Bruno-3 regulates sarcomere component expression and contributes to muscle phenotypes of myotonic dystrophy type 1

Abstract

Steinert disease, or myotonic dystrophy type 1 (DM1), is a multisystemic disorder caused by toxic noncoding CUG repeat transcripts, leading to altered levels of two RNA binding factors, MBNL1 and CELF1. The contribution of CELF1 to DM1 phenotypes is controversial. Here, we show that the Drosophila CELF1 family member, Bru-3, contributes to pathogenic muscle defects observed in a Drosophila model of DM1. Bru-3 displays predominantly cytoplasmic expression in muscles and its muscle-specific overexpression causes a range of phenotypes also observed in the fly DM1 model, including affected motility, fiber splitting, reduced myofiber length and altered myoblast fusion. Interestingly, comparative genome-wide transcriptomic analyses revealed that Bru-3 negatively regulates levels of mRNAs encoding a set of sarcomere components, including Actn transcripts. Conversely, it acts as a positive regulator of Actn translation. As CELF1 displays predominantly cytoplasmic expression in differentiating C2C12 myotubes and binds to Actn mRNA, we hypothesize that it might exert analogous functions in vertebrate muscles. Altogether, we propose that cytoplasmic Bru-3 contributes to DM1 pathogenesis in a Drosophila model by regulating sarcomeric transcripts and protein levels.

Keywords: Bruno-3; CELF1; Drosophila; Myotonic dystrophy type 1; RNA CLIP; mRNA stability.

Fig. 1.

Bru-3 is expressed in larval somatic muscles and enriched in pathological lines. (A) CELF1 sequence aligned with Bruno orthologs. lsm, linker-specific motif; RRM, RNA recognition motif. The N-terminus peptide designed to raise the antibody specifically against Bru-3 is highlighted in red. (B-E) Bru-3 expression in WT larval muscle. Immunostaining against Bru-3 (gray) shows that Bru-3 is expressed in larval somatic muscles more specifically in the sarcomeres (B,B′) as well as in granules around the nuclei, as indicated by white arrows (C). This antibody also detected weak Bru-3 expression in nuclei (B′). (D-E) Immunostaining of larval muscles with pre-immune serum shows the specificity of the antibody raised against Bru-3. (F) Densitometric measurements of Bru-3 in the nucleus and cytoplasm of different genotypes. Cytoplasmic signal measurement was performed using ImageJ as described previously by McCloy et al. (2014). DAPI was used as reporter of tissue accessibility for the staining, because sarcomeric markers expression is potentially altered in pathological contexts (see Fig. 4). *P<0.05, **P<0.01, ***P<0.001 versus WT. (G-L) Bru-3 expression in segment border muscle (SBM) of third-instar larvae. Sarcomeric immunostained Bru-3 (gray) frames the Z-line represented by Actn immunostaining (red) in WT muscle (G). Bru-3 immunostaining is also shown in Mef>bru-3(43) (H), Mef>bru-3(37) (I), Mef>960CTG (J), Mef>960CTG,Df(bru-3) (K) and Mef>bru-3RNAi (L) conditions. (M-R) Representative pictures of nuclear Bru-3 expression compared with corresponding DAPI staining (M′-R′) in third-instar larval muscle. Bru-3 immunostaining is shown in WT (M), Mef>bru-3(43) (N), Mef>bru-3(37) (O), Mef>960CTG (P), Mef>960CTG,Df(bru-3) (Q) and Mef>bru-3RNAi (R). Scale bars: 20 µM.

Fig. 2.

Bru-3 overexpression alters motility and contributes to fiber splitting. (A) Righting assay. The graph represents the average recorded time taken for the larvae of each genotype to turn over. (B-C) Assessment of overall muscle pattern and quantification of muscle abnormalities. Z-stacks of in vivo scans of muscle pattern in a single third-instar larvae abdominal segment (between A5 and A7) (B). Arrows point to splitting fibers and arrowheads to extra fibers, both represented by schemes in B′. Muscles taken into account for quantification are represented and named in WT context (B′): LT1, LT2, LT3 and LT4 refer to lateral transverse muscles; DT1, dorso transverse 1; DO1 and DO2, dorso oblique muscles. Graphs show the average number of each defect observed in vivo for each mutant line on a window of three abdominal segments (C). Color coding is as in A. *P<0.05, **P<0.01, ***P<0.001 versus Mef>lacZ. NS, not significant. ^#P<0.05, ^##P<0.01, ^###P<0.001 Mef>960CTG versus Mef>960CTG,Df(bru-3) or Mef>bru-3(37) versus Mef>bru-3(43). Scale bars: 70 µM.

Fig. 3.

bru-3 overexpression affects myoblast fusion process, but not contractility, in vivo. (A) The average length of abdominal VL3 fibers is significantly reduced in Mef>bru-3(37) and Mef>bru-3(43) lines, and in Mef>960CTG and Mef>960CTG,Df(bru-3) lines, compared with controls (Mef>lacZ and corresponding transgenic control line). (B,C) Myoblast fusion defect in DM1 condition is Bru-3 dependent. The average number of nuclei per abdominal VL3 fiber is significantly reduced in Mef>bru-3 and Mef>960CTG lines, but is rescued in Mef>960CTG,Df(bru-3) condition (B). Number of nuclei is used as an indicator of the number of fusion events during myogenesis. Fusion is therefore affected in the DM1 line in a Bru-3-dependent manner. Images of abdominal VL3 fibers representative of an altered condition [Mef>bru-3(37)] and control condition [UAS-bru-3(37)/+] (C). Nuclei were stained with anti-Lamin antibody (green). Actin was stained with phalloidin (gray). (D) Larval muscle growth was not affected in pathological contexts. The average number of sarcomeres along the VL3 fiber is represented for each genotype and represents an index of fiber growth. (E,F) Sarcomere shortening in DM1 condition is not Bru-3 dependent. Z-band profiles along VL3 fibers were assessed with phalloidin staining along a 100-µM length (E). The distance measured between two peaks gives sarcomere size. Thus, more peaks present on the profiles equates to more contracted muscle. *P<0.05, **P<0.01 versus both Mef>lacZ and respective transgenic control line. Sarcomere size, which reflects the state of contraction or relaxation of VL3 muscle, is presented on the graph for each mutant line (F). A significant reduction in the length of a sarcomere is an index of muscle hypercontraction. ***P<0.001.

Fig. 4.

Global gene expression analysis of bru-3 overexpression suggests a cytoplasmic function for Bru-3 in regulating sarcomeric transcript stability. (A) Venn diagrams of genes downregulated in Mef>bru-3(37) condition versus Mef>mblRNAi and DM1 (Mef>960CTG∩Mef>600CTG) lines show that ∼80% of transcriptomic alterations caused by Bru-3 overexpression are common to DM1 and/or mbl-attenuated lines. The diagram was generated from a list of transcripts that are >1.5-fold enriched or depleted relative to the Mef>lacZ reference. (B) Volcano plot summarizing microarray data for the Mef>bru-3(37) line versus Mef>lacZ. The horizontal axis plots the fold change on a log2 scale. The vertical axis plots the P-value on a −log10 scale. Gray dots indicate probes below the threshold. Green dots indicate probes with significantly altered expression. Black dots indicate downregulated sarcomere component probes. (C) Expression of transcripts encoding sarcomeric proteins is Bru-3 dependent. RT-qPCR on a set of mRNAs encoding sarcomeric proteins in normal condition (Mef>lacZ), pathological contexts [Mef>960CTG, Mef>bru-3(37)] and rescue condition [Mef>960CTG,Df(bru-3)]. *P<0.05, **P<0.01 versus Mef>lacZ or versus Mef>960CTG where indicated by a colored bar. ^#P<0.05, ^##P<0.01 indicate significant differences in data distribution between genotypes (Kruskal–Wallis test). (D) Sarcomeric transcript Mlc1 (red) colocalizes with Bru-3 (green) in cytoplasmic granules surrounding the DAPI-stained nuclei (blue) in WT condition. Scale bar: 20 µM. (E) qPCR of nascent Mlc-1 and mature Mlc1 transcripts indicates that despite a reduced level of mature transcripts, transcription is at the same rate between control (Mef>lacZ) and bru-3-overexpressing lines. *P<0.05 versus Mef>lacZ. (F) Sarcomeric Actn transcripts (red), similar to Mlc1 transcripts (D), colocalize with Bru-3 (green) in cytoplasmic granules. DAPI-stained nuclei are in blue. Scale bar: 60 µm.

Fig. 5.

The Bru-3 ortholog, CELF1, is relocalized in the cytoplasm of C2C12 myotubes and binds mRNA transcripts orthologous to Bru-3 targets. (A) Immunostaining against CELF1 (using 3B1 antibody) in undifferentiated C2C12 myoblasts and C2C12 myotubes after 10 days of differentiation. CELF1 signal (red) is essentially nuclear in myoblasts and accumulates in the cytoplasm upon differentiation. Nuclei are DAPI counterstained (blue). Scale bars: 20 µM. (B) Sarcomeric transcripts physically bound to CELF1 in myotubes. Strip charts show the enrichments of cross-linked mRNAs in immunopurified complexes (CLIP experiments). Enrichments on anti-CELF1 beads (gray boxes) are compared with those on IgG beads (white boxes). Student’s t-test P-values are shown above each lane of the graph. Jun mRNA (a known CELF1 target) was used as a positive control; Gapdh was used as a negative control. (C) 3′UTR of significantly-enriched sarcomeric transcripts in CLIP experiments. Destabilizing sites (Lee et al., 2010) (Table S3) are represented in red and random sites in white. Jun was used as a positive control; Gapdh was used as a negative control. *P<0.05, **P<0.01.

Fig. 6.

Modulating Bru-3 level influences Actn transcript and protein levels in an opposite manner. (A) Single-molecule FISH showing Actn transcripts in muscle cytoplasm (gray). Nuclei are visualized with DAPI (blue). Very low and lower Actn signals are detected in DM1₉₆₀ and bru-3-overexpressing lines, respectively, compared with WT. Actn mRNA is enriched in the cytoplasm of the knockdown line muscles. Scale bars: 10 µm. (B) Actn protein immunostaining in the different conditions. Scale bars: 20 µm. (C) Western blot analysis of Actn protein and protein quantification relative to alpha-tubulin (α-tub) immunoblot (n=3-4). (D) qPCR of nascent and mature Actn transcripts shows that the Actn transcription rate remains unchanged between control (Mef>lacZ) and bru-3-overexpressing lines. **P<0.01 versus Mef>lacZ. (E) Models of Bru-3 actions on Actn transcripts and on Actn protein levels in normal and pathological conditions.

Fig. 7.

Hypothetical model for dual Bru-3 role in the sarcoplasm. Sarcomeric transcripts such as Actn transcripts are exported from the nuclei to the sarcoplasm independently of Bru-3. Before being translated, they are stored in cytoplasmic granules/P-bodies. Cytoplasmic Bru-3 associates with granules and promotes release of sarcomeric mRNAs, which then undergo in situ translation in sarcomeres. Bru-3 positively regulates this in situ translation and at the same time leads to co-translational mRNA decay. The newly synthesized proteins, such as Actn protein, are immediately incorporated to the sarcomeres. Thus, according to this model, the increased levels of cytoplasmic Bru-3 detected in Bru-3-overexpressing and DM1 contexts would have double impact on sarcomeric components. Bru-3 associated with granules would favor the release of sarcomeric transcripts from these storage sites. In parallel, Bru-3 associated with sarcomeres would promote in situ translation of released transcripts and their quick subsequent decay, leading to the reduction of sarcomeric RNA levels. In such a context, proper turnover of sarcomeric components will be affected, or at least inefficient, and could contribute to muscle weakening observed in DM1.