Technical Comments

Response to Comment on "Inhibition of Hepatitis B Virus Replication by APOBEC3G"

See allHide authors and affiliations

Science  03 Sep 2004:
Vol. 305, Issue 5689, pp. 1403
DOI: 10.1126/science.1101974

Rösler et al. (1) nicely confirm that APOBEC3G blocks hepatitis B virus (HBV) replication by suppressing viral DNA synthesis and further reveal that this enzyme can occasionally edit the viral genome in the particular context of HepG2 cells. Even in this case, when APOBEC3G was overexpressed, fewer than one out of ten HBV genomes became hypermutated. Although these results demonstrate that HBV is not immune to APOBEC3G-mediated editing, we do not favor the hypothesis of Rösler et al. that cells in which blockage of HBV DNA synthesis occurs without noticeable editing lack a cofactor important for cytidine deamination or produce a suppressor of this activity. Indeed, we verified that APOBEC3G can efficiently hypermutate vif-defective human immunodeficiency virus–1 released from such cells, including Huh7 [(2) and data not shown]. Therefore, we think that the Rösler et al. data are more consistent with a model in which, in HepG2 cells, APOBEC3G is slightly less efficient at blocking HBV pregenomic RNA packaging or at destabilizing the HBV reverse transcription complex, so that minusstrand viral DNA is occasionally made. This DNA can then serve as a target for APOBEC3G-mediated editing. In Huh7 cells, the cytidine deaminase blocks DNA synthesis completely, thus depriving itself of its editing substrate.

Still, these new data lend credence to the suggestion that HBV editing might occur in certain tissues and, as such, contribute to HBV pathogenesis. HBeAg-negative HBV strains often result from a G-to-A change at the first position of a 5′-GGGG stretch in the precore coding sequence (3), a possible consequence of APOBEC3G-mediated editing because this enzyme acts preferentially on the 5′-CC dinucleotide (in the minus-strand DNA) (4). However, other naturally occurring HBV genomes exhibit a pattern of hypermutation that departs from this consensus, with a strong predominance of G-to-A changes in the 5′-GA motif (5). This suggests that, in these cases, another cytidine deaminase may be at play. APOBEC3F is a particularly attractive candidate, because it is also endowed with antiretroviral activity but markedly favors 5′-TC as its target (68).

Consistent with this model, we found that APOBEC3F can efficiently block HBV DNA synthesis (Fig. 1). We do not yet know whether it can also hypermutate the HBV genome in HepG2 or other cells, but our result already suggests that this cytidine deaminase might participate in the noncytopathic virus clearance that is observed during an acute HBV infection (9). A recent transcriptome analysis of the liver of chimpanzees acutely infected with HBV failed to document an induction of the expression of several APOBEC3 family members, including APOBEC3G, but the array used in these experiments did not seem to carry an APOBEC3F-specific probe (10). We agree that additional studies are warranted to investigate the full impact of APOBEC proteins on HBV infection.

Fig. 1.

Real-time PCR quantification of cytoplasmic core–associated HBV DNA purified from HBV-transfected Huh7 cells in the absence (–) or presence (+) of the reverse transcription inhibitor 3TC, of APOBEC3G, or of APOBEC3F, as previously described (1). NT, nontransfected.

References and Notes

Navigate This Article