An Alu Element Insertion in Intron 1 Results in Aberrant Alternative Splicing of APOBEC3G Pre-mRNA in Northern Pig-Tailed Macaques (Macaca leonina) That May Reduce APOBEC3G-Mediated Hypermutation Pressure on HIV-1

APOBEC3 family members, particularly APOBEC3F and APOBEC3G, are important cellular antiviral factors. Recently, more attention has been paid to targeting APOBEC3G for AIDS therapy. To appropriately utilize macaque animal models for the study of APOBEC3-related issues, it is important that the differences between human and macaque APOBEC3s are clarified. In this study, we identified a novel and conserved APOBEC3G pre-mRNA alternative splicing pattern in macaques, which differed from that in humans and which may reduce the APOBEC3G-mediated hypermutation pressure on HIV-1 in northern pig-tailed macaques (NPMs). Our work provides important information for the proper application of macaque animal models for APOBEC3-related issues in AIDS research and a better understanding of the biological functions of APOBEC3 proteins.

ABSTRACT

APOBEC3 family members, particularly APOBEC3F and APOBEC3G, inhibit the replication and spread of various retroviruses by inducing hypermutation in newly synthesized viral DNA. Viral hypermutation by APOBEC3 is associated with viral evolution, viral transmission, and disease progression. In recent years, increasing attention has been paid to targeting APOBEC3G for AIDS therapy. Thus, a controllable model system using species such as macaques, which provide a relatively ideal in vivo system, is needed for the study of APOBEC3-related issues. To appropriately utilize this animal model for biomedical research, important differences between human and macaque APOBEC3s must be considered. In this study, we found that the ratio of APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints was much lower in peripheral blood mononuclear cells (PBMCs) from northern pig-tailed macaques than in PBMCs from humans. Next, we identified a novel and conserved APOBEC3G pre-mRNA alternative splicing pattern in macaques, which differed from that in humans and resulted from an Alu element insertion into macaque APOBEC3G gene intron 1. This alternative splicing pattern generating an aberrant APOBEC3G mRNA isoform may significantly dilute full-length APOBEC3G and reduce APOBEC3G-mediated hypermutation pressure on HIV-1 in northern pig-tailed macaques, which was supported by the elimination of other possibilities accounting for this hypermutation difference between the two hosts.

IMPORTANCE APOBEC3 family members, particularly APOBEC3F and APOBEC3G, are important cellular antiviral factors. Recently, more attention has been paid to targeting APOBEC3G for AIDS therapy. To appropriately utilize macaque animal models for the study of APOBEC3-related issues, it is important that the differences between human and macaque APOBEC3s are clarified. In this study, we identified a novel and conserved APOBEC3G pre-mRNA alternative splicing pattern in macaques, which differed from that in humans and which may reduce the APOBEC3G-mediated hypermutation pressure on HIV-1 in northern pig-tailed macaques (NPMs). Our work provides important information for the proper application of macaque animal models for APOBEC3-related issues in AIDS research and a better understanding of the biological functions of APOBEC3 proteins.

INTRODUCTION

Hosts evolve many strategies against viral infection, including the application of host restriction factors. A key group of host restriction factors is the APOBEC3 family, which shows activity against the replication and spread of various viruses such as retroviruses, hepadnaviruses, parvoviruses, and herpesviruses (1–5). During the past 30 million years, primates have been infected by many viruses. The rapid evolution of the APOBEC3 genes has kept pace with the rapid evolution of viruses (6), indicating that the APOBEC3 gene family plays an important role in restricting virus infection and maintaining genomic stability. In primates, the APOBEC3 family is comprised of seven cytidine deaminases (APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3D, APOBEC3F, APOBEC3G [A3G], and APOBEC3H) (7–10). Among them, APOBEC3G and APOBEC3F are the most intensively researched and exhibit the strongest activity against human immunodeficiency viruses (HIVs) in vitro compared with other members (11, 12) and in vivo in humanized mouse models (13). APOBEC3G and APOBEC3F can be incorporated into nascent retroviral particles, where they induce hypermutation by deoxycytidine deamination of minus-strand cDNA during viral reverse transcription when the retroviral particle subsequently infects the next cell. A C-to-U mutation in the minus strand can lead to nascent minus-strand degradation or ultimately G-to-A hypermutation in the HIV-1 genome (14–18). The copackaging of APOBEC3 proteins such as APOBEC3G and APOBEC3F or APOBEC3G and APOBEC3H can comutate the same genomes and can cooperate to inhibit HIV-1 replication (19). Additionally, recently, a biochemical assay revealed that APOBEC3G and APOBEC3F form a heterodimer that increases the efficiency of G-to-A hypermutation in the HIV-1 genome (20).

The hypermutation dinucleotide contexts of APOBEC3G and other members of the APOBEC3 family exhibit obvious differences: APOBEC3G preferentially induces GG-to-AG hypermutation, while GA-to-AA hypermutation is preferentially mediated by other APOBEC3 proteins (21), especially APOBEC3F (22, 23). The GG-to-AG and GA-to-AA mutations often produce the stop codons TAG and TAA, leading to the premature termination of translation (24). Viral hypermutation by APOBEC3 is associated with viral diversification, viral evolution (25, 26), viral transmission, and disease progression (27).

To neutralize the antiviral activity of APOBEC3s, especially APOBEC3G, most lentiviruses encode an accessory protein, Vif, which functions by binding APOBEC3 deaminases and leading them to the E3 ubiquitin ligase complex, resulting in proteasomal degradation (28, 29). The species specificity of the interaction between Vif and APOBEC3 proteins is pronounced, although not absolute, and confirms that APOBEC3s play an important role in limiting cross-species retroviruses (30). Another important host restriction factor that suppresses or blocks retroviral cross-species transmission is TRIM5α (31). Because TRIM5-cyclophilin A (TRIM5-CypA) fusion at the TRIM5 locus in pig-tailed macaques (PTMs) results in the loss of TRIM5α in these species, PTMs are the only Old World monkeys susceptible to HIV-1 infection (32, 33). Recently, a new animal model of AIDS was successfully produced using PTMs infected with adapted macaque-tropic HIV-1 (34), which is a great step forward for AIDS research. Our team also found that HIV-1_NL4-3 successfully infected northern pig-tailed macaques (NPMs) (Macaca leonina) and formed a long-term viral reservoir in vivo (35, 36). To better optimize and appropriately utilize this research animal model, the NPM major histocompatibility complex (npmMHC) (37, 38), basic serum immunoglobulin data (39), and simian immunodeficiency virus (SIV) infection (40, 41) were previously researched.

In recent years, consideration has been given to targeting APOBEC3G for AIDS therapy (21, 42), and APOBEC3s may be a future focus of antiretroviral therapeutic strategies (43, 44). However, whether it is beneficial to directly or indirectly enhance APOBEC3-mediated mutational pressure in vivo for patients remains unclear. APOBEC3 proteins not only are capable of suppressing retrovirus replication but also may facilitate viral diversity, viral adaptation to hosts (13), and virus drug resistance in vivo (26). Furthermore, the upregulation of APOBEC3 expression may impact cellular regulation mechanisms and cause hypermutation in the human genome, bringing the risk of cancer development (45, 46). Thus, a controllable in vivo model system is needed to address certain APOBEC3-related issues. Macaques provide a relatively ideal in vivo system for the study of anti-HIV-1 drug efficacy in vivo and HIV-1 host adaptation and pathogenesis (47). In addition, type I interferon (IFN-I) signaling was significantly upregulated with viral replication in macaque models, which included restriction factors (35).

In our research, we found that the ratio of APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints was much lower in NPM peripheral blood mononuclear cells (PBMCs) than in human PBMCs. Thus, we analyzed APOBEC3 mRNA and function in NPMs and found a novel alternative splicing pattern of npmAPOBEC3G pre-mRNA generating a unique and dominant APOBEC3G mRNA isoform with a 180-bp insertion at the junction of exons 1 and 2, which was highly conserved in macaques and not found in humans. Interestingly, this 180-bp fragment contained two premature translational termination codons, which could possibly lead to severely truncated translation products made up entirely of 19 or 37 amino acids. This conjecture was supported by our Western blot experiments, which did not detect the APOBEC3G protein expressed by the aberrant APOBEC3G (abA3G) mRNA isoform. We next explored the emerging mechanism of the novel splicing pattern and found that the insertion of an Alu element into macaque APOBEC3G gene intron 1, which was not found in humans, may explain the different splicing patterns of APOBEC3G pre-mRNA between the two hosts. The aberrant splicing pattern of APOBEC3G pre-mRNA, the expression level of which was higher than that of full-length APOBEC3G in different tissues, may significantly dilute full-length APOBEC3G and reduce APOBEC3G-mediated mutational pressure in macaque PBMCs.

RESULTS

In vivo HIV-1 DNA from HIV-1-infected NPMs exhibits fewer hypermutation footprints by APOBEC3G than by other APOBEC3s.

We infected two NPMs with NL4-3 via the intravenous route and harvested their PBMCs at 21 days postinfection. Using nested PCR, we then cloned a 1.2-kb segment of the HIV-1 env gene from the total cellular DNA of these PBMCs. The overwhelming majority of mutations in the DNA clones were G-to-A mutations, which were obviously the effect of APOBEC3 deaminase activity on HIV-1 (Fig. 1). Further analysis showed that the ratio of the APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints in vivo was very low, suggesting that APOBEC3G-mediated hypermutation pressure was lower than that of other APOBEC3s in NPMs. The viral hypermutation profile found in HIV-1-infected NPMs was consistent with findings in simian foamy virus (SFV)- and xenotropic murine leukemia virus (XMRV)-infected Chinese rhesus macaques (ChRMs) (22, 48). In contrast, most findings from HIV-1- and SFV-infected humans, in which the hypermutation footprints and GG-to-AG mutations by APOBEC3G were obvious (43, 48–51), indicate the strong action of APOBEC3G in humans.

FIG 1.

In vivo hypermutation profiles of HIV-1 DNA by APOBEC3s from NPM PBMCs infected with NL4-3 at 21 days postinfection. The numbers of mutation contexts and proportions are shown in the table. All mutations are marked with a vertical line, GG-to-GA mutations are labeled in red, GA-to-AA mutations are in cyan, GC-to-AC mutations are in green, GT-to-AT mutations are in magenta, and all other mutations are in black (57). nt, nucleotides.

The in vitro ratio of APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints is much lower in macaque PBMCs than in human PBMCs.

For further study, we infected cultured human and NPM PBMCs with HIV-1 and then tested if the APOBEC3G-mediated hypermutation pressure on HIV-1 was obviously lower in NPM PBMCs than in human PBMCs.

The results showed that most of the hypermutation contexts from cultured PBMCs of human donors 1, 2, and 3 were APOBEC3G favored (66.7%, 70.9%, and 84.0%, respectively). In contrast, most of the G-to-A substitutions from cultured PBMCs of macaques 1, 2, and 3 occurred in APOBEC3G-disfavored contexts, and APOBEC3G-favored contexts accounted for only 8.9%, 15.4%, and 15.9%, respectively. These results were in accordance with the hypermutation profile in in vivo experiments. These results suggested that the impacts on HIV-1 by APOBEC3G in macaque and human PBMCs were significantly different (Fig. 2).

FIG 2.

Hypermutation profiles of HIV-1 DNA in cultured NPM and human PBMCs. The numbers of mutation contexts and proportions are shown in the tables. Mutation color-coding was applied to indicate the different mutation contexts (Fig. 1).

Elimination of other possibilities accounting for the hypermutation difference found between the two hosts.

We found that the ratio of the APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints was low in vivo and in vitro. We attempted to demonstrate the anti-HIV-1 characteristics of npmAPOBEC3G and human APOBEC3G (hAPOBEC3G).

If the npmAPOBEC3F gene expression level is considerably higher than the npmAPOBEC3G gene expression level in NPM PBMCs, there may be fewer APOBEC3G-mediated hypermutation footprints in the HIV-1 genome than APOBEC3F-mediated hypermutation footprints. To determine whether the APOBEC3G/APOBEC3F mRNA ratio of human PBMCs was higher than that of NPM PBMCs, the APOBEC3G/APOBEC3F mRNA ratios at 0 days (before infection), 1 day, 3 days, and 6 days post-HIV-1 infection were determined by quantitative real-time PCR. At 0 days (before infection), 1 day, 3 days, and 6 days postinfection, there were no significant differences in the APOBEC3G/APOBEC3F mRNA ratios between the human and NPM groups (Fig. 3A), indicating that the APOBEC3G/APOBEC3F mRNA ratio in PBMCs did not account for the hypermutation difference between the two hosts.

FIG 3.

Exclusion of other possibilities accounting for the hypermutation difference between the two hosts. (A) APOBEC3G/APOBEC3F mRNA ratios before and during infection. Error bars indicate standard deviations of the means of the data from three subjects. (B) Infectivity was titrated by second-round viral infectivity assays. HIV-1 was produced in the presence or absence of APOBEC3G proteins. Infectivity in the absence of APOBEC3G was normalized to 100%, and the presence of hAPOBEC3G served as a positive control. Error bars represent standard deviations of the means from three replicate infections. Statistical differences between mean percentages were evaluated using two-tailed Student’s t test, with a P value of <0.05 (▲) being considered significant. Luc, luciferase. (C) Hypermutation pattern induced by npmAPOBEC3G on HIV-1. There were 28 independent colonies from each cell culture. The total number of hypermutation sequences from the npmAPOBEC3G group was 12, while there were no hypermutation sequences from the control group. The number of mutation contexts and proportions are shown in the tables. Mutation color-coding indicates the different mutation contexts (Fig. 1). (D) A doxycycline-on Vif expression system was used to detect whether npmAPOBEC3G is sensitive to degradation by HIV-1 Vif. The npmAPOBEC3G/F and human APOBEC3G plasmids were transfected into Trex-hvif-15 cells, the medium was then changed, and doxycycline (Dox) was added at 100 ng/ml 6 h later. Thirty-six hours later, the Vif and APOBEC3 proteins were detected by Western blotting. (E) Hypermutation profiles of Δvif HIV-1 DNA in cultured NPM and human PBMCs. The numbers of mutation contexts and proportions are shown in the tables. Mutation color-coding was applied to indicate the different mutation contexts (Fig. 1).

To test the anti-HIV-1 characteristics of full-length npmAPOBEC3G, a full-length npmAPOBEC3G expression vector was constructed and cotransfected with pNL4-3 into 293T cells. The cotransfection supernatant was then used to measure the anti-HIV-1 activity of npmAPOBEC3G in a second-cycle infectivity assay, with an empty parental vector as a control. We observed that full-length npmAPOBEC3G effectively inhibited HIV-1 infection (Fig. 3B). To test the editing effect of full-length npmAPOBEC3G on HIV-1, TZM-bl cells in 12-well plates were infected with the cotransfection supernatant. At 48 h postinfection, TZM-bl cellular DNA was extracted and used as the template for the amplification of a 1.2-kb fragment of the HIV-1 env gene. The PCR fragments were cloned and sequenced. As expected, in the cells transfected with the full-length npmAPOBEC3G expression vector, the majority of mutations in the DNA clones were G-to-A mutations (82.2%), with most (70.5%) of the G-to-A mutations by npmAPOBEC3G being in the GG context. In the control group, however, almost no mutations were found (Fig. 3C). The above-described data indicated that the function of full-length npmAPOBEC3G was normal and that it was active against HIV-1 and able to induce HIV-1 G-to-A hypermutation mainly in the GG context.

If npmAPOBEC3G is more sensitive than hAPOBEC3G to degradation by HIV-1 Vif, fewer hypermutation footprints by npmAPOBEC3G should occur in NPM PBMCs. To test this possibility, we first used a doxycycline-on Vif expression system to detect whether npmAPOBEC3G is sensitive to degradation by HIV-1 Vif. The results in Fig. 3D show that npmAPOBEC3G is not so sensitive to degradation by HIV-1 Vif. Second, mitogen-activated PBMCs were infected with Δvif HIV-1. Cellular DNA was purified from the infected PBMCs at 6 days postinfection, and a 1.2-kb fragment of the HIV-1 env gene from the purified DNA was amplified by nested PCR, cloned into the pMD19-T simple vector, and finally sequenced. In the hypermutation sequences of the HIV-1 provirus from the cultured PBMCs of humans 1 and 2, most G-to-A mutations were GG to AG (76.2%), that is, APOBEC3G-favored mutational contexts. In the cultured PBMCs of NPMs 1 and 2, APOBEC3G-disfavored mutational contexts were more frequent than APOBEC3G-favored mutational contexts (Fig. 3E). It is worth noting that the different susceptibilities of APOBEC3G proteins from different species to degradation by HIV-1 Vif cannot explain the hypermutation difference between the two hosts.

Identification of a novel alternative splicing pattern of macaque APOBEC3G pre-mRNA.

In the experiment to obtain the full-length coding DNA sequence (CDS) of npmAPOBEC3G, we found a novel alternative splicing pattern of npmAPOBEC3G pre-mRNA generating a unique and dominant APOBEC3G mRNA isoform that included a 180-bp insertion at the junction of exons 1 and 2, which was not found in humans. This alternative splicing could not be changed by HIV-1 infection (Fig. 4A and B). Through the identification of the translation initiation site and correct reading frame of the novel npmAPOBEC3G mRNA isoform based on published full-length CDSs of APOBEC3G in other primates, we found that the 180-bp insertion segment was located 18 bp downstream of the translation initiation site and included two stop codons, which could possibly lead to severely truncated translation products made up entirely of 19 or 37 amino acids (Fig. 4C). The GT-AG splicing rule is shown in Fig. 4B. Thus, this novel and dominant APOBEC3G mRNA isoform was named an aberrant APOBEC3G mRNA isoform. To test our hypothesis regarding the possibility of severely truncated translation products, normal npmAPOBEC3G protein and aberrant APOBEC3G protein isoform expression vectors were constructed, and 293T cells were then transfected with the constructed expression vectors. Western blotting of 293T cells transfected with the aberrant APOBEC3G protein isoform expression vector did not detect the expression of the aberrant APOBEC3G protein (Fig. 4D), which indicated that this novel APOBEC3G mRNA isoform or its severely truncated translation products may be very unstable in cells. Additionally, the domination of this aberrant APOBEC3G mRNA isoform was also found in ChRMs and cynomolgus macaques (CMs) (Fig. 4E), and the aberrant APOBEC3G mRNA expression level was high in PBMC subsets, including CD4⁺ T cells, CD8⁺ T cells, and NK cells (Fig. 4F), and prevalent in other tissues (Fig. 4G), which suggested that the identified novel alternative splicing pattern of APOBEC3G pre-mRNA was conserved in macaques.

FIG 4.

Identification of a novel alternative splicing pattern of macaque APOBEC3G pre-mRNA. (A) Primate APOBEC3G gene map. Exons (not to scale in distribution) are represented as shaded gray blocks, and the 5′ and 3′ untranslated regions are represented as white blocks. The 180-bp insertion is represented as a black box. (B) Agarose gel electrophoresis of PCR amplification products shows a markedly different electrophoretic pattern between npmAPOBEC3G and hAPOBEC3G, and this pattern was not changed by HIV-1 infection. The GT-AG splicing rule in npmA3G intron 1 is shown in the sequence, and the PCR primers are marked in the corresponding exon pattern diagram. (C) A 180-bp sequence insertion into the regions between exons 1 and 2, generating different electrophoretic patterns between npmAPOBEC3G and hAPOBEC3G. The 180-bp insertion sequence is underlined. (D) 293T cells were transfected with pcDNA3.1-npmAPOBEC3G-Flag, pcDNA3.1-npm-aberrant-APOBEC3G-Flag, or negative-control pcDNA3.1-empty vector. Western blotting of 293T cells was performed using specific anti-Flag antibodies. Lane 1, pcDNA3.1-empty vector control; lane 2, pcDNA3.1-npmAPOBEC3G-Flag; lane 3, pcDNA3.1-npm-aberrant-APOBEC3G-Flag. GAPDH was used as a protein loading control. (E) Results from PCR amplification using APOBEC3G cDNA of humans, NPMs, ChRMs, and CMs (each species contained three individuals, marked 1, 2, and 3) show that the special splicing pattern is conserved in macaques. The npmAPOBEC3G and aberrant npmAPOBEC3G mRNA expression levels were determined in PBMC subsets, including CD4⁺ T cells, CD8⁺ T cells, and NK cells (F), and in different tissues (G).

An Alu element insertion into macaque APOBEC3G intron 1 results in the novel alternative splicing pattern.

We next explored the emerging mechanism of the novel splicing pattern in macaques. Comparing human, ChRM, and CM APOBEC3G gene sequences obtained from GenBank, we determined if the 180-bp fragment was joined and found that both alternative splicings agreed with the GT-AG rules; thus, the 180-bp insertion was named potential exon B. In addition, there was a marked difference between human and macaque intron 1. An Alu element insertion between potential exon B and exon 2 was detected in RM and CM intron 1 but was not found in intron 1 of humans (Fig. 5A). The same Alu element insertion pattern was also detected in npmAPOBEC3G gene intron 1 in the PCR experiment (Fig. 5B). Additionally, the Alu element insertion pattern was also found in RMs and CMs in the PCR experiment (Fig. 5C). The Alu element is about 300 nucleotides long and consists of two nonidentical monomers (52). Based on the dynamic exon definition hypothesis (53) and the polymerase II (Pol II) elongation model (54), we inferred that in the case of no Alu element insertion, RNA Pol II could be quickly processed and presented to the downstream strong 3′ splice site, thereby competing effectively with the upstream weak 3′ splice site and leading to the effective exclusion of potential exon B. However, the insertion of the 300-bp Alu element into the sequence between the 180-bp potential exon B and exon 2 would likely delay the presentation of the downstream strong 3′ splice site. This delay would provide greater opportunity and time for the upstream weak 3′ splice site splicing reaction to occur without competition from the downstream strong 3′ splice site, thus leading to the substantial inclusion of potential exon B.

FIG 5.

An Alu element inserted into macaque APOBEC3G gene intron 1. (A) Diagram of human and macaque APOBEC3G gene intron 1. The Alu element is represented by a black box. Two conserved 5′ donor sites (GT) and two conserved 3′ acceptor sites (AG) recognized by the spliceosome are shown. The PCR primers are marked in the corresponding exon pattern diagram. (B) Agarose gel electrophoresis of PCR amplification products and sequence alignment show the Alu element inserted into npmAPOBEC3G gene intron 1 but not hAPOBEC3G gene intron 1. (C) Additionally, the Alu element insertion pattern was also found in RMs and CMs.

To verify this hypothesis, we constructed the pcDNA3.1-NPMAPOBEC3G-intron1-Alu+ and pcDNA3.1-NPMAPOBEC3G-intron1-Alu− plasmids by a multiple-segment seamless cloning method. The expression of mRNAs was detected after transfection. As we hypothesized, in the case of no Alu element insertion, the aberrant APOBEC3G mRNA disappeared (Fig. 6).

FIG 6.

An Alu element insertion into macaque APOBEC3G intron 1 results in the novel alternative splicing pattern. (Top) Diagram of the minigenome of the macaque APOBEC3G gene with intron 1. The Alu element is represented by a red box. (Bottom) Agarose gel electrophoresis of PCR amplification products shows that in the case of no Alu element insertion, the aberrant APOBEC3G mRNA disappeared (left lanes). An Alu element inserted into npmAPOBEC3G gene intron 1 results in the novel alternative splicing pattern (middle lanes). However, in 293T cells without any plasmids transfected, there are no bands (right lanes).

In conclusion, our hypothesis that the novel alternative splicing pattern may reduce APOBEC3G-mediated hypermutation pressure on HIV-1 in NPMs was further supported by the elimination of other possible influencing factors, as described above.

DISCUSSION

From the ratio of APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints in vivo and in vitro, we found that the APOBEC3G hypermutation footprint was much smaller in macaque PBMCs than in human PBMCs. A similar hypermutation difference was also found in SFV-infected human and macaque hosts (48). By analyzing the SIV and HIV genomes, we found that there are no fewer GG motifs in the SIV genome than in the HIV genome, and the function of full-length APOBEC3G is normal. Thus, the fewer APOBEC3G hypermutations in PTMs are not due to the smaller number of GG motifs in the SIV genome. We then identified an aberrant APOBEC3G pre-mRNA alternative splicing pattern, with a high mRNA expression level, conserved in macaques and different from that in humans. This, together with our elimination of other possible influencing factors, supported the hypothesis that the aberrant APOBEC3G mRNA isoform may significantly reduce the full-length APOBEC3G-mediated hypermutation pressure on HIV-1 in NPMs.

Alternative splicing, an important mechanism of posttranscriptional regulation of gene expression, influences mRNA stability and diversity and generates diversified protein isoforms from a very limited number of individual genes (55, 56). Not all transcripts generated by alternative splicing will be biologically functional, but aberrant splicing is an inevitable event due to the complex and continuous splicing process and can be influenced by various physiological conditions. Most aberrant splicing transcripts are eliminated by mRNA surveillance mechanisms. However, some beneficial aberrant splicing events can be selectively fixed during evolution (57), which is the basis for the assumption that conserved splicing indicates function. However, how to distinguish between splicing errors and biofunctional splicing events is still an unresolved conundrum. It is unclear whether the above-described splicing pattern of APOBEC3G RNA in the past or present was or is beneficial to macaques, despite the splicing pattern being highly conserved.

Although splice variants of APOBEC3G in humans were not found, deletions of APOBEC3B and splice variants of APOBEC3H, other important members of the A3 family, are common in populations. The APOBEC3B deletion frequencies are significantly different in diverse ethnic populations, and interestingly, the deletion frequency increases with the out-of-Africa expansion route of humanity (58). A recent study suggested that the functional loss of APOBEC3B in humans reduced the pressure on Alu retrotransposition and thus increased human adaptability and survival in facing environmental challenges during migration from Africa (59). Human APOBEC3H represents the most evolutionarily divergent A3 gene. It has seven haplotypes (I to VII) and four splice variants (SV154/182/183/200) with differing antiviral activities and geographic distributions (60). Among them, APOBEC3H haplotype II strongly reduced the infectivity of SIVcpz because it is sensitive to only specific HIV-1 Vifs. It plays an important role in limiting SIVcpz across-species transmission to humans (61). Recent research demonstrated that the polymorphism of human A3H has relevance for HIV-1 infection and AIDS progression (62). If the new pattern of APOBEC3G in macaques was an evolutionary strategy in this species, this regulated pattern was evolutionarily sensible for cost minimization. Translation of the transcript may be prematurely terminated due to the production of only 19 or 37 amino acids (the prematurely terminated protein), which would cost much less than the downregulation pattern of the degradation of the whole protein following the product of the whole protein. Unfortunately, we are currently unable to detect the level of the whole APOBEC3G protein in macaque PBMCs due to the unavailability of an effective anti-APOBEC3G antibody.

We also explored the possible factors driving the different APOBEC3G pre-mRNA splicing patterns observed in macaques and humans. We found an Alu element insertion between potential exon B and APOBEC3G gene exon 2 in macaques, which was not found in intron 1 of humans. In primate genomes, Alu elements are abundant, repetitive, and considered the most successful retrotransposable elements. They play an important role in the regulation of gene expression, such as alternative splicing, RNA editing, and protein translation (63). Based on the dynamic exon definition hypothesis (53) and the Pol II elongation model (54), we inferred that the Alu element insertion delayed the presentation of the downstream strong 3′ splice site by Pol II and ultimately led to the emergence of the novel alternative splicing pattern in macaques, and our results corresponded to this hypothesis (Fig. 6).

In conclusion, we identified a novel and conserved APOBEC3G pre-mRNA alternative splicing pattern in macaques, which was different from that in humans and results from the insertion of an Alu element into macaque APOBEC3G gene intron 1. Our results showed that the ratio of the APOBEC3G-mediated/APOBEC3-mediated HIV-1 hypermutation footprints was much lower in NPM PBMCs than in human PBMCs, suggesting that this alternative splicing pattern generating an aberrant APOBEC3G mRNA isoform may significantly dilute full-length APOBEC3G and reduce APOBEC3G-mediated hypermutation pressure on HIV-1 in NPMs. This was further supported by the exclusion of other possibilities accounting for the hypermutation difference between the two hosts. Future studies on APOBEC3 differences, especially APOBEC3G/F, between humans and macaques will be necessary to properly use macaque animal models for APOBEC3-related issues in AIDS research and to better understand the biological functions of APOBEC3 proteins.

MATERIALS AND METHODS

Animals, cells, and viruses.

The NPMs, Chinese rhesus macaques (ChRMs), and cynomolgus macaques (CMs) in this study were from the Kunming Institute of Zoology, Chinese Academy of Sciences (KIZ, CAS). All animals were maintained in accordance with the regulations and recommendations of the Animal Care Committee of KIZ, CAS. During the study, the animals were housed at the animal biosafety level 3 laboratory, and all procedures were approved by the Institutional Committee for Animal Care in the KIZ, CAS (approval number SWKX-2013021), as described previously (39). Blood from all animals was collected via venipuncture. The PBMCs in blood samples from all macaques and HIV-1-negative humans were purified by Ficoll density centrifugation. The purified PBMCs from HIV-1-negative subjects were grown in RPMI 1640 medium (containing 10% fetal bovine serum [FBS]) with interleukin-2 (IL-2) (50 U/ml) and concanavalin A (ConA) (1 mg/ml; used for macaque PBMCs) or phytohemagglutinin (PHA) (5 μg/ml; used for human PBMCs) for 3 days. 293T cells (Type Culture Collection, Chinese Academy of Sciences [TCC CAS]), TZM-bl cells (MRC AIDS Reagent Project), and Trex-hvif-15 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% FBS, while C8166 cells (MRC AIDS Reagent Project) were grown in RPMI 1640 medium containing 10% FBS. Wild-type and Δvif HIV-1_NL4-3 were produced from 293T cells by transfection with wild-type and Δvif pNL4-3. pNL4-3 is a proviral DNA plasmid constructed from 5′ HIV-1-NY5 and 3′ HIV-1-LAV-1 sequences (64) using Lipofectamine 2000 (Lipo2000) according to the manufacturer’s instructions (Invitrogen). The wild-type and Δvif HIV-1_NL4-3 proviral plasmids were gifts from Guang-Xia Gao (Institute of Biophysics, CAS) and Yong-Hui Zheng (Department of Microbiology and Molecular Genetics, Michigan State University), respectively.

Hypermutation of viral DNA in infected PBMCs in vitro.

NPM and human PBMCs activated for 3 days, as described above, were infected with 1,840 50% tissue culture infective doses (TCID₅₀) of wild-type HIV-1 or Δvif HIV-1. Total cellular DNA from PBMCs was isolated using a DNeasy DNA isolation kit (Qiagen) at 6 days postinfection. The amplification of a 1.2-kb fragment of the HIV-1 env gene from the cellular DNA was applied for nested PCR using the following primers: 5′-CCA CTC TAT TTT GTG CAT CAG-3′ (first round, forward), 5′-CTT GGT GGG TGC TAC TCC TA-3′ (first round, reverse), 5′-CAT ATG ATA CAG AGG TAC ATA ATG TTT GGGC-3′ (second round, forward), and 5′-CAC TTC TCC AAT TGT CCC TCAT-3′ (second round, reverse) (65). The PCR conditions used for both rounds of PCR included 95°C for 15 min for a hot start, followed by 95°C for 30 s, 56°C for 30 s, and 72°C for 1 min for 30 cycles and a final extension step at 72°C for 10 min. A 2-μl sample of the first-round PCR products was used as the template for the second round of PCR. Second-round PCR products were purified by using a DNA gel extraction kit (Generay Biotech, Shanghai, China), cloned into the pMD19-T simple vector (TaKaRa, Dalian, China), and finally sequenced (Majorbio, Shanghai, China). Hypermutation analysis was performed using the HYPERMUT program (https://www.hiv.lanl.gov/content/sequence/HYPERMUT/hypermut.html).

HIV-1 inoculation of NPMs and in vivo hypermutation of viral DNA in NPM PBMCs.

HIV-1_NL4-3 produced from 293T cells by transfection (see “Animals, cells, and viruses,” above) was amplified by a brief passage through C8166 cells. PBMCs (1 × 10⁷ cells) from two healthy NPMs were infected in vitro with HIV-1_NL4-3 at a multiplicity of infection of 0.01, cultivated for 3 days, and then centrifuged at 1,500 rpm for 10 min. The harvested autologous PBMCs were suspended in 1 ml of a cell-free virus stock containing 2 ng HIV-1 p24 antigens and then inoculated via the intravenous route into the corresponding northern pig-tailed macaque in an animal biosafety level 3 (ABSL-3) laboratory. PBMCs were collected from venipuncture blood samples obtained at 3 weeks postinfection. Experiments to detect the hypermutation profiles of HIV-1 DNA in the PBMCs from infected macaques were performed (see “Hypermutation of viral DNA in infected PBMCs in vitro,” above).

Amplification of APOBEC3G mRNA.

Total RNA from human, NPM, RM, and CM PBMCs was isolated with TRIzol reagent (Invitrogen) and then reverse transcribed into cDNA using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China). The primers (forward primer 5′-TAG CCG GCA AAG GAT GAA TCC TCA AAT CAG-3′ and reverse primer 5′-GAA GTA GTA GAG GCG GGC AAC AAAG-3′) near the 180-bp insertion, named potential exon B, were designed based on the relatively conserved parts of APOBEC3G cDNA in humans, NPMs, RMs, and CMs. The PCR conditions were 95°C for 15 min for a hot start, followed by 30 cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 90 s and a final extension step for 10 min at 72°C. The PCR products were analyzed in a 2% agarose gel with ethidium bromide, purified using a DNA gel extraction kit (Generay Biotech, Shanghai, China), cloned into the pMD19-T simple vector (TaKaRa, Dalian, China), and finally sequenced (Majorbio, Shanghai, China).

Amplification of APOBEC3G gene intron 1.

Total cellular DNA from human and NPM PBMCs was isolated using a DNeasy DNA isolation kit (Qiagen). The primers (first-round forward primer 5′-CCG GTC CCC AAT GGC TCC CTC CTG CA-3′, first-round reverse primer 5′-TCC ACT AAA CAC AGC CCC CTG GAT GGC TCTA-3′, second-round forward primer 5′-TGT GGC CTG GGC AGG TTA CCC CGC-3′, and second-round reverse primer 5′-CCT GGG CAA CAA AAG CGA AAC TCC GTC TCA-3′) were designed based on the relatively conserved parts of human and NPM APOBEC3G gene intron 1. The PCR conditions used for both rounds of PCR were 94°C for 15 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 64°C for 30 s, and extension at 72°C for 40 s. The final extension step was done at 72°C for 10 min. The second-round PCR products were analyzed in a 2% agarose gel with ethidium bromide, purified using a DNA gel extraction kit (Generay Biotech, Shanghai, China), cloned into the pMD19-T simple vector (TaKaRa, Dalian, China), and finally sequenced (Majorbio, Shanghai, China).

Flow cell sorting.

In the flow cell sorting experiment, PBMCs were incubated with fluorochrome-conjugated CD3, CD4, and CD8 monoclonal antibodies (mAbs) for 30 min on ice. The samples were sorted into CD4⁺ T cells (CD4⁺ CD8⁻ CD3⁺), CD8⁺ T cells (CD4⁻ CD8⁺ CD3⁺), and NK cells (CD3⁻ CD8⁺) with a Becton, Dickinson Influx cell sorter. Anti-CD3-fluorescein isothiocyanate (FITC) (clone SP34-2) was obtained from BD Biosciences, and anti-CD4-peridinin chlorophyll protein (PerCP)-Cy5.5 (clone OKT4) and anti-CD8-phycoerythrin (PE)-Cy7 (clone RPA-T8) were obtained from BioLegend.

Quantitative real-time PCR.

Total RNA from human and NPM PBMCs was isolated and then reverse transcribed into cDNA using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China). To quantify human and macaque APOBEC3G and APOBEC3F mRNA expression levels, quantitative real-time PCR was performed in triplicate, and results were normalized to the values for endogenous glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA and ribosomal protein L13A (RPL13A) mRNA, respectively (66), using a SYBR Premix ExTaq II (Tli RNase H plus) kit. The primers and probes used in this experiment are shown in Table 1 (67).

TABLE 1.

Primers used for quantitative real-time PCR

Gene	Forward primer (5′–3′)	Reverse primer (5′–3′)
hAPOBEC3G	CACGTGAGCCTGTGCATCTTC	AAAGGTGTCCCAGCAGTGCTTA
hAPOBEC3F	GTCCTGAAACCTGGAGCCT	AGACGGTATTCCGACGAGA
npmAPOBEC3G	TACCACCCAGAGATGAGATT	GTTTCCAGAAGTAGTAGAGG
npmAPOBEC3F	CTTTAATAACAGACCCATCCTT	GTTGCCACAGAACCGAGA
Human GAPDH	GAAATCCCATCACCATCTTCCAGG	GAGCCCCAGCCTTCTCCATG
Macaque RPL13A	AAGGTGTTTGACGGCATCCC	CTTCTCCTCCAAGGTGGCTGT

To quantify APOBEC3G and abAPOBEC3G mRNA expression levels, quantitative real-time PCR was performed as described above. The primers are npm-qA3G-ab-F (5′-TCA AAT CAG TCC AGG TCC-3′), npm-qA3G-ab-R (5′-GGT TCA ATT CTC CCA GTT C-3′), npm-qA3G-F (5′-CTC AAA TCA GAA ACA TGG TGG-3′), npm-qA3G-R (5′-ACC AGC GGA GGA ATC TCA-3′), npm-RPL13A-F (5′-AAG GTG TTT GAC GGC ATC CC-3′), and npm-RPL13A-R (5′-CTT CTC CTC CAA GGT GGC TGT-3′).

Molecular cloning of NPM and human APOBEC3G.

Total RNA from healthy NPM and human PBMCs was isolated with TRIzol (Invitrogen) according to the manufacturer’s protocols and then reverse transcribed into cDNA using the PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Dalian, China).

To obtain the full-length coding DNA sequence (CDS) of NPM APOBEC3G, we designed target-specific primers (forward primer 5′-GTC AGG ACT AGC CGG CAA AGG AT-3′ and reverse primer 5′-CTT CCT TAG AGA CTG AGG CCC ATC-3′) according to the conserved cDNA parts of the noncoding regions of other primates, including humans and rhesus macaques. The PCR conditions included 95°C for 15 min for a hot start, followed by 30 cycles at 94°C for 30 s, 55°C for 30 s, and 72°C for 90 s and a final extension step at 72°C for 10 min. The PCR amplicon was cloned into the pMD19-T simple vector (TaKaRa, Dalian, China) and sequenced. Sequences were analyzed by using MEGA5.0. The CDS of NPM APOBEC3G was deposited in GenBank with accession number KX583654. Based on the CDS of npmAPOBEC3G and hAPOBEC3G (68), we designed the following primers: 5′-gcg aag ctt gcc acc ATG AAT CCT CAA ATC AGA AAC ATG GTG GA-3′ (forward; for NPM normal APOBEC3G), 5′-cct cta gag gcT CAC TTA TCG TCG TCA TCC TTG TAA TCg ccg ccG TTT CCC TGA TTC TGG AGA ATG GC-3′ (reverse; for NPM normal APOBEC3G), 5′-gcg aag ctt gcc acc ATG GAT TAC AAG GAT GAC GAC GAT AAG AAT CCT CAA ATC AGT CCA GGT CC-3′ (forward; for NPM aberrant APOBEC3G), 5′-cct cta gag gcT CAG TTT CCC TGA TTC TGG AGA ATG GC-3′ (reverse; for NPM aberrant APOBEC3G), 5′-gcg aag ctt gcc acc ATG AAG CCT CAC TTC AGA AAC ACAG-3′ (forward; for hAPOBEC3G), and 5′-cct cta gag gcT CAC TTA TCG TCG TCA TCC TTG TAA TCg ccg ccG TTT TCC TGA TTC TGG AGA AT-3′ (reverse; for hAPOBEC3G). In all sequences, capital letters are for exon sequences and lowercase letters are for artificially added sequences, including restriction sites and protection bases. The PCR conditions were 95°C for 15 min for a hot start, followed by 30 cycles at 94°C for 30 s, 60°C for 30 s, and 72°C for 90 s and a final extension step at 72°C for 10 min. The three Flag-APOBEC3G amplicons were cloned into the XbaI and HindIII restriction sites of the pcDNA3.1(+) vector (Invitrogen) and then sequenced (Majorbio, Shanghai, China) for verification.

Western blot analysis.

293T cells were transfected with pcDNA3.1-npmAPOBEC3G-Flag, pcDNA3.1-aberrant-APOBEC3G-Flag, and pcDNA3.1-empty vector in 6-well plates using Lipo2000 according to the manufacturer’s protocols (Invitrogen); harvested at 48 h posttransfection; and lysed in cell lysis buffer (catalog number P0013; Beyotime). The protein extracts were mixed with 5× SDS-PAGE loading buffer and solubilized by boiling for 15 min. The boiled protein samples were separated by SDS-PAGE, and Flag-tagged proteins and GAPDH were detected with mouse monoclonal anti-Flag antibody (Abmart) and anti-GAPDH monoclonal antibody (Proteintech), respectively, and then with horseradish peroxidase (HRP)-conjugated secondary antibody, followed by visualization using chemiluminescence detection reagents (Millipore) (69).

Cotransfection.

The wild-type HIV-1_NL4-3 proviral plasmid (3 μg) was transiently cotransfected with the APOBEC3 expression plasmid (1 μg) or the pcDNA3.1(+) control plasmid (1 μg) into 293T cells using Lipo2000 according to the manufacturer’s recommendations (Invitrogen). The cell culture was replaced at 8 h posttransfection. After 48 h, the viral supernatant was collected, filtered, and then treated with DNase I (TaKaRa, Dalian, China) at 37°C for 1 h to prevent plasmid carryover.

Second-round viral infectivity assays.

The amount of virus in the cotransfection supernatants was quantified by an enzyme-linked immunosorbent assay (ELISA) for p24 Gag content (ZeptoMetrix). TZM-bl cells seeded at a density of 1 × 10⁴ cells per well in 96-well plates were infected with equal amounts of virus. The infections were done in triplicate for 48 h. The expression of luciferase was measured with the Bright-Glo luciferase assay reagent (Promega). Statistical differences between mean percentages were evaluated using two-tailed Student’s t tests, with P values of <0.05 being considered significant.

HIV-1 editing by npmAPOBEC3G in cell culture.

We used 100 μl of the resulting viral supernatant to infect TZM-bl cells cultured on 24-well plates. At 18 h postinfection, the infected TZM-bl cells were washed with phosphate-buffered saline (PBS), and total cellular DNA was then extracted using a DNeasy DNA isolation kit (Qiagen). The extracted DNA was used in the hypermutation test of the HIV-1 env gene as described above (see “Hypermutation of viral DNA in infected PBMCs in vitro”).