www.elsevier.com/locate/yviro

Virology 333 (20

Complementary function of the two catalytic domains of APOBEC3G

Francisco Navarro, Brooke Bollman, Hui Chen, Renate Kfnig, Qin Yu,

Kristopher Chiles, Nathaniel R. LandauT

Infectious Disease Laboratory, Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, CA 92037, USA

Received 23 November 2004; returned to author for revision 7 December 2004; accepted 10 January 2005

Available online 30 January 2005

Abstract

The HIV-1 viral accessory protein Vif prevents the encapsidation of the antiviral cellular cytidine deaminases APOBEC3F and

APOBEC3G by inducing their proteasomal degradation. In the absence of Vif, APOBEC3G is encapsidated and blocks virus replication by

deaminating cytosines of the viral cDNA. APOBEC3G encapsidation has been recently shown to depend on the viral nucleocapsid protein;

however, the role of RNA remains unclear. Using APOBEC3G deletion and point mutants, we mapped the encapsidation determinant to the

Zn2+ coordination residues of the N-terminal catalytic domain (CD1). Notably, these residues were also required for RNA binding. Mutations

in the two aromatic residues of CD1 but not CD2, which are conserved in cytidine deaminase core domains and are required for RNA

binding, prevented encapsidation into HIV-1, HTLV-I and MLV. The Zn2+ coordination residues of the C-terminal catalytic domain (CD2)

were not required for encapsidation but were essential for cytidine deaminase activity and the antiviral effect. These findings suggest a model

in which CD1 mediates encapsidation and RNA binding while CD2 mediates cytidine deaminase activity. Interestingly, HTLV-I was

relatively resistant to the antiviral effects of encapsidated APOBEC3G.

D 2005 Elsevier Inc. All rights reserved.

Keywords: APOBEC3G; Vif; HIV-1; Encapsidation; Deamination

Introduction

Lentiviruses, with the exception of equine infectious

anemia virus, require the accessory protein Vif to replicate

in primary CD4+ T cells and monocytes. Vif has been shown

to relieve the inhibitory activity of cellular cytidine

deaminases of the APOBEC3 family (Sheehy et al.,

2002). APOBEC3F and APOBEC3G become encapsidated

in Dvif HIV-1 virions, causing CYU deamination of the

viral DNA that is synthesized in the subsequent round of

replication (Bishop et al., 2004; Harris et al., 2003;

Lecossier et al., 2003; Liddament et al., 2004; Mangeat et

al., 2003; Mariani et al., 2003; Wiegand et al., 2004; Zhang

et al., 2003; Zheng et al., 2004). The uracil-containing viral

cDNA molecules are degraded by cellular DNA repair

0042-6822/$ - see front matter D 2005 Elsevier Inc. All rights reserved.

doi:10.1016/j.virol.2005.01.011

T Corresponding author. Fax: +1 858 554 0341.

E-mail address: Landau@salk.edu (N.R. Landau).

enzymes prior to integration. In cells infected with wild-type

HIV-1, Vif binds to APOBEC3G, inducing its rapid

ubiquitination and degradation before it can become

encapsidated (Mehle et al., 2004; Sheehy et al., 2003;

Stopak et al., 2003; Yu et al., 2003). Vif-induced degrada-

tion of APOBEC3G is mediated by an E3 ubiquitin ligase of

the ECS type that is composed of elongins B and C, Cul5

and Rbx1 to which Vif binds (Yu et al., 2003).

APOBEC3 molecules are conserved in primate and

rodent genomes and are active against HIV-1 when

expressed in human cells (Mariani et al., 2003). However,

the interaction of Vif with APOBEC3G is species-specific

(Mariani et al., 2003). The mouse genome contains a single

APOBEC3 gene that is about ~30% identical to the human

orthologue. Mouse APOBEC3 blocked HIV-1 regardless of

the presence of Vif. Similarly, African green monkey

(AGM) and rhesus macaque APOBEC3Gs were active

against Dvif and wild-type HIV-1 (Mariani et al., 2003). The

species-specificity resulted from the inability of Vif to bind

05) 374–386

F. Navarro et al. / Virology 333 (2005) 374–386 375

noncognate APOBEC3G, which was determined by a single

amino acid of APOBEC3G at position 128 (Bogerd et al.,

2004; Mangeat et al., 2004; Schrofelbauer et al., 2004; Xu et

al., 2004).

In humans, the APOBEC family of cytidine deaminases

includes APOBEC1, APOBEC2, APOBEC3A-G and the

activation induced deaminase (AID) (Harris et al., 2002;

Jarmuz et al., 2002). APOBEC1 is an RNA editing enzyme

which in a complex with APOBEC complementing factor

deaminates a single cytosine of apolipoprotein B mRNA to

introduce a termination codon. APOBEC1 also has non-

specific cytosine deaminase activity on DNA (Harris et al.,

2002). AID is required for somatic hypermutation and class

switch recombination (Muramatsu et al., 2000; Muto et al.,

2000; Revy et al., 2000) in B lymphocytes (Jarmuz et al.,

2002).

APOBEC family members contain a domain structure

characteristic of cytidine deaminases (Jarmuz et al., 2002).

A short a-helical domain is followed by a catalytic domain

(CD), a short linker peptide and a pseudocatalytic domain

(PCD). In APOBEC3B, APOBEC3F and APOBEC3G, the

entire unit is duplicated to form the domain structure helix1-

CD1-linker1-PCD1-helix2-CD2-linker2-PCD2. Each cata-

lytic domain contains the conserved motif H-X-E-(X)27–28-

P-C-X-X-C in which the His and Cys residues coordinate

Zn2+ and the Glu serves as a proton shuttle in the

deamination reaction (Betts et al., 1994; Wedekind et al.,

2003). CD1 and CD2 contain conserved aromatic residues

located between the H-X-E and P-C-X-X-C residues (Phe70

and Tyr91 in CD1 and Phe262 and Phe282 in CD2) that in

APOBEC1 are required for RNA binding (Navaratnam et

al., 1995).

The mechanism by which APOBEC3G is encapsidated

in HIV-1 virions has been the subject of several recent

reports. There is general agreement that encapsidation is

dependent upon the nucleocapsid (NC) domain of Gag

(Alce and Popik, 2004; Cen et al., 2004; Douaisi et al.,

2004; Luo et al., 2004; Schafer et al., 2004; Zennou et al.,

2004). Interaction between APOBEC3G and NC has been

shown by GST-pull down assays (Alce and Popik, 2004;

Douaisi et al., 2004), coimmunoprecipitation experiments

(Cen et al., 2004; Luo et al., 2004; Zennou et al., 2004) and

also in a binding assay using in vitro translated proteins

(Schafer et al., 2004). NC also serves to capture the viral

genomic RNA and thus it is possible that the RNA can

contribute to APOBEC3G encapsidation. Evidence showing

that purified recombinant APOBEC3G and NC directly

interact (Alce and Popik, 2004) and that RNase A treatment

does not alter the ability to coimmunoprecipitate Gag and

APOBEC3G (Cen et al., 2004; Douaisi et al., 2004) would

suggest no requirement for RNA. However, others have

found that the interaction was sensitive to ribonucleases

(Schafer et al., 2004; Svarovskaia et al., 2004; Zennou et al.,

2004).

Here, we further defined the determinants of APO-

BEC3G that are required for encapsidation. Our results

show that the critical domain of APOBEC3G required for

encapsidation mapped to the conserved Zn2+ coordinating

residues of the N-terminal catalytic domain (CD1) and that

these residues are required for RNA binding. Mutational

analyses showed that CD1 was important for RNA binding,

homodimerization and encapsidation, while CD2 was the

catalytically active site. The findings suggest that APO-

BEC3G CD1 serves as the encapsidation and dimerization

domain and CD2 catalyzes deamination of the reverse

transcripts.

Results

Virions deleted for NC encapsidate a significant amount of

APOBEC3G

Several reports have demonstrated the dependence of

APOBEC3G encapsidation on the Gag nucleocapsid (NC)

(Alce and Popik, 2004; Cen et al., 2004; Douaisi et al.,

2004; Luo et al., 2004; Schafer et al., 2004; Svarovskaia

et al., 2004; Zennou et al., 2004). However, in our

analysis of a minimal Gag that was deleted for NC,

called Zwt, we found a small but consistent amount of

APOBEC3G (Fig. 1A). In this construct, NC is replaced

by the leucine zipper domain of the yeast transcription

factor GCN4 and lacks p1 and p6 (Accola et al., 2000).

To determine whether this might have been due to over-

expression of APOBEC3G, we tested its encapsidation

over a wide range of expression levels, comparing to

full-length construct, Pr55 (Fig. 1A). Because Zwt

expressed Vif, the APOBEC3G with a D128K mutation

which prevents the interaction with HIV-1 Vif was used

(Schrofelbauer et al., 2004). Zwt particles consistently

encapsidated about 10-fold less APOBEC3G than Pr55

over the range of concentrations. The amount of

APOBEC3G synthesized in the cells was proportional

to the amount of plasmid in the transfection. Thus, in

agreement with the findings of others (Alce and Popik,

2004; Cen et al., 2004; Douaisi et al., 2004; Luo et al.,

2004; Schafer et al., 2004; Zennou et al., 2004), NC was

an important determinant of APOBEC3G encapsidation,

but a significant amount of APOBEC3G encapsidation

occurred even in the absence of NC. Quantitation of the

viral RNA content of the Zwt virions by dot blot,

showed that even without NC a small amount of viral

RNA had been packaged (Fig. 1A, bottom). These results

suggested that APOBEC3G could be encapsidated inde-

pendent of NC, perhaps as a result of binding to RNA.

The viral RNA is not required

To determine whether this could have been due to an

association with the viral RNA, we analyzed APO-

BEC3G incorporation into virus particles that lacked

viral genomic RNA. We generated VLPs using a CMV-

Fig. 1. Encapsidation of significant amounts of APOBEC3G into NC-deleted virions and lack of requirement for viral RNA. (A) Virions composed of full-

length Gag Pr55 or deleted Gag lacking NC, p1 and p6 (diagrammed above) were generated in the presence of decreasing amounts of HA tagged APOBEC3G

expression vector. The amount of pcAPOBEC3G-HA transfected is indicated above in Ag. APOBEC3G content of the virions and transfected cells was

analyzed on immunoblots probed with anti-HA Mab. The blots were reprobed with anti-p24 MAb to determine the amount of virions. (B) Virions lacking viral

genomic RNA were tested for APOBEC3G encapsidation using the gag-pol expression vector pMDL-RRE (diagrammed above). For this experiment, the

amount of particles was normalized by p24. Arrows on the left indicate the position of APOBEC3G, CA, Pr55 and Zwt. Dot blot measurement of the viral

RNA content of the virions is shown at the bottom. Controls are included for transfection of a NL4-3 with a stop codon in Gag (Gag�) and lacking viral DNA

(empty). PhosphoImager quantitation of the band intensities is shown below.

F. Navarro et al. / Virology 333 (2005) 374–386376

driven Gag-Pol expression vector pMDL-RRE that

lacked the packaging sequences upstream the gag.

These VLPs contained wild-type amounts of APO-

BEC3G (Fig. 1B, middle) but contained no detectable

viral RNA (Fig. 1B, bottom). The strength of the signal

in the DVif particles as compared to the Gag-Pol

particles (Fig. 1B, bottom) clearly showed that APO-

BEC3G encapsidation is not dependent on the viral

genomic RNA, which is consistent with previous

findings (Cen et al., 2004; Svarovskaia et al., 2004;

Zennou et al., 2004).

APOBEC3G encapsidation requires the Zn2+ coordination

residues of CD1

The feature of APOBEC family proteins that allows

their encapsidation into retroviral virions is not known. To

map the domains of APOBEC3G required for encapsida-

tion, we generated a panel of APOBEC3G deletion mutants

(Fig. 2A) and tested their encapsidation into DVif HIV-1

particles. With the exception of the N-terminal fragment,

the deleted proteins were stably expressed in transfected

cells. For several of the deletion mutants additional bands

were seen on the blots which might correspond to products

of degradation or post-translational modifications (Fig. 2B;

the bands corresponding to the expected sizes are indicated

with an asterisk). It is of note that several species of

different mobility were detected as well for APOBEC3G

wild type. Importantly, none of the deletion mutants was

efficiently encapsidated (Fig. 2B). Deletion of as little as

the first 54 amino acids prevented encapsidation.

To further define critical amino acid residues for

encapsidation, point mutants of the conserved residues of

the zinc finger motifs in CD1 and CD2 were tested (Fig. 3A).

Fig. 2. Encapsidation of APOBEC3G deletion mutants into DVif NL4-3 virus particles. (A) Diagrams of APOBEC3G deletion mutants. The numbers indicate

amino acid positions in human APOBEC3G. All deletion constructs contain a HA-tag at the carboxy terminus of the protein. (B) APOBEC3G deletion

constructs were cotransfected with the proviral DNA into 293T cells. Cell lysates and viruses were prepared 48 h after transfection and proteins were analyzed

by immunoblotting. Full-length and mutant APOBEC3Gs were detected using an anti-HA antibody. Viral Gag proteins were detected by an anti-p24 antibody.

Comparable amounts of virus particles were used, as indicated by immunoblotting with the anti-p24 MAb. Quantitative analysis, calculated as the relative ratio

of APOBEC3G in virions as compared to cell lysates and normalized for the ratio obtained with APOBEC3G wild-type, is shown below.

F. Navarro et al. / Virology 333 (2005) 374–386 377

Mutation of the His65/Cys97/Cys100 box and the

catalytic Glu67 in CD1 (Zn1) to Ala prevented encap-

sidation (Fig. 3B). Mutation of the analogous amino

acids in CD2 (Zn2), in contrast, had no significant effect.

The double mutant (Zn1/2) was also not encapsidated

(Fig. 3B). Thus, the conserved residues of CD1 but not

CD2 were critical for encapsidation. The mutated proteins

were stably expressed although those with mutations in

CD1 were expressed with a modest decrease in steady-

state level.

Mutation of the Zn2+ coordinating Cys residues, Cys97

and Cys100, of CD1 to Ala (1CCAA) was sufficient to

substantially reduce encapsidation (Fig. 3C). The analogous

mutations in CD2 of Cys288 and Cys291 to Ala (2CCAA)

had no significant effect. Mutation of the Zn2+ coordination

Cys residues in CD1 to Ser, a more conservative change,

also substantially reduced encapsidation (Fig. 3C).

Additional mutants were tested in which two conserved

aromatic residues within the core motif that have been

shown to mediate RNA binding (Navaratnam et al., 1995)

were changed to Ala. Mutations F70A/Y91A in CD1

(1FYAA) prevented encapsidation (Fig. 3C). The analogous

change in CD2 of F262A/F282A (2FFAA) also decreased

encapsidation, although its effect was not as pronounced as

in the 1FYAA mutant (Fig. 3C). These findings suggested

that the Zn2+ coordinating residues and the conserved

residues involved in RNA binding in CD1 were critical

for APOBEC3G encapsidation.

To further test the role of the conserved residues, single

point mutants were analyzed. Single mutations in the CD1

Zn2+ coordination residues H65A, C97A and C100A all

prevented or reduced encapsidation, while mutation of the

catalytic Glu67 had no significant effect (Fig. 4A).

Similarly, mutation of the catalytic residue Glu259 had no

major effect on encapsidation (Fig. 4A). All of the proteins

were stably expressed although, changes in the Zn2+

coordination residues caused a 2- to 4-fold decrease in

steady-state level suggesting that these residues also play a

structural role in the enzyme. Taken together, these findings

suggested the Zn2+ coordination residues of CD1 are critical

for encapsidation.

The catalytic residue Glu259 in CD2 is required for

antiviral activity

The contribution of the conserved amino acids of the

zinc finger motifs in CD1 and CD2 to the antiviral activity

of APOBEC3G was determined by single-cycle luciferase

reporter virus assay. The CD1 mutants Zn1, Zn1/2,

1CCAA, 1FYAA and the CD2 mutant 2FFAA which were

poorly encapsidated had no effect on infectivity (Fig. 4B,

top). The CD2 mutants Zn2 and 2CCAA also had no

effect on infectivity, despite of being efficiently encapsi-

dated (Fig. 4B, top). Single amino acid mutations of the

Zn2+ coordination residues His65, Cys97 and Cys100 of

CD1 had no major effect on infectivity (Fig. 4B, bottom).

Fig. 3. The Zn2+ coordination residues of CD1 are required for efficient encapsidation. Virions were generated with Dvif NL4-3 (Dvif) and the mutated

APOBEC3G-HA expression vector indicated above each lane. (A) Diagram showing the altered conserved residues on each mutant. (B) APOBEC3G mutated

in the active site motif of CD1 (Zn1), CD2 (Zn2) or CD1 and CD2 (Zn1/2). APOBEC3G content was analyzed on immunoblots probed with anti-HA MAb.

The position of APOBEC3G and CA is indicated by arrows on the left. (C) APOBEC3G mutated in the Zn2+ coordination residues and aromatic conserved

RNA binding residues of CD1 and CD2. Quantitative analysis, as performed in Fig. 1, is shown for panels B and C.

F. Navarro et al. / Virology 333 (2005) 374–386378

Importantly, mutation of the CD1 catalytic residue Glu67

did not reduce antiviral activity. In contrast, mutation of

the analogous CD2 amino acid, Glu259, largely dimin-

ished the antiviral effect (Fig. 4B, bottom). These findings

suggested that the catalytic activity that deaminates the

viral genome is primarily mediated by CD2.

Cytidine deaminase activity resides in CD2

To directly test the relative contributions of the active

sites, we measured cytidine deaminase activity of the

mutant proteins. The mutant proteins were encapsidated

in virions and then released by detergent solubilization.

Cytidine deaminase activity was tested using a 5V-endlabeled oligonucleotide substrate that contained a consensus

CCCA target site (Yu et al., 2004b). The molecules were

cleaved at the deaminated dU by treatment with uracil DNA

glycosylase and NaOH and detected by autoradiography.

The 1CCAA mutant was partially active while the 2CCAA

was inactive (Fig. 5). The partial activity of 1CCAA should

be considered in light of its inefficient encapsidation. This

mutant may be almost fully active on a molar basis.

Importantly, mutation of the CD1 catalytic amino acid

Glu67 resulted in a fully active enzyme, while mutation of

the analogous CD2 amino acid in Glu259 inactivated the

enzyme. Thus, CD2 appeared to mediate deamination of

DNA.

RNA binding requires the Zn2+ coordination residues of

CD1

It has been described that APOBEC3G is able to bind

RNA in vitro (Jarmuz et al., 2002; Yu et al., 2004b). To

determine the relative contribution of the Zn2+ coordina-

tion residues of CD1 and CD2 to the ability of

APOBEC3G to bind RNA, recombinant proteins with

mutations in these residues were tested for RNA binding

by EMSA (Yu et al., 2004b). Wild-type, 1CCAA and

2CCAA mutant APOBEC3Gs were purified from E. coli

as glutathione S-transferase (GST) fusion proteins. In

Fig. 4. Mutations in the conserved Zn2+ coordinating motif of CD2 abrogate antiviral activity. (A) Immunoblot analysis of encapsidation of APOBEC3G

mutants with individual amino acid residues in the conserved Zn2+ binding motif changed to Ala in CD1 (H65A, E67A, C97A and C100A) and CD2 (E259A).

Quantitative analysis, as performed in Fig. 1, is shown below. (B) Antiviral activity of the indicated mutated APOBEC3Gs was determined by single-cycle

luciferase reporter virus assay. Negative controls include infections using supernatant from 293T cells transfected with pcDNA alone or pcAPOBEC3G alone.

F. Navarro et al. / Virology 333 (2005) 374–386 379

addition, APOBEC3A was used as a control. The two

cysteines in the active site of APOBEC3A are separated by

four amino acid residues and therefore is postulated not to

Fig. 5. APOBEC3G deaminase activity is mediated by the CD2 active site.

Virions containing wild-type or the indicated mutant APOBEC3G were

produced in 293T cells, pelleted and solubilized in detergent-containing

buffer to release the encapsidated enzyme. The lysates were incubated with

5V-(32P) labeled oligonucleotide and then cleaved at sites of CYU

deamination by treatment with UDG. As controls, virions were produced

in the absence of APOBEC3G (lane 1) or without viral plasmid (lane 2). An

oligonucleotide containing a dU in place of the target dC served as a marker

for the cleaved product (lane 8).

coordinate zinc (Jarmuz et al., 2002). The proteins were

tested for binding to oligoribonucleotides that lacked C

and U (no C/U) that contained C (C-rich) or to

deoxyoligonucleotides lacking C (no C) or containing C

(C-rich) (Fig. 6). Wild-type APOBEC3G bound to

ribooligonucleotides regardless of its C-content and bound

to the deoxyoligonucleotides that contained C. Thus, the

enzyme bound nonspecifically to RNA but bound to DNA

that is targeted by for deamination. Control GST and GST-

APOBEC3A did not bind to nucleic acid. Alteration of the

Zn2+ coordination residues in CD1 dramatically reduced

the ability of the protein to bind RNA and DNA (Fig. 6).

Alteration of CD2 had no effect on binding of oligor-

ibonucleotides that contained C (C-rich) or deoxyoligonu-

cleotides that contained C. This mutant may have a

reduced ability to bind oligoribonucleotides that lacked C

and U although this difference could be within exper-

imental variability (no C/U) (Fig. 6). We concluded that

nucleic acid binding is primarily mediated by CD1.

CD1 Zn2+ coordination residues are required for

APOBEC3G dimerization

APOBEC1, APOBEC2, APOBEC3B and APOBEC3G

are homodimers (Jarmuz et al., 2002; Lau et al., 1994). To

determine which domains of APOBEC3G are required for

Fig. 6. RNA binding by APOBEC3G is primarily determined by CD1 Zn2+ coordination residues. APOBEC3A and APOBEC3G wild-type and the mutants

1CCAA and 2CCAA were produced in E. coli as GST fusion proteins and purified by affinity chromatography. 5V-end 33P labeled oligoribonucleotide was

incubated with the indicated APOBEC3G proteins. Two RNA oligonucleotides were used, one lacked C and U (no C/U) and one was enriched for C (C-rich)

and two DNA oligonucleotides were used, one lacking C (no C) and one enriched for C (C-rich). The complex was separated on native PAGE and exposed to

X-ray film. A low mobility nonspecific band is present for all proteins incubated with the no C/U oligoribonucleotide.

F. Navarro et al. / Virology 333 (2005) 374–386380

dimerization, the ability of 1CCAA and 2CCAA to dimerize

with wild-type APOBEC3G was tested (Fig. 7). 293T cells

were cotransfected with equal amounts of mutated HA-

tagged APOBEC3G and wild-type myc-tagged APO-

BEC3G expression vector. Dimerization was assessed by

coimmunoprecipitation followed by immunoblot analysis.

Wild-type APOBEC3G-myc coimmunoprecipitated with

wild-type APOBEC3G-HA. In this analysis, the 2CCAA

protein dimerized but 1CCAA did not (Fig. 7). Thus, the

Zn2+ coordination residues in CD1 but not in CD2 are

required for dimerization of APOBEC3G.

CD1 mediates encapsidation of APOBEC3G into human T

cell leukemia virus type 1 (HTLV-I) and murine leukemia

virus (MuLV)

HTLV-I and HTLV-II do not encode Vif yet both

replicate in T cells. To determine whether HTLV-I encap-

sidated APOBEC3G and whether this affected its infectivity,

HTLV-I virions were produced in 293T using an HTLV-I

packaging construct and a luciferase encoding transfer

vector (Derse et al., 2001). The virions were produced in

Fig. 7. APOBEC3G dimerization requires CD1 Zn2+ coordination residues. 293T

vector. HA tagged APOBEC3G was wild-type, mutated in CD1 (1CCAA) or CD

fragment that includes aa 163 to 384 of APOBEC3G). Cell lysates were prepared

analyzed on immunoblots probed with anti-myc MAb. All of the proteins were ex

with empty vector (pcDNA) or APOBEC3G-myc alone (APO3G-myc).

cells cotransfected with pc-APOBEC3G-HA and APO-

BEC3G content was determined by immunoblot analysis.

Encapsidated APOBEC3G was readily detected (Fig. 8A).

The analysis did not permit absolute quantitation, but the

intensity of the APOBEC3G and CA bands were similar to

those found for HIV-1 virions. Intracellular levels of

APOBEC3G were not affected by the presence of the

HTLV-I proteins in the transfected cells, suggesting that the

virus did not induce APOBEC3G degradation. The effect of

APOBEC3G on virion infectivity was determined by

titrating the amount of pcAPOBEC3G-HA plasmid and

including VSV-G vector to generate infectious pseudotypes.

The virus was then used to infect fresh 293T cells (Fig. 8B).

For comparison, Dvif HIV-1 virions were tested in parallel.

At the highest amount of pcAPOBEC3G (2 Ag), infectivitywas reduced about 2-fold. In contrast, the infectivity of Dvif

HIV-1 was reduced about 30-fold. With 1 Ag of pcAPO-

BEC3G HTLV-I infectivity was not significantly reduced

(Fig. 8B). Thus, HTLV-I, like MuLV (Mariani et al., 2003),

appeared to be relatively resistant to APOBEC3G.

To determine whether APOBEC3G was encapsidated

into divergent retroviruses by a mechanism similar to that

cells were cotransfected with HA or myc-tagged APOBEC3G expression

2 (2CCAA) Zn2+ coordination Cys residues or was deleted for CD1 (C-

and immunoprecipitated with anti-HA Mab. The immunoprecipitates were

pressed at similar levels (not shown). As controls, the cells were transfected

Fig. 8. HTLV-I is relatively resistant to encapsidated APOBEC3G. (A) HTLV-I virions were pelleted from the culture medium of 293T cells cotransfected with

the HTLV packaging plasmid pCMVHT-Denv, the transfer vector pHTC-luc (Derse et al., 2001) and pcAPOBEC3G-HA. APOBEC3G and HTLV-I CA p19

were visualized on immunoblots with anti-HA or anti-p19 MAb. (B). VSV-G pseudotyped HTLV-I luciferase reporter viruses were produced in 293T cells by

cotransfection with decreasing amounts of APOBEC3G expression vector (indicated below in Ag of transfected plasmid). Fresh 293T cells were infected with 2

ml of virus-containing supernatant. NL-Luc HIV-1 luciferase reporter viruses were prepared and analyzed in parallel. The results are representative of three

independent experiments.

F. Navarro et al. / Virology 333 (2005) 374–386 381

for HIV-1, encapsidation of mutated APOBEC3Gs into

MuLV and HTLV-I was tested. As for HIV-1, the 1FYAA

mutant in which aromatic residues of CD1 were mutated

was not encapsidated into MuLVor HTLV-I. In contrast, the

2FFAA mutant was encapsidated in both viruses (Fig. 9).

The 2FFAA mutant was encapsidated at somewhat lower

efficiency into HTLV-I, which is consistent with the results

obtained for HIV-1 (compared to Fig. 3C). Overall,

APOBEC3G was encapsidated into MuLV and HTLV-I by

a mechanism that was similar to HIV-1.

Discussion

The findings presented here suggest a model for

APOBEC3G function in which domain 1 directs encapsi-

dation and domain 2 serves as the catalytic site. The

antiviral activity of APOBEC3G requires the function of

both domains. In support of this model, domain 1 Zn2+

coordination amino acids were required for encapsidation,

while domain 2 Zn2+ coordination amino acids were

dispensable. A stable C-terminal fragment of APOBEC3G

was not encapsidated. The active site of cytidine deaminase

domains contains an essential Glu residue. Domain 1 Glu67

was dispensable for antiviral activity while domain 2

Glu259 was required. It should be noted that E259A

retained weak antiviral activity, although no signal was

detected in the cytidine deaminase activity assay. The

residual antiviral activity could be taken to indicate an

antiviral activity separate from the deaminase activity

(Shindo et al., 2003; Turelli et al., 2004), but it is more

likely that the protein retains weak deaminase activity that is

beyond the limit of sensivity of the in vitro assay. Neither of

the catalytic Glu residues was required for encapsidation.

RNA binding by the APOBEC3G and homodimerization

were mediated by the domain 1 Zn2+ coordination residues.

It may be that a two-domain structure is needed for the

antiviral effect of APOBEC3 proteins in general, as the two

domain proteins APOBEC3B, APOBEC3F and APO-

BEC3G are active while APOBEC3A which consists of a

single domain is not (Bishop et al., 2004; Yu et al., 2004a).

The mechanism of APOBEC3G encapsidation has been

addressed in several recent reports. There is general agree-

ment that NC is an important determinant of APOBEC3G

Fig. 9. APOBEC3G encapsidation into MuLV and HTLV-I virions is dependent upon conserved aromatic residues in CD1. HTLV-I (right panel) and MuLV

(left panel) virions were generated by cotransfection of 293T cells with viral plasmid DNA and wild-type or mutant 1FYAA or 2FFAA APOBEC3G-HA

expression vector. Pelleted virions and cell lysates were analyzed on immunoblots probed with anti-HA or anti-p15 or anti-p19 MAbs as indicated at right. The

positions of APOBEC3G, MuLV CA p15 and HTLV CA p19 are shown on the left. Quantitative analysis, as performed in Fig. 1, is shown below.

F. Navarro et al. / Virology 333 (2005) 374–386382

encapsidation (Alce and Popik, 2004; Cen et al., 2004;

Douaisi et al., 2004; Luo et al., 2004; Schafer et al., 2004;

Zennou et al., 2004). However, because NC mediates viral

RNA packaging (Aldovini and Young, 1990; Gorelick et al.,

1990), it has been difficult to distinguish whether the

requirement stems from direct binding of APOBEC3G to

NC or to the viral RNA. Evidence that binding to NC was

sufficient was suggested by the finding that purified

recombinant APOBEC3G and NC directly interact (Alce

and Popik, 2004) and that RNase A treatment does not

interfere with the ability to coimmunoprecipitate Gag and

APOBEC3G from transfected cells (Cen et al., 2004;

Douaisi et al., 2004). However, others have found that the

interaction was sensitive to ribonucleases (Schafer et al.,

2004; Svarovskaia et al., 2004; Zennou et al., 2004). Viral

RNA packaging-deficient viruses incorporated normal

amounts of APOBEC3G, as shown here and by others

previously (Cen et al., 2004; Luo et al., 2004; Schafer et al.,

2004; Svarovskaia et al., 2004; Zennou et al., 2004).

However, such virions are likely to package cellular RNA

(Aronoff and Linial, 1991; Berkowitz et al., 1996) and these

could substitute for the viral RNA in facilitating APO-

BEC3G encapsidation. Although it is clear that viral RNA is

not required, our findings support a role for RNA in

APOBEC3G encapsidation. First, virions deleted for NC

consistently packaged about 10% the wild-type amount of

APOBEC3G, demonstrating that the virus has another

determinant for APOBEC3G encapsidation besides NC.

Second, mutations in APOBEC3G that affected RNA

binding blocked encapsidation. The Zn2+ coordination

residues of domain 1 which were required for RNA binding

were also required for APOBEC3G encapsidation. In

addition, the conserved aromatic residues F70 and Y91 of

CD1, which in APOBEC1 are required for RNA binding

(Navaratnam et al., 1995) were also required for APO-

BEC3G encapsidation in HIV-1, MLV and HTLV-I.

Our findings differed somewhat from those of Cen et al.,

and Luo et al., who mapped an encapsidation determinant to

the N-terminal linker domain of APOBEC3G (Cen et al.,

2004; Luo et al., 2004). In those studies, the proteins were

encapsidated into Gag VLPs and not into native HIV-1

virions as in our study, although whether this could have

influenced the findings is not clear. Consistent with our

results, Li et al., found that a D1–67 APOBEC3G deletion

mutant, that lacks the conserved His of the first conserved

Zn2+ finger, was not encapsidated into DVif HIV-1 virions

(Li et al., 2004).

The ability of NC-deleted virions to encapsidate a small

but significant amount of APOBEC3G argues against a

simple protein–protein interaction between APOBEC3G

and NC. Moreover, the broad species-specificity of the

encapsidation also argues against this. APOBEC3G is

efficiently encapsidated by divergent retroviruses including

MuLV and HTLV-1 (Harris et al., 2003; Mangeat et al.,

2003; Mariani et al., 2003), the NC proteins of which are

highly divergent. Moreover, mouse APOBEC3, which has

only ~30% homology to the human protein is also

efficiently encapsidated in HIV-1 (Mariani et al., 2003). In

addition, it seems likely that if binding to NC was the sole

determinant of APOBEC3G encapsidation, the virus would

have escaped simply by altering the NC amino acid

sequence. A role for RNA in APOBEC3G encapsidation

would provide a mechanism from which the virus could not

easily escape. A role for RNA, however, presents the

paradox of how APOBEC3G is encapsidated into packaging

site mutant virions. Such virions are thought to package

cellular RNA yet it does not seem likely that APOBEC3G

molecules could coat much of the cellular mRNA. We

F. Navarro et al. / Virology 333 (2005) 374–386 383

therefore speculate that APOBEC3G encapsidation is

mediated by a complex of NC with RNA in which

APOBEC3G binds to a composite site formed by the

complex of NC with RNA.

Lastly, we found that APOBEC3G could be encapsidated

in HTLV-I virions yet did not reduce its infectivity. HTLV-I

replicates in T cells yet does not encode a Vif-like protein

and did not appear to induce APOBEC3G degradation,

arguing that it does not encode a Vif analog. This finding

raises the question of how viruses such as HTLV-1 and

MuLV that lack Vif are able to replicate in the presence of

APOBEC3G. One possibility is that their reverse tran-

scription complex is not accessible to the encapsidated

deaminase. This further poses the question of why HIV-1

was not able to become resistant to deamination by a similar

mechanism but instead used the alternative strategy of

developing a novel gene as a means of escaping this cellular

antiviral activity.

Materials and Methods

APOBEC3G expression vectors

HA and myc tagged human APOBEC3G expression

vectors pcAPOBEC3G-HA (Mariani et al., 2003) and

pcAPOBEC3G-myc were constructed by cloning a human

APOBEC3G cDNA into the EcoRI and XhoI sites of

pcDNA3.1(+) (Invitrogen) using primers that encoded the

corresponding C-terminal epitope tag. APOBEC3G mutants

were generated by overlapping PCR using pcAPOBEC3G-

HA as template. 5V and 3V fragments were amplified

separately with a primer specific for the overlap region

and a 5V external primer 5V-Apo3G.EcoRI (5V-TAG ATC

GAATTC ATG AAG CCT CAC TTC AGA AAC ACA G)

or a 3V primer 3V-Apo3G.HA.Xho (5V-ATT GAATCT CGA

GTC AAG CGT AAT CTG GAA CAT CGT ATG GAT

AGT TTT CCT GAT TCT GGA GAATGG CCC GCA G).

The 5V and 3V fragments were mixed and amplified with the

two external primers. The amplicon was cleaved with EcoRI

and XhoI and cloned into pcDNA3.1(+). All plasmids were

confirmed by sequencing. pcAPOBEC3G D128K has been

previously described (Schrofelbauer et al., 2004).

HIV-1 and HTLV-I proviral plasmids

Single cycle Dvif HIV-1 (pNL4-3 E� R� Dvif) provirus

and Dvif HIV-1 luciferase reporter virus (pNL-Luc-E� R�

Dvif) were previously described (Schrofelbauer et al., 2004).

Pr55 encodes a full-length HIV-1 genome with a point

mutation that inactivates protease (Accola et al., 2000). Zwt

mutant provirus does not contain p1, p6 and NC was

replaced with a yeast leucine zipper domain (Accola et al.,

2000). pMDLg/p is a CMV-driven expression plasmid that

contains only the gag and pol coding sequences from HIV-1

(Dull et al., 1998). HTLV-I virions were generated with the

HTLV-I packaging plasmid, pCMVHT-Denv and transfer

vector pHTC-luc (Derse et al., 2001). MLV virions were

generated with Gag/Pol expression vector pHIT-60

(Soneoka et al., 1995).

Production of recombinant viruses and detection of

encapsidated APOBEC3G

APOBEC3G encapsidation into HIV-1, HTLV-I or MLV

virions was determined as previously described (Mariani et

al., 2003). 293T cells were maintained in Dulbecco’s

modified essential medium supplemented with 10% fetal

calf serum and antibiotics. Briefly, 293T cells were seeded

at 5 � 105 per well in 6 well plates. The following day, the

cells were cotransfected with 1 Ag each of pcAPOBEC3G-

HA and HIV-1, HTLV-I or MLV plasmid using Lipofect-

amine 2000 (Invitrogen). Two days later, virus containing

supernatants were collected and virions were ultracentri-

fuged through a 20% sucrose cushion at 150,000 �g for

1.5 h. The pellets were solubilized in 100 Al of 1% Triton-

containing buffer and the proteins in 10 Al were separated

on 4–12% PAGE. The proteins were then transferred to

PDVF membranes and probed with anti-HA MAb 16B12

(Covance) followed by horseradish peroxidase-(HRP)

conjugated sheep anti-mouse Ig (Amersham). The mem-

branes were stripped and reprobed with anti-HIV-1 human

serum J810 (provided by Dr. Douglas Richman, UCSD),

anti-p24 MAb AG3.0, anti-HTLV-1 p19 MAb (Zymed) or

anti-MLV p15 goat serum followed by HRP conjugated

anti-human, anti-mouse or anti-goat IgG (Amersham) and

developed with ECL reagents (Amersham). Bands in

Western blots were quantitated by densitometry using

FluorChem 8000 (Alpha Innotech Corporation).

Dot blot analysis

Viral genomic RNA content of wild type and recombi-

nant HIV-1 virus was determined by dot blot analysis.

Briefly, virion supernatants from transfected 293T cells

were treated with 20 Ag/ml of DNaseI, RNase-free (Roche)

for 1 h at RT and then the virions were purified by

ultracentrifugation through 20% sucrose, resuspended in

PBS and normalized by p24. RNA was extracted from

equal amounts of p24 (200 ng) using Trizol Reagent

(Invitrogen) and then transferred to a nylon membrane

using a dot blot apparatus (Schleicher and Schuell). The

membrane was prehybridized 1 h at 68 8C in Express-Hyb

Hybridization solution (Clontech). Incubation with the

radiolabeled probe (106 cpm/ml) was for 2 h at 68 8C. Asprobe we used an EcoRI/XhoI fragment of pNL4-3

(corresponding to nucleotides 5743 to 8887 of pNL4-3)

random labeled with a32P-dCTP (ICN) using High Prime

(Roche). The membrane was washed twice with 2� SSC–

0.05% SDS for 20 min at RT and once with 0.1� SSC–

0.1% SDS for 40 min at 50 8C and analyzed with a

PhosphorImager.

F. Navarro et al. / Virology 333 (2005) 374–386384

APOBEC3G dimerization assay

239T cells were cotransfected in a six well dish with 1

Ag pcAPOBEC3G-HA or its mutated derivatives and 1 AgpcAPOBEC3G-myc. Two days posttransfection, the cells

were lysed in 1.0 ml CHAPS buffer (5 mM CHAPS, 50

mM Tris, pH 8.0, 50 mM NaCl) at 4 8C for 1 h. The lysates

were cleared by microcentrifugation for 20 min at 4 8C.Lysates (250 Al) were precleared twice with 50 Al of a 1:1

slurry of protein-A sepharose for 30 min at 4 8C. Anti-HAMAb (1 Ag) was added to 200 Al of precleared lysate and

incubated 2 h at 4 8C followed by addition of 30 Al ofprotein-A sepharose beads. After 1 h, the beads were

collected and washed three times in 1 ml of CHAPS buffer

followed by two washes in 1 ml of 50 mM Tris, pH 8.0, 50

mM NaCl. The supernatant was removed and 15 Al of 2�reducing sample buffer (Invitrogen) was added. Samples

were heat to 958 for 5 min and then separated on 4–20%

SDS-PAGE. The proteins were transferred to PVDF

membranes and probed with anti-myc MAb 9E10 (Cova-

nce) followed by HRP-conjugated sheep anti-mouse Ig

(Amersham). The membranes were developed using ECL

Reagent (Amersham).

HIV-1 and HTLV-I luciferase reporter virus assay

VSV-G-pseudotpyed single-cycle HIV-1 and HTLV-I

luciferase reporter viruses were generated as previously

described (Connor et al., 1995; Derse et al., 2001). Briefly,

293T cells were cotransfected using Lipofectamine 2000

(Invitrogen) with 1 Ag of HIV-1 or HTLV-I luciferase

reporter virus, 1 Ag of wild type or mutant pcAPOBEC3G-

HA or pcDNA3.1(+) and 1 Ag of pcVSV-G. For

APOBEC3G titrations, the total amount of plasmid was

held constant by the addition of pcDNA3.1(+). Virus-

containing supernatant was harvested 2 days posttransfec-

tion filtered and the p24 concentration measured. 293T

cells (2 � 105) seeded the previous day in six-well plates

were infected with 300 Al of HIV-1 or 2 ml of HTLV-I

containing supernatant in the presence of 8 Ag/ml

polybrene (Sigma). After 6 h, medium was removed

and replaced with fresh medium. To assay for luciferase

activity, cells were harvested 72 h post infection, pelleted

and resuspended in 0.1 ml of lysis buffer (1% Triton X-

100, 50 mM NaCl, 10 mM Tris–HCl (pH 7.6), 5 mM

EDTA). Luciferase assays were performed with 20 Al ofcell extract in the Promega luciferase assay system

according to the manufacturer’s protocol. Data shown

are the mean cps +/� the standard error of the

triplicates.

APOBEC3G RNA binding assay

Vectors to produce recombinant APOBEC3 proteins

were constructed by transferring APOBEC cDNAs to

pGEX-6P-3 plasmid (Amersham) in-frame to the GST

coding sequence at the EcoRI and XhoI sites. Recombinant

GST-APOBEC3G proteins (wild-type, CD1 1CCAA and

CD2 2CCAA mutants) and GST-APOBEC3A were pro-

duced in E. coli BL21-DE3 induced with 0.5 mM IPTG.

Bacteria were pelleted and lysed in PBS containing 10 mM

EDTA/5 mM dithiothreitol/1% triton-X100. The proteins

were purified by affinity chromatography on glutathione

sepharose beads. DNA and RNA oligonucleotides

described by Yu et al. (2004b) were 5V-end labeled with33P-gATP using T4 polynucleotide kinase and unincorpo-

rated nucleotide was removed on G25 spin column. Each

protein (2 Ag) was incubated with oligonucleotide (1 � 105

cpm) in binding buffer (10 mM HEPES, 10% glycerol, 100

mM KCl, 10 mM MgCl2, 0.1 mM EDTA, 0.5 mM DTT) at

378 for 40 min and then separated by native PAGE on a 4–

20% gradient gel. The gel was dried and exposed to

Biomax MR (Kodak) film overnight.

Cytidine deaminase assay

Virions were generated by cotransfection of 293T cells

with 1:1 ratio of Dvif pNL4-3 and pcAPOBEC3G-HA.

Supernatant was harvested 3 days posttransfection, filtered

and pelleted through 2 ml of 20% (w/v) sucrose/PBS by

ultracentrifugation for 1h at 35,000 RPM in an SW41 rotor.

Pelleted virions were lysed in 100 Al buffer containing 50

mM Tris (pH 8.0), 40 mM KCl, 50 mM NaCl, 5 mM

EDTA, 10 mM DTT and 0.1% (v/v) triton-X100. Lysates

were normalized for p24 concentration. The virus lysate

(100 ng p24) was mixed with 1 � 105 cpm of 5V-endlabeled oligonucleotide with the sequence 5V-TTT TTT

TTT TTT TTT CCC GTT TTT TTT TTT TTT TTT TTT

TT T in deaminase buffer (40 mM Tris, pH 8.0, 40 mM

KCl, 50 mM NaCl, 5M EDTA, 10% (v/v) glycerol, 1 mM

DTT). After 5 h at 37 8C, the reactions were terminated by

heating to 90 8C for 5 min. The oligonucleotide was then

treated with uracil DNA glycosylase (UDG) for 30 min at

37 8C in UDG buffer (20 mM Tris, pH 8.0, 1 mM DTT)

and cleaved by treatment with 0.15 M NaOH at 37 8C for

30 min. The products were separated on 15% TBE-urea

PAGE and detected by autoradiography. A labeled marker

oligonucleotide in which the target C was replaced with U

was processed in parallel to indicate the position of the

cleaved product.

Acknowledgments

We thank David Derse for HTLV-1 packaging plas-

mids and Heinrich Gottlinger for mutant HIV-1 plasmids.

This work was supported by NIH grants DA14494 and

AI58864, a grant from the American Foundation for

AIDS Research to FN and an Elizabeth Glaser Pediatric

AIDS Foundation grant to RK. NRL is an Elizabeth

Glaser Scientist of the Elizabeth Glaser Pediatric AIDS

Foundation.

F. Navarro et al. / Virology 333 (2005) 374–386 385

References

Accola, M.A., Strack, B., Gottlinger, H.G., 2000. Efficient particle

production by minimal Gag constructs which retain the carboxy-

terminal domain of human immunodeficiency virus type 1 capsid-p2

and a late assembly domain. J. Virol. 74 (12), 5395–5402.

Alce, T.M., Popik, W., 2004. APOBEC3G is incorporated into virus-like

particles by a direct interaction with HIV-1 Gag nucleocapsid protein.

J. Biol. Chem. 279 (33), 34083–34086.

Aldovini, A., Young, R.A., 1990. Mutations of RNA and protein sequences

involved in human immunodeficiency virus type 1 packaging result in

production of noninfectious virus. J. Virol. 64 (5), 1920–1926.

Aronoff, R., Linial, M., 1991. Specificity of retroviral RNA packaging.

J. Virol. 65 (1), 71–80.

Berkowitz, R., Fisher, J., Goff, S.P., 1996. RNA packaging. Curr. Top.

Microbiol. Immunol. 214, 177–218.

Betts, L., Xiang, S., Short, S.A., Wolfenden, R., Carter Jr., C.W., 1994.

Cytidine deaminase. The 2.3 A crystal structure of an enzyme:

transition-state analog complex. J. Mol. Biol. 235 (2), 635–656.

Bishop, K.N., Holmes, R.K., Sheehy, A.M., Davidson, N.O., Cho, S.J.,

Malim, M.H., 2004. Cytidine deamination of retroviral DNA by diverse

APOBEC proteins. Curr. Biol. 14 (15), 1392–1396.

Bogerd, H.P., Doehle, B.P., Wiegand, H.L., Cullen, B.R., 2004. A single

amino acid difference in the host APOBEC3G protein controls the

primate species specificity of HIV type 1 virion infectivity factor. Proc.

Natl. Acad. Sci. U.S.A. 101 (11), 3770–3774.

Cen, S., Guo, F., Niu, M., Saadatmand, J., Deflassieux, J., Kleiman, L.,

2004. The interaction between HIV-1 Gag and APOBEC3G. J. Biol.

Chem. 279 (32), 33177–33184.

Connor, R.I., Chen, B.K., Choe, S., Landau, N.R., 1995. Vpr is required for

efficient replication of human immunodeficiency virus type-1 in

mononuclear phagocytes. Virology 206, 936–944.

Derse, D., Hill, S.A., Lloyd, P.A., Chung, H., Morse, B.A., 2001.

Examining human T-lymphotropic virus type 1 infection and replication

by cell-free infection with recombinant virus vectors. J. Virol. 75 (18),

8461–8468.

Douaisi, M., Dussart, S., Courcoul, M., Bessou, G., Vigne, R., Decroly, E.,

2004. HIV-1 and MLV Gag proteins are sufficient to recruit

APOBEC3G into virus-like particles. Biochem. Biophys. Res. Com-

mun. 321 (3), 566–573.

Dull, T., Zufferey, R., Kelly, M., Mandel, R.J., Nguyen, M., Trono, D.,

Naldini, L., 1998. A third-generation lentivirus vector with a condi-

tional packaging system. J. Virol. 72 (11), 8463–8471.

Gorelick, R.J., Nigida Jr., S.M., Bess Jr., J.W., Arthur, L.O., Henderson,

L.E., Rein, A., 1990. Noninfectious human immunodeficiency virus

type 1 mutants deficient in genomic RNA. J. Virol. 64 (7),

3207–3211.

Harris, R.S., Petersen-Mahrt, S.K., Neuberger, M.S., 2002. RNA editing

enzyme APOBEC1 and some of its homologs can act as DNA mutators.

Mol. Cell 10 (5), 1247–1253.

Harris, R.S., Bishop, K.N., Sheehy, A.M., Craig, H.M., Petersen-Mahrt,

S.K., Watt, I.N., Neuberger, M.S., Malim, M.H., 2003. DNA

deamination mediates innate immunity to retroviral infection. Cell

113 (6), 803–809.

Jarmuz, A., Chester, A., Bayliss, J., Gisbourne, J., Dunham, I., Scott,

J., Navaratnam, N., 2002. An anthropoid-specific locus of orphan

C to U RNA-editing enzymes on chromosome 22. Genomics 79 (3),

285–296.

Lau, P.P., Zhu, H.J., Baldini, A., Charnsangavej, C., Chan, L., 1994.

Dimeric structure of a human apolipoprotein B mRNA editing protein

and cloning and chromosomal localization of its gene. Proc. Natl. Acad.

Sci. U.S.A. 91 (18), 8522–8526.

Lecossier, D., Bouchonnet, F., Clavel, F., Hance, A.J., 2003. Hyper-

mutation of HIV-1 DNA in the absence of the Vif protein. Science

300 (5622), 1112.

Li, J., Potash, M.J., Volsky, D.J., 2004. Functional domains of APOBEC3G

required for antiviral activity. J. Cell. Biochem. 92 (3), 560–5672.

Liddament, M.T., Brown, W.L., Schumacher, A.J., Harris, R.S., 2004.

APOBEC3F properties and hypermutation preferences indicate activity

against HIV-1 in vivo. Curr. Biol. 14 (15), 1385–1391.

Luo, K., Liu, B., Xiao, Z., Yu, Y., Yu, X., Gorelick, R., Yu, X.F., 2004.

Amino-terminal region of the human immunodeficiency virus type 1

nucleocapsid is required for human APOBEC3G packaging. J. Virol. 78

(21), 11841–11852.

Mangeat, B., Turelli, P., Caron, G., Friedli, M., Perrin, L., Trono, D., 2003.

Broad antiretroviral defence by human APOBEC3G through lethal

editing of nascent reverse transcripts. Nature 424 (6944), 99–103.

Mangeat, B., Turelli, P., Liao, S., Trono, D., 2004. A single amino acid

determinant governs the species-specific sensitivity of APOBEC3G to

Vif action. J. Biol. Chem 279 (15), 14481–14483.

Mariani, R., Chen, D., Schrofelbauer, B., Navarro, F., Konig, R., Bollman,

B., Munk, C., Nymark-McMahon, H., Landau, N.R., 2003. Species-

specific exclusion of APOBEC3G from HIV-1 virions by Vif. Cell 114

(1), 21–31.

Mehle, A., Strack, B., Ancuta, P., Zhang, C., McPike, M., Gabuzda, D.,

2004. Vif overcomes the innate antiviral activity of APOBEC3G by

promoting its degradation in the ubiquitin-proteasome pathway. J. Biol.

Chem. 279 (9), 7792–7798.

Muramatsu, M., Kinoshita, K., Fagarasan, S., Yamada, S., Shinkai, Y.,

Honjo, T., 2000. Class switch recombination and hypermutation require

activation-induced cytidine deaminase (AID), a potential RNA editing

enzyme. Cell 102 (5), 553–563.

Muto, T., Muramatsu, M., Taniwaki, M., Kinoshita, K., Honjo, T., 2000.

Isolation, tissue distribution, and chromosomal localization of the

human activation-induced cytidine deaminase (AID) gene. Genomics

68 (1), 85–88.

Navaratnam, N., Bhattacharya, S., Fujino, T., Patel, D., Jarmuz, A.L., Scott,

J., 1995. Evolutionary origins of apoB mRNA editing: catalysis by a

cytidine deaminase that has acquired a novel RNA-binding motif at its

active site. Cell 81 (2), 187–195.

Revy, P., Muto, T., Levy, Y., Geissmann, F., Plebani, A., Sanal, O., Catalan,

N., Forveille, M., Dufourcq-Labelouse, R., Gennery, A., Tezcan, I.,

Ersoy, F., Kayserili, H., Ugazio, A.G., Brousse, N., Muramatsu, M.,

Notarangelo, L.D., Kinoshita, K., Honjo, T., Fischer, A., Durandy, A.,

2000. Activation-induced cytidine deaminase (AID) deficiency causes

the autosomal recessive form of the Hyper-IgM syndrome (HIGM2.

Cell 102 (5), 565–575.

Schafer, A., Bogerd, H.P., Cullen, B.R., 2004. Specific packaging of

APOBEC3G into HIV-1 virions is mediated by the nucleocapsid

domain of the gag polyprotein precursor. Virology 328 (2), 163–168.

Schrofelbauer, B., Chen, D., Landau, N.R., 2004. A single amino acid of

APOBEC3G controls its species-specific interaction with virion infec-

tivity factor (Vif). Proc. Natl. Acad. Sci. U.S.A. 101 (11), 3927–3932.

Sheehy, A.M., Gaddis, N.C., Choi, J.D., Malim, M.H., 2002. Isolation of a

human gene that inhibits HIV-1 infection and is suppressed by the viral

Vif protein. Nature 418 (6898), 646–650.

Sheehy, A.M., Gaddis, N.C., Malim, M.H., 2003. The antiretroviral enzyme

APOBEC3G is degraded by the proteasome in response to HIV-1 Vif.

Nat. Med 9 (11), 1404–1407.

Shindo, K., Takaori-Kondo, A., Kobayashi, M., Abudu, A., Fukunaga, K.,

Uchiyama, T., 2003. The enzymatic activity of CEM15/Apobec-3G is

essential for the regulation of the infectivity of HIV-1 virion but not a

sole determinant of its antiviral activity. J. Biol. Chem. 278 (45),

44412–44416.

Soneoka, Y., Cannon, P.M., Ramsdale, E.E., Griffiths, J.C., Romano, G.,

Kingsman, S.M., Kingsman, A.J., 1995. A transient three-plasmid

expression system for the production of high titer retroviral vectors.

Nucleic Acids Res. 23 (4), 628–633.

Stopak, K., de Noronha, C., Yonemoto, W., Greene, W.C., 2003. HIV-1 Vif

blocks the antiviral activity of APOBEC3G by impairing both its

translation and intracellular stability. Mol. Cell 12 (3), 591–601.

Svarovskaia, E.S., Xu, H., Mbisa, J.L., Barr, R., Gorelick, R.J., Ono, A.,

Freed, E.O., Hu, W.S., Pathak, V.K., 2004. Human apolipoprotein B

mRNA-editing enzyme-catalytic polypeptide-like 3G (APOBEC3G) is

F. Navarro et al. / Virology 333 (2005) 374–386386

incorporated into HIV-1 virions through interactions with viral and

nonviral RNAs. J. Biol. Chem. 279 (34), 35822–35828.

Turelli, P., Mangeat, B., Jost, S., Vianin, S., Trono, D., 2004. Inhibition of

hepatitis B virus replication by APOBEC3G. Science 303 (5665), 1829.

Wedekind, J.E., Dance, G.S., Sowden, M.P., Smith, H.C., 2003. Messenger

RNA editing in mammals: new members of the APOBEC family

seeking roles in the family business. Trends Genet. 19 (4), 207–216.

Wiegand, H.L., Doehle, B.P., Bogerd, H.P., Cullen, B.R., 2004. A second

human antiretroviral factor, APOBEC3F, is suppressed by the HIV-1

and HIV-2 Vif proteins. EMBO J. 23 (12), 2451–2458.

Xu, H., Svarovskaia, E.S., Barr, R., Zhang, Y., Khan, M.A., Strebel, K.,

Pathak, V.K., 2004. A single amino acid substitution in human

APOBEC3G antiretroviral enzyme confers resistance to HIV-1 virion

infectivity factor-induced depletion. Proc. Natl. Acad. Sci. U.S.A.

101 (15), 5652–5657.

Yu, X., Yu, Y., Liu, B., Luo, K., Kong, W., Mao, P., Yu, X.F., 2003.

Induction of APOBEC3G ubiquitination and degradation by an HIV-1

Vif–Cul5–SCF complex. Science 302 (5647), 1056–1060.

Yu, Q., Chen, D., Konig, R., Mariani, R., Unutmaz, D., Landau, N.R.,

2004a. APOBEC3B and APOBEC3C are potent inhibitors of simian

immunodeficiency virus replication. J. Biol. Chem. 279 (51),

53379–53386.

Yu, Q., Konig, R., Pillai, S., Chiles, K., Kearney, M., Palmer, S., Richman,

D., Coffin, J.M., Landau, N.R., 2004b. Single-strand specificity of

APOBEC3G accounts for minus-strand deamination of the HIV

genome. Nat. Struct. Mol. Biol. 11 (5), 435–442.

Zennou, V., Perez-Caballero, D., Gottlinger, H., Bieniasz, P.D., 2004.

APOBEC3G incorporation into human immunodeficiency virus type 1

particles. J. Virol. 78 (21), 12058–12061.

Zhang, H., Yang, B., Pomerantz, R.J., Zhang, C., Arunachalam, S.C., Gao,

L., 2003. The cytidine deaminase CEM15 induces hypermutation in

newly synthesized HIV-1 DNA. Nature 424 (6944), 94–98.

Zheng, Y.H., Irwin, D., Kurosu, T., Tokunaga, K., Sata, T., Peterlin,

B.M., 2004. Human APOBEC3F is another host factor that blocks

human immunodeficiency virus type 1 replication. J. Virol. 78 (11),

6073–6076.