Rbfox proteins regulate tissue-specific alternative splicing of Mef2D required for muscle differentiation

Abstract

Among the Mef2 family of transcription factors, Mef2D is unique in that it undergoes tissue-specific splicing to generate an isoform that is essential for muscle differentiation. However, the mechanisms mediating this muscle-specific processing of Mef2D remain unknown. Using bioinformatics, we identified Rbfox proteins as putative modulators of Mef2D muscle-specific splicing. Accordingly, we found direct and specific Rbfox1 and Rbfox2 binding to Mef2D pre-mRNA in vivo. Gain- and loss-of-function experiments demonstrated that Rbfox1 and Rbfox2 cooperate in promoting Mef2D splicing and subsequent myogenesis. Thus, our findings reveal a new role for Rbfox proteins in regulating myogenesis through activation of essential muscle-specific splicing events.

Keywords: Alternative splicing; Mef2; Muscle differentiation; Rbfox.

Fig. 1.

Rbfox proteins are putative positive regulators of Mef2D splicing and specifically associate with Mef2D pre-mRNA. (A) Position of Rbfox-binding site (Rbfox BS) with respect to α2 exon within Mef2D pre-mRNA. (B) RT-qPCR of Rbfox1 and Rbfox2 expression during muscle differentiation, relative to Gapdh. (C) Immunoblotting of Rbfox1 and Rbfox2 proteins during muscle differentiation. Tubulin is used as loading control. (D) Top, scheme indicating primer sets position inside Mef2D. Bottom, RT-qPCR of Mef2D splicing pattern during muscle differentiation. Mef2D α1 and α2 isoforms levels are relative to total Mef2D transcripts, measured using a primers set on a constitutive exon (Mef2D all isof.), and represented as fold expression over myoblast stage. (E) Top, position of control primers (Mef2D exon 2, gray) and primers specific for the Rbfox-binding site (Mef2D Rbfox BS, black) in Mef2D pre-mRNA. Bottom, UV-RIP in wild-type myotubes using anti-Rbfox1 (αRbfox1), anti-Rbfox2 (αRbfox2) and anti-IgG (αIgG-control) antibodies. Immunoprecipitated RNA was analyzed by RT-qPCR with primers spanning the Rbfox-binding site (Mef2D Rbfox BS, black), a different region of Mef2D transcript (Mef2D exon 2, grey) or the Gapdh control pre-mRNA (white), and expressed as percentage of input. MB, myoblasts; MT, myotubes; d1, d2, d3: day 1, day 2, day 3). Results are mean±s.e.m., n≥3. *P<0.05; **P<0.01 (Student's t-test).

Fig. 2.

Alteration of Rbfox1 or Rbfox2 levels affects Mef2D muscle-specific splicing. (A) RT-qPCR of Rbfox1 and Rbfox2 expression in myotubes (MT) transfected with siRNAs targeting selectively Rbfox1 (siRbfox1) or Rbfox2 (siRbfox2), a mixture of both siRNAs (siRbfox1+2) or a non-silencing control (siNSL). (B) Immunoblotting of Rbfox1 and Rbfox2 proteins from cells in A. Tubulin is used as loading control. (C) RT-qPCR of Mef2D isoforms expression in the same cells in A. (D) RT-qPCR of Rbfox1 and Rbfox2 expression in myoblasts (MB) transfected with either pIRES-Rbfox1 or pIRES-Rbfox2 or both. EV, empty vector control. (E) Immunoblotting of Rbfox1 and Rbfox2 proteins from cells as in D. Tubulin is used as loading control. (F) RT-qPCR of Mef2D splicing pattern in the same cells as in D. Rbfox1 and Rbfox2 expression levels are relative to Gapdh. Mef2D splicing isoforms levels are relative to the total amount of Mef2D transcripts, assayed by a primer set on a constitutive exon (Mef2D all isof.). Data in D and F are represented as fold expression over EV control. Results are mean±s.e.m., n≥3. *P<0.05; **P<0.01; ***P<0.001 (Student's t-test).

Fig. 3.

Rbfox1 and Rbfox2 are both required for muscle differentiation. (A) RT-qPCR of Rbfox1 and Rbfox2 expression in myotubes (MT) infected with lentiviral particles (LV) carrying a shRNA targeting either Rbfox1 (shRbfox1), Rbfox2 (shRbfox2), a mixture of both (shRbfox1+2) or a non-silencing control (shNSL). Data are normalized over Gapdh and shown as fold expression over shNSL control. (B) RT-qPCR of Mef2D isoforms expression in the same cells in A, normalized to the amount of total Mef2D transcripts (Mef2D all isof.) and shown as fold expression over shNSL control. (C) Immunoblotting analysis of endogenous Mef2D α1 and α2 isoforms in stable Rbfox double-depleted cells. Mef2D pan antibody is used as loading control. (D) Representative immunostaining of myosin heavy chain (Mhc, green) in differentiated cells as described in A. Hoechst 33342 was used to stain nuclei (blue). Scale bars: 200 µm. (E) Quantification of terminally differentiated myotubes in the cell types shown in D. (F) Immunoblotting of Mhc in the same cells as in D. Tubulin is used as loading control. Results are mean±s.e.m., n≥3. *P<0.05; **P<0.01; ***P<0.001 (Student's t-test).

Fig. 4.

The differentiation defect of Rbfox1 and Rbfox2 double-knockdown cells is selectively rescued by Mef2D α2. (A) Representative immunofluorescence images for myosin heavy chain (Mhc) or DAPI performed in C2C12 myotubes stably expressing shRNAs targeting both Rbfox1 and Rbfox2 (shRbfox1+2) or a non-silencing control (shNSL) infected with retroviruses expressing either the ubiquitous (Mef2D α1) or muscle-specific (Mef2D α2) isoforms of Mef2D. Scale bars: 100 µm. (B) Quantification of terminally differentiated myotubes in the cell types shown in A. Values are presented as mean±s.e.m., n≥3. **P<0.01; ns, not significant (Student's t-test). (C) Mef2D splicing isoforms overexpression visualized using Mef2D pan antibody.