Drosophila Transglutaminase preserves the integrity of muscle attachments with and without mechanical strain
Dylan Feist, Ziwei Zhao, David Brooks, Jared Ridder, Emma Peters, Nicole Green, Prabhat Tiwari, Erika R. Geisbrecht
TL;DR
This study shows that Transglutaminase in fruit flies helps maintain muscle attachments, especially under mechanical stress.
Contribution
The study reveals a new role for Transglutaminase in preserving muscle attachments under mechanical strain in Drosophila.
Findings
Tg RNAi knockdown or inactive Tg causes muscle–cuticle detachment in Drosophila.
MTJ stability is reduced in adults when mechanical tension increases.
Tg crosslinking provides stability to muscle attachments under tension.
Abstract
The strict control, yet dynamic nature of adhesive structures that form in the extracellular environment are crucial for the development and homoeostasis of multicellular organisms. A gradual increase in the strength of the myotendinous junction (MTJ) occurs as ligands accumulate in the extracellular matrix (ECM) and bind to opposing integrin complexes at muscle junction interfaces. Although proteomic studies of the muscle–tendon junction in mice and humans have revealed the complexity of protein classes in this extracellular environment, the functions of many ECM proteins remain elusive. To fill this gap in knowledge, we performed a sensitized genetic screen to expose MTJ-relevant genes in Drosophila melanogaster whose functions might be redundant or sensitive to mechanical strain. Aside from the expected ECM proteins that comprise the basement membrane, we uncovered functional roles…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Fig. 1
Fig. 2
Fig. 3
Fig. 4- —National Institute of Arthritis and Musculoskeletal and Skin Diseaseshttp://dx.doi.org/10.13039/100000069
- —USDA National Institute of Food and Agriculture Hatch
- —Kansas INBRE
- —Kansas State Universityhttp://dx.doi.org/10.13039/100007829
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsBlood properties and coagulation · Skin and Cellular Biology Research · Cellular Mechanics and Interactions
INTRODUCTION
Normal growth and development require proper extracellular matrix (ECM) adhesion for a wide range of biological processes, including cell migration, tissue development and immunity (Saraswathibhatla et al., 2023). The importance of the ECM is underscored by the wide range of diseases that result from the abnormal biogenesis and/or maintenance of these non-cellular components, which include connective tissue disorders, cardiovascular disease, and tumorigenesis (Iozzo and Gubbiotti, 2018; Theocharis et al., 2019). Despite extensive knowledge of the ECM in the development and physiology of multicellular organisms, mechanisms that control the remodeling and stability of this extracellular environment, especially in response to mechanical stresses, remain incompletely understood.
The myotendinous junction (MTJ), or site of muscle–tendon attachment, is an excellent model for understanding ECM biogenesis and maintenance in response to mechanical tension (Narayanan and Calve, 2021; Valdivia et al., 2017). Muscle and tendon cells arise from independent progenitor lineages and reciprocal signaling between these two cell types is a prerequisite for MTJ biogenesis in both invertebrates and mammals. Structurally, the mature MTJ consists of secreted ECM proteins, some of which are physically tethered to integrin complexes expressed on the plasma membranes of both muscle and tendon cells (Charvet et al., 2012; Schweitzer et al., 2010; Subramanian and Schilling, 2015; Tong et al., 2024). Functionally, the MTJ is the primary site for force transmission from the interior of the muscle cell, across its membrane, and to the ECM. In healthy muscle tissue, the MTJ provides resistance against mechanical stress generated during muscle contraction and any decrease in MTJ formation and/or stability leads to muscle detachment in diverse organisms (Jakobsen and Krogsgaard, 2021; Mun et al., 2022; Schweitzer et al., 2010).
Proteomic analysis of the MTJ in mice and humans has revealed a rich, heterogeneous mix of ECM proteins, although the function for many of these remains elusive (Jacobson et al., 2020; Karlsen et al., 2022; Schmidt et al., 2024). Bioinformatic approaches to define the matrisome – or the complete set of genes and proteins that comprise the ECM – show high levels of conservation across species, and it can be divided into distinct categories (Davis et al., 2019; Hynes and Naba, 2012; Statzer and Ewald, 2020). Core matrisome proteins are structural components that include collagens and glycoproteins as part of the basement membrane. Matrisome-associated proteins might share features with core matrisome proteins (ECM-affiliated), participate in ECM remodeling (ECM regulators) and/or modulate mechanical strength and stability (ECM crosslinkers). Additional subgroups responsible for maintaining the rigid, yet flexible meshwork in the ECM environment include metalloproteases and apical matrix proteins. The simpler genome of Drosophila melanogaster, with less redundancy than vertebrates, has proven instrumental in dissecting mechanisms of ECM secretion and assembly throughout development (Davis et al., 2019).
Although all Drosophila MTJs share similar structural proteins, the morphology of these muscle–tendon attachment sites differ during the embryonic–larval and pupal–adult stages to reflect distinctive functional requirements (Poovathumkadavil and Jagla, 2020). A single syncytial muscle connects to an individual tendon cell during embryonic development to provide an interdigitated interface that allows for slower, sustained contractions during larval locomotion (Green et al., 2016). In contrast, the pupal–adult MTJ is composed of multiple indirect flight muscles (IFMs) and tendon cells that allow for an increased surface area to withstand more rapid, powerful muscle contractions required during flight. Mechanical tension and spontaneous muscle twitching provide physical signals not only to initiate IFM sarcomere assembly (Weitkunat et al., 2017), but also to strengthen muscle–tendon attachment sites that link the muscles to the chitinous exoskeleton.
In this study, we carried out a sensitized genetic screen in live Drosophila larvae to uncover ECM proteins required for MTJ stability during organismal growth. We further examined the ECM crosslinking protein Transglutaminase (Tg) to understand the enzymatic and non-enzymatic roles of Tg in MTJ formation and remodeling throughout development. In addition to the established role of Tg in hemolymph coagulation and adult cuticle morphogenesis (Karlsson et al., 2004; Oliva et al., 2009; Shibata et al., 2010), we find that Tg crosslinking activity is required for MTJ stability, likely to counteract mechanical tension during pupal muscle development.
RESULTS AND DISCUSSION
Sensitized RNAi screen uncovers ECM proteins required for muscle–tendon adhesion
Fondue (fon) and Tiggrin (tig) are both Drosophila ECM proteins crucial for maintaining adhesion at muscle–muscle and muscle–tendon attachment sites (Bunch et al., 1998; Green et al., 2016). We combined Myosin heavy chain (Mhc)–GFP into mutant backgrounds for each of these alleles and examined the body wall muscle pattern in live third-instar larvae (L3). Heterozygous combinations of fon (fon^Δ24^/+) or tig (tig^X^/+) alleles displayed the typical muscle pattern in all hemisegments (Fig. 1A). Consistent with previous results (Bunch et al., 1998; Green et al., 2016), homozygous fon^Δ24^/fon^Δ24^ or tig^X^/tig^X^ mutants showed detached (yellow arrows) and/or missing muscles (yellow carets) (Fig. 1B). Despite the presence of attached muscles in fon^Δ24^/+ larvae, this 50% reduction in fon gene copy (fon^Δ24^/+) sensitized the animals towards a muscle detachment phenotype when combined with tig RNAi (Green et al., 2016). We leveraged this dosage sensitivity to screen for predicted ECM proteins in either fon^Δ24^/+ or tig^X^/+ sensitized backgrounds.
Sensitized RNAi screen for matrisome proteins involved in MTJ stability. (A,B) Larval muscles labeled with Mhc–GFP. Heterozygous fonΔ24/+ or tigX/+ larvae show a normal muscle pattern (A), whereas fon-/- or tig-/- homozygotes display detached (yellow arrows) or missing (yellow carets) muscles (B). (C) Genetic crosses and live screening strategy to identify candidate matrisome genes required for MTJ stability. Created in BioRender by Geisbrecht, E. 2026. https://BioRender.com/elz74cs. This figure was sublicensed under CC-BY 4.0 terms. (D) Stacked bar plot summarizing the number of genes showing either detached and/or missing muscles compared to the total number of genes screened for each class of ECM proteins. (E) Grouped summary plot depicting the number of genes in each category after expressing the RNAi alone (gray), or in the fon (green) or tig-sensitized (magenta) backgrounds. (F) Venn diagram summarizing the number of genes with detached and/or missing muscles in each genotype. (G) The WT pattern of L3 larval muscles at low (left panel) or high (right panel) magnification. F-actin labels muscles (magenta) and βPS integrin (cyan) denotes muscle-muscle attachment sites. (H) Tg::3xHA can be detected at muscle–muscle attachment junctions upon overexpression with da-Gal4. Images in A,B,G,H representative of n=3 experimental repeats.
The ECM protein Thrombospondin is secreted from tendon cells (Subramanian et al., 2007), whereas the main source of Fon is the fat body (Green et al., 2016). Therefore, we chose the weak, ubiquitous da-Gal4 to drive UAS-based candidate RNAi lines to assure that all cellular sources of secretion were targeted with minimal effects on viability. A Mhc-GFP transgene was combined into the background of da-Gal4 alone or the fon^Δ24^ (fon^Δ24/+^; da-Gal4) or tig^X^ (tig^X/+^; da-Gal4) heterozygotes to facilitate live screening from the first-instar (L1) through L3 larval stages using a fluorescence dissecting microscope (Fig. 1C). At total of 86 independent UAS-RNAi lines representing 67 putative ‘matrisome’ genes were selected for screening using the Gene List Annotation for Drosophila (GLAD) resource (https://www.flyrnai.org/tools/glad/web/; Davis et al., 2019; Statzer and Ewald, 2020). Selected genes fell into the following categories – metalloproteases, ECM-affiliated, ECM regulators, ECM glycoproteins, ECM crosslinkers, collagens, and apical matrix proteins. We also included genes encoding for Toll pathway members in the adhesion and/or signaling group given that we previously showed that fon genetically interacts with genes encoding for proteins in the Toll signaling pathway (Green et al., 2018) (Table S1). Genes encoding for insect-related development (chitin or chitin-related, eggshell, glue or mucins etc.) or well-studied signaling pathways with secreted components (Wnt, FGF, etc.) were excluded.
Live screening throughout larval development resulted in 40% of the candidate RNAi lines (27/67) showing missing and/or detached muscles in the sensitized fon or tig sensitized backgrounds (Fig. 1D–F; Fig. S1). As expected, ≥75% of genes encoding for basement membrane components (e.g. collagens or laminins in the ECM glycoprotein category) exhibited MTJ phenotypes (Fig. 1D,E). This finding is consistent with the enrichment of these proteins at the MTJ interface (Karlsen et al., 2022; Schmidt et al., 2024) and muscle fiber detachment resulting from genetic mutations in laminin α2 (Charvet et al., 2012). A similar trend was also evident for adhesion and signaling genes, whereby seven out of nine Toll receptors were required to maintain muscle attachments. More studies are necessary to determine whether the Toll family of proteins directly mediate cell–cell adhesion or indirectly regulate cell signaling, possibly in response to stress (Green et al., 2018). In the remaining matrisome-based classifications (metalloproteases, ECM-affiliated, ECM regulators and ECM crosslinkers), between 20 and 60% of the genes analyzed showed muscle detachment phenotypes, either due to genetic redundancy or weak knockdown of protein levels using our RNAi approach. We were surprised that no MTJ phenotype was evident in genes encoding for apical matrix proteins, including Dumpy (Dpy) and other zona pellucida domain (ZPD)-containing proteins, given that they are required in other stages of development for cuticle attachment (Bökel et al., 2005; Chu and Hayashi, 2021; Göpfert et al., 2025).
To verify whether any of the putative ECM proteins are found at sites of muscle attachment, we examined the localization of publicly available tagged proteins. For this assay, we choose to analyze muscle–muscle attachment sites at the hemisegment borders where the localization of βPS integrin is evident (Fig. 1G). Both Toll::Venus and Toll6::GFP accumulated at muscle–muscle junctions when driven with da-Gal4 (Fig. S2). As it is known that Collagen IV is secreted from fat body tissue (Pastor-Pareja and Xu, 2011), the ppl-Gal4 driver was used to express Col4a1. This GFP-tagged protein also localized to muscle–muscle attachment sites (Fig. S2), as did overexpression of Tg driven with da-Gal4 (Fig. 1H). For the remainder of this study, we focused on the role of Tg in MTJ formation and stability. Further quantification of Tg RNAi phenotypes revealed more larvae with detached or missing muscles in either fon or tig-sensitized backgrounds than da>Tg RNAi alone (Fig. S3A). Given that Fon and Tig are known substrates of Tg activity in clot formation (Karlsson et al., 2004; Lindgren et al., 2008), Tg crosslinking might also provide structural rigidity at the MTJ.
Tendon cell secretion of Tg influences muscle attachment
Tg protein has previously been reported in whole L3 larvae, but not the L1 or L2 stages via western blotting (Shibata et al., 2010). As RNA and protein expression data are publicly available (https://flybase.org/), we mined this data to gain a better understanding of Tg expression throughout development, including the embryonic stages. RNA-seq data obtained through the modENCODE Development RNA-Seq project (accession PRJNA75285; Boley et al., 2014; Graveley et al., 2011) revealed Tg mRNA expression as early as 4–6 h after egg laying (AEL) with a peak at 18–20 h (Fig. 2A). Tg transcripts decreased in the L1 stage and continued to reach a maximum in 1-day-old pupa, followed by a gradual decrease until eclosion and continuing into adulthood. A different, almost inverse, pattern of Tg was observed across the developmental proteome (accession PXD005691; Casas-Vila et al., 2017) whereby protein expression was high in larval development and again in the adult stage, possibly reflecting a transcriptional feedback mechanism.
*sr-dependent secretion of Tg is required for embryonic MTJ formation. (A) Profile of Tg mRNA or Tg protein expression across developmental stages (data from Flybase). (B) WT embryos stained with anti-Tg (magenta). Tg partially overlaps with the pattern of tendon cells marked by sr-GFP (green) in st. 13 embryos (top panel) or βPS integrin (cyan) in st. 15 (middle panel) and st. 17 (bottom panel) embryos. White asterisks mark trachea. (C) qPCR confirms knockdown of Tg mRNA in whole larvae at the L3 stage. N=3 biological replicates. Mean±s.d. (D) St. 16 embryos immunostained for Tropomyosin (TM, magenta) or βPS integrin (cyan) to visualize the somatic muscles or MASs, respectively. The average length of all MASs is shorter in sr>Tg RNAi embryos compared to controls (sr>mCh.NLS). (E) Violin plot with quantification of the ventral MASs in control and Tg RNAi embryos. Dashed lines highlight median and quartiles. Each data point represents the average length of MASs in hemisegments A1–A4 for muscles VL1–VL4 per embryo. n≥14 for each genotype. Average values reflected as median. ***P<0.0005; ****P<0.0001; n.s., not significant (Kruskal–Wallis test). (F) Amino acid alignment showing the conserved cysteine in Drosophila Tg isoforms A (dTg.A) and B (dTg.B), human TGM2 (h.TG2), and human factor XIIIa (h.FactorXIIIa). Created in BioRender by Geisbrecht, E., 2026. https://BioRender.com/jfriraa. This figure was sublicensed under CC-BY 4.0 terms. (G) Western blot and bar graph showing increased levels of protein in Tg.A, but similar expression levels of Tg in da>Tg WT or da>Tg C330S whole L3 larva. Total protein is shown as a loading control. n=3 biological replicates. Mean±s.d. (H) Violin plot with quantification of the ventral MASs upon overexpression of normal or mutant Tg. Dashed lines highlight median and quartiles. Each data point represents the average length of MASs in hemisegments A1–A4 for muscles VL1–VL4 per embryo. n≥14 for each genotype. Average values reflected as median. *P<0.05; **P<0.001; ***P<0.0001; ns, not significant (Kruskal–Wallis test). Images in B representative of n=3 experimental repeats.
To further investigate the pattern of Tg expression in embryogenesis, we developed an antibody against the C-terminus of this protein (amino acids 494–776). Tg immunoreactivity was evident in stage (st.) 13–17 embryos in the developing trachea and in epidermal stripes (Fig. 2B). Tracheal staining is consistent with a role for Tg in chitin morphogenesis as chitin is a structural component of the apical ECM (aECM) in the tracheal lumen (Öztürk-Çolak et al., 2016). Given that Tg (magenta) localization in the epidermis looks similar to the pattern of tendon cells (Nabel-Rosen et al., 1999), we confirmed overlap using the tendon cell marker stripe (sr)-GFP (green) in st. 13 embryos. As development proceeded until st. 17, Tg continued to be present in the epidermis adjacent to muscle attachment sites (MASs) marked by βPS integrin (Fig. 2B). Although βPS integrin was expressed at similar levels in all hemisegments (Fig. S3B), Tg staining was enriched in the posterior half of st. 13 embryos and became visible in anterior hemisegments as development proceeded (Fig. 2B), suggesting that Tg expression is tightly regulated.
We next explored the possibility that tendon cells are a source of Tg and are required for the formation of the embryonic MAS. Two independent Tg RNAi lines emerged from our sensitized screen, so we chose v26101 for further studies. Even though this UAS-Tg RNAi line has been published (Lin et al., 2015), we first confirmed Tg knockdown using qPCR and western blotting under control of the da promoter. Tg transcripts were reduced ∼70% in the L3 larval stage (Fig. 2C) and Tg protein was nearly undetectable in adults (Fig. S3C), validating the use of this RNAi line and the newly generated Tg antibody. To assess whether a reduction in Tg affected the formation of embryonic MASs, we utilized the tendon cell sr-Gal4 driver to induce Tg RNAi and observed shorter MASs in late-stage embryos (Fig. 2D,E). This phenotype was not enhanced by the addition of Dcr-2.
Mammalian transglutaminase 2 (TGM2) and factor XIIIa both exhibit crosslinking activity for ECM proteins and fibrin, respectively (Anokhin et al., 2020; Yao et al., 2024). To determine whether Tg crosslinking activity is important for MTJ formation in the embryo, we created transgenic flies with a mutation in a conserved cysteine residue (C330S) that is essential for catalytic activity (Fig. 2F) (Balklava et al., 2002; Lee et al., 1993). Insertion of the Tg WT (wild type) or Tg C330S transgenes into the same chromosomal location ensured similar expression levels (Fig. 2G). Expression of catalytically inactive Tg C330S, but not Tg WT, showed reduced MAS lengths in st. 16 embryos (Fig. 2H). Testing of an independent Tg line (UAS-Tg.A) with higher levels of Tg expression (Fig. 2G) also produced smaller muscle attachments (Fig. 2H). These data together hint that both loss of Tg activity or excess protein levels can both be detrimental for normal MTJ formation.
Although chitin deposition as part of the procuticle begins in late st. 16 embryos, the bulk of chitin secretion occurs in st. 17 to eventually form the differentiated cuticle (Moussian et al., 2006). The emergence of Tg expression earlier suggests a role for crosslinking that is independent of initial MTJ formation, but does not exclude a crosslinking function during the foraging larval stages. To examine this further, we characterized Tg function by assessing muscle attachment in wandering L3 larva. The inclusion of Dcr-2 was necessary to observe phenotypes, which is consistent with high levels of Tg mRNA in embryos after MTJ formation and just before larval hatching (Fig. 2A). Approximately 10% of Dcr-2, sr>Tg RNAi larva showed a peculiar curling behavior, always in a clockwise direction, at the L2 stage (Fig. 3A). Dissection of Tg knockdown individuals at the L3 stage showed either missing (yellow asterisks mark attached muscles) or detached (yellow arrows) muscles compared to Dcr-2, sr>mCh-NLS controls (Fig. 3B). Two types of muscle attachments are present in Drosophila larval muscles – direct (intrasegmental) MASs, which are characterized by muscle–tendon contacts, and indirect (intersegmental) MASs, which include both muscle–muscle and muscle–tendon attachments (Maartens and Brown, 2015; Prokop et al., 1998). Loss of Tg preferentially affected muscles that are directly attached to the cuticle (Fig. 3C), consistent with the localization of Tg in epidermal cells that will eventually secrete cuticle. However, expression of Tg WT or Tg C330S did not result in missing or detached muscles (Fig. 3C), implying that either cuticle crosslinking and/or remodeling is not necessary for larval MTJ maintenance and/or the amount of endogenous Tg present is sufficient to perform this enzymatic activity.
*Knockdown of Tg affects MTJ stability. (A) Whole-mount WT L3 larvae (1). L2 larvae with RNAi knockdown of Tg (Dcr-2, sr>Tg RNAi) showed a prominent curled phenotype (2). (B) Three hemisegments of the larval musculature in control (Dcr-2, sr>mCh-NLS) or Tg RNAi (Dcr-2, sr>Tg RNAi) L3 larva. Attached muscles are indicated by *. Missing (carets) or detached (arrows) muscles are evident in Tg RNAi fillets. (C) Scatter plot depicting the percentage of muscle detachment in direct (DO1–DO2, LT1–LT4, VO4–VO6) versus indirect (DA1–DA2, DA1–DA3, LO1, VL1–VL4) muscles as indicated in the schematic. Directly attached muscles are preferentially affected upon knockdown of Tg RNAi. Each data point represents the percentage of detached muscles per larval fillet. n>10≤44 for each genotype. Average values reflected as median. ***P<0.0001; ns, not significant (two-way ANOVA with Dunnett's multiple comparisons test post test). Schematic created in BioRender by Geisbrecht, E., 2026. https://BioRender.com/hbqh7j5. This figure was sublicensed under CC-BY 4.0 terms. Images in A representative of n=3 experimental repeats.
Increased mechanical force promotes MTJ detachment in adult IFMs
Cuticle morphology and function differ in distinct developmental stages. The larval cuticle is primarily composed of an endocuticle that is softer and allows for growth and larval contractions (Chihara et al., 1982; Moussian, 2010). In contrast, the cuticle that overlies the adult thorax consists of an endocuticle and hardened exocuticle that provides additional rigidity for flight muscle attachment. A subset of notum epithelial cells differentiates into tendon cells and serve as attachment sites for the IFMs (Schulman et al., 2015; Valdivia et al., 2017). Perturbation of jitterbug (jbug; also known as Filamin-type protein), chascon (chas) or Rho kinase (Rok) in tendon cells, or dpy and quasimodo (qsm) in the ECM affect assembly of thoracic MTJs, causing deformation of the cuticle (Chu and Hayashi, 2021; Olguín et al., 2011; Vega-Macaya et al., 2016). As Tg is detected in tendon cells in the adult thorax (Fig. 4A), we were able to use the notum deformation assay to examine the requirement for Tg during IFM formation at muscle–tendon attachment sites. Expressing either one or two copies of the UAS-Tg RNAi transgene throughout development under control of the sr promoter did not show an obvious cuticular phenotype (Fig. S4A). Analysis of bisected thoraces revealed minor irregularities at the MTJ whereby F-actin-labeled IFMs occasionally pulled away from attachment sites in regions that coincided with abnormally distributed βPS integrin (Fig. S4B).
Increased mechanical tension weakens IFM attachment. (A) sr-Gal4 is driving CD8–GFP (green) to mark the tendon cells in 2-day-old adult IFMs where muscles are labeled with phalloidin (gray). Anti-Tg immunostaining (magenta) shows labeling in tendon cells (t) and the cuticle (c). (B) Phenotypic classes present in Mhc[S1]/+, sr>Tg RNAi adult flies. Anterior is to the left. Compared to a ‘WT’ notum, small and V-like indentations (white arrowsheads) are seen in phenotypes classified as ‘mild.’ These are larger and wider in the ‘severe’ category. Phalloidin staining reveals the normal complement of six IFM fibers () in WT samples. Mild or severe phenotypes show degenerated or detached IFMs (yellow carets). (C) Grouped bar graph depicts the percentage of female or male adult flies with WT, mild or severe phenotypes shown in B. Increased tension in Mhc[S1]; sr>Tg RNAi enhances thoracic indentation phenotypes compared to sr>Tg RNAi individuals. N=3 biological replicates with a total of n=90 for each genotype. Mean±s.d. (D) Grouped bar graph of female or male notums showing WT, mild or severe phenotypes. Increased tension in Mhc[S1]; sr>Tg C330S enhances thoracic indentation phenotypes compared to Mhc[S1]; sr>Tg WT or controls. N=3 biological replicates with a total of n=90 for each genotype. Mean±s.d. (E) Schematic of optogenetics approach to induce mechanical force in the IFMs. Created in BioRender by Geisbrecht, E. (2026). https://BioRender.com/ckiahly. This figure was sublicensed under CC-BY 4.0 terms. (F) Experimental outline of temperature shifts, ATR feeding and LED exposure. C created in BioRender by Geisbrecht, E., 2026. https://BioRender.com/xd4mnoc. This figure was sublicensed under CC-BY 4.0 terms. (G) Representative confocal single plane images of sr>Tg WT RNAi IFMs subjected to optical stimulation showing either attached (−ATR, upper panel) or detached (+ATR, lower panel) IFMs (F-actin, gray) with stretched βPS integrin (magenta) staining. (H) Scatter plot showing that IFM detachment is heightened in sr>Tg WT flies after muscle contractions induced by ATR and LED exposure. n=10 for each genotype. Mean±s.d. *P<0.05; **P<0.0005; ns, not significant (Mann–Whitney test). Images in A representative of n=3 experimental repeats.
Given that genetic manipulation of Tg was not sufficient to disrupt muscle attachment, we next tested whether perturbing mechanical forces during pupal myogenesis impacted MTJ integrity by recombining a temperature-sensitive dominant hypercontraction allele of Myosin heavy chain (Mhc[S1]) (Montana and Littleton, 2004) with sr-Gal4 to drive Tg knockdown. This actomyosin-generated increase in tension not only caused mild or severe indentations in the thoracic cuticle but also resulted in IFM detachment (Fig. 4B). Quantification of these deformation phenotypes (as mild or severe as opposed to WT) was substantially increased upon Tg RNAi knockdown compared to controls and exhibited a sex-specific difference whereby males showed enhanced phenotypes compared to females (Fig. 4C). No cuticle deformation was observed in Tg RNAi alone (Fig. S4A).
We next tested whether decreased Tg activity affected IFM attachment when subjected to increased mechanical force. The sr promoter was used to induce either Tg WT or Tg C330S in flies heterozygous for Mhc[S1], followed by analysis of cuticle deformation. Indented thoraces were present in 10% of females and 25% of males of the genotype Mhc[S1]/+, sr>Tg WT (Fig. 4D), similar to our embryo data whereby excess Tg protein was detrimental for MTJ formation (Fig. 2H). Disruption of crosslinking activity through expression of catalytically inactive Tg (Tg C330S) coupled with Mhc[S1]-induced muscle tension increased the percentage of indented thoraces to ∼35% of females and 60% of males (Fig. 4D). We speculate that overexpression of catalytically inactive Tg C330S weakens the structural rigidity of the cuticle and/or the ability of the IFMs to remodel the MTJ due to disrupted or uneven distributed forces during development.
We next tested whether Tg was required for maintenance of adult IFM attachment after development. Flies expressing Tg RNAi, Tg WT or Tg C330S in tendon cells with sr-Gal4 were subjected to forced muscle contractions using an optogenetic approach in conjunction with the LexA/LexAop system (Fig. 4E). Flies were raised at 18°C to limit Gal4/UAS and LexA/LexAop expression during development. Adult flies were then shifted to 30°C for 4 days, fed the chromophore all-trans retinol (ATR) to increase sensitivity of the channelrhodopsin (CsCrimson), and subjected to 655 nm light for 5 s on–off over a 2-h period for neuronal activation (Fig. 4F). After being subjected to optical stimulation, sr>Tg WT flies expressing channelrhodopsin in motor neurons (RapGAP1>CsCrimson) with no ATR exhibited normal climbing behavior (Movie 1), whereas flies of the same genotype fed ATR dropped to the bottom of the vial (Movie 2). Although it is difficult to ascertain IFM engagement through the cuticle, LED stimulation of sr>Tg RNAi, sr>Tg WT or sr>Tg C330S flies all exhibited contraction behaviors not present in flies without the transgenes (Movies 3–7). For quantification, the average distance between the cuticle edge of the βPS integrin signal and the internal edge of the IFMs labeled by phalloidin was measured for all genotypes with and without optical stimulation (Fig. S5). Although the feeding of ATR alone could induce small detachments (<5 µm), only sr>Tg RNAi myofibrils in females were frequently observed pulling away from the cuticle with sparse βPS integrin remaining (Fig. 4G,H). These data together suggest two important conclusions. First, Tg protein is required for IFM attachment during pupal development and after adult eclosion, but only in response to increased physical force. Second, Tg crosslinking activity is only required during pupal myogenesis, but indispensable for maintenance of the adult MTJ.
Conclusions and limitations
Our sensitized screening approach was designed to uncover genes that exhibit early lethality in standard zygotic screens or might have redundant functions in MTJ formation and/or stability. Tg likely exhibits both cuticle-independent and cuticle-dependent functions. Smaller MTJs are present in sr>Tg RNAi embryos even though Tg is detected in the developing embryo prior to cuticle deposition. Tg crosslinking activity likely provides additional rigidity to counteract increased forces generated upon IFM engagement during flight. We cannot rule out scaffolding functions as there is no evidence for crosslinking activity in the mouse MTJ even though Tg protein is present (Jacobson et al., 2020). Another curious, but unexplained observation in response to tension, is the heightened occurrence of IFM detachment in males during pupal development, whereas females showed MTJ instabilty as adults. Possible factors include sex-specific differences in Tg levels, sr-Gal4 expression, or yet, undefined developmental versus adult-specific differences. Future experiments should be directed towards clarifying these unknowns.
MATERIALS AND METHODS
Drosophila stocks
Fly stocks were maintained on Bloomington Drosophila Stock Center (BDSC) Cornmeal Food (https://bdsc.indiana.edu/information/recipes/bloomfood.html) at 25°C unless otherwise specified. All UAS-RNAi lines used for screening were acquired from either the BDSC, the Vienna Drosophila Resource Center (VDRC), or the Fly Stocks of National Institute of Genetics (NIG-FLY) and are listed in Table S1. w^1118^ or y,w was used as controls. daughterless (da)-Gal4 (BL55850) was used for ubiquitous tissue expression and tendon cell-specific expression was accomplished using stripe (sr)-Gal4 (BL26663). UAS-Dcr2 (BL24648) was recombined with sr-Gal4 and used to enhance RNAi phenotypes. UAS-mCherry.NLS (BL38424) was crossed to da-Gal4 or sr-Gal4 to serve as a Gal4/UAS control. The sr-Gal4, UAS-CD8_GFP line was a gift from Talia Volk (Department of Molecular Genetics, Weizmann Institute of Science, Israel). Additional stocks from the BDSC include: *fon^Δ24^/*CyO (BL43644), *tig^x^/*CyO (BL8796), Mhc-GFP (BL38460), UAS-Toll::Venus (BL30898), UAS-Toll6::GFP (BL92993), UAS-Col4a1::GFP (BL96361), RapGAP1-LexA::GAD (BL81553), and LexAop2-CsChrimson.tdTomato (BL82183). The *Mhc[S1]/*Cyo allele was obtained from the Littleton laboratory (Department of Biology, Massachusetts Institute of Technology, USA; Montana and Littleton, 2004) and UAS-Tg.A was a gift from Shun-ichiro Kawabata (Department of Biology, Kyushu University, Japan).
Sensitized-background screen
To construct sensitized-background fly lines, fon^Δ24^/CyO or tig^X^/CyO alleles were recombined with Mhc–GFP on the second chromosome and further combined with da-Gal4 on the third chromosome to generate Mhc-GFP, fon^Δ24^/CyO; da-Gal4/da-Gal4 or Mhc-GFP, tig^X^/CyO; da-Gal4/da-Gal4. The non-sensitized Mhc-GFP; da-Gal4/da-Gal4 stable stock was also generated. A total of 8–10 virgin females from the fon- or *tig-*sensitized fly stocks were mated with 6–8 males from each candidate UAS-RNAi fly line listed in Table S1 and incubated at 29°C. Parents from the original cross were transferred to a new vial for continual screening (n>50 larva for each vial). Vials were monitored each day throughout development by two independent laboratory members using a Leica M165FC fluorescent stereoscope to screen for detached or missing muscles. UAS-RNAi candidates that showed MTJ phenotypes in the primary screen were set up again as described above (n>100 larva for each genetic cross). In this secondary screen, the matched non-sensitized genotype (i.e. da-Gal4 only) were screened at the same time as each sensitized-background genotype (i.e. fon^Δ24^+/−; da-Gal4 or tig^X^+/−; da-Gal4). At this time, wandering L3 progeny showing detached or missing muscles were selected from vials, heat-killed with 60°C deionized water and immediately imaged under a fluorescent stereomicroscope. The larvae were rotated to take ventral, dorsal and lateral images of the musculature to document any missing and/or detached muscles (Fig. S1).
qRT-PCR
y,w control or da>Tg RNAi larvae were shifted from 25°C to 31°C at the late embryo or early L1 stage. For each data point, RNA was prepared from three wandering L3 larvae of each genotype and purified using the QuantaBio Extracta Plus RNA kit (QuantaBio, Beverly, MA). cDNA was synthesized from 2000 ng of RNA using qScript XLT cDNA SuperMix kit (Quanta Biosciences, Beverly, MA). For the qPCR, 1:20 dilutions of the cDNA were combined with PowerUp SYBR Green Master Mix (Thermo Fisher Scientific, Waltham, MA, USA). The following primers were used (final concentration=0.5 µM): Tg forward 5′-CGGCGATTGGAAGTGGTGTA, reverse 5′-ATCGACCTTCAAAACGCCCA, and rp49 forward 5′-GCCCAAGGGTATCGACAACA, reverse 5′-GCGCTTGTTCGATCCGTAAC. All primers were synthesized by Integrated DNA Technologies (IDT, Stokie, IL, USA). Quantitative transcript levels were obtained using a QuantStudio 3 instrument with QuantStudio Design and Analysis software (ThermoFisher, Waltham, MA). Normalize fold changes were calculated using the 2^−ΔΔCt^ method (Livak and Schmittgen, 2001) and graphed as mean±s.d.
Embryo fixation and immunofluorescence
Male and virgin female flies of the appropriate genotypes were allowed to mate overnight at 25°C. The following day, all flies were transferred to FlyStuff embryo collection cages (#59-105; Genesee Scientific, San Diego, CA, USA) and embryos were collected onto a 35 mm Petri dish containing apple juice-agar medium at 25°C in 60% humidity. Fresh yeast paste was added to the apple juice-agar plates that were exchanged every 24 h and stored for no longer than 48 h at 4°C before fixation. Embryos were transferred from the medium plates to a 70 μm mesh collection basket using deionized water and a paintbrush. Collected embryos were rinsed with deionized water to remove contaminants before being immersed into a fresh dilution of 50% Clorox bleach (sodium hypochlorite) for 3 min. Dechorionated embryos were washed thoroughly with a 0.7% NaCl and 0.04% Triton X-100 solution, then transferred to a small glass vial containing a 1:1 ratio of heptane and 4% formaldehyde in PEM buffer (0.1 M PIPES pH 8.0, 2 mM MgSO_4_ and 1 mM EGTA). Embryos in the fixation solution were shaken at maximum speed (450 rpm) on a platform shaker for 12 min. Afterward, the fixative layer was removed, and an equal volume of methanol was added to the glass vial. Embryos were shaken vigorously for 1 min and settled embryos were transferred from the glass vial to a microcentrifuge tube using a glass Pasteur pipette. Fixed embryos were washed three times with methanol before another three washes with PBS with 0.1% Triton X-100 (PBT). Embryos were blocked at room temperature for 20 min in a blocking solution containing 5% normal goat serum in PBT, then incubated and rotated overnight at 4°C with primary antibody dilutions. Primary antibodies included mouse anti-βPS integrin CF.6G11 (1:50; Developmental Studies Hybridoma Bank, DSHB, Iowa City, IA), rat anti-tropomyosin (TM) MAC141 (1:200; Babraham Institute, Cambridge, UK), mouse anti-HA #2367 (1:50; Cell Signaling Technology, Danvers, MA), or rabbit anti-Tg #DZ41557 (1:100; BosterBio, Pleasanton, CA). The Tg antibody was generated against amino acids 494–776 of isoform Tg-PA. After incubation, the primary antibody solution was removed, and the embryos were subjected to three 5-min PBT washes before blocking at room temperature for 20 min in a PBT solution with 5% normal goat serum. Embryos were incubated and rotated at room temperature for 2 h with goat anti-mouse-IgG conjugated to Alexa Fluor 405, goat anti-rabbit IgG conjugated to Alexa Fluor 488 or goat anti-rat Alexa Fluor 594 (all 1:400; Thermo Fisher Scientific). Three 5-min PBT washes followed incubation. Embryos were allowed to settle, and PBT was removed from the microcentrifuge tube. Fluorescent mounting medium was added to the embryos and gently mixed using a pipette. A pipette was again used to transfer embryos and mounting media onto glass slides and topped with a cover slip. Imaging was performed on an LSM Zeiss 700.
Larva dissection and immunofluorescence
Wandering L3 larvae were heat-killed, pinned onto a Sylgard plate dorsal side up near the mouth hooks, and covered with a drop of PBS. After removing internal organs, the PBS was removed, and fillets were fixed in 4% formaldehyde for 15 min. The muscle carcasses were washed three times for 5 min each in PBT, blocked in 5% normal goat serum in PBT, and stained overnight at 4°C with the appropriate primary antibodies: mouse anti-βPS integrin CF.6G11 (1:10; DSHB, Iowa City, IA, USA) or mouse anti-HA #2367 (1:50; Cell Signaling Technology, Danvers, MA, USA). Tissues were incubated for 2 h at room temperature using the following secondary antibodies: goat anti-mouse-IgG or goat anti-rabbit-IgG conjugated to Alexa Fluor 488 or 594 (1:400; Thermo Fisher Scientific). To enhance GFP signal, rabbit anti-GFP conjugated to Alexa Fluor 488 (1:400; cat. no A-21311, Invitrogen) was included in the secondary antibody incubation. Muscle tissue was visualized using Alexa Fluor 488 or 594 conjugated to phalloidin (1:400, Invitrogen) to label F-actin. Maximum intensity confocal images were collected using an LSM Zeiss 700.
Adult IFM dissection and immunofluorescence
Male or female adult flies were separately anesthetized on a CO_2_ plate. The head, wings and abdomen were removed, leaving the thorax intact. The thoraces were fixed in 4% formaldehyde for 2 h at room temperature, protected from light. Following fixation, the samples were washed several times in PBT and transferred to a dissection slide. Each thorax was bisected longitudinally to expose the muscle fibers. Dissected thorax samples were washed three times in PBT for 10 min each. To block nonspecific binding, tissues were incubated in 5% normal goat serum diluted in PBT for 30 min. Samples were then incubated with anti-Tg and mouse anti-βPS integrin as described above at room temperature for 2 h. After primary antibody incubation, samples were washed three times in PBT for 10 min each and re-blocked with 5% normal goat serum for 30 min. For fluorescence detection, samples were incubated with anti-mouse-IgG or anti-rabbit-IgG Alexa Fluor 488 or Alexa Fluor 594 secondary antibodies (1:400, Thermo Fisher Scientific). F-actin was labeled with phalloidin (Alexa Fluor 488, 594 or 647; 1:400, Molecular Probes), and nuclei were counterstained with Hoechst 33342 (1:400, Invitrogen). Fluorescence images were captured using a Zeiss 700 confocal microscope.
Mhc[S1] hypercontraction experiments
The temperature-sensitive Mhc[S1]/CyO allele was combined with sr-Gal4 and the stock was maintained at 18°C. All control or experimental fly crosses were reared at 30°C. All adult flies were scored for thorax phenotypes (as WT, mild or severe). Dissected adult flies (described above) were labeled with phalloidin–Alexa Fluor 488 (1:400, Molecular Probes).
Western blotting
Western blotting was used to confirm Tg antibody specificity upon Tg RNAi knockdown (Fig. S3C) or UAS-Tg WT or UAS-Tg C330S transgene expression (Fig. 3G). Crosses to obtain progeny of the genotypes da>mCh.NLS or da>Tg RNAi were reared at 25°C and then shifted to 31°C for 2 days after eclosion before sample processing (n=3 for each genotype). For each biological replicate, three adult males were transferred to a tube containing 100 µl of 1× LiCor sample buffer (4× formulation is 125 mM Tris-HCl, pH=6.8, 50% glycerol, 4% SDS and 0.2% Orange G) (LI-COR, Lincoln, NE, USA), boiled for 3 min, homogenized, and boiled for an additional 10 min. The samples were centrifuged for 2 min at 21,000 g and loaded onto a 4–12% Bis-Tris polyacrylamide gel (Invitrogen, Thermo Fisher Scientific). Gels were run with standard MOPS running buffer in a Bio-Rad Mini-PROTEAN Tetra Cell (Hercules, CA) and transfers were performed using a BioRad Transblot Turbo System (Hercules, CA) with Bjerrum Schafer-Nielsen SDS transfer buffer to a Millipore Immobilon-F 0.45 µm PVDF membrane (Thermo Fisher Scientific). Total protein was visualized using the Revert 700 Total Protein Stain (LI-COR). The blots were blocked in LI-COR blocking buffer (LI-COR) and incubated overnight in rabbit anti-Tg (1:1000, as above) diluted in blocking buffer. Blots were washed with standard Tris-buffered saline (TBS)+0.1% SDS and incubated for 1 h at room temperature (∼22°C) with IRDye 800 CW goat anti-rabbit-IgG (1:15,000, LI-COR). After washing, the Tg RNAi blot was imaged using the LI-COR Odyssey XF (LI-COR) and normalized fold change was determined using the Analyze Gels function in ImageJ with total protein as the loading control. Graphs and statistics were generated with GraphPad Prism 10 software. Uncropped images of blots presented in this work are shown in Fig. S6.
Creation of Tg transgenic lines
Genscript synthesized cDNAs corresponding to the coding region of isoform A (FBpp0079155) for wild-type Tg (Tg WT) or mutant Tg (Tg C330S) followed by a 3× HA tag. Each cDNA was independently subcloned into the pUAST_attB vector using EcoRI/XbaI sites and sequence verified. Plasmid DNA was purified using the Qiagen Maxi kit (Hilden, Germany) and sent to Rainbow Transgenics. After generation of transgenic flies at the PhiC31 landing site (BL9736-53B2), all lines were balanced over the w; Sco/Cyo,Tb balancer stock (BL36335). Western blotting was performed as above using three whole larvae per sample to confirm equivalent levels of protein expression (Fig. 3G). LI-COR Empiria software (LI-COR) was used to calculate normalized fold change for Tg with total protein as the loading control.
Optogenetic experiments
RapGAP1-LexA::GAD; sr-Gal4/TM6, Tb virgin females were mated with: (1) UAS-TG RNAi; LexAop2-CsCrimson.tdTomato/TM6, Sb, Tb males, (2) UAS-TG WT::3xHA; LexAop2-CsCrimson.tdTomato/TM6, Sb, Tb, or (3) UAS-TG C330S::3xHA; LexAop2-CsCrimson.tdTomato/TM6, Sb, Tb males and reared at 18°C. The tdTomato(+) adults lacking the Sb and Tb markers were collected after 1 week and incubated at 30°C for 4 days to allow for transgene expression. The flies were then transferred to new fly food vials supplemented with 100 µM all trans-retinal (ATR) in the food and a Kimwipe soaked with 100 µM ATR dissolved in a 20% sucrose and 6% dextrose solution for 24 h at 30°C. ATR-fed flies were then exposed to 655 nm light (5 s on, 5 s off) for 2 h. LED-exposed flies were processed for dissection and staining according to the protocol outlined for adult IFM preparation.
Quantification and statistics
All graphs and statistical analysis were performed in GraphPad Prism 10.4.1. See Table S2 for a summary of sample size, type of statistical test and P-values.
MAS length quantification in the embryo
The length of four MASs spanning the ventral longitudinal muscle group (VL1–VL4) in hemisegments A1–A4 in each embryo were quantified using the ‘Analyze>Measure’ function in ImageJ. Specifically, signal corresponding to the entire length of either the βPS integrin or the TM signal at each MAS were measured. These four MAS measurements were then averaged and plotted as a single data point for each embryo. Data are presented as violin plots with individual data points visible. n≥12 for each genotype.
Larval muscle detachment
For all experiments, muscles in hemisegments A1–A4 were characterized as detached if the muscles were rounded and clearly separated at an attachment site or if the muscles were missing from the fillet. Direct muscle subsets scored were dorsal oblique (DO) 1–2, lateral transverse (LT) 1–4, and ventral oblique (VO) 4–6. Indirect muscle subsets scored were dorsal acute (DA) 1–3, longitudinal lateral (LL) 1, oblique lateral (LO) 1 and ventral longitudinal (VL 1–4). Percentage detachment was calculated as the number of individual detached muscles divided by the total number of muscles quantified per fillet and is presented as a single data point. Aggregate data are presented as scatter plots with individual data points. n>10≤44 for each genotype.
Indented notum phenotype
Notum phenotypes were separated into three categories: WT (no thoracic indentations), mild (a single thoracic indentation towards the anterior region of the notum) or severe (a single thoracic indentation located about halfway down the notum region). Males and females were quantified separately. Data are presented as stacked bar graphs. n≥90 for each genotype.
IFM detachment quantification
For each genotype, ten images of individual dissected thoraces were acquired with a 63× objective with 1× zoom. ImageJ was used for quantification with the InterEdgeDistance macro. For each image, user-defined lines were drawn along the F-actin edge and the corresponding internal cuticle edge (marked by βPS integrin) to measure the distance between the two structures. The mean distance output by the macro was used as the value for the final quantification plot. Males and females were quantified separately. n=10 for each sample.
Supplementary Material
10.1242/joces.264299_sup1Supplementary information
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1Anokhin, B. A., Dean, W. L., Smith, K. A., Flick, M. J., Ariëns, R. A. S., Philippou, H. and Maurer, M. C. (2020). Proteolytic and nonproteolytic activation mechanisms result in conformationally and functionally different forms of coagulation factor XIII A. FEBS J. 287, 452-464. 10.1111/febs.1504031407850 PMC 7269131 · doi ↗ · pubmed ↗
- 2Balklava, Z., Verderio, E., Collighan, R., Gross, S., Adams, J. and Griffin, M. (2002). Analysis of tissue transglutaminase function in the migration of Swiss 3T 3 fibroblasts: the active-state conformation of the enzyme does not affect cell motility but is important for its secretion. J. Biol. Chem. 277, 16567-16575. 10.1074/jbc.M 10983620011867617 · doi ↗ · pubmed ↗
- 3Bökel, C., Prokop, A. and Brown, N. H. (2005). Papillote and Piopio: Drosophila ZP-domain proteins required for cell adhesion to the apical extracellular matrix and microtubule organization. J. Cell Sci. 118, 633-642. 10.1242/jcs.0161915657084 · doi ↗ · pubmed ↗
- 4Boley, N., Wan, K. H., Bickel, P. J. and Celniker, S. E. (2014). Navigating and mining mod ENCODE data. Methods 68, 38-47. 10.1016/j.ymeth.2014.03.00724636835 PMC 4857704 · doi ↗ · pubmed ↗
- 5Bunch, T. A., Graner, M. W., Fessler, L. I., Fessler, J. H., Schneider, K. D., Kerschen, A., Choy, L. P., Burgess, B. W. and Brower, D. L. (1998). The PS 2 integrin ligand tiggrin is required for proper muscle function in Drosophila. Development 125, 1679-1689. 10.1242/dev.125.9.16799521906 · doi ↗ · pubmed ↗
- 6Casas-Vila, N., Bluhm, A., Sayols, S., Dinges, N., Dejung, M., Altenhein, T., Kappei, D., Altenhein, B., Roignant, J. Y. and Butter, F. (2017). The developmental proteome of. Genome Res. 27, 1273-1285. 10.1101/gr.213694.11628381612 PMC 5495078 · doi ↗ · pubmed ↗
- 7Charvet, B., Ruggiero, F. and Le Guellec, D. (2012). The development of the myotendinous junction. A review. Muscles Ligaments Tendons J. 2, 53-63.23738275 PMC 3666507 · pubmed ↗
- 8Chihara, C. J., Silvert, D. J. and Fristrom, J. W. (1982). The cuticle proteins of Drosophila melanogaster: stage specificity. Dev. Biol. 89, 379-388. 10.1016/0012-1606(82)90326-86173274 · doi ↗ · pubmed ↗