Heat Shock Protein Family A Member 1 Promotes Intracellular Amplification of Hepatitis B Virus Covalently Closed Circular DNA

ABSTRACT

Hepatitis B virus (HBV) contains a partially double-stranded relaxed circular DNA (rcDNA) genome that is converted into a covalently closed circular DNA (cccDNA) in the nucleus of the infected hepatocyte by cellular DNA repair machinery. cccDNA associates with nucleosomes to form a minichromosome that transcribes RNA to support the expression of viral proteins and reverse transcriptional replication of viral DNA. In addition to the de novo synthesis from incoming virion rcDNA, cccDNA can also be synthesized from rcDNA in the progeny nucleocapsids within the cytoplasm of infected hepatocytes via the intracellular amplification pathway. In our efforts to identify cellular DNA repair proteins required for cccDNA synthesis using a chemogenetic screen, we found that B02, a small-molecule inhibitor of DNA homologous recombination repair protein RAD51, significantly enhanced the synthesis of cccDNA via the intracellular amplification pathway in human hepatoma cells. Ironically, neither small interfering RNA (siRNA) knockdown of RAD51 expression nor treatment with another structurally distinct RAD51 inhibitor or activator altered cccDNA amplification. Instead, it was found that B02 treatment significantly elevated the levels of multiple heat shock protein mRNA, and siRNA knockdown of HSPA1 expression or treatment with HSPA1 inhibitors significantly attenuated B02 enhancement of cccDNA amplification. Moreover, B02-enhanced cccDNA amplification was efficiently inhibited by compounds that selectively inhibit DNA polymerase α or topoisomerase II, the enzymes required for cccDNA intracellular amplification. Our results thus indicate that B02 treatment induces a heat shock protein-mediated cellular response that positively regulates the conversion of rcDNA into cccDNA via the authentic intracellular amplification pathway.

IMPORTANCE Elimination or functional inactivation of cccDNA minichromosomes in HBV-infected hepatocytes is essential for the cure of chronic hepatitis B virus (HBV) infection. However, lack of knowledge of the molecular mechanisms of cccDNA metabolism and regulation hampers the development of antiviral drugs to achieve this therapeutic goal. Our findings reported here imply that enhanced cccDNA amplification may occur under selected pathobiological conditions, such as cellular stress, to subvert the dilution or elimination of cccDNA and maintain the persistence of HBV infection. Therapeutic inhibition of HSPA1-enhanced cccDNA amplification under these pathobiological conditions should facilitate the elimination of cccDNA and cure of chronic hepatitis B.

INTRODUCTION

Hepatitis B virus (HBV) chronically infects 296 million people worldwide and causes more than 800,000 deaths annually due to cirrhosis and hepatocellular carcinoma (HCC) (1). Although the long-term suppression of viral replication by nucleos(t)ide analogue viral DNA polymerase inhibitors reduces the risk of death due to liver diseases by 50 to 70% (2, 3), less than 5% of those treated patients achieve the loss of circulating HBV surface antigen (HBsAg), i.e., the functional cure of chronic hepatitis B (CHB) (3, 4). The antiviral drugs that induce a functional or virological cure of CHB in the majority of patients with a finite duration of therapy are unmet medical needs (5).

HBV infects hepatocytes by binding to its receptor, sodium taurocholate-cotransporting polypeptide (NTCP), on the plasma membrane, which triggers the endocytosis and sorting of virions into the late endosome networks (6, 7). The fusion of viral envelope and endosomal membranes results in the release of nucleocapsids into the cytoplasm and subsequent delivery of viral genome, a 3.2-kb partially double-stranded relaxed circular DNA (rcDNA), into the nuclei upon the disassembly of nucleocapsids at the nuclear pore complexes (8–10). The rcDNA is converted into a covalently closed circular DNA (cccDNA) by cellular DNA repair machinery for transcription of viral RNA to support the expression of viral proteins and reverse transcriptional replication of viral DNA (11, 12). Briefly, binding of viral DNA polymerase to the stem-loop structure at the 5′ terminus of pregenomic RNA (pgRNA) initiates their packaging by 120 copies of core protein dimers to form a nucleocapsid (13–15). Inside the nucleocapsid, viral DNA polymerase converts the pgRNA first to a single-stranded DNA and then to rcDNA (16). The rcDNA-containing mature nucleocapsids acquire an envelope and are secreted out of cells as virions (17). In addition, rcDNA in the cytoplasmic progeny nucleocapsids can also be delivered into the nuclei to form more cccDNA to maintain the pool of cccDNA, which is subjected to regulation by viral and host cellular factors (18–20).

cccDNA is the most stable HBV replication intermediate (21). It is refractory to the currently available antiviral medications and is the resource of viral replication rebound after the termination of long-term antiviral therapies (22, 23). Lack of knowledge on the metabolism and functional regulation of cccDNA hampers the discovery and development of novel antiviral therapeutics to eliminate or silence cccDNA. Recently, significant progress has been made in the identification of cellular proteins involved in cccDNA biosynthesis and its regulation (24). Specifically, in vitro biochemical studies suggest that the core components of DNA lagging-strand synthesis machinery are sufficient to convert recombinant rcDNA into cccDNA (25, 26). Cell-based genetic studies have identified various DNA repair proteins, including tyrosyl-DNA phosphodiesterase 2 (TDP2) (27); flap endonuclease I (FEN-1) (28); DNA polymerase κ, α, and δ (29, 30); DNA topoisomerases I and II (31); and DNA ligases 1 and 3 (32), that are required for the distinct steps of cccDNA synthesis. Moreover, using genetic and pharmacological approaches, recent studies also revealed that SAM and HD domain-containing deoxynucleoside triphosphate triphosphohydrolase 1 (SAMHD1) and ATR-CHK1 signaling pathway positively regulate cccDNA formation (33, 34), whereas retinoid X receptor alpha (RXRα) negatively regulates cccDNA synthesis (35). In this study, we found that a small-molecule compound, B02, potently enhanced cccDNA biosynthesis via the intracellular amplification pathway. Mechanistic analysis demonstrated that B02 enhancement of cccDNA amplification depends on the induction and function of heat shock protein family A (HSP70) member 1 (HSPA1). Our findings thus shed light on the critical role of the cellular chaperone protein in the maintenance of the cccDNA pool under cellular stress-related pathobiological conditions and provide clues for the development of therapeutics to accelerate the loss of cccDNA and facilitate the cure of chronic HBV infections.

RESULTS

B02, a RAD51 inhibitor, potently enhances cccDNA intracellular amplification.

In order to identify cellular proteins required for or regulating cccDNA biosynthesis, we screened a panel of commercially available small-molecule compounds targeting cellular DNA metabolic enzymes in HepAD38 cells under a synchronized rapid cccDNA synthesis condition reported previously (31). As depicted in Fig. 1A, HepAD38 cells were cultured in the absence of tetracycline (Tet) and presence of phosphonoformic acid (PFA, or Foscarnet), a reversible HBV DNA polymerase inhibitor, for 4 days to arrest HBV DNA synthesis at the stage of full-length negative-strand DNA. Tet was then added back to shut off pgRNA transcription from the HBV transgene in the cellular chromosome, and PFA was removed to resume viral DNA synthesis. Our previous study showed that the rcDNA was gradually accumulated and cccDNA became detectable at 12 h and increased in the following 12 h after the removal of PFA (31). Taking advantage of the synchronized rapid cccDNA synthesis within 24 h post-PFA removal, testing compounds were added to the culture at the removal of PFA for 24 h, and their effects on cccDNA synthesis were determined by Southern blotting hybridization. Using this assay, we discovered that the inhibitors of DNA polymerase alpha and topoisomerases I and II potently inhibited cccDNA synthesis (30, 31). However, to our surprise, although the total amounts of HBV replication intermediates and rcDNA were not altered (Fig. 1B), B02 treatment markedly increased the amounts of cccDNA with the concurrent reduction of deproteinized rcDNA (DP-rcDNA), the putative precursor of cccDNA (20, 36, 37), in a concentration-dependent manner (Fig. 1C to E). No cytotoxicity was observed at the concentration of B02 up to 30 μM at 24 h of treatment by microscopic inspection and MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide] cell viability assay (Fig. 1F).

FIG 1.

B02, a RAD51 inhibitor, potently enhances cccDNA intracellular amplification in HepAD38 cells. (A) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after removal of tetracycline (Tet). On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and treated with the indicated concentrations of B02. (B and C) Cytoplasmic HBV core DNA and cccDNA were detected by Southern blotting hybridization. For cccDNA detection, Hirt DNA was denatured at 88°C for 8 min and chilled on ice, which completely denatures DP-rcDNA into single-stranded DNA (annotated as denatured DP-rcDNA), whereas cccDNA remains a double-stranded circular DNA. The heat-denatured Hirt DNA samples were further digested with EcoRI to linearize the cccDNA into unit-length double-stranded linear DNA, designated as ccc/EcoRI. Mitochondrial DNA (mtDNA) served as a loading control of Hirt DNA. rc, relaxed circular DNA; dsl, double-stranded linear DNA; ss, single-stranded DNA. (D and E) The intensity of HBV cccDNA (ccc/EcoRI) and denatured DP-rcDNA signals on Southern blotting was measured by ImageJ and presented as relative amount to that in cells treated with DMSO. (F) The viability of HepAD38 cells treated with the indicated concentrations of B02 for 24 h under the condition illustrated in panel A was determined by an MTT assay.

To rule out the possibility that the observed B02 enhancement of cccDNA synthesis is an HepAD38 cell line-specific phenomenon, the finding was further validated in HepG2 cells transduced by recombinant adenoviruses supporting inducible HBV replication and cccDNA synthesis (10). Specially, HepG2 cells were infected with recombinant adenoviruses that Tet-inducibly transcribe wild-type HBV pgRNA or pgRNA encoding core protein with I126A substitution, designated Ad-HBV and Ad-HBVcoreI126A, respectively. As depicted in Fig. 2A (top), the cells were mock treated or treated with B02, starting at the time of PFA removal for 24 h. In agreement with that observed in HepAD38 cells (Fig. 1), B02 treatment enhanced cccDNA amplification in Ad-HBV-transduced HepG2 cells (Fig. 2A). As anticipated, the I126A mutation of the HBV core protein facilitates the nucleocapsid uncoating, which resulted in the reduced cytoplasmic rcDNA, but increased amounts of cccDNA (38) (Fig. 2A). Remarkably, B02 treatment further enhanced the cccDNA synthesis in HepG2 cells, replicating the core protein mutant HBV (Fig. 2A). Moreover, to rule out the possibility that the observed B02 enhancement of cccDNA synthesis is peculiar to the PFA-synchronized cccDNA synthesis condition, HepAD38 cells were treated with 10 μM B02, starting after 4 days of Tet removal for additional 4 days (Fig. 2B, top). Although the amounts of core-associated HBV DNA replication intermediates and rcDNA under the prolonged treatment condition were reduced due to an unknown mechanism(s) (Fig. 2B, middle), B02 treatment did increase the amount of cccDNA with the concurrent reduction of DP-rcDNA (Fig. 2B, bottom). Hence, the results presented here further confirmed that B02 treatment enhances the intracellular amplification of cccDNA.

FIG 2.

Enhancement of cccDNA intracellular amplification by B02 in Ad-HBV-infected HepG2 and HepAD38 cells without PFA synchronization. (A) HepG2 cells were infected with Ad-HBV or Ad-HBVcoreI126A at a multiplicity of infection (MOI) of 10 and cultured in the absence of Tet. We applied 2 mM PFA to culture media from day 2 to day 6 after infection to arrest HBV replication. On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and mock treated or treated with 20 μM B02. Cytoplasmic HBV core DNA and Hirt DNA were extracted and resolved by Southern blotting assays, with mtDNA as a loading control for Hirt DNA. (B) HepAD38 cells were cultured in the absence of Tet for 4 days. The cells were then mock treated or treated with 10 μM B02 for additional 4 days. Cytoplasmic HBV core DNA and Hirt DNA were extracted and resolved by Southern blotting assays. mtDNA served as a loading control for Hirt DNA.

B02 treatment does not promote the conversion of preaccumulated DP-rcDNA into cccDNA.

The accumulation of DP-rcDNA in hepatoma cells supporting HBV replication had been attributed to the low efficiency of DP-rcDNA to cccDNA conversion, possibly due to the inherently less effective DNA repair processes in the tumor cells (20, 39). The concurrent decrease of DP-rcDNA with the increase of cccDNA in B02-treated HepG2 cells evokes a hypothesis that B02 treatment promotes the conversion of DP-rcDNA to cccDNA. To test this hypothesis, HepAD38 cells were cultured for 48 h after the removal of PFA to allow the accumulation of DP-rcDNA and establishment of cccDNA. The cells were then treated with a serial concentration of B02 for 24 h. In striking contrast to the significant reduction of DP-rcDNA in hepatoma cells treated with B02 starting at the time of PFA removal (Fig. 1 and 2), B02 treatment did not alter the levels of DP-rcDNA and cccDNA under this experimental condition (Fig. 3A to C). The failure of B02 to promote the conversion of the preaccumulated DP-rcDNA into cccDNA implies that the accumulated DP-rcDNA is not the precursor of cccDNA. It is thus possible that the newly produced DP-rcDNA is short-lived and will be either immediately converted into cccDNA or misrepaired to “dead-end” protein-free rcDNA species and accumulated in cells (Fig. 3D).

FIG 3.

B02 treatment does not promote the conversion of preaccumulated DP-rcDNA into cccDNA. (A) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after Tet withdrawal. On day 6, Tet was added back to stop pgRNA transcription, and PFA was removed to resume HBV DNA synthesis and establishment of the cccDNA pool for 48 h. The cells were then mock treated or treated with 20 μM of B02 for an additional 24 h. Hirt DNA was extracted and resolved by Southern blotting assays. (B and C) The intensity of HBV cccDNA and DP-rcDNA signals on Southern blot was measured by ImageJ and presented as the relative amount to that in cells mock treated with DMSO. (D) Schematic illustration of a proposed “dead-end” DP-rcDNA model.

B02 treatment promotes the synthesis of cccDNA from both rcDNA and double-stranded linear DNA.

As depicted in Fig. 4A, cccDNA can be synthesized from either rcDNA or double-stranded linear DNA (dslDNA), a minor DNA species produced by in situ priming of positive-strand DNA synthesis (40). While the conversion of rcDNA to cccDNA perfectly repairs the cohesive ends via unknown DNA repair pathways (24), synthesis of cccDNA from dslDNA is catalyzed by the error-prone nonhomologous end-joining (NHEJ) DNA repair pathway (41), and the resulted cccDNA molecules contain indels at the junction region (42, 43). Because B02 is an inhibitor of RAD51, a key player in the homologous recombination DNA repair pathway (44), we thus speculated that B02 treatment might favor the synthesis of cccDNA from the dslDNA (Fig. 4A, right) via the NHEJ DNA repair pathway. Indeed, sequencing the gap junction region of cccDNA revealed that only 3 out of 44 (7%) cccDNA from the cells treated with control solvent (dimethyl sulfoxide [DMSO]) contains indels, whereas 11 out of 49 (22%) cccDNA from B02-treated cells harbors indels in this region (Fig. 4B), with single T insertion in the stretch of five Ts being the most frequent mutation in both mock and B02 treated cells (data not shown). However, although the frequency of dslDNA-derived cccDNA was slightly increased in B02-treated cells, the majority of cccDNA was synthesized from rcDNA in either the absence or presence of B02 treatment (Fig. 4B). In agreement with this observation, inhibition of the NHEJ pathway by DNA-PK inhibitor II did not apparently inhibit B02 enhancement of cccDNA synthesis (Fig. 4C). On the contrary, inhibitors of DNA repair enzymes essential for the conversion of rcDNA into cccDNA (25, 30, 31), such as aphidicolin (Fig. 4D) and doxorubicin (Fig. 4E), potently inhibited B02-enhanced cccDNA synthesis. Notably, the amount of cccDNA increased with various supercoiled turns, as indicated by the ladder of HBV DNA between supercoiled cccDNA and rcDNA (31, 45, 46) (Fig. 4F, lanes 2 and 5). However, the putative intermediate from rcDNA to cccDNA conversion, i.e., covalently closed minus strand rcDNA, abbreviated to cc(−)rcDNA (47), was not significantly increased by B02 treatment (Fig. 4F, lanes 2 and 5). This later observation suggests that B02 treatment efficiently accelerates the conversion of rcDNA to cccDNA without the accumulation of this intermediate.

FIG 4.

Characterization of cccDNA molecules synthesized under B02 treatment. (A) Schematic illustration of cccDNA biosynthesis pathways. (B) Sequence analyses of rcDNA gap junction region in cccDNA molecules in HepAD38 cells. (C to E) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after removal of Tet. On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and treated with B02 in combination with DNA-PK inhibitor (DNA-PKi) II (C), group B DNA polymerase inhibitor aphidicolin (APH) (D), and topoisomerase inhibitor doxorubicin (DOXO) (E). Hirt DNA was extracted and resolved by Southern blotting assays. (F) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after removal of Tet. On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and treated with DMSO or B02. Hirt DNA was extracted and either loaded directly (lanes 1 and 4) or loaded after exonucleases digestion (lanes 2 and 5). For samples loaded on lanes 3 and 6, exonuclease digested Hirt DNA was purified by phenol-chloroform followed by EcoRI digestion. Minus-strand covalently closed (cc−) rcDNA is denoted.

B02 enhancement of cccDNA synthesis does not depend on RAD51.

Previous biochemical studies demonstrated that B02 disrupts the association between single-stranded DNA and RAD51, thereby suppressing RAD51-mediated DNA strand exchange and branch migration, an essential step for DNA homologous recombination repair (44, 48). Functional inhibition of RAD51 by B02 was validated in HepAD38 cells by B02 inhibition of RAD51 recruitment to γH2A.X foci upon induction of DNA double-strand breaks by doxorubicin treatment (data not shown) (49). To investigate the role of RAD51 in B02 promotion of cccDNA synthesis, we first determined the effects of knocking down RAD51 expression by small interfering RNA (siRNA) on cccDNA synthesis in the absence or presence of B02 treatment. Successful knockdown of RAD51 expression was verified by immunoblotting assay (Fig. 5A and B). To our surprise, reduction of RAD51 protein abundance did not alter basal, as well as B02-enhanced, cccDNA intracellular amplification (Fig. 5C). In agreement with the results obtained from the siRNA knockdown assay, the levels of cccDNA and DP-rcDNA were not affected by treatment with another structurally distinct RAD51 inhibitor, RI-1 (50), or activator, RS-1 (51) (Fig. 5D). These results thus favor a hypothesis that B02 enhances cccDNA amplification through a mechanism independent of RAD51.

FIG 5.

B02 enhancement of cccDNA synthesis does not depend on RAD51. (A) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal. At day 4, the cells were reseeded in 6-well plates and transfected with 10 pmol RAD51-targeting siRNA or scramble siRNA by using Lipofectamine RNAiMAX. From 48 h to 72 h after siRNA transfection (day 6 to day 7 after Tet removal), cccDNA synthesis was resumed by removal of PFA from culture medium and treating with or without B02 (20 μM). (B) The levels of RAD51 and β-actin were determined at 72 h after siRNA transfection by Western blotting. (C) HBV core DNA and Hirt DNA were extracted and resolved by Southern blotting assays, with mtDNA as a loading control of Hirt DNA. (D) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after Tet removal. On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and treating with indicated compounds. HBV core DNA and Hirt DNA were extracted and resolved by Southern blotting assays, with mtDNA as a loading control of Hirt DNA.

B02 enhancement of cccDNA synthesis depends on protein synthesis.

Since compounds directly targeting DNA repair proteins, such as DNA polymerase α and topoisomerase II, promptly inhibited cccDNA synthesis (30, 31), we next asked whether B02 promotes cccDNA synthesis in a rapid fashion. Accordingly, a time-of-addition experiment was performed to determine the minimal time of incubation period required for B02 to elicit a detectable increase of cccDNA. As shown in Fig. 6A, B02 treatment for 8 h during the peak time of cccDNA formation, between 16 h to 24 h after PFA release, neither increased the amount of cccDNA nor reduced the level of DP-rcDNA. As anticipated, extending the B02 treatment from 16 h to 40 h after PFA release significantly increased cccDNA and reduced DP-rcDNA (Fig. 6A). These results indicate that a longer than 8-h treatment is required for B02 to enhance an ongoing cccDNA synthesis process.

FIG 6.

B02 enhancement of cccDNA synthesis depends on protein synthesis. (A) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal. From day 6, cccDNA synthesis was initiated by removing PFA for 16 h and then culturing in the presence of indicated compounds for the next 8 h or 24 h of cccDNA synthesis. Cells were harvested at 16 h, 24 h, and 40 h post-PFA removal. Hirt DNA was extracted, and HBV DNA was detected by Southern blotting assay. (B) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal. From day 6, cccDNA synthesis was allowed by removing PFA for 16 h in the presence of B02 and cycloheximide, alone or in combination. Cytoplasmic core DNA and Hirt DNA were extracted and resolved by Southern blotting assay. (C) The intensity of HBV cccDNA band was quantified by ImageJ and presented as the relative amount in comparison with that in DMSO-treated controls. Data represent 3 independent experiments (mean ± SD). Data were analyzed by two-tailed Student's t test (unpaired). *, P < 0.05.

The slow action of B02 on cccDNA synthesis favors a model that induction of effector protein expression and/or activation of a cellular response pathway are required for B02 to enhance cccDNA synthesis. In support of this hypothesis, treatment with cycloheximide (CHX), a protein translation inhibitor, significantly attenuated B02 enhancement of cccDNA synthesis (Fig. 6B and C). The results thus imply that B02 enhancement of cccDNA synthesis depends on new protein synthesis.

B02 treatment induces the expression of multiple heat shock proteins in HepAD38.

In order to identify the cellular protein(s) mediating B02 enhancement of cccDNA synthesis, RNA sequence analysis was performed with mRNA extracted from HepAD38 cells mock treated or treated with B02 (20 μM) for 7 h. The expression of more than 80 cellular genes was altered greater than 2-fold by B02 treatment (Fig. 7A). Pathway analysis shows that the top-affected genes are related to glycan biogenesis, hormone metabolism, and cell death, but not DNA repair pathways. However, multiple genes belonging to heat shock protein families were induced, among which HSPA1A/B (HSP70 family), DNAJB4 (HSP40 family), and HSPH1 (HSP110 family) were the highly induced ones. The elevated expression of selected heat shock proteins upon B02 treatment was verified by immunoblot assays (Fig. 7B). Because previous studies indicate that certain heat shock proteins participate in the DNA repair process by facilitating the function of DNA repair enzymes and/or signal transduction of DNA damage response (52–55), we thus speculated that one or more B02-induced heat shock proteins may be required for the enhanced cccDNA synthesis.

FIG 7.

B02 treatment induces the expression of multiple heat shock proteins in HepAD38. (A) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal, followed by DMSO or B02 (20 μM) treatment for 7 h. Total RNA was extracted by TRIzol, and RNA-seq analysis was performed. Fold changes of gene expression levels (log₂) are illustrated (red, elevated; blue, decreased). (B) HepAD38 cells were harvested after 24 h exposure to B02 at the indicated concentrations. Induction of HSPA1, DNAJB1, and DNAJB4 expression by B02 treatment was verified by a Western blot assay.

HSPA1 is required for B02 enhancement of cccDNA synthesis.

To identify the heat shock protein(s) required for B02 enhancement of cccDNA synthesis, each of the highly induced heat shock proteins was knocked down in HepAD38 cells by siRNA. As shown in Fig. 8A and B, knockdown of HSPA1 by siRNA HSPA1-A, but not other heat shock proteins by the respective siRNA, resulted in a significant reduction of cccDNA in the absence or presence of B02 treatment. Those results thus suggest that HSPA1 plays an important role in both the basal and B02-enhanced cccDNA synthesis. Notably, although both siRNA HSPA1-A and HSPA1-B target the coding regions shared by HSPA1A and HSPA1B transcripts that translate HSPA1 and siRNA HSPA1-B less efficiently knocked down the basal and B02-induced HSPA1 expression (Fig. 8A), and consequentially did not significantly alter cccDNA amount (Fig. 8B). In addition to the siRNA approach, we also evaluated the effects of pharmacological inhibitors that inhibit the distinct step of HSP70 catalytic cycle on cccDNA synthesis in the absence and presence of B02 treatment (Fig. 9A). VER155008, a catalytic inhibitor of HSP70 that competes with ATP binding (56), efficiently blocked B02 enhancement of cccDNA synthesis at a noncytotoxic concentration (Fig. 9B and E). Similarly, two allosteric inhibitors of HSP70 targeting the binding site of the nucleotide exchange factors, MKT-077 and its analog JG-98 (57, 58), reduced the amounts of cccDNA, particularly under the condition of B02 treatment (Fig. 9C and D), without cytotoxicity at effective concentrations (Fig. 9F and G). In summary, the genetic and pharmacological evidence presented in Fig. 8 and 9 strongly support the notion that HSPA1 plays an essential role in basal and B02-enhanced cccDNA amplification in human hepatoma cells.

FIG 8.

Knockdown of HSPA1 attenuates B02 enhancement of cccDNA intracellular amplification. (A) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal. At day 4, the cells were reseeded in 6-well plates and transfected with 10 pmol siRNA targeting indicated genes by using Lipofectamine RNAiMAX. From 48 h to 72 h after siRNA transfection (day 6 to day 7 after Tet removal), cccDNA synthesis was resumed by removal of PFA from the culture medium and mock treated or treated with B02 (20 μM). Hirt DNA was extracted and resolved by Southern blotting assays, with mtDNA as a loading control of Hirt DNA. The levels of HSPA1, DNAJB1, DNAJB4, and β-actin were determined at 72 h after siRNA transfection by Western blotting. (B) The intensity of HBV cccDNA bands was quantified by ImageJ and presented as relative amount in comparison with that in cells transfected with scramble siRNA and mock treated with DMSO. Data represent 2 independent experiments (mean ± SD). Data were analyzed by two-tailed Student's t test (unpaired). *, P < 0.05.

FIG 9.

Induction and function of HSPA1 are required for B02 enhancement of cccDNA intracellular amplification. (A) HSP70 catalytic cycle. Briefly, the J-domain protein presents the unfolded protein to HSP70-ATP. Subsequent ATP hydrolysis triggers a conformation change of HSP70 to stabilize the interaction with the unfolded protein and release the J-domain protein. The nucleotide exchange factor (NEF) will then interact with the HSP70 complex, followed by ATP recruitment and hydrolysis to release the folded protein and NEF from HSP70. Inhibitors targeting distinct steps of the HSP70 catalytic cycle are denoted. (B to D) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after removal of Tet. On day 6, cccDNA synthesis was allowed by removing PFA and treating with B02 for 24 h in the absence or presence of VER155008 (B), JG-98 (C), or MKT-077 (D). Hirt DNA was extracted and resolved by Southern blotting assays. The intensity of HBV cccDNA and DP-rcDNA bands were quantified by ImageJ and presented as relative amount to that in cells mock treated with DMSO. Data represent average values from 2 independent experiments. (E to G) Cell viability under each treatment condition was determined by cell CytoTox-Glo assay following the manufacturer’s protocol. and the results are normalized as percentage of viable cells in DMSO-treated groups. Data represent ≥3 independent experiments (mean ± SD).

Induction of HSPA1 alone is not sufficient to enhance cccDNA amplification.

To address if cccDNA intracellular amplification can be boosted by induction of HSPA1 alone, an HSP70 transcription activator, TRC051384, was applied to induce HSPA1 expression (59). Despite that HSPA1 expression was induced in a dose-dependent manner, TRC051384 treatment did not alter the amount of cccDNA in HepAD38 cells (Fig. 10A). However, culturing HepAD38 cells at 43.5°C for 2 h (heat shock) and then for 22 h at 37°C upon removal of PFA significantly increased the level of cccDNA, which can be significantly attenuated by knocking down HSPA1 expression (Fig. 10B and C). Of note, the DP-rcDNA reduction is not evident after heat shock treatment, possibly due to the relative lower degree of cccDNA enhancement than that under B02 treatment condition (Fig. 10B and D). Hence, although induction of HSPA1 expression alone cannot enhance cccDNA synthesis, the cellular stress responses induced by either B02 treatment or heat shock promote intracellular cccDNA amplification in an HSPA1-dependent manner.

FIG 10.

Induction of HSPA1 alone does not enhance cccDNA synthesis. (A) HepAD38 cells were cultured in the presence of 2 mM PFA to arrest HBV replication from day 2 to day 6 after removal of Tet. On day 6, cccDNA synthesis was allowed by removing PFA for 24 h and treating with increasing concentrations of TRC051384. The levels of HSP70 and β-actin were determined at 24 h after TRC051384 treatment by Western blotting assay. Hirt DNA was extracted and resolved by Southern blotting. (B) HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal. At day 4, the cells were reseeded in 6-well plate and transfected with 10 pmol HSPA1-targeting siRNA or scramble (−) siRNA by using Lipofectamine RNAiMAX. From 48 h to 72 h after siRNA transfection (day 6 to day 7 after Tet removal), cccDNA synthesis was resumed by removal of PFA from culture medium. At the initial 2 h after PFA removal, the cells were exposed to 43.5°C heat shock and recovered at 37°C for 22 h before harvesting. HBV core DNA and Hirt DNA were extracted and resolved by Southern blotting assays, with mtDNA as a loading control of Hirt DNA. The levels of HSPA1 and β-actin were determined at 72 h after siRNA transfection by Western blotting. (C and D) The intensity of HBV cccDNA (C) and DP-rcDNA (D) signals on Southern blotting was quantified by ImageJ and presented as the relative amount to that in cells transfected with control siRNA without heat shock treatment. Results from 5 independent experiments are presented (mean ± SD). Data were analyzed by two-tailed Student's t test (unpaired). *, P < 0.05; **, P < 0.01; ns, not significant.

B02 does not enhance cccDNA synthesis in de novo HBV infection.

To assess whether B02 treatment enhances the virion-derived rcDNA-to-cccDNA conversion during de novo HBV infection, C3A^hNTCP cells were infected by HBV under B02 cotreatment and pretreatment conditions (Fig. 11A). Upregulation of HSPA1 expression was readily detected in C3A^hNTCP cells treated by B02 (Fig. 11B). However, in striking contrast with the robust enhancement of cccDNA intracellular amplification (Fig. 1A), B02 treatment did not alter the level of cccDNA at 48 h after HBV infection (Fig. 11C).

FIG 11.

B02 does not promote de novo cccDNA synthesis. (A) C3A^hNTCP cells were mock infected or infected with HBV at 500 genome-of-equivalent (GOE). B02 (20 μM) treatment was started either at 16 h prior to or at the infection. Cells were harvested at 12 and 48 hpi. (B) HSPA1 (HSP70) and β-actin were detected by Western blotting assay. (C) Hirt DNA was extracted and resolved by Southern blotting assays, with mtDNA as a loading control.

DISCUSSION

The biosynthesis, maintenance, and function of cccDNA are essential for HBV replication and persistence (11). Although multiple cellular DNA repair proteins have been demonstrated to participate in cccDNA synthesis, the molecular pathways converting rcDNA to cccDNA and their regulation under various pathobiological conditions have not been fully understood (24). The studies reported here showed that treatment of human hepatoma cells with B02 drastically promoted cccDNA intracellular amplification (Fig. 1 and 2), but not de novo cccDNA synthesis (Fig. 11). Mechanistic analysis revealed that B02 treatment promoted the conversion of newly produced, but not preaccumulated DP-rcDNA, to cccDNA (Fig. 3). Interestingly, although the function of RAD51 can be efficiently inhibited in the hepatoma cells by B02, the enhancement of cccDNA synthesis appears to be independent of RAD51 (Fig. 5) and works, instead, through induction of a heat shock protein-mediated cellular stress response (Fig. 7 and 8). Particularly, the induction and function of HSPA1 are required for B02-enhanced cccDNA synthesis (Fig. 8 and 9). The positive regulation of cccDNA synthesis by the heat shock protein-mediated cellular stress response may be attributed to the maintenance of cccDNA pool in infected hepatocytes under certain pathobiological conditions and thus contribute to the persistence of HBV infection.

DP-rcDNA has long been considered as the precursor of cccDNA (20, 36, 37) or dead-end product of HBV replication (33, 47). Identification of cc(−)rcDNA as the intermediate of cccDNA synthesis, as well as the concurrent reduction of DP-rcDNA with the B02-induced increase of cccDNA, seems to support the precursor role of DP-rcDNA (47). However, the massive accumulation of DP-rcDNA in the cytoplasm and nuclei of hepatoma cells supporting HBV replication implies that not all the DP-rcDNA are the precursor of cccDNA. The failure to convert the preaccumulated DP-rcDNA into cccDNA by B02 treatment further suggests that the majority of accumulated DP-rcDNA is not the functional precursor of cccDNA, but most likely dead-end protein-free rcDNA derived from improper repairment of true precursor DP-rcDNA (Fig. 3D). Alternatively, the accumulated DP-rcDNA may be trapped in a subcellular compartment or incompletely uncoated capsids and cannot be repaired into cccDNA. We thus favor the model depicted in Fig. 3D, i.e., the newly produced DP-rcDNA is short-lived and will be either immediately converted into cccDNA or misrepaired (misplaced) to “dead-end” protein-free rcDNA species and accumulated in cells.

It is very interesting that B02 enhances cccDNA intracellular amplification, but not de novo synthesis. A plausible explanation is that the functional precursor of cccDNA synthesis is more abundant for intracellular amplification. The very inefficient entry of HBV virions into hepatocytes results in a small number of nucleocapsids (rcDNA) available for cccDNA synthesis (6, 7). On the contrary, large amounts of nucleocapsids containing rcDNA, i.e., mature nucleocapsids, exist in the cytoplasm of hepatoma cells supporting HBV replication (20, 60). Alternatively, our previous studies demonstrated that de novo synthesis and intracellular application of cccDNA utilize different cellular DNA polymerases and, thus, most likely distinct DNA repair pathways (30). It is therefore possible that B02 treatment only enhances the DNA repair pathway catalyzing cccDNA intracellular amplification.

Concerning the mechanism of B02 promotion of cccDNA intracellular amplification, our results rejected the involvement of RAD51. Instead, B02 treatment induced significant transcriptome change via unknown mechanisms (Fig. 7). While the pathway analysis did not reveal the apparent alteration of DNA repair-related genes, the expression of several heat shock proteins was induced by B02 treatment. siRNA knockdown of the individual heat shock protein expression identified that HSPA1 is required not only for basal but also B02-enhanced cccDNA amplification. This finding was further supported by the results showing that three HSP70 inhibitors targeting distinct steps of HSP70 catalytic cycle inhibited both basal and B02-enhanced cccDNA amplification. As a molecular chaperone, it can be speculated that HSPA1 may facilitate the folding and/or function of one or multiple cellular proteins essential for cccDNA amplification. However, the step(s) of cccDNA synthesis and cellular DNA repair proteins facilitated by the chaperone remains to be identified.

Interestingly, although required for basal and B02-enhanced cccDNA amplification, transcriptional induction of HSPA1 by TRC051384 failed to enhance cccDNA synthesis. However, the phenocopy of B02 effects on cccDNA synthesis by heat shock treatment demonstrated that, as one of the key components of cellular stress responses induced by B02 treatment or heat shock, HSPA1 plays an important role in promoting cccDNA amplification. In addition, the fact that the degree of cccDNA synthesis enhancement is not solely correlated with the levels of HSPA1 indicates that other unknown cellular factors are involved in promoting cccDNA amplification, particularly under the B02 treatment condition. Although the efficient intracellular amplification pathway plays an important role in the maintenance of the cccDNA pool in the hepatocytes infected by duck hepatitis B virus (DHBV) and woodchuck hepatitis virus (WHV) (18, 61), the function of this pathway in HBV infection of hepatocytes in culture and in vivo in humans remains controversial. However, the cccDNA intracellular amplification does occur in cultured hepatocytes and HBV-infected humanized mice (62, 63). Our study thus raises the possibility that the status of HBV cccDNA amplification might be upregulated under certain pathobiological conditions, such as inflammation, oxidative stress, and endoplasmic reticulum (ER) stress (64), to subvert the loss of cccDNA and maintain the persistence of HBV infection. It will be interesting to investigate the effects of other cellular stress responses on HSPA1 expression and cccDNA synthesis in the future. A better understanding of cccDNA metabolism and regulation will pave the way toward discovery and development of therapeutics to eliminate or inactivate cccDNA and cure of chronic hepatitis B.

MATERIALS AND METHODS

Cell culture.

Human hepatoblastoma cell line HepG2 was purchased from ATCC and cultured in Dulbecco’s modified Eagle medium (DMEM)-F12 (1:1) supplemented with 10% fetal bovine serum (FBS) (Gibco), 100 U/mL of penicillin, and 100 μg/mL of streptomycin. HepAD38 was obtained from Christoph Seeger at Fox Chase Cancer Center, Philadelphia, Pennsylvania, USA (60). Tetracycline was removed from HepAD38 culture media to initiate pgRNA transcription and HBV replication as needed. The C3A^hNTCP cell line was derived from C3A, a subclone of HepG2 (ATCC HB-8065), and stably expressing human sodium taurocholate-cotransporting polypeptide (NTCP) (65). Construction and production of recombinant adenoviruses, Ad-HBV and Ad-HBVcoreI126A, were reported previously (10).

Antibodies, chemicals, and siRNAs.

Anti-RAD51 antibody (catalog no. D4B10) and anti-β-actin antibody (catalog no. 3700) were purchased from Cell Signaling Technology. Anti-HSPA1 (catalog no. GTX111088) was purchased from GeneTex. Anti-DNAJB1 (catalog no. 13174) and anti-DNAJB4 (catalog no. 13064) antibodies were purchased from Proteintech. All antibodies were used with 1:1,000 dilutions for Western blot assays. Foscarnet, B02, RI-1, RS-1, aphidicolin, VER155008, MKT-077, TRC051384, and MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide] were purchased from MilliporeSigma. JG-98 was purchased from MedChemExpress. DNA-PK inhibitor II was purchased from Santa Cruz Biotechnology. Cycloheximide was purchased from Cell Signaling Technology. CytoTox-Glo cytotoxicity assay kit was purchased from Promega. RAD51 siRNA was purchased from Origene, and all the other listed siRNAs were purchased from Dharmacon. The sequence information of siRNA used in this study is summarized in Table 1.

TABLE 1.

Targeting sequences of the siRNA used in this study

siRNAs	Manufacturer (catalog no.)	Sequence information
Rad51 siRNA duplex sequences	OriGene (SR303966)	rGrCrUrArUrGrUrArGrCrArArArGrGrGrArArUrGrGrGrUCT
		rCrCrArArCrGrArUrGrUrGrArArGrArArArUrUrGrGrArAGA
		rArGrArUrArArUrCrCrUrArGrArGrUrCrUrUrArArArGrCAT
HSPA1-A siRNA target sequence	Dharmacon (D-005168-01-0002)	GAGAUCGACUCCCUGUUUG
HSPA1-B siRNA target sequence	Dharmacon (D-003501-03-0002)	GAUCAACGACGGAGACAAG
DNAJB1 siRNA target sequences	Dharmacon (M-012735-02-0005)	GAAAGAGCAUUCGAAACGA
		GCUCUGAUGUCAUUUAUCC
		GCGAGAUCUUCGACCGCUA
		AAAUAUUGACCAUCGAAGU
DNAJB4 siRNA target sequences	Dharmacon (M-015678-00-0005)	GGAGAAGAAUUAUUGGAUA
		GAAGAGAUAUAUAGUGGUU
		CGCAGAAGCUUAUGAAGUA
		UUGUGUGGCUGCUCAAUUA

HBV DNA extraction and Southern blot hybridization analysis.

Cytoplasmic core DNA from HBV-replicating cell lines or recombinant adenovirus-infected cells were extracted as described previously. Protein-free HBV DNA species were extracted with a modified Hirt DNA extraction procedure (20, 41). Hirt DNA samples were denatured at 88°C for 8 min and chilled on ice. Such a procedure allows the complete denaturation of DP-rcDNA into single-stranded DNA (annotated as denatured DP-rcDNA), whereas cccDNA remains a double-stranded circular DNA. The heat-denatured Hirt DNA samples were further digested with EcoRI to linearize cccDNA into unit-length double-stranded linear DNA. If needed, Hirt DNA was digested by ExoI and -III to reveal minus-strand covalently closed rcDNA as shown in Fig. 4F. Both core DNA and Hirt DNA were resolved in 1.5% agarose gel electrophoresis, transferred onto Hybond-XL membrane, and hybridized with an α-³²P-UTP-labeled minus strand-specific full-length HBV riboprobe (66).

siRNA transfection.

HepAD38 cells were seeded into 6-well plates at a density of 6 × 10⁵ cells per well of a 12-well plate and cultured in the absence of Tet. Two days later, 2 mM PFA was added to the culture media. On day 4 postseeding, the cells were reseeded and transfected with 10 pmol siRNA oligonucleotides and 1 μL Lipofectamine RNAiMAX (Life Technologies) following the manufacturer’s protocol. Two days later, 1 μg/mL Tet was added into culture medium to stop pgRNA transcription from the HBV transgene, and PFA was removed from the culture medium to resume HBV DNA synthesis and cccDNA formation. Cells were harvested 24 h later. Gene silencing efficiencies were validated by Western blot assays.

Western blot assay.

Cells grown in a 12-well plate were lysed with 125 μL NuPAGE LDS sample buffer supplemented with 2.5% 2-mercaptoethanol. Cell lysates were subjected to heat denaturing at 100°C, loaded on NuPAGE 4 to 12% Bis-Tris Gel, and run with NuPAGE MOPS (morpholinepropanesulfonic acid)-SDS running buffer. Proteins were transferred from the gel onto a polyvinylidene difluoride (PVDF) membrane using iBlot 2 dry blotting system. Membranes were blocked with TBST (50 mM Tris-HCl, pH 7.6, 150 mM NaCl, and 0.1% Tween 20) containing 5% nonfat milk for 1 h and incubated with the desired antibody overnight at 4°C. After washing with TBST, the membrane was incubated with LI-COR IRDye secondary antibodies. Membranes were again washed with TBST and imaged with LI-COR Odyssey system.

Sequencing cccDNA gap junction region.

Hirt DNA was extracted from HepAD38 cells, which was followed by 88°C heat denaturation and EcoRI digestion to linearize cccDNA. After resolving with agarose gel electrophoresis, the linearized cccDNA molecules, which run at a 3.2-kb position, were retrieved and purified to serve as the PCR templates. The rcDNA gap junction region on cccDNA molecules was amplified by PCR using 5′-CTGAATCCTGCGGACGACC-3′ as forward primer and 5′-CGTACTGAAGGAAAGAAGTC-3′ as reverse primer. The PCR amplicons were purified and ligated to pGEM-T vector. About 50 colonies were selected for sanger sequencing. The sequence was aligned with the HBV strain ayw genome (GenBank accession no. NC_003977).

RNA-seq analysis.

HepAD38 cells were cultured in the presence of 2 mM PFA from day 2 to day 6 after Tet removal, followed by DMSO or B02 (20 μM) treatment for 7 h. Total RNA was extracted by TRIzol, and transcriptome sequencing (RNA-seq) analysis was performed. RNA sample quantity was measured by UV absorbance at 260, 280, and 230 nm with a NanoDrop 1000 spectrophotometer (Thermo Scientific). RNA integrity was determined by Bioanalyzer using an RNA Nano chip instrument (Agilent Technologies). We used 1,000 ng total RNA from each sample to make an mRNA-seq library using the Illumina TruSeq stranded mRNA library kit. Specifically, mRNAs were enriched twice via poly(T)-based RNA purification beads and fragmentated at 94°C for 8 min. The first-strand cDNA was synthesized by SuperScript II and random primers at 42°C for 15 min, followed by second-strand synthesis at 16°C for 1 h. Adapters with Illumina P5 and P7 sequences, as well as indices, were ligated to the cDNA fragment at 30°C for 10 min. After Ampure bead (BD) purification, a 15-cycle PCR was used to enrich the fragments. Sample libraries were subsequently pooled and loaded to the Illumina NextSeq 500 system. Raw sequence reads were aligned to the human genome (hg38), using the STAR (version 2.6.1) aligner and Gencode version 24 gene annotation. RSEM (version 1.2.31) was used to estimate gene and transcript abundance. DESeq2 was used to statistically assess expression changes in quantified genes in different conditions. Genes with a false-discovery rate <0.05 and fold change >2 were considered significant (67–69).

HBV virion production and HBV infection assay.

HBV virions were produced from culture media of HepAD38 cells and concentrated by 8% polyethylene glycol 8000 (PEG 8000) as described previously (65). For infection, C3A^hNTCP cells were pretreated with DMEM supplemented with 3% FBS, 1× nonessential amino acids (NEAA), (Gibco), and 2% DMSO 24 h before the virus infection. The cells were then infected with HBV at 500 genome-ofzy-equivalents (GOE) in DMEM containing 3% FBS, 1× NEAA (Gibco), 2% DMSO, and 4% PEG 8000 (Sigma; catalog no. P1458). The inoculums were removed at 12 h postinfection, and cells were washed with PBS 3 times before refreshing with DMEM containing 3% FBS, 1× NEAA (Gibco), and 2% DMSO. The infected cells were harvested at 12 or 48 h postinfection (hpi).

Statistical analysis.

Signal intensity in Southern blotting was scanned with ImageJ software. Data shown in the histograms were mean ± standard deviation. Data were analyzed using analysis of variance (ANOVA) and Student’s t test. The level of significance was set at a P value of <0.05.

Data availability.

RNA-seq data were deposited to the National Center for Biotechnology Information GEO database under the accession number GSE209863.