Transcriptional activation by extradenticle in the Drosophila visceral mesoderm

Developmental Biology 290 (2006) 482 – 494 www.elsevier.com/locate/ydbio

Transcriptional activation by extradenticle in the Drosophila visceral mesoderm Brian G. Stultz a,1, Donald G. Jackson b,1,2, Mark A. Mortin c, Xiang Yang d,3, Philip A. Beachy b, Deborah A. Hursh a,* a

c

Cellular and Tissue Therapy Branch, Center for Biologics Evaluation and Research, Food and Drug Administration, HFM-730, Bldg. 29B, Rm. 1E16, 8800 Rockville Pike, Bethesda, MD 20892, USA b Department of Molecular Biology and Genetics, The Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA Laboratory of Biochemistry, National Cancer Institute and Laboratory of Molecular Genetics, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, MD 20892, USA d Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda, MD 20892, USA Received for publication 22 February 2005, revised 22 November 2005, accepted 22 November 2005 Available online 3 January 2006

Abstract decapentaplegic (dpp) is a direct target of Ultrabithorax (Ubx) in parasegment 7 (PS7) of the embryonic visceral mesoderm. We demonstrate that extradenticle (exd) and homothorax (hth) are also required for dpp expression in this location, as well as in PS3, at the site of the developing gastric caecae. A 420 bp element from dpp contains EXD binding sites necessary for expressing a reporter gene in both these locations. Using a specificity swap, we demonstrate that EXD directly activates this element in vivo. Activation does not require Ubx, demonstrating that EXD can activate transcription independently of homeotic proteins. Restoration is restricted to the domains of endogenous dpp expression, despite ubiquitous expression of altered specificity EXD. We demonstrate that nuclear EXD is more extensively phosphorylated than the cytoplasmic form, suggesting that EXD is a target of signal transduction by protein kinases. Published by Elsevier Inc. Keywords: Drosophila melanogaster; decapentaplegic (dpp); extradenticle (exd); homothorax (hth); Ultrabithorax (Ubx); Transcriptional activation; Visceral mesoderm; Phosphorylation

Introduction The Drosophila homeotic (HOX) genes and their vertebrate homologs regulate morphogenesis (Manak and Scott, 1994; Mann and Morata, 2000; McGinnis and Krumlauf, 1992). These proteins share a conserved DNA binding motif, the 60 amino acid homeodomain, and specify positional information by regulating the transcription of downstream target genes (reviewed in Biggin and McGinnis, 1997; Graba et al., 1997; Mann and Morata, 2000). Relatively few direct targets have

* Corresponding author. Fax: +1 301 827 0449. E-mail address: [email protected] (D.A. Hursh). 1 Co-authors have contributed equally to this work. 2 Current address: Bristol-Meyers Squibb, Pharmaceutical Research Institute, Princeton, NJ 08543, USA. 3 Current address: 57 Timber Rock Dr. Gaithersburg, MD 20878, USA. 0012-1606/$ - see front matter. Published by Elsevier Inc. doi:10.1016/j.ydbio.2005.11.041

been identified; they include signaling molecules (Brodu et al., 2002; Capovilla et al., 1994; Sun et al., 1995), transcription factors (Capovilla et al., 2001; Ryoo and Mann, 1999), and structural proteins (Kremser et al., 1999). The regulatory properties of HOX proteins are more specific than their differences in DNA sequence recognition. The HOX homeodomains all have glutamine at position 50 (Q50) (Burglin, 1994; Hayashi and Scott, 1990), which is an important determinant of sequence specificity. Many, including Ultrabithorax (UBX) and Abdominal-A (ABD-A), bind similar sequences (Appel and Sakonju, 1993; Ekker et al., 1994). In contrast, the non-HOX protein Bicoid (BCD) has a lysine at position 50 (K50) and recognizes a distinct DNA sequence (Treisman et al., 1989). The Drosophila extradenticle (exd) and mammalian Pbx genes encode homeodomain proteins of the PBC family. They are included in the atypical TALE class of homeo-

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

domain proteins (Burglin, 1997). exd is required for the normal expression of many HOX target genes (Chan et al., 1994; Pinsonneault et al., 1997; Popperl et al., 1995; Rauskolb and Wieschaus, 1994). In vitro studies suggest that EXD and PBX affect DNA binding by HOX proteins (Chang et al., 1995; Popperl et al., 1995; Shen et al., 1996; van Dijk and Murre, 1994) via a hexapeptide or PBC interaction domain found in most HOX proteins (In der Rieden et al., 2004; Merabet et al., 2003). EXD binding sites are critical for correct expression of several defined targets (Chan et al., 1994; Gebelein et al., 2002; Grieder et al., 1997; Pinsonneault et al., 1997; Ryoo et al., 1999; Sun et al., 1995). However, other studies suggest that, whereas EXD increases the affinity of HOX proteins for their sites (compared to nonspecific sites), it may not enhance discrimination among different HOX sites (Neuteboom and Murre, 1997). In addition, not all EXD/HOX interactions involve cooperative binding (Pinsonneault et al., 1997). To access the nucleus, PBC family proteins require association with members of another family of homeodomain proteins, the MEIS family (Abu-Shaar et al., 1999; Berthelsen et al., 1999; Jaw et al., 2000). The MEIS family is also within the TALE class and includes the vertebrate MEIS and PREP proteins and the product of the Drosophila homothorax (hth) gene (Burglin, 1997). hth and exd engender virtually identical mutant phenotypes (Azpiazu and Morata, 2002; Kurant et al., 1998; Rieckhof et al., 1997) suggesting that they function together. A requirement for HOX/HTH/EXD trimeric complexes has been demonstrated for several target genes (Ferretti et al., 2000; Gebelein et al., 2002; Jacobs et al., 1999; Ryoo et al., 1999; Shen et al., 1996), suggesting that MEIS/HTH’s role extends beyond nuclear translocation. The Drosophila TGF-h homolog decapentaplegic (dpp) is a well-defined regulatory target for HOX proteins. Expression of dpp is positively regulated by UBX in parasegment 7 (PS7) of the embryonic visceral mesoderm (VM) and repressed in more posterior parasegments by ABD-A. It is also expressed in the gastric caecae at PS3 of the VM. cis-Regulatory elements from the 5V end of the dpp gene reproduce normal expression in the VM. These elements respond to UBX, ABD-A, and EXD and contain binding sites for these proteins (Capovilla and Botas, 1998; Capovilla et al., 1994; Hursh et al., 1993; Manak et al., 1994; Rauskolb and Wieschaus, 1994; Sun et al., 1995). We previously demonstrated that UBX directly regulates dpp in PS7 of the VM using a specificity swap strategy. We mutated subsets of six UBX binding sites in a 420 bp reporter construct (PX) from binding sites for Q50 homeodomains to binding sites for K50 homeodomains. We were then able to restore the expression of these constructs by changing Q50 to K50 in the UBX protein (called UBX K50). However, we were unable to restore expression of a reporter in which all six UBX sites were altered. This suggested that an additional factor was required, and we noted that the alterations in the fully substituted PX reporter also disrupted closely apposed EXD binding sites (Fig. 1), suggesting that UBX and EXD may co-regulate dpp (Sun et al., 1995). We here demonstrate that the 420 bp PX element requires EXD binding sites for expression. We show that the element

483

Fig. 1. EXD and UBX binding sites in the 420 bp dpp PX regulatory element. (A) Four EXD binding sites (e1 – e4) and five HOX binding sites (Site 4 – Site 9) were identified by DNAse I footprinting (Sun et al., 1995). A fifth EXD site (e5) was identified later. Labels above the sequence designate sites on the plus strand; labels below the sequence designate sites on the minus strand. (B) Schematic of wild-type and mutated PX elements used to regulate expression of lacZ reporters and their in vivo expression patterns. HOX binding sites shown in panel A were mutated to TAATCCC, the optimal site for HOX K50, shown as white circles. Unaltered sites are shown as black circles. The expression of these reporter constructs in the absence (w/o) and presence (w) of UBX K50 is also shown (Sun et al., 1995). (C) Correspondence between in vivo and in vitro binding sites. The e2 site is the only match within the PX sequence to either the reported PBX1/HOXB7 optimal site (Chang et al., 1996) or to the PBX1/MEIS1 optimal site (Chang et al., 1997). Matches to the e2 site in wild-type PX and in PX4 – 9 are shown.

contains a site that binds HTH/EXD and also has a genetic requirement for hth. Using a specificity swap, we show that EXD directly regulates dpp in the VM. Simultaneous expression of EXD K50 and UBX K50 restored expression of the fully substituted PX reporter, demonstrating that UBX and EXD act together to directly activate this HOX response element. Restored expression coincided with dpp expression in PS3 and PS7 and with Ubx expression in PS7. EXD K50 expression alone restored the fully substituted construct in PS3

484

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

and PS7, although PS7 expression was not as strong as when both UBX K50 and EXD K50 were present. Furthermore, the complete absence of wild-type UBX did not prohibit restoration by EXD K50 in either PS3 or PS7. These data identify EXD as a major transcriptional activator of dpp expression in PS3 and PS7 of the VM, possibly in concert with HTH. Additionally, these experiments indicate that EXD can activate target gene expression, in vivo, in the absence of a HOX protein both in PS3, where none exist, but additionally in PS7, where UBX is known to be a direct activator of gene expression. Materials and methods Drosophila strains and genetics Fly cultures and crosses were performed according to standard methods at 25-C. The following Drosophila strains were used in this work: hth64-1 , dpp S4 , Ubx 9.22 , UAS Ubx1a, UAS dpp, UAS arm s10, and twist-Gal4; B24-Gal4 (Buff et al., 1998).

Transgenic Drosophila strains and assays Binding sites for UBX (sites 4 – 9) and EXD (sites e1 – e5) within a 420 bp PstI – XhoI (PX) fragment from the shortvein regulatory region of the dpp gene are shown (Fig. 1A). The construction of RD, BE, PX4 – 9, PX4 – 6,9 and PX5,7,9 and their transformation into Drosophila are described elsewhere, as is an hsUBX K50 transgene (Hursh et al., 1993; Sun et al., 1995). Recombinant PCR (Jones and Winistorfer, 1992) was used to alter EXD sites e1 – e5 to GGACCGGTCC to create PXe1 – e5 (Fig. S1). The PX portion of the mutagenized plasmid was subcloned into pCaSpeR-hs43-LacZ and used to create transgenic flies by standard methods. Seven independent lines were examined for this construct. The EXD K50 transgene was constructed by recombinant PCR using an EXD cDNA, subcloned into phsCaSpeR, and used in germline transformation. Induction of hs lines was performed as described (Sun et al., 1995).

Immunohistochemical detection Immunohistochemical detection of proteins was performed as described (Hursh et al., 1993; Sun et al., 1995). The following antibodies were used: hgalactosidase monoclonal antibody at 1:1000 (Gibco), Cappell h-galactosidase antibody at 1:20,000, EXD monoclonal antibody at 1:10 (gift of Rob White), SCR monoclonal antibody at 1:2000 (gift of Jim Mahaffey), rabbit anti-HTH at 1:500 (gift of Adi Salzberg), fasciclin III monoclonal antibody at 1:500 (Developmental Studies Hybridoma Bank). Donkey secondary antibodies directly conjugated to either horseradish peroxidase, Cy2, or Cy5 (Jackson Laboratory) were used at 1:1000. Embryos were examined using DIC with a Zeiss Axiophot or with a Biorad Radiance Confocal Microscope.

Electrophoretic mobility gel shift assay The homeodomains (amino acids 237 – 314) from full-length clones of EXD and EXD K50 were amplified by PCR. These were cloned into pET-28a, expressed in BL21 cells, and protein was purified on Ni2+-resin columns. Fulllength EXD and HTH proteins were expressed as above and isolated in a denaturing guanidinium lysis buffer. Following purification, proteins were refolded by step wise dialysis in 20 mM HEPES (pH 6.9), 50 mM NaCl, 2 mM MgCl2, 10% glycerol containing 4 M urea, then 2 M urea, 1 M urea, and finally 0 M urea. Binding assays for the truncated homeodomain proteins were performed in 100 mM potassium acetate, 10 mM HEPES (pH 7.0), 2 mM MgCl2, 5 mM DTT, 0.1% Triton X-100, 10% glycerol, 50 Ag/ml bovine serum albumin, dIdC, tRNA, and E. coli DNA where indicated. Bound and free probes were separated on 1% agarose 0.5 TBE gels. Full-length protein

binding assays were performed in 10 mM Tris (pH 7.5), 50 mM NaCl, 1 mM MgCl2, 0.5 mM DTT, 0.5 mM EDTA, 4% glycerol, and 0.1% Triton X-100. Bound and free probes were separated on 0.7% agarose 0.5 TBE gels. Radiolabeled double-stranded oligonucleotide probes of the following sites were used in the binding assays: wild-type UBX site 5/EXD site e2 (AGGCCTATCAATTAGCACC, only the plus strands will be indicated), mutant UBX site 5/EXD site e2 (AGGCCTAGGGATTAGCACC), mutant UBX site 8/EXD site e3 (CCAAGTAATCCCTTTGAAT), and a consensus SP1 site (TCACGGGGCGTTACCGAGT) used as a negative control. The mutant sites match compensatory changes made in creating PX4 – 9 from wild-type PX.

Antisera production Antisera were raised against residues 148 – 376 of the EXD protein. Proteins were cloned into pET15b (Novagen), expressed in E. coli strain BL21 (DE3)-pLysS (Studier et al., 1990), and purified according to the manufacturer’s protocol. Proteins were used to immunize rabbits (Hazelton).

Western blot Following heat shock and recovery, identical volumes of embryos were homogenized in 1.5 ml eppendorf tubes in 200 Al of SDS-PAGE running buffer and frozen at 70-C until use. Ten microliters of each sample was boiled 5 min, and their proteins were resolved on 10% acrylamide SDS gels and transferred to nitrocellulose by electroblotting. Protein was recognized with polyclonal EXD antibody at a 1:20,000 dilution and detected using enhanced chemiluminescence (Amersham).

Embryo fractionation Nuclei and cytoplasm of Drosophila embryos were fractionated as described (Kamakaka and Kadonaga, 1994) in the presence of 1 mM DTT, 1 mM Leupeptin, 0.2 mM PMSF, 50 mM NaF, and 1 mM NaVO3. Nuclei were solubilized in RIPA buffer with 50 mM NaF and 1 mM NaVO3.

2-D gel electrophoresis and immunoblotting Embryo or cell samples were prepared by lysis in 50 mM Tris – HCl pH 8.0, 400 mM NaCl, 0.1% SDS, 0.5% Sodium deoxycholate, 1% Triton X-100 with 50 mM NaF, and 1 mM NaVO3 followed by sonication. Nucleic acids were removed by precipitation with polyethylenimine. In vitro translations were performed using the TNT coupled transcription/translation system in the presence of 35S-Methionine according to the manufacturer’s instructions (Promega). Tube gel isoelectric focusing was done in 9.2 M urea/3.8% acrylamide gels with 0.4% (v/v) Bio-lyte pH 3 – 10 ampholytes, 0.1% (v/v) Bio-lyte pH 4 – 7 ampholytes, 1.5% (w/v) CHAPS, and 0.5% (v/v) Nonidet P40. Samples were precipitated using acetone and resuspended at 1 mg/ml in 4% (w/v) CHAPS, 0.5% (v/v) Bio-lyte pH 3 – 10 ampholytes, 0.05% SDS, and 5 M urea with 65 mM DTT. Samples were incubated at 30-C for 15 min prior to loading. Focusing was performed for 2700 V h. Immobilized pH gradient (IPG) focusing was performed using a Bio-Rad Protean IPG system with 7 cM pH 5 – 8 linear pH gradient strips according to the manufacturer’s recommendations. Second dimension electrophoresis was performed on 10% SDSpolyacrylamide gels followed by electroblotting to nitrocellulose. Anti-EXD serum (against residues 148 – 376) was used at a 1:2500 dilution in blocking buffer followed by HRP anti-rabbit IgG at 1:10,000 in blocking buffer. Bands were visualized by chemiluminescence using the SSCL substrate (Pierce).

Results dpp requires exd and hth for expression in the VM We previously defined a 420 bp minimal homeotic response element within the 5V cis-regulatory DNA of the dpp gene (the PX element in Fig. 1). We identified five EXD binding sites

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

within this element by DNase I protection assays [Sun et al., 1995; data not shown). To test whether these sites are required for PX expression, we created a construct with all five sites mutated, PXe1 – e5 (Fig. S1), and examined h-galactosidase expression driven by this construct. We never detected mesodermal expression from this construct at any stage examined, even after extensive staining times (Fig. 2A, bottom). An unsubstituted PX construct, stained in parallel, is shown for comparison (Fig. 2A, top). These data show that the EXD binding sites are absolutely required for expression of the PX construct. EXD binding sites have also been shown to be required for full DPP VM expression in other dpp constructs (Chan et al., 1994; Manak et al., 1994). Since exd may function in concert with hth, we next asked whether expression from PX requires hth. There is a potential EXD/HTH half site (Fig. 1C and see below) and additional putative DNA binding sites for HTH in the PX element. In hth mutant embryos, PX expression was missing in PS3 and substantially reduced in PS7 (Fig. 2B, bottom). In later embryos, no midgut constrictions were visible, and the gastric caecae were absent (data not shown). Fig. 2B (top) also shows a wild-type PX-bearing segregant stained in parallel for comparison. Similar mutant phenotypes and dpp expression results were seen using germline clones to produce exd

485

maternal and zygotic nulls and examining a reporter construct related in size and sequence to PX (Rauskolb and Wieschaus, 1994), suggesting that hth and exd act in concert in the VM to activate dpp in PS3 and PS7 through sites resident in the PX construct. It was also reported that expression from larger dpp cisregulatory constructs was continuous from PS3 to PS7 in exd null embryos (Rauskolb and Wieschaus, 1994). We also observed this ectopic expression in hth mutant embryos from both an 812 bp construct (BE; Sun et al., 1995) that contains PX (data not shown) and an 8.9 kb construct (RD2; Hursh et al., 1993) that also contains PX (Fig. 2C, bottom). Fig. 2C (top) also shows a wild-type segregant stained in parallel. Notice that the normal expression of the dpp reporter construct in PS3 and PS7 appears to be reduced in the homozygous hth mutant, while ectopic expression extends between them. Thus, sites not contained within the PX construct also respond to exd/hth to repress dpp. This indicates that exd and hth act on dpp both to activate gene expression in PS3 and PS7 and to repress gene expression in the parasegments in between. We do not know if exd and hth’s repressive effect on dpp is direct or indirect, unlike the activation function of EXD, which appears to be direct (see below), however, it is clear that no sites within the 420 bp PX construct are responsible for exd/hth-mediated repression.

Fig. 2. Expression of PX requires EXD and hth. (A) h-galactosidase expression in stage 14 embryos for wild-type PX (top) and PXe1 – e5 (bottom), in which all five EXD sites have been mutated. VM expression is never seen with the PXe1 – e5 construct. The PX embryo was stained in parallel to PXe1 – e5 and is thus overstained. (B) h-galactosidase expression from the PX construct in a wild-type stage 13 segregant (top) and an hth 64-1 (bottom) mutant embryo. (C) h-galactosidase expression from the 8.9 kb RD construct in wild-type stage 13 segregant (top) and stage 13 hth 64.1 mutant (bottom). The black bar indicates anterior expansion. Detection of hgalactosidase expression in all panels is by immunohistochemistry. Position of PS3 (gastric caecae) and PS7 are indicated by carets and arrowheads, respectively. Anterior is to the left.

486

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

Direct regulation of dpp by EXD To determine if the PX homeotic response element requires direct simultaneous regulation by EXD and UBX, we used a specificity swap strategy. Altering DNA binding sites for several different homeodomain proteins to BCD-like sites abolishes reporter expression. Supplying compensatorily changed proteins that recognized these altered sites restores reporter gene expression (Capovilla and Botas, 1998; Capovilla et al., 2001; Schier and Gehring, 1992). We found that altering the sequence of the five EXD binding sites in PX from the consensus ATCAATC/AAA/T (van Dijk and Murre, 1994) to ATCAATCCC does not affect binding by wild-type EXD protein in vitro (data not shown). Unlike UBX and FTZ, which have glutamine at residue 50, EXD has glycine. This residue does not make significant DNA contacts (Passner et al., 1999; Piper et al., 1999). Consistent with this result, PX constructs with substituted EXD binding sites show normal expression (data not shown). We instead used a mutated PX element designated PX4 – 9. This element has six predicted HOX protein binding sites altered from the wild-type sequences (labeled as sites 4 through 9 in Fig. 1A) to TAATCCC (Fig. S1). The alterations in PX4 – 9 disrupt UBX binding but also affect the 5V and core nucleotides in the closely juxtaposed EXD binding sites e2 and e3. EXD binding to the PX4 – 9 element is greatly reduced in vitro, and the construct fails to express in vivo. Furthermore, expression from PX4 – 9 cannot be restored by ectopically expressing UBX, UBX K50 (Sun et al., 1995), or EXD (Fig. 3E). Many types of homeodomain proteins have been shown to bind to TAATCCC when position 50 of the homeodomain is altered to lysine (Capovilla and Botas, 1998; Capovilla et al., 2001; Hanes and Brent, 1989; Schier and Gehring, 1992; Treisman et al., 1989). We therefore predicted that EXD K50 could bind to TAATCCC sites, and the PX4 – 9-lacZ expression construct was used to test whether EXD is able to directly activate dpp gene expression in the VM. Simultaneous ectopic expression of UBX K50 and EXD K50 restored expression of the PX4 – 9 construct. Restored expression is observed in both PS3 and PS7 of the visceral mesoderm (Fig. 3C). In contrast, the restoration of partially substituted reporters (PX4 – 6,9 and PX5,7,9, see Fig. 1B) by ectopic UBX K50 alone was only seen in PS7, never in PS3 (Sun et al., 1995). This restoration depends upon induction of EXD K50 and UBX K50 as no reporter expression was detected in non-heat-shocked PX4 – 9 embryos (Fig. 3J) and heat-shocked embryos bearing PX4 – 9 but lacking EXD K50 and UBX K50 (Fig. 3F). Notably, restoration only occurs in sites of endogenous dpp gene expression (Fig. 3A). Since endogenous UBX is not expressed in PS3 and UBX K50 alone never restored expression of reporter constructs in this region, we tested whether PS3 expression could be restored solely by EXD K50. Expressing EXD K50 but not wild-type EXD alone restored expression in both PS3 and PS7 (Figs. 3D and E). It was formally possible that the failure of wild-type EXD to restore PX4 – 9 expression resulted from a failure to actually induce wild-type EXD via heat shock. To test for this

possibility, we compared the overall levels of EXD/EXD K50 protein in wild-type EXD; PX4 – 9 and EXD K50; PX4 – 9 embryos with and without heat shock. Immunohistological staining of these embryos demonstrates the induction of both wild-type EXD and EXD K50, following heat shock, to comparable levels (Figs. 3G – J). Protein levels detectable on a Western blot also confirm the induction of these proteins to comparable levels (Fig. 3K). We conclude that the restoration of PX4 – 9 expression in PS3 is specific to EXD K50. The restored expression in PS3 by EXD K50 was typically as strong or stronger than expression in PS7 (compare Figs. 3C and D). This differs from wild-type dpp, which is expressed more strongly in PS7 than in PS3 (Fig. 3A). The restored expression of the PX4 – 9 reporter by simultaneous expression of EXD K50 and UBX K50 (Fig. 3C) produces relative staining intensities in PS3 and PS7 that more closely resemble wildtype dpp and the unaltered PX reporter (Fig. 3B). We interpreted our restoration data as resulting from the action of EXD K50 on the compensatorily changed UBX sites. To verify this interpretation, we performed electrophoretic mobility shift assays on oligonucleotide probes made to match the UBX site 5/EXD site 2 (e2) from both wild-type and PX4 – 9 and the UBX site 8/EXD site 3 (e3) from PX4 – 9. These sites were chosen, as mentioned above, because the compensatory changes introduced into the UBX 5 and 8 sites prevent both wild-type UBX and EXD binding (Sun et al., 1995) while rendering the UBX site a target for the altered homeodomains of UBX K50 and possibly EXD K50. Thus, these two compound binding sites have compensatorily changed UBX sites and no functional EXD sites. We also tested the unaltered, wild-type UBX site 5/EXD site 2. This latter construct contains a functional EXD site and a site that binds wild-type UBX homeodomain, but not that of UBX K50 (Sun et al., 1995). The homeodomains of EXD and EXD K50 were expressed in bacteria, purified, and tested for binding. The EXD K50 homeodomain binds strongly to both oligonucleotide probes containing compensatorily altered UBX sites and no functional EXD sites (Fig. 4A, lanes 3, 4, 7, and 8), indicating that EXD K50 is indeed binding to the compensatorily changed UBX site. EXD K50 was also capable of weakly binding the oligonucleotide probe containing the wild-type UBX site 5/ EXD site 2, which contains both normal EXD and UBX binding sites (Fig. 4A, lanes 11 and 12). EXD K50 was unable to bind an unrelated consensus SP1 binding site (data not shown). We did not observe mobility shifts for any of our probes with wild-type EXD (Fig. 4A), even though approximately equal amounts of EXD and EXD K50 protein were used in the gel shift. Inability of EXD alone to form stable protein/ DNA complexes in mobility shift experiments has been reported for many DNA targets (Ryoo and Mann, 1999; Ryoo et al., 1999; van Dijk and Murre, 1994). These data indicate that EXD K50 binds strongly to the same sites that would bind UBX K50. In addition, EXD K50 binds by itself to unaltered EXD sites, more strongly than wildtype EXD. Thus, we surmise that, in the PX4 – 9 construct, EXD K50 would be capable of binding strongly to the 5 compensa-

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

487

Fig. 3. Restoration of PX4 – 9-lacZ expression is restricted to domains of endogenous dpp expression. (A, B) Expression of dpp in the VM of stage 14 embryos detected by RNA in situ hybridization (A) and unsubstituted PX-lacZ expression (B). dpp accumulates in PS3 at the gastric caecae primordia (carets) and in PS7 at the site of the second midgut constriction (black arrowheads). (C – F) Expression of the PX4 – 9-lacZ reporter gene detected by immunostaining with a h-galactosidase antibody. (C) Induction of both EXD K50 and UBX K50 results in PX4 – 9 reporter expression in stage 14 embryos that closely resembles that of wild-type dpp. (D, E) Reporter gene expression is induced as a result of ectopic EXD K50 (D) but not wild-type EXD (E) expression alone in stage 14 embryos. Expression in PS3 is more intense than in PS7 when EXD K50 is used alone. (F) Control embryos bearing PX4 – 9 reporter, but crossed to yw and heat-shocked in parallel to panels C – E. Heat shocking wild-type EXD and EXD K50 embryos results in the accumulation of EXD protein to approximately equal levels (G – K). (G, I) Wild-type EXD; PX4 – 9 and (H, J) EXD K50; PX4 – 9 stage 13 embryos heat-shocked (G, H) and non-heat-shocked (I, J) stained with anti-EXD and h-galactosidase. Heat-shocked embryos show darker ubiquitous EXD expression than non-heat-shocked embryos. Restoration is difficult to see in panel H, due to presence of VM EXD staining. Anterior is to the left in all panels. (K) Western blot of proteins from equivalent volumes of embryos heat-shocked identically to panels G – J detected by anti-EXD.

torily altered UBX sites and more weakly to the 3 unaltered wild-type EXD sites (Fig. 1A). Ectopic expression of EXD K50 has no effect on the expression of unsubstituted PX (such as an increase in PS3 expression), therefore, it does not appear that EXD K50’s ability to act on wild-type EXD sites is primarily responsible for the homeotic-independent restoration of PX4 – 9. Furthermore, EXD K50 does not activate the PXe1 – e5 construct, indicating that there are no other sites within PX that are sufficient to mediate expression with EXD K50. We thus conclude that EXD K50’s ability to restore expression from the PX4 – 9 construct, both with and without

UBX K50, results from the interaction of EXD K50 on the compensatorily changed UBX sites. The PX enhancer contains a binding site for EXD/HTH Our electrophoretic mobility shift assay using EXD and EXD K50 homeodomains did not address the ability of the PX enhancer element to interact with EXD as EXD does not bind independently in these assays (see above). EXD will form complexes in the presence of HTH, and, given our evidence that hth was genetically required for enhancer activity, we

488

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

and that the compensatory site change introduced in the PX4 – 9 construct destroyed the EXD binding site at this location (Sun et al., 1995). Note that the EXD K50 homeodomain, run in parallel, binds both sites robustly (Fig. 4B, lanes 1 and 10). These data suggest that EXD acts with HTH on the PX enhancer from the dpp gene. Spatial localization of HTH does not restrict PX4 – 9 restoration

Fig. 4. EXD K50 strongly binds altered UBX site 5/EXD site e2 and altered UBX site 8/EXD site e3 from PX4 – 9. (A) Amino acids 237 – 314 of EXD K50 but not EXD bound altered sites UBX 5/EXD e2 and UBX 8/EXD e3, in addition to binding wild-type site UBX 5/EXD e2. Radiolabeled doublestranded DNA corresponding to altered sites UBX 5/EXD e2 (lanes 1 – 4), UBX 8/EXD e3 (lanes 5 – 8), and wild-type site UBX 5/EXD e2 (lanes 9 – 12) was bound to polypeptides and analyzed on 1% agarose nondenaturing gels. Altered sites match sequence from PX4 – 9 (see Fig. S1), while the wild-type site matches PX (Fig. 1A). No polypeptide was added to lanes 1, 5, and 9, EXD to lanes 2, 6, and 10, EXD K50 to lanes 3, 4, 7, 8, 11, and 12. Lanes 4, 8, and 12 are identical to 3, 7, and 11 but with the addition of E. coli DNA. (B) Fulllength EXD and HTH cooperatively bind wild-type but not mutant site UBX 5/ EXD e2. Radiolabeled wild-type (lanes 1 – 5) and mutant (lanes 6 – 10) site UBX 5/EXD e2 were bound to proteins and analyzed on 0.7% agarose nondenaturing gels. No protein was added to lanes 2 and 6, EXD to lanes 3 and 7, HTH to lanes 4 and 8, EXD and HTH to lanes 5 and 9, and the homeodomain of EXD K50 to lanes 1 and 10.

wished to see if EXD/HTH complexes would form on the enhancer. This necessitated using full-length EXD and HTH constructs to provide the necessary protein interaction domains to allow EXD/HTH complex formation (Ryoo et al., 1999). Full-length constructs of EXD and HTH were expressed in bacteria, purified, and tested for binding. As can be seen (Fig. 4B, lane 5), full-length EXD and HTH form a complex on the e2 site of the PX enhancer, while neither proteins produce appreciable complexes independently. We do see weak but reproducible binding of full-length EXD alone under these conditions (Fig. 4B, lane 3). Neither full-length EXD nor HTH alone nor together form complexes on the mutant e2 site found in PX4 – 9 (Fig. 4B, lanes 7 – 9), further supporting our observation that wild-type EXD does not affect this element

Although we expressed the EXD K50 and UBX K50 proteins ubiquitously, the restored expression was spatially restricted. This suggests that another factor (or activity) resident in PS3 and PS7 is also required for restoration. The ability of EXD K50 alone to restore the PX4 – 9 construct indicates that EXD is necessary and sufficient for expression. Endogenous EXD is expressed ubiquitously, but its nuclear import is regulated by HTH; therefore, we asked whether localization of HTH in the VM could be the factor restricting EXD K50 activity. Localization of HTH serves as a marker for nuclear, therefore active, EXD. To distinguish the VM from the closely apposed endoderm, embryos were simultaneously stained for HTH and fasciclin III, which is a marker for the VM (Patel et al., 1987). HTH was ubiquitously expressed throughout the VM (Figs. 5A –D). Patches of greater intensity were regularly seen in PS3 (arrows). This is in agreement with data on EXD (Aspland and White, 1997; Mann and Abu-Shaar, 1996). While increased abundance of EXD and HTH might contribute to PS3 restoration, we saw no such increase in HTH protein in PS7. Therefore, spatial restriction of HTH or nuclear EXD does not easily explain the observed restricted restoration. To ask if DPP signaling had an effect on HTH expression, we examined HTH expression in dpp S4 , a mutation that abolishes all of dpp’s expression in the VM. We used the homeotic protein Sex combs reduced (SCR) in these experiments to monitor both the position of the VM and the genotype. Scr is expressed in PS4 of the VM, and its expression expands anteriorly when dpp fails to be expressed in PS3 (Reuter et al., 1990). Loss of dpp expression reduced HTH expression in PS3 but did not abolish it (Figs. 5G and H), indicating that DPP signaling might influence HTH in PS3. However, we have carried out EXD K50 restoration of PX4 – 9 in the presence of ectopic DPP signaling and/or ectopic Wingless (WG) signaling. None of these treatments affected the spatial restriction of restoration (data not shown). We conclude that factors other than the control of EXD nuclear localization by HTH or signaling by either DPP or WG are responsible for our restricted restoration domains, although our data suggest that DPP signaling may affect the abundance of HTH in PS3 of the VM. UBX protein is not required for restoration The lack of spatial restriction of HTH at PS7 led us to ask if UBX itself could limit restoration to PS7. While wild-type UBX would not be able to act on the PX4 – 9 reporter, as all its UBX sites have been altered, it remained possible that some

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

489

Fig. 5. HTH is expressed throughout the VM; therefore, nuclear localization of EXD does not restrict restoration of PX4 – 9. Localization of HTH and fasciclin III in Stage 13 (A, B) and 14 (C, D) embryos is shown. In these panels, only half of the VM of a given embryo is shown. HTH (green) can be detected throughout the VM (A, C), as indicated by overlapping localization with fasciclin III (red) (B, D). Some regions of the VM are out of the plane of focus in panels A and B. Expression of HTH in wild-type and dpp S4 mutant embryos. In these panels, the anterior portion of both sides of the VM is shown. Elevated levels of HTH (green) expression can be seen in PS3 of the VM in wild-type embryos (E, F). In homozygous dpp S4 mutants, HTH continues to be expressed throughout the VM, but the elevated levels observed in PS3 are absent (G, H). Mutant embryos were identified by expression of SCR (red) in the VM. SCR expression expands anteriorly into PS3 in dpp S4 mutants. Arrows indicate PS3/gastric caecae primordia, bars indicate the normal extent of SCR expression, arrowheads indicate the position of PS7.

additional target genes of UBX was required to act in concert with EXD to activate the reporter in PS7. This putative target gene would by necessity be localized to PS7. To address this question, we carried out restoration experiments, either in the presence of ectopically expressed UBX or in an Ubx null mutation (Ubx 9.22 ). Expressing UBX throughout the mesoderm did not alter the spatial restriction of restoration to PS3 and PS7 (data not shown). Lack of UBX did not inhibit the ability of EXD K50 to activate the PX4 – 9 reporter (Figs. 6B, F), although the intensity of h-galactosidase expression in PS7 is reduced at early stages of the restoration (compare Figs. 6A and B). Again, this restoration requires induction of EXD K50 as nonheat-shocked embryos show no induction (Figs. 6C and D). At late embryonic stages (Stage 16), when maximal restoration is seen, both Ubx 9.22 homozygotes and their wild-type segregants show roughly equivalent amounts of restoration in both PS3 and PS7 (compare Figs. 6E and F). Therefore, an EXD protein of altered specificity and increased ability to bind DNA is capable of activating gene expression without input from HOX proteins, even in locations where such HOX proteins are required for normal gene activation, and therefore wild-type EXD may be capable of contributing to transcriptional activation under its normal binding conditions in vivo. Interestingly, we observed some loss of spatial restriction in Ubx 9.22 homozygotes at late stages. h-galactosidase expression

extends weakly from PS8 to the end of the gut in these embryos, suggesting that UBX may contribute some activity required for spatial control in our restoration experiments (Figs. 6E and F). UBX is required to induce endogenous DPP, which in turn autoregulates itself and indirectly contributes to WG expression in PS8. Lack of DPP may impede restoration in Ubx 9.22 homozygotes, while UBX’s demonstrated ability to communicate to posterior parasegments may exert some control over dpp repression in those locations. Nuclear EXD is more highly phosphorylated than the cytoplasmic protein The observation that our PX restorations were restricted to domains of known WG and DPP signaling activity led us to examine the phosphorylation of EXD protein using 2-D isoelectric focusing/SDS-PAGE. EXD from 9 –12 h embryos migrated as multiple charge isoforms (Fig. 7A). Treatment with bacteriophage lambda protein phosphatase (which dephosphorylates Ser, Thr, and Tyr) converted many of the more acidic species to more basic forms (Fig. 7B), indicating that the charge heterogeneity was at least partly due to phosphorylation. The most basic form of embryonic EXD co-migrated with in vitro translated EXD (Figs. 7C and D), indicating that this form is unphosphorylated.

490

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

Fig. 6. EXD K50 restoration of PX4 – 9 does not absolutely require UBX. (A, B) h-galactosidase expression generated by PX4 – 9, as a result of ectopic EXD K50 expression in either a wild-type stage 14 segregant (A) (identified by ectodermal staining from reporter on the balancer chromosome) or Ubx 9.22 homozygote (B). (C, D) Wild-type stage 14 segregant (C) or Ubx 9.22 homozygote (D) from same stock as used for A, B, but without heat shock. No visceral mesoderm h-galactosidase expression is seen. (E, F) Late stage 16 embryo wild-type segregant showing maximal PX4 – 9 restoration (E and F) late stage 16 Ubx 9.22 homozygote, with missing secondary midgut constriction. Note light ectopic staining seen posterior to PS7 in this genotype. Embryos in panels A, B, E, and F were all stained in same tube. Position of PS3 (gastric caecae) and PS7 are indicated by carets and arrowheads, respectively. Ectopic staining is indicated by a bracket. Anterior is to the left.

To investigate whether a connection exists between the extent of EXD phosphorylation and its subcellular localization, we first compared EXD phosphorylation in embryos of different developmental stages. During the first 3 h of development, EXD protein is mostly cytoplasmic (Aspland and White, 1997) and shows very little phosphorylation (Fig. 7E). By 6 – 9 h, significant amounts of EXD are nuclear and the protein is extensively phosphorylated (Fig. 7F). To confirm the connection between phosphorylation and nuclear localization of EXD, we prepared nuclear and cytoplasmic fractions from 6 – 9 h embryos. Cytoplasmic EXD was less extensively phosphorylated than the nuclear protein (Figs. 7G and H). We were unable to immunoprecipitate EXD using a phosphotyrosine antibody, suggesting that phosphorylation does not involve tyrosine residues (not shown). We also did not observe changes in EXD phosphorylation following ectopic expression of dpp. However, we cannot rule out subtle effects of dpp on EXD phosphorylation. Discussion EXD directly activates dpp in embryonic visceral mesoderm In previous work, we were unable to restore expression of the fully substituted PX4 – 9 reporter using UBX K50 (Sun et al., 1995). We show here that ubiquitous and simultaneous induction of UBX K50 and EXD K50 restores expression of PX4 – 9 in a manner that is similar to wild-type PX (Figs. 3B and C). Induction of EXD K50 alone also restores PX4 – 9 in these domains but changes the balance of staining intensity between them, with PS7 expression appearing less prominent (Fig. 3D).

This reflects a sufficiency of EXD K50 for activation of gene expression at both sites, but with an additional requirement for UBX K50 to achieve wild-type levels in PS7. These experiments identify EXD as a direct activator of dpp’s VM expression in both PS3 and PS7. EXD’s activation of dpp in PS3 and PS7 of the VM does not require HOX proteins No HOX proteins are expressed in PS3 (Bienz, 1994; Reuter et al., 1990). Thus, EXD K50 activates gene expression independently of HOX family proteins in this location. In cases where UBX K50 restored partially substituted PX constructs, restoration was never seen in PS3 (Sun et al., 1995), further indicating that dpp expression here does not require HOX proteins. In addition, in PS7, where it is clearly established that UBX contributes to activation of dpp expression, our results demonstrate that UBX is not absolutely required for EXD K50 to activate transcription. UBX increases the level of dpp expression, as demonstrated by the reduced PS7 expression in our EXD K50-alone restorations, but is not required for EXD function. This point is further reinforced by the ability of EXD K50 to activate PX4 – 9 gene expression, even in Ubx homozygous mutants. On simple EXD binding sites, PBX proteins have not demonstrated transcriptional activation (van Dijk et al., 1993), but our data suggest that EXD can participate in gene activation without a HOX gene. Other unidentified factors in PS3 or PS7 could also be involved, and one candidate would be HTH, which we show is genetically required for dpp’s VM expression and capable of binding to the PX element in concert with EXD. Genetic evidence for the ability of EXD/HTH to act

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

491

Fig. 7. EXD is a phosphoprotein, which is more highly phosphorylated in the nucleus. (A, B) EXD protein from 9 – 12 h embryos has up to 7 charged isoforms (A), which are converted to more basic forms following treatment with bacteriophage lambda protein phosphatase. (C, D) Comparison of EXD from 6 – 9 h embryos (C) with in vitro translated EXD (D). (E, F) EXD phosphorylation changes during embryogenesis. EXD protein from 0- to 3-h-old embryos (E) compared with 6- to 9-hold embryos (F). (G, H) EXD phosphorylation correlates with subcellular localization. Comparison of cytoplasmic (G) and nuclear (H) EXD from 6 – 9 h embryos. An additional spot does not represent an EXD isoform (G, white arrow). Two-dimensional isoelectric focusing/SDS-polyacrylamide gel electrophoresis of EXD proteins was followed by transfer to nitrocellulose and detection with EXD antisera. All gels are oriented with their basic sides to the left. Panels A, B and E – H represent gels whose pH gradients formed during their runs. Panels C and D result from gels using immobilized pH gradient strips for the focusing dimensions. Black arrows (C – H) indicate the IEF origin.

in the absence of HOX proteins has been steadily accumulating (Azpiazu and Morata, 2000; Casares and Mann, 1998; Gonzalez-Crespo and Morata, 1995; Jaw et al., 2000; Merabet et al., 2005), based on mutant phenotypes that cannot be attributed to HOX genes, and both genetic and in vitro data suggest that HTH/MEIS may have transcriptional activation capabilities (Huang et al., 2005; Inbal et al., 2001). Two models for the role of EXD in regulating HOX targets have been proposed. Our data indicate that the PX element is directly regulated by both EXD and UBX, allowing us to evaluate these two models based on our results. The Fcoselective binding_ model proposes that EXD enhances the specificity and affinity of its HOX partner for a DNA binding site (Chan and Mann, 1996; Chan et al., 1994; Chang et al., 1996; van Dijk and Murre, 1994). This model requires that EXD and HOX proteins bind cooperatively as heterodimers to closely spaced EXD and HOX binding sites. This model

predicts that the relative spacing and orientation of PBX/EXD and HOX binding sites must be tightly constrained, as has been shown by in vitro studies (Chan and Mann, 1996; Chang et al., 1996). Although the dpp cis-regulatory PX element contains multiple UBX and EXD sites identified by DNA footprinting (Sun et al., 1995), only site e2 resembles the optimal site for binding by a PBX1/HOXB7 heterodimer (Chang et al., 1996). Even this site is not a perfect match (Fig. 1C), and data in Fig. 4B indicate that this site may be more likely to bind EXD/HTH in vivo. The electrophoretic mobility shift data presented in Fig. 4A demonstrate that EXD K50 can bind TAATCCC sites that replace UBX sites, as well as unaltered EXD sites. Thus, EXD K50 restores PX4 – 9 by binding to some or all of these sites. This demonstrates that an EXD protein altered only in its binding specificity can act in vivo through sites of altered spacing and orientation and is not necessarily constrained to act in close proximity to a HOX protein.

492

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

The second model, Fwidespread binding_, proposes that EXD determines the outcome of HOX protein action (Biggin and McGinnis, 1997; Li et al., 1999). According to this model, either EXD or HOX proteins in isolation can bind DNA and act as transcriptional repressors. When both proteins are present, a complex that activates transcription is formed. For Deformed, EXD activates an otherwise silent transcriptional activation domain within the Deformed protein (Li et al., 1999). The physical association between the proteins stabilizes their binding to DNA, but they do not have to bind as heterodimers (Biggin and McGinnis, 1997; Pinsonneault et al., 1997). This model is more consistent with both the spacing of EXD and HOX sites in the dpp PX element and the apparent flexibility in the location of EXD-responsive sites observed in our experiments. However, this model predicts that independent EXD action is repressive, based on a Deformed-responsive target (Pinsonneault et al., 1997). In contrast, our data indicate that EXD can also activate reporter gene expression without a HOX partner, suggesting that repression is not the default action of EXD in the absence of HOX proteins. We have shown that EXD is a direct activator of dpp expression in the VM. In PS3, the normal action of EXD does not require input from any homeotic protein. In PS7, input from UBX is critical to achieving the correct level of gene expression, but our data do not support a model where UBX is absolutely necessary for transcriptional activation. Our data suggest that EXD can activate transcription in the absence of HOX proteins but that, in many cases, it also collaborates with HOX proteins, allowing the complex to achieve a more robust level of transcriptional activation. The current notion is that EXD is an essential cofactor for homeotic proteins. An equally tenable model for gene activation is that HOX proteins are the cofactors of EXD, imparting additional spatial regulation, site specificity, and activity to this transcriptional regulator. Nuclear EXD protein is extensively phosphorylated The striking restriction of reporter restoration to domains influenced by kinase-mediated signaling pathways led us to examine EXD protein for evidence of phosphorylation. The primary sequence of EXD contains more than 15 potential sites for various protein kinases, including Protein Kinase A (PKA) and Casein Kinase II upstream of its NLS (Pai et al., 1998). Protein kinase action is required for gene activation by PBX proteins in tissue culture cells (Ogo et al., 1995), PKA converts HOX/PBX complexes from repressors to activators on the Hoxb1 autoregulatory element (Saleh et al., 2000), and phosphorylation by PKA induces nuclear import of PBX1 independently of the PBX/MEIS nuclear localization mechanism (Kilstrup-Nielsen et al., 2003). While we are unable to establish a connection between DPP signaling and EXD phosphorylation, nonetheless, EXD clearly exists in multiple phosphoprotein forms, and the increased phosphorylation is clearly correlated to subcellular localization in Drosophila as well. Thus, EXD must be a target of kinase action, although whether this activity is solely required for nuclear translocation or for activity once in the nucleus is unresolved.

dpp’s visceral mesoderm gene expression dpp requires both its own expression and that of wg to achieve normal gene expression in the VM (Hursh et al., 1993; Thuringer and Bienz, 1993; Yang et al., 2000; Yu et al., 1996). These data led us to hypothesize that the spatial restriction observed in our restoration must be connected to DPP or WG signaling. However, our data do not support this hypothesis, and it is more likely that the major inputs generating dpp’s localization in the VM are repressive in nature. In previous work, we postulated that dpp’s spatial regulation in the VM was the result of dual modes of regulation involving both general activation and spatially specific repression and spatially restricted activation (Sun et al., 1995). The general activator has been identified as biniou (bin), a member of the FoxF/ forkhead family of transcription factors (Zaffran et al., 2001). This factor is capable of inducing dpp expression throughout the posterior half the VM, including PS7, when its action is not specifically repressed. This repression comes from multiple inputs. dpp is a direct target of posterior repression via Abd-A (Capovilla and Botas, 1998; Hursh et al., 1993). dpp is also repressed outside of PS3 and PS7 via the action of Drosophila T Cell Factor (dTCF) in the absence of WG signaling (Yang et al., 2000). The ectopic PS4 – 6 expression of longer dpp constructs in exd or hth null embryos discussed above (Fig. 2C) identifies exd and hth or a downstream target of these genes as another repressor of dpp in PS4 – 6. Such a downstream target could be teashirt, a known repressor whose VM expression is lost in exd null embryos and is expressed in PS4 –6 (Rauskolb and Wieschaus, 1994; Waltzer et al., 2001). To this model of multiple general activators and spatially specific repressors is added the spatially localized strong activator Ubx. UBX directly regulates dpp and may also have indirect inputs to dpp’s PS7 gene expression, as our reduced restoration in Ubx 9.22 null embryos indicates. Ubx is itself repressed via chromatin factors such as Polycomb (Collins and Treisman, 2000; Immergluck et al., 1990) and osa (Collins and Treisman, 2000) in the anterior midgut and Abd-A posterior to PS7 (Reuter et al., 1990; Tremml and Bienz, 1989). dpp autoregulation provides additional weak activation via inputs from SMAD proteins (XY, MAM, DAH, unpublished observations) and through DPP-mediated schnurri repression of the repressor brinker (Marty et al., 2000; Torres-Vazquez et al., 2001). Thus, dpp expression is the cumulative result of general activation constrained by spatially specific repression and augmented by spatially specific activation. Clearly, evolution has deemed the formation of the embryonic midgut of sufficient importance to create a highly buffered, reinforced system of gene expression. Acknowledgments The authors wish to thank: Jim Jaynes for hth and exd constructs, Cordelia Rauskolb for the exd cDNA, Adi Salzberg for hth mutants and antibody, Rob White for exd antibody, Jim Mahaffey for Scr antibody, Ben Sun for assistance with Drosophila embryo microinjection, Clark Riley, Janine Ptak

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

and Carol Davenport for oligonucleotide synthesis and DNA sequencing; and Steve Desiderio and Jeffrey Porter for advice on examining EXD phosphorylation. We thank Miki Fujioka and Lesley Brown for technical advice. We also thank Jim Jaynes, Jim Kennison, and Brent McCright for insightful comments on this manuscript. P.A.B. is an Investigator of the Howard Hughes Medical Institute. This research was supported in part by the Food and Drug Administration and the National Institute of Child Health and Human Development. This paper does not represent an official position of the FDA. Appendix A. Supplementary data Supplementary data associated with this article can be found in the online version at doi:10.1016/j.ydbio.2005.11.041. References Abu-Shaar, M., Ryoo, H.D., Mann, R.S., 1999. Control of the nuclear localization of Extradenticle by competing nuclear import and export signals. Genes Dev. 13, 935 – 945. Appel, B., Sakonju, S., 1993. Cell-type-specific mechanisms of transcriptional repression by the homeotic gene products UBX and ABD-A in Drosophila embryos. EMBO J. 12, 1099 – 1109. Aspland, S.E., White, R.A., 1997. Nucleocytoplasmic localisation of extradenticle protein is spatially regulated throughout development in Drosophila. Development 124, 741 – 747. Azpiazu, N., Morata, G., 2000. Function and regulation of homothorax in the wing imaginal disc of Drosophila. Development 127, 2685 – 2693. Azpiazu, N., Morata, G., 2002. Distinct functions of homothorax in leg development in Drosophila. Mech. Dev. 119, 55 – 67. Berthelsen, J., Kilstrup-Nielsen, C., Blasi, F., Mavilio, F., Zappavigna, V., 1999. The subcellular localization of PBX1 and EXD proteins depends on nuclear import and export signals and is modulated by association with PREP1 and HTH. Genes Dev. 13, 946 – 953. Bienz, M., 1994. Homeotic genes and positional signalling in the Drosophila viscera. Trends Genet. 10, 22 – 26. Biggin, M.D., McGinnis, W., 1997. Regulation of segmentation and segmental identity by Drosophila homeoproteins: the role of DNA binding in functional activity and specificity. Development 124, 4425 – 4433. Brodu, V., Elstob, P.R., Gould, A.P., 2002. abdominal A specifies one cell type in Drosophila by regulating one principal target gene. Development 129, 2957 – 2963. Buff, E., Carmena, A., Gisselbrecht, S., Jimenez, F., Michelson, A.M., 1998. Signalling by the Drosophila epidermal growth factor receptor is required for the specification and diversification of embryonic muscle progenitors. Development 125, 2075 – 2086. Burglin, T.R., 1994. A comprehensive classification of Homeobox genes. In: Duboule, D. (Ed.), Guidebook to the Homeobox Genes. Oxford Univ. Press, Oxford, UK, pp. 25 – 72. Burglin, T.R., 1997. Analysis of TALE superclass homeobox genes (MEIS, PBC, KNOX, Iroquois, TGIF) reveals a novel domain conserved between plants and animals. Nucleic Acids Res. 25, 4173 – 4180. Capovilla, M., Botas, J., 1998. Functional dominance among Hox genes: repression dominates activation in the regulation of Dpp. Development 125, 4949 – 4957. Capovilla, M., Brandt, M., Botas, J., 1994. Direct regulation of decapentaplegic by Ultrabithorax and its role in Drosophila midgut morphogenesis. Cell 76, 461 – 475. Capovilla, M., Kambris, Z., Botas, J., 2001. Direct regulation of the muscleidentity gene apterous by a Hox protein in the somatic mesoderm. Development 128, 1221 – 1230. Casares, F., Mann, R.S., 1998. Control of antennal versus leg development in Drosophila. Nature 392, 723 – 726.

493

Chan, S.K., Mann, R.S., 1996. A structural model for a homeotic protein – extradenticle – DNA complex accounts for the choice of HOX protein in the heterodimer. Proc. Natl. Acad. Sci. U. S. A. 93, 5223 – 5228. Chan, S.K., Jaffe, L., Capovilla, M., Botas, J., Mann, R.S., 1994. The DNA binding specificity of Ultrabithorax is modulated by cooperative interactions with extradenticle, another homeoprotein. Cell 78, 603 – 615. Chang, C.P., Shen, W.F., Rozenfeld, S., Lawrence, H.J., Largman, C., Cleary, M.L., 1995. Pbx proteins display hexapeptide-dependent cooperative DNA binding with a subset of Hox proteins. Genes Dev. 9, 663 – 674. Chang, C.P., Brocchieri, L., Shen, W.F., Largman, C., Cleary, M.L., 1996. Pbx modulation of Hox homeodomain amino-terminal arms establishes different DNA-binding specificities across the Hox locus. Mol. Cell. Biol. 16, 1734 – 1745. Chang, C.P., Jacobs, Y., Nakamura, T., Jenkins, N.A., Copeland, N.G., Cleary, M.L., 1997. Meis proteins are major in vivo DNA binding partners for wildtype but not chimeric Pbx proteins. Mol. Cell. Biol. 17, 5679 – 5687. Collins, R.T., Treisman, J.E., 2000. Osa-containing Brahma chromatin remodeling complexes are required for the repression of wingless target genes. Genes Dev. 14, 3140 – 3152. Ekker, S.C., Jackson, D.G., von Kessler, D.P., Sun, B.I., Young, K.E., Beachy, P.A., 1994. The degree of variation in DNA sequence recognition among four Drosophila homeotic proteins. EMBO J. 13, 3551 – 3560. Ferretti, E., Marshall, H., Popperl, H., Maconochie, M., Krumlauf, R., Blasi, F., 2000. Segmental expression of Hoxb2 in r4 requires two separate sites that integrate cooperative interactions between Prep1, Pbx and Hox proteins. Development 127, 155 – 166. Gebelein, B., Culi, J., Ryoo, H.D., Zhang, W., Mann, R.S., 2002. Specificity of Distalless repression and limb primordia development by abdominal Hox proteins. Dev. Cell 3, 487 – 498. Gonzalez-Crespo, S., Morata, G., 1995. Control of Drosophila adult pattern by extradenticle. Development 121, 2117 – 2125. Graba, Y., Aragnol, D., Pradel, J., 1997. Drosophila Hox complex downstream targets and the function of homeotic genes. BioEssays 19, 379 – 388. Grieder, N.C., Marty, T., Ryoo, H.D., Mann, R.S., Affolter, M., 1997. Synergistic activation of a Drosophila enhancer by HOM/EXD and DPP signaling. EMBO J. 16, 7402 – 7410. Hanes, S.D., Brent, R., 1989. DNA specificity of the bicoid activator protein is determined by homeodomain recognition helix residue 9. Cell 57, 1275 – 1283. Hayashi, S., Scott, M.P., 1990. What determines the specificity of action of Drosophila homeodomain proteins? Cell 63, 883 – 894. Huang, H., Rastegar, M., Bodner, C., Goh, S.L., Rambaldi, I., Featherstone, M., 2005. MEIS C termini harbor transcriptional activation domains that respond to cell signaling. J. Biol. Chem. 280, 10119 – 10127. Hursh, D.A., Padgett, R.W., Gelbart, W.M., 1993. Cross regulation of decapentaplegic and Ultrabithorax transcription in the embryonic visceral mesoderm of Drosophila. Development 117, 1211 – 1222. Immergluck, K., Lawrence, P.A., Bienz, M., 1990. Induction across germ layers in Drosophila mediated by a genetic cascade. Cell 62, 261 – 268. Inbal, A., Halachmi, N., Dibner, C., Frank, D., Salzberg, A., 2001. Genetic evidence for the transcriptional-activating function of Homothorax during adult fly development. Development 128, 3405 – 3413. In der Rieden, P.M.J., Mainguy, G., Woltering, J.M., Durston, A.J., 2004. Homeodomain to hexapeptide or PBC-interaction-domain distance: size apparently matters. Trends Genet. 20, 76 – 79. Jacobs, Y., Schnabel, C.A., Cleary, M.L., 1999. Trimeric association of Hox and TALE homeodomain proteins mediates Hoxb2 hindbrain enhancer activity. Mol. Cell. Biol. 19, 5134 – 5142. Jaw, T.J., You, L.R., Knoepfler, P.S., Yao, L.C., Pai, C.Y., Tang, C.Y., Chang, L.P., Berthelsen, J., Blasi, F., Kamps, M.P., Sun, Y.H., 2000. Direct interaction of two homeoproteins, homothorax and extradenticle, is essential for EXD nuclear localization and function. Mech. Dev. 91, 279 – 291. Jones, D.H., Winistorfer, S.C., 1992. Recombinant circle PCR and recombination PCR for site-specific mutagenesis without PCR product purification. BioTechniques 12, 528 – 535. Kamakaka, R.T., Kadonaga, J.T., 1994. The soluble nuclear fraction, a highly efficient transcription extract from Drosophila embryos. Methods Cell Biol. 44, 225 – 235.

494

B.G. Stultz et al. / Developmental Biology 290 (2006) 482 – 494

Kilstrup-Nielsen, C., Alessio, M., Zappavigna, V., 2003. PBX1 nuclear export is regulated independently of PBX – MEINOX interaction by PKA phosphorylation of the PBC-B domain. EMBO J. 22, 89 – 99. Kremser, T., Hasenpusch-Theil, K., Wagner, E., Buttgereit, D., RenkawitzPohl, R., 1999. Expression of the beta3 tubulin gene (beta Tub60D) in the visceral mesoderm of Drosophila is dependent on a complex enhancer that binds Tinman and UBX. Mol. Gen. Genet. 262, 643 – 658. Kurant, E., Pai, C.Y., Sharf, R., Halachmi, N., Sun, Y.H., Salzberg, A., 1998. Dorsotonals/homothorax, the Drosophila homologue of meis1, interacts with extradenticle in patterning of the embryonic PNS. Development 125, 1037 – 1048. Li, X., Murre, C., McGinnis, W., 1999. Activity regulation of a Hox protein and a role for the homeodomain in inhibiting transcriptional activation. EMBO J. 18, 198 – 211. Manak, J.R., Scott, M.P., 1994. A class act: conservation of homeodomain protein functions. Dev. Suppl. 61 – 77. Manak, J.R., Mathies, L.D., Scott, M.P., 1994. Regulation of a decapentaplegic midgut enhancer by homeotic proteins. Development 120, 3605 – 3619. Mann, R.S., Abu-Shaar, M., 1996. Nuclear import of the homeodomain protein extradenticle in response to Wg and Dpp signalling. Nature 383, 630 – 633. Mann, R.S., Morata, G., 2000. The developmental and molecular biology of genes that subdivide the body of Drosophila. Annu. Rev. Cell Dev. Biol. 16, 243 – 271. Marty, T., Muller, B., Basler, K., Affolter, M., 2000. Schnurri mediates Dpp-dependent repression of brinker transcription. Nat. Cell Biol. 2, 745 – 749. McGinnis, W., Krumlauf, R., 1992. Homeobox genes and axial patterning. Cell 68, 283 – 302. Merabet, S., Kambris, Z., Capovilla, M., Berenger, H., Pradel, J., Graba, Y., 2003. The hexapeptide and linker regions of the AbdA Hox protein regulate its activating and repressive functions. Dev. Cell 4, 761 – 768. Merabet, S., Ebner, A., Affolter, M., 2005. The Drosophila Extradenticle and Homothorax selector proteins control branchless/FGF expression in mesodermal bridge-cells. EMBO Rep. 6, 762 – 768. Neuteboom, S.T., Murre, C., 1997. Pbx raises the DNA binding specificity but not the selectivity of antennapedia Hox proteins. Mol. Cell. Biol. 17, 4696 – 4706. Ogo, A., Waterman, M.R., Kamps, M.P., Kagawa, N., 1995. Protein kinase A-dependent transactivation by the E2A-Pbx1 fusion protein. J. Biol. Chem. 270, 25340 – 25343. Pai, C.Y., Kuo, T.S., Jaw, T.J., Kurant, E., Chen, C.T., Bessarab, D.A., Salzberg, A., Sun, Y.H., 1998. The Homothorax homeoprotein activates the nuclear localization of another homeoprotein, extradenticle, and suppresses eye development in Drosophila. Genes Dev. 12, 435 – 446. Passner, J.M., Ryoo, H.D., Shen, L., Mann, R.S., Aggarwal, A.K., 1999. Structure of a DNA-bound Ultrabithorax – Extradenticle homeodomain complex. Nature 397, 714 – 719. Patel, N.H., Snow, P.M., Goodman, C.S., 1987. Characterization and cloning of fasciclin III: a glycoprotein expressed on a subset of neurons and axon pathways in Drosophila. Cell 48, 975 – 988. Pinsonneault, J., Florence, B., Vaessin, H., McGinnis, W., 1997. A model for extradenticle function as a switch that changes HOX proteins from repressors to activators. EMBO J. 16, 2032 – 2042. Piper, D.E., Batchelor, A.H., Chang, C.P., Cleary, M.L., Wolberger, C., 1999. Structure of a HoxB1 – Pbx1 heterodimer bound to DNA: role of the hexapeptide and a fourth homeodomain helix in complex formation. Cell 19 (96), 587 – 597. Popperl, H., Bienz, M., Studer, M., Chan, S.K., Aparicio, S., Brenner, S., Mann, R.S., Krumlauf, R., 1995. Segmental expression of Hoxb-1 is

controlled by a highly conserved autoregulatory loop dependent upon exd/pbx. Cell 81, 1031 – 1042. Rauskolb, C., Wieschaus, E., 1994. Coordinate regulation of downstream genes by extradenticle and the homeotic selector proteins. EMBO J. 13, 3561 – 3569. Reuter, R., Panganiban, G.E., Hoffmann, F.M., Scott, M.P., 1990. Homeotic genes regulate the spatial expression of putative growth factors in the visceral mesoderm of Drosophila embryos. Development 110, 1031 – 1040. Rieckhof, G.E., Casares, F., Ryoo, H.D., Abu-Shaar, M., Mann, R.S., 1997. Nuclear translocation of extradenticle requires homothorax, which encodes an extradenticle-related homeodomain protein. Cell 91, 171 – 183. Ryoo, H.D., Mann, R.S., 1999. The control of trunk Hox specificity and activity by Extradenticle. Genes Dev. 13, 1704 – 1716. Ryoo, H.D., Marty, T., Casares, F., Affolter, M., Mann, R.S., 1999. Regulation of Hox target genes by a DNA bound Homothorax/Hox/Extradenticle complex. Development 126, 5137 – 5148. Saleh, M., Rambaldi, I., Yang, X.J., Featherstone, M.S., 2000. Cell signaling switches HOX – PBX complexes from repressors to activators of transcription mediated by histone deacetylases and histone acetyltransferases. Mol. Cell. Biol. 20, 8623 – 8633. Schier, A.F., Gehring, W.J., 1992. Direct homeodomain – DNA interaction in the autoregulation of the fushi tarazu gene. Nature 356, 804 – 807. Shen, W.F., Chang, C.P., Rozenfeld, S., Sauvageau, G., Humphries, R.K., Lu, M., Lawrence, H.J., Cleary, M.L., Largman, C., 1996. Hox homeodomain proteins exhibit selective complex stabilities with Pbx and DNA. Nucleic Acids Res. 24, 898 – 906. Studier, F.W., Rosenberg, A.H., Dunn, J.J., Dubendorff, J.W., 1990. Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol. 185, 60 – 89. Sun, B., Hursh, D.A., Jackson, D., Beachy, P.A., 1995. Ultrabithorax protein is necessary but not sufficient for full activation of decapentaplegic expression in the visceral mesoderm. EMBO J. 14, 520 – 535. Thuringer, F., Bienz, M., 1993. Indirect autoregulation of a homeotic Drosophila gene mediated by extracellular signaling. Proc. Natl. Acad. Sci. U. S. A. 90, 3899 – 3903. Torres-Vazquez, J., Park, S., Warrior, R., Arora, K., 2001. The transcription factor Schnurri plays a dual role in mediating Dpp signaling during embryogenesis. Development 128, 1657 – 1670. Treisman, J., Gonczy, P., Vashishtha, M., Harris, E., Desplan, C., 1989. A single amino acid can determine the DNA binding specificity of homeodomain proteins. Cell 59, 553 – 562. Tremml, G., Bienz, M., 1989. Homeotic gene expression in the visceral mesoderm of Drosophila embryos. EMBO J. 8, 2677 – 2685. van Dijk, M.A., Murre, C., 1994. Extradenticle raises the DNA binding specificity of homeotic selector gene products. Cell 78, 617 – 624. van Dijk, M.A., Voorhoeve, P.M., Murre, C., 1993. Pbx1 is converted into a transcriptional activator upon acquiring the N-terminal region of E2A in pre-B-cell acute lymphoblastoid leukemia. Proc. Natl. Acad. Sci. U. S. A. 90, 6061 – 6065. Waltzer, L., Vandel, L., Bienz, M., 2001. Teashirt is required for transcriptional repression mediated by high Wingless levels. EMBO J. 20, 137 – 145. Yang, X., van Beest, M., Clevers, H., Jones, T., Hursh, D.A., Mortin, M.A., 2000. Decapentaplegic is a direct target of dTcf repression in the Drosophila visceral mesoderm. Development 127, 3695 – 3702. Yu, X., Hoppler, S., Eresh, S., Bienz, M., 1996. Decapentaplegic, a target gene of the wingless signalling pathway in the Drosophila midgut. Development 122, 849 – 858. Zaffran, S., Kuchler, A., Lee, H.H., Frasch, M., 2001. Biniou (FoxF), a central component in a regulatory network controlling visceral mesoderm development and midgut morphogenesis in Drosophila. Genes Dev. 15, 2900 – 2915.