Different modes of retrovirus restriction by human APOBEC3A and APOBEC3G in vivo

Abstract

The apolipoprotein B editing complex 3 (A3) cytidine deaminases are among the most highly evolutionarily selected retroviral restriction factors, both in terms of gene copy number and sequence diversity. Primate genomes encode seven A3 genes, and while A3F and 3G are widely recognized as important in the restriction of HIV, the role of the other genes, particularly A3A, is not as clear. Indeed, since human cells can express multiple A3 genes, and because of the lack of an experimentally tractable model, it is difficult to dissect the individual contribution of each gene to virus restriction in vivo. To overcome this problem, we generated human A3A and A3G transgenic mice on a mouse A3 knockout background. Using these mice, we demonstrate that both A3A and A3G restrict infection by murine retroviruses but by different mechanisms: A3G was packaged into virions and caused extensive deamination of the retrovirus genomes while A3A was not packaged and instead restricted infection when expressed in target cells. Additionally, we show that a murine leukemia virus engineered to express HIV Vif overcame the A3G-mediated restriction, thereby creating a novel model for studying the interaction between these proteins. We have thus developed an in vivo system for understanding how human A3 proteins use different modes of restriction, as well as a means for testing therapies that disrupt HIV Vif-A3G interactions.

Figure 1. Expression of A3A and A3G transgenes.

A) RT-qPCR analysis of RNA isolated from different tissues of the A3G^high and A3G^low strains. B) RT-qPCR analysis of RNA isolated from different tissues of the A3A^high and A3A^low strains. Shown for comparison for both graphs are the endogenous A3 levels in nontransgenic C57BL/6 mice (mA3), as well as A3A and A3G expression in human H9 cells and human PBMCs (average of 2 individuals). The mice used for this analysis were uninfected. Both panels are representative of 2 independent experiments with a different mouse of each genotype. Error bars denote standard deviation of technical replicates.

Figure 2. Cellular lysates from splenocytes derived from uninfected transgenic, wild type (BL/6) or mA3 knockout (KO) mice or from 293T cell lines over-expressing A3A or A3G were incubated with a 3′-fluorophore labeled 50-mer single-stranded oligonucleotide (S50) containing cytosine in the sequence context preferred by A3G (S50-CCC) or A3A (S50-TTC).

Deamination was detected by uracil excision by UDG followed by fragmentation of the resulting abasic site by NaOH and heat, resulting in a 35-mer product (P35). High levels of activity in the 293/A3A samples result in deamination at multiple potential cytosines in the S50-CCC substrate (P37). This experiment was performed several times with the same lysates, with similar results.

Figure 3. A3A and A3G restrict murine retrovirus infection in vivo.

A) Newborn mice were infected with M-MLV and 16 days post-infection, virus titers in spleens were measured. Each point represents the titer obtained from an individual mouse; the average for each group is shown by a horizontal bar. The transgenic mice were derived from 2–3 litters each; the knockout mice are the littermates of the transgenic mice. N = 12 A3G^high, 6 A3G^low, 10 A3A^high, 7 A3A^low, 42 KO and 6 WT mice. B) Five day old mice were infected with MMTV and 3 weeks post-infection, DNA was isolated from spleens and subjected to RT-qPCR with MMTV-specific primers. Each point represents the proviral DNA levels measured in splenic DNA from an individual mouse; the average for each group is shown by a horizontal bar. The transgenic mice were derived from 2–3 litters each; the knockout mice are the littermates of the transgenic mice. N = 5 A3G^high, 10 A3G^low, 12 A3A^high, 6 A3A^low and 24 KO mice. *, p≤0.0001, **, p≤.001, ***, p≤.01, NS, not significant (Mann-Whitney t test).

Figure 4. Deamination of M-MLV viral DNA and RNA in A3G transgenic mice.

Splenic DNA or RNA from virions was isolated from the infected mice described in Figure 3 and cloned and sequenced. In most cases > 10 sequences from 4 different mice were analyzed, as indicated in the figure. For the viral RNA samples, the viruses from 5 animals were pooled for sequencing. Shown are the G to A changes in the sequences; other mutations are indicated in Table 1. Red  =  GG > AG, cyan  =  GA > AA, green  =  GC > AC and magenta  =  GT > AT transitions. Red arrows denote mutation hotspots seen in viruses isolated from A3G^high and A3G^low mice; black arrows denote hotspots identified only in A3G^high mice.

Figure 5. A3G but not A3A is packaged into M-MLV and inhibits reverse transcription.

A) and B) Virions were isolated from the spleens of A3G (A), A3A (B), wild type and KO mice and EnRT assays were performed. Shown is the average of 3 independent experiments using different virus preparations; error bars show standard deviation. C) Western blot analysis of M-MLV virions from the transgenic mice, using anti-myc antisera (top panel). The blots were stripped and reprobed with anti-MLV antisera (bottom panel). Abbreviations: CA, capsid; +, extracts from 293T cells transfected with the A3A or A3G transgenes.

Figure 6. Expression of A3A and A3G in BMDCs restrict incoming murine retroviruses.

A) Infection of BMDCs with M-MLV. RT-qPCR analysis of genomic DNA with M-MLV-specific primers, normalized to GAPDH. B) MMTV infection of BMDCs with MMTV. RT-qPCR analysis of genomic DNA with MMTV-specific primers, normalized to GAPDH. Shown are the results of 2 independent experiments (open bars, expt.1; closed bars, expt. 2) with 3 technical replicates in each experiment. Error bars (standard deviation) and p values were calculated for each experiment. *, p≤.01 based on one-way ANOVA.

Figure 7. Vif counteracts A3G in transgenic mice.

Newborn mice were infected with F-MLV-2A or F-MLV-2A-vif. A) Splenic extracts and isolated virions were analyzed by western blot. A3G was detected with anti-A3G antisera. Shown is a representative western blot from individual mice (KO and A3G^high) infected with F-MLV-2A and F-MLV-2A-vif. This experiment was repeated twice with 1 additional mouse of each genotype and gave similar results. B) Virus titers in spleens were measured. Each point represents the titer obtained from an individual mouse; the average for each group is shown by a horizontal bar. The transgenic mice were derived from 1–3 litters each; the knockout mice are the littermates of the transgenic mice. N = 11 KO, 4 A3G^high and 5 A3G^low mice infected with F-MLV-2A-vif and 12 KO, 4 A3G^high and 4 A3G^low mice infected with FMLV-2A. *, p≤0.02; **, p≤0.004; NS, not significant (Mann-Whitney T test).