Hepatitis B virus DNA is subject to extensive editing by the human deaminase APOBEC3C

Authors

  • Thomas F. Baumert,

    Corresponding author
    1. Department of Medicine II, University of Freiburg, Germany
    2. Unit 748, Institut National de la Santé et de la Recherche Médicale, Strasbourg, France
    3. Louis Pasteur University, Strasbourg, France
    4. Service d'Hépatogastroentérologie, Hôpitaux Universitaires de Strasbourg, Strasbourg, France
    • Unit 748, Institut National de la Santé et de la Recherche Médicale, Louis Pasteur University, 3 Rue Koeberle, F-67000 Strasbourg, France
    Search for more papers by this author
    • fax: ++33-3 90 24 37 23

  • Christine Rösler,

    1. Department of Medicine II, University of Freiburg, Germany
    Search for more papers by this author
  • Michael H. Malim,

    1. Department of Infectious Diseases, King's College School of Medicine, London, United Kingdom
    Search for more papers by this author
  • Fritz von Weizsäcker

    Corresponding author
    1. Department of Medicine II, University of Freiburg, Germany
    • Department of Medicine 1, Schlosspark-Klinik, Humboldt University, Heubnerweg 2, D-14059 Berlin, Germany
    Search for more papers by this author
    • fax: ++49-30-3264-1300


  • Potential conflict of interest: Nothing to report.

Abstract

APOBEC3G (A3G) and APOBEC3C (A3C), 2 members of the APOBEC family, are cellular cytidine deaminases displaying broad antiretroviral activity. A3G inhibits hepatitis B virus (HBV) production by interfering with HBV replication without hypermutating the majority of HBV genomes. In contrast, A3C has little effect on HBV DNA synthesis. The aim of this study was to further dissect the mechanisms by which A3G and A3C interfere with the HBV life cycle. Immunoprecipitation experiments demonstrated that both A3G and A3C bind to the HBV core protein. A ribonuclease (RNase) treatment resulted in the nearly complete dissociation of the HBV core protein from A3G, whereas the HBV core-A3C complex was more stable. Interestingly, the majority of the newly synthesized HBV DNA genomes displayed extensive G-to-A mutations in the presence of A3C, whereas no A3C-induced HBV RNA mutations were detected. These findings support a model in which the RNA-dependent entrapment of A3G into the preassembly complex hampers subsequent steps in capsid formation. On the other hand, A3C is readily packaged into replication-competent capsids and efficiently deaminates newly synthesized HBV DNA. Conclusion: These findings demonstrate that HBV is highly vulnerable to the editing activity of an endogenous human deaminase and suggest that A3C could contribute to innate anti-HBV host responses. (HEPATOLOGY 2007.)

Hepatitis B virus (HBV), the prototypic member of the hepadnavirus family, is a major cause of liver disease worldwide, ranging from acute and chronic hepatitis to cirrhosis and hepatocellular carcinoma.1, 2 Other members of the family include the duck HBV and the woodchuck hepatitis virus.3, 4 Hepadnaviral replication involves the reverse transcription of a pregenomic hepatitis B virus RNA (pgRNA) intermediate inside nucleocapsids. This unique life cycle places HBV in the large family of retroelements, all of which require the reverse transcription of an RNA intermediate.

Recently, a novel cellular innate defense mechanism targeting a wide range of retroviruses has been identified.5 It has been shown that the propagation of human immunodeficiency virus 1 (HIV-1) strains lacking the accessory protein vif (virion infectivity factor) is suppressed in a number of nonpermissive cells and that this block is due to the expression of the cytidine deaminase APOBEC3G (A3G).6–8 Further studies have revealed that A3G induces massive C-to-U deamination of single-stranded retroviral DNA, which may result in the reduced accumulation of viral reverse transcripts and/or lethal G-to-A hypermutation.9–11 Recent evidence suggests that nonediting mechanisms may play an even more important role in A3G-mediated HIV restriction.12, 13 A3G contains an N-terminal RNA-binding domain and 2 catalytic deaminase domains in the A3G N-terminus and C-terminus, respectively.12, 14 An intact C-terminal mutator motif is essential for A3G mutator activity.14, 15

We and others have demonstrated that A3G can also markedly interfere with HBV production in transfected hepatoma cells.16, 17 In our previous study, we demonstrated that the early steps of the viral life cycle, including RNA and protein synthesis, the binding of pgRNA to the core protein, and the self-assembly of the core protein, were not affected by A3G.18 Ongoing reverse transcription in capsids that escaped the block in morphogenesis was also not significantly inhibited.18 However, A3G rendered HBV core protein–associated full-length pgRNA nuclease-sensitive and resulted in the down-regulation of replication-competent HBV nucleocapsids,18 and this suggests that A3G may interfere with HBV packaging. A3G-mediated editing did occur but was detected only in a minority of clones produced in transfected HepG2 hepatoma cells.17

APOBEC3C (A3C) is another member of the APOBEC family.5, 19 Unlike the larger A3G protein containing 2 deaminase domains, the smaller A3C protein contains a single deaminase domain. The impact of A3C on the HBV life cycle is poorly understood. A recent study by Bonvin et al.20 did not reveal a significant inhibition of HBV DNA synthesis when A3C was overexpressed in cotransfected hepatoma cells. Suspene et al.21 noted the hypermutation of a small fraction of HBV genomes in the presence of A3C.

In this study, we aimed to further dissect the mechanisms of A3G-mediated and A3C-mediated interference with the HBV life cycle. We demonstrate that the binding of A3G to the HBV core protein depends on the presence of RNA and is completely (A3G) or partially (A3C) sensitive to a treatment with RNase. The entrapment of A3G blocks proper capsid formation and subsequent HBV DNA synthesis. In contrast, A3C does not interfere with the formation of replication-competent nucleocapsids but results in extensive editing of packaged HBV DNA. These observations suggest that A3C is packaged into viral particles, and this allows it to edit newly synthesized HBV DNA. Our findings define a mechanism of AG3-induced and A3C-induced anti-HBV activity and suggest that A3C could contribute to innate anti-HBV host responses.

Abbreviations

A3C, APOBEC3C; A3G, APOBEC3G; CMV, cytomegalovirus; GFP, green fluorescent protein; HA, hemagglutinin fusion epitope; HBV, hepatitis B virus; HIV, human immunodeficiency virus; ORF, open reading frame; pgRNA, pregenomic HBV RNA; rcDNA, relaxed circular HBV DNA; ssDNA, single-stranded HBV DNA.

Materials and Methods

Plasmid Constructs.

A3G-HA is a cytomegalovirus (CMV)-driven expression vector encoding A3G fused with a hemagglutinin (HA) fusion epitope tag at its carboxy-terminal end.18 An A3C expression construct was kindly provided by J. Köck (Köck and Blum, submitted for publication). In brief, human A3C mRNA was amplified from human peripheral blood mononuclear cells by reverse transcription polymerase chain reaction (PCR). A HindIII restriction site was introduced into the forward primer upstream of the A3C start codon sequence (5′-GGGACAAGCTTATCTAAGAAGCTG-3′), and an XhoI restriction site was introduced into the reverse primer following the A3C stop codon sequence (5′-GACCTCGAGGCCCAGGGAGACCCC-3′). The PCR fragment was then ligated into pCDNA3.1 (Invitrogen). Expression construct pc-A3C-HA is a CMV-driven expression vector encoding A3C fused with an HA tag at its carboxy-terminal end. The pc-A3C-HA was generated by the replacement of A3B cDNA of expression plasmid pC-A3B-HA by A3C cDNA derived from pc-A3C. In brief, A3C cDNA was PCR-amplified from pc-A3C with forward primer 5′-GGCAGTACATCTACGTATTAG-3′ (corresponding to the CMV promoter sequence of pc-A3B and containing an SnaBI restriction site) and reverse primer 5′-CCCTCTAGACTGGAGACTCTCCCGTAGCC-3′ (containing an XbaI restriction site instead of the A3C stop codon sequence). Following the restriction digest with SnaBI and XbaI, the amplified PCR fragment was subcloned into pc-A3B-HA. The correct construction of plasmids was verified by the restriction digest, sequencing, and protein expression by western blotting. The wild-type HBV 1.3 overlength construct (genotype ayw) has been described before.22 For studying encapsidated RNA, the YMDD motif within the polymerase open reading frame of HBV 1.3 was modified to YMHD by site-directed mutagenesis (QuikChange, Stratagene).18 This mutation blocks the reverse transcription of pgRNA and prevents viral RNase H–mediated degradation of reverse-transcribed viral RNA.23 HBV core expression construct pCS1C1,24 expressing the HBV core in the absence of pgRNA, was generously provided by M. Nassal. The GFP expression vector (pGFP) and CMV expression vector (pCDNA3.1) were obtained from Invitrogen. Plasmid HBVadwR9pol−25 contains a naturally occurring missense mutation (A to C at nucleotide 2798) in the polymerase gene, which terminates HBV replication.26

Cell Culture, Transfection, and Harvesting.

Huh-7 or HepG2 human hepatoma cells were cultured and transfected as described.18 On day 3 or 4 after transfection, the cells were harvested, washed with PBS, pelleted, and resuspended in an iso-osmotic lysis buffer [140 mM NaCl, 1.5 mM MgCl2, and 50 mM Tris-HCl (pH 8.0) containing 0.5% Nonidet P-40]. Nuclei were removed by centrifugation for 5 minutes at 2000 rpm in an Eppendorf centrifuge, and the supernatant was cleared of cell debris by centrifugation for another 5 minutes at 14,000 rpm.

Analysis of Viral DNA and RNA.

An analysis of viral replication was performed by Southern blotting.18 A quantitative analysis of viral particles released into the supernatants of transfected HepG2 cells was performed by Southern blot analysis of particle-associated HBV-DNA.18 Viral particles released into the supernatant were partially purified by sedimentation through a 25% (wt/vol) sucrose TNE (10 mM Tris, 150 mM NaCl, and 1 mM ethylene diamine tetraacetic acid, pH 8.0) cushion by ultracentrifugation (140,000g for 16 hours at 4°C with a TST41 rotor; Centrikon T-1055 Ultracentrifuge, Kontron Instruments). HBV DNA from pelleted particles was isolated and analyzed by a Southern blot as described.18 HBV DNA levels were then quantified by densitometry with the FujiFilm BAS Imaging system (Fuji) and AIDA 3.20.116 software (Raytest). For the analysis of nuclease-resistant viral RNA, an aliquot of cytoplasmic lysate was treated with 5 units of micrococcal nuclease for 30 minutes at 37°C in the presence of 2 mM CaCl2. Thereafter, ethylene diamine tetraacetic acid was added to a final concentration of 5 mM, and the mixture was placed on ice before the addition of 0.7 mL chilled RLT lysis buffer and RNA purification.18 Nucleocapsid-associated viral DNA from cytoplasmic lysate was amplified with forward primer 556 (5′-TCCTTGGACTCATAAGGTGGG-3′) and reverse primer 2508 (5′-AGGTTCCACGCATGCGCTGAT-3′) by 30 cycles of 30 seconds at 94°C, 40 seconds at 55°C, and 80 seconds at 72°C. Nuclease-resistant packaged viral pgRNA from cytoplasmic lysates of cells transfected with replication-deficient YMHD-HBV was isolated with the RNeasy kit (Qiagen) and reverse-transcribed for 1 hour at 42°C with a random primer and reverse transcriptase (Roche) according to the manufacturer's protocol. HBV DNA reversed-transcribed in vitro from purified YMHD-HBV pgRNA was amplified by PCR with Taq polymerase (Roche) and HBV-specific primers as described.18 These PCR products were cloned into SphI/EcoRI of a pUC19 vector (Invitrogen). Individual clones of transformed DH5a were sequenced by SeqLab (Göttingen) with reverse primer 2201 (5′-GATTTTTTGTATGATGTG-3′) from SeqLab. The nomenclature of all primers is according to their position in the genome with respect to the core start codon.

Analysis of Core-APOBEC and APOBEC-RNA Binding by Coimmunoprecipitation Studies.

Cytoplasmatic lysates, prepared as described previously, were incubated for 4 hours at 4°C with protein A sepharose, coated with the polyclonal rabbit anti–HBV core protein antibody (BioTrend) for the immunoprecipitation of core and APOBEC proteins. Following low-speed centrifugation, the beads were washed repeatedly, and immunoprecipitated proteins were analyzed by western blotting with mouse anti-HA (16B12; Covence) for A3C and A3G or mouse anti-GFP mAb (8362-1, Clontech) for green fluorescent protein (GFP) followed by incubation with sheep anti-mouse immunoglobulin conjugated to horseradish peroxidase. The immunoprecipitated core was detected with an HBV core–specific rabbit antibody conjugated to horseradish peroxidase (PO312/PO158, a kind gift from M. Nassal). To assess whether the coimmunoprecipitation of the core, A3G, and A3C proteins was dependent on the presence of viral RNA, cytoplasmic lysates were preincubated with RNase A (0.5 μg/μL; Qiagen) for 1 hour at 37°C before immunoprecipitation.

Results

Binding of A3G and A3C to the HBV Core Protein.

To functionally characterize the mechanism of A3G inhibition of HBV production, we studied the interaction of A3G with the HBV ribonucleoprotein complex consisting of viral pgRNA, polymerase, and HBV core and cellular proteins.27 To facilitate the detection of APOBEC proteins, A3G and AC3 were HA-tagged. Although we cannot exclude that HA tagging may modify the ABOBEC structure and function in a subtle manner, the results of previous studies using HA-tagged members of the APOBEC family suggest that HA tagging does not preclude an analysis of their proper function.15, 28

Previous studies provide indirect evidence that A3G may be part of an intracellular macromolecular structure corresponding to the HBV ribonucleoprotein complex.20, 21 However, the mechanism of APOBEC-nucleocapsid interaction is not clearly understood. Following an optimization of our immunoprecipitation protocol18 using a different anti-core antibody for the immunoprecipitation of HBV nucleocapsids, we now demonstrate that A3G binds the HBV core protein in an RNA-dependent manner (Fig. 1A,B). Thus, the pretreatment of cell lysates with RNase A abolished core-A3G binding (compare lanes 4 and 9 in Fig. 1A). The expression of the HBV core protein from a CMV-driven expression vector was not sufficient for binding significant amounts of A3G under these experimental conditions (compare lanes 4 and 5 in Fig. 1A). These results indicate that (1) the expression of the HBV core protein from pgRNA facilitated the RNA-dependent recruitment of A3G and that (2) most A3G proteins associated with the HBV core protein were not packaged into intact nucleocapsids because the complex was accessible to exogenous RNase.

Figure 1.

Binding of A3G and A3C to the HBV core protein. HepG2 hepatoma cells were cotransfected with terminal-redundant, replication-competent HBV DNA (HBV1.3), HA-tagged A3C and A3G (A3C-HA and A3G-HA), and HBV core (CMV-HBVcore) expression constructs. Three days after transfection, the cells were lysed and processed for core immunoprecipitation and western blot, as described in the Materials and Methods section. (A,C) Coimmunoprecipitation of the HBV core and A3C or A3G. Cytoplasmatic lysates were incubated with an anti–HBV core protein antibody, and the immunoprecipitated proteins were analyzed by a western blot with HA-A3G–specific (upper panel) or anti–core-specific antibodies (lower panel). To assess whether the coimmunoprecipitation of the core and A3C or A3G proteins was dependent on the presence of viral RNA, an aliquot of cytoplasmic lysates was preincubated with RNase A before immunoprecipitation (lanes 6-10). (B,D) Expression of A3C, A3G, and core proteins. To control for the transfection efficiency, HBV core, A3C, and A3G protein expression was analyzed by a western blot of cell lysates with anti–HA-specific (upper panel) and anti–core-specific (middle panel) antibodies. Molecular weight markers are indicated on the left side. The numbers indicate units in kilodaltons. The detected proteins are depicted on the right side. [C,D (lower panel)] To demonstrate the specificity of HBV core immunoprecipitation, the cells were cotransfected with a GFP expression construct. The presence of GFP expressed in cell lysates and in proteins immunoprecipitated by an anti-core antibody was analyzed by a GFP-specific western blot, as described in the Material and Methods section (lower panels). The exposure time for the anti-GFP immunoblot of immunoprecipitated proteins was 30 minutes (left panel), and the exposure time for GFP expression in the total lysates was 1 minute (right panel).

Next, we studied the interaction of A3C, another member of the APOBEC protein family containing only 1 deaminase domain, with the HBV core protein. Similar to A3G (Fig. 1A), A3C was coimmunoprecipitated with the HBV core protein (Fig. 1C,D). Again, we observed that the A3C-HBV core interaction was best detectable when the HBV core was expressed from pgRNA rather than from a CMV-driven expression vector (compare lanes 4 and 5 in Fig. 1C). However, at higher levels of CMV-driven HBV core protein expression, A3C did coimmunoprecipitate even in the absence of pgRNA (data not shown). The pretreatment of lysates with RNase A also appeared to partially dissociate the A3C HBV complex. Interestingly, however, this effect was clearly less pronounced in comparison with A3G (compare lanes 4 and 9 in Fig. 1C). These findings suggest that a significant proportion of bound A3C molecules may be packaged into HBV nucleocapsids, rendering the complex resistant to nuclease attack.

Taken together, these data demonstrate that both A3G and A3C bind to the HBV core protein and that A3G recruitment is RNA-dependent. Furthermore, our data suggest that a significant proportion of A3C proteins but not A3G proteins may be packaged into intact HBV nucleocapsids. The cofactors required for A3G and A3C recruitment to the HBV core protein remain to be further characterized. Besides RNA, potentially important factors may include other viral proteins or cellular docking proteins. To address the potential role of the viral polymerase in A3C recruitment, we performed coimmunoprecipitation experiments, using a replication-deficient construct25 containing a mutation within the polymerase gene terminating replication.26 Interestingly, the A3C-core interaction was still present in cells transfected with the polymerase-mutant construct, suggesting that a fully replication-competent nucleocapsid is not required for A3C recruitment to the HBV core protein (Fig. 2).

Figure 2.

The binding of A3C to the HBV core protein is independent of an active viral polymerase. Huh-7 hepatoma cells were cotransfected with an HA-tagged A3C expression construct and replication-competent HBV DNA (adwR9) or a replication-deficient HBV DNA construct in the same adw background containing a mutation within the polymerase gene terminating viral replication (adwR9pol−).25, 26 Three days after transfection, the cells were lysed and processed for core immunoprecipitation, as described in Fig. 1. (A) After preincubation with RNase A, cytoplasmatic lysates were incubated with an anti–HBV core protein antibody, and the immunoprecipitated proteins were analyzed by a western blot with an anti–HA-specific antibody, as described in Fig. 1. (B) To control for the expression levels of transfected A3C-HA expression plasmid, the A3C-HA protein expression was analyzed by a western blot of cell lysates with an anti–HA-specific antibody. Molecular weight markers and detected A3C proteins are indicated on the left and right sides, respectively. The numbers indicate units in kilodaltons.

We next studied the consequences of APOBEC protein-nucleocapsid interactions on the accumulation of viral DNA replicative intermediates in transfected cells. Although A3G had a substantial effect on the HBV DNA levels [a 73% decrease in the intracellular HBV DNA levels versus cells cotransfected with HBV and pGFP (a mean of 2 independent experiments)], as reported previously,18, 29 A3C resulted in only a minor suppression of HBV replication [a 29% decrease in the intracellular HBV DNA levels versus cells cotransfected with HBV and pGFP (a mean of 2 independent experiments)], as shown in Fig. 3. Similar results were obtained for the effects of A3G and A3C on HBV particle release into the supernatant of transfected cells, as assayed by Southern blot analysis (Fig. 3; 88.5% and 50.5% decreases for A3G and AC3, respectively). These results suggest that A3G more efficiently inhibits HBV production than A3C, confirming and extending recent results reported by Bonvin et al.20

Figure 3.

Impact of A3C and A3G on the accumulation of cytoplasmic viral DNA replicative intermediates and viral particles released into the supernatant of transfected cells. HepG2 cells were cotransfected with terminal-redundant, replication-competent HBV DNA and expression constructs encoding GFP, A3C, or A3G as indicated in lanes 1–6 (APOBEC proteins without an HA tag). The accumulation of HBV replicative intermediates in cytoplasmic lysates (shown in lanes 1–3) was analyzed by Southern blot analysis of nuclease-resistant HBV DNA, as described in the Materials and Methods section. The release of viral particles into the culture supernatant of transfected cells (shown in lanes 4–6) was analyzed by a Southern blot of HBV DNA isolated from HBV particles purified by sucrose gradient ultracentrifugation from the culture supernatant. Molecular weight markers are indicated on the left. The numbers indicate kilobase pairs. Abbreviations: A3C/A3G, APOBEC3G/APOBEC3C expression vectors; C, cytoplasmic hepatitis B virus DNA; RC, relaxed circular DNA; S, particle-associated HBV DNA purified from cell culture supernatants; SS, single-stranded HBV DNA.

Extensive Editing of the Majority of HBV Genomes by A3C but not by A3G.

Because our data were compatible with the efficient packaging of A3C into HBV nucleocapsids, we asked the question whether packaged A3C was able to edit newly synthesized encapsidated HBV DNA. To search for A3C-mediated editing of HBV DNA in viral nucleocapsids, newly synthesized nucleocapsid-associated HBV DNA was PCR-amplified from nuclease-digested cell lysates of cotransfected Huh-7 cells, and this was followed by the sequencing of individual clones (Fig. 4). In the absence of A3C, G-to-A mutations were rare, regardless of the presence or absence of A3G, as shown previously.17, 18, 29 In contrast, in A3C expressing Huh-7 cells, the number of clones bearing G-to-A mutations and the overall number of G-to-A mutations increased dramatically, whereas other nucleotide substitutions were rare (Fig. 4).

Figure 4.

G-to-A mutations in newly synthesized HBV DNA produced in Huh-7 hepatoma cells in the presence or absence of A3C or A3G. (A) Huh-7 hepatoma cells were cotransfected with terminal-redundant, replication-competent HBV DNA and A3C or A3G expression constructs (without an HA tag). Nucleocapsid-associated HBV DNA was PCR-amplified, cloned, and sequenced with primer 5′-ACAGTAGCTCCAAATTCTTTA-3′ (about 300 nucleotides per clone). The footnotes indicate the total number of sequenced clones and the number of clones displaying G-to-A mutations. The boxes display the total number of respective mutations. (B) Nucleotide sequence of 2 individual PCR-amplified HBV clones produced in Huh-7 cells cotransfected with a replication-competent HBV construct and an A3C expression vector. Nucleocapsid-associated HBV DNA was PCR-amplified, cloned, and sequenced as described in part A. The mutations are depicted with respect to the wild-type sequence previously (wt). The asterisks represent the nucleotide identity. The numbers indicate the nucleotide positions with respect to the start codon of the core protein.

To further characterize the mechanism of A3C-mediated editing of HBV genomes, we performed parallel experiments, using an HBV variant containing a YMDD motif within the viral polymerase open reading frame modified to YMHD by site-directed mutagenesis. This mutation blocks the reverse transcription of pgRNA and its degradation by the viral RNase H.23 Because pgRNA derived from YMHD will not be reverse-transcribed in nucleocapsids, this mutant allows us to study selectively the effects of A3C on RNA editing. To study whether A3C was capable of pgRNA editing, we purified pgRNA from YMHD-HBV–transfected Huh-7 cells. YMHD-HBV pgRNA was then reverse-transcribed by the exogenous addition of reverse transcriptase in vitro, and YMHD cDNA was sequenced following PCR amplification. As shown in Fig. 5, the overall HBV RNA mutation rate was low and was similar in the absence and presence of A3C. The absence of A3C-mediated RNA editing demonstrates that A3C-mediated editing occurs at the level of HBV DNA, most likely single-stranded HBV DNA transcribed from viral pgRNA.

Figure 5.

A3C does not edit nucleocapsid-associated HBV RNA. Huh-7 hepatoma cells were transfected with HBV DNA containing a defect in the reverse transcription (HBV-YMHD) and an A3C expression construct (without an HA tag). To study whether A3C was capable of pgRNA editing, we purified pgRNA from YMHD-HBV–transfected Huh-7 cells. YMHD-HBV pgRNA was then reverse-transcribed by the exogenous addition of reverse transcriptase in vitro, and YMHD cDNA was sequenced after PCR amplification with primer 5′-ACAGTAGCTCCAAATTCTTTA-3′ (about 300 nucleotides per clone). The footnotes indicate the total number of sequenced clones and the number of clones displaying G-to-A mutations. The boxes display the total number of respective mutations.

Discussion

In summary, our findings demonstrate that the association of A3G to the HBV core protein is RNA-dependent. Our results are consistent with A3G-mediated interference with reverse transcription caused by destabilization of the preassembly complex and degradation of associated viral pgRNA by cellular RNases. Alternatively, the binding of A3G to pgRNA may block the priming of reverse transcription by, for example, blocking the respective polymerase binding sites on the viral pregenome. In contrast, the packaging of the small A3C protein did not destabilize the HBV ribonucleoprotein complex and allowed the reverse transcription of viral pregenomic RNA with extensive editing of newly synthesized DNA genomes. A possible explanation accounting for this difference may be the difference in the protein size between A3G (384 aa) and A3C (190 aa),30 with A3C containing only 1 deaminase domain and being significantly smaller than A3G.30 Alternatively, differences in the primary amino acid sequence may play an important role. Interestingly, A3C-induced deamination was observed in Huh-7 cells. In contrast, earlier results for A3G had identified A3G-induced mutations in transfected HepG2 but not in Huh-7 cells.17, 29 Conceivably, there are qualitative or quantitative differences in cellular cofactors influencing the formation and stability of the preassembly complex and replication-competent nucleocapsids.

Emerging evidence derived from the study of A3G-HIV and A3G-HLTV-1 interaction suggests that the binding of APOBEC to RNA has an important function for the facilitation of A3G and A3C packaging, in line with our results obtained for HBV. Indeed, it has been noted that the presence of RNA enhances A3G encapsidation into HIV nucleocapsids by promoting the stable association of A3G with HIV-1 nucleoprotein complexes.15

Interestingly, previous studies assessing the editing function of A3C on HBV genomes did not unravel the mechanism identified in this study.20, 21 The difference of these findings may be due to technical aspects. Thus, incomplete nuclease digestion of nonencapsidated HBV genomes in cell lysates21 or immunoprecipitated cores20 may have precluded a selective analysis of encapsidated HBV genomes. The presence and origin of nuclease [micrococcal nuclease (this study) versus deoxyribonuclease I used by Bonvin et al.20] may be important for the selective analysis of packaged HBV genomes. Alternatively, the HA tagging of A3C may influence the editing activity.

Finally, our results suggest that A3C could contribute to innate anti–HBV host responses. Two observations support the notion that this defense mechanism may play a role in virus-host interaction during HBV infection in vivo: (1) A3C is expressed within the human liver, albeit at low levels,20 and (2) edited HBV genomes containing G-to-A hypermutations are present in HBV-infected patients.31 In vivo, the A3C-induced editing of single-stranded retroviral DNA may result in defective viral particles or the emergence of viral variants. Because interferon-alpha has been shown to up-regulate A3C expression in primary human hepatocytes, it is conceivable that interferon-alpha may enhance the A3C-induced hypermutation of HBV DNA, thus contributing to interferon-induced inhibition of HBV replication. Although it is also conceivable that A3C has other prime retroviral or pararetroviral targets, our study demonstrates that HBV is highly vulnerable to the editing activity of an endogenous human deaminase.

Acknowledgements

We thank J. Köck for continuous support of this study, helpful discussions, and the gift of A3C expression plasmid; H. E. Blum for critical reading of the manuscript; and M. Nassal for valuable reagents and helpful discussions.

Ancillary

Advertisement