Abstract
RNA-binding proteins (RBPs) regulate cell physiology via the formation of ribonucleic-protein complexes with coding and non-coding RNAs. RBPs have multiple functions in the same cells; however, the precise mechanism through which their pleiotropic functions are determined remains unknown. In this study, we revealed the multiple inhibitory functions of heterogeneous nuclear ribonucleoprotein K (hnRNPK) for myogenic differentiation. We first identified hnRNPK as a lncRNA Myoparr binding protein. Gain- and loss-of-function experiments showed that hnRNPK repressed the expression of myogenin at the transcriptional level. The hnRNPK-binding region of Myoparr was required to repress myogenin expression. Moreover, hnRNPK repressed the expression of a set of genes coding for aminoacyl-tRNA synthetases in a Myoparr-independent manner. Mechanistically, hnRNPK regulated the eIF2α/Atf4 pathway, one branch of the intrinsic pathways of the endoplasmic reticulum sensors, in differentiating myoblasts. Thus, our findings demonstrate that hnRNPK plays lncRNA-associated and - independent multiple roles during myogenic differentiation, indicating that the analysis of lncRNA-binding proteins will be useful for elucidating both the physiological functions of lncRNAs and the multiple functions of RBPs.
Introduction
Long non-coding RNAs (lncRNAs), which are >200 nucleotides (nt) in length and which do not encode more than 100 amino acids, are emerging as important regulators in diverse biological processes, including transcription, splicing, RNA stability, and translation (Statello et al., 2021). LncRNAs are pervasively transcribed from the noncoding genomic DNA; cis-regulatory regions, including promoter and enhancer, introns, 3’ untranslated regions, and repetitive sequences (Chakraborty et al., 2014). LncRNAs are also expressed from the antisense direction of the coding genomic DNA (Hon et al., 2017). Thus, most of the genomic regions have the potential to express lncRNAs. Thus far, more than 260,000 lncRNA genes are registered in the LncBook, which is a curated knowledge-based database for human lncRNAs (Ma et al., 2019). Since lncRNAs exert their molecular functions by interacting with proteins, mRNAs, or microRNAs (Hitachi and Tsuchida, 2020), their molecular functions differ widely, depending on the interacting partners.
As lncRNA-interacting factors, RNA-binding proteins (RBPs) are essential to determine the molecular function of each lncRNA (Statello et al., 2021). In the human genome, more than 1,500 genes encode RBPs (Gerstberger et al., 2014). RBPs consist of ribonucleoprotein complexes together with lncRNAs to regulate various biological aspects. For example, the association of Ddx5/Ddx17 with lncRNAs, such as SRA, mrhl, MeXis, or Myoparr, is required to activate the expression of downstream genes (Caretti et al., 2006; Hitachi et al., 2019; Kataruka et al., 2017; Sallam et al., 2018). RBPs, including NONO, SFPQ, FUS, and RBM14, associate with Neat1, which is a highly abundant lncRNA in mammals, to form a large membrane-less structure paraspeckle in the nucleus (Hirose et al., 2019). A ubiquitously expressed RBP, known as human antigen R (HuR) associates with lncRNAs and regulates their stability; HuR increases the cytosolic linc-MD1 levels in skeletal muscle cells (Legnini et al., 2014), whereas it promotes the decay of lincRNA-p21 in HeLa cells (Yoon et al., 2012). Additionally, RBPs are also involved in RNA splicing, polyadenylation, RNA transport, and translation (Kelaini et al., 2021). The majority of RBPs are involved in multiple biological processes in concert with lncRNAs (Briata and Gherzi, 2020; Jonas et al., 2020; Nostrand et al., 2020), and mutations in genes coding for RBPs are associated with human genetic disorders (Gebauer et al., 2021).
Heterogeneous nuclear ribonucleoprotein K (hnRNPK), a member of the heterogeneous nuclear ribonucleoprotein family, has multiple roles, including chromatin remodeling, transcription, RNA splicing, and translation (Bomsztyk et al., 2004; Wang et al., 2020). hnRNPK acts together with lincRNA-p21, EWSAT1, and lncRNA-OG, to regulate the expression of downstream genes in mouse embryonic fibroblasts, Ewing sarcoma, and bone marrow-derived mesenchymal stem cells (Dimitrova et al., 2014; Howarth et al., 2014; Tang et al., 2018). However, the molecular function of hnRNPK in skeletal muscle cells has not been fully elucidated. LncRNA Myoparr is an essential regulator of skeletal muscle cell proliferation and differentiation (Hitachi et al., 2019). Myoparr shares the same promoter region with the myogenin gene and activates myogenin expression by promoting the interaction between Ddx17 and histone acetyltransferase PCAF (Hitachi et al., 2019). We previously identified hnRNPK as a candidate for Myoparr-associated protein in skeletal muscle cells (Hitachi et al., 2019), suggesting the unique functions of Myoparr-associated hnRNPK during myogenic differentiation. In the present study, we revealed the inhibitory role of hnRNPK as a Myoparr-associated protein in skeletal muscle cell differentiation. Moreover, by comparing the downstream genes regulated by Myoparr and hnRNPK, we also found a Myoparr-independent role of hnRNPK during myogenic differentiation. Our findings reveal that hnRNPK plays Myoparr-associated and -independent multiple roles in skeletal muscle cell differentiation and will contribute to elucidating the complex roles of RBPs in cell differentiation.
Materials and Methods
Cell cultures, siRNA transfection, and ISRIB treatment
A mouse myoblast cell line, C2C12, was cultured in Dulbecco’s modified Eagle medium (DMEM) supplemented with 10% fetal bovine serum at 37°C under 5% CO2. Myogenic differentiation was induced by replacing the medium with the differentiation medium, DMEM supplemented with 2% horse serum. C2C12 myoblasts were transfected with 50 nM of Stealth RNAi (Thermo Fisher Scientific, Waltham, MA, USA) using Lipofectamine 3000 (Thermo Fisher Scientific) according to the manufacturer’s protocol. The following siRNAs were used: Stealth RNAi siRNA negative control (Negative Control, Med GC, Thermo Fisher Scientific), stealth RNAi for Myoparr, and stealth RNAi siRNAs specific for hnRNPK (MSS205172 and MSS205173, Thermo Fisher Scientific). The siRNA sequences are listed in Supplemental Table 1. At 24 h after siRNA transfection, myogenic differentiation was induced. At 24 h or 72 h after differentiation induction, cells were collected for the analysis of RNAs and proteins. For ISRIB treatment, the differentiation medium was added either with or without 1 µM ISRIB (Cayman Chemical Company, Ann Arbor, MI, USA).
RNA isolation, reverse transcription reaction, and quantitative RT-PCR
Total RNA was extracted from C2C12 cells using ISOGEN II reagent (Nippon Gene, Tokyo, Japan) according to the manufacturer’s protocol. After DNase I (Thermo Fisher Scientific) treatment, total RNA was used for reverse transcription reaction using SuperScript III Reverse Transcriptase (Thermo Fisher Scientific) or ProtoScript II Reverse Transcriptase (New England Biolabs (NEB), Beverly, MA, USA) with random or oligo (dT) primers (Thermo Fisher Scientific). Quantitative real-time PCR was conducted using SYBR Premix Ex Taq (Takara Bio Inc., Shiga, Japan) and a Thermal Cycler Dice Real Time System TP800 (Takara Bio Inc.). The results were normalized to the Rpl26 expression. The primers used are listed in Supplementary Table 1.
Protein extraction and Western blotting
Cells were lysed in RIPA buffer (50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 0.5% sodium deoxycholate) containing protease inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 4 μg/ml leupeptin) and phosphatase inhibitors (5 mM NaF, 5 mM β-glycerophosphate, 1 mM Na3VO4). The protein concentration was measured using a Pierce BCA Protein Assay Kit (Thermo Fischer Scientific). Equal amounts of protein were used for Western blotting. The following primary antibodies were used: myogenin antibody (F5D) (sc-12732, Santa Cruz Biotechnology, Dallas, TX, USA), MHC antibody (MF20, Developmental Studies Hybridoma Bank (DSHB)), Myod antibody (CE-011A, Cosmo Bio Co. Ltd., Tokyo, Japan), hnRNPK antibody (#4675, Cell Signaling Technology (CST), Beverly, MA, USA), hnRNPK antibody (F45P9C7, BioLegend, San Diego, CA, USA), TIAR antibody (#8509, CST), and Atf4 antibody (693901, BioLegend). The following HRP-linked secondary antibodies were used: anti-mouse IgG (#7076, CST), anti-rabbit IgG (#7074,CST), and TrueBlot ULTRA anti-Ig HRP, Mouse (Rat) (18-8817-33, Rockland Immunochemicals Inc., Limerick, PA, USA). Can Get Signal Immunoreaction Enhancer Solution (Toyobo, Osaka, Japan) was used when necessary. The signal was detected with ImmunoStar LD reagent (FUJIFILM Wako, Osaka, Japan) using a cooled CCD camera system (Light-Capture, ATTO, Tokyo, Japan).
Immunofluorescence assay
Immunofluorescence analyses of C2C12 cells were performed as previously described (Hitachi et al., 2019). Briefly, 24 h or 72 h after the induction of differentiation, cells were fixed with 4% PFA and permeabilized with 0.2% Triton X-100. After blocking with 5% FBS, cells were stained with an anti-myogenin antibody (F5D, DSHB) or with an anti-MHC antibody (MF20, DSHB). The following secondary antibodies were used: Goat anti-Mouse IgG (H+L) Highly Cross-Adsorbed Secondary Antibody, Alexa Fluor488 (Thermo Fisher Scientific). Nuclei were counterstained with DAPI (Dojindo, Kumamoto, Japan). A DMI4000B microscope with a DFC350FX CCD camera (Leica, Wetzlar, Germany) was used for visualization of signals. Images were analyzed using the Image J software program (ver. 1.53a).
Identification of Myoparr-binding proteins
Myoparr-binding proteins were collected with the RiboTrap Kit (Medical & Biological Laboratories (MBL), Aichi, Japan) using BrU labeled RNAs, as described previously (Hitachi et al., 2019). BrU-labeled Myoparr and EGFP RNA were prepared using the Riboprobe System (Promega, Madison, WI, USA). Twenty-four hours after the induction of differentiation, nuclear extract was prepared from differentiating C2C12 myoblasts. The Myoparr or EGFP RNA (50 pmol) were mixed with the nuclear extract from differentiating C2C12 myoblasts for 2 h at 4°C. The RNA-protein complexes were collected by Protein G Plus Agarose (Thermo Fisher Scientific) conjugated to an anti-BrdU antibody, and proteins were eluted by adding BrdU. Purified proteins were detected by Western blotting using a specific antibody, as described above.
RNA immunoprecipitation and the RNA pull-down assay
Immunoprecipitation of endogenous RNAs (Myoparr, Xist, or Neat1) was performed using the RIP-Assay Kit (MBL) using 2 × 107 C2C12 cells 48 h after the induction of differentiation, as previously described (Hitachi et al., 2019). The following antibodies were used for RNA immunoprecipitation: normal rabbit IgG (#2729S, CST), anti-HNRNPK pAb (RN019P, MBL), or TIAR mAb (#8509, CST). After treatment with DNase I, immunoprecipitated RNAs was used for the reverse transcription reaction. The precipitation percentage (precipitated RNA vs. input RNA) was calculated by qRT-PCR using the primers listed in Supplementary Table 1.
An RNA pull-down assay was performed with a RiboTrap Kit. Various lengths of Myoparr were subcloned into a pGEM-Teasy vector (Promega). BrU-labeled Myoparr (10 pmol each) was bound to Protein G Plus Agarose conjugated to an anti-BrdU antibody and mixed with the in vitro transcribed/translated hnRNPK protein. After several washing steps, the binding of hnRNPK to Myoparr was analyzed by Western blotting using an hnRNPK antibody as described above.
Luciferase reporter assay
The upstream region of myogenin (−1650/+51) was PCR-amplified and clonedinto the pGL4.20 vector (Promega). The hnRNPK-expressing plasmid was a kind gift from Dr. H. Okano (Yano et al., 2005). Proliferating C2C12 cells were transfected with the myogenin promoter and hnRNPK-expressing plasmid using Lipofectamine 2000 (Thermo Fisher Scientific). The total amount of DNA was kept constant by the addition of the pcDNA3 vector. Cells were collected at 24 h after the induction of differentiation and dissolved in Passive Lysis Buffer (Promega). The effect of hnRNPK on myogenin promoter was measured using a Lumat LB 9507 luminometer (Berthold Technologies, Bad Wildbad, Germany) with the Dual-Luciferase Reporter Assay System (Promega) according to the manufacturer’s protocol. The pGL4.74 vector (Promega) was used as an internal control. The relative luciferase activity is shown as the firefly to Renilla luciferase ratio.
To reconstitute the complex chromatin structure and epigenetic regulation in vitro (Liu et al., 2001), the upstream regions of myogenin (−242/+51 and -1650/+51) were subcloned into the episomal luciferase vector, pREP4-luc, and the -242-Luc and -1650-Luc constructs were generated. To create the -1650Δccawmcc-Luc construct, the upstream region (−971/-902) of myogenin was deleted from the -1650-Luc construct. The region containing the putative hnRNPK-binding motif ccawmcc was identified by RBPmap (Paz et al., 2014). The episomal luciferase vector, pREB7-Rluc, was used as an internal control for the luciferase assay. The pREP4-luc and pREP7-Rluc vectors were gifts from Dr. K. Zhao (Liu et al., 2001). Subconfluent C2C12 cells were transfected with the indicated episomal luciferase vectors using Lipofectamine 2000. After myogenic induction, cells were collected and used to measure the relative luciferase activity.
Analysis of downstream genes regulated by Myoparr KD and hnRNPK KD
Downstream genes regulated by Myoparr KD and hnRNPK KD were identified as described previously (Hitachi and Tsuchida, 2019) using our RNA-Seq raw data (accession No. DRA005527). Briefly, statistical analysis of differentially expressed genes by Myoparr KD and hnRNPK KD was performed using DESeq2 ver. 1.12.4 software (Love et al., 2014) with a Wald test (cut-offs: false discovery rate (adjusted p-value, padj)< 0.05 and log 2 fold change > 0.75 or < -0.75). A pathway analysis of significantly upregulated genes coding for cytosolic aminoacyl-tRNA synthetases was performed based on the KEGG (Kanehisa et al., 2020). A Gene Ontology (GO) analysis of differentially expressed genes was performed with DAVID ver. 6.8 (https://david.ncifcrf.gov/). An enrichment analysis was performed using Metascape (Zhou et al., 2019).
Statistical analysis
Error bars represent standard deviation. Statistical analyses were performed using unpaired two-tailed Student’s t-tests. For comparisons of more than 2 groups, a one-way ANOVA followed by Tukey’s post hoc test was performed using Prism 9 (GraphPad Software, San Diego, CA, USA). Statistical significance is reported in the Figures and Figure legends. P values of < 0.05 were considered statistically significant.
Results
Identification of hnRNPK as a Myoparr-binding protein in skeletal muscle cells
Our proteomics analysis identified both hnRNPK and TIAR as candidates for Myoparr-associated proteins (Hitachi et al., 2019). To reveal whether hnRNPK and TIAR are associated with the Myoparr function during myogenic differentiation, we first examined the specific interaction of Myoparr with endogenous hnRNPK and TIAR. In vitro synthesized Myoparr was labeled with 5-bromouridine (BrU) and mixed with the nuclear extract from differentiating C2C12 myoblasts. After purification of Myoparr by immunoprecipitation with a BrdU antibody, specific binding between Myoparr and hnRNPK protein was confirmed by immunoblotting (Figure 1A). Since binding between EGFP mRNA and hnRNPK was not observed, EGFP mRNA was used as a negative control. Intracellular binding between endogenous Myoparr and hnRNPK was shown by RNA immunoprecipitation using an hnRNPK-specific antibody without a crosslinking (Figure 1B and C). This enrichment of Myoparr by hnRNPK was stronger than that of Xist or Neat1, both of which interact with hnRNPK in other cells (Chu et al., 2015; Kawaguchi et al., 2015). Although TIAR protein was retrieved by synthesized Myoparr, intracellular endogenous interaction between Myoparr and TIAR was not observed (Supplementary Figure 1A-C). These results suggested that hnRNPK has function associated with Myoparr-binding during skeletal muscle differentiation.
hnRNPK represses the expression of myogenin in differentiating myoblasts
During myogenic differentiation, the expression of Myoparr gradually increases and is required to activate the expression of myogenin (Hitachi et al., 2019). Thus, we examined the changes in the expression of hnRNPK during C2C12 cell differentiation. Although the expression of myogenin was highly increased after myogenic induction, the expression of hnRNPK gradually decreased (Figure 2A), suggesting the regulatory role of hnRNPK in the expression of myogenin. We next knocked down hnRNPK using the small interfering RNAs (siRNAs) in differentiating C2C12 cells. An immunocytochemistry analysis of the myogenin expression afterhnRNPK knockdown (KD) by two distinct siRNAs did not show a sufficient increase in the ratio of myogenin-positive cells comparing to the control (Figure 2B). However, we observed that the number of cells with high intensity of myogenin signal was increased by hnRNPK KD (Figure 2B and C). In addition, western blotting analyses showed that hnRNPK KD was associated with a significant increase in the expression of myogenin (Figure 2D-F). These results indicated that although hnRNPK KD is not enough to induce the expression of myogenin in cells where the expression of myogenin is not intrinsically present, hnRNPK KD increases the expression of myogenin in cells with intrinsic myogenin expression.
We observed that the expression of myogenin was significantly increased by hnRNPK KD (Figure 2G and H). Although the differences were not statistically significant, the Myoparr expression levels also tended to be increased by hnRNPK KD (Figure 2I). From these results, we surmise that hnRNPK negatively regulates the expression of myogenin at the transcriptional level. The effect of hnRNPK on the transcription of myogenin was examined using the myogenin-promoter-driven luciferase assay. The overexpression of hnRNPK in differentiating C2C12 cells decreased the myogenin promoter activity (Figure 2J), indicating that hnRNPK negatively regulates the expression of myogenin via the myogenin promoter.
The hnRNPK-binding region of Myoparr is required to repress myogenin expression
The hnRNPK-binding region of Myoparr was determined by RNA pull-down experiments using the various forms of Myoparr (Figure 3A). Figure 3B showed that hnRNPK bound to the full-length sense-strand of Myoparr (#3 in Figure 3B), but not to the full-length antisense-strand of Myoparr (#4). Although deletion of the 5’-region (#5 and 6) or 3’-region (#2 and 7) of Myoparr did not affect binding to hnRNPK, the deletion of the 3’-half of Myoparr (#1) completely diminished the binding to hnRNPK (Figure 3B). These results indicated that an approximately 300-nt region (613-952 nt) of Myoparr is indispensable for binding to hnRNPK. Searching the motif sequence of RBPs from the 300-nt region revealed that there are 8 ccawmcc motifs, which are recognized by hnRNPK (Figure 3C). The deletion of the motifs (660-729 nt) from the full-length of Myoparr (#8) markedly weakened the binding to hnRNPK (Figure 3B). Thus, the ccawmcc motifs on Myoparr were shown to be required for binding to hnRNPK.
To clarify whether Myoparr is involved in regulating the expression of myogenin by hnRNPK, we examined the effect of the ccawmcc motif on the myogenin promoter activity. To imitate a chromatin structure and epigenetic regulation on the plasmid DNA (Liu et al., 2001), the upstream region of myogenin (−1649 to +52) including Myoparr was cloned into an episomal luciferase vector. In accordance with our previous findings (Hitachi et al., 2019), the myogenin promoter showed high activity in the presence of Myoparr (−1650-Luc) in comparison to the -242-Luc construct, which only contains the myogenin promoter region in differentiating myoblasts (Figure 3D). Intriguingly, the activity of the -1650-Luc construct was further enhanced by the deletion of a region of approximately 70 bp (1650Δccawmcc-Luc), which corresponds to the 660-729 nt region on Myoparr (Figure 3D). These results indicate that the hnRNPK-binding region of Myoparr is required to repress the expression of myogenin during skeletal muscle differentiation.
hnRNPK inhibits skeletal muscle differentiation but is required for normal myotube formation
The inhibitory role of hnRNPK in the expression of myogenin possibly via the ccawmcc motif on Myoparr suggested that hnRNPK and Myoparr have common downstream genes. Our RNA-Seq analysis (Hitachi and Tsuchida, 2019) revealed that hnRNPK KD significantly increased the expression of 226 genes and significantly decreased the expression of 190 genes. We compared the downstream genes regulated by hnRNPK KD and Myoparr KD, and the comparative heatmap analysis showed that the genes regulated by hnRNPK KD showed the opposite direction to the Myoparr KD (Figure 4A). Twenty percent of genes (84/416) altered by hnRNPK KD overlapped with genes regulated by the Myoparr KD (Figure 4B and Supplementary Talbe 2). The intersection of these genes was 12.3-fold greater than that expected by chance (p = 1.393037 × 10−45). Although these 84 genes showed a low correlation coefficient (R=0.0758323), we observed a negative correlation trend for one segment of them, including myogenin; the expression of 50 genes belonging to this segment was increased by hnRNPK KD and decreased by Myoparr KD (red frame in Figure 4C). These genes were enriched in sarcomere organization, myofibril assembly, and muscle contraction categories in GO terms (Figure 4D), indicating that hnRNPK inhibits myogenic differentiation and maturation.
In accordance with the results of the RNA-Seq analysis, we observed a significant increase in the expression of Myod, one of master regulators of myogenesis, and Myosin heavy chain (MHC), which is a later marker of myogenic differentiation and maturation, in the early stages of differentiation with hnRNPK KD (Figure 4E), indicating the hnRNPK KD causes premature differentiation of myoblasts. Although hnRNPK KD increased MHC expression in the late stages of differentiation (Figure 4F), hnRNPK KD did not affect the percentage of differentiated cells; this was shown by the fusion index (Figure 4G) as well as the results from immunocytochemistry to detect myogenin (Figure 2C). Instead, we observed the appearance of locally spherical myotubes following hnRNPK KD. These differed from the normal tube-shaped myotubes (Figure 4G). Thus, these results indicated that hnRNPK is required for normal myotube formation, possibly through the inhibitory effect on the premature differentiation of myoblasts at early stages of differentiation.
hnRNPK represses the expression of aminoacyl-tRNA synthetases via the eIF2α/Atf4 pathway
The Venn diagram in Figure 4B indicates that 332 genes regulated by hnRNPK KD were Myoparr-independent, suggesting the Myoparr-independent role of hnRNPK in differentiating myoblasts. We performed an enrichment analysis of genes regulated by hnRNPK KD and compared them with genes regulated by Myoparr KD. As reported previously (Hitachi et al., 2019), genes related to cell cycle and cell division were only enriched in genes regulated by Myoparr KD (Figure 5A). Skeletal muscle-associated genes were regulated by both Myoparr KD and hnRNPK KD (Figure 5A). Intriguingly, genes coding for aminoacyl-tRNA synthetases were regulated specifically by hnRNPK KD (red frame in Figure 5A). In mice, there are two-types of aminoacyl-tRNA synthetases: cytosolic and mitochondrial aminoacyl-tRNA synthetase. Our RNA-Seq analysis showed that hnRNPK KD significantly increased the expression of 10 genes coding for cytosolic aminoacyl-tRNA synthetases, whereas it had little effect on the expression of genes coding for mitochondrial aminoacyl-tRNA synthetases (Supplementary Figure 2). To confirm the RNA-Seq results, we randomly picked up 5 genes coding for cytosolic aminoacyl-tRNA synthetases, and their expression changes by hnRNPK KD were verified by qRT-PCR. hnRNPK KD using two distinct siRNAs significantly increased the expression of Aars, Gars, Iars, Nars, and Sars (Figure 5B-F). This regulation was not observed following Myoparr KD (Figure 5B-F), indicating that hnRNPK regulates the expression of these genes in a Myoparr-independent manner.
The expression of genes coding for almost all cytosolic aminoacyl-tRNA synthetases is activated by transcription factor Atf4 (Harding et al., 2003; Shan et al., 2016). Thus, we examined whether hnRNPK KD altered the expression of Atf4 in skeletal muscle cells. Although not statistically significant, the Atf4 expression in differentiating myoblasts tended to be increased by hnRNPK KD (Figure 5G). Myoparr KD did not affect the expression of Atf4 (Figure 5G). The effect of hnRNPK KD was more pronounced in the expression of Atf4 protein. The amount of Atf4 protein was highly increased by hnRNPK KD (Figure 5H). We further examined the expression changes of other ATF4 target genes by hnRNPK KD in differentiating myoblasts. The expression levels of Asns and Psat1, which encode proteins related to amino acid synthesis, were significantly increased by hnRNPK KD (Supplementary Figure 3A and B). In addition, the expression levels of Chop, Chac1, and Trb3, pro-apoptosis genes, and Gadd34, an another ATF4 target gene, also tended to be increased by hnRNPK KD (Supplementary Figure 3C-F). These results suggest that hnRNPK regulates the expression of genes associated with amino acid synthesis via the expression of Atf4.
Under the condition of endoplasmic reticulum (ER) stress, Atf4 mRNA is translated more efficiently and contributes to the restoration of cell homeostasis via the regulation of cytosolic aminoacyl-tRNA synthetases (Afroze and Kumar, 2017). Thus, to reveal the molecular mechanism by which hnRNPK regulates the expression of Atf4, we finally focused on ER stress. ISRIB is an inhibitor of eIF2α, which is a downstream component of PERK signaling, one branch of the ER stress sensors. We investigated whether ISRIB treatment could suppress the increase in the expression of Atf4 induced by hnRNPK KD and found that ISRIB treatment completely rescued this hnRNPK-KD-induced increase (Figure 6A). Furthermore, ISRIB treatment abrogated the increased expression of Atf4 target genes, Aars, Gars, Iars, Nars, and Sars by hnRNPK KD (Figure 6B-F and Supplementary Figure 4A-E). Thus, these results indicate that Myoparr-independent hnRNPK function is the regulation of the eIF2α/Atf4 pathway during myogenic differentiation.
Discussion
RBPs have multiple molecular functions, including RNA splicing, transcription, translation, RNA stability, and the formation of the nuclear structure, to regulate cell proliferation, differentiation, development, and diseases (Kelaini et al., 2021). Although many RBPs have multiple functions in the cells (Briata and Gherzi, 2020; Jonas et al., 2020; Nostrand et al., 2020), it is still unclear how their pleiotropic functions are determined. In this study, we revealed novel multiple functions of hnRNPK, a member of the hnRNP family of RBPs, in skeletal muscle cells. By focusing on a lncRNA Myoparr-associated protein, we found that hnRNPK repressed the expression of myogenin, coding for one of the master regulators of muscle differentiation. Deletion of the hnRNPK-binding region of Myoparr activated the expression of myogenin. Moreover, our comparative analysis of the downstream genes of hnRNPK and Myoparr showed that the function of hnRNPK was pleiotropic. During myogenic differentiation, hnRNPK repressed the expression of a set of genes coding for cytosolic aminoacyl-tRNA synthetases via the eIF2α/Atf4 pathway. Taken together, our study revealed multiple inhibitory roles of hnRNPK in skeletal muscle cells: one was Myoparr-associated and the other was Myoparr-independent (Figure 7). Recently, Xu et al. reported that the deficiency of 36 amino acids in hnRNPK diminished C2C12 differentiation (Xu et al., 2018). However, our results provided strong evidence to support that hnRNPK has an inhibitory effect on muscle differentiation. In addition, we observed the appearance of locally spherical myotubes following hnRNPK KD. Considering the facts that dysregulated Myod expression leads to premature myogenic differentiation (Bröhl et al., 2012) and results in the formation of dysfunctional myofibers in mice (Schuster-Gossler et al., 2007), uncoordinated increases in Myod, myogenin, and MHC expression by hnRNPK KD may lead to abnormal shape of myotubes. In addition, the morphology of these myotubes closely resembles myotubes with myofibril-assembly defects (Wang et al., 2013), suggesting that hnRNPK may also be involved in the regulation of the myofibril assembly in myotubes. Therefore, despite its lncRNA-associated and - independent roles in the inhibition of myogenic differentiation, hnRNPK is apparently required for the formation of normal myotubes.
We observed that hnRNPK KD increased myogenin protein levels more robustly than myogenin mRNA. The peak expression of myogenin protein is detected 1-2 days after that of myogenin mRNA in both in vitro and in vivo myogenesis (Angelis et al., 1992; Figueroa et al., 2003), suggesting that a slight increase in myogenin mRNA by hnRNPK KD in the early stages of myogenic differentiation eventually led to a marked increase in myogenin protein. Therefore, the fine-tuning of the myogenin expression by hnRNPK at the early stages of differentiation may have a significant impact on the overall myogenic differentiation processes through the RNA-protein network. Intriguingly, despite the percentage of myogenin-positive cells was not changed, the expression of myogenin protein was increased by hnRNPK KD. We observed that hnRNPK KD only increased the number of cells with high intensity of myogenin signal. These results suggest that hnRNPK can specifically repress the expression of myogenin in a subset of responding cells, rather than by simply turning off the expression of myogenin in every myoblast. Our experiments showed that hnRNPK repressed the expression of myogenin at the transcriptional level possibly via binding to the ccawmcc motif on Myoparr, suggesting that the existence of Myoparr would be necessary for hnRNPK to inhibit the expression of myogenin. Since myogenin and Myoparr share the same promoter region (Hitachi et al., 2019), myogenin and Myoparr are likely expressed in the same cells. Thus, the expression of myogenin would not be activated in the cells without the cell-intrinsic expression of Myoparr, even if hnRNPK is depleted in every myoblast. Further studies are required to investigate the more precise molecular mechanism by which hnRNPK regulates the expression of myogenin via binding to Myoparr.
ER stress is induced by several perturbations disrupting cell homeostasis, including protein misfolding, viral infection, and changes in intracellular calcium concentration (Hetz et al., 2015). The cells recognize those stresses with three branches of ER transmembrane sensors signaling, PERK, inositol-requiring protein 1 (IRE1), and ATF6 (Afroze and Kumar, 2017). During myogenic differentiation, ATF6 signaling was activated and led to apoptosis in myoblasts (Nakanishi et al., 2005). The increased phosphorylation of eIF2α, a component of PERK signaling, was observed in myoblasts at the early stage after the induction of differentiation (Alter and Bengal, 2011). In addition, treatment with ER stress inducers enhanced apoptosis in myoblasts but led to efficient myogenic differentiation in the remaining surviving cells (Nakanishi et al., 2007). Recently, the deletion of PERK in satellite cells, which are adult muscle stem cells, was shown to inhibit myogenic differentiation and led to impaired skeletal muscle regeneration in adult mice (Xiong et al., 2017), indicating that ER stress promotes myogenic differentiation. In this study, we showed that the hnRNPK KD in differentiating myoblasts increased the expression of Atf4 and this effect was diminished by treatment with ISRIB, a specific inhibitor of eIF2α. The expression levels of a set of genes coding for cytosolic aminoacyl-tRNA synthetases, which are regulated by Atf4, were also increased by the hnRNPK KD, and were completely rescued after ISRIB treatment. The inhibitory effects of hnRNPK on Atf4 and cytosolic aminoacyl-tRNA synthetases were independent of Myoparr. Therefore, our findings suggest that hnRNPK fine-tunes the myogenic differentiation process by modulating ER stress via eIF2α/Atf4 signaling in a lncRNA-independent manner. Since hnRNPK is involved in the translational efficiency (Lynch et al., 2005; Yano et al., 2005), a decrease in hnRNPK may induce the unfolded protein response via the translational machinery in myoblasts and alter eIF2α/Atf4 signaling.
In conclusion, hnRNPK plays multiple lncRNA-dependent and -independent roles in the inhibition of myogenic differentiation. Thus, the analysis of RBPs bound to lncRNAs will be useful for elucidating both the physiological functions of lncRNAs and the complex functions of RBPs in cell differentiation. Induced ER stress, including increased PERK signaling, was observed in skeletal muscle biopsy samples from myotonic dystrophy 1 patients and in mdx mice, a model of Duchenne muscular dystrophy (Hulmi et al., 2016; Ikezoe et al., 2007). Moreover, mutations in genes coding for aminoacyl-tRNA synthetases are implicated in human neuromuscular disorders (Benarroch et al., 2020), and autoantibodies against aminoacyl-tRNA synthetases are found in autoimmune disease (Targoff et al., 1993). Collectively, downstream genes of hnRNPK are strongly associated with neuromuscular and other disorders in humans, suggesting that targeting hnRNPK to regulate the expression of these genes and signaling may become a new therapeutic strategy for human diseases.
Author Contributions
K.H. and K.T. designed research; K.H. analyzed data; K.H., Y.K., and M.N. performed research; K.H. and K.T. wrote the paper. All authors revised, edited, and read the manuscript and approved the final manuscript.
Conflict of interest
The authors declare no competing interests in association with the present study.
Figure Legends
Supplementary Table 1. Primer and siRNA sequences.
Supplementary Table 2. The value (log2 fold change) of the genes for which expression levels were increased or decreased by Myoparr KD and hnRNPK KD.
Supplementary Figure 1. Detection of the interaction between Myoparr and TIAR in differentiating C2C12 cells. (A) The interaction between synthesized Myoparr and TIAR was detected by immunoblotting using a TIAR antibody. (B) qRT-PCR for the detection of Myoparr and Xist following RNA-immunoprecipitation using a TIAR antibody. The presence or absence of reverse transcription reaction is shown by (RT+) or (RT-), respectively. Bars indicate the average of two independent experiments, and circles and triangles represent the values of each experiment. (C) Purified TIAR protein from C2C12 cells by immunoprecipitation using a TIAR antibody was confirmed by Western blotting.
Supplementary Figure 2. RNA-Seq revealed the altered expression of a group of genes coding for cytosolic aminoacyl-tRNA synthetases by hnRNPK KD. (A) The KEGG pathway diagram of cytosolic aminoacyl-tRNA biosynthesis. The gene names surrounded by the red frame indicate genes that are significantly upregulated by hnRNPK KD. (B-C) The results of the RNA-Seq analysis showing the altered expression of genes coding for cytosolic aminoacyl-tRNA synthetases (B) and mitochondrial aminoacyl-tRNA synthetases (C) by hnRNPK KD. Red and blue indicate genes that are significantly upregulated or downregulated, respectively, by hnRNPK KD.
Supplementary Figure 3. The expression changes of ATF4 target genes by hnRNPK KD in differentiating myoblasts. (A-F) The results of qRT-PCR for detecting the expression of Asns (A), Psat1 (B), Gadd34 (C), Chop (D), Chac1 (E), and Trb3 (F) following hnRNPK KD. n = 3, mean ± SD. ***p < 0.001, *p < 0.05.
Supplementary Figure 4. hnRNPK regulates the expression of a group of genes coding for aminoacyl-tRNA synthetases via the eIF2α/Atf4 pathway. (A-E) The results of qRT-PCR for detecting the altered expression of Aars (A), Gars (B), Iars (C), Nars (D), and Sars (E) following hnRNPK KD (using a different siRNA from Figure 6) and ISRIB treatment. n = 3, mean ± SD. *p < 0.05.
Acknowledgments
This work was supported in part by JSPS KAKENHI (19H03427 and 20K07315), Intramural Research Grants (2-5) for Neurological and Psychiatric Disorders of NCNP, and a Grant-in-Aid from the Mochida Memorial Foundation for Medical and Pharmaceutical Research.