Cloning of the Promoters for the -Cell ATP-Sensitive K-Channel Subunits Kir6.2 and SUR1 Rebecca Ashfield and Stephen J.H. Ashcroft

The -cell ATP-sensitive potassium channel (K-ATP channel), which regulates insulin secretion, is composed of two types of subunits: 1) a sulfonylurea receptor (SUR1) and 2) an inwardly rectifying potassium channel (Kir6.2). We have isolated clones containing 5 flanking DNA for both genes by hybridization screening of a human genomic library. Sequencing of over one kilobase of each upstream region has revealed that the putative promoters are G + C rich, with no TATA box. Several E-boxes and potential Sp1 sites are present in both promoters, and the Kir6.2 upstream region contains an Alu repeat. Using a luciferase reporter gene in transient transfection assays, we demonstrate that the upstream DNA contains promoters that are active in the -cell lines HIT T15 and MIN6. The promoters are completely inactive in the fibroblast cell line COS7 but show some activity in HepG2 (liver) and HEK293 (epithelial) cell lines. Deletion analysis suggests that a short (173–base pair [bp]) fragment of SUR1 5 -flanking sequence is sufficient for maximal promoter activity. In contrast, over 900 bp of Kir6.2 5 sequence are required for similar high level expression, and deletion of the Alu repeat results in an increase in promoter activity. Diabetes 47:1274–1280, 1998

A

TP-sensitive potassium channels (K-ATP channels) are found in excitable tissues, where they couple changes in cellular metabolism to electrical activity (1). In pancreatic -cells, the channels regulate insulin secretion; a rise in blood glucose increases the ATP:ADP ratio inside the cell, which in turn closes the K-ATP channels, resulting in membrane depolarization, Ca2+ influx, and insulin secretion (1). The K-ATP channel is a target for sulfonylureas, such as glibenclamide and tolbutamide, used in the treatment of NIDDM (2). In addition, mutations in both channel subunits give rise to the hereditary disease persistent hyperinsulinemic hypoglycemia of infancy (PHHI) (3). From the Nuffield Department of Clinical Biochemistry, John Radcliffe Hospital, Headington, Oxford, U.K. Address correspondence and reprint requests to Dr. Rebecca Ashfield, Nuffield Department of Clinical Biochemistry, John Radcliffe Hospital, Headington, Oxford OX3 9DU, U.K. E-mail: [email protected]. Received for publication 8 December 1997 and accepted in revised form 13 April 1998. ABC, ATP-binding cassette; bp, base pair; CFTR, cystic fibrosis transmembrane conductance regulator; CMV, cytomegalovirus; HNF, hepatocyte nuclear factor; K-ATP channel, ATP-sensitive potassium channel; kb, kilobase; Kir6.2, inwardly rectifying potassium channel; nt, nucleotide; PHHI, persistent hyperinsulinemic hypoglycemia of infancy; PKA, protein kinase A; PKC, protein kinase C; SUR, sulfonylurea receptor. 1274

K-ATP channels are a complex of two types of subunits: 1) an inwardly rectifying potassium channel, a member of the Kir6.x family, and 2) a sulfonylurea receptor (SUR), which belongs to the ATP-binding cassette (ABC) transporter family (4–7). The channel is believed to be an octamer, containing four subunits of each type (8,9). Different combinations of Kir6.x and SUR are possible, depending on the particular members of each family expressed in a given tissue. The cell K-ATP channel is composed of the subunits Kir6.2 and SUR1 (5,6), while cardiac and skeletal muscle channels contain SUR2A, encoded by a separate gene (7). Kir6.2 is the poreforming subunit; it is inactive in the absence of SUR1 and is inhibited by binding ATP (10). The regulatory subunit SUR1 binds antidiabetic sulfonylureas, resulting in channel closure and insulin secretion, and channel openers, such as diazoxide, which inhibit insulin secretion (4,11). PHHI is often caused by mutations in SUR1 that result in a failure to couple to, and activate, Kir6.2 (3). The genes encoding human Kir6.2 and SUR1 have been cloned, and they lie adjacent to one another on chromosome 11p15.1. The genes are very closely spaced, with only 4.5 kilobases (kb) separating the 3 end of SUR1 and the 5 end of Kir6.2 (4). Transcriptional regulation of the two genes is important for correct tissue-specific expression and to ensure that equal amounts of each subunit are produced. A channel containing one (out of four) Kir6.2 subunit not coupled to an SUR1 subunit has no activity (8,9), so an excess of Kir6.2 would produce inactive K-ATP channels containing four Kir6.2 but less than four SUR1 subunits. It is therefore possible that mutations in the SUR1 promoter resulting in lower mRNA and protein levels relative to Kir6.2 could inactivate K-ATP channels and cause PHHI. From Northern blots, SUR1 RNA is expressed only at high levels in brain and -cells, while Kir6.2 is expressed both in these tissues and in heart and skeletal muscle (4–6). We have cloned the promoter regions for SUR1 and Kir6.2 and shown them to be only partially responsible for tissue-specific expression. The promoters resemble those for ubiquitously expressed housekeeping genes, being G + C rich with no TATA box. Consensus sequences for several transcription factors are present, including Sp1 and E-boxes, which bind a family of helix-loop-helix proteins (12,13). RESEARCH DESIGN AND METHODS Molecular biology. The human genomic library, containing 3.8 106 independent clones with an average insert size of 15 kb, was obtained from Clontech. A total of 6 105 clones were screened with mouse Kir6.2 cDNA and the 5 248–base pair (bp) BamHI fragment of hamster SUR cDNA (corresponding to exon 1 and part of exon 2 of the human gene). Hybridization was at 37°C with 40% formamide, and filters were washed in 0.1% SDS, 0.1% sodium chloride–sodium citrate at room temDIABETES, VOL. 47, AUGUST 1998

R. ASHFIELD AND S.J.H. ASHCROFT

perature for 20–60 min. Four positively hybridizing clones were isolated for each probe, and DNA was isolated from these according to the Clontech genomic library protocol. Southern blots were performed on a range of restriction digests of each positive clone. Hybridizing fragments (a 4-kb HindIII-XbaI SUR1 and 3.5kb BamHI + 2.2-kb XhoI Kir6.2 fragments) were subcloned into pBluescript and sequenced. Over 1 kb of 5 -flanking DNA was isolated for each gene. Putative promoter fragments from the pBluescript constructs, containing some of the 5 -untranslated region of each gene, were cloned into pGL2-basic (Promega, Madison, WI), containing the firefly luciferase gene with no 5 promoter. Fragments were cloned as 1.4-kb XhoI-BstEII (blunted) Kir6.2 and 1.1-kb XhoIPvuII SUR1 fragments into pGL2-basic cut with XhoI and HindIII (blunted) to give pGL2-Kir6.2 and pGL-SUR1, respectively. Deletions were made by cutting these constructs with the following enzymes, filling in if necessary, and re-ligating (having removed the excised fragments): pGL-Kir6.2 was cut with SmaI to give pGLKir6.2 1, XhoI + SacII to give pGL-Kir6.2 2, and BstXI to give pGL-Kir6.2 Alu (BstX1 cuts at each end of the Alu repeat); pGL-SUR1 was cut with XhoI + AvrII to give pGL-SUR1 1 and MluI to give pGL-SUR1 2. Identification of transcription factor binding sites. A search for possible transcription factor binding sites was performed using the SIGNAL SCAN program available from the U.K. Human Genome Mapping Project software package. TATA and CCAAT boxes were identified using a general transcription factor database, while the binding sites listed in Table 1 were compiled to form a separate customized database. Primer extension analysis. Poly A+ RNA was obtained from human islets (provided by Dr. Roger James, Department of Surgery, University of Leicester) using an Oligotex Direct mRNA kit (Qiagen, Crawley, U.K.). Probes were made by 5 end labeling synthetic oligonucleotides with [ 32P]ATP (Du Pont-NEN, Boston, MA; 3,000 Ci/mmol, 10 mCi/ml). The SUR1 probe was complementary to the sequence 20–40 nucleotides (nt) downstream of the human SUR1 cDNA 5 end (CGCCCGGCCCGGCCGTCAGG) and the Kir6.2 probe to the sequence 20–48 nt downstream of the human Kir6.2 cDNA 5 end (TCCTTACCTCCACCTGGGTCCCACTTCA). Probes were annealed to 1–2 µg of poly A+ RNA at 70°C for 10 min and were cooled to 50°C, and reverse transcription was carried out for 1 h at 50°C in a volume of 25 µl in the presence of 20 mmol/l Tris-HCl (pH 8.4), 50 mmol/l KCl, 2.5 mmol/l MgCl2, 10 mmol/l dithiothreitol, 400 µmol/l each of dATP, dCTP, dGTP, and dTTP, 40 U RNasin, 50 µg/ml actinomycin D, and 200 U Superscript II reverse transcriptase (Life Technologies, Paisley, U.K.). The product was treated with RNase for 15 min at 37°C, concentrated, and loaded onto a 6% polyacrylamide gel. Transfections. DNA constructs were transiently transfected into cell lines together with a cytomegalovirus (CMV)-luciferase control plasmid (pRL-CMV; Promega; containing the sea pansy luciferase gene) as follows. COS7 and HIT T15 cells were trypsinized, spun down, and resuspended in complete RPMI 1640 medium at 1–2.5 107 cells/ml. After 5–10 min on ice, 0.4-ml aliquots were mixed with the DNA (20 µg of test and 10 µg of control plasmid) in 0.4-cm Bio-Rad (Richmond, CA) cuvettes and left for 5 min at room temperature. The cells plus DNA were remixed and electroporated at 960 µF, 250 V using a Bio-Rad Gene Pulser. After 25 min at room temperature, cells were transferred to 6- (HIT) or 10-cm (COS7) dishes containing prewarmed media. For MIN6, HEK293, and HepG2 cells, 1.5 µg of test and 0.5 µg of control DNA were transfected into cells plated on 3.5-cm dishes, 30–70% confluent, using 10 µl lipofectAMINE (Life Technologies) according to the manufacturer’s protocol; liposomes were left on cells for 5 h before being removed and replaced with 2 ml complete medium. After 48–72 h, cells were harvested and assayed with the Dual Luciferase Reporter Assay kit (Promega), which assays the test promoter luciferase (firefly) and CMV control luciferase (sea pansy) in the same cuvette. Test luciferase levels are expressed as test luciferase/control luciferase activities (arbitrary units) 103.

RESULTS

Human SUR1 and Kir6.2 genes possess G + C–rich promoter regions. The promoter regions of Kir6.2 and SUR1 were cloned from a human genomic library, using the Kir6.2 cDNA and exons 1–2 of SUR1 as probes (see METHODS for details). Over 1 kb of 5 -flanking DNA was sequenced for each gene; both upstream regions are G + C rich, 69% for Kir6.2 and 55% for SUR1. The putative SUR1 promoter (Fig. 1) has several E-boxes (binding helix-loop-helix proteins) (12), Sp1 sites (14), and a G-box similar to that in the insulin promoter, which binds the ubiquitous transcription factor MAZ/Pur-1 (15). In addition, it has predicted AP2 sites (16), two inverted CCAAT boxes, and a TATA element, although the latter is almost 1 kb 5 of exon 1. The putative Kir6.2 promoter DIABETES, VOL. 47, AUGUST 1998

FIG. 1. The sequence of 1132 bp of SUR1 5 -flanking DNA, shown in lowercase letters. Numbering is relative to the transcriptional start site, obtained from the human SUR1 cDNA sequence (GenBank accession no. L78207) and from primer extension analysis (Fig. 3). Transcribed DNA is shown in uppercase lettering, and the ATG translation start site is shown in bold. Putative regulatory elements identified using a homology search with the consensus sequences shown in Table 1 are boxed. Elements in the sense orientation are labeled above each box, and in the reverse orientation beneath each box. Restriction enzyme sites used to generate deletion constructs are indicated.

(Fig. 2) also has several E-boxes, G-boxes, Sp1 sites, and possible AP2 sites (16); there are no CCAAT or TATA boxes. It has an Alu repeat in the 306 bp immediately 5 of the transcription start site predicted from the cDNA sequence. A search for likely consensus sequences for transcription factors involved in -cell tissue-specific expression (Table 1) showed that both promoters contain possible binding sites for Pdx-1, IB1, PEA3, Pax6, and hepatocyte nuclear factor (HNF)-3 (12,13,17–22) (a description of individual factors is given in DISCUSSION). The predicted sites are shown in both orientations, and some lie in regions of DNA that may be deleted without affecting short-term promoter activity (see below). The promoter sequences have been submitted to GenBank and given the accession numbers AF053477 (Kir6.2) and AF053478 (SUR1). Mapping the transcriptional start sites for SUR1 and Kir6.2. Primer extension was performed to determine the transcriptional start sites of SUR1 and Kir6.2 (Fig. 3). The 1275

CLONING OF K-ATPCHANNELPROMOTERS

TABLE 1 Consensus transcription factor binding sites used in a homology search of the SUR1 and Kir6.2 promoters Element Sp1 E-box AP2 G-box PEA3 Pdx-1 Pax6 HNF-3 IB1 (Ripe3b)

Consensus (reference) GGGCGG (14) CANNTG (12) YCSCCMNSSS (17) Purine (G/A) rich, G>A (32) AGGAAG (33) TAAT (12) TCACGC(WTSANTKM) (34) (A)WTRTTK(R)(Y)T(Y) (35) CCTCAGCC (32)

Sites were obtained from the references given in parentheses after each sequence. Nucleotides in parentheses were not included in the search but are shown because they form part of the consensus site. N = A/C/G/T; Y = C/T; S = C/G; M = A/C; W = A/T; R = A/G; and K = G/T.

FIG. 2. The sequence of 1189 bp of Kir6.2 5 -flanking DNA, shown in lowercase letters. Numbering is relative to the transcriptional start site predicted from primer extension analysis (Fig. 3). Transcribed DNA is shown in uppercase lettering, and the ATG translation start site is shown in bold. The 5 end of a human cDNA clone (accession no. D50582) is indicated by an arrowhead (at +12). Putative regulatory elements identified using a homology search with the consensus sequences shown in Table 1 are boxed. Elements in the sense orientation are labeled above each box, and in the reverse orientation beneath each box. Restriction enzyme sites used to generate deletion constructs are indicated, and an element highly homologous to Alu repeats is underlined.

SUR1 primer was extended by 20 nt, which corresponds to the start site obtained from sequencing of a human SUR1 cDNA clone (GenBank accession no. L78207) (Fig. 1). The Kir6.2 primer was extended by 32 nt, 12 nt further than expected from the human cDNA sequence (GenBank accession no. D50582); this corresponds to a CA dinucleotide (where A is the first transcribed nucleotide). We attribute the presence of multiple shorter products to the pausing of reverse transcriptase rather than to additional start sites of transcription because of the high GC content of these sequences. The 5 -flanking regions contain promoters that confer partial tissue specificity. The putative SUR1 and Kir6.2 promoters were placed upstream of a luciferase reporter 1276

gene, and transient transfection assays were performed in a variety of cell lines. Figure 4 shows the promoter activities obtained, relative to a control CMV-luciferase construct that is active in all cell lines tested. The two sequences we used, both over 1 kb in length, gave a high level of reporter gene expression in the -cell lines HIT T15 and MIN6; the SUR1 sequence was more active than Kir6.2 in MIN6 and less active in HIT T15 cells. Both promoters were completely inactive in the cell line COS7 (monkey kidney fibroblast). The human cell lines HepG2 (liver) and HEK293 (epithelial, embryonic kidney) gave an intermediate level of expression. The 5 deletions define minimal promoter sequences. To define the amount of 5 -flanking sequence required for maximal promoter activity, a number of 5 truncations and deletions were made in the SUR1 and Kir6.2 promoters. The deleted promoters were tested for their ability to drive luciferase expression in MIN6 cells. As shown in Fig. 5A, truncation of the SURI promoter to 173 bp did not result in any decrease in luciferase activity (lane SURI 2, 3.9 arbitrary units compared with 3.5 arbitrary units for the full-length promoter, lane SURI), suggesting that this sequence contains all the promoter elements required for high level expression in -cells. Indeed, there was a small increase in promoter activity when distal sequences were removed, so these may contain a negative regulatory element. The same observations were made in both HIT T15 and HepG2 cells (data not shown). The 173bp minimal promoter sequence includes two consensus Sp1 sites and three E boxes, and it is 74% G + C rich. Figure 5B shows that, in contrast, Kir6.2 promoter activity requires at least 915 bp of sequence. Truncation of the promoter from 1189 to 915 bp resulted in a small decrease in promoter activity (compare lane Kir6.2, 1.1 arbitrary units, and lane Kir6.2 1, 0.96 arbitrary units), while further truncation to 734 bp virtually abolished promoter activity (compare lane Kir6.2 2, 0.19 arbitrary units, and the promoterless control, 0.13 arbitrary units). These data suggest that 915 bp of the Kir6.2 upstream sequence contain most of the elements required for maximal promoter activity, although additional elements may reside 5 of this region. Again, HIT T15 and HepG2 cells gave a similar result (data not shown). Figure 5B, lane Kir6.2 Alu, shows that deletion DIABETES, VOL. 47, AUGUST 1998

R. ASHFIELD AND S.J.H. ASHCROFT

A

FIG. 4. Luciferase activities of plasmids containing a luciferase gene and no promoter (control), the 1132-bp SUR1 putative promoter (SUR1) and the 1189-bp Kir6.2 putative promoter (Kir6.2) after transient transfection into the cell lines indicated. The activities are expressed relative to a CMV-luciferase control plasmid. The data shown are for typical transfections; repeat experiments gave similar results.

DISCUSSION

B

FIG. 3. A: Primer extension analysis of human islet mRNA. The Kir6.2 primer (lane Kir6.2) spans 20–48 bp 3 of the start site predicted from the cDNA sequence, and the SUR1 primer (lane SUR1) spans 20–40 bp 3 of the predicted start site. Arrows indicate the fully extended products, assumed to represent the 5 ends of Kir6.2 and SUR1 mRNAs. (A longer exposure did not reveal any bands corresponding to start sites further 5 ). Molecular weight markers are shown in lane M. B: A plot of distance moved versus log10 (marker length), from which the lengths of extended Kir6.2 and SUR1 products were calculated.

of the Alu repeat immediately 5 of the Kir6.2 gene increases promoter activity by a factor of 1.5. It therefore appears that the 300 bp just upstream of the Kir6.2 gene contain no essential promoter elements, and that the Alu repeat actually inhibits Kir6.2 transcription. DIABETES, VOL. 47, AUGUST 1998

Electrophysiological studies show that K-ATP channels are present in -cells, brain, heart, and smooth and skeletal muscle in all mammals studied, although the sensitivity to various sulfonylureas and channel openers differs between tissues (1,2). There appear to be three main subtypes of K-ATP channel: in brain/ -cell, heart/skeletal muscle, and smooth muscle (1). Binding studies with radiolabeled sulfonylurea compounds show that binding follows the tissue distribution of the channels (1,2). Recent cloning of K-ATP channel subunits has confirmed these tissue-specific differences, with the brain/ -cell type of channel composed of Kir6.2 and SUR1, and the heart/skeletal muscle type reconstituted by combining Kir6.2 and SUR2A (5–7). An alternatively spliced variant of SUR2A, SUR2B, forms a smooth muscle type of KATP channel (23). Northern blot analysis shows that the -cell lines HIT T15 (hamster), RINm5F (rat), and MIN6 (mouse) and the -cell line TC-6 (mouse) all express SUR1 at high levels, as do rat pancreatic islets. Smaller amounts are found in brain, and very low levels are found in heart and skeletal muscle (rat tissues). All other tissues tested, e.g., mouse liver and rat kidney, do not express SUR1 RNA at a level detectable by Northern blot (4,5). Kir6.2 is expressed at high levels in - and -cell lines and rat islets, heart, brain, and skeletal muscle; there is weak expression in kidney and none in liver (5,6). No data are available on the distribution of SUR1 or Kir6.2 RNA in human tissues. The results presented here show that the human SUR1 and Kir6.2 promoters drive expression of a reporter gene in -cell lines, but not in a monkey kidney fibroblast cell line (COS7). However, the data obtained from HepG2 (liver) and HEK293 (kidney epithelial) cells indicate that the promoter sequences are not entirely tissue specific, as SUR1 and Kir6.2 are not expressed in liver or epithelium, and yet there was reporter gene expression in these cell lines. There must therefore be sequences distal to these core promoters that are required for full tissue specificity, such as enhancer or silencer elements (although as SUR1 and Kir6.2 are not expressed in exactly the same tissues, the two promoters are unlikely to share the 1277

CLONING OF K-ATPCHANNELPROMOTERS

A

B

FIG. 5. A: The upper diagram represents the SUR1 promoter region and exon 1; numbering is relative to the transcriptional start site. Below are various promoter constructs, with differing amounts of 5 -flanking DNA (as indicated) driving a luciferase reporter gene. Each construct contains 93 bp of the 5 -untranslated region of SUR1. The control construct has no promoter. To the right of each construct is the luciferase activity obtained after transient transfection into MIN6 cells, expressed relative to a CMV-luciferase control. B: Same as for A, with the Kir6.2 promoter replacing that for SUR1. Each construct contains 214 bp of the Kir6.2 5 -untranslated region. A typical transfection is shown; repeat experiments gave similar results.

same enhancer or silencer). The observation that promoter deletions gave similar results in -cell lines and HepG2 (liver) cells suggests that deleted regions do not contain sequences important in tissue-specific expression and supports the hypothesis that distal control elements exist. In this respect, the SUR1 and Kir6.2 basal promoters resemble those for the potassium channel genes Kv1.3 and Kv1.4 (24,25). The precise promoter activities obtained are cell line dependent, as HIT T15 and MIN6 are both -cell lines but do not give identical results; the SUR1 promoter is more active than Kir6.2 in MIN6, but not in HIT T15 cells. Although the ratio of SUR1:Kir6.2 subunits is 1:1 in K-ATP channels (8,9), it is not surprising that these small differences exist in nonhuman cell lines when measuring reporter gene expression 1278

in transient transfection assays. Transcription factor levels may differ between cell lines, and translation efficiency and/or protein stability of SUR1 and Kir6.2 may contribute to determining the final ratio of subunits in vivo. Both promoters are of the “housekeeping” type, as they lack a TATA box in the minimal promoter region and are G + C rich, containing multiple Sp1 sites (GGGCGG) (14). Housekeeping genes are ubiquitously expressed, but an increasing number of tissue-specific genes appear to have this type of promoter, including that for cystic fibrosis transmembrane conductance regulator (CFTR), an ABC protein related to SUR1 that requires only 102 bp for maximal activity (26). The promoters for the potassium channel genes Kv1.3 and Kv1.4 are also G + C rich and TATA-less and, as mentioned above, require DIABETES, VOL. 47, AUGUST 1998

R. ASHFIELD AND S.J.H. ASHCROFT

additional elements for tissue-specific expression (24,25). The transcription start sites for SUR1 and Kir6.2 were determined by primer extension analysis. For Kir6.2, transcription begins at a CA dinucleotide, which is often the case for TATAless promoters, as this forms part of the initiator sequence, Py C A N TA Py Py. The SUR1 start site does not correspond to an initiator element. The two promoters contain several E-boxes (CANNTG), known to be important for insulin gene expression. These boxes bind a family of helix-loop-helix proteins; the insulin gene E-boxes are believed to bind E2A gene products and NeuroD/ -2 (13). The family member(s) bound by a particular Ebox is determined mainly by the two central nucleotides. For example, MyoD binds CAGCTG and E47 (an E2A protein) binds CACCTG (27). The minimal 173-bp SUR1 promoter possesses three E-boxes, two of which are the CACTTG type found in the insulin promoter, while the remaining one, and the majority of the Kir6.2 E-boxes, has the sequence CAGGTG. The SUR1 promoter does contain a consensus TATA box at –980; however, this is not involved in basal promoter activity because it lies 800 bp upstream of the minimal 173bp promoter. There are two inverted CCAAT elements (at –765 and –574), but again, these are 5 of the minimal promoter. Although not involved in basal promoter activity, it is possible that they mediate transactivation by protein kinase A (PKA), which has been shown to activate the CFTR promoter through an inverted CCAAT box (28). The SUR1 promoter also contains potential AP2 sites that are responsive to both PKA and protein kinase C (PKC) (16). It is therefore possible that SUR1 transcription is activated by either PKA or PKC. The Kir6.2 promoter has no TATA or CCAAT box, but it contains potential AP2 sites, so again, it is possible that the promoter can be activated by PKA or PKC. In addition, the SUR1 and Kir6.2 promoters contain a number of possible binding sites for several tissue-specific factors. Pdx-1 and IB1 are -cell–specific transcription factors that transactivate a number of -cell genes including insulin (13,17); Pax6 is a homeodomain/paired domain protein important in islet development (21). HNF-3 is related to HNF-1 and HNF-4 , transcription factors that are found to be mutated in maturity-onset diabetes of the young (MODY) (22,29,30). PEA3 belongs to the Ets family, and it is expressed in multiple tissues and is regulated by several protein kinases (18,19). Further studies will be necessary to determine which, if any, of these factors are important for promoter activity in vivo. For example, Pdx-1 binds to sites including the conserved motif TAAT, but nucleotides surrounding this motif are also involved in binding, and it is not possible to predict simply from the DNA sequence whether Pdx-1 will bind at a particular location. It appears that binding sites upstream of the minimal 173-bp SUR1 promoter are not required for transient expression in -cell lines, although it cannot be excluded that the sites are involved in vivo, e.g., regulating gene expression during development or in response to external stimuli. The most important elements regulating SUR1 promoter activity are likely to be the two Sp1 sites, G-box (MAZ/Pur-1 binding site), and three E-boxes (possibly binding NeuroD/E2A proteins), which reside in the 140 bp immediately 5 of exon 1. The Kir6.2 gene promoter appears to be more complex, requiring sequences up to 1 kb 5 of the gene, although the most proximal 300 bp can be deleted with no reduction in DIABETES, VOL. 47, AUGUST 1998

activity. Deletion of sequences immediately 5 of most genes removes crucial elements, resulting in abolition of promoter activity, so the Kir6.2 promoter is highly unusual in this respect. This sequence encompasses an Alu repeat, a 300-bp element related to the 7SL RNA gene, which is found scattered throughout the human genome (~105 copies) and is actively transcribed by RNA polymerase (31). It is possible that transcription of the Kir6.2 Alu element interferes with initiation of RNA polymerase II on the Kir6.2 promoter, explaining why deletion of the repeat actually increases Kir6.2 promoter activity. The region of DNA important for Kir6.2 promoter activity therefore lies upstream of the Alu repeat and contains several Sp1 sites, G-boxes, and E-boxes; the E-boxes may not bind the same proteins as the insulin and SUR1 promoter Eboxes because they differ in sequence. The contribution of these sequence elements to regulation of Kir6.2 transcription will need to be determined by mutagenesis and in vitro binding studies. It will be of particular interest to determine what is responsible for conferring tissue specificity on both the SUR1 and Kir6.2 promoters, which outwardly resemble those for ubiquitously expressed genes. ACKNOWLEDGMENTS

This work was supported by grants from the Wellcome Trust and the British Diabetic Association. The U.K. Human Genome Mapping Project provided software for DNA sequence manipulation. We thank Dr. Roger James, Department of Surgery, University of Leicester, U.K., for provision of human islets of Langerhans. REFERENCES 1. Ashcroft SJH, Ashcroft FM: Properties and functions of ATP-sensitive Kchannels. Cell Signal 2:197–214, 1990 2. Ashcroft SJH, Ashcroft FM: The sulfonylurea receptor. Biochim Biophys Acta Mol Cell Res 1175:45–59, 1992 3. Thomas PM, Cote GJ, Wohllk N, Haddad B, Matthew PM, Rabl W, Aguilar-Bryan L, Gagel RF, Bryan J: Mutations in the sulfonylurea receptor gene in familial persistent hyperinsulinemic hypoglycemia of infancy. Science 268:426–429, 1995 4. Aguilar-Bryan L, Nichols CG, Wechsler SW, Clement JP, Boyd AE III, González G, Herrera-Sosa H, Nguy K, Bryan J, Nelson DA: Cloning of the -cell high-affinity sulfonylurea receptor: a regulator of insulin secretion. Science 268:423–426, 1995 5. Inagaki N, Gonoi T, Clement JP, Namba N, Inazawa J, Gonzalez G, AguilarBryan L, Seino S, Bryan J: Reconstitution of IKATP: an inward rectifier subunit plus the sulfonylurea receptor. Science 270:1166–1170, 1995 6. Sakura H, Ämmälä C, Smith PA, Gribble FM, Ashcroft FM: Cloning and functional expression of the cDNA encoding a novel ATP-sensitive potassium channel subunit expressed in pancreatic -cells, brain, heart, and skeletal muscle. FEBS Lett 377:338–344, 1995 7. Inagaki N, Gonoi T, Clement JP, Wang CZ, Aguilar-Bryan L, Bryan J, Seino S: A family of sulfonylurea receptors determines the pharmacological properties of ATP-sensitive K+ channels. Neuron 16:1011–1017, 1996 8. Clement JP, Kunjilwar K, Gonzalez G, Schwanstecher M, Panten U, AguilarBryan L, Bryan J: Association and stoichiometry of KATP channel subunits. Neuron 18:827–838, 1997 9. Inagaki N, Gonoi T, Seino S: Subunit stoichiometry of the pancreatic -cell ATPsensitive K+ channel. FEBS Lett 409:232–236, 1997 10. Tucker SJ, Gribble FM, Zhao C, Trapp S, Ashcroft FM: Truncation of Kir6.2 produces ATP-sensitive K+ channels in the absence of the sulphonylurea receptor. Nature 387:179–183, 1997 11. Ämmälä C, Moorhouse A, Ashcroft FM: The sulphonylurea receptor confers diazoxide sensitivity on the inwardly rectifying K+ channel Kir6.1 expressed in human embryonic kidney cells. J Physiol (Lond) 494:709–714, 1996 12. Clark AR, Docherty K: Cell-specific gene expression in the islets of Langerhans: E boxes and TAAT boxes. Biochem Soc Trans 21:154–159, 1993 13. Madsen OD, Jensen PB, Larsson LI, Serup P: Transcription factors con1279

CLONING OF K-ATPCHANNELPROMOTERS

tributing to the pancreatic -cell phenotype. Horm Metab Res 29:265–270, 1997 14. Kadonaga JT, Jones KA, Tijan R: Promoter-specific activation of RNA polymerase II transcription by Sp1. TIBS 11:20–23, 1986 15. Kennedy GC, Rutter WJ: Pur-1, a zinc-finger protein that binds to purine-rich sequences transactivates an insulin promoter in heterologous cells. Proc Natl Acad Sci U S A 89:11498–11502, 1992 16. Imagawa M, Chiu R, Karin M: Transcription factor AP-2 mediates induction by two different signal-transduction pathways: protein kinase C and cAMP. Cell 51:251–260, 1987 17. Waeber G, Steinmann M, Nicod P, Bonny C: The -cell–specific control of the insulin and GLUT2 genes is mediated, in part, by the islet-brain 1 (IB1) transactivator (Abstract). Exp Clin Endocrinol Diabetes 105:A23, 1997 18. Ohagan RC, Tozer RG, Symons M, McCormick F, Hassell JA: The activity of the Ets transcription factor PEA3 is regulated by two distinct MAPK cascades. Oncogene 13:1323–1333, 1996 19. deLaunoit Y, Baert JL, Chotteau A, Monte D, Defossez PA, Coutte L, Pelczar H, Leenders F: Structure-function relationships of the PEA3 group of Etsrelated transcription factors. Biochem Mol Med 61:127–135, 1997 20. Skak K, Jensen J, Michelsen BK: Localisation of the rat GAD65 promoter: expression kinetics of GAD in the pancreas (Abstract). Exp Clin Endocrinol Diabetes 105:A31, 1997 21. Sander M, Neubüser A, Kalamaras J, Ee HC, Martin GR, German MS: Genetic analysis reveals that PAX6 is required for normal transcription of pancreatic hormone genes and islet development. Genes Dev 11:1662–1673, 1997 22. Vaisse C, Kim J, Espinosa R III, Le Beau MM, Stoffel M: Pancreatic islet expression studies and polymorphic DNA markers in the genes encoding hepatocyte nuclear factor-3 , -3 , -3 , -4 , and -6. Diabetes 46:1364–1367, 1997 23. Isomoto S, Kondo C, Yamada M, Matsumoto S, Higashiguchi O, Horio Y, Matsuzawa Y, Kurachi Y: A novel sulfonylurea receptor forms with BIR (Kir6.2) a smooth muscle type ATP-sensitive K+ channel. J Biol Chem 271:24321–24324, 1996 24. Simon M, Conley EC, Shelton PA, Gutman GA, Chandy KG: Transcription of the T-cell potassium channel Kv1.3 is regulated by a GC-rich TATA-less promoter. Cell Physiol Biochem 7:243–250, 1997 25. Wymore RS, Negulescu D, Kinoshita K, Kalman K, Aiyar J, Gutman GA, Chandy KG: Characterization of the transcription unit of mouse Kv1.4, a volt-

1280

age-gated potassium channel gene. J Biol Chem 271:15629–15634, 1996 26. Yoshimura K, Nakamura H, Trapnell BC, Dalemans W, Pavirani A, Lecocq J-P, Crystal RG: The cystic fibrosis gene has a “housekeeping”-type promoter and is expressed at low levels in cells of epithelial origin. J Biol Chem 266:9140–9144, 1991 27. Blackwell TK, Weintraub H: Differences and similarities in DNA binding preferences of MyoD and E2A protein complexes revealed by binding site selection. Science 250:1104–1110, 1990 28. Pittman N, Shue G, LeLeiko NS, Walsh MJ: Transcription of cystic fibrosis transmembrane conductance regulator requires a CCAAT-like element for both basal and cAMP-mediated regulation. J Biol Chem 270:28848–28857, 1995 29. Yamagata K, Oda N, Kaisaki PJ, Menzel S, Furuta H, Vaxillaire M, Southam L, Cox RD, Lathrop GM, Boriraj VV, Chen X, Cox NJ, Oda Y, Yano H, Le Beau MM, Yamada S, Nishigori H, Takeda J, Fajans SS, Hattersley AT, Iwasaki N, Hansen T, Pedersen O, Polonsky KS, Turner RC, Velho G, Chèvre J-C, Froguel P, Bell GI: Mutations in the hepatocyte nuclear factor-1 gene in maturity-onset diabetes of the young (MODY3). Nature 384:455–458, 1996 30. Yamagata K, Furuta H, Oda N, Kaisaki PJ, Menzel S, Cox NJ, Fajans SS, Signorini S, Stoffel M, Bell GI: Mutations in the hepatocyte nuclear factor-4 gene in maturity-onset diabetes of the young (MODY1). Nature 384:458–460, 1996 31. Weiner AM, Deininger PL, Efstratiachis A: Nonviral retroposons: genes, pseudogenes, and transposable elements generated by the reverse flow of genetic information. Annu Rev Biochem 55:631–661, 1986 32. German M, Ashcroft S, Docherty K, Edlund H, Edlund T, Goodison S, Imura H, Kennedy G, Madsen O, Melloul D, Moss L, Olson K, Permutt MA, Philippe J, Robertson BP, Rutter WJ, Serup P, Stein R, Steiner D, Tsai MJ, Walker MD: The insulin gene promoter: a simplified nomenclature. Diabetes 44:1002–1004, 1995 33. Martin ME, Piette J, Yaniv M, Tang W-J, Folk WR: Activation of the polyomavirus enhancer by a murine activator protein (AP1) homolog and two contiguous proteins. Proc Natl Acad Sci U S A 85:5839–5843, 1988 34. Epstein J, Cai J, Glaser T, Jepeal L, Maas R: Identification of a Pax paired domain recognition sequence and evidence for DNA-dependent conformational changes. J Biol Chem 269:8355–8361, 1994 35. Overdier DG, Porcella A, Costa RH: The DNA-binding specificity of the hepatocyte nuclear factor-3/forkhead domain is influenced by amino acid residues adjacent to the recognition helix. Mol Cell Biol 14:2755–2766, 1994

DIABETES, VOL. 47, AUGUST 1998