Previously, we reported that HBV replication increases expression of endogenous SIRT2, leading to deacetylation of α-tubulin (Piracha et al., 2018, 2020). To confirm the results of these previous reports (Piracha et al., 2018, 2020), we transfected Huh7 and HepG2 cells with the 1.3mer HBV wild-type (WT) ( Figures 1A, B ). HBV replicative intermediate (RI) DNA, including partially double-stranded relaxed circular relaxed circular (RC) and double-stranded linear (DL) DNA, were detected ( Figures 1A, B , bottom panels, lanes 2). As expected, endogenous expression of SIRT2 was higher in HBV-transfected cells than in mock-transfected cells (Piracha et al., 2018, 2020) ( Figures 1A, B ; third panels, lane 1 vs. lane 2). Consistent with this, α-tubulin was deacetylated in HBV-transfected cells (Piracha et al., 2018, 2020) ( Figures 1A, B , top panels, lane 1 vs. 2).

Previously, we reported that AGK2, a selective SIRT2 inhibitor (Outeiro et al., 2007), inhibits HBV replication by downregulating AKT/GSK-3β/β-catenin signaling in HBV-transfected Huh7 and HepG2 cells (Piracha et al., 2018). This inhibitory activity of AGK2, particularly its effect on α-tubulin deacetylation, was validated by (Outeiro et al., 2007). Before initiating the current experiments, we performed MTT assays to evaluate AGK2 cytotoxicity in Huh7, HepG2, HepAD38, and HepG2.2.15 cells, confirming CC₅₀ values above 100 µM for all cell lines (data not shown).This inhibitory activity of AGK2, particularly its effect on α-tubulin deacetylation, was validated by (Outeiro et al., 2007). Consistent with previous reports (Piracha et al., 2018; Yu et al., 2018), an MTT assay revealed that AGK2 was not cytotoxic at the concentrations used to treat Huh7 and HepG2 cells ( Supplementary Figure S1A ). However, during this study, slight morphological changes were observed in Huh7 cells treated with 10 µM AGK2; thus, we reduced the concentration to 2.5 µM for this cell line, which still effectively inhibited HBV replication. Next, we transfected the cells with 1.3mer HBV WT for 72 h and treated them simultaneously with AGK2. Consistent with the previous and above-mentioned results, we found that 1.3 mer HBV WT-transfected Huh7 and HepG2 cells harbored deacetylated α-tubulin and showed increased expression of SIRT2 ( Figures 2A, B ; top and third panels, lane 1 vs. 2); however, there was a significant reduction in SIRT2 expression upon treatment with AGK2 ( Figures 2A, B ; lane 2 vs. 3). Expression of HBc protein, core particle formation, and HBV DNA synthesis were reduced significantly by AGK2 treatment ( Figures 2A, B ; fourth to seventh panels, lane 2 vs. 3). A HBsAg ELISA conducted in Huh7 cells at 72 h post-transfection with HBV WT revealed that 2.5 μM AGK2 reduced HBsAg secretion significantly, while 1 μM entecavir did not ( Figure 2C , upper panel). Unlike secretion of HBsAg, secretion of HBeAg secretion by Huh7 cells was not affected by AGK2 or entecavir ( Figure 2C , lower panel). At 72 h post-transfection of HepG2 cells, the HBsAg and HBeAg ELISAs revealed that 10 μM AGK2 reduced HBsAg and HBeAg secretion significantly, whereas 1 μM entecavir did not ( Figure 2D ).

Since HBV DNA synthesis decreased in AGK2-treated HBV replicating cells ( Figures 2A, B ), we hypothesized that the transcriptional activity of HBV would also decrease. Therefore, we performed a luciferase reporter assay in AGK2-treated cells to examine HBV promoter activity ( Figures 3A, B ). The results showed that HBV transcriptional activity, including that of the enhancer II/core promoter (EnhII/Cp), the preS1 promoter, and the enhancer I/X promoter (EnhI/Xp), in Huh7 cells decreased upon treatment with AGK2 ( Figure 3A ). By contrast, there was no marked change in preS2 promoter activity following AGK2 treatment ( Figure 3A ). In AGK2-treated HepG2 cells, the activity of the EnhII/Cp, preS1, and preS2 promoters decreased ( Figure 3B ). The promoter activity of EnhI/Xp decreased, but the decrease was not statistically significant ( Figure 3B ). Consistent with results of the luciferase assays ( Figures 3A, B ), northern blotting revealed that the level of HBV pgRNA and subgenomic S mRNAs decreased in AGK2-treated, HBV-transfected Huh7 and HepG2 cells ( Figures 3C, D , lane 2 vs. 3), suggesting that inhibition of SIRT2 by AGK2 represses HBV transcription, thereby decreasing HBV replication. The level of HBx mRNA also decreased, but not significantly ( Figures 3C, D , lane 2 vs. 3).

To investigate the effect of AGK2 in the HBV infection system, we established HepG2-hNTCP-C9 cells ( Figure 4A , third panel, lanes 2–4). Briefly, HepG2‐hNTCP‐C9 stable cells were infected with 2 × 10² GEq of HBV ( Figure 4A , lanes 3–4) and treated with 10 μM AGK2 (lane 4), as described previously (Kim et al., 2022). Consistent with the above findings, and with previous reports (Piracha et al., 2018; Yu et al., 2018), we found that AGK2 was not cytotoxic in HepG2-NTCP‐C9 cells ( Supplementary Figure S1B ). In accordance with the transfection experiments ( Figures 1 , 2 ), as well as previous infection experiments (Piracha et al., 2018, 2020), expression of endogenous SIRT2 was higher in HBV-infected cells than in non-infected cells ( Figure 4A ; fourth panel, lane 2 vs. 3). Correspondingly, α-tubulin was deacetylated in HBV-infected cells ( Figure 4A ; top panel, lane 2 vs. 3). Consistent with the results of the transfection experiments ( Figure 2 ) (Piracha et al., 2018), AGK2 decreased endogenous SIRT2 levels significantly, and increased acetylation of α-tubulin ( Figure 4A ; top and fourth panels, lane 3 vs. 4). Expression of HBc protein, core particle formation, and HBV DNA synthesis were all reduced markedly by AGK2 treatment ( Figure 4A ; fifth to bottom panels, lane 3 vs. 4). Additionally, there was a significant reduction in the HBV cccDNA level in HBV-infected cells after AGK2 treatment ( Figure 4B , lane 3 vs. 4, white arrowhead). Since AGK2 decreased HBV DNA synthesis and cccDNA levels in HBV-infected cells ( Figures 4A, B ), we next examined HBV RNA levels by northern blot analysis of AGK2-treated, HBV-infected HepG2-hNTCP-C9 cells ( Figure 4C ). The levels of HBV pgRNA and subgenomic S mRNAs fell significantly in AGK2-treated cells, but not in cells treated with 1µM entecavir ( Figure 4C , lane 3 vs. lane 4 vs. lane 5).

In addition, HBsAg and HBeAg ELISAs revealed that secretion of HBsAg and HBeAg fell markedly after AGK2 treatment ( Figure 4D ). Unlike in transfected cells ( Figures 2C, D ), secretion of HBsAg and HBeAg by HBV-infected HepG2-hNTCP-C9 cells treated with 1 μM entecavir fell significantly, although to a much lesser extent than in AGK2-treated cells ( Figure 4D ). Collectively, these results indicate that AGK2 downregulates HBV cccDNA, leading to a decrease in syntheses of HBV RNAs and RI DNAs, and in secretion of HBsAg and HBeAg.

We also infected PXB^® cells with HBV and then treated them with AGK2 ( Figure 5 ). Briefly, PXB^® cells were infected with 2 × 10³ GEq of HBV ( Figure 5 ), as described previously (Kim et al., 2022; Tang et al., 2024). As shown for AGK2-treated, HBV-infected HepG2‐hNTCP‐C9 cells, AGK2 treatment of infected PXB^® cells decreased HBc protein and core particle formation significantly ( Figure 5A ). Also, secretion of HBsAg and HBeAg fell markedly after AGK2 treatment; however, entecavir did not reduce HBsAg and HBeAg secretion significantly at 5 days or 7 days post-infection ( Figure 5B ).

Next, we sought to explore the mechanism by which AGK2 inhibits viral transcription; in this context, we wanted to focus on mechanisms other than downregulation of the AKT/GSK-3β/β-catenin signaling pathway. Given that AGK2 suppresses transcriptional activity and reduces HBV RNA levels ( Figures 3 , 4C ), and the fact that overexpressed SIRT2.5 is recruited to cccDNA along with repressive HKMTs that deposit epigenetic repressive markers to suppress HBV transcription (Piracha et al., 2020), one possibility is that AGK2 inhibits viral transcription by altering the chromatin structure of cccDNA. To investigate this, we conducted a ChIP assay using HBV-infected HepG2-hNTCP-C9 cells to examine the association between AGK2 and cccDNA chromatin ( Figure 6 ). Briefly, after infecting HepG2-hNTCP-C9 cells with 2 × 10² GEq/cell of HBV and treating them with AGK2, we immunoprecipitated the chromatin using control IgG or specific antibodies. The immunoprecipitated pellet was subjected to semiquantitative PCR analysis ( Figure 6 ) using selective cccDNA primers targeting the DR1 to DR2 region, initially designed by Werle-Lapostolle et al. (2004). These primers have subsequently been widely used for cccDNA PCR (Pollicino et al., 2006; Belloni et al., 2009, 2012), including our previously published work (Piracha et al., 2020). To begin with, constant amounts of an actin gene fragment (250 bp) were amplified ( Figure 6 , bottom), confirming consistent extraction of chromatin DNA from both mock- and HBV-infected cells. Amplification of input DNA revealed that cells treated with AGK2 had the lowest levels of cccDNA ( Figure 6 , fifteenth panel, lane 2 vs. 3), corroborating the findings shown in Figure 4B . AGK2-treated cells showed lower recruitment of SIRT2 to cccDNA; the levels were roughly comparable with the level of input DNA ( Figure 6 , first panel, lane 2 vs. 3). In accordance with a previous report (Belloni et al., 2012; Piracha et al., 2020), we found that SIRT1 was recruited to cccDNA at levels approximately comparable with those of input DNA ( Figure 6 , second panel). HDAC6, an exclusively cytoplasmic protein (Hubbert et al., 2002; Piracha et al., 2020) that cannot be recruited to cccDNA, was utilized as a negative control ( Figure 6 , third panel). Even though levels of endogenous cytoplasmic SIRT2 were reduced in AGK2-treated cells ( Figures 2A and 4A ), recruitment of endogenous SIRT2 to the nucleus was approximately comparable with that of input cccDNA and SIRT1 ( Figure 6 , first and second panels, lane 2 vs. 3).

Post-translational modifications such as methylation and acetylation on H3 and H4 play a role in altering chromatin structure and regulating transcription of HBV cccDNA (Pollicino et al., 2006; Belloni et al., 2009; Levrero et al., 2009; Belloni et al., 2012). Heterochromatin-associated histone lysine modifications such as H4K20me1, H3K27me3, and H3K9me3 can induce chromatin condensation and transcriptional repression (Levrero et al., 2009; Piracha et al., 2020; Yang et al., 2020a; Andrisani, 2021). To assess whether AGK2 treatment correlates with a transcriptionally inactive cccDNA chromatin structure, we performed ChIP assays to determine deposition of repressive histone lysine methylations such as H4K20me1, H3K27me3, and H3K9me3 in AGK2-treated HBV-infected cells. Our data demonstrated that methylation of H4K20, H3K27, and H3K9 on the cccDNA minichromosome was higher in AGK2-treated cells than in control cells ( Figure 6 , fourth, sixth, and eighth panels, lane 2 vs. 3), indicating transcriptional repression upon AGK2 treatment. Overall, this finding demonstrates that AGK2 induces epigenetic modifications on cccDNA, as evidenced by increased levels of transcriptionally-repressive markers on cccDNA ( Figure 6 ). Next, we investigated recruitment of repressive HKMTs, including PR-Set7, EZH2, SETDB1, and SUV39H1 ( Figure 6 , fifth, seventh, ninth and tenth panels, lane 2 vs. 3), all of which catalyze addition of a methyl group(s), resulting PR-Set7-mediated H4K20me1, EZH2-mediated H3K27me3, or SETDB1- and SUV39H1-mediated H3K9me3 (Beck et al., 2012; Piracha et al., 2020; Yang et al., 2022; Wang et al., 2023; Wu et al., 2023) in AGK2-treated HBV-infected cells. Consistent with methylation of H4K20, H3K27, and H3K9, AGK2 treatment increased recruitment of PR-Set7, EZH2, SETDB1, and SUV39H1 to cccDNA ( Figure 6 , fourth–tenth panels, lane 2 vs. 3).

Finally, we investigated the impact of AGK2 on recruitment of host RNA polymerase II, H3, and acetylated H3 (H3K9ac and H3K14ac) to cccDNA. We noticed a significant reduction in recruitment of RNA polymerase II to cccDNA upon AGK2 treatment ( Figure 6 , thirteenth panel, lane 2 vs. 3). The pattern for acetylated H3 was similar to that for RNA polymerase II ( Figure 6 , eleventh panels, lane 2 vs. 3), while recruitment of total H3 was not affected by AGK2 ( Figure 6 , twelfth panel, lane 2 vs. 3).

The HBV wild type (WT) 1.3 mer (subtype ayw) plasmid was kindly provided by Dr. Ryu WS (Yonsei Univ, South Korea). The pcDNA6.1-human NTCP (hNTCP)-C9 plasmid was a kind gift from Dr. W. Li (National Institute of Biological Sciences, China) (Yan et al., 2012). The pCDH-hNTCP-C9 plasmid was generated from pcDNA6.1-hNTCP-C9 (Piracha et al., 2018). The luciferase reporter plasmids pGL3‐EnhII/Cp, pGL3‐EnhI/Xp, pGL3‐ preS1p, and pGL3‐preS2p were described previously (Saeed et al., 2019).

HepG2, HepG2-hNTCP-C9, HepAD38, Huh7, and HEK293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS; Gibco-BRL) and 1% (v/v) penicillin‐streptomycin (Gibco). The cells were maintained in a humidified atmosphere at 37°C/5% CO₂. HepAD38 cells were kindly provided by Christopher Seeger (Fox Chase Cancer Center) and cultured in the same medium as HepG2 cells, but with addition of 1 mg/ml G418 (G0175.0005, Duchefa Biochemie) and 2 µg/ml tetracycline (T7660-5G, Sigma-Aldrich). The tetracycline was removed to induce HBV transcription and production of virions for infection (Ladner et al., 1997). Primary human hepatocytes (PXB^®) cells were maintained in modified dHCGM medium (Yamasaki et al., 2006) (PhoenixBio USA Corporation). For transfection, Huh7 (5 × 10⁵) cells cultured in 6 cm dishes were incubated with a mixture comprising 3 μg of the 1.3mer HBV WT (ayw) or a control plasmid, 9 μg of polyethylenimine (PEI, Polysciences, 23966‐2), and 200 μl of Opti-MEM (Gibco, 31985062). In addition, HepG2 (2 × 10⁶) cells cultured in 6 cm dishes were transfected with a mixture containing 8 μg of the 1.3mer HBV WT (ayw) or a control plasmid, 40 μg of PEI, and 200 μl of Opti-MEM. Cells were inoculated with the transfection mixture at 24 h post-seeding. The mixture was replaced with fresh medium at 24 h post-transfection. At 72 h post-transfection, cells were lyzed in 0.2% NP‐40 (IGEPAL)‐TNE (10 mM Tris‐HCl [pH 8.0], 50 mM NaCl, 1 mM EDTA) buffer, as described previously (Kim et al., 2022).

The HBV virions used for infection were harvested from HepAD38 cells as described, with minor modifications (Kim et al., 2022). Stable HepG2‐hNTCP‐C9 cells were established using a lentiviral expression system as described previously (Kwon et al., 2023). Briefly, 1 × 10⁶ (in 6 cm collagen-coated (Corning, 354249) plates) or 3 × 10⁶ (in 10 cm collagen-coated plates) HepG2 or HepG2‐hNTCP‐C9 cells, respectively, were infected with 200 genome equivalents (GEq) of HBV virions per cell in medium containing 4% polyethylene glycol (PEG, Affymetrix, 25322- 68-3) and 2.5% dimethyl sulfoxide (DMSO, Sigma Aldrich, D8418). PXB^® (4 × 10⁵ cells/well in a 24 well-plate) cells were infected with 2,000 GEq of HBV. After 24 h, the cells were washed thoroughly (at least three times) with phosphate-buffered saline (PBS) and maintained in the same medium containing 2.5% DMSO. The medium was refreshed every second day.

To examine the effects of AGK2 (catalog number A8231; Sigma-Aldrich) (Outeiro et al., 2007) on HBV replication, the compound was dissolved in DMSO (4mM AGK2 stock solution) and used to treat transfected/HBV-infected cells for 72 h, or for 5 or 7 days. In brief, Huh7, HepG2, or HepG2‐hNTCP‐C9 or PXB^® cells were either transiently transfected or infected as described above. AKG2 (at the indicated concentrations) was added at the time of transfection or infection for the indicated times, followed by preparation of cell lysates as described (Kim et al., 2022). The cytotoxic effect of AGK2 were assessed in an MTT (3‐[4,5‐dimethylthiazol‐2‐yl]‐2,5‐diphenyltetrazolium bromide) assay, as described previously (Jung et al., 2015), or in an MTS assay according to the manufacturer’s instructions (Cell Proliferation; Colorimetric; ab197010).

To investigate the effects of AGK2 on formation of HBV cccDNA, HBV cccDNA was extracted from cells using the Hirt protein‐free DNA extraction procedure, as previously described (Cai et al., 2013), but with minor modifications (Kim et al., 2022). Briefly, 3 × 10⁶ HepG2 or HepG2‐hNTCP‐C9 cells, cultured on collagen‐coated 10 cm dishes, were infected with HBV and treated with 10 μM AGK2 for 7 days. Cell lysates were prepared as described (Cai et al., 2013; Kim et al., 2022), and cccDNA was precipitated using ethanol and then analyzed by Southern blotting. To further validate the authenticity of HBV cccDNA, the Hirt DNA sample was heated to 85°C for 5 min. The HBV DNA extracted from the Hirt protein-free DNA extraction sample includes cccDNA (2.1 Kbp), DL DNA (3.2 Kbp), and protein-free RC DNA (above 3.2 Kbp). Heating the Hirt DNA sample denatures the RC and DL DNAs, leaving cccDNA unaffected. Consequently, the electrophoretic mobility of the cccDNA does not change. In addition, the heat-treated DNA sample was digested with EcoR I to linearize the cccDNA, resulting in a genome-length double-stranded DNA of 3.2 Kbp.

The protein content of the cell lysates prepared in 0.2% NP‐40‐TNE buffer was measured using a Bradford assay (Kim et al., 2022). Lysates containing equal quantities of protein were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS‐PAGE) in 10% or 12% gels and transferred to PVDF membranes (Millipore). The membranes were blocked with 4% skim milk, incubated with appropriate primary antibodies, and then incubated with secondary horseradish peroxidase (HRP)‐conjugated goat anti-rabbit or HRP‐conjugated anti-mouse antibodies ( Table 1 ). Proteins were visualized by enhanced chemiluminescence (ECL, GE Healthcare Life Sciences, RPN2106). Relative band intensities were quantified using ImageJ version 1.50b. Culture supernatants from 1.3 mer HBV WT-transfected Huh7 or HepG2 cells, and from HBV infected-HepG2‐hNTCP‐C9 or PXB^® cells, were collected at the indicated times and used in enzyme-linked immunosorbent assays (ELISA) designed to detect HBsAg (WANTAI HBsAg ELISA kit, WB-2296) and HBeAg (WANTAI HBeAg ELISA kit, WB-2496).

To visualize core particles, cell lysates prepared in 0.2% NP‐40‐TNE buffer were separated by 1% native agarose gel electrophoresis (NAGE), transferred to a PVDF membrane, and immunoblotted with a rabbit polyclonal anti-HBc antibody (Jung et al., 2012) followed by an HRP-conjugated anti-rabbit secondary antibody. The core particles were visualized using ECL, and relative band intensities were quantified using ImageJ version 1.50b.

Total RNA was extracted from Huh7, HepG2, or HepG2-hNTCP-C9 cells using Trizol (Ambion, 15596018) at 3 days post-transfection or at 5 days post-infection. Next, 20 μg of total RNA was denatured at 65°C for 10 min and separated in 1% agarose gels (Invitrogen, 16500100) containing 18% formaldehyde (Sigma Aldrich, F8775) and 1 × morpholinopropanesulfonic acid (MOPS) buffer (200 mM MOPS, 10 mM EDTA, 50 mM sodium acetate [pH 7.0]). The RNA was transferred to nylon membranes (Roche, 11417240001) and hybridized at 68°C for 4 h with a random‐primed ³²P‐labeled probe specific for full‐length HBV sequences.

HBV DNA was extracted from isolated core particles at 3 days post-transfection or 7 days post-infection, electrophoresed on 1% NAGE gels, transferred to nylon membranes, and subjected to hybridization as described above. Next, in situ nucleic acid blotting was performed to analyze HBV nucleic acids (including pgRNA and RI DNAs) inside core particles. The PVDF membrane used for immunoblotting of core particles was treated for 60 s with 0.2 N NaOH, washed quickly with distilled water, UV-crosslinked (XL-1500 UV CROSSLINKER, Spectrolinker™), and subjected to hybridization as described above. Relative band intensities were quantified using ImageJ version 1.50b.

To determine the effects of AGK2 on HBV transcription, Huh7 (2.5 × 10⁵) or HepG2 (1 × 10⁶) cells were seeded onto 6-well plates and transfected with 2 μg or 4 μg the luciferase report vectors pGL3-null, pGL3-EnhII/Cp, pGL3-PreS1p or pGL3-PreS2p, and pGL3-EnhI/Xp. The cells were treated with AGK2 at the time of transfection. After 72 h, cells were lysed using 1 × luciferase cell culture lysis reagent (Promega, E153A) and luciferase activity measured by adding luciferin (Promega). Plates were read in a luminometer (Molecular Devices, EPOCH2NS).

ChIP of HBV cccDNA from cells at 7 days post-infection was conducted as described previously, with minor modifications (Piracha et al., 2020; Saeed et al., 2021). Briefly, 3 × 10⁶ HepG2‐hNTCP‐C9 cells were seeded on 10 cm collagen‐coated dishes, infected with HBV, and treated with 10 μM AGK2 as described above. At 7 days post-infection, nuclear pellets derived from cell lysates (0.2% NP‐40‐TNE) were fixed for 30 min at 4°C with 1 ml of 1% formaldehyde-containing buffer (20 mM Tris-HCl [pH 8.0], 20 mM KCl, 3 mM MgCl₂, 1 mM PMSF, 1 mM DTT). The fixed pellet was then subjected to immunoprecipitation overnight at 4°C with 1 µg of the indicated specific antibody ( Table 1 ) or with normal mouse or rabbit IgG (negative controls), as described previously (Pollicino et al., 2006; Piracha et al., 2020; Saeed et al., 2021). The immunoprecipitated protein-DNA complexes were eluted (1% SDS, 0.1 M NaHCO₃) and reverse cross-linked at 60°C for 4 h. The immunoprecipitated DNA was purified by treatment with proteinase K (Sigma Aldrich, P2308), followed by phenol-chloroform extraction and ethanol precipitation. The purified DNA was then dissolved in nuclease-free water (Invitrogen, AM9932). Input samples were prepared from sonicated chromatin solutions. Following measurement of the optical density at 260 nm (OD260), the DNA concentration was adjusted to 50 ng. Actin was utilized to facilitate equal loading of protein from lysates. The forward and reverse primers used to target actin gene were 5’-CAT GTA CGT TGC TAT CCA GGC-3’ and 5’-CTC CTT AAT GTC ACG CAC GAT-3’, respectively. The forward and reverse primers targeting cccDNA were 5’-CTC CCC GTC TGT GCC TTC T-3’ and 5’-GCC CCA AAG CCA CCC AAG-3’, respectively. Samples were subjected to both PCR (GeneAmp PCR system 2700; Applied Biosystems, Thermo Fisher Scientific).

Data are presented as the mean ± SD from a minimum of three independent experiments. Statistical comparisons of mean values were performed using Student’s t-test. The luciferase reporter assay data were analyzed using GraphPad Prism version 9.0. P values of 0.05 were considered statistically significant.