The Heteroaryldihydropyrimidine Bay 38-7690 Induces Hepatitis B Virus Core Protein Aggregates Associated with Promyelocytic Leukemia Nuclear Bodies in Infected Cells

Despite the availability of effective vaccines and treatments, HBV remains a significant global health concern, with more than 240 million individuals chronically infected. Current treatments are highly effective at controlling viral replication and disease progression but rarely cure infections. Therefore, much emphasis is being placed on finding therapeutics with new drug targets, such as viral gene expression, covalently closed circular DNA formation and stability, capsid formation, and host immune modulators, with the ultimate goal of an HBV cure. Understanding the mechanisms by which novel antiviral agents act will be imperative for the development of curative HBV therapies.

KEYWORDS antiviral agents, capsid, hepatitis B virus, mechanisms of action, virology, virus-host interactions H epatitis B virus (HBV) is a small, enveloped DNA virus of the Hepadnaviridae family that has highly specific tropism for liver cells. An estimated Ͼ240 million people worldwide are chronically infected by HBV, a condition that leads to liver disease, cirrhosis, and hepatocellular carcinoma. Currently, HBV treatment options include only nucleoside/nucleotide analogs (NUCs) and the immunomodulatory agent interferon alpha (IFN-␣). NUCs are able to suppress the virus to undetectable levels but rarely cure infections, so treatment duration is indefinite, possibly lifelong. IFN-␣ treatment is able to clear infections, but virological response is observed in only a small portion of patients, and adverse side effects are common (1)(2)(3)(4)(5)(6). New treatments, therefore, are highly desired for HBV therapy.
Bay 38-7690 induces large heterogeneous capsid assemblies in vitro. The effects of multiple HAPs on capsid assembly have been studied, and various assembly products have been observed. To study the effects of Bay 38-7690 on capsid assembly, we expressed and purified Cp (construct C150, which has the N-terminal 149-amino-acid Cp assembly domain with the 3 natural cysteines mutated to alanines and an added C-terminal cysteine) (16,(64)(65)(66). Purified C150 (18 M dimer) was assembled in the presence of 1% dimethyl sulfoxide (DMSO) or 10 M Bay 38-7690 by addition of 50 mM HEPES (pH 7.5) with 500 mM NaCl and incubation at room temperature (RT) for 1 h, and assemblies were imaged by transmission electron microscopy (TEM). Bay 38-7690 caused large heterogeneous assemblies, similar to those reported to be induced by Bay 41-4109 ( Fig. 2A) (21). Next, C150 (7.5 M dimer) was assembled in the presence of 0 to 50 M Bay 38-7690 and analyzed by dynamic light scattering (DLS). The size of capsid aggregates increased in a dose-dependent manner upon Bay 38-7690 treatment, similarly to results previously reported for the HAPs Bay 41-4109 and GLS4 ( Fig. 2B) (24).
Bay 38-7690 induces cellular Cp loss. While the biophysical effects of HAPs on in vitro capsid assembly have been studied, their impacts on Cp and viral components in HBV-infected cells remain incompletely understood. We cultured HepAD38 cells in the absence of Tet, treating them with Bay 38-7690 (0.25 to 5 M) every 2 days for 4 days, and assessed cell lysates for Cp content by Western blotting. Similarly to treatment with other HAPs (10), Bay 38-7690 treatment suppressed Cp in a dose-dependent manner, and consistent with previous reports, 3TC had no significant effect on the amount of Cp (10) (Fig. 3A). To evaluate the timing of Cp loss, HepAD38 cells were induced for 48 h and treated with 5 M Bay 38-7690 for 0 to 96 h, and cell lysates were assessed for Cp by Western blotting. We observed a time-dependent decrease in the amount of Cp, with the vast majority being lost by 48 h posttreatment (Fig. 3B).
Bay 38-7690 induces nuclear aggregation of Cp. The Western blot results in Fig. 3 suggested that Cp may be degraded upon HAP treatment as previously observed (10). Interestingly, however, we did not observe complete disappearance of Cp, even after long treatment durations and high concentrations of Bay 38-7690. Since no single-cell imaging studies have been conducted to visualize the effect of HAP treatment on HBV Cp, we carried out a series of experiments to assess the state of the nondegraded Cp in treated cells. First, HepAD38 cells were treated with DMSO or 5 M Bay 38-7690 in the absence of Tet for 4 days, fixed, and stained for Cp. We found that Cp was dispersed throughout the cell nuclei and cytoplasm without compound treatment and that nearly all cells contained Cp due to the clonality of HepAD38 cells (Fig. 4A). We further observed in Bay 38-7690-treated cells that nearly all cytoplasmic Cp seemed to dissipate by 48 h posttreatment (Fig. 4B), and surprisingly, the remaining Cp appeared in  nuclear aggregates that occurred in a time-and dose-dependent manner ( Fig. 4B to E). Consistent with these findings, we observed a marked increase in the ratio of Cpassociated fluorescence in the nucleus versus the cytoplasm as a function of time after initiation of treatment ( Fig. 4D) and inhibitor concentration (Fig. 4E).
In order to confirm that the compound effect was indeed an HBV-specific event and not an artifact of the HBV-expressing HepAD38 cells, we transfected HepG2 cells with an HBV-expressing plasmid (pHBV), treated them with DMSO or 5 M Bay 38-7690, and stained them for Cp. Bay 38-7690 exerted the same Cp aggregating effect as observed in HepAD38 cells ( Fig. 5A and C). Furthermore, the nuclear aggregation was not observed in HepG2 cells transfected with pHBV harboring the T109M HAP resistance mutation in Cp (18), demonstrating that the effect is due to direct binding of Bay 38-7690 to Cp ( Fig. 5B and C). We also tested the effect of Bay 38-7690 on Cp in HepG2 cells overexpressing the HBV receptor sodium taurocholate cotransporting polypeptide (NTCP; HepG2-NTCP) (67,68) infected with HBV from the supernatant of HepAD38 cells and treated as described above. In this fully infectious system, Bay 38-7690 again induced nuclear Cp foci, although the aggregates were notably smaller than those formed in HepAD38 and pHBV-transfected HepG2 cells (Fig. 6). The size difference is likely due to the fact that Cp is not overexpressed in HepG2-NTCP as it is in the other two systems.
Bay 38-7690-induced aggregation of Cp is independent of HBV replication. Upon verifying that the observed Cp aggregation was relevant in multiple contexts including full infection, we wished to test if the effect was reliant on HBV replication, or if it required only Cp-compound interactions. To this end, we transfected HepG2 cells with a full-length Cp-overexpressing plasmid (pC183), treated them with DMSO or 5 M Bay 38-7690, and stained them for Cp. While Cp was dispersed throughout the cells without treatment, Cp in compound-treated cells was aggregated similarly to virusinfected cells treated with compound, demonstrating that Bay 38-7690-induced aggregation is independent of viral replication (Fig. 7).
Bay 38-7690-induced Cp loss may not depend on nuclear factors. We further considered the effect of inhibitor on Cp loss and the presence of Cp aggregates in the nucleus. Due to the large size of the foci, it is unlikely that the aggregates are transported across the nuclear membrane. Furthermore, we found that Cp aggregates are present in the cytoplasm as well as in the nucleus at early times posttreatment (Fig. 4B, 8 h and 24 h) but that nearly all cytoplasmic Cp seems to dissipate by 48 h posttreatment (Fig. 4B), suggesting that cytoplasmic, but not nuclear, Cp aggregates are degraded. We therefore hypothesized that cytoplasmic factors are primarily responsible for Bay 38-7690-induced degradation. To test this, we cultured HepAD38 cells in the presence or absence of aphidicolin (2 g/ml), treated them with 5 M Bay 38-7690 in the absence of Tet for 4 days, and stained them for Cp. Aphidicolin was added to arrest cell division and presumably block transfer of large aggregates to the nucleus upon breakdown of the nuclear envelope. As expected, aphidicolin treatment decreased the number of cells (Fig. 8). Furthermore, aphidicolin-treated cells still contained nuclear Cp aggregates as in Bay 38-7690-treated cells without aphidicolin. Notably, while an average of 15.7% of untreated cells was positive for Cp after Bay  38-7690 treatment, 79.7% of aphidicolin-treated cells were Cp positive after Bay 38-7690 treatment (P ϭ 0.02 for 2 independent experiments). Taken together, the data from Fig. 4 and 8 suggest that (i) nuclear Cp aggregates are likely formed from Cp that was present in the nucleus before treatment or escaped compound-induced degradation by trafficking to the nucleus before aggregating, (ii) Cp degradation occurs primarily in the cytoplasm, and (iii) degradation of nuclear Cp aggregates may occur upon exposure to cytoplasmic proteasomes during cell division.
Bay 38-7690-induced aggregates associate with PML NBs. Because of the highly focused nature of the aggregates, we tested whether the foci were associated with a specific nuclear compartment. Due to the morphology and distribution of the aggregates, as well as the reported association of PML NBs with HBV and Cp (52-59), we chose to investigate nucleoli and PML NBs as potential sites of Cp aggregation. HepAD38 cells treated with 5 M Bay 38-7690 for 4 days in the absence of Tet were fixed and stained for Cp, PML, and nucleolin (for visualization of nucleoli). Cp aggregates colocalized with PML protein foci, but not with nucleolin, indicating that the aggregates were associated with PML NBs (Fig. 9A). Furthermore, the Cp aggregates in pHBV-transfected HepG2 and HBV-infected HepG2-NTCP cells were also associated with PML NBs (Fig. 9B and C), indicating that the Bay 38-7690-induced association of Cp with PML NBs is relevant in a fully infectious context.

DISCUSSION
The present study provides mechanistic information for HAPs at the virological level and complements previous primarily biophysical studies on the effects of HAPs and other HBV capsid assembly effectors (CAEs) on in vitro capsid assembly. We show that Bay 38-7690 affects in vitro HBV capsid assembly in a manner consistent with other HAPs and recapitulate data that may support a role for proteasomes in Cp degradation upon HAP treatment of HBV-infected cells, although we did not directly assay for proteasome involvement. Importantly, we report the novel finding that Bay 38-7690 promotes the formation of Cp aggregates associated with PML NBs. Interestingly, HAP-induced Cp loss seems to occur primarily in the cytoplasm, although PML NBs have been reported to contain proteasomal subunits (32). Our data suggest that loss of nuclear Cp is dependent on cell division, likely due to exposure of the aggregates to cytoplasmic proteasomes after breakdown of the nuclear envelope. If proteasomes are indeed involved in compound-induced Cp loss, these data are interesting not only in the context of virology but also as general cell biology, as our findings imply that nuclear and cytoplasmic protein degradation machineries are distinct and may have substrate biases or kinetic differences. Our future studies will partially focus on the differential effects that Bay 38-7690 (and possibly other HAPs) has on nuclear versus cytoplasmic protein.
Our data suggest that Bay 38-7690, and possibly other HBV CAEs, have pleiotropic effects and may also affect steps of the virus life cycle that involve PML NBs. We have found that a HAP CAE alters the composition of PML NBs by causing local aggregation of Cp, which in turn may affect the roles of PML NBs in apoptosis, DNA damage response, and cellular senescence. Ongoing studies focus on the specific effects of Bay 38-7690 and other CAEs in these cellular functions. Drug-induced accumulation of HBV Cp at PML NBs and potential disruption of such cellular processes could impart detrimental effects and possible cell death specifically to HBV-infected cells, thus enhancing the prospects of HBV eradication.

MATERIALS AND METHODS
Compounds. Bay 38-7690 was synthesized as described previously (10, 20, 60-62), 3TC was obtained through the U.S. NIH AIDS Reagent Program, and aphidicolin was purchased from Sigma-Aldrich. qPCR analysis of HBV DNA. Total cellular DNA was extracted using the QIAamp DNA blood minikit (Qiagen). Forward and reverse primers for total HBV DNA quantification were 5= CCTGGTTATCGCTGGA TGTGT 3= and 5= GGACAAACGGGCAACATACCTT 3=, respectively (70). DNA in a 10-l reaction volume was subjected to amplification by denaturation at 95°C for 15 min, followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s using ABsolute qPCR SYBR green mix (Thermo Scientific) in a PikoReal real-time PCR system. A standard curve was generated with dilutions of pHBV (69).
HBV Cp purification. A gBlock gene fragment coding for the 149-amino-acid assembly domain of HBV capsid protein with an added C-terminal cysteine (C150) (16,(64)(65)(66) with NdeI and BamHI restriction sites was synthesized by Integrated DNA Technologies and cloned into the pET11a expression vector (Novagen). HBV C150 was expressed and purified as previously described (16,(64)(65)(66), with minor modifications. The C150 expression plasmid was transformed into Escherichia coli BL21(DE3), grown at 37°C to an optical density at 600 nm (OD 600 ) of~0.8, and induced for 3 h with 1 mM isopropyl-␤-Dthiogalactopyranoside (IPTG) at 37°C. Cells were pelleted and resuspended in 50 mM Tris (pH 7.5), 1 mM EDTA, 20 mM 2-mercaptoethanol (2-ME), 1 mM phenylmethylsulfonyl fluoride (PMSF), 150 g/ml lysozyme, and 0.2 mg/ml DNase I. The suspension was incubated on ice for 30 min and lysed by sonication. Polyethylenimine (PEI) was added to a final concentration of 0.15% (wt/vol) to precipitate DNA, and the lysate was centrifuged at 16,000 ϫ g for 1 h. Ammonium sulfate was added to the supernatant to 40% saturation. The solution was gently stirred for 1 h and then centrifuged at 16,000 ϫ g for 1 h. The pellet was resuspended in buffer A (100 mM Tris [pH 7.5], 100 mM NaCl, 10 mM 2-ME) tõ 10 mg/ml and centrifuged at 16,000 ϫ g for 20 min, and the supernatant was loaded onto a buffer A-equilibrated HiLoad 26/60 Superdex 200 preparation-grade (GE Healthcare) column and eluted at 2.5 ml/min. Fractions were pooled based on the chromatogram and SDS-PAGE, concentrated tõ 5 mg/ml, and dialyzed into buffer N (50 mM sodium bicarbonate [pH 9.6], 10 mM 2-ME). Solid urea was added to 3 M and stirred for 1 h at 4°C. The solution was loaded onto a buffer N-equilibrated HiLoad 26/60 Superdex 200 preparation-grade column and eluted at 2.5 ml/min. Fractions containing the C150 dimer were pooled, concentrated, and stored at Ϫ80°C. Final protein concentration was determined spectrophotometrically.
Transmission electron microscopy. C150 (18 M dimer) in buffer N was assembled in the presence of 1% DMSO or 10 M Bay 38-7690 by addition of an equal volume of 100 mM HEPES (pH 7.5) with 1 M NaCl and incubation at room temperature (RT) for 1 h. Assemblies were absorbed to glow-discharged carbon-coated 200 mesh copper grids (Electron Microscopy Sciences), stained with 2% uranyl acetate, and imaged with a JEOL JEM-1400 transmission electron microscope.
Dynamic light scattering. C150 (7.5 M dimer) was assembled in the presence of 0 to 50 M Bay 38-7690 as described above. Assemblies were analyzed with a Protein Solutions DynaPro MS dynamic light scattering system. Data were collected at 4°C with 20 intervals of 5 s for each sample. The particle sizes were calculated from the diffusion coefficient by using the Stokes-Einstein equation with the cumulant method.
Transfections and infections. For transfections, cells were grown on 12-mm collagen-coated coverslips (neuVitro) until they reached~80% confluence. Cells were transfected with plasmids using Fugene 6 transfection reagent (Promega) according to the manufacturer's instructions. The final concentration of all plasmids was 0.6 g/ml.
For HBV infections, HepG2-NTCP cells (3 ϫ 10 4 ) were plated on 12-mm collagen-coated coverslips and treated with 1.5% DMSO for 24 h. Virus inoculation was performed with 400 HBV genome equivalents per plated cell in the presence of 4% PEG 8000 and 1.5% DMSO. The next day, cells were washed with PBS and DMEM, and DMEM-10% FBS-1.5% DMSO was added. Medium was refreshed every 2 days. Fresh medium with compound was added at 5 and 7 days postinfection (dpi), and cells were fixed for immunofluorescence analysis at 9 dpi.
Confocal microscopy and image analysis. Images were taken with a Leica TCS SP8 MP inverted spectral confocal laser scanning microscope. For quantification purposes, cells in 96-well image plates (BD Falcon) were fixed and stained as described above, and automated imaging was conducted with a Zeiss LSM 510 Meta confocal microscope with Autostage, Multitile, and MultiTime series 4.0.31 beta software, as previously described (71); imaging was carried out using a 40ϫ objective, capturing 9 images per well per experiment. Images were processed using CellProfiler (72)(73)(74).