Enhanced myogenesis through lncFAM-mediated recruitment of HNRNPL to the MYBPC2 promoter

Abstract

The mammalian transcriptome comprises a vast family of long noncoding (lnc)RNAs implicated in physiologic processes such as myogenesis, through which muscle forms during embryonic development and regenerates in the adult. However, the specific molecular mechanisms by which lncRNAs regulate human myogenesis are poorly understood. Here, we identified a novel muscle-specific lncRNA, lncFAM71E1-2:2 (lncFAM), which increased robustly during early human myogenesis. Overexpression of lncFAM promoted differentiation of human myoblasts into myotubes, while silencing lncFAM suppressed this process. As lncFAM resides in the nucleus, chromatin isolation by RNA purification followed by mass spectrometry (ChIRP-MS) analysis was employed to identify the molecular mechanisms whereby it might promote myogenesis. Analysis of lncFAM-interacting proteins revealed that lncFAM recruited the RNA-binding protein HNRNPL to the promoter of MYBPC2, in turn increasing MYBPC2 mRNA transcription and enhancing production of the myogenic protein MYBPC2. These results highlight a mechanism whereby a novel ribonucleoprotein complex, lncFAM-HNRNPL, elevates MYBPC2 expression transcriptionally to promote myogenesis.

INTRODUCTION

With the advent of high-throughput sequencing technologies, the human transcriptome has been found to comprise only a small fraction of protein-coding messenger (m)RNAs; the vast majority are noncoding (nc)RNAs with other functions (1–3). Among this large class, long noncoding (lnc)RNAs, spanning >200 nucleotides, are increasingly recognized as having pivotal regulatory functions in the cell, influencing processes like replication, differentiation, senescence, apoptosis and responses to a range of damaging and activating stimuli (4–6). Given their influence on cellular processes, lncRNAs have been implicated in many developmental processes, including neurogenesis, adipogenesis, and myogenesis; they also participate in disease processes like cancer, neurodegeneration, cardiovascular disease, diabetes and sarcopenia (7–11).

LncRNAs are highly heterogeneous in size and structure, they may reside in the cytoplasm or the nucleus (12–14), and their subcellular localization is closely linked to their molecular functions (13,15–17). Cytoplasmic lncRNAs can associate with cytosolic granules, ribosomes, as well as the endoplasmic reticulum and mitochondria; through these interactions, they regulate mRNA transport, stability, and translation, as well as protein stability, post-translational modification and function (6,18–20). Nuclear lncRNAs associate with specialized domains like paraspeckles, nucleoli, and the lamina, as well as with chromosomes, chromatin domains and gene regions; accordingly, they modulate nuclear processes like chromatin organization, and RNA transcription and splicing (16,21–23). Nuclear lncRNAs can control large-scale organization of entire chromosomes or segments of chromosomes (e.g., telomeres), but they can also have localized influence on chromatin organization by altering local DNA methylation and histone acetylation and/or methylation (24). LncRNAs directly influence transcription through proteins that bind enhancers or promoters and either promote or suppress the transcription machinery (25). Additionally, nuclear lncRNAs can influence pre-mRNA splicing and assemble proteins and RNAs in nuclear bodies.

Many lncRNAs have been found to be potent drivers of skeletal myogenesis in mammals, a key developmental process through which muscle forms during embryonic development and regenerates in the adult (8). During myogenesis, muscle stem cells (satellite cells) first differentiate into myoblasts, and myoblasts then progress through differentiation to fuse into multinucleated myotubes. The early stages of myogenesis are controlled by several transcription factors, including MYOD, MYF5, MYOG, MRF4, MEF2A and MEF2C (26). The later stages of myogenesis include events during which mononucleated myoblasts fuse into multinucleated myotubes, cytoskeletal proteins reorganize, and the myosin heavy chain (MYH) proteins are expressed and function in muscle contraction. Although the functions of most lncRNAs remain unexplored, several myogenic lncRNAs have been found to function through their association with microRNAs (∼21-nt ncRNAs) that typically repress the stability and/or translation of mRNAs with which they are partially complementary (27). For example, lincMD1 enhances myogenesis by binding and thereby neutralizing miR-135 and miR-133 (28); MALAT1 regulates myogenesis by serving as a miR-133 sponge and thereby increasing the production of miR-133 targets, and also by regulating the transcription of Myod mRNA (29); Malat1 regulates myogenic differentiation and muscle regeneration by modulating MYOD transcriptional activity (30); and lnc-mg promotes myogenesis by suppressing miR-125b function (31). In addition, lncMyoD regulates myoblast differentiation by suppressing the translation of IMP2 (32), the promoter-associated lncRNA Myoparr is required for myogenic differentiation (33), and the circular lncRNA circSamd4 can regulate myogenesis by preventing the binding of myogenic transcriptional repressors PURA and PURB (34).

In this study, we report a novel human lncRNA named lncFAM71E1-2:2 (lncFAM), located at chr19:50486810–50487638 (hg19, Ensembl v82), which increases markedly during myogenesis and is highly abundant in human myoblasts but not in other cell types. We find that silencing lncFAM significantly reduces myotube formation and delays myogenesis, while overexpressing lncFAM accelerates myogenic differentiation. Analysis of lncFAM, a primarily nuclear RNA, by ChIRP-MS (chromatin isolation by RNA purification followed by mass spectroscopy) revealed that it interacts functionally with the RNA-binding protein (RBP) heterogeneous nuclear ribonucleoprotein L (HNRNPL). The lncFAM-HNRNPL ribonucleoprotein (RNP) complex binds to the MYBPC2 promoter and increases transcription of the MYBPC2 mRNA, which encodes a key component of skeletal muscle involved in myogenesis. In sum, we have identified a novel lncRNA that strongly promotes myogenesis by enhancing the transcription of the myogenic protein MYBPC2.

MATERIALS AND METHODS

Cell culture, transfection and myogenic differentiation

Immortalized human AB1167 and AB678 myoblasts were developed and cultured as described (35,36). Briefly, human myoblasts were cultured in growth medium (‘GM’, equal volume mixture of Ham's F10 medium supplemented with 20% FBS and Promocell Skeletal Muscle Cell Growth Medium) and were induced to differentiate by growth to high density followed by replacement of the GM with differentiation medium (‘DM’, DMEM supplemented with 2% horse serum). For silencing experiments, control small interfering RNA (Ctrl siRNA), lncFAM siRNA, MYBPC2 siRNA, or HNRNPL siRNA were transfected at a final concentration of 50 nM siRNA using Lipofectamine RNAiMAX (Thermo Fisher Scientific), typically 24 h before inducing differentiation.

Immunofluorescence

Cells were fixed using 4% paraformaldehyde in PBS (pH 7.4) for 20 min, washed with PBS, incubated for 5 min in 0.2% Triton X-100 in PBS, and then washed once more with PBS. Fixed cells were then incubated in the same solution containing primary antibody at 4°C for 20 h, washed with PBS, and incubated for 2 h at 25°C in PBS containing FITC- or rhodamine-conjugated secondary antibody. An antibody recognizing MYH/MHC (B-5, Santa Cruz Biotechnology) was used to detect MYH by immunofluorescence; staining with DAPI (4′,6-diamidino-2-phenylindole) was used to identify nuclei.

Reverse transcription (RT), real-time quantitative (q)PCR analysis, and copy number estimations

Total RNA from cultured cells was isolated using the Direct-zol RNA^TM MiniPrep kit (Zymo Research), which includes a digestion step using DNase I. For cDNA synthesis, reverse transcription (RT) was performed using Maxima Reverse Transcriptase (Thermo Fisher Scientific) following the manufacturer's protocol. Quantitative (q)PCR analysis of RNAs was carried out following the manufacturer's instructions for KAPA SYBR FAST ABI Prism qPCR kit (KAPA Biosystems) with RNA-specific primers (Supplementary Table S1). RT-qPCR reactions were performed on QuantStudio 5 Real-Time PCR System (Thermo Fisher Scientific) with a cycle setup of 2 min at 95°C and 40 cycles of 5 sec at 95°C plus 20 sec at 60°C; the fold change in abundance was calculated by the 2^−ΔΔCt method. In qPCR amplification reactions, control ‘RT minus’ (‘RT-’) reactions were routinely included. The number of molecules of lncFAM per cell (copy number) in myoblasts and nucleus-equivalent in myotubes were calculated by comparing lncFAM Ct values with the Ct values of a transcript of known abundance, GAPDH mRNA (estimated to be present at ∼1300 copies per AB678 cell and nucleus-equivalent, by comparing to a standard curve).

Single-molecule fluorescence in situ hybridization (smFISH)

The lncFAM RNA was imaged using a smFISH published protocol (37). Briefly, 25 probes, each 20 nt in length and designed to bind to the entire length of lncFAM, were ordered from LGC Biosearch Technologies with a 3′ NH2 modification. The probes were pooled, coupled with Texas Red dye, and the coupled fraction purified using high-performance liquid chromatography (HPLC). The labeled probes were then used to hybridize myoblasts and differentiated myotubes grown on glass coverslips. The coverslips were washed to remove unbound probes, stained with DAPI to visualize nuclei, and mounted on glass slides for imaging. Series of z stacks were obtained using the 100 × oil objective in a Nikon Ti-E inverted fluorescence microscope equipped with a PIXIS:1024B charge-coupled device (CCD) camera (Princeton Instruments) and Metamorph imaging software. The z stacks were merged to composite, and overlays were obtained using the NIH ImageJ software.

Lentivirus construction and transduction

Lentiviruses expressing lncFAM or MYBPC2, and an empty vector were purchased from GeneCopoeia. For transduction, AB678 cells were inoculated in a six-well plate and cultured until they reached 70–80% confluence. The control lentivirus and lentiviruses expressing lncFAM deletions of segments S3 or S6 (made and amplified in-house), full-length lncFAM or MYBPC2 were added to the culture at multiplicity of infection (MOI) of 1 for 24 h in the presence of polybrene. The cells were then differentiated by using the differentiated medium for analysis 24 h after transduction.

Western blot analysis and subcellular fractionation

Total protein lysates were prepared in RIPA buffer (25 mM Tris–HCl [pH 7.6], 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS [Thermo Fisher Scientific]) containing protease inhibitors. Nuclear and cytoplasmic fractions of AB678 were prepared following the manufacturer's instructions (NE-PER kit, Thermo Fisher Scientific). The cytoplasmic and nuclear fractions were subjected to analysis of protein, RNA, and ribonucleoprotein complexes. Proteins were size-separated by SDS-PAGE and transferred onto nitrocellulose membrane (Bio-Rad). For western blot analysis, primary antibodies were employed that recognized MYH/MHC (B-5) and HSP90 (F-8) (Santa Cruz Biotechnology), MEF2C (D80C1, Cell Signaling Technology), MYOG (Santa Cruz Biotechnology), HNRNPL (4D11, Santa Cruz Biotechnology), nucleolin (anti-C23, Santa Cruz Biotechnology) and MYBPC2 (PAB19214, Abnova). After incubation with the appropriate secondary antibodies, protein signals were developed using chemiluminescence.

Creatine kinase activity

Creatine kinase (CK) activity was determined in cell lysates by using the EnzyChrom creatine kinase assay kit (BioAssay Systems) following the manufacturer's protocol. Briefly, cell lysates (1–2 μg) were incubated with 10 μl substrate solution, 100 μl assay buffer and 1 μl enzyme mix at 37°C for 20 min; reactions were read 20 and 40 min later at 340 nm. CK activity was calculated by the equation CK = (OD40min – OD20min/ODCALIBRATOR – ODH2O) × 150, and expressed as ‘units per μg of total protein’ or ‘fold change’.

Chromatin isolation by RNA purification (ChIRP) combined with MS/western blot

Thirty 15-cm dishes of AB678 cells were used per ChIRP-MS experiment. Cell harvesting, disruption and ChIRP analysis were performed as described (38). Cells were crosslinked in 3% formaldehyde for 30 min, quenched in 125 mM glycine for 5 min and incubated in lysis buffer (50 mM Tris–HCl [pH 7.5], 10 mM EDTA, 1% SDS, protease inhibitor, and RNase inhibitor). Cell extracts were sonicated in Bioruptor^® (on: 30 s; off: 45 s; Intensity: strong) for 15 min and centrifuged at max speed (16 000 × g). Supernatants were collected and pre-cleared with streptavidin magnetic beads (Dynabeads MyOne Streptavidin C1, Thermo Fisher Scientific) for 45 min at 37°C. Supernatants were incubated with biotinylated antisense oligonucleotides (ASOs, designed at LGC Biosearch Technologies^TM) directed at lncFAM or lacZ (negative control) in hybridization buffer (50 mM Tris–HCl [pH 7.5], 1 mM EDTA, 1% SDS, 750 mM NaCl, 15% formamide, protease inhibitors, and RNase inhibitor) for 16 h at 37°C. The probes were then pulled down using streptavidin magnetic beads and washed five times in wash buffer (2 × SSC, 0.5% SDS) at 37°C. The bound proteins were eluted and reverse-crosslinked by boiling in SDS sample buffer (95°C for 30 min) for mass spectrometry and western blot analysis.

Mass spectrometry

After ChIRP assay, mass spectrometry analysis was performed by Poochon Scientific. Briefly, after in-gel trypsin digestion, peptides were analyzed by liquid chromatography-coupled tandem MS (LC-MS/MS) using a QExactive hybrid quadrupole orbitrap mass spectrometer (Thermo Scientific) with a Dionex UltiMate 3000 RSLCnano system. Peptide identification and protein assembly were performed on a Thermo Proteome Discoverer 1.4.1 platform. Tandem mass spectrometry datasets were analyzed against corresponding UniProtKB/Swiss-Prot database using the SEQUEST and Percolator algorithms. The results of the proteomic analysis are shown in the Supplementary Table S2.

Crosslinked RNP immunoprecipitation (CLIP) assay

Cells were fixed with 0.1% formaldehyde (10 min, 25°C) followed by quenching with 125 mM glycine for 5 min. After washes with cold PBS, samples were incubated in lysis buffer (50 mM Tris–HCl [pH 7.5], 150 mM KCl, 0.1% SDS, 1% Triton-X, 5 mM EDTA, 0.5% sodium deoxycholate, 0.5 mM DTT, protease inhibitors, and RNase inhibitor) for 20 min at 4°C, sonicated in a Bioruptor^® (on: 30 s; off: 45 s; intensity: strong) for 10 min, and centrifuged at max speed (16 000 × g). The supernatants were pre-cleared with Protein G beads (Dynabeads, Thermo Fisher Scientific) and the beads were discarded. An aliquot of the supernatant was saved as the total input control. The remaining supernatant was incubated with 5 μg of antibodies from Santa Cruz Biotechnology, recognizing HNRNPL (4D11) or nucleolin (Anti-C23), or with control IgG (Thermo Fisher Scientific); after rotation for 16 h at 4°C, 30 μl Protein G Sepharose™ 4 Fast Flow (GE Healthcare) beads were added and rotated for 2 h at 4°C and washed with lysis buffer. The IP beads and input samples were reverse-crosslinked in RCL buffer (2% N-lauroylsarcosine, 10 mM EDTA, 5 mM DTT in PBS without Mg²⁺ or Ca²⁺) with Proteinase K and RNase inhibitor for 45 min at 50°C with shaking. RNA was extracted using TriPure isolation reagent (Roche) and subjected to RT-qPCR analysis to measure the presence of specific RNAs.

In vitro transcription of lncFAM

To synthesize lncFAM, an in vitro transcription assay was employed to synthesize biotinylated overlapping fragments (∼150 nt) using the MEGAshortscript T7 kit (Thermo Fisher Scientific) following the manufacturer's instructions. The DNA templates used for the transcription assay were obtained by RT followed by conventional PCR amplification. Synthetic non-overlapping biotinylated RNA fragments (30–40 nt) spanning the entire sequence of lncFAM were custom-made from IDT.

Biotin pulldown analysis

AB678 whole-cell lysates (100 μg) were incubated with 100 ng of recombinant (in vitro-transcribed) biotinylated RNAs (fragments A through G) or synthetic non-overlapping biotinylated RNAs (fragments 1 through 6) covering the full sequence of lncFAM (30 min, 25°C) with rotation. Complexes were isolated with streptavidin magnetic beads (Dynabeads, Thermo Fisher Scientific) and the pulldown material was assessed by western blot analysis using anti-HNRNPL antibodies.

ChIP-qPCR analysis

AB678 cells (∼10⁷) were crosslinked with 1% formaldehyde for 10 min at 25°C and quenched by 125 mM Glycine. Cells were washed and collected by centrifugation at 300 × g for 5 min at 4°C. Cells were lysed in 10 ml of LB1 (50 mM HEPES–KOH [pH 7.5], 140 mM NaCl, 1 mM EDTA, 0.25% Triton X-100, 0.5% NP-40, 10% glycerol, and protease inhibitors) and incubated with rotation at 4°C for 10 min. Following centrifugation at 800 × g for 5 min at 4°C, the pellets were washed with 10 ml of LB2 (10 mM Tris–HCl [pH 8.0], 200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, protease inhibitors) by incubating them with rotation at 4°C for 10 min. Following centrifugation at 800 × g for 5 min at 4°C, cell pellets were resuspended in 1 ml of buffer LB3 (10 mM Tris–HCl [pH 8.0], 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% Na-Deoxycholate, 0.5% N-lauroylsarcosine, and protease inhibitors) and sonicated with Covaris S220 (Intensity 140 W, Duty Cycle 5%, cycles per burst 200, time 15 min). Triton X-100 was added to the resulting lysate to reach a final concentration of 1% Triton X-100, cleared by centrifugation for 10 min at maximum speed, and then incubated with 30 μl of washed Protein G beads (Dynabeads, Thermo Fisher Scientific) for 6 h at 4°C. The cleared lysate was then incubated with 2–10 μg HNRNPL (4D11, Santa Cruz Biotechnology) for 16 h at 4°C, washed twice with buffer FA (50 mM HEPES–KOH [pH 7.5], 140 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, and protease inhibitors), once with high-salt wash buffer (50 mM HEPES–KOH [pH 7.5], 500 mM NaCl, 1 mM EDTA, 0.1% Triton X-100, and protease inhibitors), twice with LiCl wash buffer (10 mM Tris–HCl [pH 8], 250 mM LiCl, 1 mM EDTA, 0.5% NP-40, and protease inhibitors), and twice with cold TE buffer with 0.1% NP-40. Immunocomplexes were eluted in ChIP elution buffer (50 mM Tris–HCl [pH 8.0], 10 mM EDTA, 1% SDS) with incubation for 30 min at 65°C, and crosslinking in the eluents was reversed by incubating at 65°C for 16 h. Immunoprecipitated DNA was treated with RNase A at 37°C for 1 h, followed by Proteinase K treatment at 37°C for 1 h. Immunoprecipitated DNA was extracted by ChIP DNA Clean & Concentrator (D5205, Zymo Research) and resuspended in 100 μl elution buffer. Primers for target genomic loci (Supplementary Table S1) were used for qPCR analysis.

RNA-seq analysis

Paired-end sequencing was performed on an Illumina NovaSeq 6000 sequencer. The FASTQ files were extracted using bcl2fastq v2.20.0.422, trimmed for adapter sequences using Cutadapt v1.18, and aligned to human genome hg19 Ensembl v82 using STAR software v2.4.0j. FeatureCounts was used to create gene counts. The DESeq2 package version 1.30.0 (39) in R (version 4.0.3) was used to carry out the differential expression analysis. Statistical testing was performed using the Wald test, and those transcripts with Benjamini-Hochberg adjusted P-values <0.05 and absolute log₂ fold change >1 were considered to be differentially expressed. The Gene Ontology (GO) pathway enrichment analysis was performed using the R package clusterProfiler version 4.0 (40). The RNA-seq data were deposited in GSE202793 (41) and GEO SuperSeries GSE215343.

RNA-sequencing analysis using the Oxford Nanopore Technology

For long-read RNA sequencing using the Oxford Nanopore Technology, AB678 myoblasts were induced into differentiation, and RNA was extracted 48 h later. High-quality total RNA (200 ng) was used to prepare sequencing libraries using a cDNA-PCR Sequencing Kit (Oxford Nanopore, Cat. #: SQK-PCS111) following the manufacturer's instructions. The quality and quantity of the sequencing libraries were assessed using the 2100 Bioanalyzer System (Agilent) and the Qubit 3 Fluorometer (Invitrogen). The library was sequenced in a PromethION 24 (Oxford Nanopore) instrument and basecalling was performed using the software Guppy version 6.01. Subsequently, Minimap2 version 2.17 (42) was employed to align the reads to hg19 using the -ax splice parameter, and StringTie2 version 2.2.1 was run in assembly mode using the -mix parameter (43,44) to generate the GTF file which was used as input for SQANTI3 (45). The transcripts identified by Nanopore sequencing are summarized in Supplementary Table S3.

Rapid amplification of cDNA ends (RACE)

The 3′ rapid amplification of cDNA ends (RACE) was performed using the 5′/3′ RACE Kit, 2nd Generation (Sigma) following the manufacturer's instructions. The primer sequences for 3′RACE of lncFAM were GAAGACCAGCAACTGCTAAGA and CTAAGACAGAGACTTCGAGGGA.

Statistical analysis

All experiments were repeated at least 3 times unless otherwise stated. Quantitative data are represented as the means ± SEM and compared statistically by unpaired Student's t-test, using Prism GraphPad (9.0). Statistical significance is indicated in the figures as * (P < 0.05), ** (P < 0.01) and *** (P < 0.001). Dedicated statistics were applied to RNA-seq data analysis as specified in the legends.

RESULTS

lncFAM increases markedly during human myogenesis

To investigate human myogenesis, we employed human myoblast lines AB678 and AB1167 (36,46), which proliferate in growth media (GM), but when switched to differentiation media (DM) containing 2% horse serum (Materials and Methods), they enter myogenic differentiation and gradually fuse over the following 48–72 h. Multinucleated myotubes were seen by phase-contrast microscopy and by monitoring the expression of myosin heavy chain (MYH), a marker of late stages of myogenesis, using immunofluorescence (IF) (Figure 1A). Western blot analysis was also used to track the expression levels of differentiation markers, including early (MYOG and MEF2C) and late (MYH) myogenesis markers, across the differentiation program (Figure 1B). The levels of mRNAs encoding these proteins, as measured by reverse transcription followed by quantitative PCR (RT-qPCR) analysis, followed similar patterns of abundance as the myogenic proteins (Figure 1C). Another widely used indicator of skeletal myogenesis, creatine kinase (CK) activity (47), similarly increased with differentiation in both AB678 and AB1167 cells (Figure 1D). These experiments served as framework for subsequent analysis of myogenesis.

Figure 1.

Characterization of human myoblasts during myogenesis. (A) AB678 and AB1167 human myoblasts were placed in differentiation conditions until myotubes were visible by phase-contrast microscopy (top) and fluorescence microscopy to detect MYH signal (bottom, green); nuclei were detected by DAPI staining (blue). (B) The levels of the early differentiation markers MYOG and MEF2C proteins, and the late differentiation marker MYH were assessed by western blot analysis of total protein lysates prepared from AB678 and AB1167 myoblasts at different times (h), after placement in differentiation medium (Diff.). (C) The levels of MYOD, MYOG and MEF2C mRNAs, as well as normalization control GAPDH mRNA, were measured by reverse transcription followed by quantitative PCR (RT-qPCR) analysis at different times during differentiation of AB678 and AB1167 myoblasts. (D) Creatine kinase activity levels were assessed at the times shown during myogenic differentiation of AB678 (top) and AB1167 (bottom) human myoblasts. (E) Volcano plot representation of the RNAs found to be differentially abundant after RNA-seq analysis of differentiating (24 h) versus proliferating (0 h) AB678 human myoblasts. Top, total RNA; bottom, lncRNAs. The x-axis represents log2 fold changes between the two groups, the y-axis represents the –log₁₀ adjusted p-value of gene expression variation. Red dots represent elevated (log2 fold change > 1), blue dots represent reduced (log₂ fold change < –1). All RNAs above the threshold (adjusted P-value < 0.05) show significant changes during differentiation in AB678 human myoblasts. (F) RT-qPCR analysis of lncFAM levels (normalized to GAPDH mRNA levels) at the times shown during differentiation of human AB678 myoblasts. (G) Quantification of copy numbers of lncFAM in AB678 myoblasts during differentiation, shown as number of molecules per cell/nucleus-equivalent (Materials and Methods). Data in (D, F) are the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001. Other data are representative of three or more biological replicates.

To comprehensively identify differentially expressed lncRNAs during human myogenesis, we collected AB678 cells before differentiation (time 0) and 24 h into differentiation, when gene expression programs were changing robustly. We analyzed RNA profiles by systematic RNA-sequencing analysis [Materials and Methods, GSE202793 (41) and GEO SuperSeries GSE215343] and visualization of both total RNA and lncRNA changes (Figure 1E). Upon validation by RT-qPCR analysis, one of the RNAs most strongly upregulated with myogenesis, with ∼2000-fold increase in differentiating AB678 myoblasts (36 h), was a previously uncharacterized human lncRNA, lnc-FAM71E1-2:2, hereafter lncFAM (Figure 1F). We first assessed the sequence of lncFAM in human myoblasts by 3′ rapid extension of cDNA ends (RACE), and then employed Nanopore RNA-sequencing for more accurate annotation (GSE215343; Materials and Methods). Both the 3′RACE and Nanopore sequencing results indicated that the main isoform of lncFAM in skeletal muscle has the same sequence as ENST00000595005.1, with the exception of a slightly longer isoform that fully encompasses ENST00000595005.1. Although little is known about lncFAM (ENST00000595005.1), it spans 626 nucleotides and is quite abundant at peak times of expression, reaching ∼860 copies per nucleus-equivalent (Materials and Methods) in AB678 cultures by 36 h into the differentiation program (Figure 1G). Given its strong, time-dependent rise, its absolute overall levels of expression, and its specific presence in human myoblasts (Supplementary Figure S1), we further investigated the role of lncFAM in human myogenesis.

lncFAM promotes myogenesis in human myoblasts

To study the function of lncFAM in myogenesis, we first sought to modulate lncFAM levels. We began by studying if loss of lncFAM influenced human myogenesis. RT-qPCR analysis confirmed that transfection of small interfering (si)RNAs directed at lncFAM for 24 h followed by initiating a myogenic program revealed that silencing lncFAM delayed the rise in production of myogenic markers MEF2C and MYOG mRNAs, particularly early in myogenesis (Figure 2A). Along with this delay, phase-contrast microscopy at different times during differentiation (Figure 2B, left) and immunofluorescence analysis of MYH levels at 48 h (Figure 2B, right) indicated that silencing lncFAM attenuated myotube formation and reduced myogenesis. Myotube formation by 72 h was quantified by measuring the fusion index, which dropped from ∼80% to ∼35% when lncFAM was silenced, while the average number of nuclei per myotube declined from ∼25 to ∼3 in the lncFAM siRNA-transfected cells (Figure 2C). Consistent with these changes, the expression levels of myogenic protein markers MYOG, MEF2C, and MYH (as detected by western blot analysis) were markedly reduced in lncFAM-silenced cells (Figure 2D), and creatine kinase activity, another marker of myogenesis, was similarly reduced after silencing lncFAM (Figure 2E). We confirmed these observations by employing two additional lncFAM siRNAs, both of which also reduced lncFAM, as determined by RT-qPCR analysis (Supplementary Figure S2A), and reduced myotube formation as assessed by immunofluorescence staining for MYH (Supplementary Figure S2B). Notably, silencing lncFAM (Supplementary Figure S2C) in the human myoblast line AB1167 also attenuated myotube formation, as shown by a strong reduction in MYH signals (Supplementary Figure S2D). Taken together, these results indicate that lncFAM is necessary for the timely progression of human myogenesis, as silencing lncFAM reduced many features of myogenesis, including morphological changes (myotube formation), biochemical markers (creatine kinase activity) and the levels of myogenic proteins MYOG, MEF2C and MYH.

Figure 2.

Silencing lncFAM reduced human myogenesis. (A) Ctrl siRNA or lncFAM-directed siRNA were transfected into AB678 myoblasts, and 24 h later the culture was switched to differentiation medium and RNA was collected at 16 and 36 h. RT-qPCR analysis was performed to measure the levels of lncFAM as well as differentiation markers MEF2C and MYOG mRNAs; RNA data were normalized to the levels of GAPDH mRNA, encoding a housekeeping protein. (B) Cells were transfected and processed as described in (A), and 36, 48, and 60 h into myogenic differentiation, cultures were visualized by phase-contrast microscopy (left); at 72 h into differentiation, myotubes were visualized by fluorescence microscopy to detect MYH signal (green) and nuclei were detected by DAPI staining (blue) (right). (C) In cultures processed as described in (B) at 72 h, fusion index and number of nuclei per myotube were quantified from five fields per experimental group. (D) AB678 myoblasts were transfected as in (A); 16, 24, and 48 h after inducing differentiation, the levels of early myogenic markers MYOG and MEF2C, as well as the late myogenic marker MYH and the loading control HSP90 were assessed in whole-cell lysates by western blot analysis. (E) AB678 cells were transfected with Ctrl or lncFAM siRNAs and 24 h later they were placed in differentiation medium for an additional 48 h, whereupon the levels of creatine kinase activity were measured (Materials and Methods). Data in (A, C, E) represent the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

To complement the loss-of-function data, we transduced myoblasts with a lentivirus (LV) expressing lncFAM. By 16 h of differentiation after lncFAM overexpression, the levels of myogenic markers MYOG and MEF2C mRNAs, as measured by RT-qPCR analysis, were markedly elevated (Figure 3A). Similarly, morphological evidence of myotube formation, as determined by both phase-contrast microscopy and MYH immunofluorescence staining (Figure 3B, C), revealed that myogenesis was enhanced in the presence of high levels of lncFAM. Myotube formation was further monitored by quantifying fusion indices at 48 h (which increased from ∼20% to 60%) and by measuring the number of nuclei, which increased per myotube from ∼2 to ∼9 after lncFAM overexpression (Figure 3D). Collectively, the expression levels of myogenic marker mRNAs and myotube formation increased strongly after lncFAM overexpression. In sum, loss- and gain-of-function experiments revealed that lncFAM promoted human myogenesis.

Figure 3.

Overexpression of lncFAM enhances human myogenesis. (A) Proliferating AB678 myoblasts were infected with lentiviruses, either control (LV Ctrl) or expressing lncFAM (LV lncFAM), at 1 MOI; 24 h later, they were switched to differentiation medium and the levels of lncFAM and myogenic marker transcripts MYOG and MEF2C mRNAs (normalized to the levels of control GAPDH mRNA) were measured by RT-qPCR analysis. (B, C) In cultures treated as described in (A), myotube formation at 48 h after inducing differentiation was determined by phase-contrast microscopy (B), and by immunofluorescence staining of MYH (red) along with nuclear staining using DAPI (blue) (C). (D) The fusion index and number of nuclei per myotube from cultures in (B) at 48 h into myogenesis were quantified; five fields were assessed per experiment. Data in (A, D) represent the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

Nuclear lncFAM associates with heterogenous nuclear ribonucleoprotein L (HNRNPL)

To further study the function of lncFAM in promoting human myogenesis, we first examined its subcellular localization (17). Nuclear and cytoplasmic fractions were prepared (Materials and Methods) and the distribution of lncFAM in each compartment was assessed by RT-qPCR analysis. Control RNAs preferentially localized in the cytoplasm [7SL; (48)] and the nucleus [MALAT1; (49)] were also monitored to ensure proper subcellular fractionation. As shown, lncFAM mainly localized in the nucleus in both proliferating and differentiating cultures (Figure 4A). Moreover, the smFISH analysis showed that lncFAM was predominantly present in the nucleus and its expression increased markedly in myotubes compared to myoblasts (Figure 4B) (37).

Figure 4.

Nuclear lncFAM interacts with RNA-binding protein HNRNPL. (A) Following fractionation of nuclear and cytoplasmic lysates from AB678 cultures that were proliferating (Prolif.) and differentiated for 24 h (Diff.), the levels of lncFAM, as well as control cytoplasmic lncRNA 7SL and control nuclear lncRNA MALAT1 were measured by RT-qPCR analysis. (B) smFISH analysis depicting the nuclear distribution of lncFAM and its relative abundance in AB1167 myoblasts and myotubes. (C) Schematic of the ChIRP-MS workflow. Briefly, AB678 cells were crosslinked with 3% formaldehyde for 30 min and sonicated. After incubating with four biotinylated antisense oligomers (ASOs) directed at lncFAM (or control LacZ-directed ASOs), the complexes were pulled down by using streptavidin-coated magnetic beads. Interacting proteins were then eluted for analysis by mass spectrometry. (D) Top proteins interacting with lncFAM from differentiating (48 h) AB678 cells, as identified by ChIRP-MS analysis. (E) Left, RT-qPCR analysis of the enrichment of lncFAM in each ASO pulldown group (relative to the control LacZ biotin ASO pulldown) from the ChIRP experiment. Right, western blot analysis of the enrichment of HNRNPL in lncFAM biotin-ASO samples relative to control LacZ biotin ASO; the RBP nucleolin (NCL) was included to test for background binding by other RBPs. (F) ASOs separated into two sets, Odd and Even, and incubated each set of ASOs with lysates of AB678 cells at 48 h of differentiation. The enrichment in lncFAM in the Odd and Even ASO pulldown groups by ChIRP analysis was measured by RT-qPCR analysis (left); the enrichment in HNRNPL in lncFAM biotin-Odd and biotin-Even pulldowns was detected by western blot analysis (right). (G) Analysis of the interaction of lncFAM with HNRNPL by crosslinking RNP immunoprecipitation (IP) assays using anti-HNRNPL or control IgG antibodies. The enrichment of lncFAM in the IP was determined by RT-qPCR analysis (left), using the levels of GAPDH mRNA for normalization, and the IP results were assessed by western blot analysis (right) of HNRNPL, including NCL as a control RBP. (H) In AB678 cultures undergoing myogenesis for the times shown, the levels of HNRNPL (top, including loading control GAPDH) and HNRNPL mRNA (bottom, normalized to the levels of control GAPDH mRNA) were assessed by western blot and RT-qPCR analyses, respectively. (I) AB678 cells were transfected with Ctrl siRNA or HNRNPL siRNAs (#1 or #2); 24 h later, they were placed in differentiation medium for an additional 48 h, whereupon the levels of HNRNPL, MYOG, and MYH (and loading control HSP90) were determined by western blot analysis. Data in (A, E-H) represent the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

To begin to ascribe function to lncFAM, we undertook an approach aimed at identifying the proteins that interact with it. Given that lncFAM is in the nucleus, we focused on lncFAM-associated proteins by performing Chromatin Isolation by RNA Purification combined with Mass Spectroscopy [ChIRP-MS, Figure 4C (50,51)]. Briefly, cells were crosslinked with 3% formaldehyde, lysed, and homogenized by sonication; four complementary biotinylated antisense oligonucleotides (ASOs) designed to bind to lncFAM were then added to the lysate and used to specifically pull down lncFAM-interacting complexes, including chromatin, protein, and RNA, by using magnetic streptavidin-coated beads. The pulldown efficiency is shown in Supplementary Figure S3, and the interacting proteins were further identified by mass spectrometry (MS, Supplementary Table S2). As negative controls, we used lacZ ASOs. We focused on the proteins that were selectively enriched in the lncFAM ASOs pulldown samples and not in the control lacZ ASO pulldown samples. Among them, we selected one of the top candidate proteins for further analysis, heterogenous nuclear ribonucleoprotein L (HNRNPL), which is a nuclear RNA-binding protein (RBP, Figure 4D) that was recently reported to promote myogenic differentiation by binding ncRNAs (52,53). We validated the interaction between lncFAM and HNRNPL by performing ChIRP again, this time followed by western blot analysis. As shown in Figure 4E, RT-qPCR analysis verified that lncFAM was enriched in the lncFAM biotin-ASO pulldown and not in the control lacZ biotin-ASO pulldown (Figure 4E, left), and, importantly, HNRNPL was highly enriched in the lncFAM ASO pulldown, but not in the lacZ biotin-ASO pulldown (Figure 4E, right); the RBP nucleolin (NCL), included here as a negative control because it did not bind lncFAM, was not detected in the pulldown material. To test the specificity of the biotinylated ASOs, we split them into two groups, Odd and Even, and confirmed the interaction of lncFAM with HNRNPL by ChIRP analysis. As anticipated, lncFAM was enriched in the pulldown of both Odd and Even, but not in the lacZ pulldown pool (Figure 4F, left), and HNRNPL was enriched in both lncFAM pulldowns (Figure 4F, right).

We used an orthogonal approach, namely ribonucleoprotein (RNP) crosslinking (CL) followed by immunoprecipitation (IP) analysis, to confirm the interactions between HNRNPL and lncFAM. Briefly, at 48 h into differentiation, AB678 cells were crosslinked by exposure to 0.1% formaldehyde, and RNP complexes were detected by IP using an antibody that recognizes the protein component of the RNP (anti-HNRNPL antibody) and the bound RNAs were quantified by RT-qPCR analysis. As shown in Figure 4G (left), lncFAM was highly enriched in the HNRNPL IP samples, but not in the control IgG IP samples. The successful IP of HNRNPL was visualized by western blot analysis of the IP samples (Figure 4G, right); NCL, a protein that did not bind lncFAM, was again included as a negative control.

To study if this interaction might be functionally important, we measured the levels of HNRNPL mRNA and HNRNPL protein (by RT-qPCR and western blot analyses, respectively) in AB678 cultures and found only minor changes during differentiation (Figure 4H). However, silencing HNRNPL by transfection of two different siRNAs directed at HNRNPL mRNA revealed that reducing HNRNPL also lowered the levels of markers of differentiation MYOG and MYH at 48 h of differentiation (Figure 4I); an attenuated differentiation was also observed when utilizing a third siRNA directed at HNRNPL mRNA (Supplementary Figure S4A) followed by MYH immunofluorescence staining (Supplementary Figure S4B). We also confirmed the effect of HNRNPL silencing on myogenesis in AB1167 cells. Briefly, RT-qPCR analysis indicated that the three different HNRNPL siRNAs significantly reduced HNRNPL mRNA levels (Supplementary Figure S4C) as well as myotube formation (Supplementary Figure S4D) in AB1167 cells. Together, these results indicate that human lncFAM interacts with the RBP HNRNPL, and this association appears to be important for human myogenesis.

Impact of the lncFAM–HNRNPL complex in human myogenesis

To map the region(s) of interaction between HNRNPL and lncFAM, we transcribed a series of biotinylated RNA fragments spanning the length of lncFAM, using as a template the corresponding PCR fragments (Figure 5A, top). We then incubated the purified biotinylated lncFAM fragments (A through G) with AB678 cell lysates, and pulled them down using streptavidin-coated beads. The interaction of HNRNPL with each of these fragments was tested by western blot analysis of the proteins in the pulldown material. As shown, fragments C through G had higher affinity for HNRNPL (Figure 5B; including a negative control ‘Neg’, which was beads only). Therefore, we prepared additional biotinylated fragments (each between 30 and 43 nt in length) spanning the C-G segment (positions 301–527 on lncFAM; Figure 5A, bottom). The RNA segments showing the strongest affinity for HNRNPL were found to be fragment 3 (nucleotides 377–414 on lncFAM) and fragment 6 (nucleotides 498–527 on lncFAM) by western blot analysis of the pulldown material (Figure 5C).

Figure 5.

Mapping and functional analysis of the lncFAM region where HNRNPL interacts. (A) Schematic of biotinylated RNA segments of lncFAM (fragments A through G) that were synthesized by in vitro transcription, each spanning ∼150 nt (top). Further subdivision of the lncFAM region 301–527 into an additional 6 synthetic biotinylated RNAs (fragments 1 through 6), each spanning ∼40 nt. (B, C) Whole-cell lysates from AB678 cultures at 48 h into differentiation were incubated with biotinylated fragments A through G (B) or with biotinylated fragments 1 through 6 (C); the complexes were pulled down using streptavidin beads, and the presence of HNRNPL among the bound proteins was determined by western blot analysis. ‘Input’ (lysate only); ‘Neg.’ (negative control), beads incubated with lysate but without biotinylated RNA. (D) Lentiviral vectors (LV) were prepared that expressed no insert (empty vector, EV), wild-type lncFAM [lncFAM(WT)], or lncFAM bearing a deletion in either fragment 3 or fragment 6 [lncFAM(Δ3) or lncFAM(Δ6)]. Twenty-four hours after infection with the LVs, differentiation was induced, and 24 and 48 h after inducing differentiation the levels of the overexpressed lncFAM (full-length or deletion mutants) were measured by RT-qPCR analysis and normalized to control GAPDH mRNA levels. (E) Cells were processed as in (D), and the levels of MYH protein and loading control HSP90 were assessed by western blot analysis 24 and 48 h later. (F) In cells processed as described in (D), myotube formation was assessed 72 h into differentiation in all four infection groups by fluorescence microscopy to detect MYH (red) and nuclei (blue). (G) The fusion index and number of nuclei per myotube were quantified after assessing five fields per experiment in cells processed as described in (F). Data in (D) and (G) represent the means ± SEM from at least three independent experiments. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

To test the possible impact of HNRNPL binding at fragments 3 and 6, we prepared a battery of lentiviral vectors: an empty vector control (EV), and vectors expressing either full-length wild-type [LV lncFAM(WT)], or lncFAM lacking fragments 3 or 6 [LV lncFAM(Δ3), LV lncFAM(Δ6), respectively]. These vectors expressed the inserts efficiently and comparably, as determined by RT-qPCR analysis (Figure 5D). Assessment of myogenesis by monitoring the levels of MYH revealed that overexpressing lncFAM(WT) promoted MYH expression, but in cells expressing lncFAM(Δ3) or lncFAM(Δ6), MYH levels were comparable to those in the control [LV (EV)] culture (Figure 5E). Along with these changes, myotube formation was found to increase robustly by 72 h in the population infected with LV lncFAM(WT), but remained at the level of control cultures [LV (EV)] after infecting with LV lncFAM(Δ3) or LV lncFAM(Δ6) (Figure 5F). These differences were also reflected in quantifications of fusion indices and average numbers of nuclei per myotube at 72 h; compared with control populations [LV (EV)], strong increases in these myogenic parameters were observed in the LV lncFAM(WT) populations, but these were reduced significantly if LV lncFAM(Δ3) or LV lncFAM(Δ6) were expressed instead (Figure 5G). These data collectively support the notion that lncFAM associates through specific RNA regions with HNRNPL, and the interaction contributes to myogenesis.

Nuclear-retained lncFAM promotes MYBPC2 transcription

Given that nuclear lncRNAs often function as regulators of the transcription of nearby genes (22,25), we investigated if genes near the locus where lncFAM resides (chr19:50486810–50487638; hg19, Ensembl v82) were affected by lncFAM expression levels. From the UCSC genome browser, six genes, MYBPC2, FAM71E1, EMC10, JOSD2, ASPDH and LRRC4B, were localized in close proximity to lncFAM (Figure 6A). To test if lncFAM regulates these nearby genes, we silenced lncFAM and the levels of expression of mRNAs encoded by these genes were measured by RT-qPCR analysis. As shown, among these six genes, only the levels of MYBPC2 mRNA declined significantly after lncFAM silencing, while the levels of FAM71E1, EMC10, JOSD2, ASPDH or LRRC4B mRNAs did not (Figure 6B). Skeletal muscle myosin-binding protein C (MYBP-C) is an important component of the sarcomere, the smallest unit for modulating muscle contractility (54). Three isoforms of MYBP-C are expressed in slow-twitch and fast-twitch skeletal muscle, as well as in cardiac muscle [encoded by MYBPC1, MYBPC2 and MYBPC3 (55–58)]. We confirmed that MYBPC2 mRNA levels increased at 24 h of differentiation in AB678 myoblasts by RNA-seq analysis (Supplementary Figure S5A) and extended our measurement of MYBPC2 mRNA levels by RT-qPCR analysis to later time points (Figure 6C). The levels of MYBPC2 protein, as assessed by western blot analysis, also increased dramatically with differentiation (Figure 6D). The changes in MYBPC2 mRNA levels appeared to be driven at least in part by increased transcription of the MYBPC2 gene, as the changes in MYBPC2 pre-mRNA levels, which served as a surrogate of the rate of transcription of the MYBPC2 gene, increased with the same pattern as MYBPC2 mRNA (Figure 6E; a different set of primers verified this pattern, Supplementary Figure S5B, C).

Figure 6.

Silencing of lncFAM attenuates transcription of nearby MYBPC2 gene. (A) Six genes (gray) in the vicinity of the lncFAM locus (red) on human chromosome 19. Arrows, direction of transcription. (B) lncFAM was silenced in AB678 cells, and 24 h later differentiation was induced for an additional 24 h. The levels of expression of MYBPC2, JOSD2, FAM71E1, EMC10, ASPDH and LRRC4B mRNAs were quantified by RT-qPCR analysis at the times shown after inducing differentiation; GAPDH mRNA were also quantified and used for normalization. (C–E) AB678 human myoblasts were cultured in differentiation conditions and the levels of MYBPC2 mRNA (C), MYBPC2 protein (D), and MYBPC2 pre-mRNA (E) were determined by RT-qPCR analysis, normalized to GAPDH mRNA levels (C,E), and western blot analysis including GAPDH as a loading control (D). (F) From AB678 cells transfected and induced to differentiate as described in (B) for 16 h, RNA was collected, subjected to RNA-seq analysis, and represented (volcano plot); MYBPC2 mRNA and other RNAs are highlighted. (G) Cells were transfected and induced to differentiate as explained in (B) and the levels of MYBPC2 pre-mRNA were quantified by RT-qPCR analysis at 16, 24 and 36 h after inducing differentiation. (H) AB678 myoblasts were infected with either LV Ctrl or with LV lncFAM (1 MOI) and 24 h later, they were switched to differentiation medium; the levels of MYBPC2 mRNA levels were measured by RT-qPCR analysis at 16 and 48 h (and normalized to the levels of GAPDH mRNA). Data in (B,C,E,G,H) represent the means ± SEM from at least three independent experiments. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001. Other data are representative of three or more biological replicates.

In further agreement with the idea that lncFAM promoted MYBPC2 transcription, RNA-seq data confirmed that MYBPC2 mRNA was significantly reduced after silencing lncFAM in human myoblasts at 16 and 48 h into the differentiation program (Figure 6F, Supplementary Figure S5D). Additionally, the levels of the nascent transcription, as assessed by measuring MYBPC2 pre-mRNA levels by RT-qPCR analysis, also declined in myoblasts after silencing lncFAM (Figure 6G). Gene Ontology (GO) analysis of the altered transcriptome after silencing lncFAM for 16 or 48 h was carried out to begin to understand the biological function of lncFAM. As shown in Supplementary Figure S6, we found that mRNAs encoding components of the sarcomere (the contractile unit of skeletal muscle) were significantly reduced after silencing lncFAM, supporting the notion that lncFAM may play an important role in muscle contraction, and underscoring the established role of MYBPC2 in sarcomere function and muscle contractility.

Conversely, overexpression of lncFAM in AB678 myoblasts that had been differentiating for 24 h led to a rise in the levels of MYBPC2 mRNA, as determined by RT-qPCR analysis; the increase was particularly apparent at 16 h, while at 48 h the expression levels of MYBPC2 mRNA may have reached saturation (Figure 6H). To help interpret the above findings, we determined the biological role of MYBPC2 in human myoblasts. MYBPC2 was silenced separately using two different siRNAs; as shown in Supplementary Figure S7A, the levels of MYH, a late marker of differentiation, were suppressed by 72 h, when MYBPC2 was silenced, even if silencing was incomplete. Similarly, myotube formation was significantly reduced when MYBPC2 was silenced for 24 h in AB678 myoblasts before they were switched to differentiation medium and cultured for an additional 72 h (Supplementary Figure S7B). Quantification of myotubes indicated that silencing of MYBPC2 reduced the fusion index from ∼80% to ∼70% after 72 h of differentiation, and that the average number of nuclei per myotube declined from ∼12 to ∼3 (Supplementary Figure S7C). A third MYBPC2 siRNA (Supplementary Figure S8A) confirmed that silencing MYBPC2 reduced myotube formation at 72 h, as measured by MYH immunofluorescence (Supplementary Figure S8B); all three MYBPC2 siRNAs had similar effects on AB1167 cells (Supplementary Figure S8C, D). In sum, lncFAM facilitates the transcription of MYBPC2 mRNA, which in turn increases MYBPC2 protein levels and promotes myogenesis.

lncFAM modulates HNRNPL binding to the MYBPC2 locus

We then investigated the molecular mechanism whereby lncFAM promotes the transcription of MYBPC2 mRNA. HNRNPL was found to associate with several lncRNAs to promote the transcription of specific genes (52,59,60). To test the hypothesis that the lncFAM-HNRNPL complex may be needed to activate MYBPC2 transcription during human myogenesis, we first determined if HNRNPL was able to modulate the levels of MYBPC2 mRNA by RT-qPCR analysis. As shown in Figure 7A, siRNA-mediated silencing of HNRNPL in AB678 cells for 24 h (as shown in Figure 4G), reduced the levels of MYBPC2 mRNA by 16 and 48 h into differentiation (Figure 7A, middle). The levels of MYBPC2 pre-mRNA were also significantly lower at both times of differentiation (Figure 7A, right), supporting the notion that transcription of the MYBPC2 gene was reduced when HNRNPL was silenced.

Figure 7.

lncFAM promotes HNRNPL binding to the MYBPC2 promoter region. (A) AB678 cells were transfected with control siRNA or HNRNPL siRNA; 24 h later they were placed in differentiation medium for an additional 16 or 48 h, whereupon the expression levels of HNRNPL mRNA (to monitor silencing efficiency), MYBPC2 mRNA, and MYBPC2 pre-mRNA were measured by RT-qPCR analysis. GAPDH mRNA levels were quantified for normalization. (B) Schematic of the segments of the MYBPC2 promoter locus amplified using specific primer pairs. (C) Proliferating AB678 myoblasts as well as AB678 cultures induced to differentiate into myotubes for 48 h. After harvesting, chromatin IP (ChIP) was performed using anti-HNRNPL antibody, followed by PCR analysis across four MYBPC2 promoter regions (or ACTN promoter region, as a control) in proliferating relative to differentiated cultures. Input: 10% of lysates. (D) AB678 myoblasts were transfected with Ctrl or lncFAM siRNAs; 24 h later, they were placed in differentiation medium for 48 h, whereupon ChIP-qPCR analysis was performed as explained in (C). Data in (A,C,D) represent the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

We next performed chromatin IP (ChIP) followed by qPCR detection of four MYBPC2 promoter regions (Figure 7B). Two different conditions were analyzed, proliferating myoblasts and differentiated myotubes (at 48 h), when lncFAM was strongly elevated (Figure 1F). ChIP-qPCR analysis, performed using an antibody directed at HNRNPL (normalized to Input), indicated that HNRNPL associated with the most proximal MYBPC2 promoter region [region 1 (R1)], closest to the transcription start site (TSS), but the other regions, including R2-4 and a region from an unrelated promoter (ACTN), were not enriched. At R1, binding was elevated in cultures 48 h into differentiation (‘Differentiated’) relative to Proliferating cultures (Figure 7C). To further test if lncFAM promotes the binding of HNRNPL to the MYBPC2 promoter, we transfected siRNAs to silence lncFAM 24 h before placing cells in differentiation medium, and 48 h later we evaluated HNRNPL binding to the MYBPC2 promoter by ChIP-qPCR analysis. As shown in Figure 7D, the interaction between HNRNPL and the MYBPC2 promoter region R1 was significantly reduced in lncFAM-silenced AB678 cells, further supporting the view that lncFAM recruited HNRNPL to the region ∼100 nt upstream of the MYBPC2 TSS (Figure 7B); western blot analysis confirmed the successful IP of HNRNPL in ChIP samples (not shown). Besides promoter regions, lncRNAs can also recruit transcriptional regulators to enhancer elements. We therefore examined if lncFAM could regulate HNRNPL binding to the enhancer regions of the MYBPC2 gene [as identified using the registry of candidate cis-regulatory Elements (cCREs) derived from ENCODE; Supplementary Figure S9A]. As shown by ChIP-qPCR analysis, HNRNPL bound the MYBPC2 enhancer region at sites E1961732 and E1961741 in differentiated conditions; other enhancer sites were either not enriched (Supplementary Figure S9B) or not detectable.

Ectopic expression of MYBPC2 partially rescues the loss of myogenesis by silencing lncFAM

To further determine if lncFAM silencing reduced myogenesis by lowering MYBPC2 production, we investigated if ectopically overexpressing MYBPC2 was able to rescue the effect of lncFAM silencing on myogenesis. We overexpressed MYBPC2 in AB678 myoblasts using a lentivirus, and 24 h after infection we silenced lncFAM by transfecting siRNAs. Twenty-four hours after that, cells were incubated in differentiation medium for 24 h, and the levels of lncFAM (Figure 8A, left) and MYBPC2 mRNA (Figure 8A, right) were measured by RT-qPCR analysis (Figure 8A); 72 h into differentiation, myotube formation was determined by assessing the expression patterns of myotube marker MYH using immunofluorescence (red) (Figure 8B). As anticipated, silencing lncFAM [lncFAM siRNA + LV (EV)] reduced MYH signal intensities compared to the Ctrl siRNA population [Ctrl siRNA + LV (EV)]. Importantly, however, ectopic MYBPC2 overexpression partially rescued the effects of silencing lncFAM, as determined by the increased MYH signals in the (lncFAM siRNA + LV MYBPC2) population relative to the [lncFAM siRNA + LV (EV)] population (Figure 8B). In addition, the fusion index (Figure 8C, left) and the numbers of nuclei per myotube (Figure 8C, right) were also partially restored when re-expressing MYBPC2 ectopically, as determined by comparing the (lncFAM siRNA + LV MYBPC2) to the [lncFAM siRNA + LV (EV)] populations. Taken together, the reduced myogenesis in cells with silenced lncFAM is partially rescued by ectopically overexpressing MYBPC2. Thus, we propose a model in which the increased expression of the novel human lncRNA lncFAM promotes myotube formation at least in part by binding and recruiting HNRNPL to the MYBPC2 regulatory regions (promoter and enhancer), and transcriptionally increasing the production of the myogenic protein MYBPC2 during the myogenic developmental program (Supplementary Figure S10).

Figure 8.

Effect of ectopic overexpression of MYBPC2 on myotube formation after silencing lncFAM. (A) AB678 myoblasts were infected with lentiviruses (LV) expressing MYBPC2 and 24 h later, lncFAM siRNAs were transfected; 24 h after that, AB678 cells were induced to differentiate and 24 h into differentiation, the levels of lncFAM (left) and MYBPC2 mRNA (right) were assessed by RT-qPCR analysis. (B) Cells were processed as in (A) and 72 h into differentiation, myotube formation was assessed by immunofluorescent detection of the myotube marker MYH (red); DAPI (blue) was used to visualize nuclei. (C) Quantification of myotube formation by fusion index and number of nuclei per myotube from five fields per experiment. Data in (A,C) represent the means ± SEM from three or more biological replicates. Significance was established using Student's t-test. *P < 0.05; **P < 0.01; ***P < 0.001.

DISCUSSION

We report the identification and analysis of lncFAM, a novel human lncRNA expressed from the lncFAM71E1-2:2 locus in human myoblasts, but not in other cell types studied. As we discuss here, lncFAM is abundant, mainly nuclear, and critically implicated in promoting human myogenesis. To systematically annotate the lncFAM sequence in skeletal muscle, we performed long-read RNA sequencing using the Oxford Nanopore technology. In total, we detected 6 different isoforms, five novel and one matching the annotated transcript. The annotated lncFAM RNA was the most abundant isoform with the exception of one isoform that was twice as abundant and slightly longer, although it fully encompassed the annotated isoform studied here; the annotated lncFAM comprised one-third of the total pool of six lncFAM isoforms expressed (Supplementary Table S3; GSE215343). Molecular analysis of lncFAM revealed its specific association with HNRNPL to form a lncRNP complex that facilitates HNRNPL binding to the MYBPC2 promoter region, and increases the transcription of MYBPC2 mRNA, in turn elevating the production of MYBPC2 expression and supporting myogenesis.

Skeletal myogenesis is a dynamic and complex process whereby myoblasts differentiate into multinucleated myotubes during embryonic development and in regenerating adult muscle (61). Impairments in different segments of myogenic progression can lead to muscle disease, such as muscle atrophy, muscle dystrophy, and sarcopenia (62,63). There is growing recognition that lncRNAs may be implicated in a number of muscle diseases. For example, lnc-31, a lncRNA abundant in the skeletal muscle of Duchenne muscular dystrophy (DMD) patients, was found to promote myoblast proliferation and delay differentiation (64), while lincMD1, a lncRNA that is highly abundant in differentiated myotubes and promotes myogenesis, is found in low abundance in myoblasts derived from persons with DMD (28). In a mouse model of muscle atrophy, overexpression of murine lncRNA MAR1, a promoter of myogenesis, successfully restored muscle loss in mice (65).

Nuclear lncRNAs participate in a variety of processes, including chromatin organization and promotion or repression of transcription factor activity (2,22,66). In examples from cancer biology, lncTCF7 recruits the chromatin-remodeling complex SWI/SNF to the TCF7 promoter in hepatocellular carcinoma (67), and lncRNA BCAR4 interacts with RBPs SNIP1 and PNUTS, restoring p300-dependent histone acetylation and RNA-polymerase II activity, while also facilitating breast cancer metastasis (68). Some lncRNAs function at DNA regions proximal to the site of lncRNA transcription; for example, low-dose irradiation induced the lncRNA PARTICLE to form a DNA-lncRNA triplex at the MAT2A promoter and recruit transcription-repressive complex proteins for methylation in breast cancer (69). Other lncRNAs function at distant areas of the genome; for example, the lncRNA PURPL interacts with the protein MYBBP1A, thereby preventing the formation of the p53-MYBBP1A complex that promotes p53 stability, and in turn causing p53 levels to decline, an effect that was linked to tumorigenicity in colorectal cancer (70).

Evidence is also accumulating that RBPs associating with lncRNAs may themselves elicit transcriptional regulation. In particular, the complexes that result from the association of HNRNPL with different lncRNAs can regulate the transcription of different genes through binding to different promoter regions. For example, in macrophages, lncTHRIL promotes TNF production by recruiting HNRNPL to the TNF promoter (59), and in retinoblastoma and bladder cancer cells, lncRNA RBAT1 associates with HNRNPL at the E2F3 promoter region and promotes the transcription of E2F3 mRNA (60). Interestingly, one study found that HNRNPL binds to lncRNAs functioning as super enhancer RNAs (seRNAs), including seRNAs that were transcribed by MYOD, and drives myogenic differentiation in mouse (52). In this regard, lncFAM-HNRNPL is one of the first complexes reported to control transcription in human myogenesis. Given the high abundance of lncFAM during myogenic differentiation, we anticipate that future studies will uncover other interacting partners and functions for this lncRNA in myogenesis.

Interestingly, besides HNRNPL, myogenic transcription factors such as MYOD and MYOG also have a potential binding affinity for the MYBPC2 promoter region, as determined using the PROMO tool (71) to predict transcription factor binding sites. Although the MYBPC2 promoter region harbors 4 and 11 binding sites for MYOD and MYOG, respectively, whether the lncFAM-HNRNPL complex may function together with MYOD and MYOG to promote MYBPC2 transcription is not known at present. Our initial results indicated that MYOD and MYOG do not bind lncFAM, as they were not identified by ChIRP-MS analysis, but perhaps alternative or less stringent experimental forms of detection may reveal interactions with lncFAM, possibly transient and/or as part of large complexes.

As our studies advance, it will be important to explore whether lncFAM influences myogenesis in animal models of muscle regeneration. Unfortunately, there is no mouse lncFAM ortholog, in keeping with the poor conservation of many lncRNAs across species. However, given growing evidence that primate-specific lncRNAs can affect mouse biological functions (72,73), future studies will examine whether lncFAM can promote mouse myogenesis in culture, as well as mouse muscle development, strength, and regeneration after injury. Complementary studies will also address whether an as-yet unknown mouse lncRNA performs a function homologous to that of lncFAM to promote the transcription of Mybpc2 mRNA during murine myogenesis, as well as whether HNRNPL is similarly implicated in this paradigm. Further analysis of the function of MYBPC2 in myogenesis, as well as investigation of the mechanisms that control MYBPC2 production in this process are also warranted. In this regard, in mouse myogenesis, Mybpc2 mRNA levels significantly increased after silencing the mouse lncRNA SYISL (74) although it is not yet known if HNRNPL is involved in this regulation. Access to experimental animal models and a deeper understanding of how lncFAM and HNRNPL regulate MYBPC2 production during myogenesis, will permit us to evaluate if this paradigm can be exploited in pathologies of human muscle development.

DATA AVAILABILITY

The RNA-seq data were deposited in GEO (https://www.ncbi.nlm.nih.gov/geo/) under accession number GSE202793 (41) and SuperSeries GSE215343.

SUPPLEMENTARY DATA

Supplementary Data are available at NAR Online.

FUNDING

NIH IRP [Z01-AG00394]. Funding for open access charge: NIA IRP, NIH.

Conflict of interest statement. None declared.