N6-methyladenosine modification of HCV RNA genome regulates cap-independent IRES-mediated translation via YTHDC2 recognition

Significance

N6-methyladenosine (m⁶A) modification is an emerging topic of RNA biology, and its functional roles are being investigated. m⁶A modification occurs in RNAs cotranscriptionally and regulates RNA stability and translation. m⁶A methylation has been identified in RNA virus genomes and the RNA transcripts of DNA viruses. These m⁶A modifications regulate the viral life cycle in various aspects and pathogeneses associated with individual viral infections. Here, we demonstrated that m⁶A modification of the internal ribosome entry site (IRES) region of the hepatitis C virus (HCV) recruits YTHDC2, which in concert with the cellular La antigen supports the HCV IRES-mediated translation. Our study shows the functional roles of m⁶A modification and YTHDC2 in the IRES-mediated translation.

Abstract

Hepatitis C virus (HCV) infections are associated with the risk of progression to fibrosis, cirrhosis, and hepatocellular carcinoma. The HCV RNA genome is translated by an internal ribosome entry site (IRES)-dependent mechanism. The structure and function of the HCV IRES have been investigated by both biological and biophysical criteria. Recently, the role of N6-methyladenosine (m⁶A) in cellular RNA and viral transcripts has been intensely investigated. The HCV RNA genome is m⁶A-methylated, and this modification regulates the viral life cycle. In this study, we investigated the role of m⁶A modification of the HCV genome in the IRES-dependent translation function by mutating m⁶A consensus motifs (DRACH) within the IRES element in stem–loop III and IV regions and studied their effect on translation initiation. There are several DRACH motifs within the IRES element. Of these, the DRACH motif at nucleotide (nt) 329-333, located about 7 nt upstream of initiator AUG (iAUG) codon, regulates IRES-mediated translation initiation. Mutational analysis showed that m⁶A methylation of the adenosine at nt 331 is essential for the IRES-dependent translation. m⁶A reader protein YTHDC2, containing the RNA helicase domain, recognizes m⁶A-methylated adenosine at nt 331 and, in concert with the cellular La antigen, supports HCV IRES-dependent translation. The RNA helicase dead YTHDC2 (E332Q) mutant failed to stimulate HCV translation initiation. This report highlights the functional roles of m⁶A modification and YTHDC2 in the HCV IRES-dependent translation initiation, thus offering alternative therapeutic avenues to interfere with the infectious process.

Most hepatitis C virus (HCV) infections are chronic and are associated with the development of hepatocellular carcinoma (HCC) (1). HCV infections are easily controlled today by effective direct-acting antivirals, but cured individuals may still carry the risk of liver disease pathogenesis, including the development of HCC. The HCV RNA genome, which encodes a large open reading frame of about 330 amino acids, is directly translated upon its entry into the cytoplasm of infected hepatocytes (2). The translation is cap independent and is mediated by an internal ribosome entry site (IRES) element of ∼340 nucleotides (nt) that resides in the 5′-untranslated region (UTR) (3–8). HCV IRES, a complex RNA structure, has been the subject of the most rigorous and intense investigations at both the biological and structural levels (8–11). Most notable among those are the cryogenic electron microscopy, chemical probing, and NMR spectroscopy studies that defined the key structural elements and their distinct role in IRES translation (7, 8, 10–12). Several structural domains as well as long-range cis-acting elements have been implicated in the IRES-mediated translation initiation (6, 13–15). The IRES stem–loop structure that spans from nt 40 to 372 of the 5′-UTR and continues into the core-encoding sequences consists of stem–loop (SL) I to IV domains with initiator AUG (iAUG) located at nt 342 or 345 (8, 14). In addition to these SL domains, an RNA pseudoknot (PK), the tertiary structure located upstream of the initiator codon in the SL-IV domain, was defined by mutagenesis and functional data (9, 10, 16). Any perturbation of the PK and those in domains IIId and IIIe disrupted IRES-mediated translation initiation (8, 10, 16, 17). Domains IIId and IIIe of HCV IRES directly bind the 40S ribosomal subunit to initiate translation (8, 10, 17). Moreover, several noncanonical RNA binding proteins have also been reported to assist in IRES-mediated translation initiation (14, 18). Most notably, it has been reported that the human La autoantigen directly binds the IRES structure in the context of the GGAC motif located a few nt upstream of the iAUG to support HCV IRES translation (19–24). Another unique feature of the HCV IRES element is that it harbors binding sites for microRNA-122 (miR-122), which regulates RNA stability, replication, and translation (14, 25, 26).

N6-methylation of adenosine modification of RNA is an emerging topic in RNA biology, and its functional roles are continuously being investigated (27–32). N6-methyladenosine (m⁶A) is the most abundant internal RNA modification that occurs cotranscriptionally in the context of a consensus RRACH/DRACH (D = G, A, or U; R = G or A; H = A, U, or C). m⁶A methylation is reversibly catalyzed by an m⁶A “writer” (METTL3, METTL14, and WTAP complex) and “eraser” (FTO or ALKBH5) (28, 29). m⁶A “reader” (YTH domain family [YTHDF] proteins) proteins recognize m⁶A-modified RNA and regulate RNA stability, turnover, and translation (30, 31). m⁶A modification of RNA virus genome likely occurs coincidentally during RNA replication which may be viewed as a post-transcriptional event. The current m⁶A mapping approach relies on methylated RNA immunoprecipitation and sequencing analysis using m⁶A-specific antibodies (30). Generally, RNAs are highly structured, which can inhibit m⁶A site recognition by m⁶A antibodies. Also, immunoprecipitated multiple different m⁶A-containing RNA fragments generate duplicate reads during high-throughput sequencing analysis. Thus, the current m⁶A mapping techniques have limitations to quantify m⁶A status at specific sites.

Importantly, this modification is frequently identified in transcripts of DNA viruses as well as genomes of RNA viruses (32–40). The HCV genome, like other viruses, is modified by m⁶A at several sites (35). Gokhale et al. identified about 19 m⁶A-methylated regions in the HCV genome and investigated their functional relevance in the viral life cycle (35). In this study, we focused on three conserved m⁶A sites in the structured SL-III and SL-IV regions. We mutated the A residue in the three DRACH motifs, nt 154 (MT-154), 245 (MT-245), and 331 (MT-331), and investigated their functional role in translation initiation. Our results clearly identify the adenosine at nt 331 in the HCV IRES region was m⁶A methylated, and this modification dictated IRES translation. m⁶A binding protein YTHDC2, containing helicase activity known to assist in translation, specifically supported the IRES-mediated translation initiation in cooperation with the La antigen by recognition of m⁶A-methylated nt 331 adenosine of HCV RNA. On the contrary, YTHDC2 failed to recognize the MT-331 of HCV RNA and did not support their translation initiation. Further, we identified that YTHDC2 enhances HCV-dependent IRES translation by resolving the RNA secondary structure. These results highlight the significant roles of m⁶A modification and YTHDC2 in the cap-independent HCV IRES-mediated translation.

Results

m⁶A Modification of the HCV IRES Regulates HCV Translation Initiation.

Overall, m⁶A modification negatively regulates the HCV life cycle (35). Each m⁶A modification within the large HCV RNA genome may have a different function depending on its location in the HCV genome. For instance, those located in the 3′ end of the genome regulate RIG-I sensing activity (41). Similarly, the m⁶A sites within the E1 region of HCV RNA are relevant to virion maturation (35). The roles of the other m⁶A methylation sites of the HCV genome remain to be characterized. In this study, we focused on the functions of m⁶A methylation motifs located within the HCV IRES element. m⁶A peaks within the IRES element were identified between nt 275 and 340 of the HCV IRES region (35). We identified five DRACH motifs in the HCV IRES element (Fig. 1A). Of these, the three conserved motifs are located within the SL-III and SL-IV regions, known for 40S ribosome binding and eIF3 interactions (Fig. 1A and SI Appendix, Fig. S1). We generated mutants MT-154, MT-245, and MT-331, each indicating the nucleotide position of the DRACH motif, in the full-length HCV genome (JFH-1). First, we investigated whether these mutations affect viral protein expression. Huh7 cells were infected with the same amounts of HCV wild-type (WT) and m⁶A site-mutant infectious particles. Interestingly, viral protein and RNA levels were reduced in MT-331-infected cells compared to WT infection, whereas MT-154- and MT-245-infected cells did not show any notable decrease in HCV protein and RNA levels (Fig. 1B and SI Appendix, Fig. S2A). HCV genome levels of WT and m⁶A mutant-infected cells were analyzed at 16 h postinfection and were not changed to any significant levels, suggesting that MT-331 of HCV IRES did not affect viral infectivity (SI Appendix, Fig. S2B). Because the secondary structure of SL-IV could result in the helical distortion from MT-331, we generated a compensatory mutant (CM; MT-331-CM) plasmid, in which U was mutated to G to restore base pairing (Fig. 1C). Importantly, the viral protein and genome levels were not restored in HCV MT-331-CM infected cells compared to MT-331-infected cells, suggesting that the altered HCV IRES RNA secondary structure by MT-331 did not affect the viral life cycle (Fig. 1D and SI Appendix, Fig. S2 C and D). We further confirmed these results using HCV replicon systems (42). MT-331 and MT-331-CM were generated in J6/JFH-p7-Rluc2A plasmid, and these mutated HCV RNAs were transfected in Huh7 cells (Fig. 1E and SI Appendix, Fig. S2E). Keeping in with our results using HCV infection, the luciferase activities of MT-331 and the MT-331-CM HCV replicon did display a decrease in HCV IRES translation at 48 and 72 h posttransfection compared to WT, but MT-331 and MT-331-CM did not affect HCV IRES-mediated luciferase activity at 24 h posttransfection. Because in vitro transcribed HCV RNAs carry no m⁶A methylation, translation of the HCV replicon may have occurred from unmethylated RNA until 24 h. To confirm this, we performed m⁶A-methylated RNA immunoprecipitation (MeRIP; SI Appendix, Fig. S2F). Cellular genes CREBBP and HPRT represent as m⁶A-positive and -negative controls, respectively (43). In the MeRIP analysis, m⁶A modifications of the HCV genome were relatively lower at 24 h posttransfection, while m⁶A methylation levels of the HCV genome were induced at 48 h posttransfection, suggesting that HCV RNA gained m⁶A methylation during HCV replication and this modification induced HCV IRES translation after 48 and 72 h posttransfection (Fig. 1E and SI Appendix, Fig. S2F). Both HCV infection and HCV replicon transfection results demonstrate that m⁶A methylation of adenosine at nt 331 positively regulate the viral life cycle.

Fig. 1.

The m⁶A motif at nt 329 to 333, located upstream of initiator AUG, controls HCV IRES-dependent translation. (A) The proposed secondary structure of HCV IRES depicting four stem–loop regions (SL-I to SL-IV) and other structural domains. The DRACH motifs are shown in red (nt 51 to 55, 152 to 156, 183 to 187, 243 to 247, and 329 to 333). Three adenosines, nucleotides 154, 245, and 331, are mutated to cytosine. miR-122 binding sites are shown in gray, and the AUG initiator codon is shown in green. The m⁶A peaks identified from MeRIP sequencing are shown in blue. (B) Huh7 cells were infected with the indicated HCV particles. After 72 h, cellular lysates were assayed by immunoblotting (Left). HCV core protein levels were normalized with GAPDH levels using ImageJ software (Right). (C) The DRACH motif (329-333 nt) of SL-IV is shown in red. The mutation of the m⁶A site (blue; MT-331) in the SL-IV is predicted to create a bulge. The compensatory mutation was generated in MT-331 to restore the secondary structure of SL-IV. (D) The indicated HCV particles were infected in Huh7 cells. After 72 h, cells were harvested to analyze immunoblotting. (E) In vitro transcribed WT, MT-331, or MT-331-CM RNA of J6/JFH-p7-Rluc2A was transfected in Huh7 cells. The cells were harvested at 24, 48, or 72 h after transfection to analyze luciferase activity. (F) Schematic representation of a dual-luciferase reporter expressing firefly luciferase (Fluc) and Renilla luciferase (Rluc) by cap- and IRES-dependent translation, respectively (Top). Huh7 cells were transfected with the indicated dual-luciferase vectors. At 48 h posttransfection, cells were harvested to analyze luciferase activity. Rluc activity was normalized with Fluc. In B, E, and F, the error bars are the SDs of three independent experiments, each involving triplicate assays. The P values are calculated via an unpaired Student’s t test. **P < 0.01; n.s., not significant by unpaired Student’s t test.

Next, we used the dicistronic HCV IRES dual-luciferase (firefly by cap-dependent translation; Renilla by IRES-dependent translation) plasmid system to focus on the effect of m⁶A methylation on HCV IRES translation. We confirmed that HCV IRES RNAs generated from dicistronic plasmids were m⁶A modified (SI Appendix, Fig. S2G). The reduced m⁶A modification levels of MT-331 RNA were observed compared to WT, suggesting that adenosine at 331 nt of HCV IRES luciferase RNA is m⁶A methylated. Importantly, MT-331 reduced HCV IRES luciferase activity compared to WT, MT-154, and MT-245 (Fig. 1F). This result was consistent with those described above using HCV infection and replicon RNA transfection (Fig. 1 B and F), suggesting that m⁶A methylation of adenosine at nt 331 positively regulates the viral life cycle by inducing the HCV IRES translation activity.

We next investigated the role of HCV IRES mutations in an in vitro translation system using HeLa cell extracts. To synthesize m⁶A-modified RNA, N6-methyladenosine 5′ triphosphate was added to the reaction mixture during in vitro transcription as described previously (SI Appendix, Fig. S2 I and J) (44). These RNAs derived from HCV IRES luciferase vectors were subjected to in vitro translation using HeLa cell extracts as well as transfection into Huh7 cells. At the very outset, the IRES translation level was dramatically increased in m⁶A-methylated HCV IRES RNA of both in vitro (HeLa cell extracts) and cultured Huh7 cells compared to unmethylated RNAs (SI Appendix, Fig. S2 K and L). In unmethylated HCV IRES translation, MT-331 and MT-331-CM did not affect HCV IRES translation (black bars in SI Appendix, Fig. S2 K and L), further supporting that the alteration of RNA secondary structure by MT-331 did not affect HCV IRES translation activity. However, in m⁶A-methylated HCV IRES RNAs, MT-331 and MT-331-CM modestly reduced m⁶A-methylated HCV IRES translation both in HeLa cell extracts and in Huh7 cells transfected by these RNAs (blue bars in SI Appendix, Fig. S2 K and L). This reduced effect of mutations can be explained by the inefficient/random incorporation of N6-methyladenosine during in vitro transcription as the accurate m⁶A modification sites of RNAs were unknown (SI Appendix, Fig. S2 I and J) (44). Of note is the similarity in the moderate reduced HCV IRES translation level by MT-331 in in vitro synthesized HCV IRES RNAs containing m⁶A either introduced by transfection or as the template for in vitro translation using HeLa cell extracts (SI Appendix, Fig. S2 K and L), suggesting that it is the aberrant m⁶A modification of in vitro synthesized mutant RNAs that failed to reach the reduced levels seen in the HCV infection system (Fig. 1). Combined, these results, using myriad approaches and strategies, clearly point to the specific m⁶A modification at nt 331 playing a pivotal role in the HCV IRES translation initiation.

m⁶A Methyltransferases Induce HCV IRES-Mediated Translation Initiation.

The m⁶A methylation is a dynamic internal modification that is reversibly catalyzed by methyltransferases (METTL3/14) and demethylases (FTO or ALKBH5) (28). Cytoplasmic RNA genomes of RNA viruses are likely methylated by cytoplasmic methyltransferases. We next investigated whether m⁶A methyltransferases (METTL3/14) and demethylase (FTO) affect HCV IRES-mediated translation. Contrary to expectations, HCV protein levels were up-regulated in both WT and MT-331-infected cells by the silencing of METTL3/14 enzymes (Fig. 2A). Previously, Gokhale et al. showed that m⁶A modification occurs in several regions (about 19 regions) of the HCV genome and m⁶A methylation in the HCV E1-E2 region has a negative effect on the HCV life cycle (35). Thus, the results of increased viral protein levels due to the depletion of METTL3/14 could be derived from m⁶A-modified sites other than adenosine 331-nt. To confirm this, we investigated whether the silencing of METTL3/14 affects HCV RNA stability (Fig. 2B). Huh7 cells depleted with METTL3/14 were transfected with in vitro transcribed HCV GND RNA from JFH-1 GND plasmids. JFH-1 GND contains dead viral polymerase activity and therefore does not replicate. As shown in Fig. 2B, the silencing of METTL3/14 increased HCV RNA stability, suggesting that the silencing of METTL3/14 positively regulates the HCV life cycle in the full-genome context. These results imply that the HCV life cycle can be increased or decreased depending on where the m⁶A modification of the HCV genome occurs. Although the depletion of METTL3/14 induced HCV protein expression in the HCV-infected cells (Fig. 2A), in the case of dicistronic HCV IRES luciferase reporter transfected cells, the HCV IRES-mediated translation displayed a consistent decrease by the depletion of METTL3/14 in WT, MT-154, and MT-245 luciferase plasmid transfected cells (Fig. 2 C and D and SI Appendix, Fig. S3A). Importantly, HCV IRES translation activity of MT-331 was not affected by the silencing of METTL3/14, suggesting that methylation of adenosine at nt 331 is important for IRES-dependent translation initiation.

Fig. 2.

The silencing of METTL3/14 reduces HCV IRES-mediated translation. (A) Huh7 cells were transfected with control (Ctrl), METTL3/14, or FTO siRNA and infected 4 h later with WT or MT-331 HCV particles for 72 h. The indicated proteins were analyzed by immunoblotting. (B) Huh7 cells were transfected with control (Ctrl) or METTL3/14 siRNA. After 24 h, cells were transfected with HCV GND RNA. Cells were harvested at 0, 4, 8, and 12 h based on 12 h posttransfection, and relative levels of remaining HCV GND RNA were analyzed by qRT-PCR. (C and D) Huh7 cells were transfected with control (Ctrl) or METTL3/14 siRNA and transfected 4 h later with the indicated dual-luciferase plasmids for 48 h, and lysates were analyzed for luciferase activity (C) and immunoblotting (D). In B and C, the error bars represent the SDs of three independent experiments. The P values are calculated via an unpaired Student’s t test. **P < 0.01.

YTHDC2 Is Essential for HCV IRES Translation Initiation.

The m⁶A binding proteins (YTH domain family proteins) recognize m⁶A-modified cellular and viral RNAs and regulate their RNA stability and translation activity (30). Of the YTH family of proteins, only the YTHDC2 protein contains RNA helicase domain and is known to bind m⁶A motifs and affect translation by resolving structural constraints of translating RNAs (45). Here, we investigated whether YTHDC2 affects m⁶A-mediated HCV IRES translation. Interestingly, the silencing of YTHDC2 decreased viral protein levels in WT, MT-154, and MT-245 HCV-infected cells (Fig. 3A and SI Appendix, Fig. S3B). Also, the depletion of YTHDC2 decreased the HCV IRES-mediated luciferase activities of WT, MT-154, and MT-245, not affecting the cap-dependent translation (Fig. 3 B and C and SI Appendix, Fig. S3C). However, the viral protein level and luciferase activity of the MT-331 remained low regardless of the silencing of YTHDC2 proteins (Fig. 3 A and B), indicating that YTHDC2 regulates HCV IRES translation depending on the single DRACH motif (at 329-333 nt). Next, we focused on the RNA helicase activity of YTHDC2 in HCV IRES-dependent translation because it has been reported that the RNA helicase domain of YTHDC2 is essential for translation of m⁶A-modified cellular mRNA by resolving the RNA secondary structure (45, 46). We employed the helicase dead YTHDC2 (E332Q) mutant to transfect the HCV-infected cells (Fig. 3D and SI Appendix, Fig. S3D). Importantly, HCV core protein expression levels were increased in YTHDC2 WT transfected cells but were significantly decreased in the YTHDC2-E332Q transfected cells. Because decreased viral protein expression by YTHDC2-E332Q may be from different RNA recognition activities between WT and E332Q MT, we performed RNA immunoprecipitation using lysates from HCV-infected YTHDC2 WT or E332Q-expressing cells. E332Q mutation of YTHDC2 did not affect recognition of HCV RNA (SI Appendix, Fig. S3 E and F), suggesting that RNA helicase activity of YTHDC2 is essential for HCV IRES-mediated translation. We further confirmed this result using luciferase assay systems. HCV IRES-mediated luciferase activity was stimulated by YTHDC2 WT transfection, while YTHDC2 E332Q MT transfected lysates displayed reduced luciferase activity (Fig. 3 E and F and SI Appendix, Fig. S3G). In the case of MT-331, HCV IRES luciferase activity was not affected by YTHDC2 WT and E332Q MT (Fig. 3E). These results unequivocally suggest that YTHDC2 may be involved in resolving the secondary structure of the HCV IRES region via its helicase domain to support HCV IRES-mediated translation by recognition of the m⁶A-methylated adenosine at nt 331.

Fig. 3.

YTHDC2 enhances HCV IRES-mediated translation via the m⁶A-modified (DRACH) motif at nt 329 to 333, whereas helicase dead YTHDC2 mutant (E332Q) is ineffective in HCV IRES translation. (A) Control (Ctrl) or YTHDC2 siRNA was transfected into Huh7 cells. After 4 h, cells were infected with the indicated HCV infectious particles for 72 h. The indicated proteins were analyzed by immunoblotting. (B and C) Huh7 cells were transfected with Control (Ctrl) or YTHDC2 siRNA. After 4 h, cells were transfected with the indicated dual-luciferase plasmids for 48 h, before luciferase assay (B). The indicated proteins were analyzed by immunoblotting (C). (D) Huh7 cells were transfected with the control vector (Ctrl), FLAG-YTHDC2, or FLAG-YTHDC2 E332Q. After 4 h, cells were infected with HCV infectious particles for 72 h. The indicated proteins were analyzed by immunoblotting (Left). HCV core protein levels were normalized with GAPDH levels using ImageJ software (Right). (E and F) Huh7 cells were transfected with Control (Ctrl), FLAG-YTHDC2, or FLAG-YTHDC2 E332Q plasmid. After 4 h, cells were transfected with the indicated dual-luciferase plasmids for 48 h, and lysates were subjected to luciferase assay (E) and immunoblotting (F). In B, D, and E, the error bars represent the SDs of three independent experiments. The P values are calculated via an unpaired Student’s t test. **P < 0.01; n.s., not significant by unpaired Student’s t test.

We also investigated the effect of the other YTHDF proteins (YTHDF1, 2, and 3) on HCV IRES-dependent translation. The silencing of YTHDF1 to YTHDF3 proteins did not affect viral protein levels in MT-331-infected cells or WT infection (SI Appendix, Fig. S4A), and HCV IRES-mediated luciferase activity of all m⁶A mutants also was not affected by YTHDF proteins (SI Appendix, Fig. S4 B and C). These results are consistent with the study of Gokhale et al., who reported that YTHDF proteins reduced viral particle production but did not affect HCV translation and replication (35).

YTHDC2 Interacts with the La Protein in the Context of m⁶A-Methylated DRACH motif at 329-333 nt.

The role of the human La antigen in HCV translation was previously described in several reports. The La antigen helps to position the AUG start codon of the HCV genome in the entry of the 40S subunit interaction (21–24). Therefore, we investigated whether YTHDC2 interacts with the cellular La antigen to induce the HCV IRES translation. Coimmunoprecipitation experiments were performed using lysates from HCV-infected or noninfected cells. The interaction between YTHDC2 and the La antigen was observed in HCV-infected cells (Fig. 4A), whereas YTHDC2 did not interact with the La antigen in the absence of the HCV genome (SI Appendix, Fig. S5A). Because both YTHDC2 and La antigen are RNA binding proteins, the interaction between YTHDC2 and the La antigen may be caused by HCV RNA-mediated interaction. To confirm this, we analyzed the interaction between YTHDC2 and the La antigen in the presence of RNase A during HCV infection (Fig. 4B). RNase A treatment abolished YTHDC2 interaction with the La antigen, suggesting that the interaction between YTHDC2 and the La antigen is mediated by the presence of HCV RNA.

Fig. 4.

YTHDC2 interacts with La antigen in the context of the DRACH motif at 329-333 nt position in the HCV IRES element. (A) HCV-infected cells were transfected with FLAG-La plasmid for 48 h. Total cell lysates were immunoprecipitated with anti-FLAG, followed by immunoblotting for the indicated proteins. (B) HCV-infected cells were transfected with FLAG-La plasmid for 48 h. Cell lysates were immunoprecipitated with anti-FLAG in the presence or absence of RNase A, followed by immunoblotting for the indicated proteins. (C–F) Huh7 cells transfected with the FLAG-La expression vector were infected with the indicated HCV infectious particles for 72 h. Before harvesting, cells were irradiated with UV for RNA–protein cross-linking. Cellular lysates were immunoprecipitated using anti-FLAG or anti-YTHDC2 antibody. The indicated proteins were analyzed by immunoblotting (C and D). Enriched HCV genome levels were normalized by input HCV genome levels using qRT-PCR (E and F). In E and F, the error bars represent the SDs of three independent experiments. The P values are calculated via an unpaired Student’s t test. ***P < 0.001; n.s., not significant by unpaired Student’s t test.

After determining that YTHDC2 interacts with the La antigen during HCV infection, we next investigated whether m⁶A methylation at nt 331 of HCV IRES affects the interaction between YTHDC2 and the La antigen. We used the ultraviolet (UV)-mediated cross-linking method to produce RNA–protein complexes (47). Interestingly, we observed that the YTHDC2 interaction with the La antigen did not occur in the lysates from cells infected with MT-331 (Fig. 4 C and D). qRT-PCR analysis of these immunoprecipitated lysates showed that MT-331 dramatically reduced YTHDC2 interaction with HCV RNA compared to WT (Fig. 4E and SI Appendix, Fig. S5B), suggesting that m⁶A methylation of adenosine at nt 331 of HCV IRES RNA leads to interaction between YTHDC2 and La to support HCV IRES-dependent translation. However, the recognition of HCV genome by the La antigen was not affected by MT-331 (Fig. 4F), demonstrating that the La antigen binding to RNA occurs independently of m⁶A methylation at nt-331 of HCV RNA. These results together demonstrate that YTHDC2 and La bind with each other in the context of the HCV RNA genome and YTHDC2 recognition of the HCV RNA genome requires m⁶A methylation of the DRACH motif, at nt 329 to 332, to potently affect IRES-dependent translation initiation.

m⁶A Modification of the DRACH Motif at nt 329 to 333 Is Important in La-Mediated HCV Initiation.

Having found that YTHDC2 interacts with the La antigen via m⁶A methylation of nt 331 of HCV IRES SL-IV, we investigated whether the interaction between the La antigen and YTHDC2 affects HCV IRES translation activity. The silencing of the La antigen decreased HCV protein expression levels, while the ectopic La expression increased HCV core protein levels (Fig. 5 A and B and SI Appendix, Fig. S6 A and B) consistent with previous reports (21–24). Importantly, these results were not observed in MT-331 HCV-infected cells. We further investigated the effect of the La antigen on HCV IRES translation using a luciferase reporter assay. Similar results were observed in that the HCV IRES-mediated luciferase activity was induced by La antigen expression, but La antigen did not affect MT-331 HCV IRES-mediated luciferase activity (Fig. 5 C and D and SI Appendix, Fig. S6 C and D). We further analyzed whether YTHDC2 expression affects HCV IRES translation activity in the absence of the La antigen. Fig. 5 E and F show that YTHDC2 overexpression did not induce HCV core protein expression and HCV IRES luciferase activity in the La antigen-depleted cells. These data collectively suggest that the m⁶A-methylated DRACH motif at the nt 329 to 333 position recruits the YTHDC2 protein, which then assists in HCV IRES translation initiation, perhaps by interacting with the La antigen and other factors. Further, these results illustrate the notion that YTHDC2 interaction with the La antigen is important in HCV IRES-dependent translation activity.

Fig. 5.

La antigen–mediated IRES translation is dependent upon the DRACH motif at 329-333 nt position in the HCV IRES element. (A) The La antigen–depleted Huh7 cells were infected with the indicated HCV infectious particles for 72 h, prior to immunoblotting. (B) The FLAG-La plasmid transfected Huh7 cells were infected with the indicated HCV infectious particles and analyzed 72 h later by immunoblotting assays. (C) Huh7 cells treated with control (Ctrl) or La antigen siRNA were transfected with the indicated dual-luciferase plasmids for 48 h, before luciferase assay. Rluc activity was normalized with Fluc (Top), and the indicated proteins were analyzed by immunoblotting (Bottom). (D) Huh7 cells treated with FLAG-La or control plasmid were transfected with the indicated dual-luciferase plasmids. After 72 h, cells were harvested to assess luciferase assay (Top). The indicated proteins were analyzed by immunoblotting assays (Bottom). (E) HCV-infected cells were transfected with FLAG-YTHDC2 plasmids or La siRNA for 48 h, prior to immunoblotting. (F) FLAG-YTHDC2 plasmid or La siRNA was transfected into Huh7 cells. After 12 h, the cells were transfected with HCV IRES luciferase vector for 48 h, before luciferase assay (Left) and immunoblotting (Right). In C, D, and F, the error bars represent the SDs of three independent experiments. The P values are calculated via an unpaired Student’s t test. *P < 0.05; **P < 0.01; n.s., not significant by unpaired Student’s t test.

Discussion

Regulatory functional roles of m⁶A modifications have increasingly become an intense focus of RNA biology. They include development, differentiation, sex determination, obesity, metabolism, stress, viral infections, and cancer (27–29, 32, 48). We previously defined several functional consequences of m⁶A modifications of hepatitis B (HBV) and HCV RNAs (33, 34, 41, 49, 50). In this study, we focused on the role of m⁶A modification in HCV IRES-dependent translation. The HCV IRES element harbors critical structural motifs termed domains IIId and IIIe and the tertiary RNA PK structure to regulate cap-independent translation. Our results show that the DRACH motif located at nt 329 to 333, in the vicinity of the PK structure, is m⁶A methylated, and this modification is distinctly recognized by YTHDC2, which in concert with the human La antigen mediates HCV IRES translation initiation. Generally, m⁶A-modified RNAs are recognized by m⁶A reader YTH domain family proteins, which regulate RNA stability and translation (30, 31). YTHDC2 is the only member of the YTH domain family protein which contains the helicase domain (46). A previous report showed that YTHDC2 is highly expressed in the testis and has a critical role in spermatogenesis (51). However, YTHDC2 protein is expressed in many other human tissues, which suggests that the role of YTHDC2 is not limited to germ cells. Recently, the role of YTHDC2 in m⁶A-methylated mRNA translation has emerged (45, 46, 51). The recognition of m⁶A-methylated mRNA by YTHDC2 promotes translation of a small subset of mRNAs and this activation requires RNA helicase activity of YTHDC2. In this respect, YTHDC2 most likely helps to resolve these structures to promote translation initiation because the HCV IRES region is highly structured. The La antigen and YTHDC2 do not interact with each other in the absence of HCV RNA (Fig. 3A), while in HCV-infected cells, YTHDC2 binds to the La antigen via m⁶A-methylated adenosine at nt 331 of the HCV genome. (Fig. 3B). Because the La antigen was shown previously to regulate IRES-mediated translation of both viral and cellular IRES elements (19–23), m⁶A modification of cellular IRES may help this activity via recruitment of YTHDC2. Interaction between the La antigen and YTHDC2 could be one of the mechanisms accelerating IRES translation by m⁶A modification. The role of m⁶A modification in 5′ cap-independent translation was described (44), but this study described the specific role of m⁶A modification in the cap-independent IRES-mediated translation and further identified the required recognition of YTHDC2 RNA binding protein in mediating this stimulatory activity. In addition to its role in m⁶A-modified mRNA translation, YTHDC2 is also known to be involved in the degradation of mRNA. YTHDC2 interacts with the 5′ to 3′ exonuclease XRN1 to accelerate mRNA decay (46). However, 5′-UTR of the HCV genome is protected by miR-122 from the exonuclease activity of XRN1 (26). It is conceivable that HCV has an effective translation strategy to use only the helicase function of YTHDC2 for increasing translation and precludes the RNA decay ability of YTHDC2-associated XRN1 by the resident miR-122 on the IRES element.

Overall, about 0.6% of adenosines in cellular RNAs are m⁶A methylated, with an average of two or three m⁶A sites per transcript, and this modification ratio makes m⁶A the most abundant modification in eukaryotic mRNA (52, 53). However, it is important to be aware that the status of m⁶A at a given transcript is low in many cellular RNAs. In the case of the HCV genome, ∼0.16% of adenosines are m⁶A methylated, which indicates that only 6% of the total HCV genome gains m⁶A methylation (35). It was similarly observed with other viral RNAs, such as HIV, Zika, and HBV (33, 35, 36, 38). Although the 1 to 5% m⁶A methylation ratio in viral and host RNAs is not high, this level of m⁶A appears sufficient to regulate their dynamics. Because m⁶A demethylases can convert methylated adenosine to adenosine in viral and cellular RNAs, m⁶A modification represents a reversible dynamic process (33–36, 54). Notably, cellular mRNA abundance is not correlated with protein expression, which suggests that only a portion of mRNA participates in translation (55). Also, viral replication is not accompanied by all viral genomes. Thus, the 1 to 5% ratio of m⁶A can affect the fate of viral RNAs as well as cellular RNAs. It is realized that current computational methods have limitations, in that they do not adequately represent transcript abundance in mapping m⁶A sites and therefore provide different results on m⁶A occupancy.

Since the RNA-dependent RNA polymerase of HCV has no proofreading function, the HCV genome is easily mutated and rapidly evolves during the HCV infections as quasispecies (56). In this respect, the maintenance of the conserved m⁶A site(s) in the HCV genome gives an evolutionary advantage to the virus because m⁶A modification may either induce or reduce the HCV life cycle to establish persistent infection. The m⁶A modification of the HCV genome has diverse functions in the viral life cycle. For instance, the m⁶A methylation of the E1 region inhibits HCV particle production (35), the m⁶A motif of the NS5B region from nt 8,764 to 8,788 disrupts RIG-I sensing activity (41), and as shown here, m⁶A modification of the IRES SL-IV region increases HCV IRES translation. However, the role of the other m⁶A sites remains to be defined. Gokhale et al. showed that all m⁶A-modified motifs of HCV RNA are not recognized by YTHDF proteins and ∼50% of YTHDF protein binding sites identified on HCV RNA overlapped with m⁶A peaks of the HCV genome. These results suggest that the other cytoplasmic m⁶A reader proteins may be involved in the HCV life cycle and YTHDF proteins may regulate the HCV life cycle in an m⁶A-independent manner. Many m⁶A reader proteins are known to interact with RNA binding proteins to regulate RNA functions (50, 57). Thus, identifying m⁶A reader protein–protein interactions enriched during HCV will be important in the future to understand the role of m⁶A modifications in the viral life cycle.

In the development of HCV therapeutics targeting host factors, our research may provide host target candidates. Despite the remarkable efficacy (about 98%) of multiple direct-acting antivirals in eliminating the viral infection, the risk of HCC development and treatment failure in a subset of patients still remains a challenge (58). For this reason, therapies designed to target host factors offer future novel opportunities in treating liver disease pathogenesis.

These studies describe the unique role of a single DRACH motif located next to an iAUG codon near the tertiary/secondary structural elements of HCV IRES in cap-independent translation initiation. Furthermore, the key role is played by an m⁶A binding protein YTHDC2 whose specific recognition is needed to carry out this function with the aid of the human La autoantigen.

Materials and Methods

Plasmids, Antibodies, and Reagents.

HCV pJFH-1 and GND plasmids were obtained from Dr. Takaji Wakita (National Institute of Infectious Diseases, Tokyo, Japan). MT-154, MT-245, MT-331, and MT-331-CM of pJFH-1 plasmids were generated by site-directed mutagenesis. J6-JFH-p7-Rluc2A plasmid was a kind gift from Dr. Charles M. Rice (Rockefeller University, New York, NY) (42). MT-331 and MT-331-CM of J6-JFH-p7-Rluc2A plasmids were established by site-directed mutagenesis. FLAG-La (pLJM60-Ssb) and HCV dual-luciferase (pFR_HCV_xb) plasmids were purchased from Addgene. The FLAG sequences were inserted in each pcDNA3.1 YTHDC2 or E332Q MT plasmid obtained from Dr. Shu-Bing Qian (Cornell University, Ithaca, NY) using recombinational cloning methods. Antibodies were obtained from the following suppliers: anti-HCV core (ab2740), anti-HCV NS5A (ab13833), and anti-FTO (sb92821) antibodies from Abcam; anti-YTHDF3 (SC-379119) and anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH; SC-47724) antibodies from Santa Cruz Biotechnology; anti-METTL3 (15073-1-AP) antibody from Proteintech Group; anti-METTL14 (HPA038002) antibody from Sigma-Aldrich; anti-La antigen (2081), anti-YTHDF1 (86463), anti-YTHDF2 (80014), anti-YTHDC2 (35440), and anti-FLAG (14793) antibodies from Cell Signaling Technology; and anti-m⁶A antibody from Synaptic Systems. The anti-HCV core and NS5A antibodies were diluted to a 1:500 ratio in 2.5% bovine serum albumin (BSA) buffer, and the other antibodies were diluted to a 1:1,000 ratio in 5% BSA buffer for immunoblotting. The ON-TARGET plus small interfering RNAs (siRNAs) of METTL3 (L-005170-02-0005), METTL14 (L-014169-02-0005), FTO (L-004159-01-0005), YTHDC2 (L-014220-01-0005), La (L-006877-01-005), YTHDF1 (L-018095-02-0005), YTHDF2 (L-021009-02-0005), and YTHDF3 (L-017080-01-0005) were obtained from Dharmacon.

Cell Culture and Transfection.

The Huh7 cells were obtained from ATCC. Huh7 cells seeded at 2 × 10⁶ cells in a 100 mm dish were incubated overnight and transfected with 10 μg of plasmids using Mirus TransIT-LT1. Huh7 cells were transfected with 10 μM siRNA using Lipofectamine RNAiMAX reagent (Thermo Fisher Scientific) according to the manufacturer’s protocol. In vitro transcribed HCV RNAs were electroporated in Huh7 cells using Bio-Rad Gene Pulser Xcell.

In Vitro Transcription.

Full-length HCV and J6/JFH-p7-Rluc2A RNAs were synthesized from XbaI-cut pJFH-1 and J6/JFH-p7-Rluc2A plasmids using the MEGAscript T7 transcription kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. The HCV IRES luciferase sequence containing T7 promoter amplified from HCV dual-luciferase plasmids was used for in vitro transcription. N6-methyladenosine 5′ triphosphate (Sigma-Aldrich) was added in the reaction mixture to synthesize RNA containing m⁶A.

Virus Production and Cell Infection.

The in vitro transcribed HCV genome was transfected into Huh7 cell. After 72 h, the culture medium was collected and concentrated using an Amicon ultra-centrifugal filter (Merck). For infection, Huh7 cells were incubated with the concentrated culture medium at a multiplicity of infection (MOI) of 0.5 for 4 h, and cells were washed with PBS.

Methylated RNA Immunoprecipitation.

Total RNA was incubated with anti-m⁶A antibody (Synaptic Systems) conjugated to Protein G Dynabeads (Thermo Fisher Scientific) in MeRIP buffer (50 mM Tris⋅HCl [pH 7.4], 150 mM NaCl, 1 mM ethylenediamine tetraacetic acid [EDTA], and 0.1% Nonidet P-40) containing RNase inhibitor (Thermo Fisher Scientific) at 4 °C. After 4 h, RNA–bead complexes were washed with MeRIP buffer five times. RNA was eluted from beads using MeRIP buffer containing 6.7 mM m⁶A 5′-monophosphate sodium salt (Sigma-Aldrich). Eluted RNA was isolated using TRIzol (Invitrogen).

qRT-PCR.

Total RNA and immunoprecipitated RNA were extracted using TRIzol reagent (Invitrogen). The complementary DNA (cDNA) library was synthesized from extracted RNA using the iScript cDNA Synthesis kit (Bio-Rad). The qPCR was analyzed with Ssoadvanced Universal probe supermix or Ssoadvanced SYRB supermix (Bio-Rad) using the following primers: HCV genome (forward primer, 5′-GCG​TCT​AGC​CAT​GGC​GTT​AGT​ATG​AGT​GTC-3′; reverse primer, 5′-ACC​ACA​AGG​CCT​TTC​GCG​ACC​CAA​CAC​TAC-3′; HCV TaqMan probe, 5′-FAM- CTGCGGAACCGGTGAGTACAC-BHQ-3′), GAPDH (forward primer, 5′-AAC​CTG-​CCA​AGT​ACG​ATG​ACA​TC-3′; reverse primer, 5′-GTA​GCC​CAG​GAT​GCC​CTT​GA-3′; GAPDH TaqMan probe, 5′-TCC​GAC​GCC​TGC​TTC​ACC​ACC​TTC​T-3′), and HCV IRES luciferase RNA (forward primer, 5′-GTA​CAC​CGG​AAT​TGC​CAG​GA-3′; reverse primer, 5′-TGC​GGA​CCA​GTT​ATC​ATC​CG-3′). Relative HCV genome levels were normalized to GAPDH levels using the delta-delta Ct method.

Western Blotting and Immunoprecipitation.

The Nonidet P-40 lysis buffer (1% Nonidet P-40, 50 mM Tris⋅HCl, pH 8.0, 150 mM NaCl) supplemented with a protease inhibitor and a phosphatase inhibitor (Thermo Fisher Scientific) was placed into cell pellets. After centrifugation, clarified lysates were quantified using the DC protein assay (Bio-Rad). For immunoprecipitation, extracted cell lysates were incubated with Anti-FLAG M2 Magnetic Beads (Sigma-Aldrich) for 2 h on a rotator at 4 °C. For some experiments (as indicated in the text), 0.05 μg/μL RNase A (Sigma-Aldrich) was added to the cell extract and incubated at room temperature for 30 min before the immunoprecipitation. Immunoprecipitated and extracted cell lysates were loaded in sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS/PAGE) and transferred to a nitrocellulose membrane (Bio-Rad).

RNA–Protein UV Cross-Linking Assay.

Before harvesting, plated cells were washed with ice-cold PBS and then exposed to 245 nm UV light for 250 mJ/cm. Ice-cold PBS was added to a culture plate, and cells were harvested. Cell pellets were lysed with SDS lysis buffer (0.5% SDS, 50 mM Tris⋅HCl, pH 6.8, 1 mM EDTA, 1 mM DTT, 150 mM NaCl) supplemented with a protease inhibitor and RNase inhibitor (Thermo Fisher Scientific). Purified lysates were immunoprecipitated with Anti-FLAG M2 Magnetic Beads (Sigma-Aldrich) or a YTHDC2 antibody–protein G bead complex for 2 h on a rotator at 4 °C. The bead complex was washed with Nonidet P-40 lysis buffer five times. After the wash step, beads were resuspended in Nonidet P-40 lysis buffer. The immunoprecipitated lysates were extracted from half of the bead mixture. The other half of the bead mixture was treated with proteinase K for 90 min at 37 °C. Immunoprecipitated RNA was extracted using TRIzol (Invitrogen).

In Vitro Translation.

One microgram of in vitro transcribed HCV IRES luciferase RNA was used for in vitro translation using a 1-Step Human High-Yield Mini IVT Kit or Retic Lysate IVT Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. The in vitro translation activity was analyzed by luciferase assay.

Dual-Luciferase Assay.

Huh7 cells seeded at 3 × 10⁵ cells/well in a six-well plate were incubated overnight and transfected with 1 μg HCV dual-luciferase vector or in vitro transcribed J6/JFH-p7-Rluc2A. Cells were lysed with Glo Lysis buffer (Promega). Extracted cell lysates were used for dual-luciferase assay using the Dual-Glo luciferase assay system (Promega) according to the manufacturer’s protocol.

Data Availability

All study data are included in the article and SI Appendix.