Differential Regulation of Gene Expression by PITX2 Isoforms*

THE JOURNAL OF BIOLOGICAL CHEMISTRY © 2002 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 277, No. 28, Issue of July 12, pp. 25001–25010, 2002 Printed in U.S.A.

Differential Regulation of Gene Expression by PITX2 Isoforms* Received for publication, February 20, 2002, and in revised form, April 3, 2002 Published, JBC Papers in Press, April 10, 2002, DOI 10.1074/jbc.M201737200

Carol J. Cox‡, Herbert M. Espinoza‡, Bryan McWilliams‡, Kimberly Chappell‡, Lisa Morton‡, Tord A. Hjalt§, Elena V. Semina§, and Brad A. Amendt‡¶ From the ‡Department of Biological Science, The University of Tulsa, Tulsa, Oklahoma 74104-3189 and the §Department of Pediatrics, The University of Iowa, Iowa City, Iowa 52242

Three major PITX2 isoforms are differentially expressed in human, mice, zebrafish, chick, and frog tissues. To demonstrate differential regulation of gene expression by these isoforms we used three different promoters and three cell lines. Transient transfection of Chinese hamster ovary, HeLa, and LS-8 cell lines revealed differences in PITX2A and PITX2C activation of the PLOD1 and Dlx2 promoters, however, PITX2B is inactive. In contrast, PITX2B actives the pituitary-specific Prolactin promoter at higher levels than either PITX2A or PITX2C. Interestingly, co-transfection of either PITX2A or PITX2C with PITX2B results in a synergistic activation of the PLOD1 and Dlx2 promoters. Furthermore, PITX2 isoforms have different transcriptional activity dependent upon the cells used for transfection analysis. We have isolated a fourth PITX2 isoform (PITX2D) expressed only in humans, which acts to suppress the transcriptional activity of the other PITX2 isoforms. Electrophoretic mobility shift assays and glutathione S-transferase pull-down experiments demonstrated that all isoforms interact with PITX2D and that PITX2B forms heterodimeric complexes with PITX2A and PITX2C. Our research provides a molecular basis for differential gene regulation through the expression of PITX2 isoforms. PITX2 isoform activities are both promoter- and cell-specific, and our data reveal new mechanisms for PITX2-regulated gene expression.

The PITX genes are members of the bicoid class of the homeodomain proteins. These have a lysine residue at position nine of the third helix and are especially noteworthy for a role in both DNA and RNA binding (1–3). PITX2 was identified by positional cloning of the 4q25 locus in patients with AxenfeldRieger syndrome (4). Patients diagnosed with classical Rieger syndrome have PITX2 mutations, mostly clustered in the homeodomain (3–5). The mouse Pitx2 gene was subsequently cloned from a pituitary library (6). This gene has been cloned by other groups and assigned various names (Ptx2, Otlx2, Brx1, and ARP1) (7–9). Pitx2 has been shown to be expressed in the brain, heart, pituitary, mandibular and maxillary regions, eye, and umbilicus (4, 6, 7). Recent reports using genetic and epigenetic studies and Pitx2 knockout mice have demonstrated * This work was supported by Grant 1-RO1-DE13941 from the NIDCR, National Institutes of Health (to B. A. A.) and by the Fight For Sight Research Division of Prevent Blindness America (Postdoctoral Grant PD99018 to T. A. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ¶ To whom correspondence should be addressed: Dept. of Biological Sciences, The University of Tulsa, 600 S. College Ave., Tulsa, OK 74104-3189. Tel.: 918-631-3328; Fax: 918-631-2762; E-mail: [email protected]. This paper is available on line at http://www.jbc.org

that this gene product is required for the proper development of the embryo (6, 10 –19). Several laboratories have shown that Pitx2 is a mediator of left-right signaling in vertebrates. Epigenetic studies suggested a role for Pitx2 in the determination of vertebrate heart and gut looping (10, 11, 13–16). The analysis of Pitx2⫺/⫺ homozygous knockout mice reveals that Pitx2 is required for normal heart morphogenesis, development of the mandibular and maxillary facial prominences, and normal tooth and pituitary development (17–19). However, Pitx2⫺/⫹ heterozygous mice display certain defects in embryogenesis as seen with the homozygous mice (17, 19). A small fraction of heterozygous mice exhibit anterior chamber defects of the eye and heart defects (17). These mice also failed to close the ventral body wall consistent with omphalocele found in Rieger patients (17). Rieger syndrome results from haploinsufficiency consistent with some of the defects seen in Pitx2⫺/⫹ heterozygous mice (17, 20). We and others have identified three major PITX2 isoforms produced by alternative splicing and use of different promoters (4, 6, 9, 21, 22). PITX2A and PITX2B are generated by alternative splicing mechanisms, and PITX2C uses an alternative promoter located upstream of exon 4 (see Fig. 1 below). All isoforms contain dissimilar N-terminal domains, whereas the homeodomain and C-terminal domains are identical. The Cterminal domain contains a highly conserved 14-amino acid region described in the homeobox genes Otp and aristaless (4, 6), which is called the OAR (Otp and aristaless) domain. Recent reports have provided evidence of Pitx2 isoform regulation in left-right asymmetry (23–25). Epigenetic and genetic studies reveal that tissue and organ developments are differentially regulated by Pitx2a and Pitx2c isoforms. In the chick it appears that Pitx2c plays a crucial role in the left-right axis determination and rightward heart looping during chick embryogenesis (25). However, these researchers were unable to detect the Pitx2b isoform in chicks. Zebrafish are somewhat different in that Pitx2a has a greater impact on cardiac symmetry than Pitx2c (23). In Zebrafish Pitx2c is asymmetrically expressed in the left dorsal diencephalon and developing gut, whereas Pitx2a is seen in the left heart primordium. Eloquent experiments in mice that were defective in Pitx2a and Pitx2b expression demonstrate that different organs have distinct requirements for Pitx2c dosage (24). These researchers have shown that lower levels of Pitx2c expression were required for cardiac atria and higher levels for duodenum and lung development. In contrast, other investigators have reported expression of Pitx2c and Pitx2b but not Pitx2a in mice and frogs (26). They report overlapping and distinct patterns of Pitx2 expression in the lateral plate mesoderm, heart, gut, cement gland, head mesenchyma, pituitary gland, branchial arches, myotome, and muscles. Pitx2c and Pitx2b were expressed in the head region of mice, and overlapping expression patterns were seen in the brain of frogs. However, they report only Pitx2c expression was

25001

25002

PITX2 Isoform Transcriptional Activities

observed during heart development in both the mouse and frog (26). Other experiments in mice have shown Pitx2 expression in the odontogenic epithelium, and it is the first transcriptional marker of tooth development (27). More recently, we have shown that Pitx2 protein is restricted to the developing dental epithelium (28). Altogether these data demonstrate that Pitx2 isoforms are required either separately or in overlapping domains and in different doses to regulate normal vertebrate heart, lung, brain, tooth, pituitary, and gut development. However, the biochemical/molecular mechanisms of these effects have not been determined. Target genes for PITX2 have been described for the pituitary; the Prolactin gene is synergistically activated by Pit-1 and PITX2 (3). Other pituitary-specific Pitx2 target genes have also been described (29). However, we have now identified two genes outside of the pituitary that are specifically regulated by PITX2. We have shown that PITX2 regulates procollagen lysyl hydroxylase (PLOD) and Dlx2 gene expression (30, 31). The PLOD1 gene encodes an enzyme responsible for hydroxylizing lysines in collagens that plays a role in specifying the extracellular matrix and provides a foundation for the morphogenesis of tissues and organs. The Dlx2 gene encodes a transcription factor expressed in the mesenchymal and epithelial cells of the mandibular and maxillary regions and expressed in the diencephalon. Dlx2, a member of the distal-less gene family, has been established as a regulator of branchial arch development (32, 33). Homozygous mutants of Dlx2 have abnormal development of forebrain cells and craniofacial abnormalities in developing neural tissue. Dlx genes exhibit both sequential and overlapping expression, implying that temporo-spatial regulation of Dlx genes are tightly regulated (34). Pitx2 and Dlx2 genes are expressed in the same tissues early during development. These reports establish that the PITX2 family of bicoid-like homeodomain genes are key regulators of important development processes and are required to regulate specific genes during embryogenesis. Our studies demonstrate differential activation of the PLOD1 and Dlx2 promoters by PITX2 isoforms in several cell lines. We demonstrate synergism between PITX2 isoforms, and their activities appear to be promoter-dependent. We report the identification of a new PITX2 isoform (PITX2D) generated by the PITX2C alternative promoter and differential splicing (Fig. 1). We have only observed this isoform expressed in humans, and it was identified from a human craniofacial library. The PITX2D isoform acts to down-regulate the transcriptional activities of PITX2A and PITX2C. All isoforms can form homodimers, and heterodimers are formed with PITX2B. We demonstrate new regulatory mechanisms for the fine-tuning of PITX2 transcriptional activity that is required for normal development. Our data provide a molecular/biochemical basis for the developmental regulation of organ and tissue development by PITX2 isoforms reported in humans, zebrafish, chicks, frogs, and mice. MATERIALS AND METHODS

Expression and Purification of GST-PITX2 Fusion Proteins—The PITX2A isoform was PCR-amplified from a cDNA clone as described (3). The PITX2B, PITX2C, and PITX2D isoforms were PCR-amplified from cDNA clones provided by Drs. Elena Semina and Jeff Murray (Department of Pediatrics, University of Iowa). The 5⬘-primers all contained the initiation codon and a unique SalI site, whereas the 3⬘-antisense primer contained PITX2 sequences downstream of the stop codon and a unique NotI site (5⬘-GTACTGCAGATGCGGCCGCAGCATAATTCCCAGTC-3⬘) to facilitate cloning into the pGEX6P-2 GST1 vector (Amersham Bio-

1 The abbreviations used are: GST, glutathione S-transferase; EMSA, electrophoretic mobility shift assay; CMV, cytomegalovirus; CHO, Chinese hamster ovary cells.

sciences) as previously described (3, 35). The 5⬘-primers were unique for each isoform and consisted of the following: PITX2B (5⬘-CGTCGTCGACATGGAGACCAATTGTCGC-3⬘), PITX2C (5⬘-CGTCGTCGACATGAACTGCATGAAAGGC-3⬘), PITX2D (5⬘-CGTCGTCGACATGTCCACACGCGAAGAA-3⬘). All pGST-PITX2 plasmids were confirmed by DNA sequencing. The plasmids were transformed into BL21 cells. Proteins were isolated as described previously (3). PITX2 proteins were cleaved from the GST moiety using 80 units of PreScission Protease (Amersham Biosciences) per milliliter of glutathione-Sepharose. The cleaved proteins were analyzed on SDS-polyacrylamide gels and quantitated by the Bradford protein assay (Bio-Rad). Electrophoretic Mobility Shift Assay—Complementary oligonucleotides containing the Dlx2 bicoid and bicoid-like sites with flanking partial BamHI ends were annealed and filled with Klenow polymerase to generate 32P-labeled probes for EMSAs, as described previously (31). Standard binding assays were performed as previously described (35). The bacteria-expressed and -purified PITX2 proteins were used in the assays at the indicated amounts. The samples were electrophoresed, visualized, and quantitated as described previously, except quantitation of dried gels was performed on the Molecular Dynamics Storm PhosphorImager (Amersham Biosciences) (31). GST-PITX2 Pull-down and Western Blot Assays—Immobilized GSTPITX2D fusion protein was prepared as described above and suspended in binding buffer (20 mM Hepes, pH 7.5, 5% glycerol, 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1% milk, and 400 ␮g/ml ethidium bromide). Purified bacteria-expressed PITX2 proteins (200 ng) were added to 5 ␮g of immobilized GST-PITX2D fusion proteins or GST in a total volume of 100 ␮l and incubated for 30 min at 4 °C. The beads were pelleted and washed four times with 200 ␮l of binding buffer. The bound proteins were eluted by boiling in SDS-sample buffer and separated on a 12.5% SDS-polyacrylamide gel. Approximately 200 ng of purified PITX2 proteins were analyzed in separate Western blots. Following SDS-gel electrophoresis, the proteins were transferred to polyvinylidene difluoride filters (Millipore), immunoblotted, and detected using PITX2 antibody P2R10 (28, 31) and ECL reagents from Amersham Biosciences. Cloning of the PITX2D Isoform—The homeobox sequence of the PITX2 gene was PCR-amplified using the primers, sense, 5⬘-caggggaagaatgaggacgt-3⬘, and antisense, 5⬘-gaagccattcttgcatagct-3⬘, and the PITX2A plasmid as a template (4). The 175-bp fragment containing sequences of the first exon of the PITX2D isoform was PCR-amplified using the primers, sense, 5⬘-ctgagctgcggcaaggc-3⬘, and antisense, 5⬘ggcagccctgacagagatg-3⬘, and the PITX2D plasmid as a template (this report). The PCR fragments were separated by electrophoresis in agarose gel and extracted from the gel, and DNA was cleaned using a Qiagen gel extraction kit and labeled with 32P using a random-prime labeling kit (Roche Molecular Biochemicals) according to the manufacturer’s protocols. The following cDNA libraries were screened: human craniofacial (constructed from mRNA derived from the craniofacial region of human embryos ranging from 42- to 53-day gestation (36), mouse embryonic carcinoma (Stratagene), and mouse 15-day embryo (Novagene). The hybridization, washing, exposure, identification of the positive clones, excision, and sequencing procedures were performed as previously described (4). The exon-intron boundaries were identified by comparison of the identified cDNA and the PITX2 genomic sequence. By using the 266-bp homeobox sequence of the PITX2 as a probe, we identified multiple positive clones that can be divided into seven groups: sequences of PITX2 isoform A, PITX2 isoform B, PITX2 isoform C, PITX2 isoform D, partial PITX2 sequences that can be attributed to any isoform, and various sequences belonging to either the PITX1 or PITX3 genes. The PITX2 A–C isoforms were described before (4, 6, 9, 21). The PITX2D isoform was only identified from the human craniofacial library: two independent clones were isolated from the 3 ⫻ 106 clones examined. The PITX2D sequence consists of three exons: the first exon (178 bp), which was found to be located 230-bp upstream of the first exon of the PITX2C isoform as it was identified from the human craniofacial library; the second exon (77 bp) representing a partial sequence of the PITX2 exon 5 lacking 129 of its 5⬘-nucleotides; and the third exon (1258 bp) that is equivalent to the PITX2 exon 6 (Fig. 1). By searching GenBankTM, we identified that the first exon of the PITX2D isoform was found to be a part of the first exon of the PITX2C isoform in two independent submissions: IMAGE 3937807 clone isolated from the lung library and ARP1C (PITX2C) cDNA. Sequences at the exon/intron junctions for the PITX2D were identified for the 5⬘-splice site, GGGCTGCCGC/gt, and for the 3⬘-splice site, cactttcc/AGAGGAACAGC. It is notable that the nucleotides at positions ⫺1 and ⫺2 of the 3⬘-splice site (CC) do not correspond with the conserved sequence identified as AG. The PITX2D isoform is predicted to encode the 205-amino acid protein


FIG. 1. PITX2 major isoforms found in humans. A, genomic organization of the PITX2 gene; intron sizes are shown on the top, and exon sizes are at the bottom; exons are numbered. B, the protein structure is shown with the location of the homeodomain (HD) and 14-amino acid conserved OAR domain. Checkered and stippled boxes denote the differences in the N-terminal region of the isoforms. The exons that code for the respective proteins are shown below each isoform. PITX2C and PITX2D RNA is transcribed using an internal promoter shown as a striped box flanking exon 4. with the initiation codon for methionine (ATG) located at the beginning of the second helix sequence of the homeobox. Additional screening of the above described libraries with the fragment containing the first exon sequence of the PITX2D isoform failed to identify any clones from the mouse cDNA libraries as well as any additional clones from the human craniofacial cDNA library. Expression and Reporter Constructs—Expression plasmids containing the cytomegalovirus (CMV) promoter linked to the PITX2 DNA were constructed in pcDNA 3.1 MycHisC (Invitrogen) (35). Constructions of the Prolactin, Dlx2, and PLOD1 promoter plasmids have been previously described (3, 30, 31). All constructs were confirmed by DNA sequencing. A CMV ␤-galactosidase reporter plasmid (CLONTECH) was co-transfected in all experiments as a control for transfection efficiency. Cell Culture, Transient Transfections, Luciferase, and ␤-Galactosidase Assays—CHO, HeLa, and LS-8 cells were cultured in Dulbecco’s modified Eagle’s medium supplemented with 5% fetal bovine serum and penicillin/streptomycin in 60-mm dishes and transfected by electroporation. CHO, HeLa, and LS-8 cells were mixed with 2.5 ␮g of expression plasmids, 5 ␮g of reporter plasmid, and 0.5 ␮g of CMV ␤-galactosidase plasmid plated in 60-mm culture dishes and fed with 5% fetal bovine serum and Dulbecco’s modified Eagle’s medium. Electroporation of CHO cells was performed at 360 V and 950 microfarads (Bio-Rad), and electroporation of HeLa cells was at 220 V and 950 microfarads. The cells were fed 24 h prior to transfection. LS-8 cells were transfected by electroporation as previously described (31). Transfected cells were incubated for 24 h then lysed and assayed for reporter activities and protein content by Bradford assay (Bio-Rad). Luciferase was measured using reagents from Promega. ␤-Galactosidase was measured using the Galacto-Light Plus reagents (Tropix Inc.). All luciferase activities were normalized to ␤-galactosidase activity. Expression of transiently expressed PITX2 proteins has been previously demonstrated (35). RESULTS

PITX2 Isoforms Differentially Regulate the Dlx2, PLOD1, and Prolactin Promoters—To investigate the transcriptional activities of the three major PITX2 isoforms, we used three different naturally occurring promoters linked to the luciferase gene and assayed their activity in three cell lines representing different tissues. We have identified the only downstream targets of PITX2 outside of the pituitary, and we compared the

25003

levels of PITX2 activation of these promoters to the pituitaryspecific Prolactin promoter. We have previously shown that the PITX2A isoform can activate all three promoters (30, 31, 35). We compared the activities of the PITX2 A, B, and C isoforms using the full-length Dlx2–3276 and minimal Dlx2–200 promoters in CHO cells (Fig. 2A). PITX2A activates the Dlx2–3276 promoter at 30-fold, PITX2C at 22-fold, but surprisingly PITX2B demonstrates only 2- to 3-fold activation of this promoter compared with empty expression vector transfected as a control (Fig. 2B). Each isoform has limited activity when transfected with the Dlx2–200 minimal promoter, which we use as a control to demonstrate specific PITX2 activity. The result with PITX2B was surprising because all isoforms contain identical homeodomain and C-terminal regions (Fig. 1). We and others (25, 29, 35) have shown that the PITX2 C-terminal region contains a transcriptional activation domain. Thus, this was an unexpected result and is probably due to the presence of different PITX2 N-terminal sequences. Another explanation for this result would include reduced expression or stability of PITX2B in our transfected CHO cells. To address this we analyzed our transfected CHO lysates for PITX2 protein expression and found equal expression of all isoforms (Fig. 2C). Thus, these results suggest that the N-terminal region of PITX2B may negatively regulate its transcriptional activity in CHO cells using the Dlx2 promoter. We next asked if these PITX2 isoforms would differentially regulate the PLOD1 promoter. Two PLOD1 promoter constructs were made and linked to the luciferase gene (Fig. 3A). PITX2A activated the full-length PLOD1–261 promoter at 10fold, whereas PITX2C revealed increased activity at 17-fold and PITX2B demonstrated only 2-fold activation in transfected CHO cells, compared with empty vector transfection (Fig. 3B). Although these results again reveal that PITX2B has little activity in contrast to experiments with the Dlx2 promoter, PITX2C is more active than PITX2A when assayed in the same cell line (p ⬍ 0.05). These results indicate that the activities of the PITX2 isoforms, though subtle, are promoter-dependent. To further analyze the differences in PITX2 isoform activities, we have used the pituitary-specific Prolactin promoter. It has been previously reported that the three major PITX2 isoforms, including PITX2B, all activated several pituitary promoters at similar levels (29, 37). We co-transfected the naturally occurring Prolactin promoter with the PITX2 isoforms and found that PITX2B gave the greatest activation with this promoter in CHO cells (Fig. 4). PITX2A and PITX2C activated the Prolactin promoter at slightly less levels. Thus, our data agree with other investigators using the Prolactin promoter. Altogether these results demonstrate that the PITX2 A, B, and C isoforms have different transcriptional activities that are promoter-dependent. However, PITX2B is only active using the Prolactin promoter in CHO cells. DNA Binding Activities of the PITX2 Isoforms—A possible explanation for the differential transcriptional regulation would be due to differences in DNA binding activities by the isoforms. We next asked if these PITX2 isoforms bound the Dlx2 bicoid DNA element (5⬘-TAATCC-3⬘) at similar activities. We expected that they would, because they all contain identical homeodomains; however, the different N termini could influence their binding activities. We have previously reported that the PITX2A C-terminal tail can interact with the N terminus to modulate its DNA binding activity (35, 37). Therefore, we speculated that this interaction could be either disrupted or enhanced depending on the relative structure of the N terminus. However, we demonstrate that each isoform binds the Dlx2 bicoid element with similar activities and all form homodimers (Fig. 5). We assayed increasing protein concentrations from 80

25004


FIG. 2. PITX2 isoform activation of the Dlx2 promoter in CHO cells. A, schematic of the Dlx2 promoter constructs used in transient transfection assays showing the location of bicoid and bicoid-like DNA elements; Bcd, bicoid and bicoid-like sequences. B, CHO cells were transfected with either the Dlx2–3276- or Dlx2–200 luciferase reporter genes (5 ␮g). The cells were co-transfected with the CMV-PITX2 isoform expression plasmids or the CMV plasmid without PITX2 (⫺) (2.5 ␮g). To control for transfection efficiency, all transfections included the SV-40 ␤-galactosidase reporter (0.5 ␮g). Cells were incubated for 24 h then assayed for luciferase and ␤-galactosidase activities. The activities are shown as mean -fold activation compared with the Dlx2 promoter plasmids without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from ten independent experiments for panel B). The mean Dlx2 promoter luciferase activity with PITX2 expression was about 100,000 light units per 15 ␮g of protein, and the ␤-galactosidase activity was about 70,000 light units per 15 ␮g of protein. C, Western blot of transfected CHO cell lysates using the PITX2 antibody. CHO cell lysates from transfection experiments in panel B (10 ␮g) were tested for PITX2 isoform expression. As a control, 500 ng of bacterial-expressed PITX2A was used to show the correct migration of the transient-expressed PITX2 isoform proteins. CHO cells co-transfected with the Dlx2 promoter construct and empty expression vector were used as the mock control.

to 240 ng for each isoform (Fig. 5). Thus, the lack of PITX2B activity using the Dlx2 and PLOD1 promoters was not due to a disruption of its DNA binding activity, because both promoters contain multiple bicoid elements. Cell-specific Regulation of PITX2 Isoform Transcriptional Activity—Because all three major PITX2 isoforms contain different N-terminal amino acid sequences, we asked if cellular factors might influence their activities. As we discussed in the introduction the PITX2 isoforms are expressed in distinct and overlapping domains during human, frog, mouse, zebrafish, and chick development. We reasoned that these isoforms might have different activities in cell lines derived from different tissues and vertebrate species, due to PITX2 interacting factors. Transfection of HeLa cells with the Dlx2 promoter revealed changes in PITX2 isoform activities compared with CHO transfections presented in Fig. 2B. In HeLa cells PITX2C was more active than PITX2A: compare the 5-fold activation of

Dlx2–3276 for PITX2C to the 3-fold activation for PITX2A (p ⬍ 0.05) (Fig. 6). In CHO cells PITX2A was more active than PITX2C using the Dlx2 promoter (p ⬍ 0.05). However, PITX2B remained inactive in HeLa cells. These data clearly demonstrate a difference in PITX2 isoform activity based on the transfected cell line. We further analyzed this response in a tooth epithelial cell line, LS-8, which we have previously shown reduces the activity of the Dlx2 promoter when co-transfected with PITX2A, compared with CHO cells (31). In this cell line the activities of all three major PITX2 isoforms were similar, albeit with low activation as we have previously reported (Fig. 7). LS-8 cells endogenously express Msx2, which can antagonize PITX2 activation of the Dlx2 promoter (31). Furthermore, we have shown that the LS-8 cell line contains factors that complex with PITX2A to presumably regulate its activity. The reduced PITX2 activation of the Dlx2 promoter in LS-8 cells is currently under investigation. Reverse transcription-PCR ex-


FIG. 3. PITX2 isoform activation of the PLOD1 promoter in CHO cells. A, schematic of the PLOD1 promoter constructs used in transient transfection assays showing the location of bicoid and bicoidlike DNA elements. B, CHO cells were transfected with either the PLOD1–261 or PLOD1–2561 luciferase reporter genes (5 ␮g). The cells were co-transfected with the CMV-PITX2 isoform expression plasmids or the CMV plasmid without PITX2 (⫺) (2.5 ␮g). All transfection assays were performed as described in Fig. 2. The activities are shown as mean -fold activation compared with the PLOD1 promoter plasmids without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from nine independent experiments).

FIG. 4. PITX2 isoform activation of the Prolactin promoter in CHO cells. CHO cells were transfected with the prolactin 2.5-luciferase reporter plasmid and co-transfected with the CMV-PITX2 isoform expression plasmids or the parental CMV plasmid without PITX2 (⫺). All transfection assays were performed as described in Fig. 2. The activities are shown as mean -fold activation compared with the Prolactin promoter plasmids without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from eight independent experiments).

FIG. 5. PITX2 isoforms have similar binding activities. PITX2 proteins (80, 160, and 240 ng) were incubated with the Dlx2 bicoid consensus sequence (TAATCC) as the radioactive probe. The EMSA experiments were analyzed in 8% native polyacrylamide gels. The free and bound forms of DNA were quantitated using the Molecular Dynamics Storm PhosphorImager (Amersham Biosciences). The free probe, bound, and dimer complexes are indicated.

periments have identified the expression of Pitx2a and Pitx2c but not Pitx2b in this cell line (31). Altogether, these data demonstrate a cell-specific regulation of PITX2 isoform transcriptional activity.

25005

FIG. 6. PITX2 isoform activation of the Dlx2 promoter in HeLa cells. HeLa cells were transfected with either the Dlx2–3276- or Dlx2– 200 luciferase reporter genes (5 ␮g). The cells were co-transfected with the CMV-PITX2 isoform expression plasmids or the CMV plasmid without PITX2 (⫺) (2.5 ␮g). To control for transfection efficiency, all transfections included the SV-40 ␤-galactosidase reporter (0.5 ␮g). Cells were incubated for 24 h then assayed for luciferase and ␤-galactosidase activities. The activities are shown as mean -fold activation compared with the Dlx2 promoter plasmids without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from eight independent experiments).

FIG. 7. PITX2 isoform activation of the Dlx2 promoter in the LS-8 tooth epithelial cell line. LS-8 cells were transfected with either the Dlx2–3325- or Dlx2–250 luciferase reporter genes. The cells were co-transfected with the CMV-PITX2 isoform expression plasmids or the CMV plasmid without PITX2 (⫺). To control for transfection efficiency, all transfections included the CMV ␤-galactosidase reporter. Cells were incubated for 24 h then assayed for luciferase and ␤-galactosidase activities. The activities are shown as mean -fold activation compared with the Dlx2 promoters without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from six independent experiments). The mean Dlx2–3325 luciferase activity with PITX2 expression was about 5,000 light units per 15 ␮g of protein, and the ␤-galactosidase activity was about 40,000 light units per 15 ␮g of protein.

Transcriptional Synergism by PITX2 Isoforms—PITX2 isoforms are co-expressed in different combinations during vertebrate development. All three major Pitx2 isoforms are expressed in the pituitary and craniofacial region, whereas other organs and tissues express combinations of the isoforms that can be species-dependent. To examine the effect of these isoforms acting together to regulate gene expression, we transfected CHO cells with combinations of PITX2A, 2B, and 2C. PITX2A activated the Dlx2 promoter at 30-fold and PITX2B at 2-fold compared with empty vector, but surprisingly co-transfection of both yielded a synergistic 67-fold activation of the Dlx2 promoter (Fig. 8). Co-transfection of PITX2C and PITX2B also revealed a synergistic 63-fold activation, whereas co-transfection of PITX2A and PITX2C resulted in only an additive effect (Fig. 8). These data suggest that PITX2B can interact with PITX2A and 2C to increase their transcriptional activities. To determine if CHO cells provided factors that enhanced the PITX2B interactions and transcriptional activity, we also cotransfected HeLa cells. In HeLa cells we observed similar results compared with transfected CHO cells. PITX2A activated the Dlx2 promoter at 3-fold, whereas PITX2B was inactive; however, co-transfection of both PITX2A and 2B yielded a 7-fold activation (Fig. 9). Co-transfection of PITX2B and PITX2C activated the Dlx2 promoter in HeLa cells at 7-fold compared with empty vector control transfection (Fig. 9). However, co-transfection of PITX2A and PITX2C resulted in an

25006


FIG. 10. Transcriptional synergism of PITX2 isoforms using the Prolactin promoter in CHO cells. CHO cells were transfected as described in Fig. 4, using combinations of the PITX2 isoform expression plasmids.

FIG. 8. Transcriptional synergism of PITX2 isoforms in CHO cells. A, CHO cells were transfected with either the Dlx2–3276- or Dlx2–200 luciferase reporter genes (5 ␮g). The cells were co-transfected with combinations of the CMV-PITX2 isoform expression plasmids or the CMV plasmid without PITX2 (⫺) (2.5 ␮g each). To control for transfection efficiency, all transfections included the SV-40 ␤-galactosidase reporter (0.5 ␮g). Cells were incubated for 24 h then assayed for luciferase and ␤-galactosidase activities. The activities are shown as mean -fold activation compared with the Dlx2 promoter plasmids without PITX2 expression and normalized to ␤-galactosidase activity (⫾S.E. from ten independent experiments).

FIG. 9. Transcriptional synergism of PITX2 isoforms in HeLa cells. HeLa cells were transfected as described in Fig. 8.

additive effect similar to CHO cells (Fig. 9). Interestingly, we observe less dramatic synergistic effects through PITX2 isoform interactions using the Prolactin promoter in CHO cells. Co-transfection of PITX2A and PITX2B resulted in a 15-fold activation, PITX2A and PITX2C co-transfection resulted in a 16-fold activation, and PITX2B and PITX2C resulted in a 21-fold activation (Fig. 10). Surprisingly, with the Prolactin promoter, all three isoforms can synergize with each other, whereas with the Dlx2 and PLOD1 promoters co-transfection of PITX2A and PITX2C resulted in additive activation (data not shown for PLOD1). Clearly, these isoforms can interact and significantly increase promoter activity. These data, when taken together, reveal a mechanism for the combinatorial role these isoforms can play when expressed in the same tissues during development. This type of mechanism would rapidly activate genes required for normal development, and the control of PITX2 isoform expression would work to tightly control gene expression. Furthermore, although PITX2B appears inactive by itself, we demonstrate a critical role for this isoform in activation of the Dlx2 and PLOD1 genes. PITX2 Isoforms Heterodimerize—Because we have shown transcriptional synergy between PITX2A and 2C with PITX2B, we next asked if this was due to physical interactions between these isoforms. To address this question we used EMSAs to determine if heterodimers were formed when combinations of the isoforms were mixed together and allowed to bind to the Dlx2 bicoid probe. We have previously demonstrated, using GST-PITX2 pull-down assays, that PITX2 isoforms can physi-

cally interact (31). When PITX2A and 2B were mixed in equal amounts (80 ng) we observed an increase in DNA binding through the formation of heterodimers and little increase in monomer binding (Fig. 11). Quantitation of the gels revealed that most of the PITX2B bound as a heterodimer complex with PITX2A. This is visualized as a large slower migrating complex above the PITX2A monomer band (Fig. 11). Interestingly, mixing equal amounts of PITX2A and 2C resulted in a slight increase in heterodimer formation however; the increase in overall binding resulted from an increase in each protein binding as a monomer (Fig. 11). This can be seen by two separate faster migrating bands, which run at the same position as each separate monomer protein. Mixing equal amounts of PITX2C and PITX2B resulted in an increase in heterodimer formation (Fig. 11). Overall these results reveal that the PITX2B isoform appears to facilitate dimerization with PITX2A and PITX2C. These data suggest that the transcriptional synergism observed between PITX2A and PITX2C with PITX2B occurs through the ability of the PITX2B isoform to physically interact with the other two isoforms. The PITX2D Isoform Inhibits the Transcriptional Activity of PITX2A and PITX2C—We have identified a new PITX2 isoform from a human craniofacial library. It is made by alternative splicing of a transcript produced from the internal promoter located in intron 3, which also produces the PITX2C isoform (Fig. 1). PITX2D results from splicing of exon 4a to a cryptic 3⬘-splice site in exon 5, which produces a truncated homeodomain and complete C-terminal tail. We have shown that this isoform does not bind to DNA as expected, because it does not contain a functional homeodomain (Fig. 5). However, because it is expressed with the other isoforms, we asked if it had a functional activity with respect to the other isoforms. CHO cells were co-transfected with the Dlx2 promoter and PITX2 expression plasmids. As expected, PITX2D has no transcriptional activity when transfected with the Dlx2 promoter (Fig. 12A). However, when co-transfected with PITX2A, PITX2D caused a 3-fold reduction in PITX2A transcriptional activity of the Dlx2–3276-luc promoter, from 30- to ⬃10-fold in CHO cells (Fig. 12A). A 2-fold reduction of PITX2C transcriptional activity was observed when co-transfected with PITX2D, from 25- to 12-fold (Fig. 12A). PITX2D had no effect on the transcriptional activity of PITX2B, which we have shown is not active with this promoter. Co-transfection of the PITX2 isoforms with the minimal Dlx2–200-luc plasmid revealed little activation, and PITX2D only minimally inhibited this activation (Fig. 12A). These data reveal that PITX2D can negatively regulate the transcriptional activities of PITX2A and PITX2C isoforms. A possible explanation for these results could involve a mechanism where the PITX2D RNA inhibits the translation of the other PITX2 isoforms in CHO cells. We performed a Western blot of transfected CHO cell lysates to determine if PITX2A expression and protein stability were affected by PITX2D. We assayed two concentrations of CHO lysates and found that


25007

FIG. 13. PITX2D suppresses Dlx2 promoter activation in HeLa cells. HeLa cells were transfected as in Fig. 6 except that PITX2D (2.5 ␮g) was co-transfected were indicated with the other PITX2 isoform plasmids.

FIG. 11. PITX2B forms dimers with the other PITX2 isoforms. PITX2 proteins (80 ng) were incubated either separately or in combinations with the Dlx2 bicoid consensus sequence (TAATCC) as the radioactive probe. The EMSA experiments were analyzed in 8% native polyacrylamide gels. The free and bound forms of DNA were quantitated using the Molecular Dynamics Storm PhosphorImager (Amersham Biosciences). The free probe, bound, and dimer complexes are indicated.

FIG. 12. PITX2D suppresses Dlx2 promoter activation by the other PITX2 isoforms. CHO cells were transfected as in Fig. 2 except that PITX2D (2.5 ␮g) was co-transfected where indicated with the other PITX2 isoform plasmids. B, Western blot of transfected CHO cell lysates using the PITX2 antibody. CHO cell lysates from transfection experiments in panel A (10 and 20 ␮g) were tested for PITX2A expression. As a control, 1 ␮g of bacterial expressed PITX2A was used to show the correct migration of the transient expressed PITX2A protein. Our PITX2 antibody will not recognize the PITX2D isoform, because the antibody epitope flanks the homeodomain and lies in the N-terminal region shared by the other three PITX2 isoforms (28, 31). This epitope is missing in the PITX2D isoform. CHO cells co-transfected with the Dlx2 promoter construct and empty expression vector were used as the mock control.

co-expression of PITX2D with PITX2A had no effect on PITX2A protein expression or stability (Fig. 12B). We also found that PITX2C protein expression was unaffected by co-expression of PITX2D (data not shown). Our PITX2 antibody recognizes an N-terminal epitope shared by PITX2A, B, and C isoforms however, it will not recognize PITX2D because the epitope is lost. All of our PITX2 expression plasmids contain a C-terminal Myc tag that allows us to observe PITX2D expression using the Myc

antibody (data not shown). These results suggest that factors specific for CHO cells might interact with PITX2D to facilitate its repression of PITX2A and 2C transcriptional activity. To address this possibility we transfected HeLa cells and observed similar repression of PITX2A and 2C activity using the Dlx2–3276-luc reporter plasmid by PITX2D as seen in CHO cells (Fig. 13). Thus, this repressive effect of PITX2D does not appear to be due to specific factors associated with a specific cell line. The repressive effects by PITX2D were observed with other promoter constructs, including the PLOD1 and Prolactin promoters demonstrating that this effect is not restricted to a specific promoter (data not shown). PITX2D Physically Interacts with PITX2A and PITX2C Isoforms—Our transfection results indicate that PITX2D could physically interact with the other PITX2 isoforms to attenuate their activity. We used GST-PITX2D pull-down assays to determine if PITX2A and PITX2C isoforms could interact with PITX2D. GST-PITX2D was immobilized to Sepharose 4B beads (Amersham Biosciences) and incubated with bacteria-expressed and -purified PITX2 isoforms. As a control the purified PITX2 isoform proteins were also incubated with GST beads. PITX2A and PITX2C were able to bind to immobilized GSTPITX2D (Fig. 14). As a control we show that GST beads alone did not bind PITX2A or C (Fig. 14). These data clearly demonstrate that the PITX2D isoform can physically interact with the other PITX2 isoforms. These experiments corroborate our previous experiments demonstrating that PITX2 isoforms interact through their C-terminal tails (31). Because all PITX2 isoforms contain identical C-terminal tails, our data demonstrate that each isoform has the capability to interact with other isoforms. Another explanation for the suppression of PITX2A and PITX2C activity by PITX2D might be due to the inability of a PITX2A/2D or PITX2C/2D complex to bind DNA. We performed EMSAs where we mixed PITX2A and PITX2C with PITX2D and found neither a loss of binding or increased PITX2A or PITX2C binding activity (data not shown). Clearly, the easiest explanation is that PITX2D binds factors that are essential for PITX2 activity, thereby sequestering that factor from interacting with PITX2 isoforms. Although this is a possibility, we speculate that PITX2D directly binds to PITX2A and PITX2C to inhibit their activity. This mechanism is analogous to our previous report demonstrating that the C-terminal 39-amino acid peptide can also inhibit the transcriptional activity of PITX2A (35). DISCUSSION

Gene expression can be regulated by alternatively spliced transcription factors. Alternative splicing of transcription fac-

25008


FIG. 14. PITX2D directly interacts with PITX2A and PITX2C. GST-PITX2D pull-down assay with bacterial-expressed and -purified PITX2A and PITX2C proteins. The PITX2A and PITX22C isoforms bind to PITX2D demonstrating that PITX2D can physically interact with the other isoforms. The bound isoforms were detected by Western blot using the PITX2 antibody. As a control, GST beads were incubated with purified PITX2A and PITX2C proteins to demonstrate the specificity of binding to PITX2D.

tors provides a mechanism for the fine-tuning of gene expression during development. Three major PITX2 isoforms have been isolated and shown to differentially regulate organogenesis. However, the molecular mechanism for this development preference of the different PITX2 isoforms was unknown. We have been studying the mechanism of PITX2 transcriptional regulation and have recently identified several genes that are regulated by PITX2 (28, 31, 35). The results from the present study reveal a promoter- and cell-dependent activation by the three major PITX2 isoforms. Our recent identification of a fourth minor PITX2 isoform expressed in humans adds another level of regulation to the transcriptional activity of PITX2. In the brain, craniofacial region, and pituitary, which express all three major PITX2 isoforms, the interactions between PITX2 isoforms provide a mechanism to tightly regulate gene expression controlled by PITX2. Our research provides several mechanisms by which PITX2 isoforms may interact to both activate and repress gene expression. We have shown in this report and previously (31, 35) that the PITX2 isoforms can heterodimerize. Furthermore, we have shown that PITX2A can synergize with Pit-1 to activate the Prolactin promoter (3, 35). All of the major PITX2 isoforms can interact with Pit-1 to synergistically activate the Prolactin promoter (29).2 Our results demonstrate that the PITX2 isoforms interact to specifically regulate Prolactin gene expression. We have provided new insights into pituitary development by demonstrating that the three major PITX2 isoforms interact to significantly upregulate Prolactin expression. Thus, the levels and combinations of PITX2 isoform expression contribute to the dosageresponse model proposed for pituitary and other organ development (17, 24). Interestingly, using the Dlx2 promoter whose gene product is 2 C. J. Cox, H. M. Espinoza, B. McWilliams, K. Chappell, L. Morton, T. A. Hjalt, E. V. Semina, and B. A. Amendt, unpublished observations.

expressed during craniofacial and brain development, we find that PITX2A is more active than PITX2C whereas PITX2B has minimal activity. However, PITX2B, when co-transfected with PITX2A or PITX2C isoforms, results in synergistic activation of the Dlx2 promoter. It has been reported that all three major PITX2 isoforms are expressed in the brain and craniofacial region (37). Thus, even though a different promoter regulates PITX2C expression its level of expression is similar to PITX2A and 2B in the brain. Therefore, the location of PITX2 isoform expression in regions of the developing embryo could have important effects on PITX2 target gene expression such as Dlx2. During embryogenesis, we speculate based on our research that the combination and levels of PITX2 isoform expression will greatly influence Dlx2 gene expression. We demonstrate this type of regulation by using different cell lines, which reveal differences in PITX2 isoform activities. PITX2A activates the Dlx2 promoter more strongly than PITX2C in CHO cells whereas, PITX2C activates the Dlx2 promoter more strongly in HeLa cells compared with PITX2A. However, PITX2B is inactive in both cell lines. Furthermore, we use a tooth epithelial cell line to assay PITX2 isoform activities and observe a different ratio of PITX2 isoform transcriptional activities. These data reveal a preference for PITX2 isoform transcriptional activity based on the cells used in our transfection experiments. We propose that PITX2 isoform transcriptional activity will be similarly different in tissues of the developing embryo. In contrast to Dlx2 and Prolactin expression by PITX2 isoforms, the PLOD1 promoter is activated more strongly by PITX2C than PITX2A in CHO cells. Similar to Dlx2 the PLOD1 promoter is not activated by PITX2B. However, the Dlx2 and PLOD1 promoters are activated synergistically by the combination of PITX2A/2B and PITX2C/2B (Figs. 8 and 9).2 Interestingly, PLOD1 has been shown to be expressed in the heart, and because PITX2C has been reported to be the major isoform expressed in the heart our results corroborate the function of PITX2C in heart development. Furthermore, our laboratory has recently identified a heart-specific gene that is regulated by PITX2 isoforms, however, only PITX2C can synergistically activate this promoter in the presence of another heart-specific transcription factor.3 Through the use of three naturally occurring promoters we are able to demonstrate specific differences in PITX2 isoform activities. We have shown that the activities of the major PITX2 isoforms are dependent on the specific promoter they activate and the cells in which they are expressed. Our research provides the first functional mechanism for differential gene expression by PITX2A, 2B, and 2C isoforms. It has been shown that other transcription factor isoforms can differentially regulate gene expression. The Pax-5 gene produces four isoforms as a result of alternative splicing, which can act to either positively or negatively regulate gene expression (38). Interestingly, an alternative form of Pax5, termed Pax5e, does not bind DNA but causes an increase in Pax5a activity. Thus, there is precedence for alternatively spliced transcription factors that do not bind DNA but can exert a regulatory effect on the other isoforms. However, our data reveal a negative regulatory mechanism for the PITX2D isoform. The Pit-1 transcription factor gene produces alternatively spliced products that regulate Prolactin gene expression (39). Because the major PITX2 isoforms only differ in their N termini, we speculate that the N terminus must play a role in the differential transcriptional activities of these isoforms. However, we have shown that each isoform binds DNA simi-

3 C. J. Cox, H. M. Espinoza, B. McWilliams, K. Chappell, L. Morton, T. A. Hjalt, E. V. Semina, and B. A. Amendt, manuscript in preparation.

PITX2 Isoform Transcriptional Activities larly, thus, the N terminus must interact with tissue- and/or cell-specific factors to regulate their activities. Our data indicate that this occurs because we observed transcriptional activation differences in several cell lines. To address the functional properties of the different N termini we have cotransfected each N-terminal-specific peptide with the PITX2 isoforms (data not shown). Our rationale was that if the N termini were binding specific factors then expression of the N-terminal peptides would sequester cellular factors by binding them and thus allow for differential regulation by the wild-type isoforms. Interestingly, when we co-transfected the PITX2A isoform with the PITX2A and PITX2B N-terminal peptides, we observed a 2-fold increase in PITX2A transcriptional activity.2 These data suggest that the N termini of the PITX2 isoforms may be binding factors, which regulate their activities. However, co-transfection of the 2C N-terminal peptide had no effect on PITX2A activity. We did not present these results in this report because they are difficult to interpret. However, they do provide clues that the N terminus of each PITX2 isoform may bind cell-specific factors that regulate their activities. Functions of PITX2D—Our transfection data clearly demonstrate that isoform PITX2D has the ability to act as a transcriptional suppressor. We demonstrated that PITX2D inhibits PITX2A and 2C activation of the Dlx2 promoter. Furthermore, this inhibition occurs in CHO, HeLa, and LS-8 cells, demonstrating that it was not cell-specific. We also observe this inhibitory effect with the PLOD1 and Prolactin promoters.2 PITX2D has a truncated homeodomain, which is derived from the use of a cryptic 3⬘-splice site. This isoform appears to be produced as a result of aberrant splicing in humans. We isolated this isoform from a human craniofacial library, and the specific tissue or organ distribution of this isoform has not yet been determined. However, we have repeatedly seen this isoform in human craniofacial libraries, and we speculate that it may have important functions in regulating PITX2 transcriptional activity in humans. PITX2D does not bind DNA, and we have shown that it does not inhibit the expression of PITX2 isoforms in the transfected cell lines. However, it does physically interact with the major PITX2 isoform proteins, which appear to be the mechanism by which it inhibits the activity of the other isoforms. Interestingly, this isoform acts in a very similar manner to that of our previously reported PITX2 Cterminal 39-amino acid peptide, which acts to inhibit PITX2A transcriptional activity in transfected cells (35). In that report we demonstrated that the PITX2 C39 peptide bound to PITX2 to inhibit its transcriptional activity and, similarly, in this report we demonstrate that PITX2D also binds PITX2A and PITX2C isoforms. The mechanism of this suppressive effect is currently unknown, and we are investigating its action. But, interestingly, we and others (25, 29, 35) have reported the existence of a transactivation domain in the C terminus of PITX2. Thus, if the PITX2D protein forms a heterodimer with the other isoforms, it should be able to activate the promoters in our transfection assays because it contains the C-terminal transactivation domain. Because mixing PITX2D with each PITX2 isoform does not inhibit their DNA binding activity (data not shown), then suppression of PITX2A and PITX2C transcriptional activity is not caused by a loss of DNA binding activity. One explanation for the suppressive effect would involve PITX2D-binding cellular factors required for PITX2 activity. We and others (35, 40) have shown that the C-terminal region of PITX2 is important for protein-protein interaction and that it binds cellular factors. Thus, PITX2D may act to suppress the activity of

25009

the other PITX2 isoforms by “soaking up” factors that normally bind to PITX2. We have titrated PITX2D DNA concentrations in our transfection experiments and have not observe a corresponding decrease in PITX2 activation upon increased PITX2D DNA concentrations.2 Thus, it appears unlikely that PITX2D binds a factor essential for PITX2 activity. We are currently investigating the mechanism of this novel PITX2 isoform. Developmentally, it does provide an interesting mechanism for the regulation and fine-tuning of PITX2 transcriptional activity, which appears to be required for the normal morphogenesis of several organs. In summary, our studies provide evidence that PITX2 isoforms differentially activate genes involved in development. We provide a molecular basis for organ/tissue development by PITX2 isoforms, where the expression of PITX2 isoforms can greatly influence gene expression. Furthermore, we provide evidence for the regulation of PITX2 isoform transcriptional activity in a cell-dependent manner. Lastly, we demonstrate a new mechanism for the regulation of PITX2 transcriptional activation through the action of a novel PITX2 isoform. Acknowledgments—We thank John Hall and Crystal (Zoe) Hansen for excellent technical assistance and Drs. Jeffrey C. Murray and Andrew F. Russo (University of Iowa) for reagents and helpful discussions. We also thank Drs. Paul Sharpe and Bethan Thomas (King’s College, University of London) for the Dlx2 promoter plasmid. REFERENCES 1. Dubnau, J., and Struhl, G. (1996) Nature 379, 694 – 699 2. Lamonerie, T., Tremblay, J. J., Lanctot, C., Therrien, M., Gauthier, Y., and Drouin, J. (1996) Genes Dev. 10, 1284 –1295 3. Amendt, B. A., Sutherland, L. B., Semina, E., and Russo, A. F. (1998) J. Biol. Chem. 273, 20066 –20072 4. Semina, E. V., Reiter, R., Leysens, N. J., Alward, L. M., Small, K. W., Datson, N. A., Siegel-Bartelt, J., Bierke-Nelson, D., Bitoun, P., Zabel, B. U., Carey, J. C., and Murray, J. C. (1996) Nat. Genet. 14, 392–399 5. Amendt, B. A., and Semina, E. V., and Alward, W. L. M. (2000) Cell. Mol. Life Sci. 57, 1652–1666 6. Gage, P. J., and Camper, S. A. (1997) Hum. Mol. Genet. 6, 457– 464 7. Mucchielli, M., Martinez, S., Pattyn, A., Goridis, C., and Brunet, J. (1996) Mol. Cell. Neurosci. 8, 258 –271 8. Kitamura, K., Miura, H., Yanazawa, M., Miyashita, T., and Kato, K. (1997) Mech. Dev. 67, 83–96 9. Arakawa, H., Nakamura, T., Zhadanov, A. B., Fidanza, V., Yano, T., Bullrich, F., Shimizu, M., Blechman, J., Mazo, A., Canaani, E., and Croce, C. M. (1998) Proc. Natl. Acad. Sci. 95, 4573– 4578 10. Campione, M., Steinbeisser, H., Schweickert, A., Deissler, K., van Bebber, F., Lowe, L. A., Nowotschin, S., Viebahn, C., Haffter, P., and Kuehn, M. R., Blum, M. (1999) Dev. 126, 1225–1234 11. Yoshioka, H., Meno, C., Koshiba, K., Sugihara, M., Itoh, H., Ishimaru, Y., Inoue, T., Ohuchi, H., Semina, E. V., Murray, J. C., Hamada, H., and Noji, S. (1998) Cell 94, 299 –305 12. Semina, E. V., Reiter, R. S., and Murray, J. C. (1997) Hum. Mol. Genet. 6, 2109 –2116 13. Logan, M., Pagan-Westphal, S. M., Smith, D. M., Paganessi, L., and Tabin, C. J. (1998) Cell 94, 307–317 14. Piedra, M. E., Icardo, J. M., Albajar, M., Rodriguez-Rey, J. C., and Ros, M. A. (1998) Cell 94, 319 –324 15. Ryan, A. K., Blumberg, B., Rodriguez-Esteban, C., Yonei-Tamura, S., Tamura, K., Tsukui, T., de la Pena, J., Sabbagh, W., Greenwald, J., Choe, S., Norris, D. P., Robertson, E. J., Evans, R. M., Rosenfeld, M. G., and Belmonte, J. C. I. (1998) Nature 394, 545–551 16. St. Amand, T. R., Ra, J., Zhang, Y., Hu, Y., Baber, S. I., Qiu, M., and Chen, Y. P. (1998) Biochem. Biophys. Res. Commun. 247, 100 –105 17. Gage, P. J., Suh, H., and Camper, S. A. (1999) Development 126, 4643– 4651 18. Lu, M., Pressman, C., Dyer, R., Johnson, R. L., and Martin, J. F. (1999) Nature 401, 276 –278 19. Lin, C. R., Kioussi, C., O’Connell, S., Briata, P., Szeto, D., Liu, F., IzpisuaBelmonte, J. C., and Rosenfeld, M. G. (1999) Nature 401, 279 –282 20. Flomen, R. H., Gorman, P. A., Vatcheva, R., Groet, J., Barisic, I., Ligutic, I., Sheer, D., and Nizetic, D. (1997) J. Med. Genet. 34, 191–195 21. Gage, P. J., Suh, H., and Camper, S. A. (1999) Dev. Biol. 210, 234 22. Kitamura, K., Miura, H., Miyagawa-Tomita, S., Yanazawa, M., Katoh-Fukui, Y., Suzuki, R., Ohuchi, H., Suehiro, A., Motegi, Y., Nakahara, Y., Kondo, S., and Yokoyama, M. (1999) Development 126, 5749 –5758 23. Essner, J. J., Branford, W. W., Zhang, J., and Yost, H. J. (2000) Development 127, 1081–1093 24. Liu, C., Liu, W., Lu, M., Brown, N. A., and Martin, J. F. (2001) Dev. 128, 2039 –2048 25. Yu, X., St. Amand, T. R., Wang, S., Li, G., Zhang, Y., Hu, Y., Nguyen, L., Qiu, M., and Chen, Y. (2001) Development 128, 1005–1013 26. Schweickert, A., Campione, M., Steinbeisser, H., and Blum, M. (2000) Mech. Dev. 90, 41–51 27. Mucchielli, M.-L., Mitsiadis, T. A., Raffo, S., Brunet, J.-F., Proust, J.-P., and

25010


Goridis, C. (1997) Dev. Biol. 189, 275–284 28. Hjalt, T. A., Semina, E. V., Amendt, B. A., and Murray, J. C. (2000) Dev. Dyn. 218, 195–200 29. Tremblay, J. J., Goodyer, C. G., and Drouin, J. (2000) Neuroendocrinology 71, 277–286 30. Hjalt, T. A., Amendt, B. A., and Murray, J. C. (2001) J. Cell Biol. 152, 545–552 31. Green, P. D., Hjalt, T. A., Kirk, D. E., Sutherland, L. B., Thomas, B. L., Sharpe, P. T., Snead, M. L., Murray, J. C., Russo, A. F., and Amendt, B. A. (2001) Gene Expr. 9, 265–281 32. Qiu, M., Bulfone, A., Martinez, S., Meneses, J. J., Shimamura, K., Pedersen, R. A., and Rubenstein, J. L. R. (1995) Genes Dev. 9, 2523–2538 33. Thomas, B. L., Liu, J. K., Rubenstein, J. L. R., and Sharpe, P. T. (2000) Development 127, 217–224

34. Liu, J. K., Ghattas, I., Liu, S., Chen, S., and Rubenstein, J. L. R. (1997) Dev. Dyn. 210, 498 –512 35. Amendt, B. A., Sutherland, L. B., and Russo, A. F. (1999) Mol. Cel. Biol. 19, 7001–7010 36. Padanilam, B. J., Stadler, H. S., Mills, K. A., McLeod, L. B., Solursh, M., Lee, B., Ramirez, F., Buetow, K. H., and Murray, J. C. (1992) Hum. Mol. Genet. 1, 407– 410 37. Smidt, M. P., Cox, J. J., van Schaick, H. S. A., Coolen, M., Schepers, J., van der Kleij, A. M., and Burbach, J. P. H. (2000) J. Neurochem. 75, 1818 –1825 38. Lowen, M., Scott, G., and Zwollo, P. (2001) J. Biol. Chem. 276, 42565– 42574 39. Day, R. N., and Day, K. H. (1994) Mol. Endocrinol. 8, 374 –381 40. Szeto, D. P., Ryan, A. K., O’Connell, S. M., and Rosenfeld, M. G. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 7706 –7710