Postsynaptic chromatin is under neural control at the neuromuscular junction

Abstract

In adult skeletal muscle, the nicotinic acetylcholine receptor (AChR) specifically accumulates at the neuromuscular junction, to allow neurotransmission. This clustering is paralleled by a compartmentalization of AChR genes expression to subsynaptic nuclei, which acquire a unique gene expression program and a specific morphology in response to neural cues. Our results demonstrate that neural agrin-dependent reprogramming of myonuclei involves chromatin remodelling, histone hyperacetylation and histone hyperphosphorylation. Activation of AChR genes in subsynaptic nuclei is mediated by the transcription factor GABP. Here we demonstrate that upon activation, GABP recruits the histone acetyl transferase (HAT) p300 on the AChR epsilon subunit promoter, whereas it rather recruits the histone deacetylase HDAC1 when the promoter is not activated. Moreover, the HAT activity of p300 is required in vivo for AChR expression. GABP therefore couples chromatin hyperacetylation and AChR activation by neural factors in subsynaptic nuclei.

Figure 1

Chromatin is decondensed in subsynaptic nuclei at the NMJ. Electron microscopy of subsynaptic (A) and extrasynaptic (B) areas of Tibialis anterior muscle of 5-week-old OF1 male mice. The electron density of subsynatic nucleus (SN) located just beneath the nerve ending (NE) appears lower than that of the extrasynaptic nucleus (EN). This is directly correlated with the higher chromatin condensation in ES nucleus. Scale bar, 2 μm.

Figure 2

Chromatin is hyperacetylated and hyperphosphoacetylated in subsynaptic nuclei. Immunofluorescence on isolated mouse Tibialis anterior muscle fibres using antibodies specific for H3 modifications. (A, B) AcH3: histone H3 acetylated on K9 and K14. (C, D) PAcH3: histone H3 phosphorylated on S10 and acetylated on K14. The NMJ was stained with Alexa488 α-Bungarotoxin (BGT) and nuclei with Hoechst 33258. Scale bars, 50 μm (A, C) and 10 μm (B, D). Dashed lines represent the limits of the muscle fibres.

Figure 3

Extrasynaptic expression of neural agrin induces hyperacetylation and hyperphosphoacetylation at ectopic synapses. Ectopic synapses were induced by injecting recombinant agrin into adult mouse Soleus muscle fibres. Muscle fibres were isolated 14 days later and histone modification level was analysed by immunofluorescence as described in Figure 2 with antibodies specific for AcH3 (A) and PAcH3 (B). Scale bar, 20 μm. Dashed lines represent the limits of the muscle fibres.

Figure 4

Neuregulin induces histone acetylation and phosphorylation on AChR ɛ gene promoter. (A) Neuregulin induces histone acetylation in cultured myotubes. C2C12 myotubes were treated with neuregulin (for indicated time) and 30 μg of whole-cell extracts was analysed by Western blot using antibodies specific for histone H3 modifications (AcH3: histone H3 acetylated on K9 and K14; PH3: histone H3 phosphorylated on S10) and β-tubulin as loading control. Diagrams correspond to Western blot quantification using Image J software. (B) Neuregulin induces modification of histone H3 on the AChR ɛ gene promoter. ChIP experiments were performed on C2C12 myotubes treated with neuregulin. Histone modification levels on the cJun and AChR ɛ promoter genes were analysed by ChIP using AcH3- and PAcH3-specific antibodies. The data are means±s.e.m. of three independent experiments; P<0.05 (t-test).

Figure 5

HDAC1 level is reduced in the presence of neuregulin. (A) Myotube cultures were treated with neuregulin for the indicated time and were analysed by immunofluorescence using an anti-HDAC1 antibody. Scale bars, 20 and 2.5 μm for magnifications. Dashed lines symbolize the plasma membrane of the myotubes. (B) Differentiated myotube cultures were treated with neuregulin (indicated time) and 30 μg of whole-cell extracts was analysed by Western blot using antibodies specific for HDAC1 and β-tubulin as loading control. (C) HDAC1 is removed from the AChR ɛ promoter in the presence of neuregulin. ChIP experiments were performed with an anti-HDAC1 antibody on control myotubes and on myotubes treated for 6 h with neuregulin. The data correspond to the means±s.e.m. of three independent experiments; P<0.01 (t-test).

Figure 6

p300 HAT activity is required for the expression of AChR ɛ gene. (A) Neuregulin induces the recruitment of p300 on the AChR ɛ promoter in cultured myotubes. The ChIP experiments were performed with an antibody specific for p300 concomitantly to those described in Figure 5C. The data correspond to the means±s.e.m. of three independent experiments; P<0.05 (t-test). (B) The CBP/p300 HAT inhibitor LysCoA inhibits the expression of the AChR ɛ gene in vivo. NaCl (0.9 %), lysine (1 mM) and CoA (1 mM) or LysCoA (1 mM) were injected and electroporated in Tibialis anterior muscles and gene expression was evaluated 1 day later by real-time RT–PCR. The data correspond to the means±s.e.m. of the results obtained with five mice. P<0.05 (t-test). The experiment was reproduced three times. In addition, similar results were obtained when gene expression was measured 2 days after electroporation. (C) LysCoA represses AChR ɛ promoter-dependent transcriptional activation. A reporter construct in which the AChR ɛ promoter was placed upstream of the luciferase gene was electroporated into Tibialis anterior muscle with increasing doses of LysCoA. Whole-muscle luciferase activity was measured 1 day after electroporation. The data correspond to the means±s.e.m. of the results obtained with three mice. P<0.05 (t-test). The experiment was reproduced five times. In addition, similar results were also obtained when luciferase expression was measured 2 and 3 days after electroporation.

Figure 7

GABP selectively interacts with p300 or HDAC1. (A) GABPα interacts with p300 and HDAC1. C2C12 cells were cotransfected with GABPα and Flag-p300 or Flag-HDAC1 expression vectors. Twenty-four hours following transfection, cell lysates were prepared and the immunoprecipitation performed with an anti-GABPα antibody. The fractions were subsequently separated on SDS–PAGE, followed by immunoblotting with anti-Flag and anti-GABPα antibodies. The experiment was repeated three times. (B) Endogenous GABPα interacts with endogenous p300 and HDAC1, and neuregulins favour the recruitment of p300. Untreated and neuregulin-treated myotube (6 h) lysates were immunoprecipitated with an anti-GABPα antibody and the immune complex was analysed by immunoblotting with anti-p300 or anti-HDAC1 antibodies. The experiment was repeated three times.

Figure 8

A model for synaptic gene regulation at the NMJ. Agrin and neuregulin are accumulated in the basal lamina of the synaptic cleft and activate their muscle receptor to induce local activation of intracellular signalling pathways, which in turn activate the transcription factor GABP. In addition, they induce histone hyperacetylation and hyperphosphoacetylation, which participate in chromatin decondensation. The recruitment of p300 on synaptic genes by GABP in subsynaptic nuclei favours chromatin hyperacetylation and decondensation. Conversely, in extrasynaptic nuclei, GABP recruits the histone deacetylase HDCA1 on synaptic gene promoters, thereby promoting chromatin compaction.