Atonal (Ato)/Math (Mammalian atonal homolog) family proneural proteins are key regulators of neurogenesis in both vertebrates and invertebrates. In the Drosophila eye, Ato is essential for the generation of photoreceptor neurons. Ato expression is initiated at the anterior ridge of the morphogenetic furrow but is repressed in the retinal precursor cells behind the furrow to prevent ectopic neurogenesis. We show that Ato repression is mediated by the conserved homeobox proteins BarH1 and BarH2. Loss of Bar causes cell-autonomous ectopic Ato expression, resulting in excess photoreceptor clusters. The initial ommatidial spacing at the furrow occurs normally in the absence of Bar, suggesting that the ectopic neurogenesis within Bar mutant clones is not due to the lack of Notch(N)-dependent lateral inhibition. Targeted misexpression of Bar is sufficient to repress ato expression. Furthermore, we provide evidence that Bar represses ato expression at the level of transcription without affecting the expression of an ato activator, Cubitus interruptus(Ci). Thus, we propose that Bar is essential for transcriptional repression of ato and the prevention of ectopic neurogenesis behind the furrow.

The generation of neurons is one of the most fundamental events in development of all nervous systems. Neurogenesis in Drosophila is regulated by two major classes of genes: proneural and neurogenic. Proneural genes such as ato and achaete-scute complex (ASC)encode transcription factors of the basic helix-loop-helix (bHLH) family and are expressed in small subsets of populations of uncommitted cells. Only one cell in each equally competent group of proneural cells is selected to become the founder neuron or neuroblast (Brunet and Ghysen, 1999). ato is an essential proneural gene for the formation of photoreceptor neurons and chordotonal organs(Jarman et al., 1994; Jarman et al., 1995), whereas the ASC is important for the generation of sensory organ precursors(SOPs) that give rise to sensory bristles(Ghysen and Dambly-Chaudiere,1988; Romani et al.,1989; Cubas et al.,1991; Skeath and Carroll,1991). Neurogenic genes, however, are involved in N-mediated lateral inhibition, which prevents other neighboring cells from adopting a neuronal fate (Dietrich and Campos-Ortega, 1984; Schweisguth et al., 1996; Artavanis-Tsakonas et al.,1999).

An important question is how proneural gene expression is regulated to either promote or inhibit neurogenesis. The Drosophila eye is an excellent genetic model with which to address this question as the waves of repetitive neurogenesis can be easily identified at the cellular level and analyzed using powerful genetic tools. The adult compound eyes are composed of∼800 unit eyes, or ommatidia, each of which contains eight photoreceptor(R1-R8) neurons and 12 accessory cells(Wolff and Ready, 1993). Retinal differentiation begins as the morphogenetic furrow moves anteriorly from the posterior margin of the eye imaginal disc during larval development(Wolff and Ready, 1993). Posterior to the furrow, R8 founder cells are specified first, and then they recruit other photoreceptors in the sequential order of R2/5, R3/4 and finally R1/6/7 (Wolff and Ready,1993). Therefore, the selection of the founder R8 cell in each ommatidium is a critical step in retinal neurogenesis.

Neurogenesis of R8 founder cells from the undifferentiated retinal epithelium of the eye disc requires the function of the proneural gene ato (Jarman et al.,1994). The dynamic pattern of ato expression can be divided into four steps (Frankfort and Mardon, 2002) (see Fig. 1C). During stage 1, Ato is expressed as a stripe pattern across the disc in the most anterior region of the furrow. Next, just posterior to the stripe, Ato is expressed in columns of repetitive groups of about 20 cells called intermediate groups (stage 2). At stage 3, Ato is expressed in subsets of two or three cells called R8 equivalence groups. In the final step (stage 4), Ato expression becomes restricted to a single R8 founder cell. The sequential restriction of Ato-expressing cells from the initial stripe of cells (stage 1) to the founder R8 neuron (stage 4) is strictly regulated by the interaction of positive and negative signaling(Dokucu et al., 1996; Baker and Yu, 1998; Frankfort and Mardon, 2002). At stage 1, Hedgehog (Hh) and N signaling act positively on the expression of Ato. During stage 2, Scabrous (Sca) are important for the formation of regularly spaced intermediate groups. During stages 3 and 4, N-mediated lateral inhibition, Rough (Ro) and Senseless (Sens) are involved in the selection of single R8 founder cells.

The repression of Ato activity is required for specification of the proper number of photoreceptor neurons. The mechanisms by which the R8 founder cells are generated from the initial stripe of Ato-expressing cells are well understood. However, the mechanism by which the initial expression domain of Ato, which is regulated through 3′-regulatory elements of ato(Sun et al., 1998), is abruptly repressed behind the furrow has not been well characterized. It is possible that strong downregulation of Ato expression may depend on a timely expression of a specific repressor(s) behind the furrow of the eye disc. Such a repressor is likely to be expressed posterior to the furrow in a complementary pattern to the Ato expression, and its function should be essential for proper control of neurogenesis in the eye.

One of the candidate genes involved in ato repression is Bar, which encodes two functionally redundant homeodomain proteins,BarH1 and BarH2. Bar proteins are expressed in undifferentiated retinal precursor cells behind the furrow in the developing eye disc(Higashijima et al., 1992),showing a complementary expression pattern to Ato (see Fig. 1D,E). They are also expressed in two specific photoreceptors, R1 and R6, and other accessory cells(see Fig. 1F). Studies on Bar function in the eye have mostly been focused on its roles in the differentiation of R1/R6 and pigment cells(Higashijima et al., 1992; Hayashi et al., 1998), but little is known about its role in the population of undifferentiated cells. Interestingly, there have been hints that Bar may play a role in photoreceptor patterning. For example, ubiquitous Bar overexpression by heat-shock (hs)-Bar or a duplication of BarH1 gene in Bar1 mutant leads to downregulation of Hh,Decapentaplegic (Dpp) and/or Ato expression in the eye disc, resulting in the furrow-stop phenotype and reduced eyes(Heberlein et al., 1993; Epps et al., 1997; Hayashi and Saigo, 2001). However, it has not been determined whether the Bar gene is directly related to ato repression.

To understand the precise function of Bar in retinal neurogenesis,we analyzed the regulatory relationships between Bar and ato. We show that Bar is required to repress Ato expression behind the furrow. Ectopic induction of Ato in the absence of Bar is sufficient to form ectopic photoreceptor clusters. Furthermore, ato-repression is mediated at the transcriptional level through the regulatory elements of ato. This function of Bar on ato-repression is independent of Ci, a mediator of Hh signaling(Ingham, 1998) known to activate ato expression. Therefore, Bar expression is crucial for the control of ato and neurogenesis in the eye. The Bar class homeobox genes are evolutionarily conserved(Jones et al., 1997; Saito et al., 1998; Bulfone et al., 2000; Patterson et al., 2000). Recent studies suggest that mammalian Bar class genes may function as regulators for the expression of bHLH proneural genes during neurogenesis(Saito et al., 1998). Hence,our study may contribute to understand the mechanism of Bar homolog function in vertebrate neurogenesis.

Drosophila stocks

The following mutant and transgenic flies were used in this study: Df(1)B263-20(Higashijima et al., 1992), UAS-BarH1M13, UAS-BarH2F11(Sato et al., 1999), 3F:5.8, 5F:7.2, 5F:9.3 ato-lacZ (Sun et al.,1998) and lozenge (lz)-Gal4(Crew et al., 1997). Other strains were obtained from the Bloomington Stock Center.

Generation of loss-of-function mosaic clones and misexpression studies

Bar loss-of-function clones were generated using Df(1)B263-20 with the FLP/FRT system(Xu and Rubin, 1993). First instar larvae from the cross between yw Df(1)B263-20FRT18A/FM7 females and w Ubi-GFPS65T FRT18A; hs-FLP3males were treated for 1 hour at 37°C and incubated at room temperature until dissection. For misexpression of Bar, progeny from the cross between dpp-Gal4/TM6B males and UAS-BarH1M13 (or UAS-BarH2F11) females were cultured at 25°C until dissection at the third instar larval stage. lz-Gal4 was used to express Bar (or CiFL) in the basal cells.

Immunocytochemistry

Third instar eye imaginal discs were dissected in phosphate-buffered saline(PBS) on ice, fixed in 2% paraformaldehyde-lysine-periodate fixative and stained as described (Carroll and Whyte,1989). The following primary antibodies were used in this study:mouse anti-Arm (1:200; Developmental Studies Hybridoma Bank [DSHB]), mouse anti-Boss (1:2000; DSHB), mouse anti-Sca (1:200; DSHB), mouse anti-β-gal(1:250; Promega), mouse anti-dpERK (1:250; Sigma), mouse anti-Elav (1:10;DSHB), rabbit anti-Ato (1:5000) (Jarman et al., 1995), rabbit anti-BarH1 (1:20)(Higashijima et al., 1992),rabbit anti-GFP (1:2000; Molecular Probes), rat anti-CiFL (1:10)(Motzny and Holmgren, 1995),guinea pig anti-Ato (1:1000) (Hassan et al., 2000), guinea pig anti-Dlg (1:1000; provided by P. Bryant)and guinea pig anti-Sens (1:1000) (Nolo et al., 2000). Secondary antibodies were anti-mouse-CY3,anti-mouse-fluorescein isothocyanate (FITC), anti-rabbit-CY3, anti-rabbit-FITC and anti-guinea pig-CY5 (Jackson Immunochemicals). Fluorescent images were scanned using Zeiss LSM laser-scanning confocal microscope and processed with Adobe Photoshop.

Bar expression in basal undifferentiated cells is complementary to Ato expression

Each ommatidium of the adult compound eye consists of eight photoreceptors(Fig. 1A) that are generated by the proneural function of Ato expressed within and anterior to the furrow in the eye disc (Fig. 1B,C)(Jarman et al., 1994). The domain of Ato expression is juxtaposed to the Bar-expressing undifferentiated cells behind the furrow (Fig. 1D,E). Although Bar is also expressed in R1 and R6 photoreceptors(Fig. 1F)(Higashijima et al., 1992),this study focuses specifically on the Bar expression in the undifferentiated cells and the Ato expression in adjacent anterior cells. These Bar-expressing undifferentiated cells will be referred as the `basal cells' as their nuclei stay in the basal region while photoreceptor cell nuclei migrate apically,although cell bodies of both cell types are connected to the top and bottom of the eye disc epithelium (Fig. 1F). Nuclei of Ato-expressing cells are located basally during the stages 1 and 2, but migrate apically as they become R8 founder neurons posterior to the furrow (Fig. 1C,D).

Bar is necessary and sufficient to repress Ato expression

The complementary pattern of Bar and Ato expression raises the possibility that Bar may be involved in inhibition of Ato expression in the basal cells behind the furrow. To test this possibility, we generated loss-of-function clones of Bar using a Bar deficiency, Df(1)B263-20, with the FLP/FRT system(Xu and Rubin, 1993). Df(1)B263-20 uncovers BarH1 and BarH2,along with forked (f), Fimbrin (Fim), a broken 297 retrotransposon, and an uncharacterized transcript X2(Higashijima et al., 1992; Sato et al., 1999). It has been shown that defects in wing and leg discs associated with Df(1)B263-20 can be fully rescued by overexpression of either BarH1 or BarH2 (Sato et al.,1999; Kojima et al.,2000). To test further whether Df(1)B263-20can be used as a Bar mutant in the eye disc, we generated Df(1)B263-20 loss-of-function clones with a background of overexpression of BarH1 (or BarH2) gene by lozenge(lz)-Gal4, which drives Gal4 expression in the basal cells and in the R1, R6 and R7 photoreceptor cells behind the furrow(Flores et al., 1998). The expression of either BarH1 or BarH2 under the control of lz-Gal4 in the basal cells strongly rescued the eye phenotypes found in the Df(1)B263-20 loss-of-function clones (data not shown). Therefore, Df(1)B263-20 was used as a null mutant of two Bar genes in the eye for the following studies.

We then examined the expression of Ato in the mutant cells of Barloss-of-function clones (Fig. 2A-C). The expression level of Ato was strongly elevated in the basal cell layer and the photoreceptor cell layer within Barloss-of-function clones behind the furrow(Fig. 2C; data not shown). In the basal layer, all undifferentiated cells showed Ato upregulation(Fig. 2C, white arrow). Additionally, many individually singled-out Ato+ cells showed even stronger Ato expression in the basal and apical layers than other basal cells(Fig. 2C, yellow arrow).

To further confirm that Bar functions as a negative regulator of Ato expression, we misexpressed Bar with a dpp-Gal4 driver using the Gal4/UAS system (Brand and Perrimon, 1993). Dpp is normally expressed in the furrow of the eye disc and in the ventral sector of the antennal disc(Blackman et al., 1991). The dpp-Gal4 enhancer trap precisely reflects the dpp expression pattern in the antennal disc but not in the eye disc(Fig. 2D). In the eye disc, it shows only the early pattern of dpp expression in the posterior lateral margin, but is not expressed in the furrow(Chanut and Heberlein, 1997). Misexpression of BarH1 or BarH2 using dpp-Gal4(dpp-Gal4>BarH1 or dpp-Gal4>BarH2) strongly blocked furrow migration and eliminated photoreceptor differentiation in the ventral margin of the eye disc (Fig. 2E, arrowhead; data not shown). By contrast, Ato expression and photoreceptor differentiation were normal in the dorsal margin of the dpp-Gal4>BarH1 (or dpp-Gal4>BarH2) eye disc. Examination of Bar expression by dpp-Gal4>BarH1 revealed that BarH1 was ectopically expressed in the ventral margin but not in the dorsal margin for an unknown reason. Furthermore, ectopic Bar expression dramatically downregulated Ato expression in the ventral sector of the antennal disc where dpp-Gal4 drives Bar expression(Fig. 2E,F, arrows). Taken together, both loss-of-function and misexpression studies suggest that Bar is not only required but also sufficient to downregulate Ato expression.

Ectopic Ato+ cells are sufficient to recruit other photoreceptors

As shown in Fig. 2C, the elevated expression of Ato within Bar mutant cells could eventually resolve into single Ato+ cells, as individual R8 cells are selected from equivalence groups of Ato+ cells near the furrow. Interestingly, similar ectopic Ato expression has been observed in loss-of-function clones of Drosophila homologs of the transcriptional coactivator complex subunits, blind spot (bli; TRAP240) and kohtalo (kto; TRAP230)(Treisman, 2001). However,singled-out Ato+ cells within TRAP loss-of-function clones could not further differentiate into mature R8 cells, thus resulting in failure to form photoreceptor clusters(Treisman, 2001).

To examine whether ectopic singled-out Ato+ cells within Bar loss-of-function clones differentiate into mature R8 photoreceptor cells, we checked the process of R8 cell differentiation using early or late R8 cell-specific markers(Fig. 3). Sca is an early R8 cell marker normally expressed in the intermediate groups and the R8 cells within a few columns posterior to the furrow(Fig. 3C, bracket)(Mlodzik et al., 1990; Lee et al., 1996). Sca was ectopically expressed within Bar loss-of-function clones behind the furrow (Fig. 3A-C). We also checked the expression of Sens, which is expressed in the R8 equivalence groups and all singled-out R8 photoreceptor cells within and behind the furrow(Nolo et al., 2000; Frankfort et al., 2001). The number of cells expressing Sens was increased within Barloss-of-function clones (Fig. 3D-F). Finally, we checked the expression of bride-of-sevenless(Boss), a late differentiation marker for R8 cells(Fig. 3G-I). The number of Boss+ cells was increased within Bar loss-of-function clones (Fig. 3I). Taken together, increased numbers of Sca+, Sens+ and Boss+ cells within Bar loss-of-function clones suggest that ectopic singled-out Ato+ cells are able to differentiate into mature R8 photoreceptor cells.

As R8 cells are essential for the recruitment of other photoreceptors, we asked whether the ectopic R8 cells generated by loss of Bar could form ectopic photoreceptor clusters. We examined the formation of photoreceptor clusters using two different markers, Elav and Armadillo (Arm)(Fig. 3J-O). Elav is a neuron-specific marker expressed in the nuclei of all photoreceptor cells, and Arm is localized in the adherens junctions of each photoreceptor cell(Izaddoost et al., 2002; Pellikka et al., 2002). We observed that there are excess photoreceptor clusters stained by anti-Elav and anti-Arm antibodies within Bar mutant clones(Fig. 3L,O). These observations indicate that ectopic Ato+ cells within Barloss-of-function clones can differentiate into mature R8 cells(Fig. 3A-I), recruit other photoreceptor cells, and finally form ectopic photoreceptor clusters(Fig. 3J-O). To check whether the cells recruited into photoreceptor clusters within Barloss-of-function clones are correctly specified, we examined the expression of different photoreceptor-specific markers: Rough (Ro) for R2/5 cells,mδ0.5-lacZ for R4 cells, Lz for R1/6/7 cells and Prospero (pros) for R7 cells. We observed Ro+, mδ0.5-lacZ+,Lz+ and Pros+ cells within Bar loss-of-function clones (data not shown), suggesting that photoreceptor cells recruited into ommatidial clusters within Bar loss-of-function clones correctly differentiate into R1 to R8 cells. Therefore, unlike TRAP, Bar is specifically required for Ato repression but not for the photoreceptor recruitment.

Bar is independent of lateral inhibition

The fact that Bar mutant cells with the elevated Ato expression could eventually resolve into single Ato+ R8 cells(Fig. 2C) suggests that N-mediated lateral inhibition may function within Barloss-of-function clones behind the furrow. In addition to lateral inhibition,Sca and Epidermal Growth Factor Receptor (Egfr) seem to be essential for interommatidial spacing between the intermediate groups of Ato+cells in the furrow (Baker et al.,1990; Baker and Yu,1997; Chen and Chien,1999; Baonza et al.,2001; Frankfort and Mardon,2002). The expression of Sens in the R8 equivalence groups within Bar loss-of-function clones showed regularly spaced staining(Fig. 4A-C), suggesting that loss of Bar does not affect the initial spacing of intermediate groups. Thus, Sca, Egfr and N signaling pathways involved in ommatidial spacing may function normally near the furrow within Barloss-of-function clones.

To test this possibility, we examined the activity of Sca, Egfr and N signaling pathways within Bar loss-of-function clones using different markers, such as Sca, dpERK (dual phosphorylated extracellular signal regulated kinase), Delta (Dl), and Enhancer of split [E(spl)]. As mentioned earlier, Sca is expressed in subsets of Ato-expressing cells of the intermediate groups and the R8 cells. The expression of Sca in the intermediate groups within Bar loss-of-function clones was normal with regular spacing (data not shown), although ectopic Sca+ cells were generated in Bar mutant regions behind the furrow(Fig. 3A-C). The Egfr-mediated MAP kinase signaling pathway is also involved in ommatidial spacing(Chen and Chien, 1999; Baonza et al., 2001). We also found that the staining pattern of dpERK, the activated MAP kinase of the Ras signaling cascade, was normal near the furrow within Bar mutant cells(Fig. 4D-F). This suggests that Egfr signaling pathways function near the furrow to produce regularly spaced intermediated groups within Bar loss-of-function clones. However,dpERK expression in one cell per cluster in 6-8th columns posterior to the furrow was strongly reduced (Fig. 4F). This high-level dpERK has been shown to be expressed in the R7 cells (Kumar et al., 1998). As R1/R6 cells contribute to R7 cell specification via N signaling(Tomlinson and Struhl, 2001),the mis-specification of R1/R6 cells in Bar loss-of-function clones might contribute to the loss of dpERK expression in R7 cells. Finally, we examined the expression of Dl and E(spl) near the furrow to test if the N signaling pathway functions within Bar loss-of-function clones to select initial intermediate groups and R8 cells. Both Dl and E(spl) were normally expressed within Bar loss-of-function clones and showed no significant difference from the adjacent wild-type tissues (data not shown). By contrast, Sca, dpERK, Dl and E(spl) proteins were ectopically expressed within Bar mutant cells away from the furrow(Fig. 3A-C, Fig. 4D-F; data not shown). Thus, Bar mutant cells with elevated Ato expression away from the furrow can eventually resolve into single Ato+ cells(Fig. 2C).

Bar is required for transcriptional repression of ato

As Bar is a DNA-binding homeodomain transcription factor, ectopic induction of Ato in Bar loss-of-function clones suggests that Bar is required to repress ato expression at the transcription level. However, it is equally possible that Bar may be involved in destabilizing Ato protein rather than repressing ato transcription. To test whether Bar is necessary for transcriptional repression of ato, we used ato-lacZreporters in which lacZ is under the control of two regulatory regions of ato, 3F:5.8 or 5F:9.3 (Sun et al., 1998). The 5.8 kb of ato 3′ sequence(3F:5.8) specifies the initial stripe of atoexpression and is Ato-independent (stage 1; Fig. 1C, Fig. 5A). By contrast, the 9.3 kb of ato 5′ sequence (5F:9.3) is responsible for ato expression in the equivalence groups and the R8 founder cells, but is auto-regulated by Ato itself (stages 2-4; Fig. 1C, Fig. 5E,I).

First, we examined β-galactosidase (β-gal) activity of 3F:5.8 ato-lacZ within Bar mutant clones(Fig. 5B-D). Indeed, we could see strongly increased β-gal expression within Barloss-of-function clones comparing with surrounding wild-type regions(Fig. 5D, arrows), suggesting that Bar represses ato expression at the level of transcription through 3′-regulatory region of ato.

Next, we examined the β-gal activity of 5F:9.3 ato-lacZ within Bar mutant clones to test whether Bar may also repress ato transcription through 5′-regulatory region of ato (Fig. 5F-H). In the wild-type eye disc, antibody staining of β-gal in the 5:9.3 ato-lacZ eye disc shows precise co-localization with Sens in the R8 equivalence groups and the singled-out R8 cells behind the furrow (Fig. 5E). The expression level of 5F:9.3 ato-lacZ was dramatically elevated within Bar loss-of-function clones(Fig. 5H), suggesting that Bar might also repress ato transcription through 5′-regulatory elements of ato. However, as 5F:9.3 is dependent on Ato for driving its expression(Sun et al., 1998), it is possible that increased β-gal expression of 5F:9.3 ato-lacZ within Bar mutant cells might be caused by increased endogenous expression of Ato protein, which can activate 5F:9.3 enhancer, rather than by direct transcriptional derepression through 5F:9.3.

To test these two possibilities, we ectopically expressed BarH1 (or BarH2)in the antennal disc using the dpp-Gal4 driver(Fig. 5I-L). The expression of ato in the antennal disc is regulated by 2.1 kb enhancer regions of 7.2 kb upstream to the ato-coding region and is Ato independent(Sun et al., 1998). The 5F:9.3 and 5F:7.2 ato-lacZlines include the antenna-specific enhancer regions and thus expressβ-gal even in the ato1 mutant discs(Sun et al., 1998). The ato1 encodes a non-functional Ato protein because of a mutation in the DNA-binding domain (Jarman et al., 1994). Misexpression of BarH1 or BarH2 showed the strong repression of β-gal expression of 5F:9.3 or 5F:7.2 ato-lacZ in the ventral sector of the antennal disc both in the wild-type and the ato1 mutant backgrounds(Fig. 5J-L, arrows; data not shown). These observations suggest that Bar represses atotranscription directly through the 5′-regulatory region of ato,probably as well as by autoregulation. Consistent with this interpretation,some Bar mutant cells were lacZ positive but Sens negative(Fig. 5M-P, arrows), indicating that increase in number of lacZ+ cells within Barmutant clones might be due to the direct transcriptional derepression by the loss of Bar. Taken together, these data suggest that Bar represses ato expression at the transcription level through 3′- and 5′-regulatory regions of ato.

Bar repression of ato is CiFL-independent

After furrow is initiated, Hh is secreted anteriorly from differentiating photoreceptors to activate the expression of target genes within the furrow(Dominguez and Hafen, 1997; Strutt and Mlodzik, 1997; Borod and Heberlein, 1998; Dominguez, 1999; Greenwood and Struhl, 1999). In the absence of Hh signal, Ci exists as an inactive cleaved form, whereas in the presence of Hh, the full-length active form of Ci (CiFL) is produced (Chen et al., 1999; Methot and Basler, 1999; Price and Kalderon, 1999). CiFL is strongly expressed anterior to the furrow but expressed at a low level posterior to the furrow (Fig. 6A,B) (Strutt and Mlodzik,1997). Ato is one of the targets of CiFL in the stripe of the furrow (Fig. 6A,C)(Dominguez, 1999), showing an overlapping expression with CiFL(Fig. 6A,D).

The expression domain of CiFL was complementary to the Bar expression domain along the furrow (Fig. 6E-H). This raised the possibility that Bar may inhibit CiFL expression behind the furrow and thereby indirectly repress ato expression. To test this, we examined the expression of CiFL within Bar mutant cells(Fig. 6I-L). We found that expression of CiFL was not elevated(Fig. 6J) but that of Ato was upregulated (Fig. 6K) in Bar loss-of-function clones. Bar loss of function clones located within and anterior to the furrow also showed no changes in CiFL expression (data not shown). This result strongly suggests that ato-repression by Bar is independent of CiFL.

We have shown that Bar is essential for the repression of proneural gene ato expression in the basal cells behind the furrow. These basal cells have been considered as the source for retinal cells with the little known function in morphogenesis. Our findings reveal an important novel function of Bar and the basal cells in early neuronal patterning.

Bar represses ato transcription through the regulatory regions of ato

Ato expression is highly elevated in the absence of Bar behind the furrow(Fig. 2C), suggesting that Bar is necessary for downregulation of Ato expression. Furthermore, we showed that Bar repressed ato expression at the transcriptional level through both 3′- and 5′-regulatory regions of ato(Fig. 5). The 9.3 kb of ato 5′ sequence (5F:9.3) has been shown to be responsible for ato expression in the equivalence groups and the R8 founder cells in Ato-dependent manner (stages 2-4; Fig. 1C, Fig. 5E)(Sun et al., 1998). Sca and Egfr-mediated MAP kinase signaling may inhibit this enhancer function of 5′-regulatory element of ato within interommatidial regions to establish regularly spaced intermediate groups(Chen and Chien, 1999). By contrast, the 5.8 kb of ato 3′ enhancer(3F:5.8) is only activated anterior to the furrow to drive the initial stripe of ato expression (stage 1; Fig. 1C, Fig. 5A)(Sun et al., 1998). How this enhancer activity is inhibited posterior to the stage 1 Ato domain of the eye disc is unknown. Our results now indicate that the initial stripe (stage 1) of ato expression driven by 3′-regulatory element is strongly inhibited by Bar behind the furrow.

Ectopically elevated Ato expression within Bar loss-of-function clones was sufficient to induce the formation of mature ectopic photoreceptor clusters (Fig. 3). This suggests that Bar mutations specifically eliminate the repression of initial ato expression with little effects on subsequent steps of photoreceptor recruitment. This is consistent with the observations that Sca,Egfr signaling and N-mediated lateral inhibitions function properly within Bar loss-of-function clones (Fig. 4). Therefore, the major role of Bar during retinal neurogenesis appears to be the inhibition of initial stripe ato expression through 3′-regulatory elements of ato behind the furrow.

Bar is a DNA-binding homeodomain transcription factor. Mammalian homolog Barx2 was shown to bind directly to regulatory elements of several neural cell adhesion molecules, which contains target sites including the core sequence CCATTAGPyGA (Jones et al.,1997). Interestingly, the 5′F9.3 and 3′F5.8 regulatory regions of ato also have multiple potential Bar binding sties containing the same core sequence (data not shown), suggesting that Bar may directly bind to these target sites of ato regulatory elements and repress ato transcription.

It is important to note that CiFL induced by Hh signaling can activate ato expression(Dominguez, 1999). Furthermore, Bar and CiFL are expressed complementarily each other(Fig. 6E-H). These observations raise the possibility that ato repression by Bar may be mediated by Bar repression of CiFL. However, our results indicate that Bar function is independent of CiFL(Fig. 6I-L), supporting that the primary cause of ato repression behind the furrow may be a direct function of Bar as a repressor rather than indirect effects of the removal of the activator, CiFL. Furthermore, overexpression of CiFLby the lz-Gal4 driver in the presence of Bar did not activate ato expression (data not shown), indicating that Bar-mediated ato repression is epistatic to an overexpression of CiFLactivator.

Based on our findings, we propose a model of Bar function in retinal neurogenesis as summarized in Fig. 7. Ato is expressed within the furrow and is required for the generation of R8 founder neurons. Bar homeodomain proteins are expressed in the basal cells behind the furrow and represses ato expression,showing a complementary expression pattern to Ato. This function of Bar on ato-repression occurs independent of CiFL, a transcriptional activator of ato. Rather, Bar may directly repress ato transcription by binding to 3′- and 5′-regulatory regions of ato through its potential binding sites.

Dual function of Bar as transcriptional activator or repressor for different proneural genes

Our finding that Bar inhibits the expression of the proneural gene ato as a transcriptional repressor in the eye disc(Fig. 5) raises the interesting issue of whether Bar can also repress the expression of other proneural genes in the eye or other tissues. Misexpression of BarH1 or BarH2using a dpp-Gal4 driver showed increased expression of proneural gene scute in the eye and wing discs rather than repressing its expression, generating more sensory bristles (data not shown)(Sato et al., 1999). By contrast, the deficiency in Bar caused the loss of sensory interommatidial bristles in the eye (J.L. and K.-W.C., unpublished). These results suggest that Bar can act as an activator for the expression of ASC proneural genes to generate bristle sensory neurons. Therefore,Bar can act as transcriptional activator as well as repressor for different proneural genes depending on developmental contexts in Drosophila. This dual function of Bar was also observed in mammalian Barx2(Edelman et al., 2000). Barx2 has activator and repressor domains in the C- or N-terminal regions,respectively. Mbh1, another mammalian homolog of the Barclass genes, functions as either activator or repressor for the expression of neural bHLH genes in cell culture system(Saito et al., 1998). Therefore, the dual function of transcriptional activation and repression may be a general property of Bar family homeodomain proteins in the control of expression of neural target genes. These opposite actions of Bar may be dependent on the binding of their specific partners to the activator or repressor domain (Edelman et al.,2000).

It has been shown that loss of groucho (gro) results in increased ato expression behind the furrow of the eye disc(Chanut et al., 2000). Gro is a member of E(spl) complex to repress the expression of proneural genes during N-mediated lateral inhibition (Heitzler et al., 1996). It is interesting to note that Bar family homeodomain proteins have a conserved ∼10 amino acid motif termed the FIL domain at the N-terminal region of homeobox (Saito et al., 1998; Patterson et al., 2000). This domain shows sequence similarity to the core region of the engrailed homology-1 (eh1) domain in Engrailed (En) repressor(Smith and Jaynes, 1996),which can directly interact with Gro co-repressor through its eh1 motif(Jiménez et al., 1997). Therefore, Bar may interact with Gro through its FIL domain for its repressor function.

Conserved role of Bar family proteins

Bar class homeodomain proteins are evolutionarily highly conserved from Drosophila to human. Vertebrate Bar homologs include Xenopus XBH1 and XBH2(Patterson et al., 2000),mouse and human Barhl1 and Barhl2(Bulfone et al., 2000), rat Mbh1 [same gene as Barhl2(Saito et al., 1998)], and murine and human Barx1 and Barx2 genes(Jones et al., 1997). Although in vivo function of Bar homologs has not been extensively analyzed,some members of the Bar class homeobox genes may be involved in the genesis and fate specification of neuronal cells. A mammalian homolog, Mbh1, is expressed in a complementary pattern to Mash1, a homolog of ASC, in the rat eye(Saito et al., 1998). Hence,Mbh1 may be involved in inhibition of Mash1 expression, similar to the ato repression by Drosophila Bar proteins.

In vertebrate eye development, a mammalian homolog of ato, Math5(and/or Xath5), is crucial for the generation of retinal ganglion cells, which are the first neurons to arise and therefore may be analogous to the R8 founder cells in the Drosophila eye(Kanekar et al., 1997; Wang et al., 2001). The essential role of Math5 in the genesis of ganglion cells suggests that Math5 plays Ato-like proneural function in vertebrate eye development. It will be interesting to see whether a specific Bar homolog(s) may be involved in the repression of Math5 as the Drosophila Bar inhibits ato expression. In addition, Bar class genes are attractive candidates for many human genetic disorders, including Joubert syndrome and Rieger syndrome (Hjalt and Murray,1999; Bulfone et al.,2000; Blair et al.,2002). The new function of Drosophila Bar in negative regulation of neurogenesis may provide insights into the function of Bar family genes in vertebrates and the molecular basis of human diseases associated with altered Bar function.

We thank Hugo Bellen, Peter Bryant, Robert Holmgren, Yuh-Nung Jan, Tetsuya Kojima, Graeme Mardon, Kaoru Saigo, the Bloomington Drosophila Stock Center and the Developmental Studies Hybridoma Bank for providing flies and antibodies. We are grateful to Kyung-Ok Cho, Ya-Chieh Hsu, Sang-Chul Nam,Gregg Roman and Amit Singh for comments and discussion. Confocal microscopy was supported by a grant from the NIH to D. B. Jones. This work was supported by a grant from the NIH to K.-W.C.

Artavanis-Tsakonas, S., Rand, M. D. and Lake, R. J.(
1999
). Notch signaling: cell fate control and signal integration in development.
Science
284
,
770
-776.
Baker, N. E., Mlodzik, M. and Rubin, G. M.(
1990
). Spacing differentiation in the developing Drosophila eye: a fibrinogen-related lateral inhibitor encoded by scabrous.
Science
250
,
1370
-1377.
Baker, N. E. and Yu, S.-Y. (
1997
). Proneural function of neurogenic genes in the developing Drosophila eye.
Curr. Biol.
7
,
122
-132.
Baker, N. E. and Yu, S.-Y. (
1998
). The R8-photoreceptor equivalence group in Drosophila: fate choice precedes regulated Delta transcription and is independent of Notch gene dose.
Mech. Dev.
74
,
3
-14.
Baonza, A., Casci, T. and Freeman, M. (
2001
). A primary role for the epidermal growth factor receptor in ommatidial spacing in the Drosophila eye.
Curr. Biol.
11
,
396
-404.
Blackman, R. K., Sanicola, M., Raftery, L. A., Gillevet, T. and Gelbart,W. M. (
1991
). An extensive 3′ cis-regulatory region directs the imaginal disk expression of decapentaplegic, a member of the TGF-β family in Drosophila.
Development
111
,
657
-665.
Blair, I. P., Gibson, R. R., Bennett, C. L. and Chance, P. F. (
2002
). Search for genes involved in Joubert syndrome:evidence that one or more major loci are yet to be identified and exclusion of candidate genes EN1, EN2, FGF8, and BARHL1.
Am. J. Med. Genet.
107
,
190
-196.
Borod, E. R. and Heberlein, U. (
1998
). Mutual regulation of decapentaplegic and hedgehog during the initiation of differentiation in the Drosophila retina.
Dev. Biol.
197
,
187
-197.
Brand, A. H. and Perrimon, N. (
1993
). Targeted gene expression as a means of altering cell fates and generating dominant phenotypes.
Development
118
,
401
-415.
Brunet, J. F. and Ghysen, A. (
1999
). Deconstructing cell determination: proneural genes and neuronal identity.
BioEssays
21
,
313
-318.
Bulfone, A., Menguzzato, E., Broccoli, V., Marchitiello, A.,Gattuso, C.,Mariani, M., Consalez, G. G. Martinez, S., Ballabio, A. and Banfi, S. (
2000
). Barhl1, a gene belonging to a new subfamily of mammalian homeobox genes, is expressed in migrating neurons of the CNS.
Hum. Mol. Genet.
9
,
1443
-1452.
Carroll, S. B. and Whyte, J. S. (
1989
). The role of the hairy gene during Drosophilamorphogenesis-stripes in imaginal discs.
Genes Dev.
3
,
905
-916.
Chanut, F. and Heberlein, U. (
1997
). Role of decapentaplegic in initiation and progression of the morphogenetic furrow in the developing Drosophila retina.
Development
124
,
559
-567.
Chanut, F., Luk, A. and Heberlein, U. (
2000
). A screen for dominant modifiers of ro(Dom), a mutation that disrupts morphogenetic furrow progression in Drosophila, identifies groucho and hairless as regulators of atonal expression.
Genetics
156
,
1203
-1217.
Chen, C. K. and Chien, C. T. (
1999
). Negative regulation of atonal in proneural cluster formation of Drosophila R8 photoreceptors.
Proc. Natl. Acad. Sci. USA
96
,
5055
-5060.
Chen, C.-H., von Kessler, D. P., Park, W., Wang. B., Ma, Y. and Beachy,P. A. (
1999
). Nuclear trafficking of Cubitus interruptus in the transcriptional regulation of hedgehog target gene expression.
Cell
98
,
305
-316.
Crew, J. R., Batterham, P. and Pollock, J. A.(
1997
). Developing compound eye in lozenge mutants of Drosophila: Lozenge expression in the R7 equivalence group.
Dev. Genes Evol.
206
,
481
-493.
Cubas, P., de Celis, J. F., Campuzano, S. and Modolell, J.(
1991
). Proneural clusters of achaete-scute expression and the generation of sensory organs in the Drosophila imaginal wing disc.
Genes Dev.
5
,
996
-1008.
Dietrich, U. and Campos-Ortega, J. A. (
1984
). The expression of neurogenic loci in imaginal epidermal cells of Drosophila melanogaster.
J. Neurogenet
.
1
,
315
-322.
Dokucu, M. E., Zipursky, S. L. and Cagan, R. L.(
1996
). Atonal, Rough and the resolution of proneural clusters in the developing Drosophila retina.
Development
122
,
4139
-4147.
Dominguez, M. (
1999
). Dual role for Hedgehog in the regulation of the proneural gene atonal during ommatidia development.
Development
126
,
2345
-2353.
Dominguez, M. and Hafen, E. (
1997
). Hedgehog directly controls initiation and propagation of retinal differentiation in the Drosophila eye.
Genes Dev.
11
,
3254
-3264.
Edelman, D. B., Meech, R. and Jones, F. S.(
2000
). The homeodomain protein Barx2 contains activator and repressor domains and interacts with members of the CREB family.
J. Biol. Chem.
275
,
21737
-21745.
Epps, J. L., Jones, J. B. and Tanda, S. (
1997
). oroshigane, a new segment polarity gene of Drosophila melanogaster, functions in hedgehog signal transduction.
Genetics
145
,
1041
-1052.
Flores, G. V., Daga, A., Kalhor, H. R. and Banerjee, U.(
1998
). Lozenge is expressed in pluripotent precursor cells and patterns multiple cell types in the Drosophila eye through the control of cell-specific transcription factors.
Development
125
,
3681
-3687.
Frankfort, B. J. and Mardon, G. (
2002
). R8 development in the Drosophila eye: a paradigm for neural selection and differentiation.
Development
129
,
1295
-1306.
Frankfort, B. J., Nolo, R., Zhang, Z., Bellen, H. J. and Mardon,G. (
2001
). senseless repression of rough is required for R8 photoreceptor differentiation in the developing Drosophila eye.
Neuron
32
,
403
-414.
Ghysen, A. and Dambly-Chaudiere, C. (
1988
). From DNA to form: the achaete-scute complex.
Genes Dev.
2
,
495
-501.
Greenwood, S. and Struhl, G. (
1999
). Progression of the morphogenetic furrow in the Drosophila eye: the roles of Hedgehog, Decapentaplegic and the Raf pathway.
Development
126
,
5795
-5808.
Hassan, B. A., Bermingham, N. A., He, Y., Sun, Y., Jan, Y. N.,Zoghbi,H. Y. and Bellen, H. J. (
2000
). atonal regulates neurite arborization but does not act as a proneural gene in the Drosophila brain.
Neuron
25
,
549
-561.
Hayashi, T., Kojima, T. and Saigo, K. (
1998
). Specification of primary pigment cell and outer photoreceptor fates by BarH1 homeobox gene in the developing Drosophila eye.
Dev. Biol.
200
,
131
-145.
Hayashi, T. and Saigo, K. (
2001
). Diversification of cell types in the Drosophila eye by differential expression of prepattern genes.
Mech. Dev.
108
,
13
-27.
Heberlein, U., Wolff, T. and Rubin, G. M.(
1993
). The TGFβ homolog dpp and the segment polarity gene hedgehog are required for propagation of a morphogenetic wave in the Drosophila retina.
Cell
75
,
913
-926.
Heitzler, P., Bourouis, M., Ruel, L., Carteret, C. and Simpson,P. (
1996
). Genes of the Enhancer of split and achaete-scute complexes are required for a regulatory loop between Notch and Delta during lateral signalling in Drosophila.
Development
122
,
161
-171.
Higashijima, S.-i., Kojima, T., Michiue, T., Ishimaru, S.,Emori, Y. andSaigo, K. (
1992
). Dual Barhomeo box genes of Drosophila required in two photoreceptor cells, R1 and R6, and primary pigment cells for normal eye development.
Genes Dev.
6
,
50
-60.
Hjalt, T. A. and Murray, J. C. (
1999
). The human BARX2 gene: genomic structure, chromosomal localization, and single nucleotide polymorphisms.
Genomics
62
,
456
-459.
Ingham, P. W. (
1998
). Transducing Hedgehog: the story so far.
EMBO J.
17
,
3505
-3511.
Izaddoost, S., Nam, S.-C., Bhat, M. A., Bellen, H. J. and Choi,K.-W. (
2002
). Drosophila Crumbs is a positional cue in photoreceptor adherens junctions and rhabdomeres.
Nature
416
,
178
–183.
Jarman, A. P., Grell, E. H., Ackerman, L., Jan, L. Y. and Jan,Y. N. (
1994
). atonal is the proneural gene for Drosophila photoreceptors.
Nature
369
,
398
-400.
Jarman, A. P., Sun, Y., Jan, L. Y. and Jan, Y. N.(
1995
). Role of the proneural gene, atonal, in formation of Drosophila chordotonal organs and photoreceptors.
Development
121
,
2019
-2030.
Jiménez, G., Paroush, Z. and Ish-Horowicz, D.(
1997
). Groucho acts as a corepressor for a subset of negative regulators, including Hairy and Engrailed.
Genes Dev.
11
,
3072
-3082.
Jones, F. S., Kioussi, C., Copertino, D. W., Kallunki, P.,Holst, B. D. andEdelman, G. M. (
1997
). Barx2, a new homeobox gene of the Bar class, is expressed in neural and craniofacial structures during development.
Proc. Natl. Acad. Sci. USA
94
,
2632
-2637.
Kanekar, S., Perron, M., Dorsky, R., Harris, W. A., Jan, L. Y.,Jan, Y. N. and Vetter, M. L. (
1997
). Xath5participates in a network of bHLH genes in the developing Xenopusretina.
Neuron
19
,
981
-994.
Kojima, T., Sato, M. and Saigo, K. (
2000
). Formation and specification of distal leg segments in Drosophila by dual Bar homeobox genes BarH1 and BarH2.
Development
127
,
769
-778.
Kumar, J. P., Tio, M., Hsiung, F., Akopyan, S., Gabay, L.,Seger, R., Shilo,B. Z., and Moses, K. (
1998
). Dissecting the roles of the Drosophila EGF receptor in eye development and MAP kinase activation.
Development
125
,
3875
-3885.
Lee, E. C., Hu, X., Yu, S. Y. and Baker, N. E.(
1996
). The scabrous gene encodes a secreted glycoprotein dimer and regulates proneural development in Drosophilaeyes.
Mol. Cell. Biol.
16
,
1179
-1188.
Ma, C., Zhou, Y., Beachy, P. A. and Moses, K.(
1993
). The segment polarity gene hedgehog is required for progression of the morphogenetic furrow in the developing Drosophila eye.
Cell
75
,
927
-938.
Methot, N. and Basler, K. (
1999
). Hedgehog controls limb development by regulating the activities of distinct transcriptional activator and repressor forms of Cubitus interruptus.
Cell
96
,
819
-831.
Mlodzik, M., Baker, N. E. and Rubin, G. M.(
1990
). Isolation and expression of scabrous, a gene regulating neurogenesis in Drosophila.
Genes Dev.
4
,
1848
-1861.
Motzny, C. K. and Holmgren, R. (
1995
). The Drosophila cubitus interruptus protein and its role in the wingless and hedgehog signal transduction pathways.
Mech. Dev.
52
,
137
-150.
Nolo, R., Abbott, L. A. and Bellen, H. J.(
2000
). Senseless, a Zn finger transcription factor, is necessary and sufficient for sensory organ development in Drosophila.
Cell
102
,
349
-362.
Patterson, K. D., Cleaver, O., Gerber, W. V., White, F. G. and Krieg, P. A. (
2000
). Distinct expression patterns for two Xenopus Bar homeobox genes.
Dev. Genes Evol.
210
,
140
-144.
Pellikka, M., Tanentzapf, G., Pinto, M., Smith, C., McGlade, C. J., Ready,D. F. and Tepass, U. (
2002
). Crumbs, the Drosophila homologue of human CRB1/RP12, is essential for photoreceptor morphogenesis.
Nature
416
,
143
-149.
Price, M. A. and Kalderon, D. (
1999
). Proteolysis of Cubitus interruptus in Drosophila requires phosphorylation by protein kinase A.
Development
126
,
4331
-4339.
Romani, S., Campuzano, S., Macagno, E. R. and Modolell, J.(
1989
). Expression of achaete and scute genes in Drosophila imaginal discs and their function in sensory organ development.
Genes Dev.
3
,
997
-1007.
Saito, T., Sawamoto, K., Okano, H., Anderson, D. J. and Mikoshiba, K. (
1998
). Mammalian BarH homologue is potential regulator of neural bHLH genes.
Dev. Biol.
199
,
216
-225.
Sato, M., Kojima, T., Michiue, T. and Saigo, K.(
1999
). Bar homeobox genes are latitudinal prepattern genes in the developing Drosophila notum whose expression is regulated by the concerted functions of decapentaplegic and wingless.
Development
126
,
1457
-1466.
Schweisguth, F., Gho, M. and Lecourtois, M.(
1996
). Control of cell fate choices by lateral signaling in the adult peripheral nervous system of Drosophila melanogaster.
Dev. Genet.
18
,
28
-39.
Skeath, J. B. and Carroll, S. B. (
1991
). Regulation of achaete-scute gene expression and sensory organ pattern formation in the Drosophila wing.
Genes Dev.
5
,
984
-995.
Smith, S. T. and Jaynes, J. B. (
1996
). A conserved region of engrailed, shared among all en-, gsc-, Nk1-, Nk2- and msh-class homeoproteins, mediates active transcriptional repression in vivo.
Development
122
,
3141
-3150.
Strutt, D. I. and Mlodzik, M. (
1997
). Hedgehog is an indirect regulator of morphogenetic furrow progression in the Drosophila eye disc.
Development
124
,
3233
-3240.
Sun, Y., Jan, L. Y. and Jan, Y. N. (
1998
). Transcriptional regulation of atonal during development of the Drosophila peripheral nervous system.
Development
125
,
3731
-3740.
Tomlinson, A. and Struhl, G. (
2001
). Delta/Notch and Boss/Sevenless signals act combinatorially to specify the Drosophila R7 photoreceptor.
Mol. Cell
7
,
487
-495.
Treisman, J. E. (
2001
). Drosophilahomologues of the transcriptional coactivation complex subunits TRAP240 and TRAP230 are required for identical processes in eye-antennal disc development.
Development
128
,
603
-615.
Wang, S. W., Kim, B. S., Ding, K., Wang, H., Sun, D., Johnson,R. L.,Klein, W. H. and Gan, L. (
2001
). Requirement for math5 in the development of retinal ganglion cells.
Genes Dev.
15
,
24
-29.
Wolff, T. and Ready, D. F. (
1993
). Pattern formation in the Drosophila retina. In
The Development of Drosophila melanogaster
(ed. Bate, M. and Martinez-Arias, A.),pp.
1277
-1326. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press.
Xu, T. and Rubin, G. M. (
1993
). Analysis of genetic mosaics in developing and adult Drosophila tissues.
Development
117
,
1223
-1237.