Proneural Basic Helix-Loop-Helix Proteins and Epidermal Growth Factor Receptor Signaling Coordinately Regulate Cell Type Specification and cdk Inhibitor Expression during Development

Abstract

Cell differentiation and cell cycle exit are coordinately regulated during development; however, the molecular logic underlying this regulation is not known. The Drosophila cdk inhibitor Dacapo (Dap) is one of the key cell cycle regulators that exhibit dynamic expression during development and contribute to the developmental regulation of the cell cycle. In this study, regulation of Dap expression during cell type specification was investigated. The expression of Dap in the R2 and R5 precursors of the developing eye and in the newly recruited leg disc femoral sense organ precursors was found to be controlled by the epidermal growth factor receptor signaling-regulated transcription factor Pointed (Pnt) and the proneural basic helix-loop-helix proteins Atonal (Ato) and Daughterless (Da). We show that Pnt, Ato, and Da regulate Dap expression directly through their respective binding sites precisely at the time when these transcription factors function to specify neural fates. These results show that Dap expression is directly regulated by developmental mechanisms that simultaneously control cell type specification. This is potentially a general mechanism by which the expression of key cell cycle regulators is coordinated with differentiation during normal development. The direct regulation of key cell cycle regulators by the differentiation factors ensures coordinated regulation of cell cycle and differentiation.

Cell differentiation and cell cycle exit are coordinately regulated during development; however, the molecular logic underlying this regulation is not known. Although several cell cycle regulators have been found to be regulated by developmental cues, there has been no molecular characterization of the developmental mechanisms that regulate the expression of cell cycle regulators to allow a mechanistic understanding of how cell differentiation and cell cycle regulation are coordinated during development.

The Drosophila developing eye is a well-established system to study the coordination between cell cycle regulation and differentiation. The adult Drosophila eye is derived from the so-called eye imaginal discs, which proliferate and differentiate during the larval and pupal stages. Photoreceptor differentiation is initiated within a region referred to as the morphogenetic furrow (MF), which is marked by an indentation in the third larval eye disc. The MF progresses from the posterior part of the eye disc toward the anterior end. Cells in the anterior divide asynchronously, whereas cells in the MF are arrested at G₁ and initiate photoreceptor differentiation by forming regularly spaced preclusters. These preclusters will eventually develop into ommatidia, which are the units of the fly compound eye. Posterior to the MF, cells that are not in the clusters undergo a single round of synchronous division, the second mitotic wave, before they exit from the cell cycle (32). In the developing eye, cell cycle exit of the differentiating photoreceptor neurons is controlled by the redundant function of RBF and Dacapo (Dap) (13). Dap is expressed in and just posterior to the MF, where photoreceptor cell differentiation initiates (8, 22). Studies of the cis-regulatory elements that regulate Dap expression identified many independent enhancers, including an eye enhancer, Dap-HB (23, 24).

There are eight photoreceptor cells in each ommatidium, referred to as R1 to R8. During photoreceptor cell differentiation, R8 is determined first, followed by R2 and R5. Specification of the R8 photoreceptor cell requires the proneural genes Atonal (Ato) and Daughterless (Da) (4, 18). Ato and Da are basic helix-loop-helix (bHLH) transcription factors that function together as a heterodimer. Ato is initially expressed at the anterior edge of the MF and becomes progressively refined to clusters of decreasing cell numbers. Immediately before R8 specification, Ato is detected in clusters of two to three cells, which are referred to as “R8 equivalence groups” (9, 18). Later, Ato is maintained only in the R8 photoreceptors while Pointed (Pnt), an Ets family transcription factor regulated by epidermal growth factor receptor (EGFR) signaling, is required for the differentiation and cell cycle arrest of R2 and R5 photoreceptors (34). In addition, EGFR signaling is also required for the recruitment of all subsequent photoreceptors (10).

In this study, we investigated the regulation of Dap expression during cell type specification. We show that a Dap eye disc enhancer (Dap-HB) (23) is active specifically in the R2 and R5 precursors in the developing eye and surprisingly also in the newly recruited leg disc femoral sense organ precursors (SOPs). We demonstrate that Dap-HB enhancer activity in these two cell types is regulated by Ato, Da, and Pnt through the E-box and Pnt binding sites, respectively. Removal of either Ato-Da or Pnt blocks Dap expression in vivo. Similarly, mutations of either the E-box or Pnt binding site significantly decreased Dap-HB enhancer activity. As Ato-Da and EGFR signaling has been shown or implicated in both leg disc femoral SOP recruitment and R2 and R5 specification (18, 34-36), these results indicate that the combination of differentiation factors directly regulates Dap expression while simultaneously functioning in cell fate specification. This is potentially a general mechanism by which dynamic expression of a cdk inhibitor is controlled and coordinated with cell fate specification during normal development.

MATERIALS AND METHODS

Fly stocks and antibodies.

The following fly strains were used in this study: Dap-HB-lacZ (23), UAS-Ato and UAS-Da (5), dap⁴, dap⁴⁴⁵⁴ (8, 22), da¹⁰ (4), ato³ (19), and pnt^5484Δ19 (27). The following antibodies were used: anti-Dap (7), anti-PntP1 (1), anti-Ato (19), and anti-Bar (16). Anti-β-galactosidase (mAb40-1a), Elav, and Rough (ro-62C2A8) were obtained from the Developmental Studies Hybridoma Bank. The genotypes used in this study were (i) GMRGal4/+; UAS-Ato UAS-Da/+, (ii) GMRGal4/+; UAS-Ato UAS-Da/Dap-HB-LacZ, (iii) GMRGal4/UAS-Yan; UAS-Ato UAS-Da/Dap-HB-LacZ, (iv) GMRGal4/UAS-Da; UAS-Ato UAS-PntP1/Dap-HB-LacZ, (v) GMRGal4/+; UAS-Ato UAS-Da/Dap-HB mut1-2-GFP, (vi) Df(2L)da¹⁰ FRT40A/UbiGFP FRT40A, (vii) Df(2L)da¹⁰ FRT40A/UbiGFP FRT40A; Dap-HB-lacZ/+, (viii) Df(2L)da¹⁰ Su(H)^Δ47 FRT40A/UbiGFP FRT40A, (ix) Df(2L)da¹⁰ Su(H)^Δ47 FRT40A/UbiGFP FRT40A; Dap-HB-lacZ/+, (x) hsFLP; FRT82B, ato³/FRT82B Ubi-GFP RpS3, (xi) hsFLP; Dap-HB-GFP/+; FRT82B, ato³/FRT82B, Arm-lacZ (cytoplasmic), and (xii) hspFLP; Dap-HB-GFP/+; FRT82B, pnt^5484Δ19/FRT82B, Arm-lacZ (cytoplasmic).

BrdU incorporation, immunohistochemistry, and in situ hybridization.

Eye discs were dissected, incubated with bromodeoxyuridine (BrdU; final concentration, 75 μg/ml) at room temperature for 60 min, washed with phosphate-buffered saline (PBS), and fixed with 4% paraformaldehyde in PBS, followed by postfixing with 4% paraformaldehyde in PBS-0.6% Tween 20. The discs were washed with DNase I buffer, followed by incubation with DNase I (100 U/500 μl) for 1 h. Mouse anti-BrdU antibody (Becton Dickinson) was used at a 1:50 dilution. Immunohistochemistry and in situ hybridization were performed essentially as previously described (11).

Gel shift assay and generation of transgenic reporter lines.

His₆-tagged Da (amino acids 347 to 710) and Ato (amino acids 7 to 312) were expressed in bacteria, purified, and renatured together. Gel shift assays were carried out as previously described (12), except that oligonucleotides with mutated E-box binding sites were also labeled as probes. The sequences of the WT (wild-type) and Mut (mutant) probes are as follows: WT1, TCAAAGCTGACTGCAGCTGTTTCAGCTCCTCAT; Mut1, TCAAAGCTGACTGTAAATGTTTCAGCTCCTCAT; WT2, CTATAAATGCCAACAGCTGTGACTCCGCTTTGT; Mut2, CTATAAATGCCAATAAATGTGACTCCGCTTTGT; WT3, ACCGGAAATGCAGCAGCTGACTTTGATTTGGCA; Mut3, ACCGGAAATGCAGTAAATGACTTTGATTTGGCA; E-CG, GCCGTCCACGTGCCACAATCTGGGAA; Mut-CG, GCCGTCTAAATGCCACAATCTGGGAA. The WT and mutant E-box sites are underlined. A PhosphorImager from Molecular Dynamics was used to scan the image, and ImageQuant was used to quantify the intensities of specific gel shift bands. Gel shift assays of Ato-Da with Pnt were carried out with proteins derived by in vitro transcription and translation with a kit from Promega.

The same mutations in the E-box binding sites shown here were introduced into the Dap-HB enhancer by PCR to generate specific E-box-mutated Dap-HB enhancers. The WT and mutated Dap-HB enhancers were sequenced and cloned into the pHStinger vector (2) to generate the new Dap-HB transgenic lines. At least two independent transgenic lines were examined for each Dap-HB construct, and no significant variations in expression were observed.

RESULTS

The Dap-HB enhancer targets Dap expression in R2 and R5 progenitor cells in the developing eye.

To characterize Dap expression in the developing eye, we cloned the Dap-HB enhancer into the pHStinger vector (2), which uses green fluorescent protein (GFP) as a reporter. As shown in Fig. 1A, in situ hybridization showed that the WT Dap-HB enhancer conferred GFP reporter expression in some cells within and just posterior to the furrow, consistent with the reported expression of Dap-HB (23). On the other hand, GFP could be detected throughout the posterior part of the eye disc because of the extended stability of the nuclear GFP (Fig. 1B). We took advantage of the stability of the GFP reporter to determine the precursor cell types in which the Dap-HB enhancer was active. Staining of Dap-HB-GFP eye discs with anti-Elav antibody revealed that the GFP reporter was expressed in two cells of each photoreceptor cell cluster (Fig. 1C), suggesting that the Dap-HB enhancer targets Dap expression only in a subset of photoreceptor precursors. As Bar Homeo-box proteins can be used to specifically identify R1 and R6 photoreceptors (16), staining with the anti-Bar antibody revealed that the Dap-HB-GFP reporter was not expressed in R1 or R6 photoreceptor precursors (data not shown). Similarly, staining with anti-Senseless antibody showed that the Dap-HB-GFP reporter was not expressed in R8 photoreceptors either (data not shown). The Rough Homeo-box protein is broadly expressed in the MF. In the posterior, Rough is initially expressed in the R2 and R5 photoreceptor neurons and later expanded to R2, R3, R4, and R5 cells (9, 20). Staining with anti-Rough antibody revealed that the Dap-HB-GFP reporter overlapped with the developing R2 and R5 photoreceptor neurons (Fig. 1D to F). Therefore, Dap-HB is the enhancer that drives Dap expression specifically in R2 and R5 precursors of the developing eye.

FIG. 1.

The Dap-HB enhancer is active in the R2 and R5 precursors of the developing eye and contains multiple E-box binding motifs that are regulated by Atonal and Daughterless. The level of Dap-HB-GFP reporter expression, as determined by in situ hybridization, is shown in panel A, and the level of GFP is shown in panel B. The GFP observed in the posterior part of the eye disc was due to the stability of the GFP reporter protein. (C) Staining with anti-Elav antibody showed that two cells per ommatidial cluster express GFP driven by Dap-HB. (D to F) Dap-HB drives reporter expression in R2 and R5 precursor cells. The Dap-HB-GFP reporter was expressed in cells (R2 and R5) that were labeled by the anti-Rough antibody (arrowheads in panels D to F). The anti-Rough antibody also stained additional cells (R3 and R4) in the posterior that did not have GFP expression. Rough was reported to be expressed initially in the R2 and R5 cells and later in the R2, R3, R4, and R5 cells posterior to the furrow (9, 20). (G) Gel shift assay showing that the Atonal-Daughterless heterodimer bound strongly and specifically to the CAGC/GTG type of E-box sites in the Dap-HB enhancer. WT1 to WT3 (lanes 1, 3, and 7) and Mut1 to Mut3 (lanes 2, 4, and 8) indicate probes with WT or mutated CAGC/GTG type of E-box binding sites, respectively. E-CG (lane 5) and Mut-CG (lane 6) indicate probes with WT and mutated E-box binding sites that are different from consensus Ato-Da binding sites. The same amounts of Ato-Da protein were added for the gel shift experiment. An arrow points to the Ato-Da-DNA complex. The sequences of the WT and E-box-mutated probes are shown in Materials and Methods. (H to K) Ectopic expression of Atonal and Daughterless in the posterior of the developing eye with the GMR-GAL4 driver increased Dap protein levels (H and I), as well as Dap-HB enhancer activity (J and K). The levels of Dap protein in WT (H) or GMR-Gal4 UASAto-Da (I) eye discs were detected by anti-Dap antibody staining. Activity of the Dap eye disc enhancer, Dap-HB, was determined by anti-β-galactosidase staining, which is shown in red (J and K). WT and Ato-Da indicate eye discs from Dap-HB/+ and GMR-Gal4/+; Dap-HB/UASAto UASDa larvae, respectively.

The Dap-HB enhancer contains multiple Ato-Da binding sites.

Inspection of the Dap-HB sequences revealed three of the CAGC/GTG type of E-box binding motifs that are conserved between D. melanogaster and D. pseudoobscura (see Materials and Methods for the E-box sequences). The CAGC/GTG type of E-box motifs are binding sites for the heterodimers of Da with Ato and Ac/Sc family of bHLH transcription factors (17, 25). As both Ato and Da are expressed in the developing eye, we carried out gel shift assays to determine if Ato-Da heterodimers can bind to the observed E-box binding sites in the Dap-HB enhancer. As shown in Fig. 1G, the Ato-Da heterodimer can bind to all three of the conserved CAGC/GTG type of E-box binding sites, with site 2 (WT-2) being the strongest (Fig. 1G, lanes 1, 3, and 7). In contrast, Ato-Da does not bind strongly to a nonconsensus Ato-Da binding site (Fig. 1G, lane 5). As point mutations in each of the E boxes disrupted the binding of Ato-Da to the respective sites (Fig. 1G, lanes 2, 4, and 8), we conclude that Ato-Da binds specifically to all three of the CAGC/GTG types of E-box motifs in the Dap-HB enhancer.

Ato and Da are sufficient to activate Dap expression in the developing eye.

In the developing Drosophila eye, Ato is highly expressed at the anterior edge of the furrow and becomes refined to clusters of decreasing numbers of cells. Before Ato is restricted to the future R8 cells, it is expressed in the two- to three-cell clusters of the R8 equivalence group that are the R8 and R2 and R5 precursors (9, 18). The expression pattern of Ato is consistent with the idea that the Dap-HB enhancer is regulated by the Ato and Da proteins. Therefore, the effects of ectopically activating Ato-Da in the developing eye were characterized with the UAS-Gal4 system (3) to determine if Ato and Da are sufficient to increase the endogenous Dap protein level and Dap-HB enhancer activity.

Dap is expressed in and just posterior to the MF in the developing eye (7, 23). Ectopic expression of Ato and Da in the posterior of the developing eye disc with the GMR-Gal4 driver led to significant accumulation of Dap protein in the posterior (Fig. 1H and I). In addition, ectopic expression of Ato and Da in the posterior region of the eye disc also strongly induced ectopic Dap-HB enhancer activity (Fig. 1J and K). Interestingly, strong upregulation of Dap-HB enhancer activity was observed in the posterior close to the MF but not in the far posterior (Fig. 1K, indicated by an arrow), even though Ato and Da were expressed throughout the posterior of the eye disc. These observations suggest that while ectopic Ato-Da expression is sufficient to activate the Dap-HB enhancer in some eye disc cells, additional factors are also likely involved in Ato-Da-mediated activation of Dap-HB.

Requirements of Da and Ato for Dap expression in the developing eye.

Da is a general-class bHLH protein that can form complexes with different cell type-restricted bHLH proteins, including Ato (25). To determine if Da regulates Dap in vivo, the effect of removing Da on the expression of Dap was studied by generating clones of da mutant cells in a da heterozygous background. This technique generates somatic clones of homozygous mutant cells from heterozygous precursors, resulting in adjacent populations of WT and mutant cells that can be easily identified (33). As shown in Fig. 2A to C, da mutant cells are identified by the absence of GFP, and the level of Dap protein was significantly decreased in da mutant clones in the eye disc (n = >10). These results showed that Da is required for the endogenous levels of Dap in the developing eye disc. To further determine if Da regulates Dap at the level of transcription, the effect of removing Da on the activity of the Dap eye disc enhancer Dap-HB was determined. As shown in Fig. 2G to I, Dap-HB enhancer activity was completely blocked in da mutant clones, indicating that Da is required for the transcriptional activation of Dap in the eye disc. Therefore, we conclude that the general-class bHLH protein Da is required for the expression of Dap in the developing eye.

FIG. 2.

Daughterless and Ato are required for endogenous Dap expression in the developing eye. (A to F) The level of Dap protein in da or ato mutant clones was determined by staining with anti-Dap antibody. Decreased Dap protein levels (red staining in panels B, C, E, and F) were observed in the da single-mutant clones (A to C), as well as in the ato single-mutant clones (D to F). These mutant clones are identified by lack of GFP. (G to L) Daughterless and Ato are required for Dap-HB enhancer activity. The Dap-HB-lacZ reporter was used in panels G to I. The da single-mutant clone was identified by lack of GFP. No Dap-HB-lacZ reporter expression (red in panels H and I) was observed in the da mutant clones. The Dap-HB-GFP reporter was used in panels J to L. ato mutant clones are identified by the lack of cytoplasmic β-galactosidase staining (red in panels J and L), and no Dap-HB-GFP reporter (green in panels K and L) was detected in the ato mutant clones.

To determine if the cell type-restricted bHLH protein Ato is also required for Dap expression in vivo, the effects of removing Ato on Dap protein levels and Dap-HB enhancer activity were determined. As ato mutant clones were quite small even in the Minute background, large numbers of eye discs with ato mutant clones were analyzed. ato mutant clones consistently showed a significant decrease in Dap protein levels (arrows in Fig. 2D to F, >10 discs with clones examined). As the Dap-HB-lacZ reporter used above is on the same chromosome on which Ato also resides, we used a Dap-HB-GFP reporter on the second chromosome to examine if removing Ato blocks Dap-HB enhancer activity. As shown in Fig. 2J to L, expression of the GFP reporter was not observed in ato mutant clones (marked by lack of cytoplasmic β-galactosidase staining in Fig. 2J to L, >10 discs with clones examined). These results indicate that both Ato and Da are required for Dap expression in the developing eye. As Ato is expressed in the R8 equivalence groups, which are the R2, R5, and R8 precursors, it is conceivable that Ato and Da can regulate Dap expression directly in these cells. Experiments to determine whether Ato and Da regulate Dap-HB directly by mutating the E-box binding sites are described below.

Da and Ato regulate the Dap-HB enhancer directly through the E-box binding sites.

Our observation that removing Da and Ato blocks Dap-HB enhancer activity in the developing eye can be explained either by a direct requirement of Ato and Da for Dap-HB enhancer activity or by an indirect requirement of Da and Ato for the initiation of R2 and R5 photoreceptor cell differentiation. To determine if Ato and Da regulate Dap-HB enhancer activity directly, the effects of mutating the observed E-box binding sites on Dap-HB enhancer activity were examined in the developing eye. As shown in Fig. 3A to C, mutation of E-box binding sites 1 and 2 completely blocked Dap-HB enhancer activity while mutation of E-box binding site 3 also significantly decreased Dap-HB enhancer activity. These results demonstrate that the conserved E-box binding sites are required for Dap-HB enhancer activity in the developing eye disc.

FIG. 3.

Dap-HB enhancer activity requires both the E-box binding sites and a Pnt binding site near E-box binding site 3. (A to E) Effect of mutating the E-box binding sites or the putative Pnt binding site on Dap-HB enhancer activity. The levels of GFP reporter protein from eye discs with WT or mutated Dap-HB enhancers are shown. mE1-2, mE3, mT1, and mT2 indicate Dap-HB enhancers with specific E-box or Pnt binding sites mutated as indicated in panel I. (F to H) Overexpression of Ato-Da in the posterior of the eye disc failed to induce the Dap-HB enhancer with mutated E-box binding sites 1 and 2. The level of Ato is shown by immunostaining (red in panels F and H), and the activity of the Dap-HB enhancer with mutated E-box sites is shown in green (G and H). (I) Diagram of different transgenic Dap-HB enhancers. Filled circles represent E-box binding sites, and filled squares indicate putative Pnt binding sites. The symbol × represents Dap-HB enhancers with mutations in the indicated binding sites.

Since Ato and Da are required for the initiation of photoreceptor cell differentiation, the observed regulation of Dap by ectopically activating Ato-Da shown in Fig. 1 could potentially be due to an indirect effect of expressing Ato-Da on eye development. To determine if the ability of Ato and Da to activate Dap-HB expression is mediated directly by their binding to the observed E-box binding sites, the effect of ectopic expression of Ato and Da on the activity of the Dap-HB enhancer with mutated E-box binding sites (Dap-HB-mut) was determined. As shown in Fig. 3F to H, expression of Ato and Da failed to induce the Dap-HB-mut1-2 enhancer, which has E-box binding sites 1 and 2 mutated, even though it did induce WT Dap-HB enhancer activity (Fig. 1J and K). Taken together, these results indicate that Ato-Da activates the Dap-HB enhancer directly through the E-box binding sites.

The Dap-HB enhancer contains a Pnt binding site adjacent to an E-box binding site.

EGFR signaling has been shown to be required for the cell cycle arrest and differentiation of R2 and R5 cells (34). Our finding that the Dap-HB enhancer is expressed specifically in R2 and R5 precursor cells prompted us to examine if the Dap-HB enhancer is also regulated by Pnt, a transcription factor regulated by EGFR signaling. Examination of the Dap-HB enhancer revealed two putative Pnt binding sites (Fig. 4A, T1 and T2). Gel shift assays were carried out to determine if either the T1 or T2 site can bind to Pnt-P1 protein in vitro. As shown in Fig. 4B, the WT but not mutant T2 site could compete for binding of the Pnt protein (lanes 4 to 6) while neither the WT nor the mutant T1 site could compete for binding of the Pnt protein. These observations suggest that only T2 is an actual Pnt binding site.

FIG. 4.

Ato-Da and PntP1 can bind to the Dap-HB enhancer simultaneously. (A) Diagram of Dap-HB enhancer and nucleotide sequences of the probe used. The sequences of the Pnt binding site (T2) and the E-box motif (E3) are shown in capital letters. (B) Gel shift experiments showing that Pnt protein binds specifically to the T2 site but not the T1 site. In vitro-translated Ato-Da and PntP1 bound to the probe with both T2 and E3 binding sites (lanes 1 and 4), and this binding was competed away specifically by WT E3 and T2 oligonucleotides (lanes 2 and 5) but not by mutated E3 and T2 oligonucleotides (lanes 3 and 6). In contrast, neither the WT nor the mutated T1 oligonucleotide was able to compete for PntP1 binding (lanes 7 to 9). (C) Gel shift assay showing that Ato-Da and Pnt can bind simultaneously to the probe with both T2 and E3 binding sites. A slower-migrating complex was observed when in vitro-translated Ato-Da and PntP1 were mixed in the gel shift assay (lanes 10 to 12). This slower-migrating complex was competed away specifically by an oligonucleotide containing either the T2 or the E3 binding site (lanes 12 to 15). Binding was significantly diminished by anti-Ato (lane 16) or anti-PntP1 (lane 18) antibodies. Con1 and Con2 indicate equal amounts of control sera for specificity of antibody binding to Ato and PntP1, respectively.

Interestingly, the Pnt binding site T2 and the E-box binding site E3 are only 4 bases apart (Fig. 4A), similar to the reported Ato recruitment enhancer (36), in which Ato, Da, and Pnt bind synergistically to adjacent sites. Furthermore, it was shown that Ato and Pnt physically interact in vitro (36). To determine if Pnt and Ato-Da can simultaneously bind to the adjacent Pnt and E-box binding sites in the Dap-HB enhancer, an oligonucleotide containing both the T2 and E3 binding sites of the Dap-HB enhancer was used as a probe in gel shift assays. As shown in Fig. 4C, a slower-migrating complex was observed when Ato and Da were added together with Pnt (lanes 10 to 12). This slower-migrating complex likely contains Ato-Da, as well as Pnt, since it was significantly diminished by an anti-Ato antibody and an anti-Pnt antibody (Fig. 4C, lanes 12 and 16 to 19) and was competed away by an oligonucleotide containing either the T2 or E3 site individually (Fig. 4C, lanes 13 and 14). In conclusion, these results indicate that the Dap-HB enhancer has a Pnt binding site that allows the simultaneous binding of Ato-Da and Pnt to adjacent sites. To determine if there is synergism between Ato-Da and Pnt in binding to adjacent sites in our experiments, we quantified each of the gel shift bands and found that the intensity ratios of Pnt/free probe, Ato-Da/free probe, and Pnt-Ato-Da/free probe were 23.4% ± 0.8%, 10.2% ± 0.2%, and 4.80% ± 0.29%, respectively. If Ato-Da and Pnt bind DNA completely independently, one would expect the Pnt-Ato-Da/free probe ratio to be around 23.4% ± 0.8% × 10.2% ± 0.2% = 2.38% ± 0.08%. Therefore, there is some level of synergistic binding observed in our gel shift experiments.

Dap-HB enhancer activity also requires Pnt.

Both Pnt-P1 and Pnt-P2 were shown to contribute to both the G₁ arrest and differentiation of the R2 and R5 photoreceptors (34). Previous results indicated that Dap protein levels are reduced but not eliminated in pnt mutant clones (13). As there are potentially multiple enhancers that regulate Dap expression in the developing eye and since Dap-HB is the one that drives expression specifically in the R2 and R5 precursors, we were interested in determining whether Pnt is required for Dap-HB enhancer activity. As shown in Fig. 5A to C, no GFP reporter from the Dap-HB enhancer was observed in pntP1 mutant clones (n, = >15). As a few of the R2 and R5 photoreceptors can still differentiate in pntP1 mutant clones (34), our observations suggest that Pnt proteins likely play a direct role in activation of the Dap-HB enhancer rather than affecting Dap-HB enhancer activity indirectly by blocking differentiation. In support of the idea that Pnt proteins are directly required for Dap-HB enhancer activity, mutation of the T2 Pnt binding site in the Dap-HB enhancer greatly reduced enhancer activity (Fig. 3E). In contrast, disruption of the T1 site, which does not bind Pnt in vitro, had no effect on Dap-HB enhancer activity (Fig. 3D).

FIG. 5.

Pointed is required for Dap-HB enhancer activity. (A to C) Pnt is required for Dap-HB enhancer activity. The Dap-HB-GFP reporter (green in panels B and C) was not observed in pntP1 mutant clones (marked by absence of β-galactosidase, shown in red in panels A and C). (D to K) Effects of expressing Ato, Da, PntP1, and Yan either alone or in different combinations with the GMR-Gal4 driver. Coexpression of Ato and Da led to synergistic activation of the β-galactosidase reporter driven by the Dap-HB enhancer (D to G). While expression of PntP1 together with Ato-Da had little additional effect (I), expression of Yan, an inhibitor of Pnt, blocks activation of the Dap-HB enhancer by Ato-Da (K).

To further determine the requirement of Pnt for Dap-HB enhancer activity, the effects of ectopic expression of Ato, Da, Pnt-P1, and the EGFR signaling antagonist Yan either alone or in different combinations were compared. As shown in Fig. 5D to K, expression of Da, Pnt-P1, or Yan alone had no significant effect on Dap-HB enhancer activity while ectopic expression of Ato showed very weak enhancement of Dap-HB enhancer activity. Interestingly, while strong synergistic activation of the Dap-HB enhancer was observed when Ato and Da were expressed together, no obvious additional expansion of Dap-HB expression was observed when Pnt-P1 was expressed together with Ato and Da (Fig. 5G and I). This observation suggests that the inability of Ato and Da to activate Dap-HB in the posterior end of the eye disc is not due to lack of Pnt proteins. In addition, it is possible that the Pnt protein is not limiting in cells where Ato and Da are capable of activating the Dap-HB enhancer, which would result in a similar activation of Dap-HB by Ato-Da and Ato-Da-Pnt. In support of the idea that Pnt is required for activation of Dap-HB by Ato and Da, expression of Yan blocked this activation (Fig. 5G and K). Taken together, these observations show that Pnt proteins are required in addition to Ato and Da for activation of the Dap-HB enhancer in the developing eye disc.

The Dap-HB enhancer is also active in the leg femoral chordotonal organ SOPs.

The above results indicate that Dap expression in R2 and R5 precursors is regulated by the differentiation factors that specify the fate of R2 and R5. In addition to R2 and R5 determination, EGFR signaling and Ato were also shown to be involved in leg femoral SOP specification (36). This prompted us to examine if Dap-HB is also expressed in the femoral SOPs of the developing leg disc during SOP specification. As shown in Fig. 6A, the GFP reporter was expressed in a small cluster of cells in the leg disc. As the leg femoral SOPs can be marked by Senseless (26, 36), our observation that the GFP reporter driven by the Dap-HB enhancer coincided with cells that are marked by Senseless (Fig. 6A to C) indicates that Dap-HB is expressed in a subset of femoral SOPs. In addition, GFP expressed from the Dap-HB enhancer also overlaps Ato (Fig. 6D to F), which is expressed in the overlying proneural cluster and in the newly recruited SOPs but not in mature SOPs (35). We conclude that Dap-HB is an enhancer element that drives Dap expression in newly recruited femoral SOPs but not in mature SOPs. Staining with anti-Pnt-P1 antibody showed that the GFP reporter also overlaps a subset of the Pnt-P1-expressing SOPs (Fig. 6G and H), consistent with the idea that the Dap-HB enhancer expression in leg femoral SOPs is coordinately regulated with SOP recruitment by Pnt and Ato-Da transcription factors. To examine further the requirement of Ato-Da and Pnt for Dap-HB expression in the leg SOPs, GFP reporter activity of Dap-HB enhancers harboring different mutations was studied. Mutation of either the E-box binding site E3 or the adjacent Pnt binding site T2 significantly diminished Dap-HB enhancer activity in the leg disc femoral SOPs (Fig. 6I to K). In addition, mutations in the first two E-box binding sites completely abolished Dap-HB enhancer activity in these cells (Fig. 6L). Therefore, the Dap-HB enhancer requires both the E-box binding sites and the Pnt binding site for full enhancer activity in leg disc femoral SOPs.

FIG. 6.

Dap-HB is expressed in leg disc femoral SOPs during their recruitment. (A to H) Expression of Dap-HB-GFP (green) in leg disc femoral SOPs relative to that of Senseless (red in panels B and C), Ato (red in panels E and F), and PntP1 (red in panels G and H) is shown. Arrows point to SOPs with Dap-HB-GFP expression, and arrowheads in panels E and F point to Ato staining in femoral proneural cluster cells (out of this focal plane). (I to L) Dap-HB enhancer activity in femoral SOPs requires both the E-box binding sites and the Pnt binding site. Mutation of either the E3 binding site (J) or the T2 binding site (K) significantly decreased Dap-HB enhancer activity in leg disc femoral SOPs. Mutations of both E1 and E2 binding sites (L) abolished Dap-HB enhancer activity. (M to R) Dap-HB-GFP expression in leg discs corresponded to femoral SOPs that have high levels of Dap protein (M to O) and are cell cycle arrested, as shown by a lack of BrdU incorporation of Dap-HB-GFP-expressing SOPs (arrows in panels P to R). Arrowheads in panels P to R showed intense BrdU incorporation by some of the mature SOPs.

To determine if Dap-HB reporter expression in the leg disc corresponds to endogenous Dap protein, Dap levels were visualized by antibody staining. As shown in Fig. 6M to O, GFP reporter expression corresponds to femoral SOP cells with a high level of Dap protein, and the level of Dap protein decreases in the more mature SOPs. To determine if Dap-HB expression corresponds to cell cycle arrest, BrdU incorporation in leg discs was carried out. As shown in Fig. 6P to R, cells with GFP expression do not incorporate BrdU, consistent with the idea that Dap is transiently induced during SOP recruitment to ensure cell cycle arrest. Interestingly, some very strong BrdU incorporation is observed next to the GFP-expressing cells. Double labeling showed that these very strong BrdU-incorporating cells are also labeled with Senseless (data not shown), suggesting that these are mature SOP cells that have reentered the cell cycle.

DISCUSSION

Our analysis of the Dap-HB enhancer showed that Dap expression is regulated by cell-intrinsic transcription factors (the bHLH proteins Ato and Da) and the signaling-activated transcription factor Pnt in R2 and R5 photoreceptor precursors and leg disc femoral SOPs. Interestingly, these same transcription factors that control Dap-HB activity are also known to control the differentiation of cells in which Dap-HB is active. In the case of leg disc femoral SOP recruitment, Ato was previously shown to function as the competence factor and EGFR signaling as the trigger for femoral SOP recruitment (35, 36). Our results showed that the Ato and Pnt binding sites are required for the transient induction of Dap in SOPs, suggesting that Ato and Pnt coordinately regulate expression of Dap and other genes involved in SOP recruitment (such as Ato autoregulation). In the case of R2 and R5 photoreceptor cell specification, Ato is detected in the two- to three-cell R8 equivalence groups with apically localized nuclei before R8 specification (9). It is possible that once the R8 photoreceptor is specified, the other cells in the R8 equivalence groups will become the R2 and R5 photoreceptors. This is consistent with the observation that ato¹ mutant cells have impaired differentiation of R2 and R5 photoreceptors (18). In addition, diphosphorylated Erk staining, which indicates mitogen-activated protein kinase activation, is detected in the Ato-expressing proneural clusters (6, 21, 30), suggesting that the Ato-expressing R8 equivalence group already has some level of EGFR signaling. Consistent with this, Pnt is known to be required for the differentiation of R2 and R5 photoreceptors and our results showed that Pnt was also required for the induction of Dap-HB enhancer activity in these cells. It is possible that Pnt and Ato-Da proteins activate the Dap-HB enhancer immediately following the R8 equivalence group stage when there is enough Ato remaining in the R2 and R5 precursors to induce activation of the Dap-HB enhancer in conjunction with Pnt. The lack of Dap-HB expression in R8 cells could be due to the presence of a repressor or the lack of another cofactor that is also required for activation of the Dap-HB enhancer in this cell type.

It is tempting to speculate that the coordinated control of cell differentiation and Dap expression we observed is a general mechanism by which Dap transcription is regulated in different cell types. If this is the case, one might expect Dap to be regulated by multiple distinct regulatory elements corresponding to different cell types even in the same developing tissue. In support of this, Dap-HB was found to be expressed only in R2 and R5 precursors in the developing eye disc and there are likely additional enhancers that drive Dap expression in other cell types. Similarly, dissection of Dap expression in the embryonic epidermis also showed that multiple enhancers were used in the control of Dap expression in different regions of the epidermis (24). While previous studies showed that Dap is regulated by many cis-acting elements and suggested that Dap is under developmental control (23, 24), no clear mechanism was derived from those studies. The present work provides the first in-depth analysis of the expression of Dap in two well-defined models; we show that, in both cases, the transcription of Dap is regulated by both Ato-Da and EGFR signaling, which are also the factors involved in controlling the differentiation of the respective cell types.

It should be pointed out that our results do not exclude the possibility that Ato might activate another bHLH protein that can dimerize with Da and activate Dap-HB expression in conjunction with Pnt in the R2 and R5 precursors. However, given the specific physical interaction between Ato and Pnt (36) and the similar organization of the E-box and Pnt binding sites between the Ato autoregulation enhancer and the Dap-HB enhancer, we favor the idea that Ato-Da activates Dap directly, at least through the E-box 3 site.

Importantly, the ability of bHLH proteins to coordinately induce cell differentiation and cell cycle inhibitor expression appears to be an evolutionarily conserved mechanism. In this study, we showed that the bHLH proteins Ato and Da coordinate Dap expression with photoreceptor cell differentiation and SOP recruitment. Previous studies of myocyte differentiation in mammalian tissue culture systems have shown that the myogenic bHLH proteins coordinately regulate muscle cell differentiation and induction of the cdk inhibitor p21 (14, 15, 28). As large numbers of bHLH proteins have been implicated in the cell fate determination of diverse cell types in both mammals and flies (31), it is possible that these bHLH proteins may also control cell cycle inhibitor expression through a similar mechanism and that the inhibitors of the bHLH proteins such as the Id (inhibitor of differentiation) family of proteins (29) may control cell proliferation, at least in part, by antagonizing the induction of the Cip/Kip family of cdk inhibitors by the bHLH proteins.

While Dap was found to be expressed transiently just before cell cycle exit in diverse cell types (8, 22), analysis of the effect of removing Dap showed that Dap is not always essential for the G₁ arrest of those cells. In the developing photoreceptor preclusters, for example, it was shown RBF and Dap function redundantly for the G₁ arrest of the precluster cells (13). As the G₁ arrest of these cells during cell fate specification is essential for normal development, multiple mechanisms have likely evolved to ensure the cell cycle arrest of these cells before differentiation. Therefore, the induction of Dap during cell fate specification potentially provides a fail-safe mechanism to ensure the G₁ arrest of the differentiating cells.