CORE Metadata, citation and similar papers at core.ac.uk

Provided by Elsevier - Publisher Connector Current Biology, Vol. 15, 166–170, January 26, 2005, ©2005 Elsevier Ltd All rights reserved. DOI 10.1016/j.cub.2004.12.068 Antiviral Function of APOBEC3G Can Be Dissociated from Activity

Edmund N.C. Newman,1 Rebecca K. Holmes,1 in rpoB can be evaluated by screening for the frequency Heather M. Craig,1 Kevin C. Klein,2 of acquired rifampicin resistance (rifR) relative to cell Jaisri R. Lingappa,2 Michael H. Malim,1 viability. Whereas mutations made in the N-terminal resi- and Ann M. Sheehy1,* dues had very little effect on acquired rifampicin resis- 1Department of Infectious Diseases tance, each mutation made in the C terminus dramati- Guy’s, King’s, and St. Thomas’ School of Medicine cally decreased the frequency of rifR bacterial colonies King’s College London (e.g., the H257R mutant-induced rifampicin resistance .(fold relative to wild-type APOBEC3G-17ف London SE1 9RT is decreased United Kingdom Not unexpectedly, the mutational activity of the double- 2 Department of Pathobiology point mutants was also significantly suppressed. These University of Washington results implicate the C-terminal as necessary Seattle, Washington 98195 for the DNA-mutating function of APOBEC3G and seem to suggest that the N-terminal active site is dispensable for this activity. It is also worth noting that the N-terminal active site contains a proline at position 2 rather than Summary the expected alanine or valine. It is possible that the positioning of this proline distorts the local structure of 2ϩ The antiretroviral activity of the cellular APO- the active site, perturbing the Zn coordination. Our BEC3G [1] has been attributed to the excessive deami- results support the inference that the C-terminal region nation of cytidine (C) to uridine (U) in minus strand is primarily responsible for the DNA-mutating function reverse transcripts, a process resulting in guanosine of APOBEC3G. (G) to adenosine (A) hypermutation of plus strand These mutant were next tested in a single- [2–5]. The HIV-1 Vif counteracts APO- round infection assay for their ability to regulate vif- ⌬ BEC3G by inducing proteasomal degradation and ex- deficient ( vif) HIV-1 infection (Figure 1B). Surprisingly, clusion from virions through recruitment of a cullin5 relative to the wild-type APOBEC3G, none of the single- ECS E3 complex [6–13]. APOBEC3G point mutant proteins were significantly affected in their belongs to the APOBEC , members of ability to function as antiviral factors. Although titrating which possess consensus (H/C)-(A/V)-E-(X)24-30-P- down the expression of the mutant proteins ameliorated C-(X)2-C cytidine deaminase motifs [14]. Earlier analy- their antiviral activity (data not shown), all of the mutants ses of APOBEC-1 have defined specific residues that retained more than 60% of the antiviral activity of the are important for zinc coordination, proton transfer, wild-type protein under the experimental conditions and, therefore, catalysis within this motif [15]. Because used here. Additionally, each mutant was examined for APOBEC3G contains two such motifs, we used site- activity against wild-type HIV-1; all exhibited Vif-sensi- directed mutagenesis of conserved residues to assess tivity and were therefore unable to suppress wild-type each region’s contribution to anti-HIV-1 activity. Sur- virion infectivity significantly. Analyzed together with the prisingly, whereas either the N- or C-terminal domain bacterial editing assay, this rather unexpected result could confer antiviral function in tissue culture-based suggests that impairment of the editing function of APO- infectivity assays, only an intact C-terminal motif was BEC3G does not necessarily inhibit APOBEC3G’s ability essential for DNA mutator activity. These findings re- to suppress HIV-1 infection in this context. veal the nonequivalency of APOBEC3G’s N- and When the double point mutant proteins were exam- C-terminal domains and imply that APOBEC3G-medi- ined for their effects on infectivity, the loss of APO- ated DNA editing may not always be necessary for BEC3G function was far more dramatic. In fact, the ob- antiviral activity. Accordingly, we propose that APO- served loss of activity for each double mutant exceeded BEC3G can achieve an anti-HIV-1 effect through an what might be expected from the corresponding single undescribed mechanism that is distinct from cytidine point mutant data. For example, whereas E67Q and . E259Q each retain 80%–90% antiviral activity, the corre- sponding double mutant was inactive. These results support the idea that while the highly conserved resi- Results and Discussion dues of one “deaminase domain” may be dispensable, the matching residues in the other are then essential. Initially, wild-type and mutant APOBEC3G (apolipopro- Thus, antiviral activity appears to be dependent on one tein B mRNA-editing enzyme-catalytic polypeptide-like of the domains being intact, but with little discrimination 3G) proteins (Table 1) were assayed for their ability to as to which one. function as DNA mutators in an E. coli-based assay In light of the suggestion that the antiviral activity of (Figure 1A). This assay is a readout of the ability of a APOBEC3G can be mediated through a mechanism that transduced to mutate the RNA polymerase B gene is independent of its cytidine deaminase function, we (rpoB) in the KL-16 strain of bacteria [16, 17]. Mutations next sequenced viral cDNAs recovered from cells chal- lenged with ⌬vif virions produced in the presence of *Correspondence: [email protected] wild-type or mutant APOBEC3G proteins [18]. The muta- Editing-Independent Inhibition of HIV by APOBEC3G 167

Table 1. Summary of APOBEC3G Mutations

Consensus: H-(A/V)-E-(X)24–30-P-C-(X)2-C N-Terminal Active Site C-Terminal Active Site aa Change

H-P-E-(X)24–30-P-C-(X)2-C H-A-E-(X)24–30-P-C-(X)2-C Wild-type

R-P-E-(X)24–30-P-C-(X)2-C H-A-E-(X)24–30-P-C-(X)2-C H65R

H-P-E-(X)24–30-P-C-(X)2-C R-A-E-(X)24–30-P-C-(X)2-C H257R

R-P-E-(X)24–30-P-C-(X)2-C R-A-E-(X)24–30-P-C-(X)2-C H65R and H257R

H-P-Q-(X)24–30-P-C-(X)2-C H-A-E-(X)24–30-P-C-(X)2-C E67Q

H-P-E-(X)24–30-P-C-(X)2-C H-A-Q-(X)24–30-P-C-(X)2-C E259Q

H-P-Q-(X)24–30-P-C-(X)2-C H-A-Q-(X)24–30-P-C-(X)2-C E67Q and E259Q

H-P-E-(X)24–30-P-S-(X)2-C H-A-E-(X)24–30-P-C-(X)2-C C97S

H-P-E-(X)24–30-P-C-(X)2-C H-A-E-(X)24–30-P-S-(X)2-C C288S

H-P-E-(X)24–30-P-S-(X)2-C H-A-E-(X)24–30-P-S-(X)2-C C97S and C288S

H-P-E-(X)24–30-P-C-(X)2-S H-A-E-(X)24–30-P-C-(X)2-C C100S

H-P-E-(X)24–30-P-C-(X)2-C H-A-E-(X)24–30-P-C-(X)2-S C291S

H-P-E-(X)24–30-P-C-(X)2-S H-A-E-(X)24–30-P-C-(X)2-S C100S and C291S The consensus sequence of the active site for cytidine deaminases is shown at the top. Mutations were made with oligonucleotide site- directed mutagenesis (Invitrogen). The underlined letters indicate the change in for the corresponding mutant.

tions identified in cloned cDNAs are catalogued in Table cently implicated as being important for the packaging 2. As shown previously, analysis of the mutational spec- of APOBEC3G into virions [22–25]. We used continuous tra confirms that wild-type APOBEC3G induces high sucrose gradients to isolate virions made in the pres- ,of all G residues ence of either wild-type or mutant APOBEC3G proteins %13ف) levels of G to A hypermutation mutated). The rates of mutation induced by the N-ter- and then we examined their contents by immunoblot minal mutants (H65R and C97S) also correlated with the [1]. As seen in Figures 2A and 2B, the mutant proteins antiviral activity, although the actual rates were some- were readily detectable in purified virions (upper panels) what lower. To our surprise, and in good agreement with and were present at approximately similar levels. There- the results of the E. coli mutator assays (Figure 1A), we fore, although the cytidine deaminase activity of the were unable to detect any significant mutational activity C-terminal mutants is not required for their ability to in any of the infections carried out in the presence of suppress virion infectivity, we were unable to dissociate the C-terminal mutants (Table 2), despite the fact that the ability of these proteins to be packaged into assem- these proteins displayed strong antiviral effects, ranging bling virions from their antiviral function. Finally, it could from 53%–97% (Figure 1B; Table 2). These data suggest be argued that editing-deficient APOBEC3G mutants that an antiviral function of APOBEC3G can be exerted interfere with infection in a more fundamental manner independently of its DNA deaminase activity. than the wild-type protein (for instance, by severely di- We considered the possibility that disruption of the minishing reverse transcription). We consider this expla- C-terminal active site created mutant proteins in which nation to be unlikely: First, extensively edited reverse this domain acquired a dominant-interfering phenotype transcripts were still readily recovered in the mixing ex- over the N-terminal site, such that the assumed lack of periment discussed earlier (Table S1), and second, pre- enzymatic activity of the N-terminal site was, instead, liminary measurements of viral cDNA levels indicate that due to suppression. Recently, a number of groups have the early phases of reverse transcription are as efficient experimentally shown the dimerization capabilities of in the presence of these C-terminal mutants as they are several APOBEC family members [19–21]. We exploited with wild-type APOBEC3G (data not shown). these observations to ask whether mutant carboxy-ter- Previous reports of APOBEC3G-induced hypermuta- minal domains can influence the activity of editing- tion have correlated cytidine deaminase activity with active domains that may be in close spatial proximity. antiviral function [3, 5, 26]. Our results with cultured cell Expression constructs for wild-type APOBEC3G and the systems and a panel of mutant APOBEC3G proteins E259Q mutant were cotransfected (at a 1:6 ratio) into indicate that this is not always the case. In particular, 293T cells, and both virion infectivity and cDNA editing we demonstrate that, under the described experimental capacity were examined (see Table S1 in the Supple- conditions, the cytidine deamination function of APO- mental Data available with this article online). The pres- BEC3G can be separated from antiviral activity. We note ence of the E259Q protein had no effect on the extent of that our study differs from these earlier reportings in editing mediated by the wild-type APOBEC3G protein, that we employed proteins that had not been epitope arguing against the idea of a dominant-negative effect. tagged. In our experience, we have observed that a Because the C-terminal mutants fail to edit viral cDNA terminal tag can not only interfere with mutant protein in target cells, we questioned whether this protein was expression but can impair function in the E. coli-based still packaged into HIV-1 virions. In APOBEC-1, several assay and during virus infection (unpublished data). Our residues within the active site have been shown to also data raise the possibility that the APOBEC3G-induced be critical for RNA binding [15]. If these conserved resi- viral-cDNA hypermutation that has been observed by dues within the C-terminal active site of APOBEC3G many groups may be an observable side effect and play a similar role, then their replacement could also may not always be central to the antiviral function of influence RNA binding, an attribute that has been re- APOBEC3G. In other words, the finding that APOBEC3G Current Biology 168

Figure 1. The Cytidine Deaminase Activity of APOBEC3G Is Not Required for an Antiviral Effect (A) The ability of wild-type or mutated APO- BEC3G proteins to mutate E. coli DNA. APO- BEC3G-expressing vectors were transformed into bacteria, and the resulting induction of rif- ampicin resistance was quantitated by count- ing rifR colonies (relative to viable cell counts). Wild-type APOBEC3G typically induced rif- fold relative to the-30ف ampicin resistance by vector control. The mutation rates for the mu- tant APOBEC3G proteins are plotted as a per- centage of that seen with wild-type protein (set at 100%). Error bars represent the stan- dard deviation. The lower panel shows immunoblots confirm- ing comparable expression of the mutant APOBEC3G proteins in bacteria. (B) Antiviral function of APOBEC3G mutant proteins. vif-deficient (⌬vif) HIV-1 was pro- duced in 293T cells transiently transfected with APOBEC3G expression vectors (1:1 plasmid ratio of proviral and APOBEC3G ex- pression plasmid). Virus-containing superna- tants were harvested at 24 hr and quantitated with a p24Gag ELISA (enzyme-linked immuno- sorbent assay). Normalized inocula were used to infect the C8166-CCR5/HIV-LTR CAT indicator cell line, in which infectivity is corre- lated with the induction of CAT activity [30], and the antiviral activity of each mutant pro- tein is presented as a percentage of that seen for wild-type APOBEC3G. Here, the wild-type APOBEC3G was HA-tagged at its C terminus: This modification does not interfere with the assayable antiviral function of this protein. Error bars represent the standard deviation. The lower panel is an immunoblot confirming expression of the mutant APOBEC3G pro- teins in producer cell lysates.

variants that lack cytidine deaminase function can still antiviral effects. For instance, different domains of APO- impose a significant antiviral effect indicates the exis- BEC3G may function independently in its role as an tence of an additional cellular pathway (or mechanism) antiviral factor. Structure-function analyses of AID (acti- for resisting HIV-1 infection. vation-induced cytidine deaminase), another member of Recent studies examining the impact of APOBEC3G the APOBEC family, have led to the identification of two on hepatitis B viral (HBV) infection also suggest that separable domains independently responsible for the APOBEC3G may exert its antiviral effect in multiple ways primary functions of this protein, class-switch recombi- [27, 28]. Specifically, these investigations showed that nation and somatic hypermutation, both of which are although APOBEC3G exhibited a strong suppressive ef- key processes in antibody diversification [29]. Although fect against HBV, the small amount of DNA found associ- the mutations that illustrate this dissociation are not ated with viral cores lacked G to A mutations (at least within the core cytidine deamination domain of AID, the in the Huh7 hepatoma cell line). These and our results results support the premise that different domains within underscore the need for continuing reevaluation of the AID may function independently. possible mechanisms involved in APOBEC3G-mediated With respect to APOBEC3G, it would be interesting Editing-Independent Inhibition of HIV by APOBEC3G 169

Table 2. The Loss of Detectable Hypermutation Does Not Block APOBEC3G Antiviral Activity Average Average Antiviral Total GtoA GtoA Activity Clones Total Bases Total G to A Other Mutations Mutations APOBEC3G (%) Sequenced Sequenced Mutations Mutations Mutations per Sequence per 100 bp

None 0 10 6,500 3 1 2 0.1 0.02 Wild-type 100 8 5,200 163 155 8 19 2.98 H257R 54 10 6,500 3 0 3 0.0 0.00 E259Q 97 10 6,500 1 1 0 0.1 0.02 C288S 77 10 6,500 8 8 0 0.8 0.12 C291S 53 9 5,850 6 1 5 0.1 0.02 H65R 87 16 10,400 149 136 13 9 1.31 C97S 84 10 6,500 119 112 7 11 1.72 vif-deficient (⌬vif) HIV-1 virions were produced in 293T cells transiently transfected with the APOBEC3G expression constructs and were then used to infect the C8166-CCR5/HIV-1 LTR CAT indicator cells. At 24 hr, infectivity was measured as an induction of CAT activity, and total DNA was isolated from the infected cells and subjected to high-fidelity PCR (Advantage HF-2 PCR kit, Clontech) with HIV-1-specific oligonucleotides. The PCR product was cloned into pBluescript and sequenced. The number of bases sequenced is indicated. It should be noted that the reported antiviral activity was from this single experiment, but the activities all fall within the average antiviral activities reported in Figure 1B.

to identify point mutations that disrupt the antiviral effect Received: August 10, 2004 without perturbing the cytidine deamination activity. At Revised: November 12, 2004 present, it is interesting to speculate that the hypermuta- Accepted: November 15, 2004 Published: January 26, 2005 tion that has been seen as a hallmark of APOBEC3G antiviral function may instead just be a tangible experi- References mental readout. Alternatively, deamination may in fact contribute to the suppression of viral infection, whereas 1. Sheehy, A.M., Gaddis, N.C., Choi, J.D., and Malim, M.H. (2002). the other antiviral effects of APOBEC3G remain to be Isolation of a human gene that inhibits HIV-1 infection and is suppressed by the viral Vif protein. Nature 418, 646–650. elucidated. In sum, it will be crucial to understand how 2. Harris, R.S., Bishop, K.N., Sheehy, A.M., Craig, H.M., Petersen- APOBEC3G can act as an HIV-1 inhibitor in the absence Mahrt, S.K., Watt, I.N., Neuberger, M.S., and Malim, M.H. (2003). of cytidine deamination. Identifying this mechanism may DNA deamination mediates innate immunity to retroviral infec- uncover an unidentified host cell antiviral pathway and tion. Cell 113, 803–809. suggest new directions for novel therapeutics. 3. Mangeat, B., Turelli, P., Caron, G., Friedli, M., Perrin, L., and Trono, D. (2003). Broad antiretroviral defence by human APO- BEC3G through lethal editing of nascent reverse transcripts. Supplemental Data Nature 424, 99–103. A supplemental table is available at http://www.current-biology. 4. Lecossier, D., Bouchonnet, F., Clavel, F., and Hance, A.J. (2003). com/cgi/content/full/15/2/166/DC1/. Hypermutation of HIV-1 DNA in the absence of the Vif protein. Science 300, 1112. Acknowledgments 5. Zhang, H., Yang, B., Pomerantz, R.J., Zhang, C., Arunachalam, S.C., and Gao, L. (2003). The cytidine deaminase CEM15 in- This work was supported by the United Kingdom Medical Research duces hypermutation in newly synthesized HIV-1 DNA. Nature Council, the Guy’s and St Thomas’ Charitable Foundation, the Bio- 424, 94–98. technology and Biological Sciences Research Council, the Royal 6. Stopak, K., de Noronha, C., Yonemoto, W., and Greene, W.C. Society (A.M.S.), and the American Foundation for AIDS Research (2003). HIV-1 Vif blocks the antiviral activity of APOBEC3G by (J.R.L.). M.H.M. is an Elizabeth Glaser Scientist supported by the impairing both its translation and intracellular stability. Mol. Cell Elizabeth Glaser Pediatric AIDS Foundation. 12, 591–601.

Figure 2. Incorporation of Mutant APOBEC3G Protein into Virions (A and B) Both wild-type and mutant APOBEC3G proteins are efficiently packaged into virions. The purified virions analyzed in (A) and (B) are from independent experiments, and wild-type APOBEC3G packaging was examined within each experiment. The upper panels are immunoblots verifying the presence of APOBEC3G proteins within virions produced from transfected 293T cells (as for Figure 1B). Viral supernatants were harvested at 24 hr after transfection, and virions were isolated by continuous sucrose gradient purification [1]. E67Q and E259Q were also packaged to similar levels (data not shown). The lower panels are immunoblots probed with an antibody specific for p24Gag (loading control). Current Biology 170

7. Sheehy, A.M., Gaddis, N.C., and Malim, M.H. (2003). The antiret- vity of HIV-1 virion but not a sole determinant of its antiviral roviral enzyme APOBEC3G is degraded by the proteasome in activity. J. Biol. Chem. 278, 44412–44416. response to HIV-1 Vif. Nat. Med. 9, 1404–1407. 27. Turelli, P., Mangeat, B., Jost, S., Vianin, S., and Trono, D. (2004). 8. Marin, M., Rose, K.M., Kozak, S.L., and Kabat, D. (2003). HIV-1 Inhibition of hepatitis B virus replication by APOBEC3G. Science Vif protein binds the editing enzyme APOBEC3G and induces 303, 1829. its degradation. Nat. Med. 9, 1398–1403. 28. Rosler, C., Kock, J., Malim, M.H., Blum, H.E., and von Weiz- 9. Mehle, A., Strack, B., Ancuta, P., Zhang, C., McPike, M., and sacker, F. (2004). Comment on “Inhibition of hepatitis B virus Gabuzda, D. (2004). Vif overcomes the innate antiviral activity replication by APOBEC3G”. Science 305, 1403. of APOBEC3G by promoting its degradation in the ubiquitin- 29. Shinkura, R., Ito, S., Begum, N.A., Nagaoka, H., Muramatsu, M., proteasome pathway. J. Biol. Chem. 279, 7792–7798. Kinoshita, K., Sakakibara, Y., Hijikata, H., and Honjo, T. (2004). 10. Conticello, S.G., Harris, R.S., and Neuberger, M.S. (2003). The Separate domains of AID are required for somatic hypermuta- Vif protein of HIV triggers degradation of the human antiretroviral tion and class-switch recombination. Nat. Immunol. 5, 707–712. DNA deaminase APOBEC3G. Curr. Biol. 13, 2009–2013. 30. Simon, J.H., and Malim, M.H. (1996). The human immunodefi- 11. Mariani, R., Chen, D., Schrofelbauer, B., Navarro, F., Konig, R., ciency virus type 1 Vif protein modulates the postpenetration Bollman, B., Munk, C., Nymark-McMahon, H., and Landau, N.R. stability of viral nucleoprotein complexes. J. Virol. 70, 5297– (2003). Species-specific exclusion of APOBEC3G from HIV-1 5305. virions by Vif. Cell 114, 21–31. 12. Yu, X., Yu, Y., Liu, B., Luo, K., Kong, W., Mao, P., and Yu, X.F. (2003). Induction of APOBEC3G ubiquitination and degradation by an HIV-1 Vif-Cul5-SCF complex. Science 302, 1056–1060. 13. Kao, S., Khan, M.A., Miyagi, E., Plishka, R., Buckler-White, A., and Strebel, K. (2003). The human immunodeficiency virus type 1 Vif protein reduces intracellular expression and inhibits pack- aging of APOBEC3G (CEM15), a cellular inhibitor of virus infec- tivity. J. Virol. 77, 11398–11407. 14. Bhattacharya, S., Navaratnam, N., Morrison, J.R., Scott, J., and Taylor, W.R. (1994). Cytosine nucleoside/nucleotide deami- nases and apolipoprotein B mRNA editing. Trends Biochem. Sci. 19, 105–106. 15. MacGinnitie, A.J., Anant, S., and Davidson, N.O. (1995). Muta- genesis of -1, the catalytic subunit of the mammalian apolipoprotein B mRNA editing enzyme, reveals distinct do- mains that mediate cytosine nucleoside deaminase, RNA bind- ing, and RNA editing activity. J. Biol. Chem. 270, 14768–14775. 16. Petersen-Mahrt, S.K., Harris, R.S., and Neuberger, M.S. (2002). AID mutates E. coli suggesting a DNA deamination mechanism for antibody diversification. Nature 418, 99–103. 17. Harris, R.S., Petersen-Mahrt, S.K., and Neuberger, M.S. (2002). RNA editing enzyme APOBEC1 and some of its homologs can act as DNA mutators. Mol. Cell 10, 1247–1253. 18. Bishop, K.N., Holmes, R.K., Sheehy, A.M., Davidson, N.O., Cho, S.J., and Malim, M.H. (2004). Cytidine deamination of retroviral DNA by diverse APOBEC proteins. Curr. Biol. 14, 1392–1396. 19. Liddament, M.T., Brown, W.L., Schumacher, A.J., and Harris, R.S. (2004). APOBEC3F properties and hypermutation prefer- ences indicate activity against HIV-1 in vivo. Curr. Biol. 14, 1385– 1391. 20. Jarmuz, A., Chester, A., Bayliss, J., Gisbourne, J., Dunham, I., Scott, J., and Navaratnam, N. (2002). An anthropoid-specific locus of orphan C to U RNA-editing on chromosome 22. Genomics 79, 285–296. 21. Wiegand, H.L., Doehle, B.P., Bogerd, H.P., and Cullen, B.R. (2004). A second human antiretroviral factor, APOBEC3F, is sup- pressed by the HIV-1 and HIV-2 Vif proteins. EMBO J. 23, 2451– 2458. 22. Svarovskaia, E.S., Xu, H., Mbisa, J.L., Barr, R., Gorelick, R.J., Ono, A., Freed, E.O., Hu, W.S., and Pathak, V.K. (2004). Human APOBEC3G is incorporated into HIV-1 virions through interac- tions with viral and nonviral . J. Biol. Chem. 279, 35822– 35828. 23. Luo, K., Liu, B., Xiao, Z., Yu, Y., Yu, X., Gorelick, R., and Yu, X.F. (2004). Amino-terminal region of the human immunodeficiency virus type 1 nucleocapsid is required for human APOBEC3G packaging. J. Virol. 78, 11841–11852. 24. Zennou, V., Perez-Caballero, D., Gottlinger, H., and Bieniasz, P.D. (2004). APOBEC3G incorporation into human immunodefi- ciency virus type 1 particles. J. Virol. 78, 12058–12061. 25. Schafer, A., Bogerd, H.P., and Cullen, B.R. (2004). Specific pack- aging of APOBEC3G into HIV-1 virions is mediated by the nu- cleocapsid domain of the gag polyprotein precursor. Virology 328, 163–168. 26. Shindo, K., Takaori-Kondo, A., Kobayashi, M., Abudu, A., Fuku- naga, K., and Uchiyama, T. (2003). The enzymatic activity of CEM15/Apobec-3G is essential for the regulation of the infecti-