MicroRNAs (miRNAs) are naturally occurring small RNA molecules that modulate gene expression by binding to partially complementary target sites usually located in the 3'UTR of protein coding transcripts1. They have been implicated in biological functions like tissue differentiation, establishment of cell fate during development, apoptosis and oncogenesis 2345. The cellular miRNA, hsa-miR-32 has been shown to directly interfere with the replication of primate foamy virus in HeLa cells and to reduce viral RNA levels6. Another cellular miRNA, miR-122a, involved in cellular stress response and modulated by interferon beta, can also influence the susceptibility to Hepatitis C virus 789. Earlier, we had predicted sites in the HIV-1 genome that can be potentially targeted by human encoded miRNAs using consensus target prediction, and we had proposed the possibility that the cellular levels of these miRNAs may determine disease progression following HIV-1 infection10. Here we experimentally confirmed our computational predictions by demonstrating that the expression of specific cellular miRNAs can reduce target protein expression and HIV-1 replication in cultured human cells.

We used a consensus approach to predict targets in the HIV-1 genome for all human miRNAs. Five miRNAs (hsa-miR-29a, 29b, 149, 324–5p and 378) that were prioritized from the predictions by miRanda11 and further refined using RNAhybrid12, MicroInspector13, and DIANA-MicroT14, showed putative targets in four HIV-1 genes. These included two miRNAs of highly related sequence (hsa-miR-29a and hsa-miR-29b) that could potentially bind to the region coding for the accessory protein, Nef of HIV-1 (Fig. 1A). The predicted target site of hsa-miR-29a and 29b, located 407 bases into the nef transcript, is highly conserved in the sequences from all clades of HIV-1 (A, B, C, D, F and H), with the exception of the outlier group clade O (see Additional file 1). Previous studies have shown that Nef is critical to the progression of HIV-1 infection 15. Nef is expressed early during the HIV-1 life cycle; and it represses CD4, promotes the release of infectious HIV virions16 and establishes a persistent HIV-1 infection. The HIV-1 variant found in a cohort of individuals who failed to show signs of disease manifestation even after 10 years of virus infection harbored a deletion in nef. Nef deleted viruses thus fail to produce symptoms of acquired immunodeficiency syndrome (AIDS). This is consistent with recent studies that Nef is necessary for efficient viral replication and pathogenesis17. As nef overlaps with the 3' long terminal repeat (LTR), the inhibition of Nef may also lead to transcriptional repression by interference of the 3'LTR function18. Furthermore, miRNA N367 which has been shown to inhibit Nef has also been reported to downregulate the transcription and replication of HIV-1 19. The importance of nef in HIV-1 infection prompted us to experimentally test the anti-HIV-1 potential of hsa-miR-29a and 29b.

To address whether the appropriate miRNAs are expressed in cells susceptible to HIV-1 infection, we first tested for the presence of hsa-miR-29a and 29b in PBMCs isolated from healthy volunteers. We also probed for hsa-miR-29c, as this novel sequence-related miRNA has been added to the hsa-miR-29 family since our original prediction (Fig. 1B). Hwang et al. have earlier reported that miR-29c could not be detected in HeLa cells20. However, we detected mature hsa-miR-29a, b and c in PBMCs, using primer extension of species specific probes against 29a, b and c (Fig. 1C, upper panel). To identify cells suitable for overexpression of these miRNAs or their depletion using anti-miRNA molecules, we determined the endogenous levels of hsa-miR-29a, b and c in a variety of cell lines. The epithelial derived HeLa cell line has been used widely for reporter assays. We found that HeLa cells expressed 29 a, b and c miRNAs (Fig. 1C, middle panel) at levels comparable to PBMCs. HEK293T cells used extensively for HIV-1 single cycle replication studies, on the other hand, did not show detectable levels of miRNA (Fig. 1C, lower panel). However, in later experiments we could observe expression of hsa-miR-29a, b in HEK293T cells using miRNA specific qRT-PCR Taqman assays. HEK293T cells, therefore, appeared to be appropriate for artificially expressing the miRNAs and for studying anti-HIV-1 potential in single cycle replication experiments. In addition, cell line specific miRNA expression profiling studies have reported that 29 a, b and c are expressed in Jurkat cells 21.

Reporter assays using the nef target region fused to the 3'UTR of luciferase gene in pMIR-REPORT™ Luciferase (luc) were used next to test the suppressive ability of the cellular miRNAs. The reporter-target fusion construct (luc-nef) was co-transfected with a control beta-galactosidase expression plasmid into HeLa cells which naturally express hsa-miR-29a and 29b. Luciferase activity from the reporter plasmid bearing the target region showed a three fold reduction in expression (Fig. 2A) as compared to control luc-vector transfection. Luciferase RNA levels did not show a significant reduction suggesting that the repression was post-transcriptional (Fig. 2B).

Locked nucleic acid (LNA)-modified anti-miRNA molecules interfere with miRNA function in a highly specific manner22. We used LNA-modified anti-miRNA oligonucleotides against hsa-miR-29a and 29b (Fig. 2C) to "knock-down" the cellular miRNAs. Such knock-down should restore reporter activity from the luc-nef fusion construct. Indeed, co-transfection of the luc-nef reporter construct with LNA-modified anti-miR29a and b partially restored the reporter activity in HeLa cells (Fig. 2D and 2E). LNA-modified anti-miR29a was the more effective of the two, restoring reporter activity to 2.5 times that of untreated controls. LNA-modified anti-miR29a and b had no effect on the luciferase vector without the fused nef target sequence, supporting the specificity of the LNA-miRNA interactions (Fig. 2D and 2E).

We next either over-expressed or down regulated hsa-miR-29a and hsa-miR-29b to test the effects of these miRNAs on virus replication. We cloned the pre-miRNA for hsa-miR-29a and hsa-miR-29b into a mammalian expression vector, pEGFP-N3. This construction was designed to express the miRNA under the control of the CMV immediate-early promoter, as a GFP-fusion transcript (Fig. 3A). Expression of the mature hsa-miR-29a and 29b was confirmed by quantitative real-time PCR (Fig. 3B). The inserted pre-miRNA disrupted the 5' UTR region of the EGFP reporter gene. In the absence of splice sites flanking the pre-miRNA, the current model of miRNA processing23 predicts that miRNA processing would result in a reduced expression of the downstream reporter gene. Indeed, EGFP expression was drastically reduced in the miRNA expressing clones (Fig 3C). We then used these clones to over-express the miRNAs in HEK293T and Jurkat cells transfected with either a Nef expression vector or an HIV-1 molecular clone. In order to study the effect of the miRNA expressed from the EGFP vectors, we co-transfected pre-miRNA-EGFP fusion clones of hsa-miR-29a and b, along with pCDNA-HA-Nef into HEK293T cells and analyzed Nef expression by immunoblotting. As shown in Fig. 4A, both the hsa-miR-29a and 29b expressing clones reduced the expression of Nef, with hsa-miR-29a being highly active. This result indicates that hsa-miR-29a and 29b inhibit Nef expression.

To analyze the effect of these miRNAs on virus replication, we next co-transfected the miRNA-expressing clones with HIV-1 molecular clone pNL4.3 into HEK293T cells. 48 hours post-transfection, the cells were lysed for analysis of Nef expression, and the cell supernatant was used for HIV-1 p24 antigen capture ELISA to quantitatively assess virus production. Over-expression of hsa-miR-29a significantly inhibited both Nef expression and virus production, while modest inhibition was observed with hsa-miR-29b (Fig. 4B, upper and lower panels). This data clearly show that hsa-miR-29a inhibits Nef expression and viral replication in HEK293T cells. As T cells are the primary target of HIV-1, we then used human CD4+ T cell line, Jurkat, for analyzing the role of hsa-miR-29a and 29b in virus replication. Jurkat cells were nucleofected with the miRNA-expressing clones with the pNL4.3 viral clone and analyzed for Nef and p24 expression. In Jurkat T cells, both hsa-miR-29a and 29b clones significantly inhibited Nef expression (Fig. 4C, upper panel) and virus production (Fig. 4C, lower panel). These results, taken together, showed that the two human miRNAs not only inhibit Nef expression but also virus replication.

In the over-expression studies, hsa-miR-29a consistently showed higher efficacy in reducing viral replication and Nef levels. To investigate the physiological role of cell endogenous hsa-miR-29a on HIV-1 replication, we used LNA modified anti-miR29a to downregulate cellular miR29a in HEK293T cells. As shown in Figure 5A, the LNA modified anti-miR29a resulted in a three fold reduction of the cellular miR29a level. We observed a corresponding increase in virus production, monitored using HIV-1 p24 antigen capture ELISA (Fig. 5B). We used a mock LNA-modified oligonucleotide as a control to control for the specificity of the anti-miR29a oligonucleotide. Anti-miR29a transfected cells showed a two fold increase in viral levels compared to the mock LNA control.

Although Nef was initially reported to be a negative factor for HIV-1, later results from several laboratories including ours 2425 have found that Nef enhances HIV replication17. Interfering with Nef expression is expected to decrease viral replication. Thus, the findings that the inhibition of Nef by ectopic over expression of miR29a and 29b reduced viral replication and that the suppression of endogenous miR29a by anti-miR29a LNA increased viral replication are wholly consistent.

Accumulating evidence indicates that miRNAs of both viral and host origin may influence host-virus interaction in a variety of ways: as direct modulators of viral replication, as factors affecting viral susceptibility, and as indirect modulators of cellular genes that influence viral propagation 26272829. In this regard, artificial inhibition of the miRNA processing machinery using siRNAs against Dicer and Drosha has been shown to result in faster replication of HIV-1 in PBMCs30. Dicer was also shown to be important in cellular resistance to infection by Vesicular Stomatitis Virus and influenza A virus since cells with Dicer defective alleles or cells with knockdown of Dicer exhibited hypersusceptibility to infection by these viruses 313233. A recent report posited that different stages of HIV-1 progression starting with infection followed by the transition from latency to activated replication appears to be associated with discrete expression profiles of cellular miRNAs34. That study demonstrated a region common to the 3'UTRs of all HIV-1 transcripts, except nef, which is targeted by a cluster of five cellular miRNAs. These miRNAs were suggested to collectively aid in the establishment of viral latency. Based on those observations, one could reason that a panel of miRNA inhibitors might activate latent HIV-1 infection 34. Compatible with such reasoning, Wang et al. have indeed recently demonstrated that the suppression of anti-HIV-1 miRNAs in monocytes facilitates HIV-1 infectivity, while the increase in macrophages of miRNAs that target HIV-1 inhibited viral replication35. Our current results are consistent with the emerging concept that augmenting the expression of cellular anti-viral miRNAs can be a useful strategy in developing anti-HIV-1 therapeutics. In addition expression levels of natural anti-HIV miRNAs may also be useful in studying susceptibility to infection.

miRNAs are nodal molecules in an intricate network of host-virus interactions that form a chain of strategies and counter strategies developed by the virus and the host26. Taken together, the range of interactions between the HIV-1 and host cells suggests that miRNAs may be involved in fine tuning the transition from latency to activation, the clearance of latent HIV-1 reservoirs, and the reduction of virion production. Cellular miRNAs with anti-viral roles may have additional roles in host cellular functions. Anti-HIV-1 therapeutics based on the regulation of miRNA levels will have to address how these changes perturb normal cellular homeostasis.

The authors declare that they have no competing interests.

SKB, VS, MH designed the study. JA, SZK, KS, PR, AG performed experiments. BP, DM analyzed the data and SKB, BP, DM wrote the paper.