Regulation by Homeoproteins: A Comparison of Deformed-Responsive Elements

Corresponding author: James W. Mahaffey, Department of Genetics, North Carolina State University, Raleigh, NC 27695-7614., jim_mahaffey{at}ncsu.edu (E-mail)

Communicating editor: T. C. KAUFMAN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Homeotic genes of Drosophila melanogaster encode transcription factors that specify segment identity by activating the appropriate set of target genes required to produce segment-specific characteristics. Advances in understanding target gene selection have been hampered by the lack of genes known to be directly regulated by the HOM-C proteins. Here we present evidence that the gene 1.28 is likely to be a direct target of Deformed in the maxillary segment. We identified a 664-bp Deformed Response Element (1.28 DRE) that directs maxillary-specific expression of a reporter gene in transgenic embryos. The 1.28 DRE contains in vitro binding sites for Deformed and DEAF-1. The Deformed binding sites do not have the consensus sequence for cooperative binding with the cofactor Extradenticle, and we do not detect cooperative binding to these sites, though we cannot rule out an independent role for Extradenticle. Removing the four Deformed binding sites renders the 1.28 DRE inactive in vivo, demonstrating that these sites are necessary for activation of this enhancer element, and supporting the proposition that 1.28 is activated by Deformed. We show that the DEAF-1 binding region is not required for enhancer function. Comparisons of the 1.28 DRE with other known Deformed-responsive enhancers indicate that there are multiple ways to construct Deformed Response Elements.

DISTINCT morphological structures exist along the anterior-posterior axes of animals. In Drosophila melanogaster the homeotic complex (HOM-C) genes are integral components of the pathways that ascribe different identities to cells along this axis (reviewed by MCGINNIS and KRUMLAUF 1992 ). The HOM-C has been conserved throughout evolution and is found in all animals examined (MCGINNIS and KRUMLAUF 1992 ; KRUMLAUF 1994 ; MANAK and SCOTT 1994 ). In Drosophila the HOM-C genes are found in two complexes: the Bithorax complex and the Antennapedia complex (LEWIS 1978 ; BENDER et al. 1983 ; SANCHEZ-HERRERO et al. 1985 ; AKAM 1989 ; KAUFMAN et al. 1990 ; KESSEL and GRUSS 1990 ; LUFKIN et al. 1992 ; RAMIREZ-SOLIS et al. 1993 ). The genes of the Bithorax complex (Ultrabithorax, abdominal A, and Abdominal B) specify segment identity in the posterior thorax and abdominal region. The genes of the Antennapedia complex (labial, proboscipedia, Deformed, Sex combs reduced, and Antennapedia) specify segmental identity in the head and anterior thorax.

The products of the homeotic genes (homeoproteins) function as transcription factors (reviewed by HAYASHI and SCOTT 1990 ). They contain a highly conserved 60-amino-acid DNA-binding domain known as the homeodomain (MCGINNIS et al. 1984 ; SCOTT and WEINER 1984 ; reviewed by SCOTT et al. 1989 ). Homeoproteins are thought to specify segmental identity by activating the appropriate battery of target genes that ultimately produce segment-specific characteristics (reviewed by ANDREW and SCOTT 1992 ; BOTAS 1993 ; MORATA 1993 ). Yet, how this is achieved remains unknown. For several reasons it is unclear how the specificity of target gene selection is attained. Different homeoproteins recognize very similar DNA sequences in vitro and have nearly identical binding affinities (DESPLAN et al. 1988 ; HOEY and LEVINE 1988 ; AFFOLTER et al. 1990 ; FLORENCE et al. 1991 ; DESSAIN et al. 1992 ; EKKER et al. 1994 ; WALTER et al. 1994 ). Further, a single homeoprotein can recognize a variety of DNA sequences (though almost all have the core sequence TAAT). On certain sites with TAAT (or subtle variants of that sequence) an immediately upstream site, TGAT, leads to cooperative heterodimer binding of the Hox protein with the Extradenticle (Exd) protein (CHAN and MANN 1996 ). This interaction enhances the sequence specificity of Hox DNA binding and appears to be required for some Hox proteins to function as transcriptional activators (LI et al. 1999B ). Hox/Exd heterodimer binding sites are found in a subset of Hox response elements, but it is still unclear whether Exd function is required on all response elements or only on some.

To understand HOM-C specification of axial patterning, it is necessary to identify downstream target genes that are controlled by specific homeoproteins. Identification of similarly regulated target genes would allow comparisons of the regulating enhancers, and this could lead to the identification of important cues in target gene regulation. There have been several attempts to systematically identify downstream target genes (GOULD et al. 1990 ; GOULD and WHITE 1992 ; WAGNER-BERNHOLZ et al. 1991 ; GRABA et al. 1992 ; MAHAFFEY et al. 1993 ; FEINSTEIN et al. 1995 ; MASTICK et al. 1995 ; BOTAS and AUWERS 1996 ). Unfortunately, the regulatory regions of only a few target genes have been characterized in sufficient detail to identify homeotic response elements. These include the regulatory regions of teashirt (tsh), decapentaplegic (dpp), and Deformed. tsh controls head vs. trunk development, and high levels of tsh expression in the thoracic epidermis require Antennapedia function (FASANO et al. 1991 ; MCCORMICK et al. 1995 ). Transcriptional regulation of tsh by Antennapedia is probably direct and involves sequences in addition to the homeoprotein binding sites. dpp is a member of the transforming growth factor (TGF)-ß family of proteins and has been shown to be directly regulated by multiple homeoproteins in the embryonic midgut (PADGETT et al. 1987 , PADGETT et al. 1993 ; CAPOVILLA et al. 1994 ; MANAK et al. 1994 ; SUN et al. 1995 ). The homeotic gene Deformed is itself a target of HOM-C regulation through autoregulatory activation during embryonic development, and several autoregulatory elements have been identified that function as Deformed-responsive maxillary enhancers (REGULSKI et al. 1991 ; ZENG et al. 1994 ; GROSS and MCGINNIS 1996 ). In addition, two potential targets of Deformed have been identified, Distal-less (O'HARA et al. 1993 ) and 1.28 (MAHAFFEY et al. 1993 ), respectively.

In this article, we describe results of experiments indicating that the Drosophila 1.28 gene is likely to be a direct target of the Deformed homeoprotein. This allows us to compare activation of 1.28 with autoregulation of Deformed, where a molecular basis for Deformed regulation has been studied extensively. These comparisons provide evidence that activation of different target genes by a single homeoprotein may involve distinct pathways.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Transformation constructs:
Genomic DNA fragments were subcloned into pBluescript II KS+ (pBS), then into pHSS7. This vector allows fragments to be excised with NotI ends, which permits cloning into the NotI site of the transformation-reporter vector pHZ-white. The 1.28 DRE was subcloned into pBS in a three-part ligation. A 285-bp EcoRI fragment and a 379-bp EcoRI-BsrBI fragment were subcloned into the EcoRI-SmaI sites of pBS. The resulting construct was excised from pBS by digestion with BamHI and HindIII and subcloned into pHSS7, then into pHZ-white. Plasmids were purified using QIAGEN'S (Valencia, CA) plasmid midi kit. DNA was ethanol precipitated and resuspended in injection buffer (SPRADLING and RUBIN 1982 ; ASHBURNER 1989 ) at a concentration of 400 ng/µl. Drosophila transformation followed the procedure of ROBERTSON et al. 1988 . Numbers of independent fly lines for each construct are as follows: 1.28 DRE, 8; 1.28 mut1-4, 4; 1.28 DRE Deformed binding region, 10; 1.28 DEAF-1 binding region, 3; chimera 1, 5; chimera 2, 3.

DNase I footprint assays:
Deformed protein was produced in Escherichia coli and purified according to DESSAIN et al. 1992 . DNase I footprinting experiments were carried out as described in HEBERLEIN et al. 1985 . DEAF-1 protein purification and footprint reactions followed the protocol of GROSS and MCGINNIS 1996 .

Site-directed mutagenesis:
Site-directed mutagenesis followed the protocol of KUNKEL et al. 1987 . The 1.28 DRE was subcloned into pBS as described above. The pBS clone was transformed into CJ236 and selected on ampicillin (Amp) and chloramphenicol (CAM). Cells from this culture were patched onto LB plates containing 50 µg/ml Amp and 10 µg/ml CAM and were allowed to grow for 5 hr. A small loop of cells was used to inoculate 10 ml 2x YT containing 50 µg/ml Amp, 10 µg/ml CAM, and 0.25 µg/ml uridine. After 40 min at 37°, the helper phage M13K07 was added. After an additional 30 min, kanamycin was added to a final concentration of 70 µg/ml and the culture was allowed to grow overnight at 37°. Phage were isolated by precipitation with 4% PEG/0.5 M NaCl. The phage were resuspended in 100 µl TE, phenol-extracted, and precipitated. A 10-fold molar excess of mutagenizing primer was added to template DNA. Primer extension reactions included 5 units of Klenow, 0.5 mM dNTPs, and 800 units of T4 ligase. The primer extension reaction was transformed into SURE cells (Stratagene, La Jolla, CA).

In situ hybridization:
Embryos for whole-mount in situ localization of 1.28 or ß-gl transcripts were dechorionated and fixed following the procedure of TAUTZ and PFEIFLE 1989 . In situ hybridization analysis used ribonucleotide probes generated with an RNA transcription kit (Stratagene) and DIG-11-UTP (Boehringer Mannheim, Indianapolis). Hybridization was carried out using modifications to the method of TAUTZ and PFEIFLE 1989 . Anti-DIG-AP (Boehringer Mannheim) was used to detect hybridization.

Chimeric enhancer construction:
The 120-bp module E element was subcloned into the HindIII site of pBS and oriented so the Deformed binding region could be amplified using the forward primer and the DEAF-1 binding region could be amplified using the reverse primer. Primer 120 Deformed was designed to amplify the Deformed binding region of module E and has the sequence 5'-GGAAGCTTCGCCAGTCGGTTGG-3'. Primer 120 DEAF-1 was designed to amplify the DEAF-1 binding region and has the sequence 5'-GGAAGCTTGGGCACATTTCTT-3'. Each primer has a HindIII site added to the end to simplify the subcloning of the fragments. PCR was conducted using the following conditions: 95° for 40 sec, 52° for 1 min, 72° for 3 min for 25 cycles. Five microliters of each reaction was cut with 20 units of HindIII and subcloned into pBS. Positive clones were sequenced using a Sequenase kit (United States Biochemical, Cleveland) to confirm identity and to ensure that no changes were incorporated during PCR. For chimera 1, the 60-bp module E DEAF-1 binding region was digested with HindIII and an aliquot of this digestion was ligated with the HindIII-digested Deformed binding region of 1.28 in pBS and transformed into bacteria. Colonies were picked onto an LB/ampicillin grid plate, grown overnight, and transferred to Magna membrane (Micron Separations, Inc., Westborough, MA) or Nytran (Schleicher & Schuelle; Keene, NH). The filters were probed with module E DNA, which was labeled using the Multiprime DNA labeling system (Amersham, Arlington Heights, IL) and [-³²P]dCTP (New England Nuclear Life Science Products, Boston). The orientations of inserted fragments were determined via restriction digests. Chimera 2 was constructed similarly, except that the 60-bp module E Deformed binding region was subcloned as an EcoRI-XhoI fragment adjacent to the DEAF-1 binding region of 1.28.

Isolation of P-element derivative lines:
To mobilize the P element within the 1.28 gene, the 1.28^P chromosome was crossed to a 2-3 transposase source following the method of ROBERTSON et al. 1988 . Two pairs of primers for PCR were designed to detect deletions in adjacent 1.28 genomic DNA either 5' or 3' to the original P-element insertion point. Deletions were detected by lack of the specific PCR product and verified by Southern blot analysis following standard molecular techniques (SAMBROOK et al. 1989 ).

Electrophoretic mobility shift assay with fragments of 1.28 and module C:
Probes for electrophoretic mobility shift assay were generated by PCR, gel-purified, cut with the enzyme AgeI, and filled in with [³²P]dCTP and Klenow. The 1.28 fragment corresponds to nucleotides 393–536 from Fig 4A. The module C fragment corresponds to nucleotides 680–805 from the 2.7-kb Deformed epidermal autoregulatory element (Fig 4; ZENG et al. 1994 ).

• 1.28 upper-strand primer: 5'-ACTACCGGTGCAGCGCTTC TTAGACTTTG;

  
  

  View larger version (9K): In this window In a new window Download PPT slide   Figure 1. Molecular map of the 1.28 DRE. The top map shows the initial 4-kb SalI fragment originally shown to direct expression in the maxillary segment. The enlargement shows the position of the 664-bp 1.28 DRE. The potential starts and direction of transcription are shown by arrows, and the insertion position of the enhancer trap P element is marked. Bases are numbered from the right EcoRI site to indicate the subfragments described in this study.

  
  

  View larger version (141K): In this window In a new window Download PPT slide   Figure 2. Comparison of 1.28 expression and ß-gal expression generated by 1.28 DRE reporter constructs. Arrows point to the expression in the maxillary segment. (A and B) Expression of the endogenous 1.28 gene in the maxillary segment at stages 12 and 14 of embryogenesis, respectively. (C) The 1.28 DRE directs ß-gal expression in the maxillary segment beginning at stage 14 in a subset of endogenous 1.28-expressing cells. (D) ß-Gal expression from the 1.28 mut1-4 enhancer in a stage 16 embryo. Expression from 1.28 mut1-4 is weak and delayed with respect to the 1.28 DRE. ß-Gal expression detected by in situ hybridization with ß-gal antisense probe. In all figures, anterior is to the left.

  
  

  View larger version (86K): In this window In a new window Download PPT slide   Figure 3. Deformed and DEAF-1 bind to the 1.28 DRE. (A) DNase I footprint of the 1.28 DRE by bacterially expressed Deformed protein. G denotes sequencing lane. Increasing amounts of Deformed protein from left to right (0, 100, and 500 ng, respectively). Deformed binds to the four sites indicated by bars. The sequence of each Deformed binding site is shown in Fig 4. (B) DNaseI footprint of the 1.28 DRE by DEAF-1 protein. G denotes sequencing lane. Increasing amounts of bacterial DEAF-1 protein from left to right (0, 1, and 5 ng, respectively). In each panel, the top strand is labeled. The sequence of the top strand is shown in Fig 4.

  
  

  View larger version (48K): In this window In a new window Download PPT slide   Figure 4. (A) Sequence of the 1.28 DRE. The Deformed binding sites are shown as boxes and are marked DBS1–4. DEAF-1 binding sites are underlined, with the bases forming the imperfect inverted repeat marked with an asterisk. Two potential transcription start sites are indicated by arrows with "start" written above them. The P-element insertion point is shown. Note that the Deformed and DEAF-1 binding sites do not overlap. The four Deformed binding sites are within the EcoRI restriction fragment, and the eight DEAF-1 binding sites are located within the BsrBI-EcoRI restriction fragment. (B) Alignment of the module E (top) and 1.28 DRE (bottom) inverted repeat sequences. Identical bases are marked with an an asterisk. The sequence identified as necessary for expression of module E by LI et al. 1999A is shown in boldface type. Note that this sequence is not found in the inverted repeat region nor elsewhere in the 1.28 DRE.

• 1.28 bottom-strand primer: 5'-ACTACCGGTGCCTCAGCAA ACTAGCG;

• module C upper-strand primer: 5'-ATCACCGGTAAATTCG AATTGAATTTTGGCGG;

• module C bottom-strand primer: 5'-ACTACCGGTCAAAATTTCACAAGATACAACGC.

Binding reactions (20 µl) were performed by incubating labeled DNA fragments (80,000 cpm per lane) with protein translation products in the binding buffer (NEUTEBOOM and MURRE 1997 ), with 0.5 µg poly(dIdC) as a competitor. After 15 min of incubation at room temperature, complexes were resolved on a 5% polyacrylamide gel. Deformed and Exd proteins were produced using the TNT Quick Coupled in vitro transcription/translation system (Promega, Madison, WI). Deformed protein was made from pAR-Dfd (JACK et al. 1988 ), and Exd was produced from pSP64ATG-Exd (VAN DIJK et al. 1993 ).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Identification of a 1.28 Deformed Response Element:
Maxillary expression of 1.28 begins at about stage 10 of embryogenesis in the posterior lateral epidermis of the maxillary lobes, continues through stage 16, and then declines during later development (Fig 2A and Fig B; see also MAHAFFEY et al. 1993 ). 1.28 is also expressed in the gut and the anterior and posterior spiracles. We demonstrated previously that expression of 1.28 in the maxillary epidermis is Deformed dependent; embryos lacking functional Deformed protein do not express 1.28 in the maxillary segment, though expression in other regions of the embryo is not altered (MAHAFFEY et al. 1993 ). Furthermore, we demonstrated that a 4.0-kb fragment from the 1.28 gene was able to direct maxillary-specific expression of a ß-gal reporter construct in transgenic flies.

In the present study we examine portions of this 4.0-kb DNA fragment to identify smaller fragments containing maxillary enhancer elements. Fragments from the 4.0-kb maxillary enhancer were inserted into the P-element vector pHZ-white and tested for enhancer activity in transgenic embryos. Multiple independent transgenic fly lines were obtained for each construct (see MATERIALS AND METHODS). A 664-bp fragment referred to as the 1.28 DRE (1.28 Deformed Response Element) activates reporter expression in a manner similar to the entire 4.0-kb fragment. The position of this fragment with respect to the initial 4-kb enhancer is depicted in Fig 1. The 1.28 DRE directs ß-gal expression in the maxillary segment beginning at stage 14 (Fig 2C). ß-Gal accumulation coincides with endogenous 1.28 expression. Although the pattern of ß-gal expression from the1.28 DRE is similar to the endogenous 1.28 gene, it does not completely reproduce 1.28 gene expression. The endogenous gene is expressed earlier and in more cells of the maxillary lobe. Maxillary expression by the 1.28 DRE is dependent upon the presence of functional Deformed protein; ß-gal is not observed in maxillary lobes of mutants lacking Deformed (data not shown). Like the endogenous gene, the 1.28 DRE directs expression of the reporter in the developing gut beginning at stage 12; this gut expression is independent of Deformed.

The 1.28 DRE contains in vitro Deformed binding sites that are required for enhancer function in vivo:
If Deformed regulation of 1.28 is direct, then the Deformed protein should bind to sequences within the 1.28 gene, and the binding sites should be required to enhance transcription. DNase I footprint analysis was performed to identify potential Deformed binding sites within the 1.28 DRE, using full-length, bacterially expressed Deformed protein (REGULSKI et al. 1991 ). Four Deformed binding regions (DBS1–4) were identified within the 1.28 DRE (Fig 3A); the sequences of the sites are indicated in Fig 4A. The sites all contain the core sequence ATTA/TAAT common to most HOM-C protein binding sites.

To test whether these Deformed binding sites are required for reporter activation in vivo, we used site-directed mutagenesis to change the ATTA/TAAT core and one base on either side to G-C-rich sequences at all four Deformed-binding sites. DNaseI footprinting demonstrated that Deformed could no longer bind to these sites (data not shown). The altered fragment (1.28 mut1-4) was cloned into pHZ-white, transformed into flies, and embryos containing this enhancer were assayed for ß-gal accumulation. This 1.28 mut1-4 element does not direct expression of ß-gal in maxillary cells of stage 14 embryos as does the wild-type 1.28 DRE. We did notice some ß-gal expression from the 1.28 mut1-4 beginning at about stage 16, when a low level of ß-gal could be detected in the maxillary epidermis (Fig 2D). This level of ß-gal is significantly lower than that observed for the wild-type 1.28 DRE and is comparable to levels observed from the pHZ-white vector alone (HAERRY and GEHRING 1997 ; data not shown). We conclude that the sequences forming the in vitro binding sites are necessary for enhancer function in vivo.

Further dissection of the 1.28 DRE:
Several epidermal Deformed-responsive enhancers have been described, some identified within the autoregulation region of Deformed (module A-F; ZENG et al. 1994 ), and one created by changing two bases within a labial response element so that this element now responds to Deformed [repeat 3 (TA); CHAN et al. 1997]. Along with Deformed, there is evidence for a role of several other proteins in activation of these enhancer elements. The DEAF-1 protein was identified by its ability to bind to a region of the 120-bp Deformed autoregulatory enhancer, module E (ZENG et al. 1994 ; GROSS and MCGINNIS 1996 ). Also, the Extradenticle protein (Exd) has been shown to be required for activation of some Deformed-responsive maxillary enhancers (CHAN et al. 1997 ; PINSONNEAULT et al. 1997 ). LI et al. 1999A have recently identified a sequence found at several Deformed-responsive maxillary enhancers that likely forms binding sites for unknown factors that are required for activation of these enhancers in vivo. This sequence was first identified within an imperfect inverted region of the module E enhancer. Since expression of 1.28 and autoregulation of Deformed occur in similar cells of the maxillary epidermis and at a similar stage of embryogenesis, it is possible that one or more of these factors contributes to 1.28 enhancer function. We next investigated whether any of these factors might be involved in activation of the 1.28 DRE.

We first looked for similarities between the 1.28 DRE and the Deformed autoregulation enhancer. The DEAF-1 protein has been shown to bind to module E, where there are several binding sites for this protein. Using DNaseI footprint analysis, we determined that a 300-bp region within the 1.28 DRE contains eight DEAF-1 binding sites (Fig 3B). The sequences of the regions protected from DNaseI digestion by DEAF-1 are indicated in Fig 4A. These binding sites ranged from 10 to 53 bp in length and include the imperfect inverted repeat. The larger protected region likely contains multiple binding sites for the DEAF-1 protein. It is interesting to note that the Deformed and DEAF-1 binding sites do not overlap. The Deformed binding sites are located within the 3' half of this enhancer, and the DEAF-1 binding sites lie within the 5' half.

The 1.28 DRE can be divided at an EcoRI restriction site to give a 285-bp fragment containing the Deformed binding sites and a 379-bp fragment with the DEAF-1 region (Fig 1 and Fig 4A). We tested these two separate regions for enhancer function in vivo by cloning them into pHZ-white, transforming them into flies, and staining embryos to detect ß-gal. The 285-bp fragment containing the four Deformed binding sites is sufficient to direct maxillary expression of the reporter gene in vivo, and this expression is indistinguishable from the complete 1.28 DRE (data not shown). Conversely, the 379-bp fragment containing the DEAF-1 binding region does not direct maxillary expression of the reporter gene at any stage; therefore, this fragment repressed the stage 16, weak, background expression usually seen with the pHZ-white vector alone (data not shown). We conclude that the information needed to direct maxillary expression during stages 14 and 15 is contained entirely within the 285-bp Deformed-binding portion of the 1.28 DRE.

Exd functions as a cofactor at many Hox enhancers, including some Deformed-responsive maxillary enhancers. We used electrophoretic mobility shift assays to determine whether or not the Exd protein could bind to the 1.28 DRE (Fig 5). As a control, we chose module C from the 2.7-kb Deformed epidermal autoregulatory enhancer (ZENG et al. 1994 ). Module C contains a consensus Deformed/Exd heterodimer binding site (TGATTAAT). A fragment containing this region from module C binds Deformed alone, and a trimeric complex between Deformed, Exd, and the module C DNA probe is observed upon addition of Exd to the binding reaction. No trimeric complex was observed in lanes with Deformed, Exd, and the 1.28 DRE probe, though Deformed protein could form a stable complex with the 1.28 DRE (Fig 5). We did not detect binding of Exd to the 1.28 DRE either alone or as a complex with Deformed.

View larger version (86K): In this window In a new window Download PPT slide   Figure 5. Deformed/Exd protein binding to Hox binding sites in the 1.28 element. Electrophoretic mobility shift assays were performed as described in MATERIALS AND METHODS. Deformed protein alone formed a stable complex with the 1.28 DRE fragment that included all Deformed binding sites identified in the 1.28 DRE. This binding was not enhanced by addition of Exd. A fragment of the 2.7-kb Deformed epidermal autoregulatory enhancer module C (ZENG et al. 1994 ), which contains a consensus Deformed/Exd heterodimer binding site (TGATTAAT), was bound by Deformed alone with a slightly higher affinity than the 1.28 fragment. Upon addition of Exd to the binding reaction, a trimeric complex among Deformed, Exd, and the module C DNA probe was observed (indicated by arrow). A trimeric complex was not observed in lanes with Deformed, Exd, and the 1.28 DRE DNA probe. Control lysate was added to the binding reactions to keep the total lysate volume in each reaction at 2 µl.

Other yet-unknown factors are likely to be involved as partners with Hox proteins in target gene activation. These factors would have binding sites that are required at specific enhancers. A candidate for such a binding site is present in an imperfect inverted repeat sequence found in the Deformed autoregulatory enhancer, module E (ZENG et al. 1994 ; GROSS and MCGINNIS 1996 ; LI et al. 1999A ). A similar imperfect inverted repeat sequence is located within the 1.28 DRE (see Fig 4A and Fig B). Since this inverted repeat is located within the DEAF-1 binding region, it was removed from the 1.28 DRE in the experiments described above, indicating that this region does not participate in activating the 1.28 DRE in reporter constructs.

The 1.28 gene was originally identified in an enhancer trap screen (MAHAFFEY et al. 1993 ). In this enhancer trap line the P element is inserted into the DEAF-1 binding region of the endogenous 1.28 DRE (see P-element insertion site in Fig 1 and Fig 4). Insertion of this element causes a significant reduction in 1.28 expression in the maxillary segment (Fig 6B). We mobilized the P element to create deletions in adjacent genomic DNA (ROBERTSON et al. 1988 ). P-element-derivative lines were screened by PCR and Southern blot analysis to identify lines where mobilization of the P element created deletions in the region upstream of the 1.28 gene but not extending into the transcribed portion of the gene. A deletion of 1 kb upstream of the original P-element insertion site was created in derivative line 13 (data not shown). This deletion removes the two DEAF-1 binding regions upstream of the P-element insertion site, including the imperfect inverted repeat (see Fig 4 for the positions of these sites). Expression of the 1.28 in wild type and line 13 embryos is indistinguishable (Fig 6A and Fig C, respectively). This further indicates that the imperfect inverted repeat is not required for the expression of 1.28.

View larger version (47K): In this window In a new window Download PPT slide   Figure 6. The imperfect inverted repeat is not required for in vivo 1.28 expression in the maxillary segment. All embryos are about stage 12. (A) 1.28 mRNA expression in the maxillary segment of a wild-type embryo. (B) 1.28 mRNA expression in the maxillary segment of the enhancer trap line 1.28P. Note that the expression is significantly reduced compared to wild-type 1.28 expression. (C) In P-element derivative line 13, 1.28 mRNA expression is indistinguishable from the wild-type expression pattern.

Further comparisons of the 1.28 DRE and module E:
The overall composition of binding sites making up the 1.28 DRE and module E is quite similar. Both enhancers contain nonoverlapping Deformed and DEAF-1 binding regions. Both enhancers function similarly in vivo, but module E directs reporter expression earlier and in more cells of the maxillary segment than does the 1.28 DRE (Fig 8A and Fig B, respectively). To determine whether the differences are attributable to specific regions of these enhancers, we constructed chimeric enhancers composed of portions of each. Maps of the chimeras are shown in Fig 7. Using these chimeric enhancers we could address whether the Deformed and DEAF-1-binding regions of the two enhancers are similar in function and ascertain which regions are responsible for the differences in enhancer strength, that is, in the timing and number of cells expressing the reporter.

View larger version (17K): In this window In a new window Download PPT slide   Figure 7. Schematic drawing of chimeric enhancer transformation constructs. Deformed binding sites within the 1.28 DRE are indicated by solid boxes and the module E Deformed binding site is shown as an open box. The DEAF-1 binding region of the 1.28 DRE is indicated by a solid oval. The module E DEAF-1 binding region is indicated with an open oval. Arrows denote orientation of DNA fragments. Restriction sites are as follows: B, BsrBI; H, HindIII; R, EcoRI; X, XhoI. H* indicates a HindIII site generated by PCR (see MATERIALS AND METHODS).

View larger version (145K): In this window In a new window Download PPT slide   Figure 8. Comparison between Deformed response elements and chimeric enhancers. All embryos are about stage 14 and arrows indicate maxillary expression. (A) Maxillary expression of ß-gal directed by module E. (B) 1.28 DRE generates maxillary ß-gal expression at a lower level compared to module E. (C) Chimera 1 directs ß-gal expression in the maxillary segment at higher levels compared to the 1.28 DRE. (D) Chimera 2 does not express ß-gal at any stage. ß-Gal expression was detected by RNA in situ hybridization.

Chimera 1 was created so that the 1.28 Deformed binding sites were fused to regions 5 and 6 of module E, which contain DEAF-1 binding sites (Fig 7 C). When assayed after transformation, chimera 1 functions as a maxillary enhancer and drives expression at levels comparable to module E (Fig 8C and A, respectively). Expression of the reporter was earlier, at higher levels, and in a few more cells than the intact 1.28 DRE. Chimera 2 had the 1.28 DEAF-1 binding region linked 5' to the Deformed-binding portion of module E (Fig 7D). In contrast to chimera 1, chimera 2 did not activate maxillary expression at any stage (Fig 8D). Furthermore, the 1.28 DRE DEAF-1 binding region again appears to repress even the slight background expression observed with pHZ-white. These results indicate that the Deformed and DEAF-1 binding regions of the two enhancers are not identical. module E appears to contain a site or sequence that increases enhancer activity when fused to other Deformed binding sites. This result supports the conclusion of LI et al. 1999A , that a sequence enhancing maxillary expression is present in this fragment. Furthermore, this demonstrates that the function can be transferred to another response element.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

We previously demonstrated that a 4.0-kb fragment from the 1.28 gene was able to direct maxillary-specific expression of ß-gal from a reporter construct in transgenic flies (MAHAFFEY et al. 1993 ). In the present study we examined fragments of this 4.0-kb DNA element to identify smaller maxillary enhancer elements. We identified a 664-bp DNA fragment, encompassing the 1.28 start of transcription and containing in vitro binding sites for Deformed, that functions as a maxillary enhancer. That the sequences forming the in vitro Deformed binding sites are required for enhancer function in vivo supports our proposition that the 1.28 gene may be a direct target of Deformed in the maxillary segment. However, we cannot absolutely rule out that one or more of these sites may be interacting with other factors in vivo.

Activation of 1.28 and autoregulation of Deformed occur simultaneously in many cells along the posterior-lateral edge of the maxillary segment beginning at about stage 10. This suggests that common factors may play roles in activating these two genes. Clearly, Deformed is likely a component of both pathways, since Deformed is required for 1.28 expression and autoregulation of Deformed (BERGSON and MCGINNIS 1990 ; MAHAFFEY et al. 1993 ; ZENG et al. 1994 ). However, though Deformed binds to sites within the 1.28 DRE and the Deformed autoregulatory enhancer module E (ZENG et al. 1994 ), the enhancers are not equivalent. The sequences forming the Deformed-binding region of module E will not function as a maxillary enhancer without the addition of other elements. This is similar to the repeat 3[TA] maxillary enhancer as this enhancer requires Deformed and Exd binding sites, and binding to each half-site is required for maxillary expression of the reporter (CHAN et al. 1997 ). In the 1.28 DRE, it appears that the 285-bp Deformed-binding region is sufficient to direct maxillary expression of a reporter gene. The difference between the 1.28 DRE and other Deformed response elements does not appear to be the number of Deformed binding sites within each enhancer, since multimerization of module E, such that there are four Deformed binding sites, does not eliminate the requirement for additional regulatory elements.

From the lack of heterodimer consensus sites and from our protein-binding studies, we suspect that activation of the 1.28 DRE does not require binding of the Exd cofactor as a heterodimer with Deformed, though we cannot rule out an independent role for Exd in activating this element. LI et al. 1999B suggest that the interaction between Exd and Deformed at a heterodimer binding site leads to exposure of the transcriptional activation domain of the Deformed protein and thereby activation of the target gene. If in fact the 1.28 DRE can activate expression without Exd, then either some other unidentified factor must lead to exposure of this activation domain, or perhaps certain arrangements of Deformed binding sites can alleviate this requirement. In some cases Exd may function when not in the heterodimer Hox/Exd arrangement. The Deformed module E enhancer does not contain the consensus TGATTAAT Exd/Deformed binding sequence, but an Exd binding site is located several bases from the Deformed binding site. The sequence forming this Exd binding site appears to be important at some level because eliminating this sequence reduces maxillary expression of a reporter gene, though maxillary expression is not eliminated (PINSONNEAULT et al. 1997 ). At this time, we do not know if such an independent role for Exd is needed to activate the 1.28 gene. Regulation of the endogenous gene is complex and likely requires other regulatory factors and binding sites. It is possible, if not likely, that there are other maxillary enhancer elements at the 1.28 gene.

As mentioned above, to function as a maxillary enhancer module E requires at least one sequence in addition to the Deformed and Exd binding sites. This sequence is found in an imperfect inverted repeat (ZENG et al. 1994 ; GROSS and MCGINNIS 1996 ; LI et al. 1999A ). Site-directed mutagenesis of this imperfect inverted repeat abolishes module E enhancer function. Though it seemed noteworthy that a similar imperfect inverted repeat sequence is located within the 1.28 DRE (Fig 4B), that sequence is not required for 1.28 DRE enhancer function, and deletion of this repeat has no consequence on expression of the endogenous 1.28 gene. Attaching the module E inverted repeat sequence to the Deformed binding portion of the 1.28 DRE does increase activity of the 1.28 DRE, indicating that this sequence can function in a heterologous enhancer. We favor the idea that the module E inverted repeat region contains a binding site or sites for other unknown factors, and that these factors act to enhance maxillary-specific expression. LI et al. 1999A have shown that such a site is likely to be within the inverted repeat sequence of module E. They propose that factors bind to sequences GGC and AAAGC of the module E repeat. This sequence is not present in the 1.28 DRE, suggesting again that regulation through these two enhancers uses different mechanisms.

The DEAF-1 protein was initially hypothesized to be an activator involved in Deformed autoregulation because it bound tightly to the inverted repeat region of module E. However, accumulating evidence indicates that this may not be the case. The DEAF-1 binding site is located in region 6 of module E (LI et al. 1999A ). Though this region is necessary for maxillary enhancer function, eliminating the DEAF-1 binding does not alter the ability of this fragment to be a maxillary enhancer. In our studies, we find that the DEAF-1 binding region of the 1.28 DRE does not enhance maxillary expression of either the 1.28 DRE or the module E Deformed binding sites, and furthermore, in both cases this region suppresses the weak, endogenous activity often observed for the pHZ-white reporter alone. In our studies, the DEAF-1 binding region appears to, at least under some circumstances, act as a negative element. DEAF-1 perhaps does play a role in expression, as a repressor.

Hox response elements are quite complex. Along with the Hox protein-binding region, binding sites for other factors are necessary, integrating other cell identity cues such as proximal/distal position and timing of expression. We have shown here that there are apparently multiple ways to construct a Deformed response element. This likely reflects the many different mechanisms of Hox activation of target genes. Clearly, as more target genes and enhancers are examined, more and different mechanisms will be found. But commonalties are likely as well. Whether the differences indicate a diversity of ways to activate target genes or reflect a role of integrating other determining events in cell-type specification remains to be seen.

	FOOTNOTES

¹ Present address: Wyeth-Lederle Vaccines and Pediatrics, Sanford, NC 27330.
² Present address: Center for Neurobiology and Behavior, Columbia University, New York 10032.

	ACKNOWLEDGMENTS

We thank K. Matthews and the Bloomington Stock Center for supplying various fly stocks. Our sincere thanks go out to the other members of the Mahaffey lab for helpful comments concerning the manuscript and the research leading to it, especially to Mike Griswold for his exceptional editing ability. A.V. was supported by a Howard Hughes Medical Institute predoctoral fellowship. This research was supported by National Science Foundation grant IBN95-14246 to J.W.M.

Manuscript received March 8, 2000; Accepted for publication June 1, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AFFOLTER, M., A. PERCIVAL-SMITH, M. MULLER, W. LEUPIN, and W. J. GEHRING, 1990  DNA binding properties of the purified Antennapedia homeodomain. Proc. Natl. Acad. Sci. USA 87:4093-4097.

AKAM, M., 1989  Drosophila development: making stripes inelegantly. Nature 341:282-283.

ANDREW, D. J. and M. P. SCOTT, 1992  Downstream of the homeotic genes. New Biol. 4:5-15.

ASHBURNER, M., 1989 Drosophila: A Laboratory Handbook. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

BENDER, W., P. SPIERER, and D. S. HOGNESS, 1983  Chromosomal walking and jumping to isolate DNA from the Ace and rosy loci and the Bithorax-complex in Drosophila melanogaster.. J. Mol. Biol. 168:17-33.

BERGSON, C. and W. MCGINNIS, 1990  An autoregulatory enhancer element of the Drosophila homeotic gene Deformed.. EMBO J. 9:4287-4297.

BOTAS, J., 1993  Control of morphogenesis and differentiation by HOM/Hox genes. Curr. Opin. Cell Biol. 5:1015-1022.

BOTAS, J. and L. AUWERS, 1996  Chromosomal binding sites of Ultrabithorax homeotic proteins. Mech. Dev. 56:129-138.

CAPOVILLA, M., M. BRANDT, and J. BOTAS, 1994  Direct regulation of decapentaplegic by Ultrabithorax and its role in Drosophila midgut morphogenesis. Cell 76:461-475.

CHAN, S.-K and R. S. MANN, 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. USA 93:5223-5228.

CHAN, S.-K, H.-D. RYOO, A. GOULD, R. KRUMLAUF, and R. S. MANN, 1997  Switching the in vivo specificity of a minimal Hox-responsive element. Development 124:2007-2014.

DESPLAN, C., J. THEIS, and P. H. O'FARRELL, 1988  The sequence specificity of homeodomain-DNA interaction. Cell 54:1081-1090.

DESSAIN, S., C. T. GROSS, M. A. KUZIORA, and W. MCGINNIS, 1992  Antp-type homeodomains have distinct DNA binding specificities that correlate with their different regulatory functions in embryos. EMBO J. 11:991-1002.

EKKER, S. C., D. G. JACKSON, D. P. VON KESSLER, B. I. SUN, and K. E. YOUNG et al., 1994  The degree of variation in DNA sequence recognition among four Drosophila homeotic proteins. EMBO J. 13:3551-3560.

FASANO, L., L. RODER, N. CORE, E. ALEXANDRE, and C. VOLA et al., 1991  The gene teashirt is required for the development of Drosophila embryonic trunk segments and encodes a protein with widely spaced zinc finger motifs. Cell 64:63-79.

FEINSTEIN, P. G., K. KORNFELD, D. S. HOGNESS, and R. S. MANN, 1995  Identification of homeotic target genes in Drosophila melanogaster including nervy, a proto-oncogene homologue. Genetics 140:573-586.

FLORENCE, B., R. HANDROW, and A. LAUGHON, 1991  DNA-binding specificity of the fushi tarazu homeodomain. Mol. Cell. Biol. 11:3613-3623.

GOULD, A. P. and R. A. WHITE, 1992  Connectin, a target of homeotic gene control in Drosophila. Development 116:1163-1174.

GOULD, A. P., J. J. BROOKMAN, D. I. STRUTT, and R. A. WHITE, 1990  Targets of homeotic gene control in Drosophila. Nature 348:308-312.

GRABA, Y., D. ARAGNOL, P. LAURENTI, V. GARZINO, and D. CHARMOT et al., 1992  Homeotic control in Drosophila: the scabrous gene is an in vivo target of Ultrabithorax proteins. EMBO J. 11:3375-3384.

GROSS, C. T. and W. MCGINNIS, 1996  DEAF-1, a novel protein that binds an essential region in a Deformed response element. EMBO J. 15:1961-1970.

HAERRY, T. E. and W. J. GEHRING, 1997  A conserved cluster of homeodomain binding sites in the mouse Hoxa-4 intron functions in Drosophila embryos as an enhancer that is directly regulated by Ultrabithorax. Dev. Biol. 186:1-15.

HAYASHI, S. and M. P. SCOTT, 1990  What determines the specificity of action of Drosophila homeodomain proteins? Cell 63:883-894.

HEBERLEIN, U., B. ENGLAND, and R. TJIAN, 1985  Characterization of Drosophila transcription factors that activate the tandem promoters of the alcohol dehydrogenase gene. Cell 41:965-977.

HOEY, T. and M. LEVINE, 1988  Divergent homeo box proteins recognize similar DNA sequences in Drosophila. Nature 332:858-861.

JACK, T., R. REGULSKI, and W. MCGINNIS, 1988  Pair-rule segmentation genes regulate the expression of the homeotic selector gene, Deformed.. Genes Dev. 2:635-651.

KAUFMAN, T. C., M. A. SEEGER and G. OLSEN, 1990 Molecular and genetic organization of the Antennapedia gene complex of Drosophila melanogaster, pp. 309–362 in Advances in Genetics: Genetic Regulatory Hierarchies in Development, Vol. 27, edited by P. R. F. WRIGHT.

KESSEL, M. and P. GRUSS, 1990  Murine developmental control genes. Science 249:374-379.

KRUMLAUF, R., 1994  Hox genes in vertebrate development. Cell 78:191-201.

KUNKEL, T. A., J. D. ROBERTS, and R. A. ZAKOUR, 1987  Rapid and efficient site-specific mutagenesis without phenotypic selection. Methods Enzymol. 154:367-382.

LEWIS, E. B., 1978  A gene complex controlling segmentation in Drosophila. Nature 276:565-570.

LI, X, A. VERAKSA, and W. MCGINNIS, 1999a  A sequence motif distinct from Hox binding sites controls the specificity of a Hox response element. Development 126:5581-5589.

LI, X, C. MURRE, and W. MCGINNIS, 1999b  Activity regulation of a Hox protein and a role for the homeodomain in inhibiting transcriptional activation. EMBO J. 18:198-211.

LUFKIN, T., M. MARK, C. P. HART, P. DOLLE, and M. LEMEUR et al., 1992  Homeotic transformation of the occipital bones of the skull by ectopic expression of a homeobox gene. Nature 359:835-841.

MAHAFFEY, J. W., D. F. JONES, J. A. HICKEL, and C. M. GRISWOLD, 1993  Identification and characterization of a gene activated by the deformed homeoprotein. Development 118:203-214.

MANAK, J. R. and M. P. SCOTT, 1994  A class act: conservation of homeodomain protein functions. Development 120(Suppl):61-77.

MANAK, J. R., L. D. MATHIES, and M. P. SCOTT, 1994  Regulation of decapentaplegic midgut enhancer by homeotic proteins. Development 120:3605-3619.

MASTICK, G. S., R. MCKAY, T. OLIGINO, K. DONOVAN, and A. J. LOPEZ, 1995  Identification of target genes regulated by homeotic proteins in Drosophila melanogaster through genetic selection of Ultrabithorax protein-binding sites in yeast. Genetics 139:349-363.

MCCORMICK, A., N. CORE, S. KERRIDGE, and M. P. SCOTT, 1995  Homeotic response elements are tightly linked to tissue-specific elements in a transcriptional enhancer of the teashirt gene. Development 121:2799-2812.

MCGINNIS, W. and R. KRUMLAUF, 1992  Homeobox genes and axial patterning. Cell 68:283-302.

MCGINNIS, W., C. P. HART, W. J. GEHRING, and F. H RUDDLE, 1984  Molecular cloning and chromosome mapping of a mouse DNA sequence homologous to homeotic genes of Drosophila. Cell 38:675-680.

MORATA, G., 1993  Homeotic genes of Drosophila. Curr. Opin. Genet. Dev. 3:606-614.

NEUTEBOOM, S. T. and C. MURRE, 1997  Pbx raises the DNA binding specificity but not the selectivity of antennapedia Hox proteins. Mol. Cell. Biol. 17:4696-4706.

O'HARA, E., B. COHEN, S. M. COHEN, and W. MCGINNIS, 1993  Distal-less is a downstream gene of Deformed required for ventral maxillary identity. Development 117:847-856.

PADGETT, R. W., R. D. ST. JOHNSTON, and W. M. GELBART, 1987  A transcript from a Drosophila pattern gene predicts a protein homologous to the transforming growth factor-beta family. Nature 325:81-84.

PADGETT, R. W., J. M. WOZNEY, and W. M. GELBART, 1993  Human BMP sequences can confer normal dorsal-ventral patterning in the Drosophila embryo. Proc. Natl. Acad. Sci. USA 90:2905-2909.

PINSONNEAULT, J., B. FLORENCE, H. VAESSIN, and W. MCGINNIS, 1997  A model for extradenticle function as a switch that changes HOX proteins from repressors to activators. EMBO J. 16:2032-2042.

RAMIREZ-SOLIS, R., H. ZHENG, J. WHITING, R. KRUMLAUF, and A. BRADLEY, 1993  Hoxb-4 (Hox-2.6) mutant mice show homeotic transformation of a cervical vertebra and defects in the closure of the sternal rudiments. Cell 73:279-294.

REGULSKI, M., S. DESSAIN, N. MCGINNIS, and W. MCGINNIS, 1991  High-affinity binding sites for the Deformed protein are required for the function of an autoregulatory enhancer of the Deformed gene. Genes Dev. 5:278-286.

ROBERTSON, H. M., C. R. PRESTON, R. W. PHILLIS, D. M. JOHNSON-SCHLITZ, and W. K. BENZ et al., 1988  A stable genomic source of P element transposase in Drosophila melanogaster.. Genetics 118:461-470.

SAMBROOK, J., E. F. FRITSCH and T. MANIATIS, 1989 Molecular Cloning: A Laboratory Manual, Ed. 2. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SANCHEZ-HERRERO, E., I. VERNOS, R. MARCO, and G. MORATA, 1985  Genetic organization of Drosophila bithorax complex.. Nature 313:108-113.

SCOTT, M. P. and A. J. WEINER, 1984  Structural relationships among genes that control development: sequence homology between the Antennapedia, Ultrabithorax, and fushi tarazu loci of Drosophila. Proc. Natl. Acad. Sci. USA 81:4115-4119.

SCOTT, M. P., J. W. TAMKUN, and G. W. HARTZELL, III, 1989  The structure and function of the homeodomain. Biochim. Biophys. Acta 989:25-48.

SPRADLING, A. C. and G. M. RUBIN, 1982  Transposition of cloned P elements into Drosophila germ line chromosomes. Science 218:341-347.

SUN, B., D. A. HURSH, D. JACKSON, and P. A. BEACHY, 1995  Ultrabithorax protein is necessary but not sufficient for full activation of decapentaplegic expression in the visceral mesoderm. EMBO J. 14:520-535.

TAUTZ, D. and C. PFEIFLE, 1989  A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98:81-85.

VAN DIJK, M. A., P. M. VOORHOEVE, and C. MURRE, 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. USA 90:6061-6065.

WAGNER-BERNHOLZ, J. T., C. WILSON, G. GIBSON, R. SHUH, and W. GEHRING, 1991  Identification of target genes of the homeotic gene Antennapedia by enhancer detection. Genes Dev. 5:2467-2480.

WALTER, J., C. A. DEVER, and M. D. BIGGIN, 1994  Two homeo domain proteins bind with similar specificity to a wide range of DNA sites in Drosophila embryos. Genes Dev. 8:1678-1692.

ZENG, C., J. PINSONNEAULT, G. GELLON, N. MCGINNIS, and W. MCGINNIS, 1994  Deformed protein binding sites and cofactor binding sites are required for the function of a small segment-specific regulatory element in Drosophila embryos. EMBO J. 13:2362-2377.

This article has been cited by other articles:

				D. E. Reed, X. M. Huang, J. A. Wohlschlegel, M. S. Levine, and K. Senger DEAF-1 regulates immunity gene expression in Drosophila PNAS, June 17, 2008; 105(24): 8351 - 8356. [Abstract] [Full Text] [PDF]

				C. M. Walsh and S. B. Carroll Collaboration between Smads and a Hox protein in target gene repression Development, October 15, 2007; 134(20): 3585 - 3592. [Abstract] [Full Text] [PDF]

				L. K. Robertson, D. B. Bowling, J. P. Mahaffey, B. Imiolczyk, and J. W. Mahaffey An interactive network of zinc-finger proteins contributes to regionalization of the Drosophila embryo and establishes the domains of HOM-C protein function Development, June 15, 2004; 131(12): 2781 - 2789. [Abstract] [Full Text] [PDF]

				K. Koh, S. M. Peyrot, C. G. Wood, J. A. Wagmaister, M. F. Maduro, D. M. Eisenmann, and J. H. Rothman Cell fates and fusion in the C. elegans vulval primordium are regulated by the EGL-18 and ELT-6 GATA factors -- apparent direct targets of the LIN-39 Hox protein Development, March 13, 2003; 129(22): 5171 - 5180. [Abstract] [Full Text] [PDF]