Myogenin controls via AKAP6 non-centrosomal microtubule-organizing center formation at the nuclear envelope

Abstract

Non-centrosomal microtubule-organizing centers (MTOCs) are pivotal for the function of multiple cell types, but the processes initiating their formation are unknown. Here, we find that the transcription factor myogenin is required in murine myoblasts for the localization of MTOC proteins to the nuclear envelope. Moreover, myogenin is sufficient in fibroblasts for nuclear envelope MTOC (NE-MTOC) formation and centrosome attenuation. Bioinformatics combined with loss- and gain-of-function experiments identified induction of AKAP6 expression as one central mechanism for myogenin-mediated NE-MTOC formation. Promoter studies indicate that myogenin preferentially induces the transcription of muscle- and NE-MTOC-specific isoforms of Akap6 and Syne1, which encodes nesprin-1α, the NE-MTOC anchor protein in muscle cells. Overexpression of AKAP6β and nesprin-1α was sufficient to recruit endogenous MTOC proteins to the nuclear envelope of myoblasts in the absence of myogenin. Taken together, our results illuminate how mammals transcriptionally control the switch from a centrosomal MTOC to an NE-MTOC and identify AKAP6 as a novel NE-MTOC component in muscle cells.

Keywords: AKAP6; MTOC; cell biology; centrosome; developmental biology; human; microtubules; mouse; muscle differentiation; myogenin.

Figure 1.. Myogenin is required for microtubule-organizing center (MTOC) protein localization to the nuclear envelope.

(A, B) C2C12 cells were differentiated for 1 day and immunostained for the myogenic regulatory factors (MRFs) MyoD or myogenin (Myog) and nesprin-1α (A) or PCM-1 (B). Orange asterisks: MRF+/PCM-1+ nuclei; yellow asterisks: MRF+/nesprin-1α+ nuclei; arrowheads: MRF+/nesprin-1α- nuclei; arrows: MRF-/nesprin-1α+ nuclei. (C) Quantification of (A) and (B). (D) Quantification of MyoD and Myog in relation to nesprin-1α showing that not all nesprin-1α+ nuclei are myogenin+. (E, F) C2C12 myoblasts were transfected with negative control (si-ctrl) or myogenin siRNA (si-Myog) and differentiated for 2 days. Immunostaining (E) and subsequent quantification (F) shows that myogenin depletion affects nuclear envelope localization of PCM-1, PCNT, and AKAP9 but not of nesprin-1α. 95% CI of differences si-Myog vs. si-ctrl = –3.11% to 0.98% (nesprin-1α+), –11.31% to –7.22% (PCM-1+), –7.53% to 3.43% (PCNT+), and –7.52% to –3.42% (AKAP9+). (G, H) C2C12 myoblasts were transfected with si-ctrl or Pcm1 siRNA (si-PCM1) and differentiated for 2 days. PCNT was detected by immunostaining (G) and subsequent quantification (H) showed that PCM-1 depletion reduces PCNT nuclei. 95% CI of differences si-PCM1 vs. si-ctrl = –6.6% to –3.4% (full), –3.78% to –0.59% (partial). Scale bars (A, B, E, G): 20 µm. Data (C, D, F, H) are represented as individual biological replicates (n = 3), together with mean ± SD. ns: p>0.05; *p<0.05; ***p<0.001.

Figure 1—figure supplement 1.. Intermediate stages of microtubule-organizing center (MTOC) protein localization to the nuclear envelope.

(A) C2C12 cells were differentiated for 24 hr and PCM-1 was detected by immunostaining. Representative images of different intermediate stages of PCM-1 recruitment are shown. Scale bar: 10 µm. (B) A more detailed quantification of PCNT+ and AKAP9+ nuclei in differentiating C2C12 (treated as described in Figure 1F) showing that myogenin depletion affects also partial recruitment of PCNT and AKAP9. (C) C2C12 myoblasts were transfected with si-ctrl or Pcm1 siRNA (si-PCM1) and differentiated for 2 days. AKAP9 was detected by immunostaining and subsequent quantification showed that PCM-1 depletion does not affect AKAP9+ nuclei. Data (B, C) are represented as individual biological replicates, together with mean ± SD. ns: p<0.05; ***p<0.001; n = 3.

Figure 1—figure supplement 2.. PCM-1 exhibits a centrosomal localization pattern at nesprin-1+ nuclei in myogenin-depleted C2C12 cells.

C2C12 myoblasts were transfected with negative control (si-ctrl) or myogenin siRNA (si-Myog) and differentiated for 2 days. In si-ctrl cultures, PCM-1 frequently localizes to the envelope of highly nesprin-1+ nuclei (yellow arrow). In si-Myog cultures, PCM-1 can be found in a centrosomal pattern near nesprin-1+ nuclei (yellow arrowhead). Scale bar: 10 µm.

Figure 2.. Myogenin expression is sufficient to induce nuclear envelope microtubule-organizing center (NE-MTOC) formation in non-muscle cells.

(A) NIH3T3 fibroblasts were transfected with constructs encoding GFP, MyoD-GFP or myogenin-GFP (Myog-GFP). After three days, PCM-1 localization was assessed by immunostaining. Arrows indicate nuclei of transfected cells which have recruited PCM‑1. Scale bars: 10 µm. (B) Quantification of (A) demonstrating that myogenin induces nuclear envelope localization of PCM‑1 more efficiently than MyoD. Data are represented as individual biological replicates (n = 3), together with mean ± SD. ***: p < 0.001; 95% CI of difference Myog-GFP vs. MyoD-GFP = 30.99% to 49.22%. n = 3. (C-H) NIH3T3 Tet-ON mScarlet or MYOG-2A-mScarlet (MYOG-mScarlet) cells were treated with doxycycline (Dox) for three days. After immunostaining, nuclear envelope localization of PCM-1 (C-D), PCNT (E-F), and AKAP9 (G-H) was analyzed and quantified. Data are depicted as violin plots. Red line indicates the median, dotted lines indicate the 25% and 75% percentile. ***: p < 0.001. Scale bars: 20 µm (I) Immunostaining of MYOG-mScarlet cells treated with Dox for three days showing the presence of nesprin‑1α+ nuclei. Scale bars: 20 µm (J) RT-PCR analysis of MYOG-mScarlet cells in the absence of Dox (-Dox) or treated with Dox for the indicated time points demonstrating that nesprin‑1α is upregulated upon myogenin expression. Gapdh was used as equal input control. (K) ChIP-PCR analysis of Dox-treated MYOG-mScarlet cells using an anti-myogenin antibody or an IgG1 control showing that myogenin binds an E-box in the nesprin-1α promoter region. (L-M) Immunostaining of α-tubulin and subsequent quantification of nuclear envelope coverage after 30s of microtubule regrowth following cold-induced microtubule depolymerization in mScarlet or MYOG-mScarlet cells treated with Dox for three days. Data are depicted as violin plots. Red line indicates the median, dotted lines indicate the 25% and 75% percentile. ***: p < 0.001. Scale bars: 20 µm. N numbers indicate total number of analyzed nuclei pooled from three biological replicates.

Figure 2—figure supplement 1.. Myogenin is required for MyoD-induced nuclear envelope recruitment of PCM-1.

(A) Three days after ectopic expression of MyoD-GFP in NIH3T3 fibroblasts, immunostaining revealed that nuclear PCM-1 localization is associated with high myogenin (Myog) expression (yellow arrowhead). (B) NIH3T3 cells were treated with negative control (si-ctrl) or myogenin (si-Myog) siRNA and 24 hr later transfected with a construct encoding MyoD-GFP. Immunofluorescence staining indicates that myogenin is required for MyoD-induced PCM-1 nuclear recruitment. (C) C2C12 cells were transfected with the indicated siRNAs and differentiated for 2 days. RT-PCR analysis reveals that depletion of MyoD reduces myogenin levels in differentiating myoblasts but not vice versa. Scale bars (A, B): 10 µm.

Figure 2—figure supplement 2.. Myogenin targets are specifically precipitated in induced MYOG-mScarlet cells.

Chromatin immunoprecipitation (ChIP) was performed on the indicated lysates using the anti-myogenin antibody or an isotype control (IgG). The relative abundance of (A) the Syne1 α-isoform promoter region, (B) an intronic region of Syne1 (negative control), or (C) the desmin promoter region (positive control) were determined by qPCR. Data are represented as individual biological replicates (n = 3), together with mean ± SD. ns: p>0.05; ***p<0.01; ****p<0.001.

Figure 2—figure supplement 3.. Microtubule depolymerization in mScarlet and MYOG-mScarlet cells.

Immunostaining of α-tubulin in mScarlet or MYOG-mScarlet cells treated with doxycycline (Dox) for 3 days following cold-induced microtubule depolymerization without regrowth. Scale bar = 20 µm.

Figure 3.. Myogenin expression attenuates the centrosomal microtubule-organizing center (MTOC).

(A–E) mScarlet or MYOG-mScarlet cells were stimulated with doxycycline (Dox) for 3 days and PCNT (A), Cep135 (C), and γ-tubulin (A, C) were detected by immunostaining. Quantification shows that PCNT (B) and γ-tubulin (E) intensities at the centrosome are reduced upon myogenin induction while Cep135 intensity (D) is not significantly affected. Single-channel images of Pcnt, γ-tubulin, and Cep135 are false-colored to visualize different intensities. Data are shown as violin plots. The red line indicates the median, and dotted lines indicate the 25% and 75% percentile. ns: p>0.05; ***p<0.001. Scale bars = 5 µm. N numbers indicate the total number of analyzed centrioles (y-tubulin foci) pooled from four biological replicates. (F) Immunostaining of α-tubulin and γ-tubulin in Dox-stimulated mScarlet or MYOG-mScarlet cells after 30 s of microtubule regrowth. Intensity-based color coding of α-tubulin shows that microtubule growth from centrioles is reduced after myogenin induction. Scale bars: 5 µm.

Figure 4.. The nesprin-1α interaction partner AKAP6 is a potential mediator of myogenin-induced nuclear envelope microtubule-organizing center (NE-MTOC) formation.

(A) Scheme illustrating the bioinformatics workflow used to identify potential myogenin downstream candidates. (B) Venn diagram depicting the numbers of genes matching criteria for the individual data sets and for intersection of data sets. Criteria for myogenin ChIP-seq data (red): Genes where myogenin binding was detected at the promoter region; criteria for C2C12 RNA-seq data (green) and for microarray data of rat heart development (blue): upregulated genes. (C) ChIP-PCR analysis of doxycycline (Dox)-treated MYOG-mScarlet cells using an anti-myogenin antibody or an IgG1 control showing that myogenin binds an E-box in the Akap6β promoter region. (D) RT-PCR analysis of MYOG-mScarlet cells in the absence of Dox (-Dox) or treated with Dox for the indicated time points demonstrating that Akap6β is upregulated upon myogenin expression. The two bands for Akap6β derive from alternative splicing of the first exon of Akap6β, which results in an ~200 bp insertion in the 5’ untranslated region. Gapdh was used as equal input control. Please note that the same samples and Gapdh control were used as in Figure 2J. (E) C2C12 cells were differentiated for 2 days, and immunostaining shows that all AKAP6⁺ nuclei are also nesprin-1α⁺. Scale bar: 20 µm. (F) High-resolution Airyscan image of (E). Arrowhead indicates AKAP6 localized at the cytoplasmic side of nesprin-1α signal. Arrow marks nesprin-1α that is localized at the nuclear side of AKAP6 signal. Scale bar: 0.5 µm. (G) Myoblasts from healthy donors (wt) and from patients carrying a mutation in the SYNE1 gene (SYNE1^-/-) were differentiated for 4 days. Immunostaining analysis showed that loss of nesprin-1α is associated with loss of AKAP6 from the nuclear envelope in differentiated myotubes (troponin I). Scale bars: 10 µm.

Figure 4—figure supplement 1.. The myogenin antibody specifically precipitates the Akap6β promoter region in induced MYOG-mScarlet cells.

Indicated lysates were probed using the anti-myogenin antibody or an isotype control (IgG). The relative abundance of (A) the Akap6β promoter region or (B) an intronic region of Akap6 was determined by qPCR. Data are represented as individual biological replicates (n = 3), together with mean ± SD. ns: p>0.05; ****p<0.001.

Figure 4—figure supplement 2.. Quantification of nesprin-1α and AKAP6 signal at the nuclear envelope.

Differentiated C2C12 cells were immunostained for nesprin-1α and AKAP6. Signal intensity was quantified as line scan profiles across the nuclear envelope. 10 line scan profiles were obtained in regular intervals along the nuclear periphery and pooled from 10 nuclei. Data are shown as mean ± SD.

Figure 5.. AKAP6 is required for the nuclear envelope localization of microtubule-organizing center (MTOC) proteins.

(A) C2C12 cells were differentiated for 2 days. Immunostaining shows that all PCM-1+ nuclei are also AKAP6+. (B) Quantification of AKAP6+ and PCM-1 nuclei in C2C12 cells treated with negative control (si-ctrl) or Akap6 (si-Akap6) siRNA after 2 days of differentiation indicates that AKAP6 is required for nuclear envelope localization of PCM-1. Data are represented as individual biological replicates (n = 3), together with mean ± SD. 95% CI = 6.21% to 8.74%; 95% CI = 4.63% to 9.43%. (C) MYOG-mScarlet cells were treated with si-ctrl or si-Akap6 and subsequently treated with doxycycline (Dox) for 3 days. Image analysis revealed that myogenin-induced localization of PCM-1 to the nuclear envelope is AKAP6-dependent. (D) Quantification of (C). (E, F) Quantification of PCNT (E) and AKAP9 (F) nuclear coverage in Dox-stimulated MYOG-mScarlet cells treated with si-ctrl or si-Akap6. (G) Co-immunoprecipitation (IP) of PCM-1 from MYOG-mScarlet but not from mScarlet lysate (L) using an anti-AKAP6 antibody. (H, I) Enriched C2C12 myotubes (troponin I) were transfected with si-ctrl or si-Akap6 and immunostaining demonstrates that AKAP6 is required for maintaining nuclear envelope localization of PCM-1 (H) as well as PCNT and AKAP9 (I). Scale bars (A, H) 20 µm, (C, I) 10 µm. Data (D–F) are shown as violin plots. The red line indicates the median, and dotted lines indicate the 25% and 75% percentile. N numbers indicate the total number of analyzed nuclei pooled from three biological replicates. *p<0.05.; **p<0.01; ***p<0.001.

Figure 5—figure supplement 1.. AKAP6 depletion does not affect nesprin-1α.

Enriched C2C12 myotubes were transfected with non-targeting (si-ctrl) or Akap6 (si-Akap6) siRNA and immunostained for AKAP6 and nesprin-1α. Scale bars: 10 μm.

Figure 5—figure supplement 2.. AKAP6 or nesprin-1α depletion does not prevent centrosome attenuation in MYOG-expressing cells.

(A, B) mScarlet or MYOG-mScarlet transfected with the indicated siRNAs were stimulated with doxycycline (Dox) for 3 days. PCNT (A) and γ-tubulin (B) immunostaining intensities were quantified. Data are shown as violin plots. The red line indicates the median, and dotted lines indicate the 25% and 75% percentile. N numbers are the total number of analyzed centrioles pooled from three biological replicates. Note that here PCNT and γ-tubulin intensities were measured simultaneously.

Figure 5—figure supplement 3.. MyoD induces nesprin-1α and AKAP6.

(A) NIH3T3 cells were transfected with MyoD-GFP. Two days after transfection, nesprin-1α and AKAP6 were detected via immunostaining. Scale bar: 10 µm. (B) NIH3T3 cells were transfected with non-targeting siRNA (si-ctrl) or myogenin siRNA (si-Myog) and 1 day later transfected with MyoD-GFP. Indicated microtubule-organizing center (MTOC) components were detected using immunostaining and positive nuclei were quantified. Data are represented as mean ± SD. n = 3. ns: p>0.05; **p<0.01; ***p<0.001.

Figure 6.. AKAP6 is required for correct nuclear positioning in myotubes.

(A) Scheme illustrating the role of the nuclear envelope microtubule-organizing center (NE-MTOC) in myonuclear positioning and the potential impact of AKAP6 depletion. (B) Enriched C2C12 myotubes (troponin I) were transfected with negative control (si-ctrl) or Akap6 (si-Akap6) siRNA. The upper si-Akap6 panel shows a representative image of a myotube with misaligned nuclei, and the lower si-Akap6 panel shows nuclei overlapping inside a myotube. (C) Quantification of (B). Data are represented as individual biological replicates (n = 3), together with mean ± SD. *p<0.05, 95% CI of difference si-Akap6 vs. si-ctrl = 10.42% to 52.25% (left graph); 95% CI = 9.65% to 55.02% (right graph). (D) Enriched C2C12 myotubes (troponin I) were transfected with si-ctrl or si-Akap6 and subsequently subjected to a nocodazole-based microtubule (α-tubulin) regrowth assay. Image analysis showed that AKAP6 depletion abrogated microtubule nucleation at the nuclear envelope. (E) Enriched C2C12 myotubes (troponin I) transfected with si-ctrl or si-Akap6 were immunostained for the dynein regulator p150glued. Image analysis showed that AKAP6 depletion reduces p150glued signal at the nuclear envelope. Scale bars (B) 20 µm, (D) 10 µm, and (E) 5 µm.

Figure 6—figure supplement 1.. AKAP6 depletion does not affect steady state microtubule organization but reduces detyrosinated microtubules.

(A, B) Enriched C2C12 myotubes (troponin I) were transfected with negative control (si-ctrl) or Akap6 (si-Akap6) siRNA and subsequently analyzed for overall microtubule (α-tubulin) structure (A) or detyrosinated microtubules (detyr α-tubulin). Scale bars: 10 µm.

Figure 7.. Myogenin preferentially induces microtubule-organizing center (MTOC)-associated isoforms of Syne1 and Akap6.

(A, C) Schematic representation of the murine Syne1 (A) and Akap6 (C) gene and derived transcripts. Exons are indicated by gray rectangles and the first exon of each transcript is marked by color. E-boxes (myogenin consensus sites) inside putative promoters are indicated as yellow boxes and small black arrows mark the primers used for qPCR. (B, D) Myogenin chromatin immunoprecipitation (ChIP) from doxycycline (Dox)-stimulated MYOG-mScarlet cells followed by qPCR for the indicated E-boxes shows that myogenin preferentially binds the promoter regions upstream of Syne1 α-isoform and Akap6 β-isoform transcripts. (E, F) Luciferase assay testing the activity of the indicated Akap6 (E) or Syne1 (F) promoters in the presence of GFP or myogenin-GFP (MYOG-GFP). (G) Overexpression of nesprin-1α-mCherry alone or together with AKAP6β-GFP in undifferentiated (myogenin-negative) C2C12 myoblasts. Co-expression of nesprin-1α and AKAP6β is sufficient for nuclear envelope recruitment of endogenous PCM-1. Scale bars: 20 µm. Data (B, D–F) are represented as individual biological replicates (n = 3), together with mean ± SD. ns: p>0.05; **p<0.01; ***p<0.001.

Figure 7—figure supplement 1.. Ectopic co-expression of nesprin-1α and AKAP6β is not sufficient for microtubule-organizing center (MTOC) function at the nuclear envelope.

Proliferating (myogenin-negative) C2C12 myoblasts were transfected with nesprin-1α-mCherry and AKAP6β-GFP. Two days after transfection, microtubule regrowth assays were performed. Scale bar: 10 µm.

Figure 8.. Schematic overview of the role of myogenin and AKAP6 in nuclear envelope microtubule-organizing center (NE-MTOC) formation.

Myogenin induces expression of AKAP6β that connects MTOC proteins like PCNT, AKAP9, and PCM-1 to the nuclear membrane protein nesprin-1α, whose expression can be induced by myogenin as well as MyoD. Depletion, overexpression, and co-immunoprecipitation experiments suggest that AKAP6β acts as an adapter between MTOC proteins and nesprin-1α. Yet, other proteins might be involved and the here presented protein complex at the nuclear envelope is hypothetical. At the same time, myogenin is sufficient to attenuate centrosomal MTOC function. AKAP6-dependent anchoring of MTOC proteins as well as microtubule nucleation from the nuclear envelope are required for correct positioning of nuclei inside differentiating myotubes.