Growth in sarcomere width and length is imbalanced in bru1^M3 mutant IFM

We used the bru1^M3 allele to perform a detailed analysis of IFM development from 48 h APF through 1 d adult in thorax hemi-section preparations. We reasoned that a phenotypic null mutant might produce a stronger phenotype than the bru1-IR knockdown used in previous studies [14,36]. bru1^M3 mutant myofibrils are severely degraded in 1 d adult, but are structurally intact until 90 h APF. However, already at 60 h APF mutant myofibrils are split and irregular, and neighboring myofibrils appear partially aligned (Fig 1A). From 48 to 90 h APF, sarcomeres in control w¹¹¹⁸ flies grow significantly in length, but sarcomeres in bru1^M3 do not (Fig 1A and 1D), indicating that growth in sarcomere length in bru1^M3 IFM is arrested after 48 h APF. By contrast, already at 60 h APF in bru1^M3, myofibrils are significantly wider than in wild type, and myofibril width in bru1^M3 IFM continues to increase significantly from 48 h to 90 h APF (Fig 1E). This suggests that although bru1^M3 sarcomeres do not grow in length after 48 h APF, they do grow in width.

To better understand the myofibril growth defects in bru1^M3 mutant IFM between 48 h to 90 h APF, we performed an ultrastructural time course analysis using transmission electron microscopy (TEM). Sarcomeres in bru1^M3 IFM were significantly shorter and thicker than w¹¹¹⁸ sarcomeres at 60 h, 72 h, and 90 h APF (Fig 1F, S1G and S1H Fig), confirming our observations from thorax hemi-sections (Fig 1A, 1D and 1E). Notably, already at 48 h APF, split myofibrils as well as discontinuous and misaligned z-discs were evident in bru1^M3 IFM (Fig 1F and 1G, S1I and S1J Fig), indicating that myofibril defects in bru1 mutants arise earlier than previously reported [14,36]. Myofibril splitting and z-disc defects were also detected in bru1^M3 at the 60 h, 72 h, and 90 h time points (Fig 1F and 1G, S1I and S1J Fig). Peculiarly, starting from 60 h APF, we could identify cytoplasmic components and even mitochondria trapped in the middle of otherwise continuous myofibrils (Fig 1F), suggestive of an underlying defect in radial growth or myofibril fusion. This data thus reveals 2 novel aspects of the bru1^M3 phenotype in IFM that we explore in more detail below: (1) during myofibril maturation in the absence of functional Bru1, there is an imbalance between growth in sarcomere length and width; and (2) defects in bru1^M3 mutant myofibril structure are already evident at 48 h APF.

Misregulation of gene expression and splicing lead to protein expression defects in bru1^M3 muscle

We next investigated the molecular phenotype underlying the myofibril defects observed in bru1^M3 IFM. We performed mRNA-Seq and whole proteome mass spectrometry on IFM dissected from 0- to 24-hour-old (1 d adult) wild-type w¹¹¹⁸ or bru1 mutant flies, to evaluate changes on both the RNA and protein levels (Fig 1H–1J and S1 Table). A differential expression analysis with DESeq2 revealed hundreds of significant changes in gene expression in the mRNA-Seq data (Fig 1H). Up-regulated genes were enriched for biological process gene ontology (GO) terms such as “muscle attachment,” “sarcomere organization,” “actin cytoskeletal organization,” “actin filament capping,” and “establishment of RNA localization” (Fig 1K). Down-regulated genes were in contrast enriched for terms such as “translation,” “cation transport,” and “oxidation-reduction process.” Using DEXSeq, we further detected hundreds of significant changes in exon use in bru1^M3 versus w¹¹¹⁸ IFM (Fig 1I), reflecting changes in alternative splicing as well as alternative promoter use. Notably, both up-regulated and down-regulated exons were enriched for GO terms such as “sarcomere organization,” “actin cytoskeleton organization,” “muscle contraction,” and “calcium-mediated signaling” (Fig 1K). This likely reflects isoform switches in structural genes, as for example sarcomere proteins (SPs) display both up- and down-regulated exons (Fig 1I), which we investigate in more detail below. Interestingly, on the gene level, SPs are mostly up-regulated in bru1^M3 versus w¹¹¹⁸ IFM, potentially reflecting transcriptional compensation in response to changes in isoform use (Fig 1H). We conclude that loss of Bru1 function leads to changes in both gene expression and alternative splicing.

We complimented our transcriptomic data with proteomics from 1 d adult IFM to evaluate if mRNA-level changes translate to altered protein expression. We grouped detected peptides into protein groups, such that peptides from the same gene that are unique to different protein isoforms form distinct protein groups. Analysis of the proteomics data revealed significant changes in the expression of hundreds of protein groups, with a bias toward down-regulation (Fig 1J). Down-regulated proteins were enriched in GO terms such as “muscle system process,” “muscle contraction,” “electron transport chain,” and “oxidation-reduction process,” while up-regulated proteins were enriched in “tissue development” and “intracellular receptor signaling process” (Fig 1K). As in the exon use analysis, we see up- and down-regulation of different sets of protein groups from SPs (Fig 1J). We then tested if there is a relationship between the changes in gene expression at the mRNA and protein level, and observed a weak but positive correlation for all significantly changed protein groups (Pearson’s R² = 0.34, Spearman’s R² = 0.41) (Fig 1L). Interestingly, we saw that both changes in gene expression as well as changes in exon use correlated with differential expression on the protein level (Fig 1L and S1K Fig). We also noted that different categories of genes show distinct patterns of regulation on the mRNA and protein level. Cytoskeletal genes such as SPs, genes involved in actin cytoskeleton organization, and microtubule-associated genes tend to be up-regulated at the gene level, but show both up- and down-regulation in the DEXSeq and proteomics data (S2A Fig). Mitochondrial and fibrillar core genes [14], by contrast, have up- and down-regulated exons, but are down-regulated on both the gene and protein level (S2A Fig). We conclude that Bru1-mediated changes to the IFM transcriptome are complex and indeed alter both protein and protein isoform expression.

Protein-level misexpression reflects alternative splicing changes in bru1^M3 IFM

Reasoning that the strong changes in structural gene expression at both the mRNA and protein level might contribute to the myofibril defects in bru1^M3 IFM, we analyzed sarcomere protein expression in greater detail. We noted that while 39% (9 of 23) of significantly differentially expressed (DE) sarcomere proteins were differentially regulated on the mRNA level (compared to 36% (184 of 514) of all significantly DE proteins), 74% (17 of 23) displayed significantly differential exon use (compared to 84 of 514 or 16% of all proteins) (Fig 2A and S2A Fig). We therefore compared changes in gene expression, exon use, and protein isoform expression in SPs (Fig 2B and S2 Table). We found a clear correspondence between changes in mRNA expression or splicing and altered protein expression level. For some SPs, for example Fhos, Ilk, or Mlp84B, gene expression changes match observed protein-level changes (Fig 2B). For other SPs, for example Mlp60A, wupA, Mhc, bt and Tm1, we see a clear switch in isoform expression, where changes in exon use match observed protein isoform changes (Fig 2B). This data also illustrates the breadth of sarcomere assembly processes impacted in bru1 mutant muscle, as misexpressed SPs include previously identified regulators of thin and thick filament growth, structural components of the z-disc and M-line, components of the integrin and cell adhesion machinery, and regulators of actomyosin interactions [68].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Misexpression of muscle-type-specific protein isoforms results in hypercontraction and abnormal contractile dynamics in bru1^-/- IFM.

(A) Venn diagram of the overlap between SPs differentially expressed at the gene level (purple), the protein level (green) or that have differentially used exons (orange). (B) Hierarchical clustering and heat map of the coordinated changes in SP gene expression (top, DESeq2 log₂FC), exon use (middle, top 10 DEXSeq log₂FC values), and protein expression (bottom, Perseus Difference of the top 3 protein groups). Black dot denotes p ≤ 0.05. (C) Summary heatmap of semi-quantitative RT-PCR results (S2 Fig) verifying muscle-type specificity of select alternative splice events in Strn-Mlck, wupA, Mhc, Zasp52, sls, and Tm1 in wild-type fly muscles, and loss or gain of those events in bru1^M3 IFM (not detected by PCR, white; detected by PCR, blue; weak band, light blue; strong band, dark blue). (D) Boxplot of gene, exon, and protein-level expression changes in genes with associated hypercontraction phenotypes (S1 and S2 Tables). Significantly DE proteins in bru1^M3 IFM are labeled; blue dot denotes p ≤ 0.05. (E–J) Confocal Z-stack images of DLM myofibers in w¹¹¹⁸, bru1^M3 and bru1^M3, Mhc [10] from 1 d and 3 d adults. Myofiber loss and hypercontraction in bru1^-/- IFM is alleviated in the Mhc [10] background (G, J). Thorax boundaries in (F), dashed line; phalloidin stained actin, gray; scale bar = 100 μm. (K) Quantification of myofiber tearing and detachment phenotypes from (E–J) at 90 h APF, 1 d and 3 d adults. N > 40 myofibers from at least 10 flies for each individual genotype and time point. (L) Snapshots from live movies of Talin-YPet labeled DLMs at 48 h and 72 h APF from w¹¹¹⁸ and bru1^M3 animals. Time 0.0 (magenta) is overlaid with time +0.65 s (green; at 48 h APF) or +16 s (green; at 72 h APF). A complete overlap (white) depicts no movement. Scale bar = 50 μm. (M) Quantification of distance of maximum myofiber extension at 48 h APF in control and bru1^M3. Boxplots are shown with Tukey whiskers, significance by unpaired t test (***, p < 0.001). (N) Quantification of spontaneous contraction events per fiber per 10 min in control and bru1^M3. At 72 h APF, bru1^M3 DLM fibers continue to undergo slow, unidirectional extension. N > 50 fibers/10 animals for each genotype and time point. Error bars = SEM. Underlying data can be found in S2 Table and Fig 2 Source Data files as listed in S6 Table. APF, after puparium formation; DE, differentially expressed; DLM, dorsal-longitudinal myofiber; IFM, indirect flight muscle; SP, sarcomere protein.

https://doi.org/10.1371/journal.pbio.3002575.g002

To independently confirm changes in alternative splicing and protein isoform expression in bru1^M3 IFM, we used GFP-tagged reporters and RT-PCR to test a panel of events in Strn-Mlck, wupA, Mhc, Zasp52, sls, and Tm1 (Fig 2C). We confirmed the loss of expression of Strn-Mlck isoform R (exons 23 and 25) at both the mRNA and protein level in bru1^M3 IFM (Fig 2C, S2C, S2F, S2G, S2G’, S2H and S2H’ Fig), which was previously shown to result in IFM hypercontraction [14]. In wupA, which encodes Troponin I [69], we confirmed an isoform switch from the IFM- to the tubular-specific termination (exon 4) at both the mRNA and protein levels (Fig 2C, S2D, S2I, S2J, S2J’, S2K and S2K’ Fig). We found that an alternative termination of Mhc that is used at early stages of IFM development and in tubular muscle (exon 37) [44,66] is expressed at increased levels in 1 d adult bru1^M3 IFM at both mRNA and protein levels (Fig 2C, S2E, S2L, S2M, S2M’, S2N and S2N’ Fig). Interestingly, this Mhc isoform is incorporated uniformly across the width of the myofibril in bru1^M3, instead of restricted to the center of the myofibrils as observed in control IFM (S2M, S2M’, S2N and S2N’ Fig). We further could confirm a switch in exon use in Zasp52 exon 14/15, sls exon 10 and Tm1 exon 28/29 and 30 (Fig 2C, S2O, S2P and S2Q Fig). This independent dataset validates our transcriptome and proteome analyses and confirms that Bru1 regulates alternative splicing and protein isoform expression of sarcomere proteins in IFM.

bru1^M3 mutant myofibers experience abnormal contractility throughout development

Several of the proteins we verified to have altered isoform expression in bru1^M3 mutants share a function in regulating actomyosin interactions, suggesting that normal contractile dynamics may be altered after loss of Bru1. Specifically, when we curated a list of genes with reported hypercontraction phenotypes in fly muscle, we noticed significant changes in alternative splicing and protein isoform expression in Mhc, Tm1, wupA, RyR, Strn-Mlck, and up (TnT) (Fig 2D, S2H’, S2K’ and S2N’ Fig). Misregulated actomyosin interactions can lead to hypercontraction, a condition characterized by aberrant contractility and short, thick sarcomeres [70–72]. In our time course analysis, we noticed a dramatic increase in the severity of the bru1^M3 phenotype at late stages of pupal development (Fig 1A, 1D and 1E). From 80 h APF to 1 d adult, w¹¹¹⁸ sarcomeres grow in length while sarcomeres in bru1^M3 shorten significantly and more than double in width (Fig 1D and 1E). Additionally, while IFM myofibers are attached at 90 h APF in bru1^M3 mutants, at 1 day 71% and by 3 days 97% of myofibers are torn or completely detached (Fig 2E, 2F, 2H, 2I and 2K). A hypercontraction phenotype can be rescued by minimizing actomyosin forces, and we confirmed that bru1^M3 myofibers remain attached in an Mhc¹⁰ mutant background in both 1- and 3-day adults (Fig 2G, 2J and 2K). This verifies that myofiber loss is indeed myosin-activity dependent, consistent with previous RNAi results [14,36]. We conclude that from 80 h APF, bru1^M3 myofibrils experience hypercontraction as characterized by progressive shortening and thickening of sarcomeres and eventual myofiber loss in adult flies.

Actomyosin-dependent tension also plays an important role during early stages of IFM development to organize and refine sarcomere structure [73,74]. Spontaneous contractions, or twitches, are evident at 34 h APF shortly after sarcomere assembly, reach peak intensity around 48 h APF, and are suppressed by 72 h APF as IFM myofibrils develop stretch-activation properties [42]. To determine if actomyosin contractility is disrupted in bru1^M3 mutants during IFM development, we performed live imaging of IFM labeled with Talin-YPet. At 48 h APF, we could detect twitch events in both bru1^M3 and control w¹¹¹⁸ myofibers (Fig 2L and 2N and S1 Movie). Strikingly, contractions in bru1^M3 myofibers were stronger than in w¹¹¹⁸ myofibers, resulting in a greater displacement of the myofiber tip (Fig 2M), and occurred more frequently (Fig 2N). At 72 h APF, we were unable to detect any movement in w¹¹¹⁸ myofibers (Fig 2L and 2N). Unexpectedly, in 17 of 89 bru1^M3 myofibers (19%), we observed a slow twitch movement (Fig 2L and 2N and S2 Movie). These fibers slowly contracted over a period of 16 s, as compared to less than 1 s at 48 h APF, and did not efficiently reextend. We conclude that actomyosin interactions are atypical throughout IFM development in bru1 mutants, with abnormal contractility first contributing to sarcomere organization defects and then resulting in hypercontraction in late pupa and adult flies.

Formation of hollow myofibrils is driven by misregulated radial growth and myofibril fusion

Another subset of SPs misexpressed in bru1^M3 IFM, including Fhos, sals, sls, Mlp84B, Unc-89, tmod, and Zasp52, are known regulators of sarcomere growth and integrity [68]. These genes caught our attention, as we observed splitting and an abnormal balance between growth in length and width in bru1^M3 sarcomeres (Fig 1A and 1F). We also noticed that not just SPs, but more broadly actin cytoskeleton and microtubule-associated genes, are misregulated in bru1^M3 (Fig 3A and S2A Fig). When we examined actin gene expression, we found altered expression ratios in our mRNA-Seq data that we could further confirm by RT-qPCR (Fig 3B and 3C, and S3A Fig), including a dramatic up-regulation of cardiac actin Act57B. Previous studies have shown that actin genes have differential abilities to integrate into the growing sarcomere and misexpression of cardiac Act57B disrupts IFM function [75,76]. We reasoned that the altered ratios in actin gene expression in bru1^M3 IFM, together with changes in key cytoskeletal regulators, might lead to aberrant sarcomere growth.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Hollow myofibril formation in bru1 mutants is an active process resulting from defective expression and splicing of cytoskeletal genes and aberrant sarcomere growth.

(A) Boxplot of gene, exon, and protein level expression changes between 1 d adult bru1^-/- and w¹¹¹⁸ IFM in genes from GO term “actin cytoskeleton.” Blue dot denotes p ≤ 0.05. (B) Heatmap of actin gene expression in bru1-IR, bru1^M3, leg and TDT as compared to wild-type IFM. (C) RT-qPCR verification of Actin gene expression levels between bru1^M3 and wild-type w¹¹¹⁸ IFM. (D) Confocal micrographs of DLM myofibril cross-sections in w¹¹¹⁸ and bru1^M3 at 48 h, 60 h, 72 h, 80 h, 90 h APF, and 1 d adult. Phalloidin stained actin, gray; scale bar = 5 μm. (E, F) Quantification of myofibril width (E) and density (F) in (D). bru1^M3 mutant myofibrils are significantly wider than in wild type at 48 h APF (0.48 ± 0.04 μm versus 0.37 ± 0.02 μm, ANOVA p = 0.001), and there are fewer myofibrils (0.76 ± 0.02 μm versus 1.4 ± 0.2 myofibrils per μm², ANOVA p < 0.001). Boxplots are shown with Tukey whiskers, outlier data points marked as black dots. Significance determined by ANOVA with post hoc Tukey (**, p < 0.01; ***, p < 0.001). (G) Quantification of myofibril structural morphology in (D). bru1^M3 myofibrils progressively form hook and ring structures starting from 60 h APF. N > 10 animals per genotype and time point. Error bars = SEM. (H–I”) Deconvoluted confocal images at 90 h APF of Fln-Gal4 driven UASp-GFP-Actin88F incorporation into control (H-H”) and bru1^M3 (I-I”) sarcomeres. Both xy- and zy-projections are shown. Fln-Gal4 expression from approximately 56 h APF results in a box-like pattern of GFP-Act88F incorporation (H’-H”) into growing wild-type sarcomeres, which is abnormal in bru1^M3 (I’-I”). GFP, green; phalloidin stained actin, magenta; scale bar = 5 μm. (J–K”) Deconvoluted confocal images at 90 h APF of Mhc-weeP26-GFP incorporation into control (J-J”) and bru1^M3 (K-K”) sarcomeres. Expression of GFP-labeled Mhc isoforms containing exon 37 is restricted to early developmental stages in IFM, resulting in a dot-like pattern flanking the M-line in wild-type sarcomeres (J’-J”) which is disrupted in bru1^M3 (K’-K”). weeP26-GFP, green; phalloidin stained actin, magenta; scale bar = 5 μm. (L) Quantification of myofibril number per fiber bundle in w¹¹¹⁸ and bru1^M3 at 48 h, 60 h, 72 h, 80 h, and 90 h APF. There is a significant reduction from 60 h to 80 h APF in bru1^M3 (45 ± 9 to 24 ± 5 myofibrils per bundle, ANOVA, p < 0.001) but not in w¹¹¹⁸ IFM (69 ± 15 to 68 ± 14 myofibrils per bundle, ANOVA, p = 0.30). Plot represents the mean ± SEM. Significance determined by ANOVA and post hoc Tukey (ns, not significant; **, p < 0.01). N > 8 animals per genotype and time point. Underlying data can be found in Fig 3 Source Data and the RNA-Seq data tables as listed in S6 Table. APF, after puparium formation; DLM, dorsal-longitudinal myofiber; GO, gene ontology; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g003

To evaluate defects in lateral or radial sarcomere growth in greater detail, we performed a cross-section time course analysis of bru1^M3 IFM development. In wild-type w¹¹¹⁸ IFM, myofibrils are organized into distinct bundles and grow uniformly in the radial direction from 48 h to 1 d adult, maintaining a consistent, circular appearance in cross-section (Fig 3D). Myofibrils in bru1^M3 flies are also organized into bundles, but already at 48 h APF, mutant myofibrils are significantly wider than in wild type (Fig 3D and 3E), and there are significantly fewer myofibrils (Fig 3F). Strikingly, although bru1^M3 myofibrils appear uniform and circular at 48 h APF, by 60 h APF there is variability among the sizes of individual myofibrils and some appear more oval or hook-like than circular (Fig 3D and 3G). As the myofibrils develop from 60 h to 90 h APF, this hooked phenotype becomes more pronounced, affecting 8.7 ± 2.8% of myofibrils at 60 h APF and 31% to 35% of myofibrils at 72 h, 80 h, and 90 h APF (Fig 3G). Concurrently, we see the progressive development of rings, or hollow myofibrils, affecting 12.6 ± 2.2% of myofibrils at 72 h APF and 33.9 ± 2.2% of all myofibrils at 90 h APF (Fig 3D and 3G). In 1 d adults, bru1^M3 myofibrils are highly irregular and display a strong atrophic phenotype. Taken together, this analysis reveals a distinct temporal progression in the appearance of myofibril defects, suggesting atypical and irregular radial growth during myofibril maturation. Moreover, such growth could provide a mechanism whereby cytoplasmic components or even organelles become trapped in the middle of myofibrils, as we observed in our TEM data (Fig 1F).

To test if sarcomere growth is indeed aberrant in bru1^M3 IFM, we adapted an approach to visualize thin filament growth dynamics using temporally restricted expression of a GFP-tagged actin [65]. Flight muscles predominantly express Act88F, which is incorporated into thin filaments both laterally and radially as sarcomeres grow [65,77]. We expressed UAS-GFP-Act88F starting around 56 h APF using Flightin-Gal4 (Fln-G4), which allowed us to monitor actin incorporation over the critical time period for hollow myofibril formation. In control sarcomeres from 90 h APF flies, we observed a box-like labeling pattern, reflecting GFP-Act88F integration both laterally as well as at the plus and minus ends of the thin filament (Fig 3H, 3H’ and 3H”, and S3B–S3B”‘ Fig and S3 Movie). In bru1^M3 mutants, this pattern is significantly altered. GFP-Act88F is incorporated weakly at the z-disc and strongly and uniformly across the M-line (Fig 3I, 3I’ and 3I”, and S3C–S3C”‘ Fig and S4 Movie). This is striking, as sarcomeres do not grow appreciably in length in bru1^M3 mutants, thus indicating a defect either in thin filament stability or capping protein function. Moreover, the strong lateral incorporation of GFP-Act88F coupled with exclusion from the central core of the sarcomere observed in control flies is not evident in bru1^M3 (Fig 3H” and 3I”), even though mutant sarcomeres exhibit excessive radial growth from 60 h to 90 h APF (Fig 3D and 3E). We interpret this to reflect loose packing of the thin filament lattice allowing actin integration across the sarcomere, as well as radial integration of other unlabeled actin isoforms. These data reveal a clear defect in thin filament growth dynamics, and identify separable phenotypes in processes controlling growth in sarcomere length and width.

We next evaluated the localization of Mhc^Wee-P26-GFP, to test if thick filament growth is also disrupted in bru1^M3 myofibrils. In control sarcomeres, Mhc^Wee-P26-GFP is localized in 2 central dots flanking the M-line (Fig 3J, 3J’ and 3J”, S3D and S3E–S3E”” Fig and S5 Movie), because a developmental switch in Mhc isoform expression results in incorporation of an unlabeled Mhc isoform after 48 h APF [66]. This isoform switch is partially impaired in bru1^M3 IFM, so Mhc^Wee-P26-GFP is continuously expressed [44], resulting in GFP incorporation across the thick filament (Fig 3K, 3K’ and 3K”, and S3F Fig). Interestingly, Mhc^Wee-P26-GFP can be seen to form an irregular and asymmetric hook-like pattern at the M-line in bru1^M3 sarcomeres (Fig 3K”, S3F and S3G–S3G”” Fig and S6 Movie), indicative of M-line misalignment or instability. We also observed myofibrils with 2 or more seemingly distinct Mhc^Wee-P26-GFP dots (Fig 3K” and S6 Movie), suggesting that large, irregular myofibrils may result from myofibril fusion. Logically, neighboring myofibrils undergoing abnormal radial growth might grow into one another and fuse, contributing to the formation of hollow myofibril structures. To test this, we calculated the number of myofibrils per bundle in our cross-section data (Fig 3D) and found a significant reduction from 60 h to 80 h APF in bru1^M3 but not in w¹¹¹⁸ IFM (Fig 3L). Interestingly, this partial fusion of myofibrils does not reflect a switch in the fiber-type fate of the IFM, as the majority of tubular genes are not misexpressed in bru1^M3 IFM (S3H Fig). Similarly, while there is misuse of tubular exons in bru1^M3 IFM (S2D, S2E and S2O–S2Q Fig), many tubular exons are not strongly affected (S3H Fig), indicating that loss of Bru1 does not result in a complete switch in fiber-type-specific splicing, but rather loss of IFM-specific splice events. Taken together, our data support a mechanism whereby altered expression ratios of actins and cytoskeletal regulators impairs sarcomere growth in length, but concurrently promotes aberrant radial growth leading to myofibril fusion and the formation of hollow myofibrils during myofibril maturation in bru1^M3 IFM.

Developmental up-regulation of fibrillar genes is impaired in bru1-IR IFM

While investigating fiber-type-specific exon misregulation, we noticed that genes and exons that are normally expressed preferentially in mature IFM are down-regulated in bru1^M3 IFM (S3I Fig). As CELF1 is known to promote embryonic splice events in vertebrate muscle [78], we wondered if developmentally regulated transitions in gene expression and splicing are also disrupted in IFM lacking Bru1. To explore temporal-dependent dynamics in gene expression and exon use, we analyzed an mRNA-Seq time course in control and bru1-IR muscle at 24 h, 30 h, and 72 h APF and 1 d adult. We confirmed via qPCR that bru1-IR (Mef2-Gal4 driven RNAi against bru1) results in a significant, 56-fold decrease in bru1 mRNA levels in IFM (S7A Fig). We noted that at all 4 time points, we could detect significant changes in gene expression and exon use in bru1-IR IFM (Fig 4A and S4A Fig and S3 Table). There was a dramatic increase in the number and magnitude of changes at both the gene and exon level in bru1-IR at 72 h APF and 1 d adult (Fig 4A and S4A Fig), which mirrors the increase in severity of cellular phenotypes in bru1^M3 IFM from 80 h APF to 1 d adult (Figs 1A, 1F, 3D and 3G). These results show that the number and severity of gene and exon regulatory defects progressively increases in bru1-IR IFM as development proceeds.

Drosophila IFM has been shown to undergo a developmental transition in gene expression that is necessary to establish the stretch-activation mechanism characteristic of mature flight muscle [42]. We next evaluated if this transition is disrupted in bru1-IR IFM. We evaluated specific subsets of genes previously shown to change across IFM development, including SPs, fibrillar muscle genes, and actin cytoskeleton genes. While we saw down-regulation of individual SPs and cytoskeleton genes, almost all core fibrillar muscle genes are strongly down-regulated in 1 d adult bru1-IR IFM (Fig 4B, S4B and S4C Fig). This is also reflected in GO term enrichments across the time course. On the gene expression level, at pupal time points we found enrichment for terms such as “negative regulation of translation,” “cell surface receptor signaling pathway,” and “behavior” at 24 h APF, “regulation of myofibril size” at 30 h APF, and “regulation of muscle system development” and “translation” at 72 h APF (Fig 4C). In 1 d adult bru1-IR IFM, we saw enrichment for cytoskeletal and muscle terms such as “sarcomere organization,” “actin cytoskeleton organization,” and “actomyosin structure organization,” as well as terms related to metabolism and translation such as “oxidation-reduction process,” “cytoplasmic translation,” and “metabolic process.” To test genome-wide if this switch in gene expression during IFM maturation is disrupted, we identified all genes in our mRNA-Seq time course that are significantly regulated between 24 h APF and 1 d adult in control IFM (adjusted p-value <0.05), and plotted their temporal change in expression in bru1-IR IFM (Fig 4E). We found that developmental changes in gene expression in bru1-IR IFM are correlated with those observed in control IFM (correlation value = 0.9107848), suggesting that bru1-IR IFM does undergo a temporal switch in gene expression. However, the magnitude of up- or down-regulation is often reduced in bru1-IR IFM, and a subset of genes shows regulation in the opposite direction between 24 h APF to 1 d adult (Fig 4E and S4A Fig), suggesting the switch is not as clean as in control IFM. We verified this switch in expression of Mlp60A and Mlp84B, which show strong up-regulation at the gene level in bru1-IR IFM from 72 h APF to 1 d adult (S5A Fig), using GFP fosmid reporters (S5F and S5G Fig). We conclude that loss of Bru1 selectively impairs up-regulation of fibrillar muscle genes, but not the entire maturation program during IFM development.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Knockdown of bru1 disrupts temporal dynamics of gene expression and alternative splicing necessary for maturation of flight muscle.

(A) Plot of the number of genes (magenta) and exons (green) significantly differentially expressed (p ≤ 0.05, abs(log₂FC) ≥ 2) in bru1-IR versus control IFM at 24 h, 30 h, 72 h APF, and in 1 d adult. (B) Boxplot of changes in gene expression (DESeq2) and exon use (DEXSeq) in sarcomere proteins and fibrillar muscle genes across the bru1-IR IFM time course. Blue dot denotes p ≤ 0.05. (C) Heatmap of select Biological Process GO term enrichments in significantly regulated genes (DESeq2, p-adj ≤ 0.05) and exons (DEXSeq, p-val ≤ 0.05) in bru1-IR IFM at 24 h, 30 h, 72 h APF, or in 1 d adult, or at all 4 time points. (D) Heatmap (left) of all exons significantly DE (DEXSeq, p-val ≤ 0.05) at all time points in bru1-IR versus control IFM. The fifth column shows the temporal change in use of the same exons in w¹¹¹⁸ IFM from 24 h APF to 1 d adult. Exons are identified as belonging to ribosomal subunit, microtubule associated, contractile fiber, sarcomere proteins, or myosin complex gene categories (right, black boxes). (E) Left: Plot of the log₂FC values of all genes differentially expressed (DESeq2, p-adj ≤ 0.05) in w¹¹¹⁸ IFM (black dots) from 24 h APF to 1 d adult, ordered by control log₂FC, and the corresponding change in the same genes in bru1-IR IFM (orange dots). Right: Violin plots comparing control (gray) and bru1-IR (orange) expression of strongly up-regulated (log₂FC ≥ 2) and down-regulated (log₂FC ≤ -2) temporal-switch genes. (F) Left: Plot of the log₂FC value of all exons differentially expressed (DEXSeq, p-val ≤ 0.05) in w¹¹¹⁸ IFM (black dots) from 24 h APF to 1 d adult, and the corresponding change in those same exons in bru1-IR IFM (orange dots). Right: Violin plots comparing exon use of strongly up-regulated (log₂FC ≥ 2) and down-regulated (log₂FC ≤ -2) temporal-switch exons in control (gray) and bru1-IR (orange) IFM. Underlying data can be found in S3 Table and the RNA-Seq data tables as listed in S6 Table. APF, after puparium formation; GO, gene ontology; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g004

A developmental switch to mature splice isoforms is blocked in IFM lacking Bru1

Based on the developmental transition in gene expression, we hypothesized that a similar transition in alternative splicing exists in wild-type IFM. We next investigated if such a splicing transition exists, and if it is disrupted in IFM lacking Bru1. The Biological Process GO terms enriched in the hundreds of DE exons at all 4 time points in our mRNA-Seq time course (Fig 4A) included “actin cytoskeleton organization,” “actin filament-based process,” “actomyosin structure organization,” “sarcomere organization,” and “cytoplasmic translation” (Fig 4C). We therefore started by looking at exon use dynamics in core fibrillar muscle genes and SPs, as several mature, IFM-specific protein isoforms have been reported in these categories [14,66,79,80]. In total, at 24 h, 30 h, or 72 h APF or at 1 d adult, we saw significant changes in 222 exons from 73 core fibrillar muscle genes and 413 exons from 56 SPs in bru1-IR IFM (S4D Fig). Strikingly, exons that were up-regulated in bru1-IR IFM were down-regulated in control IFM between 24 h APF to 1 d adult, while exons that were down-regulated in bru1-IR IFM were up-regulated in control IFM between 24 h APF to 1 d adult.

To expand beyond SPs and fibrillar genes, we next identified a set of 91 exons that are significantly misregulated in at least 3 of 4 time points in bru1-IR IFM. These exons became more strongly misregulated as development proceeds and belonged to genes encoding ribosomal subunits, microtubule-associated genes, SPs, contractile fiber, and myosin complex genes (Fig 4D). When we looked at the temporal change in use of these exons in control muscle from 24 h APF to 1 d adult, we found that all of the exons up-regulated in bru1-IR IFM are normally down-regulated in 1 d adult muscle. Likewise, the exons down-regulated in bru1-IR IFM are normally up-regulated between 24 h APF and 1 d adult in control IFM (Fig 4D). We then identified all exons in our mRNA-Seq time course that are significantly regulated between 24 h APF and 1 d adult in control IFM (Fig 4F). We observed a clear temporal switch in exon use, reflecting mainly alternative splice events and alternative 3’ UTRs, but also alternative promoter use (S3 Table). Strikingly, when we plotted the temporal change in use of these exons in bru1-IR IFM, we observed a reduction in coordinated developmental regulation. We confirmed these mRNA-level changes at the protein level for wupA, Kettin, and Clip190 using GFP-tagged reporters under the control of endogenous regulatory elements (S5A and S5C–S5E Fig). This analysis reveals that loss of Bru1 results in a block in the temporal shift in exon use during IFM development, including increased expression of tubular-preferential isoforms. Thus, the molecular defect that underlies the myofibril growth and hypercontraction defects observed during myofibril maturation in bru1-IR and bru1^M3 IFM is a failure to transition to expression of mature, muscle-type-specific splice isoforms.

Loss of Bru1 leads to cytoskeletal organization defects at early stages of myogenesis

Although published Bru1 phenotypes in IFM are only reported after 48 h APF [14,36], our data above show that gene expression and splicing of structural genes are already abnormal in bru1-IR IFM from 24 h APF, prior to myofibril formation, and that abnormal sarcomere structure and contractility are already evident in bru1^M3 myofibers at 48 h APF. We therefore investigated the possibility that Bru1 also has a function in early IFM development. We extended our cross-section time course to include time points at 24 h, 26 h, 28 h, 30 h, 32 h, and 34 h APF, a time window during which the developing IFM myofibers undergo dramatic cytoskeletal reorganization in preparation for myofibril formation from 30 to 34 h APF [81,82]. Already at 24 h APF, we detected a defect in actin cytoskeleton organization in bru1^M3 myofibers (Fig 5A). In control w¹¹¹⁸ IFM at 24 h APF, F-actin is tightly organized at the sarcolemma, while in bru1^M3 myofibers it forms a more diffuse meshwork throughout the sarcoplasm that fails to condense as tightly as in wild type (Fig 5B and 5C). By 26 h APF in w¹¹¹⁸ myofibers, F-actin organizes into uniformly distributed cables at the sarcolemma, which then subdivide and migrate toward the center of the myofiber from 28 to 30 h APF. In bru1^M3 myofibers this process is abnormal, with F-actin organizing into many smaller, unequally distributed cables with impaired migration (Fig 5C and 5D). Both w¹¹¹⁸ and bru1^M3 myofibers undergo myofibrillogenesis from 30 to 34 h APF, but there are fewer myofibrils present in bru1^M3 myofibers by 34 h APF and those myofibrils are significantly larger in diameter than w¹¹¹⁸ myofibrils (Fig 5C, 5E and 5F). Progressive organization of the actin cytoskeleton during IFM myogenesis leads to myofiber compaction in addition to generating the tension necessary to drive myofibrillogenesis [63,73,74]. At both 24 h and 30 h APF, myofibers are longer in bru1^M3 than in w¹¹¹⁸ (Fig 5G and 5H), demonstrating that IFM compaction is mildly impaired and consistent with a cytoskeletal defect in myofibers lacking Bru1. We conclude that Bru1 is necessary during early IFM development to regulate cytoskeletal organization dynamics that support proper myofibril formation.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Early cytoskeletal rearrangement and myofiber compaction are abnormal in bru1 mutant IFM.

(A) Confocal images of DLM cross-sections at 24 h APF. Phalloidin stained F-actin (gray) reveals progressive condensation of the actin network into cables (arrow head) near the sarcolemma in w¹¹¹⁸, but a meshwork (arrow) in bru1^M3. Scale bar = 10 μm. (B) Quantification of actin network structure in (A). N > 10 for each genotype. (C) Cross-section time-course of early cytoskeletal rearrangements and pre-myofibril formation in w¹¹¹⁸ and bru1^M3 at 24 h, 26 h, 28 h, 30 h, 32 h, and 34 h APF. Scale bar = 5 μm. Irregular actin condensation at 24 h and cable splitting at 26 h and 28 h APF is evident in bru1^M3, prior to pre-myofibril formation at 30–32 h APF. (D) Quantification of actin cable number per myotube in w¹¹¹⁸ and bru1^M3 at 26 h and 28 h APF. Significance determined by ANOVA and post hoc Tukey (ns, not significant; ***, p-val < 0.001). (E, F) Quantification of myofibril width (E) and density (F) in w¹¹¹⁸ and bru1^M3 at 30 h, 32 h, and 34 h APF. The first sarcomere growth defects in bru1^M3 are already detected at 34 h APF, when the pre-myofibrils are fully formed. Significance determined by ANOVA and post hoc Tukey (ns, not significant; **, p-val < 0.01; ***, p-val < 0.001). All boxplots are shown with Tukey whiskers, outlier data points marked as black dots. (G) Confocal Z-stack images of DLMs in w¹¹¹⁸ and bru1^M3 at 24 h, 30 h, 40 h, and 48 h APF. DLM fibers undergo compaction at 30 h APF, followed by re-extension and fiber growth. Phalloidin stained actin, gray; scale bar at 24, 30 h APF = 50 μm, at 40, 48 h APF = 100 μm. (H) Quantification of the DLM fiber length in (G). bru1^M3 DLM fibers fail to fully compact. Significance determined by ANOVA and post hoc Tukey (ns, not significant; *, p-val < 0.05; ***, p-val < 0.001). Underlying data can be found in Fig 5 Source Data as listed in S6 Table. APF, after puparium formation; DLM, dorsal-longitudinal myofiber; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g005

Temporally restricted RNAi reveals a requirement for Bruno1 in early and late myogensis

The data we present above suggest that Bru1 is required both during early and late stages of IFM myogenesis. To test the temporal requirement of Bru1 in muscle development, we performed RNAi with a panel of 5 previously characterized Gal4 enhancer lines with distinct temporal expression patterns in IFM (Fig 6A, S6A and S7 Figs). In contrast to Mef2-Gal4, which is expressed in all muscle throughout development, salm-Gal4, Act88F-Gal4, UH3-Gal4, and Fln-Gal4 are expressed in IFM from approximately 16 h APF, 24 h APF, 34 h APF, and 56 h APF, respectively [77,80,83,84]. Him-Gal4 is expressed in myoblasts and is down-regulated in IFM by 30 h APF [42]. bru1-IR led to a strong reduction in bru1 mRNA with all drivers tested (S8A Fig), but did not impair adult viability or eclosion (S6B Fig). We then compared phenotypes of bru1-IR driven by Gal4 lines active during early myogenesis (bru1-IR^Him, bru1-IR^salm, bru1-IR^Act88F) to those active after myofibrillogenesis is completed (bru1-IR^UH3 and bru1-IR^Fln).

Knockdown of bru1 during early myogenesis with Him-Gal4, salm-Gal4, or Act88F-Gal4 resulted in severe behavioral and structural defects. We focus below on bru1-IR^Him as representative of bru1 RNAi knockdown phenotypes during early myogenesis. bru1-IR^Him flies were flightless (Fig 6B), and the percent of detached IFM myofibers increased between 90 h APF and 1 d adult, reflecting a progressive hypercontraction phenotype (Fig 6C and 6D), and 1 d adult bru1-IR^Him myofibrils were torn and degenerate, and at both 90 h APF and 1 d adult sarcomeres were significantly shorter and wider than control sarcomeres (Fig 6D–6F). Him-Gal4 driven expression of UAS-GFP or UAS-GFP-RNAi did not affect sarcomere length or width (S7A–S7O Fig), confirming this phenotype is specific to knockdown of bru1. Interestingly, although Him-Gal4 turns off at 30 h APF, while salm-Gal4 and Act88F-Gal4 are expressed continuously starting from 16 h and 24 h APF, respectively, they all produce a similar bru1-IR phenotype characterized by short, thick sarcomeres and progressive hypercontraction mediated loss of adult IFM myofibers (Fig 6B–6F and S6C–S6G Fig). We confirmed that bru1-IR^Him results in a loss of Bru1 protein at 24 h and 30 h APF, indicating that RNAi is effective at early pupal stages, and from 48 h APF Bru1 protein is again detectable and presumably functional in the nuclei of bru1-IR^Him flies (S8M–S8Q Fig). This data shows that decreased expression of Bru1 during early stages of IFM development is sufficient to produce severe myofibril and sarcomere phenotypes.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Temporally restricted RNAi demonstrates a functional requirement for Bru1 during early myogenesis.

(A) Scheme of Him-Gal4 and UH3-Gal4 expression timing during IFM myogenesis. Gradient color of the bar indicates the strength of expression. Time points as marked. (B) Quantification of flight ability in Him-Gal4 driven bru1-IR (bru1-IR^Him) (N > 130). (C) Quantification of myofiber ripping and detachment phenotypes in control and bru1-IR^Him at 90 h APF and 1 d adult (N > 40). (D) Confocal projections of control and bru1-IR^Him hemithoraxes (upper, scale bar = 100 μm) and single-plane images of myofibrils (lower, scale bar = 5 μm) in control (top) and bru1-IR^Him (bottom) at 48 h and 90 h APF and 1 d adult. (E, F) Quantification of sarcomere length (E) and myofibril width (F) from (D). Boxplots are shown with Tukey whiskers, outlier data points marked as black dots. Significance determined by ANOVA and post hoc Tukey (ns, not significant; ***, p-val < 0.001). (G) Quantification of flight ability in UH3-Gal4 driven bru1-IR (bru1-IR^UH3) (N > 110). (H) Quantification of myofiber integrity in control and bru1-IR^UH3 at 90 h APF and 1 d adult (N > 40). (I) Confocal projections of hemithoraxes (upper, scale bar = 100 μm) and single-plane images of myofibrils (lower, scale bar = 5 μm) in control (top) and bru1-IR^UH3 (bottom) at 48 h and 90 h APF and 1 d adult. The severity of the bru1-RNAi associated myofibril phenotype is comparable between (D) and (I). (J, K) Quantification of sarcomere length (J) and myofibril width (K) from (I). Significance determined by ANOVA and post hoc Tukey (ns, not significant; ***, p-val < 0.001). (L) Single-plane cross-section images of DLM myofibrils in control w¹¹¹⁸, bru1-IR^Him, and bru1-IR^UH3 at 48 h, 72 h, and 90 h APF. Scale bar = 5 μm. (M) Quantification of myofibril width (top) and density (bottom) in (L). Significance determined by ANOVA and post hoc Tukey (ns, not significant; ***, p-val < 0.001). (N) Quantification of myofibril structure from (L), showing the ratio of normal dot morphology (white) to abnormal hook (cyan) and ring (blue) structures. Fewer hooks and rings form in early expressed bru1-IR^Him as compared to bru1-IR^UH3. Error bars = SEM. (O) Quantification of mean myofibril number per bundle in (L). bru1-IR^UH3 IFM forms the correct number of myofibrils while bru1-IR^Him IFM does not, but myofibril fusion is more extensive after 48 h APF in bru1-IR^UH3. bru1-IR^UH3 IFM have a near wild-type number of 62 ± 12 myofibrils per bundle at 48 h APF, but by 72 h APF the number of 26 ± 6 myofibrils per bundle is less than the 32 ± 7 observed with bru1-IR^Him. Error bars = SEM. F-actin in (D, I, L) stained with phalloidin. Underlying data can be found in Fig 6 Source Data as listed in S6 Table. APF, after puparium formation; DLM, dorsal-longitudinal myofiber; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g006

We then examined bru1 knockdown phenotypes with the UH3-Gal4 and Fln-Gal4 drivers that are expressed after myofibril formation. While bru1-IR^UH3 produces flightless flies with ripped and detached IFM myofibers (Fig 6G–6I), bru1-IR^Fln flies are able to fly and have intact IFM (S6C, S6D, S6J-J’, and S6P-P’). Although bru1 mRNA levels are significantly decreased by both bru1-IR^UH3 and bru1-IR^Fln (S8A Fig), Bru1 protein is still detected in bru1-IR^Fln myofibers at 90 h APF, but not in bru1-IR^UH3 myofibers (S8D–S8H Fig). This can be explained if Bru1 protein perdures for some time after RNAi is induced, or if Bru1 protein turnover decreases at later pupal stages. We therefore focused our subsequent analysis on bru1-IR^UH3. Like we observed in bru1-IR^Him as well as in bru1^M3 IFM, sarcomeres in bru1-IR^UH3 IFM are shorter and thicker than control sarcomeres at 90 h APF and 1 d adult, and myofibrils are torn and display a progressive loss of sarcomere architecture (Fig 6I, 6J and 6K). However, when examined in cross-section, bru1-IR^UH3 produces stronger radial growth defects than bru1-IR^Him (Fig 6L). bru1-IR^UH3 myofibrils are significantly wider in diameter than control or bru1-IR^Him myofibrils at both 72 h and 90 h APF (Fig 6M), and proportionally more myofibrils appear as rings instead of hooks or dots already at 72 h APF (Fig 6N). In addition, although bru1-IR^UH3 IFM have a near wild-type number of myofibrils per bundle at 48 h APF, a large drop in myofibril number is observed between 48 h and 72 h APF (Fig 6O). This suggests that the recovering level of Bru1 protein after 48 h APF in bru1-IR^Him promotes a somewhat less severe radial overgrowth phenotype than the absence of Bru1 protein from 48 h APF onwards in bru1-IR^UH3. These data support a requirement for Bru1 during myofibril maturation after 48 h APF to regulate sarcomere growth dynamics.

Levels of Bru1 expression during early myogenesis impact myofibril formation and growth

Our data indicate that phenotypes related to Bru1 function in early and late myogenesis may be separable, where early function defines the number and organization of myofibrils while later function regulates sarcomere radial and lateral growth. To test this hypothesis, we performed temporally regulated rescue experiments, as continuous Bru1 overexpression with Mef2-Gal4 causes severe phenotypes [14,36]. We generated a UAS-Bru1 (isoform A) rescue construct (S9A Fig), and then optimized expression conditions with our Gal4 panel (S9B–S9P Fig). Overexpression with Him-Gal4 and Fln-Gal4, the 2 Gal4 drivers with the narrowest expression windows, produced surviving flies with intact IFM (S9B–S9D Fig) and enabled us to evaluate Bru1 function in the early 0 to 30 h APF and late 56 h APF to adult time windows, respectively.

To evaluate Bru1 function during early IFM development, we performed overexpression and rescue experiments with Him-Gal4. We first confirmed that in a bru1^M3 background, Him-Gal4 drives expression of Bru1 at 24 h APF, but not at 48 h or 90 h APF (S10A and S10B Fig), demonstrating the temporal-restriction of Bru1 expression to early myogenesis. Overexpression of Bru1 in a wild-type background is sufficient to produce short, thick sarcomeres and loss of flight ability, but IFM myofibers remain attached (Fig 7A–7F and 7M–7P, and S10C–S10F Fig). This demonstrates that overexpression of Bru1 produces the same sarcomere phenotype as loss-of-function, indicating that Bru1 is dosage sensitive. In cross-sections, overexpression of Bru1 with Him-Gal4 leads to a reduced myofibril density, a decreased number of myofibrils per bundle and myofibrils of variable thickness at 72 h and 90 h APF, but is not sufficient to produce hollow myofibrils (Fig 7Q, 7R, 7V and 7W, and S10I Fig). Unexpectedly, IFM overexpressing Bru1 have a prominent hole in the middle of the myofiber at 48 h APF (Fig 7R–7R”), which we interpret to reflect failed migration of actin cables to the center of the fiber from 28 to 30 h APF (Fig 5C). These data demonstrate that the dosage of Bru1 before 30 h APF is crucial to regulate the number and regularity of myofibrils and can further influence sarcomere growth during later developmental stages.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Expression of Bru1 restricted to early myogenesis fails to rescue and exacerbates bru1^M3 myofibril phenotypes.

(A–L”) Confocal projections of hemithorax (scale bar = 100 μm) and single-plane images of myofibril structure (scale bar = 5 μm) in control (A–C”); Bru1 overexpression (Him-Gal4 > UAS-Bru1, D–F”); early temporal rescue (Him-Gal4 > UAS-Bru1, bru1^M3/M3, G–I”); and mutant bru1^M3 (J–L”). Time points at 48 h, 90 h APF and 1 d adult as labeled. Phalloidin stained actin, gray. (M, N) Quantification of flight ability (M) and DLM fiber integrity (N) in (B, E, H, K). Genotypes are denoted by symbols: wild type, white square; bru1^M3, red square; Him-Gal4, tan triangle; UAS-Bru1, magenta diamond. N > 40 fibers for each genotype. (O, P) Quantification of sarcomere length (O) and myofibril width (P) in (C, F, I, L). Boxplots are shown with Tukey whiskers, outlier data points marked as black dots. Significance determined by ANOVA and post hoc Tukey (ns, not significant; *, p-val < 0.05; **, p-val < 0.01; ***, p-val < 0.001). (Q–T”) Single-plane cross-section images of myofibril structure at 48 h, 72 h, and 90 h APF in control (Q–Q”), Bru1 overexpression (R–R”), early temporal rescue (S–S”), and mutant bru1^M3 IFM (T–T”). Magnified image of selected area (white rectangle) shown in lower right corner. Overexpression of Bru1 with Him-Gal4 produces holes in the center of IFM myofibers (yellow asterisks) and cannot rescue later formation of hollow myofibrils; DAPI, green; phalloidin stained actin, magenta; scale bar = 10 μm (Q–T, Q’–T’), 20 μm (Q”–T”) and 5 μm (zoom-in sections). (U, V) Quantification of myofibril density (U) and myofibril structure (V) in (Q–T”). Genotypes, boxplot, and significance as above. Error bars in V = SEM. (W) Quantification of mean myofibril number per fiber bundle in (Q–T”). Him-Gal4 driven Bru1 can partially rescue the number of myofibrils formed in bru1^M3 myofibers before 48 h APF, but cannot rescue myofibril fusion and hollow myofibril formation after 48 h APF. Error bars = SEM. Underlying data can be found in Fig 7 Source Data as listed in S6 Table. APF, after puparium formation; DLM, dorsal-longitudinal myofiber; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g007

We then investigated if Him-Gal4 mediated expression of Bru1 is sufficient to rescue the bru1^M3 phenotype. Instead of rescuing, expression of Bru1 during early developmental stages increased the severity of the bru1^M3 phenotype. Rescue flies are flightless and a show greater proportion of flies with ripped or detached IFM at both 90 h APF and 1 d adult than bru1^M3 (Fig 7G–7N, S10C and S10D Fig). Sarcomeres in rescue IFM are significantly shorter than both control and bru1^M3 sarcomeres at 90 h APF (Fig 7O), are as wide as mutant sarcomeres (Fig 7P), and show myofibril tearing and loss of sarcomere architecture in 1 d adults (Fig 7G–7I”). In cross-section, Him-Gal4 > Bru1 rescue flies form hollow myofibril structures that are larger and more irregular than those formed by bru1^M3 myofibrils at 90 h APF (Fig 7S–7T”). Interestingly, while the myofibril density and developmental progression of hollow myofibril formation mirrors that observed in bru1^M3 (Fig 7U and 7V), rescue flies have more myofibrils per bundle at 48 h APF than bru1^M3 (Fig 7W), indicating that Bru1 expression may partially rescue some early myofibril defects. Rescue flies also have a hole in the middle of the IFM myofiber that is evident at 48 h, 72 h, and 90 h APF (Fig 7S–7S”). We confirmed the presence of this hole using hematoxylin and eosin staining, as well as the absence of such a hole in wild-type or bru1^M3 myofibers (S10G and S10H Fig). Taken together, these data demonstrate: (1) that Bru1 has a function during early stages of IFM myogenesis; (2) myofibril phenotypes are sensitive to Bru1 expression level before 48 h APF; (3) loss and gain of Bru1 can produce similar phenotypes; and (4) misregulation of Bru1 in early development also impacts later stages of myofibril maturation.

Rescue of Bru1 during late development is insufficient to rescue myofibril structural defects

To evaluate Bru1 function during late IFM development, we performed overexpression and rescue experiments with Fln-Gal4. We first verified by antibody staining that at 90 h APF in a bru1^M3 background, Fln-Gal4 drives expression of UAS-Bru1 (S1A and S1B Fig). Overexpression of Bru1 with Fln-Gal4 in a wild-type background was sufficient to produce flightless flies, but myofibers remained attached and did not display hypercontraction-related tearing (Fig 8A–8F’ and S11C Fig). Although myofibril and sarcomere structure were largely intact, sarcomeres were significantly shorter and thinner than in the control and 52.4% were frayed in 1 d adults (Fig 8A–8F’ and 8O–8Q, and S11C–S11E Fig). We conclude that overexpression of Bru1 during late IFM development with Fln-Gal4 is sufficient to produce flightlessness due to mild sarcomere defects, but not strong enough to reproduce the bru1^M3 phenotype.

We then evaluated if expression of Bru1 from 56 h APF with Fln-Gal4 can rescue the bru1^M3 phenotype. Strikingly, whereas 71.0% of myofibers are ripped or detached in 1 d adult bru1^M3 flies, only 27.4% are ripped or detached in Fln-Gal4 rescue flies (Fig 8A–8L’ and 8N). In addition, while 91.5% of myofibrils are degraded, degenerate, or frayed in bru1^M3 IFM, only 20.7% of myofibrils are affected in the rescue (Fig 8Q). However, the rescue flies are still flightless and their sarcomeres are also significantly shorter and thicker than control sarcomeres (Fig 8L’, 8M, 8O and 8P). Cross-section analysis revealed that hollow myofibril formation in the Fln-Gal4 rescue flies proceeds to the same extent as in bru1^M3 IFM, and myofibril density and diameter, the percent of myofibrils converted to hook and ring-like structures, as well as the number of myofibrils per bundle is not significantly different from the bru1^M3 mutant (Fig 8R–8U and S10F Fig). This indicates that expression of Bru1 with Fln-Gal4 is sufficient to produce a partial rescue of hypercontraction-related myofiber tearing and loss of sarcomere architecture, but is insufficient to rescue sarcomere and myofibril growth defects. To further test this conclusion, we performed semi-quantitative RT-PCR to validate the rescue of individual alternative splice events we have previously shown to be regulated by Bru1 [14,43]. We found that events in Strn-Mlck, sls, Tm1, and Mhc are completely rescued (Fig 8V and 8W, S11H and S11I Fig), while events in wupA and Zasp52 are partially rescued (Fig 8X and S11G Fig). We conclude that expression of Bru1 with Fln-Gal4 is sufficient to restore Bru1-mediated splicing of key structural genes and partially alleviate hypercontraction-related phenotypes seen in bru1^M3 IFM. However, this late-stage rescue cannot repair preexisting cytoskeletal structural and growth defects leading to a continued imbalance in growth in sarcomere length and width, abnormal radial growth of the myofibril, and associated functional deficits.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Expression of Bru1 restricted to late myogenesis partially rescues bru1^M3 myofibril phenotypes and restores alternative splicing defects.

(A–L’) Confocal projections of hemithorax and single-plane images of myofibril structure at 90 h APF and 1 d adult in control (A–C’); Bru1 overexpression (Fln-Gal4 > UAS-Bru1, D-F’); late temporal rescue (Fln-Gal4 > UAS-Bru1, bru1^M3/M3, G-I’); and mutant bru1^M3 (J–L’). The severity of myofiber detachment and torn myofibril phenotypes is partially rescued in Fln-Gal4 > UAS-Bru1, bru1^M3 IFM. Dashed line outlines the thorax boundary in (K). Frayed (cyan), degenerate (blue), or degraded (purple) myofibrils, arrowheads; Phalloidin stained actin, gray; scale bar = 100 μm (thorax); scale bar = 5 μm (IFM). (M, N) Quantification of flight ability (M) and DLM fiber integrity (N) in (B, E, H, K) (N > 50). Genotypes are denote by symbols: wild type, white square; bru1^M3, red square; Fln-Gal4, orange triangle; UAS-Bru1, magenta diamond. (O, P) Quantification of sarcomere length (O) and myofibril width (P) in (C–L’). Boxplots are shown with Tukey whiskers, outlier data points marked as black dots. Significance determined by ANOVA and post hoc Tukey (ns, not significant; **, p-val < 0.01; ***, p-val < 0.001). (Q) Quantification of myofibril phenotypes present in (C’–L’) (N > 70). (R) Single-plane cross-section images of myofibril structure in DLM from control, late temporal rescue, and mutant bru1^M3 at 90 h APF. Phalloidin stained actin, gray; scale bar = 5 μm. (S–U) Quantification of cross-section myofibril density (S), myofibril structure (T), and number of myofibrils per bundle (U) in (R). Fewer hollow myofibrils (rings) develop in Fln-Gal4 > UAS-Bru1, bru1^M3 IFM by 90 h APF. Significance determined as above, error bars in (T) = SEM. (V–X) Semi-quantitative RT-PCR verification of Bru1 regulated alternative splice events in Strn-Mlck (V), sls (W), and wupA (X). Representative gel images and quantification of percent exon use in control fibrillar IFM, tubular leg and jump (tergal depressor of the trochanter, TDT) muscle, and mutant bru1^M3 or late rescue Fln-Gal4 > UAS-Bru1, bru1^M3 IFM. Error bars = SD. Scheme on the right of alternative isoforms with primer locations, color coding consistent between scheme and bar plot. 3′-UTR regions in light beige. Exon numbering according to FB2021_05. Underlying data can be found in Fig 8 Source Data and Gels as listed in S6 Table. APF, after puparium formation; DLM, dorsal-longitudinal myofiber; IFM, indirect flight muscle.

https://doi.org/10.1371/journal.pbio.3002575.g008

Fly stocks and crosses

Experimental work with Drosophila melanogaster was approved under German §15 GenTSV (license number 55.1-8791-14.1099). Fly stocks were maintained using standard culture conditions at room temperature. Fly food was made by combining 16 L water, 1,300 g corn flour, 150 g soy flour, 1,300 g molasses, 130 g agar, 300 g yeast, and 650 g malt extract in a water-jacketed cooker. After food was sufficiently cool, 415 ml of 10% Nipagin and 295 ml acid mix (3% phosphoric acid, 21% propionic acid) were added. Food was distributed in vials and bottles with a peristaltic pump, dried at room temperature and stored at 4°C till use. All experimental crosses were kept in a 27°C incubator.

w¹¹¹⁸ alone or in combination with the relevant Gal4 driver (Gal4 x w¹¹¹⁸) was used as the wild-type control. We additionally validated the Gal4 x w¹¹¹⁸ control by verifying that flight ability, fiber integrity, and sarcomere length and width were not altered in the Gal4 driver crossed to w¹¹¹⁸ as compared to the Gal4 driver crossed to UAS-GFP (gift from Ulrike Gaul) or UAS-GFP-RNAi (Bloomington 41551) (S7 Fig). The bru1^M2 and bru1^M3 alleles were generated by insertion of a selectable 3xP3-DsRed cassette using a CRISPR-Cas9 approach [146]. The bru1^M2 mutant contains a cassette insertion upstream of bru1 exon 12, and has been described previously [43]. bru1^M3 is described in this manuscript. The UAS-Bru1 line was generated by amplifying the full-length bru1-RA transcript from w¹¹¹⁸ using RT-PCR (primer sequences are available in S4 Table), cloning the cDNA into the pUAST-attB transformation vector containing a 5× UAS-hsp70 promoter region and a SV40 polyadenylation sequence, and integrating at the attP-86Fb landing site [147] with ϕC31 integrase (S8A Fig). RNAi against bru1 (bru1-IR) was achieved with a previously characterized hairpin (GD41568) [14,43] obtained from the Vienna Drosophila Resource Center (VDRC). Df(2L)BSC407 is a deficiency allele on chromosome 2L that includes the bru1 locus and was obtained from the Bloomington Drosophila Stock Center (BDSC). Mhc¹⁰ is a TDT and IFM-specific amorphic myosin mutant [148]. Gal4 driver lines used in this study include: Mef2-Gal4 [149], which constitutively drives in all muscles; Him-Gal4 [42], which drives strongly in myoblasts and at early stages of IFM development until about 30 h APF; salm-Gal4 [84], which expresses in IFM starting from approximately 16 h APF; Act88F-Gal4 [77], which drives strongly in IFM starting at about 24 h APF; UH3-Gal4 [80], which expresses specifically in IFM starting at around 36 h APF; and Fln-Gal4 [83], which is expressed IFM-specifically from about 56 h APF (S6A Fig). weeP26 is a GFP-trap line inserted in the intron between Mhc exon 36 and 37, which only tags the long isoform of Mhc that terminates in exon 37 after the insertion [44,66,150]. Act88F localization was tracked using UASp-GFP-Act88F [75]. GFP-tagged fosmid reporter fly lines included Strn-Mlck-GFP (strn⁴, which tags IFM-specific isoform R) [14], wupA-GFP (fTRG925), Kettin-GFP (fTRG477), Clip190-GFP (fTRG156), Mlp84B-GFP (fTRG678), and Mlp60A-GFP (fTRG709) [151]. Talin-YPet is a CRISPR-mediated endogenous tag that strongly localizes to tendon attachment sites at muscle tips [152]. All fly lines and reagents are listed in the Resources S5 Table.

Generation of the bru1^M3 CRISPR allele

bru1^M3 was generated by CRISPR-mediated genome modification, using a previously described approach [146]. Two sgRNAs targeting sequences in the intron between bru1 exon 17 and 18 and after the 3′-UTR downstream of bru1 (S1A Fig and S4 Table) were screened in S2 cells for cutting efficacy and then co-injected with a 3x-P3 DsRed integration cassette flanked by approximately 1 kilobase long arms homologous to the genome sequence immediately upstream and downstream, respectively, of the sgRNA cut sites. The integration cassette contains a strong splice-acceptor followed by a 3 frame stop and a poly-adenylation sequence. Instead of the intended deletion, our genome modification resulted in the insertion of the cassette into the intron just upstream of exon 18, the last coding exon shared by all bru1 isoforms (S1A and S1B Fig). Although we could detect increased expression of bru1 mRNAs that include upstream exons 12 to 14, splicing into exon 18 was dramatically reduced, while splicing from exon 17 into the inserted construct containing the SV40 polyadenylation sequence was strongly increased (S1C Fig). Splicing into the integrated cassette in bru1^M3 results in deletion of exon 18, which encodes the terminal 88 amino acids (aa516-604 of bru1-RA) in the third RRM domain (the extended RRM3 domain comprises aa471-604 [39]). bru1^M3 is therefore a truncation allele that results in early termination after exon 17, which affects all bru1 isoforms (S1D Fig).

Behavioral assays

Flight ability was assayed as described previously [153]. Briefly, N > 30 adult male flies were collected under CO₂ on day 1 after eclosion, recovered overnight at 27°C and introduced into a 1-meter-long cylinder divided into 5 zones. Flies with a “normal flight” ability landed in the top 2 zones, “weak fliers” in the middle two, and “flightless” males fell to the bottom. Eclosion competence was assayed by counting the number of adult flies that emerged from their pupal cases. At least 60 live pupae were unbiasedly selected after 48 h APF and monitored until eclosion.

Immunofluorescence staining

Pupae of the desired genotype were tightly staged as follows: newly pupated 0 h APF flies (motionless, transparent-white body color and everted spiracles) were selected and sorted by sex, and males were transferred to a wetted filter paper in a 60 mm Petri dish and maintained at 27°C until the desired time point. Pupae and adult flies were dissected and stained as described previously [154]. For early pupae (before 48 h APF), we performed open-book dissections by removing the ventral half of the pupa to expose the developing IFM. For late pupal (after 48 h APF) to adult fly stages, we bisected the thorax sagittally after fixation to allow visualization of IFM. All samples were fixed in 4% PFA in 0.5% PBS-T (1× PBS + Triton X-100) for 30 to 60 min. For primary antibody staining, samples were blocked in 5% normal goat serum in 0.5% PBS-T for 90 min at room temperature, and then incubated overnight at 4°C with rabbit anti-Bru1 (1:500) [43] or rabbit anti-GFP (1:1,000, Abcam ab290). All antibodies are listed in the Resources S5 Table. After washing 3 times in 0.5% PBS-T for 10 min at room temperature, samples were incubated for 2 h at room temperature with secondary antibody Alexa 488 goat anti-rabbit IgG or rhodamine-phalloidin (1:500, Invitrogen, Molecular Probes). Samples were washed 3 times in 0.5% PBS-T and mounted in Vectashield containing DAPI.

Cryosectioning and histological staining

Pupae were staged as above. Cryosections were performed as described previously [42]. Samples were removed from the pupal case and fixed in 4% PFA in 0.5% PBS-T overnight at 4°C, and then incubated overnight in 30% sucrose in 0.5% PBS-T at 4°C on a rocking shaker. Pupae or 1 d adult thoraxes were arranged ventral side down in a vinyl specimen mold (Sakura Finteck), embedded in Tissue-Tek O.C.T. (Sakura Finteck) and snap-frozen on dry ice. Blocks were stored at −80°C, and then sectioned from anterior to posterior at 30 μm on a cryostat (Leica). Sections were collected on glass slides coated with 0.44 mM chromium potassium sulfate dodecahydrate in 1% gelatin to avoid detachment of the tissue during subsequent washing steps. Slides were post-fixed for 5 min in 4% PFA in 0.5% PBS-T at room temperature, washed 2 times in 0.5% PBS-T, and stained with rhodamine-phalloidin (1:500) for 2 h at room temperature in a humidity chamber protected from light. Slides were washed 3 times in 0.5% PBS-T, mounted with Fluoroshield containing DAPI (Sigma), and imaged on a Leica SP8X WLL upright confocal.

For hematoxylin-eosin staining, slides were post-fixed in formalin for 10 min, washed in 35°C running tap water, and rinsed in distilled water. Slides were incubated with Harris Hematoxylin stain (Roth) for 1 min, rinsed with distilled water, and placed under running tap water for 7 min to remove excess stain. After rinsing with distilled water, slides were stained in 1% Eosin solution (Apotheke Klinikum der Universität München) for 2 min, rinsed in distilled water, and dehydrated in 70% EtOH, 96% EtOH, and 99% EtOH in consecutive 5-min incubation steps. Slides were placed in xylenes (Roth) and mounted with Eukitt (Orsatec). After drying, slides were imaged on an Olympus IX83 inverted microscope with a 0.95 NA 40× objective.

Confocal microscopy

Laser-scanning confocal images were acquired on a Leica SP8X WLL upright confocal using Leica LAS X software in the Core Facility Bioimaging at the LMU, Biomedical Center (Martinsried, DE). Sagittal hemi-thorax sections were imaged with an HCPL FLUOTAR 10×/0.30 objective to resolve the whole fiber morphology and with an HCPL APO 63×/1.4 OIL CS2 objective to capture myofibril and sarcomere structure. Fibers of early pupae (before 48 h APF) were imaged with an HCPL APO 20×/0.75 IMM CORR CS2 objective. Samples stained with rabbit anti-Bru1 were processed in replicates using same antibody mix and were imaged with the same laser gain settings. 3D stereo projection movies of GFP-Act88F and Mhc^Wee-P26-GFP myofibril incorporation were assembled with Leica LAS X software from 0.1 mm confocal Z-stacks. Samples for the live imaging of spontaneous flight muscle contractions at 48 h and 72 h APF were prepared as previously described [155], and 10-min live recordings were obtained at a frame rate of 0.65 s/frame with an HCPL APO 40×/1.30 WATER CS2 objective. Samples for developmental GFP-tagged fosmid reporter expression at 48 h, 72 h, and 1 d adult were stained with the same antibody mix and images acquired with the same laser settings on a Zeiss LSM 780 confocal microscope using a Plan-APOCHROMAT 100×/1.46 oil immersion objective lens.

Transmission electron microscopy

w¹¹¹⁸ or bru1^M3 pupae were staged as described above to 48 h, 60 h, 72 h, or 90 h APF, and removed from the pupal case. Approximately 70% of the abdomen was carefully removed in 0.1 M cacodylate buffer using fine scissors, and thoraxes were incubated in glutaraldehyde fix solution (4% glutaraldehyde in 0.1 M cacodylate solution with 3% sucrose) for 2 h at room temperature on a rocking shaker. After pre-fixation, pupae were cut sagittally with a #C35 Feather microtome blade (Feather), and then fixed overnight in glutaraldehyde fix solution on a rocker at 4°C. Samples were post-fixed in osmium tetroxide (1% in cacodylate buffer) and stained with uranyl acetate (1% in water), dehydrated in a graded acetone series, and embedded in spurr low-viscosity epoxy resin following standard protocols. Samples were sectioned at 1 μm on an ultramicrotome (RMC MT 7000) with a diamond knife as previously described [156]. Sections were collected on glass slides, stained with Richardson blue and evaluated on an Olympus BX61VS light microscope until the sectioning depth of the IFM was reached. Then, the ultrathin sections were made at 60 to 70 nm, collected on TEM grids, and post-stained with uranyl acetate (1% in water) and lead citrate to enhance contrast. Images were acquired on an FEI Morgagni transmission electron microscope at 80 kV with a side-mounted SIS Megaview 1K CCD camera. At least 2 individual samples per genotype and time point were analyzed.

RNA isolation, RT-PCR, and RT-qPCR

All primers are listed in S4 Table. Whole thorax samples were prepared from 10 or more flies by removing the head, wings, and abdomen in a drop of pre-cooled 1× PBS using a fine scissors. Dissected IFM (>30 flies) and TDT (>60 flies) muscle samples were prepared from 1 d adult flies as described previously [44]. Legs were removed from >15 flies using fine scissors. Samples were snap-frozen in 50 μl of TRIzol (TRIzol Reagent; Ambion) on dry ice and stored at −80°C. Total RNA was isolated using TRIzol-Chloroform per the manufacturer’s guidelines. RNA samples were treated with DNaseI (New England Biolabs) and concentration was assessed using a Qubit 2.0 Fluorometer (Invitrogen). For normalization purposes, equivalent quantities of total RNA were used for cDNA synthesis with the LunaScript RT SuperMix Kit (New England Biolabs). For RT-PCR, 1 μg of cDNA was amplified with Phusion polymerase for 32 cycles and separated on a standard 1% agarose gel together with a 100 bp or 1 kb ladder (New England Biolabs). Ribosomal protein L32 (RpL32, RP49) served as an internal control for some reactions. Semi-quantitative analysis of gel band intensity to determine differences in exon use was performed using the “gel analysis” feature in Fiji. Band intensity was normalized against RpL32 to compare expression levels, or was used to calculate percent exon usage as: 100× (individual band intensity)/Σ (intensity of all bands observed for same primer pair).

For RT-qPCR, total RNA was extracted from IFM dissected from 150 flies using TRIzol. Reverse transcription and cDNA generation were performed with the LunaScript RT SuperMix kit, starting with 1 μg of cDNA. cDNA was diluted from 1:8 to 1:20, depending experimentally on the expression level of the target. Samples were assayed using SYBRgreen on a QuantStudio 3 (Applied Biosystems by Thermo Fisher Scientific) with an extension temperature of 60°C for 40 cycles. Samples were normalized against either RpL32 or Vha44, and fold change was calculated using the 2^-ΔΔCT method [157]. Data were plotted in GraphPad Prism, and significance was calculated from at least 3 replicates using either a Student’s t test (for a single sample versus control) or ANOVA (for more than 2 samples).

Image analysis

Image analysis was performed in Image J/Fiji [158]. For every confocal-based assay, 10 to 15 images were acquired from >10 adult flies or pupae. For TEM, more than 20 images were acquired from 2 to 3 animals. DLM fiber integrity and length were measured from Z-stack projections of hemi-thorax samples based on rhodamine-phalloidin staining. Fiber length was measured as distance (μm) from anterior to posterior tip of a single fiber using the freehand drawing tool in Fiji. The same tool was used on TEM images, after setting the scale based on the inset in each image, to measure the sarcomere length and myofibril width as the distance between or across Z-disks, respectively. For confocal images, sarcomere length and myofibril width of sagittal sections, as well as myofibril diameter and density of cross-sections, were measured based on phalloidin (F-actin) staining using MyofibrilJ [42] (https://imagej.net/MyofibrilJ). The number of myofibrils per bundle was quantified in a semi-automated manner as follows: the boundaries of individual fiber bundles were determined from the Z-stack and at least 3 separate and complete bundles from >10 flies were cropped, analyzed with MyofibrilJ, and then manually corrected for accuracy (myofibril “dots” were added or subtracted based on manual inspection of the MyofibrilJ output). Each “hook” or “ring” structured myofibril was counted as one. The number of myofibrils classified as “dots,” “hooks,” or “rings” were manually counted per defined cross-sectional area. The number of actin cables per fiber (at 26 h and 28 h APF) was determined manually based on rhodamine-phalloidin staining. Bru1 signal intensity was analyzed as follows: the nuclei were selected based on DAPI staining using the wand tool in Fiji, and relative Bru1 signal in the nuclei (overlapping the DAPI positive regions) was quantified as the mean intensity of Bru1 staining divided by the mean intensity of the absolute background.

Quantification of spontaneous contractions in developing DLM has been described previously [42] and was modified here as follows. Live movies were obtained for 10 min, and “twitches” in each visible DLM fiber at 48 h APF were classified as single, double, or triple contraction events, and resulted in displacement of the myofiber and return to the original resting position. The range of each single spontaneous contraction at 48 h APF was measured using the freehand drawing tool in Fiji as the distance between the tip of the fiber prior to and after the extension. At 72 h APF, contraction events quantified in the bru1 mutant were slower than “twitches” and resulted in a one-directional myofiber extension, i.e., without return to the original resting position. All data were tabulated in Excel, and plotting and statistical analysis were performed in GraphPad Prism 8.4.0, using ANOVA or unpaired Student’s t test.

Deconvolution

Deconvolution was performed in ImageJ using the Diffraction PSF 3D and FFTJ–DeconvolutionJ plugins. Images were obtained at a resolution of 23.2848 pixels per micron and a corresponding voxel size of 0.0429 × 0.0429 × 0.2014 micron³ for deconvolution and 3D projection. YZ stacks were generated with interpolation using the Reslice option native to ImageJ. A digital point spread function (PSF) was generated with the following settings: IR = 1.518, NA = 1.40, 580 nm, pixel spacing 23.28 and corresponding slice spacing of 4.97 units, with a Rayleigh resolution of 10.68 pixels. Deconvolution was performed without resizing at 32-bit, double precision, and with a gamma setting of 0.01/0.03.

Proteomics

IFM were dissected as described previously [44], and 30 flies per genotype were dissected from w¹¹¹⁸ and bru1^M2 at 72 h APF and 1 d adult, and 4 biological replicates were prepared per genotype. Samples were processed according to the manufacturer’s instructions using the PreOmics iST Sample Preparation Kit (Preomics, #0000.0061) and analyzed by the Protein Analysis Unit (ZfP) at the LMU Biomedical Center as follows. Desalted peptides were injected in an Ultimate 3000 RSLCnano system (Thermo) and separated in a 25-cm analytical column (75 μm ID, 1.6 μm C18, IonOpticks) with a 100-min gradient from 5% to 60% acetonitrile in 0.1% formic acid. The effluent from the HPLC was directly electrosprayed into a Qexactive HF (Thermo) operated in data-dependent mode to automatically switch between full scan MS and MS/MS acquisition. Survey full scan MS spectra (from m/z 375–1,600) were acquired with resolution R = 60,000 at m/z 400 (AGC target of 3 × 10⁶). The 10 most intense peptide ions with charge states between 2 and 5 were sequentially isolated to a target value of 1 × 10⁵, and fragmented at 27% normalized collision energy. Typical mass spectrometric conditions were: spray voltage, 1.5 kV; no sheath and auxiliary gas flow; heated capillary temperature, 250°C; ion selection threshold, 33,000 counts. MaxQuant 1.6.14.0 [159] was used to identify proteins and quantify by LFQ with the following parameters: Database, Uniprot_AUP000000803_Dmelanogaster_Isoforms_20210325.fasta; MS tol, 10 ppm; MS/MS tol, 20 ppm; Peptide FDR, 0.1; Protein FDR, 0.01 Min. peptide Length, 5; Variable modifications, Oxidation (M); Fixed modifications, Carbamidomethyl (C); Peptides for protein quantitation, razor and unique; Min. peptides, 1; Min. ratio count, 2. Additional analysis was performed in Perseus [160]. We filtered the data to retain biologically relevant protein groups with missing intensities between mutant and control samples (MNAR values) by requiring at least 3 replicates in either control or mutant to contain a value, and we imputed missing values by replacement with a constant value (lowest observed intensity– 1). Differential expression was tested by t test with FDR = 0.05. Results were exported and further visualization performed in R.

mRNA-Seq and transcriptome bioinformatic analysis

For transcriptome analysis, IFM were dissected from 1 d adult (0 to 24 h after eclosion) w¹¹¹⁸ and bru1^M3 flies as described previously [44]. Two replicates of IFM from 100 flies were dissected per genotype. RNA was isolated using TRIzol and sent to LC Sciences (Houston, TX) for sequencing. After quality verification, poly-A mRNA selection and library construction, samples were sequenced as stranded, 150 bp paired-end on an Illumina HiSeq to a depth greater than 70 million reads. Transcriptome data from bru1-IR and control IFM at 24 h, 30 h, 72 h, and 1 d adult was generated previously [14,42], and was reanalyzed as part of this manuscript.

Sequence data was mapped with STAR to ENSEMBL genome assembly BDGP6.22 (annotation dmel_r6.32 (FB2020_01)). Files were indexed with SAMtools and processed through featureCounts. Downstream analysis and visualization were performed in R using packages listed in the S5 Table. Differential expression was analyzed at the gene level with DESeq2 and at the exon level with DEXSeq. Differential exon use can reflect alternative splicing as well as alternative promoter use. Both packages were additionally used to generate normalized counts values. We employed previously annotated sets of genes, including sarcomere proteins [42], genes with an RNAi phenotype in muscle [153], core fibrillar genes regulated by Spalt major and differentially expressed between IFM and tubular muscle [14], mitochondrial proteins [104,161], and all genes with the Flybase GO term “muscle contraction,” “actin cytoskeleton,” or “actin cytoskeleton organization.” Genes with a hypercontraction phenotype in Drosophila muscle were curated by hand from the literature. Full lists of all gene categories are available in S2 Table. Plots were generated using ggplot2 or ComplexHeatmap. Unless otherwise stated, we used an adjusted p-value threshold of 0.05 for significance. If a log₂FoldChange threshold was used in a specific analysis, it is specified in the legend for the relevant figure panel.