Summary
The genesis of syncytial muscles is typically considered as a paradigm for an irreversible developmental process. Notably, transdifferentiation of syncytial muscles is naturally occurring during Drosophila development. The ventral longitudinal heart-associated musculature (VLM) arises by a unique mechanism that revokes the differentiated fate from the so-called alary muscles and comprises at least two distinct steps: syncytial muscle cell fragmentation into single myoblasts and direct reprogramming into founder cells of the VLM lineage. Here we provide evidence that the mesodermal master regulator twist plays a key role during this reprogramming process. Acting downstream of Drosophila Tbx1 (Org-1) in the alary muscle lineage, Twist is crucially required for the derepression of the Hippo pathway effector Yki and thus for the initiation of syncytial muscle dedifferentiation and fragmentation. Subsequently, cell-autonomous FGFR-Ras-MAPK signaling in the resulting mono-nucleated myoblasts is maintaining Twist expression, thereby stabilizing nuclear Yki activity and inducing their lineage switch into the founder cells of the VLM.
Introduction
Lineage restriction is an essential concept in multicellular organisms. On the one hand, a lower fate commitment enables a high regenerative potential, but on the other hand, a higher fate commitment prevents uncontrolled growth and cancer. During Drosophila embryonic development early fate decisions within the blastoderm are initiated by the bHLH transcription factor Twist, which is crucial for mesoderm specification (Thisse et al., 1988) and is required for the activation of mesoderm specific factors (Furlong et al., 2001; Leptin, 1991) such as the Drosophila FGFR heartless (htl), required for mesodermal cell migration (Beiman et al., 1996; Gisselbrecht et al., 1996; Shishido et al., 1993; Shishido et al., 1997), and Mef2 initiating muscle differentiation (Bour et al., 1995; Lilly et al., 1995; Taylor et al., 1995). Although Twist was reported to promote embryonic myogenesis in Drosophila it appears to play the opposite role during vertebrate myogenesis, where primary Twist function is proposed in preventing the premature differentiation of myogenic cells (Baylies and Bate, 1996; Hebrok et al., 1994; Rohwedel et al., 1995; Spicer et al., 1996).
Drosophila twist is known to become restricted after its early mesodermal expression to adult muscle precursors (AMPs), muscle stem cells from which the adult somatic musculature will arise during metamorphosis (Bate et al., 1991). Herein we report that twist expression additionally persists in the embryonic and larval heart-associated syncytial alary muscles (AMs), which connect the embryonic heart to internal organs and the exoskeleton (Bataillé et al., 2020; Boukhatmi et al., 2014). Our recent work has shown that during metamorphosis a subset of these AMs exhibits an extraordinary level of cell fate plasticity. Escaping histolysis, the three anterior muscle pairs undergo an in vivo direct lineage reprogramming process that includes their dedifferentiation, fragmentation into mononucleated alary muscle derived cells (AMDCs) and re-differentiation into the ventral longitudinal musculature of the heart (VLM) (Schaub et al., 2015). We demonstrate that after the onset of metamorphosis twist expression persists during AM to VLM transdifferentiation in the AM lineage and that the reprogramming process crucially depends on Twist function. Mechanistically, the induction of trans-differentiation depends on the function of the Drosophila homolog of Tbx1 (optomotor-blind-related-gene-1, org-1), whereas the initiation of syncytial muscle dedifferentiation and fragmentation requires derepression of the transcriptional effector of the Hippo pathway, Yorkie (Yki) (Schaub et al., 2015; Schaub et al., 2019). We find that twist expression during lineage reprogramming is maintained by Org-1 and that Twist function is required for the derepression of Yki function in the anterior AMs. This identifies Twist as a crucial factor that links these two mechanisms. We show that Twist together with the Drosophila FGFR Heartless is essential for the reprogramming process of the AMDCs into the founder myoblasts of the VLMs and demonstrate that this lineage switch requires the stabilization of nuclear Yki localization in the AMDCs by active Ras-MAPK signaling. Altogether, our results show that a Twist/Yki/FGFR axis plays a central role in the regulatory network that mediates syncytial muscle cell fate plasticity and initiates a lineage switch during cellular reprogramming. Of note, vertebrate orthologs of Twist, Yki and FGFR are well known to be involved in the reprogramming processes that transform myogenic cells into rhabdomyosarcoma (Goldstein et al., 2007; Li et al., 2019; Tremblay et al., 2014). Our results connect the function of this set of regulators within a naturally occurring muscle lineage reprogramming event. Apart from adding mechanistic insights into this process they are potentially also relevant for neoplastic reprogramming.
Results
Twist is expressed in the alary muscle lineage and is required for AM to VLM lineage reprogramming
Twist expression is found during early embryonic development in the whole mesoderm, where its expression is regulated by upstream enhancer elements that are bound by the morphogen Dorsal (Jiang et al., 1991; Thisse et al., 1991). During later developmental stages somatic twist expression is known to be restricted to the AMPs that are specified via the rhomboid-triggered EGF signaling pathway (Bate et al., 1991; Figeac et al., 2010; Thisse et al., 1988). Unexpectedly, during an enhancer dissection analysis at the twist locus, reporter constructs driven by the complete downstream intergenic region (twist.Ko6-GFP, Fig. 1A) were found to include expression in the embryonic alary muscles (Dominik Müller, Jakob Bartle-Schultheis and Manfred Frasch, unpublished data; Fig. 1B). Visualization of the Twist protein together with an alary muscle-specific reporter (tupAME-GFP) during embryonic stages clearly show a localization of Twist in the nuclei of the embryonic alary muscles (Fig. 1C,C’). Moreover, twist.Ko6-GFP can be detected in larval alary muscles as well as in the VLM and the adult alary muscles (Fig. 1D-F). Live imaging demonstrated that the twist.Ko6-GFP reporter is active during the whole AM to VLM transdifferentiation process in the alary muscle lineage (video S1). The analysis of the expression pattern of a GFP-tagged version of Twist under endogenous control (twi::GFP) revealed that it is expressed in the org-1-RFP positive AMs, after dedifferentiation and fragmentation in the AMDCs as well as in the primordia of the VLM (Fig. 2A-C), indicating a potential role of Twist during alary muscle reprogramming. To dissect a possible requirement of twist function during VLM generation we induced its AM specific knock-down by RNAi or by inducible CRISPR using org-1-GAL4 during metamorphosis. The analyses of the resulting phenotypes revealed a strong perturbation of VLM formation and alary muscle reprogramming (Fig. 2E,H). Accordingly, live imaging demonstrated that upon RNAi mediated twist knock-down alary muscle dedifferentiation and fragmentation is severely impaired (video S2). Past work has shown that Twist function during embryonic myogenesis in Drosophila as well as vertebrates depends on dimer choice and protein-protein interactions (Franco et al., 2011). Interestingly, the downregulation of the Drosophila E-protein Daughterless, which has been proposed to inhibit somatic myogenesis via heterodimerization with Twist (Castanon et al., 2001), by RNAi does not cause a significant phenotype during AM lineage reprogramming (Fig. 2H). In contrast, the elevation of the transcript levels of twist and nautilus, the Drosophila MyoD homolog (Michelson et al., 1990), using org-1-GAL4 during lineage reprogramming abolishes VLM formation (Fig. 2F,G,H), indicating a dose-dependent, mechanistical role for bHLH factor interaction during transdifferentiation. These results demonstrate an important role for Twist function in mediating lineage plasticity and reversal of differentiated, syncytial muscle fate during AM lineage programming.
Twist links Org-1 and Yki function during AM dedifferentiation
The induction of AM to VLM transdifferentiation crucially requires a transcriptional program that is activated by the AM lineage-specific induction of the Drosophila Tbx1 homolog Org-1 and its direct target, the Islet1 homolog Tailup (Schaub et al., 2015). To test for any functional connections between Org-1, Tup and Twist we analyzed the expression pattern of twi::GFP in genetic backgrounds that produce AM-specific org-1 or tup knock-downs using org-1-GAL4. Strikingly, loss of Org-1 or of Tup leads to the nearly complete abolishment of twi::GFP expression in the org-1-RFP positive alary muscles (Fig. 3B,C), indicating that twist represents a downstream target of the Org-1 regulatory cascade during AM transdifferentiation. To further dissect the regulatory function of Org-1 on the twist cis-regulatory region (CRM) we reanalyzed the twist.Ko6 downstream intergenic region and were able to identify an enhancer region that drives GFP reporter expression in the VLM lineage (twiVLM-GFP, Fig. 3D,F), suggesting a role of this CRM in twist regulation during AM lineage reprogramming. Org-1 as well as Tup are capable to significantly bind twiVLM in vivo (Fig. 3D) (Kudron et al., 2018) and the mutation of its predicted Org-1 binding sites (twiVLM-orgI-IIImut-GFP, Fig. 3E) leads to nearly complete abolishment of reporter activity in the VLM lineage (Fig 3G), thus demonstrating direct regulation of twist via this CRM by Org-1 during AM to VLM transdifferentiation.
Yki represents the transcriptional effector of the Hippo network and its function as a transcriptional co-activator of Scalloped (Sd) is regulated by phosphorylation, thereby inhibiting its nuclear translocation (Goulev et al., 2008; Huang et al., 2005; Wu et al., 2008; Zhao et al., 2008). During AM to VLM transdifferentiation, Org-1 and Tup induce the derepression of the Yki/Sd complex in the syncytial AMs and thereby the activation of its target genes, thus leading to the dedifferentiation of the AMs into mononucleated cells and the induction of lineage reprogramming (Schaub et al., 2019). To dissect a functional connection between Twist and Yki during AM transdifferentiation we forced the co-expression of a phosphorylation resistant, constitutively active version of Yki together with RNAi against twist by using org-1-GAL4. Strikingly, we observed that constitutively active Yki caused a significant rescue of VLM formation in Twist knock-down backgrounds (UAS-dsRNA-twi, UAS-yki.S168A; p≤0.01; Fig. 4A-C), strongly suggesting that Twist is required for Yki derepression during alary muscle reprogramming. Taken together, these results support a mechanism that requires Org-1-dependent Twist function for the derepression of Yki in the alary muscles and concomitant activation of Yki/Sd target gene programs, which trigger alary muscle dedifferentiation.
Nuclear Twist and Yki are stabilized by FGFR signaling during muscle lineage reprogramming
During the establishment of the adult abdominal somatic musculature the Twist positive adult muscle precursors show expression of the Drosophila FGFR Heartless (Htl), which mediates the induction of founder myoblast fate in some of them (Dutta et al., 2005). Stainings for htl transcript as well as live imaging of the activity of a GAL4 line (htl-GAL4) (Jenett et al., 2012) driven by an htl enhancer containing previously characterized early embryonic Twist-binding sites (Stathopoulos et al., 2004; Zinzen et al., 2009), revealed that htl is expressed during AM to VLM transdifferentiation in the alary muscles as well as in the AMDCs (Fig. 4D,E; Fig. S1). Since after induction of RNAi against htl (UAS-dsRNA-htl) or the forced expression of a dominant negative version of Htl (UAS-htl.DN) no org-1-RFP positive muscles remain in pharate adults (Fig. 4A,F; Schaub et al. 2015), we deduce that activated FGFR signaling induces molecular events that are crucially required for alary muscle lineage reprogramming. To further investigate the role of FGFR signaling during AM to VLM transdifferentiation we performed live imaging experiments, thus demonstrating that mis-expression of a dominant negative version of Htl (UAS-htl.DN) using org-1-GAL4 has no influence on alary muscle dedifferentiation and fragmentation but inhibits reprogramming of the org-1-RFP positive AMDCs into the founder cells of the VLM (video S3). Since reprogramming of the AMDCs crucially depends on the activation of Yki targets (Schaub et al., 2019), we asked if Yki activity in the AMDCs may be connected with activated Htl signaling. Our analyses revealed that the loss of VLM phenotypes generated via the activation of HtlDN (UAS-htl.DN) can be significantly rescued by co-expression of a constitutive active Yki (UAS-htl.DN; UAS-yki.S168A, p≤0.001, Fig. 4A,F,G), thus suggesting that activated FGFR signaling is required for the maintenance of nuclear Yki function during AMDC reprogramming. Interestingly, HtlDN (UAS-htl.DN) also leads to the loss of twi::GFP expression in the AMDCs, but not in the alary muscles before fragmentation (Fig. 4H,I). Taken into account that constitutive active Yki is epistatic over Twist (Fig. 4A,C), this suggests a regulatory interaction between FGFR, Twist and Yki function. Since it has been shown that Hippo signaling in wing discs can be negatively regulated via activated EGFR-Ras-MAPK signaling through the Ajuba LIM protein Jub (Reddy and Irvine, 2013), we speculated that sustained derepression of Yki in the AMDCs may depend on an analogous mechanism mediated by FGFR-Ras-MAPK signaling. Whereas induction of RNAi against EGFR or forced expression of the EGFR antagonist Argos has no phenotypic consequences (Fig. 4A), org-1-GAL4 induced expression of dominant negative versions of Drosophila Ras (UAS-Ras85D.DN) as well as mammalian Ras proteins (UAS-Ras.DN) or downregulation of the Drosophila MAPK Rolled (Rl, UAS-dsRNA-rl) lead to the loss of VLM formation and AMDC reprogramming (Fig 4A,J). Strikingly, these phenotypes can be significantly rescued by co-expression of phosphorylation resistant Yki during AMDC reprogramming (p≤0.001, Fig. 4A,K), pointing towards a FGFR-Ras-MAPK-dependent regulation of nuclear Yki. The downregulation of Ajuba (UAS-dsRNA-jub) using org-1-GAL4 provokes the abolishment of VLM generation (Fig. 4A), but unexpectedly these phenotypes cannot be significantly rescued by the constitutive activation of Yki targets, suggesting additional roles of Jub during AMDC reprogramming. Thus, we propose a model in which FGFR-Ras-MAPK signaling is required for the stabilization of nuclear Twist and Yki during reprogramming of the AMDCs into the VLM founder cells.
Discussion
Twist function during embryonic muscle development in Drosophila has been described to possess dominant myogenic properties (Baylies and Bate, 1996), whereas vertebrate Twist is excluded from the myotome during embryogenesis and has been implicated in inhibiting myogenesis (Te and Reggiani, 2002). Twist expression and function in terminally differentiated striated, syncytial muscles has not been described yet. Here we show that Twist expression persist in the syncytial striated alary muscles after terminal differentiation and during metamorphosis in Drosophila. The alary muscles are a special type of striated muscles that are clearly distinguished from the body wall muscles (Bataillé et al., 2020; Boukhatmi et al., 2014), a feature that may be functionally connected to their persistent twist expression. During metamorphosis, the alary muscles demonstrate an extraordinary amount of cellular fate plasticity. Whereas the posterior larval alary muscles are remodeled into the adult alary muscles, the anterior ones generate the heart-associated ventral longitudinal musculature (VLM) by transdifferentiation (Schaub et al., 2015). Twist function and cellular plasticity have been connected in the literature mainly due to its role as one of the transcriptional master regulators during the transition from epithelial to mesenchymal states (EMT), crucially required for embryonic morphogenesis and often linked to metastasis in various cancer models (Lu and Kang, 2019). In terms of myogenic cell fate plasticity it is known that prolonged Twist expression negatively regulates muscle cell differentiation during embryonic myogenesis as well as adult flight muscle development in Drosophila (Anant et al., 1998; Cripps and Olson, 1998; Domsch et al., 2020), whereas forced expression of vertebrate Twist proteins can even reverse myogenic differentiation in vitro by unwiring the myogenic regulatory network from its targets (Hjiantoniou et al., 2008; Li et al., 2019; Liu et al., 2017; Mastroyiannopoulos et al., 2013) and appear to play a role during the regeneration of craniofacial muscles (Zhao et al., 2020). Here we provide evidence that Twist mediates the reversal of syncytial cell fate uncovering an unknown function of Twist in Drosophila muscles during metamorphosis.
It has been shown that sustained activation of the Hippo effector YAP inhibits the terminal myogenic differentiation program and that its Drosophila counterpart, Yki, induces reversal of syncytial differentiation during AM lineage reprogramming (Schaub et al., 2019; Sun et al., 2017; Watt et al., 2010). Here we demonstrate that Twist function is required for the derepression of Yki during AM transdifferentiation, connecting the Twist regulatory network to the activation of Yki target genes required for myogenic fate plasticity. Numerous studies have provided evidence for roles of YAP and TAZ, the vertebrate homologs of Yki, in EMT transdifferentiation and proposed that the induction of EMT upon forced Twist expression in cultured epithelial cells is connected to the inactivation of the Hippo pathway (Pan, 2010; Wang et al., 2016). These two mechanisms could be linked via the activation of JNK signaling and its transcriptional effector AP-1, both shown to be required for Twist-induced EMT as well as for Yki target activation (Nam et al., 2015; Sahu et al., 2015; Schaub et al., 2019; Sun and Irvine, 2013). Thus we speculate that Twist-induced abrogation of Yki target gene repression may represent a common molecular mechanism mediating cell fate plasticity in mesodermal cells. Twist function and FGFR signaling have been shown to be closely connected during development of mesodermal derivatives in various species (Beiman et al., 1996; Bloch-Zupan et al., 2001; Gisselbrecht et al., 1996; Harfe et al., 1998; Johnson et al., 2000; Pfeifer et al., 2014; Shishido et al., 1993; Shishido et al., 1997). During the genesis of the Drosophila adult musculature, FGFR signaling is required in Twist positive myoblasts for the specification of the abdominal founder cells that will seed the respective myofibers (Dutta et al., 2004; Dutta et al., 2005). We show that induction of FGFR signaling is required in another population of Twist positive myoblasts, i. e. those derived from the fragmentation of Twist positive syncytial alary muscles (AMDCs) during metamorphosis, to induce their lineage switch to the founder cells of the ventral heart-associated musculature (VLM). Furthermore, our data link active FGFR/MAPK signaling with Twist-mediated sustained de-repression of Yki in these cells. This connects Twist function to the induction of a potential Twist/Yki/FGFR feed-forward pathway driving AMDC lineage reprogramming. This is reminiscent of a mechanism proposed to be required for survival in cancer cells (Hua et al., 2016; Rizvi et al., 2016).
Taken together we propose a model in which Twist acts as a key regulator of alary muscle lineage reprogramming by coupling spatially and temporally specific transcriptional inputs to the de-repression of Yki targets (Fig. 4L). At the onset of metamorphosis, Org-1/Tbx1 and Tup/Isl1 expression are induced in the anterior alary muscles by the combinatorial inputs of the Hox protein Ultrabithorax and the ligand bound receptor of Ecdysone (Schaub et al., 2015). This leads to the stabilization of Twist expression only in the AMs fated to undergo transdifferentiation. Subsequently, the integration of the Org-1, Tup and Twist induced transcriptional networks leads to the de-repression of Yki/Sd target genes (e. g. Myc) required for the induction of muscle dedifferentiation and alary muscle fragmentation (Schaub et al., 2019). Once the alary muscles have fragmented into single myoblasts, the subsequent activation of FGFR/MAPK signaling in the AMDCs stabilizes the expression of Twist-induced Yki/Sd target genes, possibly via a regulatory Twi/Yki/FGFR feed-forward loop, and mediates reprogramming of the AMDCs into the founder cells of the VLM.
Materials and methods
Drosophila stocks and culture conditions
Standard Drosophila melanogaster genetics were performed at 25 °C. The following stocks were used: [org-1-GAL4, org-1-RFP; twi::GFP], [org-1-RFP; twist.Ko6-GFP], tupAME-GFP (Boukhatmi et al., 2014), [org-1-GAL4, org-1-RFP], [org-1-GAL4, org-1-RFP; hand-GFP], [org-1-GAL4, org-1-RFP; UAS-Cas9.P2], [org-1-GAL4, org-1-RFP; UAS-yki.S168A] (Schaub et al., 2019), UAS-dsRNA-org-1 (62953), UAS-dsRNA-tup (51763), UAS-dsRNA-twi (51164), UAS-htl.DN-B40; UAS-htl.DN.-B61 (5366), UAS-dsRNA-htl-2 (35024), UAS-dsRNA-EGFR-1 (60012), UAS-dsRNA-EGFR-2 (25781), UAS-argos.M; UAS-argos.M (5363), UAS-Ras85D.N17 (4845), UAS-Ras.N17 (4846), UAS-dsRNA-rl (34855), UAS-dsRNA-jub (32923), GMR93H07-htl-GAL4 (40669), UAS-His::RFP (derived from 56555) and twi::GFP (79615) were obtained from the Bloomington Drosophila Stock Center. UAS-dsRNA-htl-1 (6692) UAS-dsRNA-da-1 (51297), UAS-dsRNA-da-2 (51300) and UAS-dsRNA-da-3 (105259) were obtained from the Vienna Drosophila Resource Center (VDRC).
Generation of transgenic Drosophila stocks
To create UAS-t::gRNA-twi4x we utilized the system described in (Port and Bullock, 2016). Using 1 ng pCFD6 (Addgene, Cat. No: 73915) as template, we amplified PCR fragments with the respective primers containing the gRNA sequences (Table S1) and utilizing Q5 Hot Start High-Fidelity 2X Master Mix (New England Biolabs, Cat. No: M0494L). After gel-purification of the PCR fragments with the QIAquick Gel Extraction Kit (Qiagen, Cat. No: 28704), these were assembled with BbsI-HF (New England Biolabs, Cat. No: 3539S) digested pCFD6 backbone to generate UAS-t::gRNA-twi4x using the NEBuilder® HiFi DNA Assembly Cloning Kit (New England Biolabs, Cat. No: R5520S), following the instructions provided by the manufacturers.
To create twiVLM-GFP, the respective genomic region (2R:23052698..23054095) was amplified from yw DNA with the primers KpnI-twiVLM-fw (GGGGT ACCCCCAGTAAGGCAAATTGCTCAG) and XhoI-twiVLM-rev (CCGCTCGAGCGGAACTTGC CTTGTCCTTCGTC) utilizing Q5 Hot Start High-Fidelity 2X Master Mix (New England Biolabs, Cat. No: M0494L). The resulting PCR product was gel purified with the QIAquick Gel Extraction Kit (Qiagen, Cat. No: 28704) and cloned into KpnI/XhoI of pH-Stinger-AttB (Jin et al., 2013) using T4 DNA ligase (New England Biolabs, Cat. No: M2020L).
For the analogous creation of twiVLM-orgI-IIImut-GFP the Org-1 binding sites within the twiVLM sequence were predicted with Target Explorer (Sosinsky et al., 2003) using a positional weight matrix generated by SELEX with Org-1 (Jin et al., 2013) and mutated via site-directed mutagenesis.
Transgenes of the constructs were generated by utilizing the phiC31/attP/attB system (Bischof et al., 2007) and performed by a commercial embryo injection service (BestGene Inc, Chino Hills, CA 91709, USA) to insert UAS-t::gRNA-twi4x at the attP40 landing site or to insert twiVLM-GFP as well as twiVLM-orgI-IIImut-GFP into the attP2 landing site. Potential transformants were selected by mini-white expression.
Fluorescent antibody staining
For antibody stainings of embryos the standard fixation protocol (Wieschaus and Nüsslein-Volhard, 1986) was applied. For dissection of pupal stages with microsurgery scissors the pupal cases were first removed and prefixation with 3,7 % formaldehyde for 30 minutes was performed. Pharate adults were dissected without prefixation. The dissected animals were fixed in 3,7 % formaldehyde for 10 minutes and were washed three times for 30 minutes in PBT (PBS + 0.1 % Tween©20).
Fixed embryos and pupae were blocked in 10 % Bovine Serum Albumin (Serva, Cat. No: 11930.04) for 1 hour. Samples were incubated for overnight (embryos) or two days (pupae) at 4 °C with the respective primary antibodies: Rabbit polyclonal anti-GFP (Rockland, Cat. No: 600-401-215, 1:500), Goat polyclonal anti-GFP (GeneTex Cat. No: GTX26673, 1:1000), Rabbit polyclonal anti-RFP (Rockland, Cat. No: 600-401-379S, 1:500), Rat polyclonal anti-Tropomyosin (MAC141, abcam, Cat. No: ab50567, 1:200), Rat polyclonal anti-Org-1 ((Schaub et al., 2012), 1:100), Guinea pig polyclonal anti-Twist (a gift of Katrin Domsch given by Eileen Furlong, 1:200), Rabbit polyclonal anti-Twist (a gift of Manfred Frasch, 1:500). To visualize filamentous actin Phalloidin conjugated either with Atto 488 (1:1000, Sigma-Aldrich, Cat. No: 49409) or Atto 647N (1:1000, Sigma-Aldrich, Cat. No: 65906) was added to the antibody solution. After the incubation with the primary antibodies the embryos and pupal dissections were washed three times for 30 minutes in PBT and incubated with the respective secondary antibodies: Alexa Fluor 488 conjugated Donkey Anti-Goat (abcam Cat. No: ab150129), Alexa Fluor 488 conjugated Goat Anti-Rabbit (ThermoFisher Scientific Cat. No: A-11008), DyLight 549 conjugated Goat anti-Rabbit (Jackson ImmunoResearch, Cat. No: 111-505-003) Alexa Fluor 555 conjugated Donkey Anti-Rabbit (abcam Cat. No: ab150074), Cy3 conjugated Donkey anti-Guinea pig (Jackson ImmunoResearch, Cat. No: 706-165-148), DyLight 649 conjugated Goat anti-Rat (Jackson ImmunoResearch, Cat. No: 112-495-003) in 1:200 dilution overnight at 4 °C. Finally, the stained animals were washed three times for 30 minutes in PBT and were embedded into Vectashield containing DAPI (Vectorlabs, Cat. No: H-1200).
In situ hybridization chain reaction
To detect heartless mRNA in pupae we performed in situ hybridization chain reaction (HCR) as described in (Choi et al., 2018). For this purpose, we purchased a kit containing a custom htl DNA probe set and a fluorescently labelled DNA HCR amplifier as well as hybridization, wash and amplification buffers from Molecular Instruments (Los Angeles, CA 90041, USA).
For dissection of pupal stages with microsurgery scissors the pupal cases were first removed and prefixation with 3,7 % formaldehyde for 30 minutes was performed. The dissected animals were fixed in 3,7 % formaldehyde for 10 minutes, washed three times for 10 minutes in PBT (PBS + 0.1 % Tween©20), followed by a 5 minute wash in 50 % Methanol/PBT, a 5 minute wash in 100 % Methanol and another 5 minute wash in 50 % Methanol/PBT. The dissected pupae were rehydrated by three 5 minute washes in PBT followed by three 30 minute washes in PBT. After post-fixation with 3,7 % formaldehyde for 30 minutes the specimens were washed for three times with PBT for 5 minutes. After 30 minutes of prehybridization at 37°C in hybridization buffer, dissected pupae were hybridized with 0,4 pmol of the htl DNA probe set in 100 µl hybridization buffer for 16 hours at 37 °C. To remove unbound probes, the specimens were washed four times for 15 minutes with probe wash buffer at 37 °C, followed by three washes with 5xSSCT (5xSSC + 0.1 % Tween©20) at room temperature. For signal amplification the embryos and pupae were subjected to preamplification in amplification buffer for 30 minutes, followed by incubation with 6 pmol of each snap cooled, Alexa Fluor 647 labelled amplification hairpin in 100 µl amplification buffer for 16 hours at room temperature. To remove excess hairpins, two washes with 5xSSCT for 5 minutes and two washes with 5xSSCT for 30 minutes were performed. The pupae stained for htl mRNA localization were subjected to antibody stainings as described above.
Sample analysis
Confocal Z-stacks of fixed or in vivo specimen were acquired with a Leica SP5 II (10x/0.4 HC PL APO Air, 20x/0.7 HC PL APO Glycerol, 63x/1.3 HC PL APO Glycerol). Maximum Projections of the Z-stacks were performed with FIJI/ImageJ (v1.52v) (Schindelin et al., 2012).
In vivo time-lapse imaging
Stage P3 pupae of the respective genotypes were aligned on a strip of double-faced adhesive tape connected to a slide, covered with a drop of Halocarbon oil 700 (Halocarbon Products Corp, Sigma Aldrich Cat. No: H8898) and a coverslip. Time-lapse series were acquired on a Leica SP5 II confocal system using a HC PL APO 10x/0.4 Air. Acquisition was done over a time course of 48 hours with the following settings: Pinhole 1.4 AU, 1.2x optical zoom, scan speed 200 Hz, Z-stack about 30-35 sections with a step size of 2-4 µm and time intervals of 10 min per stack. Movies were generated using Leica Application Suite X (LAS X, Leica Microsystems) and with FIJI/ImageJ (v1.52v) (Schindelin et al., 2012).
Statistical analysis
VLM phenotype frequency was quantified with living pharate stages of the respective genotypes that were aligned on a microscope slide. With the aid of a fluorescence equipped microscope (Nikon, Eclipse 80i) the org-1-RFP lineage marker in the living animals was visualized and phenotypes with fully differentiated VLM (proper aligned and elongated muscle fibers) were scored versus phenotypes ranging from missing VLM to defective VLM differentiation (no aligned muscle fibers). The Chi square test function of Microsoft Excel (Microsoft) was used for statistical analysis of the samples and statistical significance was defined by p≤0.05. Fluorescent signal quantification and illustration was performed with FIJI/ImageJ (v1.52v) (Schindelin et al., 2012).
Competing interests
The authors declare no competing financial interests.
Author contributions
Conceptualization: Christoph Schaub; Investigation: Marcel Rose, Jakob Bartle-Schultheis, Katrin Domsch, Ingolf Reim and Christoph Schaub ; Writing – Original Draft: Christoph Schaub; Writing – Review & Editing: Marcel Rose, Jakob Bartle-Schultheis, Katrin Domsch, Ingolf Reim and Christoph Schaub; Funding Acquisition: Christoph Schaub; Supervision: Christoph Schaub
Funding
This work was funded by the Deutsche Forschungsgemeinschaft (DFG) (SCHA 2091/1-1 to C.S.).
Supplemental information
Figure S1.Activity of htl expression during AM/VLM transdifferentiation. Flies carrying htl-GAL4 (GMR93H07) and UAS-His::RFP in addition to twi.Ko6-GFP were imaged during metamorphosis. A-J show stills from the same image series in developmental order (dorsal view). Arrows indicate htl-GAL4-driven expression of Histone-RFP in (A, B) either AMs prior to abdominal compaction or (C-J) in AMDCs and the developing VLM. (Note that only some portions are optically accessible from the outside particularly at early pre-pupal stages.) Expression is also observed in the central heart and various flanking abdominal skeletal muscles (both GFP-negative) as well as the developing flight muscles (GFP-positive, to the left in C-J).
Video S1.Dorsal view of a pupa that carries twist.Ko6-GFP and org-1-RFP. The movie starts at late stage P3 and demonstrates the co-expression of twist.Ko6-GFP and org-1-RFP during the transdifferentiation process of the larval alary muscles into the ventral longitudinal musculature of the heart. At the beginning of the movie, the syncytial org-1-RFP and twist.Ko6-GFP positive anterior alary muscles initiate dedifferentiation into mesen-chymal, mononucleated myoblasts. These keep close contact to each other and migrate towards the dorsal vessel where they start re-differentiating into the ventral longitudinal heart muscles elongating along the heart tube towards the posterior. In addition, twist.Ko6-GFP can be detected in the remodeling org-1-RFP negative larval posterior alary muscles, the dorsal adult muscle precursors (AMPs) as well as the hindgut musculature.
Video S2.Dorsal view of a pupa that carries hand-GFP and org-1-RFP and in which org-1-GAL4 mediated RNAi against twist was induced (org-1>>UAS-dsRNA-twi). The movie starts at late stage P3 and demonstrates that knock-down of twi prevents the alary muscles from dedifferentiating into mononucleated myoblasts, suggesting that Twist function is indispensable for the induction of this essential process during alary muscle trans-differentiation.
Video S3.Dorsal view of a pupa that carries hand-GFP and org-1-RFP and in which a dominant negative version of the Drosophila FGFR Htl was induced (org-1>>UAS-htl.DN). The movie starts at late stage P3 and shows that interference with FGFR mediated signaling has no impact on the fragmentation of the alary muscles into mononucleated myoblasts but demonstrates that transdifferentiation is arrested under these conditions after alary muscle dedifferentiation. This suggests an essential role of Htl-signaling during the lineage reprogramming of the alary muscle derived myoblasts into the founders of the VLM during transdifferentiation.
Acknowledgements
We thank the Drosophila community for providing us with antibodies and fly stocks and in particular Manfred Frasch (Friedrich-Alexander-Universität Erlangen-Nürnberg) for his gifts of unpublished materials. We are also grateful to the Bloomington Drosophila Stock Center and the Vienna Drosophila Resource Center for providing fly stocks as well as to the Optical Imaging Centre Erlangen (OICE) for their support and facilities.
Footnotes
Author list updated; htl expression data during lineage reprogramming included; Figure 4 revised; Supplemental files updated