Enhanced Virus Translation Enables miR-122-Independent Hepatitis C Virus Propagation

ABSTRACT The 5′ untranslated region (UTR) of the hepatitis C virus (HCV) genome forms RNA structures that regulate virus replication and translation. The region contains an internal ribosomal entry site (IRES) and a 5′-terminal region. Binding of the liver-specific microRNA (miRNA) miR-122 to two binding sites in the 5′-terminal region regulates viral replication, translation, and genome stability and is essential for efficient virus replication, but its precise mechanism of action is still unresolved. A current hypothesis is that miR-122 binding stimulates viral translation by facilitating the viral 5′ UTR to form the translationally active HCV IRES RNA structure. While miR-122 is essential for detectable replication of wild-type HCV genomes in cell culture, several viral variants with 5′ UTR mutations exhibit low-level replication in the absence of miR-122. We show that HCV mutants capable of replicating independently of miR-122 display an enhanced translation phenotype that correlates with their ability to replicate independently of miR-122. Further, we provide evidence that translation regulation is the major role for miR-122 and show that miR-122-independent HCV replication can be rescued to miR-122-dependent levels by the combined impacts of 5′ UTR mutations that stimulate translation and by stabilizing the viral genome by knockdown of host exonucleases and phosphatases that degrade the genome. Finally, we show that HCV mutants capable of replicating independently of miR-122 also replicate independently of other microRNAs generated by the canonical miRNA synthesis pathway. Thus, we provide a model suggesting that translation stimulation and genome stabilization are the primary roles for miR-122 in promoting HCV. IMPORTANCE The unusual and essential role of miR-122 in promoting HCV propagation is incompletely understood. To better understand its role, we have analyzed HCV mutants capable of replicating independently of miR-122. Our data show that the ability of viruses to replicate independently of miR-122 correlates with enhanced virus translation but that genome stabilization is required to restore efficient HCV replication. This suggests that viruses must gain both abilities to escape the need for miR-122 and impacts the possibility that HCV can evolve to replicate outside the liver.

. Thus, the extreme 59 end of the viral genome is a multifunctional regulatory sequence that harbors structures required for both viral replication and translation.
An emerging hypothesis suggests a role for miR-122 as an RNA chaperone that modulates the structure of the 59 UTR and the activity of the HCV IRES (10,11,28,29). RNA structure predictions suggest that nucleotides 1 to 42 preceding viral IRES may modify the thermodynamics of the RNA IRES structure and that annealing of miR-122 may shift the thermodynamics to favor the formation of the active IRES RNA structure (Fig. 1). While this hypothesis remains to be confirmed using biophysical or biochemical methods, indirect support stems from the analysis of HCV mutants capable of miR-122-independent replication (10,11). The first model of miR-122-independent replication was a bicistronic subgenomic RNA that contains an encephalomyocarditis virus (EMCV) IRES and suggested that altered translation regulation alleviates the need for miR-122 for viral propagation (30). Moreover, full-length HCV genomes having nucleotide substitutions or deletions in the 59-terminal region have been found that exhibit miR-122-independent replication. While replication is detectable, all of the models of miR-122-independent replication support relatively inefficient replication and are still dependent on miR-122 for wild-type (WT) levels of replication (10,(31)(32)(33)(34)(35).  Mutations SLI SLII SLI Wild-type U4C/G28A/C37U HCV SLII miR-122 The mechanisms of miR-122-independent replication of these HCV variants are still unknown. Prediction software suggests that the 59 UTR RNA of the mutants has a greater propensity to form the active HCV IRES even in the absence of miR-122 (10, 11) (Fig. 1), but a report has also highlighted the possibility of binding of other microRNAs to the 59 UTR of the viral genome (36).

HCV 5'UTR
In this study, we have developed new mutant HCV variants capable of replicating in the absence of miR-122 and used these and previously published mutants to investigate the mechanism of miR-122-independent replication. Our results suggest that mutants capable of miR-122-independent replication display enhanced translation activity compared to the wild-type HCV in the absence of miR-122 and that the translation efficiency is linked with miR-122-independent replication efficiency. We also characterized the relative contributions of translation and genome stability on miR-122-independent HCV replication and found that translation stimulation is essential for HCV propagation and that enhanced translation and genome stabilization can rescue wild-type levels of HCV replication in the absence of miR-122. Finally, we confirmed that HCV mutants capable of replicating independently of miR-122 also replicate independently of other miRNAs generated by Drosha and the canonical miRNA biogenesis pathways. Thus, our data support the notion that miR-122 has roles in both viral translation regulation and genome stabilization and suggest that these two mechanisms of action are sufficient to support wild-type levels of replication.

RESULTS
Some HCV genomes having point mutations across the miR-122 binding sites can replicate independently of miR-122. Several studies have used viruses having mutations to the miR-122 binding sites to investigate the impact of miR-122 on virus replication (24,36,37). For example, we and others have used S1 position 3 (S1p3) and S1p4 (C26G and U25A) mutants to study the impact of abolishing miR-122 annealing. In addition, other studies have used an array of miR-122 binding site mutants (S2p1, -p2, -p3, -p4, and -p5) to determine the impact of miR-122 annealing at each nucleotide (37) and on the ability of viruses to replicate and sometimes revert (38). However, these mutants have not been assessed in transient replication assays for their impact on miR-122-dependent and miR-122-independent replication. In light of more recent evidence that point mutations at miR-122 binding sites facilitate miR-122-independent replication of HCV, we wanted to revisit these mutants to assess whether they are capable of miR-122-independent replication (10,31). To investigate whether mutations in the miR-122 binding sites alter virus translation or allow the virus to replicate in the absence of miR-122, we mutated each nucleotide of both the miR-122 binding sites individually to their complementary nucleotide and tested for virus replication in the absence of miR-122 in Huh 7.5 miR-122 knockout (KO) cells. Altogether, we created 12 mutants: 6 for site 1, S1p2 to S1p7, and 6 for site 2, and S2p2 to S2p7 ( Fig. 2A). Replication of these mutants was also tested in the presence of miR-122 to verify their replication efficiency when one miR-122 site was bound. We first noticed that C26G, also known as S1p3 (24), allows for miR-122-independent replication. This mutation has been used frequently to model abolition of miR-122 binding to S1, but our data indicate an additional impact on the miR-122-independent replication ability of this genome and suggest that additional interpretation might be necessary for experiments using this mutant virus. In addition, U25A (S1p4) and A38U (S2p6) also induced miR-122-independent replication. That U25A stimulated miR-122-independent replication was interesting since U25C had previously been shown to also have this effect (31). This prompted us to also test U25G, and we observed that it also induced miR-122-indepenedent replication. Thus, any nucleotide at position 25 except the wild-type U allows for miR-122-independent replication.
59 UTR mutations that induce miR-122-independent replication also stimulate viral translation. To test our hypothesis that 59 UTR mutations that enable miR-122-independent replication would enhance translation efficiency compared to that of the wild-type virus, we compared the translation efficiency of mutant genomes capable or incapable of miR-122-independent replication with that of wild-type viral genomes in Huh 7.5 miR-122 KO cells. We assessed mutant genomes from this ( Fig. 2A) and other (10,31,34) studies. To focus on virus translation in this assay, we assessed Renilla luciferase (RLuc) expression from replication-defective reporter virus mutants (Rluc GNN) and a firefly luciferase (Fluc) mRNA (T7 mRNA) as an internal control. The Renilla-to-firefly luciferase ratio was measured to calculate the translation efficiency and is expressed as percent luciferase expression relative to the translation by wild-type viral RNA (Fig. 4A).
In support of our hypothesis, the translation assays showed that HCV mutants capable of replicating independently of miR-122 displayed higher translation efficiency than wild-type virus and mutants incapable of miR-122-independent replication (Fig. 4B). The level of enhanced translation also correlated with the efficiency of miR-122-independent replication in Huh 7.5 miR-122 KO cells (r = 0.9; P , 0.001) (Fig. 4C). For example, the U4C/G28U/ C37U and HCV-S2-GGCGUG mutants exhibited the highest translation efficiency among the mutants and the highest miR-122-independent replication efficiency, mutants with intermediate translation efficiency (G28A, G28del, and G28C) and the U25C mutant exhibited intermediate miR-122-independent replication, and mutants that exhibited lower translation than other mutants (C26del, G28U, and C30U/A34G) replicated poorly in the absence of miR-122. Finally, the translation efficiencies of mutants incapable of replicating independently of miR-122 (A34G and G33C) were similar to that of wild-type virus. The result supports our hypothesis that enhanced translation efficiency by 59 UTR mutant viruses allows HCV to replicate independently of miR-122 and suggest a role for miR-122 in viral translation regulation. In addition, mutation of miR-122 binding site 2 from CACUCC to GGCGUG (HCV-S2-GGCGUG) (34) and CACUCC to GUGAGG (HCV-S2-GGCGUG) also enhanced translation Wild-type Wild-type +miR-122

(U25C)
AC CU  FIG 3 Structure prediction models of HCV J6/JFH-1 (p7Rluc2a) wild type and miR-122 binding site 1 and site 2 mutants. The sequences from nucleotides 1 to 117 of wild-type HCV and these mutants were analyzed using the "RNA structure" online tool, and the 4 lowest free-energy structure models are presented. These structures are prediction models and have not been experimentally validated.  ability compared with that of the virus having a wild-type 59 UTR (Fig. 4B), suggesting that the S2 region modulates the ability of the 59 UTR to regulate virus translation and virus dependence on miR-122.
Mutations do not enhance virus translation and replication when miR-122 is bound. To exclude the possibility that the mutations that allow for miR-122-independent replication simply enhance overall virus translation and replication ability regardless of miR-122 binding, we assessed the replication and translation of a set of mutants with and without miR-122. We expected that the mutations would primarily impact translation and replication in the absence of miR-122 and have little effect in the presence of miR-122. For these experiments, we chose viruses with mutations outside the miR-122 binding sites so that miR-122 annealing efficiency would be consistent and not a confounding factor. In Fig. 5A we show the replication and translation of the G28 mutants in the presence and absence of miR-122. We also included a wild-type control and the G33C and A34G mutants, which did not exhibit miR-122-independent replication. In the absence of miR-122, translation is correlated with replication efficiency (r = 0.786; P . 0.033), but in the presence of miR-122, translation and replication do not correlate (r = 0.078). Thus, the mutations enabling miR-122-independent replication do not impact the general translation and replication abilities of the viral RNAs but specifically impact translation and replication when miR-122 is absent and thus modulate the reliance on miR-122. HCV RNA translation and replication in the presence of miR-122 was analyzed using nonparametric Spearman correlation. All data are presented as the averages of three or more independent experiments. Error bars indicate the standard deviations of the means, and asterisks indicate significant differences. The significance was determined by using t-distribution. ns, not significant.
but the relative contribution of each role to overall HCV propagation is still unknown. In this study, we have correlated translation efficiency with the ability of mutant viral RNAs to replicate independently of miR-122. However, mutation-induced miR-122-independent replication is still 10-fold lower than miR-122-dependent replication. We hypothesize that lower miR-122-independent replication efficiency is due to lack of genome stabilization by miR-122 annealing. To test this, we analyzed whether genome stabilization by other means will enhance miR-122-independent replication of the mutant genomes. These experiments also allowed for separate analyses of the relative contributions of miR-122-induced genome stabilization and genome translation to genome replication. To stabilize the genome, we knocked down cellular RNA-degrading enzymes, pyrophosphatases DOM3Z and DUSP11 and exonuclease XRN1, shown previously to enhance miR-122-independent HCV replication (25,42). To assess the impact of genome stabilization on miR-122-independent replication of full-length viral genomes, we assessed rescued miR-122-independent replication of the wild-type versus two mutant HCV genomes (U4C/G28A/C37U and U25C) (Fig. 6A) following knockdown of XRN1, DOM3Z, and DUSP11. Protein knockdown was confirmed by Western blotting (Fig. 6B). Our results showed that knockdown of XRN1, DOM3Z, and DUSP11 rescued miR-122-independent replication of the U4C/G28A/C37U and U25C mutants to levels similar to those seen in the presence of miR-122 (Fig. 6C, mutant HCV1miControl1siDOM3Z, siXRN1, and siDUSP11 versus mutant HCV1miR-1221siControl). miR-122-independent replication of the wild-type virus was also enhanced by knockdown of XRN1, DOM3Z, and DUSP11 but did not reach miR-122-dependent levels (Fig. 6C, wild-type HCV1miControl1siXRN1, siDUSP11, and siDOM3Z versus wild-type HCV1miR-1221siControl). From this experiment, we suggest that enhanced translation derived from 59-terminal point mutations can completely rescue viral replication to miR-122-dependent levels when the RNA is stabilized, and thus, enhanced translation and genome stabilization can completely compensate for the role of miR-122 in the viral life cycle. We also used this assay to attempt to separate the relative impact of miR-122 on viral translation and genome stability. Knockdown of XRN1, DOM3Z, and DUSP11, which stabilizes the viral genome, increased miR-122-independent luciferase expression of wild-type, U4C/G28A/C37U, and U25C viruses almost 10-fold compared to luciferase levels without knockdown (Fig. 6D). Thus, we conclude that miR-122-induced RNA stabilization accounts for about a 10-fold increase in replication. In addition, we assessed the relative impact of translation stimulation on HCV replication by comparing luciferase levels from the wild type versus mutants in DOM3Z, XRN1, and DUSP11 knockdown cells (Fig. 6D). In this experiment, we eliminated the need for miR-122 in viral genome stabilization by knocking down DOM3Z, XRN1, and DUSP11, thus allowing the experiment to isolate the impact of the translation-enhancing mutations on virus replication. In this case, the mutant viruses replicated about 100-fold greater than wild-type virus, and we conclude that miR-122induced translation stimulation accounts for about a 100-fold increase in HCV replication. While we acknowledge that enhanced genome stabilization will also indirectly affect virus protein expression levels, our results suggest that translation stimulation is the primary mechanism by which miR-122 promotes HCV propagation and that genome stabilization is an important but less potent mechanism. That enhanced translation and genome stabilization rescues can completely compensate for the absence of miR-122 also suggests that there may be no other roles for miR-122 in promoting HCV replication.
Analysis of miR-122-Independent HCV Replication Journal of Virology independent small RNAs to HCV replication in miR-122 KO cells. However, the fact that miR-122-independent replication levels were similar between Drosha KO and miR-122 KO cells suggests that any impact is minor.

DISCUSSION
In this study, we investigated miR-122-independent replication of HCV in an effort to understand the role of miR-122 in promoting the HCV life cycle. We have used several virus mutants having 59 UTR mutations reported by us and others that support miR-122-independent replication, and our analyses suggest that translation stimulation may be the primary role for miR-122 in promoting the virus life cycle and that genome stabilization has a less potent impact but is still important. Our data also suggest that miR-122 does not have other roles in virus replication since enhancing translation and genome stabilization by other methods reinstates miR-122-dependent levels of HCV replication in the absence of miR-122 for some mutant virus genomes.
In this study, we have explored the translation efficiency of several mutants that replicate independently of miR-122 and found that all mutant viruses capable of miR-122-independent replication also had enhanced translation abilities and that the translation strength of the mutant genomes correlated with miR-122-independent replication abilities (Fig. 4). This suggests that translation stimulation by miR-122 contributes to HCV replication promotion, and mutants with enhanced translation efficiency can compensate for the lack of miR-122. This supports our previous finding that altered translation regulation may compensate for miR-122's function in viral propagation (30) based on miR-122-independent replication of bicistronic JFH-1 subgenomic replicon (SGR) RNAs in which viral protein expression is driven by the EMCV IRES instead of the HCV IRES. Translation stimulation was proposed several years ago as a mechanism by which miR-122 promotes the HCV life cycle, but the small stimulation of translation observed, approximately 2-fold, appeared insufficient to account for the dramatic effect on virus replication. However, a recent study on translation and replication dynamics of picornavirus (positive-strand RNA virus) suggested that the efficiency of initial viral translation and production of a sufficient amount of protein to initiate genome replication are important to establish an infection, and in cases in which there is an insufficient translation and failed replication, the virus reinitiates translation for another attempt to establish an infection (44). The study also suggested that viral translation occurs at multiples phases of the viral life cycle. Hence, the seemingly small translation stimulation by miR-122 as detected in translation assays could have a significant and compounding role throughout the viral life cycle.
A second proposed mechanism by which miR-122 promotes the HCV life cycle is by stabilizing the viral genome, and our results indicate that miR-122-induced genome stabilization has a less potent impact than translation stimulation but is still important for efficient virus propagation. While we were not able to directly compare the stabilities of the mutant viral RNAs in cells, we indirectly assessed the contribution of genome stability to mutant RNA replication by analyzing it in the absence of cellular RNA degradation proteins. We found that miR-122-independent replication of mutant viruses was rescued to miR-122-dependent levels by knockdown of exonuclease XRN1 and pyrophosphatases DOM3Z and DUSP11, suggesting that genome stabilization allows mutants genomes having mutations that enhanced translation to replicate to miR-122-dependent levels. This supports the role of miR-122 in both translation stimulation and genome stabilization and the notion that wild-type levels of miR-122-independent replication can be achieved by providing these roles using alternative methods. Further, we were also able to determine the relative contributions of translation stimulation and genome stabilization to ;100-fold and ;10-fold, respectively. Although we suggest translation as a major contributor to miR-122's function in viral propagation, and miR-122-induced genome stability is a minor but essential function in HCV propagation, we also suggest that HCV genome translation and stability are overlapping functions and one event can contribute to other. However, our data suggest a lack of an additional role of miR-122 in viral genome replication.
We and others have proposed a model that binding of miR-122 to the 59 UTR induces the formation of the translationally active RNA IRES secondary structure (10,11,28,29). The proposed model suggests that the 59 UTR RNA structure is dynamic and forms one or more translationally unfavorable structures (SLII Alt ) in the absence of miR-122 and that miR-122 annealing modifies the thermodynamics to favor the formation of the translationally active SLII structure. Based on extensive structure-function analyses and X-ray crystallography, nucleotides 40 to 372 of the HCV RNA form an IRES structure in solution (4,45). Our model proposes that the 59 terminal nucleotides 1 to 42, including the miR-122 binding sites, modulates the structure of the IRES; however, X-ray crystallography experiments were done using RNAs that lacked the 59-terminal region containing the miR-122 binding sites (4,45). We propose that 59 UTR point mutants that promote miR-122-independent replication also shift the equilibrium to favor the formation of SLII, even in the absence of miR-122; in support of this hypothesis, the mutants we identified to have enhanced translation are also predicted to favor the formation of the canonical SLII prediction structures in the absence of miR-122 (10, 28) ( Fig. 1 and 3). In contrast, however, some of the mutants, specifically the site 2 HCV-S2-GGCGUG, HCV-S2C-GUGAGG, and A38U mutants, were predicted to form other RNA structures, suggesting that the mechanism might be more complex. In addition, the model does not consider the roles of RNA binding proteins recruited or displaced by miR-122 which have been proposed to also modulate 59 UTR structure and activity (28). Thus, confirmation or refinement of the model will require biophysical analyses of the RNA structures generated by these mutants and investigation of RNA protein binding.
In the process of studying miR-122-independent replication of these mutants, we have identified one 59 UTR mutant, S2p5 (C39G), that does not replicate well even in the presence of miR-122. We speculate that this mutation may affect the structure and function of the complement strand of the 59 UTR during genome replication. A study by Friebe and Bartenschlager showed that the RNA secondary structure of the negative strand affects viral RNA replication (12). Thus, we suggest that mutations in the 59 UTR of the viral genome may affect miR-122 annealing and the secondary structure of both strands and this could limit the tolerance of the 59 UTR to point mutations. Thus, we suggest that miR-122-independent replication of HCV is dictated by the multifunctional nature of the 59 UTR and its complementary strand.
It is also interesting that while numerous HCV 59 UTR point mutants capable of miR-122-independent replication have been isolated in the lab (Fig. 2B), few are found in patients' samples. Only the G28A mutant was found in hepatic and extrahepatic tissue of patients as well as in cell culture, supporting miR-122-independent replication and suggesting evolutionary pressure to retain a dependence on miR-122, perhaps to restrict HCV replication to the liver. In addition, nonprimate hepaciviruses have at least one conserved miR-122 binding site on their genome, suggesting an evolutionary relationship to retain dependence on miR-122 (34,46).
A recent report suggests that HCV replication can be supported by other miRNAs (miR-504-3p, miR-574-5p, and miR-1236-5p) in a miR-122-like manner and by miR-25-5p and miR-4730 in a non-miR-122-like manner (36). To investigate the possible effects of these miRNAs on our miR-122-independent HCV replication models, we assessed replication of several 59 UTR mutants in Huh 7.5 Drosha KO cells, cells lacking the canonical miRNA biogenesis pathway and thus lacking the miRNAs listed above and miR-122. We found that the mutant viruses replicating in the absence of miR-122 could also replicate in Huh 7.5 Drosha KO cells, suggesting that replication is not supported by these miRNAs in our assays. However, we cannot exclude the possibility that miR-122-independent replication might be supported by small RNAs generated by noncanonical pathways such as mirtrons, snoRNA-derived miRNAs, and tRNA-derived miRNAs (47,48).
In conclusion, we have established that viral variants that replicate independently of miR-122 display enhanced translation and that genome replication in the absence of miR-122 can be completely rescued when both translation and genome stability are addressed by other means. Finally, we show evidence that translation stimulation is a major function of miR-122-induced promotion of HCV, whereas genome stability has a secondary function.