ABSTRACT
In Drosophila, neural stem cells or neuroblasts (NBs) acquire different identities according to their site of origin in the embryonic neuroectoderm. Their identity determines the number of times they will divide and the types of daughter cells they will generate. All NBs divide asymmetrically, with type I NBs undergoing self-renewal and generating another cell that will divide only once more. By contrast, a small set of NBs in the larval brain, type II NBs, divides differently, undergoing self-renewal and generating an intermediate neural progenitor (INP) that continues to divide asymmetrically several more times, generating larger lineages. In this study, we have analysed the origin of type II NBs and how they are specified. Our results indicate that these cells originate in three distinct clusters in the dorsal protocerebrum during stage 12 of embryonic development. Moreover, it appears that their specification requires the combined action of EGFR signalling and the activity of the related genes buttonhead and Drosophila Sp1. In addition, we also show that the INPs generated in the embryo enter quiescence at the end of embryogenesis, resuming proliferation during the larval stage.
INTRODUCTION
Drosophila neuroblasts (NBs) represent an excellent experimental model for studying neural stem cell (NSC) biology. Indeed, the study of Drosophila NBs has contributed significantly to our understanding of asymmetric cell division and cell fate specification, particularly regarding the temporal factors and quiescent stages involved, etc. (reviewed by Homem and Knoblich, 2012). One possible limitation of Drosophila NBs as an experimental model is their simplicity, particularly when compared with the complexity in the mammalian cerebral cortex, where NSCs use highly sophisticated mechanisms to generate the huge number of neurons that constitute the grey matter of the mammalian brain (reviewed by Lodato and Arlotta, 2015). However, the recent discovery of Drosophila type II NBs extends this model and its use in analysing how the diverse cell types found in central nervous systems (CNSs) are generated.
Drosophila NBs divide asymmetrically, generating a self-renewing cell and an intermediate cell that then divides once more to generate two cells that will differentiate into neurons or glial cells (type I NBs). More recently, a small set of NBs was identified in the Drosophila brain, type II NBs, and when these type II NBs divide they also generate a self-renewing cell and an intermediate neural progenitor (INP). After maturation, this latter cell undergoes several asymmetric divisions (Bello et al., 2008; Boone and Doe, 2008; Bowman et al., 2008) and, therefore, type II NBs generate much larger lineages than the type I NBs, giving rise to about one quarter of the neurons in the adult Drosophila brain. However, what is most interesting is that type II NB division resembles that of the radial glial cells in the outer subventricular zone of the mammalian cerebral cortex (Haubensak et al., 2004; Noctor et al., 2004). In addition, genetic analysis of INP maturation identified several genes that, when mutated, generate large malignant tumours in the Drosophila brain (Bello et al., 2006; Betschinger et al., 2006; Eroglu et al., 2014; Lee et al., 2006; Wang et al., 2009; Weng et al., 2010). Thus, the study of Drosophila type II NBs promises not only to improve our understanding of how the complexity of the mammalian cerebral cortex is generated, but also the origin of malignant brain tumours.
There are two stages of neurogenesis in the Drosophila CNS and the NBs generated during embryonic neurogenesis produce the CNS required for larval life. At the end of embryogenesis, most of the embryonic NBs that enter into a quiescent state later resume proliferation during larval neurogenesis, generating the more complex CNS required for adult life. Recent studies have addressed several aspects of the biology of type II NBs during larval development (Jiang and Reichert, 2014). However, very little is known of the origin of type II NBs or their behaviour during embryonic development. Hence, several interesting questions arise: are these NBs generated in the embryo or in the early larva?; what genetic code specifies type II as opposed to type I NBs?; do these NBs divide and generate INPs as do larval type II NBs?; do they express the same set of temporal factors as the larval type II NBs?; and do embryonic INPs die at the end of embryogenesis or do they become quiescent and resume proliferation in the larva?
In this study, we have addressed all these questions and have found that type II NBs are generated in three clusters during stage 12 of embryogenesis. Their specification requires the combined action of the EGFR pathway, and the related buttonhead (btd) and Drosophila Sp1 (the Drosophila homologue of the human Sp1 gene) genes. Moreover, these NBs generate INPs that express the same temporal factors as those expressed by INPs during larval development. Finally, we show that embryonic INPs become quiescent at the end of embryogenesis and that they resume proliferation in early larval development.
RESULTS
Drosophila type II NBs differ from type I NBs in that they do not express the proneural gene asense (ase) (Bowman et al., 2008). Thus, type II NBs were identified using either double immunostaining with antibodies against Deadpan (Dpn, a pan-neuroblast marker) and Ase, or with the pan-neuroblast driver worniu (wor)-Gal4 and ase-Gal80 (Neumüller et al., 2011). Indeed, this latter combination has been used previously to identify type II NBs in larval brains (Neumüller et al., 2011). The use of wor-Gal4 and ase-Gal80 in combination with 20xUAS-6xCherry (hereafter referred to as wach) results in strong Cherry expression in type II NBs. To confirm this method is valid to identify type II NBs during embryogenesis, we labelled wach embryos with antibodies against Dpn and Ase. All the Cherry+ NBs and no other NBs were Dpn+/Ase–, and therefore type II NBs (Fig. S1 and Table S1). Although wach labelling enables type II NBs to be unequivocally identified, the main limitation of this construct is that whereas Wor is detected from the time of NB delamination, the Cherry marker is not expressed until stage 13.
Type II NBs are generated in stage 12 embryos
There are eight type II NBs present in each lobe of the larval brain, six in the dorsomedial domain and two in a dorsolateral region (Bello et al., 2008). We first investigated whether NBs are present in the embryo and, if so, where and when they are generated during embryonic neurogenesis. As such, we labelled wach embryos with antibodies against Dpn and Ase, identifying type II NBs at different embryonic stages. Some type II NBs were seen in a dorsomedial position of the brain at stage 11 (Fig. 1A,A′); by stage 12, the full set of eight type II NBs (Fig. 1B-B‴) was evident, grouped into three clusters containing three, three and two NBs. Later in development, these three clusters became more conspicuous as anterior dorsomedial (aDM), posterior dorsomedial (pDM) and dorsolateral (DL) clusters, the latter located in a more ventral position. Therefore, we conclude that type II NBs are generated in the anterior-most segment of the brain (protocerebrum), although we failed to observe a previously reported placode-like structure in this area (de Velasco et al., 2007; Urbach and Technau, 2004).
Generation of Drosophila type II NBs during embryonic development. (A-F′) Dorsal views of Drosophila embryo heads at early stage 11 (A,A′), stage 12 (B-B‴), stage 13 (C,C′), stage 15 (D,D′), stage 16 (E,E′) and stage 17 (F,F′). For each stage: brain is indicated in grey; blue circles indicate the type II NBs labelled for Dpn and Ase; red circles indicate type II NBs observed with wach. Stage 11 (A,A′) and 12 (B-B‴) wach embryos were stained for Ase (green) and Dpn (blue), early stages at which wach is not yet expressed. Type II NBs (arrowheads) lie close to the ventral brain midline, which appears to be dorsal at this stage due to involution of the head segments. (B′-B‴) Sequential focal planes in the z-axis are shown and the numbers indicate the location of the NBs in the brain rather than the sequence of NB delamination. (C-F′) The expression of wach starts at stage 13 and from then onwards, type II NBs and their progeny can be easily identified. The persistence of strong Cherry expression enables neurons and their axons to be visualised. (G) Clusters of Cherry-expressing cells in a wach first instar larva. The progeny of the eight type II NBs are grouped into three clusters of NBs: NBs 1-3, aDM cluster; NBs 4-6, pDM cluster; and NBs 7 and 8, DL cluster. The yellow lines indicate the brain midline and anterior is on the left in all the embryos shown. A, anterior; P, posterior.
Generation of Drosophila type II NBs during embryonic development. (A-F′) Dorsal views of Drosophila embryo heads at early stage 11 (A,A′), stage 12 (B-B‴), stage 13 (C,C′), stage 15 (D,D′), stage 16 (E,E′) and stage 17 (F,F′). For each stage: brain is indicated in grey; blue circles indicate the type II NBs labelled for Dpn and Ase; red circles indicate type II NBs observed with wach. Stage 11 (A,A′) and 12 (B-B‴) wach embryos were stained for Ase (green) and Dpn (blue), early stages at which wach is not yet expressed. Type II NBs (arrowheads) lie close to the ventral brain midline, which appears to be dorsal at this stage due to involution of the head segments. (B′-B‴) Sequential focal planes in the z-axis are shown and the numbers indicate the location of the NBs in the brain rather than the sequence of NB delamination. (C-F′) The expression of wach starts at stage 13 and from then onwards, type II NBs and their progeny can be easily identified. The persistence of strong Cherry expression enables neurons and their axons to be visualised. (G) Clusters of Cherry-expressing cells in a wach first instar larva. The progeny of the eight type II NBs are grouped into three clusters of NBs: NBs 1-3, aDM cluster; NBs 4-6, pDM cluster; and NBs 7 and 8, DL cluster. The yellow lines indicate the brain midline and anterior is on the left in all the embryos shown. A, anterior; P, posterior.
Cherry expression was first detected in early stage 13 embryos (Fig. 1C,C′), and its strong expression meant that the Cherry gene product can be inherited by most, if not all, the progeny of the NB. Therefore, this marker allows the expansion of the clusters to be tracked and the axon projections of the neurons generated to be visualised (Fig. 1C-F′). The three clusters identified remained compact and clearly evident. Moreover, at the end of embryogenesis a neuronal projection from the DL cluster formed a fascicle that bifurcated, with one branch joining the aDM cluster and the other joining projections from the pDM cluster. Both branches projected their axons into the interhemispheric commissure (Fig. 1G and Movie 1).
EGFR signalling specifies type II NBs
In the ventral nerve cord, Epidermal growth factor receptor (EGFR) signalling is active in the ventral-most neuroectoderm prior to NB specification, propagating lateral signals that specify different fates in a concentration-dependent manner (Skeath, 1998). In the brain, EGFR activation has been visualized by labelling the phosphorylated form of mitogen-activated protein kinase (pMAPK), which is active at early developmental stages in the ventral neuroectoderm and the adjacent mesoderm (Dumstrei et al., 2002). From stage 10 onwards, pMAPK is confined to small cell clusters in the ventral-most neuroectoderm (Jussen et al., 2016) and as this is where type II NBs appear, we asked whether the EGFR pathway is involved in specifying these NBs.
We first examined the presence of type II NBs in wach embryos that carried mutations in: (1) the EGFR ligand encoded by the spitz (spi) gene; (2) the protease required for Spi secretion encoded by the rhomboid (rho) gene; and (3) the transcription factor targeted by the EGFR pathway, encoded by the pointed P1 (pntP1) gene. Significantly, in all three mutants virtually no type II NBs formed (Fig. 2A-F″, Fig. S2 and Table S1). In order to suppress any effects caused by global changes in these mutant embryos, we examined the effects of silencing the EGFR pathway using a neuroectoderm-specific driver [scabrous (sca)-Gal4] and either: (1) a dominant-negative form of EGFR (sca-Gal4 UAS-EGFRDN); or (2) a constitutively active form of the yan (also known as anterior open) gene (sca-Gal4 UAS-yanCA), a transcription factor that represses EGFR targets. Activation of the EGFR pathway phosphorylates Yan, which is then degraded (Rebay and Rubin, 1995); thus, misexpression of yanCA should abolish the downstream effects of EGFR signalling. Most, if not all, type II NBs were absent after silencing EGFR signalling using either method (Fig. S3 and Table S1). Finally, we evaluated what effect the activation of the EGFR pathway had by analysing PntP1 expression in the neuroectoderm of late stage 11 embryos. In the neuroectoderm, PntP1 was expressed by a strip of cells that extended along the midline of the protocerebrum, the same area where type II NBs appeared, and this expression was maintained at later stages (Fig. 2G-H‴). We did not observe PntP1 expression in type I NBs. In the light of these data, we conclude that activation of the EGFR pathway by the Spi ligand in the dorsal-most neuroectoderm of the protocerebrum is required to specify type II NBs.
Specification of type II NBs requires activation of the EGFR pathway. (A) Diagram of the positions of the three clusters of Cherry+ type II NBs in a stage 16 wach embryo. (B-D) Cherry expression in stage 16 wach (B), wach rho (C) and wach pntP1 (D) embryos. Most type II NBs are absent in C and D. (E-F″) Ase (green) and Dpn (red) expression in stage 15 wild-type (E) and spi1 (F-F″) embryos. In E, several type II NBs can be observed in this focal plane (arrowheads), whereas no type II NBs are detected in spi mutants. (G-H‴) Expression of PntP1 (green), Ase (red) and Dpn (blue) in late stage 11 (G-G‴) and stage 14 (H-H‴) wild-type embryos. There is a strip of neuroectodermal cells expressing PntP1 in the area where type II NBs appear (arrowheads in G,G″ and H,H″). For a more precise description of the genotypes see Materials and Methods. The white lines indicate the midline of the brain: A, anterior; P, posterior; white arrowheads (B), aDM and pDM clusters; arrow, DL cluster. Anterior is on the left in all the embryos shown.
Specification of type II NBs requires activation of the EGFR pathway. (A) Diagram of the positions of the three clusters of Cherry+ type II NBs in a stage 16 wach embryo. (B-D) Cherry expression in stage 16 wach (B), wach rho (C) and wach pntP1 (D) embryos. Most type II NBs are absent in C and D. (E-F″) Ase (green) and Dpn (red) expression in stage 15 wild-type (E) and spi1 (F-F″) embryos. In E, several type II NBs can be observed in this focal plane (arrowheads), whereas no type II NBs are detected in spi mutants. (G-H‴) Expression of PntP1 (green), Ase (red) and Dpn (blue) in late stage 11 (G-G‴) and stage 14 (H-H‴) wild-type embryos. There is a strip of neuroectodermal cells expressing PntP1 in the area where type II NBs appear (arrowheads in G,G″ and H,H″). For a more precise description of the genotypes see Materials and Methods. The white lines indicate the midline of the brain: A, anterior; P, posterior; white arrowheads (B), aDM and pDM clusters; arrow, DL cluster. Anterior is on the left in all the embryos shown.
To confirm these conclusions, we analysed the presence of type II NBs after ectopic EGFR activation in: (1) yan mutant embryos (wach yan); (2) embryos misexpressing a constitutively activated EGFR isoform (sca-Gal4 UAS-EGFRCA); and (3) embryos misexpressing an active form of the small GTPase-binding protein ras (sca-Gal4 UAS-rasV12). More type II NBs were detected in wach yan embryos and the three wild-type clusters were apparently fused (Fig. 3A-C′ and Table S1). Although wach is not normally expressed in stage 12 embryos due to a delay in Cherry activation, in stage 12 wach yan embryos a single cluster of Cherry+ cells was evident in each hemilobe, each containing an average of five type II NBs (data not shown). This average increased to 9.8 in stage 13 embryos, whereas in wach embryos there was only a mean of 6.2 type II NBs at this stage. Moreover, in sca-Gal4 UAS-EGFRCA embryos there was a mean of 10 type II NBs at stage 13 and in sca-Gal4 UAS-rasV12 embryos, a mean of 21.3 type II NBs at stage 13 (Fig. 3D-F″ and Table S1).
Ectopic activation of EGFR produces additional type II NBs. (A) Diagrams showing the dorsal (A) and lateral (A′) views of an embryo head. The central nervous system is shown in grey and the blue square indicates the area shown in B′ and C′. (B-C′) Dorsal (B,C) and lateral (B′,C′) views of stage 16 wach (B,B′) and wach yan (C,C′) embryos. All three clusters of Cherry+ cells appear to be fused together and to contain more cells. (D-F″) Ase (green) and Dpn (blue) expression in stage 15 wild-type (D-D″), sca-Gal4 UAS-EGFRCA (E-E″) and sca-Gal4 UAS-rasV12 (F-F″) brains. Whereas single type II NBs were observed in wild-type brains (D-D″), there were clusters of type II NBs in E-F″. The areas in which additional type II NBs appear (red arrowheads) are the same in the yan mutant (C), although this effect does not extend to the whole brain. The white lines indicate the brain midline: white arrowheads, aDM and pDM clusters; arrows, DL clusters; A, anterior; P, posterior; D, dorsal; V, ventral. Anterior is on the left in all the embryos shown.
Ectopic activation of EGFR produces additional type II NBs. (A) Diagrams showing the dorsal (A) and lateral (A′) views of an embryo head. The central nervous system is shown in grey and the blue square indicates the area shown in B′ and C′. (B-C′) Dorsal (B,C) and lateral (B′,C′) views of stage 16 wach (B,B′) and wach yan (C,C′) embryos. All three clusters of Cherry+ cells appear to be fused together and to contain more cells. (D-F″) Ase (green) and Dpn (blue) expression in stage 15 wild-type (D-D″), sca-Gal4 UAS-EGFRCA (E-E″) and sca-Gal4 UAS-rasV12 (F-F″) brains. Whereas single type II NBs were observed in wild-type brains (D-D″), there were clusters of type II NBs in E-F″. The areas in which additional type II NBs appear (red arrowheads) are the same in the yan mutant (C), although this effect does not extend to the whole brain. The white lines indicate the brain midline: white arrowheads, aDM and pDM clusters; arrows, DL clusters; A, anterior; P, posterior; D, dorsal; V, ventral. Anterior is on the left in all the embryos shown.
These results confirm our hypothesis that Spi ligand-activated EGFR signalling is required to specify type II NBs. Nevertheless, although the EGFR pathway was activated in the whole neuroectoderm, not all NBs adopted a type II identity, suggesting that additional factors are required.
The related buttonhead and Sp1 genes are required to specify type II NBs
The buttonhead (btd) gene and the adjacent related Sp1 gene encode zinc-finger proteins with partially redundant functions (Schöck et al., 1999; Wimmer et al., 1996). In larval type II NBs, Btd promotes the generation of INPs (Xie et al., 2014), although its role in the development of the embryonic brain has not yet been studied. We investigated whether btd and/or Sp1 are needed to specify type II NBs, first examining their expression in the embryonic brain. As antibodies against these transcription factors were not available, we used the btd-Gal4 and Sp1:green fluorescent protein (GFP) reporter lines. Reporter expression was very similar in both lines (btd-Gal4 UAS-Cherry and Sp1:GFP) and they were expressed in the dorsal-most neuroectoderm of stage 11 embryos, a signal that was maintained in later stages (Fig. 4A-E″) and was reminiscent of PntP1 expression (Fig. 2G-H‴). We tested the role of btd and Sp1 in embryos carrying a small genetic deletion that removes both genes, and then in embryos mutant for either btd or Sp1. Most type II NBs are missing in wach btd Sp1 (Df(1) btd Sp1; wach/+) embryos and similar phenotypes were obtained for the mutants of each single gene (btdXG or Sp1HR; Fig. 4F-H and Table S1). Thus, btd and Sp1 expression in the brain neuroectoderm appears to be required to specify type II NBs.
buttonhead and Sp1 are required to specify type II NBs. (A-D) Expression of Ase (green) and Cherry (red) in stage 11 (A,B) and stage 12 (C,D) btd-Gal4 UAS-Cherry embryo brains: (A,C) dorsal views; (B,D) lateral views. (E-E″) GFP (green), Cherry (red) and Dpn (blue) expression in a stage 15 wach Sp1:GFP embryo. (F) Cherry expression in a stage 15 Df(1)btd Sp1; wach embryo where Cherry is mostly absent. (G-J) Ase (green) and Dpn (red) expression in btdXG (G), Sp1HR (H), sca-Gal4 UAS-Sp1 (I) and sca-Gal4 UAS-btd UAS-Sp1 (J) embryos. Magnification cation of the boxed areas in G,H,J are shown on the right of each figure. Asterisks indicate type II NBs that persist in these genotypes (G,H) or extra type II NBs (J). The aDM, pDM and DL are shown at higher magnification in I. Anterior is on the left in all the embryos shown.
buttonhead and Sp1 are required to specify type II NBs. (A-D) Expression of Ase (green) and Cherry (red) in stage 11 (A,B) and stage 12 (C,D) btd-Gal4 UAS-Cherry embryo brains: (A,C) dorsal views; (B,D) lateral views. (E-E″) GFP (green), Cherry (red) and Dpn (blue) expression in a stage 15 wach Sp1:GFP embryo. (F) Cherry expression in a stage 15 Df(1)btd Sp1; wach embryo where Cherry is mostly absent. (G-J) Ase (green) and Dpn (red) expression in btdXG (G), Sp1HR (H), sca-Gal4 UAS-Sp1 (I) and sca-Gal4 UAS-btd UAS-Sp1 (J) embryos. Magnification cation of the boxed areas in G,H,J are shown on the right of each figure. Asterisks indicate type II NBs that persist in these genotypes (G,H) or extra type II NBs (J). The aDM, pDM and DL are shown at higher magnification in I. Anterior is on the left in all the embryos shown.
We did not detect any effect of misexpressing btd or Sp1 under the control of a neuroectodermal driver (sca-Gal4 UAS-btd and sca-Gal4 UAS-Sp1; Fig. 4I, S4A and Table S1), and when btd was misexpressed with a constitutive active version of ras (sca-Gal4 UAS-btd UAS-rasV12) there was no enhancement of the phenotype observed by misexpressing rasV12 alone (Fig. S4B-B″ and Table S1). However, we observed extra type II NBs when btd and Sp1 were simultaneously misexpressed (sca-Gal4 UAS-btd UAS-Sp1; Fig. 4J and Table S1). Finally, btd expression in the neuroectoderm did not appear to require activation of the EGFR pathway, because in rho mutant embryos the expression of btd-Gal4 is not affected (btd-Gal4/+; rhoM3 UAS-Cherry/rhoM3; Fig. S4C-C″). Hence, we conclude that either the three factors (EGFR, Btd and Sp1) have to be misexpressed to activate type II NB fate in other domains of the embryonic brain, or, in addition to them, other factors must also be required.
Type II NBs generate INPs during embryonic development
Most neurons generated by type II NBs give rise to the central complex, an area of the adult brain formed by a neuropil that serves as an integration centre implicated in several adult behaviours (Izergina et al., 2009; Wolff et al., 2015; Young and Armstrong, 2010). Type II NBs probably owe their existence to the need to generate a large number of neurons in a short period of time, yet larval life does not require such complex machinery as that needed in the eclosed animal. Therefore, we assessed whether type II NBs generate INPs during embryonic development.
To address this issue, we identified the INPs in the clusters of Cherry+ cells of wach embryos through their co-expression of Ase and Dpn, apparently identifying INPs in all clusters (Fig. 5A-C‴). The presence of INPs in these clusters was confirmed in the R9D11-Gal4 line, which was generated by cloning a cis-regulatory fragment from the earmuff (erm) gene into a Gal4 plasmid that is expressed by most, if not all, mature larval INPs but not by type II NBs or other brain cells (Bayraktar et al., 2010; Pfeiffer et al., 2008). Three clusters of Cherry+ cells were seen adjacent to the type II NB clusters in the brain of R9D11-Gal4 UAS-Cherry embryos, indicating that type II NBs do indeed generate INPs during embryonic development. Given the perdurance of the Cherry product in these clusters, Cherry+ cells must represent INPs, GMCs and neurons. To confirm the presence of INPs within the clusters, we found that Ase and Dpn were expressed in several cells (Fig. 5D-D‴). We further confirmed that these cells divided asymmetrically by assessing the presence of Miranda (Mira), which is indicative of asymmetric division (Fig. 5E-F″). Thus, we conclude that embryonic type II NBs divide to generate INPs.
Type II neuroblasts generate INPs during embryonic development. (A-D‴) Expression of Ase (green), Cherry (red) and Dpn (blue) in the aDM (A-A‴), pDM (B-B‴) and DL (C-C‴) clusters of stage 15 wach embryos, and in a DL cluster of a R9D11-Gal4 UAS-Cherry embryo at the same stage (D-D‴). The arrows indicate the mature INPs (Dpn+ Ase+) and the arrowheads indicate the type II NBs (Dpn+ Ase−). In A-C‴, the type II NBs lie in the Cherry clusters, whereas in D-D‴ they lie in the proximity of the cluster. (E-F″) Expression of Mira (green) in Cherry clusters from R9D11-Gal4 UAS-Cherry embryos; the arrows indicate the asymmetric Mira distribution in the cortex of INPs. Anterior is on the left in all the embryos shown.
Type II neuroblasts generate INPs during embryonic development. (A-D‴) Expression of Ase (green), Cherry (red) and Dpn (blue) in the aDM (A-A‴), pDM (B-B‴) and DL (C-C‴) clusters of stage 15 wach embryos, and in a DL cluster of a R9D11-Gal4 UAS-Cherry embryo at the same stage (D-D‴). The arrows indicate the mature INPs (Dpn+ Ase+) and the arrowheads indicate the type II NBs (Dpn+ Ase−). In A-C‴, the type II NBs lie in the Cherry clusters, whereas in D-D‴ they lie in the proximity of the cluster. (E-F″) Expression of Mira (green) in Cherry clusters from R9D11-Gal4 UAS-Cherry embryos; the arrows indicate the asymmetric Mira distribution in the cortex of INPs. Anterior is on the left in all the embryos shown.
Type II NBs express the same temporal series of factors during embryonic and larval development
Type I NBs generate different progeny through the sequential expression of a series of transcription factor genes: hunchback (hb), krüppel (kr), Pdm, castor (cas) and grainy head (grh) (Baumgardt et al., 2009; Brody and Odenwald, 2000; Grosskortenhaus et al., 2005; Isshiki et al., 2001; Kambadur et al., 1998). Expression of this series of genes stops when NBs enter quiescence at the end of embryogenesis, and it is reactivated when NBs resume proliferation during larval development, following the expression of cas and seven up (svp) (Kao et al., 2012; Maurange et al., 2008; Zhu et al., 2006). This makes the generation of a number of different neuronal types possible.
Larval type II NBs also express a temporal series of factors that include the genes Dichaete (D), cas and svp (Bayraktar and Doe, 2013). In addition, INPs sequentially express a different series of factors that diversify the neuronal types, including the D, grh and eyeless (ey) genes. Neuronal diversity is enhanced through the combined action of these series of genes (Bayraktar and Doe, 2013). Thus, we asked whether embryonic type II NBs express the same series of factors as type I NBs, and whether all INPs express the same series of factors or whether the INPs generated from NBs that express different temporal factors in turn express different factors.
When the expression of Hb, Kr, Pdm, Cas and Grh was assessed in wach embryos from stage 12, no Hb or Kr expression was observed, and Pdm was expressed only transiently; yet Cas and Grh were expressed by all type II NBs (Fig. 6A-D″). We also detected the expression of Nab, a transcriptional co-factor and a target of Cas that is required for the NBs to enter quiescence (Tsuji et al., 2008) and that defines a window within the Cas temporal window (Fig. 6C-C″) (Baumgardt et al., 2009). Hence, we conclude that type II NBs express only the later elements of the temporal series of factors, probably because they are generated late in embryogenesis.
Expression of temporal factors in embryonic type II NBs and INPs. (A-D″) Cherry (red), Pdm (green, A-A″), Cas (green, B-B″), Nab (green, C-C″) and Grh (green, D-D″) expression in the aDM (A,B,C,D), pDM (A′,B′,C′,D′) and DL (A″,B″,C″,D″) clusters of wach embryos; the arrowheads indicate type II NBs. (E-F″) Cherry (red), Dpn (blue), D (green in E-E″) and Ey (green in F-F″) expression in the aDM (E,F), pDM (E′,F′) and DL (E″,F″) of clusters from R9D11-Gal4 UAS-Cherry embryos. The green and blue channels are shown on the right of each figure and the stage of the embryos is indicated in each figure. Anterior is on the left in all the embryos shown.
Expression of temporal factors in embryonic type II NBs and INPs. (A-D″) Cherry (red), Pdm (green, A-A″), Cas (green, B-B″), Nab (green, C-C″) and Grh (green, D-D″) expression in the aDM (A,B,C,D), pDM (A′,B′,C′,D′) and DL (A″,B″,C″,D″) clusters of wach embryos; the arrowheads indicate type II NBs. (E-F″) Cherry (red), Dpn (blue), D (green in E-E″) and Ey (green in F-F″) expression in the aDM (E,F), pDM (E′,F′) and DL (E″,F″) of clusters from R9D11-Gal4 UAS-Cherry embryos. The green and blue channels are shown on the right of each figure and the stage of the embryos is indicated in each figure. Anterior is on the left in all the embryos shown.
When we evaluated D and Ey expression in INPs, they were clearly expressed in embryonic INPs following the same logic as in larval INPs. Young INPs (stage 14) expressed D and when they were older (stage 15), they expressed Ey (Fig. 6E-F″). Therefore, we found no differences between the temporal series expressed by embryonic and larval NBs and INPs.
Type II NBs in embryos are the same as those in larvae
Most NBs in the brain enter quiescence, including type II NBs, and, indeed, we observed the same number of type II NBs in embryos and larvae. However, to confirm this feature we used the FLEXAMP memory cassette that allows the complete lineage of embryonic type II NBs to be permanently labelled (Bertet et al., 2014; Harris et al., 2015). Recombination of the cassette was specifically activated in type II NBs using wach when keeping the embryos at 29°C from stage 11 to the end of embryogenesis (Fig. 7A,B). Early larvae were then shifted to 17°C to prevent recombination in the NBs that may be generated in larvae. To identify the NBs within the clones, the larvae were shifted back to 25°C in the mid-third instar to activate Cherry expression and in the late third instar, clones of only eight NBs labelled with GFP were observed (Fig. 7C,C′). Thus, we conclude that the eight type II NBs generated in the embryo are the same as those observed in the larva. As a control experiment, embryos and larvae of the same genotype were maintained at 17°C until mid-third instar larvae and then shifted to 25°C: no GFP expression was detected (Fig. 7D,E).
Embryonic and larval type II NBs are the same. (A) The logic of the FLEXAMP memory cassette experiment to immortalize the lineage of the embryonic type II NBs. (B) Diagram showing the stages of development in which temperature shifts were made. The stages in which wor-Gal4 (black line) and the FLEXAMP cassette (green line) are active are indicated. (C,C′) GFP (green) and Cherry (red) expression in the brain of a third instar wach larva. (C′) Red channel of the left hemibrain. GFP expression using the FLEXAMP system allows the whole lineage of type II NBs to be labelled from the time that cassette recombination by wor-Gal4 is activated. Cherry expression labels type II NBs and, given the persistence of the Cherry product, the progeny generated from mid-third instar larvae remain labelled. Type II NBs (asterisks) can be recognized by their position in the clusters and their size. (D,E) Control experiment in which embryos were collected at 25°C, shifted to 17°C and then shifted to 25°C to allow the activation of the 20xUAS-6xCherry. The activation of the cassette recombination was not detected. The midline is indicated. Anterior is upwards in all the embryos shown.
Embryonic and larval type II NBs are the same. (A) The logic of the FLEXAMP memory cassette experiment to immortalize the lineage of the embryonic type II NBs. (B) Diagram showing the stages of development in which temperature shifts were made. The stages in which wor-Gal4 (black line) and the FLEXAMP cassette (green line) are active are indicated. (C,C′) GFP (green) and Cherry (red) expression in the brain of a third instar wach larva. (C′) Red channel of the left hemibrain. GFP expression using the FLEXAMP system allows the whole lineage of type II NBs to be labelled from the time that cassette recombination by wor-Gal4 is activated. Cherry expression labels type II NBs and, given the persistence of the Cherry product, the progeny generated from mid-third instar larvae remain labelled. Type II NBs (asterisks) can be recognized by their position in the clusters and their size. (D,E) Control experiment in which embryos were collected at 25°C, shifted to 17°C and then shifted to 25°C to allow the activation of the 20xUAS-6xCherry. The activation of the cassette recombination was not detected. The midline is indicated. Anterior is upwards in all the embryos shown.
INPs enter quiescence at the end of embryonic development and resume proliferation in the larva
At the end of embryonic neurogenesis, NBs can either enter quiescence, ready to resume proliferation in first instar larvae, or they can die by programmed cell death (PCD). As INPs behave similarly to type I NBs, we investigated whether embryonic INPs (eINPs) also enter quiescence at the end of embryogenesis, as do embryonic type I NBs, or whether they die by PCD. We labelled wach embryos at different stages with an antibody against Dcp-1, which recognizes the cleaved form of the caspase 1 effector. As no Dcp1 expression was evident in the clusters of Cherry+ cells at any stage of embryonic development, we conclude that embryonic INPs do not die by PCD (Fig. 8A).
Embryonic INPs enter quiescence and resume proliferation in early second instar larvae. (A) Dcp1 (green) and Dpn (blue) expression in a 20 h wach brain, showing separate Dcp1 and Cherry expression. No cell death was detected in the aDm, pDM (arrowheads) and DL (arrow) clusters. (B,C) EdU (green) and Cherry (red) labelling in the brain of a late first instar (B) and an early second instar (C) wach larva. EdU is not detected in the type II NB clusters during late first instar, yet several cells in the Cherry+ clusters express EdU in early second instar larvae, indicating that NBs and INPs start to cycle. In C, a closer view of each cluster is shown on the right and the NBs in the clusters are identified by their larger size (asterisks). (D) The number of EdU-expressing cells in each cluster, identifying NBs within each cluster by their larger size. (E-F″) EdU (green) and Cherry (red) labelling in the brain of early second instar R9D11-Gal4 UAS-Cherry tub-Gal80ts larvae. Arrowheads indicate EdU-positive cells within the Cherry-expressing clusters. (G-H″) EdU labelling (green) in eINP lineages permanently labelled with the FLEXAMP cassette (red). The clones are outlined and the EdU-labelled cells are indicated (asterisks). (I) Early first instar larvae were shifted from 29°C to 17°C to repress the expression of R9D11-Gal4 and prevent the activation of Cherry in larval INPs. EdU is expressed in several cells in each cluster. The lines indicate the brain midline in A-C. Anterior is towards the top in all the embryos shown.
Embryonic INPs enter quiescence and resume proliferation in early second instar larvae. (A) Dcp1 (green) and Dpn (blue) expression in a 20 h wach brain, showing separate Dcp1 and Cherry expression. No cell death was detected in the aDm, pDM (arrowheads) and DL (arrow) clusters. (B,C) EdU (green) and Cherry (red) labelling in the brain of a late first instar (B) and an early second instar (C) wach larva. EdU is not detected in the type II NB clusters during late first instar, yet several cells in the Cherry+ clusters express EdU in early second instar larvae, indicating that NBs and INPs start to cycle. In C, a closer view of each cluster is shown on the right and the NBs in the clusters are identified by their larger size (asterisks). (D) The number of EdU-expressing cells in each cluster, identifying NBs within each cluster by their larger size. (E-F″) EdU (green) and Cherry (red) labelling in the brain of early second instar R9D11-Gal4 UAS-Cherry tub-Gal80ts larvae. Arrowheads indicate EdU-positive cells within the Cherry-expressing clusters. (G-H″) EdU labelling (green) in eINP lineages permanently labelled with the FLEXAMP cassette (red). The clones are outlined and the EdU-labelled cells are indicated (asterisks). (I) Early first instar larvae were shifted from 29°C to 17°C to repress the expression of R9D11-Gal4 and prevent the activation of Cherry in larval INPs. EdU is expressed in several cells in each cluster. The lines indicate the brain midline in A-C. Anterior is towards the top in all the embryos shown.
We then assessed whether eINPs enter quiescence at the end of neurogenesis to then resume proliferation in larva. As such, we labelled the S-phase of the cell cycle using EdU (5-ethynyl-2 deoxyuridine) in late first/early second instar wach larvae. There was no EdU labelling within the Cherry clusters in the late first instar larvae, even though other NBs in the brain at that stage had already exited quiescence and started cycling (Fig. 8B). However, the cluster of Cherry+ cells in early second instar larvae contained several cells with EdU and type II NBs, which can be identified by their size (Fig. 8C,D). Since the ganglia were incubated in EdU for only a short period (30 min), these EdU-expressing cells were probably not generated in the larvae and therefore, they may correspond to eINPs that have remained quiescent before resuming proliferation. To confirm this possibility, we generated embryos with a tub-Gal80ts; UAS-Cherry; R9D11-Gal4 genotype, in which Gal80ts is active at 17°C and inactive at 25°C. We collected these embryos at 25°C and shifted them to 29°C at stage 11 to increase the Gal4 activity; then we shifted them to 17°C at the early first instar larval stage to block the Gal4 activity (Fig. 8I). Thus, Cherry expression was activated exclusively in the embryonic INPs and it was blocked in the larval INPs. In these larvae, EdU was incorporated into cells in the Cherry cluster, which indicates that eINPs divide. As a control, we checked that embryos of this genotype maintained at 17°C do not express Cherry and that they start to do so when shifted to 29°C as second instar larvae. To further confirm this result, we permanently labelled eINPs and their progeny with GFP using the FLEXAMP memory cassette (see Materials and Methods), and EdU incorporation was observed in GFP-expressing clones (Fig. 8G-H″). These results confirm that eINPs enter quiescence at the end of embryonic neurogenesis and that they resume proliferation in early second instar larvae.
DISCUSSION
The discovery of type II NBs in the brain of larval Drosophila has raised many questions, including how their specification is controlled genetically, as well as what mechanisms are involved in the generation of INPs, a distinctive feature of these NBs. A major limitation to analysing NBs in the fly has been the inability to identify larval counterparts to the embryonic NBs. To resolve this problem, a detailed map of the molecular markers of NBs in the early embryo brain has been established (Urbach and Technau, 2003), and these markers have since been traced to later embryonic and larval stages (Lacin and Truman, 2016; Sprecher et al., 2007; Younossi-Hartenstein et al., 2006). In the current study, we investigated the embryonic origin of larval type II NBs. The combined use of wor-Gal4, ase-Gal80 and 20xUAS-6xCherry, although with some limitations related to the delayed activation of Gal4, enabled us to clearly identify the embryonic origin, proliferative behaviour and axonal projections of type II NBs. We found that type II NBs appear during early embryonic stage 12 in three clusters. These groups of NBs and their progeny remain clustered together during embryonic and early larval development, projecting axon fascicles that later enable the type II NBs to cross the brain lobes via the interhemispheric commissure.
One interesting issue that we did not address is whether the NBs in each cluster generate the same or different progeny. It has been suggested that there are several neuroectodermal placodes in the dorsomedial protocerebrum of stage 12 embryos (de Velasco et al., 2007; Hwang and Rulifson, 2011). According to the molecular markers expressed, these placodes seem to have different fates, and it is tentative to suggest that these placodes are the source of the different clusters of type II NBs. However, an in-depth lineage analysis will be necessary to address such issues, as previously carried out for the mushroom body NBs (Kunz et al., 2012; Lee et al., 1999).
We did investigate whether embryonic type II NBs generate INPs when they divide or whether they behave more like type I NBs, and whether they change their behaviour on exiting from quiescence during early larval stages. The expression of the INP-specific driver R9D11-Gal4 showed that type II NBs generate INPs when they divide and the expression of Mira confirms that they divide asymmetrically. Most neurons generated from type II NBs will form the central complex, an important integrative centre for adult behaviour. However, the role of the embryonic progeny of type II NBs in the larva remains to be determined. Strikingly instead of dying by PCD after several divisions, as larval INPs do, embryonic INPs become quiescent and they resume proliferation in the early second instar larva. Thus, it will be interesting to study what distinct roles these INPs and their progeny play in larval development.
We further analysed how these type II NBs are specified. From a set of genes expressed in the dorsomedial region of the brain (Urbach and Technau, 2003), relevant phenotypes were only found for mutants of the EGFR pathway, and of the related btd and Sp1 genes (data not shown). The data indicate that the combined action of the EGFR pathway, which is activated by the Spi ligand, and that of the related btd and Sp1 genes is necessary to specify type II NB fate. It is interesting to note that, despite the strong sequence homology of the Btd and Sp1 genes, and their similar expression patterns, both are required to specify type II NBs. This is similar to the role of these genes in leg development where they are also both required (Córdoba et al., 2016). Nevertheless, it remains to be determined how the three distinct groups of NBs (aDM, pDM and DL) are specified and what distinct features they have.
Finally, we analysed the temporal series of factors expressed by embryonic type II NBs and INPs. With respect to the series of genes expressed by most type I NBs (Hb-Kr-Pdm-Cas-Nab-Grh), type II NBs appear not to express Hb and Kr, although Kr is expressed by neurons in the Cherry clusters (data not shown). Furthermore, Pdm is expressed transiently in all three clusters, followed by the expression of Cas, Nab and Grh. This result is not surprising, given that NBs delaminating during the later stages of embryogenesis (e.g. NB3-3 and NB5-5) fail to express the earliest factors in the temporal series (Benito-Sipos et al., 2010; Tsuji et al., 2008). Thus, the factors expressed in a temporal series by type II NBs do not seem to differ from those expressed by type I NBs. However, we have detected expression of D and Ey in eINPs, which confirms that eINPs express the same factors as larval INPs. As we have not been able to distinguish between embryonic INPs generated during different temporal windows, we cannot say whether all INPs express the same temporal series of genes or whether this is determined by the factors expressed by the NB at the time the INP is generated. In an article published while this manuscript was being reviewed (Walsh and Doe, 2017), the authors report that, unlike our observations, there was no expression of Ey in the embryonic INPs. This apparent contradiction could be explained if INPs generated by the same NB in different temporal windows express either a different temporal cascade or bypass some of the components of the cascade.
MATERIALS AND METHODS
Drosophila strains
We used the following alleles to analyse the wild-type and mutant phenotypes: y w, Df(1)btd Sp1 (a deficiency that removed the btd and Sp1 genes (Estella and Mann, 2010); btdXG (Cohen and Jürgens, 1990); Sp1HR (Córdoba et al., 2016); rhoM3 (Díaz-Benjumea and García-Bellido, 1990); pntD86 (Fuchs et al., 2012); yan884 (Schober et al., 2005); spi1, Df(3L)ED225 and PBac[Sp1:eGFP.S] (FlyBase).
The Gal4/Gal80 lines used were: sca-Gal4, wor-Gal4, ase-Gal80 (Neumüller et al., 2011); R9D11-Gal4 (Pfeiffer et al., 2008); btd-Gal4 (Estella and Mann, 2010); and tub-Gal80ts.
The Upstream Activation Sequence (UAS) lines used were: 20xUAS-6xcherry:HA (Bloomington Drosophila Stock Center, #52267 and #52268); UAS-btd and UAS-Sp1 (Estella and Mann, 2010); and UAS-EGFRDN, UAS-EGFRCA, UAS-rasV12, UAS-yanCA (FlyBase).
The following fly stocks were used:
wach (wor-Gal4 ase-Gal80 20xUAS-6xCherry:HA/CyO or wor-Gal4 ase-Gal80; 20xUAS-6xCherry:HA).
wach btd Sp1 (Df(1)btd Sp1/FM7, Act>GFP; wor-Gal4 ase-Gal80 20xUAS-6xCherry:HA/CyO).
wach yan (yan884 wor-Gal4 ase-Gal80/CyO, Act>GFP; 20xUAS-6xCherry:HA).
wach rho (wor-Gal4 ase-Gal80; rhoM3 20xUAS-6xCherry:HA/TM3, Act>GFP).
wach pntP1 (wor-Gal4 ase-Gal80; 20xUAS-6xCherry:HA pntP1D86/TM3, Act>GFP).
wach Sp1:GFP (Sp1:GFP; wor-Gal4 ase-Gal80; 20xUAS-6xCherry:HA).
btd-Gal4 UAS-Cherry rho (btd-Gal4/FM7, Act>GFP; rhoM3 20xUAS-6xCherry:HA/TM3, Act>GFP).
w; 13xLexAop2-IVS-myr:GFP Actcp4.6>dsFRT<nlsLexAp5; tub-Gal80ts and
5xUAS-FlpPest; R9D11-Gal4 (Bertet et al., 2014; Harris et al., 2015).
Immunostaining and confocal microscopy
Embryos and larvae were collected, fixed and stained as described previously (Benito-Sipos et al., 2010), and mounted in Vectashield mounting medium. The primary antibodies used in this study were: rabbit anti-Ase (1:400), rat anti-Ase (1:400), rabbit anti-Dpn (1:500), rat anti-Dpn (1:500) and rat anti-Miranda (1:1000; for further details see the section ‘Antibody production’); rabbit anti-Nab (1:500) (Terriente et al., 2007); rabbit anti-Pdm1 (1:500) (Terriente et al., 2008); rabbit anti-Pnt1 (1:500, provided by J. B. Skeath, Washington University, St Louis, MO, USA); rabbit anti-Cas (1:500) (Kambadur et al., 1998); rat anti-Grh (1:100, provided by S. Thor, Linköping University, Sweden); guinea pig anti-Hb (1:200, provided by I. Miguel-Aliaga, Imperial College, London, UK); rabbit anti-Kr (1:500, provided by P. Carrera, Max Planck Institute, Göttingen, Germany); rabbit anti-Dichaete (1:1000, provided by S. Russell, Cambridge University, UK); mouse anti-Eyeless (1:50, Developmental Studies Hybridoma Bank, DSHB); anti-Sex lethal (1:50, DSHB, #M18); mouse anti-GFP (1:50, DSHB, #12A6); chicken anti-GFP (1:200, Upstate, #AB16901); and rabbit anti-cleaved Dcp1 (1:200, Cell Signaling #9578).
Labelling with anti-Sex lethal allowed the sex of the embryos to be determined. Images were captured on a Zeiss LSM510 or Zeiss LSM710 confocal microscope with a z-resolution of 1.5 µm, and they were processed with Fiji and Adobe Photoshop software. Three-dimensional brain reconstructions and movies were generated using a Nikon A1R confocal microscope with a z-resolution of 1.0 µm.
EdU labelling
Ganglia from individuals at the relevant stages were dissected out, attached to poly-L-lysine-coated slides, incubated in 0.2 mM 5-ethynyl-2 deoxyuridine (EdU) for 30 min and fixed with 4% paraformaldehyde (PFA) for 20 min. The Click-iT reaction was carried out according to the manufacturer's instructions (Fisher Scientific, #C10337 and #C10339), and the slides were then immunostained as described previously (Benito-Sipos et al., 2010) and mounted in Vectashield mounting medium.
FLEX-AMP memory cassette
This system allows cells and their progeny to be permanently labelled with GFP through selective activation of FRT/Flp-mediated recombination in a few cells using a specific Gal4 line. Once recombination of the cassette is activated, controlled by the use of a thermosensitive Gal80 line, cells permanently express GFP (Actcp4.6>nlsLexAp5>13xLexAop2-IVS-myr:GFP).
To label the complete lineage of embryonic type II NBs, female 5xUAS-FlpPest; wor-Gal4 ase-Gal80; 20xUAS-6xCherry:HA flies were crossed with w; 13xLexAop2-IVS-myr:GFP Act5cp4.6>dsFRT< nlsLexAp5; tub-Gal80ts males. Embryos were collected at 25°C for 3 h, shifted to 29°C for 15 h to selectively activate LEX-cassette recombination in type II NBs, and then shifted to 17°C to prevent cassette activation in new larval NBs. Mid-third instar larvae were finally shifted to 25°C to activate Cherry expression. In a control experiment, embryos were collected at 25°C for 3 h, shifted to 17°C, then shifted to 25°C in mid-third instar and dissected in late third instar.
To label embryonic INPs, 5xUAS-FlpPest; R9D11-Gal4 females were crossed with w; 13xLexAop2-IVS-myr:GFP Actcp4.6>dsFRT<nlsLexAp5; tub-Gal80ts males. Embryos were collected at 25°C for 3 h then shifted to 29°C for 15 h to drive R9D11-Gal4 expression and LEX cassette recombination, and then shifted to 17°C until they became early second instar larvae (to prevent cassette activation in larval INPs). The larvae were then incubated with EdU for 30 min and fixed.
Antibody production
To generate anti-Ase antibodies, two rats and two rabbits were injected with the synthetic peptide CLSDESMIDAIDWWEAHAPKSNGACTNLSV, corresponding to a fragment in the C-terminal domain of the putative Ase protein. The terminal Cys residue was added to couple the peptide to the keyhole limpet hemocyanin (Klh) carrier protein. After five immunizations, animals were bled and the resulting sera were tested on the embryonic CNS.
To generate antibodies against Dpn, an internal BamHI fragment (base pairs 594-1363) encoding amino acids 109-365 of the predicted Dpn protein (Bier et al., 1992) was cloned into the pRSET-B vector and expressed in bacteria (E. coli BL21-DE3). After lysis of the bacteria, the insoluble fusion protein was incubated for 1 h in 8 M urea in phosphate-buffered saline and centrifuged at 10,000 rpm (6708 g) for 5 min at 4°C. The protein was renatured by stepwise dialysis against decreasing concentrations of urea, and injected into two rats and two rabbits. After five immunizations, the animals were bled and the resulting sera were tested on the embryonic CNS.
To generate an antibody against Miranda, two rats were injected with the synthetic peptide KAKLKRFNDVDVAIC that corresponds to the N-terminal amino acids 5-19 of the putative Mira protein. After three immunizations, animals were bled and the resulting sera were tested on the embryonic CNS.
Acknowledgements
We are most grateful to Antonio Baonza, Sonsoles Campuzano, Pilar Carrera, Carlos Estella, Jurgen Knoblich, Ward Odenwald, Steve Russell, Jim Skeath, James Truman, Gerd Vorbrüggen, the Bloomington Drosophila Stock Center and the Developmental Studies Hybridoma Bank for fly stocks and antibodies. We also thank Chris Doe for sharing unpublished results, Cristina López Rosso, Leila Maestro Paramio, Lucía Mendoza Lupiañez and Raul González for their help with several of the experiments, and Alicia Estacio for her comments on the manuscript. The monoclonal antibodies against Ey, GFP and Sxl were obtained from the Developmental Studies Hybridoma Bank, created by the NICHD (part of the NIH) and maintained in the University of Iowa (Department of Biology, Iowa City, IA 52242, USA).
Footnotes
Author contributions
Conceptualization: F.J.D.-B.; Methodology: F.J.D.-B.; Validation: F.J.D.-B.; Formal analysis: F.J.D.-B.; Investigation: J.-A.A., F.J.D.-B.; Writing - original draft: F.J.D.-B.; Supervision: F.J.D.-B.; Funding acquisition: F.J.D.-B.
Funding
This work was supported by a grant from the Ministerio de Economía y Competitividad (BFU2014-53761-P) to F.J.D.-B. and by institutional grants from the Fundación Ramón Areces and Banco de Santander to the CBMSO.
References
Competing interests
The authors declare no competing or financial interests.