An RNAi based screen in Drosophila larvae identifies fascin as a regulator of myoblast fusion and myotendinous junction structure
A strength of Drosophila as a model system is its utility as a tool to screen for novel regulators of various functional and developmental processes. However, the utility of Drosophila as a screening tool is dependent on the speed and simplicity of the assay used.
Here, we use larval locomotion as an assay to identify novel regulators of skeletal muscle function. We combined this assay with muscle-specific depletion of 82 genes to identify genes that impact muscle function by their expression in muscle cells. The data from the screen were supported with characterization of the muscle pattern in embryos and larvae that had disrupted expression of the strongest hit from the screen.
With this assay, we showed that 12/82 tested genes regulate muscle function. Intriguingly, the disruption of five genes caused an increase in muscle function, illustrating that mechanisms that reduce muscle function exist and that the larval locomotion assay is sufficiently quantitative to identify conditions that both increase and decrease muscle function. We extended the data from this screen and tested the mechanism by which the strongest hit, fascin, impacted muscle function. Compared to controls, animals in which fascin expression was disrupted with either a mutant allele or muscle-specific expression of RNAi had fewer muscles, smaller muscles, muscles with fewer nuclei, and muscles with disrupted myotendinous junctions. However, expression of RNAi against fascin only after the muscle had finished embryonic development did not recapitulate any of these phenotypes.
These data suggest that muscle function is reduced due to impaired myoblast fusion, muscle growth, and muscle attachment. Together, these data demonstrate the utility of Drosophila larval locomotion as an assay for the identification of novel regulators of muscle development and implicate fascin as necessary for embryonic muscle development.
KeywordsNuclear movement Drosophila Myoblast fusion Myotendinous junction Fascin Myogenesis
Skeletal muscle has a distinctive architecture that is generated by a unique set of developmental phases. Making the myofiber syncytium requires the fusion of mononucleated myoblasts. In both Drosophila and mammalian systems, individual myoblasts invade growing myotubes and deposit their nucleus into the common cytoplasm to drive myotube growth [1, 2]. Myoblast fusion is an actin-dependent process that is reminiscent of a cancer cell invading a tissue during metastasis . The mononucleated myoblast extends a protrusive invadapodia-like structure that makes possible the penetration of the myotube and the mixing of cytoplasm. Many factors and signaling pathways that regulate myoblast fusion have been identified [2, 4, 5]. However, many of the genes necessary for invadapodia-like structures in other contexts have not yet been implicated in myoblast fusion, suggesting that additional regulators remain to be identified. One glaring omission from the categories of proteins that have been identified as regulators of myoblast fusion is proteins that stabilize filopodia. Invadapodia are filopodia-like structures [6, 7], and they require several factors that are known to stabilize filopodia. Furthermore, although loss-of-function data is lacking, dominant negative mutants of the formin Diaphanous inhibit myoblast fusion and may suggest that filopodia are essential for myoblast fusion [8, 9]. Thus, it is likely that additional factors that have been identified as capable of stabilizing filopodia for the purpose of protrusion and invasion in other contexts contribute to myoblast fusion.
Beyond myoblast fusion, there are several features of muscle development that either require, or have been hypothesized to require, precise regulation of the actin cytoskeleton, including the positioning of nuclei and the development of the myotendinous junction (MTJ). To date, evidence for actin-dependent nuclear movement in muscle is restricted to the squeezing of nuclei to the periphery of the muscle [10, 11], although it has been proposed that actin may contribute to the movement of nuclei along the length of the muscle [12, 13]. The role of actin in MTJ development is more established. MTJs are integrin-based adhesions that transmit force from the muscle to the skeleton . The initial formation involves extension of filopodia-like structures from the muscle cell that interact with the tendon cell before forming a stable and somewhat rigid attachment that enables effective force transmission [15, 16]. All of these processes require linear actin-cables. The similarity in the actin-based structures suggests that the same molecular components may contribute to each of these aspects of muscle development. Therefore, it is critical to determine how newly identified genes and proteins contribute to each process.
Because the developmental path and final architecture of muscle cells is conserved from Drosophila to humans, flies provide a genetically tractable and inexpensive model for the identification of genes that are necessary for muscle development. Indeed, many screens for regulators of muscle development have been completed. Researchers have used adult locomotion  and embryonic muscle structure  as indicators of muscle development. Although these strategies have proven effective, they each have drawbacks. Analysis of embryonic muscle structure is labor intensive and requires significant expertise in muscle biology. Analysis of adult locomotion is limited because the disruption of many genes is lethal during pupation. Therefore, we have developed a simple assay for muscle function based on the larval locomotion. We have used this assay to screen for novel regulators of muscle function and identified fascin as one such regulator. Subsequent cell biological analysis implicates fascin as a regulator of myoblast fusion and MTJ structure.
All stocks were grown under standard conditions at 25 °C. The Fascin sn28 allele was a generous gift from Tina Tootle (University of Iowa). All UAS-RNAi Drosophila lines were purchased from Bloomington Drosophila Stock Center. UAS-RNAi constructs were driven specifically in the mesoderm using twist-GAL4, which drives expression in the mesoderm from stage 10 of embryonic development through stage 13 of embryonic development, DMef2-GAL4 that drives expression in the muscles from stage 12 through adulthood, or MHC-GAL4 which drives expression in muscle from stage 17 of embryonic development through adulthood.
Larval locomotion assay
We performed a modified version of the previously used assay that has been used to measure larval locomotion in individual larvae . Virgins expressing DMef2-GAL4 were mixed with males that carried the UAS-RNAi for 1 h in a vial with standard Drosophila food. After 1 h, adults were moved to a new vial and the embryos laid during the 1 h period were aged for 5 days until they were third-instar larvae (L3). Larvae were then floated from the food by the addition of 15% sucrose. Using a paintbrush, larvae were moved to a plate with wet yeast. After all of the genotypes had been collected, 10 larvae of each RNAi were moved to an arena that consisted of 3% agarose dyed black with standard food color poured over the top of a 96-well plate cover. Movement of larvae toward a stick dipped in ethyl butyrate was captured using an iPhone mounted above the arena. The speed of each larva was then analyzed using ImageJ.
Stage 16 embryos were picked and placed on an agarose plate. The selected embryos were incubated at 20 °C overnight. After incubation, the number of unhatched embryos were counted to determine embryonic lethality and the L1 larvae from the hatched embryos were transferred to a vial of standard fly food. Larvae were incubated at 25 °C until larvae began to pupate and eclose. Adult flies were counted to determine pupal lethality and removed from the vials of standard fly food upon eclosure. After 1 week of no eclosures, the number of pupal cases were counted to determine larval lethality.
Preparation of embryos
Embryos were collected at 25 °C and were dechorionated by submersion in 50% bleach for 4 min. Embryos were then fixed in a solution of equal parts heptane and 10% formalin (Sigma, Product # HT501128). Fixation lasted for 20 min during which time the embryos were placed on an orbital shaker that rotated at a rate of 250/min. Following fixation, the formalin and heptane were removed and replaced with a solution of equal parts methanol and heptane. The embryos were vortexed for 1 min to devitellinize the embryos. Embryos were stored in methanol at − 20 °C until immunostaining.
Preparation of larvae
Dissection of larvae was carried out as previously described  with minor modifications. The primary difference being that the buffer used was modified to increase the preservation of muscle structure. The modified dissection buffer was 100 mM PIPES (Sigma-Aldrich, P6757), 115 mM d-sucrose (Fisher Scientific, BP220-1), 5 mM trehalose (Acros Organics, 182550250), 10 mM sodium bicarbonate (Fisher Scientific, BP328-500), 75 mM potassium chloride (Fisher Scientific, P333-500), 4 mM magnesium chloride (Sigma-Aldrich, M1028), and 1 mM EGTA (Fisher Scientific, 28-071-G). Larvae were then fixed with 10% formalin (Sigma-Aldrich, HT501128) for 20 min. Briefly, dissection involved lateral cuts at the anterior and posterior end of the larva that encompassed 70% of larval circumference. These were followed by a longitudinal cut through the dorsal surface of the animal that connected the two lateral cuts. The intestines, other internal tissues, and neurons were then removed, and the flaps of tissue composed of epidermis and muscle were pinned down and fixed. For fixation, larvae were incubated in a solution of 10% formalin in PBS for 20 min.
Staining of embryos and larvae was identical. Antibodies were used at the following dilutions: rabbit anti-dsRed (1:400, Clontech 632496), rat anti-tropomyosin (1:200, Abcam ab50567), mouse anti-GFP (1:50, Developmental Studies Hybridoma Bank GFP-G1), mouse anti-integrin βPS (1:50, Developmental Studies Hybridoma Bank CF.6G11), mouse anti-Fascin (1:25, Developmental Studies Hybridoma Bank sn 7C, generous gift from Tina Tootle), and mouse anti-αTubulin (1:200, Sigma-Aldrich T6199). Conjugated fluorescent secondary antibodies used were Alexa Fluor 555 donkey-anti-rabbit (1:200), Alexa Fluor 488 donkey-anti-rat (1:200), and Alexa Fluor 647 donkey-anti-mouse (1:200) (all Life Technologies) and Alexa Fluor 488 donkey-anti-mouse (1:200, Life Technologies). Furthermore, Acti-stain 555 phalloidin (1:400, Cytoskeleton PHDH1-A) and Hoechst 33342 (1 μg/ml) were used on larvae. Embryos and larvae were mounted in ProLong Gold (Life Technologies, P36930).
All microscopy was performed on a Zeiss LSM700 with an oil-immersion × 40 APOCHROMAT, 1.4 NA objective. All images of embryos, and the image of fascin localization in Fig. 2, were acquired with a 1.0× optical zoom. All other larval images were acquired with a 0.5× optical zoom. Image tiling was necessary to acquire images of the full larval muscles and was completed using the tiling function in the ZEN software that controls the microscope.
All statistics were performed using GraphPad Prism. All data sets were compared to appropriate controls by a Student’s t test. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.
Analysis of nuclear position in larvae
Although the field has traditionally measured the distance between nuclei [18, 20, 21], this measurement does not account for changes in muscle size and nuclear number. We have therefore modified this measurement to determine how evenly nuclei are spaced within a muscle . First, the area and length of the muscle was measured. Next, the position and number of nuclei is calculated by using the multipoint tool in ImageJ to place a point in the center of each nucleus. The position of each nucleus is used to calculate the actual internuclear distance. The maximal internuclear distance is determined by taking the square root of the muscle area divided by the nuclear number. This value represents the distance between nuclei, if internuclear distance was fully maximized. The ratio between the actual internuclear distance and the maximal internuclear distance ratio was then used to determine how evenly nuclei were distributed. This method normalizes the internuclear distance to both nuclear count and muscle area which leads to a more representative means of comparison between muscles, larvae, and genotypes. All viable (not torn) ventral longitudinal (VL3) muscles were measured from each larva. At least four larvae from one experiment were measured for each genotype. Statistical analysis was performed with Prism 4.0 (GraphPad). Student’s t test was used to assess the statistical significance of differences in measurements between experimental genotypes and controls.
Analysis of nuclear position in embryos
The position of nuclei was measured in stage 16 embryos. This is the latest stage before cuticle development blocks the ability to perform immunofluorescence microscopy. Embryos were staged based primarily on gut morphology as previously described . At stage 16, the nuclei are reliably positioned adjacent to the muscle ends, and disruptions in this positioning can be easily determined as previously described [21, 23, 24]. Images, acquired as described above, were processed as maximum intensity projections of confocal z-stacks using ImageJ. The position of the nuclei was determined by using the line function in ImageJ to measure the distance between either the dorsal end of the muscle and the nearest nucleus or the ventral end of the muscle and the nearest nucleus. All four LT muscles were measured in four hemisegments from each embryo. At least 20 embryos from at least two independent experiments were measured for each genotype with each data point representing the average for all muscles measured within a single embryo.
Analysis of nuclear count in embryos
The number of nuclei was counted using apRed fluorescence in the 4 lateral transverse (LT) muscles of stage 17 embryos, once nuclei had separated from their clusters and into easily distinguished individual nuclei. Only hemisegments in which there were four lines of nuclei, corresponding to 4 LT muscles, were counted. Data points indicate the number of apRed positive nuclei within a single hemisegment. A total of at least 49 hemisegments from at least 19 individual embryos were counted.
Analysis of nuclear count in larvae
The number of nuclei in the VL3 muscles were counted using Hoechst staining. Data points indicate the number of nuclei within an individual muscle. Careful analysis of z-stacks was used to ensure that nuclei were in the muscle that they were attributed to.
Analysis of muscle attachment in embryos
Integrin accumulation at the myotendinous junction of stage 16 embryos were measured in dorsal muscle 2 (DO2). Z-stack projection images that extended through the entire MTJ were used for these data. Integrin fluorescence intensity line scans were measured within a 10 ×10 μm box at the segment border using plot profile function in ImageJ. The width of the fluorescence peak composed of fluorescence intensities greater than 25% of the maximum intensity (75% fluorescence intensity peak) was measured. Each data points indicate the width of the 75% fluorescence intensity peak for a single myotendinous junction. A total of at least 50 MTJs from 20 different embryos were analyzed.
Analysis of muscle size in larvae
The area of the VL3 muscles were measured using the polygon selection tool in ImageJ as previously described . Data points indicate the size of an individual muscle.
Analysis of general muscle architecture
Qualitative muscle phenotype analysis was completed on embryos of each genotype. All analysis was based on the immunofluorescence staining pattern of Tropomyosin in stage 16 embryos. The frequency of the following phenotypes were scored: the number of free myoblasts in an embryo that indicated a defect in myoblast fusion (small, unfused circles stained by anti-tropomyosin) and the number of muscles in each hemisegment (> 4 defined as extra muscles, < 4 defined as missing muscles) indicating gross abnormalities in the specification of muscle tissue. For analysis of the unfused myoblasts, embryos were grouped into bins with a width of 5 and the first bin centered on zero.
Fascin is necessary for muscle function
For the proof-of-concept screen, we expressed RNAi against 82 genes (Additional file 1: Table S1). The selected genes included those expected to impact muscle function (e.g., Dystrophin, Msp300) and many for which we did not have a prediction. The speed of locomotion in larvae that expressed each RNAi was compared to control larvae in which DMef2-GAL4 drove the expression of mCherry RNAi. The disruption of 12/82 genes significantly altered larval locomotion compared to control larvae, indicating that these 12 genes regulate muscle function. Of these 12 genes, disruption of 5 caused larvae to move faster compared to controls, and the disruption of 7 caused the larvae to move more slowly than controls (Fig. 1b). RNAi directed against the expression of singed (sn) which encodes for the actin-binding protein fascin [26, 27] caused the greatest decrease in larval locomotion. Therefore, we investigated the impact that fascin depletion had on muscle structure to identify the mechanism by which fascin regulates muscle function.
Fascin is localized to the nucleus in Drosophila muscle
An emergent regulator of muscle function is the position of the many myonuclei within a single cytoplasm [12, 13, 29]. Based on the localization of fascin to the nucleus, we hypothesized that fascin may regulate nuclear movement during muscle development. Consistent with this hypothesis, fascin interacts with the nuclear envelope protein Nesprin-2 and regulates nuclear movement in migrating fibroblasts  and is necessary for the positioning of nuclei in developing Drosophila oocytes . To determine whether nuclear position was affected in Drosophila muscle, we crossed apRed, a marker for the nuclei in the lateral transverse (LT) muscles  into the sn 28 Drosophila and measured the position of nuclei as previously described . In sn 28 mutants, nuclei were closer to the ventral end of the muscle compared to controls (Fig. 2b–d). To determine whether the effect on nuclear position was muscle autonomous, we transiently expressed RNAi specifically in the mesoderm during the early stages of muscle development using Twist-GAL4 or in a more sustained manner using DMef2-GAL4 and measured the position of the nuclei. Muscle-specific depletion of fascin had no impact on nuclear position (Fig. 2e–h). Together, these data indicate that although nuclei are closer to the muscle end in fascin mutants, this is not regulated by fascin expressed in muscle during embryonic muscle development in Drosophila.
Fascin regulates myoblast fusion
The typical hemisegment from a control embryo has four LT muscles. Seventy percent of control embryos had four LT muscles in every hemisegment, and 30% of controls had at least one hemisegment with greater than four LT muscles. In sn 28 mutants, 70% of embryos had four LT muscles in every hemisegment and 19% of embryos had at least one hemisegment with greater than four LT muscles. Additionally, 19% of sn 28 embryos had at least one hemisegment with fewer than four LT muscles (Fig. 3b), and were therefore missing LT muscles. Eight percent of embryos had at least one hemisegment with an extra muscle and one hemisegment with a missing muscle and were thus counted in both categories, pushing the total embryos over 100%. Because the absence of muscles can indicate a defect in myoblast fusion, we counted the number of nuclei that were incorporated into the LT muscles per hemisegment. This number was reduced from a mean of 26 in the controls to a mean of 22 in the sn 28 mutants (Fig. 3c). Consistent with this, there was an increase in unfused myoblasts. In controls, the median number of free myoblasts per embryo was 1, and that number increased to 7.5 in sn 28 mutants (Fig. 3d). Additionally, 70% of control embryos had two or fewer identifiable unfused myoblasts whereas 75% sn 28 mutant embryos had at least three unfused myoblast and 50% of sn 28 mutant embryos eight or more unfused myoblasts. Based on the missing muscles and the abundance of unfused myoblasts in mutant embryos, we tested the viability of the sn 28 mutants and found that there was significant lethality during both the embryonic and larval stages (Fig. 3e). These data indicate that fascin regulates myoblast fusion.
To determine whether these phenotypes were muscle autonomous, we used the GAL4/UAS system to deplete fascin specifically from the developing mesoderm and muscle of the Drosophila embryo. The expression of a UAS-sn RNAi (fascin RNAi) was driven with each of three GAL4 drivers. Twist-GAL4 was used to drive RNAi expression in the early mesoderm, DMef2-Gal4 was used to drive RNAi expression slightly later in muscle with sustained expression throughout development, and MHC-GAL4 was used to drive RNAi expression beginning at the final stage of embryonic development and continuing throughout development. We then examined the general muscle structure in stage 16 embryos as we had done for sn 28 mutant embryos. There were no defects in muscle morphology when MHC-GAL4 was used to drive RNAi expression suggesting that fascin must be depleted early during development to have significant impact (Additional file 2: Figure S1).
Fascin-dependent fusion effects are evident in larvae
Fascin regulates muscle attachment
One of the many strengths of Drosophila as a model system is its utility as a tool to identify novel regulators of specific biological functions. This ability utilizes the immense genetic tools that are available and requires simple and fast assays to screen many mutants and/or RNAi lines. In this work, we adapted a published larval-tracking assay  to perform a proof-of-concept screen for muscle function. We identified 12 genes that regulate muscle function, either positively or negatively. We continued these experiments by examining the mechanism by which singed, Drosophila fascin, regulated muscle function because fascin-depletion had the strongest effect on muscle function.
We used a combination of mutant alleles and tissue-specific expression of RNAi against fascin to demonstrate that fascin regulates both myoblast fusion and the structure of the MTJ. Fascin is well-described as a protein that can bundle F-actin filaments and increase their strength, and the strength of actin-based cellular protrusions . Furthermore, by this mechanism, fascin contributes to cellular invasions associated with cancer metastasis [34, 35]. Myoblast fusion requires a similar organization of protrusive F-actin structures that invade the growing myotube. The most surprising aspect of the myoblast fusion data is the relatively minor effect that fascin has compared to other genes necessary for myoblast fusion [31, 36, 37, 38]. The reason for this is not clear. One possibility is that maternal loading provides sufficient fascin to facilitate the initial rounds of fusion. Alternatively, perhaps the final fusion events require greater protrusive force and only then does the function of fascin become critical.
The contribution of fascin to MTJ structure is consistent with previously published data. Fascin contributes to filopodia formation  and MTJ development is dependent on filopodia-like extensions. Furthermore, although the MTJ forms as a smooth attachment during pupation , the MTJ in the embryo is dynamic . Thus, perhaps fascin is continually necessary for the turnover and the integrity of the MTJ.
Perhaps the most intriguing feature of these data is the temporal separation of fascin-dependent myoblast fusion and fascin-dependent MTJ stability. This conclusion is based on our finding that the time and duration of fascin depletion determines the phenotype that will emerge. Transient depletion of fascin during early stages of muscle development disrupted myoblast fusion but not MTJ structure. Conversely, later, and sustained depletion of fascin affected MTJ structure, but not myoblast fusion. These data are important because they demonstrate that although both fusion and MTJ structure require fascin function, they are not codependent features of muscle development.
It is not clear whether either function is more critical than the other. Certainly, sustained depletion of fascin, which disrupts MTJ integrity has a greater effect on animal survival than does the transient depletion that disrupts fusion. However, this conclusion is limited because the impact that a small reduction in nuclear number has on muscle organization is not clear. Reduced nuclear numbers do correlate with reduced muscle size [40, 41] and therefore likely cause reduced muscle function. Data in embryos indicated that DMef2-GAL4-mediated expression of fascin RNAi only affected MTJs and would therefore allow us to isolate the impact of the MTJ versus the impact of myoblast fusion. However, we see that in larvae, there is a reduction in the number of nuclei per muscle. Because there is no repair of embryonic and larval muscles in Drosophila, we suspect that this reduction is based in muscle damage that may be linked to improper attachments and poor mechanical stability. However, further work is necessary to understand the mechanism by which nuclei are lost so that the impact of individual fascin-dependent functions can be determined.
In total, we have used a larval locomotion assay to identify novel regulators of muscle function. Furthermore, we have described that the strongest hit from the screen, fascin, contributes to muscle function by regulating myoblast fusion and MTJ structure. Both of these functions are consistent with known biochemical abilities for fascin and suggest that fascin has multiple essential functions for muscle development that are separated in time. More generally, this work outlines a new strategy for the identification of genes and pathways that can be manipulated to either increase or decrease muscle function.
The authors would like to thank Tina Tootle (University of Iowa) for providing the fascin sn28 allele and her advice. Drosophila stocks obtained from the Bloomington Drosophila Stock Center (NIHP400D018537) were used in this study. The integrin betaPS antibody developed by D. Brower at Harvard Medical School was obtained from the Developmental Studies Hybridoma Bank, created by the NICHD of the NIH and maintained at the University of Iowa, Department of Biology, Iowa City, IA 52242, USA.
This work was funded by grants from the American Heart Association to E.S.F. and by startup funds provided by Boston College to E.S.F.
Availability of data and materials
All data are presented in the manuscript. Raw data, if requested, will be provided by the corresponding author. Similarly, all reagents that are not commercially available will be made available upon request to the corresponding author.
JMC and TRM designed, completed, and analyzed the experiments and contributed to the writing of the manuscript. RM designed, completed, and analyzed the experiments related to larval locomotion. JM completed and analyzed the experiments. CHH and DS contributed to the development of the tracking assay. ESF designed the experiments, interpreted the results, and wrote the manuscript. All authors read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 10.D’Alessandro M, Hnia K, Gache V, Koch C, Gavriilidis C, Rodriguez D, et al. Amphiphysin 2 orchestrates nucleus positioning and shape by linking the nuclear envelope to the actin and microtubule cytoskeleton. Dev Cell. 2015;35:186–98. https://doi.org/10.1016/j.devcel.2015.09.018.CrossRefPubMedGoogle Scholar
- 19.Louis M, Piccinotti S, Vosshall LB. High-resolution measurement of odor-driven behavior in Drosophila larvae. J Vis Exp. 2008;11:638. https://doi.org/10.3791/638.
- 22.Collins MA, Mandigo TR, Camuglia JM, Vazquez GA, Anderson AJ, Hudson CH, et al. Emery-Dreifuss muscular dystrophy-linked genes and centronuclear myopathy-linked genes regulate myonuclear movement by distinct mechanisms. Mol Biol Cell. 2017; https://doi.org/10.1091/mbc.E16-10-0721.
- 39.Auld AL, Roberts SA, Murphy CB, Camuglia JM, Folker ES. J Cell Sci. 2018;131(6). https://doi.org/10.1242/jcs.205807.
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.