In eukaryotes, the ability of DNA-binding proteins to act as transcriptional repressors often requires that they recruit accessory proteins, known as corepressors, which provide the activity responsible for silencing transcription. Several of these factors have been identified, including the Groucho (Gro) and Atrophin (Atro) proteins in Drosophila. Here we demonstrate strong genetic interactions between gro and Atro and also with mutations in a third gene, scribbler (sbb), which encodes a nuclear protein of unknown function. We show that mutations in Atro and Sbb have similar phenotypes, including upregulation of the same genes in imaginal discs, which suggests that Sbb cooperates with Atro to provide repressive activity. Comparison of gro and Atro/sbb mutant phenotypes suggests that they do not function together, but instead that they may interact with the same transcription factors, including Engrailed and C15, to provide these proteins with maximal repressive activity.

---

AN essential feature of all eukaryotic cells is the ability to transcribe only a subset of the genes present in their genomes. Whether a gene is transcribed or not is largely dependent upon the presence and activity of specific regulatory transcription factors, which bind in a sequence-specific manner to adjacent cis-regulatory elements. These regulatory transcription factors can act either as activators and promote transcription or as repressors and inhibit it. To function as repressors, these transcription factors often need to recruit accessory proteins known as corepressors, which provide the activity necessary to prevent transcription of a particular gene.

Several types of corepressors, including Groucho (Gro), C-terminal binding protein, Atrophin, N-CoR, and SMRT, have been identified (reviewed in GASTON and JAYARAMAN 2003). These corepressors will prevent transcription when in proximity to a specific promoter or activators; how they do this and how close they have to be to the promoter- or activator-binding sites can vary among corepressors. However, in terms of mechanism, a common theme is the ability to modify chromatin structure, often by recruiting histone deacetylase complexes; deacetylation results in compaction of chromatin, which presumably hinders access of regulatory and general transcription factors, thus shutting down transcription (DAVIE and DENT 2004).

Some transcription factors have been shown to recruit more than one corepressor; for example, the Hairy and Brinker proteins from Drosophila have recruitment motifs for both CtBP and Gro (PAROUSH et al. 1994; JIMENEZ et al. 1997; POORTINGA et al. 1998; ZHANG and LEVINE 1999; PHIPPEN et al. 2000; HASSON et al. 2001). In theory, the ability to recruit more than one corepressor can provide at least two possible advantages: the first is quantitative—that is, it may increase the repressive activity of the transcription factor; and the second is qualitative—that is, one corepressor may be more effective than another at repressing transcription of a specific gene. In reality, it is often more difficult to understand exactly why a particular transcription factor recruits more than one corepressor. For Brk, CtBP and Gro are largely redundant (HASSON et al. 2001; WINTER and CAMPBELL 2004), while for Hairy there is evidence that it may selectively recruit each corepressor for repression of specific targets (BIANCHI-FRIAS et al. 2004).

In this article we present a genetic study that investigates the links among three different proteins that appear to act as corepressors: Gro, Atro, and another nuclear protein, Scribbler, the specific function of which was previously unknown. New mutations in the genes encoding these factors were identified in a genetic screen for enhancers of mutations in the aristaless (al) gene. al encodes a transcription factor expressed at the presumptive tip of the leg and antenna, which, in combination with C15, another transcription factor expressed in the same cells, directs the differentiation of the structures found there in adult appendages, the arista and tarsal claws (SCHNEITZ et al. 1993; CAMPBELL and TOMLINSON 1998; CAMPBELL 2005; KOJIMA et al. 2005).

Gro belongs to the Gro/TLE family of related proteins and is one of the most widely studied corepressors (FISHER and CAUDY 1998; CHEN and COUREY 2000). It has been shown to interact with many different transcriptional repressors including Hairy, Runt, Engrailed (En), Even Skipped (Eve), Huckebein (Hkb), and Brk (PAROUSH et al. 1994; FISHER et al. 1996; ARONSON et al. 1997; JIMENEZ et al. 1997; TOLKUNOVA et al. 1998; GOLDSTEIN et al. 1999; HASSON et al. 2001; ZHANG et al. 2001) via a short interaction motif present in these proteins, of which there are two general types: the WRPW tetrapeptide class and the eh1/GEH octapeptide class (FISHER et al. 1996; ARONSON et al. 1997; TOLKUNOVA et al. 1998; GOLDSTEIN et al. 1999; JIMENEZ et al. 1999). These proteins require Gro to have maximal repressive activity, although most can repress by other mechanisms (PAROUSH et al. 1994; FISHER et al. 1996; ARONSON et al. 1997; JIMENEZ et al. 1997; GOLDSTEIN et al. 1999; HASSON et al. 2001; KOBAYASHI et al. 2001; ZHANG et al. 2001).

Drosophila Atro (aka Grunge) is a homolog of the Atrophin-1 protein of humans (ERKNER et al. 2002; ZHANG et al. 2002), named because the expansion of a polyglutamine tract is the cause of an inherited neuronal degenerative disease, dentatorubral-pallidoluysian atrophy (KOIDE et al. 1994; NAGAFUCHI et al. 1994). Atro mutants in Drosophila have a range of defects, including aberrant segmentation, neurogenesis, and dorsoventral patterning that are associated with derepression of several genes (ERKNER et al. 2002; ZHANG et al. 2002). The role of Atro as a corepressor was revealed by its direct interaction with the Eve and Hkb transcription factors, which appeared to be required for the normal repressive activities of these proteins. Further, both Drosophila Atro and human Atrophin-1 can repress transcription when fused to a heterologous DNA-binding domain (ZHANG et al. 2002). Atro is a multifunctional protein and, in addition to acting as a corepressor, appears to be involved in two other processes. First, it acts as a positive regulator of Hox gene expression and has been classed as a member of the trithorax group of genes (KANKEL et al. 2004). Second, it has also been shown to function in the cytoplasm, where it interacts with the cytoplasmic domain of the atypical cadherin Fat, and is required for establishment of planar cell polarity in the eye and wing (ZHANG et al. 2002; FANTO et al. 2003).

The scribbler (sbb) gene (aka brakeless and master-of-thickveins) encodes a novel nuclear protein of unknown function (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001). sbb mutations were uncovered through three different phenotypes: (1) abnormal turning behavior in larvae (YANG et al. 2000); (2) an axon-targeting defect in the eye where some photoreceptors project to the wrong location in the optic lobe (RAO et al. 2000; SENTI et al. 2000); and (3) deregulation of expression of the thickveins (tkv) gene in the wing imaginal disc (FUNAKOSHI et al. 2001). The sbb locus encodes two different isoforms: a short one, SbbA, of 929 residues, and a longer one, SbbB, of 2302 residues, which is a C-terminal extension of SbbA (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001). Both isoforms are composed largely of low-complexity sequence with only a single predicted functional domain, a zinc finger, in SbbB and none in SbbA. However, the C terminus of SbbA, extending into SbbB, has significant sequence homology to novel vertebrate proteins (predicted from ESTs), indicating that this is an important functional domain. Rescue experiments demonstrated that the SbbA isoform is capable of rescuing most mutant phenotypes (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001).

Mutations in sbb result in derepression of at least two genes, tkv in the wing and runt (run) in the eye (FUNAKOSHI et al. 2001; KAMINKER et al. 2002). tkv encodes a BMP receptor and is downregulated in the medial regions of the wing imaginal disc (TANIMOTO et al. 2000), but not in sbb mutant cells where levels are equivalent to those in the lateral regions (FUNAKOSHI et al. 2001). Curiously, the pattern of an enhancer trap in sbb mirrors that of tkv (FUNAKOSHI et al. 2001). However, elsewhere sbb appears to be uniformly expressed, including in the eye (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000). Each ommatidium of the eye has eight photoreceptors and the transcription factor Run is expressed only in photoreceptors R7 and R8. It appears to be important for photoreceptors to project their axons to the medulla in the optic lobe; axons from R1 to R6 project to the lamina instead (KAMINKER et al. 2002). Loss of sbb leads to ectopic expression of run in R2 and R5, resulting in axons from all six outer photoreceptors, R1–R6 projecting to the medulla (RAO et al. 2000; SENTI et al. 2000; KAMINKER et al. 2002). The fact that Sbb is ubiquitously expressed in the eye but represses run in R2 and R5, and not in R7 and R8, indicates either that its activity is modulated or that it could act as a corepressor for a transcription factor expressed in R2 and R5.

Here we show that Atro and sbb interact genetically and that mutations in each have very similar phenotypes in some tissues, including derepression of the same genes, raising the possibility that Sbb and Atro function together in a corepressor complex. In addition, these genetic studies indicate that Atro/Sbb are required, at least in part, for the repressive activity of two homeodomain transcription factors, C15 and Engrailed. We also show that gro interacts genetically with sbb and Atro mutants, but that the phenotype of gro mutants is distinct from that of Atro or sbb, indicating that Gro and Atro/Sbb function independently. The genetic interactions can, however, be explained if both Gro and Atro/Sbb associate with the same transcription factors, again including En and C15, and if both are required for the maximal activity of these factors.

Fly strains:

Flies carrying the following existing alleles or transgenes were used: al^ice, al¹³⁰ (In(2L)al¹³⁰), al^ush (not in FlyBase; CAMPBELL 2005), hs-flp (P{hsFLP}22), FRT82B (P{ry[+t7.2] = neoFRT}82B), Ubi-GFP (P{Ubi-GFP(S65T)nls}3R), M(2)201, M(3)95A (RpS3^Plac92), sbb⁶ (previously, mtv⁶), Atro³⁵ (Gug³⁵), hs-GFP (on 2R is Avic/GFP^hs.T:Ivir\HA, on 3L is Avic/GFP^{hs.T:Hsap\MYC}), gro^E48, arm-lacZ (EcoI/lacZ^arm.PV), M(2)IK (not in FlyBase; ZECCA and STRUHL 2002), M(3)i⁵⁵ (RpS17⁴), en^E (Df(2L)en^E), FRTG13 (P{w[+mW.hs] = FRT(w[hs])}G13), ovoD (P{w[+mC] = ovoD1-18}2R), FRT42D (P{ry[+t7.2] = neoFRT}42D), y+ (y^+t7.7), FRT2A (P{w[+mW.hs] = FRT(w[hs])}2A), tkv-lacZ (tkv⁰⁴⁴¹⁵), dpp-lacZ (dpp¹⁰⁶³⁸), hs-CD2 (P{hsp70-CD2.J}), mwh¹, y¹. Unless indicated otherwise in parentheses, all genotypes are as denoted in FlyBase (http://flybase.bio.indiana.edu), where more information on each can be found.

al enhancer screen:

This screen has been described previously and enhancers were characterized as mutations in the Egfr and C15 genes (CAMPBELL 2005).

Clonal analysis in adults and imaginal discs:

Homozygous mutant clones in imaginal discs and adults were generated in imaginal discs by hs-flp/FRT-induced mitotic recombination (XU and RUBIN 1993). Clones were generated in the second or early third instar of larvae with the following genotypes:

y hs-flp; FRT42D sbb⁶/FRT42D hs-GFP M(2)IK y+

y hs-flp; FRT42D en^E/FRT42D hs-GFP M(2)IK y+

hs-flp; dpp-lacZ FRT42D en^E/FRT42D hs-GFP M(2)IK y+

y hs-flp; Atro³⁵ FRT2A/M(3)i⁵⁵ y+ FRT2A

hs-flp; mwh¹ Atro³⁵ FRT2A/M(3)i⁵⁵ hs-GFP FRT2A

hs-flp; tkv-lacZ/+; Atro³⁵ FRT2A/M(3)i⁵⁵ hs-GFP FRT2A

hs-flp; FRT82B gro^E48/FRT82B Ubi-GFP M(3)95A

y hs-flp; dpp-lacZ/+; FRT82B gro^E48/FRT82B Ubi-GFP M(3)95A

y hs-flp; tkv-lacZ/ +; FRT82B gro^E48/FRT82B hs-CD2 M(3)95A

hs-flp; FRT42D sbb⁶/FRT42D arm-lacZ M(2)60E; FRT82B gro^E48/FRT82B Ubi-GFP M(3)95A.

Clones were identified in adults by loss of y+ and in discs by the loss of GFP or CD2 expression.

Germline clones:

Embryos mutant for both maternal and zygotic sbb were generated by the dominant autosomal germline clone technique (CHOU and PERRIMON 1992) using females with the genotype hs-flp; FRTG13 sbb⁶/FRTG13 ovoD. Clones were induced in late third instar larvae and the resulting adults were crossed to sbb⁶/Cyo males. It was not possible to identify which progeny received the sbb⁶ or the Cyo chromosome, but, as almost exactly 50% died (with a specific defect), it was assumed that these were homozygous for sbb⁶ and that the wild-type sbb gene on Cyo could rescue the loss of maternal contribution in the remaining 50%.

Immunostaining and cuticle preps:

Dissection and staining of imaginal discs were carried out by standard techniques. The following antibodies were used: anti-Al (rat; 1:1000) (CAMPBELL et al. 1993); anti-B (rabbit, 1:5) (HIGASHIJIMA et al. 1992); anti-ß-gal [rabbit, Cappell, 1:2000; mouse, Promega (Madison, WI) 1:200], anti-Ci (2A1, rat, 1:1) (SLUSARSKI et al. 1995); anti-En (ascites, 1:500) (PATEL et al. 1989); and anti-Run (1:500) (KOSMAN et al. 1998). Secondary antibodies were from Jackson Immunoresearch (West Grove, PA; Cy5 conjugates at 1:200) and Molecular Probes (Eugene, OR; Alexa 488 and Alexa 568 conjugates at 1:500). Tissue from adult flies was mounted in GMM (LAWRENCE and JOHNSTON 1986). First instar cuticle preps were prepared by an initial clearing in actic acid:glycerol (4:1; 30 min at 60°, 24 hr at room temp) and then mounting in a 3:1 mixture of CMCP-10 mounting medium (Polysciences):lactic acid.

A genetic screen for enhancers of al uncovers mutations in gro, Atro, and sbb:

The al gene encodes a homeodomain transcription factor that is expressed in the center of the leg and antennal imaginal discs of Drosophila (CAMPBELL et al. 1993; SCHNEITZ et al. 1993) (Figure 1A). This region gives rise to the distal-most structures of the respective appendage, the claw organ in the adult leg, and the arista in the antenna; in al mutants, both of these structures are absent or reduced (Figure 2, A and B) (CAMPBELL and TOMLINSON 1998). Previously, we described a genetic screen that identified genes operating upstream, Egfr, and in parallel, C15, to al (CAMPBELL 2005). EGF-receptor signaling is required for activation of al (CAMPBELL 2002; GALINDO et al. 2002), while C15 encodes another homeodomain protein, which is expressed in the same cells as Al. Al and C15 function together to regulate gene expression in the center of the leg and antennal discs (CAMPBELL 2005; KOJIMA et al. 2005).

View larger version (50K): In this window In a new window Download PPT slide   FIGURE 1.— Bar (B) expression (red) in leg discs containing gro, Atro, and sbb mutant clones. Clones are identified by the loss of a ubiquitous marker, GFP (green) in Bii, Cii, Dii, and Eii, ß-gal (blue) in Eii, and thus appear black. (A) In wild-type leg discs, Al (green) is expressed in a central domain and is surrounded by a ring of B expression. The Al domain is divided (white line) into two halves corresponding to the anterior and posterior compartments, which are almost identical in size. (Aii) Magnification of the central region showing complete absence of B in the center apart from two spots of expression corresponding to the tarsal claws. (B) B is partially derepressed in gro mutant clones. A clone bisects the center; its position and straight margin indicate that it comprises all of the anterior compartment in the center (the non-B-expressing cells; asterisk); this is smaller than the wild-type posterior, indicating that B expression has expanded into the center, also evidenced by the B domain being wider than in wild-type discs. In the very center (indicated by an asterisk in B and shown magnified in Biii), B is still almost completely repressed in gro mutant cells, although there is actually some very weak B expression in some cells here (Biii). (C) Atro clones are similar to gro mutant clones. This disc also contains a mutant clone that bisects the center; here the posterior (non-B-expressing cells; asterisk) is composed of Atro mutant cells and is smaller than the wild-type anterior due to expansion of B into the center. (Ciii) Also, although B is almost completely repressed in Atro mutant cells in the very center, there is some weak derepression, which is actually more apparent than in the gro mutant cells in Biii. (D) There is occasional expansion of the B domain into the center (arrow) in sbb mutant clones. (E) This disc contains clones of both sbb and gro; in the very center there is a clone of cells mutant for both sbb and gro (arrowhead). B is not deregulated in this clone.

 

View larger version (81K): In this window In a new window Download PPT slide   FIGURE 2.— al enhancer screen and the phenotype of sbb and Atro mutant legs and antenna. (A–H) Adult antennae. (A) Wild-type with the arista labeled (ar). (B) alice mutant in which the arista is almost completely absent. (C) Weaker al130/ush mutant that has an almost full-length arista. (D–F) Also al130/ush mutants, but now heterozygous for AtroEal15 (D), sbbEal17 (E), and groEal13 (F). These mutations dominantly enhance the arista phenotype of al130/ush mutants. (G) Antenna almost entirely composed of AtroEal15 mutant clones; the arista is fatter and shorter with reduced lateral branches. (H) Antenna almost entirely composed of sbb6 mutant clones; the arista is very similar to that in G. (I–K) Tips of adult legs. (I) sbbEal17 homozygote, which has one normal claw (Ii) and associated pulvillus (p) and a defective claw (Iii) and a normal associated pulvillus. (J) Leg containing AtroEal15 clones. (Ji) This claw is completely mutant and is much narrower than the adjacent claw (Jii), which is composed of wild-type cells. (K) Claw from a sbb6 mutant clone; this is also narrower. (L) Adult prothoracic legs composed almost entirely (apart from the claws) of AtroEal15 clones. The legs have fused and segmentation of the medial and proximal regions is almost absent. (M) Prothoracic legs from the sbbEal17 homozygote; these are also fused and have defective segmentation. A partial duplication is also present. (N–Q) The margin from adult wings containing Atro35 or sbb6 mutant clones. (N and O) Anterior, proximal margin that normally develops the triple row of bristles, including a large peg-like bristle; wild-type bristles are indicated (TR). These large bristles are narrower in Atro and sbb mutant tissue, and longer in Atro– tissue. (P and Q) Posterior margin that normally develops posterior row bristles that do not have a socket (PR in P). These bristles have a socket and are longer in Atro (P) and sbb (Q) mutant tissue.

 
This screen also uncovered six additional enhancers: 4, 13, 15, 17, 37, and 73 (Figure 2, C–F), which corresponded to three different complementation groups—4/17, 13/37, and 15/73—and not to Egfr or C15. Mapping by recombination and/or deficiencies and subsequent complementation testing against known genes in potential locations of these enhancers revealed that 4/17 is allelic to sbb, 13/37 to groucho (gro), and 15/73 to Atrophin (Atro). Null alleles and deficiencies covering sbb [Df(2L)PC4] and gro [Df(3R)Espl22] also act as enhancers. In contrast, Atro³⁵, a null allele (ERKNER et al. 2002), is a weaker enhancer than 15 or 73, suggesting that these may have dominant negative activity. The six enhancers have been renamed as sbb^Eal4, sbb^Eal17, gro^Eal13, gro^Eal37, Atro^Eal15, and Atro^Eal73.

sbb and Atro mutations have similar effects on patterning of the aristae and claws:

Next we compared the phenotype of al mutations to those of gro, Atro, and sbb mutations in a wild-type al background, first in adults and then in imaginal discs. Strong al mutants survive to the adult stage in which both the arista (Figure 2B) and the claw organ (consisting of a pair of claws, a pair of pulvilli, and an empodium) are absent (CAMPBELL and TOMLINSON 1998). Of the six enhancers, direct analysis of homozygotes was possible only in sbb^Eal17 because the other five enhancers were either embryonic or larval lethal as homozygotes. Similarly, previously characterized null alleles of Atro and gro, including Atro³⁵ and gro^E58, are embryonic lethal, while sbb nulls, such as sbb⁶, survive embryogenesis but die during the larval phase (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001; ERKNER et al. 2002; ZHANG et al. 2002; KANKEL et al. 2004). sbb^Eal17 thus is weaker than a null allele and sequencing revealed a premature stop in this mutant, truncating the long isoform (SbbB) at residue 1620 and eliminating 682 residues at the C terminus, the sequence of the short isoform (SbbA) being normal (Figure 3A).

View larger version (22K): In this window In a new window Download PPT slide   FIGURE 3.— Sbb and C15 proteins. (A) sbb encodes two protein isoforms, SbbA and SbbB, which are made up largely of low-complexity sequence, except for three regions that show similarity to predicted vertebrate proteins (indicated as 1, 2, and 3) (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001). The mutation in sbbEal17 introduces a stop codon after residue 1620, truncating the SbbB isoform but leaving the SbbA isoform unchanged. (Bi) The C15 protein contains a eh1-type Gro-interaction motif situated near the N terminus. (Bii) Vertebrate homologs of C15 also possess this motif at the N terminus.

 
sbb^Eal17 homozygotes actually die as pharate adults, and this permitted examination of their antennae and legs, revealing their aristae to be wild type (not shown), but their legs occasionally had a severely reduced tarsal claw (Figure 2I; 7% had at least one claw affected); the other structures of the claw organ, the pulvilli and empodium, were unaffected. sbb^Eal17/6 flies also survive to pharate adults in which defective claws are found at a higher frequency (24%). The effect of complete loss of Sbb was examined in sbb⁶ homozygous clones and was found to result in severe reduction in the size, particularly the width, of claws composed of mutant tissue (Figure 2K). Patterning of sbb⁶ mutant antennae was also disrupted so that the aristae were reduced in length and fatter (Figure 2H); in addition, some sense organs on the third antennal segment, including the sensilla trichodea, were absent or lacking setae (not shown). The claw phenotype is similar to that found in the hypomorphic al¹³⁰ mutant (CAMPBELL and TOMLINSON 1998), suggesting that Al activity is reduced in the absence of Sbb. However, the arista phenotype is distinct from that in al hypomorphs, which have aristae that resemble those in wild type, but are shorter and thinner (STERN and BRIDGES 1926).

Atro mutant clones (both Atro^Eal15 and Atro³⁵) at the tip of the antenna and leg produced phenotypes very similar to that of sbb⁶ clones, i.e., shorter, fatter aristae and reduced claws resembling bristles (Figure 2, G and J). In contrast, as already reported, gro mutant clones (both gro¹³ and gro^E48) are not recovered in adults (HEITZLER et al. 1996).

gro, Atro, and sbb mutants show partial derepression of B:

Al, in combination with C15, represses expression of genes such as Bar (B) in the center of the leg and antennal discs; in wild-type discs B is absent from the center and is expressed in a ring surrounding Al (Figure 1A); in al mutants, however, B is expressed throughout the center (CAMPBELL et al. 1993; KOJIMA et al. 2000; CAMPBELL 2005; KOJIMA et al. 2005). Given that Gro and Atro are corepressors, an obvious possibility was that one or both interact with Al and/or with C15 to facilitate repression of B. Previous studies indicated that Al alone does not act as a transcriptional repressor and, in fact, can activate expression of B in the absence of C15 (CAMPBELL 2005). Consistent with this, no Gro interaction motifs could be identified in Al. C15, however, possesses an eh1 class Gro interaction motif first identified in the En homeodomain protein (Figure 3B) (TOLKUNOVA et al. 1998; JIMENEZ et al. 1999); this motif is conserved among vertebrate homologs of C15 (Figure 3Bii). More direct studies are required to determine if Atro physically interacts with C15 because, although Atro has been shown to interact directly with other transcription factors, the specific sequence required has not yet been characterized (ZHANG et al. 2002).

To determine if Gro and/or Atro are required to provide repressive activity to C15/Al, B expression was examined in leg discs containing gro and Atro mutant clones. This revealed that there was some derepression of B in both gro and Atro mutant cells in the center (Figure 1, B and C), but that this was much less dramatic than with al or C15 mutant clones (CAMPBELL 2005). More specifically, the B expression domain expands into the center in gro and Atro mutant cells, but there is always a small region in the center where B is still repressed (Figure 1, B and C). sbb mutant clones also have minor effects on B expression, although this appeared to be less consistent and severe than with gro or Atro clones (Figure 1D). Thus, loss of Gro or Atro does result in partial derepression of B, but with a much weaker effect than loss of Al or C15, demonstrating that Al and C15 can still repress B in the absence of either Gro or Atro corepressors, although not as effectively. One possibility is that Gro and Atro act partially redundantly in this respect so that, in the absence of both, Al and C15 will not be able to repress B. Testing this by generating gro Atro double-mutant clones is technically difficult as each is on a different arm of the third chromosome. However, further experiments, described below, indicate that Atro requires Sbb to provide repressor activity to transcription factors; this would also account for the very similar phenotypes of Atro and sbb clones at the tip of the leg and antenna. Consequently, the phenotype of sbb gro double-mutant clones was investigated (technically not that difficult as each is on a separate chromosome) and it was found that B was still repressed in the very center (Figure 1E). This indicates that Atro may function as a corepressor in the absence of Sbb or that Al and C15 can use other mechanisms to repress B in addition to Atro/Sbb and Gro.

Atro and sbb mutations have similar phenotypes outside the tip of the leg and antenna:

As already described, sbb and Atro mutant clones produce similar phenotypes at the tip of the leg and antenna; this is also true in other locations in the adult. First, patterning of the dorsal region of the head around the ocelli is disrupted in both sbb and Atro mutants. In wild-type adults, this region, known as the ocellar triangle, consists of three well-spaced ocelli, six to eight small interocellar bristles, and two large ocellar bristles (Figure 4A). In sbb^Eal17 homozygotes and sbb^Eal17/6 adults, there is a compression of this region, with fusion of ocelli and loss or mispositioning of the interocellar and ocellar bristles (Figure 4, B and C). sbb⁶ or Atro³⁵ mutant clones result in the loss of almost all of these elements, with a single ocellus often being the only structure remaining (Figure 4, D and E). Second, it has been reported already that Atro mutant clones can generate fusion of segments in the leg (ERKNER et al. 2002; KANKEL et al. 2004) (Figure 2L) and this is also true for sbb mutant clones (not shown). This is also seen quite dramatically in 5% of the legs from sbb^Eal17 homozygotes (Figure 2M; curiously, this phenotype is stronger in sbb^Eal17 homozygotes than in sbb^Eal17/6 adults). The most dramatic fusions are actually seen between adjacent prothoracic legs; this is also true for Atro clones (Figure 2L). Third, both Atro and sbb clones produce a similar phenotype at the margin of the wing. In proximal anterior regions, the triple row of wild-type wings includes a large, wide central bristle; in both Atro and sbb mutant clones these bristles are thinner, and in Atro, but not sbb, clones they are longer (Figure 2, N and O). In the posterior of wild-type wings, the hairs at the margin do not have a socket, but sbb and Atro mutant hairs do possess a small socket and are slightly longer than wild-type margin hairs. This is indicative of a partial transformation to the anterior where all the bristles at the margin possess a socket.

View larger version (61K): In this window In a new window Download PPT slide   FIGURE 4.— Defects in patterning of the ocellar triangle in sbb, Atro, gro, and en mutant tissue and in trans-heterozygotes. (A) In wild-type adults this region of the dorsal head comprises three ocelli (arrowheads), two large ocellar bristles (b), and six to eight smaller interocellar bristles (i). (B and C) In sbbEal17 homozygotes and sbbEal17/6 adults, this region is compressed, resulting in fusion of ocelli, loss of interocellar bristles, and loss and mispositioning of the ocellar bristles. (D and E) Heads containing sbb6 and Atro35 mutant clones result in more severe patterning defects in which almost all of this region is lost; arrows mark a single ocellus (note that it is not possible to identify mutant tissue in the region affected because the marker used, y, marks only the bristles and no mutant interocellar or ocellar bristles are ever identified). (F and G) Trans-heterozygotes among sbbEal17, AtroEa;15 or groEal13 also have defects in patterning in this region, but these are less extreme. (H and I) Trans-heterozygotes between enL11 and Atro and gro also have defective ocellar triangles; here the number of interocellar bristles is reduced. (J) enL11 mutant clones result in almost complete loss of the ocellar triangle; here a single enlarged ocellus remains.

 
It should be noted, however, that sbb and Atro mutant clones do not always produce similar phenotypes. For example, Atro mutant clones have been shown to disrupt planar polarity (ZHANG et al. 2002; FANTO et al. 2003); sbb mutant clones have no effect on bristle or hair orientation in the wing (not shown).

Atro and sbb mutations have identical effects on tkv and run expression in imaginal discs:

The similarity between the phenotypes of sbb and Atro mutant clones in adults suggested that Sbb and Atro may function together in a corepressor complex. This predicted that any gene derepressed in cells lacking Atro would be derepressed in cells lacking Sbb and vice versa. This was tested in imaginal discs: previous studies have revealed that loss of Sbb results in upregulation of tkv expression in the anterior compartment of the wing and misexpression of run in the eye (FUNAKOSHI et al. 2001; KAMINKER et al. 2002). It was found that Atro mutant clones show an identical phenotype (Figure 5, A, B, D, and E), supporting the argument that Sbb is required for Atro to function as a corepressor.

View larger version (97K): In this window In a new window Download PPT slide   FIGURE 5.— Comparison of Atro and gro mutant clones in imaginal discs. Clones appear black and are identified by loss of a ubiquitous marker (red: GFP or CD2). (A) In wild-type wing discs, tkv is expressed at relatively high levels posterior to the compartment boundary (arrow); however, anterior to the boundary, levels are very low and then increase laterally. (B) Wing disc containing Atro mutant clones (black); tkv (green) is upregulated in clones in the anterior (arrowheads). (C) Wing disc containing gro mutant clones; these have no effect on tkv expression (green). (D and E) Eye disc containing Atro mutant clones. E is a magnification of the box in D. In wild-type tissue (red), Run (green) is expressed in a single photoreceptor per ommatidium (yellow arrowhead is R8; in older ommatidia Run is also expressed in R7, which is in a different plane; photoreceptors are identified by blue Elav staining). In mutant tissue there are three Run-positive photoreceptors per ommatidium (white arrowhead). (F and G) Eye disc containing gro mutant clones. G is a magnification of the box in F. Mutant tissue (white arrowhead) is largely devoid of Run-expressing cells except at the margin, although mutant cells do express the neural-specific Elav marker. (H–J) Antennal discs. (H) In wild-type antennal discs Run is expressed in a small number of cells in three clusters in the third segment. (I) This pattern appears almost identical in this disc containing Atro mutant clones. (J) gro clones, however, result in a dramatic increase in the number of cells expressing Run.

 

sbb mutant embryos have a similar but weaker phenotype than Atro mutants:

A significant difference between Atro and sbb mutants is that null Atro mutants are embryonic lethal while null sbb mutants survive embryogenesis and die as larvae (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001; ERKNER et al. 2002; ZHANG et al. 2002; KANKEL et al. 2004). Although Atro mutants are lethal, the severity of this lethal phenotype is made significantly worse by removing maternal contribution with germline clones (ERKNER et al. 2002; ZHANG et al. 2002; KANKEL et al. 2004). Consequently, we tested whether sbb mutant embryos may be rescued by maternal contribution and found that loss of both maternal and zygotic sbb resulted in embryonic lethality (embryos lacking only maternal contribution survive embryogenesis); and although there was some variability, these embryos had a distinct segmental phenotype consisting of the deletion of three or more abdominal segments (Figure 6, A and B). Analysis of En expression, which marks the posterior of each segment, revealed that most mutant embryos had defects in alternate segments, primarily A4, -6, and -8, corresponding to even-numbered En stripes in parasegments 10, 12, and 14 (Figure 6, C and D). This is similar to the phenotype of Atro mutant embryos, which also have defects in even-numbered En stripes (KANKEL et al. 2004). However, this Atro mutant phenotype is associated with embryos that are lacking maternal but not zygotic contribution; embryos that lack both have a very severe phenotype and in many cases appear as if they have not been fertilized (ERKNER et al. 2002; ZHANG et al. 2002; KANKEL et al. 2004). Consequently, although the sbb embryonic phenotype is very similar to that of Atro mutants, it is not as severe.

View larger version (56K): In this window In a new window Download PPT slide   FIGURE 6.— sbb germline clones. (A and B) Cuticle preps of first instar larvae. (A) Wild-type with eight abdominal segments (1–8). (B) sbb mutant; the number of abdominal segments is reduced to four. (C) Wild-type germband extended embryo stained for En expression. (D) sbb maternal and zygotic mutant embryo stained for En. There are defects in some of the stripes in the abdomen (asterisks).

 

Dominant genetic interactions among Atro, sbb, and gro:

Although the initial complementation tests revealed that the Atro mutations picked up as enhancers of al were not alleles of sbb because they were on different chromosomes, it was noted that sbb/+; Atro/+ trans-heterozygous adults had weak phenotypes similar to homozygous single mutants of Atro and sbb. This was most profound in the dorsal head where patterning of the ocellar triangle was invariably disrupted; this was manifested in fusion of ocelli, mispositioning of ocellar bristles, and loss of interocellar bristles (Figure 4F). This interaction was observed between Atro and sbb alleles uncovered in the al enhancer screen and also between the extant alleles sbb⁶ and Atro³⁵ and so was not associated with preexisting mutations in the stocks used in this screen. It proved possible to quantitate this phenotype by simply counting the number of interocellar bristles in adults of different genotypes (Figure 7). Wild-type adults have an average of 7.3 (±1.1) interocellar bristles (Figure 7A) and, surprisingly, adults heterozygous for either sbb or Atro mutations (including the likely null alleles, Atro³⁵ and sbb⁶) have a significant reduction in this number, indicating that each has a very weak dominant phenotype, presumably due to haplo-insufficiency, at least for the null alleles (Figure 7B). This number is significantly reduced further in sbb/+; Atro/+ trans-heterozygotes (Figure 7C).

View larger version (49K): In this window In a new window Download PPT slide   FIGURE 7.— Genetic interactions among sbb, Atro, gro, and en revealed by a reduction in the number of interocellar bristles in adult trans-heterozygotes. The genotype of the adults is given below the histogram. (A) Wild type. (B) Heterozygotes for only a single mutation. (C–E) Trans-heterozygotes; i.e., the adults also had a wild-type copy of mutations 1 and 2. To determine if there was a significant change in the number of interocellar bristles, Student t-tests were performed comparing the number of bristles to that of both mutations in single heterozygotes (in B). Almost all of the genotypes showed a significant reduction in the number of bristles. Error bars indicate the standard error of the mean.

 
Further analysis revealed that gro mutations also interacted dominantly with sbb and Atro mutations (Figure 4G), so that gro heterozygous adults also have a significant reduction in the number of ocellar bristles (Figure 7B) and this is reduced further in both sbb/+; gro/+ and Atro +/ + gro trans-heterozygotes (Figure 7D).

sbb/Atro mutants have phenotypes distinct from those of gro mutants:

The dominant genetic interactions between sbb and Atro can be explained if Sbb and Atro proteins function together; this can also account for the very similar phenotypes of sbb and Atro homozygous clones in imaginal discs and adults described above. This led to the question of whether the dominant genetic interactions between gro and Atro or sbb can also be explained if Atro, Sbb, and Gro function together as a corepressor complex. This was tested by comparing the phenotype of gro mutant clones in imaginal discs to those of Atro and sbb. As already described, both Atro and sbb clones have an identical effect on tkv and run expression in the wing and eye disc, respectively. In contrast, it was found that loss of gro had no effect on tkv expression in the wing (Figure 5C), while gro mutant clones in the eye had a phenotype different from that of Atro and sbb, resulting primarily in loss of run expression rather than ectopic expression (Figure 5, F and G). During the course of this study, however, it was found that run was ectopically expressed in gro mutant clones located in the antennal disc. In wild-type antennal discs, Run is expressed in a small number of cells in three clusters in the presumptive third antennal segment; in gro mutant clones, the number of Run-expressing cells in this disc increases dramatically (Figure 5, H and J). In contrast, Run expression in antennal discs containing Atro or sbb mutant clones is indistinguishable from wild type (Figure 5I and data not shown). Thus, there are clear differences between the effects of the loss of Atro/Sbb and the loss of Gro on gene expression so that, for example, Atro-dependent repression of tkv in the wing is independent of Gro, while Gro-dependent repression of run in the antenna is independent of Atro/Sbb.

Dominant genetic interactions between Gro/Atro/sbb and en:

What is the basis of the dominant genetic interactions between gro and Atro or sbb? One possibility is that Gro and Atro/Sbb interact with the same transcription factor that represses gene expression in the presumptive ocellar triangle and that both Gro and Atro/Sbb are required for maximal activity of this factor. This is the same explanation used for Al and C15 to explain why gro, Atro, and sbb all act as enhancers of al. A candidate for this transcription factor required for patterning the ocellar triangle was En, which has been reported to be expressed in this region (ROYET and FINKELSTEIN 1996) and is known to have a Gro interaction motif, the original eh1 motif (JIMENEZ et al. 1997, 1999; TOLKUNOVA et al. 1998). In addition, the posterior wing margin phenotype of Atro and sbb mutant clones is consistent with a weak posterior-to-anterior transformation that could be explained by a reduction in En activity, En being required to establish posterior cell fates (GARCIA-BELLIDO and SANTAMARIA 1972).

This possibility was tested first by determining if there were any dominant genetic interactions between en and gro, Atro, or sbb [the en allele used in these studies was Df(2L)en^E, which takes out both en and its partially redundant partner, invected (inv)] (TABATA et al. 1995). Unlike gro, Atro, and sbb mutations, en^E is not haplo-insufficient and heterozygotes have a normal number of interocellar bristles (Figure 7A). However, en^E trans-heterozygotes with some, but not all, gro, Atro, and sbb alleles show a significant reduction in the number of these bristles (Figure 4, H and I; Figure 7E). This possibility was tested further by examining the phenotype of en^E clones in the head and it was found that these resulted in almost complete loss of the ocellar triangle with a single ocellus remaining (Figure 4J).

Derepression of one but not other En-target genes in gro, Atro, or sbb mutant cells:

The genetic interactions between en and gro, Atro, and sbb predicted that loss of En in the presumptive ocellar triangle in the eye imaginal disc might result in derepression of some of the same genes in gro and/or Atro and sbb mutant clones. A direct test of this is not possible at present because there are no known targets of En in the presumptive ocellar triangle. Consequently, we decided to study this question in the wing imaginal disc where several targets of En have been identified, including decapentaplegic (dpp) and cubitus interruptus (ci); En is expressed in the posterior compartment of the wing, while ci is expressed throughout the anterior and dpp in a stripe immediately anterior to the compartment boundary (Figure 8, A, F, and I) (BROWER 1986; EATON and KORNBERG 1990; MASUCCI et al. 1990; ORENIC et al. 1990; RAFTERY et al. 1991; SANICOLA et al. 1995; SCHWARTZ et al. 1995; TABATA et al. 1995; ZECCA et al. 1995). Consistent with previous studies, we find that both ci and dpp are ectopically expressed in en mutant clones in the posterior (Figure 8, F and I). We have also identified another En target in the wing—al; unlike the leg, where al is expressed in both anterior and posterior compartments, in the wing it is restricted to the anterior (Figure 8A). As for ci and dpp, clonal analysis reveals that this restriction is En dependent (Figure 8B).

View larger version (75K): In this window In a new window Download PPT slide   FIGURE 8.— Deregulation of gene expression in en, gro, Atro, and sbb mutant clones in wing discs. Clones are identified by the loss of a ubiquitous marker (red or blue). (A–E) Al expression (green). (A) Wild-type disc. En (red) is expressed in the posterior (to the left in all discs); Al is expressed in the anterior in a fan-shaped pattern [the slight overlap with En is probably explained by the anterior expansion of En in the late third instar (BLAIR 1992)]. (B) Disc almost entirely composed of en clones showing expansion of the Al expression domain into the posterior (arrow), forming a mirror image of that in the anterior. (C) Disc almost completely composed of gro clones showing expansion of the Al domain into the posterior, but the domain here is smaller than in the anterior (expression in the anterior is also expanded slightly). (D and E) Al is also ectopically expressed in posterior Atro and sbb clones (arrows), but this is much more restricted compared to expression in the anterior. (F–H) Ci (green) is normally expressed only in the anterior but is ectopically expressed in posterior en clones (F, arrows). However, expression is normal in discs almost entirely composed of gro (G) or Atro (H) clones. (I–K) dpp is normally expressed only in a stripe immediately anterior to the compartment boundary. As for Ci, dpp is misexpressed in posterior en clones (I), but loss of gro (J) or Atro (K) has no effect on expression in the posterior. (L and M) Disc containing both gro (loss of blue) and sbb (loss of red) clones; cells mutant for both appear as black. M is a magnification of the box in L. In the posterior, Ci is very weakly ectopically expressed in cells mutant for both gro and sbb (arrow).

 
Next we tested whether repression of ci, dpp, and al by En was dependent upon Gro, Atro, or Sbb. Both ci and dpp expression is unaffected in gro, Atro, or sbb single-mutant clones in the posterior (Figure 8, G, H, J, and K and not shown), indicating that En can repress these targets completely in the absence of one of these factors. In contrast, there is some ectopic expression of al in gro, Atro, or sbb mutant clones (Figure 8, C–E). However, this derepression is only partial so that the ectopic al expression is not as extensive as with loss of en. These results indicate that En can repress ci and dpp completely even in the absence of Gro, Atro, or Sbb, but that repression of al is partially dependent upon Gro, Atro, or Sbb. Similar to Al/C15-dependent repression of B in the leg, one possibility was that Gro and Atro/Sbb may act redundantly (for ci and dpp) or partially redundantly (for al) in En-dependent repression. As already mentioned, generating gro Atro double-mutant clones is technically difficult, but we did analyze sbb gro double-mutant clones and found that there was very weak derepression of ci in the posterior in cells lacking both Sbb and Gro (Figure 8, L and M). This shows that Gro and Sbb are at least partially redundant, but that En is still very active in the absence of both. As for Al/C15, we can conclude either that En uses mechanisms to repress that do not involve Gro and Atro/Sbb or that Atro has some activity in the absence of Sbb. Resolution of this will await analysis of Atro gro double-mutant clones.

Sbb is required for Atro activity:

Previous studies demonstrated that Atro acts as a corepressor in Drosophila, the most convincing of these being the demonstration that fusion of Atro to a heterologous DNA-binding domain confers repressive activity to the chimera (ZHANG et al. 2002). Atro has been shown to interact directly with two transcription factors, Even-Skipped (Eve) and Huckebein, and the repressive ability of Eve is compromised in Atro mutants (ZHANG et al. 2002), probably accounting for the loss of en expression in even-numbered parasegments in Atro mutant embryos (KANKEL et al. 2004).

Our studies here are consistent with Atro acting as a corepressor as we show that several genes, including run, tkv, al, and B, are completely or partially derepressed in Atro mutant clones in imaginal discs (Figure 1C; Figure 5, B and E; Figure 8D), suggesting that transcriptional repressors required to silence these genes recruit Atro. Atro-dependent repression of B in the center of the leg disc is very likely due to interaction with the transcription factor C15, which is expressed in the center of the leg and is required for repression of B (CAMPBELL 2005; KOJIMA et al. 2005). Similarly, Atro-dependent repression of al in the posterior of the wing is very likely due to interaction with En, which is expressed in the posterior and required to exclude al from this compartment (Figure 8, A and B). At present it is unclear which transcription factors recruit Atro to repress run in the eye or tkv in the wing, although a strong candidate for run would be the Rough homeodomain protein (TOMLINSON et al. 1988), which is expressed in the same cells, R2 and R5, that exhibit ectopic run expression in Atro mutant clones. Whether Atro can, in fact, bind directly to C15, En, and possibly Rough, needs to be tested biochemically, as previous studies with Eve and Hkb (ZHANG et al. 2002) did not identify a possible interaction motif for Atro nor do sequence comparisons among C15, En, Eve, and Hkb suggest a common motif.

The sbb gene encodes a nuclear protein with unknown function (RAO et al. 2000; SENTI et al. 2000; YANG et al. 2000; FUNAKOSHI et al. 2001). sbb mutations have many different phenotypes affecting multiple tissues. Here we show that sbb and Atro interact very strongly genetically (Figure 4F; Figure 7) and that many of the phenotypes of sbb mutants are very similar to those of Atro mutants, including derepession of run, tkv, al, and B in imaginal discs (Figure 1, C and D; Figure 5, B and E) (FUNAKOSHI et al. 2001; KAMINKER et al. 2002). Thus, Atro is unable to silence these genes in the absence of Sbb, suggesting that it is required for Atro activity either to recruit Atro to transcription factors or possibly to assist binding of these factors to DNA. As these transcription factors appear to function normally in some respects in the absence of Sbb (or Atro), it appears more likely that Sbb and Atro function together in a corepressor complex.

One problem with the proposal that Atro activity is dependent upon Sbb is that the phenotypes of Atro and sbb mutants are not identical. For example, embryos lacking both maternal and zygotic Atro have a very severe, almost uncharacterizable phenotype (ERKNER et al. 2002; ZHANG et al. 2002; KANKEL et al. 2004), while embryos lacking both maternal and zygotic Sbb have a much less severe phenotype, characterized by a reduced number of abdominal segments (Figure 6B), that is similar to that of embryos lacking only maternal Atro. This could be explained if Atro is partially active in the absence of Sbb, or if it is dependent upon Sbb for repression of some genes but not others. Alternatively, the difference between Atro and sbb mutant phenotypes could be related to Atro having functions other than that of a corepressor. It is has been implicated in positive regulation of Hox gene expression (KANKEL et al. 2004), and it also functions in the cytoplasm to control planar cell polarity (ZHANG et al. 2002; FANTO et al. 2003). Our analysis of sbb mutants does not reveal any potential involvement of Hox gene expression or planar cell polarity and, consequently, if Sbb is required only for Atro to act as a corepressor, then it is not surprising that Atro and sbb mutant phenotypes are not identical. Further experiments are required to determine the nature of the Atro dependence on Sbb for transcriptional repression and how direct any interactions might be.

Some transcription factors may recruit both Gro and Atro/Sbb for maximal activity:

We originally uncovered mutations in sbb and Atro in a genetic screen for enhancers of al. As already mentioned, it is likely that they act as enhancers because they are utilized by the C15 transcription factor to repress genes such as B; C15 is expressed in the same cells as Al and it is thought that they bind together to regulate gene expression (CAMPBELL 2005; KOJIMA et al. 2005). We have also uncovered strong genetic interactions among sbb, Atro, and en mutations, which could be explained if En also recruits Atro/Sbb.

Curiously, our genetic studies also revealed strong interactions among gro, sbb, and Atro. This could be explained if Gro was also required for Atro activity; i.e., all three proteins may form a corepressor complex. However, this appears to be unlikely because, in contrast to the similar phenotypes of sbb and Atro mutants, there are several distinct differences among the phenotypes of gro mutants and those of sbb and Atro mutants. For example, repression of tkv in the anterior of the wing is dependent on both Sbb and Atro but not on Gro (Figure 5, B and C) (FUNAKOSHI et al. 2001), while repression of run in the antennal disc is dependent upon Gro but not upon Atro or Sbb (Figure 5, H and I; data not shown). This suggests that a specific transcription factor recruits Atro/Sbb to repress tkv in the wing and another transcription factor recruits Gro to repress run in the antenna. The identity of these transcription factors remains to be uncovered.

In some cases gro mutants do have a similar phenotype to those of Atro and sbb; this includes partial derepression of al expression in the posterior of the wing and B in the center of the leg (Figure 1, B and C; Figure 8, C and D). This can be explained if C15 (expressed in the center of the leg) and En (expressed in the posterior of the wing) recruit both Gro and Atro/Sbb and if each imparts some but not all the repressive activity to these transcription factors. Consistent with this, both C15 and En possess eh1-type Gro-interaction motifs (Figure 3) (JIMENEZ et al. 1997, 1999; TOLKUNOVA et al. 1998) and previous studies have revealed that En can repress in the absence of Gro (TOLKUNOVA et al. 1998). As already mentioned, further biochemical studies are required to determine if C15 and En can indeed recruit Atro.

At present it is unclear whether Atro and Gro provide all the repressive activity to C15 and En; this will await the generation of Atro gro double-mutant clones. We have, however, analyzed sbb gro double-mutant clones and these reveal that some targets of C15 and En are still at least partially repressed, although En activity appears to be somewhat compromised following the simultaneous loss of Sbb and Gro, in comparison to loss of one of these alone (Figure 1E; Figure 8M). Either Atro has some activity in the absence of Sbb or C15 and En can use mechanisms other than recruitment of Gro and Atro to repress transcription. Many transcription factors have been shown to have the ability to repress by several mechanisms; for example, although Brk recruits both CtBP and Gro, it can repress some genes in the absence of both, using additional repression domains (WINTER and CAMPBELL 2004).

Why do C15 and En need to recruit both Gro and Atro? En can repress some genes completely in the absence of either Gro or Atro, for example, ci and dpp in the wing (Figure 8, G, H, J, and K). However, for repression of al, the activity of En is clearly reduced in the absence of either, indicating that it needs to recruit both to completely repress this gene (Figure 8, C and D). This would suggest a quantitative explanation; i.e., En recruits both Gro and Atro to increase its activity, rather than to allow it to repress specific genes repressed more efficiently by one or the other. This is consistent with the suggestion that both corepressors function via a similar mechanism: both Gro and a vertebrate homolog of Atro have been shown to recruit a histone deacetylase (CHEN et al. 1999; ZOLTEWICZ et al. 2004). The recruitment of both may decrease histone acetylation to a level that cannot be achieved with either alone.

We thank A. Bodnar for technical assistance and T. Welch for some preliminary studies. The screen for enhancers of al was begun in Andrew Tomlinson's lab and G.C. thanks him for support. We are indebted to Pascal Heitzler for the al^ush allele. We thank the following people for materials used in this study: C. Alonso, S. Cohen, B. Dickson, S. Kerridge, R. Holmgren, T. Kojima, K. Matthews, K. Cook and the Bloomington Stock Center, Y. Rao, J. Reinitz, K. Senti, M. Sokolowski, G. Struhl, T. Tabata, and M. Zecca. We thank D. Chapman and B. Stronach for comments on the manuscript. This work was supported by grants from the National Institutes of Health (GM60368) and the March of Dimes (no. 1-FY02-176) to G.C.

ARONSON, B. D., A. L. FISHER, K. BLECHMAN, M. CAUDY and J. P. GERGEN, 1997 Groucho-dependent and -independent repression activities of Runt domain proteins. Mol. Cell. Biol. 17: 5581–5587.[Abstract]

BIANCHI-FRIAS, D., A. ORIAN, J. J. DELROW, J. VAZQUEZ, A. E. ROSALES-NIEVES et al., 2004 Hairy transcriptional repression targets and cofactor recruitment in Drosophila. PLoS Biol. 2: E178.[CrossRef][Medline]

BLAIR, S. S., 1992 Engrailed expression in the anterior lineage compartment of the developing wing blade of Drosophila. Development 115: 21–33.[Abstract]

BROWER, D. L., 1986 Engrailed gene expression in Drosophila imaginal discs. EMBO J. 5: 2649–2656.[Medline]

CAMPBELL, G., 2002 Distalization of the Drosophila leg by graded EGF-receptor activity. Nature 418: 781–785.[CrossRef][Medline]

CAMPBELL, G., 2005 Regulation of gene expression in the distal region of the Drosophila leg by the Hox11 homolog, C15. Dev. Biol. 278: 607–618.[CrossRef][Medline]

CAMPBELL, G., and A. TOMLINSON, 1998 The roles of the homeobox genes aristaless and Distal-less in patterning the legs and wings of Drosophila. Development 125: 4483–4493.[Abstract]

CAMPBELL, G., T. WEAVER and A. TOMLINSON, 1993 Axis specification in the developing Drosophila appendage: the role of wingless, decapentaplegic, and the homeobox gene aristaless. Cell 74: 1113–1123.[CrossRef][Medline]

CHEN, G., and A. J. COUREY, 2000 Groucho/TLE family proteins and transcriptional repression. Gene 249: 1–16.[CrossRef][Medline]

CHEN, G., J. FERNANDEZ, S. MISCHE and A. J. COUREY, 1999 A functional interaction between the histone deacetylase Rpd3 and the corepressor groucho in Drosophila development. Genes Dev. 13: 2218–2230.[Abstract/Free Full Text]

CHOU, T. B., and N. PERRIMON, 1992 Use of a yeast site-specific recombinase to produce female germline chimeras in Drosophila. Genetics 131: 643–653.[Abstract]

DAVIE, J. K., and S. Y. DENT, 2004 Histone modifications in corepressor functions. Curr. Top. Dev. Biol. 59: 145–163.[Medline]

EATON, S., and T. B. KORNBERG, 1990 Repression of ci-D in posterior compartments of Drosophila by engrailed. Genes Dev. 4: 1068–1077.[Abstract/Free Full Text]

ERKNER, A., A. ROURE, B. CHARROUX, M. DELAAGE, N. HOLWAY et al., 2002 Grunge, related to human Atrophin-like proteins, has multiple functions in Drosophila development. Development 129: 1119–1129.[Abstract/Free Full Text]

FANTO, M., L. CLAYTON, J. MEREDITH, K. HARDIMAN, B. CHARROUX et al., 2003 The tumor-suppressor and cell adhesion molecule Fat controls planar polarity via physical interactions with Atrophin, a transcriptional co-repressor. Development 130: 763–774.[Abstract/Free Full Text]

FISHER, A. L., and M. CAUDY, 1998 Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and invertebrates. Genes Dev. 12: 1931–1940.[Free Full Text]

FISHER, A. L., S. OHSAKO and M. CAUDY, 1996 The WRPW motif of the hairy-related basic helix-loop-helix repressor proteins acts as a 4-amino-acid transcription repression and protein-protein interaction domain. Mol. Cell. Biol. 16: 2670–2677.[Abstract]

FUNAKOSHI, Y., M. MINAMI and T. TABATA, 2001 mtv shapes the activity gradient of the Dpp morphogen through regulation of thickveins. Development 128: 67–74.[Abstract]

GALINDO, M. I., S. A. BISHOP, S. GREIG and J. P. COUSO, 2002 Leg patterning driven by proximal-distal interactions and EGFR signaling. Science 297: 256–259.[Abstract/Free Full Text]

GARCIA-BELLIDO, A., and P. SANTAMARIA, 1972 Developmental analysis of the wing disc in the mutant engrailed of Drosophila melanogaster. Genetics 72: 87–104.[Abstract/Free Full Text]

GASTON, K., and P. S. JAYARAMAN, 2003 Transcriptional repression in eukaryotes: repressors and repression mechanisms. Cell. Mol. Life Sci. 60: 721–741.[CrossRef][Medline]

GOLDSTEIN, R. E., G. JIMENEZ, O. COOK, D. GUR and Z. PAROUSH, 1999 Huckebein repressor activity in Drosophila terminal patterning is mediated by Groucho. Development 126: 3747–3755.[Abstract]

HASSON, P., B. MULLER, K. BASLER and Z. PAROUSH, 2001 Brinker requires two corepressors for maximal and versatile repression in Dpp signalling. EMBO J. 20: 5725–5736.[CrossRef][Medline]

HEITZLER, P., M. BOUROUIS, L. RUEL, C. CARTERET and P. SIMPSON, 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.[Abstract]

HIGASHIJIMA, S., T. KOJIMA, T. MICHIUE, S. ISHIMARU, Y. EMORI et al., 1992 Dual Bar homeo 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.[Abstract/Free Full Text]

JIMENEZ, G., Z. PAROUSH and D. ISH-HOROWICZ, 1997 Groucho acts as a corepressor for a subset of negative regulators, including Hairy and Engrailed. Genes Dev. 11: 3072–3082.[Abstract/Free Full Text]

JIMENEZ, G., C. P. VERRIJZER and D. ISH-HOROWICZ, 1999 A conserved motif in goosecoid mediates groucho-dependent repression in Drosophila embryos. Mol. Cell. Biol. 19: 2080–2087.[Abstract/Free Full Text]

KAMINKER, J. S., J. CANON, I. SALECKER and U. BANERJEE, 2002 Control of photoreceptor axon target choice by transcriptional repression of Runt. Nat. Neurosci. 5: 746–750.[Medline]

KANKEL, M. W., D. M. DUNCAN and I. DUNCAN, 2004 A screen for genes that interact with the Drosophila pair-rule segmentation gene fushi tarazu. Genetics 168: 161–180.[Abstract/Free Full Text]

KOBAYASHI, M., R. E. GOLDSTEIN, M. FUJIOKA, Z. PAROUSH and J. B. JAYNES, 2001 Groucho augments the repression of multiple Even skipped target genes in establishing parasegment boundaries. Development 128: 1805–1815.[Abstract]

KOIDE, R., T. IKEUCHI, O. ONODERA, H. TANAKA, S. IGARASHI et al., 1994 Unstable expansion of CAG repeat in hereditary dentatorubral-pallidoluysian atrophy (DRPLA). Nat. Genet. 6: 9–13.[CrossRef][Medline]

KOJIMA, T., M. SATO and K. SAIGO, 2000 Formation and specification of distal leg segments in Drosophila by dual Bar homeobox genes, BarH1 and BarH2. Development 127: 769–778.[Abstract]

KOJIMA, T., T. TSUJI and K. SAIGO, 2005 A concerted action of a paired-type homeobox gene, aristaless, and a homolog of Hox11/tlx homeobox gene, clawless, is essential for the distal tip development of the Drosophila leg. Dev. Biol. 279: 434–445.[CrossRef][Medline]

KOSMAN, D., S. SMALL and J. REINITZ, 1998 Rapid preparation of a panel of polyclonal antibodies to Drosophila segmentation proteins. Dev. Genes Evol. 208: 290–294.[CrossRef][Medline]

LAWRENCE, P. A., and P. JOHNSTON, 1986 Methods of marking cells, pp. 229–242 in Drosophila: A Practical Approach, edited by D. B. ROBERTS. IRL Press, Oxford.

MASUCCI, J. D., R. J. MILTENBERGER and F. M. HOFFMANN, 1990 Pattern-specific expression of the Drosophila decapentaplegic gene in imaginal disks is regulated by 3' cis-regulatory elements. Genes Dev. 4: 2011–2023.[Abstract/Free Full Text]

NAGAFUCHI, S., H. YANAGISAWA, E. OHSAKI, T. SHIRAYAMA, K. TADOKORO et al., 1994 Structure and expression of the gene responsible for the triplet repeat disorder, dentatorubral and pallidoluysian atrophy (DRPLA). Nat. Genet. 8: 177–182.[CrossRef][Medline]

ORENIC, T. V., D. C. SLUSARSKI, K. L. KROLL and R. A. HOLMGREN, 1990 Cloning and characterization of the segment polarity gene cubitus interruptus Dominant of Drosophila. Genes Dev. 4: 1053–1067.[Abstract/Free Full Text]

PAROUSH, Z., R. L. FINLEY, JR., T. KIDD, S. M. WAINWRIGHT, P. W. INGHAM et al., 1994 Groucho is required for Drosophila neurogenesis, segmentation, and sex determination and interacts directly with hairy-related bHLH proteins. Cell 79: 805–815.[CrossRef][Medline]

PATEL, N. H., E. MARTIN-BLANCO, K. G. COLEMAN, S. J. POOLE, M. C. ELLIS et al., 1989 Expression of engrailed proteins in arthropods, annelids, and chordates. Cell 58: 955–968.[CrossRef][Medline]

PHIPPEN, T. M., A. L. SWEIGART, M. MONIWA, A. KRUMM, J. R. DAVIE et al., 2000 Drosophila C-terminal binding protein functions as a context-dependent transcriptional co-factor and interferes with both mad and groucho transcriptional repression. J. Biol. Chem. 275: 37628–37637.[Abstract/Free Full Text]

POORTINGA, G., M. WATANABE and S. M. PARKHURST, 1998 Drosophila CtBP: a Hairy-interacting protein required for embryonic segmentation and hairy-mediated transcriptional repression. EMBO J. 17: 2067–2078.[CrossRef][Medline]

RAFTERY, L. A., M. SANICOLA, R. K. BLACKMAN and W. M. GELBART, 1991 The relationship of decapentaplegic and engrailed expression in Drosophila imaginal disks: Do these genes mark the anterior-posterior compartment boundary? Development 113: 27–33.[Abstract]

RAO, Y., P. PANG, W. RUAN, D. GUNNING and S. L. ZIPURSKY, 2000 brakeless is required for photoreceptor growth-cone targeting in Drosophila. Proc. Natl. Acad. Sci. USA 97: 5966–5971.[Abstract/Free Full Text]

ROYET, J., and R. FINKELSTEIN, 1996 hedgehog, wingless and orthodenticle specify adult head development in Drosophila. Development 122: 1849–1858.[Abstract]

SANICOLA, M., J. J. SEKELSKY, S. ELSON and W. M. GELBART, 1995 Drawing a stripe in Drosophila imaginal disks: negative regulation of decapentaplegic and patched expression by engrailed. Genetics 139: 745–756.[Abstract]

SCHNEITZ, K., P. SPIELMANN and M. NOLL, 1993 Molecular genetics of aristaless, a prd-type homeo box gene involved in the morphogenesis of proximal and distal pattern elements in a subset of appendages in Drosophila. Genes Dev. 7: 114–129. (erratum: Genes Dev. 7(5): 911).[Abstract/Free Full Text]

SCHWARTZ, C., J. LOCKE, C. NISHIDA and T. B. KORNBERG, 1995 Analysis of cubitus interruptus regulation in Drosophila embryos and imaginal disks. Development 121: 1625–1635.[Abstract]

SENTI, K., K. KELEMAN, F. EISENHABER and B. J. DICKSON, 2000 brakeless is required for lamina targeting of R1–R6 axons in the Drosophila visual system. Development 127: 2291–2301.[Abstract]

SLUSARSKI, D. C., C. K. MOTZNY and R. HOLMGREN, 1995 Mutations that alter the timing and pattern of cubitus interruptus gene expression in Drosophila melanogaster. Genetics 139: 229–240.[Abstract]

STERN, C., and C. V. BRIDGES, 1926 The mutants of the extreme left end of the second chromosome of Drosophila melanogaster. Genetics 11: 503–530.[Free Full Text]

TABATA, T., C. SCHWARTZ, E. GUSTAVSON, Z. ALI and T. B. KORNBERG, 1995 Creating a Drosophila wing de novo, the role of engrailed, and the compartment border hypothesis. Development 121: 3359–3369.[Abstract]

TANIMOTO, H., S. ITOH, P. TEN DIJKE and T. TABATA, 2000 Hedgehog creates a gradient of DPP activity in Drosophila wing imaginal discs. Mol. Cell 5: 59–71.[CrossRef][Medline]

TOLKUNOVA, E. N., M. FUJIOKA, M. KOBAYASHI, D. DEKA and J. B. JAYNES, 1998 Two distinct types of repression domain in engrailed: one interacts with the groucho corepressor and is preferentially active on integrated target genes. Mol. Cell. Biol. 18: 2804–2814.[Abstract/Free Full Text]

TOMLINSON, A., B. E. KIMMEL and G. M. RUBIN, 1988 rough, a Drosophila homeobox gene required in photoreceptors R2 and R5 for inductive interactions in the developing eye. Cell 55: 771–784.[CrossRef][Medline]

WINTER, S. E., and G. CAMPBELL, 2004 Repression of Dpp targets in the Drosophila wing by Brinker. Development 131: 6071–6081.[Abstract/Free Full Text]

XU, T., and G. M. RUBIN, 1993 Analysis of genetic mosaics in developing and adult Drosophila tissues. Development 117: 1223–1237.[Abstract]

YANG, P., S. A. SHAVER, A. J. HILLIKER and M. B. SOKOLOWSKI, 2000 Abnormal turning behavior in Drosophila larvae: identification and molecular analysis of scribbler (sbb). Genetics 155: 1161–1174.[