• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jvirolPermissionsJournals.ASM.orgJournalJV ArticleJournal InfoAuthorsReviewers
J Virol. Aug 2011; 85(16): 8197–8207.
PMCID: PMC3147987

Polymorphism in Human APOBEC3H Affects a Phenotype Dominant for Subcellular Localization and Antiviral Activity [down-pointing small open triangle]

Abstract

The APOBEC3 family of cytidine deaminases is part of the innate host defense targeted toward retroviruses and retroelements. APOBEC3H is the most distantly related member of the family and carries functional polymorphisms in current human populations. Haplotype II of APOBEC3H, which is more commonly found in individuals of African descent, encodes a protein with the highest antiviral activity in cells, whereas the other haplotypes encode proteins with weak or no antiviral activity. Here, we show that the different human APOBEC3H haplotypes exhibit differential subcellular localizations, as the haplotype I protein is mostly found in the nucleus and the haplotype II protein is mostly localized to the cytoplasm. The determinant responsible for this phenotype maps to a single amino acid that is also important for APOBEC3H protein stability. Furthermore, we show that the cytoplasmic localization is dominant over nuclear localization, by using fusion proteins of APOBEC3H. Our data support a model in which the APOBEC3H protein encoded by haplotype II is actively retained in the cytoplasm by interacting with specific host factors, whereas the less active protein encoded by haplotype I is allowed to enter the nucleus by a passive mechanism. Together, cytoplasmic localization and its link with protein stability correlate with the ability of APOBEC3H to inhibit HIV replication, providing a mechanistic basis for the differential antiviral activities of different APOBEC3H haplotypes.

INTRODUCTION

The APOBEC3 (apolipoprotein B mRNA-editing catalytic polypeptide) genes encode a family of cytidine deaminases that exhibit antiviral activity against a wide range of retroviruses and retroelements by both deaminase-dependent and -independent mechanisms (14, 13). Rodents carry a single APOBEC3 gene, while the same locus in primates has expanded to give rise to seven genes, encompassing three conserved zinc-coordinating motifs (Z1 to Z3): APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3DE, APOBEC3F, APOBEC3G, and APOBEC3H. Most of the APOBEC3 genes have evolved under adaptive (positive) selection over primate evolution, which suggests that they are engaged in “genetic conflict” with the pathogens or endogenous retroelements that they help control (18, 21). This genetic conflict is exemplified by the APOBEC3 interaction with the viral Vif protein. All lentiviruses except equine infectious anemia virus encode Vif, which specifically targets the APOBEC3 proteins for ubiquitination and subsequent degradation through the proteasome (14, 22, 29). As a result, HIV-1 is immune to the antiviral effects of APOBEC3 and can efficiently replicate in cells expressing APOBEC3.

APOBEC3H is the most distantly related member of the human APOBEC3 locus and encodes the only Z3-type cytidine deaminase of the locus (5, 11). APOBEC3H is also unique in its extent of functional human polymorphisms. Four major haplotypes (I to IV) have been reported to exist in human populations at frequencies greater than 10% (8, 17, 24, 26). The haplotypes differ greatly in their antiviral activity and in their expression levels. Haplotype II encodes the most active form of APOBEC3H, while haplotype I is much less active, and haplotypes III and IV are completely inactive. The antiviral activity of the APOBEC3H proteins is directly correlated with their turnover rate in the cell, with the most stable protein encoded by haplotype II being the most active against vif-deficient HIV-1 (17). Loss of APOBEC3H protein stability occurred twice in the human lineage as a result of two independent single-nucleotide polymorphisms (SNPs) at positions 15 (Del15N) and 105 (R105G).

Our lab has previously shown that an individual who is homozygous for haplotype II, which encodes the most stable protein, harbors HIV-1 variants encoding vif genes that can better neutralize APOBEC3H than those encoded by viruses from individuals who carry only one copy of haplotype II or none (12). Consistent with a role of APOBEC3H in HIV target cells, APOBEC3H mRNA expression is highly upregulated in stimulated CD4+ T cells by about 22-fold (20). Moreover, a recent study suggested that the most active haplotype of APOBEC3H is associated with reduced viral load in early, untreated HIV-infected individuals (7). Hence, the APOBEC3H genotype of HIV-infected individuals may have an impact on the evolution of their viruses, which suggests an important in vivo role for APOBEC3H polymorphisms in HIV infection.

In addition to APOBEC3H functional divergence in human populations, this gene has also had a dynamic evolutionary history in primates. All primates except humans and chimpanzees encode an APOBEC3H protein that is localized to the cytoplasm due to the presence of a putative nuclear export signal (NES) in the C terminus of the protein (17). After the common ancestor of humans and chimpanzees diverged from gorillas, a mutation was fixed in the APOBEC3H gene which led to the truncation of the C-terminal end of the sequence and the eventual loss of exclusive cytoplasmic retention. Interestingly, macaque APOBEC3H demonstrates an impaired ability to restrict HIV-1 when its C-terminal region encoding NES is removed, which suggests that cytoplasmic localization might be important for the antiviral activity of primate APOBEC3H proteins (17).

Previous studies have shown that the haplotype I protein has a cell-wide distribution, with predominantly nuclear staining in some cases (9, 11, 16, 17, 31). However, here we found that different variants of human APOBEC3H differ dramatically in their intracellular localization; while we confirmed that the protein encoded by APOBEC3H haplotype I is primarily localized in the nucleus, we found that the protein encoded by APOBEC3H haplotype II is primarily localized in the cytoplasm of transfected cells. Moreover, we mapped the determinant of APOBEC3H subcellular localization to the same amino acid that is responsible for its protein stability. Using a combination of experiments that target APOBEC3H to different cellular compartments and by making fusion proteins between different haplotypes, we found that the cytoplasmic localization of the protein encoded by haplotype II is dominant over nuclear localization of that encoded by haplotype I. Our data suggest that the haplotype II-encoded protein (the most active APOBEC3H haplotype in human populations) is actively retained in the cytoplasm by potentially interacting with specific host factors, while the less active protein encoded by haplotype I appears to passively diffuse into the nucleus. Therefore, the APOBEC3H proteins found in individuals from diverse world populations are localized to different cellular compartments by different mechanisms, and their interactions with host factors could play an essential role in their defense against viral pathogens.

MATERIALS AND METHODS

APOBEC3H cloning, expression constructs, and plasmids.

Cloning of macaque, chimpanzee, and untagged human APOBEC3H (haplotypes I and II) cDNAs into pcDNA3.1 (Invitrogen) was previously described (12, 17, 18). To create the myc-tagged pyruvate kinase fusion proteins with APOBEC3H, human APOBEC3H haplotypes I and II were cloned into the EcoRI/Xba sites of pcDNA1/Amp-myc-PK construct. Human APOBEC3H haplotypes I and II were also amplified with a 5′ or 3′ primer carrying a flexible linker sequence, GGT GGT GGT GGT GGC GCC (Gly-Gly-Gly-Gly-Gly-Ala), and the products were mixed and amplified in an overlapping PCR to create an APOBEC3H-5GlyAla-APOBEC3H fragment, which was then cloned into the EcoRI/XhoI sites of pcDNA3.1. The simian virus 40 (SV40) nuclear localization signal (NLS; PKKKRKV) was added to either the N or C terminus of human APOBEC3H. APOBEC3H mutants (hapI K121D, hapII R105G, hapII D178E, and hapII R105G D121K) were created by introducing point mutations into the human APOBEC3H sequence by site-directed mutagenesis using the QuikChange kit (Stratagene), and the entire insert was resequenced.

Cells, transfections, and Western blot analysis.

HEK293T cells were maintained in Dulbecco's modified Eagle's medium, 1% penicillin-streptomycin (Pen/Strep), 10% bovine growth serum at 37°C in a CO2 incubator. SupT1 cells were maintained similarly in RPMI with 1% Pen/Strep, 10% bovine growth serum. Transfections were performed with the TransIT-LT1 transfection reagent (Mirus Bio) at a reagent:plasmid DNA ratio of 3:1. For Western blot analysis, cells were lysed in ice-cold radioimmunoprecipitation assay buffer (150 mM sodium chloride, 1.0% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, and 50 mM Tris; pH 8.0) with protease inhibitors (Roche Complete Mini, EDTA-free tablets), resolved by either 10% or 12% SDS-PAGE, transferred to polyvinylidene difluoride, and probed with antibodies. A 1:1,000 dilution of human APOBEC3H antibody (P1H6-1), a 1:5,000 dilution of human APOBEC3G antibody, a 1:500 dilution of myc antibody, and a 1:2,000 dilution of actin antibody were used (12).

Immunofluorescence.

HeLa cells plated onto coverslips in a 6-well dish at 1 × 105 cells/ml were transfected the following day with 2 μg pcDNA3.1/APOBEC3H expression vectors. Twenty-four hours posttransfection, slips were fixed in 4% paraformaldehyde, permeabilized in chilled 1:1 100% methanol-acetone, and blocked in 10% fetal bovine serum–phosphate-buffered saline (PBS). Untagged APOBEC3H proteins were detected using monoclonal antibody against APOBEC3H (P1H6-1) (12) at a 1:50 dilution and a secondary anti-mouse antibody that was conjugated to Texas Red at a 1:100 dilution (Invitrogen). Nuclei were stained with 4′,6-diamidino-3-phenylindole (DAPI; diluted to 300 nM in PBS; Invitrogen) and mounted onto slides with SlowFade Gold antifade reagent (Invitrogen), and images were collected using an Olympus DeltaVision microscope.

For quantification of immunofluorescence staining, HeLa cells were plated in a 96-well imaging plate (Falcon BD BioCoat) at 8 × 104 cells/ml and transfected the next day with 0.125 μg pcDNA3.1/APOBEC3H expression vectors. Cells were prepared for analysis the same way in the 96 wells as on coverslips and were stained with Cellomics whole-cell stain Green (Thermo Scientific; resuspended in 50 μl of dimethyl sulfoxide, and diluted 1:1,000 in PBS) before images were collected on a Thermo Scientific ArrayScan VTI HCS reader. The colocalization protocol of the Cellomics cell analysis software was used to determine the nuclear and cytoplasmic boundaries of cells, based on DAPI and whole-cell staining, respectively, and to identify the cytoplasmic region by subtracting the nuclear region from the whole-cell region. The average intensity of APOBEC3H expression in each cell compartment was then calculated as the mean ratio of the average nuclear versus cytoplasmic staining of APOBEC3H by Texas Red. The mean ratio calculated for each cell was then plotted in histograms, and an unpaired 2-sample t test was carried out to compare the histogram distributions of the human APOBEC3H proteins.

Velocity sedimentation.

For analysis of APOBEC3-containing complexes, 293T cells were lysed in ice-cold NP-40 buffer with EDTA and protease inhibitors (0.1 M Tris [pH 7.4], 0.1 M NaCl, 10 mM EDTA, 0.625% NP-40, and 50 mM K-acetate), and 100 μl of cell lysate was layered onto step gradients containing 900 μl each of 10%, 15%, 20%, 30%, and 50% sucrose in NP-40 buffer with EDTA and subjected to velocity sedimentation in a Beckman SW55-Ti at 45,000 rpm (163,000 × g) for 37 min at 4°C as previously described (25). Fractions (400 μl) were serially collected from the top of the gradient, precipitated with cold acetone, and analyzed by immunoblotting with antibodies to APOBEC3G or APOBEC3H. Secondary anti-rabbit or anti-mouse IgG antibodies with IRDye conjugate were then used at a 1:15,000 dilution to visualize and quantify band intensities of APOBEC3 from the gradient fractions.

Viral infectivity assays.

Single-round HIV-1 infectivity assays were performed as previously described (18, 28). All assays were performed by transfection of 1.25 × 105 293T cells in 24-well plates with an approximately 1:1 ratio of pcDNA3/APOBEC plasmid (200 to 250 ng) to 250 ng of pLai3ΔenvLuc2 (28) or Δvif proviral plasmid (18). Virus equivalent to 2 nanograms of p24CA was used to infect 4 × 104 SupT1 cells in a 96-well plate in the presence of 20 μg/ml DEAE-dextran. After 48 h, cells from triplicate infections were lysed in 100 μl of Bright-Glo luciferase assay reagent (Promega) and results were read on a luminometer.

RESULTS

Differential subcellular localizations of the active human APOBEC3H proteins.

Previous analysis showed that the subcellular localization of APOBEC3H in primate evolution has changed between the chimpanzee/human lineage and all other primates examined. Despite the C-terminal truncation of human APOBEC3H, we hypothesized that the haplotype II protein may exhibit NES-independent cytoplasmic localization, which would allow it to better carry out its role as an antiviral effector. Since APOBEC3H localization might be functionally relevant for its antiviral activity, we asked whether variations found in the human APOBEC3H gene might affect the subcellular localization of the different APOBEC3H proteins. In order to examine the subcellular localization of these proteins, HeLa cells were transfected with APOBEC3H expression plasmids, and the subcellular localization of the protein was examined after staining with an APOBEC3H-specific monoclonal antibody. As previously reported, rhesus macaque APOBEC3H is mainly cytoplasmic while chimpanzee APOBEC3H is mostly nuclear (Fig. 1). Consistent with previous reports, we also confirmed that human APOBEC3H protein encoded by haplotype I is distributed in both the nucleus and cytoplasm, but mostly in the nucleus (Fig. 1). Surprisingly, we found that the APOBEC3H haplotype II-encoded protein is predominantly cytoplasmic, bearing strong similarity to the subcellular distribution of macaque APOBEC3H (Fig. 1). The marked differences in localization between the human APOBEC3H proteins encoded by haplotypes I and II are particularly striking, given that these alleles differ by only three amino acid changes.

Fig. 1.
Intracellular localization of the protein encoded by APOBEC3H haplotype I is different from that of the protein encoded by APOBEC3H haplotype II. Deltavision images of HeLa cells transfected with rhesus macaque, chimpanzee, and human APOBEC3H (haplotypes ...

In order to carry out a more quantitative and comprehensive analysis of the subcellular localization of APOBEC3H in populations of expressing cells, we developed an assay originally designed for high-throughput imaging and quantitative analysis of fixed cells transfected in a 96-well format and analyzed with an ArrayScan reader. In this assay, the nuclear and whole-cell regions are stained with DAPI (blue) and whole-cell stain (green), respectively, and the subcellular localization of APOBEC3H is represented as the ratio of average intensity staining of Texas Red in the nucleus versus that in the cytoplasm, which is the region identified by nuclear subtraction from the whole cell (Fig. 2A). A mean nuclear-to-cytoplasmic ratio of APOBEC3H average intensity staining (total APOBEC3H intensity divided by the area of the corresponding compartment) of greater than 1.0 indicates predominant nuclear localization of the APOBEC3H protein, whereas a mean ratio of less than 1.0 indicates predominant cytoplasmic localization of the protein. The mean ratio of nuclear to cytoplasmic staining for chimpanzee APOBEC3H is 1.253 (760 cells measured), which confirms its nuclear localization observed qualitatively with a Deltavision microscope (Fig. 2B). In contrast, rhesus macaque APOBEC3H has a mean ratio of 0.7495 (458 cells measured), which is indicative of a cytoplasmic protein (Fig. 2B). The mean ratio of nuclear to cytoplasmic staining for the protein encoded by APOBEC3H haplotype I is 1.290 (544 cells measured), which confirms that the protein encoded by APOBEC3H haplotype I has a pronounced nuclear localization (Fig. 2B). In contrast to haplotype I, the mean ratio of nuclear to cytoplasmic staining for the protein encoded by APOBEC3H haplotype II is 0.8168 (448 cells measured), which shows that the haplotype II protein is mostly localized to the cytoplasm, similar to the active macaque APOBEC3H protein (Fig. 2B). The difference in the average APOBEC3H staining in the nucleus versus cytoplasm between the cells transfected with haplotype I and those with haplotype II was statistically significant (P < 0.001), and the results were reproducible between independent experiments, with a standard deviation of about 5% (3.3 to 8.6%). Moreover, the differential localization of the APOBEC3H proteins was independent of the amount of plasmid used to transfect HeLa cells (Table 1; 30 to 300 ng of APOBEC3H). The haplotype I protein is mostly nuclear, with a mean ratio ranging from 1.239 to 1.391, while the haplotype II protein is cytoplasmic, with a mean ratio ranging from 0.8831 to 0.9498. Therefore, the protein encoded by the most active APOBEC3H haplotype currently found in the human population has a different subcellular localization than a less active haplotype, as demonstrated by both qualitative and quantitative imaging approaches.

Fig. 2.
Quantitative analysis of APOBEC3H localization confirmed that the haplotype I and II proteins are targeted to different subcellular compartments. (A) Cellomics cell analysis software was used to determine the nuclear and cytoplasmic boundaries of cells ...
Table 1.
Localization of human APOBEC3H proteins is not affected by plasmid concentration

We also compared the total average staining intensity of APOBEC3H in the whole cell with the mean ratio of average intensity staining in the nucleus over cytoplasm for the proteins encoded by haplotypes I and II. The haplotype I-encoded protein demonstrates a strong positive correlation between the concentration of the protein, represented by APOBEC3H staining intensity in the whole cell, and nuclear localization of the protein (Fig. 2C) (Pearson coefficient, 0.596; P < 0.001), suggesting that when more of the haplotype I protein is produced in the cell, then it is more likely it will be found in the nucleus. This trend suggests a concentration-dependent nuclear transport mechanism, potentially passive diffusion. In contrast, we did not observe a similar correlation for those two parameters for the haplotype II protein, which is mostly found in the cytoplasm regardless of its concentration in the cell (Fig. 2C) (Pearson coefficient, −0.075; P = 0.113). This phenotypic difference between the APOBEC3H proteins suggests that cytoplasmic retention of the haplotype II protein is mediated by an active mechanism and is not random. Hence, the most active APOBEC3H protein is localized to a different cellular compartment than the less active protein.

Differential subcellular localization of human APOBEC3H proteins is linked to the amino acid determinant for protein stability.

The proteins encoded by haplotypes I and II differ by only 3 amino acids at positions 105, 121, and 178 (Fig. 3A). The polymorphic changes at positions 105 and 121 have been previously reported to be the determinants for the stability of the protein and APOBEC3H sensitivity to Vif, respectively (8, 12, 17, 24, 26, 30). On the other hand, the polymorphism at position 178 has no known function. We hypothesized that either one or more of these changes are responsible for the differential subcellular localization of the active APOBEC3H proteins. Single and double mutations were introduced into the sequences of APOBEC3H haplotypes I and II at positions 105, 121, and 178 to determine if a single or a combination of changes is required to change the localization pattern of the proteins. The mutants were overexpressed in 293T cells to check for their stability by immunoblotting (Fig. 3B). The APOBEC3H proteins with R105 were stabilized, whereas the ones with G105 were unstable, as indicated by their lower expression levels. Remarkably, a single mutation at position 105 in the protein encoded by haplotype I, from glycine to arginine, was also sufficient to change APOBEC3H nuclear localization to a distinct cytoplasmic localization, as shown by the Cellomics analysis of the transfected cells (mean ratio of 0.9303 for average intensity staining of hapI G105R in the nucleus over cytoplasm, compared with 1.290 of hapI) (Fig. 3C). A second mutation, K121D, which is responsible for APOBEC3H sensitivity to Vif, in combination with G105R slightly changed the localization of the mutant further (mean ratio of 0.8634) (Fig. 3C). The K121D mutation alone had no significant impact on the ability of the haplotype I-encoded protein to enter the nucleus (mean ratio of 1.200, compared to 1.290 for hap lI) (Fig. 3C). Similarly, when haplotype I-specific changes are introduced into haplotype II, R105G alone is sufficient to confer nuclear localization on the protein encoded by haplotype II (mean ratio of 1.146 for average intensity staining of hapII R105G, compared with 0.8168 for hapII) (Fig. 3C). Double mutations R105G D121K changed the local- ization of the mutant APOBEC3H further (Fig. 3C; mean ratio of 1.404). However, the change at position 178 had no effect on the localization of the haplotype II-encoded protein (Fig. 3C; mean ratio of 0.8116, compared to 0.8168 for hap II). Therefore, the polymorphic change at position 105 that is responsible for protein stability is the major determinant for subcellular localization of APOBEC3H and is distinct from the amino acid responsible for its differential sensitivity to Vif at position 121.

Fig. 3.
The determinant for protein stability at position 105 is sufficient to change the subcellular localization of APOBEC3H. (A) Schematic of polymorphic changes in human APOBEC3H that differ between haplotypes I and II (amino acid numbers are listed above ...

Passive diffusion of the protein encoded by haplotype I into the nucleus.

In order to determine the mechanistic basis for the nuclear distribution of the haplotype I protein, we asked if haplotype I contains an active NLS by using an established assay. Pyruvate kinase (PK) is normally excluded from the nucleus due to its large size but, if haplotype I carries an NLS, the fusion protein between PK and APOBEC3H should be transported into the nucleus. We overexpressed the fusion proteins of PK and APOBEC3H haplotypes in 293T cells and examined their subcellular localizations in HeLa cells. The fusion protein encoded by myc-PK-hap I had lower expression than that encoded by myc-PK-hap II, which was similar to the difference in stability between the human APOBEC3H proteins (Fig. 4A). Results from the Cellomics analysis of transfected cells indicated that the myc-PK-hap I fusion protein was localized to the cytoplasm, with a mean ratio of 0.8074 for the average intensity staining of APOBEC3H in the nucleus over cytoplasm, which resembled the cytoplasmic localization of myc-PK alone (Fig. 4B; 757 cells analyzed). The fusion protein of myc-PK-hap II is also found in the cytoplasm, with a mean ratio of 0.8116 (Fig. 4B; 502 cells analyzed). Therefore, haplotype I does not appear to encode a nuclear localization signal. Rather, adding PK potentially makes the haplotype I protein too large to passively diffuse into the nucleus. Taken together with the observation that the haplotype I protein migrates into the nucleus in a concentration-dependent manner (Fig. 2C), the data support a passive diffusion mechanism for the nuclear distribution of the less active APOBEC3H protein due to its small size (21.7kDa).

Fig. 4.
The haplotype I protein does not encode a nuclear localization signal that can target pyruvate kinase to the nucleus. (A) Western blot analysis of myc-PK-APOBEC3H fusion proteins after transfection in 293T cells. Note that 12 times more lysate was loaded ...

Cytoplasmic retention of the haplotype II-encoded protein is dominant over nuclear transport of the haplotype I-encoded protein.

To characterize the mechanism by which the haplotype II-encoded protein is retained in the cytoplasm, we created double APOBEC3H fusion constructs, in which two copies of haplotype I or haplotype II or two different haplotypes are linked together by a flexible linker. HeLa cells were transfected with these constructs, and subcellular localization of these fusion proteins was measured. The double haplotype I-encoded protein had a similar phenotype as the single haplotype I-encoded protein and was found mostly in the nucleus, with a mean average APOBEC3H intensity ratio of 1.162 (Fig. 5A; 587 cells analyzed). These data suggest that doubling the size of APOBEC3H (to 43.7kDa) does not exclude the haplotype I-encoded protein from the nucleus, likely because only proteins much greater than 40 kDa are prevented from freely entering the nucleus (27). Similarly, double haplotype II-encoded protein resembled its single counterpart and was localized to the cytoplasm, with a mean average APOBEC3H intensity ratio of 0.7617 (Fig. 5A; 543 cells analyzed). On the other hand, the chimeric APOBEC3H proteins, haplotype I-haplotype II and haplotype II-haplotype I, were both found in the cytoplasm, with mean average APOBEC3H intensity ratios of 0.8763 (488 cells analyzed) and 0.8146 (556 cells analyzed), respectively (Fig. 5A). These chimeric APOBEC3H proteins are confined to the cytoplasm regardless of the presence of the haplotype I sequence, which demonstrates that the cytoplasmic retention of the haplotype II-encoded protein is dominant over passive diffusion of the haplotype I-encoded protein into the nucleus. These data suggest that the haplotype II-encoded protein is exported to or retained in the cytoplasm by an active mechanism. To further confirm this phenotype in a more relevant cell type for HIV infection, we also examined the localization of the haplotype II protein in a T-cell line and found that the most active APOBEC3H protein was mostly localized to the cytoplasm in SupT1 cells (data not shown).

Fig. 5.
Cytoplasmic localization and stability of the haplotype II protein exert dominant effects on nuclear localization and rapid turnover of the less-active haplotype I protein. (A) Cellomics analysis of HeLa cells transfected with double APOBEC3H proteins ...

Next, we asked whether the instability of the haplotype I protein is dominant over the stability of the haplotype II protein or vice versa. One might expect that if the unstable haplotype I protein contains a signal for proteasome-mediated degradation, then the fusion haplotype I-linker-II protein would also be unstable. However, our lab previously showed that treatment of cells expressing the unstable human APOBEC3H proteins (haplotype I) with a proteasome inhibitor, MG132, does not completely rescue the stability of the protein to the same level as the stably expressed macaque APOBEC3H (18). In fact, what we found here is that when the haplotype I-encoded protein was fused to the haplotype II-encoded protein, the chimeric proteins retained the expression level of the stably expressed haplotype II protein (Fig. 5B; compare hap I-linker-II and hap II-linker-I to hap II-linker-II). Thus, stability, rather than rapid degradation, of the APOBEC3H-encoded proteins is dominant.

Consistent with increased steady-state protein levels of the fusion constructs carrying the haplotype II sequence, the proteins encoded by haplotype II-linker-II, I-linker-II, and II-linker-I are all highly active against vif-deficient HIV-1, as infectivity is reduced from 100% to 0.11 to 0.33% in the presence of APOBEC3H (Fig. 5B). As a result, the stability and activity of the APOBEC3H protein encoded by haplotype II can override the instability and inactivity of the haplotype I-encoded protein. Thus, both the cytoplasmic localization and stability of haplotype II are linked and have dominant effects over nuclear localization and instability of haplotype I.

Active retention of the most active human APOBEC3H protein in the cytoplasm.

To further explore the possibility that an active mechanism is involved in the cytoplasmic retention of the haplotype II protein, we added the NLS from SV40 T antigen to haplotypes I and II to redirect the localization of APOBEC3H proteins. If our hypothesis is true, the presence of an NLS would be insufficient to completely target the haplotype II protein to the nucleus. We overexpressed NLS-APOBEC3H fusion proteins in cells and examined their expression and subcellular localization. Similar to the different expression levels of the human APOBEC3H proteins, the NLS-haplotype I protein was also more unstable than the NLS-haplotype II protein regardless of the position of the NLS (Fig. 6A). We also tested the activity of NLS-APOBEC3H fusion proteins against vif-deficient HIV-1 and found that the addition of NLS caused both the haplotype I and II proteins to lose some antiviral activity but not completely (Fig.6B). Compared with the haplotype I protein alone (mean ratio, 1.060), both the NLS-hap I and hap I-NLS proteins were localized to the nucleus to a greater extent, with mean ratios of 1.903 and 1.754, respectively, and their images also exhibited an exclusively nuclear staining of APOBEC3H in the cell (Fig.6C; 445 and 320 cells analyzed). Interestingly, unlike the haplotype I proteins with NLS, NLS-hap II and hap II-NLS were not completely targeted to the nucleus, with mean ratios of 1.134 and 1.221, respectively (Fig.6C; 260 and 406 cells analyzed), and their images exhibited a more diffused staining of APOBEC3H in both the cytoplasm and nucleus. Taken together, these data reinforce the idea that the haplotype I-encoded protein enters the nucleus by passive diffusion and therefore can be easily converted to exclusive nuclear localization with the help of an NLS. On the other hand, there is a strong determinant for export or cytoplasmic retention in the haplotype II sequence that presents a competing force against nuclear import mediated by the SV40 NLS. As a result, a larger fraction of the haplotype II-encoded protein remains in the cytoplasm, even in the presence of an NLS, compared to the haplotype I protein.

Fig. 6.
Cytoplasmic retention of human APOBEC3H encoded by haplotype II is likely due to its interactions with specific host factors. (A) Western blot analysis of APOBEC3H fusion proteins with NLS after transfection in 293T cells. (B) Inhibition of HIV vif by ...

Next, we asked whether there are gross differences between the types of ribonucleoprotein complexes where human APOBEC3H proteins are found in the cell. Human APOBEC3G associates with high-molecular-mass (HMM) complexes in resting CD4+ T cells and T cell lines but dissociates into a low-molecular-mass (LMM) form upon activation or RNase A treatment (25). It is possible that the haplotype II protein is retained in the cytoplasm by interacting with a molecular mass complex that is much larger than that associated with the haplotype I protein. We resolved APOBEC3-containing complexes by using a velocity sedimentation gradient and found that the majority of both APOBEC3H proteins associated with HMM complexes, similar to APOBEC3G (Fig.7A, fractions 2 to 7). This is consistent with a previous study which showed the haplotype I protein associates in an HMM form that is similar in size to the APOBEC3G complex (24). RNase A treatment of the lysates converted APOBEC3H from an HMM form to an LMM form, which suggests that both haplotype-encoded proteins form complexes that are sensitive to RNase A (Fig.7A, fractions 1 to 3). Similarly, quantification of the amount of APOBEC3 protein found in each fraction resulted in line graphs with overlapping peaks for APOBEC3G and APOBEC3H haplotypes with and without RNase treatment (Fig.7B). Hence, the haplotype II protein does not differentially associate with a larger molecular mass complex that excludes it from entering the nucleus. Taken together with the observation that addition of an NLS to the haplotype II protein does not redirect it to exclusive nuclear localization (Fig.6B), we propose a model in which the haplotype II protein is actively retained in or exported to the cytoplasm by interacting with specific host factors.

Fig. 7.
Both the haplotype I and II proteins are associated with similar-sized molecular mass complexes. (A) Cell lysates from APOBEC3G- or -3H-expressing 293T cells with or without RNase A treatment were subjected to velocity sedimentation, fractionated (lanes ...

DISCUSSION

Here, we have presented evidence for differential subcellular localization of proteins encoded by two active APOBEC3H alleles currently circulating in human populations. The haplotype II protein, which potently restricts HIV-1, is found mostly in the cytoplasm, similar to other APOBEC3 proteins with anti-HIV activity, whereas the less active version of APOBEC3H encoded by haplotype I is predominantly localized to the nucleus. We further demonstrated that the amino acid responsible for APOBEC3H localization in the cell is the same residue that determines its protein stability. In addition, cytoplasmic retention of the haplotype II protein is strong enough to partially overcome nuclear import by the SV40 NLS and to confer cytoplasmic localization to the haplotype I protein. Together, we propose a model for differential localization of the active human APOBEC3H proteins, in which the more stable protein encoded by haplotype II stays in the cytoplasm by interactions with other cellular factors, whereas the less stable haplotype I protein enters the nucleus by passive diffusion.

Previously, another group reported that the intracellular localization of the protein encoded by haplotype I was not different from that of the protein encoded by haplotype II (19), although examination of the single image that was published seems to show more nuclear staining for the APOBEC3H haplotype I protein. Use of a high-throughput and quantitative imaging analysis method that surveyed more than hundreds of cells at once and presented their localization in a normal distribution allowed us to characterize the localization profiles of the APOBEC3H proteins more accurately. This novel assay clearly demonstrated that the majority of the haplotype II protein is exclusively localized to the cytoplasm, whereas the haplotype I protein has a cell-wide distribution.

Additionally, we showed that stability and cytoplasmic localization of APOBEC3H are linked to the same polymorphic residue at position 105, but it is unclear whether those two properties are functionally linked. Cytoplasmic APOBEC3H proteins are always more stable and active (e.g., the haplotype II protein, the haplotype I G105R mutant, the double haplotype II, and the chimeric APOBEC3H proteins), while the less stable and inactive proteins are usually nuclear (e.g., the haplotype I protein, the haplotype II R105G mutant, and the double APOBEC3H protein encoded by haplotype I-linker-I). One hypothesis would be that rapid turnover of the unstable protein encoded by haplotype I specifically takes place in the nucleus, and therefore cytoplasmic retention determines protein stability of APOBEC3H. We investigated protein levels of the myc-PK-haplotype I fusion and the haplotype II protein with SV40 NLS, which are artificially targeted to the cytoplasm and nucleus, respectively, and made two observations that argue against this hypothesis. The fusion protein of myc-PK-hap I was not stabilized to the same level as myc-PK-hap II when it was excluded from the nucleus (Fig.4A). Moreover, both NLS-hap II and hap II-NLS were expressed to similar levels as the haplotype II protein alone, which suggests that the source of degradation is not specifically located in the nuclear region (Fig.6A). Taken together, we found that the stability of APOBEC3H is determined by a mechanism distinct from its subcellular localization.

The localization of APOBEC3 proteins could have an impact on its activity against HIV-1, a virus that replicates in the cytoplasm, and sensitivity to the cytoplasmic Vif protein. However, we observed that the NLS-hap II and hap II-NLS proteins are still highly active against vif-deficient HIV-1, suggesting that the haplotype II protein is still active in the nucleus (Fig.6B). One caveat with this experiment is that the haplotype II protein is only partially localized to the nucleus in the presence of NLS. In addition, the mutation at position 121 alone, which affects Vif sensitivity, does not impact subcellular localization of APOBEC3H, although it has a minor effect in the presence of the mutation at position 105 (Fig.3C). Hence, the location of APOBEC3H proteins in the cell does not determine their major activity against HIV-1 and sensitivity to HIV-1 Vif.

In this study, we also explored possible mechanisms that could account for APOBEC3H differential localization. We showed that the less stable haplotype I protein does not encode an NLS (Fig.4B) but instead enters the nucleus in a concentration-dependent manner (Fig.2C), which can be overcome by the cytoplasmic localization of the haplotype II protein (Fig.5A). These data support a model in which the haplotype I protein migrates to the nucleus by passive diffusion. On the other hand, because the SV40 NLS can only partially target the haplotype II protein to the nucleus (Fig.6C), a cytoplasmic retention signal or an export signal might actively retain the most stable APOBEC3H protein in the cytoplasmic compartment. Moreover, when we fused haplotype I to haplotype II, the artificial dimer showed cytoplasmic localization, which is also consistent with active retention of haplotype II in the cytoplasm (Fig.5A). However, since both human APOBEC3H proteins associate with HMM complexes and readily dissociate into an LMM form upon RNase A treatment (Fig.4), it is unlikely that the haplotype II protein is retained in the cytoplasm by differentially interacting with a larger complex. Therefore, we hypothesize that the most active human APOBEC3H protein is actively kept in the cytoplasm by interacting with specific host factors via its polymorphic site 105R. These host proteins are either export factors or cellular factors with a signal for export or cytoplasmic retention. Although there is no evidence for a classical leucine-rich NES (10, 27) in the sequence of human APOBEC3H (data not shown), we cannot rule out the possibility that the haplotype II protein is exported into the cytoplasm by a mechanism distinct from the CRM1-dependent nuclear export pathway, such as the exportin 7 pathway (15). Similar to human APOBEC3H encoded by haplotype II, APOBEC3G carries cytoplasmic retention determinants in its amino-terminal half and is targeted to the cytoplasm by a mechanism other than CRM1-dependent nuclear export (23).

Further experiments will need to be carried out to identify and characterize interacting partners of APOBEC3H in the cytoplasm that are specific to the haplotype II protein. A previous study identified several RNA-binding proteins and small polymerase III RNAs in APOBEC3-associated ribonucleoprotein complexes that were unique interacting partners of APOBEC3G but not APOBEC3F, which could potentially play a role in their functional differences (6). Also, there might be polymorphisms in the host factors that interact with the APOBEC3H proteins that could potentially modulate their subcellular localization. One way to address this is to examine the localization of endogenous APOBEC3H proteins in the cells of individuals that possess haplotype I or II. Furthermore, it is possible that APOBEC3H interaction with host factors not only confers cytoplasmic retention of the haplotype II protein but also protects it from getting rapidly degraded. This hypothesis is based on the evidence that the double APOBEC3H proteins encoded by haplotype I-linker-II and haplotype II-linker-I are stabilized compared to the protein encoded by two copies of haplotype I, likely because the chimeric proteins can now interact with cellular factors via the determinant 105R of the haplotype II protein (Fig.5A). Interestingly, the degradation machinery to which the unstable APOBEC3H protein is sensitive is not present in bacterial cells, since both the haplotype I and II proteins are stably expressed in Escherichia coli (12, 18).

The evolutionary history of APOBEC3H subcellular localization is dynamic in the hominid lineages. Although the APOBEC3H protein of chimpanzees carries an arginine at position 105, it is a nuclear deaminase that is stably expressed in the cell, unlike the stable cytoplasmic human protein encoded by haplotype II. Chimpanzees might have acquired a compensatory mutation in their APOBEC3H gene that confers nuclear localization since divergence from the most recent common ancestor with humans. On the other hand, the APOBEC3H protein of the human ancestor is localized to the cytoplasm most likely is due to the presence of 105R (data not shown). This property of the human APOBEC3H protein to be localized to the cytoplasm is further maintained in individuals of African descent who carry haplotype II at a high frequency (>50%). However, certain groups in the human populations, such as Caucasians and Asians, acquired a polymorphism of R105G in their APOBEC3H gene after the split from their common human ancestor, which allows the unstably expressed protein to move into the nucleus. Interestingly, a recent study has described one additional active APOBEC3H haplotype (V) that exists in human populations at a frequency of 20% (26). In contrast to haplotype II, haplotype V is not only found in the African American population but also in the Caribbean and Chinese populations, which were believed to carry mainly the unstable haplotypes previously. Haplotypes V encodes an arginine at position 105 and therefore is likely localized to the cytoplasm, suggesting that an active cytoplasmic APOBEC3H protein might be more widely distributed in populations than we previously thought.

It is likely that the differential subcellular localization of the APOBEC3H proteins contributes to their functional differences. For example, the cytoplasmic protein encoded by haplotype II is more efficiently packaged into HIV-1 in the absence of a functional Vif. Moreover, the localization of the most active cytidine deaminase into the cytoplasm rather than the nucleus may have important implications for prevention of deleterious mutations to the host genome. In addition, it is possible that particular viral pathogens that replicate in the nucleus might have driven the loss of APOBEC3H cytoplasmic retention in the chimpanzee lineage specifically. Finally, the study of APOBEC3 variations in the human population allows us to better understand how the host interacts with viruses on a population level and to make predictions on how previous selective pressures might have shaped the functions of these antiviral genes.

ACKNOWLEDGMENTS

We thank Julio Vazquez and Dave McDonald from the FHCRC Scientific Imaging core for help with using the Thermo Scientific ArrayScan VTI HCS reader and the DeltaVision microscope and with analyzing Cellomics data; the FHCRC Genetic Analysis core; and Jaisri Lingappa for advice on velocity sedimentation. We thank Masahiro Yamashita, Molly OhAinle, Oliver Fregoso, Nisha Duggal, Efrem Lim, Alex Compton, and Patrick Mitchell for comments on the manuscript.

This work was supported by NIH grant R37 AI30937.

Footnotes

[down-pointing small open triangle]Published ahead of print on 8 June 2011.

REFERENCES

1. Albin J. S., Harris R. S. 2010. Interactions of host APOBEC3 restriction factors with HIV-1 in vivo: implications for therapeutics. Expert. Rev. Mol. Med. 12:e4. [PMC free article] [PubMed]
2. Bishop K. N., et al. 2004. Cytidine deamination of retroviral DNA by diverse APOBEC proteins. Curr. Biol. 14:1392–1396. [PubMed]
3. Bishop K. N., Verma M., Kim E. Y., Wolinsky S. M., Malim M. H. 2008. APOBEC3G inhibits elongation of HIV-1 reverse transcripts. PLoS Pathog. 4:e1000231. [PMC free article] [PubMed]
4. Bogerd H. P., et al. 2006. Cellular inhibitors of long interspersed element 1 and Alu retrotransposition. Proc. Natl. Acad. Sci. U. S. A. 103:8780–8785. [PMC free article] [PubMed]
5. Conticello S. G., Thomas C. J., Petersen-Mahrt S. K., Neuberger M. S. 2005. Evolution of the AID/APOBEC family of polynucleotide (deoxy)cytidine deaminases. Mol. Biol. Evol. 22:367–377. [PubMed]
6. Gallois-Montbrun S., et al. 2008. Comparison of cellular ribonucleoprotein complexes associated with the APOBEC3F and APOBEC3G antiviral proteins. J. Virol. 82:5636–5642. [PMC free article] [PubMed]
7. Gourraud P. A., et al. 2011. APOBEC3H haplotypes and HIV-1 pro-viral vif DNA sequence diversity in early untreated human immunodeficiency virus-1 infection. Hum. Immunol. 72:207–212. [PMC free article] [PubMed]
8. Harari A., Ooms M., Mulder L. C., Simon V. 2009. Polymorphisms and splice variants influence the antiretroviral activity of human APOBEC3H. J. Virol. 83:295–303. [PMC free article] [PubMed]
9. Kinomoto M., et al. 2007. All APOBEC3 family proteins differentially inhibit LINE-1 retrotransposition. Nucleic Acids Res. 35:2955–2964. [PMC free article] [PubMed]
10. Kutay U., Güttinger S. 2005. Leucine-rich nuclear-export signals: born to be weak. Trends Cell Biol. 15:121–124. [PubMed]
11. LaRue R. S., Jónsson S. R., et al. 2008. The artiodactyl APOBEC3 innate immune repertoire shows evidence for a multi-functional domain organization that existed in the ancestor of placental mammals. BMC Mol. Biol. 9:104. [PMC free article] [PubMed]
12. Li M., Wu L., Emerman M. 2010. The range of human APOBEC3H sensitivity to lentiviral Vif proteins. J. Virol. 84:88–95. [PMC free article] [PubMed]
13. Malim M. H., Emerman M. 2008. HIV-1 accessory proteins: ensuring viral survival in a hostile environment. Cell Host Microbe 3:388–398. [PubMed]
14. Mehle A., et al. 2004. Vif overcomes the innate antiviral activity of APOBEC3G by promoting its degradation in the ubiquitin-proteasome pathway. J. Biol. Chem. 279:7792–7798. [PubMed]
15. Mingot J. M., Bohnsack M. T., Jäkle U., Görlich D. 2004. Exportin 7 defines a novel general nuclear export pathway. EMBO J. 23:3227–3236. [PMC free article] [PubMed]
16. Muckenfuss H., et al. 2006. APOBEC3 proteins inhibit human LINE-1 retrotransposition. J. Biol. Chem. 281:22161–22172. [PubMed]
17. OhAinle M., Kerns J. A., Li M. M., Malik H. S., Emerman M. 2008. Antiretroelement activity of APOBEC3H was lost twice in recent human evolution. Cell Host Microbe 4:249–259. [PMC free article] [PubMed]
18. OhAinle M., Kerns J. A., Malik H. S., Emerman M. 2006. Adaptive evolution and antiviral activity of the conserved mammalian cytidine deaminase APOBEC3H. J. Virol. 80:3853–3862. [PMC free article] [PubMed]
19. Ooms M., Majdak S., Seibert C., Harari A., Simon V. 2010. The localization of APOBEC3H variants in HIV-1 virions determines their antiviral activity. J. Virol. 84:7961–7969. [PMC free article] [PubMed]
20. Refsland E. W., et al. 2010. Quantitative profiling of the full APOBEC3 mRNA repertoire in lymphocytes and tissues: implications for HIV-1 restriction. Nucleic Acids Res. 38:4274–4284. [PMC free article] [PubMed]
21. Sawyer S. L., Emerman M., Malik H. S. 2004. Ancient adaptive evolution of the primate antiviral DNA-editing enzyme APOBEC3G. PLoS Biol. 2:E275. [PMC free article] [PubMed]
22. 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. 9:1404–1407. [PubMed]
23. Stenglein M. D., Matsuo H., Harris R. S. 2008. Two regions within the amino-terminal half of APOBEC3G cooperate to determine cytoplasmic localization. J. Virol. 82:9591–9599. [PMC free article] [PubMed]
24. Tan L., Sarkis P. T., Wang T., Tian C., Yu X. F. 2009. Sole copy of Z2-type human cytidine deaminase APOBEC3H has inhibitory activity against retrotransposons and HIV-1. FASEB J. 23:279–287. [PMC free article] [PubMed]
25. Thielen B. K., et al. 2007. T cells contain an RNase-insensitive inhibitor of APOBEC3G deaminase activity. PLoS Pathog. 3:1320–1334. [PMC free article] [PubMed]
26. Wang X., et al. 2011. Analysis of human APOBEC3H haplotypes and anti-human immunodeficiency virus type 1 activity. J. Virol. 85:3142–3152. [PMC free article] [PubMed]
27. Wente S. R., Rout M. P. 2010. The nuclear pore complex and nuclear transport. Cold Spring Harb. Perspect. Biol. 2:a00562. [PMC free article] [PubMed]
28. Yamashita M., Emerman M. 2004. Capsid is a dominant determinant of retrovirus infectivity in nondividing cells. J. Virol. 78:5670–5678. [PMC free article] [PubMed]
29. Yu X., et al. 2003. Induction of APOBEC3G ubiquitination and degradation by an HIV-1 Vif-Cul5-SCF complex. Science 302:1056–1060. [PubMed]
30. Zhen A., Wang T., Zhao K., Xiong Y., Yu X. 2010. A single amino acid difference in human APOBEC3H variants determines HIV-1 Vif sensitivity. J. Virol. 84:1902–1911. [PMC free article] [PubMed]
31. Zielonka J., et al. 2009. Restriction of equine infectious anemia virus by equine APOBEC3 cytidine deaminases. J. Virol. 83:7547–7559. [PMC free article] [PubMed]

Articles from Journal of Virology are provided here courtesy of American Society for Microbiology (ASM)

Formats:

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...

Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...