Murine Leukemia Virus P50 Protein Counteracts APOBEC3 by Blocking Its Packaging

MLV has existed in mice for at least a million years, in spite of the existence of host restriction factors that block infection. Although MLV is considered a simple retrovirus compared to lentiviruses, it does encode proteins generated from alternatively spliced RNAs. Here, we show that P50, generated from an alternatively spliced RNA encoded in gag, counteracts APOBEC3 by blocking its packaging. MLV also encodes a protein, glycoGag, that increases capsid stability and limits APOBEC3 access to the reverse transcription complex (RTC). Thus, MLV has evolved multiple means of preventing APOBEC3 from blocking infection, explaining its survival as an infectious pathogen in mice.

ABSTRACT

Apolipoprotein B editing enzyme, catalytic polypeptide 3 (APOBEC3) family members are cytidine deaminases that play important roles in intrinsic responses to retrovirus infection. Complex retroviruses like human immunodeficiency virus type 1 (HIV-1) encode the viral infectivity factor (Vif) protein to counteract APOBEC3 proteins. Vif induces degradation of APOBEC3G and other APOBEC3 proteins and thereby prevents their packaging into virions. It is not known if murine leukemia virus (MLV) encodes a Vif-like protein. Here, we show that the MLV P50 protein, produced from an alternatively spliced gag RNA, interacts with the C terminus of mouse APOBEC3 and prevents its packaging without causing its degradation. By infecting APOBEC3 knockout (KO) and wild-type (WT) mice with Friend or Moloney MLV P50-deficient viruses, we found that APOBEC3 restricts the mutant viruses more than WT viruses in vivo. Replication of P50-mutant viruses in an APOBEC3-expressing stable cell line was also much slower than that of WT viruses, and overexpressing P50 in this cell line enhanced mutant virus replication. Thus, MLV encodes a protein, P50, that overcomes APOBEC3 restriction by preventing its packaging into virions.

IMPORTANCE MLV has existed in mice for at least a million years, in spite of the existence of host restriction factors that block infection. Although MLV is considered a simple retrovirus compared to lentiviruses, it does encode proteins generated from alternatively spliced RNAs. Here, we show that P50, generated from an alternatively spliced RNA encoded in gag, counteracts APOBEC3 by blocking its packaging. MLV also encodes a protein, glycoGag, that increases capsid stability and limits APOBEC3 access to the reverse transcription complex (RTC). Thus, MLV has evolved multiple means of preventing APOBEC3 from blocking infection, explaining its survival as an infectious pathogen in mice.

INTRODUCTION

Apolipoprotein B editing enzyme, catalytic polypeptide 3 (APOBEC3) family members are cytidine deaminases that play important roles in the intrinsic response to retrovirus infection. The human genome encodes seven APOBEC3 genes, while the mouse genome encodes a single gene (1). Human APOBEC3G was initially discovered because of its ability to restrict vif-deficient human immunodeficiency virus type 1 (HIV-1). In cells infected with vif-deficient HIV-1, APOBEC3G protein is packaged into progeny virions, making it available to subsequently act in target cells (2). After target cell infection, APOBEC3G deaminates deoxycytidine residues in the viral negative-strand DNA during reverse transcription, resulting in G to A hypermutation in the coding strand. APOBEC3 proteins also inhibit replication of HIV-1 by cytidine deaminase-independent mechanisms, such as blocking reverse transcription and integration (3–9). Vif binds APOBEC3 proteins in virus producer cells and redirects them to ubiquitination and degradation in the proteasome, thereby preventing their incorporation in virions and protecting the viral genome from mutation (10–12).

Mouse APOBEC3 restricts murine retroviruses, including mouse mammary tumor virus (MMTV) and Friend, Moloney, and AKR murine leukemia viruses (MLVs) (13–16). This has been demonstrated in cell culture, as well as in vivo through the use of Apobec3 knockout (KO) mice (13–15, 17). Mouse APOBEC3 does not induce DNA hypermutation of most murine retroviruses. Instead, mouse APOBEC3 restricts mouse retroviruses at an early replication step, likely by binding reverse transcriptase (RT) and inhibiting reverse transcription (5, 18–21). Most, if not all, wild-type (WT) MLVs are partially resistant to restriction by mouse APOBEC3; although infection levels are higher in Apobec3 KO mice, the virus nevertheless replicates in wild-type mice (14, 17, 22). The MLV glycosylated Gag (glycoGag) protein contributes to this resistance. glycoGag enhances the stability of viral cores and prevents APOBEC3 access to the RTC (19, 23, 24). Core stability conferred by glycoGag also blocks access of other host restriction factors such as nucleic acid sensors to the reverse transcription complex (19, 25). HIV-1 core stabilization has similarly been shown to restrict access of restriction factors to this virus (26, 27). In contrast, the MMTV reverse transcriptase (RT) has evolved to impede access of mouse APOBEC3 to transiently exposed minus DNA strands by increasing the rate of reverse transcription (28).

MLV is considered a simple retrovirus, and thus far, no Vif-like protein has been identified. However, viral proteins derived from alternatively spliced mRNAs are produced. Besides the full-length viral genomic and env RNAs, a 4.4-kb viral RNA transcript was identified in Friend MLV (FMLV)- and Moloney MLV (MMLV)-infected cells (29). This viral RNA, termed the alternative splice donor site (SD′) RNA, results from an alternative splice donor site in the gag region that uses the env splice acceptor site (Fig. 1). SD′ mutant viruses replicated more slowly than wild-type viruses, and they induced altered tumorigenesis in mice (28, 30, 31). SD′ RNA is specifically incorporated into virions, and two viral proteins, namely P50, which uses the bona fide Gag start codon, and P60, which initiates from the CUG codon used by glycoGag and is expressed at lower levels than P50, are encoded by this alternatively spliced RNA (32) (Fig. 1). The association of overexpressed P50 with cell membranes and virions suggested that P50 plays a role in virion assembly (32). Taken together, these results imply a role for either the SD′ RNA or its encoded proteins in viral replication. But the exact function(s) of P50 and P60 are largely unknown.

FIG 1.

(A) Map of the splice variants found in MMLV and FMLV. Diagram of the genomic, env, and alternatively spliced RNAs and P50 and P60 proteins. SD, canonical splice donor site; SD′, alternative splice donor site; SA, splice acceptor site. (Modified from references 29 and 32 with permission.) (B) Reverse transcription-PCR (RT-PCR) analysis of RNA from MMLV-, M1-, FMLV-RFP-, and F1-RFP-infected cells, using primers that detect the SD′-generated splice variant, the Env protein (MMLV), or red fluorescent protein (RFP) (FMLV). (C and D) MMLV wild type (WT), M1, FMLV WT, and F1 mutant virus loads in infected mice. Mice were infected with equal amounts of viruses and sacrificed at 16 days postinfection (dpi), and virus titers in spleens were measured. (C) Virus titers in the spleens of APO^+/+ and APO^−/− mice infected with 2 × 10⁴ IC/mouse MMLV and M1 viruses. (D) Virus titers in the spleens of APO^+/+ and APO^−/− mice infected with 2 × 10³ IC/mouse RFP-tagged FMLV and F1 viruses. Error bars indicate the standard deviation (SD) of the mean of each series shown in panels C and D. Significance was determined by unpaired two-tailed t tests. ****, P ≤ 0.0001; **, P ≤ 0.002; *, P ≤ 0.03; NS, not significant. Each point represents the data from an individually infected mouse; the number of mice analyzed is shown above the x axis.

Here, we provide insight into how MLV uses P50 to counteract APOBEC3 restriction. We show that the replication of SD′ mutant viruses is attenuated in wild-type mice compared to that in Apobec3 knockout (KO) mice. We also found that P50 interacts with APOBEC3 protein and blocks its packaging into virions. These data suggest a novel means by which MLV counteracts this restriction factor.

RESULTS

SD′ mutant MLVs are restricted by APOBEC3.

SD′ spliced RNA is easily detected in both MMLV and FMLV-infected cells (Fig. 1B). The SD′ RNA encodes two viral proteins, P50, which includes the matrix protein (MA) p12 and the first 110 amino acids (aa) of capsid (CA) in frame with the last 115 aa of integrase (IN) and an additional and less abundant protein, p60, which is ∼10 kDa larger because it uses the upstream CUG glycoGag start codon (Fig. 1A) (29, 32–34). Hence, the N-terminal portions of P50 and P60 are the same as those in Gag and glycoGag, respectively (encoded by the nucleotides up to the splice junction in the CA-encoding region). We first examined the replication of SD′ mutant viruses in Apobec3 KO (APO^−/−) and wild type (APO^+/+) mice, using mutant viruses generated in both the Moloney (MMLV) and Friend (FMLV) backbone. M1 is a mutant virus derived from MMLV, with 3 nucleotide mutations near the SD’ site that leave the Gag amino acid sequence unchanged (Fig. 1A). SD′ RNA was undetectable in M1-infected cells (29) (Fig. 1B). We infected APO^+/+ mice and APO^−/− mice with the same amount of M1 and WT viruses (2 × 10⁴ infectious center [IC] units) and found that although infection levels were slightly lower in APO^−/− mice, M1 virus titers were 5.5-fold lower than those of wild-type MMLV in APO^+/+ mice (Fig. 1C).

To determine if this defect in counteracting APOBEC3 occurred in additional MLV strains with SD′ mutations, we next tested a mutant virus derived from FMLV, called F1, which has the same 3-nucleotide silent mutation near the SD′ site as M1 and no SD′ RNA expression (Fig. 1A) (29). To facilitate viral titer assays, we generated red fluorescent protein (RFP)-tagged F1 and FMLV by engineering RFP into FMLV-2A, which has a picornavirus 2A peptide followed by a cloning site downstream of env; we previously used this construct to express HIV Vif in an infectious virus (35, 36). Neonatal APO^+/+ mice and APO^−/− mice were injected with equal amounts of F1 or FMLV viruses (2 × 10³ ICs). The F1 virus was slightly attenuated compared to the wild-type virus in APO^−/− mice (F1 titers were on average about 0.7-log lower than FMLV titers in APO^−/− mice); this is probably due to effects of the mutation on the spliced env mRNA levels, which are reduced (29). However, the difference between the WT and F1 virus was close to 2 logs in APO^+/+ mice (Fig. 1D). Thus, both M1 and F1 mutant viruses replicated more poorly in APO^+/+ mice than did WT viruses.

M1 mutant viruses do not revert to WT in vivo.

glycoGag mutant virus has a point mutation that generates a stop codon in the gPr80 reading frame (UAU to UAG). This mutation reverted to WT in BL/6 and BALB/c but not in APO^−/− mice, and reversion was concomitant with increased virus load by 6 weeks postinfection (19, 37, 38). The M1 mutant virus has three nucleotide changes near the SD′ site. To investigate whether APOBEC3 also puts selective pressure on the M1 mutant causing it to revert to WT, we infected APO^+/+ and APO^−/− mice with M1 mutant virus. At 8 weeks after infection, DNA was isolated from the spleens, PCR was performed using primers that flank the SD′ acceptor region, and the PCR products were sequenced. M1 virus did not revert to wild type in either APO^+/+ or APO^−/− mice, nor did we see an increase in M1 virus titer, suggestive of reversion (Table 1). We suspect that the 3-nucleotide mutation found in M1 virus, as opposed to the single-nucleotide change in glycoGag virus, greatly diminishes the likelihood of RT misincorporation and reversion.

TABLE 1.

M1 mutant virus does not revert to WT in APO^−/− and APO^+/+ mice^a

Strain	No. of mice with revertant/total no. of mice	Sequence	Log10 IC/106 cells
APO−/−	0/6	ACGC	4.5 ± 0.10
APO+/+	0/6	ACGC	3.3 ± 0.08

^aThe wild-type M-MLV SD′ site sequence is UAGG.

P50 partially rescues SD′ mutant virus replication.

To determine if P50 counteracts mouse APOBEC3, we did a rescue experiment. We generated stable NIH 3T3 cells expressing either green fluorescent protein (GFP)-tagged mouse APOBEC3 or GFP alone (Fig. 2A). The APOBEC3-expressing and GFP control stable cell lines were infected with FMLV or F1 mutant virus, and virus isolated from the culture medium was subjected to Western blot (WB) analysis using anti-MLV antisera (Fig. 2A), and titer was determined (Fig. 2B). F1 titers were lower in than FMLV titers in the GFP-expressing stable cell line, consistent with the mild attenuation of the virus in vivo (Fig. 2B). However, as was seen in vivo in APO^+/+ mice, the F1 titers in the APOBEC3-expressing stable cell line were lower than those in the GFP-expressing stable cell line, while the FMLV titers were not affected by the presence of mA3 (Fig. 2B).

FIG 2.

P50 expression partially restores the replication of F1-RFP mutant virus. (A) Western blots (WBs) showing the expression of GFP-tagged APOBEC3 (top panel) in FMLV- and F1-infected cells. The bottom panel is a Western blot using anti-MLV antisera and shows the virions produced from these cells. (B) Virus titers of the supernatants of infected cell lines stably expressing mAPOBEC3-GFP (mA3) or control (GFP). Bar shows the average ± standard deviation of 3 independent experiments, represented by individual points. One-way analysis of variance (ANOVA) with Tukey’s multiple-comparison test was used to determine significance. ***, P ≤ 0.0002; **, P ≤ 0.001; ns, not significant. (C) Fluorescence microscopy of cells transiently transfected with the P50-GFP, P60-GFP, and GFP expression vectors. (D) Viral titers in the supernatants of infected stable cell lines transfected with the indicated plasmids. The infected mouse APOBEC3 (mA3)-expressing stable cell line was transfected with GFP, P50-GFP, or P60-GFP expression vectors; supernatants were collected after 72 h and used for IC assays. Shown are the averages ± standard deviation of 3 independent experiments, represented by individual points. One-way ANOVA with Tukey’s multiple comparison was used to determine significance. **, P ≤ 0.01.

We then transiently transfected GFP, GFP-P50, or GFP-P60 expression plasmids into the F1- and FMLV-infected cell lines stably expressing APOBEC3. At 72 h posttransfection, the transfection frequency was assessed by immunofluorescence microscopy, the media were collected and titers were determined. Transfection efficiencies were roughly similar with the 3 constructs, although the GFP-alone construct expressed at higher levels in individual cells (Fig. 2C). P50 expression significantly increased F1 but not FMLV titers (Fig. 2D). The lack of complete rescue is likely because not all of the cells expressed P50-GFP (Fig. 2C). Cotransfection of the P60 or GFP plasmid did not significantly increase infection levels. Thus, P50 expression can rescue F1 mutant virus replication.

P50 does not act by blocking APOBEC3-mediated G-to-A mutations in proviral DNA.

It is well recognized that human APOBEC3 proteins restrict HIV and other retroviruses mainly by means of cytidine deamination, although they also restrict infection by deaminase-independent means (3–9). In contrast, although mouse APOBEC3 retains functional cytidine deaminase activity, it restricts both MLV and MMTV mostly by cytidine-deaminase-independent means, likely by inhibiting reverse transcription. One possible mechanism by which P50 might relieve APOBEC3 restriction would be blocking its deaminase activity. To determine whether P50 protein played a role in the resistance to mouse APOBEC3-mediated deamination, we isolated DNA from MMLV- and M1-infected splenocytes and sequenced a portion of the env gene, which we previously showed was highly deaminated by human APOBEC3G but not by mouse APOBEC3 in vivo (35). No G-to-A mutations were observed in viral DNA isolated from either APO^+/+ or APO^−/− mice (Table 2). Thus, P50 does not block APOBEC3 by preventing its deamination activity.

TABLE 2.

G-to-A mutations in MLV M1 proviral DNA^a

Strain	No. of mice	No. of clones sequenced
		Total	With G-to-A mutation	Without mutation
APO−/−	4	35	0	35
APO+/+	4	45	0	45

^aMouse APOBEC3 does not induce G-to-A mutations in M1 proviral DNA isolated from MLV M1 virus-infected splenocytes of APO^−/− or APO^+/+ mice. A 549-bp segment from the env gene of MLV was amplified and sequenced.

P50 interacts with APOBEC3.

We first found that the MLV IN and APOBEC3 proteins bound each other in coimmunoprecipitation (co-IP) assays (Fig. 3A). Since both the P50 and P60 proteins contained the C terminus of IN, we wondered whether they would also bind APOBEC3. Since P50 is more abundant than P60 in MLV-infected cells and was more effective at rescuing infection by F1 virus (Fig. 2B), we focused on P50. We amplified the P50 coding sequence (CDS) from MMLV producer cells and subcloned it into the myc-BioID2-MCS pCDNA3.1 plasmid, which contains an Myc tag and BioID2 tag upstream of the multiple cloning site (MCS). This protein was abundantly expressed after transfection of the vector into 293T cells (Fig. 3B).

FIG 3.

Coimmunoprecipitation of APOBEC3 and MLV P50. (A) 293T cells were cotransfected with Myc-tagged IN construct and FLAG-tagged APOBEC3. Lysates were immunoprecipitated, and Western blots (WBs) were probed with anti-Myc and anti-FLAG antibodies. (B) 293T cells were cotransfected with Myc-tagged MMLV P50 and FLAG-tagged full-length or truncated mouse APOBEC3 expression constructs, diagrammed above the WBs. Lysates were immunoprecipitated with an anti-Myc antibody, and WBs were probed with an anti-FLAG antibody. An expression construct encoding red fluorescent protein (RFP) was used as a control. Shown are representative WB of 3 independent experiments.

Mouse APOBEC3 protein has two conserved zinc-coordinating domains, referred to as the cytidine deaminase domains (CD). The N-terminal CD1 of mouse APOBEC3 possesses the cytidine deaminase activity, while the C-terminal CD2 domain is essential for the encapsidation (39). We subcloned full-length mouse APOBEC3, as well as the N-terminal and C-terminal CD domains, into FLAG-tagged expression vectors; FLAG-tagged red fluorescent protein (RFP) served as a control (Fig. 3B). By co-IP experiments, we found that P50 interacted with full-length as well as the C-terminal domain of mouse APOBEC3, but not the N-terminal domain (Fig. 3B). Thus, P50 binds mouse APOBEC3 via its CD2 encapsidation domain.

MLV infection does not induce APOBEC3 degradation.

To overcome the antiviral effect of APOBEC3 proteins, in particular APOBEC3G and APOBEC3F, HIV-1 encodes Vif, which binds APOBEC3G and induces its ubiquitination and proteasomal degradation. To determine whether MLV uses a similar mechanism to overcome restriction by mouse APOBEC3, we infected the cell line stably expressing the GFP-tagged APOBEC3 with RFP-tagged FMLV, and at 24 h postinfection (hpi), the cells were observed by fluorescence microscopy. Virtually all of the cells were positive for both APOBEC3-GFP and FMLV-RFP, and there was no diminution in the APOBEC3-GFP signal in the infected cells (Fig. 4A). The cell lysates were collected and analyzed by Western blotting. The level of APOBEC3 protein was not altered by FMLV infection (Fig. 4B).

FIG 4.

MLV infection does not induce mAPOBEC3 degradation. (A and B) NIH 3T3 cells stably expressing GFP-tagged APOBEC3 were infected with FMLV at a multiplicity of infection (MOI) of 1. At 24 hpi, extracts were prepared and analyzed by (A) immunofluorescence and (B) WB. Line indicates GFP-mAPOBEC3 protein; the other bands are nonspecific cellular proteins that react with the anti-GFP antisera. (C) 3T3-N43 cells (cell line infected with MMLV) were transfected with the HA-tagged APOBEC3, APOBEC3G, or RFP expression vectors, and at 24 h posttransfection, extracts were analyzed by WB. The control is a nonspecific band recognized by the anti-HA antisera. Shown in panels B and C are representative WBs of 3 independent experiments. (D) Extracts made from the spleens of three FMLV-infected and uninfected APOBEC3^+/+ mice were each prepared and analyzed by WB for APOBEC3, MLV, and GAPDH. As a control, spleen extract from an uninfected APO^−/− mouse, as well as extract from NIH 3T3 cells stably expressing APOBEC3-GFP, were also included. NS, nonspecific band detected by the anti-APOBEC3 antisera, as previously reported (18, 40).

We obtained similar results by transiently transfecting 3T3-N43 cells (NIH 3T3 cells infected with MMLV) with a mouse APOBEC3 expression plasmid; a plasmid expressing human APOBEC3G was also tested. Mouse APOBEC3 levels were the same in MMLV-infected cells and in uninfected, APOBEC3-expressing NIH 3T3 cells (Fig. 4C). APOBEC3G levels were also not altered.

We also examined whether in vivo infection altered endogenous APOBEC3 levels. Lysates made from the spleens of FMLV-RFP-infected mice at 16 days postinfection (dpi) were subjected to Western blot analysis with anti-mouse APOBEC3 antibodies (40). The level of APOBEC3 was not affected by FMLV infection (Fig. 4D). Thus, neither Friend nor Moloney MLV induce APOBEC3 degradation.

P50 does not disrupt RT-APOBEC3 interaction.

Mouse APOBEC3 restricts FMLV, MMLV, or MMTV by blocking reverse transcription. We and others previously showed that APOBEC3 binds MLV RT and that this binding was not dependent on RNA; thus, the ability of mouse APOBEC3 to inhibit MMLV reverse transcription may depend on its ability to bind RT (20, 41). To investigate whether P50 disrupted RT-APOBEC3 interaction, we transiently transfected APOBEC3 (FLAG tag) and MMLV RT (Myc tag) expression vectors and carried out co-IPs with RT and APOBEC3 in the presence of increasing amounts of GFP-tagged P50; as a control, co-IPs were carried out in the presence of the same amounts of GFP alone. Whether P50 was present or not, the amount of the coimmunoprecipitated APOBEC3 and MLV RT was similar (Fig. 5). In the control experiments, smaller amounts (0.2 or 0.5 μg) of transfected GFP-expressing vector also did not affect the interaction, while transfection with the largest amount (≥1 μg) decreased the expression of both MLV-RT and APOBEC3, likely due to competition for transcription or translation factors (Fig. 5). Thus, the P50 does not disrupt RT-APOBEC3 interaction.

FIG 5.

P50 does not disrupt the interaction between APOBEC3 and MLV RT. 293T cells were cotransfected with Myc-tagged reverse transcriptase (RT), FLAG-tagged mAPOBEC3, and increasing amounts of GFP-tagged P50. Lysates were immunoprecipitated with an anti-Myc antibody, and Western blots were probed with the indicated antibodies. Shown is a representative WB of 3 independent experiments.

P50 blocks APOBEC3 packaging.

HIV-1 Vif induces APOBEC3G degradation and prevent its packaging. Mouse APOBEC3 is also incorporated into MLV virions. Our results suggested that rather than induce its degradation, P50 blocked APOBEC3 by other means. To investigate the underlying mechanism, we first carried out colocalization analysis of mouse APOBEC3 and viral P50. We transfected the mouse APOBEC3-GFP stable cell line with an expression plasmid encoding HA-tagged P50, and expression was examined by fluorescence microcopy. We found that P50 and APOBEC3 were colocalized in the cytoplasm and that APOBEC3 localization was not changed by coexpression of P50 (Fig. 6A).

FIG 6.

More mAPOBEC3 is packaged in M1 virus than in WT MMLV. (A) Colocalization of HA-P50 with APOBEC3-GFP. HA-tagged P50 was expressed in an APOBECE-GFP 3T3 stable cell line, and the cells were stained with anti-HA. (B) Western blot of the virions isolated from APO^−/− and APO^+/+ mice infected with MMLV WT or M1 mutant virus and probed with anti-APOBEC3 (upper) or anti-MLV (lower) antisera. Shown is a representative blot. The faint bands in the APO^−/− lanes are background bands detected by the antisera, as previously reported (18). Shown below is quantitative analysis of the amount packaged APOBEC3 protein normalized to the viral p30 protein, using ImageJ analysis software (NIH) and averaged from three independent experiments. Shown is the average ± standard deviation for each series. Significance was determined by two-tailed unpaired t test. **, P ≤ 0.004. (C) Western blot of the virions isolated from M1- or MMLV (WT)-infected stable cell lines transfected with the HA-P50 plasmid and probed with anti-HA (upper), anti-MLV (middle), or anti-APOBEC3 (A3) antisera. Shown is a representative blot of four different experiments.

To determine whether SD′ protein affected packaging of APOBEC3 into virions in vivo, 2-day-old APO^+/+ and APO^−/− mice were infected with MMLV and M1 viruses, and at 16 dpi, virions were isolated from spleens and analyzed by Western blots for viral proteins and packaged APOBEC3. We found that substantially more APOBEC3 was packaged into M1 than into MMLV virions (Fig. 6B).

Finally, we tested whether P50 was packaged into virions. APOBEC3-GFP stable cells were transiently transfected with HA-P50 and then infected with MMLV or M1 virus. The virions produced by these cells were analyzed by Western blotting for the presence of P50. A similar amount of hemagglutinin (HA)-P50 protein was packaged into virions when APOBEC3 present (Fig. 6C). Similar to what was seen with virions produced in vivo, M1 virions packaged more APOBEC3 than did MMLV virions. However, the low levels of APOBEC3-GFP packaged into either MMLV or M1 virions in vitro does not allow us to rule out that there is competition for packaging in virions produced in vivo.

DISCUSSION

In this study, we provide insight into the function of MLV P50, an MLV protein produced from an alternatively spliced RNA, and provide a mechanism by which P50 counteracts the restrictive effects of APOBEC3 on MLV infection. By combining in vivo and in vitro infection with biochemical analysis, we found that P50 interacts with mouse APOBEC3, preventing its packaging and thereby blocking APOBEC3-mediated restriction of MLV. This means of blocking APOBEC3 restriction is similar to that of the foamy virus Bet protein, which also prevents APOBEC3 packaging without inducing its degradation (42).

The SD′-mutant viruses replicated to lower levels in APO^+/+ mice than did WT viruses, demonstrating the important role that the protein(s) generated from the alternatively spliced message plays in counteracting APOBEC3. Although both P50 and P60 can be produced from the SD′ RNA, our data suggest that P50 is more important for counteracting APOBEC3, as follows: (i) overexpressed P50 but not P60 rescued the F1 virus (Fig. 2D), and (ii) our previous results showed that APOBEC3 is packaged at similar levels in WT and glycoGag mutant viruses; since P60 and glycoGag use the same CUG start codon, P60 is not produced by the glycoGag mutant virus (19). Taken together, this strongly suggests that P50 but not p60 prevents APOBEC3 packaging.

One possibility was that one of the viral SD′-encoded proteins acted as a Vif-like protein. By cotransfection, co-IP experiments, we demonstrated that P50 interacts with the C-terminal CD2 domain of mouse APOBEC3. Although this binding did not affect the stability of mouse APOBEC3, it did affect APOBEC3 packaging. Considering that the C terminus of APOBEC3 is necessary for its incorporation into viruses, it is not surprising that P50 binding to this domain affected APOBEC3 packaging. Whether this occurs through blocking APOBEC3 interaction with NC or RNA or both is under investigation.

To infect host cells, viruses must be able to counteract the antiviral responses of host cells imposed on them. Host APOBEC3 family proteins play important roles in intrinsic responses to infection by retroviruses. Some complex retroviruses, such as HIV, simian immunodeficiency virus (SIV), and feline immunodeficiency virus, encode Vif proteins that directly bind APOBEC3 proteins and lead to their proteasomal degradation in virus-producing cells (10–12). However, other retroviruses, like MLV, have no apparent Vif-like genes. Instead, MLV glycoGag protein stabilizes the viral core and prevents APOBEC3 access to the RTC as well as to cytosolic sensors of foreign DNA (19). MLV is a so-called simple retrovirus carrying only 3 genes in its genome. That the SD′ RNAs derived from an alternative 3′ splice acceptor site in gag produce a protein, P50, that counteracts APOBEC3, suggests that our notion of “simple” and “complex” retroviruses needs reexamination.

Although both Vif and P50 prevent the packaging of APOBEC3, those two proteins still have distinct differences in their actions. Vif does not specifically incorporate into virions, but P50 does. Vif induces APOBEC3G protein degradation, but P50 does not induce mouse APOBEC3 degradation. Moreover, Vif is more efficient at preventing APOBEC3G packaging into HIV-1 than is P50 in preventing APOBEC3 packaging into MLV. All of these differences imply that P50 not as powerful as Vif protein in counteracting APOBEC3 proteins. This may explain why MLV possesses two mechanisms to overcome the restriction by mouse APOBEC3, P50 to prevent mouse APOBEC3 packaging and glycoGag to affect the capsid stability and sequester packaged APOBEC3 away from the RTC. Individually, neither of those two means is completely effective, whereas when they are combined, MLV can more efficiently overcome the restriction imposed by APOBEC3 (Fig. 7).

FIG 7.

Model showing how glycoGag and P50 block APOBEC3 from restricting MLV. In WT viruses, P50 in the cytoplasm sequesters APOBEC3 (A3) and prevents it from being packaged, and the stable glycoGag-containing capsids prevent APOBEC3 from accessing the RTC. In the SD′ mutant, more APOBEC3 is packaged, thereby decreasing reverse transcription. In the glycoGag mutant virus, the capsid is loosened, and packaged APOBEC3 more easily accesses the RTC.

The results presented here show that P50 binds APOBEC3 and prevent its packaging. The potential SD′ site is highly conserved in many mammalian simple retroviruses. A similar mechanism may be used by other simple retroviruses that replicate in the presence of APOBEC3 but lack vif-like genes.

MATERIALS AND METHODS

Ethics statement.

All mice were housed according to the policy of the Animal Care Committee of the University of Illinois at Chicago, and all studies were performed in accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The experiments performed with mice in this study were approved by the committee (UIC ACC protocol no. 18-168).

Mice.

APO^−/− mice and APO^+/+ control mice were bred at the University of Illinois at Chicago and were previously described (13).

Cell culture and transfection.

NIH 3T3 cells and 293T cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS), l-glutamine, and penicillin/streptomycin. Lipofectamine 3000 (Invitrogen) and Lipofectamine 2000 (Invitrogen) were used for 3T3 cells and 293T transfections, respectively.

Virus isolation.

FMLV-RFP WT, F1-RFP MLV, MMLV, and M1 MLV were isolated from the supernatants of stably infected NIH 3T3 cells (cells in which infection is allowed to spread to 100% of the culture and maintained in this state thereafter), as previously described (19, 43). For virus isolation from spleen, spleen tissues were homogenized and resuspended in 10 ml DMEM medium, 10% fetal bovine serum (FBS), nonessential amino acids, and penicillin-streptomycin. The medium was passed through a 0.45-μm filter, treated with 20 U/ml DNase I (Sigma) at 37°C for 30 min, and pelleted through a 25% sucrose cushion. After resuspension, titers of FMLV-RFP WT, F1-RFP MLV, MMLV WT, and M1 MLV were determined on NIH 3T3 cells (see “Virus titers,” below). Viruses were subjected to reverse transcriptase quantitative PCR (RT-qPCR), and the number of viruses was estimated from the amount of virus-specific RNA using primers located in the env gene (MMLV F primer, 5′-CCTACTACGAAGGGGTTG-3′; MMLV R primer, 5′-CACATGGTACCTGTAGGGGC-3′; FMLV F primer, 5′-AAGTCTCCCTCCGCC-3′; and FMLV R primer 5′-AGTGCCTGGTAAGCTCCCTGT-3′). The titer of each preparation was also determined on NIH 3T3 cells to determine infectivity. Representative levels of virion RNAs and IC units for different preparations of viruses used in this study are found in Table 3. Viruses were also analyzed by Western blotting with anti-MLV antiserum (polyclonal goat anti-MLV antibody; NCI Repository). Viral infectivity ratios were calculated by quantifying the p30 signal on WBs and normalizing the viral titers to this arbitrary unit. The ratios for the viruses used in this study are as follows: MMLV, 1.02 × 10⁵ IC/p30; M1, 6.1 × 10⁴ IC/p30; FMLV-RFP, 1.4 × 10⁵ IC/p30; and F1-IC, 4.9 × 10³ PFU/p30. In all in vivo and in vitro experiments, the same number of IC units was used for infection, as specified in the text.

TABLE 3.

Representative infectivity to particle ratios^a

Virus	IC/ml	Virus/ml	Infectivity/particle ratio
MMLV	4.5 × 105	4.48 × 108	1.00 × 10−3
M1	2.0 × 105	1.24 × 109	1.61 × 10−4
FMLV-RFP	1.4 × 105	1.03 × 108	1.36 × 10−3
F1-RFP	2.2 × 104	1.11 × 108	1.98 × 10−4

^aSee Materials and Methods for details.

Virus titers.

MMLV WT and M1 infection levels in the spleens of the infected mice or the supernatants of infected cells were determined by IC assays using a focal immunofluorescence assay, as previously described (14). Briefly, NIH 3T3 cells were infected with 10-fold serial dilutions of virus. At 4 days postinfection, the plates were stained a monoclonal antibody (538) that recognizes the Env protein. After staining with fluorescein-conjugated secondary antibody, the colonies of green cells were quantified by automated counting using a Keyence fluorescence microscope. Viral titers (ICs) were calculated from the numbers of fluorescent colonies corrected for the dilution factors of the viral stocks in each plate. FMLV-RFP and F1-RFP infection levels were also determined by IC assays. Instead of using antibody, ICs were counted directly using the Keyence fluorescence microscope.

Identification of the SD′-spliced RNA in infected cells.

Total RNA was extracted from mock-infected or NIH 3T3 cells infected with the indicated viruses. Relative semiquantitative PCR was performed. For MMLV and M1, primers P50 F1 (5′-ATGGGCCAGACTGTTACC-3′) and P50 B1 (5′-GGGGGCCTCGCGGGTTAA-3′) were used to detect SD′ RNA. These primers also weakly hybridize to the FMLV SD′ RNA. The env primers described above (“Virus isolation”) were used to detect the spliced env RNA. For FMLV-RFP and F1-RFP, the primers FMLV SDF (5′-GGATGACTGCCAACAGCT-3′) and FP50 B1 (5′-TTAGGAGGTCCCGCGGGT-3′) were used to detect SD env RNA. RFPF1 (5′-AGACTCGAGGCCTCCTCCGAGGACGTC-3′) and RFPB1 (5′-GTGGAATTCTTAGGCGCCGGTGGAGTG-3′) were used to detect RFP RNA.

In vivo infections.

For MMLV WT and M1 infections, 2-day-old mice were infected by intraperitoneal injection of 2 × 10⁴ ICs of WT MMLV or M1 and harvested at 16 dpi, as previously described (19). For FMLV-RFP and F1-RFP infection, 2 × 10³ ICs were injected into 2-day-old mice.

Revertant analysis.

APO^−/− and APO^+/+ mice (2 days old) were infected with M1 and sacrificed at 8 weeks. DNA was isolated from spleens, and the M1 region was PCR amplified and sequenced as previously described (19). Briefly, DNA was isolated from spleens, and the M1 region was amplified from 100-ng samples of DNA with the primers MMLV SDF (GACTCTGCTGACCGGAGA) and MMLV SDB (CTTTCTCCCAATGTCTGG). DNA was also isolated from spleens of noninfected and MMLV-infected APO^+/+ mice, and the M1 region was PCR amplified with the same primers. The PCR products were loaded into the agarose, and the gel was run. The M1 region band was only found in PCR products from infected mice and not in those from uninfected ones. The band was purified and subjected to sequencing with the same primers. The appearance of revertants in samples was assessed by the sequence traces for nucleotides by SeqMan software (Lasergene). We judged by the overlapping peak size on the y axis. We had the following criteria: (i) the sequencing data were clean, and no nonspecific data were detected in the regions we analyzed, and (ii) the percentage of revertants was determined by the relative height of the overlapping peaks and had to be more than 20% of the height of the highest peaks. This method was successfully used previously (19, 41).

Expression constructs.

To generate the GFP-pcDNA vector, the GFP CDS was amplified by PCR using the primers GFP F1 (5′-CAGGCTAGCCACCATGGTGAGCAAGGGC3-3′, containing an NheI restriction site and Kozak sequence before the translation start codon) and GFP B1 (5′-TAACTCGAGACCTCCACCCTTGTACAGCTCGTCCATGCC-3′, with an XhoI restriction site and 3-GLY linker at the 3′ end). The CDS was then subcloned into the FLAG-HA-pcDNA3.1 vector (Addgene plasmid no. 52535). GFP-mAPOBEC3-pcDNA was generated by cloning the mouse Apobec3 CDS into GFP-pcDNA. To generate the FLAG-pcDNA vector, forward (5-CTAGCCACCATGGACTACAAAGACGATGACGATAAAGGTGGAGGTGGAGGTTCTAGAC-3′) and reverse (5′-TCGAGTCTAGAACCTCCACCTCCACCTTTATCGTCATCGTCTTTGTAGTCCATGGTGG-3′) oligonucleotides encoding the FLAG tag sequence were inserted into the FLAG-HA-pcDNAPOBEC3.1 vector, which was cut by NheI and XhoI. The full-length mouse APOBEC3 CDS was amplified by PCR using mA3 F1 (5′-AGCCTCGAGATGGGACCATTCTGTCT-3′) and mA3 B1, (5′-GTGGAATTCTCAAGACATCGGGGGTCC-3′). The N-terminal CDS fragment of APOBEC3 was amplified using mA3 F1 and mA3 B2 (5′-GTGGAATTCTCACATTCGCCTCAGAATCTC-3′). The C-terminal CDS fragment of APOBEC3 was amplified using mA3 F2 (5′-AGCCTCGAGCTGAGGCGAATGGACCCG-3′) and mA3 B1. The amplified fragments were subcloned into the FLAG-pcDNA vector. The HA-pcDNA vector was generated using a similar method to that used to generate the FLAG-pcDNA vector (primers HA sense, 5′-CTAGCCACCATGTACCCATACGATGTTCCAGATTACGCTGGTGGAGGTGGATCTAGAC-3′, and HA reverse, 5′-TCGAGTCTAGATCCACCTCCACCAGCGTAATCTGGAACATCGTATGGGTACATGGTGG-3′). The P50 CDS was amplified using MMLV P50 F1 (5′-ATGGGCCAGACTGTTACC-3′) and P50 B1 (5′-GGGGGCCTCGCGGGTTAA-3′) and was then subcloned into HA-pcDNA or pcDNA3.1 mycBioID (Addgene plasmid no. 35700). The IN CDS was amplified using IN F2 (5′-AGCCTCGAGATAGAAAATTCATCACCC-3′) and IN B2 (5′-GTGGAATTCGGGGGCCTCGCGGGTT-3′) and was then subcloned into pcDNA3.1 mycBioID. To generate the FMLV-RFP vector, the CDS of RFP was amplified by PCR using primers RFP F1 (5′-CTAGCGGCCGCAATGGCCTCCTCCGAGGA-3′) with an added NotI restriction and RFP B1 (5′-TGCGGCCGCTTAGGCGCCGGTGGAGTG-3′), also with a Not I restriction site, followed by NotI digestion and subcloning into the NotI site of FMLV-2A-vif (19). F1-RFP was generated using a site-directed mutagenesis kit from NEB. The primers F1F (5′-CCCAACGAGGACGCAACCACCTAGTC-3′) and F1B (5′-TGTTGTAGTCCCAGTCGG-3′) were used for the mutation. The Myc-tagged MLV RT construct was a gift from Stephen Goff. The GFP-p50 and GFP-p60 expression constructs were previously described (29, 32). All plasmids were sequenced prior to use.

Western blot analysis.

Affinity-purified polyclonal rabbit anti-mouse APOBEC3 antibody has been previously described (40). Polyclonal goat anti-MLV antibody (NCI Repository), anti-Gapdh, mouse anti-Myc, mouse anti-HA, rabbit anti-HA, rabbit anti-GFP, horseradish peroxidase (HRP)-conjugated anti-rabbit (Cell Signaling Technology), and mouse anti-FLAG, anti-goat and anti-mouse antibody (Sigma-Aldrich) were used for detection, using either Amersham ECL Prime Western blotting detection reagent (GE Healthcare Life Sciences) or Pierce ECL Western blotting substrate (Thermo Scientific).

Coimmunoprecipitation of APOBEC3 and P50.

HEK293 cells were transfected with plasmids expressing the Myc-tagged P50 and FLAG-tagged APOBEC3 proteins. At 24 h after transfection, cells were lysed in cell lysis buffer (Cell Signaling) containing 20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM Na₂EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM-glycerophosphate, 1 mM NAPOBEC3VO4, 1 g/ml leupeptin, and a Halt protease and phosphatase inhibitor cocktail (Thermo Scientific). Supernatants were incubated with monoclonal anti-Myc antibody (Cell Signaling) at 4°C with gentle rotation for 4 h, and then protein A/G Plus-agarose (Santa Cruz Biotechnology) was added and incubated at 4°C with gentle rotation overnight. Following immunoprecipitation, the beads were washed four times with lysis buffer. The immunoprecipitated proteins were analyzed by immunoblot analysis using rabbit anti-HA or mouse anti-Myc antibodies (Cell Signaling).

Stable cell line generation.

3T3 cells were transfected with pCDNA-GFP or pCDNA-GFP-mAPOBEC3 plasmids, then were selected in DMEM medium containing 500 μg/ml G418 for 2 weeks. The GFP-positive cells were singly sorted into a 96-well plate using a MoFlo Astrios cell sorter and cultured for about 3 weeks in DMEM containing 100 μg/ml G418 medium to generate stable cell lines. The expression of GFP or GFP-mAPOBEC3 in stable cell lines was assessed by microscopy and Western blot analysis. Cells highly expressing GFP were selected for experiments.

Immunofluorescence.

In total, 5 × 10⁵ cells per well were seeded into 6-well plates. The next day, cells were transfected with 2 μg of the HA-P50 plasmid. At 24 h posttransfection, 2 × 10⁵ transfected cells per well were seeded into 4-chamber culture slides (Millicell EZ slide; Millipore). The next day, cells were rinsed with ice-cold phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 15 min at room temperature, which was followed by permeabilization with 0.3% Triton X-100. The cells were subjected to immunofluorescence staining with anti-HA (1:1600) antibody (Cell Signaling) overnight at 4°C. The cells were then washed with cold PBS three times for 5 min and incubated with Alexa 568-labeled anti-rabbit secondary antibody (1:1,000) (Invitrogen) at room temperature for 1 h. The cells were examined by fluorescence microscopy (Keyence).

Incorporation of APOBEC3 into virions.

For immunoblot analysis of APOBEC3 packaged in virions made in vivo, virus was isolated from spleens of infected mice as described in “Virus isolation.” Viruses were then analyzed by Western blotting with anti-MLV and anti-APOBEC3 antibodies. The bands were quantified using ImageJ (NIH), and APOBEC3 levels were normalized to p30 CA levels. For APOBEC3 packaged into virions made in vitro, GFP or GFP-APOBEC3 stable cell lines were infected with MMLV or M1 mutant virus for 2 weeks, then 6 × 10⁶ cells were seeded in 10-cm dishes. The next day, the cells were transiently transfected with 20 mg HA-P50 expression plasmids. Twenty-four hours after transfection, viruses were isolated from media by sucrose cushions. The cells were also collected and lysed by the lysis buffer. Both the viruses and cell lysates were analyzed by Western blotting with anti-HA, anti-MLV, and anti-APOBEC3 antibodies.

Statistical analysis.

Data shown are the averages of at least 3 independent experiments, or as otherwise indicated in the figure legends. Statistical analysis was performed using GraphPad Prism 8.1 software. Tests used to determine significance are indicated in the figure legends. All raw data have been deposited in the Mendeley data set found at https://doi.org/10.17632/f86f8wppd6.1.