Protein O-Fucosyltransferase 1 Expression Impacts Myogenic C2C12 Cell Commitment via the Notch Signaling Pathway

Abstract

The Notch signaling pathway plays a crucial role in skeletal muscle regeneration in mammals by controlling the transition of satellite cells from quiescence to an activated state, their proliferation, and their commitment toward myotubes or self-renewal. O-fucosylation on Notch receptor epidermal growth factor (EGF)-like repeats is catalyzed by the protein O-fucosyltransferase 1 (Pofut1) and primarily controls Notch interaction with its ligands. To approach the role of O-fucosylation in myogenesis, we analyzed a murine myoblastic C2C12 cell line downregulated for Pofut1 expression by short hairpin RNA (shRNA) inhibition during the time course of differentiation. Knockdown of Pofut1 affected the signaling pathway activation by a reduction of the amount of cleaved Notch intracellular domain and a decrease in downstream Notch target gene expression. Depletion in Pax7⁺/MyoD⁻ cells and earlier myogenic program entrance were observed, leading to an increase in myotube quantity with a small number of nuclei, reflecting fusion defects. The rescue of Pofut1 expression in knockdown cells restored Notch signaling activation and a normal course in C2C12 differentiation. Our results establish the critical role of Pofut1 on Notch pathway activation during myogenic differentiation.

INTRODUCTION

In response to injury, adult skeletal muscle has a remarkable ability to regenerate through skeletal muscle adult stem cells called satellite cells. They participate in postnatal muscle growth and regeneration. When activated by stimuli such as injury or exercise, satellite cells enter the cell cycle and begin to proliferate (1). Most cells commit to a myoblast cell fate for fusion and fiber formation, while some participate in the self-renewal of satellite cells. After birth, cell commitment to a myogenic program is regulated by the expression of Pax3 and Pax7, which control activation of the myogenic determination factor MyoD (2, 3) and allows Myog expression, necessary for the formation of multinucleated cells (4). Mice knocked out for Pax7 completely lack satellite cells, and their skeletal muscle mass is severely impacted (5). In MyoD^−/− mice, muscle regeneration is also impaired due to a defect in the ability of satellite cells to initiate the process of differentiation (6). Deletion of Pax7 in mouse myoblasts (MB) was shown to diminish the expression of MyoD by 25% but had no impact on Myog (7). Thus, the ratio of Pax7 to MyoD is critical in cell fate determination (8). Quiescent satellite cells were demonstrated to be Pax7⁺/MyoD⁻, whereas proliferative cells were Pax7⁺/MyoD⁺, and differentiated cells were Pax7⁻/MyoD⁺. In vitro, myogenic cell lines like murine C2C12 cells are able to reproduce different steps of muscle cell differentiation, similar to the reactivation of satellite cells (9). Distinct cell types can be followed during C2C12 differentiation: proliferating myoblasts, multinucleated cells called myotubes (MT) expressing muscle-specific structural proteins like myosin and troponin, and mononucleated cells, also known as reserve cells (RC). These encompass two distinct populations, where some are committed to self-renew the quiescent progenitors and others are intended for future fusion processes (10).

Among genes controlling the behavior and fate of stem cells, Notch signaling pathway members are essential for the proper progress of myogenesis (11, 12). The canonical Notch signaling pathway is initiated by the interaction between a member of the DSL (for Delta, Serrate, and Lag-2) ligand family and a member of the Notch transmembrane receptors (13). Four receptors (Notch1 to Notch4) and five ligands (Dll1, Dll3, Dll4, Jagged1, and Jagged2) are known in mammals; however, the reason why a receptor interacts with a specific ligand and what precise biological functions they control are still unclear (11). Binding of ligands to Notch receptors leads to an enzymatic cleavage of Notch, which is released in the cytoplasm as the active form of the protein, also called NICD (Notch intracellular domain), to finally translocate into the nucleus (14). Activation of Notch target genes, such as the Hes and Hey family members of basic helix-loop-helix (bHLH) transcription factors, inhibits myogenic differentiation (15). In C2C12, this inhibition results from two molecular mechanisms. In a CBF1/RBP-Jκ-dependent mechanism, NICD switches CBF1/RBP-Jκ from a transcriptional repressor to an activator inducing Hes1 transcription and the subsequent decrease of MyoD (16). A CBF1-independent mechanism contributes to a more general cellular differentiation and does not antagonize MyoD activity (17–19). The ratio between cells intended to fuse and reserve cells was demonstrated to be controlled by the Notch signaling pathway, as well as the activation of reserve cells (10). Furthermore, NICD directly regulates Pax7 expression through CBF1/RBP-Jκ in satellite cells, and MyoD^−/− mouse myoblasts upregulate Pax7 due to the activated Notch pathway (8). As a cross-inhibitory interaction between Pax7 and MyoD exists, every change in the relative amount of transcriptional factors, partly controlled by Notch activity, will affect cell fate determination (20). Numerous actors participate in the modulation of Notch pathway activation (11). For example, the expression of ligands and Notch receptors on the same cell can attenuate the signaling in a cell-autonomous manner. In C2C12 cells, the asymmetrical shedding of Dll1 ligands with more ADAM (a disintegrin and metalloprotease)-mediated cleavages in reserve cells (Pax7⁺) than in myotubes (Pax7⁻) participates in the cell determination (9).

The phenotype of Pofut1^−/− mouse embryos is more severe than the Notch1-null phenotype and resembles embryos lacking downstream components of Notch signaling, like CBF1/RBP-Jκ (21). Interactions between Notch receptors and ligands depend on the activity of the protein O-fucosyltransferase 1(Pofut1) (22), which O-fucosylates the C²X₄(S/T)C³ (where C² and C³ are the second and third cysteines, respectively) consensus motif of some epidermal growth factor (EGF)-like domains (23). Both active Pofut1 and O-fucosylation of Notch are required for the canonical Notch signaling by Delta1 or Jagged1 (24). Mutation on the O-fucosylation site in mouse Notch1 EGF12 results in a hypomorphic allele affecting the Notch-ligand interaction (25). Notch signaling is also influenced by β1,3-N-acetylglucosaminyltransferases of the Fringe family (Lunatic, Manic, and Radical Fringe) (26, 27) and a β1,4-galactosyltransferase (28), which elongate the O-linked fucose. The presence of a terminal sialic acid was shown on Notch1 O-fucosylglycan (29), but no information was reported in favor of its involvement in Notch signaling. A global transcriptomic approach based on glycogene analysis brought our attention to these enzymes in the context of the onset of myogenesis in C2C12 (30). The inverting glycosyltransferase Pofut1 was first characterized in mammalian CHO cells (31), and subsequent investigations demonstrated its presence in all principal metazoan genomes as a single-copy gene (32). Its expression is ubiquitous in the organism and confirmed to be present in skeletal muscles (22, 32). Unlike the Golgian fucosyltransferases involved in N-glycosylation, Pofut1 is an endoplasmic reticulum (ER)-resident glycoprotein (33). To identify the role of Pofut1 during myogenesis, a stable murine C2C12 cell line with knockdown of Pofut1 (Po⁻) was created. Semiquantitative real-time reverse transcription-PCR (RT-PCR) and Western blot analyses were performed to profile the expression of Notch signaling actors and some key myogenic players during differentiation of C2C12 cells. Phenotypic studies and coimmunostaining experiments were also completed. Our results provide evidence that Po⁻ cells, compared to wild-type C2C12 cells, present a disturbed myogenic program with an increased fusion index and earlier expression of myogenic regulatory factors (MRFs), resulting in depletion of progenitor cells. The peculiar Pofut1 knockdown C2C12 phenotype is linked to an attenuation of the Notch signaling pathway. In disturbing the ratio between Pax7 and MyoD, it provokes an earlier differentiation with impaired progression into the myogenic process.

MATERIALS AND METHODS

C2C12 cell culture.

The C2C12 cell line, established from the leg muscle of an adult C3H mouse (American Type Culture Collection [ATCC], Manassas, VA), was cultured in a growth medium (GM) with Dulbecco's modified Eagle's medium (DMEM; Gibco, Life Technologies, Carlsbad, CA) supplemented with 10% fetal calf serum (Eurobio, Courtaboeuf, France), 4 mM l-glutamine, 50 units/ml penicillin, and 50 μg/ml streptomycin (at 37°C and 5% CO₂). Cells were plated at a density of 1.5 × 10⁴ cells/cm². After 48 h, growth medium was removed, and differentiation was induced by the addition of differentiation medium (DM), which is DMEM supplemented with 2% horse serum, 4 mM l-glutamine, 50 units/ml penicillin, and 50 μg/ml streptomycin. Medium was routinely changed every 24 h. For each sample, cells were harvested after trypsinization (0.125% trypsin, 0.125 mM EDTA) for 5 min at 37°C. Samples named myoblasts (MB) corresponded to C2C12 cells at 0 h, defined by replacement of GM with DM, and myoblasts in differentiation (MBd) corresponded to C2C12 cells from 12 h to 48 h in DM. Reserve cells (RC) were isolated using a short trypsinization (0.1% trypsin, 0.1 mM EDTA; 30 s) that specifically removed myotubes (MT) and left only RC adherent to the flask (35). The two cell types were collected at 72 h and 120 h after DM addition.

Generation of Pofut1 knockdown C2C12 cell line.

The Pofut1 knockdown was obtained using electroporation by stably transfecting C2C12 cells with a pSilencer 2.1-U6-hygromycin construct containing an insert encoding a small interfering RNA (siRNA) directed against the mouse Pofut1 exon 3 (kind gift of P. Stanley, Albert Einstein College of Medicine, New York, NY). Selection for Pofut1 short hairpin RNA (shRNA) C2C12 clones was performed using 500 μg/ml hygromycin B (Thermo Fisher Scientific, Waltham, MA). The Pofut1 shRNA C2C12 clone 5 was chosen for further study and is here referred to as the Po⁻ cell line. A control C2C12 cell line (Ctrl shRNA) was obtained by electroporation of a pSilencer 2.1-U6-neomycin vector containing a negative-control siRNA template (Ambion, Life Technologies) into C2C12 cells. It was selected with a neomycin concentration of 1 mg/ml (Roche Applied Science, Mannheim, Germany). Electroporation-mediated transfections were performed on C2C12 proliferating myoblasts using a GenePulserX cell transfection system (Bio-Rad, Hercules, CA). Briefly, 3 × 10⁶ cells were suspended in 1 ml of phosphate-buffered saline (PBS) containing 3 μg of plasmid and then electroporated at 340 V for 13 ms. The electroporated cells were cultured in 100-mm petri dishes in GM and selected with the appropriate antibiotics.

Generation of Pofut1 knockdown C2C12 cell lines reexpressing Pofut1.

The Pofut1 knockdown cell lines reexpressing Pofut1 were obtained by exogenous expression of the Pofut1 mutated coding sequence (CDS) using pLJM1-EGFP (where EGFP is enhanced green fluorescent protein) lentiviral vector (Addgene, Inc., Cambridge, MA). Three knock-in lentivectors were generated by replacing EGFP with the mouse Pofut1 (no mutation, 0M) or a Pofut1 mutated CDS (with one or three mutations, indicated as 1M or 3M, respectively). The mutagenesis was performed by PCR to generate pLJM1-Pofut1 expression vectors that contained one (1M) or three (3M) silent mutations and checked by sequencing. The mutations were introduced in exon 3 of the Pofut1 gene at the 21-bp Pofut1 siRNA target sequence. For virion production, packaging HEK293T cells were seeded at 1 × 10⁶ into a 100-mm tissue culture dish in GM without antibiotics and incubated for 24 h (at 37°C and 5% CO₂). At 50% of cell confluence, cells were washed with 1× PBS, and 4 mM l-glutamine-supplemented DMEM was added to proceed to transfection. For each pLJM1-Pofut1 construct, 2.6 μg was cotransfected with 26 μl of packaging vector using Mission lentiviral packaging mix (Sigma-Aldrich Corp., St. Louis, MO) and 16 μl of Lipofectamine reagent (Invitrogen, Life Technologies, Carlsbad, CA) according to the manufacturer's instructions. Cells were incubated for 16 h, and then the medium was changed. After 24 h of virion production, the medium was removed, centrifuged for 5 min at 3,000 rpm, and filtered through a 0.45-μm-pore-size filter. To generate stable Pofut1-reexpressing Po⁻ cell lines, 3 × 10⁵ cells were seeded into six-well plates. These cells were infected 24 h later with viral particles using 8 μg/ml of Polybrene. Po⁻ cell lines that stably inserted lentivectors were selected 48 h after infection with 2 μg/ml puromycin over 5 days.

Phenotypic studies.

Cells were fixed in PBS containing 4% paraformaldehyde (PFA) for 20 min and dried overnight in 70% ethanol. Mono- and multinucleated cells were counted after hematoxylin-eosin staining in three independent experiments. For each sample, six randomized visual fields at a magnification of ×100 were studied using ImageJ, version 1.45s, software (36). The fusion index was calculated by dividing the number of nuclei in myotubes by the total number of nuclei. Myotubes were categorized into five different groups according to the number of nuclei per myotube (<5, 5 to 10, 10 to 20, 20 to 40, and >40).

Semiquantitative real-time RT-PCR.

Total RNA was extracted using an RNeasy minikit (Qiagen, Inc., Hilden, Germany). Integrity and quantity of total RNA were measured using an Agilent 2100 Bioanalyser (Agilent Technologies, Santa Clara, CA) and a NanoDrop 1000 instrument (NanoDrop Technologies, Wilmington, DE), respectively. A high-capacity cDNA reverse transcription kit (Applied Biosystems, Life Technologies) was used for the conversion of 10 μg of total RNA to single-stranded cDNA suitable for quantitative PCR applications. Gene expression was determined by semiquantitative real-time RT-PCR on an ABI Prism 7900 Sequence Detector System (SDS) using the TaqMan probe-based chemistry (Applied Biosystems). 6-Carboxyfluorescein (6-FAM) was used as a reporter. The amplification reactions were performed with 2 ng of cDNA according to the manufacturer's instructions. A TaqMan low-density array microfluidic card (Applied Biosystems) preloaded with 20 probes was used. Validation of probes and primer specificities were assessed by Applied Biosystems (data available upon request).

Data analysis.

Gene expression data were collected using the SDS, version 2.2.2, software (Applied Biosystems). The comparative threshold cycle (C_T) method (ΔΔC_T) (37) was used to quantify the relative abundance of each mRNA. This method uses a calibrator sample in order to enable comparison of the gene expression levels of different samples. When the proliferation stage was studied, the calibrator sample corresponded to cells just after plating, whereas in the differentiation stage, we used C2C12 cells just before the addition of DM (0 h) as the calibrator sample. Relative quantity (RQ) values reflect expression changes in the sample of interest compared to the calibrator sample, after normalization with 18S and Gapdh reference genes. They were calculated only for a threshold cycle (C_T) lower than 37 (data available upon request). Statistical analyses were performed by comparison of each differentiation time relative to 0 h of wild-type (WT) C2C12 cells, which was set as 1.

Western blot analysis.

Proteins were extracted from C2C12 cells in radioimmunoprecipitation assay (RIPA) lysis buffer containing 50 mM Tris-HCl, pH 8, 150 mM NaCl, 0.5% sodium deoxycholate, 1% NP-40, 0.1% SDS, and protease inhibitor cocktail (Complete; Roche Diagnostics). The suspension was then centrifuged (12,000 × g for 10 min) at 4°C. Supernatant was recovered, and protein quantity was estimated using a bicinchoninic acid (BCA) protein assay (Sigma-Aldrich Corp.) with bovine serum albumin as a standard. Fifty or hundred micrograms of extracted proteins was separated under denaturing and reducing conditions on a 12% polyacrylamide gel and then transferred to a Hybond C-Extra Nitrocellulose membrane (GE Healthcare, Buckinghamshire, United Kingdom). Nonspecific antibody binding was prevented using blocking buffer for 1 h at room temperature with 5% bovine serum albumin in 0.1% Tween 20-Tris-buffered saline (TBST; 50 mM Tris-HCl, 150 mM NaCl, 0.1% Tween 20, pH 7.6) for cleaved NICD and desmin detection or with 5% nonfat dry milk in 0.1% TBST for the other antibodies. Incubation with the primary antibodies diluted in blocking buffer was performed overnight at 4°C, followed by three washes of 5 min each in 0.1% TBST. Blots were incubated with horseradish peroxidase (HRP)-coupled secondary antibody in 2.5% nonfat dry milk in 0.1% TBST for 1 h at room temperature. After three washes in 0.1% TBST, immunoblots were revealed by enhanced chemiluminescence using BM Chemiluminescence Western blotting substrate (peroxidase [POD]) (Roche Applied Science) and exposed to a film (Hyperfilm ECL; GE Healthcare). The following antibodies were used: cleaved Notch1 (Val1744) from Cell Signaling Technology (Danvers, MA), desmin (Y-20) from Santa Cruz Biotechnology, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) from R&D Systems (Minneapolis, MN), MyoD1 (clone 5.8A) and MyoG (clone F5D) from Dako Cytomation (Glostrup, Copenhagen, Denmark), myosin (skeletal, fast MY-32) from Sigma-Aldrich Corp., Pax7 from Developmental Studies Hybridoma Bank (University of Iowa, IA), and Pofut1 rabbit antiserum raised against the peptide CNLAPSHWPPEKRVAY and purified by Agro-Bio (La Ferté St-Aubin, France). For quantification, analysis of band intensities was carried out using ImageJ, version 1.45s, software. All band intensities were normalized to the corresponding Gapdh or pan-cadherin band intensity. Then, band intensities were compared to the WT C2C12 myoblast level at 0 h, which was set as 1, except for the experiment shown in Fig. 7 where Po⁻ cells were used as reference.

FIG 7.

Rescue of the Pofut1 knockdown C2C12 cell line. (A) Schematic representation of the pLJM1 lentiviral vector expressing the coding region of WT Pofut1 (0M) or Pofut1 containing one (1M) or three (3M) silent mutations. Mutations are indicated in italic and bold letters. Pofut1, NICD, and Gapdh were detected by Western blotting in Po⁻, 0M, 1M, 3M, and WT C2C12 cells. Band intensities were calibrated relative to Po⁻ cells. (B) Photographs of Po⁻ and Pofut1 rescue cell lines during differentiation. (C) Percentages of fusion for all cell lines relative to Po⁻ cells. (D) Focus on multinucleated cells characteristic of each cell line. (E) Western blot analyses of Pax7, MyoD, and Myog in myoblasts for all cell lines. Histograms represent quantification of Pax7, MyoD, and Myog levels in 0M, 1M, 3M, and WT C2C12 cells normalized to Gapdh and calibrated with Po⁻.*, P < 0.05; **, P < 0.01; ***, P < 0.001.

Biotinylation of cell surface proteins and Notch detection.

The detection of the Notch receptor on the cell surface was inspired by the experimental process described by Rampal et al. (38). C2C12 cell lines were plated in 100 mm tissue culture dishes and collected the next day with a cell scraper. Cells were washed three times with PBS (pH 8.0) and counted on a Malassez slide. For each of the C2C12 cell lines, 5 × 10⁶ cells were incubated with 2 mM EZ-link sulfo-N-hydroxysuccinimide-biotin (sulfo-NHS-biotin) (Pierce Chemical Company, Rockford, IL) or with PBS (as a control) at room temperature for 30 min. The biotinylation reagent was then removed, and cells were washed with 100 mM glycine-PBS (pH 8.0) for 15 min to quench the biotinylation reaction. Cells were lysed in RIPA lysis buffer. Then, lysates were cleared by centrifugation (12,000 × g for 10 min) at 4°C and incubated overnight at 4°C with 50 μl of magnetic nanoparticles conjugated to streptavidin (MagCellect Streptavidin Ferrofluid; R&D Systems) previously equilibrated with RIPA buffer. The membrane proteins were isolated from the total protein fraction using a magnetic rack. Streptavidin-bound biotinylated proteins were captured by the magnet during 15 min at 4°C, leaving the nonbiotinylated proteins in the supernatant. The washed magnetic streptavidin beads were placed in Laemmli buffer, boiled for 5 min at 95°C, and briefly vortexed prior to separation of proteins on a gradient (4 to 12%) of Novex Tris-glycine precast gel electrophoresis gels (Invitrogen, Life Technologies). The correct isolation of biotinylated membrane proteins from the other nonbiotinylated proteins was checked by Western blotting using streptavidin-HRP and revealed with the 3,3′-diaminobenzidine tetrahydrochloride (DAB) peroxidase substrate. Anti-Notch1 (H-131) raised against amino acids 20 to 150 within the extracellular domain of human Notch1 (Santa Cruz Biotechnology, Santa Cruz, CA) was used to detect Notch1 at the C2C12 cell surface. Anti-pan-cadherin (Sigma-Aldrich Corp.) and anti-GAPDH (R&D Systems) antibodies were used to control the correct separation of membrane proteins.

Cell proliferation assay.

C2C12 myoblasts were seeded at 2,000 cells per well in GM into 96-well plates. They were incubated for 12 h (at 37°C and 5% CO₂). Twenty microliters of MTS [3-(4,5)-dimethylthiazol-2-yl)-5-(3-carboxymethoxy phenyl)-2-(4-sulfophenyl)-2H-tetrazolium salt] solution (Cell Titer 96 Aqueous Non-Radioactive cell proliferation assay; Promega Corp., Madison, WI) in a 200-μl final volume was added at 12, 24, 48, 72, or 120 h of the time course study. The plates were then incubated for 1 h, and the absorbance at 490 nm of formazan, a product from the bioreduced MTS, was measured using an enzyme-linked immunosorbent assay (ELISA) plate reader (FLUOstar Omega; BMG Labtech, Ortenberg, Germany).

Immunofluorescence studies.

C2C12 cell lines were fixed with 4% PFA-PBS for 20 min and permeabilized with 0.1% Triton X-100–PBS for 30 min at 4°C. To minimize nonspecific reactions, cells were saturated for 1 h at room temperature using PBS with 20% serum from the same species in which the secondary antibody was raised and then washed three times with PBS. Immunolabeling was performed with the primary antibodies MyoD (C-20) from Santa Cruz Biotechnology, GRP94 from Abcam (Cambridge, United Kingdom), Pax7, or Pofut1 overnight at 4°C; cells were washed three times with 0.1% Tween 20-PBS and incubated for 1 h at room temperature with polyclonal Alexa Fluor-conjugated secondary antibodies (Molecular Probes, Life Technologies, Eugene, OR). After three washes with 0.01% Tween 20-PBS, nuclei were labeled using 1 μg/ml 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich Corp.). Cells were rinsed three final times before being mounted on slides with Mowiol 4-88 mounting medium and sealed with glass coverslips. For cellular detection of incorporated 6-alkynyl fucose (Invitrogen, Life Technologies), cell lines were seeded at 1.5 × 10⁴ cells in eight-well tissue culture chambers (Sarstedt AG & Co., Nümbrecht, Germany) and incubated for 48 h in untreated GM or medium containing 200 μM 6-alkynyl fucose. Cells were rinsed with 1× PBS, fixed, and permeabilized. The Cu(I)-catalyzed azide-alkyne cycloaddition (CuAAC), or “click chemistry,” reaction was performed at room temperature for 1 h using azido-biotin and a Click-iT protein reaction buffer kit, according to the manufacturer's instructions (Invitrogen, Life Technologies). Molecules modified by alkynyl fucose were detected by Alexa Fluor 594-streptavidin (Invitrogen, Life Technologies). All the coimmunostaining and colocalization studies were performed using confocal microscopy (LSM 510 META; Zeiss) with Zeiss LSM software, version 4.2.0.121. Photographs were taken using an epifluorescence microscope (Leica DMI4000B MM AF Imaging System) powered by MetaMorph (Universal Imaging Corp., Downingtown, PA). The numbers of Pax7⁺/MyoD⁻, Pax7⁺/MyoD⁺, and Pax7⁻/MyoD⁺ cells were counted and expressed as percentage of the total numbers of immunostained cells. Data from multiple fields were pooled to give a population mean (±standard error of the mean [SEM]) of triplicate independent experiments.

DAPT cell treatment.

Wild-type C2C12 and Po⁻ cells were induced to differentiate for 3 days in DM in the presence of dimethyl sulfoxide (DMSO) and 0.5, 1, 2.5, 5, or 10 μM N-[N-(3,5-difluorophenacetyl)-l-alanyl]-(S)-phenylglycine t-butyl ester (DAPT) (Calbiochem, San Diego, CA). Afterwards, cells were stained by hematoxylin-eosin to determine their fusion index and were also used for protein extraction. A total of 100 μg of each extract was analyzed by Western blotting using the cleaved Notch1 (Val1744) antibody (Cell Signaling Technology).

Statistical analysis.

All of the experiments were performed at least three times, and results are reported as the means ± SEMs. Statistical comparisons were performed using two-tailed t tests implemented in Prism, version 5.03 (GraphPad Software, Inc., San Diego, CA). A P value of 0.05 or less was considered statistically significant.

RESULTS

A stable C2C12 cell line knockdown for Pofut1.

In the present work, we evaluated the role of protein O-fucosyltransferase 1, Pofut1, responsible of O-fucose addition, during the onset of myogenic differentiation. Seven stable Pofut1 shRNA C2C12 clones were generated. Six of them showed significant regulation of Pofut1 protein levels, decreasing from a factor of 1.9 to 6.2 (Fig. 1A). Moreover, they showed similar transcriptional profiles for genes involved in the Notch signaling pathway and in myogenic differentiation (data not shown). Among them, clone 5 displayed the strongest downregulation of Pofut1; this clone was named Po⁻ and was used in the subsequent work. At 0 h, Po⁻ showed reduced Pofut1 mRNA expression of about 40% relative to the WT or control shRNA C2C12, and it still remained significantly lower during the first 120 h of the differentiation process (Fig. 1B). Interestingly, Pofut1 was significantly more expressed in RC than in MT, independently of the considered cell lines (Fig. 1B). These expression profiles were confirmed at the protein level with a higher Pofut1 amount in WT and control shRNA C2C12 cells than in Po⁻ cells from 0 to 48 h of myogenic differentiation (Fig. 1C). Surprisingly, Pofut1 was poorly expressed in MT and did not show any significant difference in expression levels between cell lines, whereas Po⁻ myoblasts and reserve cells presented less than 50% of Pofut1 expression relative to WT or control shRNA C2C12 myoblasts and reserve cells, respectively (Fig. 1C). The Pofut1 differential expression linked to cell type could reveal a modulator role for Pofut1 in the context of myoblast differentiation.

FIG 1.

Characterization of the Pofut1 knockdown C2C12 cell line. (A) Pofut1 expression in MB of Pofut1 shRNA C2C12 clones. Histograms showed Pofut1 expression relative to Gapdh for each clone compared to that in WT C2C12 cells. (B) Relative quantities of Pofut1 expression in WT C2C12 and Po⁻ cells. In both cell lines, fold changes are expressed relative to the value at 0 h of WT C2C12 cells. (C) Western blot analyses of Pofut1 expression in WT C2C12, control shRNA, and Po⁻ cells. Histograms represent quantification of Pofut1 band intensity relative to Gapdh. All ratios were calibrated relative to time zero of WT C2C12 cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Activation of Notch signaling pathway during myogenic differentiation.

To determine if the Notch signaling pathway was affected by Pofut1 knockdown, the presence of the active form of Notch receptor was followed by Western blotting using a Val1744 antibody that reveals only cleaved NICD (39). In WT and control shRNA C2C12 cells, NICD was detected from 0 h to 48 h of the differentiation process, and it was more strongly expressed in RC than in MT, particularly at 120 h (Fig. 2A). This finding confirms the previous study by Sun et al. (9), where NICD was detected during the first 3 days of differentiation, with nearly restricted expression in RC at day 3. The NICD profile in Po⁻ cells was significantly lowered by 34 to 87% than in WT and control shRNA C2C12 cells depending on the differentiation time considered. As previously noted in the Pofut1 expression profile (Fig. 1C), NICD expression in RC is greater than in MT, with a decrease by at least half in Po⁻ RC compared to WT and control shRNA RC (Fig. 2A). Regardless of the cell lines, no significant difference in NICD levels was noticed in MT. In Po⁻ cells, down-expression of Pofut1 could alter the proteolytic cleavage following a weaker interaction between Notch and ligands present on adjacent cells. But the smaller amount of cleaved Notch detected in Po⁻ cells could also be related to a lesser presence of Notch receptors at the cell surface. In order to test this eventuality, levels of Notch receptor at the cell surface were investigated in C2C12 cell lines using an anti-Notch1 extracellular domain (NECD) antibody on proteins extracted after cell surface protein biotinylation (Fig. 2B). Notch1 was detected only in the enriched fraction of membrane proteins for each C2C12 cell line (Fig. 2Ba). Probing for common membrane and cytosolic protein markers showed that membrane proteins such as cadherins and cytosolic proteins like Gapdh were highly enriched in the membrane and cytosolic fractions, respectively. No significant difference was observed in Notch cell surface expression between Po⁻ and WT or control shRNA C2C12 cells (Fig. 2Bb).

FIG 2.

Notch signaling pathway activation and cell surface receptor expression. (A) Detection of cleaved Notch intracellular domain (NICD) by Western blotting, during myogenic differentiation, for WT C2C12, control shRNA, and Po⁻ cells. Histograms represent quantification of NICD as previously calculated. (B, a) Western blot detection of the Notch1 extracellular domain (NECD), pan-cadherin, and Gapdh expression, after a cell surface biotinylation process, on the streptavidin-retained fraction (lanes 1) and flowthrough fraction (lanes 2). Labeling with biotin and correct separation of protein fractions were checked using streptavidin-HRP and revealed by DAB. (b) Histograms representing quantification of Notch receptor at the surface of WT C2C12, control shRNA, and Po⁻ myoblasts calculated by NECD/pan-cadherin ratio. (C) Gene expression of Rbpj (a), encoding a nuclear partner of NICD, and Hes1 (b), encoding a transcriptional repressor regulating myogenesis in WT C2C12 cells and Po⁻ cells. In both cell lines, fold changes are expressed relative to time zero for WT C2C12 cells. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Nevertheless, the presence of cleaved Notch does not prove that the Notch signaling pathway is fully activated since several intracellular partners can regulate NICD action toward the transcription of its target genes. Following potential cytoplasmic interactions, NICD forms a transcriptional coactivation complex in the nucleus (40). We analyzed the expression of three genes encoding essential proteins that compose this complex: Rbpj (the DNA binding transcription factor, nuclear activator of the pathway), Maml1 (Mastermind-like 1 transcriptional coactivator), and Ep300 (a histone acetyltransferase). Maml1 and Ep300 expression levels were different only between RC and MT for both WT C2C12 and Po⁻, with greater expression in RC (data available upon request). However, the expression of Rbpj decreased throughout the differentiation process (Fig. 2Ca). At 120 h in WT C2C12 cells, Rbpj expression had decreased by 66% and 42% for MT and RC, respectively. In Po⁻ cells, the phenomenon was accentuated with a decrease of 47% at the beginning of the time course compared to the WT level and a significant reduction by 45% in RC at 120 h.

Final activation of Notch signaling can be measured by expression of some Hairy/Enhancer of split genes encoding transcriptional inhibitors (41). We analyzed the expression of Hes1, Hes5, Hes7, Hey1, Hey2, and HeyL, which are targets for NICD (42), and Hes6, which is upregulated during myogenic differentiation (30, 43). Expression of the Hes5, Hes7, and Hey2 genes was not detected regardless of the time course of the C2C12 differentiation process. Although not necessary in this context, they cannot be ruled out as actors of early developmental processes like somitogenesis, as demonstrated for Hes7 (44). Expression levels of Hes1, Hes6, Hey1, and HeyL were modified during WT and Po⁻ cell differentiation. Regardless of the C2C12 cell line considered, Hes1, Hey1, and HeyL were more expressed in RC (Fig. 2Cb) (data available upon request), coinciding with the presence of NICD and expression of Rbpj, Maml1, and Ep300. In Po⁻ cells, Hes1 expression was significantly reduced (except in RC), compared to WT C2C12 cells (reduction between 53% and 71%). Although both Hes1 and Hey1 had inhibitory activities during myogenesis, directly targeting promoters of myogenic genes or interacting with key myogenic factors, only Hes1 seemed to directly interact with MyoD and sequester it (41, 45). Hes6 expression was enhanced during the first 120 h of the differentiation process and was higher in MT than in RC but without a significant difference between levels in WT and Po⁻ cells (data available upon request).

We showed that cleaved NICD, components of the coactivation complex, and some transcription factors inhibiting myogenic differentiation were upregulated in RC versus MT. Downregulation of Pofut1 significantly lowered the quantity of cleaved NICD and the expression levels of Rbpj and Hes1 but without modification of the cell surface expression pattern of Notch1.

Pofut1 reduction affects the timing of myogenic differentiation through Notch activation.

Although observed phenotypes and fusion indexes (Fig. 3A and B) of WT and control shRNA C2C12 cells did not show significant differences during the first 120 h of differentiation process, the downregulation of Pofut1 in Po⁻ cells induced more elongated mononucleated cells at 0 h and clearly more multinucleated cells (MT) from 48 h. These observations were also noted for Pofut1 shRNA clones 2, 4, and 6, where Pofut1 expression was more than 60% diminished (Fig. 1A and 3A and B). For clones 3 and 7 where the Pofut1 decrease was less marked (Fig. 1A), the quantity of slender multinucleated cells was also less pronounced, and no significant difference in fusion indexes was seen after 48 h (Fig. 3A and B). Only clone 10 did not show any difference compared to the WT and control shRNA C2C12, in agreement with an absence of Pofut1 deregulation (Fig. 1A). Globally, it appeared that the differentiation process was advanced for all clones, except number 10. In Po⁻ cells, we found that the fusion process was already active around 24 h earlier than normal. Taken together, these results demonstrate that Pofut1 is a key actor in the timing of myogenic differentiation. To characterize the peculiar MT phenotype of Pofut1 knockdown, the number of nuclei per multinucleated cell was counted in Po⁻, WT, and control shRNA C2C12 cell lines. No significant difference was observed during the first 72 h (data not shown). At 120 h, while WT and control shRNA C2C12 cells had a mean of 9 and 8 nuclei per MT, respectively, Po⁻ cells had a significantly lower number of nuclei, with an average of only 4 nuclei per MT. To take into account the heterogeneity in number of nuclei per MT, they were categorized into 5 groups (Fig. 3C). For each group, Po⁻ cells had significantly more MT with a small number of nuclei (68% versus 39% in WT C2C12) than MT with a number of nuclei greater than 5. For example, 16% of WT C2C12 MT had 10 to 20 nuclei, whereas this percentage was only of 5% in Po⁻ cells. No MT with more than 40 nuclei was observed in Po⁻ cells at 120 h. Therefore, when Pofut1 is knocked down, myoblast fusion is affected during differentiation, with an earlier start but a significantly reduced capacity of cell fusion into multinucleated myotubes. To better characterize phenotypes of multinucleated cells in the three cell lines, their diameters and perimeters were measured (Fig. 3D). No significant difference was observed between WT and control shRNA C2C12 cells. In Po⁻ cells, the MT diameter was significantly smaller (7 μm at 48 h and 9 μm at 120 h) than that in WT C2C12 cells (15 μm at 48 h and 25 μm at 120 h). The MT perimeter was significantly increased in Po⁻ cells at 48 h and 72 h but not the total area (data not shown). From at least 48 h of differentiation, MT in Po⁻ cells adopted a particularly stretched shape with a narrower diameter than in WT and control shRNA C2C12 cells (Fig. 3D). At day 5, WT MT were 227 ± 7 μm long, fusiform, thick, and elongated in several directions, whereas in Po⁻ cells they were 471 ± 23 μm long and slender, with only a few y shaped (Fig. 3E). So, in Po⁻ cells, the myoblast-myoblast fusion (primary step of fusion) was not particularly disturbed, with no significant difference in the number of nuclei per myotube during the first 72 h compared to WT C2C12, but the myoblast-myotube fusion (secondary step of fusion) was deeply affected with less myoblast fusion, which normally contributes to increase myotube diameter.

FIG 3.

Pofut1 knockdown in C2C12 cells promotes myogenic differentiation but alters myotube phenotype. (A) Photographs of WT C2C12, control shRNA, Po⁻, and various Pofut1 shRNA clones. (B) Fusion indexes of WT C2C12, control shRNA, Po⁻, and other Pofut1 shRNA clones. (C) Percentages of myotube populations categorized by the number of nuclei for WT C2C12, control shRNA, and Po⁻ cells at 120 h. (D) Phenotypic measurements of myotube diameter and perimeter at three times of the differentiation process. (E) Myotube photographs and lengths observed at 120 h. Cl, clone. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

To rule out the possibility that alteration of the differentiation process is due to a change in cell proliferation prior to DM addition and thus to an accelerated entry into differentiation, an MTS assay was performed in GM until cells reached confluence. No significant difference in the absorbance at 490 nm was observed between the WT, control shRNA C2C12, and clones downregulated for Pofut1 from 0 h to 120 h (data not shown). Consequently, no pronounced alteration was observed in the rhythm of divisions of the cell lines. So, the earliest appearance of MT in Pofut1 knockdown cells (Fig. 3) was not due to a difference in the proliferation process leading to an earlier confluence and consequently to the start of the differentiation process.

As Notch signaling is involved in cell fate commitment, an alteration of the NICD cleavage process could be responsible for the earlier differentiation in Pofut1 knockdown cells. To challenge this hypothesis, the γ-secretase inhibitor DAPT, which prevents proteolytic cleavage of NICD, was used during the first 72 h of WT, control shRNA, and Po⁻ C2C12 cell differentiation in a range of concentrations from 0.5 μM to 10 μM (Fig. 4). As previously shown by the data in Fig. 2, there was more of the active form of Notch receptor in WT C2C12 than in Po⁻ cells even when cells were treated with DMSO (Fig. 4A). Compared to untreated WT C2C12 cells, DAPT-treated C2C12 cells showed a gradual decrease in cleaved NICD from a concentration of 0.5 to 10 μM (Fig. 4A). C2C12 cells treated with 2.5 μM DAPT had a level of cleaved NICD close to that of the Po⁻ sample (Fig. 4A). From a DAPT concentration of 5 μM, less NICD was present than in Po⁻ cells until it was virtually undetectable at 10 μM DAPT (Fig. 4A). In WT and control shRNA C2C12 cells, DAPT treatment appeared to alter the entry into differentiation, whereas Po⁻ cells did not seem to be affected by the inhibition of γ-secretase at any DAPT concentration (Fig. 4B). From 0.5 to 10 μM DAPT, WT and control shRNA C2C12 cell lines showed an increase in myotube formation compared to untreated cells. At 2.5 μM DAPT, multinucleated cells were significantly increasing, to the level of the Po⁻ phenotype cells. These observations were confirmed with fusion index analysis of the three cell lines at 72 h of myogenic differentiation (Fig. 4C). C2C12 cells treated with 0.5 to 10 μM DAPT showed fusion indexes higher than control DMSO-treated C2C12 cells, and no significant difference was observed with untreated control shRNA regardless of the DAPT concentration (Fig. 4C). Both of these cell lines presented a lower fusion index than Po⁻ cells (around 0.25 versus 0.4; P < 0.001) under the control DMSO condition, but the difference tended to become gradually less significant when the DAPT treatment rose from 0.5 μM (around 0.28 versus 0.4; P < 0.01) with 1 μM (around 0.30 versus 0.4; P < 0.05) to 10 μM (around 0.41 versus 0.44; not significant) (Fig. 4C). This trend correlates with the significant reduction in NICD cleavage (Fig. 4A). Consequently, a decrease in NICD cleavage by γ-secretase leads to an earlier entry in the differentiation process, resulting in an increase in the fusion index. Reduction in Pofut1 expression would not affect the proliferative capacity of myogenic progenitor cells but would impact determination of cells during the early phase of differentiation, acting on the processing of NICD.

FIG 4.

Earlier differentiation in Pofut1 knockdown C2C12 cells relies on the Notch signaling pathway. (A) Cleaved NICD detected by Western blotting for untreated C2C12 and Po⁻ cells and C2C12 cells treated with different concentrations of DAPT. (B) Photographs of WT C2C12, control shRNA, and Po⁻ cells at 72 h of differentiation with DMSO or DAPT treatment at various concentrations. (C) Fusion indexes at 72 h. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Pofut1 knockdown decreases Pax7 and disrupts the expression of myogenic markers.

To assess the alteration of the differentiation program due to Pofut1 knockdown, Pax7 and myogenic markers were studied. Our semiquantitative RT-PCR and Western blot experiments showed a clear downregulation of Pax7 in MT compared to RC in WT C2C12 cells and a decrease during progression of differentiation (Fig. 5A and B) (data available upon request). A strong downregulation of Pax7 was observed in Po⁻ cells (Fig. 5A) where its expression significantly decreased from 12 h. Pax7 RQ was reduced in RC by a factor of 70.3 and 25.7 at 72 h and 120 h, respectively, compared to RC of WT C2C12 cells. These results are consistent with immunodetection using mouse anti-Pax7 antibody (Fig. 5B).

FIG 5.

A premature myogenic differentiation program in Pofut1 knockdown C2C12 cells. (A) Expression of Pax7 in WT C2C12 and Po⁻ cells. (B) Western blot analyses of Pax7, MyoD, Myog, and MyHC. Band intensities were normalized with Gapdh and calibrated using the value at time zero of WT C2C12 cells, except for MyHC, where the time point 72 h was used for calibration. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

As a mechanism of reciprocal inhibition between Pax7 and the MRFs MyoD and Myog was previously demonstrated (20), we followed the expression of these two MRFs. In WT C2C12 cells, MyoD was expressed during the first 120 h, with a progressive downregulation within the first 48 h of the myoblast differentiation process. Strong expression of MyoD was detected in MT (Fig. 5B). In Po⁻cells, the most noticeable features were the accelerated downregulation of MyoD occurring 24 h earlier and an increase of MyoD expression in RC, both compared to WT C2C12 cells. Myog detection by Western blotting showed a mutually exclusive pattern relative to Pax7 pattern in WT C2C12 cells (Fig. 5B). Pax7 was detected from 24 h and mostly in MT. For Po⁻ cells, if Myog was still more strongly detected in MT, its expression was more sustained and faintly detectable from 0 h. Finally, myosin heavy chain (MyHC), which was only faintly detected at 72 h in WT MT, was strongly revealed in Po⁻ MT (Fig. 5B). mRNA expression of myocyte enhancer factors (MEF), which can potentiate action of MRFs in myogenesis, was also followed. Although Mef2b was not detected, significant differences were noticed in MT versus RC for only Mef2a, Mef2c, and Mef2d, regardless of the cell line used (data available upon request).

Colocalization of Pax7 and MyoD was performed by cell immunofluorescence staining during the first 24 h of the differentiation process (Fig. 6A). At the start of differentiation, three cell populations were detected: Pax7⁺/MyoD⁻ (self-renewing cells), Pax7⁺/MyoD⁺ (proliferating cells) and Pax7⁻/MyoD⁺ (differentiating cells). In WT C2C12, their respective proportions were 74.4%, 24.1% and 1.5% (Fig. 6B). During the course of myogenic differentiation, MyoD will be expressed in a subpopulation of Pax7⁺/MyoD⁻ cells, and then Pax7 will be lost in a subpopulation of Pax7⁺/MyoD⁺ cells committed to future fusion process. In Po⁻ cells, the proportions of Pax7⁺/MyoD⁻, Pax7⁺/MyoD⁺, Pax7⁻/MyoD⁺ cell populations were 20.2%, 74% and 5.8%, at the start of the differentiation process (Fig. 6B). The pool of progenitor cells, Pax7⁺/MyoD⁻, was drastically depleted predominantly in favor of Pax7⁺/MyoD⁺ population. At 24 h, the Pax7⁻/MyoD⁺ population clearly prevailed (68%) over the Pax7⁺/MyoD⁺ population (31%). Therefore, in Po⁻ cells, the balance between Pax7, a marker of proliferating myoblasts and self-renewal of progenitors, and MyoD, a marker of myogenically determined cells, is modified from the beginning of the differentiation process in favor of the differentiation marker, potentially decreasing the pool of progenitor cells.

FIG 6.

Early commitment toward differentiating cells in Pofut1 knockdown C2C12 cells. (A) Coimmunostaining for Pax7 (green) and MyoD (red) of WT C2C12 and Po⁻ cells from 0 h to 24 h. (B) Percentages of Pax7⁺/MyoD⁻, Pax7⁺/MyoD⁺, and Pax7⁻/MyoD⁺ cell populations in WT C2C12 and Po⁻ cells during onset of the differentiation process.

Rescue of the Pofut1 knockdown phenotype with shRNA-resistant Pofut1.

To confirm that the Po⁻ cell phenotype was due only to deregulation of Pofut1 expression, Po⁻-derived clones were produced by reestablishing Pofut1 expression using the coding region of murine Pofut1, mutated to be Pofut1 shRNA resistant. Three constructs were generated using a lentiviral vector containing Pofut1 CDS under a cytomegalovirus (CMV) promoter. The 21-bp of Pofut1 shRNA target sequence was left intact (0M) or was mutated to introduce one (1M) or three (3M) silent mutations (Fig. 7A). Po⁻ cells were transfected with one of the three constructs, and three stable cell lines were established. In Po⁻ and 0M cells, Pofut1 expression levels were similar (Fig. 7A), indicating the efficiency of Pofut1 shRNA to knock down exogenous wild-type Pofut1. In contrast, in 1M and 3M cell lines, expression of Pofut1 was resistant to silencing by Pofut1 shRNA. In 1M cells, Pofut1 protein expression was reestablished and comparable to the WT C2C12 level, whereas in 3M cells Pofut1 was significantly expressed 3.8 times more than in Po⁻ cells and 1.5 times more than in WT cells. As we previously showed, Pofut1 impacted the Notch signaling activation. Both Po⁻ and 0M cells showed comparable levels but less cleaved NICD than the WT or 1M C2C12 cells, indicating that Notch activation was not modified by introducing the lentivector containing shRNA-sensitive Pofut1 CDS (Fig. 7A). In 3M cells, a higher quantity of cleaved NICD was detected that was around 3.4 times more than in Po⁻ and 0M cells and around 1.1 times more than under normal biological conditions (WT C2C12 cells) (Fig. 7A). Therefore, expression of exogenous Pofut1, which is resistant to mRNA degradation induced by the Pofut1 shRNA, enabled the increase of cleaved NICD.

During myogenic differentiation, Po⁻ and 0M cells showed comparable phenotypes, particularly with long and slender myotubes at 120 h (Fig. 7B). Cell fusion indexes in 0M relative to Po⁻ cells were not significantly different (Fig. 7C). At 24 h, the 1M cell line presented a reduction in cell fusion of 72% compared to Po⁻ cells and of 52% at 120 h, values which were not significantly different from those of WT C2C12 cells. The low fusion indexes of the 3M cell line, corresponding to a decrease of 93% and 84% in comparison to Po⁻ cells at 24 h and 120 h, respectively, were significantly different from the WT C2C12 level. A major defect of the myogenic fusion process was found in 3M cells, where very few myotubes were observed at 120 h; most of them were small and rounded (Fig. 7D). In contrast to Po⁻ and 3M cells, 1M myotubes were similar in shape and nucleus content to WT C2C12 and 0M cells at 120 h.

The initial status of the myogenic program was evaluated in MB at 0 h for Po⁻, 0M, 1M, 3M, and WT C2C12 cell lines by Western blotting detection of Pax7, MyoD, and Myog (Fig. 7E). The severe depletion of Pax7 observed in Po⁻ cells was also found in 0M cells. This marker of quiescent and proliferating cells was reexpressed in cell lines expressing an exogenous shRNA-resistant Pofut1 CDS. Pax7 was reexpressed 3.5 and 5.4 times more in 1M and 3M myoblasts than in Po⁻ cells, respectively, and at a level similar to that of WT C2C12 cells. The myogenic determination marker MyoD was detected at equal levels in Po⁻ and 0M cells and was significantly less in 1M and WT C2C12 cells (Fig. 7E). In 3M MB, less MyoD was present (around 38% decrease relative to the level in Po⁻), which could explain their less differentiated phenotype. Myog was expressed in Po⁻ and 0M cells but not in 1M and 3M cells (Fig. 7E). However, although these expression profiles were sufficient to reestablish a cell fusion process comparable to WT C2C12 one for 1M, it was not the case for 3M (Fig. 7C).

The rescue of Pofut1 knockdown in Po⁻ cells with exogenous shRNA-resistant Pofut1 restored a normal fusion process and reactivation of the Notch signaling pathway. It showed that the early entrance of myoblasts into the differentiation program was due to Pofut1, probably through the O-fucosylation of Notch receptors.

Pofut1 localization and O-fucosylation profile in C2C12 cell lines.

In WT, control shRNA, Po⁻, and Pofut1 rescue cell lines, Pofut1 mainly colocalized with GRP94, an endoplasmic reticulum chaperone, in agreement with the ER-resident status of Pofut1 (33) (Fig. 8A). Immunofluorescence staining showed that Pofut1 intensity in C2C12, control shRNA, and 1M cells was similar, and Pofut1 colocalized at around 62% with GRP94. Compared to these three cell lines, Po⁻ and 0M showed a diminished mean intensity of Pofut1 labeling and colocalized with GRP94 at around 91% with a peculiar restricted perinuclear distribution (Fig. 8A). Finally, 3M cells displayed higher intensity of Pofut1 labeling than WT, shRNA, or 1M cells. Pofut1 only partly colocalized with GRP94 (around 30%); some Pofut1 may be distributed outside the ER (Fig. 8A).

FIG 8.

Fluorescence imaging of Pofut1 and fucosyl conjugates in C2C12 cell lines using copper(I)-catalyzed azide-alkyne cycloaddition. (A) Immunofluorescence labeling of Pofut1 (red), GRP94 endoplasmic reticulum protein (green), and nuclei (blue) in C2C12, control shRNA, Po⁻, and Pofut1 rescue cell lines. Yellow indicates colocalization between Pofut1 and GRP94. (B) Staining of O-fucosyl glycans after click chemistry reaction (+CC) using azido-biotin revealed by streptavidin (red), GRP94 (green), and nuclei (blue) in WT C2C12, control shRNA, Po⁻, and Pofut1 rescue cell lines. A negative control without the cycloaddition reaction (−CC) was also performed. Colocalization of alkynyl fucose labeling with GRP94 was visualized in yellow.

Microscopic visualization of fucosyl-conjugates in WT, Po⁻, control shRNA, and Pofut1 rescue C2C12 cell lines was performed by a copper(I)-catalyzed cycloaddition of 6-alkynyl fucose with biotinylated azide and fluorescent streptavidin detection (Fig. 8B). As 6-alkynyl fucose incorporation was more efficiently performed in living cells onto EGF-like repeats and trombospondin repeats (TSRs) than in N-linked glycans (34, 46), its detection could give a good overview of Pofut1 enzymatic activity. Furthermore, costaining with GRP94 allowed us to discriminate O-fucosylation occurring in the ER from fucosylation taking place in Golgi apparatus (for mucins and N-glycans) (Fig. 8B). Control without cycloaddition showed a very low background, indicating that the labeling was specific to alkynyl fucose detection. Wild-type C2C12, control shRNA, and 1M cells showed similar levels of detection of incorporated alkynyl fucose, and the labeling colocalized with GRP94 (Fig. 8B). In Po⁻ and 0M cells, incorporation of fucose was less important but still mainly colocalized with GRP94 in a perinuclear fashion as already noticed for Pofut1 staining (Fig. 8A). In cells overexpressing Pofut1, the 3M line, a significant increase of alkynyl fucose incorporation was found compared to the levels in the WT C2C12, control shRNA, and 1M cells. As also seen for Pofut1 repartition (Fig. 8A), a part of incorporated labeled fucose did not costain with GRP94 and seemed to reside outside the ER (Fig. 8B). In each of the cell lines, Pofut1 and incorporated fucose profiles clearly overlapped. To conclude, the only fucose incorporation in macromolecules occurring in the ER, the O-fucosylation, was diminished in Po⁻ and 0M cells, where Pofut1 was also decreased, indicating that the defects observed in the course of myogenic differentiation were mainly due to the alteration of this original posttranslational modification.

DISCUSSION

The early steps of myogenesis are controlled by expression of transcription factors acting on myoblast proliferation (i.e., Pax7) and those playing roles during determination, fusion, and differentiation processes (i.e., MRFs). The Notch signaling pathway governs the expression of these transcription factors and allows proliferation of myoblasts and inhibition of their differentiation, replenishment of the satellite cell pool, and the sustaining of their quiescent state in the adult. In the context of myogenic differentiation of C2C12 cells, our present data show that Pofut1, an enzyme responsible for O-fucosylation of proteins, is crucial for Notch pathway activation, and, consequently, through the Pax7/MRFs ratio, it influences the major cell fate decisions in the cell line, i.e., myoblast, reserve cell, and myotube (Fig. 9).

FIG 9.

Proposed model of Pofut1 involvement in commitment of myogenically differentiated cells through activation of the Notch signaling pathway. (A) In wild-type C2C12 cells, myoblasts express Pofut1 at a basal level to correctly O-fucosylate the Notch extracellular domain. During differentiation, Pofut1 is differentially expressed with a high level in RC, leading to high activation of the signaling pathway responsible for proliferation and progenitor self-renewal. In parallel, a lower level of Pofut1 could affect NICD cleavage, resulting in cell fusion at the origin of MT. The cleaved NICD quantity impacts the Pax7/MyoD expression balance, resulting in a cell fate decision between RC and MT. (B) In the Pofut1 knockdown context, the lower Pofut1 basal level in myoblasts would alter the correct O-fucosylation of Notch receptor and prevent the correct cleavage of NICD. The limited signaling pathway activation prematurely disrupts the Pax7/MyoD expression balance, resulting in depletion of progenitors and earlier myoblast fusion, causing later myotube defects.

The moderate downregulation of Pofut1 expression in our Pofut1 shRNA C2C12 clones (around 40% less than in WT C2C12 cells) underlines the important role of Pofut1 in this context. Indeed, Pofut1 function is never complemented by another enzyme nor by its paralogue, Pofut2 (47). Interestingly, the cax (for compact axial skeleton) mutation in mouse, which creates a Pofut1 hypomorphic allele due to an intracisternal A particle (IAP) insertion, more deeply affects Pofut1 expression without altering viability (48). However, somitogenesis is disorganized to a lesser extent than in Pofut1^−/− embryos (21), but no experiment was conducted on cax adult skeletal muscle properties. To activate the canonical Notch pathway, a correct O-fucosylation of receptors is needed (24), whereas ligand O-fucosylation seems dispensable, as demonstrated for mouse Dll1 (49). We show that the Pofut1 transcripts and proteins are expressed at higher levels in RC than MT in WT C2C12 and Po⁻ cells. We could hypothesize that changes in Pofut1 expression will contribute to hypofucosylation of the Notch receptor or to the global absence of O-fucosylation on some Notch EGF-like repeats. These changes in glycosylation profile could affect NECD conformation and consequently Notch pathway activation.

Twenty potential O-fucosylation sites on EGF-like repeats are present on mouse Notch1, and 13 are known to be modified by O-fucose (50). Notch2 and Notch3 receptors have 21 and 15 potential O-fucosylation sites, respectively, and are probably predominantly O-fucosylated. Among Notch1 EGF-like repeats, EGF12 in the ligand binding region and EGF26 and EGF27 in the Abruptex region identified in fly have been reported to modulate the binding affinity of Notch for its ligands (51). The difference in Pofut1 expression observed in myotubes compared to reserve cells would not only prevent O-fucosylation of some EGF-like domains but also change ligand binding affinities to Notch receptors. We could also envisage that a limited O-fucosylation would not be sufficient to change the structural conformation of Notch receptors to uncover the ligand-interacting area, in a way proposed by Pei and Baker (52). They assume that EGF-like repeats from the Abruptex region of fly Notch (EGF22 to EGF27), which contains 36 EGF-like repeats like to mouse Notch1, and 23 potential O-fucosylation sites, bind to those of the ligand-binding domain (EGF11 and EGF12) in the absence of an interacting ligand. The state of O-fucosylation may affect the local folding of EGF-like domains in NECD and the accessibility to ADAM proteases (53). Moreover, O-glucosylation, which also contributes to Notch glycosylation, had been proposed to maintain fly NECD in a conformation that is permissive for signaling (54). However, we demonstrate that correct expression of Pofut1, responsible of O-fucosylation of Notch receptors, is indispensable for correct cleavage of Notch.

Knockdown of Pofut1 in C2C12 cells significantly decreases Pax7 expression at the mRNA and protein levels during the 120-h time course of the differentiation process while accelerating downregulation of MyoD and upregulation of Myog compared to WT levels. The consequence is an earlier appearance of MT, around 24 h before their appearance in the WT. The Pax7/MyoD ratio has a key role in cell fate determination (1). A low Pax7/MyoD ratio commits cells to myogenic differentiation, a high ratio permits renewal of the pool of quiescent progenitor cells (satellite cells in vivo), and an intermediate ratio favors proliferation (20). The Pax7/MyoD ratio is under the dependence of Notch pathway activation. Indeed, overexpressed NICD upregulates Pax7 at both the mRNA and protein levels, resulting in enhanced self-renewal of satellite cells (8). Inhibition of Notch signaling leads to a downregulation of Pax7, resulting in satellite cell depletion and improved terminal differentiation (55). NICD directly regulates Pax7 expression through CBF1, whereas MyoD expression is mediated by Hes and Hey family proteins whose expressions are dependent on NICD/CBF1 control (56). In a Pofut1-limited expression context, a majority of cells are Pax7⁺/MyoD⁺ at the start of differentiation at a level three times greater than in WT C2C12 cells. It is at the expense of Pax7⁺/MyoD⁻ cells, i.e., progenitor cells. Indeed, if the low NICD level present in Po⁻ allows proliferation of Pax7⁺/MyoD⁺ cells, it is not sufficient to increase the Pax7 expression necessary to self-renew the Pax7⁺/MyoD⁻ pool, especially because the NICD half-life is only about 3 h (57). Consequently, the myogenic differentiation route is favored by the loss of Pax7 and by Myog induction. As Pax7 expression is negatively regulated by microRNAs (miRNAs) 1, 206, and 486, which are themselves controlled by MyoD expression in myoblast differentiation (58, 59), a quantification of these miRNAs should be conducted in our cell lines.

The earlier entrance in the differentiation process in Po⁻ cells results in elongated and slender myotubes. They are characterized by a small number of nuclei, indicating that the primary fusion of myoblasts, forming nascent myotubes, was not altered but that the secondary fusion, involving nuclear accretion and changes in size of the myotubes, was. Various extracellular, cell surface, and intracellular proteins, such as cadherins, integrins, ADAMs, nuclear factor of activated T cells (NFAT), and interleukin-4 (IL-4), influence the fusion of myoblasts (60, 61). Po⁻ myoblasts could present altered expression for these factors, which would subsequently impair the lateral fusion necessary to increase myotube diameter and formation of characteristic y-shaped myotubes.

An additional role of Pofut1 as a chaperone in the folding, secretion, and endocytosis of Notch receptors was previously proposed (62, 63). Unlike analyses in flies, where Notch receptors are not correctly addressed to plasma membrane in the absence of Pofut1 (64, 65), the majority of Notch receptors in Pofut1-null mouse cells are transferred to the cell membrane, with some, however, sequestered in the ER (24, 66). We show that Pofut1 is probably not confined only to the ER in WT C2C12, an observation which is strengthened by the analysis of Henningsen and collaborators (67) of Pofut1's presence in the C2C12 secretome and its decrease as differentiation progresses. Although no difference in the quantity of Notch1 exposed at the cell surface was seen, immunolocalization in Po⁻ cells of Notch receptors and characterization of their interactions with ligands on neighboring cells will help us to understand if, in addition to O-fucosylation of Notch, Pofut1 had a definitive role in its chaperoning.

Finally, we cannot rule out the possibility that Pofut1 knockdown affects targets other than Notch receptors in the context of myogenic differentiation as over 100 predicted targets of Pofut1 were identified in silico (71). However, only 15 proteins were biologically demonstrated to be modified by O-fucose, and only agrin, for which O-fucose is involved in acetylcholine receptor aggregation in neuromuscular development, was detected in C2C12 MT (68); but this could not explain our Po⁻ phenotype. Moreover, O-fucosylation is not the sole type of glycosylation known to occur on the Notch1 ECD. Although N-glycosylation (six potential sites) of Notch does not appear to have a peculiar role in Notch signaling pathway (23), it would be of interest to test the impact of O-glucosylation (69) and O-GlcNAcylation (70) deregulation in the context of myogenic differentiation.