Cloning of the mouse Pax4 gene promoter and identification of a pancreatic beta cell specific enhancer

Molecular and Cellular Endocrinology 170 (2000) 79 – 89 www.elsevier.com/locate/mce

Cloning of the mouse Pax4 gene promoter and identification of a pancreatic beta cell specific enhancer Wenzhong Xu a, Liam J. Murphy a,b,* b

a Department of Physiology, Uni6ersity of Manitoba, Winnipeg, Canada R3E 0W 3 Department of Internal Medicine, Uni6ersity of Manitoba, Winnipeg, Canada R3E 0W 3

Received 26 April 2000; accepted 17 July 2000

Abstract Pax4 encodes a paired-box transcription factor and is essential for the differentiation of islet cells since the Pax4 homozygous mutant mice lack mature b and d cells. However, little is known about the transcriptional regulation of the Pax4 gene. We isolated and sequenced a 2.4-kb mouse genomic DNA fragment containing the 5% flanking sequence of Pax4 and identified a previously unrecognized intron. Primer extension revealed that this TATA-less promoter had only one transcription start site. The promoter activity of this fragment with various deletion mutants when tested in b and non-b cell lines indicated the presence of a b-cell specific enhancer in the region, −1858 to −1954 bp. DNase 1 footprinting and gel retardation assays indicated that nuclear proteins from bHC3 cells interacted with two sequences which contained putative CdxA/Nkx.2 and GATA-1,-2 binding sites. Site-directed mutagenesis indicated that both of these regions were necessary for b-cell specific enhancer activity. © 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Pax4; DNase 1 footprinting; b-cell specific enhancer; Nkx 2.2; CdxA

1. Introduction During embryonic life and in the post-natal period, a subset of pluri-potential cells present in the pancreatic duct undergo a series of differentiation steps that give rise to four endocrine cell types; a, b, d and PP cells which produce glucagon, insulin, somatostatin and pancreatic polypeptide, respectively (Alpert et al., 1988; LeDouarin, 1988). Recent studies using homologous recombination gene disruption have revealed a series of genes, which are necessary for the endocrine pancreatic development. These include Isl-1 (Ahlgren et al., 1997), Pdx-1 (Offield et al., 1996), Beta2/neuroD (Naya et al., 1997), Nkx-2.2 (Sussel et al., 1998) and the paired-box genes Pax4 and Pax6 (Sosa-Pineda et al., 1997; St-Onge et al., 1997). These genes appear to be involved in different stages of the differentiation of pancreatic endocrine precursors. * Corresponding author. Tel.: +1-204-7893779; fax: + 1-2047893940. E-mail address: [email protected] (L.J. Murphy).

Cell differentiation is the result of differential gene expression orchestrated by transcription factors. The Pax gene family of developmental transcription factors share conserved DNA-binding regions with a helixturn-helix motif (Walter et al., 1991). They are expressed in a tissue specific manner and play important role in embryogenesis (Dahl et al., 1997). Pax4 is expressed early in the primordial pancreas but is later restricted to b cells (St-Onge et al., 1997). Inactivation of the Pax4 gene results in the absence of the insulinproducing b cells and the somatostatin-producing d cells in the newborn mouse, which dies within 3 days of birth (Sosa-Pineda et al., 1997). Normally, the earliest insulin staining pancreatic precursor cells are present at E8.5–E9 (Gittes and Rutter, 1992). In the Pax4 null mutant mouse embryos, insulin staining is apparent at this stage indicating that Pax4 expression is not required for generation of these early b cells. However, the onset of expression of Pax4 in the pancreas, around E9.5 and the absence of mature b cells in pancreas of Pax4 mutant newborn mice suggest an important role of Pax4 in the survival and differentiation of these early b cells, and the differentiation of other islet cell precur-

0303-7207/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0303-7207(00)00329-4

80

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

sors in the developing endocrine pancreas (Sosa-Pineda et al., 1997). Investigation of the transcriptional regulation of Pax4 is important to our understanding of the early events in the differentiation of pancreatic b cell progenitors. Although both the mouse and human Pax4 cDNAs had been cloned recently (Inoue et al., 1998; Matsushita et al., 1998; Tao et al., 1998), little is known about the Pax4 gene promoter in either species. In this study, we isolated the 5% flanking sequence of the mouse Pax4 gene and used gene transfer and deletional analysis to characterize the Pax4 promoter in b and non-b cell lines.

2. Experimental procedures

GTCTCAGGGTC-3%, complementary to a portion of the first exon of the mouse Pax4 gene (Matsushita et al., 1998), was end-labeled with 32P using T4 polynucleotide kinase. The 32P-labeled primer, 2×105 cpm, was hybridized with 50 mg of bHC3 RNA or yeast tRNA in 30 mg of hybridization solution (40-mM PIPES, pH 6.4, 400-mM NaCl, 1-mM EDTA, 80% formamide) overnight at 37°C after denaturation of the mixture at 80°C for 10 min. After ethanol precipitation, the pellet was dissolved in 25 ml of buffer; 20-mM Tris, pH 8.4, 50-mM KCl, 2.5-mM MgCl2, 10-mM DTT, 0.4-mM dNTPs and 200 U of Superscript II (GibcoBRL, Burlington, Ont., Canada), and incubated at 45°C for 50 min. The extension products were analyzed on a 6% sequencing gel in parallel with a DNA sequence reaction using the same primer.

2.1. Library screening

2.4. DNA sequencing

A mouse liver genomic library (Clontech, Palo Alto, CA, USA) was used to clone the promoter fragment. In order to increase the screening efficiency, the phage containing Pax4 gene in the library was enriched as described previously (Israel, 1993). PCR using a sense primer, 5%-GTGAATGGCCGGCCCCTTCCTCTG, and an antisense primer, 5%-TGGGTACAAAGCCCTTCAGTGCAA, based upon the reported mouse Pax4 cDNA sequence (Sosa-Pineda et al., 1997) were used to test aliquots of the library and to generate a probe for bacteriophage plaque hybridization. The 600bp PCR fragment was confirmed by sequence analysis after cloning into pCR2.1 vector (Invitrogen, San Diego, CA, USA). This cloned fragment was used as a probe to screen the enriched library. Hybridization was carried out in 600-mM NaCl and 60-mM sodium citrate containing 50% formamide, 0.1% SDS, 2-mM EDTA, 0.125 mg/ml salmon sperm DNA, 1 mg/ml polvinypyrrolidone, 1 mg/ml bovine serum albumin, 1 mg/ml Ficoll, and the 32P-labeled 600-bp probe at 45°C overnight.

DNA fragments were sequenced on both strands using either T7 Sequenase kit according to the manufacture’s instructions (US Biochemical, Cleveland, OH, USA) or an ABI automated DNA Sequencer with a dye terminator kit (ABI, Foster City, CA, USA).

2.2. Restriction and southern blot Phage DNA (10 mg) was digested with various restriction enzymes. Digested samples were analyzed by electrophoresis in a 0.8% agarose gel and transferred onto nylon membrane. The membrane was probed with the radiolabeled 600-bp PCR fragment. Hybridization was carried out under the same conditions as the library screening.

2.3. Primer extension Total RNA, extracted from bHC3 cells using the acid guanidine/phenol –chloroform method was used as a template. An oligonucleotide, 5%-CAGGAAGAG-

2.5. DNA constructs A 5.5-kb Bgl II–Bgl II fragment which hybridized with the PCR generated probe and an overlapping 1.8-kb EcoR1–EcoR1 fragment were cloned into the BamHI or EcoR1 site of pBluescript II SK vector, respectively. The fragment, containing − 1116 to +51 bp upstream of the Pax4 transcription start site, was excised by digestion of the 5.5-kb clone with Not I and PvuII, and inserted into the Sma1 site of promoterless luciferase reporter plasmid pGL2-basic (Promega, Madison WI, USA Fig. 1) to generate the p-1116Luc construct. To generate the p-2488Luc construct, the p-1116Luc plasmid was first linearized by Spe I, then partially digested with Nsi I. The resulting 6.3-kb fragment was ligated to the 1.4-kb Spe1-Nsi1 fragment that had been excised from the 1.8-kb EcoR1–EcoR1 clone which overlapped the − 1116 bp sequence. The 5%deleted constructs were generated as shown in Fig. 1. The −681 and − 420-bp constructs were generated by cloning EcoR1–PvuII and HincII–PvuII fragments into Sma1 site of pGL2-basic, respectively. The −135 and − 42 bp constructs were generated by inserting Bsu361–BglII and HaeIII–BglII fragments of p1116Luc into the Sma1/BglII site of pGL2-basic. Various fragments were isolated from restriction digests and were cloned upstream of the SV40 promoter in the pGL2/promoter plasmid or upstream of the TK minimal promoter in the pTKLuc reporter vector. All fragments were isolated by restriction endonuclease digestion except the F10 fragment, which generated by

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

PCR amplification. The orientations of all these constructs were confirmed by restriction digestion or sequence analysis.

2.6. Cell culture The bHC3 and bHC9 cell lines were provided by Dr D. Hanahan, University of California, San Francisco. All other cell lines were purchased from the American Type Culture Collection, Manassas, VA, USA. The bHC3, bHC9 and HIT-15 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 10% fetal calf serum (FCS) at 37°C in a humidified atmosphere containing 5% CO2 whereas the NIT-1 cells were grown in F-12K medium with 10% FCS. The COS-1, Rat-1, NIH/3T3 and MCF-7 cell lines were cultured in DMEM-F12 medium with 10% FCS. All cell culture reagents were obtained from GibcoBRL (Burlington, Ont., Canada).

2.7. Transient transfection and luciferase assays The bHC3, bHC9, HIT-15 and NIT-1 cells were transfected with reporter constructs by the cationic liposome-mediated method (Pera and Kessel, 1998). NIH/3T3, COS-1, Rat-1 and MCF-7 cells were transfected by the calcium phosphate method. The plasmids, pGL2-SV40 and pGL2-basic were used as positive and negative controls, respectively. Each reporter construct DNA (2.5 mg per 35-mm dish) was used for each

81

transfection. To normalize for transfection efficiency, 0.05 mg of pCH110, a b-galactosidase expression plasmid, was used. Cell extracts were prepared 48 h after transfection and assayed in duplicate or triplicate for luciferase activity as described previously (Pera and Kessel, 1998). All transfection experiments were repeated on three or more occasions independently.

2.8. Preparation of nuclear extracts and DNase 1 footprinting Nuclear extracts were prepared from bHC3 and NIH/3T3 cells as described elsewhere (Dignam et al., 1983). The pTKLuc/F9 clone was digested with BamHI or XhoI endonucleases and labeled by filling in with Klenow enzyme and the labeled fragment was excised subsequently from the plasmid by either XhoI or BamHI, respectively, to generate coding and noncoding strand probes labeled at only one end. The radiolabeled 212 bp F9 fragment containing the sequence − 2070 to − 1858 bp was incubated with varying amounts of nuclear extract for 20 min at room temperature. Freshly diluted DNase 1 (GibcoBRL, Burlington, Ont., Canada) was added and the reaction mix was incubated at room temperature for 30 s. The reaction was stopped, extracted with phenol:chloroform and ethanolprecipitated. The DNA was resuspended in formamide loading buffer dye and resolved on a 6% denaturing sequencing gel.

Fig. 1. The restriction map of the mouse genomic clone containing the Pax4 gene promoter and the structure of the 5% end of the mouse Pax4 gene. Coding exons are shown as solid boxes whereas untranslated regions as open boxes. The various reporter constructs and their relative luciferase activities in bHC3 cells have been compared with the promoterless pGL2-basic plasmid, arbitrarily attributed a value of 1. Data, which have been corrected for transfection efficient by co-transfected with pCH110, expressing b-galactosidase, represent the mean 9S.E.M. for three or more independent experiments.

82

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

2.9. Electrophoretic mobility shift assay Oligonucleotides corresponding to the putative CdxA/Nkx2 binding site, 5%-TCTGAGTTAATGTATAATTGTGA-3%, the putative GATA-1,2 binding site, 5%-GTGAGCAGATGGCGGGGGCTGGCAG-3% and the putative ADR1 binding site, 5%-GGCAGCAGCCTGGGGGCTGGGCACA-3% and their 3%– 5% complements were synthesized. The oligonucleotides were annealed and labeled with g-32P-ATP using T4 polynucleotide kinase. Nuclear extract from bHC3 and NIH/ 3T3 was preincubated in presence or absence of competitor DNAs on ice for 15 min prior to the addition of  0.5 pmol of radiolabeled probe. The reaction mixture, in a total volume of 30 ml, containing 5 mg of nuclear protein, 40 000 cpm of radiolabeled probe, 13-mM Hepes, pH 7.9, 75-mM KCl, 1-mM MgCl2, 12% glycerol, 0.1-mM EDTA, 0.3-mM dithiothretitol, 0.1-mM PMSF and 0.5 mg poly(dI:dC), was incubated at room temperature for 20 min and loaded onto a 5% polyacrylamide gel (acrylamide/bisacrylamide, 44:1) in 0.5XTBE.

2.10. Site-directed mutagenesis PCR was used to mutate the putative transcription factor binding sites. A reverse primer complementary to the TK promoter was used together with the mutagenic primers. In addition to the mutated sequence, each primer included a BamHI site to facilitate directional cloning. The PCR was performed for 25 cycles of denaturation at 96°C for 1 min, annealing at 55°C for 1 min and extension at 72°C for 1 min. The PCR generated fragments were isolated from an agarose gel, digested with BamHI and HindIII and cloned into the pT81Luc vector. Each clone was sequenced to confirm that the expected mutation had been introduced and then tested for enhancer activity as described above.

3. Results

3.1. Cloning of the Pax4 promoter and identification of the transcription start site Two positive clones were obtained from screening about 10 000 phage plaques from the enriched mouse genomic library. Phage DNA of one clone was prepared and subjected to restriction endonuclease digestion and Southern blot analysis. The restriction map is shown in Fig. 1. The 5.5-kb Bgl II – Bgl II and the 1.8-kb EcoR1–EcoR1 fragments were cloned into pBluescript II SK vector and partial sequence was obtained. Although the sequence of a full-length mouse Pax4 cDNA has been reported previously (Matsushita et al., 1998), the genomic structures of the 5% end had

not been elucidated nor had the transcription site been reported. Ten exons and nine introns have been identified previously in mouse Pax4 gene (Inoue et al., 1998). We compared the genomic sequence with the cDNA sequence, and found another intron approximately 2 kb in length. The previously reported exon 1 containing the 5% untranslated region, the translation start site and a small amount of coding sequence appears to represent exon 1 and 2. The previously unidentified exon 1, contains 148 nucleotides of 5% untranslated sequence. Therefore, the mouse Pax4 gene consists of 11 exons and ten introns. The transcription start site was determined by primer extension (Fig. 2) and by comparison of the sequence of the genomic fragment and the cDNAs generated by 5% RACE (Inoue et al., 1998). According to our calculations, the transcription start site would be located 2.2 kb upstream of the translation start site and would result in a 5% untranslated region of 244 bp.

3.2. Sequence analysis of the Pax4 promoter The sequence of the −2488 bp 5% upstream fragment of the mouse Pax4 gene was determined and submitted to GenBank (AF118441). It lacked a classical TATA box in the proximal region but contained a potential CCAAT box at −35 bp upstream. When the mouse Pax4 promoter sequence − 2488 to +51 bp was compared with sequences in GenBank a high match, up to 86%, was found with the region 33596–36097 of the human cosmid clone AC000359 (Matsushita et al., 1998). This region is 2500 bp upstream of the human Pax4 gene coding sequences in this cosmid, indicating that the 5% structure of the mouse and human Pax4 genes are likely to be similar.

3.3. Deletion analysis of Pax4 promoter A PvuII site was identified at + 51 bp down stream of the transcription start site in exon I. The −1116 to + 51 bp BglII–PvuII fragment was subcloned into pGL2-basic. The −2488 bp construct (Fig. 1) was generated by ligation of the 1.8-kb EcoR1–EcoR1 fragment with the EcoR1 digested − 1116 bp fragment. Various 5% deletion mutants were constructed. The ability of each of these reporter constructs to promote transcription of the luciferase gene was tested in bHC3 cells. As shown in Fig. 1, the − 2488 to +51 bp fragment produced a luciferase activity of 9.8-fold above the background activity seen with transfection of the promoterless reporter plasmid. Maximal promoter activity was observed with the p-135Luc construct. The promoter activity was decreased  50% by inclusion of the sequences contained within the region −136 to − 420 bp. These data are consistent with the existence of a repressor element or elements in the − 135 and

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

83

Fig. 2. Primer extension analysis of the mouse Pax4 gene. An oligonucleotide with the sequence 5%-CAGGAAGAGGTCTCAGGGTC-3% was used with RNA from bHC3 cells as a template. The arrowhead indicates the transcription start site. The corresponding nucleotide sequence of the genomic fragment is shown. The potential CCAAT box is underlined.

− 420 bp fragment. Sequences contained in the −421 to −1116 bp fragment had little effect on promoter activity. However, the construct containing the sequence − 1117 to −2488 had activity similar to the p-135Luc promoter construct, suggesting that this region contains an enhancer element or elements which either acts directly or negate the repressor activity contained in the − 136 to −420 bp fragment.

3.4. Comparison of Pax4 promoter acti6ity in b and non-b cell lines The relative luciferase activities of the Pax4 constructs were tested in pancreatic b cell lines, bHC3, bHC9, HIT-15 and NIT-1 cells, and in the non-b cell lines, COS-1, NIH/3T3, Rat-1 and MCF-7. In order to correct the differences in transfection and promoter efficiency in different cell lines, the data were expressed as relative luciferase activity, that is, as a percentage of the activity seen in cells transfected with the pGL2-pro-

moter vector containing luciferase gene driven by the SV40 promoter. As a negative control, the promoterless pGL2-basic plasmid was also tested. In each of the b cell lines, the relative luciferase activity of the various reporter construct was greater consistently than in the non-b cell lines (Fig. 3). For example, the relative luciferase activity of the p-1116Luc reporter construct was approximately 25–30% in the b cell lines and was lower considerably in the non-b cell lines; 1.1190.56, 1.549 0.15, 7.749 3.73 and 17.39 3.7% for COS-1, MCF-7 Rat-1 and NIH/3T3 cells, respectively. An additional difference observed between the b cell and non-b cell lines was the absences of repressor activity in the − 136 to − 430 bp fragment in the non-b cell lines (Fig. 3). Furthermore, in the b cell lines, the p-2488Luc construct had significantly higher promoter activity than the p-1116Luc construct. The opposite was true in the case of the non-b cell lines. Because of this difference, we chose to examine the − 1117 to − 2488 bp region in more detail.

84

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

3.5. Identification of a beta cell specific upstream enhancer The various fragments from the region − 1116 to − 2488 bp were subcloned upstream of the heterologous SV40 promoter and tested in b and non-b cell lines (Fig. 4). Comparison of fragments F1, F2, F3 and F4 indicated that the b cell specific enhancer was likely to reside in the region −2070 to − 1494 bp. Fragments F7 and F9 containing the sequences −2070 to −1692 bp and −2070 to −1860 bp, respectively, were found to enhance SV40 promoter activity 12.8and 7.6-fold in bHC3 cells compared with an approximately 2-fold enhancement in the NIH/3T3 cells. The enhancer activities of these fragments were also tested using another heterologous promoter, the thymidine kinase promoter (Fig. 5). An even greater b cell specific enhancer effect was observed using this promoter. The enhancer activity of the F7 fragment was apparent in both forward and reverse orientations. The F10 fragment, which contained the sequence − 1954 to − 1859 bp enhanced the activity of the TK promoter almost 200-fold in bHC3 cells but had only a minimal effect in NIH/3T3 cells (Fig. 6). RT-PCR was used to demonstrate expression of Pax4 in bHC3 cells. The predicted 1-kb fragment from the

coding region of Pax4 was observed when RNA from bHC3 cells was used as the substrate. No Pax4 band was apparent when RNA from NIH/3T3 cells was used whereas b-actin was amplified easily in both cell lines (data not shown).

3.6. Analysis of the upstream Pax4 promoter region by DNase 1 footprinting Nuclear extracts from both bHC3 and NIH/3T3 cells were analyzed in a footprinting assay using the F9 fragment containing the sequence − 1858 to −2070 bp. DNase 1 protection was apparent with nuclear extracts from both cell lines. However, differences in the fragmentation patterns were observed consistently with nuclear extracts from the two cell lines in four regions (Fig. 7). Since deletional analysis suggested that the b-cell specific enhancer resided in the sequence − 1954 to − 1859 bp (fragment F10, Fig. 6), we chose to concentrate on this region. Two protected regions, R2 and R3 resided within this fragment. For these two regions, DNase 1 protection was apparent on both the coding and non-coding strands but differences between the nuclear extracts from the two cell lines was marked particularly on the coding strand.

Fig. 3. Comparison of the promoter activities of various deletion reporter constructs in various cell lines. The transfection efficiency was corrected in same cell lines by co-transfection with pCH110, and in different cell lines by comparison to the SV40 positive promoter control. The relative activity has been expressed as a percentage of the activity seen with the SV40-driven luciferase construct pGL2-promoter plasmid in each cell line. Note the different scales used with different cell lines. Data represent the mean 9S.E.M. for three or more independent experiments.

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

85

Fig. 4. Comparison of the enhancer activities of various fragments of the Pax4 promoter using the heterologous SV40 promoter. Fragments were subcloned upstream of the SV40 minimal promoter in pSV40Luc and tested in various cell lines. Data represent the mean9 S.E.M. for three or more independent experiments and have been expressed relative to activity seen with the pSV40Luc plasmid arbitrarily attributed a value of 1 in each cell line. N.A. indicates not assayed.

Fig. 5. Comparison of the enhancer activities of the F7 fragments of the Pax4 promoter using the heterologous TK promoter. The fragment was subcloned in both orientations upstream of the TK promoter and tested in various cell lines. Data represent the mean 9S.E.M. for three or more independent experiments and have been expressed relative to activity seen with the pTKLuc plasmid arbitrarily attributed a value of 1 in each cell line.

3.7. Analysis of the upstream Pax4 promoter region by EMSA Overlapping oligionucleotides probes corresponding to the F10 fragment containing regions R2 and R3, identified in the DNase 1 footprinting assay were synthesized and analyzed in a gel shift assay using nuclear

extracts from bHC3 and NIH/3T3 cells (Fig. 8). Nuclear extract from bHC3 cells retarded probe A, corresponding to − 1952 to − 1929 bp (lane 2, Fig. 8). The specificity of the bands observed was verified by the competition with 10- and 100-fold excess unlabeled probe (lanes 3 and 4) whereas no competition was apparent with the non-specific oligonucleotide. Nuclear

86

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

extract from NIH/3T3 cells also caused some retardation of the DNA but the apparent size of the complex was smaller than that seen with nuclear extract from bHC3 cells and the intensity of the bands observed were much less. Probe B, corresponding to − 1930 to − 1905 bp, showed a different pattern of interaction with nuclear extract from bHC3 and NIH/3T3 cells. A number of non-specific retarded bands was apparent with extract from bHC3 cells (data not shown). Intense specific interactions were observed between nuclear extract from bHC3 cells and probe C corresponding to sequence − 1913 to −1888 bp and containing R3, (Fig. 8, lower panel). There was a weak, more rapidly migrating band apparent with nuclear extracts from NIH/3T3 cells, which was attenuated by unlabeled probe (Fig. 8, lower panel, right arrow).

3.8. Mutagenesis of the upstream Pax4 enhancer element The DNA sequence of the F10 fragment was analyzed for known transcription factor binding sites using TFSEARCH software with a score threshold of 85%. The R2 region contained a putative CdxA site, an adjacent CdxA/Nkx2 binding site and a GATA-1/-2 binding site. The matrix score for these transcription factor binding sites was 98.6, 87.2 and 92.9, respectively. The R3 region contained a putative ADR1 binding site (matrix score 98.5). All of these sequences with the exception of the ADR1 site were conserved in the human Pax4 promoter sequence (Fig. 9). The putative recognition sequences for these transcription factors

were analyzed by site-directed mutagenesis. Each of these regions was mutated and the mutants were tested for their ability to enhance the TK promoter in bHC3 cells and NIH/3T3 cells. Disruption of either the R2 region markedly reduced the enhancer activity in bHC3 cells but had no significant effect on the weak enhancer activity observed in the NIH/3T3 cells (Fig. 9). In contrast, disruption of the R3 region had minimal effect on the enhancer activity in bHC3 cells but, interestingly, had a significant effect on the small amount of enhancer activity observed in NIH/3T3 cells.

4. Discussion The development of the endocrine pancreas involves a complex interaction between various transcription factors as the progenitor cells differentiate into mature endocrine cells. These include Nkx-2.2 and Pax4 (SosaPineda et al., 1997; Sussel et al., 1998). Disruption of either of these genes results in failure of development of mature functional islets. The interaction between these genes and other homeodomain genes involving the development of the endocrine pancreas has as yet received little attention but is clearly important in our understanding of the differentiation of b cell progenitors. Homologous disruption of the Pax4 gene results in morphologically normal mice, which lack mature b and d cells but have abundant glucagon producing a cells present in disordered islet-like clusters (Sosa-Pineda et al., 1997). Although mature b cells are absent in the Pax4 null mutant mice, insulin staining cells are appar-

Fig. 6. Comparison of the enhancer activities of F7 and subfragments of the Pax4 promoter using the heterologous TK promoter. The F7 fragment and smaller subfragments of the promoter were tested in the NIH/3T3 and bHC3 cell lines. Data represent the mean9S.E.M. for three or more independent experiments and have been expressed relative to activity seen with the pTKLuc plasmid arbitrarily attributed a value of 1 in each cell line.

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

87

promoter indicated that multiple enhancer/repressor elements are likely to reside in the proximal 5% region of this gene. While a minimal promoter fragment containing the proximal 135 bp gave maximal activity, constructs containing an additional  300 bp of 5% sequence had approximately half of this activity (Fig. 1). These data are consistent with the presence of a repressor element or elements in the region −136 to −420 bp. However, constructs containing the sequences −1116 to −2488 had similar activity to the minimal promoter suggesting that this more distal region contained an element, which either enhanced transcription directly or was able to overcome the repressor activity present in the region − 136 to − 420 bp. Both tissue-specific enhancers and repressor elements are probably involved in the restriction of expression of Pax4 to different subsets of pancreatic endocrine cells. We have demonstrated the presence of an upstream b cell specific enhancer element in the Pax4 gene. This resides in a 62-bp fragment, which contains putative recognition sites for three groups of transacting factors. The DNase 1 footprinting and gel retardation experiments indicating that there are nuclear proteins present

Fig. 7. DNase 1 footprinting analysis of the enhancer region of the Pax4 promoter. The F9 fragment was labeled on the coding and noncoding strand and incubated with 10 or 20 mg of nuclear protein from bHC3 and NIH/3T3 cells and then digested with DNase 1. Differences in the banding pattern between nuclear extracts from the two cell lines was observed in three regions.

ent at E10.5 in the primordial pancreas (Sosa-Pineda et al., 1997) indicating that maturation, rather than commitment, of b and d cell progenitors was defective in these mice. Functional differentiation of a cells is preserved in these Pax4 null mutant mice possibly because of redundancy or alternative pathways. In this regard, it is of interest that Pax6 null mutant mice have defective a cell differentiation (St-Onge et al., 1997). In this manuscript, we have determined the transcription start site and the structure of the 5% end of the Pax4 gene and demonstrated the existence of the previously unidentified exon 1. Deletion analysis of the Pax4 gene

Fig. 8. Electrophoretic mobility shift assay of the enhancer region of the Pax4 promoter. Nuclear extract from bHC3 or NIH/3T3 cells was incubated with radiolabeled double stranded oligonucleotide probes in the presence or absence of 10- or 100-fold unlabeled probe or non-specific oligonucleotide (NS). Probe A corresponds to the putative CdxA and CdxA/Nkx2 binding sites. Probe C corresponds to the putative ADR1 binding site. The NS competitor nucleotides used in the upper and lower panel were unlabeled probe B and probe A, respectively.

88

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

Fig. 9. Site-directed mutagenesis of the enhancer region of the Pax4 gene. The F10 fragment and various mutants were tested for their ability to enhance the TK promoter activity in bHC3 and NIH/3T3 cells. The upper panel shows the sequence of F10 and the mutants tested. For comparison, the sequence of the human Pax4 gene promoter is shown. In the lower panel, the relative luciferase activity of the F10 fragments and its mutants subcloned into the pTKLuc plasmid is shown. Data represent the mean 9 S.E.M. for three or more independent experiments and have been expressed relative to activity seen with the pTKLuc plasmid arbitrarily attributed a value of 1 in each cell line.

in b cell extracts capable of interacting with these sequences. The R2 region at the 5% end of the 62-bp fragment contained putative recognition sites for the CdxA and Nkx-2 group of trans-acting factors as an incomplete palindrome, GTTAATGTATAATTG. Sitedirected mutagenesis indicated that this region was important in the b cell specific enhancement of promoter activity. Furthermore, nuclear extracts from the b cell line, which expressed Pax4, but not the fibroblast cell line was able to retard this fragment. The R2 region also contained a putative binding site for the GATA-1,-2 group of transcription factors, which was also important in the b cell enhancer effect. A 10-fold reduction in the enhancer activity was observed when either the CdxA/Nkx-2 or the GATA-1,-2 region was mutated. The Nkx homeodomain proteins are a large family of transcription factors with individual members of this family involved in the differentiation of various cell lineages. For example, Nkx 2.1 is involved in differentiation of thyroid and pulmonary epithelium (Pera and Kessel, 1998) while Nkx 2.5 is essential for differentiation of the dorsal mesoderm into cardiac tissue (Tanaka et al., 1999). Targeted disruption of the Nkx 2.2 gene in mice results in neonatal hyperglycemia and death (Sussel et al., 1998). These mice lack insulin-staining b cells and have reduced numbers of a and PP cells in their islets. The Nkx 2.2 null mutant mouse islets also con-

tain a large number of cells, which do not stain for any hormone but do express some b cell markers and, presumably, represent incompletely differentiated b cells progenitors. Thus, like Pax4, Nkx 2.2 is required for differentiation of islet progenitor cells into mature b cells. Interestingly, an interaction between Nkx 2.5 and GATA-4 has been demonstrated in expression of the murine A1 adenosine receptor in cardiac tissue (Rivkees et al., 1999) and Nkx 2.5 and GATA-4 have been shown to be mutual cofactors in transcription of atrial natriuretic factor in cardiac cells (Shiojima et al., 1999). Our data indicate that the two regions containing the putative CdxA/Nkx.2 and GATA-1,-2/myoD binding sequences are both important in b-cell type-specific expression of mouse Pax4 gene since mutagenesis of either one attenuates the b-cell enhancer effect. Since this region is identical in the human Pax4 gene sequence, it is likely that a similar situation pertains to Pax4 expression in humans. CdxA is a chicken homeobox-containing gene related to caudal in Drosophila and its expression extends medially and caudally in the chick embryo as formation of the gut tube progresses (Ishii et al., 1997). It is not expressed in the developing pancreatic duct in the chick (Ishii et al., 1997). The murine equivalents include three genes related to caudal, Cdx-1, Cdx-2/3 and Cdx-4. Interestingly, Cdx-2/3 is involved in proglucagon gene expression both in islets and intestinal endocrine cells

W. Xu, L.J. Murphy / Molecular and Cellular Endocrinology 170 (2000) 79–89

(Laser et al., 1996; Jin et al., 1997). It is possible that the tissue specific enhancer sequence that we have identified in the Pax4 promoter binds a member of the Cdx family rather than Nkx 2.2. Unlike Nkx 2.2 knock-out mice, Cdx-1 knock-out mice are viable and are not diabetic (Subramanian et al., 1995). Cdx-2/3 null mutants die between 3.5 and 5.5 days of embryonic life and while heterozygotes exhibit a variable phenotype, diabetes does not appear to be a feature (Chawengsaksophak et al., 1997). The phenotypic manifestations Cdx-4 disruption have not been reported yet. Further studies are needed to determine the exact identity of the factors which interact with the b-cell specific enhancer regions of the Pax4 gene. In addition, the functional interaction between these two regions and the transcription factors, which bind to them, warrants further investigation.

Acknowledgements This research was supported by the Canadian Diabetes Association. LJM is a recipient of an MRC Senior Scientist award and an endowed Research Professorship in Metabolic Diseases.

References Ahlgren, U., Pfaff, S.L., Jessell, T.M., Edlund, T., Edlund, H., 1997. Independent requirement for ISL1 in formation of pancreatic mesenchyme and islet cells. Nature 385, 257–260. Alpert, S.D., Hanahan, D., Teitelman, G., 1988. Hybrid insulin genes reveal a devlopmental lineage for pancreatic endocrine cells and imply a relationship with neurons. Cell 53, 295–308. Chawengsaksophak, K., James, R., Hammond, V.E., Kontgen, K., Beck, F., 1997. Homeosis and intestinal tumours in Cdx2 mutant mice. Nature 386, 84–87. Dahl, E., Koseki, H., Balling, R., 1997. Pax genes and organogenesis. Bioessays 19, 755 – 765. Dignam, J.D., Lebovitz, R.M., Roeder, R.G., 1983. Accurate transcription by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucl. Acids Res. 11, 1475–1489. Gittes, G.K., Rutter, W.J., 1992. Onset of cell-specific gene expression in the developing mouse pancreas. Proc. Natl. Acad. Sci. USA 89, 1128 – 1132. Inoue, H., Nomiyama, J., Nakai, K., Matsutani, A., Tanizawa, Y., Oka, Y., 1998. Isolation of full-length cDNA of mouse PAX4 gene and identification of its human homologue. Biochem. Biophys. Res. Commun. 243, 628–633. Ishii, Y., Fukuda, K., Saiga, H., Matsushita, S., Yasugi, S., 1997. Early specification of intestinal epithelium in the chicken embryo: a study on the localization and regulation of CdxA expression. Dev. Growth Differ. 39, 643–653. Israel, D.I., 1993. A PCR-based method for high stringency screening of DNA libraries. Nucleic Acids Res. 21, 2627–2631.

.

89

Jin, T., Trinh, D.K., Wang, F., Drucker, D.J., 1997. The caudal homeobox protein cdx-2/3 activates endogenous proglucagon gene expression in INR1-G9 islet cells. Mol. Endocrinol. 11, 203 – 209. Laser, B., Media, P., Constant, I., Philippe, J., 1996. The caudal-related homeodomain protein Cdx-2/3 regulates glucagon gene expression in islet cells. J. Biol. Chem. 271, 28984 – 28994. LeDouarin, N.M., 1988. On the origin of pancreatic endocrine cells. Cell 53, 169 – 171. Matsushita, T., Yamaoka, T., Otsuka, S., Moritani, M., Matsumoto, T., Itakura, M., 1998. Molecular cloning of mouse paired-boxcontaining gene (Pax)-4 from an islet cell line and deduced sequence of human Pax-4. Biochem. Biophys. Res. Commun. 242, 176 – 180. Naya, J.F., Huang, H.P., Qiu, Y., Mutoh, H., DeMayo, F.J., Leiter, A.B., Tsai, M.J., 1997. Diabetes, defective pancreatic morphogenesis, and abnormal enteroendocrine differentiation in BETA2/ neuroD-deficient mice. Genes Dev. 11, 2323 – 2334. Offield, M.F., Jetton, T.L., Labosky, P.A., Ray, M., Stein, R.W., Magnuson, M.A., Hogan, B.L.M., Wright, C.V.E., 1996. PDX-1 is required for pancreatic outgrowth and differentiation of the rostral duodenum. Development 122, 983 – 995. Pera, E.M., Kessel, M., 1998. Demarcation of ventral territories by the homeobox gene NKX2.1 during early chick development. Dev. Genes Evol. 208, 168 – 171. Rivkees, S.A., Chen, M., Kulkarni, J., Browne, J., Zhao, Z., 1999. Characterization of the murine A1 adenosine receptor promoter, potent regulation by GATA-4 and Nkx2.5. J. Biol. Chem. 274, 14204 – 14209. Shiojima, I., Komuro, I., Oka, T., Hiroi, Y., Mizuno, T., Monzen, K., Aikawa, R., Akazawa, H., Yamazaki, T., 1999. Context-dependent transcriptional cooperation mediated by cardiac transcription factors Csx/Nkx-2.5 and GATA-4. J. Biol. Chem. 274, 8231 – 8239. Sosa-Pineda, B., Chowdhury, K., Torres, M., Oliver, G., Gruss, P., 1997. The Pax4 gene is essential for differentiation of insulin-producing beta cells in the mammalian pancreas. Nature 386, 399– 402. St-Onge, L., Sosa-Pineda, B., Chowdhury, K., Mansouri, A., Gruss, P., 1997. Pax6 is required for differentiation of glucagon-producine alpha-cells in mouse pancreas. Nature 387, 406 –409. Subramanian, V., Meyer, B.I., Gruss, P., 1995. Disruption of the murine homeobox gene Cdx1 affects axial skeletal identities by altering the mesodermal expression domains of Hox genes. Cell 83, 641 – 653. Sussel, L., Kalamaras, J., Hartigan-O’Connor, D.J., Meneses, J.J., Pedersen, R.A., Rubenstein, J.L., German, M.S., 1998. Mice lacking the homeodomain transcription factor Nkx2.2 have diabetes due to arrested differentiation of pancreatic beta cells. Development 125, 2213 – 2221. Tanaka, M., Chen, Z., Bartunkova, S., Yamasaki, N., Izumo, S., 1999. The cardiac homeobox gene Csx/Nkx2.5 lies genetically upstream of multiple genes essential for heart development. Development 126, 1269 – 1280. Tao, T., Wasson, J., Bernal-Mizrachi, E., Behn, P.S., Chayen, S., Duprat, L., Meyer, J., Glaser, B., Permutt, M.A., 1998. Isolation and characterization of the human PAX4 gene. Diabetes 47, 1650 – 1653. Walter, C., Guenet, J.L., Simon, D., Deutsch, U., Jostes, B., Goulding, M.D., Plachov, D., Balling, R., Gruss, P., 1991. Pax: amurine multigene family of paired box-containing genes. Genomics 11, 424 – 434.