at a dose of 50 µg/ml, no significant increase in the inhibitory activities of VEL and SAK was observed (data not shown). Therefore, a dose of 25 µg/ml was selected as the optimally active dose for the time-course study. Among the three selected time-points (day 1, 3 and 5), the maximal inhibition rate of SAK and VEL on HBsAg synthesis was ~58.8 and ~56.4%, respectively, on day 5 (Fig. 2). In comparison, LAM and QRC inhibited HBsAg by ~86.4 and ~84.5%, respectively. Notably, since cell treatment with flavonoids at the maximal dose also enhanced cell proliferation and overgrowth-mediated apoptosis (data not shown), the assay was carried out at day 5.

Synthesis of HBeAg is a serological gold marker of HBV DNA replication in patients with HBV (21). Therefore, the inhibitory effect of SEK and VEL (25 µg/ml, each) on HBeAg expression in treated HepG2.2.15 cells was further analyzed. Of the analyzed time-points (day 1, 3 and 5), the maximal inhibition rate in HBeAg production was ~55.5% by SAK and ~52.4% by VEL on day 5 (Fig. 3). Comparatively, LAM and QRC suppressed HBeAg generation by ~64 and ~62%, respectively. As aforementioned, since flavonoids could promote cell overgrowth and apoptotic death, the assay was performed on day 5.

The two isolated anti-HBV active flavonoids, VEL and SAK, were virtually docked into the binding pocket of POL and CORE proteins. The results revealed good re-alignments of the ligands. Docking of LAM and HAP generated complexes with good docking energies and orientations, thus indicating a good docking protocol (Fig. 4; Table II). Owing to their common flavonoid structure, VEL and SEK acquired relatively similar alignment and orientations inside the binding site of POL (Fig. 5; Table II) and CORE (Fig. 6; Table II). In addition, both flavonoids shared interactions with key active residues of the target proteins. Notably, similar to the negative charges of the triphosphate group, which significantly interacted with positive charged residues at Arg23 and Lys14 in the LAM-POL complex, and in coordination with Mg⁺² (Fig. 4; upper panel), the oxygen atoms of VEL and SEK showed the same interactions in the VEL-POL and SAK-POL complexes, respectively (Fig. 5). The VEL-POL complex was further stabilized by π-cation with Lys14 and π-stacking with Phe70 (Fig. 5; upper panel). In addition to the POL catalytic ‘Tyr-Met-Asp-Asp’ motif residues, other surrounding residues, such as those at Ser67 and Ala68, could also be involved in the stability of the VEL-POL and SAK-POL complexes. Nonetheless, LAM (standard) showed a more potent binding affinity compared with both VEL and SAK, which could be due to its more efficient electrostatic interactions (Table II).

Regarding docking with the HBV-CORE protein, both VEL and SAK formed complexes with very close poses (Fig. 6), while VEL displayed a higher binding affinity compared with SEK (Table II). The two ligands shared H-bonding with Ala132 and π-stacking with Trp102. Notably, HAP (standard) could also interact with Trp102 through H-bonds. Other surrounding residues, such as those at Val124 and Ser106 could also contribute to the VEL-HAP and SAK-HAP complex stabilities. Taken together, the molecular docking data suggested that the activities of VEL and SEK against HBV could be mediated by the inhibition of the viral POL and CORE proteins.