- Split View
-
Views
-
Cite
Cite
Soline Chanet, Nicolas Vodovar, Véronique Mayau, François Schweisguth, Genome Engineering-Based Analysis of Bearded Family Genes Reveals Both Functional Redundancy and a Nonessential Function in Lateral Inhibition in Drosophila, Genetics, Volume 182, Issue 4, 1 August 2009, Pages 1101–1108, https://doi.org/10.1534/genetics.109.105023
- Share Icon Share
Abstract
Lateral inhibition mediated by Notch receptor signaling regulates the determination of sensory organ precursor cells (SOPs) in Drosophila. The selection of SOPs from proneural cluster cells appears to rely on a negative feedback loop linking activation of the Notch receptor to downregulation of its ligand Delta within each cell. The molecular basis of this regulatory feedback mechanism is not known. Here, we have tested the role of the Bearded (Brd) family genes in this process. The Drosophila genome encodes eight Brd family members that interact with the E3 ubiquitin ligase Neuralized (Neur) and act as inhibitors of Neur-mediated Delta signaling. Genome engineering technologies were used to create specific deletions of all eight Brd family genes. We find that the Brd family genes mα, m4, and m6 encoded by the Enhancer of split Complex (E(spl)-C) are dispensable for Drosophila development and that deletion of the five Brd family genes encoded by the Brd Complex only reduces viability. However, deletion of all Brd family genes results in embryonic lethality. Additionally, the mα, m4, and m6 genes act redundantly with the other five Brd family genes to spatially restrict Notch activation in stage 5 embryos. These data reveal that the Brd family genes have an essential but redundant activity. While the activity of all eight Brd genes appears to be dispensable for SOP determination, clone border studies indicate that both the relative activity levels of Neur and Brd family members influence competition for the SOP fate during lateral inhibition. We propose that inhibition of Neur–Delta interaction by Brd family members is part of the feedback loop that underlies lateral inhibition in Drosophila.
SIGNALING through the Notch receptor is widely used in animal development to control cell fate choices and regulate pattern formation. One of the best-known functions of Notch is to mediate lateral inhibition, a patterning process that regulates the formation of differentiated structures at regular spatial and/or temporal intervals. In Drosophila, inhibitory cell–cell interactions mediated by the Notch receptor are responsible for the regular spacing of sensory bristles at the surface of the dorsal thorax (Simpson 1990). Sensory organ precursor cells (SOPs) are singled out from within groups of proneural cluster (PNC) cells that can differentiate as either epidermal cells or as SOPs (Ghysen and Dambly-Chaudiere 1989). Selection of SOPs among PNC cells is classically viewed as the result of a competition for adoption of the SOP fate (Heitzler and Simpson 1991). All PNC cells are thought to initially express similar levels of the receptor Notch and its ligand Delta (Dl). As a consequence, all cells in the cluster may initially inhibit one another. However, activation of Notch in a given cell is thought to decrease that cell's potential to become a SOP and to also reduce its ability to send back the Dl signal that activates Notch at the surface of neighboring cells. This negative feedback loop therefore ensures that a single cell emerges as a winner of this competition. This model whereby PNC cells compete for the adoption of the SOP fate is supported by clonal studies demonstrating that the level of Notch activity within a cell can influence the fate of its neighbors (Heitzler and Simpson 1991).
Downregulation of Dl in response to Notch receptor activation has been proposed to be under transcriptional control, as originally shown for the feedback loop operating between the Lin-12 receptor and its ligand Lag-2 (Wilkinsonet al. 1994). In this model, transcriptional activation of the bHLH genes of the Enhancer of split Complex (E(spl)-C) by activated Notch results in the downregulation of proneural gene activity, thereby leading to a downregulation of Dl gene transcription (Heitzleret al. 1996). This molecular scenario is, however, not supported by the analysis of Dl transcription: Dl transcription levels in emergent and newly specified SOPs were found to be similar to those observed in neighboring PNC cells (Parkset al. 1997). This observation therefore argues against negative regulation of Dl transcription as a central mechanism for competition. Thus, the molecular basis underlying the negative feedback linking Notch activation to Dl inhibition remains unknown.
The activity of Dl is positively regulated at the post-transcriptional level by the E3 ubiquitin ligase Neuralized (Neur) and Neur-dependent Dl signaling is essential for proper SOP specification (Boulianneet al. 1991; Laiet al. 2001; Pavlopouloset al. 2001; LeBorgne and Schweisguth 2003; LeBorgneet al. 2005). Recent studies have shown that regulation of Dl activity by Neur is antagonized by proteins of the Bearded (Brd) family. Molecular data indicate that Brd family proteins physically interact with Neur and antagonize its interaction with Dl (Levitenet al. 1997; Laiet al. 2000a,b; Bardin and Schweisguth 2006; DeRenziset al. 2006). Additionally, most Brd family genes are expressed in PNC cells (Wurmbachet al. 1999; Laiet al. 2000a,b). While transcriptional regulation studies are suggestive of a role of these genes in lateral inhibition (Singsonet al. 1994; Castroet al. 2005), whether inhibition of Neur by Brd family proteins regulates lateral inhibition has not been tested using loss-of-function analysis.
The Brd family consists of the proteins BobA, BobB, Tom, Brd, and Ocho that are encoded by the Bearded Complex (Brd-C) and of the proteins mα, m4, and m6 that are encoded by the E(spl)-C. The structurally related protein m2 is not considered here as a Brd family member as it does not include the Neur binding motif and does not antagonize Notch (Laiet al. 2000b; Bardin and Schweisguth 2006). To investigate the function of the Brd family genes, we have used here genomic engineering approaches to delete each of these genes. Our analysis demonstrates that Brd family genes act redundantly in the embryo and suggests that inhibition of Neur by Brd family members participates in the negative feedback loop linking Notch activation to downregulation of Dl within a cell during lateral inhibition.
MATERIALS AND METHODS
Drosophila stocks:
The P-elements RBe00084 and XPd08311 from the Exelixis collection (https://drosophila.med.harvard.edu/) were used to generate the Df(3)E(spl)δ-6 deficiency using Flp-mediated recombination as described in (Thibaultet al. 2004). The Dp(3;2)E(spl)δ-8, Dp(3;2)E(spl)δ-8Δα46, Dp(3;2)Brd-C, and Dp(3;2)CG13466 were generated using phiC31-mediated site-specific integration (Venkenet al. 2006; Bischofet al. 2007). The DpE(spl)δ-8 and DpE(spl)δ-8Δα46 were micro-injected in vas-phiC31-zh-2A; ZH-attP-51D embryos (Bischofet al. 2007). Micro-injection of DpBrd-C and DpCG13466 into vas-phiC31-zh-2A; ZH-attP-58A embryos was performed by BestGene (http://www.thebestgene.com). Other stocks used in this study include: Df(3)Brd-C1 (Bardin and Schweisguth 2006), UAS-mα (Apidianakiset al. 1999), neurIF65 (FlyBase), and tub>GFP, y+>Gal4 (Pinalet al. 2006).
Clones of mα overexpression were generated in hs-flp; UAS-mα; tub>GFP, y+>Gal4 pupae. Mitotic clones were induced by a 45-min heat-shock at 36.5° in first and second instar larvae of the following genotypes (numbering as in Figure 4):
(1) Control wild-type clones for 3L: hs-flp tub-Gal4 UAS-GFP;; FRT2A/tub-Gal80 FRT2A.
(2) Df(3)Brd-C1 loss-of-function clones: hs-flp tub-Gal4 UAS-GFP;; FRT2A Df(3)Brd-C1/FRT2A tub-Gal80.
(3) Df(3)Brd-C1 mutant clones in mα m4 m6 triple mutant flies: hs-flp tub-Gal4 UAS-GFP; Dp(3;2)E(spl)δ-8Δα46; Df(3)Brd-C1 FRT2A Df(3)E(spl)δ-6/tub-Gal80 FRT2A Df(3)E(spl)δ-6.
(4) Control wild-type clones for 2R: hs-flp tub-Gal4 UAS-GFP; FRT42B/FRT42B tub-Gal80.
(5) mα, m4, m6 loss-of-function clones: hs-flp tub-Gal4 UAS-GFP; FRT42B Dp(3;2)E(spl)δ-8Δα46/FRT42B Dp(3;2)E(spl)δ-8 tub-Gal80; Df(3) E(spl)δ-6.
(6) Control wild-type clones for 3R: hs-flp tub-Gal4 UAS-GFP;; FRT82B ubi-nls-GFP/FRT82B tub-Gal80.
(7) neur mutant clones: hs-flp tub-Gal4 UAS-GFP;; FRT82B ubi-nls-GFP neurIF65/FRT82B tub-Gal80.
Molecular biology:
Duplications were generated from BACs RPCI-98-13F13 and RPCI-98-01H12 (http://bacpac.chori.org) cloned into the attB-P[acman]-Ap vector using recombineering mediated gap repair as described in Venkenet al. (2006). The PCR-amplified 5′ and 3′ homology arms were cloned into attB-P[acman]-Ap using Not1 and EcoR1. Resulting plasmids were linearized using BamHI, digested with ExoSAP (USB) and used for recombineering with RPCI-98-13F13 and RPCI-98-01H12 in Escherichia coli SW102 as described at http://recombineering.ncifcrf.gov/Protocol.asp.
The following primers were used to PCR amplify the 5′ and 3′ homology arms subsequently used to generate duplications:
DpE(spl)_UF: CCCGGCCGTTAAACCAAGTTCACCTCTC
DpE(spl)_UR: CAATAAAGGGATCCCTTGTTTAATGCGGATAACG
DpE(spl)_DF: AACAAGGGATCCCTTTATTGGGATGTTGGGAG
DpE(spl)_DR: CGAATTCAGATCTGTACATGTTCTTCAGG
DpBrd_UF: CCGCGGCCGCTGGGGTTTCTTGCAACCAAC
DpBrd_UR: CATTTTGTGGATCCCTTATTAGTGGTAAGGGCAG
DpBrd_DF: CATATAAGGGATCCACAAAATGTTGGGGTAACAGC
DpBrd_DR: CCGAATTCCTATTGTTACCCTCTTCTAGG
DpCG13466_UF: CCCGGCCGTAGTGTGCCGGTTGTGTGTG
DpCG13466_UR: AACCCCAGGATCCGACTCCCACAACAGCGAGAG
DpCG13466_DF: GAGTCGGATCCTGGGGTTTCTTGCAACCAAC
DpCG13466_DR: CCGAATTCCTTATTAGTGGTAAGGGCAG.
The mα, m4, and m6 genes were deleted from the RPCI-98-13F13 using recombineering with a positive/negative selection galK cassette (Warminget al. 2005).
The following primers were used to PCR amplify the 5′ and 3′ homology arms used to delete the mα, m4, and m6 genes:
mα-UF: GAATGCCCATTAGGAATAC
mα-UR: TGTGTGTAACCGCTCGAGCGACAGAGAGG
mα-DF: CGCTCGAGCGGGTTACACACAACAAGTAG
mα-DR: CAATGACCCAGTTAGATAG
m4-UF: CATCACCTCAAATGTATGC
m4-UR: GTTTCTCCACTCGTATGAAATGGGCCTTCTC
m4-DF: TTTCATACGAGTGGAGAAACCCGAAGCCGAG
m4-DR: GGCGCCTCCGTTCGAGCGTTG
m6-UF: GCAAGTGCAATATGTGTC
m6-UR: ACAGTGACGAGTGCCACATGTGGCGGAC
m6-DF: ACATGTGGCATCGTCACTGTTTGTTACGTGG
m6-DR: TCAGGCCTGAAGCTTGAGAATC.
Immunostainings and RNA in situ hybridization:
The following antibodies were used: guinea pig anti-Sens (1:3000; H. Bellen), rabbit anti-GFP (Molecular Probes; 1:1000). RNA in situ hybridization of stage 5 embryos was performed using standard procedures with DIG-labeled RNA probes. Embryos were genotyped using the TM3 hb-lacZ balancer.
RESULTS
Brd family genes of the E(spl)-C are not essential:
To genetically study the function of the mα, m4, and m6 genes, we have used combinations of molecularly defined deletions and duplications. We first generated a 41-kb deletion, Df(3)E(spl)δ-6, that removes all four Brd genes and five of the seven bHLH repressors (Figure 1A). This deletion is associated with a strong neurogenic phenotype in embryos (not shown). This phenotype was fully rescued by a 46-kb duplication covering the entire E(spl)-C locus, DpE(spl)δ-8, which was designed and generated from a BAC carrying the E(spl)-C genomic DNA by gap repair in E. coli and integrated within the fly genome using phiC31-mediated transgenesis (Venkenet al. 2006; Bischofet al. 2007) (Figure 1A). Indeed, DpE(spl)δ-8 Df(3)E(spl)δ-6 embryos are viable (only 10% of these embryos do not hatch, n = 115, as compared with 2%, n = 273, for wild-type control embryos). We further engineered this duplication using recombineering in E. coli to precisely delete the mα, m4, and m6 genes, from the TATA box to the end of the 3′-UTR included, resulting in DpE(spl)δ-8Δα46 (Figure 1A). This mutant duplication rescued the neurogenic mutant phenotype associated with the E(spl)-C deletion, demonstrating that this duplication is functional and that the neurogenic phenotype results from the loss of the bHLH genes (not shown). Additionally, only 6% (n = 299) of the embryos triply mutant for the mα, m4, and m6 genes fail to hatch, and flies homozygous for the E(spl)-C deletion and carrying one copy of the mutant duplication are viable and fertile with no detectable morphological phenotype. We therefore conclude that the mα, m4, and m6 genes are not essential.
Functional redundancy between Brd family genes:
To test whether the mα, m4, and m6 genes are dispensable due to genetic redundancy with other genes of the Brd family, we have produced a genetic background deleted of all Brd family genes. In a first step, we have analyzed the phenotype associated with the loss of the Brd-C genes BobA, BobB, Tom, Brd, and Ocho. The Brd-C1 deletion that removes the Brd-C and truncates CG13466 is largely lethal: 60% (n = 121) of Brd-C1 individuals die as embryos and only 6% reach adulthood. This lethality is partially rescued by a genomic duplication of the Brd-C but not by the CG13466 duplication (Figure 1B): 20% (n = 133) of the DpBrd-C Brd-C1 embryos gave adult flies that are fertile and can be kept as a stock whereas DpCG13466 Brd-C1 embryos gave no escapers. We conclude that the reduced viability associated with the Brd-C1 deletion is due to the loss of the Brd-C genes.
We then generated embryos mutant for all eight Brd family genes, i.e., homozygous for the Brd-C1 and Df(3)E(spl)δ-6 deletions with one copy of the mutant DpE(spl)δ-8Δα46 duplication and found that this genotype is 100% embryonic lethal (n > 120). The wild-type DpE(spl)δ-8 duplication was used as a positive control: 39% (n = 120) of embryos homozygous for the Brd-C1 and Df(3)E(spl)δ-6 deletions with one copy of the wild-type DpE(spl)δ-8 duplication fail to hatch and only 2% reach pupal stages. The reduced viability associated with this genotype is likely to result from the loss of the Brd-C genes since reduced viability was also observed with the Brd-C1 deficiency. Together, these data clearly demonstrate that Brd family genes have an essential and redundant function in the embryo.
Brd family genes act redundantly to restrict the domain of Notch activity along the dorsoventral axis in early embryos:
Previous studies have shown that genes of the Brd-C are collectively required to restrict Dl signaling to mesodermal cells in stage 5 embryos (Bardin and Schweisguth 2006; DeRenziset al. 2006). At this stage, neur-dependent Dl endocytosis in mesodermal cells results in the activation of Notch in a single row of cells on either side of the mesoderm. These cells express the Notch target gene single-minded gene (sim) and form the mesectoderm (Figure 2, A and F) (Martin-Bermudoet al. 1995; Morel and Schweisguth 2000; Cowden and Levine 2002; Morelet al. 2003; Bardin and Schweisguth 2006; DeRenziset al. 2006) (Figure 2, A, B, and G). Brd family genes encoded by both complexes are expressed in nonmesodermal cells at this stage (Nagelet al. 2000; Zaffran and Frasch 2000). We first confirmed that Brd-C genes contribute to restrict the expression of sim to the mesectoderm (Figure 2, C and G) (Bardin and Schweisguth 2006; DeRenziset al. 2006). We next investigated the role of the mα, m4, and m6 genes in this process. We found that loss of mα, m4, and m6 gene activities had no detectable effect on sim expression (Figure 2D). However, deletion of these genes strongly enhanced the Brd-C phenotype: sim transcripts were detected in 3–5 rows of cells dorsal to the mesoderm in embryos mutant for all eight Brd family members (Figure 2E). This ectopic expression of sim was strongly suppressed by the loss of zygotic neur activity, indicating that Brd family genes inhibit the activity of neur (Figure 2F). Of note, the neur loss of sim expression phenotype was suppressed by the zygotic loss of Brd family genes (compare Figure 2, B and F). We interpret this suppression to suggest that maternally provided Neur is sufficient to activate Notch only in the complete absence of all Brd family antagonists. Together, these data demonstrate that Brd family genes of the E(spl)-C and Brd-C act redundantly to restrict the spatial domain of Neur-dependent Dl signaling, hence Notch receptor activation, in the early embryo.
Brd family genes are not required for SOP determination:
Previous studies have shown that Brd family genes of the E(spl)-C and Brd-C are expressed in PNC cells and are positively regulated by both proneural factors and Notch signaling (Singsonet al. 1994; Castroet al. 2005). These observations have suggested that Notch activation within a cell may result in Neur inhibition via the transcriptional regulation of Brd family genes (Bardin and Schweisguth 2006). To test whether the loss of Brd gene activity actually results in increased Dl activity accompanied by a SOP loss phenotype, we have studied the bristle phenotype of flies lacking some or all of the Brd family genes. No SOP and/or bristle loss was detected in Brd-C mutant clones, Brd-C homozygous mutant escaper flies, triple mα m4 m6 mutant flies or in clones of cells mutant for all eight Brd family genes (Figure 3). These observations indicate that Brd family genes are not strictly required for SOP determination. We speculate that other mechanisms, including transcriptional repression of Notch target genes by Hairless and Su(H), may act to buffer Notch signaling activity in SOPs (Banget al. 1995; Castroet al. 2005).
Brd family genes act during SOP selection:
To further test the possible role of the Brd family genes in lateral inhibition, we have used a fate competition assay (Heitzler and Simpson 1991). This assay compares the relative ability of cells of different genotypes to compete for the adoption of the SOP fate along mosaic clone borders. For instance, cells with a twofold higher level of Dl activity were shown to be more likely to inhibit their neighbors than to be inhibited by them and most often emerge as winners of this competition. Here, we have examined the genotypes of SOPs along clone borders separating cells with varying copy numbers of Brd family genes in mosaic pupae at 16–18 hr after puparium formation (APF). We found that cells homozygous for the Brd-C1 deletion are significantly more likely to become SOPs than cells with one or two copies of the Brd-C (65%, n = 145; Figure 4, A and D). Similarly, cells that are triply mutant for the mα, m4, and m6 genes are more likely to become SOPs than cells with one or two copies of each of these genes (60%, n = 194; Figure 4, B and D). For technical reasons, we could not study clone borders separating cells mutant for all Brd family genes and cells with one or two copies of each Brd family gene. These data indicate that the level of Brd family gene activity influences the choice of cell fate along mosaic border. We therefore conclude that Brd family genes of the E(spl)-C and Brd-C act in the lateral inhibition process prior to stable commitment to the SOP fate.
The level of Neur within a cell influences the fate of neighboring cells:
Since Brd family members likely act during SOP selection by inhibiting Neur in PNC cells, we predict that the relative levels of neur activity should also influence the outcome of this competition, albeit in a manner opposite to the one seen for the Brd family genes. While previous clonal studies have indicated that neur is required for lateral inhibition (Yehet al. 2000; LeBorgne and Schweisguth 2003), the influence of neur gene dosage on sensory cell fate has not been studied. A role for Neur during competition is not necessarily expected since neur transcripts and Neur proteins have been detected in SOPs but not in PNC cells (Boulianneet al. 1991; LeBorgne and Schweisguth 2003). Indeed, this expression pattern suggests that neur may only act in singled out SOPs to reinforce lateral signaling and that it does not act when PNC cells still compete for the SOP fate. Alternatively, it is conceivable that neur is expressed at low levels when PNCs compete for the SOP fate and is eventually upregulated in singled-out SOPs. Consistent with this view, expression of a neur enhancer-trap line is detected, albeit infrequently, in epidermal cells located next to SOPs (Huanget al. 1991). To functionally test whether neur acts during the competition phase prior to the stable commitment to the SOP fate, we have used the fate competition assay described above and have examined the genotypes of SOPs along clone borders separating cells with one or two wild-type copies of the neur gene. Cells with two copies of the neur gene were found to be more likely to become SOPs than cells with one copy (67%, n = 69; Figure 4, C and D). Thus, cells with a twofold reduction in neur activity appear to send a weaker inhibitory signal and are less likely to win the competition. We conclude that the level of neur activity influences the choice of cell fate along mosaic border and that neur acts to regulate the singling out of SOPs. Together, our results suggest that inhibition of Neur by Brd family members participates in SOP selection by modulating the activity of Neur in response to Notch activation within each PNC cell.
DISCUSSION
Recent advances in genome engineering now permit the genetic analysis of complex gene families in Drosophila (Venkenet al. 2006; Bischofet al. 2007; Venken and Bellen 2007). In this study, we have addressed the function of the eight Brd family proteins that interact with Neur and inhibit Neur-dependent Dl signaling. We have used FLP/FRT deletion, BAC recombineering, and phiC31-mediated integration technologies to delete specific sets of Brd family members. Our analysis reveals for the first time that Brd family genes have an essential function in the embryo and that they act in a redundant manner. Indeed, deletion of the mα, m4, and m6 genes has no major effect on viability and fertility while the combined loss of the BobA, BobB, Tom, Brd, and Ocho genes reduces viability and has a weak effect on Notch target gene expression in early embryos. However, deletion-based inactivation of all eight Brd family genes known to encode direct Neur interactors results in fully penetrant embryonic lethality that is associated with ectopic Neur-dependent Notch signaling in early embryos. This study therefore establishes that Brd family genes act in a partly redundant manner to antagonize the activity of Neur. Functional redundancy is not, however, strict and the strength of the phenotypes associated with the progressive loss of Brd family genes may be, at least in part, dosage dependent. For instance, loss of Brd-C genes has milder effect on both viability and sim ectopic expression than the complete loss of all eight Brd family genes.
Our genetic analysis revealed that cells with lower levels of Brd family gene activity relative to their neighbors are more likely to win the competition and to adopt the SOP fate in the pupal notum. Conversely, cells with lower levels of neur activity relative to their neighbors are more likely to loose competition and to differentiate as epidermal cells. These data indicate that Neur-dependent Dl signaling regulates the singling out of SOPs and that inhibition of Neur by Brd family members operates during the process of SOP selection. Since expression of Brd family genes is regulated by Notch (Wurmbachet al. 1999; Laiet al. 2000b; Castroet al. 2005), we propose that inhibition of Neur by Brd family members is one of the mechanisms whereby activation of Notch in a given cell results in the downregulation of Dl. Thus, our study provides experimental support for a novel molecular mechanism underlying the negative feedback loop mechanism operating during lateral inhibition in Drosophila (Figure 5). While our clone border analysis indicates that Brd family genes influence SOP selection, it is likely that the molecular mechanism proposed here based on the activity and regulation of the Brd family genes acts in parallel with other feedback mechanisms since SOPs are properly selected in the absence of all Brd family genes.
Several features of our model deserve mention. First, this model involves a single transcriptional step followed by one post-transcriptional regulatory step. Second, the Notch transcriptional target genes are small, typically less than 1 kb, and encode small cytoplasmic proteins, indicating that the transcription/translation time delay between Notch receptor activation and Delta inhibition is minimal. Third, Notch activates several functionally redundant target genes at once. This may therefore serve to amplify the signal of Notch activation. We speculate that these features may be important for rapid and efficient feedback regulation.
A single Brd family gene has been identified in other insect genomes. So, how could we explain the increase in gene copy number in Drosophila? First, functional diversification may be associated with this increase. For instance, unlike other Brd family members, m6 is expressed predominantly in muscles, suggesting that it may have acquired a specific function in this tissue. Similarly, the Brd and Bob proteins lack motifs 3 and 4, suggesting that their molecular activity and/or localization may slightly differ (Laiet al. 2000b). Second, an increase in the number of Notch target genes acting in parallel may have been selected as a means to selectively amplify one specific response of the genome to Notch activation. Finally, this increase may have been evolutionarily selected as a response to constraints exerted internally on the gene regulatory network by miRNAs acting to inhibit gene expression (Laiet al. 2005). This interpretation is consistent with the notion that miRNAs provide a mechanism for internal selection favoring the emergence of a stable complex gene network.
Acknowledgements
We thank S. Artavanis-Tsakonas, K. Basler, H. Bellen, J. Bischof, E. Lai, F. Pichaud, K. Venken, the Bloomington Stock Center, the Drosophila Genomics Resource Center, and the Developmental Studies Hybridoma Bank for flies, antibodies, and plasmids. We are particularly thankful to J. Bishoff and K. Venken for generously sharing reagents and expertise. We are grateful to A. Bardin for generating the Df(3)E(spl)δ-6 deletion and for her many inputs into this project. We thank all lab members for discussion and comments on the manuscript. This work was funded by the Institut Pasteur, Ecole Normale Supérieure, Centre National de la Recherche Scientifique, Agence Nationale de la Recherche (grant no. 05-BLAN-0277), and Association pour la Recherche Contre le Cancer (grant no. 3415). S.C. received a fellowship from the Ministere de l'Education Nationale de la Recherche et de Technologie.
References
Apidianakis, Y., A. C. Nagel, A. Chalkiadaki, A. Preiss and C. Delidakis,
Bang, A. G., A. M. Bailey and J. W. Posakony,
Bardin, A. J., and F. Schweisguth,
Bischof, J., R. K. Maeda, M. Hediger, F. Karch and K. Basler,
Boulianne, G. L., A. de la Concha, J. A. Campos-Ortega, L. Y. Jan and Y. N. Jan,
Castro, B., S. Barolo, A. M. Bailey and J. W. Posakony,
Cowden, J., and M. Levine,
De Renzis, S., J. Yu, R. Zinzen and E. Wieschaus,
Ghysen, A., and C. Dambly-Chaudiere,
Heitzler, P., M. Bourouis, L. Ruel, C. Carteret and P. Simpson,
Heitzler, P., and P. Simpson,
Huang, F., C. Dambly-Chaudiere and A. Ghysen,
Lai, E. C., R. Bodner, J. Kavaler, G. Freschi and J. W. Posakony,
Lai, E. C., R. Bodner and J. W. Posakony,
Lai, E. C., G. A. Deblandre, C. Kintner and G. M. Rubin,
Lai, E. C., B. Tam and G. M. Rubin,
Le Borgne, R., A. Bardin and F. Schweisguth,
Le Borgne, R., and F. Schweisguth,
Leviten, M. W., E. C. Lai and J. W. Posakony,
Martin-Bermudo, M. D., A. Carmena and F. Jimenez,
Morel, V., R. Le Borgne and F. Schweisguth,
Morel, V., and F. Schweisguth,
Nagel, A. C., Y. Apidianakis, I. Wech, D. Maier, C. Delidakis et al.,
Parks, A. L., S. S. Huppert and M. A. Muskavitch,
Pavlopoulos, E., C. Pitsouli, K. M. Klueg, M. A. Muskavitch, N. K. Moschonas et al.,
Pinal, N., D. C. Goberdhan, L. Collinson, Y. Fujita, I. M. Cox et al.,
Simpson, P.,
Singson, A., M. W. Leviten, A. G. Bang, X. H. Hua and J. W. Posakony,
Thibault, S. T., M. A. Singer, W. Y. Miyazaki, B. Milash, N. A. Dompe et al.,
Venken, K. J., and H. J. Bellen,
Venken, K. J., Y. He, R. A. Hoskins and H. J. Bellen,
Warming, S., N. Costantino, D. L. Court, N. A. Jenkins and N. G. Copeland,
Wilkinson, H. A., K. Fitzgerald and I. Greenwald,
Wurmbach, E., I. Wech and A. Preiss,
Yeh, E., L. Zhou, N. Rudzik and G. L. Boulianne,
Zaffran, S., and M. Frasch,