bHLH family proteins control the timing and completion of transition from neuroepithelial cells into neural stem cells

During brain development, neural stem cells produce a large number of a wide variety of neurons, and these neurons establish complex neuronal circuits to make the brain function correctly. For sufficient amounts of neurons to construct a functional brain, sufficient amounts of neural stem cells must be born because the number of neural stem cells reflects the total number of neurons in the brain. In the mammalian central nervous system, neural stem cells, which are called radial glial cells, arise from neuroepithelial cells of a neural tube early in development (Miranda-Negrón and García-Arrarás, 2022). Before these neural stem cells emerge, the neuroepithelial cell population must be expanded to secure the necessary amount of neural stem cells for successful brain development. Although the timing of the transition from the proliferative to neurogenic phase must be strictly determined (Dwyer et al., 2016; Fernández et al., 2016), the whole picture of the molecular mechanisms that regulate the phase transition in neuroepithelial cells is still unclear.

The fly visual system is known as one of the best models for neural developmental research because it shares structural and developmental features with the mammalian cerebral cortex (Ramón y Cajal and Sánchez, 1915; Sanes and Zipursky, 2010; Hasegawa et al., 2011; Bertet et al., 2014; Chen et al., 2016; Suzuki et al., 2016a). The fly visual center consists of four different ganglia, including the medulla, which is the largest component. Because the medulla contains more than 100 types of 40,000 neurons (Fischbach and Dittrich, 1989; Hofbauer and Campos-Ortega, 1990; Hasegawa et al., 2011), its neural circuit is sufficiently complex to serve as a model for the mammalian cerebral cortex, and its developmental process can be examined at the single neuron level. The developing medulla shows subdivision into specific zones (Hasegawa et al., 2011; Gold and Brand, 2014; Suzuki and Sato, 2014; Erclik et al., 2017; Valentino and Erclik, 2022), birth-order dependent cell fate determination (Li et al., 2013; Suzuki et al., 2013; Zhu et al., 2022; Konstantinides et al., 2022; Tang et al., 2022) and integration of neurons derived from two different progenitor pools (Dearborn and Kunes, 2004; Bertet et al., 2014; Chen et al., 2016; Suzuki et al., 2016a,b). All these processes also occur in the mammalian cerebral cortex during development.

The medulla primordium is composed of a single sheet of neuroepithelial cells that divide symmetrically to expand the population during the first half of larval development (Ceron et al., 2001; Hofbauer and Campos-Ortega, 1990). At the early stage of the second half, neuroepithelial cells that occupy the medial edge of the developing medulla initiate differentiation into medulla neural stem cells, usually termed neuroblasts in Drosophila (Egger et al., 2007). The differentiation from neuroepithelial to neural stem cells occurs in waves from the medial edge of the developing medulla to the lamina precursor, and it progresses in a single cell row width, which is called the proneural wave (Fig. 1; Yasugi et al., 2008; Sato et al., 2013). As the current location of the proneural wave can be visualized by the expression of Lethal of Scute (L'sc), which is one of the members of proneural transcription factors (Skeath and Carroll, 1992; Cabrera et al., 1994; Yasugi et al., 2008), several groups have identified regulatory signaling pathways involved in the wave progress by examining effects of genetic manipulations they have carried out on the expression pattern of L'sc: JAK-STAT, EGFR, Notch, Fat-Hippo and steroid hormone signaling pathways (Yasugi et al., 2008, 2010; Reddy et al., 2010; Kawamori et al., 2011; Wang et al., 2011a,b; Lanet et al., 2013; Pérez-Gómez et al., 2013; Morante et al., 2013; Contreras et al., 2018; Zhou et al., 2019; Shard et al., 2020). Although extensive studies have been carried out in recent years, many questions remain: How do those signaling pathways interact? What kinds of genes are regulated downstream of each signaling pathway? What is the key factor(s) that switches on those signaling at the optimal timing?

In the present study, we show that TCF4/Daughterless (Da), which is a member of the basic helix-loop-helix (bHLH) transcription factor, is essential for the completion of differentiation from neuroepithelial to neural stem cells. We also found that Extramacrochaetae (Emc), which is a member of the bHLH transcription factor family but lacks a DNA-binding domain, is another cell-intrinsic regulator of proneural wave progression. Expression of L'sc was found more laterally in the emc mutant homozygous clones, suggesting that Emc negatively regulates proneural wave progress. The expression pattern of Emc was changed by the genetic manipulation of EGFR or Notch signaling, suggesting that its expression is under the control of these signaling pathways. As Emc is known as an antagonist of other bHLH transcription factors, including Da and the Achaete-Scute complex (Benezra et al., 1990; Ellis et al., 1990; Garrell and Modolell, 1990; Cabrera et al., 1994), we examined whether Da is also involved in proneural wave progression. As expected, expression of L'sc was found more medially in the da mutant homozygous clones than in the surrounding cells, suggesting that Da accelerates proneural wave progression. This is similar to the roles of Emc and Da in the progress of morphogenetic furrow in the eye imaginal disc (Brown et al., 1995; Cadigan et al., 2002; Bhattacharya and Baker, 2011). We also searched for direct regulators of proneural wave progression under the control of Da with the nanoDamID technique (Tang et al., 2022) and identified Notch as one of the downstream effectors of Da. Altogether, our results demonstrate that bHLH proteins themselves play central roles in the transition from the proliferative to the neurogenic phase in the medulla primordium. In particular, Da is essential in determining whether to move to the next state or stay in the current state.

The differentiation from neuroepithelial cells into neural stem cells proceeds in a mediolateral direction in the medulla primordium during the latter half of larval development (Fig. 1A-B), suggesting that genes expressed in neuroepithelial cells during this period are highly likely to be involved in the regulation of differentiation. To identify genes involved in the regulation of the differentiation from neuroepithelial cells to neural stem cells, we examined the expression of transcription factors and found that both Emc and Da were expressed in the neuroepithelial cells that were visualized with Cubitus interruptus (Ci) expression and in the cells of transition zone visualized using L'sc expression (Fig. 1C-F and Fig. S1). Emc expression tends to be weakened in the transition zone, but Da is not (Fig. 1G). Both Emc and Da were abundant in the majority of neural stem cells in the developing medulla (Fig. 1H-H′). Notably, Da was expressed strongly but Emc was expressed only slightly in the neural stem cells located most laterally. Because these neural stem cells are the newest, we called them newborn neural stem cells (nbNSCs) hereafter. Da and Emc were also expressed in almost all neurons in the developing medulla (Fig. S1B,C).

Since EGFR and Notch signaling pathways are recognized as central regulators of proneural wave progress (Sato et al., 2013; Yasugi et al., 2010), we next examined whether Emc and Da are downstream effectors of these pathways. We compared the expression of Emc or Da to that of components of each signaling pathway. Weak Emc expression overlapped with that of Delta, which is one of the ligands in the Notch signaling pathway, in the transition zone visualized with L'sc (Fig. 2A,A′). Emc also overlapped with Notch (N) or Pointed P1 (PntP1), a downstream effector of the EGFR signaling pathway (O'Neill et al., 1994; Gabay et al., 1996), in the neuroepithelial cells (Fig. 2B-C′). Notch signaling was highly activated in Da- or Emc-positive neuroepithelial cells, as indicated by E(spl)mγ-GFP (Fig. 2D-D′ and Fig. S1F), which is a Notch signaling sensor (Almeida and Bray, 2005). Da expression overlapped with PntP1 or N expression in neuroepithelial cells and with Delta in the transition zone (Fig. 2E-G′). None of these molecules was abundantly expressed in the nbNSCs (Fig. 2H). In summary, both expression of Emc and Da coincide with that of components of EGFR and Notch signaling pathways in the neuroepithelial cells and transitional zone in the developing medulla.

We found that Emc shows a triphasic expression pattern across the developing medulla: Emc expression is moderate, weak and abundant in L'sc-negative neuroepithelial cells, L'sc-positive neuroepithelial cells and neural stem cells, respectively (Fig. 2H). This expression pattern implies that Emc has negative impacts on the proneural wave progress. Therefore, we examined the relationship between Emc and the EGFR or Notch signaling pathway, as well as the roles of Emc in the regulation of proneural wave progression.

We examined the expression pattern of Emc when the EGFR or Notch signaling pathway was disrupted. Emc and L'sc expression were lost by pnt knockdown in a subset of the neuroepithelial cells and in cells in the transition zone (Fig. 3A, n=22/23; Fig. S3). Emc still appeared in the neuroepithelial cells in Delta mutant clones where L'sc expression was found more laterally (Fig. 3B, n=11/14). However, prolonged Emc expression was found when the constitutively active form of N, which lost an extracellular domain of N (Go et al., 1998), was induced ectopically (Fig. 3C, n=21/21). These results suggest that Emc expression is under the control of both EGFR and Notch signaling pathways.

As Emc expression was affected by the genetic manipulation of the EGFR or Notch signaling pathways, it is possible that this molecule is involved in the regulation of neuroepithelial to neural stem cell transition. We therefore examined the roles of Emc in the proneural wave progress. In emc mutant clones, L'sc expression appeared more laterally (Fig. 3D, n=28/31), as observed in the Delta mutant (Fig. 3B, n=11/14). As expected, neural stem cells were also precociously found in more lateral regions in the emc mutant clones (Fig. 3E, n=12/17), which was never observed in the control clones. The results suggest that Emc plays suppressive roles in proneural wave progress.

We also examined whether Emc is involved in the regulation of the proneural wave via feedback mechanisms to EGFR or Notch signaling pathways. Delta was also found more laterally, coinciding with L'sc expression in emc mutant clones (Fig. 3F, n=11/13). Expression of E(spl)m8-GFP, another reporter of Notch signaling activity, was downregulated in emc mutant clones (Fig. 3G, n=19/24). For PntP1, its expression was precociously found in emc mutant clones (Fig. 3H, n=17/18). Although Hairy is also known as a negative regulator of proneural genes (Skeath and Carroll, 1991; Orenic et al., 1993), L'sc was normally observed in hairy mutant clones (Fig. S2, n=13/13).

Emc exhibits suppressive effects on the transcriptional activity of proneural transcription factors, especially Da (Cabrera et al., 1994). As shown above, Da is highly abundant in the neuroepithelial cells of the medulla primordium (Figs 1 and 2). Does Emc also antagonize Da in the neuroepithelial cells of the developing medulla? We examined whether Da shows opposite effects on proneural wave progress compared with Emc. In da mutant clones, L'sc expression was not terminated but was continuously observed more medially (Fig. 4A, n=49/49). Consistently, neural stem cells completely disappeared in da mutant clones in the lateral region of the surface of the medulla primordium (Fig. 4B, n=42/43). However, neural stem cells still appeared and showed tumor-like overgrowth in da mutant clones created in the medial region (indicated by a yellow arrow in Fig. 4B). This could be due to disruption of asymmetric division in da mutant neural stem cells, as reported previously (Yasugi et al., 2014). To test this possibility, we examined whether neural stem cells still appear in the cortex of da mutant. We found that neural stem cells ectopically appeared and occupied da mutant clones created in the medulla cortex (Fig. 4C, n=41/41). Thus, expression of L'sc is prolonged and consequently the transition from neuroepithelial cells to neural stem cells is disrupted in the absence of Da.

The loss-of-function analyses of Da or Emc indicate that both are involved in the regulation of the transition from neuroepithelial to neural stem cells by controlling L'sc expression. As Emc lacks DNA-binding domains, it is unlikely to directly regulate gene expression. We therefore focused on Da and examined whether it is one of the direct regulators of L'sc expression by elucidating target genes of Da using a newly developed nanoDamID technique (Tang et al., 2022). Successful nanoDamID analysis requires the use of a highly specific Gal4 driver to induce nanoDam expression specifically in the cell type of interest. We reconfirmed the expression pattern of ogre-gal4 (Fig. 4D-D′), which was reported as a specific Gal4 driver in the developing medulla (Hakes et al., 2018; Tang et al., 2022). ogre-Gal4 was strongly expressed in neuroepithelial cells of the developing medulla, as visualized with GFP, but not in the central brain. We chose ogre-Gal4 to induce nanoDam to identify genome-wide binding targets of Da in the neuroepithelial cells of the developing medulla. Over-representation analysis on the Da targets showed enrichment of genes involved in the developmental process, especially neural development (Fig. 4E), and those involved in Hippo, Notch and MAPK signaling pathways (Fig. 4F). We found that Da directly binds to the l'sc locus in neuroepithelial cells (Fig. 4G and Fig. S4A). Da binding at the l'sc locus includes the enhancers that are active in the central nervous system [Kvon et al., 2014; REDfly (http://redfly.ccr.buffalo.edu/) provided by Center for Computational Research in the University at Buffalo (Gallo et al., 2006, 2011; Halfon et al., 2008; Rivera et al., 2019)]. Because Da binds to E-box sequences (5′-CACCTG-3′ and 5′-CAGCTG-3′) that are widely conserved (Ohsako et al., 1994), we investigated whether these enhancers contain the motif. Consistently, an E-box motif was found in one of the enhancers (Fig. 4G).

We found that Da also bound to the N locus at a much higher frequency than to the l'sc locus (Fig. 5A, Fig. S4B, and Table S1). The Da-binding sites at the N locus included the enhancer sequences that were shown to be active in the developing medulla [FlyLight (https://flweb.janelia.org/cgi-bin/flew.cgi) provided by Janelia Research Campus (Pfeiffer et al., 2008, 2010; Jenett et al., 2012)]. These enhancers also contain the E-box motifs (Fig. 5A).

The Da-binding profile indicates that N expression is also directly regulated by Da. We investigated whether N expression is truly under the control of Da by mosaic analysis. When da mutant clones were induced in the lateral region, N appeared ectopically with prolonged expression of L'sc (Fig. 5B, n=14/15). This ectopic induction of N was also observed in the da mutant clones of the medial region (Fig. 5B, n=15/15). In contrast to N, Delta was normally expressed even in da mutant clones (Fig. S5A, n=23/23). We found a small subset of cells ectopically expressing both L'sc and PntP1 in da mutant clones (Fig. 5C, n=16/26). On the other hand, E(spl)mγ-HLH-GFP was persistent more medially in da mutant clones (Fig. 5D, n=16/16). These results suggest that L'sc expression was not terminated because of prolonged N expression in the da mutant. We then tested whether this is true by ectopic expression of a full-length (N^full) or a constitutively active form of N (N^act). In N^full expressing clones, L'sc expression was still observed more medially (Fig. 5E, n=60/63), as observed in da mutant clones (Figs 4A and 5B). More severely, a large part of the clone expressing N^act was occupied by L'sc-positive cells at the expense of neural stem cells (Fig. 5F, n=29/29).

We conducted ectopic expression experiments to determine whether the ectopic expression of da induces precocious termination of Notch signaling activation. When we induced da expression ectopically, E(spl)m8-LacZ was expressed more laterally but diminished immediately next to L'sc-positive cells (Fig. 5G-G′, n=12/12). As expected, cells expressing both L'sc and N (Fig. 5H-I, n=10/12), and neural stem cells (Fig. 5J, n=20/21) appeared more laterally after ectopic expression of da. Delta was also expressed precociously in L'sc-positive cells that appeared more laterally in the region where da was ectopically induced (Fig. S5B, n=11/11). In summary, the results suggest that Notch signaling is inactivated via suppression of N expression itself by Da during the transition from neuroepithelial to neural stem cells.

Da heterodimerizes with acheate-scute complex (AS-C) proteins to activate gene expression (Cabrera and Alonso, 1991; Vaessin et al., 1994). Among the members of AS-C, Scute (Sc), Ase and L'sc are good candidates that act as partners of Da because they are expressed in the developing medulla (Egger et al., 2007). We used deficiency lines in which different regions of the AS-C locus were deleted as described previously (Yasugi et al., 2008). Briefly, Df(1)260-1 lost all four bHLH-proneural genes, acheate (ac), sc, l'sc and ase, while Df(1)sc10-1 lost only ac and sc. Df(1)ase1 could be used as an ase null mutant because it only lost ase. When we induced Df(1)260-1 clones, N expression was prolonged in the clones at the expense of neural stem cell emergence (Fig. 6A, n=22/24). This effect was not phenocopied in either Df(1)sc10-1 (n=18/18, Fig. S5C) or Df(1)ase1 clones (n=15/15, Fig. S5D), indicating that both Sc and Ase are dispensable for the suppression of N. These results suggest that L'sc is also involved in the regulation of N expression for successful differentiation from neuroepithelial cells to neural stem cells.

The loss-of-function experiments on as-c suggest that L'sc is essential for the termination of N expression. We next examined the expression pattern of N by L'sc misexpression. N expression was precociously terminated on the surface of the developing medulla and, instead, neural stem cells emerged earlier (Fig. 6B, n=15/15). Notably, many neural stem cells were ectopically observed in the lamina (Fig. 6C, n=15/15).

The precise timing of the differentiation from neuroepithelial cells to neural stem cells must be determined to ensure the necessary amount of neural stem cells for building the functional nervous system during development. To date, the molecular mechanisms that regulate the timing of differentiation have not been the main focus in developmental neurobiology. However, recent studies have shown that dysregulation of the timing of differentiation from neuroepithelial cells to neural stem cells causes fetal hydrocephalus (Duy et al., 2022, 2023), which has classically been attributed to disrupted cerebrospinal fluid homeostasis (Aschoff et al., 1999). Understanding the molecular mechanisms that determine the timing of differentiation will be necessary to discover therapeutic targets for this disorder. Here, we identified Emc and Da as new components of the regulatory machinery involved in the timely differentiation from neuroepithelial to neural stem cells. In particular, we showed that Da suppresses N expression, which ensures that neuroepithelial cells in the transitional state differentiate into neural stem cells (Fig. 7A).

We found that Emc and Da are expressed in almost all of the neuroepithelial cells, neural stem cells and neurons (Fig. 1). In emc mutant clones, the proneural wave progress was accelerated (Fig. 3D). PntP1 expression was found precociously but E(spl)m8-GFP expression was lost in emc mutant clones (Fig. 3G,H). These results suggest that EGFR signaling was activated but Notch signaling was inactivated early in the medulla development in emc mutant. As EGFR and Notch signaling act positively and negatively, respectively (Sato et al., 2013; Yasugi et al., 2010), the proneural wave progress is more likely and actually enhanced in emc mutant.

Loss of Emc resulted in opposite phenotypes to that of Da: L'sc expression and the emergence of neural stem cells were precociously found in emc mutant clones (Fig. 3D,E), but L'sc expression was prolonged and neural stem cells were lost in da mutant clones (Fig. 4A,B). The antagonistic effects of Emc towards Da have been well studied in different tissues, such as wing imaginal discs (Cabrera et al., 1994; Wang and Baker, 2015a) or eye imaginal discs (Brown et al., 1995; Cadigan et al., 2002; Bhattacharya and Baker, 2011; Spratford and Kumar, 2015). Especially in the developing eye imaginal disc, the progress of morphogenetic furrow is accelerated in emc mutant or da-overexpressing clones (Brown et al., 1995; Cadigan et al., 2002; Bhattacharya and Baker, 2011). The present results indicate that this relationship between Emc and Da is also found in the developing medulla. This result is reasonable because there is a single E (Da) and ID protein (Emc) in Drosophila (Wang and Baker, 2015b). We speculate that Emc binds to Da in neuroepithelial cells in the developing medulla and consequently inhibits Da functions to suppress precocious differentiation from neuroepithelial cells to neural stem cells (Fig. 7A). The loss of Hairy, another bHLH suppressor (Ohsako et al., 1994), did not result in such inhibitory effects on Da, which could be achieved by Emc as an ID protein. This is also the case in the regulation of the morphogenetic furrow in the developing eye imaginal disc. The morphogenetic furrow normally progresses without affecting either Da or Emc expression in hairy mutant clones (Murre et al., 1989).

We observed that L'sc expression was not only delayed but also prolonged more medially in the da mutant clones (Figs 4A and 5B,C), as observed in an active form of N-overexpressing clones (Figs 5E,F and 7B). This finding is consistent with the observations that N was ectopically expressed and Notch signaling was also persistently activated in da mutant clones (Fig. 5B,D). Our finding that Da binds CNS enhancer regions in the N locus with remarkably high frequency indicates that Da suppresses N expression directly (Fig. 5A). The present results also suggest that this relationship between Da and N is applicable in NSCs as ectopic N expression was observed in the da mutant clones immediately next to the boundary between the developing medulla and the central brain (Fig. 5B). Another bHLH transcription factor must be involved in the regulation of N expression because Da heterodimerizes with it to regulate target gene expression (Murre et al., 1989; Cabrera and Alonso, 1991). One of the candidates is L'sc, as N ectopically appeared in Df(1)260-1 clones, as observed in da mutant clones (Fig. 6A). This finding is also supported by ectopic expression experiments of L'sc, which showed precocious termination of N expression (Fig. 6B). However, as the transition from neuroepithelial cells to neural stem cells is delayed in Df(1)260-1 clones (Yasugi et al., 2008), the ectopic expression of N found in Df(1)260-1 clones could be a secondary consequence of the delay in the transition. It is necessary to elucidate in detail the relationship between L'sc and N in the regulation of the transition from neuroepithelial cells to neural stem cells in future studies.

Our Da-binding profiling indicates that L'sc is also one of the direct targets of Da (Fig. 4E). Because L'sc expression was not lost but prolonged in da mutant clones, Da plays suppressive roles in L'sc expression rather than inductive roles (Figs 4A and 5B,C). Da could successfully suppress L'sc expression only in nbNSCs but not in the transitional zone (TZ) because L'sc is expressed in the TZ even in the presence of Da (Figs 1E-H′ and 7A). As ectopic L'sc was not observed in da mutant neural stem cells or neurons, this regulation is unlikely to be maintained in those cells (Figs 4C and 5B). Thus, the present results suggest that nbNSCs are different from other neural stem cells and that they are still differentiating: nbNSCs can be visualized by a neural stem cell marker but differentiated incompletely. Recently, a new type of cell located in the medial-most TZ, named transitory epithelial-like neural stem cells (epi-NSCs), was reported (Shard et al., 2020). The epi-NSCs are more like neuroepithelial cells than neural stem cells because they show the expression of a neuroepithelial cell marker instead of a typical neural stem cell marker (E-cadherin⁺, Dpn⁻). As nbNSCs show neural stem cell-like features (Dpn⁺, Ci⁻), they are different from epi-NSCs.

Thus, we conclude that Da is necessary to terminate L'sc expression by suppressing N expression in nbNSCs for completion of the transitional state (Fig. 7). In mammals, TCF4, an ortholog of Da, is highly abundant in the neuroepithelium of the telencephalon or hippocampus during embryonic development (Jung et al., 2018), and is involved in the control of the number of neural stem cells or intermediate progenitors (Fischer et al., 2014; Li et al., 2019; Teixeira et al., 2021). Our present results raise the possibility that TCF4 also plays essential roles in the regulation of differentiation from neuroepithelial to neural stem cells in mammals. TCF4 has been examined as an essential causative gene of Pitt–Hopkins syndrome (Amiel et al., 2007), which is characterized by developmental delay and intellectual disability (Pitt and Hopkins, 1978). Understanding the roles of TCF4 in the regulation of differentiation from neuroepithelial to neural stem cells will help identify therapeutic targets for this syndrome.

All fly lines were reared on standard Drosophila medium at 25°C unless otherwise noted. Fly larvae were dissected at the wandering late third instar. RNA interference (RNAi) experiments were conducted with the Gal4/UAS system (Brand and Perrimon, 1993). The fly lines used were Oregon-R (wild type, Kyoto, 105-669), UAS-cd8gfp (BDSC, 5130), UAS-myr-cd8gfp (BDSC, 32200), UAS-dicer2 (BDSC, 24650), pxb-gal4 (Suzuki et al., 2013), GMR29C07 (ogre-gal4; BDSC, 49340), FRT80B (BDSC, 1988), emc¹ FRT80B (Kyoto, 108-282), h²² FRT80B (a gift from Makoto Sato, Kanazawa University, Japan), da¹⁰ FRT40A (Kyoto, 108-281), hs-flp; act>yellow>gal4 UAS-gfp (Kyoto, 107-724), hs-flp; ubi-gfp FRT40A (BDSC, 5629), hs-flp; ubi-mrfp FRT40A (BDSC, 34500), hs-flp; RpS17⁴ arm-lacZ FRT80B (Kyoto, 118-135), hs-flp; ubi-gfp FRT80B (BDSC, 5630), hs-flp; ubi-gfp FRT82B (BDSC, 5188), Df(1)260-1 FRT19A (Kyoto, 118-134), Df(1)ase1 FRT19A (Kyoto, 118-133), Df(1)sc10-1 FRT19A (Kyoto, 118-131), emc⁰⁴³²² (emc-lacZ, BDSC, 11629), emc-gfp (BDSC, 55835), da-gfp (BDSC, 55836), UAS-da (BDSC, 51669), UAS-l'sc (BDSC, 51670), UAS-NanoDam (Tang et al., 2022), UAS-N^act (Go et al., 1998), UAS-N^full (BDSC, 52309), UAS-pnt RNAi (pnt^JF02227; BDSC, 31936), Delta^RevF10 FRT82B (Makoto Sato), E(spl)mγ-HLH-gfp (Almeida and Bray, 2005), E(spl)m8-HLH-gfp (BDSC, 68186), m8-lacZ (BDSC, 26786), tll-gfp (BDSC, 30874) and tub-gal80^ts (McGuire et al., 2003; Kyoto, 130-454).

We generated homozygous clones for emc¹, da¹⁰, h²², Delta^RevF10, Df(1)260-1, Df(1)ase1 or Df(1)sc10-1 with the FLP-FRT system (Golic, 1991; Xu and Rubin, 1993). emc¹ FRT80B/TM6B or E(spl)m8-HLH-gfp; emc¹ FRT80B/TM6B flies were crossed to hs-flp;; arm-lacZ Rps17⁴ FRT80B/TM6B flies. da¹⁰ FRT40A/CyO or hs-flp; da¹⁰ FRT40A/CyO flies were crossed to hs-flp; ubi-gfp FRT40A or ubi-mrfp FRT40A; E(spl)mγ-HLH-gfp/TM6B flies, respectively. h²² FRT80B/TM3 flies were crossed to hs-flp;; ubi-gfp FRT80B/TM3 flies. Delta^RevF10 FRT82B/TM3 flies were crossed to hs-flp;; ubi-gfp FRT82B/TM3 flies. Female virgin Df(1)260-1 FRT19A, Df(1)ase1 FRT19A or Df(1)sc10-1 FRT19A flies were crossed to hs-flp tub-gal80 FRT19A/Y; act-gal4 UAS-gfp/CyO flies (Lee and Luo, 1999, 2001). flp expression was induced by heat shock at 3 days after egg laying (37°C, 60 min).

For ectopic expression of da or l'sc, tub-Gal80ts; pxb-Gal4 UAS-cd8gfp/TM6B flies were crossed to UAS-da, E(spl)m8-HLH-lacZ; UAS-da or UAS-l'sc flies and raised at 16°C to reduce lethality. The larvae were then transferred to 29°C 1 day before dissection. For the ectopic expression of N, UAS-N^act or UAS-N^full flies were crossed to hs-flp; act>yellow>gal4 UAS-gfp flies. flp expression was induced by heat shock at 3 days after egg laying (34°C, 30 min).

Immunohistochemistry was performed as described previously (Hasegawa et al., 2011) using ibidi Mounting Medium (Ibidi, ib50001). In this study, neural stem cells, neuroepithelial cells and neurons were visualized with Dpn (or Ase), Ci and Elav expression, respectively. The following primary antibodies were used: rabbit anti-Ase [1:1000; generated against the peptide CLSDESMIDAIDWWEAHAPKSNGA (Brand et al., 1993; Álvarez and Díaz-Benjumea, 2018) conjugated with keyhole limpet hemocyanin in this study], guinea pig anti-Dpn (1:1600; James Skeath, Washington University, St Louis, MO, USA), guinea pig anti-L'sc (1:2000; Suzuki et al., 2013), rabbit anti-PntP1 (1:500; James Skeath), rabbit anti-Emc (1:400; Yuh Nung Jan, University of California, San Francisco, CA, USA), rabbit anti-Da (1:500; Yuh Nung Jan), mouse anti-LacZ (1:250; Promega, Z378), chicken anti-LacZ (1:500; AVES lab, BGL-1010), rabbit anti-GFP (1:1600; MBL, 598MS), chicken anti-GFP (1:400; AVES lab, GFP-1010) and sheep anti-GFP (1:500; Bio-Rad, 4745-1051). The following monoclonal antibodies were obtained from the Developmental Studies Hybridoma Bank (DSHB): mouse anti-Delta (1:10), mouse anti-N (1:10), rat anti-Ci (1:8) and rat anti-Elav (1:50). All secondary antibodies were used at 1:200. The secondary antibodies used were anti-guinea pig Alexa 594 (Jackson ImmunoResearch Laboratories, 706-585-148), anti-guinea pig Alexa 647 (Jackson ImmunoResearch Laboratories, 706-605-148), anti-mouse Cy3 (Jackson ImmunoResearch Laboratories, 715-165-151), anti-mouse Cy5 (Jackson ImmunoResearch Laboratories, 715-175-151), anti-mouse Alexa 488 (Jackson ImmunoResearch Laboratories, 715-545-151), anti-rabbit Alexa 594 (Jackson ImmunoResearch Laboratories, 711-585-152), anti-chicken Alexa 488 (Jackson ImmunoResearch Laboratories, 703-545-155), anti-chicken Alexa 647 (Jackson ImmunoResearch Laboratories, 703-605-155), anti-rabbit Alexa 488 (Thermo Fisher, A-11034), anti-rat Alexa 647 (Thermo Fisher, A-21247) and anti-sheep Alexa488 (Thermo Fisher, A-11015). Confocal images were acquired using an Olympus FV-1000 and analyzed with Fiji (Schindelin et al., 2012) or FluoRender (Wan et al., 2017).

To conduct nanoDamID for Da, we crossed the flies carrying da::gfp with UAS-NanoDam, tub-Gal80ts/CyO dfd-yfp; ogre-gal4/TM6B and raised them at 16°C to restrict the expression of NanoDam for 10-11 days. The larvae were transferred to 29°C 1 day before dissection. Brains were dissected from larvae at the wandering late third instar stage, rinsed with PBS and stored at −80°C until genomic DNA extraction. Four replicate experiments were performed and 50-60 brains were used for each replicate to extract genomic DNA. w¹¹¹⁸ flies were used instead of da::gfp flies for control experiments.

NanoDam samples were processed to prepare each library as described previously (Marshall et al., 2016; Otsuki and Brand, 2019; Tang et al., 2022). Sequencing was performed using an Illumina HiSeq 2500 (150 bp pair-end, GenWiz/Azenta). Sequencing datasets were analyzed using the damidseq_pipeline as described previously (Marshall and Brand, 2015) and aligned to Drosophila genome annotation release 6. Peaks were identified through find_peak script (FDR<0.01) and assigned to genes with peak2genes script (https://github.com/owenjm/find_peaks). The NanoDam binding tracks were visualized with Integrative Genomics Viewer (IGV, version 2.16.0) software (Robinson et al., 2011). Enrichment analysis was performed through WebGestalt (Liao et al., 2019; https://www.webgestalt.org). Distribution of the E-box motifs (5′-CACCTG-3′ and 5′-CAGCTG-3′) was examined using the JASPAR database (Mathelier et al., 2016) and CiiiDER (Gearing et al., 2019).

We thank Antonio Baonza, Andrea H. Brand, Sarah Bray, Yuh Nung Jan, Justin Kumar, Makoto Sato and Jim Skeath for antibodies and fly lines. We are grateful to Catherine Davidson of Brand lab for sharing the DamID experiment with us. We thank the Bloomington Stock Center, the Vienna Drosophila RNAi Center, NIG-fly and the Kyoto DGRC for fly strains, and the DSHB and Asian Distribution Center for Segmentation Antibodies for antibodies. We also thank all of the members of the Suzuki lab at Ibaraki University.

The peer review history is available online at https://journals.biologists.com/dev/lookup/doi/10.1242/dev.202630.reviewer-comments.pdf