Viruses have developed diverse non-immune strategies to counteract host-mediated mechanisms that confer resistance to infection. The Vif (virion infectivity factor) proteins are encoded by primate immunodeficiency viruses, most notably human immunodeficiency virus-1 (HIV-1). These proteins are potent regulators of virus infection and replication and are consequently essential for pathogenic infections in vivo. HIV-1 Vif seems to be required during the late stages of virus production for the suppression of an innate antiviral phenotype that resides in human T lymphocytes. Thus, in the absence of Vif, expression of this phenotype renders progeny virions non-infectious. Here, we describe a unique cellular gene, CEM15, whose transient or stable expression in cells that do not normally express CEM15 recreates this phenotype, but whose antiviral action is overcome by the presence of Vif. Because the Vif:CEM15 regulatory circuit is critical for HIV-1 replication, perturbing the circuit may be a promising target for future HIV/AIDS therapies.

The human cytidine deaminase APOBEC3G edits both nascent human immunodeficiency virus (HIV) and murine leukemia virus (MLV) reverse transcripts, resulting in loss of infectivity. The HIV Vif protein is able to protect both viruses from this innate restriction to infection. Here, we demonstrate that a number of other APOBEC family members from both humans and rodents can mediate anti-HIV effects, through cytidine deamination. Three of these, rat APOBEC1, mouse APOBEC3, and human APOBEC3B, are able to inhibit HIV infectivity even in the presence of Vif. Like APOBEC3G, human APOBEC3F preferentially restricts vif-deficient virus. Indeed, the mutation spectra and expression profile found for APOBEC3F indicate that this enzyme, together with APOBEC3G, accounts for the G to A hypermutation of proviruses described in HIV-infected individuals. Surprisingly, although MLV infectivity is acutely reduced by APOBEC3G, no other family member tested here had this effect. It is especially interesting that although both rodent APOBECs markedly diminish wild-type HIV infectivity, MLV is resistant to these proteins. This implies that MLV may have evolved to avoid deamination by mouse APOBEC3. Overall, our findings show that although APOBEC family members are highly related, they exhibit significantly distinct antiviral characteristics that may provide new insights into host-pathogen interactions.

The human protein apolipoprotein B mRNA-editing enzyme-catalytic polypeptide-like-3G (APOBEC3G), also known as CEM-15, mediates a newly described form of innate resistance to retroviral infection by catalyzing the deamination of deoxycytidine to deoxyuridine in viral cDNA replication intermediates. Because DNA deamination takes place after virus entry into target cells, APOBEC3G function is dependent on its association with the viral nucleoprotein complexes that synthesize cDNA and must therefore be incorporated into virions as they assemble in infected cells. Here we show that the HIV-1 virion infectivity factor (Vif) protein protects the virus from APOBEC3G-mediated inactivation by preventing its incorporation into progeny virions, thus allowing the ensuing infection to proceed without DNA deamination. In addition to helping exclude APOBEC3G from nascent virions, Vif also removes APOBEC3G from virus-producing cells by inducing its ubiquitination and subsequent degradation by the proteasome. Our findings indicate that pharmacologic strategies aimed at stabilizing APOBEC3G in HIV-1 infected cells should be explored as potential HIV/AIDS therapeutics.

CEM15/APOBEC3G is a cellular protein required for resistance to infection by virion infectivity factor (Vif)-deficient human immunodeficiency virus (HIV). Here, using a murine leukemia virus (MLV)-based system, we provide evidence that CEM15/APOBEC3G is a DNA deaminase that is incorporated into virions during viral production and subsequently triggers massive deamination of deoxycytidine to deoxyuridine within the retroviral minus (first)-strand cDNA, thus providing a probable trigger for viral destruction. Furthermore, HIV Vif can protect MLV from this CEM15/APOBEC3G-dependent restriction. These findings imply that targeted DNA deamination is a major strategy of innate immunity to retroviruses and likely also contributes to the sequence variation observed in many viruses (including HIV).

The May 28, 2003 immediate early online version of this article (Cell 113, 803-809, 13 June 2003) contained one sentence that did not appear in the printed version. In the results subsection entitled "CEM15/APOBEC3G Is Incorporated into MLV Virions," a bracketed sentence appeared in the following context: The CEM15/APOBEC3G-mediated suppression of HIV infection is thought to be accomplished by protein transferred as a virion component from virus producing cells into target cells (curiously, such physical transfer of CEM15/ APOBEC3G appears uninhibitable by Vif) (Sheehy et al., 2002). The bracketed text was removed prior to publication of the definitive printed and corresponding online versions of the manuscript. It was our intention that this correction should have occurred in all versions of the article. The authors and Cell Press apologize for any inconvenience that may have been caused.

The antiretroviral activity of the cellular enzyme APOBEC3G has been attributed to the excessive deamination of cytidine (C) to uridine (U) in minus strand reverse transcripts, a process resulting in guanosine (G) to adenosine (A) hypermutation of plus strand DNAs. The HIV-1 Vif protein counteracts APOBEC3G by inducing proteasomal degradation and exclusion from virions through recruitment of a cullin5 ECS E3 ubiquitin ligase complex. APOBEC3G belongs to the APOBEC protein family, members of which possess consensus (H/C)-(A/V)-E-(X)24-30-P-C-(X)2-C cytidine deaminase motifs. Earlier analyses of APOBEC-1 have defined specific residues that are important for zinc coordination, proton transfer, and, therefore, catalysis within this motif. Because APOBEC3G contains two such motifs, we used site-directed mutagenesis of conserved residues to assess each region's contribution to anti-HIV-1 activity. Surprisingly, whereas either the N- or C-terminal domain could confer antiviral function in tissue culture-based infectivity assays, only an intact C-terminal motif was essential for DNA mutator activity. These findings reveal the nonequivalency of APOBEC3G's N- and C-terminal domains and imply that APOBEC3G-mediated DNA editing may not always be necessary for antiviral activity. Accordingly, we propose that APOBEC3G can achieve an anti-HIV-1 effect through an undescribed mechanism that is distinct from cytidine deamination.

Retroviral DNA can be subjected to cytosine-to-uracil editing through the action of members of the APOBEC family of cytidine deaminases. Here we demonstrate that APOBEC-mediated cytidine deamination of human immunodeficiency virus (HIV) virion RNA can also occur. We speculate that the natural substrates of the APOBEC enzymes may extend to RNA viruses that do not replicate through DNA intermediates. Thus, cytosine-to-uracil editing may contribute to the sequence diversification of many viruses.

The E2A gene products, E12 and E47, are required for proper B cell development. Mice lacking the E2A gene products generate only a very small number of B220+ cells, which lack immunoglobulin DJ(H) rearrangements. We have now generated mice expressing either E12 or E47. B cell development in mice expressing E12 but lacking E47 is perturbed at the pro-B cell stage, and these mice lack IgM+B220+ B cells in both bone marrow and spleen. IgM+B220+ B cells can be detected, albeit at significantly reduced levels, in the bone marrow and spleen of mice lacking E12. Ectopic expression of both E12 and E47 in a null mutant background shows that E12 and E47 act in concert to promote B lineage development. Taken together, the data indicate that both E12 and E47 allow commitment to the B cell lineage and act synergistically to promote B lymphocyte maturation.