Logo of nihpaAbout Author manuscriptsSubmit a manuscriptNIH Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Dev Biol. Author manuscript; available in PMC Nov 15, 2009.
Published in final edited form as:
Dev Biol. Nov 15, 2008; 323(2): 248–260.
Published online Aug 22, 2008. doi:  10.1016/j.ydbio.2008.08.012
PMCID: PMC2653429

Maternal Groucho and bHLH repressors amplify the dose-sensitive X chromosome signal in Drosophila sex determination


In Drosophila, XX embryos are fated to develop as females, and XY embryos as males, because the diplo-X dose of four X-linked signal element genes, XSEs, activates the Sex-lethal establishment promoter, SxlPe, whereas the haplo-X XSE dose leaves SxlPe off. The threshold response of SxlPe to XSE concentrations depends in part on the bHLH repressor, Deadpan, present in equal amounts in XX and XY embryos. We identified canonical and non-canonical DNA-binding sites for Dpn at SxlPe and found that cis-acting mutations in the Dpn-binding sites caused stronger and earlier Sxl expression than did deletion of dpn implicating other bHLH repressors in Sxl regulation. Maternal Hey encodes one such bHLH regulator but the E(spl) locus does not. Elimination of the maternal corepressor Groucho also caused strong ectopic Sxl expression in XY, and premature Sxl activation in XX embryos, but Sxl was still expressed differently in the sexes. Our findings suggest that Groucho and associated maternal and zygotic bHLH repressors define the threshold XSE concentrations needed to activate SxlPe and that they participate directly in sex signal amplification. We present a model in which the XSE signal is amplified by a feedback mechanism that interferes with Gro-mediated repression in XX, but not XY embryos.

Keywords: Hes, X:A ratio, genetic switch, helix-loop-helix, scute, repression, WRPW, X chromosome-counting


Dose-sensitive promoters respond to small differences in regulatory protein concentrations to produce large differences in gene expression. In some instances differential concentrations of activators alone appear to set promoters into their appropriate expression states, but the general rule is that the enhancers controlling switch-like promoters integrate concentration-dependent inputs from both activators and inhibitors to establish precise boundaries of expression (see Mannervik et al., 1999; Barolo and Posakony, 2002; Clyde et al., 2003; Ochoa-Espinosa et al., 2005). In the developing nervous systems of flies and vertebrates, for example, antagonistic interactions between negatively and positively-acting proteins of the basic-helix-loop-helix, bHLH, family define the sharp boundaries of gene expression required for specification of neural precursor cells (reviewed in Massari and Murre, 2000). Similar antagonistic interactions between bHLH proteins and their associated cofactors are hypothesized to play important roles in the specification of the alternative male and female fates in Drosophila.

Chromosomal sex determination in Drosophila is a textbook example of how two-fold changes in transcriptional regulatory protein concentrations can elicit different developmental outcomes (reviewed by Cline and Meyer, 1996; Ashburner et al., 2005). In the fly, the collective dose of four X chromosome-linked signal element genes, XSEs, conveys X chromosome dose to the master regulatory gene Sex-lethal, Sxl (Cline, 1993; Erickson and Quintero, 2007). In XX embryos the double XSE dose directs the transient activation of the Sxl establishment promoter, SxlPe, initiating a positive autoregulatory splicing loop that operates on pre-mRNAs produced from the constitutive promoter, SxlPm, thereby maintaining Sxl in the on (female) state for the remainder of its life (Cline, 1984; Bell et al., 1991; Keyes et al., 1992; Nagengast et al., 2003). In XY embryos, the single dose of XSEs leaves SxlPe inactive, precluding functional splicing of SxlPm-derived transcripts and thereby directing the male fate.

Three of the four XSE genes encode transcription factors that directly regulate SxlPe. The two strongest XSEs, scute and sisA, encode bHLH and bZIP activators, while runt encodes the founding member of the RUNX class of DNA binding proteins (Cline, 1988; Cline and Meyer, 1996; Ashburner et al., 2005). Although the dose-sensitive XSE proteins are of central importance in the X-counting process, their direct action at SxlPe requires additional protein factors. Maternally-supplied daughterless protein, for example, interacts with Scute to form the DNA binding bHLH heterodimer, Sc/Da, while maternally supplied STAT, and presumably, Bicoid stability factor, bind directly to SxlPe to facilitate expression (Yang et al., 2001; Bosch et al., 2006; Avila and Erickson, 2007; De Renzis et al., 2007). How these and other factors work to effectively amplify the two-fold difference in XSE dose into an all-or-nothing response at SxlPe is unknown. Cooperative or combinatorial interactions among the XSE and maternal activators in protein assembly, DNA binding, or via interactions with the general transcription machinery, have been offered as possible explanations of how male and female XSE concentrations might be reliably distinguished at SxlPe (Cline, 1993; Erickson and Cline, 1993; Yang et al., 2001). Other models, however, focus on the means by which negative regulators might amplify the difference in XSE protein concentrations to generate a reliable sex-determining signal (see Parkhurst et al., 1990; Schutt and Nothiger, 2000; Gilbert, 2006).

Three negative regulators of SxlPe have been identified: the maternally supplied extramachrochetae (emc) and groucho (gro) products and the zygotically expressed product of the autosomal gene deadpan (dpn) (Younger-Shepherd et al., 1992; Paroush et al., 1994; Barbash and Cline, 1995). Emc is an HLH protein that lacks a basic DNA-binding domain and exerts its inhibitory effects by forming heterodimers with bHLH activators, such as Scute and Da, thereby preventing them from binding to DNA (Massari and Murre, 2000; Campuzano, 2001). While emc apparently plays a minor role in sex determination (Younger-Shepherd et al., 1992), loss of maternal gro has been reported to cause male embryos to express female levels of Sxl protein, suggesting that Gro-mediated repression of SxlPe may be essential for distinguishing X chromosome dose (Paroush et al., 1994). Gro is the archetypal example of the widely-distributed Gro/TLE family of transcriptional corepressors, that are recruited to DNA by virtue of their interactions with several different groups of sequence-specific DNA binding proteins; including bHLH repressors such as Dpn (reviewed in (Fisher and Caudy, 1998; Chen and Courey, 2000; Buscarlet and Stifani, 2007; Fischer and Gessler, 2007).

The dpn gene was identified as an autosomal sex signal element, or ASE, because it functions as a zygotically expressed negative regulator of Sxl (Younger-Shepherd et al., 1992; Barbash and Cline, 1995). Present in equal amounts in XX and XY embryos the dpn product is needed to properly assess the male XSE dose as evidenced by the finding that loss of dpn function causes some XY cells to activate SxlPe and adopt the inappropriate female fate (Younger-Shepherd et al., 1992; Barbash and Cline, 1995). Dpn is a member of the Hairy-Enhancer of split, HES, family of bHLH repressors (reviewed in Fisher and Caudy, 1998; Massari and Murre, 2000; Iso et al., 2003; Fischer and Gessler, 2007; Kageyama et al., 2007). HES proteins and the closely related HEY family (HES with YRPW) bind to the “E-box” CACGTG and the related sequence CACGCG, the later being the optimal sequence for Hairy and Dpn (Ohsako et al., 1994; Van Doren et al., 1994). HES factors also bind with reduced affinity to the "N-box" CACRAG suggesting that there is a range of allowable in vivo target sites.

HES proteins repress transcription by several different mechanisms. Best understood is the recruitment of the corepressor Gro to DNA via the C-terminal peptide sequence, WRPW, present in all HES family members (Paroush et al., 1994; Fisher et al., 1996; Fisher and Caudy, 1998). Some HES proteins recruit other corepressors such as CtBP and Sir2 to DNA and there is evidence that mutual antagonism between different corepressors can influence HES protein function (Poortinga et al., 1998; Zhang and Levine, 1999; Bianchi-Frias et al., 2004). Repression may also be mediated directly by competition with activators for DNA binding or by sequestering bHLH activators into inactive heterodimers (Fisher and Caudy, 1998; Fischer and Gessler, 2007; Kageyama et al., 2007). Most of these schemes have been invoked to explain how Dpn might function during sex determination (Paroush et al., 1994; Dawson et al., 1995; Jimenez et al., 1997), but none have been examined in detail.

Although Dpn is the only known DNA-binding repressor of SxlPe, loss of dpn function has a relatively mild effect, causing low-level ectopic activation of SxlPe in a subset of male nuclei (Barbash and Cline, 1995). Given the efficiency of HES/Gro-mediated repression in other contexts (Barolo and Levine, 1997; Zhang and Levine, 1999; Courey and Jia, 2001) and the presence of two canonical CACGCG Dpn-binding sequences at SxlPe (Hoshijima et al., 1995; Winston et al., 1999), it is not clear why Dpn has such a modest effect on sex determination. One possibility is that Dpn function could be modulated, perhaps by chemical modification (Karandikar et al., 2005), or by competition with other DNA binding proteins (Yang et al., 2001; Louis et al., 2003). A second possibility is that additional repressors negatively regulate SxlPe: an explanation consistent with the report that loss of maternal gro function leads to high-levels of Sxl protein in XY embryos (Paroush et al., 1994).

To better understand the role of transcriptional repression in primary sex determination we characterized the cis-acting promoter elements recognized by Dpn, and analyzed the effects of maternal gro on SxlPe. Our studies revealed that SxlPe contains three functional Dpn DNA-binding sites, including one with the non-canonical sequence CACACT. Mutations in the Dpn-binding sites had stronger and earlier effects on SxlPe than did a null dpn mutation, suggesting that additional bHLH repressors regulate SxlPe. We found that the Hey locus encodes one such maternal-effect repressor of SxlPe, but that the E(spl)m3 gene, which had previously been proposed to regulate Sxl (Dawson et al., 1995; Poortinga et al., 1998), does not. The gro product influences SxlPe earlier and more strongly than does dpn, suggesting that the initial concentrations of XSE proteins needed to activate SxlPe in XX embryos are defined by Gro-mediated repression and then modulated upward to compensate for rising XSE levels in XY embryos. We propose a model for SxlPe regulation in which the XSE signal is amplified by a positive feedback mechanism that inhibits Gro-mediated repression in XX, but not XY, embryos.

Materials and Methods

Plasmids, mutagenesis, and P-element transformation

The GST-Dpn bHLH plasmid was made by inserting a PCR fragment encoding amino acids 1–108 into the Bam H1 and Eco RI sites of PGEX-2TK. To make the MBP-Dpn plasmid the entire dpn coding sequence was cloned as a PCR fragment into the Eco RI and Hin dIII sites of pMAL-c2. The dpn-VP16 cell culture expression plasmid carried dpn codons 1–108 fused to the VP16 activation domain (residues 410–490) plus a Kozak sequence in PAct5CPPA (Han et al., 1989). The minimal −94 bp SxlPe-luciferase plasmid has been described (Yang et al., 2001). The 4X Dpn site-firefly luciferase reporters have the following sequences between the Xho I and EcoR I sites of the −94 bp SxlPe-Fluc plasmid: 1,2-CCCACGCGACGCCCACGCGAGCCCACGCGACGCCCACGCGAC; 3-GGCACACTTCTGGCACACTTCCGCACACTTCTGGCACACTTC (3m has CACcCT); 4-GCCACGTTCCAGCCACGTTCCGCCACGTTCCTGCCACGTTCC (4m has CAaGcT).

P-element transformation vectors were based on pCaSpeR-AUG-βgal and carried SxlPe sequences from −1.45 kb to +44 bp derived from wild-type or mutated variants of plasmid pG01 (Yang et al., 2001). Point mutations were made by site-directed oligonucleotide mutagenesis and confirmed by DNA sequencing. Mutated sequences were as follows: site 1, CACtgG; site 2, CtCGaG; site 3, CACcCT; site 4, CAaGcT. P-element transformants were obtained from w1118 flies by co-injection with the pTurbo transposase source.

Protein expression and purification

To produce GST-Dpn bHLH and MBP-Dpn proteins, BL21(DE3) cells carrying the corresponding expression plasmid were grown in LB at 21° to an OD600 of 0.3 and induced with 0.1 mM IPTG for 1–2 hr. Cell pellets were suspended in 1/40 culture volume of 20 mM Hepes, 0.6 M NaCL, 0.5 mM EDTA, 1%(v/v) NP-40, 2mM DTT, pH = 7.9 and lysed by sonication. After 10 min centrifugation at 10,000 × g, supernatants were diluted with one volume of 20 mM Hepes pH = 7.9. GST-Dpn bHLH was purified to homogeneity using glutathione-agarose beads (Sigma) and MBP-Dpn using amylose affinity resin (New England BioLabs).

DNase I footprinting and electrophoretic mobility-shift assays (EMSA)

For DNase I footprinting, probes were made by PCR amplification with one 32P end-labeled primer, and gel-purified. Approximately 104 cpm of probe was included in a 20 µl reaction containing: 15 mM Hepes, 50 mM KCL, 1 mM EDTA, 2mM DTT, 7.5% (v/v) glycerol, 0.1% (v/v) NP-40, 1µg polydI:polydC, 5 µg bovine serum albumin pH = 7.9 and indicated units of Dpn fusion protein. One Dpn unit equaled 15 nM (~10 ng of GST-Dpn bHLH domain or 20 ng of MBP-Dpn). After 30 min at 21° 0.05 units of DNase I (Epicentre) was added. After two min 80 µl 0.1M EDTA, 1.0 M NaCl was added to stop the reaction. Samples were phenol:CHCl3 extracted, ethanol precipitated, dissolved in 80% formamide, 0.01 N NaOH, 1 mM EDTA, and heated to 90° for 5 min before loading on 6% polyacrylamide/8M urea gels. Msp I-cut 32P-labeled pBR322 served as size standards. For EMSA, double-strand oligonucleotides were 32P-5′-end-labeled with polynucleotide kinase and then filled in using unlabeled dNTPs. Competitor oligonucleotides were blunt-ended, but not labeled. The indicated units of GST-Dpn bHLH were incubated for 30 min with 5× 104 cpm probe and then electrophoresed on pre-run 0.25X TBE/4% polyacrylamide gels at 21°C. In competition experiments, unlabeled probes were added immediately after labeled probes. Probe sequences are listed in Table 1.

Table 1
SxlPe oligonucleotides tested for Dpn DNA-binding in EMSA

Cell culture, immunohistochemistry and in situ hybridization

Cultivation, transfection, and assay of Schneider L2 cells were according to (Han et al., 1989). One µg of DNA was used per plate and included: 0.1 µg of firefly luciferase Dpn-binding site reporter, 0.1 µg actin5Cp-dpn-VP16 expression construct, 0.1 µg of SV40 -Renilla luciferase reporter to control for transfection efficiency (pRL-SV40 Promega), and carrier DNA. Luciferase activity was determined using a Dual-Luciferase assay kit (Promega) and a Berthold Lumat LB9501 luminometer.

Embryos were prepared for immunocytochemistry according to (Patel, 1994). Anti-Sxl mouse antibody was used as described (Erickson and Quintero, 2007). All embryos were stained with DAPI to visualize DNA and mounted in 70% glycerol. In situ hybridization was done using standard procedures including NBT/BCIP staining (Lehmann and Tautz, 1994). Briefly, digoxigenin-labeled RNA probes complementary to Sxl exon E1, or lacZ sequences were prepared using in vitro transcription of plasmid or PCR-derived templates (Avila and Erickson, 2007; Erickson and Quintero, 2007). Sxl exon E1 probes detect both SxlPe-derived mRNA and Pe-derived nascent transcripts, the later visible as dots of staining within nuclei (Shermoen and O'Farrell., 1991; Erickson and Cline, 1993; Barbash and Cline, 1995; Erickson and Cline, 1998; Erickson and Quintero, 2007). For X-linked genes, or transgenes, the number of nuclear dots corresponds to the number of X chromosomes. Embryo cell cycles were determined by nuclear density (Foe et al., 1993). Nuclei change in appearance through the cell cycle and we used this to closely stage embryos in cycles 11–13 (Edgar et al., 1994). Times within cycle 14 were estimated by nuclear shape and length, and by the extent of membrane furrow invagination (Foe et al., 1993; Grosshans et al., 2003). In wild-type females SxlPe expression begins during cycle 12. In typical embryo collections, only one quarter of cycle 12 embryos (one half of XX embryos) express Sxl and many of those express in a mosaic pattern with individual nuclei exhibiting one, two, or no nuclear dots, reflecting stochastic activation of the promoter during cycle 12 (Erickson and Cline, 1998; Erickson and Quintero, 2007). For heymat- we observed that 10/21 cycle 12 embryos exhibited Sxl staining from both X chromosomes in most, or all, nuclei. The number of Sxl-expressing cycle 12 heymat- embryos was not significantly different from wild-type (expected 5–6 with expression in some nuclei), but was consistent with our qualitative assessments of elevated staining levels in heymat- XX embryos, and is thus suggestive of a repressive effect of maternal Hey on SxlPe activation in females.

Fly culture and genetics

Flies were grown on standard medium in uncrowded conditions at 25°C. Mutations and chromosomes are described: http://flybase.bio.indiana.edu. Null alleles used: Δdpn2 (Df(2R)dpn-2) (Barbash and Cline, 1995), groE48 (Jennings et al., 2006), Df(3R)E(spl)P11, E(spl), HLHmγ, HLHmβ, HLHm3, HLHm5, HLHm7, HLHm8 (Nagel et al., 2004), and Df(2L)Exel6042, Side. The P(Bac) insertion allele heyf06656 is homozygous lethal but may retain partial function. Hey is located at position 44A2 on chromosome 2R. The FRT42B Heyf06656 chromosome was made by selecting P{FRT(whs)}G13 L+ recombinant progeny of + PBac{w+mC}Heyf06656 +/ P{FRT(whs)}G13 + L females and screening for rare flies with slightly darker eye color than the P{FRT(whs)}G13 parent. Darkest eyed flies were confirmed to carry P{FRT(whs)}G13 PBac{w+mC}Heyf06656. Germline clones (Chou and Perrimon, 1996) were generated following heat treatment of female larvae of the following genotypes: P{hsFLP}1, y1 w1118/ w1118; P{neoFRT}82B ry506 groE48/P{neoFRT}82B P{ovoD1-18}3R and P{hsFLP}1, y1 w1118/ w1118; P{neoFRT}82B Df(3R)E(spl)P11/P{neoFRT}82B P{ovoD1-18}3R and P{hsFLP}12, y1 w/y w1118; P{FRT(whs)}G13 hey/P{FRT(whs)}G13 P{ovoD1-18}2R. Females bearing recombinant germlines were crossed to w1118/Y or Sxlf1 males and their gromat- or heymat- progeny analyzed. Crosses between gromat- females and males with the deletion allele Sxlf7bO produced too few embryos for analysis. The hb-hairyen transgene (Jimenez et al., 1997) was generously provided by G. Jimenez (IBMB-CSIC-PCB, Barcelona), dpn alleles were from T. Cline (University of California, Berkeley), E(spl)P11 was a gift of A. Preiss (University of Hohenheim), FRT82B groE48 was provided by P. Simpson (University of Cambridge). Other fly stocks, including those used for FLP/FRT recombination, were provided by the Bloomington Drosophila stock center.


Dpn binds canonical and non-canonical sites at SxlPe

To identify Dpn-binding sites at SxlPe we expressed a full length Dpn-maltose-binding protein fusion and used the pure MBP-Dpn to DNase I footprint the 1.4 kb region of SxlPe sufficient to confer high-level female-specific expression (Estes et al., 1995). We found three protected regions in the proximal 400 bp of SxlPe (Fig. 1A). One region was centered on the two canonical Dpn-binding sites located at −110 and −121 bp (Hoshijima et al., 1995; Winston et al., 1999). The other protected regions were centered at −160 and −330 bp where no sequences match identified HES protein-binding sites (Fig. 2) suggesting that Dpn, like the bHLH activator Sc/Da (Yang et al., 2001), binds non-canonical sites at SxlPe.

Figure 1
Binding of Dpn to canonical and non-canonical DNA sequences at SxlPe
Figure 2
Location of protein-binding sites at SxlPe

To identify the non-canonical sequences mediating Dpn binding, we carried out of a series of gel-mobility shift assays using a purified 6X His-tagged Dpn bHLH domain fusion protein. Oligonucleotides containing the previously characterized tandem sites 1 and 2 produced two gel-shifted complexes corresponding to dimeric and tetrameric Dpn:DNA complexes (Winston et al., 1999) and mutations in the site 1 and 2 core sequences eliminated Dpn binding (Fig. 1B, Table 1). Consistent with the quantitative study of (Winston et al., 1999), we found no evidence for cooperative binding to tandem sites 1 and 2 by the Dpn bHLH domain. To determine the sequences of Dpn-binding sites 3 and 4, we examined a series of overlapping oligonucleotides for their ability to bind the Dpn bHLH domain (Table 1). We found that Dpn bound to oligonucleotides 3 and 3C containing the sequence CACACT but not to the similar fragment 3Cm carrying the single base change CACcCT (Fig. 1B, Table 1). Similarly, we found that Dpn bound to oligos 4 and 4C but not to 4L or 4R suggesting that the central CACGTT sequence is the core sequence for Dpn site 4. Consistent with this inference, mutations that changed the sequence to CAaGcT prevented Dpn binding in the gel-shift assay (Fig. 1C, Table 1). The distal portion of SxlPe has a second CACGTT sequence at −1006. We found that the Dpn bHLH protein bound an oligonucleotide containing this distal site further supporting our conclusion that CACGTT is a Dpn-binding site (Table 1). The distal site 5 was likely missed in our footprinting assays because it was too close to the ends of the probes.

The three different core sequences exhibited a range of Dpn-binding affinities in the DNase I protection experiments. Consensus sites 1 and 2 were always protected at lower Dpn concentrations than was site 3. Dpn-binding site 3 in turn, was protected by lower Dpn concentrations than was site 4 suggesting that the overall binding affinities are sites 1, 2 > 3 > 4. To further test the relative binding affinities of the Dpn sites, we performed DNA binding competition experiments. We found that Dpn could be competed off the tandem consensus sites 1 and 2 by oligonucleotides containing single sites 1, 3, or 4, but not by a mutant site 1 sequence (Fig. 1C and unpublished data). Based on the footprinting and gel-shift data we estimate that Dpn-binding sites 1 and 2 are bound with approximately four-fold greater affinity than is site 3, which in turn is bound two to five times more tightly than site 4 (Fig. 1 and unpublished data). Binding to the non-canonical site 3 and site 4 sequences is not specific to Dpn, because the related protein Side (CG10446), when used at the same concentration as Dpn, bound the same sequences with similar relative affinities (unpublished data).

In vitro defined Dpn sites bind HES proteins in vivo

We employed three assays to determine whether the Dpn-binding sites we identified in vitro can be recognized by Dpn or related HES proteins in vivo. First, we asked whether Dpn could bind artificial promoters carrying multimers of the predicted Dpn-binding sites in cultured cells. Next, we asked if ectopic hairy protein could bind the predicted sites in embryos, and finally, we asked whether the predicted Dpn-binding sites mediated repression of SxlPe-lacZ reporters in otherwise normal embryos.

To analyze Dpn binding in Schneider L2 cells, we created an activator form of Dpn containing the Dpn bHLH domain fused to the VP16 activation domain (Jimenez et al., 1999) and assayed for the ability of Dpn-VP16 to activate transcription from promoters carrying four tandem copies of the predicted Dpn-binding sites (Fig. 3A). When Dpn-VP16 was expressed from the Actin5C promoter it stimulated transcription from a luciferase reporter plasmid carrying four copies of the canonical CACGCG core sequence upstream of the otherwise inactive minimal SxlPe promoter. Plasmids carrying four copies of the site 3 CACACT or site 4 CACGTT core sequences supported levels of Dpn-VP16 activated transcription nearly equivalent to those seen with consensus sites. Point mutations in sites 3 and 4 blocked activation, confirming that these non-canonical sequences can mediate Dpn-binding in cultured cells.

Figure 3
Canonical and non-canonical DNA sequences mediate HES protein-binding at SxlPe

To determine if the Dpn-binding sites can mediate HES protein-binding and transcriptional repression in embryos, we created a series of transgenic 1.4 SxlPe-lacZ reporters carrying mutations in the predicted Dpn-binding sites and assayed their effects in vivo. We first asked if the reporters could mediate repression by an ectopically expressed version of Hairy that carries the Gro-interacting repression domain from the engrailed protein (Jimenez et al., 1999). In this assay, first employed on endogenous Sxl (Parkhurst et al., 1990), zygotic expression of Hairy-Engrailed from the anteriorly-expressed hunchback promoter causes anterior-specific repression of target genes carrying HES protein-binding sites. We found that Hairy-En repressed SxlPe-lacZ even when both canonical Dpn-binding sites 1 and 2 were mutated (12) although the degree of repression was less than seen with wild-type SxlPe-lacZ fusions (Fig. 3B). These findings indicate that Dpn-binding sites 1 and 2 bind Hairy-En, but also suggest that other, non-canonical, sequences can mediate Hairy DNA-binding in vivo. Those non-canonical sites appear to be at least one of Dpn-binding sites 3 and 4, because the (34) SxlPe-lacZ transgenes were also less effectively repressed by Hairy-En than was wild-type SxlPe-lacZ, and because mutations in all four Dpn-binding sites (1234) eliminated nearly all Hairy-En mediated repression (Fig.3B).

As a third test of the functions of the predicted Dpn-binding sites, we asked whether mutations affecting individual or multiple sites increased expression from SxlPe-lacZ transgenes, as would be expected if the sites normally mediate repression by Dpn or other HES proteins. We focused on male embryos because they do not express detectable cytoplasmic lacZ mRNA from wild-type SxlPe-lacZ transgenes (Estes et al., 1995; Bosch et al., 2006; Avila and Erickson, 2007). We found that mutations affecting Dpn-binding sites 1, 2, or 3, led to ectopic SxlPe-lacZ expression in male embryos (Fig. 3C), confirming that these three sites mediate repressor-binding at SxlPe. A Dpn-binding site 4 mutation, in contrast, did not cause ectopic SxlPe-lacZ expression in males (Table 2), suggesting that the weakest in vitro Dpn-binding sites may not mediate repression in vivo. In the following sections we explore the function of Dpn-binding sites 1, 2, and 3 in relation to the actions of Dpn and other HES proteins, as well as those of the corepressor, Gro, in the sex-specific regulation of SxlPe.

Table 2
Summary of expression of 1.4 kb SxlPe-lacZ transgene lines

The corepressor Gro is a potent negative regulator of SxlPe

Maternally supplied Gro interacts with several different types of DNA-binding proteins, including Hairy and Dpn, to repress transcription in the early embryo (Jimenez et al., 1997; Fisher and Caudy, 1998; Chen and Courey, 2000; Buscarlet and Stifani, 2007). Paroush et al., (1994) identified gro as a negative regulator of Sxl, reporting that loss of maternal gro function caused strong ectopic activation of Sxl in males that rendered male and female embryos indistinguishable with respect to Sxl protein levels. Because equality of Sxl expression between the sexes would have important implications for the mechanism of X chromosome counting, as well as for maintenance expression of this X-linked regulator of dosage compensation, we examined the effects of maternal gro on both SxlPe activity and on Sxl protein levels. Staining with anti-Sxl antibody confirmed that XY embryos derived from mothers with groE48 germline clones (hereafter gromat-), express Sxl protein in most or all cells, but also revealed, contrary to the initial report, that Sxl levels were higher in XX than in XY embryos at all stages (Fig. 4). The observed sex differences in Sxl staining could not be accounted for by gene copy number as Sxlf1/Sxl+ females carrying only one functional Sxl allele still stained more darkly than their Sxl+/Y brothers (Fig. 4). We found similar effects on SxlPe-derived mRNA, with females always staining more intensely than males (Fig. 5), demonstrating that Sxl retains some ability to differentiate between male or female XSE gene doses in the absence of maternal gro.

Figure 4
Sxl protein in gromat- embryos. Embryos from mothers bearing groE48 germline clones were immunostained stained for Sxl. Embryonic stages are mid-cellularization (left) and gastrulation (right). (Top panels) XX and XY embryos bearing normal doses of the ...
Figure 5
Time course of SxlPe activation in wild-type, Δdpn2, and maternal groE48 mutant embryos

gro has a stronger and earlier effect on SxlPe than does dpn

Loss of maternal gro raises ectopic male Sxl protein levels well above those present in dpn mutants (Younger-Shepherd et al., 1992; Paroush et al., 1994; Barbash and Cline, 1995; Fig. 4 and unpublished data). To understand the differential effects of gro and dpn on Sxl transcription, we asked how elimination of maternal gro and zygotic dpn functions altered the timing of SxlPe activation and the levels of mRNA using in situ hybridization to measure nascent and mature Sxl transcripts (Fig. 5).

In wild-type XX embryos, SxlPe is expressed from nuclear cycle 12 through the first minutes of cycle 14 (Barbash and Cline, 1995; Avila and Erickson, 2007; Erickson and Quintero, 2007). In normal XY embryos the promoter remains silent. We observed that the Δdpn2 deletion had no detectable effect on SxlPe activity in XX embryos, but that the dpn deletion caused sporadic and weak ectopic Sxl expression in XY embryos beginning in cycle 13 (Fig. 5; Barbash and Cline, 1995). During early cycle 14, SxlPe became active in more XY nuclei, but this caused only a modest and non-uniform accumulation of SxlPe-derived mRNA, consistent with the low-level accumulation of ectopic SXL (Barbash and Cline, 1995).

In contrast, loss of maternal gro function caused earlier, stronger, and more uniform effects on SxlPe than did deletion of dpn (Fig. 5, Table 2). We observed ectopic Sxl expression in many nuclei in cycle 11 XX embryos and in occasional nuclei in cycle 12 XY embryos. Every XX gromat- nucleus expressed SxlPe throughout cycle 12, and every XY nucleus expressed SxlPe by the end of cycle 13. As a consequence, Sxl mRNA was present at relatively low, but uniform, levels in XY embryos and at slightly elevated levels in XX females. Nacent SxlPe transcripts were detected until about 15 min into cycle 14 in both sexes suggesting that maternal gro does not significantly affect the timing of the shut-off of SxlPe.

The finding that maternal gro has stronger and earlier effects on SxlPe than does dpn could be explained in several ways: by the involvement of additional HES-related proteins, by the involvement of yet other types of Gro-interacting proteins, or by indirect effects of the pleiotropic gro gene in the germline or early zygote. One way to distinguish between these possibilities is to ask what effects mutations in the Dpn-binding sites have on SxlPe activity. If Dpn is the only HES-type repressor to regulate SxlPe, or if Gro acts indirectly, then the effects of mutations in the Dpn-binding sites should equal those of dpn null alleles. On the other hand, if additional HES proteins repress SxlPe, the cis-acting changes should exert a stronger effect than dpn mutations because they would block the actions of all repressors utilizing those DNA-binding sites.

Dpn-binding site mutations affect SxlPe more than the loss of dpn protein

Comparison of the male embryos carrying Dpn-site mutant SxlPe-lacZ reporters shown in Fig. 3C with the ectopic expression of endogenous Sxl in the Δdpn2 male in Fig. 5 immediately suggests that the cis-acting binding site mutations have stronger effects on SxlPe than does loss of dpn. However, this simple comparison is potentially misleading because wild-type 1.4 kb SxlPe-lacZ transgenes do not precisely mimic the normal promoter. Specifically, wild-type SxlPe-lacZ transgenes exhibit low-level activation in XY embryos and are expressed earlier in XX embryos than is endogenous SxlPe (Bosch et al., 2006)Table 2). To determine if the relatively strong lacZ expression from the Dpn-binding site mutant transgenes implicated other bHLH repressors in Sxl regulation, or if it was instead caused by the loss of Dpn-binding to already derepressed transgenes, we compared the effects of the Δdpn2 mutation on SxlPe-lacZ expression with those of the Dpn-binding site mutations. We found, that while Δdpn2 elevated expression of a typical SxlPe-lacZ reporter more than it did the endogenous Sxl locus, the effects of most Dpn-binding site mutations were stronger still. The 1, 2, 1 2, and 123 4 Dpn-site mutations caused SxlPe-lacZ to be expressed in more nuclei at earlier times and at higher overall levels than did Δdpn2 (Table 2). Transgenes carrying Dpn-binding site 3 and 34 mutations expressed ectopic lacZ at levels and times similar to the wild-type SxlPe-lacZ reporter in Δdpn2 mutants. This could indicate that Dpn binds only to site 3, but we favor the simpler idea that this non-canonical sequence is less effective at mediating repression than are sites 1 and 2.

A search for other bHLH repressors of SxlPe

Our findings that mutations in the cis-acting Dpn-binding sites led to earlier and higher levels of SxlPe-lacZ expression than did loss of dpn protein, suggests that other bHLH proteins bind these sequences to repress SxlPe. We used a genetic approach to identify the missing proteins by examining mutants with defects in known, or predicted, bHLH repressors for alterations in SxlPe expression (Moore et al., 2000; Ledent and Vervoort, 2001).

We began with E(spl)m3, a maternally supplied HES-family repressor previously cited as a negative regulator of SxlPe (Dawson et al., 1995; Poortinga et al., 1998). We found that embryos derived from mothers whose germlines lacked E(spl)m3, expressed SxlPe in a completely wild-type pattern (data not shown). There was no ectopic activation of SxlPe in XY embryos, and XX embryos expressed SxlPe at normal levels with normal timing. Homozygous mutant embryos were also wild-type for Sxl expression indicating that any zygotically expressed E(spl)m3 was without effect on SxlPe. The deletion allele we used, Df(3R)E(spl)P11, also removes the E(spl), E(spl)mγ, E(spl)mβ, E(spl)m5, E(spl)m7, and E(spl)m8 loci (Nagel et al., 2004), eliminating seven HES proteins as maternal or zygotic regulators of SxlPe. The protein most similar to Dpn is Side (CG10446) (Moore et al., 2000). We examined several Side deletion mutants for dominant maternal and recessive zygotic effects on SxlPe, but found none, consistent with reports that Side is not expressed maternally, or in the early embryo (Tomancak et al., 2002; Chintapalli et al., 2007; but see Moore et al., 2000). We did not analyze Side for recessive maternal effects because we expected the relatively large Side deletions to be cell lethal in germline clones.

Maternal Hey negatively regulates SxlPe

The Hey gene encodes a protein related to Dpn, Hairy, and E(spl), but which lacks the characteristic C-terminal Gro-binding WRPW motif (Kokubo et al., 1999; Leimeister et al., 1999). Instead, Hey and its mammalian homologs posses a YRPW motif that appears not to interact with Gro/TLE proteins (Davis and Turner, 2001; Iso et al., 2001; Fischer and Gessler, 2007; Kageyama et al., 2007). Nonetheless, Hey proteins can potentially interact with Gro as they form heterodimers with several different HES proteins, including Dpn (Iso et al., 2001; Giot et al., 2003; Chintapalli et al., 2007). The resulting Hey/HES heterodimers appear to bind DNA with higher affinity than the individual homodimers (Iso et al., 2001). The single available mutation, Heyf06656, is a recessive lethal caused by a P(Bac) insertion in the 1st intron. To examine the effects of Hey on SxlPe, we recombined Heyf06656 onto an FRT-containing chromosome and generated Heyf06656 germline clones (Chou and Perrimon, 1996). We found that 100% of Heymat- XY progeny expressed SxlPe ectopically during cycles 13 and 14 but that Sxl expression was spatially variable, with about half the nuclei in each XY embryo expressing SxlPe (Fig. 6). There was no observable accumulation of Sxl mRNA in XY embryos, consistent with the lack of a dominant maternal effect on male viability. SxlPe activity also appeared to be affected in XX Heymat- progeny as we noticed an increase in the proportion of cycle 12 XX embryos that expressed SxlPe, and an increase in the proportion of active nuclei in the expressing embryos (Fig. 6, see Materials and Methods).

Figure 6
Maternal Hey negatively regulates SxlPe

The identification of Hey as a maternally-supplied bHLH repressor of SxlPe, fulfills an important prediction of our experiments: that bHLH repressors in addition to Dpn regulate the on-or-off control of SxlPe. The involvement of maternal Hey and gro are also in keeping with the hypothesis that maternal repressors are integrated parts of the mechanism by which XSE concentrations, rather than X:A ratios, are sensed in the embryo (Cline, 1993; Erickson and Cline, 1993; Barbash and Cline, 1995, Wrischnik et al., 2003; Erickson and Quintero, 2007). Whether the relatively weak effects of Heyf06656 are explained by partial Hey protein function, or whether yet other HES family repressors regulate SxlPe remains to be determined.


SxlPe switches on in females because XX embryos have twice the amount of XSE activators as XY embryos. How this two-fold difference in XSE proteins is converted into an all-or-nothing transcriptional response at SxlPe is the central question in primary sex determination. The traditional concept of the sex determination signal as the X chromosome to autosome ratio, X:A, led to the hypothesis that the male/female difference in XSE proteins is amplified through the actions of inhibitors encoded by autosomal signal elements, or ASEs (see Schutt and Nothiger, 2000; Gilbert, 2006). In this view, Dpn and other ASE proteins amplify the signal by preferentially titrating XSE proteins in XY embryos and by competing with XSE proteins for binding to SxlPe (Parkhurst et al., 1990; Paroush et al., 1994; Schutt and Nothiger, 2000; Louis et al., 2003). An alternative idea, based on the thesis that XSE dose is the sex-determining signal, and on the finding that dpn is the only significant ASE, is that signal-amplification might occur primarily through combinatorial interactions between XSE activators and their maternally-supplied cofactors at SxlPe (Cline, 1993; Erickson and Cline, 1993; Barbash and Cline, 1995; Yang et al., 2001; Wrischnik et al., 2003; Erickson and Quintero, 2007). Repression by DNA-binding proteins is important in combinatorial schemes, but as a kind of fine-tuning control, rather than as the primary cause of dose-sensitivity.

A full understanding of the role of negative regulators in the dose-sensitive control of SxlPe requires the identification and characterization of the cis-regulatory sequences controlling repressor binding as well as the trans-acting factors working through those sites. In the following paragraphs we discuss our findings that Dpn, and other, presumably maternal, bHLH proteins bind SxlPe and act in conjunction with the corepressor, Gro, to define and maintain the threshold concentrations of XSE proteins needed to activate SxlPe. Our data suggest that neither the classical notion of amplification by titration, nor the activator-centered alternative, adequately explain how XSE dose is assessed. Rather they indicate that repression at the level of DNA, or chromatin, is a central aspect of XSE signal amplification. We conclude with a model for how Gro-mediated repression could be modulated by XSE function to generate the dosesensitive control of SxlPe.

Canonical and non-canonical bHLH repressor-binding sites at SxlPe

Although SxlPe has two typical DNA-binding sites for HES family proteins (Hoshijima et al., 1995; Winston et al., 1999), their role in Sxl regulation in vivo had not been examined. Our analysis confirmed that the canonical CACGCG sites centered at -108 and -119, bind HES-family repressors in the embryo, but also revealed that a non-canonical site 3, CACACT, at -160 mediates repression in its normal promoter context. Although CACACT had not been previously reported as a HES-binding site, considerable evidence points to the in vivo importance of DNA-binding sites with less than optimum binding affinity. N-boxes, CACNAG, bind HES proteins with lower affinity than the optimal CACG(T/G)G sequences, but are known to mediate repression of several genes in mammalian cells, and the variant CACGCA appears to bind control repression of Math1 in mice (Iso et al., 2003). The same applies to bHLH activators as illustrated by our finding that the bHLH activator Sc/Da exerts most of its dose-sensitive effects at SxlPe through non-canonical DNA-binding sites (Yang et al., 2001).

Cis-acting mutations implicate additional bHLH repressors in Sxl regulation

We found that mutations in the Dpn-binding sites had stronger and earlier effects on SxlPe activity than did complete loss of dpn function (Table 2). The simplest explanation for this finding is that additional bHLH repressors work through the same sequences as Dpn to control SxlPe. The additional repressors seem likely to be maternally supplied. This argument is based on timing; the cis-acting Dpn-site mutations can affect SxlPe-lacZ expression in XX embryos as early as nuclear cycle 10 or 11, when few zygotic genes are active, and on the results of sensitive and unbiased genome-wide genetic screens that showed dpn to be the only zygotically expressed inhibitor of SxlPe of any significance (Barbash and Cline, 1995; Wrischnik et al., 2003).

Hey is a maternal repressor of SxlPe

The prediction that bHLH repressors other than Dpn regulate SxlPe was confirmed by our discovery that maternal Hey functions as a negative regulator of SxlPe. Befitting its maternal origins, hey acts earlier than dpn, as evidenced by increased Sxl expression in cycle 12 XX embryos and ectopic activation in cycle 13 XY Heymat- mutant embryos. However, Heymat- mutants, unlike dpn embryos, accumulate no detectable Sxl protein in males suggesting either that the single available Hey mutation is not a null allele or that still other bHLH repressors regulate SxlPe. The later possibility is also suggested by the finding that mammalian Hey homologs do not appear to interact directly with Gro/TLE proteins (Iso et al., 2001). One promising candidate bHLH repressor is Her (Hes-related, CG5927). Her protein is encoded on the X chromosome and the gene is maternally expressed (Moore et al., 2000), placing this WRPW-containing HES family member in the correct cellular context to regulate SxlPe. Unfortunately no Her deletions or point mutations are currently available to test its possible function at Sxl.

Gro-dependent repression predominates at SxlPe

The first indication that repression is likely to be a quantitatively important part of primary sex determination was the finding that XY gromat- embryos expressed high-levels of ectopic Sxl protein (Paroush et al., 1994). This initial study of gro and Sxl was limited in scope because X-ray induction of germline clones could generate only a limited number of gromat- embryos. Using high efficiency FLP/FRT-mediated recombination (Chou and Perrimon, 1996) we analyzed in detail the effects of maternal gro on Sxl protein and on SxlPe activity. Our findings confirmed that loss of maternal gro leads to ectopic SXL in XY embryos and showed that this is caused by activation of SxlPe in XY embryos. Our results differed from the initial study in one important respect. Whereas Paroush et al., (1994) reported that SXL levels were indistinguishable in XY and XX gromat embryos, we found that Sxl mRNA and protein were expressed at higher levels in XX embryos at all stages of embryogenesis, even when corrected for the copy number of the X-linked Sxl gene. This has important implications for function, as it means that SxlPe responds differently to the one-X and two-X doses of XSEs even in the absence of gro-mediated repression. The ability of the promoter to distinguish XX from XY is also evident from our finding that SxlPe was always activated at least one cycle earlier in female than in male embryos when repression was compromised or eliminated (Fig. 5 and Fig. 6, Table 2).

The best evidence that the pleiotropic gro protein acts directly at SxlPe, rather than on than other maternal or zygotic genes that influence SxlPe activation, is that maternal groE48 and the 1234 Dpn-binding site mutations have nearly identical effects on SxlPe, eliciting premature activity in XX embryos and ectopic expression in XY cells (Table 2). While the somewhat derepressed state of the 1.4 kb SxlPe-lacZ transgenes prevented precise comparisons, our data suggest that most, if not all, of the repressive effects of maternal gro, and of the cis-acting repressor sites, can be explained by the recruitment of Gro to SxlPe by bHLH proteins. This suggests that several other hypothesized methods of HES-mediated repression, including competition between Dpn and Sc/Da for DNA-binding (Louis et al., 2003), or orange-domain dependent inhibition of Scute function by Dpn (Dawson et al., 1995) are likely to have little quantitative importance at SxlPe, unless such interactions are also directly related to Gro function (see below). The predominant corepressive role of Gro is also consistent with the findings that the corepressors dCtBP and Sir2, which can associate with HES proteins, do not influence Sxl expression (Poortinga et al., 1998; Zhang and Levine, 1999; Astrom et al., 2003).

Inhibition by sequestration of activators?

A means of repression that is independent of Gro and DNA binding is titration, or the sequestration of activators into non-functional heterodimers. Long a staple of models for how the the X:A ratio might be read (see Parkhurst et al., 1990; Schutt and Nothiger, 2000; Gilbert, 2006), titration schemes have found mathematical corroboration (Louis et al., 2003), but little experimental support. To our knowledge, the only evidence for sequestration of an XSE by an ASE protein is a non-reciprocal two-hybrid interaction between Dpn and SisA (Liu and Belote, 1995; Louis et al., 2003)—an interaction that we did not observe with a different two-hybrid system (Fields and Song, 1989; unpublished data). Negative regulation at the level of DNA, in contrast, is supported by the known functions of the proteins, by the initial stochastic activation pattern of each copy of SxlPe (Erickson and Cline, 1998), and by the strong effects of maternal gro and the Dpn-binding site mutations. Nonetheless, our data do leave open the possibility that some XSE signal amplification could occur via sequestration of activators. If so, we suggest that maternally-supplied Emc, the sole example of an inhibitor with a demonstrated ability to heterodimerize with an XSE protein (Campuzano, 2001), is likely the amplifying factor, rather than Dpn or an undiscovered ASE.

Groucho and the control of the SxlPe switch

SxlPe responds to threshold concentrations of XSE activators. Loss of maternal gro, or of Dpn-binding site function, causes premature onset of Sxl transcription in XX embryos and strong ectopic expression in XY embryos. Loss of dpn protein function, in contrast, has virtually no effect on Sxl in females while causing relatively late, and low-level, Sxl expression in males. These findings suggest that Gro and associated maternal repressors directly mediate the initial activation threshold at SxlPe, and that the same factors, plus the ASE protein Dpn, then act to maintain the threshold at appropriate values throughout the X-counting process (see Erickson and Cline, 1993; Barbash and Cline, 1995). An important mechanistic point is that while Gro is not needed for SxlPe to sense male/female differences in XSE doses, it is required to convert the differences into a robust all-or-nothing transcriptional response. How might Gro, acting at the level of DNA, or chromatin, amplify the XSE signal and ensure proper operation of the SxlPe switch?

The predominant model for Gro corepressor function; recruitment to DNA by repressors, oligomerization, spreading, and recruitment of histone deactylases, to generate extended regions of inactive chromatin explains how Gro can function as a dominant long-range repressor (Barolo and Levine, 1997; Chen and Courey, 2000; Martinez and Arnosti, 2008). The notion of potent long-range silencing, however, fits poorly with our understanding of Sxl regulation. First, short-range repression should suffice at SxlPe. The repressor-binding sites are located close to the transcription initiation site, and they can mediate effective repression of Sxl by ectopic derivatives of Hairy that carry Gro-independent ‘short-range’ repression domains (Jimenez et al., 1997). Second, Gro-mediated repression at SxlPe is dynamic, reversible, and relatively weak. Established early in both sexes, repression is overcome in XX embryos during cycle 12. Even in XY embryos, where SxlPe normally remains inactive, loss of dpn function causes a partial reversal of repression during cycles 13 and 14. Transient repression by Gro is not unique to Sxl. As discussed by Jennings et al., (2007), and Martinez and Arnosti, (2008) reversible Gro-mediated local repression is commonly found at loci that are expressed in dynamic developmental contexts suggesting that Gro likely represses transcription by more than one mechanism.

Models for Gro-mediated repression invoking interactions with the mediator complex or RNA polymerase (see Buscarlet and Stifani, 2007) fit better with aspects of Sxl regulation, but, like the dominant-silencing model, do not offer ready explanations for how Gro might control the switch-like response of SxlPe. In contrast, a recent model for Gro function invoking direct associations between Gro and chromatin as a necessary step in repression (Sekiya and Zaret, 2007) appears to be both compatible with transient local repression and suggestive of a means by which Gro might ‘amplify’ the XSE signal.

Sekiya and Zaret’s (2007) key finding was that the mammalian Gro/TLE protein, Grg3, represses transcription by creating a 3 to 4 nucleosome region of poorly accessible chromatin that inhibits binding by transcriptional activators. Surprisingly, Grg3 is not recruited directly by DNA-binding repressors. Instead, Grg3 first associates with chromatin via interactions with histones to form an open nucleosome array. Gro-interacting transcription factors, including the bHLH protein Hes-1, then bind their DNA sites in the array enabling Grg3 recruitment and formation of the repressive chromatin complex (Sekiya and Zaret, 2007). We propose that the requirement that Gro bind nucleosomal histones, combined with Gro’s low affinity for highly acetylated chromatin (Edmondson et al., 1996; Chen and Courey, 2000) provides the elements of a possible feedback mechanism that could work in the early embryo to amplify the female/male difference in XSE proteins into a reliable developmental signal (Fig. 7).

Figure 7
Model for dose-sensitive regulation of SxlPe

A model for Gro-mediated amplification of dose-sensitive signals

The basic tenets of our model for SxlPe regulation are: 1) The initial threshold XSE concentration needed to activate SxlPe is set by the translation products of maternally-supplied gro mRNA acting in conjunction with the products of maternally-supplied mRNAs encoding bHLH repressors. 2) The initial SxlPe activation threshold is crossed first in XX embryos because they possess twice the amount of XSE proteins present in XY embryos. 3) Activation of Sxl transcription leads to acetylation of histones at SxlPe. Histone acetylation decreases the ability of Gro to bind chromatin reducing Gro’s ‘repression potential’ and allowing the XX dose of XSE proteins to more effectively stimulate transcription from SxlPe. 4) In XY embryos, continued translation of maternal mRNAs and the activation of zygotic dpn adjust the SxlPe activation threshold upward so that it remains above the XSE concentrations present in male embryos in cycles 13 and 14 (Fig. 7). The net result is a form of signal amplification via positive feedback. Once initiated in XX embryos, Sxl transcription gains in strength from the interacting effects of rising XSE levels and decreased potential for Gro-mediated repression. The initial failure to activate SxlPe in XY embryos, in contrast, leaves Gro function unabated, so that the single-X dose of XSEs can never exceed the growing SxlPe activation threshold.

Our model for operation of the SxlPe switch, with its emphasis on signal amplification by modulation of corepressor function, is distinct from traditional titration schemes (Parkhurst et al., 1990; Schutt and Nothiger, 2000; Gilbert, 2006), and from composite models invoking titration, DNA-binding site competition, or interactions between multiple activators (Yang et al., 2001; Louis et al., 2003). Its most novel aspect is the feedback mechanism in which high XSE protein concentrations and transcription from SxlPe inhibit Gro function in females. The specific proposal that histone acetylation, occurring as a consequence of transcription and XSE activator binding (reviewed in (Shahbazian and Grunstein, 2007) inhibits Gro-mediated repression is speculative but based on the finding that the yeast Gro/TLE protein Tup1 does not bind highly acetylated histones (Edmondson et al., 1996; see Chen and Courey, 2000; Sekiya and Zaret, 2007). Feedback regulation, however, need not be limited to chromatin modifications. XSE proteins could also decrease the repression potential of Gro by competing with Dpn, Hey, and other repressors for overlapping DNA-binding sites, or by direct interference with Gro or repressor function. The C-terminal VWRPY motif of the XSE protein Runt can interact with Gro, raising the possibility that much of Runt’s positive role at SxlPe is due to its ability to directly antagonize repression (Aronson et al., 1997).

One question our model does not directly address is what prevents stochastic fluctuations in XSE levels from causing stable activation of SxlPe in some XY nuclei? The one nuclear cycle lag in Sxl activation seen in XY compared to XX nuclei when repression is compromised by mutation (Fig. 5 and Fig. 6, Table 2), hints that literal two-fold differences in XSE concentrations may be sufficient to reliably signal an on-or-off response for a limited period of time. The regulatory scheme may also provide a kind of double-check against activation due to random variations in XSE levels. Stable expression of SxlPe would require not only that the promoter be activated, but also that it be turned on at sufficiently high levels to establish the feedback mechanism. XX cells meet both criteria, but the occasional XY nucleus that surpassed threshold XSE levels would likely fail to reinforce the initial event because the single Xs of it and its neighbors would supply insufficient XSE products to do so (Gregor et al., 2007). On the other hand, the discriminatory power of the system would likely be increased by even a small increase in the relative female/male XSE signal prior to the onset of feedback regulation. Plausible early amplification mechanisms include titration of Scute by maternal Emc, and combinatorial effects due to multiple XSE activator-binding sites (Wang et al., 1999; Louis et al., 2003; Veitia, 2003).

Although our focus here is on Sxl, the idea that transcriptional activation could be a kind of feedback control of Gro-activity may be applicable to other genes and systems that respond to small or transient changes in regulatory proteins. As discussed by (Jennings et al., 2007) Gro acts in a dynamic fashion to sharpen spatial expression boundaries during segmentation and to precisely control periodic patterns expression in neuroblast multiplication and during vertebrate somitogenesis. A reversible feedback mechanism relying on general properties of transcriptional activation rather than specific interactions might have considerable evolutionary flexibility.


Our colleague Hong Lu was killed in automobile crash December 1 2002. Her friendship, humanity, and insight are greatly missed. We thank the Bloomington Drosophila stock center, G. Jimenez (IBMB-CSICPCB, Barcelona), T. Cline (University of California, Berkeley), A. Preiss (University of Hohenheim), and P. Simpson (University of Cambridge) for providing fly stocks. K. Maggert and A. Gonzalez provided stimulating discussions and helpful comments on the manuscript. This work was supported in its early stages by American Cancer Society Grant RPG-97-079-01-DB, and later, by National Institutes of Health Grant GM063606 to J.W.E.


Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.


  • Aronson BD, Fisher AL, Blechman K, Caudy M, Gergen JP. Groucho-dependent and -independent repression activities of Runt domain proteins. Mol Cell Biol. 1997;17:5581–5587. [PMC free article] [PubMed]
  • Ashburner M, Golic K, Hawley R. Drosophila A Laboratory Handbook. Cold Spring Harbor, New York: Cold Spring Harbor Press; 2005.
  • Astrom SU, Cline TW, Rine J. The Drosophila melanogaster sir2+ gene is nonessential and has only minor effects on position-effect variegation. Genetics. 2003;163:931–937. [PMC free article] [PubMed]
  • Avila FW, Erickson JW. Drosophila JAK/STAT pathway reveals distinct initiation and reinforcement steps in early transcription of Sxl. Curr Biol. 2007;17:643–648. [PubMed]
  • Barbash DA, Cline TW. Genetic and molecular analysis of the autosomal component of the primary sex determination signal of Drosophila melanogaster. Genetics. 1995;141:1451–1471. [PMC free article] [PubMed]
  • Barolo S, Levine M. hairy mediates dominant repression in the Drosophila embryo. EMBO J. 1997;16:2883–2891. [PMC free article] [PubMed]
  • Barolo S, Posakony JW. Three habits of highly effective signaling pathways: principles of transcriptional control by developmental cell signaling. Genes Dev. 2002;16:1167–1181. [PubMed]
  • Bell LR, Horabin JI, Schedl P, Cline TW. Positive autoregulation of Sex-lethal by alternative splicing maintains the female determined state in Drosophila. Cell. 1991;65:229–239. [PubMed]
  • Bianchi-Frias D, Orian A, Delrow JJ, Vazquez J, Rosales-Nieves AE, et al. Hairy transcriptional repression targets and cofactor recruitment in Drosophila. PLoS Biol. 2004;2:E178. [PMC free article] [PubMed]
  • Bosch JR, Benavides JA, Cline TW. The TAGteam DNA motif controls the timing of Drosophila pre-blastoderm transcription. Development. 2006;133:1967–1977. [PubMed]
  • Buscarlet M, Stifani S. The 'Marx' of Groucho on development and disease. Trends Cell Biol. 2007;17:353–361. [PubMed]
  • Campuzano S. Emc, a negative HLH regulator with multiple functions in Drosophila development. Oncogene. 2001;20:8299–8307. [PubMed]
  • Chen G, Courey AJ. Groucho/TLE family proteins and transcriptional repression. Gene. 2000;249:1–16. [PubMed]
  • Chintapalli VR, Wang J, Dow JA. Using FlyAtlas to identify better Drosophila melanogaster models of human disease. Nat Genet. 2007;39:715–720. [PubMed]
  • Chou TB, Perrimon N. The autosomal FLP-DFS technique for generating germline mosaics in Drosophila melanogaster. Genetics. 1996;144:1673–1679. [PMC free article] [PubMed]
  • Cline TW. Autoregulatory functioning of a Drosophila gene product that establishes and maintains the sexually determined state. Genetics. 1984;107:231–277. [PMC free article] [PubMed]
  • Cline TW. Evidence that sisterless-a and sisterless-b are two of several discrete "numerator elements" of the X:A sex determination signal in Drosophila that switch Sex-lethal between two alternative stable expression states. Genetics. 1988;119:829–862. [PMC free article] [PubMed]
  • Cline TW. The Drosophila sex determination signal: how do flies count to two?. [Review] Trends Gen. 1993;9:385–390. [PubMed]
  • Cline TW, Meyer BJ. Vive la difference: males vs females in flies vs worms. Annu Rev Genet. 1996;30:637–702. [PubMed]
  • Clyde DE, Corado MS, Wu X, Pare A, Papatsenko D, et al. A self-organizing system of repressor gradients establishes segmental complexity in Drosophila. Nature. 2003;426:849–853. [PubMed]
  • Courey AJ, Jia S. Transcriptional repression: the long and the short of it. Genes Dev. 2001;15:2786–2796. [PubMed]
  • Davis RL, Turner DL. Vertebrate hairy and Enhancer of split related proteins: transcriptional repressors regulating cellular differentiation and embryonic patterning. Oncogene. 2001;20:8342–8357. [PubMed]
  • Dawson SR, Turner DL, Weintraub H, Parkhurst SM. Specificity for the hairy/enhancer of split basic helix-loop-helix (bHLH) proteins maps outside the bHLH domain and suggests two separable modes of transcriptional repression. Mol Cell Biol. 1995;15:6923–6931. [PMC free article] [PubMed]
  • De Renzis S, Elemento O, Tavazoie S, Wieschaus EF. Unmasking activation of the zygotic genome using chromosomal deletions in the Drosophila embryo. PLoS Biol. 2007;5:e117. [PMC free article] [PubMed]
  • Edgar BA, Sprenger F, Duronio RJ, Leopold P, O'Farrell PH. Distinct molecular mechanisms regulate cell cycle timing at successive stages of Drosophila embryogenesis. Genes Dev. 1994;8:440–452. [PubMed]
  • Edmondson DG, Smith MM, Roth SY. Repression domain of the yeast global repressor Tup1 interacts directly with histones H3 and H4. Genes Dev. 1996;10:1247–1259. [PubMed]
  • Erickson JW, Cline TW. A bZIP protein, SISTERLESS-A, collaborates with bHLH transcription factors early in Drosophila development to determine sex. Genes Dev. 1993;7:1688–1702. [PubMed]
  • Erickson JW, Cline TW. Key aspects of the primary sex determination mechanism are conserved across the genus Drosophila. Development. 1998;125:3259–3268. [PubMed]
  • Erickson JW, Quintero JJ. Indirect Effects of Ploidy Suggest X Chromosome Dose, not the X:A Ratio, Signals Sex in Drosophila. PLoS Biol. 2007;5:e332. [PMC free article] [PubMed]
  • Estes PA, Keyes LN, Schedl P. Multiple response elements in the Sex-lethal early promoter ensure its female-specific expression pattern. Mol. Cell Biol. 1995;15:904–917. [PMC free article] [PubMed]
  • Fields S, Song O. A novel genetic system to detect protein-protein interactions. Nature. 1989;340:245–246. [PubMed]
  • Fischer A, Gessler M. Delta-Notch--and then? Protein interactions and proposed modes of repression by Hes and Hey bHLH factors. Nucleic Acids Res. 2007;35:4583–4596. [PMC free article] [PubMed]
  • Fisher AL, Caudy M. Groucho proteins: transcriptional corepressors for specific subsets of DNA-binding transcription factors in vertebrates and invertebrates. Genes Dev. 1998;12:1931–1940. [PubMed]
  • Fisher AL, Ohsako S, Caudy M. 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. 1996;16:2670–2677. [PMC free article] [PubMed]
  • Foe VA, Odell GM, Edgar BA. Mitosis and Morphogenesis in the Drosophila embryo: Point and Counterpoint. In: Bate M, Martinez-Arias A, editors. The Development of Drosophila melanogaster. Cold Spring Harbor: Cold Spring Harbor Press; 1993. pp. 149–300.
  • Gilbert S. Developmental Biology. Sunderland, MA: Sinauer Associates, Inc.; 2006.
  • Giot L, Bader JS, Brouwer C, Chaudhuri A, Kuang B, et al. A protein interaction map of Drosophila melanogaster. Science. 2003;302:1727–1736. [PubMed]
  • Gregor T, Tank DW, Wieschaus EF, Bialek W. Probing the limits to positional information. Cell. 2007;130:153–164. [PMC free article] [PubMed]
  • Grosshans J, Muller HA, Wieschaus E. Control of cleavage cycles in Drosophila embryos by fruhstart. Dev Cell. 2003;5:285–294. [PubMed]
  • Han K, Levine MS, Manley JL. Synergistic activation and repression of transcription by Drosophila homeobox proteins. Cell. 1989;56:573–583. [PubMed]
  • Hoshijima K, Kohyama A, Watakabe I, Inoue K, Sakamoto H, et al. Transcriptional regulation of the Sex-lethal gene by helix-loop-helix proteins. Nuc. Acids Res. 1995;23:3441–3448. [PMC free article] [PubMed]
  • Iso T, Kedes L, Hamamori Y. HES and HERP families: multiple effectors of the Notch signaling pathway. J Cell Physiol. 2003;194:237–255. [PubMed]
  • Iso T, Sartorelli V, Poizat C, Iezzi S, Wu HY, et al. HERP, a novel heterodimer partner of HES/E(spl) in Notch signaling. Mol Cell Biol. 2001;21:6080–6089. [PMC free article] [PubMed]
  • Jennings BH, Pickles LM, Wainwright SM, Roe SM, Pearl LH, et al. Molecular recognition of transcriptional repressor motifs by the WD domain of the Groucho/TLE corepressor. Mol Cell. 2006;22:645–655. [PubMed]
  • Jennings BH, Wainwright SM, Ish-Horowicz D. Differential in vivo requirements for oligomerization during Groucho-mediated repression. EMBO Rep. 2007 [PMC free article] [PubMed]
  • Jimenez G, Paroush Z, Ish-Horowicz D. Groucho acts as a corepressor for a subset of negative regulators, including Hairy and Engrailed. Genes Dev. 1997;11:3072–3082. [PMC free article] [PubMed]
  • Jimenez G, Verrijzer CP, Ish-Horowicz D. A conserved motif in goosecoid mediates groucho-dependent repression in Drosophila embryos. Mol Cell Biol. 1999;19:2080–2087. [PMC free article] [PubMed]
  • Kageyama R, Ohtsuka T, Kobayashi T. The Hes gene family: repressors and oscillators that orchestrate embryogenesis. Development. 2007;134:1243–1251. [PubMed]
  • Karandikar UC, Shaffer J, Bishop CP, Bidwai AP. Drosophila CK2 phosphorylates Deadpan, a member of the HES family of basic-helix-loop-helix (bHLH) repressors. Mol Cell Biochem. 2005;274:133–139. [PubMed]
  • Keyes LN, Cline TW, Scheld P. The primary sex determination signal of Drosophila acts at the level of transcription. Cell. 1992;68:993–943. [PubMed]
  • Kokubo H, Lun Y, Johnson RL. Identification and expression of a novel family of bHLH cDNAs related to Drosophila hairy and enhancer of split. Biochem Biophys Res Commun. 1999;260:459–465. [PubMed]
  • Ledent V, Vervoort M. The basic helix-loop-helix protein family: comparative genomics and phylogenetic analysis. Genome Res. 2001;11:754–770. [PMC free article] [PubMed]
  • Lehmann R, Tautz D. In Situ Hybridization to RNA. In: Goldstein LSB, Fyrberg EA, editors. Drosophila melanogaster: Practical Uses in Cell and Molecular Biology. Vol. 44. San Diego: Academic Press; 1994. pp. 576–597.
  • Leimeister C, Externbrink A, Klamt B, Gessler M. Hey genes: a novel subfamily of hairy-and Enhancer of split related genes specifically expressed during mouse embryogenesis. Mech Dev. 1999;85:173–177. [PubMed]
  • Liu Y, Belote JM. Protein-protein interactions among components of the Drosophila primary sex determination signal. Mol Gen Genet. 1995;248:182–189. [PubMed]
  • Louis M, Holm L, Sanchez L, Kaufman M. A theoretical model for the regulation of Sex-lethal, a gene that controls sex determination and dosage compensation in Drosophila melanogaster. Genetics. 2003;165:1355–1384. [PMC free article] [PubMed]
  • Mannervik M, Nibu Y, Zhang H, Levine M. Transcriptional coregulators in development. Science. 1999;284:606–609. [PubMed]
  • Martinez CA, Arnosti DN. Spreading of corepressor linked to action of long-range repressor Hairy. Mol Cell Biol. 2008 [PMC free article] [PubMed]
  • Massari ME, Murre C. Helix-loop-helix proteins: regulators of transcription in eukaryotic organisms. Mol Cell Biol. 2000;20:429–440. [PMC free article] [PubMed]
  • Moore AW, Barbel S, Jan LY, Jan YN. A genomewide survey of basic helix-loop-helix factors in Drosophila. Proc Natl Acad Sci U S A. 2000;97:10436–10441. [PMC free article] [PubMed]
  • Nagel AC, Maier D, Krauss S, Mezger M, Preiss A. Neurogenic phenotypes induced by RNA interference with bHLH genes of the Enhancer of split complex of Drosophila melanogaster. Genesis. 2004;39:105–114. [PubMed]
  • Nagengast AA, Stitzinger SM, Tseng CH, Mount SM, Salz HK. Sex-lethal splicing autoregulation in vivo: interactions between SEX-LETHAL, the U1 snRNP and U2AF underlie male exon skipping. Development. 2003;130:463–471. [PubMed]
  • Ochoa-Espinosa A, Yucel G, Kaplan L, Pare A, Pura N, et al. The role of binding site cluster strength in Bicoid-dependent patterning in Drosophila. Proc Natl Acad Sci U S A. 2005;102:4960–4965. [PMC free article] [PubMed]
  • Ohsako S, Hyer J, Panganiban G, Oliver I, Caudy M. Hairy function as a DNA-binding helix-loop-helix repressor of Drosophila sensory organ formation. Genes Dev. 1994;8:2743–2755. [PubMed]
  • Parkhurst SM, Bopp D, Ish-Horowicz D. X:A ratio, the primary sex determining signal in Drosophila is transduced by helix-loop-helix proteins. Cell. 1990;63:1179–1191. [PubMed]
  • Paroush Z, Finley RLJ, Kidd T, Wainwright SM, Ingham PW, et al. Groucho is required for Drosophila neurogenesis, segmentation, and sex determination and interacts directly with hairy-related bHLH proteins. Cell. 1994;79:805–815. [PubMed]
  • Patel NH. Imaging Neuronal Subsets and Other Cell Types in Whole-Mount Drosophila Embryos and Larvae Using Antibody Probes. In: Goldstein LSB, Fyrberg EA, editors. Methods in Cell Biology. Vol. 44. San Diego: Academic Press; 1994. pp. 446–487. [PubMed]
  • Poortinga G, Watanabe M, Parkhurst SM. Drosophila CtBP: a Hairy-interacting protein required for embryonic segmentation and hairy-mediated transcriptional repression. Embo J. 1998;17:2067–2078. [PMC free article] [PubMed]
  • Schutt C, Nothiger R. Structure, function and evolution of sex-determining systems in Dipteran insects. Development. 2000;127:667–677. [PubMed]
  • Sekiya T, Zaret KS. Repression by Groucho/TLE/Grg proteins: genomic site recruitment generates compacted chromatin in vitro and impairs activator binding in vivo. Mol Cell. 2007;28:291–303. [PMC free article] [PubMed]
  • Shahbazian MD, Grunstein M. Functions of site-specific histone acetylation and deacetylation. Annu Rev Biochem. 2007;76:75–100. [PubMed]
  • Shermoen AW, O'Farrell PH. Progression of the cell cycle through mitosis leads to the abortion of nascent transcripts. Cell. 1991;67:303–310. [PMC free article] [PubMed]
  • Tomancak P, Beaton A, Weiszmann R, Kwan E, Shu S, et al. Systematic determination of patterns of gene expression during Drosophila embryogenesis. Genome Biol. 2002;3 RESEARCH0088. [PMC free article] [PubMed]
  • Van Doren M, Bailey A, Esnayra J, Ede K, Posakony JW. Negative regulation of proneural gene activity: hairy is a direct transcriptional repressor of achaete. Genes Dev. 1994;8:2729–2742. [PubMed]
  • Veitia RA. A sigmoidal transcriptional response: cooperativity, synergy and dosage effects. Biol Rev Camb Philos Soc. 2003;78:149–170. [PubMed]
  • Wang J, Ellwood K, Lehman A, Carey MF, She ZS. A mathematical model for synergistic eukaryotic gene activation. J Mol Biol. 1999;286:315–325. [PubMed]
  • Winston RL, Millar DP, Gottesfeld JM, Kent SB. Characterization of the DNA binding properties of the bHLH domain of Deadpan to single and tandem sites. Biochemistry. 1999;38:5138–5146. [PubMed]
  • Wrischnik LA, Timmer JR, Megna LA, Cline TW. Recruitment of the proneural gene scute to the Drosophila sex-determination pathway. Genetics. 2003;165:2007–2027. [PMC free article] [PubMed]
  • Yang D, Lu H, Hong Y, Jinks TM, Estes PA, et al. Interpretation of X chromosome dose at Sex-lethal requires non-E-box sites for the basic helix-loop-helix proteins SISB and daughterless. Mol Cell Biol. 2001;21:1581–1592. [PMC free article] [PubMed]
  • Younger-Shepherd S, Vaessin H, Bier E, Jan LY, Jan YN. deadpan an essential pan-neural gene encoding an HLH protein, acts as a denominator element in Drosophila sex determination. Cell. 1992;70:911–921. [PubMed]
  • Zhang H, Levine M. Groucho and dCtBP mediate separate pathways of transcriptional repression in the Drosophila embryo. Proc Natl Acad Sci U S A. 1999;96:535–540. [PMC free article] [PubMed]
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...