HIV-1 nef suppression by virally encoded microRNA

Background MicroRNAs (miRNAs) are 21~25-nucleotides (nt) long and interact with mRNAs to trigger either translational repression or RNA cleavage through RNA interference (RNAi), depending on the degree of complementarity with the target mRNAs. Our recent study has shown that HIV-1 nef dsRNA from AIDS patients who are long-term non-progressors (LTNPs) inhibited the transcription of HIV-1. Results Here, we show the possibility that nef-derived miRNAs are produced in HIV-1 persistently infected cells. Furthermore, nef short hairpin RNA (shRNA) that corresponded to a predicted nef miRNA (~25 nt, miR-N367) can block HIV-1 Nef expression in vitro and the suppression by shRNA/miR-N367 would be related with low viremia in an LTNP (15-2-2). In the 15-2-2 model mice, the weight loss, which may be rendered by nef was also inhibited by shRNA/miR-N367 corresponding to suppression of nef expression in vivo. Conclusions These data suggest that nef/U3 miRNAs produced in HIV-1-infected cells may suppress both Nef function and HIV-1 virulence through the RNAi pathway.


Background
The human immunodeficiency virus (HIV), which infect humans cause acquired immunodeficiency syndrome (AIDS), which has reached pandemic levels in some societies, especially those in Southern Africa and Southeast Asia [1]. Given the immensity of HIV pandemic, the development of a rather safe and cheap, effective thera-peutics, has become the main focus [2]. Several strategies attempted to control the spread of AIDS have not shown major breakthrough and the vaccines have shown little promise as far as their efficacy is concerned. However, one approach used extensively in other diploid organisms, which now has tremendous potential to encourage antiviral defense against HIV appears to be double stranded RNA-dependent post-transcriptional gene silencing or RNA interference (RNAi).
RNAi is a defense mechanism against aberrant transcripts that may be produced during viral infection and mobilization of transposons [3,4]. The RNAi pathway has been implicated in silencing transposons in the C. elegans germline [5,6], silencing stellate repeats in the Drosophila germline, and the response against invading viruses in plants [7]. Post-transcriptional regulation by RNAi is mediated by small non-coding RNAs (~25-nucleotides; nt). Small interfering RNAs (siRNAs) are short RNA duplexes that direct the degradation of homologous transcripts [8]. In contrast, the single stranded microRNAs (miRNAs) bind to 3' untranslated regions of mRNA with complementarity of 50 to 85% to give translational repression without target degradation [9]. The mature miRNA (~25-nt) is produced by processing of ~70-nt precursor stem-loop hairpin RNAs (Pre-miRNA) by Dicer [10,11]. At the moment several human diseases, including spinal muscular atrophy (Paushkin et al., 2002), fragile X mental retardation [13,14] and chronic lymphocytic leukemia [15] have been identified as illnesses in which miRNAs or their machinery might be implicated. However, up until now there has been no clear-cut scientific proof that establishes the exact correlation between miRNAs and human infectious diseases such as AIDS.
One of the human immunodeficiency virus type 1 (HIV-1) coding accessory genes, nef, is located at the 3' end of the viral genome and partially overlaps the 3'-long terminal repeat (LTR). The nef gene is uniquely conserved in HIV-1, HIV-2 and simian immunodeficiency virus (SIV) and is not essential but important for viral replication in vivo [16]. The nef gene is expressed during HIV infection and often accounts for up to 80% of HIV-1 specific RNA transcripts during the early stages of viral replication [2]. Our own investigations have shown that defective variants of nef dsRNA containing the 3'-LTR regions, obtained from long-term non-progressor (LTNP) AIDS patients, actually inhibited the transcription of HIV-1 [17]. Furthermore, cis-expression of mutated F12-HIV-1 nef inhibits replication of highly productive NL43-HIV-1 strain, which is not related to down-regulation of CD4 [18,19]. It has been demonstrated that F12 nef gene cloned from the provirus of naturally occurring HUT-78 T cells infected with the supernatant of the peripheral blood mononuclear cells (PBMCs) of an HIV-seropositive non-producer patient, induces a block of viral replication [19]. Thus, it has been suggested that nef RNAs may be a cis-regulatory factor for HIV-1 replication [20].
In the current study, we have established the link between miRNAs and HIV infections by demonstrating that nefderived miRNAs are produced in HIV-1-infected cells. The results presented here show that nef short hairpin RNAs (shRNAs) corresponding to the nef miRNAs efficiently block RNA stability or mRNA translation, perhaps an indication that HIV-1 regulates its own replication by using nef miRNAs.

Identification of a candidate of miRNAs in HIV-1-infected cells
Very recently, the Epstein-Barr virus (EBV)-encoded miR-NAs were identified. Thus, during the preliminary stages of this study, our curiosity was fixed on the need to find out if indeed there was any relationship between nef miR-NAs and HIV-1-infected cells. To achieve this purpose, we extracted total RNA from HIV-1 IIIB strain persistently infected MT-4 T cells and northern blot analysis was performed using eight probes against the nef coding region, as shown in Figure 1A. Analyses using several anti-sense probes, small RNA molecules approximately ~25-nt in size were detected as well as HIV-1 major transcripts, 9.1, 4.3 and 1.8 kb bands (Fig. 1B). Similar results were obtained with total RNA from HIV-1 SF2 strain infected MT-4 T cells (data not shown). RNA samples treated with a mixture of the single stranded specific RNases A and T1 also generated ~25-nt RNAs that hybridized in northern blots with the sense probes against the same nef region. However HIV-1 major transcripts were not detected (data not shown), indicating that the structure of the small RNA molecules could be double-stranded RNAs (miRNAs). Some variability was observed when the quantity of the miRNAs was compared with the total of the major transcripts. A maximum of 3.2% of miRNAs was detected by using #367 probe when compared with total HIV-1 transcripts, and the minimum of 0.3% was detected by using probe# 90 (Fig. 1B).
To randomly clone the nef miRNA, ~25-nt RNAs were gel purified, cloned and sequenced. The sequences from the nef miRNA clones were 5'-acugaccuuuggauggugcuucaa-3' or similar ones, corresponded to the nucleotides approximately 420 to 443 conserved region of nef (miR-N367). The most notable feature of this analysis is that it has proven beyond reasonable doubts that nef-derived miR-NAs are produced in HIV-1 infected cells.

Inhibition of Nef by plasmids-encoding siRNA/miRNA
To examine inhibition of nef expression by the nef miRNA, we constructed eight shRNAs homologous to the native miRNA or probes used in Figure 1A [21]. Although it has been reported that three to four mutations in the sense strand derived from miRNA could have the potential to control unmutated 21-nt stem loop [22], we investigated whether the native shRNA-expressing plasmid can effectively reduce nef gene expression or not ( Fig. 1C and  1D). We used egfp or luc gene (pH1/siegfp or luc) as a Detection of HIV-1 nef miRNAs and inhibition of Nef expression by nef RNAs positive or negative control. All the shRNA-expressing plasmids including the controls were co-transfected into Jurkat T cells with pYM2.2 and cell fluorescence resulting from the expression of EGFP reporter gene was quantified by flow cytometry. The sinef176, 190, 367/miR-N367 and control siegfp all showed efficient reduction, but the sinef 007, 084, 299, 468 and 580 constructs gave only modest reductions, and no suppression was observed with si(-) and luc ( Fig. 1D and Table 1). Immunoblot analysis using anti-Nef rabbit serum also confirmed the inhibition of Nef-EGFP expression by sinef 176, 190 and 367/miR-N367 (Fig. 1E).

Inhibition of Nef expression by STYLE vector-encoding nef siRNA/miR-N367
To assess the effects of nef miR-N367 in vivo, we constructed a prototype foamy virus (PFV)-based live vector. The PFV vector expressed HIV-1 SF2 nef gene as a reporter and the STYLE vector expressed shRNA as effectors. The full-length nef gene was inserted into the bel-2 portion of a PFV clone (22) in frame to obtain pPFV/nef ( Fig. 2A). The pPFV/nef was transfected into BHK cells and treated with the histone deacetylase inhibitor, trichostatin A (TA) [23]. The viral supernatant, which contained approximately 5 × 10 6 infectious units (IFU), was collected at 72 h after transfection. For preparation of the STYLE vectors to deliver the shRNAs, the env gene portion of pSKY3.0 was replaced with the shRNA expression cassette under the control of the H1 promoter. The pSTYLE was pro-duced ( Fig. 2A), and transfected into the FFV envelopeexpressing packaging cells, CRFK sugi clone # 6, in the presence of TA. The transfected CRFKsugi clone #6 had 99% FFV Env positive cells when analyzed by flow cytometry. The viral supernatant with a titer of ca. 1 × 10 5 IFU was collected at 72 h post-transfection.
The expression of viral mRNAs and integrated DNAs from either the PFV/nef or STYLE vectors was confirmed by infection of Jurkat T cells. The mRNAs and genomic DNA were extracted from the infected cells at 2 weeks postinfection. The PFV/nef-expressed gag and nef mRNAs and the STYLE-expressed gag mRNA were detected after amplification of these regions using reverse transcription (RT)-PCR. The integration of the DNAs into the genome of Jurkat T cells was also confirmed by PCR of the LTR, gag and/or nef regions (Fig. 2B). The control SKY3.0 and PFVinfected cells were both negative for nef mRNA and integrated DNA (Fig. 2B). The integrated DNA was also detected by southern blot analysis with genomic DNA of either PFV/nef or STYLE-infected cells (data not shown). Expression of shRNAs (~22-nt) was also confirmed in STYLE-infected cells by northern blot analysis (data not shown). Expression of Nef protein in PFV/nef-infected cells was also detected with specific rabbit anti-Nef serum in immunoblots (data not shown).
To evaluate whether the STYLE encoding siRNA could inhibit the expression of the nef gene in cultured human T   (Fig. 1D and 1E) were selected for this experiment. The EGFP-positive cells were counted by flow cytometry at 48 h after transfection. Expression of Nef-EGFP fusion protein was reduced drastically following treatment with either the STYLE-176 (74 + 3.2) or 190 (51 + 4.2) and also reduced with 367/miR-N367 (32 + 2.3%). Reduction was insignificant with either the STYLE-si(-) (0 + 0.7) or STYLE-luc (7 + 0.9%) controls (Fig. 2C).

Inhibition of nef expression in human T cells by nef siRNAs
The in vitro inhibitory effects of STYLE encoding nef siRNA on HIV-1-infected cells were evaluated in Luc assays and using MT-4 T cells persistently infected with HIV-1 IIIB. Cultivation of the STYLE infected cells for 72 h followed by transfection with the pLTR SF2 and culture for another 48 h showed that STYLE-176, 190 and 367/miR-N367 all significantly (p < 0.005) suppressed Luc activity when compared to controls (Fig. 2D). HIV-1 p24 Gag was also significantly inhibited in the culture supernatant by infection with STYLE-176, 190 and 367/miR-N367 when compared to controls (p < 0.001) (Fig. 2D). These data suggested that shRNA/miR-N367 could inhibit HIV-1 transcription and replication in intact HIV-1-infected human T cells.
Jurkat T cells that had been transduced with nef shRNA for 48 h were infected with the PFV/nef. Semi-quantitative RT-PCR analysis revealed that while treatment with STYLE-190 dramatically reduced the expression of both nef and gag mRNAs of the PFV/nef, the expression of nef mRNA was also drastically suppressed by STYLE-176 and 367/miR-N367 (Fig. 2E). However the STYLE-si(-) and luc controls showed ~10% suppression of nef and ~20% suppression of gag mRNAs (Fig. 2E), which was probably a result of interference following super-infection. Nonetheless, both nef transcription and PFV/nef replication were substantially inhibited by STYLE-176, 190 and 367/miR-N367.

Inhibition of Nef expression by siRNA/miR-N367 in mice
Since different host gene products are required for siRNAmediated RNAi and miRNA-mediated translational repression with let-7 and lin-4 in C. elegans, the two RNAs may not have the same functions in vivo [24]. To test this point, we investigated the efficacy of miR-N367 using STYLE-367 in mammalian tissues. The study mice were group 1 = PFV/nef-infected (n = 6); group 2 = PFV/nef and control STYLE-luc infected (n = 6); group 3 = PFV/nef and STYLE-367-infected (n = 8); and group 4 = STYLE-367infected (n = 6). Identical study groups were used for both Balb/c and C3H/Hej mouse strains. Nef protein expressing lymphocytes were quantified by histochemical analysis using F3 Nef monoclonal antibody (mAb) or anti-Nef rabbit serum 2 days after PFV/nef infection. Nef protein was detected by immunofluoresence assay in the subcapsular area of the spleens of groups 1 or 2 Balb/c mice, but not groups 3 or 4 ( Fig. 3A and Table 2). No positive cell staining was observed using normal rabbit serum as a primary antibody (Fig. 3A). To test the expression of nef, nested RT-PCR was also done on day 2 to evaluate the degree of nef mRNA expression in the spleen, liver, adipose tissues and hematopoietic cells in groups 1-4. The nef mRNA was significantly expressed in liver and hematopoietic cells of Balb/c mice in groups 1 and 2, but not in the group 3 animals that were STYLE-367 infected ( Table  2). Tissues from group 4 did not show any nef bands after RT-PCR (data not shown).
Because extracellular Nef is internalized into human and mouse lymphocytes and macrophages [25][26][27], we examined putative Nef receptor molecule (Ner) expression with 305 mAb [27] in both mouse and human tissues by histochemical analysis. In mice, 305 mAb positive lymphocytes were detected in the subcapsular area of the spleens by immunofluoresence assay (Fig. 3B) and liver and hematopoietic cells ( Table 2) in groups 1, 2 and 3, indicating that detection of antigen by 305 mAb was not altered by Nef expression. In HIV-1 uninfected humans, the 305 mAb positive cells were detected by immunoperoxidase staining in spleen (red pulp), tonsillar follicle (germinal center), liver (Kupffer cells), salivary gland (germinal center and adipose cells), bronchi (smooth muscle cells), lung (stroma cells), thyroid gland (colloid), heart muscle (smooth muscle cells), prostate gland (smooth muscle cells), colon mucosa (intestinal absorptive and muscle cells), testis (basement membrane of tubuli seminiferi), adrenal gland (adipose cells), and brain (cerebrum cortex and cortical cells) ( Fig. 3B and Table 2).
Since Nef suppressed PPARγ expression and reduced fatty acid levels in vitro [29][30][31][32], we monitored the expression of PPARγ mRNA and body weights of mice. Significant PPARγ mRNA expression in intestinal adipose tissue of group 3, but not group 1 and 2, was detected on day 2 ( Table 2). All Balb/c mice in group 1 showed sedation and a drastic loss of weight from days 1 to 3 (day 1, p = 0.003; day 2, p = 0.021; day 3, p = 0.032 relative to mice in group 3) (Fig. 3C). Similar results were obtained in group 2 (Fig.  3C). However, group 3 mice infected with STYLE-367 did not appear to be sedated and had no drastic loss of weight (Fig. 3C). The group 4 animals, which were not infected with the PFV/nef but treated with STYLE-367, had no changes in either behavior or weight (Fig. 3C). In longitudinal examinations done during the post-infection period, the animals in groups 1 and 2 had recovered the lost weight (Fig. 3D). Similar results were obtained in group 2 from day 1 to 5 (day 1, p = 0.037; day 3, p = 0.044; day 5, p = 0.048 relative to mice in group 3) in the C3H/ In vivo effects of miR-N367 Hej mouse groups (Fig. 3D). To assess the above in vivo results, expression of nef mRNA was examined in adipose tissues ( Table 2). As shown in Table 2, although mRNAs of nef and gag were not detected in mouse adipose tissues, 305 mAb and anti-Nef rabbit serum positive staining cells were detected in mouse group 1 and 2 adipose tissues ( Fig. 3E and Table 2). Considering that the 305 mAb positive staining adipocytes appeared in mouse as well as human tissues (Fig. 3E and Table 2), these data suggest that the interaction between 305 and soluble Nef detected in adipose tissues may be responsible for the weight loss observed in mice.
In this study, whereas siRNA has been reported to inhibit hepatitis B virus replication in vivo (33)(34), our results show that nef-derived miRNAs are produced in HIV-1 infected cells, and support the possibility that miRNA and siRNA may be functionally identical, at least in a retrotransposon such as HIV. Recent studies have revealed that miRNAs and siRNAs could block mRNA expression by similar mechanisms [9] and that siRNAs could function as miRNAs [35] and EBV-encoded miRNAs were found [36].
Our results reported here are consistent with these previous observations and are suggestive of the fact that nef miR-N367 could regulate nef expression even in vivo. In our unpublished data, HIV-1 LTR promoter activity was inhibited by miR-N367 (nt number 379 to 449 of SF2 nef, 71-nt) expression, of which activity was dependent on negative responsive element (NRE) of U3 region (our unpublished data). Although no mismatch shRNA against region #367 was active, the miR-N367 from HIV-1 genome may have some mismatches and effectively inhibit HIV-1 transcription. Further the effects of siRNAs of Tc1, in particular those to the terminal inverted repeats derived from read-through transcription of entire transposable elements, were presented for silencing transposase gene expression by RNAi machinery in germ lines of C. elegans, [37]. Taken together, it could be implied from these and our other results that miRNAs produced in HIV-1-infected cells can efficiently block not  only Nef function but also HIV-1 replication through RNAi, which renders persistently low pathogenic infection latent as observed in an LTNP of 15-2-2 (see Table 1). It is equally important that although the weight loss reported here occurred only temporarily in vivo, however the inserted nef gene in the foamy retrotransposon may represent miRNAs which could inhibit nef mRNA expression by presumably an identical mechanism to that observed of siRNAs. Thus, RNAi might serve as a new sequence-specific therapeutic arsenal in AIDS prevention and possibly treatment.
Overall, our results indicate that nef shRNA transduced into T cell line inhibited HIV-1 transcription. Further, nef miRNAs could be produced from infected T cells and can block the trans-activity of Nef as well as HIV-1 replication on its own via the cis-action of nef. These functions of nef via RNAi pathways may allow persistently low pathogenic or latent infection as observed in HIV-infected non-progressors. Cumulatively, these data suggest that Nef may be involved in both viral replication and the disease progression, the findings, which may facilitate new strategies for HIV control in vivo.

Patient details
Patient selection is showed in Table 1. These SF2 (HIV-1 subtype B prototype) was included as a control nef sequence, because of the inclusion of viruses, which were also subtype B. The SF2 contained full-length nef reading frame as indicated in Table 1. Patients 1-3-3, 4-2-1 (Table  1) are rapid progressors infected with HIV-1. These patients were infected in 1984-1985 and died within 3 years of primary infection with >1 × 106 viral copies and CD4+ T cell count of 75 and 110/ml blood. Patients 15-2-2 and 16-1-1 (Table 1) are slow progressors, who were infected in 1984 and have survived HIV-1 infection with high and stable CD4+ T cell counts (690 and 760/ml blood) with low (<5000 copies) plasma viremia. All these patients acquired virus through homosexual sex. JW95-1 (Table 1) is a boy who was infected from his mother via breast feeding. The child was infected in 1983 and has survived disease free with high CD4+ T cell count (890/ml blood) with undetectable viremia. Human samples were obtained from a donor after informed consent.  [34]. The pPFV/nef (10 µg) was transfected into BHK cells and pSTYLE/si (10 µg) was transfected into CRFKsugi cells with Lipofectin Reagent. The transfected cells were cultured for 72 hr, and the viral supernatant was collected and filtered through a 0.45 µm pore size Millex-GP filter (Millipore, Bedford, MA). Vector stocks were stored at -70°C prior to use. Viral titers were determined as described previously [21]. Cells were infected with PFV/ nef and/or SKY/si at an m.o.i. of ca. 0.1 in the presence of 4 µg/ml of polybrene and infected cells were cultured at a density of 1 × 10 6 cells per ml for 3 days.

Cells and viruses
The details of plasmid constructs and the primer sequences used in cloning strategies are shown in supplementary file (see Additional file: 1).

Flow cytometric analysis
Flow cytometry was performed with a FACS Calibur (Becton Dickson, San Jose, CA) as described previously [17].

Luc assay and Immunoblotting
Firefly Luc assay was performed using the Luciferase Assay System (Promega) as described previously [17]. Immunoblotting was performed essentially as described previously by Otake et al. [28].

P24 ELISA
The concentration of p24 supernatant was determined by an antigen capture assay (Beckman Coulter, Fullerton, CA) according to the manufacturer's instructions.

Confocal laser microscopy analysis
Confocal laser microscopy analysis was performed as described previously [28].

In vivo studies and tissue analyses
Balb/c and C3H/Hej mice were raised under specific pathogen-free (SPF) conditions. Mice were infected with 1 ml of 10 5 IFU of PFV/nef and STYLE-367 by intravenous (i.v.) injection. RT-PCR analyses were performed 2 days after infection. For histological analysis, cryostat sections were prepared from both human and mouse tissues. The fixed sections were rinsed with PBS and incubated with 5% BSA for at least 1 hr to inhibit nonspecific binding of antibodies. Sections were incubated overnight at 4°C with anti-Nef rabbit serum, F3 or 305 mAb, and incubated with peroxidase or FITC conjugated secondary antibodies. The washed sections were incubated in 0.03% 3,3'diaminobenzidine (Sigma) solution in 0.05 M Tris buffer with 0.01% H 2 O 2 for development of peroxidase activity. After counterstaining with hematoxylin or methylgreen, the sections were dehydrated and mounted.

Statistical methods
Data were analysed using a one-way ANOVA analysis with a post-hoc Fisher's test. P values of 0.05 or more were determined for that of cut off.