ABSTRACT
Myosin1D (Myo1D) has recently emerged as a conserved regulator of animal LR asymmetry that governs the morphogenesis of the central LR Organizer (LRO). In addition to Myo1D, the zebrafish genome encodes the closely related Myo1G. While Myo1G also controls LR asymmetry, we show that it does so through an entirely different mechanism. Myo1G promotes the Nodal-mediated transfer of laterality information from the LRO to target tissues. At the cellular level, Myo1G is associated with endosomes positive for the TGFβ signaling adapter SARA. myo1g mutants have fewer SARA-positive Activin receptor endosomes and a reduced responsiveness to Nodal ligands that results in a delay of left-sided Nodal propagation and tissue-specific laterality defects in organs that are most distant from the LRO. Beyond LR asymmetry, Myo1G promotes signaling by different Nodal ligands in other biological contexts. Our findings therefore identify Myo1G as a novel positive regulator of the Nodal signaling pathway.
INTRODUCTION
Left-Right (LR) asymmetries in the positioning and shape of different tissues are found in both protostome and deuterostome lineages and critically required for human organ function1. In spite of the importance of LR asymmetry, our understanding of the mechanisms that govern this third body axis remains fragmentary. A particularly striking feature of LR asymmetry is the fact that an evolutionary conserved mechanism of symmetry breaking has long remained elusive. Although Nodal proteins of the Transforming Growth Factor β superfamily have long been known to control LR asymmetry in all deuterostome and some protostome species2,3, it is only recently that the unconventional type 1 Myosin Myosin1D (Myo1D) has emerged as a potentially universal regulator of animal LR asymmetry4-7. Here, we identify a close orthologue of Myo1D, Myosin1G (Myo1G) as a novel positive regulator of the Nodal signaling pathway.
Seminal studies in the mouse revealed the existence of a central LR Organizer (LRO) in which the Planar Cell Polarity (PCP)-dependent orientation of motile cilia promotes the generation of a directional symmetry-breaking fluid flow8-10. Symmetry-breaking cilia-driven fluid flows are also present in other species including fish and frogs11,12. Already within the vertebrate phylum, the LROs of birds and reptiles do however lack motile cilia and rely - at least in chick - on lateralized cell flows to trigger symmetry breaking13,14. Additional mechanisms implicated in LR asymmetry include ion flows15 and Actin-dependent chiral cell remodeling16,17. While an increasing number of studies indicate that Actin- and PCP-dependent pathways lie at the core of a symmetry-breaking toolbox4,5,18-22, our understanding of the evolutionary conservation of the mechanisms controlling LR asymmetry remains fragmentary.
In vertebrates, Nodal ligands convey laterality information from the central LRO to different target tissues1,2. Nodal ligands propagate on the left side of the embryo by inducing their own expression, allowing them to propagate from the posteriorly located LRO to more anterior target tissues 23. In species with a LRO bearing motile cilia, Nodal is expressed initially in a bilaterally symmetric fashion at the LRO, together with the TGFβ signaling antagonist Dand524. Upon establishment of a ciliary LRO flow, dand5 transcripts are degraded on the left side of the LRO25-27, allowing Nodal to travel to the left lateral plate mesoderm and propagate by autoinduction.
Nodal ligands induce cellular responses through ligand/receptor complexes that comprise TGFβ type I and II receptors and the co-receptor Cripto/Oep28. Nodal ligand binding causes type II receptors to phosphorylate and activate their type I counterpart. A population of endosomes positive for the TGFβ signaling adapter Smad Anchor for Receptor Activation (SARA) promotes signal transduction by allowing Activin/Nodal receptors to recruit their transcriptional downstream mediators SMAD2 & 329. Upon phosphorylation by activated type I receptors, SMAD2 & 3 associate with SMAD4 to enter the nucleus and activate target genes23. As Nodal ligands are highly potent, a tight regulation of Nodal signaling is essential not only for embryonic development but also to avoid tumorigenesis23,30. Lefty proteins act as feed-back inhibitors of Nodal signaling that prevent the formation of productive ligand/receptor complexes23,31. In LR asymmetry, Lefty expression at the embryonic midline is important to form a midline barrier that prevents the spreading of left-sided Nodal ligands to the contralateral side32,33.
The requirement of Nodal ligands for LR asymmetry is however not universally conserved1 and a number of protostomian species, including the fruitfly Drosophila, altogether lack nodal homologues. Studies in Drosophila identified Myo1D as a master regulator of LR asymmetry34,35. In contrast to the central LRO of vertebrate organisms that governs LR asymmetry of all lateralized organs, Drosophila myo1d acts in a local, tissue-autonomous fashion to control genital and visceral laterality18,35. Of particular interest, studies in frogs, fish and humans showed that Myo1D is also required for vertebrate LR asymmetry4-7.
Zebrafish Myo1D is required for the establishment of a functional symmetry-breaking ciliary LRO flow5. In addition to myo1d, the fish genome harbors the closely related gene myosin1g (myo1g). Although myo1g mutations impair laterality and enhance the defects of myo1d mutants, we show that Myo1G acts independently of the LRO flow, through an entirely different mechanism. We provide evidence that Myo1G represents a novel positive regulator of the Nodal signaling pathway whose function is essential for the Nodal-mediated transfer of laterality information.
RESULTS
Myosin1G mutants present tissue-specific Left-Right asymmetry defects
Myo1D controls cilia orientation in the LRO to promote the generation of a symmetry-breaking LRO flow4-6. The closely related protein Myo1G (79% amino acid similarity) is also required for zebrafish LR asymmetry but has no detectable effect on the LRO flow5, suggesting that different type I Myosins regulate LR asymmetry through distinct mechanisms. To address this issue, we performed a detailed characterization of myo1g single and myo1d; myo1g double mutants.
myo1g single mutants present defects in the leftward jogging of cardiac progenitors, the penetrance of which is further enhanced in myo1d; myo1g double mutants (Fig. 1a). To study the effect of myo1g on brain laterality, we analyzed the expression of the Nodal ligand cyclops/nodal related 2 (cyc/ndr2), its feed-back antagonist lefty1 (lft1) and its transcriptional effector pitx2 which display predominantly left-sided expression in the dorsal epithalamus of wild-type embryos33,36,37. In contrast to the mild defects observed in myo1d mutants (Fig. 1b, Supplementary Fig. 1a, b), myo1g single mutants displayed a significantly higher proportion of brain laterality defects (Fig. 1b, Supplementary Fig. 1a, b). The penetrance of brain laterality defects in myo1d; myo1g double mutants is similar to the one observed in myo1g single mutants, confirming the predominant role of myo1g in brain laterality (Fig. 1b, Supplementary Fig. 1a).
a,a’ Quantification of cardiac jogging indicates that myo1g mutants present laterality defects that are enhanced in myo1d;myo1g double mutants (a). Concomitant inactivation of the LRO flow (through dnaaf1 mutation) reveals that myo1d/g mutations enhance the cardiac jogging defects of flow-deficient animals (a’). b Brain asymmetry is impaired in myo1g and myo1d;myo1g mutants. c myo1g mutants do not show visceral LR defects. L: Liver, G: Gut, P: Pancreas. d myo1d/g inactivation enhances the brain laterality phenotypes of LRO flow-deficient dnaaf1 mutants. e,f Visceral laterality phenotypes of dnaaf1 mutants are unaffected by myo1d/g inactivation. b Frontal views of pitx2 expression at 30 somites, dorsal up. c,f Dorsal views of foxa1 expression at 48h, anterior up. Scale bars: 50 µm.
In contrast to the effect of myo1g on brain laterality, analysis of liver, pancreas and gut laterality using the endodermal marker foxa1 failed to reveal visceral LR asymmetry defects in myo1g single mutants (Fig. 1c). The observation that myo1d; myo1g double mutants present visceral laterality defects that are similar to myo1d single mutants (Fig. 1c) confirms that myo1g is dispensable for the establishment of visceral laterality.
Myo1D is required for LRO morphogenesis and the generation of a ciliary fluid flow4-6. Accordingly, myo1d mutants present defects at the level of all lateralized organs (Fig. 1a-c, Supplementary Fig. 1a, b). myo1g loss of function yields no discernable LRO flow defects5 and affects only in a subset of organs (Fig. 1a-c, Supplementary Fig. 1a, b), raising the question whether myo1g may control LR asymmetry through a flow-independent and potentially tissue-specific regulation of organ laterality, similar to the situation described for Drosophila myo1d18,35.
Myosin1G controls Left-Right asymmetry independently of the Left-Right Organizer flow
To directly test if Myosin1 proteins exert LRO flow-independent functions in LR asymmetry, we investigated whether the LR asymmetry defects of animals that lack a LRO flow could be further modified by myo1 gene inactivation. To this aim, we generated double and triple mutants to simultaneously inactivate myo1d & g and the essential regulator of ciliary motility dnaaf1/lrrc5038. Embryos that completely lack a LRO flow, as is the case for dnaaf1 mutants, display a distinctive randomization of cardiac, brain and visceral laterality where LR asymmetry is properly established in roughly one half of the population (situs solitus) but inverted in the other (situs inversus, Fig. 1a’, d, e). Only a small fraction of the embryos that lack a LRO flow display an altogether loss of LR asymmetry (i.e. absence of cardiac jogging and brain laterality markers, visceral situs ambiguus, Fig. 1a’, d, e).
In contrast, animals that lack both a LRO flow and myo1 function display a different phenotype, where the heart primordium fails to jog to either the left or the right side of the animal in most embryos (Fig. 1a’). dnaaf1; myo1g double mutants additionally present a lack of asymmetric pitx2 expression in the dorsal epithalamus that contrasts with the randomization of lateralized gene expression observed in dnaaf1 single mutants (Fig. 1b, d). In contrast to the effect observed at the levels of the heart and brain, the visceral phenotypes of dnaaf1 mutant animals are unaffected by the loss of myo1g (Fig.1 e, f), confirming that myo1g is dispensable for visceral organ laterality.
These findings provide evidence for a novel, LRO flow-independent, function of Myosin1 proteins in LR asymmetry. The observations that i) dnaaf1; myo1g double mutants present a more pronounced loss of brain laterality then dnaaf1; myo1d mutants (Fig. 1d) and that ii) dnaaf1; myo1d; myo1g triple mutants are generally similar to dnaaf1; myo1g double mutants (Fig. 1a’, d, e) suggest that Myo1G exerts a predominant role in the flow-independent control of LR asymmetry.
Myosin1G is required for Nodal pathway gene expression
Myo1 proteins could act in different ways to ensure a tissue-specific control of embryonic LR asymmetry. First, zebrafish myo1d & g could act in an organ-intrinsic fashion to promote chiral morphogenesis as in Drosophila18,35. Second, Myo1 activity could be required for the Nodal-mediated propagation of laterality information from the central LRO to different target tissues.
Already prior to the first morphological manifestations of asymmetric cell movement, the heart primordium displays asymmetries in gene expression in response to Nodal signaling from the Left Lateral Plate Mesoderm (LLPM)39,40. Of particular interest, the cardiac primordia of myo1g single and myo1d; myo1g double mutants present a reduced left-sided expression of the Nodal downstream target and feedback inhibitor lefty2 (lft2) that could reflect impaired Nodal signaling (Fig. 2a). Accordingly, the expression of southpaw (spaw), the zebrafish Nodal ligand responsible for left-sided Nodal signaling, is reduced and extends less anteriorly in the LLPM of myo1g single and myo1d; myo1g double mutants, while being affected to a lesser degree in myo1d single mutants (Fig. 2b, Supplementary Fig. 2a). In further accordance with impaired Nodal signaling, myo1-deficient animals display a reduced expression of the Nodal-targets pitx2 and elovl6 in the LLPM (Fig. 2c, Supplementary Fig. 2b, c) and a reduced extension of the Nodal feedback inhibitor lft1 in the notochord that provides a molecular midline barrier for lateralized Nodal signaling (Fig. 2d).
a myo1g single and myo1d; myo1g double mutants fail to display asymmetric lft2 expression in the cardiac primordium. b-d myo1d/g mutants display a reduced anterior propagation of the expression of the Nodal ligand spaw (b, see also Supplementary Fig. 2a), the Nodal effector pitx2 (b, see also Supplementary Fig. 2b) and the Nodal feed-back inhibitor lft1 (d). e,e’ Left-sided expression of Spaw rescues the cardiac jogging (e) but not the cardiac looping (e’) defects of myo1g mutants. a Dorsal views at 22 somites, anterior up. b-d lateral views at 18 somites, anterior left, dorsal up. Scale bars: a 50 µm, b,c 100 µm.
Unilateral Nodal expression restores cardiac laterality in myosin1g mutant animals
If the LR asymmetry defects of myo1g mutant animals are due to impaired Nodal signaling, restoring left-sided Nodal signaling should allow to rescue cardiac laterality. To test this hypothesis, Spaw and GFP RNAs were co-injected into a single blastomere at the two cell stage. By the end of gastrulation, the GFP tracer allowed to select animals in which the progeny of the injected blastomere was restricted to either the left or the right side of the embryo. In accordance with a potential requirement for myo1g in Nodal signaling, left-sided Spaw expression allowed to increase the percentage of myo1g mutants that present a proper leftward cardiac jogging (Fig. 2e). In contrast, the chirality of cardiac looping, a process that subsequently generates the atrial and ventricular chambers and occurs largely independently of Nodal signaling41 was not restored by left-sided Spaw expression (Fig. 2e’). Similarly, myo1g mutants in which Spaw-injected cells ended up on the right side of the embryo failed to display a restoration of embryonic laterality (Supplementary Fig. 2d, d’).
The Left-Right organizer flow and myosin1 genes control Nodal propagation
Through its ability to promote the unilateral degradation of transcripts encoding the Nodal signaling antagonist Dand5, the LRO flow enables the left-sided propagation of nodal expression25-27. Our observation that Myo1G and (to a lesser degree) Myo1D act to promote propagation of spaw expression (Fig. 2, Supplementary Fig. 2) raises the question whether the enhanced laterality defects of embryos that lack both a LRO flow and myo1 gene function (Fig. 1) could be due to cumulative effects on nodal gene expression? To address this issue, we performed a comparative quantitative analysis of spaw expression in the Lateral Plate Mesoderm (LPM) of embryos that lack a LRO flow (due to dnaaf1 inactivation) as well as myo1d & g activities.
In wild-type control embryos, spaw extends anteriorly up to the level of the heart and brain primordia in the left LPM, while its expression is either entirely absent or only restricted to the posterior-most LPM on the right side of the embryo (Fig. 3a). Morpholino-mediated knock-down of dnaaf1 (Fig. 3a) or its genetic inactivation (Fig. 3b) cause a reduction in the left-sided extension of spaw which is likely due to a failure to downregulate dand5 on the left side of the LRO. Additionally, dnaaf1-deficient animals present a roughly symmetric expression of spaw in the right LPM (Fig. 3a, b). Simultaneous inactivations of myo1d/g and dnaaf1 cause a reduction of the anterior extension of spaw expression on both the left and the ride side the animal (Fig. 3a, b), demonstrating thereby that Myosin1 protein exert a flow-independent control of nodal ligand expression. Similar results were obtained using dnaaf1 morphants or mutants, although quantitatively stronger effects are observed upon use of stable genetic mutants compared to transient morpholino knock-down. In accordance with our morphological analysis of embryonic laterality that suggested a predominant role of Myo1G in the LRO flow-independent control of LR asymmetry (Fig. 1), the inactivation of myo1g has a stronger effect on spaw expression in the LPM of dnaaf1-depleted animals then the loss of function of myo1d (Fig. 3a, b).
a,b Quantification of spaw extension in the Left (green dots) and Right (red dots) LPM of 18 somites stage LRO flow-deficient dnaaf1 morphant (a) or dnaaf1 mutant (b) embryos. myo1d/g loss of function causes a significant reduction of the antero-posterior extension of spaw expression in both the Left and the Right LPM. c,c’ Double in situ hybridization for spaw and the cardiac marker cmlc2 (see also Supplementary Fig. 3) reveals that spaw expression reaches the cardiac primordium in most WT control and LRO flow-deficient dnaaf1 morphant (c) or dnaaf1 mutant (c’) embryos, but fails to do so upon inactivation of myo1g.
Nodal expression fails to reach the cardiac primordium in myosin1g mutants
Spaw-mediated Nodal signaling is required to transmit laterality information from the LRO to target tissues. The zebrafish LRO, Kupffer’s Vesicle11 is located at the posterior tip of the notochord. Among the different tissues undergoing chiral morphogenesis, the visceral organ primordia are closest to the LRO while heart and brain primordia are located more anteriorly at increasing distances. As our experiments show that myo1g mutants present no defects in visceral LR asymmetry but increasingly severe phenotypes in the more anterior heart and brain (Fig. 1a-c), we wondered whether the reduced extension of left-sided spaw expression (Fig. 2b) may result in a failure to reach more anteriorly located organ primordia. To test this hypothesis, we performed two colour in situ hybridization to simultaneously visualize spaw expression and the cmlc2-positive cardiac primordium.
Our analysis reveals that by the 22 somites stage, spaw expression has reached the cardiac primordium in most wild-type embryos (Fig. 3c, c’, Supplementary Fig. 3a, a’). Similarly, spaw extends up to the level of the heart primordium on either the left, the right or both sides of the embryo in most animals that are mutant or morphant for the LRO flow regulator dnaaf1 (Fig. 3c, c’, Supplementary Fig. 3a, a’). In contrast, spaw expression fails to reach the cardiac primordium in a significant fraction of myo1g mutants, providing thereby a potential explanation for their cardiac jogging defects (Fig. 3c, c’, Supplementary Fig. 3a, a’). Compound inactivations of dnnaf1 and myo1g result in near complete failure of spaw expression to reach the cardiac primordium (Fig. 3c, c’, Supplementary Fig. 3a, a’), in accordance with the predominant lack of cardiac jogging that is observed in these animals (Fig. 1a’).
myosin1g mutants display a temporal delay in spaw expression
To investigate the mechanism through which myosin1 genes contribute to the LRO flow-independent regulation of Nodal signaling, we performed a time-course analysis of spaw expression during development. As myo1d contributes to both the regulation of the LRO flow5 and the flow-independent control of nodal expression (Fig. 3), we focused our analysis on myo1g, which plays a predominant role in the flow-independent control of Nodal signaling (Fig. 1, Fig. 3).
In wild-type embryos, spaw expression is initiated bilaterally in the cells that surround the LRO by the 6 somites stage (Fig. 4a). As development proceeds, spaw LRO levels increase until at around the 12 somites stage expression also becomes detectable in the left LPM where the ligand then propagates through auto-induction to reach more anterior target tissues (Fig. 4a). Analysis of spaw expression in myo1g mutants revealed a reduction in the initial induction of spaw expression at the level of the LRO and the subsequent propagation to the LPM (Fig. 4a). In contrast to the loss of myo1g function, a lack of LRO-flow upon depletion of dnaaf1 is without effect on the initial induction of spaw expression at the LRO (Supplementary Fig. 4a).
a Time course analysis of spaw expression indicates that initiation of spaw expression at the LRO and propagation to the Left LPM (black arrowhead) are delayed in myo1g mutants. b,b’ qPCR analysis of spaw expression confirms that spaw expression is significantly reduced in myo1g mutants (b). Conversely, spaw expression increases in mutants for the myo1d/g antagonist myo1cb (b’, see also Supplementary Fig. 4b). c myo1g mutants present a reduced rate of anterior-ward propagation of lft1 expression in the notochord (see also Supplementary Fig. 6a). d,e myo1g mutants display a weaker induction of the nodal target gene lft1 in response to Spaw overexpression. While high amounts (20 pg) of Spaw RNA induce a similar lft1 induction in myo1g mutants and wild-type siblings, myo1g-deficient embryos present a reduced response to moderate amounts (10 pg) of Spaw RNA (d, ectopic expression indicated by arrow, see Supplementary Fig. 6b for quantification). qPCR analysis confirms that equal amounts of Spaw RNA induce a reduced lft1 induction response in myo1g mutants (e). f Conversely, the overexpression of Myo1G potentiates the capacity of low amounts (5 pg) of Spaw RNA to induce ectopic lft1. a vegetal views of the LRO, anterior up. d animal pole views. Scale bars: 100 μm.
The observation that myo1g mutants present a reduced spaw expression at the LRO was confirmed by quantitative qRT-PCR (Fig. 4b). Studies in Drosophila and zebrafish revealed that Myosin1C (Myo1C) proteins can act as Myo1D/G antagonists5,42. While our analysis failed to reveal any morphological LR asymmetry defects in maternal zygotic myo1Cb mutants, gene expression analysis uncovered a mild upregulation of spaw at the LRO (Fig. 4b’, Supplementary Fig. 4b), supporting the functional relationship between Myo1D/G agonists and their Myo1Cb antagonist.
myosin1g is dispensable for Left-Right Organizer formation
The finding that zebrafish myo1d is required for LRO morphogenesis5,6 raises the question whether the loss of myo1g may similarly cause general defects in LRO morphogenesis that would ultimately result in reduced Nodal signaling at the LRO. Our analysis of different markers genes involved in LRO specification and function does however not support this hypothesis. Analysis of the endodermal markers sox17 and sox32 indicates that the specification and clustering of LRO precursor cells occurs normally in myo1g mutants (Supplementary Fig. 5a, b). In accordance with the fact that myo1g controls LR asymmetry independently of the LRO flow, myo1g loss of function has no effect on the expression of the ciliary motility genes foxj1a, dnah9 and odad1 (Supplementary Fig. 5c, d, e).
Myosin1G promotes Nodal signaling
In mice and zebrafish, Nodal expression at the LRO is initially induced by Notch signaling43, and then further upregulated through the capacity of Nodal ligands to induce their own expression44. While our analysis of the Notch target genes her4.1 and her15.1 suggests that myo1g mutants present normal Notch signaling levels (Supplementary Fig. 5f, g), the reduced expression of spaw at the LRO (Fig. 4a) is similar to the one reported in spaw mutants44.
If myo1g mutants present a defect in Spaw autoinduction, this should result not only in an initial delay in the appearance of spaw expression in the LPM (Fig. 4a), but also in a delayed subsequent anterior propagation. Accordingly, time course analysis of lft1 expression at the notochordal midline barrier reveals that myo1g mutants present a nearly two-fold reduction in the rate of Nodal target gene propagation (Fig. 4c, Supplementary Fig. 6a).
As Nodal ligands induce their own expression23, the reduced propagation of spaw expression in myo1g mutants could be indicative of a defect in Nodal signal transduction. To establish if Myo1G is important for Spaw signaling, we misexpressed Spaw in germ ring stage embryos that lack endogenous spaw expression and analyzed its capacity to upregulate the Nodal target gene lft1. While high doses (20 pg) of Spaw readily induce ectopic lft1 expression in both WT and myo1g mutant embryos (Fig. 4d, Supplementary Fig.6b), myo1g mutants present a reduced response to moderate (10 pg) doses of Spaw RNA (Fig. 4d, Supplementary Fig. 6b). Analysis of lft1 expression by quantitative RT-PCR reveals that while Spaw is still able to significantly induce lft1 in myo1g mutants, the observed effect is weaker than in homozygous WT sibling controls (Fig. 4e, Cohen’s d effect size = 1.27 for myo1g mutants versus 4.48 for WT controls).
Taken together, our observations suggest that Myo1G, while not strictly required for Spaw signal transduction, is essential to promote full strength Nodal signaling. Injecting wild-type Myo1G RNA into myo1g mutants significantly rescues the capacity of Spaw to induce lft1 expression, demonstrating the specificity of the observed effect (Supplementary Fig. 6c). To confirm that Myo1G promotes Spaw signaling, a lower amount of Spaw RNA (5 pg), that is on its own barely capable of inducing ectopic lft1 expression, was co-injected with wild-type Myo1G RNA into WT animals. qRT-PCR analysis shows that Myo1G overexpression promotes the capacity of this subliminal amount of Spaw to induce lft1 expression (Fig. 4f).
myosin1g regulates Activin receptor trafficking
How does Myo1G promote Spaw signaling? Proteomic studies identified Myo1G on exosomes, suggesting that this factor may be implicated in exovesicular secretion45. To determine whether Myo1G is required for Spaw ligand secretion, we took advantage of a functional GFP-Spaw fusion construct that has previously been used to visualize Spaw secretion46. GFP-Spaw RNA injection into wild-type or myo1g mutant animals results in a similar labeling of the extracellular space (Fig. 5a, b), suggesting that Myo1G is unlikely to control Spaw ligand production and secretion.
a,b Spaw-GFP localization is similar in WT (n=20) and myo1g mutants (n=17). H2B-RFP was injected as a tracer to ascertain that embryos that had received equal amounts of RNA. c A constitutively activated form of the Nodal signal transducer SMAD2 (CA-SMAD2) elicits similar responses in WT and myo1g mutants. d Myo1G-GFP is detected at the cell cortex and in intracellular compartments (n = 10). e,e’,e’’ Myo1G-GFP is present on endosomes positive for the TGF® signaling adapter SARA (see also Supplementary Fig. 7a,b). f-j myo1g mutants present a reduced number of endosomes positive for the Nodal receptors Acvr2Aa-GFP (f,g,j) and Acvr2Ba-GFP (h,i,j). k,l myo1g mutants and WT siblings present a similar number of CD44a-positive endosomes (see also Supplementary Fig. 7c). m-p myo1g mutants present a lower number of endosomes that are positive for both AcvrIIAa and SARA (n). Conversely, myo1g loss of function causes an increase in the fraction of SARA-negative AcvrIIAa endosomes (p, arrowheads in m,n) despite having a similar number of SARA endosomes (Supplementary Fig. 7d). a,b,d-i,k-m,o animal pole views, germ ring stage. Data points in j,n,p represent the mean number of endosomes per cell for a particular embryo (see Supplementary material for complete statistical information). Scale bars: 10 μm.
While cytoplasmic Myo1C exerts important roles in membrane trafficking47, nuclear isoforms of mammalian Myo1C can regulate TGFβ-responsive gene expression48. To determine if Myo1G controls the SMAD-mediated transcriptional downstream response to Spaw signaling, we injected RNA encoding a Constitutively Activated variant of SMAD2 (CA-SMAD2) into wild-type and myo1g mutant animals and analyzed the effect on lft1 target gene induction by qRT-PCR. In contrast to the reduced induction of lft1 that is observed upon Spaw overexpression in myo1g mutants (Fig. 4e), CA-SMAD2 elicited a similar induction of lft1 expression in myo1g-deficient animals (Fig. 5c, Cohen’s d effect size = 3.39 for myo1g mutants versus 3.62 for WT controls).
The pharmacological Myosin antagonist Pentachloropseudilin (PCIP) inhibits TGFβ signaling by regulating the membrane trafficking of TGFβ type II receptors49. In accordance with a potential function in the membrane trafficking of cell surface receptors, Myo1G-GFP localizes to both the cell cortex and to intracellular, potentially endosomal, compartments (Fig. 5d).
As endosomal compartments positive for the TGFβ signaling adapter SARA promote Nodal signal transduction29, we investigated if Myo1G-GFP positive intracellular compartments correspond to SARA endosomes. Strikingly, use of an established mRFP-SARA construct50 revealed that in 24/24 embryos, SARA positive compartments were always associated with Myo1G-GFP. Both standard laser scanning microscopy (Fig. 5e) and Airyscan super-resolution microscopy (Supplementary Fig. 7a, b) revealed that SARA-positive compartments are often part of larger, Myo1G-positive structures, suggesting that Myo1G may be important for the biology of TGFβ signaling endosomes.
SARA endosomes promote Nodal signaling by providing a subcellular platform that enables TGFβ receptors to activate downstream SMADs29. As Myosins have been linked to TGFβ type II receptor trafficking in other biological contexts49, we analyzed the effect of myo1g loss of function on the two Nodal type II receptors AcvrIIAa and AcvrIIBa. GFP-tagged versions of the two proteins indeed revealed a significant reduction of the number of Activin receptor-positive endosomes in myo1g mutants (Fig. 5f-j). In murine lymphocytes, Myo1G regulates the endocytic trafficking of the adhesion protein CD44 51, a molecule that has, in other biological contexts, been shown to regulate TGFβ signaling by acting as Hyaluronan receptor 52. In contrast to the situation observed for AcvrIIAa & Ba, myo1g mutants and their wild-type siblings present similar numbers of CD44a-positive endosomes (Fig. 5k, l, Supplementary Fig. 7c), supporting the specificity of the observed Nodal receptor trafficking defects.
The observations that myo1g mutants present a reduced number of AcvrII endosomes (Fig. 5f-j) and that Myo1G is found on SARA-positive compartments (Fig. 5e, Supplementary Fig. 7a, b) raise the question whether Myo1G may be required for the formation of AcvrII/SARA-positive endosomes. Accordingly, myo1g mutants present both a reduction in the absolute number of SARA/AcvrIIAa-positive compartments (Fig. 5m-o) and a higher fraction of AcvrIIAa compartments that lack the signaling endosome marker SARA (Fig. 5m, n, p). In contrast, myo1g loss of function has no effect on the absolute number of SARA-positive endosomes (Supplementary Fig. 7d). Taken together, these findings suggest that Myo1G may ensure full strength Nodal signaling by promoting the formation of SARA/Nodal receptor-positive endosomes.
Myosin1G promotes Southpaw-independent Nodal signaling
As myo1g mutants present a normal expression of LRO specification and differentiation markers (Supplementary Fig. 5), but reduced levels of the TGFβ superfamily ligand spaw (Fig. 4), we investigated the effect of myo1g loss of function on the LRO expression of other TGFβ signaling components.
To induce a biological response, Spaw heterodimerizes with the TGFβ superfamily member GDF353. Our analysis shows that not only spaw itself, but also the expression of its partner gdf3 is reduced in myo1g mutants (Fig. 6a). The Cerberbus/Dan family protein Dand5 antagonizes Spaw signaling at the LRO24. myo1g mutants not only present lower levels of spaw/gdf3, but also diminished dand5 expression (Fig. 6b). Of particular interest, myo1g mutants display a reduced dand5 expression already at the tail bud stage, 2 hours before the onset of spaw expression (Fig. 6b), suggesting that the function of Myo1G is not limited to Spaw signaling.
a myo1g mutants present a reduced expression of the TGFβ ligands gdf3 and cyc and the Nodal feed-back antagonist lft1 in the LRO/tail bud region. b myo1g mutants display a reduced expression of the Nodal signaling antagonist dand5. c myo1g mutants display reduced lft1 induction in response to ectopic Cyc expression (arrow indicates ectopic expression). d qPCR indicates that myo1g mutants present a significant decrease in the endogenous expression levels of the Nodal target gene lft1. e 8 somites stage myo1g mutants present a reduction in the antero-posterior extension of lft1 expression in the anterior brain. a,b vegetal views of the LRO/tail bud region, anterior up. c animal pole views of germ ring stage embryos. e dorsal view of the anterior brain, anterior up. Scale bars: a 50 µm, b,c,e 100 µm.
Prior to spaw expression, the nodal ligand cyc and its target and feed-back inhibitor lft1 are expressed in the LRO region (Fig. 6a)33,36. The observation that myo1g mutants present a reduction in the early LRO expression of cyc and lft1 (Fig. 6a) suggests that, in addition to Spaw, Myo1G may potentiate signaling by other Nodal ligands. Accordingly, myo1g mutants display a reduced lft1 induction in response to ectopic Cyc (Fig. 6c) or Squint/Nodal-related-3 (Sqt/Ndr3, Supplementary Fig. 8), the third zebrafish nodal homologue.
While the morphological LR asymmetry defects of myo1g mutants could be solely due to defective Spaw signaling, quantitative analysis of the Nodal downstream gene lft1 provided evidence that Myo1G also promotes endogenous signaling by other Nodal ligands in different contexts. First, qRT-PCR revealed that the germ ring stage lft1 expression which is induced by Cyc and Sqt is significantly reduced in myo1g mutants (Fig. 6d). Second, 8 somites stage myo1g mutants present a reduced expression of lft1 in the anterior neurectoderm, which lacks spaw but expresses cyc36,37.
Taken together, our findings identify Myo1G as a novel positive regulator of the Nodal signaling pathway that is essential for LR asymmetry and potentiates responses elicited by different Nodal ligands in different biological contexts.
DISCUSSION
A striking feature of LR asymmetry is that different species use seemingly different mechanisms for the determination of this third body axis3. Only recently has the unconventional type I Myosin Myo1D, which was initially identified as a regulator of Drosophila laterality34,35, been identified as an evolutionarily conserved regulator of animal LR asymmetry4-7. While studies in fish and frogs uncovered an essential role of Myo1D in LRO morphogenesis4-6, several observations suggested that additional functions of Myosin1 proteins in LR asymmetry remain still to be uncovered. First, previous studies had identified a function of Myo1D in the central LRO of fish and frogs, a biological structure that has no equivalent in Drosophila, where Myo1D ensures a local, organ-specific control of chiral morphogenesis. Second, in contrast to the unique myo1d gene present in flies, vertebrate genomes harbor not only myo1d, but also its close homologue myo1g. We present an in-depth analysis of the function of this second myosin1 homologue in zebrafish and uncover a novel essential function of this gene in chiral morphogenesis that is different from the reported function of myo1d 4-6.
Myo1D controls the symmetry-breaking ciliary fluid flow in the central LRO4-6. Accordingly, myo1d loss of function causes defects in all lateralized organs4-6. In contrast, our findings show that myo1g is required for the chiral morphogenesis of the heart and brain, but dispensable for visceral laterality (Fig. 1, Supplementary Fig. 1). To specifically determine whether Myo1G exerts an LRO flow-independent function, we inactivated myo1g in the context of animals that lack the ciliary motility gene dnaaf1 and therefore have no LRO flow. Lack of a LRO flow in dnaaf1 mutants causes a distinctive randomization of LR asymmetry, in which the heart, brain and viscera develop either normally (situs solitus) or as their mirror image (situs inversus). In contrast, a different phenotype is observed in dnaaf1 myo1g double (or dnaaf1 myo1d myo1g triple mutants) in which cardiac and brain laterality are altogether lost (Fig. 1). These findings establish an essential role of Myo1G in the flow-independent, tissue-specific control of LR asymmetry.
How does Myo1G exert this tissue-specific control? Myo1G could be involved in the local, organ-specific execution of chiral morphogenesis, like Drosophila myo1d18,35. Alternatively, Myo1G could be involved in the transmission of laterality information from the central LRO to different target tissues. While our experiments do not allow to rule out the first possibility, myo1g mutants present a reduced propagation of the Nodal ligand spaw and a reduction in the expression of different Nodal target genes (Fig. 2, Supplementary Fig.2). myo1g mutant laterality defects can moreover be rescued through unilateral Spaw expression (Fig. 2e).
The observation that myo1g mutants can be rescued through Nodal overexpression shows that Nodal signaling is reduced but not abolished in these animals. Quantitative analysis of Nodal target indeed reveals a reduced responsiveness to Nodal ligands in mutant embryos (Fig. 4). In accordance with a disruption of Nodal signal transduction, Myo1G is found on endosomes that are positive for the TGFβ signaling adapter SARA and myo1g mutants present a reduced number of Nodal-receptor positive SARA endosomes (Fig. 5).
The observation that Myo1G is found on SARA endosomes and regulates Nodal receptor trafficking raises the question whether, beyond LR asymmetry, Myo1G may promote Nodal signaling in other biological contexts. In accordance with this hypothesis, our experiments show that myo1g mutants present a reduced responsiveness to the Nodal ligands Cyc and Sqt (Fig. 6c, Supplementary Fig. 8) and a reduction of endogenous Nodal target gene expression levels in domains that are independent from the spaw, the nodal gene entirely dedicated to LR asymmetry (Germ ring stage blastoderm margin, 8 somites stage forebrain, Fig. 6d, e).
Taken together, our findings identify Myo1G as a general positive regulator of Nodal signaling whose function is specifically required for LR asymmetry. Our work establishes for the first time a link between unconventional type 1 Myosins that are emerging as major regulators of animal laterality, and Nodal signalling which has long been known to be the key pathway regulating vertebrate LR asymmetry.
METHODS
Zebrafish strains and embryo maintenance
Embryos were raised in 0.3X Danieau medium (17.4mM NaCl, 0.21mM KCl, 0.12mM MgSO4, 0.18mM Ca (NO3)2, 1.5mM Hepes, pH 7.6) at 28.5 °C., and staged according to standard criteria54. If necessary, 1-phenyl-2-thiourea (Sigma) was added at 30 mg/l to prevent embryonic pigmentation.
myo1d/g inactivations were performed using the previously reported myo1dtj16b, myo1dtj16c and myo1gtj18b alleles5. All presented data were obtained using Maternal Zygotic (MZ) single or double mutants. Allele specific PCR was used to identify the WT myo1d allele (forward primer 5′-AGAGTGGAGCTGGAAAAACAGA-3′, reverse primer 5′-CCCATCCCTCGTGTGAAACTAAATCAC-3′, 339 bp amplicon) as well as the mutant alleles tj16b (forward primer 5′-TGGAGCTGGAAAAAGGCTCGT-3′, reverse primer 5′-CCATCACTGCAGCAGAAATGAGAG-3′, 133 bp amplicon) and tj16c (forward primer 5′-GTGGAGCTGGAAAAAGGCTATAC-3′, reverse primer 5′-CCATCACTGCAGCAGAAATGAGAG-3′, 145 bp amplicon). The allele-specific reverse primers 5′-TCTCATACAGTTCTCTTCCCCTAG-3′ (tj18b, 115 bp amplicon) and 5′-CTCATACAGTTCTCTTCCCCTGTAG-3′ (WT, 120 bp amplicon) were used with the generic forward primer 5′-GAGAAGAGTCGTATCTACACCTTC-3′ to genotype myo1g mutant fish.
myo1Cb inactivation was performed using the myo1Cbsa16637 allele from the Zebrafish Mutation Project (http://www.sanger.ac.uk/resources/zebrafish/zmp/) obtained from the Zebrafish International Resource Center. The sa16637 allele introduces a premature stop codon at the 228th amino acid position. A generic forward primer 5’-GTCACATCCTGAACTACCTGCTAG-3’ was used along with a mutant specific reverse primer 5’-TATTACCAGTATCTGGTCAAG-3’ (164 bp amplicon) to identify the mutant allele. The WT allele was identified using the generic forward primer with the WT specific reverse primer 5’-CAGTACCAGTATCTGGTCAAG-3’ (164 bp amplicon).
dnaaf1 was inactivated using the dnaaf1tm317bmutant allele38. The forward primer 5’-GCAAGCTTTGCACGCTTAATGTCTC-3’ and reverse primer 5’ - AACACTGGAGAATGTTTGTGAC - 3’ were used to amplify the tm317b mutant allele (199 bp amplicon). The dnaaf1 WT allele was identified using the forward primer 5’-GCAAGCTTTGCACGCTTAATGTCTC-3’ and reverse primer 5’- CACACTGGAGAATGTTTGTGAC-3’ (199 bp amplicon). Beyond 24 hrs of development, dnaaf1 mutants can be identified through the oval phenotype that is diagnostic for ciliary mutations.
Plasmid generation
The myo1g ORF was amplified from mixed stage pool of cDNAs using primers 5’-GATCCCATCGATTCGATGGCGGAGCTGGAGGGCTTG-3’ and 5’-AGGCTCGAGAGGCCTTACTGGGGCAGGAGTAAGG-3’ and cloned into the pCS2+ vector using Gibson assembly mix (NEB). Bold letters in the primer sequences indicate Gibson overhangs that are also present in the pCS2+ sequence. For generating the myo1g-GFP construct, the myo1g ORF was amplified from the myo1g-pCS2 construct using the primer pair 5’-GCAGGATCCCATCGATTCGACAGTAAACATGGCGGAGCTGGAGGGCTTG-3’ and 5’-ACCATGGACCCTCCGCTGGTGCCCTGGGGCAGGAGTAAGGTAAATC-3’, and was ligated onto pCS2-GFP.
CD44a was amplified using the primers 5’-ATCCCATCGATTCGACAGTAAACATGTGGACTTTGTTATTTGTAGTGTT-3’ and 5’-ACCATGGACCCTCCGCTGGTGCCCATTAAATATTCTTTTTCGTGTTCA-3’ and ligated into pCS2-eGFP. For acvr2aa, the following primer pairs were used 5’-GGATCCCATCGATTCGACAGTAAACATGGGACCTGCAACAAAGCTGGC-3’ and 5’-ACCATGGACCCTCCGCTGGTGCCTAGACTAGACTCCTTTGGGGGATA-3’, and for acvr2ba, the forward primer 5’-GATCCCATCGATTCGACAGTAAACATGTTCGCTTCTCTGCTCACTTTGG-3’ was used with the reverse primer 5’-ACCATGGACCCTCCGCTGGTGCCGATGCTGGACTCTTTGGGCGG-3’ to amplify the ORFs, which were ligated into pCS2-eGFP. The her4.1-pBSK construct used to generate in situ probe was cloned using the primer pair 5’-GTCGACGGTATCGATAAGCCACACAGCAATGACTCCTAC-3’ and 5’-CTAGAACTAGTGGATCCCCCTTAAGTCTACCAGGGTCTCC-3’. her15.1 was amplified using the forward primer 5’-GTCGACGGTATCGATAAGCGCTCAGAGAAACAGCATCTCTCC-3’ and reverse primer 5’-CTAGAACTAGTGGATCCCCCCTCCACAGGAGTTCAACATTGAC-3’ and cloned into pBSK. A Squint in situ probe was amplified using the primers 5’-CGAGGTCGACGGTATCGATAAGCACATGTTTTCCTGCGGGC-3’ and 5’-CTCTAGAACTAGTGGATCCCCCGTTTGAAGAATCAGTGGCAGC-3’ and cloned into pBSK.
RNA and Morpholino injections
mRNAs were synthetized using the SP6 mMessage mMachine kit (Ambion). RNAs were diluted in 0.1M KCl 0.2% Phenol Red. The following constructs and quantities were used: Acvr2Aa-GFP-pCS2+ (25 pg, this study), Acvr2Ba-GFP-pCS2+ (25 pg, this study), CA-SMAD2-pCS2+ (20pg55). CD44a-GFP-pCS2+ (50 pg, this study). GFP-Spaw-pCS2+ (20pg46). Histone2B-mRFP-pCS2+ (12.5 pg56). mRFP-SARA-pCS2+ (25 pg50). Myo1G-pCS2+ (50pg, this study). Myo1G-GFP-pCS2+ (50pg, this study). For Spaw57, Cyclops58 and Squint36 different concentrations used in individual experiments are indicated in the figures.
The previously reported dnaaf1 Morpholino 5’-ATGCACTGTAATTTACCAAGTCAGG-3’40 was injected at a concentration of 500 µM diluted in 1x Danieau 0.2% Phenol Red.
For rescuing the cardiac jogging defects of myo1g mutant by Spaw mRNA injection, a mix of Spaw and GFP RNA was co-injected into one blastomere of two cell stage embryos. At bud stage embryos with a unilateral segregation of GFP expressing cells were selected using a fluorescent dissection scope (Leica M205FA), and grown further to score for cardiac jogging and looping phenotypes.
RNA in situ hybridization
Whole mount RNA in situ hybridizations were performed as previously described (Thisse and Thisse, 2008). For the following genes probes were transcribed from previously reported plasmids: spaw-pGEMT57, lefty1-pBSK31, lefty2-pBSK33, pitx2c-pBSK59, foxa1-pBSK60, cyclops-pBSK36, squint-pBSK (this study), dand5-pBSK24, sox32-pBSK61, sox17-pBSK62, odad1-pME18S-FL363, dnah9-pCRII64, foxj1a-pBSK65, cmlc2-pCS266, gdf3-pBSK63, Her4.1-pBSK (this study), Her15.1-pBSK (this study). The elovl6 probe was transcribed from a PCR product containing a T7 promotor sequence at the 3’ end. elovl6 was amplified from genomic DNA using the forward primer 5’–CCCGTCCCATGTGCAGAACATTG–3’ and the reverse primer 5’–GGTGTCCATTGTGCTCGTGTGTCTCCCTATAGTGAGTCGTATTACGC– 3’.
qPCR analysis
qPCR was performed using PowerUP SYBR Green Master Mix (Applied Biosystems) in an Applied Biosystems Step-One PCR system. Individual reactions were performed in triplicates to account for pipetting errors. For sample preparation, whole cell mRNA was isolated from 50 embryos using TRI-Reagent (Sigma). Reverse transcription was performed on 2.5 µg of RNA using Superscript III (Invitrogen) to generate cDNA. Fold changes in gene expression were normalized to the internal control gene 36b4. The primers used for the amplification reactions are as follows: lefty1: forward 5’-AGAGGAGTTTGGGTCTAGTGG-3’, reverse 5’-TACGGAGAGAGGAAATGCG-3’. Spaw: forward 5’-TGACTTCGTCCTGAGCTTGA-3’, reverse 5’-TCAAGCTCAGGACGAAGTCA - 3’. 36b4: forward 5-ACGTGGAAGTCCAACTACT-3’, reverse: 5’-GTCAGATCCTCCTTGGTGA-3’. For estimating relative gene expression, the Ct values at 40 cycles of qPCR amplification were used according to the ΔΔCT method67. Individual data points in figures documenting qPCR experiments correspond to technical replicates. Complete statistical information including numbers of biological and technical replicates is provided in the Supplementary statistical information file.
Immunocytochemistry
Dechorionated embryos were fixed for 1.5 hours at Room temperature in PEM (80 mM Sodium-Pipes, 5 mM EGTA, 1 mM MgCl2) - 4% PFA - 0.04% TritonX100 and then washed 2 × 5 min in PEMT (PEM - 0.2% TritonX100), 10 min in PEM 50 mM NH4Cl, 2 × 5 min in PEMT.
Microscopy and image analysis
Embryos were mounted in 0.75% low melting agarose (Sigma) in glass bottom dishes (Mattek) for confocal imaging. Imaging was performed using Laser scanning confocal microscopes (Zeiss LSM710, 780 and 880) using 40 x Water immersion (NA1.1) objectives. Airyscan super-resolution imaging was performed on a Zeiss LSM 880 system. In situ hybridizations were documented on a Leica M205 microscope with a Lumenera Infinity camera. Image analysis was performed using ImageJ (http://rbs.info.nih.gov/ij/).
Statistical Analyses
Appropriate statistical tests were selected for each experiment based on the nature of the comparison (bi-or multifactorial, ordinal or categorical data), data distribution and variance. Statistical analysis and representations were performed using R. Complete informations regarding the applied statistical tests, test statistics, sample sizes and displayed error bars for all experiments is provided in the supplementary statistical information file.
AUTHOR CONTRIBUTIONS
The genetic analysis of myosin1 function in zebrafish Nodal signaling was entirely performed by A.J.K. F.B. generated reagents used in the present study. M.F. designed the study, performed experiments, analyzed the data and wrote the manuscript.
COMPETING INTERESTS
The authors declare no competing interests.
Supplementary Information
a,b myo1g loss of function impairs the asymmetric expression of the Nodal ligand cyc (a) and the Nodal target gene lft1 (b) in the dorsal epithalamus. a,b Frontal views of the brain, dorsal up.
a-c myo1g mutants display a reduced expression of the Nodal ligand spaw (a), the Nodal effector pitx2 (b) and the Nodal target gene elovl6 (c) in the Left Lateral Plate Mesoderm (LLPM, black arrows). d,d’ Right-sided expression of Spaw rescues neither the cardiac jogging (d) nor the cardiac looping (d’) defects of myo1g mutants. The embryos displayed in d,d’ are derived from the same series of experiments also used in Fig. 2e,e’. a-c Dorsal views of 22 somites stage embryos, anterior up.
a,a’ Double in situ hybridization for spaw (purple) and the cardiac marker cmlc2 (red) shows that spaw expression reaches the cardiac primordium on the left, right or both sides of the embryo in most WT control and LRO flow-deficient dnaaf1 morphant (a) or dnaaf1 mutant (a’) embryos. In contrast, spaw expression fails to reach the cardiac primordium in a large fraction of myo1g single mutants (a,a’), myo1g mutant dnaaf1 morphants (a) or myo1g dnaaf1 double mutants (a’). The dataset used in this figure is the same that is also displayed in Fig. 3c, c’. Dorsal views of 22 somites stage embryos, anterior up.
a The 8 somites stage expression of spaw in the LRO region is unaffected by the loss of the LRO flow (dnaaf1 morphants), but reduced in myo1g mutants. b Conversely, 6 somites stage mutants for the myo1d/g antagonist myo1Cb present a premature expression of spaw in the LRO region. Vegetal views of the LRO region, anterior up. Scale bars: 100 µm.
a In situ hybridization for the endodermal marker sox17 shows that myo1g loss of function has no effect on the specification and behavior of the dorsal forerunner cells (black arrows) that are the precursors of the zebrafish LRO. b Accordingly, myo1g mutants present a normal expression of the endodermal marker sox32 in the LRO/Kupffer’s Vesicle of 8 somites stage embryos. c-e The ciliary motility genes foxj1a (c), dnah9 (d) and odad1 (e) are expressed normally in myo1g mutant LROs. f,g Analysis of the Notch signaling targets her4.1 (f) and her15.1 (g) indicates that myo1g loss of function has no effect on Notch signaling at the LRO. a Dorsal views, anterior up. b-g Vegetal views of the LRO region, anterior up. Scale bars: a,f,g 100 µm. b 20 µm. c-e 50 µm.
a In situ hybridization shows that myo1g mutants present a reduced anterior propagation of the Spaw target gene lft1. Lateral views, anterior to the left. Scale bar: 100 μm. The displayed embryos are part of the data set used to estimate the rate of lft1 propagation in Fig. 4c. b myo1g mutants display a weaker induction of the nodal target gene lft1 in response to Spaw overexpression. Quantification for the experimental dataset displayed in Fig. 4d. c Wild-type Myo1G RNA injection significantly enhances the capacity of Spaw to induce lft1 expression in myo1g mutants.
a,b Airy scan super-resolution microscopy indicates that the signaling endosome marker SARA localizes to sub-domains of large Myo1G-positive compartments. Animal pole views of germ ring stage WT embryos. Scale bars: 5 µm. c myo1g mutants and their WT siblings display a similar number of CD44a endosomes. Quantification of the dataset displayed in Fig. 5k,l. d myo1g mutants and their WT sibling display a similar number of SARA endosomes. Quantification of the dataset displayed in Fig. 5m-p.
myo1g mutants display a weaker induction of the nodal target gene lft1 in response to Sqt overexpression. While high amounts (5 pg) of Sqt induce similar ectopic lft1 expression, a reduced lft1 induction is observed in response to moderate amounts (1.2 pg) of Sqt RNA.
SUPPLEMENTARY STATISTICAL INFORMATION
Complete statistical information for the experiments reported in different display items
The table indicates the number of embryos in which spaw reaches the cardiac primordium in the Left LPM (Left), Right LPM (Right), Left and Right LPM (Bilateral), remains posterior to the primordium (Posterior) or is altogether absent (Absent), as displayed in Supplementary Fig. 3a. For statistical analysis, the former three and latter two categories are separately pooled into two categories for which spaw either reaches (Sum reach) or does not reach (Sum not reach) the primordium, as displayed in Fig. 3c.
The table indicates the number of embryos in which spaw reaches the cardiac primordium in the Left LPM (Left), Right LPM (Right), Left and Right LPM (Bilateral), remains posterior to the primordium (Posterior) or is altogether absent (Absent), as displayed in Supplementary Fig. 3a’. For statistical analysis, the former three and latter two categories are separately pooled into two categories for which spaw either reaches (Sum reach) or does not reach (Sum not reach) the primordium, as displayed in Fig. 3c’.
Supplementary Figure 2d: Cardiac jogging and looping in Spaw-injected myo1g mutants
Embryos expressing Spaw on the Left (Fig. 2e) and Right (Supplementary Fig. 2d) were generated in the same series of experiments. Embryo numbers and statistics for this Supplementary Fig. 2d are therefore displayed in the table provided for Fig. 2e.
Supplementary Figure 3a: southpaw propagation to the cardiac primordium in dnaaf1 Morpholino-injected myo1g mutants
Displayed embryos belong to the same data set also displayed in Fig. 3c. Embryo numbers are therefore displayed in the table provided for Fig. 3c.
Supplementary Figure 3a’: southpaw propagation to the cardiac primordium in dnaaf1 myo1g single and double mutants
Displayed embryos belong to the same data set also displayed in Fig. 3c’. Embryo numbers are therefore displayed in the table provided for Fig. 3c’.
ACKNOWLEDGEMENTS
This study was supported by an ARC project grant (PJA20181208167) and the ANR DroZeMyo (ANR-17-CE13-0024-02) (MF). AJK benefited from a 4th year PhD fellowship from La Ligue Contre le Cancer. Confocal microscopy was performed with the help of the iBV PRISM imaging platform. We thank M.Gonzalez-Gaitan, C.P.Heisenberg, T.Lepage, S.Lopes, and C. & B.Thisse for the sharing of reagents. We are grateful to S.Polès, R.Rebillard and G.Dupuy for technical assistance and excellent fish care.
