Pregenomic RNA Launch Hepatitis B Virus Replication System Facilitates the Mechanistic Study of Antiviral Agents and Drug-Resistant Variants on Covalently Closed Circular DNA Synthesis

ABSTRACT

Hepatitis B virus (HBV) replicates its genomic DNA by reverse transcription of an RNA intermediate, termed pregenomic RNA (pgRNA), within nucleocapsid. It had been shown that transfection of in vitro-transcribed pgRNA initiated viral replication in human hepatoma cells. We demonstrated here that viral capsids, single-stranded DNA, relaxed circular DNA (rcDNA) and covalently closed circular DNA (cccDNA) became detectable sequentially at 3, 6, 12, and 24 h post-pgRNA transfection into Huh7.5 cells. The levels of viral DNA replication intermediates and cccDNA peaked at 24 and 48 h post-pgRNA transfection, respectively. HBV surface antigen (HBsAg) became detectable in culture medium at day 4 posttransfection. Interestingly, the early robust viral DNA replication and cccDNA synthesis did not depend on the expression of HBV X protein (HBx), whereas HBsAg production was strictly dependent on viral DNA replication and expression of HBx, consistent with the essential role of HBx in the transcriptional activation of cccDNA minichromosomes. While the robust and synchronized HBV replication within 48 h post-pgRNA transfection is particularly suitable for the precise mapping of the HBV replication steps, from capsid assembly to cccDNA formation, targeted by distinct antiviral agents, the treatment of cells starting at 48 h post-pgRNA transfection allows the assessment of antiviral agents on mature nucleocapsid uncoating, cccDNA synthesis, and transcription, as well as viral RNA stability. Moreover, the pgRNA launch system could be used to readily assess the impacts of drug-resistant variants on cccDNA formation and other replication steps in the viral life cycle.

IMPORTANCE Hepadnaviral pgRNA not only serves as a template for reverse transcriptional replication of viral DNA but also expresses core protein and DNA polymerase to support viral genome replication and cccDNA synthesis. Not surprisingly, cytoplasmic expression of duck hepatitis B virus pgRNA initiated viral replication leading to infectious virion secretion. However, HBV replication and antiviral mechanism were studied primarily in human hepatoma cells transiently or stably transfected with plasmid-based HBV replicons. The presence of large amounts of transfected HBV DNA or transgenes in cellular chromosomes hampered the robust analyses of HBV replication and cccDNA function. As demonstrated here, the pgRNA launch HBV replication system permits the accurate mapping of antiviral target and investigation of cccDNA biosynthesis and transcription using secreted HBsAg as a convenient quantitative marker. The effect of drug-resistant variants on viral capsid assembly, genome replication, and cccDNA biosynthesis and function can also be assessed using this system.

INTRODUCTION

Hepatitis B virus (HBV) chronically infects 296 million people worldwide, and approximately 20% of the chronic HBV carriers will die from cirrhosis, hepatocellular carcinoma or liver failure, if not treated (1, 2). HBV is the prototype member of Hepadnaviridae family with a narrow host range (human being and chimpanzees) and hepatocyte tropism (3). Binding of HBV to sodium taurocholate cotransporting polypeptide (NTCP) on the plasma membrane of hepatocyte triggers endocytosis of the virion particle. The nucleocapsid is released into the cytoplasm upon the fusion of viral envelope and endosomal membrane (4, 5). HBV genomic DNA in the nucleocapsid is transported into the nucleus after the uncoating of nucleocapsid at the nuclear pore complex (6, 7). The partially double-stranded, relaxed circular DNA (rcDNA) genome is subsequently converted into a covalently closed circular DNA (cccDNA) by cellular DNA repair machinery (8, 9). The cccDNA associates with nucleosomes and assembles into a minichromosome (10), which is subsequently transcribed to yield the viral pregenomic RNA (pgRNA) and subgenomic RNA that can be translated to express core protein (Cp), DNA polymerase (Pol), HBV surface antigen (HBsAg), and X protein (HBx) (3). The reverse transcriptional replication of HBV DNA utilizes pgRNA as the template and begins by the binding of Pol to the stem-loop (ɛ) structure at the 5′ terminus of pgRNA (11, 12) to initiate the packaging of Pol-pgRNA complex by 120 copies of Cp dimers to form a nucleocapsid (13). Inside the nucleocapsid, Pol converts the pgRNA first to a single-stranded DNA and then to rcDNA (14, 15). The rcDNA-containing mature nucleocapsids can either acquire an envelope and be secreted out of cells as infectious virions or deliver the rcDNA into the nucleus to amplify cccDNA pool (16, 17).

HBV replication, as well as the evaluation and mechanistic studies of antiviral drugs, has primarily been investigated in human hepatoma cells transiently transfected with plasmid-based replicons or in hepatoma cell lines with stably integrated HBV DNA in cellular chromosomes to support constitutive or tetracycline-inducible transcription of HBV RNA and subsequent DNA replication (18–20). Although these systems support efficient HBV DNA replication and virion production, the presence of the large amount of transfected replicon DNA or integrated HBV transgene interferes with robust analysis of HBV replication, particularly the biosynthesis and function of cccDNA. The development of HBV infection cell culture systems with primary human hepatocytes and NTCP-expressing human hepatoma cells enables the investigation of HBV entry, de novo cccDNA synthesis, and transcriptional activation by HBx. However, the lack of efficient HBV virion production and cell-to-cell spread in the infected cultures prevents the selection and mechanistic analysis of drug-resistant variants (21).

Hepadnaviral pgRNA not only serves as a template for reverse transcriptional replication of viral genomic DNA but also expresses Cp and Pol, which are sufficient to support viral replication (22). It is thus conceivable that introduction of pgRNA into the cytoplasm of hepatocytes will initiate viral DNA replication and cccDNA synthesis, which is subsequently transcribed to yield all the viral RNA, followed by the synthesis of viral proteins and secretion of infectious virions. Indeed, Huang and Summers elegantly proved this hypothesis 3 decades ago by showing that expression of duck hepatitis B virus (DHBV) pgRNA from a Sindbis viral vector in the cytoplasm of hepatoma cells led to the robust viral replication and secretion of infectious virions (23). Recently, Yu et al. showed that transfection of in vitro-transcribed HBV pgRNA into Huh7.5 cells initiated HBV DNA replication, cccDNA synthesis, and secretion of surface antigen (HBsAg) (24). Encouraged by these findings, we further characterized HBV replication and HBsAg secretion kinetics in pgRNA-transfected Huh7.5 cells and demonstrated that the robust and synchronized HBV replication occurs within 48 h posttransfection is suitable for the precise mapping of HBV replication steps, from capsid assembly to cccDNA formation, targeted by distinct antiviral agents. On the contrary, treatment of cells starting at 48 h posttransfection allows assessment of antiviral mechanisms acting on post-DNA replication steps, such as uncoating of mature nucleocapsid and cccDNA transcription, with secreted HBsAg as a convenient quantitative marker. The complementary features of the early and late phases of HBV replication in pgRNA-transfected hepatoma cells should facilitate the discovery and development of antiviral drugs with novel mechanisms, including the targeting of HBx to silence cccDNA minichromosomes. Moreover, as demonstrated in this report, introduction of specific mutation(s) into pgRNA allows reverse genetic investigation of the HBV replication mechanism, as well as evaluation of drug-resistant variants on viral capsid assembly, genome replication, and cccDNA biosynthesis and function.

RESULTS

In vitro synthesis of pgRNA and establishment of a pgRNA-launch HBV replication system in Huh7.5 cells.

In order to assess the HBV pgRNA launch system, plasmids encoding wild-type pgRNA (WT), pgRNAs with an inactive Pol containing mutations at catalytic motif YMDD (Pol-YMVV), pgRNA with disrupted 5′ stem-loop (ɛ) structure (Δɛ), or a mRNA expressing Cp alone were constructed by insertion of a synthesized HBV DNA complementary to the full-length pgRNA without or with the indicated mutations or with HBV DNA corresponding to the Cp open reading frame into plasmid pUC59 immediately downstream of the T7 promoter (Fig. 1A). Plasmids were linearized by restriction enzyme digestion, and then viral RNAs were prepared via in vitro transcription, followed by in vitro 5′ capping and 3′ poly(A) tailing. The in vitro-transcribed RNA was transfected into Huh7.5 cells, and the production of viral replication intermediates was monitored over time (Fig. 1B). As revealed by a particle gel assay, HBV capsids were assembled in cells transfected with all four RNA species, in agreement with the fact that Cp is competent for the self-assembly of empty capsids in cells (25, 26). However, encapsidated pgRNA was only detectable in cells transfected with WT pgRNA or pgRNA/Pol-YMVV, which is consistent with the essential role of 5′ ɛ in pgRNA packaging (27) and the finding that polymerase activity is not required for pgRNA binding and encapsidation (28, 29). HBV DNA replication and cccDNA synthesis only occurred in cells transfected with WT pgRNA. Importantly, secretion of HBsAg was observed starting at 4 days posttransfection (dpt), which indicates that the cccDNA formed in the pgRNA-transfected cells was transcriptionally active (Fig. 1C). Further analysis of HBV particles in the culture media by a native agarose gel electrophoresis-based particle gel assay confirmed that HBV capsids were released from the cells transfected with each of the four RNA species (Fig. 1D, Capsid), whereas envelope particles were only secreted from the cells transfected with WT pgRNA (Fig. 1D, HBsAg). HBV DNA hybridization further showed while the empty capsids were released from all the transfected cultures, HBV DNA-containing capsids (nucleocapsids) were only released from cells transfected with WT pgRNA (Fig. 1D, DNA). The faint HBV DNA signal detected in the envelope particles suggests the secreted envelope particles from WT pgRNA-transfected cells are primarily subviral particles (SVPs). Altogether, these results clearly demonstrate that transfection of in vitro-transcribed pgRNA into Huh7.5 cells initiates a robust HBV DNA replication, establishing transcriptionally active cccDNA minichromosomes capable of supporting the expression of viral envelope proteins and secretion of HBsAg subviral particles.

FIG 1.

Establishment of a pgRNA-launch HBV replication system. (A) Schematic presentation of plasmids and the structure of their transcribed HBV RNA. (B) Huh7.5 cells were transfected with the indicated RNA transcripts. Culture medium was refreshed at 6 h posttransfection (hpt). The cells were harvested at 2, 4, 6, or 8 days posttransfection (dpt). The cytoplasmic HBV capsids and encapsidated viral RNA were detected by using a particle gel assay. HBV DNA replication intermediates and cccDNA were detected by Southern blotting hybridization. For cccDNA analysis, Hirt DNA preparations were denatured at 88°C for 8 min to convert DP-DNA into single-stranded DNA, followed by digestion with EcoRI to convert cccDNA into unit-length double-stranded linear DNA (3.2 kb). (C) Cell culture medium was harvested at 2, 4, 6, and 8 dpt for HBsAg detection by enzyme-linked immunosorbent assay (ELISA). Means and standard deviations from a representative biologically triplicate experiment are presented. (D) Culture media harvested from 4 to 8 dpt were pooled. HBV-derived particles in the culture media were concentrated by ultracentrifugation. Subviral particles (SVPs)/virions, secreted capsids, and their associated DNA were separated by 1.0% agarose gel electrophoresis and detected by immunoassay with the indicated antibodies, as well as hybridization with an [α-³²P]UTP-labeled riboprobe specifically hybridizing HBV minus-strand DNA. The transfected Huh7.5 cells were maintained in DMEM supplemented with 10% FBS, 0.1 mM NEAA, and 1× penicillin-streptomycin. rcDNA, relaxed circular DNA; dslDNA, double-stranded linear DNA; ssDNA, single-stranded DNA; ccc/EcoRI, EcoRI linearized cccDNA.

Introduction of two extra nucleotides at the 5′ terminus of pgRNA results in the preferential synthesis of double-stranded linear HBV DNA.

Careful inspection of HBV DNA replication intermediates presented in Fig. 1B (Core DNA panel) revealed that the mature double-stranded HBV DNA in pgRNA-transfected cells comigrated with 3.2-kb linear HBV DNA marker, and only a trace amount of HBV DNA was detected beyond the 3.2-kb double-stranded linear DNA band, presumably the circularized HBV genome, i.e., relaxed circular DNA (rcDNA). This result implies that the double-stranded linear DNA (dslDNA), but not rcDNA, was the predominant viral DNA species synthesized in the pgRNA-transfected cells.

As illustrated in Fig. 2A, the reverse transcriptional replication of HBV DNA begins with the binding of Pol at the 5′ ɛ of pgRNA to prime minus strand DNA synthesis. After copying of three nucleotides from the lower (priming) bulge of ɛ, Pol and the three-nucleotide DNA are translocated and annealed to the direct repeat 1 (DR1) region at the 3′ terminal redundancy sequence of pgRNA to resume the elongation of minus strand DNA (11, 27) (Fig. 2A, step a). While the reverse transcriptase domain of Pol copies the pgRNA into minus strand DNA, the RNase H domain of Pol cleaves the pgRNA after it has been copied (30) (step b). At the completion of minus strand DNA synthesis, the 17- to 18-nucleotide RNA oligonucleotide at the 5′ terminus of pgRNA escapes RNase H digestion (step e) and is translocated with Pol, annealing to the DR2 region to prime the synthesis of plus-strand DNA (31, 32) (step f). Upon the elongation of plus-strand DNA to the 5′ terminus of minus-strand DNA, a template switch circularizes HBV DNA and continues copying the minus-strand DNA to synthesize rcDNA (33, 34) (step g). However, a failure of Pol and RNA oligonucleotide translocation results in in situ priming of plus-strand DNA synthesis at the 3′ terminus of minus-strand DNA (step c) and the synthesis of dslDNA (35) (step d). In HBV-infected hepatocytes, the translocation of RNA primer and Pol is preferred, and only 5 to 10% of plus-strand DNA synthesis is primed in situ to yield dslDNA. By analyzing the sequence of plasmid for pgRNA transcription, we realized that the presence of the T7 promoter introduced two extra nucleotides (GG) to the 5′ terminus of the in vitro-transcribed pgRNA (Fig. 2B). This may result in the RNA oligonucleotide with two additional nucleotides being cleaved by RNase H at the completion of minus-strand DNA synthesis (36). Removal of two nucleotides from the 3′ end of RNA oligonucleotide will reduce its complementarity to DR2 region and may result in the failure or reduced efficiency of primer translocation and synthesis of dslDNA by in situ priming of plus-strand DNA synthesis (Fig. 2A). To investigate this possibility, three constructs encoding pgRNAs with deletion of two or three nucleotides from the 5′ terminus were generated to eliminate the potential effect of the extra GG introduced by in vitro transcription (Fig. 2B). In support of our hypothesis, and similar with the observation made in HepAD38 cells, all three pgRNAs with 5′ deletions of two to three nucleotides (resulting in GAC, GG, or GC at the transcriptional start site) predominantly produced rcDNA in transfected Huh7.5 cells, while the original full-length pgRNA (FL) predominantly produced dslDNA (Fig. 2C). Interestingly, comparing to HBV DNA replication intermediates in HepAD38 cells, Huh7.5 cells transfected with each of the in vitro-transcribed pgRNA accumulated much higher levels of single-stranded HBV DNA species. A plausible explanation to this phenomenon is that the A-to-G substitution or extra GG introduction at the 5′ terminus of the in vitro-transcribed pgRNA resulted in the change of the 3′-terminal sequence of minus-strand DNA, which may affect the template switching during plus-strand DNA elongation. Therefore, while the reduced RNA primer translocation results in the predominance of dslDNA, the interruption of template switching during plus-strand DNA synthesis may account for, at least in part, the apparent predominance of ssDNA over rcDNA (Fig. 2C). To minimize the change from the authentic pgRNA, GAC-pgRNA, which has only one nucleotide substitution (A to G) introduced by in vitro transcription, was used as WT pgRNA for all subsequent studies. The original WT pgRNA was renamed as pgRNA-DSL and can be used to investigate the biosynthesis of dslDNA, as well as the formation and function of dslDNA-derived cccDNA (37).

FIG 2.

Introduction of two extra nucleotides at the 5′ terminus of pgRNA results in the synthesis of primarily double-stranded linear HBV DNA. (A) Schematic presentation of the genesis of dslDNA and rcDNA from encapsidated pgRNA in nucleocapsids. (a) The pgRNA (dash line) is the template for the synthesis of minus-strand DNA from the 3′ direct repeat 1 (DR1) after priming at the 5′ ɛ structure of pgRNA. (b) The RNase H activity of polymerase protein (P) degrades the pgRNA during the synthesis of minus-strand DNA (thin solid line) and leaves a 17-nucleotide in length of RNA oligonucleotide at the 5′ end of pgRNA, which serves as the primer for plus-strand synthesis. (c to d) In vitro-transcribed full-length (FL) pgRNA has two extra “G” (purple) at its 5′ terminus from T7 promoter sequence. The extension at the 5’ terminus results in the removal of two additional nucleotides (blue with strikethrough) at the 3′ of RNA primer in the DR1 region by RNase H. The primer with shortened 3′ sequence cannot efficiently anneal with DR2 region and thus initiates the synthesis in situ at the 3′ terminus of minus-strand DNA to generate double-stranded linear dslDNA. (e to g) The RNA primer generated in the minus-strand DNA synthesis from authentic or 5′ shortened pgRNA (depicted in panel B of this figure) contains the complete sequence of DR1 (e), which allows the RNA primer to be translocated to the DR2 region at the 5′ end of minus-strand DNA to initiate the synthesis of plus-strand DNA (f), which is subsequently circularized to rcDNA (g). (B) Schematic presentation of constructs that transcribe pgRNA with the indicated 5′ sequences. T7 promoter sequence was present in the orange frame, and the transcription starts at the underlined G in the promoter sequence. The HBV pgRNA sequence is shown in the blank frame with the DR1 sequence shown in red. The gray bases with strikethrough are deleted in the indicated pgRNA transcripts. (C) Huh7.5 cells were transfected with the indicated pgRNA transcript and harvested at 24 hpt. HepAD38 cells were cultured in the absence of tetracycline (tet) for 6 days. Core DNA was extracted and detected by Southern blotting hybridization with an [α-³²P]UTP-labeled riboprobe specifically hybridizing HBV minus-strand DNA.

Production of HBsAg by pgRNA-transfected hepatoma cells depends on HBx protein expression and its interaction with DDB1.

Compared to plasmid-based HBV replicon-transfected hepatoma cells and the stable cell lines, HBV replication initiated by pgRNA transfection in Huh7.5 cells allows the analysis of cccDNA biosynthesis and transcription regulation without interference from HBV replicon or transgene DNA (20, 37). HBx is known to interact with DNA damage binding protein 1 (DDB1) and recruit the CUL4-DDB1 ubiquitin ligase machinery to promote the degradation of the cccDNA-associated SMC5/6 complex to activate HBV transcription (38, 39). Structural and functional analysis showed that HBx binds DDB1 through an α-helical motif spanning residues Ile88 to Leu100. The conserved Arg96 (R96) residue in the α-helical motif forms two hydrogen bonds with DDB1 and is essential for HBx interaction with DDB1 (40). The substitution of R96 with A or E abolishes its interaction with DDB1 and results in the failure of SMC5/6 complex degradation and cccDNA transcriptional activation (41). To investigate whether HBx is required for the transcriptional activation of cccDNA minichromosomes in the pgRNA launch system, Huh7.5 cells were transfected with WT pgRNA, pgRNA/Pol-YMVV, or mutant pgRNA deficient in HBx expression (ΔHBx) or expressing HBxR96A. On day 2 and day 6 posttransfection, capsids, HBV DNA replication intermediates, and the cccDNA were analyzed. Mitochondrial DNA (mtDNA) was also included as an internal control for cccDNA extraction. As shown in Fig. 3A, similar amounts of HBV capsids, DNA replication intermediates (core DNA), and cccDNA were accumulated in cells transfected with WT pgRNA, pgRNA-ΔHBx, or pgRNA-HBxR96A. As anticipated, HBV capsids, but not HBV DNA, were detected in cells transfected with the replication defective pgRNA/Pol-YMVV. On the contrary, HBsAg was predominantly detectable in the culture medium of cells transfected with WT pgRNA (Fig. 3B). These results strongly suggest that, while HBV DNA replication and cccDNA synthesis did not depend on HBx protein, the production of HBsAg is strictly dependent on HBV DNA replication and significantly enhanced by the expression of HBx protein. Furthermore, the results also indicate that HBx interaction with DDB1 is important for HBsAg secretion.

FIG 3.

Production of HBsAg by pgRNA-transfected hepatoma cells depends on HBx protein and its interaction with DDB1. Huh7.5 cells were transfected with in vitro-transcribed pgRNA of WT, HBx mutants (△HBx or HBxR96A), or Pol-YMVV mutant. Culture medium was refreshed (and harvested) every other day, and the cells were harvested at the indicated times posttransfection. For panels A and C, cytoplasmic capsids were detected by a native agarose gel-based particle gel assay. HBV DNA replication intermediates and cccDNA were detected by Southern blotting hybridization. Mitochondrial DNA (mtDNA) served as a loading control for cccDNA. The loading amount of HBV DNA were normalized to cell numbers in the long-term experiment (panel C). HBsAg in culture medium was measured by ELISA. Means and standard deviations from six biological replicates are presented in panels B and D.

However, the decrease in HBV capsids and core DNA and cccDNA levels from day 2 to day 6 in pgRNA-transfected cells (Fig. 3A) implied that the cccDNA in Huh7.5 cells did not transcribe a sufficient amount of pgRNA to support continuous HBV Cp and Pol expression and DNA replication. Due to the existence of residual input HBV pgRNA and degraded RNA fragments, cccDNA-derived HBV transcripts cannot be detected by a Northern blotting hybridization assay with confidence (data not shown). To further investigate the longevity of cccDNA and its transcriptional activity in the pgRNA launch system, the cells were cultured for 12 days post-pgRNA transfection. In agreement with the results presented above, both HBV core DNA and cccDNA declined as a function of culture duration, regardless of HBx expression (Fig. 3C). However, the levels of secreted HBsAg gradually increased starting on day 4 and peaking at day 10 post-pgRNA transfection (Fig. 3D). Interestingly, a much lower level of HBsAg was also detected in the media of cells transfected with pgRNA-ΔHBx with kinetics similar to that observed in the cultures transfected with WT pgRNA. In summary, these results imply that the cccDNA in pgRNA-transfected Huh7.5 cells cannot be stably maintained and that cccDNA transcription can be activated by HBx protein to express detectable levels of HBsAg. However, the molecular mechanism of the unsustainable cccDNA maintenance and viral replication in the pgRNA launch system remains to be investigated.

Characterization of cccDNA synthesized in pgRNA-transfected Huh7.5 cells.

Previous studies demonstrated that conversion of rcDNA into cccDNA by cellular DNA repair machinery perfectly repairs the junctions in both strands of rcDNA and results in the production of authentic cccDNA molecules (9, 42). On the contrary, synthesis of cccDNA from dslDNA is catalyzed by the error-prone nonhomologous end-joining (NHEJ) DNA repair pathway (8, 37), which results in the deletion and/or insertion at the junction region during dslDNA circularization (43–45). Here, we characterized the cccDNA pool in pgRNA-transfected cells by deep sequencing analysis, with cccDNA extracted from HBV-infected human hepatoma cells expressing NTCP (C3A^hNTCP) and primary human hepatocytes (PXB) transduced by a recombinant adenoviral vector expressing authentic HBV pgRNA (Ad-HBV) as controls (6). The results presented in Fig. 4 and Table 1 show that, in agreement with published results from hepadnavirus-infected hepatoma cells and liver tissues (43, 44), approximately 95% of cccDNA in HBV-infected hepatocytes had authentic junction sequences, and 5% of cccDNA had a deletion/mutation at the junction region. Also, as anticipated, Ad-HBV-transduced primary human hepatocytes accumulated predominantly (94%) authentic cccDNA. However, it is rather interesting that at day 2 posttransfection, 12.9 and 31.6% of cccDNA in Huh7.5 cells transfected with in vitro-transcribed WT- and pgRNA-DSL harbored the deletions and/or insertions at the junction region, respectively. Intriguingly, the proportion of cccDNA with junction mutations increased dramatically from 2 to 8 days post-pgRNA transfection. Those results indicate although the G-to-A substitution at the 5′ terminus of WT pgRNA restores rcDNA production (Fig. 2C), a significant fraction of cccDNA in the WT pgRNA-transfected cells are still synthesized from dslDNA. On the contrary, although pgRNA-DSL-transfected cells accumulated much less rcDNA and predominantly dslDNA (Fig. 2C), the majority of the cccDNA had authentic junction sequences and thus were most likely derived from the trace amount of rcDNA. Considering the dynamic change of WT and mutant cccDNA proportions in pgRNA-transfected cells, it is possible that the synthesis of cccDNA from rcDNA occurs faster and predominantly earlier, whereas the conversion of dslDNA into cccDNA has slower kinetics, which leads to the accumulation of mutant cccDNA at a later stage after pgRNA transfection.

FIG 4.

Deep sequence analysis of end-joining region of cccDNA. (A) Schematic illustration of the amplicon covering cccDNA junction for next-generation sequencing. (B) Proportions of WT cccDNA and cccDNA with mutations at the junction region from Ad-HBV-transduced PXB cells and harvested at day 6 postransduction, HBV-infected C3A^hNTCP cells at day 10 postinfection, Huh7.5 cells transfected with pgRNA, or pgRNA-DSL at 2 and 8 dpt, respectively. The DNA sequence data were aligned to wild-type genotype D genomic DNA (GenBank accession number U95551.1). The quantitative analysis was based on the sequencing data presented in Table 1.

TABLE 1.

Sequencing statistics for HBV cccDNA junction^a

Parameter	No.
	Ad-HBV	Infection	RC-2dpt	RC-8dpt	DSL-2dpt	DSL-8dpt
Total HBV reads	101,740	127,268	183,100	139,996	123,638	153,354
WT reads	96,264	119,043	159,382	100,694	84,504	81,416
Insertions	2,301	2,931	12,908	11,274	12,368	35,768
Deletions	2,141	3,174	7,222	24,326	23,350	30,785
Insertions and deletions	71	71	416	1,001	1,227	2,600
Base changes	963	2,049	3,172	2,701	2,189	2,785

^aColumn headings: Ad-HBV, primary human hepatocytes (PXB) transduced with recombinant Ad-HBV and harvested at day 6 posttransduction; infection, C3A^hNTCP infected with HBV derived from HepAD38 cells at day 10 postinfection; RC-2dpt and RC-8dpt, Huh7.5 cells transfected with WT pgRNA and harvested at days 2 and 8 posttransfection, respectively; DSL-2dpt and DSL-8dpt, Huh7.5 cells transfected with pgRNA-DSL and harvested at days 2 and 8 posttransfection, respectively.

HBV DNA replication is robust and synchronized early following pgRNA transfection.

The transient nature and distinct kinetics of HBV replication (early phase) and HBx-dependent HBsAg secretion (late phase) in the pgRNA launch HBV replication system may provide a unique opportunity to conveniently determine the effects of antiviral drugs on the different steps of the HBV replication cycle. To this end, we further characterized the early-phase kinetics of HBV replication in pgRNA-transfected Huh7.5 cells. As shown in Fig. 5A, Cp and capsids were detectable at 3 h posttransfection (hpt), While Cp level peaked at 12 hpt, capsids reached peak levels at 24 hpt. Single-stranded DNA (ssDNA) was detectable at 6 hpt and reached its peak at 24 hpt. The rcDNA and cccDNA became detectable at 24 hpt, and the cccDNA amount increased further in the next 24 h. A very similar HBV DNA replication kinetics was also observed in Huh7.5 cells transfected with pgRNA-DSL (Fig. 5B). These results indicate that the transfected pgRNA initiated a robust viral protein expression and capsid assembly in a relative short period time, which allows the viral DNA replication and cccDNA synthesis to occur almost in a synchronized fashion between 3 and 24 hpt. This system should thus allow for the precise determination of the effects of antiviral agents at each of HBV replication step between capsid assembly and cccDNA synthesis.

FIG 5.

Robust HBV DNA replication and cccDNA synthesis occurs upon transfection of pgRNA in Huh7.5 cells. Huh7.5 cells were transfected with in vitro-transcribed WT pgRNA (A) or pgRNA-DSL (B). Cells were harvested at 3, 6, 12, 24, and 48 hpt. The expression of HBV core protein (Cp) was measured by a Western blot assay with a rabbit polyclonal antibody HBc-170A. β-Actin served as a loading control. The capsids were resolved by 1.8% agarose gel electrophoresis, transferred to a nylon membrane, and detected by a mouse monoclonal antibody against HBV core protein. HBV core DNA and protein-free DNA (after heat denaturalization and EcoRI digestion) were detected by Southern blotting hybridization. rcDNA, relaxed circular DNA; dslDNA, double-stranded linear DNA; ssDNA, single-stranded DNA; ccc/EcoRI, EcoRI linearized cccDNA.

Effects of representative antiviral agents on HBV replication in the pgRNA launch HBV replication system.

To demonstrate the feasibility of the pgRNA launch HBV replication system in antiviral drug mechanistic studies, we tested the effects of a few antiviral agents with distinct antiviral mechanisms on HBV replication. The antiviral agents include type I and type II HBV capsid assembly modulators (CAMs) GLS4 (a 50% effective concentration [EC₅₀] of 64 nM in the reduction of core DNA) (46) and AB-506 (EC₅₀ of 35 to 77 nM in the reduction of core DNA) (47), respectively, HBV DNA polymerase inhibitor entecavir (ETV; EC₅₀ of 4 nM in the reduction of core DNA) (48), and HBV RNA destabilizer RG7834 (EC₅₀ of 2.8 nM in the reduction of secreted HBsAg) (49). The cells transfected with either WT pgRNA or pgRNA-DSL were treated with the antiviral compounds at the schedule depicted in Fig. 6A and were harvested at 48 hpt. Consistent with its mode of action, ETV did not affect Cp and capsids but potently inhibited HBV DNA replication in cells transfected with either WT pgRNA (Fig. 6A) or pgRNA-DSL (Fig. 6B). Only a small amount of incomplete single-stranded HBV DNA can be detected in ETV-treated cells (Fig. 6B and C). Since viral replication was a prerequisite for cccDNA synthesis from the pgRNA launch system, ETV treatment inhibited viral DNA synthesis and consequentially prevented the formation of cccDNA. Interestingly, although RG7834 induces the degradation of pgRNA and the preS1/preS2 envelope RNA in HBV-infected cells by inhibiting the noncanonical poly(A) polymerases PAPD5 and PAPD7 (50, 51), RG7834 treatment did not apparently affect any of HBV parameters detected. These results indicate that the compound did not apparently destabilize the transfected pgRNA in Huh7.5 cells. Concerning the two CAMs, while GLS4 induces HBV Cp dimers to assemble noncapsid aggregates that are subsequently degraded in cells (52, 53), AB-506 induces the assembly of structurally altered empty capsids without packaging of pgRNA (47). In agreement with their respective mode of action, GLS4 treatment reduced the level of Cp and inhibited the formation of capsids, whereas AB-506 treatment induced the accumulation of capsids with faster electrophoretic mobility, an indication of the global structure alteration of capsids (54). The residual amount of HBV DNA in GLS4- and AB-506-treated cells should be synthesized in the small amounts of capsids formed before the addition of CAMs and/or due to the incomplete inhibition of nucleocapsid assembly. These results also indicate that in addition to inhibiting pgRNA packaging, both GLS4 and AB-506 did not apparently inhibit the reverse transcriptional synthesis of viral DNA (55). Moreover, the undetectable amounts of rcDNA and detectable amounts of cccDNA in GLS4- and AB-506-treated cells are consistent with the findings that CAMs can induce the uncoating of mature nucleocapsids and rcDNA nuclear import, thus promoting cccDNA synthesis via the intracellular amplification pathway (6, 56).

FIG 6.

Effects of representative antiviral agents on HBV replication and cccDNA synthesis in a pgRNA launch HBV replication system (A) Schedule of antiviral agent treatment is depicted. Huh7.5 cells transfected with in vitro-transcribed WT pgRNA (B) or pgRNA-DSL (C) were treated with the indicated antiviral agents and collected at 48 hpt (blue arrow). Intracellular HBV core protein (Cp) was detected by a Western blot assay with the rabbit polyclonal antibody HBc-170A. β-Actin served as a loading control. The capsids were separated by 1.8% agarose gel electrophoresis, transferred to a nylon membrane, and detected by a mouse monoclonal antibody against HBV core protein. HBV DNA replication intermediates and cccDNA were detected by a Southern blotting hybridization assay.

To validate our interpretation on the residual HBV DNA observed in CAM-treated cells, the effects of CAMs on Cp, capsids, and HBV DNA were more carefully monitored within 24 hpt. As shown in Fig. 7, AB-506 treatment starting at 3 hpt did not affect the gradual increase in Cp and quickly induced the assembly of capsids with faster electrophoretic mobility compared to those from mock-treated cells. However, the amounts of HBV DNA replication intermediates, such as partially single-stranded DNA, single-stranded (SS) DNA, partially double-stranded DNA, dslDNA and rcDNA, were generally in proportion with the amounts of slower-migrating (normal) capsids between the mock-treated and AB-506-treated cells. These results imply that the residual amounts of HBV DNA in AB-506-treated cells are most likely due to incomplete inhibition of pgRNA packaging. On the contrary, GLS4 treatment stopped the increase in Cp and capsids and instead reduced the electrophoretic mobility of capsids. The latter finding is in agreement with our previous report that GLS4 binds to preassembled capsids to reduce the capsid electrophoresis mobility, presumably due to the induction of capsid global structure alteration (52). These results thus indicate that the capsids in GLS4-treated cells are most likely assembled before GLS4 action. The slight increase in single-stranded DNA in GLS4-treated cells could be due to the elongation of incomplete single-stranded DNA, but not the assembly of new nucleocapsids in the presence of GLS4.

FIG 7.

Effects of CAM treatment on the kinetics of Cp, capsids, and viral DNA replication intermediate accumulation in pgRNA-transfected Huh7.5 cells. (A) Schedule of CAM treatment. Huh7.5 cells were transfected with WT pgRNA and mock treated (DMSO) or treated with AB-506 (1 μM) or GLS4 (500 nM). Cells were harvested at 6, 12, and 24 hpt (blue arrows). (B) Intracellular HBV core protein was detected by a Western blot assay with a rabbit polyclonal antibody HBc-170A. β-Actin served as a loading control. The results were derived from three separate gels, as indicated. Intracellular capsids were detected by using a native agarose gel-based particle gel assay. HBV DNA replication intermediates were detected by Southern blotting.

To further assess the feasibility of pgRNA launch HBV replication system in CAM antiviral evaluation, we tested the effects of AB-506 and GLS4 on three clinically relevant mutant core proteins—CpT33N, CpI105T, and CpT109M. While CpT33N confers resistance to AB-506 and GLS4 by >300-fold (57, 58), CpI105T confers resistance to AB-506 and GLS4 by 20- and 2-fold (47), respectively. CpT109M does not confer apparent resistance to either AB-506 or GLS4 (58) but assembles capsids with a distinct phenotype. As anticipated, pgRNA encoding these Cp mutations remained sensitive to the inhibition by ETV, leading to suppression of HBV DNA replication (Fig. 8B). In agreement with previous reports, CpT33N mutation conferred complete resistance to both AB-506 and GLS4 compared to the WT in terms of Cp levels, capsids assembly, and viral DNA replication. Also consistent with previous reports, CpI105T mutation conferred partial resistance to AB-506 induction of faster mobility capsids and inhibition of HBV DNA replication but remained sensitive to GLS4 inhibition of capsid assembly and viral DNA replication. It is very interesting that, unlike the WT Cp that predominantly supports the assembly of slower-migrating capsids, CpT109M mutation leads to the assembly of slightly faster-migrating capsids even in the absence of compound treatment. Although the mutation did not confer apparent resistance to AB-506 and GLS4 inhibition of viral DNA replication, AB-506 treatment partially reversed the phenotype of assembled capsids toward the dominance of slower-migrating capsids. These latter results indicate that although the CpT109M mutation altered the structures of capsids assembled in the presence of AB-506, it does not significantly interfere with its inhibition of HBV DNA synthesis.

FIG 8.

Effects of clinically relevant Cp mutations on antiviral activity of representative antiviral agents in pgRNA-transfected Huh7.5 cells. (A) Schedule of antiviral agent treatment is depicted. Huh7.5 cells transfected with in vitro-transcribed pgRNA expressing WT or the indicated Cp mutants were treated with the indicated antiviral agents. Cells were harvested at 48 h post-pgRNA transfection (hpt) (blue arrow). (B) Intracellular HBV core protein was detected by a Western blot assay with a rabbit polyclonal antibody HBc-170A. β-Actin served as a loading control. The results were derived from two separate gels, as indicated. The capsids were separated by 1.8% agarose gel electrophoresis, transferred to a nylon membrane, and detected by a mouse monoclonal antibody against HBV core protein. HBV core DNA was measured by a qPCR assay, and the core DNA level was normalized to mock-treated cells transfected with the respective pgRNA. Means and standard deviations from four biological replicates are presented. **, P < 0.01 (versus mock-treated cells transfected with WT or respective Cp mutant HBV replicon).

Treatment of pgRNA-transfected cells after cccDNA establishment to evaluate the effects of antiviral drugs on capsid disassembly, cccDNA function, and viral RNA stability.

Our extensive characterization of HBV replication and HBsAg secretion kinetics in pgRNA-transfected cells demonstrated that robust HBV DNA synthesis was initiated within 24 hpt, with cccDNA reaching peak levels at ~48 hpt (Fig. 3 and 5). However, HBsAg production only became detectable at 96 hpt (Fig. 5). These results indicate that while cccDNA was synthesized primarily between 24 and 48 hpt, activation of cccDNA transcription and secretion of HBV subviral particles took an additional 1 to 2 days. In addition, due to the quick degradation of input pgRNA, new capsid assembly did not occur in a significant amount after 48 hpt. However, large amounts of assembled capsids containing various HBV DNA replication intermediates remained in the cells (Fig. 3). Therefore, treatment of pgRNA-transfected cells starting at or after 48 hpt provides a unique opportunity to investigate the effects of antiviral drugs on capsid stability, cccDNA transcriptional activation, and viral RNA stability, largely uncoupled from nucleocapsid formation and ongoing DNA synthesis.

Accordingly, as depicted in Fig. 9A, pgRNA-transfected Huh7.5 cells were mock treated or treated with the indicated antiviral agents, starting at 48 hpt for 6 days. The reduced accumulation of the fast-migrating capsids under AB-506 treatment, as well as the negligible reduction of capsids and the reduced capsid electrophoretic mobility under GLS4 treatment, further confirmed the notion that new capsid assembly did not occur in a significant amount by 2 dpt. Moreover, the results presented in Fig. 9B and C show that AB-506 and GLS4 treatment reduced the amounts of partially double-stranded DNA and rcDNA. However, while AB-506 treatment increased the amount of cccDNA and secreted HBsAg by ~2-fold at 6 dpt, GLS4 treatment did not significantly alter the amounts of cccDNA and secreted HBsAg. Interestingly, RG7834 treatment completely abolished the secretion of HBsAg, which is in agreement with its activity of destabilizing de novo-transcribed viral RNA in the nucleus (51, 59) through inhibition of the noncanonical poly(A) polymerases 5 or 7 (PAPD5 or PAPD7) (50, 60, 61). The reduction in HBV replication intermediates and HBsAg secretion in ETV-treated cells is most likely due to its nonspecific inhibition of cellular metabolism or cell growth under this experimental condition (Fig. 9C).

FIG 9.

Effects of antiviral agents on nucleocapsid stability and HBsAg secretion in the late phase of pgRNA launch HBV replication in Huh7.5 cells. (A) Schedule of antiviral agent treatment. Huh7.5 cells were transfected with in vitro-transcribed WT pgRNA and treated with the indicated antiviral agents, starting at 48 hpt. Huh7.5 cells were transfected with pgRNA expressing an inactive DNA polymerase (Pol-YMVV) as a negative control. Culture media were harvested at 4, 6, and 8 dpt (orange arrows). Cells were harvested at 8 dpt (blue arrow). (B) Intracellular capsids were detected by native agarose gel-based particle gel assay. HBV DNA replication intermediates and cccDNA were detected by Southern blotting. (C) HBsAg levels in the culture media were measured by ELISA from days 4, 6, and 8 post-pgRNA transfection (dpt). Means and standard deviations from eight biological replicates are presented.

Cp mutations conferring resistance to CAM misdirection of capsid assembly/pgRNA packaging also confer resistance to CAM induction of mature nucleocapsid uncoating and cccDNA synthesis.

CAMs bind to a hydrophobic HAP pocket in the Cp dimer-dimer interface to misdirect the assembly of Cp dimers and prevent the encapsidation of pgRNA and subsequent viral DNA replication. As demonstrated here and as reported previously, CAMs also induce the premature disassembly of double-stranded DNA-containing capsids to enhance intracellular DNA amplification (Fig. 9) (56). Treatment of pgRNA-transfected Huh7.5 cells after 2 dpt allows investigation of whether the Cp mutations conferring resistance to CAM misdirection of capsid assembly/pgRNA packaging also confer resistance to CAM induction of mature nucleocapsid uncoating and intracellular cccDNA amplification.

The results presented in Fig. 10 show that, as anticipated, AB-506 and GLS4 treatment induced the reduction of partially double-stranded DNA and rcDNA, as well as a slight increase of cccDNA and HBsAg secretion (≤2-fold) in cells transfected with WT pgRNA. However, the CpT33N mutation completely prevented AB-506- and GLS4-induced rcDNA reduction and cccDNA amplification. In agreement with the results shown in Fig. 8, CpI105T mutation conferred resistance to AB-506, but not GLS4. As expected, both AB-506 and GLS4 induced rcDNA reduction and a slight increase of cccDNA in cells transfected with pgRNA expressing Cp with T109M mutation. Interestingly, GLS4 treatment induced a slower migration of WT capsids and CpI105T capsids, but not CpT33N capsids. In addition, GLS4 only reduced the electrophoretic mobility of slow migrating capsids in cells expressing CpT109M. These results collectively indicate that, along with modulating encapsidation of pgRNA, CAMs disrupt the integrity of mature nucleocapsids and alter the global structure of assembled capsids through interaction with the same amino acid residues in the HAP pocket (52). Consistent with its mode of action, RG7834 treatment completely abolished the secretion of HBsAg from cells transfected with either WT or mutant Cp-expressing pgRNAs.

FIG 10.

Effects of clinically relevant Cp mutations on antiviral activity of representative antiviral agents in the late phase of pgRNA initiated HBV replication in Huh7.5 cells. (A) Schedule of antiviral agent treatment is depicted. Huh7.5 cells were transfected with in vitro-transcribed pgRNA expressing either WT or the indicated Cp mutants and treated with the indicated antiviral agents starting at 48 hpt. Culture media and cells were harvested at 6 dpt (purple arrow). (B) Intracellular capsids were detected by using a native agarose gel-based particle gel assay. HBV DNA replication intermediates and cccDNA were detected by Southern blotting. HBsAg levels in the supernatants were measured by ELISA. Means and standard deviations from a representative biologically triplicate experiment are presented. *, P < 0.05; **, P < 0.01 (versus mock-treated cells transfected with WT or respective Cp mutant HBV replicon).

DISCUSSION

We demonstrated in this report that transfection of pgRNA into Huh7.5 cells initiated a robust viral DNA replication and establishment of transcriptionally active cccDNA (Fig. 1 and 3). Interestingly, the kinetics of dslDNA accumulation in pgRNA-DSL-transfected cells are similar that of rcDNA accumulation in WT pgRNA-transfected cells (Fig. 5). These results indicate that although the synthesis of rcDNA requires two extra steps, RNA primer translocation, and template switch/genome circularization (Fig. 2A), rcDNA synthesis is not significantly slower than the synthesis of dslDNA. While the dslDNA is converted into cccDNA via the very powerful, but error prone NHEJ DNA repair pathway (37, 62), rcDNA is converted into cccDNA via the consecutive and perfect repair of minus-strand and then plus-strand DNA nicks by multiple host cellular DNA repair proteins (63–67). Interestingly, sequence analysis of cccDNA in pgRNA-transfected cells demonstrated differential accumulation kinetics of rcDNA-derived WT cccDNA and dslDNA derived cccDNA with deletions/mutations at the junction region (Fig. 4). Apparently, the molecular mechanism underlying this intriguing phenomenon should be further investigated. Moreover, it was shown previously that dslDNA, but not rcDNA, is the primary precursor of integrated HBV DNA in host cellular chromosomes (68). The integrated DNA can be detected shortly after viral infection of hepatocytes (69) and a significant portion of hepatocytes contain the integrated DNA in chronic HBV carriers (70). The integrated DNA is responsible for the majority part of HBsAg production in HBeAg-negative chronic HBV carriers (71, 72) and plays important roles in HBV-induced tumorigenesis (73). Based on those prior findings, the pgRNA-DSL launch replication may thus provide a unique opportunity for dissecting the mechanism of HBV DNA integration and exploring the possible means of therapeutic intervention.

It had been shown that cccDNA synthesized during de novo HBV infection of hepatocytes is transcriptionally silenced by association with host cellular SMC5/6 complexes. Leaked transcription of 0.7-kb HBx RNA, but not other viral RNA species, from silenced cccDNA leads to the expression of HBx protein. HBx interaction with DDB1 recruits CUL4-DDB1 ubiquitin ligase machinery to induce degradation of the SMC5/6 complex (38, 39), allowing for transcription of the cccDNA minichromosome. Due to the presence of an integrated HBV genome, or of input HBV DNA in plasmid-transfected cells, the transcriptional activation and epigenetic regulation of cccDNA synthesized from rcDNA in progeny nucleocapsids (i.e., the intracellular amplification pathway) cannot been investigated without the interference of transgenes. We discovered recently that synthesis of cccDNA from de novo infection versus intracellular amplification pathways differs in nucleocapsid uncoating at distinct subcellular compartments and in requirements for distinct cellular DNA polymerases (6, 66, 74). It remains to be determined whether cccDNA synthesized via the intracellular amplification pathway is also silenced by SMC5/6 complex and requires HBx for transcriptional activation. By leveraging the unique attributes of the pgRNA launch HBV replication system, we clearly demonstrated here that, similarly to the cccDNA synthesized from de novo infection pathway, cccDNA formed through the intracellular amplification pathway also relies on HBx activation via interaction with DDB1 (Fig. 3).

However, in contrast to the cccDNA formed from de novo infection of hepatocytes, the amount of cccDNA formed in pgRNA-transfected Huh7.5 cells begins to gradually decrease from 2 to 4 dpt. Capsid assembly and DNA synthesis also did not occur at significant amounts after 2 to 4 dpt (Fig. 1, 3). Although the HBx-dependent transcriptional activation of cccDNA in pgRNA-transfected Huh7.5 cells led to the secretion of HBsAg (subviral particles), no evidence suggests that pgRNA/pre-C mRNA transcription and HBeAg secretion occurred at a significant level. Accordingly, the transient HBV replication in pgRNA-transfected Huh7.5 cells is most likely due to unstable and/or incomplete transcriptional activation of cccDNA. Uncovering the underlying molecular mechanisms may lead to the establishment of a cell culture system in which cccDNA supports persistent HBV replication and virion production.

It was found recently that, through direct binding of the cellular zinc finger protein ZCCHC14 to the posttranscriptional regulatory element (PRE) located in the overlapped region of 3.5-, 2.4-, and 2.1-kb HBV RNA, PAPD5, or PAPD7 are recruited to viral RNA to extend the 3′ poly(A) tail and prevent its degradation (50, 75, 76). Inhibition of PAPD5 and PAPD7 polymerase activity by RG7834 accelerates the degradation of HBV RNA and inhibits the secretion of HBsAg (51). As anticipated, RG7834 treatment of pgRNA-transfected Huh7.5 cells completely abolished HBsAg secretion (Fig. 9 and 10), presumably by accelerating the degradation of the viral envelope RNA transcribed from cccDNA. Interestingly, treatment of Huh7.5 cells between 24 h before and 48 h after pgRNA transfection with RG7834 apparently did not alter the levels of Cp, capsids, viral DNA, and cccDNA (Fig. 6). This result indicates that RG7834 treatment did not accelerate the decay of transfected pgRNA. Although these observations are consistent with the HBV RNA degradation being localized to the nucleus, it is possible that other factors may be at play, and thus the differential effects of RG7834 on transfected in vitro-transcribed pgRNA versus viral RNA-transcribed de novo from cccDNA still warrant further investigation.

Although the pgRNA launch HBV replication in Huh7.5 cells does not support continual viral replication, the robust and almost synchronized viral DNA replication within 24 hpt, coupled with the rapid establishment of the cccDNA pool between 24 and 48 h post-pgRNA transfection, permits detailed molecular analysis of HBV replication events from capsid assembly to cccDNA synthesis (Fig. 5). For instance, this experimental condition is suitable for the identification and mechanistic analysis of host cellular proteins that regulate HBV replication and cccDNA synthesis. We also demonstrate in this report the effects of representative antiviral agents on Cp stability, capsid assembly and disassembly (uncoating), viral DNA replication, and cccDNA formation (Fig. 6 to 10). More importantly, although no significant viral DNA replication and cccDNA synthesis take place after 48 hpt, the transcriptional activation of cccDNA and HBsAg secretion in the late phases of the pgRNA launch replication system provide a unique opportunity to investigate the mechanism of cccDNA transcriptional silence and HBx-mediated activation. This system should also allow for the identification of novel antiviral agents that induce cccDNA silencing, with secreted HBsAg as a convenient quantitative readout.

Finally, multiple CAMs are currently in clinical development for treatment of chronic hepatitis B (77). By binding to a hydrophobic pocket (HAP pocket) between Cp dimer-dimer interface, CAMs misdirect the assembly of Cp dimers into noncapsid polymers or empty capsids devoid of pgRNA, thus precluding the synthesis of viral DNA. In addition, CAMs can also bind to assembled capsids to alter the global structure of capsids or induce the premature uncoating of double-stranded DNA-containing nucleocapsids (56, 78). Although the premature uncoating of incoming virion nucleocapsids results in the degradation of viral DNA in the cytoplasm and inhibition of cccDNA synthesis, the accelerated uncoating of the cytoplasmic progeny nucleocapsids usually results in the degradation of majority of the uncoated HBV DNA by cytoplasmic nucleases, including TREX1 (6), a small fraction of uncoated HBV DNA are imported into the nuclei, as indicated by the presence of protein-free DNA (PF-DNA) or deproteinized DNA (Dp-DNA) (Fig. 9B and 10B), for cccDNA synthesis (56, 58). However, although it is apparent that the removal of viral DNA polymerase from rcDNA ought to be essential, whether the experimentally detected PF-DNA or DP-rcDNA in HBV replicating cells are true precursor of cccDNA synthesis remain controversial (79–81). Nevertheless, whether CAM treatment accelerates or inhibits the intracellular cccDNA amplification is determined by the effects of three confounding factors, i.e., the subcellular location of uncoating, extent of cytoplasmic nuclease digestion and efficiency of nuclear import of uncoated HBV DNA (6, 56). The differential effect of AB-506 and GLS4 on cccDNA amplification as observed in the results presented in (Fig. 9 and 10) is most likely due to the difference in the subcellular location and extent of mature nucleocapsid uncoating induced by the two CAMs in Huh7.5 cells.

Using a panel of CAM-resistant Cp mutations, we demonstrated recently that these Cp mutations confer resistance not only to CAM misdirection of capsid assembly and viral DNA synthesis but also to CAM-induced capsid structural alterations, mature nucleocapsid uncoating, and inhibition of de novo cccDNA synthesis (52). Taking advantage of the pgRNA launch HBV replication system, we further demonstrated that CAM-resistant mutations also conferred resistance to CAM-induced disassembly of cytoplasmic progeny nucleocapsids and acceleration of intracellular cccDNA amplification (Fig. 9 and 10). These results collectively support the notion that through interaction with the same Cp residues in the HAP pocket, CAMs disrupt not only the assembly of capsids to inhibit pgRNA packaging and viral DNA replication but also the structure and function of assembled capsids to modulate nucleocapsid uncoating and cccDNA synthesis via both de novo infection and intracellular amplification pathways.

MATERIALS AND METHODS

Cell cultures and chemicals.

Human hepatocellular carcinoma cell line Huh7.5 was obtained from Charlie M. Rice at Rockefeller University in New York and cultured in Dulbecco modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum (FBS), l-glutamine, nonessential amino acids, penicillin, and streptomycin (82). HepAD38 cell line was obtained from Christoph Seeger at Fox Chase Cancer Center, Philadelphia, PA, and was maintained in DMEM/F-12 (1:1) supplemented with 10% FBS, penicillin, streptomycin, and 1 μg/mL of tetracycline (19). To induce pgRNA transcription, HepAD38 cells were cultured in the absence of tetracycline. The C3A^hNTCP cell line was derived from C3A, a subclone of HepG2 (ATCC HB-8065), and stably expressed human sodium taurocholate cotransporting polypeptide (NTCP) (83). PXB cells, the human hepatocytes prepared from chimeric mice (PXB mice) with humanized livers that are highly repopulated by human hepatocytes were purchased from PhoenixBio (Higashi-Hiroshima, Japan) and cultured with Advanced DMEM/F-12 (Life Tech, catalog number 12634-010) supplemented with 2% FBS, SingleQuots (Lonza, catalog number CC-4175), 2% dimethyl sulfoxide (DMSO), 2 mM glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. Entecavir (ETV) and capsid assembly modulators AB-506 and GLS4 were obtained from Arbutus Biopharma, Inc. RG7834 was purchased from Medkoo Biosciences (catalog number 563793).

Plasmids.

All the plasmids used in in vitro transcription of pgRNA or mRNA for expression of HBV core protein were synthesized at GenScipt, Piscataway, NJ, USA. Briefly, HBV DNA complementary to pgRNA (a genotype D isolate, GenBank accession number U95551.1), spanning nucleotides 1820 to 1916 was synthesized and cloned into plasmid pUC59 at the immediate downstream of T7 promoter to yield pUC59-pgRNA. Substitution of polymerase D540D541 codons (GATGAT) with GTTGTT resulted in plasmid pUC59-pgRNA/Pol-YMVV. Replacement of GTTC at nucleotides 1864 to 1867 with TAAG results in plasmid pUC59-pgRNA/Δε. Substitution of HBx starting codon (ATG) with TTG and R96 codon (AGG) with GCG yielded plasmids pUC59-pgRNA/ΔHBx and pUC59-pgRNA/HBxR96A, respectively. HBV DNA complementary to the core protein coding region (nucleotides 1903 to 2454) was synthesized and cloned into pUC59 at the immediate downstream of T7 promoter to yield plasmid pUC59-HBc.

Synthesis and purification of HBV pgRNA.

In vitro transcription of HBV RNA from plasmid templates was performed as described previously, with minor modifications (24). Briefly, the pUC59-derived plasmids described above were linearized by AseI digestion. The reaction was terminated by adding a 1/20 volume of 0.5 M EDTA, and DNA was precipitated by adding a 1/10 volume of 3 M sodium acetate and 2 volumes of ethanol. In vitro transcription was performed using T7 mScript standard mRNA production system by following the protocol provided by the manufacturer (CellScript, catalog number C-MSC100625). Upon completion of the transcription reaction, the DNA templates were removed by DNase I digestion at 37°C for 15 min. RNA was purified using an RNeasy minikit (Qiagen, catalog number 74014) according to the manufacturer’s instructions, including the optional on-column DNase digestion step (Qiagen, catalog number 79254). Capping and polyadenylation were performed following the Cap 1 mRNA protocol described in the user’s manual and the RNA was again purified using the RNeasy minikit without on-column DNase digestion.

RNA transfection.

Huh7.5 cells were seeded at 2.5 × 10⁵ cells per well in 12-well plates 2 days before transfection. The medium was changed to 1 mL of DMEM containing 1.5% FBS and 0.1 mM nonessential amino acids (NEAA; Gibco, catalog number 11140050) per well just before transfection. For each well, 0.5 μg of HBV pgRNA was mixed with 2.5 μL of Lipofectamine 2000 (Fisher Scientific, catalog number 11668019) in 250 μL of Opti-MEM reduced serum medium (Fisher Scientific, catalog number 51985034), followed by incubation at room temperature for 5 min. The mixture was then added to cells and spinoculated by centrifugation at 1,000 × g for 30 min at 37°C. Six hours later, the media were replaced with DMEM maintenance medium supplemented with 3% FBS, 0.1 mM NEAA, 1% GlutaMAX (Gibco, catalog number 35050079), 1× insulin-transferrin-selenium (Gibco, catalog number 41400045), and 1× penicillin-streptomycin (Gibco, catalog number 15140122). Culture medium and cells were harvested at the indicated times posttransfection.

HBV DNA extraction and Southern blot hybridization analysis.

Cytoplasmic core DNA from both HepAD38- and pgRNA-transfected Huh.75 cells was extracted as described previously (20). Protein-free HBV DNA were extracted with a modified Hirt DNA extraction procedure (84). Both core DNA and Hirt DNA were resolved by using 1.3% agarose gel electrophoresis, transferred onto Hybond-XL membrane, and hybridized with an [α-³²P]UTP-labeled minus-strand specific full-length HBV riboprobe.

Western blot assay.

Huh7.5 cells in each well of 12-well plate were lysed by 150 μL of 1× LDS loading buffer (Invitrogen, catalog number NP0007). Cell lysate was boiled at 100°C for 20 min and resolved in a NuPAGE 12% Bis-Tris protein gel (Invitrogen, catalog number NP0342PK2) using MOPS SDS running buffer (Genscript, catalog number M00138). Proteins in the gel were transferred onto a polyvinylidene difluoride (polyvinylidene difluoride) membrane (Thermo Fisher, catalog number IB24001). The membrane was blocked with TBS buffer (Thermo Fisher, catalog number BP2471400) containing 0.1% Tween 20 and 10% nonfat milk and then probed with rabbit polyclonal antibody against the C-terminal 14-amino-acid peptide HBV Cp (85). The bound antibody was revealed by horseradish peroxidase (HRP)-conjugated secondary antibodies and imaged using the Bio-Rad ChemiDoc touch imaging system.

Particle gel assay.

Intracellular or secreted HBV particles were detected by a native agarose gel electrophoresis-based particle gel assay as reported previously, with minor modifications (20, 86). Briefly, Huh7.5 cells in 12-well plates were lysed by the addition of 300 μL/well of lysis buffer (50 mM Tris-HCl [pH 7.5], 1 mM EDTA, 150 mM NaCl, 1% NP-40) and incubated for 10 min at room temperature. After removal of the nuclei by centrifugation at 10,000 × g for 10 min at 4°C, 18 μL of the cell lysate was mixed with 2 μL of 10× DNA loading buffer and loaded into 1.8% native agarose gel. Culture media harvested from day 4 to day 8 posttransfection were pooled and loaded onto a 20% sucrose cushion and centrifuged at 28,000 rpm (SW28) at 4°C for 16 h. Pellets were resuspended with TNE buffer (10 mM Tris-HCl [pH 7.6], 100 mM NaCl, 1 mM EDTA) and loaded into a 1% native agarose gel. Capsids or secreted virions and subviral particles were transferred onto nitrocellulose membrane with TNE buffer overnight. The membrane was fixed with 2.5% formaldehyde in phosphate-buffered saline (PBS) at room temperature for 20 min, followed by fixation with 50% methanol for an additional 20 min, followed in turn by washing with water twice. The membrane was blocked with TBS containing 0.1% Tween 20 and 10% nonfat milk and probed with an antibody against core protein (Santa Cruz, catalog number sc-52406) or HBsAg (Fitzgerald, catalog number 10-H05H). The bound antibody was revealed by anti-mouse HRP secondary antobody and visualized with a Bio-Rad Imager (Bio-Rad). Capsid-associated viral DNA and encapsidated RNA were detected with [α-³²P]UTP-labeled full-length HBV positive-strand and negative-strand RNAs, respectively.

ELISA.

HBsAg in culture medium was quantified using a commercial HBsAg chemiluminescence immunoassay kit (Autobio, China) according to the manufacturer’s instructions.

Quantification of HBV DNA by qPCR assay.

Intracellular HBV core DNA was quantified by a real-time PCR assay using LightCycler 480 SYBR green I Master PCR kit (Roche) with primers 5′-GGCTTTCGGAAAATTCCTATG-3′ (sense) and 5′-AGCCCTACGAACCACTGAAC-3′ (antisense). The PCR was carried out as follows: denaturation at 95°C for 5 min, followed by 40 cycles of amplification consisting of 95°C for 15 s and 60°C for 30 s.

HBV infection.

For HBV infection, C3A^hNTCP cells were seeded into collagen-coated 12-well plates at a density of 1.5 × 10⁶ cells per well and cultured in complete DMEM containing 2% DMSO. One day later, the cells were infected with HBV prepared from HepAD38 cell culture media at a multiplicity of infection (MOI) of 1,000 in DMEM containing 4% PEG-8000. The inocula were removed at 24 h, and the cells were washed three times with PBS. The infected cultures were maintained in complete DMEM containing 2% DMSO to for 10 days postinfection, and the medium was refreshed every 3 days.

Construction and transduction of recombinant adenovirus.

Construction of the recombinant adenovirus expressing wild-type HBV pgRNA (Ad-HBV) and the preparation of Ad-HBV were performed as described previously (6). PXB cells were transduced with Ad-HBV at an MOI of 10 and cultured in the absence of tetracycline for 6 days.

Next-generation sequencing for cccDNA junction.

Hirt DNA was extracted and treated with T5 exonuclease (NEB, catalog number M0363S) at 37°C for 1 h (84). The cccDNA junction region was amplified using AccuPrime Taq DNA polymerase (Thermo Fisher, catalog number 12339016) with the primers 5′-CCGACCTTGAGGCATACTT-3′ and 5′-GGAACCCACCGAAACCCCG-3′ (Fig. 4A). The PCR product was resolved in 1% agarose gel, and a single band (211 bp) was extracted using a MinElute gel extraction kit (Qiagen, catalog number 28606). The recovered amplicon DNA was sent for next-generation sequencing at Azenta, Inc.

Statistical analysis.

Statistical significance was calculated using ordinary one-way analysis of variance performed with SPSS. The error bars in the figures denote standard deviations.