A novel role for the RNA-binding protein FXR1P in myoblasts cell-cycle progression by modulating p21/Cdkn1a/Cip1/Waf1 mRNA stability

Abstract

The Fragile X-Related 1 gene (FXR1) is a paralog of the Fragile X Mental Retardation 1 gene (FMR1), whose absence causes the Fragile X syndrome, the most common form of inherited intellectual disability. FXR1P plays an important role in normal muscle development, and its absence causes muscular abnormalities in mice, frog, and zebrafish. Seven alternatively spliced FXR1 transcripts have been identified and two of them are skeletal muscle-specific. A reduction of these isoforms is found in myoblasts from Facio-Scapulo Humeral Dystrophy (FSHD) patients. FXR1P is an RNA-binding protein involved in translational control; however, so far, no mRNA target of FXR1P has been linked to the drastic muscular phenotypes caused by its absence. In this study, gene expression profiling of C2C12 myoblasts reveals that transcripts involved in cell cycle and muscular development pathways are modulated by Fxr1-depletion. We observed an increase of p21--a regulator of cell-cycle progression--in Fxr1-knocked-down mouse C2C12 and FSHD human myoblasts. Rescue of this molecular phenotype is possible by re-expressing human FXR1P in Fxr1-depleted C2C12 cells. FXR1P muscle-specific isoforms bind p21 mRNA via direct interaction with a conserved G-quadruplex located in its 3' untranslated region. The FXR1P/G-quadruplex complex reduces the half-life of p21 mRNA. In the absence of FXR1P, the upregulation of p21 mRNA determines the elevated level of its protein product that affects cell-cycle progression inducing a premature cell-cycle exit and generating a pool of cells blocked at G0. Our study describes a novel role of FXR1P that has crucial implications for the understanding of its role during myogenesis and muscle development, since we show here that in its absence a reduced number of myoblasts will be available for muscle formation/regeneration, shedding new light into the pathophysiology of FSHD.

Figure 1. Microarray analysis of Fxr1-depleted C2C12 myoblasts.

(A) Quantitative RT-PCR reveals a strong reduction of Fxr1 mRNA in C2C12 cells transfected with siRNA against Fxr1 compared to siControl-transfected cells. (B) Western-blot analysis of untransfected (UT) and siFxr1-transfected cells (siFxr1) revealed with the antibody #3FX recognizing all isoforms of FXR1P reveals a strong depletion of all isoforms of FXR1P (short, medium and long) compared to control (siCtl), while the levels of FXR2P protein (asterisk, *) remain unchanged. β-tubulin (β-tub) signal is used to verify equal loading of lanes. (C) Immunofluorescence analysis of FXR1P (red) subcellular distribution in siControl and siFxr1-transfected cells, using polyclonal #830 anti-FXR1P antibodies. Nuclei were counterstained with DAPI (blue) and merge images are shown in the right panel. The same exposure time was used for both image captures and reveal a strong depletion in FXR1P signal in siFxr1-transfected cells compared to control (siControl). Scale bar: 15 µm. (D) Confocal micrographs of C2C12 cells immunostained for FXR1P reveal a nucleocytoplasmic distribution of FXR1P. Please note the nuclear dot-like structures containing FXR1P. Slice depth: 1 µm, scale bar: 15 µm. (E) Volcano plot showing the distribution of differentially expressed transcripts between C2C12 cells transfected with siRNA against Fxr1 versus siControl-transfected cells. Log of the fold of change (LogFC) is plotted against the B-statistic value for each transcript. A subset of 9 transcripts selected for further validation by Quantitative-RT PCR (Fxr1, Cdk15, Sema7a, Mctp2, Asgr1, Hgf, p21, Dusp6 and Lbh) are highlighted. Significantly down- and up-regulated genes are shown in green and red, respectively. (F) Quantitative-RT PCR analysis of a subset of mRNAs confirm that Sema7a, Mctp2, Asgr1, p21, Hgf, Dusp6, Lbh, MyoD and Myog are significantly upregulated in Fxr1-depleted C2C12 myoblast, while Cdk15 is downregulated, confirming the microarray analysis. Data are presented as means ± SEM of n = 4 experiments.

Figure 2. Fxr1-depletion does not impair myoblasts viability but specifically induces accumulation in G0/G1 phase to the detriment of mitosis.

(A) PI incorporation in living siFxr1- or siControl-transfected cultures and subsequent FACS analysis was performed to show that viability of the culture is not affected by Fxr1-depletion. (B) MTT colorimetric assay show that the proliferation abilities of C2C12 cells are significantly impaired by Fxr1-depletion. (C) FACS analysis of the Propidium Iodure-stained DNA content of C2C12 cells transfected with siControl or siFxr1. Cells were analysed in asynchronous conditions or following synchronisation treatment for 8 hrs with the cell cycle blocker mimosine (late G1) followed by 16 hrs release in normal growth medium (D). In asynchronous conditions, cell cycle distribution is similar in siControl or siFxr1 transfected cells. Synchronisation of cells allows detecting significant differences in the distribution of the cells in the various cell cycle absence of FXR1P: increase in the G0/G1 proportion and decrease in the G2/M. Data are presented as means ± SEM of n = 4 experiments, with FACS analysis of a minimal cell population of 15,000 for each condition and each experiment. The asterisk (*) indicates p<0.05 of a Mann & Whitney test.

Figure 3. Knockdown of FXR1P induces premature cell cycle exit of myoblasts.

(A) Immunofluorescence analysis of C2C12 cells transfected with siControl or siFxr1. Nuclei are stained with DAPI (blue) and cells expressing the proliferation marker Ki67 are labelled with FITC antibody (green). Scale bar: 75 µm. (B) Quantification of the number of DAPI-stained nuclei. (C) Quantification of the number of Ki67-positive cells over total number of nuclei quantified in (B). Quantification was performed using a macro developed with the ImageJ software. Data presented are mean of n = 4 experiments with analysis of 10 optical fields for each condition and each experiment. The asterisk (*) indicates p<0.05 for the Student T-test.

Figure 4. FXR1P depletion in C2C12 cells and in myoblasts derived from FSHD myopathic patients biopsies contributes to a consistent increase in p21 mRNA that translates into enhanced p21 protein levels.

(A) Quantitative RT-PCR reveals a significant increase of P21 mRNA in FSHD myoblats relative to control individuals. Data are presented as means ± SEM of n = 3 individuals/group. (B) Representative western-blot of p21 protein levels in siControl (siC) or siFxr1 (siFx)-transfected C2C12 cells. Densitometric quantification of western-blots reveal that depletion of FXR1P by siRNA transfection (siFxr1) leads to a significant increase of p21 protein levels relative to siControl-transfected cells. Data are presented as means ± SEM of n = 4 experiments. (C) Representative western-blot of P21 protein levels in FSHD patients and control individuals. Densitometric quantification of western-blots reveals that muscle biopsies of FSHD patients display a significant increase of P21 protein relative to controls. Data are presented as means ± SEM of n = 3 individuals/group. The asterisks * and ** indicate respectively p<0.05 and p<0.01 of the Mann & Whitney test.

Figure 5. FXR1P overexpression in Fxr1-depleted C2C12 cells restores p21 mRNA levels to normal.

(A) Western-blot analysis (upper panel) of C2C12 cells transfected with empty pTL1 vector or pTL1.FXR1 Isoe (pTL1.Isoe) construct indicate a strong expression of FXR1P long isoform Isoe in transfected myoblasts. Quantitative RT-PCR (lower panel) reveals a significant decrease of p21 mRNA levels in C2C12 myoblasts overexpressing FXR1 Isoe, as compared to control. Data are presented as means ± SEM of n = 3 independent experiments. (B) Western-blot analysis (upper panel) of C2C12 cells transfected with control siRNA (siC) or siFxr1 (siFx) and empty pTL1 vector or a mutated version of pTL1.FXR1 Isoe (pTL1.Isoe*) bearing 4 mismatches in siFxr1 recognition sequence indicate a reexpression of FXR1P long isoform Isoe in Fxr1-depleted transfected myoblasts. In the western blot FXR2P is indicated by (*) Quantitative RT-PCR (lower panel) reveals a significant increase of p21 mRNA levels in C2C12 myoblasts transfected with siFxr1 (siFx) and the empty vector (pTL1), as compared to control. This increase is restored to normal levels when FXR1P Isoe expression is rescued by transfection of pTL1.FXR1 Isoe. Data are presented as means ± SEM of n = 3 independent experiments. The asterisks * indicate p<0.05 of the Wilcoxon paired test, ns indicates non significance.

Figure 6. FXR1P selectively binds in vitro to the distal portion of p21 mRNA 3′ UTR and associates in vivo with p21 mRNA.

(A) Scheme of the various portions of p21 mRNA 3′ UTR (α, β and γ) used for in vitro binding assays. Note that the α fragment contains a characterized ARE motif. (B) Nitrocellulose filter binding assays to determine the portion of p21 mRNA bound by FXR1P. Radiolabeled mRNA probes were incubated with increasing concentrations of recombinant FXR1P Isoe protein, the amount of radioactive probes recovered on filters after binding reaction is then plotted against the concentration of proteins. The portion of FMR1 mRNA called N19 (known to be bound by FXR1P) and its truncated version (N19Δ35) were used as controls. This reveals that the distal portion of p21 3′UTR (γ fragment) and N19 are selectively bound by FXR1P. Both the α and β fragments from the 3′UTR of p21 remain at background levels comparable to N19Δ35 binding to FXR1P. (C) Western-blot analysis of UV-crosslinking and immunoprecipitation (CLIP) assay performed on C2C12 lysates using polyclonal antibodies raised against the C-terminus of FXR1P (#830) and control rabbit IgG (R). Input lysates (lanes 1 & 2, Input, 1/50^th), immunoprecipitates (lanes 3 & 4, IP, 1/5^th) and post-immunoprecipitation supernatants (lanes 5 & 6, post, 1/50^th) were probed for FXR1P using the 3FX antibody. A selective enrichment in FXR1P medium and long isoforms is observed in #830 immunoprecipitate (lane 3), concomitant with a depletion in these isoforms in the post-immunoprecipitation supernatant (lane 4) as compared to corresponding controls (lane 3 & 5). (D) RT-PCR analysis of mRNAs associated with FXR1P complexes. RNA was extracted from input and immunoprecipitate fractions described in (C), and used as template for RT-PCR. RT-PCR products obtained from inputs and immunoprecipitations respectively from control with rabbit IgG (Lanes 1, 3) and immunoprecipitation of FXR1P using #830 (Lanes 2, 4) were separated and visualized by agarose gel electrophoresis. This reveals that p21 mRNA is selectively enriched in the #830 immunoprecipitates, while the mRNAs encoding the myogenic determination factors Myogenin and MyoD or the unrelated mRNA encoding β-tubulin are not recovered in any immunoprecipitates. The symbol # indicates aspecific PCR products corresponding to β-tub primers dimers. DNA molecular weight markers presented on the gels are respectively 100, 200, 300, 400, 500, 600, 800 and 1000 bp.

Figure 7. The γ portion of p21 mRNA 3′UTR modulates the stability of the mRNA that is potentiated by FXR1P depletion.

(A) Scheme of the constructs bearing various portions of p21 mRNA 3′UTR (α, β and γ) used for luciferase assays. (B) Effect of p21 3′UTR−α, −β and −γ fragments on Renilla luciferase (Ren) mRNA levels in C2C12 cells transfected with control siRNAs (siControl) or siRNAs targeting Fxr1 (siFxr1). Quantitative RT-PCR analysis of the levels of Ren mRNA normalised to Firefly (Luc) mRNA relative to the empty construct are presented. In siControl cells, only the γ fragment significantly increased Ren mRNA levels, this effect is potentiated by Fxr1 depletion with siFxr1. In contrast, the α and β fragment have no effect on Ren mRNA levels, in the presence or absence of FXR1P. The results are presented as the means ±SEM of 4 experiments. (C) Effect of p21 3′UTR and its α, β and γ fragments on Renilla Luciferase activity in C2C12 cells transfected with control siRNAs (siControl) or siRNAs targeting Fxr1 (siFxr1). Results presented here represent the mean of the ratio of Luc-FL, Luc-α, Luc-β and Luc-γ to Luc-empty signal. In siControl cells, only the γ fragment significantly increased luciferase activity. In siFxr1 transfected cells compared to controls, the β and γ fragments increased luciferase activity, while the α fragment has no effect. However, the amplitude of variation is greater with the γ fragment and this effect is potentiated by Fxr1 depletion. Six independent experiments in triplicate for each transfection were quantified. For each transfection, Renilla was normalized to Firefly luciferase activity. RLU, relative luciferase units. (D) Fxr1-depletion increases the stability of endogenous p21 mRNA. C2C12 transfected with siControl (empty squares) or siFxr1 (black squares) were treated with the transcription inhibitor actinomycin D for 8 hrs. p21 mRNA levels were determined by quantitative RT-PCR at several time points and normalised to levels before treatment (t0). Percentage of remaining mRNA is plotted using a semi-log scale. Data presented represent the mean of n = 3 experiments. The asterisks * indicate p<0.05 of the Mann & Whitney test, while # and ## indicate respectively p<0.05 and p<0.01 of the Wilcoxon test.

Figure 8. The γ portion of p21 mRNA 3′UTR contains an evolutionary conserved G-quadruplex structure with mRNA stabilization properties.

(A) Cation-dependent termination of reverse transcription in the 3′-UTR full-length (FL) or γ fragment of p21 mRNA. Strong and weak pauses of reverse transcriptase (RT) are, respectively, indicated by large and thin arrows. Numbers correspond to positions of RT pauses, position +1 being the first nucleotide following the stop codon. (B) Localization and conservation of the G-quadruplex structure detected in (A) on the sequences of p21 3′UTR from Mus musculus (Mmu) and Homo sapiens (Hsa). (C) Scheme of the constructs used for luciferase assays bearing the conserved G-quadruplex of γ p21 mRNA (boxed) and two versions where the G-quadruplex has been deleted partially (Δ9) and fully (Δ38). (D) Effect of p21 3′UTR G-quadruplex and its deletions on Renilla luciferase (Ren) mRNA stability in C2C12 cells. Quantitative RT-PCR analysis of the levels of Ren mRNA normalised to Firefly (Luc) mRNA relative to the levels of the empty construct. The γ fragment bearing the G-quadruplex significantly increases Ren mRNA levels relative to empty vector. Partial or full deletion of the G-quadruplex sequence strongly increases Ren mRNA levels, both relative to empty vector and to the G-quadruplex bearing fragment. The results are presented as the mean of 4 experiments (±SEM). The asterisks * and # indicate p<0.05 respectively of the Wilcoxon test or of the Mann & Whitney test.