Human cells contain an enzyme, APOBEC3G (hA3G), that induces profound resistance to infection by certain retroviruses. The hA3G protein restricts HIV-1 by two mechanisms: obstruction of DNA synthesis and hypermutation. The first mechanism involves blocking the HIV-1 reverse transcriptase so that DNA synthesis cannot proceed. In the second mechanism, deamination of cytosine residues in minus-strand DNA results in G-to-A mutation in the coding strand of the provirus. HIV-1 encodes a protein, Vif, that blocks the effects of hA3G by binding to it and promoting its degradation in the proteasome. _____While the interaction of human A3G with HIV-1 has been a central object of investigation in many laboratories, it is clear that other members of the APOBEC3 family can also have antiviral effects, that APOBEC3 family members are present in many mammalian species, and that different viruses have distinct patterns of sensitivity to the different APOBEC3 isoforms. Human A3G restricts HIV-1 in the absence of Vif (delta-Vif) and, apparently, Moloney murine leukemia virus (MLV) by a combination of G:A hypermutation and interference with DNA synthesis. Mouse APOBEC3 (mA3) restricts HIV-1 much like hA3G (except that mA3 is unaffected by Vif). In contrast, restriction of MLV by mA3, while significant, is far less potent than in these other cases, and is not accompanied by detectable G:A hypermutation. ____These observations raise two interrelated questions: (1) what is the mechanism of MLV restriction by mA3, and (2) how does MLV specifically evade hypermutation, which constitutes a significant portion of APOBEC3 restrictive power? We are using both biochemical and genetic approaches to address these questions. This study should clarify how viral proteins interact with APOBEC3s and how the two mechanisms are regulated in restriction of HIV-1. _____In order to determine whether MLV particles whose infectivity was blocked by mA3 can initiate DNA synthesis in new host cells, we generated MLV with different amounts of mA3 or hA3G by cotransfection of viral clones with the respective APOBEC3 plasmid. Cells were infected with these particles, and infected cultures were assayed in parallel for infection (using firefly luciferase as a reporter) and for minus-strand strong-stop DNA synthesis by real-time PCR. We found that there were many particles carrying hA3G that initiated DNA synthesis but failed to complete infection, while there were negligible levels of such particles in the mA3 case. Presumably, the particles that made DNA but failed to infect the cells were inactivated largely by the hypermutation activity of hA3G, which was lacking in the MLV:mA3 combination. Results with delta-Vif HIV-1, with either mA3 or hA3G, were similar to those of MLV with hA3G. Thus, mA3 blocks MLV infection either at, or before, the initiation of DNA synthesis. It is conceivable that mA3 acts by preventing the MLV from entering the cell. We have recently developed an MLV entry assay and will test this possibility. ____By a careful bioinformatic analysis of the sequences of endogenous MLVs, Jern et al. (PLoS Genet. 3:2014-2022, 2007) found that genomes from the polytropic class of endogenous MLVs (PMVs) show all the hallmarks of mA3-induced hypermutation, and were evidently deposited in the mouse germline by MLVs that, unlike the laboratory MLV isolate Moloney MLV, were susceptible to mA3-induced mutation. The difference in susceptibility between the PMV progenitor and Moloney MLV should enable us to localize the region of the genome governing susceptibility to hypermutation. To test the possibility that this region is in Gag, we cloned the gag gene from PMV19, a specific endogenous MLV: this genome has the unmutated, consensus PMV gag coding sequence, although it contains G:A mutations elsewhere. We constructed a chimeric MLV with the gag gene from PMV19 and the remainder of the genome from Moloney MLV. Virions produced from this chimera were fully as infectious as Moloney MLV particles, and had the same pattern of relative susceptibilities to mA3 and hA3G as Moloney MLV. Thus, the presence of the PMV19 gag gene in this chimera had no effect on APOBEC3 sensitivity, and the sensitivity of the parental PMV to hypermutation was not due to its gag gene. _____To better understand the molecular properties of mA3, we expressed it in insect cells fused to glutathione-S-transferase, and partially purified the fusion protein under nondenaturing conditions. This protein exhibited cytidine deaminase activity, as expected. We found several biochemical differences from the better-characterized hA3G. First, the domain organization is reversed relative to hA3G, as the enzymatic activity resides in the N-terminal domain, while the C-terminal domain is inactive but required for encapsidation. This was originally inferred by Hakata and Landau from experiments in mammalian cells (J. Biol. Chem. 281:36624-36631, 2006). Second, there is a deamination gradient, from 5' to 3', in oligonucleotides incubated with mA3 in vitro; this gradient is opposite that seen with hA3G. However, during reverse transcription in vivo, deamination by mA3 (like that by hA3G) increases towards the 3' end of the viral genome, reflecting a 3' to 5' gradient in minus-strand DNA: this polarity presumably reflects the sequence of events during reverse transcription, not the migration of the enzyme along the substrate. We are continuing our efforts to prepare highly purified, enzymatically active mA3 in insect cells. _____In collaboration with Dr. Xinhua Ji (Macromolecular Crystallography Laboratory, NCI), we constructed a model of mA3 structure. We noted several basic residues and aromatic residues in the C-terminal domain that might line an RNA-binding groove. Mutagenesis confirmed that some of these residues are essential for mA3 encapsidation into MLV or delta-Vif HIV-1; this is consistent with the idea that APOBEC3 proteins are packaged by binding to viral RNA. _____Since mA3 fails to induce G:A mutations in MLV, it seemed possible that its encapsidation in MLV somehow leads to the destruction of its cytidine deaminase activity. We tested this hypothesis by lysing MLV or delta-Vif HIV-1 virions carrying either mA3 or hA3G. We found that the particles all contained similar levels of APOBEC3 (judged by immunoblotting against HA, a shared epitope tag) and all exhibited similar levels of enzymatic activity. We also tested the possibility that the encapsidated mA3 is not within the core of the mature particle; however, like hA3G and capsid protein, it is retained in pelletable material when virions are centrifuged through 10% Igepal. Thus, active enzyme is present in the interior of the particles; we still do not know why it fails to deaminate cytidines during reverse transcription in MLV.
| Nair, Smita; Sanchez-Martinez, Silvia; Ji, Xinhua et al. (2014) Biochemical and biological studies of mouse APOBEC3. J Virol 88:3850-60 |
| Nair, Smita; Rein, Alan (2014) Antiretroviral restriction factors in mice. Virus Res 193:130-4 |
| Stavrou, Spyridon; Nitta, Takayuki; Kotla, Swathi et al. (2013) Murine leukemia virus glycosylated Gag blocks apolipoprotein B editing complex 3 and cytosolic sensor access to the reverse transcription complex. Proc Natl Acad Sci U S A 110:9078-83 |
| Sanchez-Martinez, Silvia; Aloia, Amanda L; Harvin, Demetria et al. (2012) Studies on the restriction of murine leukemia viruses by mouse APOBEC3. PLoS One 7:e38190 |
| Rein, Alan; Datta, Siddhartha A K; Jones, Christopher P et al. (2011) Diverse interactions of retroviral Gag proteins with RNAs. Trends Biochem Sci 36:373-80 |
