End game: Getting the most out of microRNAs

Viruses are notorious for their ability to usurp cellular pathways for their own benefit, often by exploiting host factors in new and unusual ways. Hepatitis C virus (HCV), a causative agent of chronic liver disease, is no exception. As well as using host proteins and lipids, the virus is known to enlist an abundant liver-specific microRNA, miR-122, to aid in its replication (1). MicroRNAs are short (≈22-nt) sequences that typically bind, through complementary ≈6-nt “seed” sites, to the 3′ noncoding regions (NCRs) of certain cellular mRNAs. By recruiting the RNA-induced silencing factor complex (RISC), microRNA binding generally leads to translational suppression and/or degradation of the target transcript. HCV interactions with microRNAs, however, seem to defy all conventions. miR-122 binds the viral genome at not one but two sites. These tandem target sequences are located in the 5′ NCR rather than at the 3′ end, and recruitment of miR-122 does not repress translation but enhances viral replication through as-yet-unclear mechanisms. In PNAS, Machlin et al. (2) uncover a further unique feature of the miR-122–HCV association. Their work reveals that miR-122 binds the viral genome in a complex structure that requires not only the seed sequence but also functionally important associations beyond the seed site. The complex RNA interaction seems to be indispensable for HCV replication but is not required for the actions of miR-122 on cellular mRNAs. This model opens up new potential mechanisms for miR-122 in the HCV life cycle.

In pursuit of understanding the role of miR-122 in HCV replication, Machlin et al. (2) hypothesize that specific regions of the microRNA required for viral growth can be uncovered. Dissecting the two miR-122 binding sites on the HCV RNA requires careful experimental design. By changing one critical nucleotide in site 1 or 2 of the viral genome, transfected miR-122 bearing the complementary change can be directed to a particular target, whereas the other site binds only the endogenous microRNA (3). Using this system, a series of miR-122 mutants is directed to each binding site, and the impact on viral replication is monitored by Northern blot analysis. Unexpectedly, miR-122 nucleotides 15 and 16, which are outside the canonical seed sequence (nucleotides 2–8), are shown to be critical for HCV replication when targeted to either genomic site. By engineering compensatory mutations, Machlin et al. (2) go on to demonstrate that these miR-122 nucleotides likely form direct base-pairing interactions with phylogenetically conserved HCV sequences (Fig. 1A). Most interestingly, the additional interactions of miR-122 upstream of site 1 result in binding to the extreme 5′ end of the HCV genome, converting a free terminus into an RNA duplex with 3′ overhang. Deletion or modification of the protruding region of miR-122 (nucleotides 20–23) severely impacts HCV replication. Remarkably, this 3′ tail and nucleotides 15–16 are not required for the conventional function of miR-122 as a microRNA or siRNA against a reporter transcript (Fig. 1B). These results suggest that the unusual tandem “bulge and tail” structure formed by miR-122 binding to the HCV genome might contribute to the unique outcomes of this interaction.

In terms of microRNA biology, these findings present a few surprises. Although nucleotides 13–16 of other microRNAs have been shown to supplement recognition of the seed site sequences (4), binding to the extreme 5′ end of a target RNA is unprecedented. Furthermore, the requirement for microRNA nucleotides 20–23 is quite unusual. It is possible that these nucleotides are needed to stabilize the HCV-bound microRNA, perhaps by recruiting factors that protect from 3′–5′ exonucleases or other degradation schemes (Fig. 1C). Alternatively, this sequence may form a localization element that is important for HCV replication, such as the nuclear targeting sequence identified in the 3′ end of miR-29b (5). The entire miR-122 sequence is unusually highly conserved among mammals, and swapping of sequences between diverse microRNAs may help shed light on the role of the 3′ sequences in viral and cellular functions.

More effort is needed to define the precise molecular complex at the 5′ end of the HCV genome and to understand the function of these new layers of intricacy. Although the genetic argument for the bulge and tail interaction is compelling, biochemically defining the structure may yield additional insights. The interaction with miR-122 mediated by site 1 spans the HCV RNA structure 5′-SL1 (Fig. 1A). This secondary structure has been shown to be dispensable for translation but absolutely required for RNA replication (6). The present findings provoke the question of whether 5′-SL1 functions solely to create a landing pad for miR-122 or whether it harbors additional independent functions. The former scenario is supported by the fact that the structural context of 5′-SL1 is more important than the sequence of the loop (7). Although HCV is unusual in hijacking a cellular microRNA as part of its genomic architecture, elaborate terminal structures are not uncommon among RNA viruses. Picornavirus protein Vpg is covalently attached to the 5′ terminus of the genome during RNA synthesis. Flavivirus genomes undergo cyclization through complementary motifs near the 5′ and 3′ termini. In influenza viruses, the genomic segments form panhandle structures mediated by the interaction of 5′ and 3′ sequences; interestingly, abundant virally encoded small RNAs complementary to the 5′ regions have recently been uncovered (8, 9). Terminal higher-order genomic structures are known to have essential roles in viral replication, such as initiation of transcription, translation, and regulation of RNA synthesis (10–12). Although considerable uncertainty surrounds the mechanism of action of miR-122 for HCV replication, a slight enhancement of HCV translation has been demonstrated (13, 14), and the elongation phase is known to be independent of miR-122 action (15, 16). Whether miR-122 interaction directly influences initiation of RNA synthesis and/or other HCV life cycle events is not clear (Fig. 1C).

Alternatively, or in addition to directly affecting replication, complex with miR-122 may increase the stability of the HCV genome (Fig. 1C). It is intriguing that the protruding 3′ end of miR-122 might serve to mask the 5′ terminus of the HCV RNA. Indeed, 5′ triphosphate moieties have been shown to activate the innate immune-sensing molecule RIG-I (17). Investigating HCV replication in cell lines deficient for innate immune function in the presence or absence of miR-122 would help define the role of the microRNA in genome protection. Understanding the native structure of the HCV 5′ end over the course of infection will be important in evaluating this hypothesis. Is the viral RNA already bound with miR-122 in a virus particle? Does miR-122 loading occur upon uncoating of the genome? Does miR-122 remain associated with the plus strand RNA throughout the viral life cycle? To date, experiments have been conducted by transfection of in vitro-transcribed RNA rather than through virus infection. By generating infectious viruses encoding miR-122 seed site mutation(s) via complementation in a producer cell, it would be interesting to monitor HCV infection in cell lines that express either wild-type or mutant microRNAs to begin to understand the impact of miR-122 on the incoming HCV RNA.

The role of the RISC complex in HCV–miR-122 interactions also remains to be determined. Typically, RISC is associated with microRNA-bound mRNA and mediates the cleavage or translational repression of the target. Knockdown experiments have suggested the importance of RISC factors, such as Argonaute (Ago) 1–4, for HCV replication (2, 18, 19). However, the direct involvement of Ago proteins in miR-122 function for HCV replication is still unclear (Fig. 1C). Machlin et al. (2) show that miR-122 mutants unable to support HCV replication were still able to suppress a reporter mRNA, indicating that loading of mutant miR-122 into RISC was intact. However, the effect of Ago knockdown on miR-122 function for HCV replication seems perplexing. Although Machlin et al. (2) find that exogenous miR-122 can enhance HCV replication upon knockdown of Ago 1 and 2, others have observed that the rescue of mutant HCV by exogenous miR-122 with a compensatory mutation was hindered by Ago 2 depletion (19). If Ago proteins are involved in miR-122 binding to the HCV RNA, their conventional role in suppressing and cleaving the target mRNA must be uncoupled from this interaction. Alternatively, it is possible that the Ago–miR-122–viral genome complex excludes other RISC components and/or includes unknown factors to modify its normal functions.

New insights into the complex between miR-122 and the viral genome may have implications for HCV therapeutics. Locked nucleic acid antagonists of miR-122 have recently shown promising antiviral activity in animal models of HCV (20). The identification of sequences specifically involved in HCV interactions hints at the possibility of antagonizing miR-122’s role in viral replication without compromising the regulation of host mRNA targets. Future experiments that uncover additional molecular details of this unusual virus–host interaction are eagerly awaited.