Characterization of the chromosomal gene and promoter for human ...

Vol. 269,No. 14,Issue of April 8,pp. 10891-10898, 1994 Printed in U.S.A.

THEJOURNAL OF BIOLOGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Characterization of the Chromosomal Gene and Promoter for Human Insulin-like Growth Factor BindingProtein-5” (Received for publication, September 24, 1993, and in revised form, December 29, 1993)

other five members of the IGFBPfamily (1,4).Additional studTo better understand the regulation of insulin-like growth factor binding protein-5 (IGFBP-5) expression, ies found that IGFBP-5 was expressed by a wide variety of we cloned theIGFBP-5 gene from human genomic librartissues (1,4). Similar to other IGFBPs, IGFBP-5 may potentiies and identified a region in the 5’ flanking sequence ate (2) or inhibit ( 5 , 6)IGF action, with different effects probwhich functions as a promoter. The human IGFBP-6 gene ably dependingon post-translational modifications of IGFBP-5 is divided into four exons which, primarily due to a and firston the specific cell targets of IGF action. intron of-25 kilobases, span -33 kilobases of DNA. The ability of IGF peptides to affect a target tissue may be Southern analysis identified a single copy of the IGFBP-5 regulated by multiple mechanisms, among them the local regugene in the haploid human genome, and several indelation of tissue IGFBP levels. In the case of IGFBP-5, this pendentmapping strategies found thisgenetightly mechanism may be important since physiologic andor hormolinked with, and in opposite transcriptional orientation nal influences strongly regulate IGFBP-5 expression in certain to, the IGFBP-2 gene at chromosomal region 2q33-34. or cells (7-11). In some instances IGFBP-5 protein levtissues Primer extension studies identified the IGFBP-5 mRNA els appear to be regulatedby post-translational events (9,11). cap site 772 base pairs (bp) 5’ to the first nucleotide of the However, in other cases IGFBP-5 expression is clearly regutranslation start codon. Analysis of the 5”flanking sequence identified a potential TATA element beginning 33 lated at the level of mRNA abundance (7-11). Thus, factors bp 5’ to the mRNA cap site. When a DNA fragment con- regulating IGFBP-5 mRNA abundance in a particular tissue taining this cap site and 461 bp of upstream sequencemay wasprofoundly influence the effect of IGF peptides on that placed 5’ to the chloramphenicol acetyltransferase re- tissue. Multihormonal regulation of IGFBP-5 mRNA levels may be porter gene and transfected into MDA-MB-468 human regubreast cancer cells, it directed chloramphenicol acetyl-controlled at the level of gene transcription, similar to the transferase expression in an orientation-specific manlation of IGFBP-1 expression by multiple hormones (7-15). As ner, suggesting that this region contains elements essena first step toward analyzinghow IGFBP-5 gene transcription tial forIGFBP-5 promoter activity. may be regulated by hormonal and other influences, we have characterized the gene organization and mRNA cap site for human IGFBP-5 (hIGFBP-51, and we have identified a region Insulin-like growth factor binding protein-5(IGFBP-5)’ was originally identified in andpurified from rat serum (11, human bone (21, and medium conditioned by human osteoblast-like cells (3) as a protein which bound IGF peptides with high affinity. Isolation and characterization of human and rat IGFBP-5 cDNA clones confirmed that thesequence of this binding protein is unique but homologous to sequences encoding the

of 5”flanking sequence with structural and functional characteristics typical for a eukaryotic gene promoter. EXPERIMENTALPROCEDURES General Methods-Methods used are from Maniatis et al. (16) unless

otherwise stated. Oligonucleotidesforsequencing,primerextension, and polymerase chain reaction (PCR) amplifications were synthesized by National Biosciences(Plymouth,MN), DNA International (LakeOs* This work was supportedby Caroline Weiss Law Karolinska-Baylor wego, OR), Symbicom AB (Umel, Sweden),or the Departmentof CliniResearch Fellowship (to S. V. A.), National Institutes of Health Grant cal Genetics, KarolinskaInstitute (Pharmacia Gene Assembler). cDNA Clones-The hIGFBP-5 cDNA clones BP-6.1 and BP-6.12 (4) R 0 1 DK-38773(to D. R. P.), the Swedish Medical Research Council, the Swedish Cancer Foundation,the Swedish Medical Society,the Magnus were used as hybridization probes and for DNA sequencing. Baaed on Bergvall Foundation, the Claes Grochinsky Foundation, the Nilssonconsensus nomenclature for IGFBPs (17), these cDNA clones will be Ehle Foundation, “Svenska Sallskapet for Medicinsk Forskning“ and referred to here as BP-5.1 and BP-5.12, respectively. The 5.4-kb BP-5.12 “Forenade Liv“ Mutual Group Life Insurance Company. The costs of clone was cleaved with BgZII; cDNA fragments of 3.8, 1.0, and 0.6 kb publication of this article were defrayedin part by the paymentof page were isolated and used as hybridization probes after 32P-labeling. The charges. This article must therefore be hereby marked “advertisement” 3.8-kb 3”fragment was subcloned into pSP73 and then sfquenced by in accordance with 18 U.S.C. Section 1734 solelyto indicate this fact. the dideoxy chain termination method (18). The nucleotide sequence(s) reported in this paper hasbeen submitted Isolation of Genomic Clones-The 1.7-kb cDNA clone BP-5.1(4)was to the GenBankmIEMBL Data Bank with accession number(s) L27556labeled with 32P and used to screen a human leukocyte genomiclibrary 27560. /I Present address:LXR Biotechnology, Inc., 1401Manna Way South, constructed in phage EMBL 3 (Clontech, Palo Alto, CA) as described previously (19). A single AIGFBP-5 clone, hgBP5-17, was plaque puriRichmond, CA 94804. ** To whom correspondence should be addressed: Texas Children’s fied and subjected to Southern analysis using as probe a 32P-labeled, Hospital, Clinical Care Center, MC#3-2482,6621 Fannin St., Houston, 455-bp BgZII fragment of the BP-5.1 cDNA clone, which contains 43 bp of 5’-untranslated sequence and the first 408 bp of hIGFBP-5-coding TX 77030. The abbreviations used are: IGFBP-5, insulin-like growth factor sequence. Fragments of interest were subcloned, and DNA sequence binding protein-5;h, human; PCR, polymerase chain reaction; kb, kilo- analysis was performed as described above. base(s);bp, base paifis); PFGE, pulsed field gel electrophoresis; CAT, The genomic clone phBP5-BS6.6, derived from the AIGFBP-5 clone chloramphenicol acetyltransferase. hgBP5-17, was 32P-labeled and used as hybridization probe to screen a

10891

10892

hIGFBP-5 Gene and Promoter

human cosmid (pWE15)librarypreparedfromplacental DNA (20) was 180x 180 s for 20 h and then 90 x 90 s for a further 20 h. Southern which was kindlyprovided by Dr. Glen Evans, theSALK Institute, San analysis was performedas described above, using 32P-labeled5' and 3' Diego, CA. Cosmid DNA was isolatedfrom purified positive clones by a regions of the hIGFBP-2 and -5 genes in succession to probe the same mini-lysate procedure (21) and subjected to Southern analysis using the filter. AnXbaIISmaI fragment from the phIGFBP2-XE6.4 subclone (26) three BglII fragments from theBP-5.12 cDNA as probes. Selected cos- was the Fj"IGFBP-2 probe, the phBP5E4.6 subclone was the 5'-IGmid fragments hybridizing with the 32P-labeledBglII probes were sub- FBP-5 probe, and PCR products from cosmid clone chBP2-2.4 (26) and cloned in pGEM3Zff+), and areas of interest in the subclones were subclone phBP5-E2.5 served as 3'-IGFBP-2 and 3'-IGFBP-5 probes, sequenced as described above. respectively. Before rehybridization, the filter was incubated in 0.1 x Southern Analysis-Total human DNA was isolated from peripheral SSC, 0.1% SDS for 3 x 10 min at 100 "C and autoradiographed to leukocytes by standard methods (22). Samplesof total human DNA, of document removal of previous probe. cosmid and A genomic DNA, or of DNA from somatic human-rodentcell Plasmid Construction-Plasmid phBP5-HBl1.1, derived from AIGhybrids and from normal Chinese hamster and mouse cells (NIGMS FBP-5 clone hgBP5-17, was cleaved with HindIII and PstI. A 484-bp Coriell Cell Repositories) were cleaved with restriction endonucleases; fragment spanning from -461 to 23 bp relative to themFWA cap site cleaved DNA was electrophoresed and transferred t o filters, and the was inserted in the sense orientation at the PstI and HindIII sites 5' to filters prehybridized, hybridized, and autoradiographed as described the bacterial chloramphenicol acetyltransferase (CAT) reporter gene in (22). Filters were washed 0.1 in x SSC, 0.1% SDS at 65 "C after hybrid- the promoterless pCAT(An) plasmid (27), creating p461CAT. The proization to cDNA or genomic probes, and washed in 6 x SSC at 65 "C moter fragment was then cut ofout p461CAT with BamHI and inserted after hybridization to oligonucleotide probes. into the pCAT(An) cloning cassette at the same BamHI sitesto create PCR Analysis-PCR used 0.25 p~ of each primer, 0.2 m~ of each the antisense promoter construct p461CAT(AS). Both constructs were dNTP, and 1.5 units of Taq-polymerase (Perkin-Elmer Cetus) in 45 mM sequenced to determine orientation. KCl, 1.0 mM MgCI,, 10 mM Tris-HCI, pH 8.4 at 70 "C, 0.1% Tween 20; Cell Culture and DNAZkansfection-MDA-MB-468 human breast cancer cells, obtained from Dr. Douglas Yee, University of Texas Health final volume was 50 pl. PCR products, analyzed in and isolated from agarose gels, were used to estimate the size of introns or of the 3'Science Center at San Antonio, were grown in Dulbecco's modified Eauntranslated region of IGFBP-5 or as 32P-labeled probes. gle's medium containing 10%fetal calf serum ina 5% CO, atmosphere To determine chromosomal localization, PCR was performed in the (28). For transfection, MDA-MB-468 cells were seeded a t 1-2 x lo6 above buffer; 50 ng of human DNA, 50 ng of hamster DNA, and 75 ng cells/60-mm plate and transfected 18 h later by calcium phosphate of DNA from somatic human-rodentcell hybrids serveda s templates in precipitation (29) as described (30). Each plate of cells was transfected separate reactions, whileoligonucleotides BP5-9 and BP5-18 (see Figs. with 10 pg of plasmid DNA, and 3 pgof pRSVL plasmid containing the 1 and 3) served as primers for each reaction. Samples were incubated Rous sarcoma virus promoter 5' to theluciferase gene was cotransfected for 3 min at 95"C followed by 35 cycles at 95 "C for 1 min, 55 "C for 1 to control for transfection efficiency (13,30, 31). After incubation in min, and 72"C for 3 min. PCR products were separated in1% agarose serum-containing mediumfor 22 h, cells were collected and theircytosol gels next toDNA length markers. assayed for protein content, and for CAT and luciferase activity, a s Fluorescent in Situ Hybridization-Slides with human metaphase described (32). chromosomes, prepared using lymphocyte cultures from normal conChloramphenicol Acetyltransferase and Luciferase Assays-CAT astrols (23), were prehybridized and hybridized as described (24); plassays were performed according to the method of Gorman et al. (33), and of de Wet et al.(31), mids phBP5-E4.6, phBP5E2.5, and phBP5-E5.6 (Fig. 2) were usedas luciferase assays were performed after the method probes after labeling with biotin-16-dUTP (Boehringer Mannheim) by as previously described (30). nick translation. The three plasmids werepooled in a mixture of 50% Primer Extension of Native IGFBP5-mRNA-MDA-MB-468 cells, formamide, 2x SSC, 1% Tween 20,10% dextran sulfate,1pg of human seeded a t a density of 2 x lo6 cells/100 mm plate, were harvested after Cot-I DNA (Life Technologies, Inc.), and 9 pg of salmon sperm DNA a 40-h incubation in serum-containing medium; total RNA was then were denaturedat 75 "C for 5 min and then preannealed a t 37 "C for 30 isolated using the acid guanidinium thiocyanate-phenol-chloroformexmin. Hybridization was performed at 37 "C overnight. The signal was traction method (34). Primer extension reactions were performed using made fluorescent, amplified, and the chromosomes counterstained as a kit (Promega) and following the recommendations of the manufacdescribed (23). Results were analyzedin a confocal laser scanning mi- turer except for modifications mentioned below. Fifty pg of MDA-MBcroscope (Leica) followed by destaining of the metaphasechromosomes 468 totalRNA was denatured at85 "C for 5 min andmixed with 1pmol and subsequent quinacrinefluorescence by quinacrine (QFQ)-banding. of oligonucleotide primer 5:PE(Fig. 6) which had been end-labeled with 32Pby T4-kinase. The mixture was incubatedat 80 "C for 5 min, cooled PulsedField Gel Electrophoresis (PFGE) andPFGE ProbesCultured lymphoblastoidcells from two normal humans were embedded over 15 min to 55 "C, and then kept at 27"C for 10 min. The 5:PE in agarose plugs and proteinase K treated as described (25) and then primer was extended a t 42 "C for 30 min using 30 units of avian myeloblastosis virus reverse transcriptase per reaction. The sample was cleaved with restriction enzymes BssHII, EagI, MluI, NaeI, NarI, NotI, 0.5 M EDTA followed by ethanol NruI, SacII, orSfiI. Fragments and size markers were separated usingfurther treated with RNase A and PFGE as described (25); PFGE was performed a t 170 V and pulse time precipitation as described (19),and was then analyzed ona 6.5% poly2342 ACATGTGCAT ATTTCATTCC CCAGGCAGAC ATTTTTTAGA AATCAATALA T G C C C C M T A TTGGAAAGAC TTGTTCTTCC ACGGTGACTA CAGTACATGC TGAAGCGTGC CGTTTCAGCC TCAATTTGTA AGTAGCGCAC GAGCCTCTGT GGGGGAGGAT AGGCTGAAAA AAAAMGTGG GCTCGTIZP ATCTACAGGACTCCATATAG TCATATATAG GCATATAAAT 2462 CTC-T 2502 CTATGCTTTT TCTTTGTTTT TTTCTTTCTT CCTTTCTTTC AAAGGTTTGC A T T M C T T T T CAAAGTAGTT CCTATAGGGG CATTGAGGAG CTTCCTCATT CTGGGAAAAC TGAGAAAACC 2702 CATATTCTCC T M T A C M C C CGTAATAGCA TTTTTGCCTG CCTCGAGGCA GAGTTTCCCG TGAGCAATAA ACTCAGCTTT TTTGTGGGGC ACAGTACTGG ATTTGACAGT GATTCCCCAC 2822 GTGTGTTCAT CTGCACCCAC CGAGCCAGGC AGAGGCCAGC CCTCCGTGGT GCACACAGCA CGCGCCTCAG TCCATCCCAT TTTAGTCTTT AAACCCTCAG GAAGTCACAG TCTCCGGACA 2942 CCACACCACA TTGAGCCCAA CAGGTCCACG ATGGATCCAC CTAGTCCCAC CCCAGCCTTT TTCTTTCATC TGAACAGAAT GTGCATTTTT GGRAGCCTCC CTCACTCTCC ATGCTGGCAG 3062 AGCAGGAGGG AGACTGMGT MGAGATGGC AGAGGGAGAT GGTGGCAAAA AGGTTTAGAT GCAGGAGAAC AGTAAGATGG ATGGTTCCGG CCAGAGTCGA TGTGGGGAGG RACAGAGGGC 3102 TGAAGGGAGA GGGGGCTGAC TGTTCCATTC TAGCTTTGGC ACMAGCAGC AGAAAGGGGG M A A G C C A A T AGAAATTTCC TTAGCTTCCC CACCATATGT ATTTTCRTGG ATTTGAGAGG 3302 AAAGAGAGGA AMTGGGGGA ATGGGTTGCA M A T A G A A A T G A G C T T M T C CAGGCCGCAG AGCCAGGGRA GGTGAGTAAC CTTAGGAGGG TGCTAGACTT TAGAAGCCAG ATAGGAAGAA 3422 TCAGTCTAAA CTGGCCATGC TTTGGMGGG ACAAGACTAT GTGCTCCGCT GCCCACCTTC AGCCTGCAAT GAGGGACTGA GGCCCACGAG TCTTTCCAGC TCTTCCTCCA TTCTGGCCAG 3542 TCCCTGCATC CTCCCTGGGG TGGAGGATGG A A G G M A G C T GGGACAAGCA GGGAACGCAT GATTCAGGGA TGCTGTCACT CGGCAGCCAG ATTCCGAAAC TCCCATTCTC CAATGACTTC 3662 C T C A A C C M T GGGTGGCCTT GTGACTGTTC TTTAAGGCTG AAGATATCCA GGAAAGGGGG CTTGGACACT GGCCAAGGAG ACCCCTTCGT GCTGTGGACA CAGCTCTCTT CACTCTTTGC 3182 TCATGGCATG ACACAGCGGA GACCGCCTCC AACAACGAAT TTGGGGCTAC G M G A G G A A T A G C G A M M G CAAATCTGTT TCAACTGATG GGAACCCTAT AGCTATAGAA CTTGGGGGCT 3902 ATCTCCTATG CCCCTGGACA GGACAGTTGG CTGGGGACAG GAGAAGTGCT CAATCTTCAT GAGACRAAGG GGCCCGATCA AGGCAGCCAC AAGGCCTTGA CCTGCCGAGT CAGCATGCCC 4022 CATCTCTCTC GACAGCTGTC CCCTAAACCC M C T C A C G T T TCTGTATGTC TTAGGCCAGT ATCCCAAACC TCTTCCACGT CACTGTTCTT TCCACCCATT CTCCCTTTGC ATCTTGAGCA 4142 GTTATCCAAC TAGGATCTGC CRAGTGGATA CTGGGGTGCC ACTCCCCTGA G A A M G A C T G A G C C A G G M C TACAAGCTCC CCCCACATTC CTCCCAGCCT GGACCTAATT CTTGAGAGGG 4262 GCTCTCTCTT CACGGACTGT GTCTGGACTT TGAGCAGGCT TCTGCCCCTT GCGTTGGCTC TTTGCTGCCA GCCATCAGGT GGGGGATTAG AGCCTGGTGT AAGTGCGCCA GACTCTTCCG 4382 GTTTCCAAAG TTCGTGCCTG CGAACCCAAA CCTGTGAGTC TCTTCTGCAT GCAGGAGTTT CTCCTGGGCA GCTGGTCACT CCCCAGAGAA GCTGGGCCTT CATGGACACA TGGAACTAAG 4502 CCTCCCAAAT GGGAGTTCTG GCTGAGCCCA GGGTGGGGAG ATCCTGGGAA GGGAGGCACT GGAGGAAGAC GGCACCTCTT CCCCCATGGC AGGGTGTGAG GGAGGCAGGT TTGGAATGGT 4622 GCGAGTATGG CAATCTAAGC AGGGGTCTGG TCTCTTTGAC TCCAGGCTCG CTTTGGCCGA CTGTCTGCTC ACCCAGAGAC CTTGGACTCC GGACTATCCA TGGCTCCGAA TCTAAGTGCT 4142 GCCCACTCCC ATGCTCACAC CCACAGAAGG TCTTCCCATC CCCTTTAGAT TCGTGCCTCA CTCCACCAGT GAGGAAGATG CCTCTGTCTT TCCCACGACT GCCAGGAGAT AGGGAAGCCC 4862 AGCCAGGACT GACCCTCCTT CCTCCAGCCT GCCCTGACCC ACCTGGCAAA GCAGGGCACA TGGGGAGGAA GAGACTGGAA CCTTTCTTTG ACAGCCAGGC CTAGACAGAC AGGCCTGGGG 4902 ACACTGGCCC ATGAGGGGAG GAAGGCAGGC GCACGAGGTC CAGGGAGGCC CTTTTCTGAT CATGCCCCTT CTCTCCCACC CCATCTCCCC ACCACCACCT CTGTGGCCTC CATGGTACCC 5102 CCACAGGGCT GGCCTCCCCT AGAGGGTGGG CCTCAACCAC CTCGTCCCGC CACGCACCGG TTAGTGAGAC AGGGCTGCCA CGCAACCGCC AAGCCCCCCT CAAGGTGGGA CAGTACCCCG 5222 GACCCATCCA CTCACTCCTG AGAGGCTCCG GCCCAGAATG GGAACCTCAG AGAAGAGCTC TAAGGAGAAG AAACCCChTA GCGTCAGAGA GGATATGTCT GGCTTCCAAG AGAAAGGAGG 5342 CTCCGTTTTG CAAAGTGGAG GAGGGACGAG GGACAGGGGT TTCACCAGCC AGCAACCTGG GCCTTGTACT GTCTGTGTTT TTAAAACCAC TAAAGTGCAA GAATTACATT GCACTGTTTC 5462 TCCRCTTTTT ATTTTCTCTT AGGCTTTTGT TTCTATTTCA AACATACTTT CTTGGTTTTC TAATGGAGTA TATAGTTTAG TCATTTCACA GACTCTGGCC TCCTCTCCTG AAATCCTTTT 5502 GGATGGGGAA AGGGAAGGTG GGGAGGGTCC GAGGGGAAGG GGACCCCAGC TTCCCTGTGC CCGCTCACCC CACTCCACCA GTCCCCGGTC GCCAGCCGGA GTCTCCTCTC TACCGCCACT 5702 GTCACACCGT AGCCCACATG GATAGCACAG TTGTCAGACA AGATTCCTTC AGATTCCGAG TTGCTACCGG TTGTTTTCGT TGTTGTTGTT GTTGTTTTTC TTTTTCTTTT TTTTTTTGAA 5022 GACAGCAATA ACCACAGTAC ATATTACTGT AGTTCTCTAT AGTTTTACAT ACATTCATAC CATAACTCTG TTCTCTCCTC TTTTTTGTTT TCAACTTTAA AAACAAAAAT AAACGATGAT y 5942 AATCTTTACT GGTGAAAAGG ATGG.YIU YTCAACAA ATGCAACCAG T T T G T G A G Y

FIG.1.3'-Untranslated sequence of hIGFEiP-6 cDNA. Sequence from hIGFBP-5cDNAclone BP-5.12 begins withthe nucleotide that follows the last thymidine residue reported previously (4). Nucleotides are numbered on the left; numbers refer tothe distance, in bp, 3' to the mRNA cap site. An AAATAAA polyadenylation signal, poly(A) motif, and two ATITA motifs are underlined and in bold print. Within the sequence, oligonucleotides BP5-19 and BP5-21 span from bp 3117 to 3137 and from bp 4445 to 4462, respectively, while BP5-16, BP5-18, and BP5-20 are complementary to sequence from bp 5979 to5999, bp 3157 to 3175,and bp 4485 to 4504, respectively.

hIGFBP-5 Gene and Promoter

hgBP5-17 I - , ' -) phBP5-BS6.6 phBPS-E8ll.l

10893

H

chBP5-9.2

I

I

phBP5-El. 6 phBP5-E2.5 phBP5-E5.6

hIGFBP-5 Gene

I 1

I

El

0.4 1.1

EcoRI cleavage

-

1 1

I,

E2 E3

95

4.6

195

25

E4

5.6

e.1

FIG.2. Structure of the hIGFBP-5 chromosomalgene. Inserts of genomic A-clone hgBP5-17 and cosmid clone chBP5-9.2 appear as bold lines. Fragments subcloned from hgBP5-17 to create phBP5-BS6.6 and phBP5-HBI1.l, and fragments subcloned from chBP5-9.2 to create phBP5-E4.6, phBP5-E2.5 and phBP5-E5.6 appear as fine lines.A schematic mapof the hIGFBP-5 gene presents exons 1 4 (El-E4) as filled boxes and introns 1-3 a s lines between exons.An EcoRI restriction map of the chBP5-9.2 cosmid insert is at the bottom; numbers represent fragment lengths in kb. -20 15 49

83

93

FIG.3. Sequence of the hIGFBP-5 gene. DNA sequence of coding and 3'-untranslated regions appears as upper case letters while sequence of introns appears a s lower case letters. The estimated size of each intron is shown, as is the size of a n unsequenced gap in the 3'-untranslated reqion which spans 3095 bp in hIGFBP5 cDNA BP-5.12 (4). Amino acids, given in three-letter code above the DNA sequence, are numberedon the left. The sequence of oligonucleotide 5-9 is underlined. The3'poly(A) motif appearing in bold print is the last sequenceshared by hIGFBP-5 cDNA and genomic clones.

127

161

170 204

210

244

acrylamide sequencing gel alongside the sequencing reaction products from the 5:PE primer and phBP5-HBll.l. Primer Extensionof mRNA Expressed by p46lCAT-Total RNA was isolated from MDA-MB-468 cells harvested 22 h after transfection with p461CAT. Sixty pgof this RNA was usedin individual primer extension experiments as described above, except that 1 pmol of oligonucleotide CATPE (5'-TATCAACGGTGGTATATCCAGTGAT-3') served as primer, and samples were incubated a t 70 "C for 5 min followed by cooling over 15 min to 55 "C. Samples were analyzed on a 6.5% polyacrylamide sequencing gel alongside sequencing reaction products from the CAT PE primer andp461CAT. Northern Analysis of hZGFBP-5 mRNA-Total RNA, isolated from MDA-MB-468 cells as described above, was used for Northern analysis as described previously (13,30). The 455-bp BglII fragment of the BP5.1 cDNA clone was labeled with 32P,and lo7 countdmin were used as probe.

RESULTS

Complete Sequence of hZGFBP-5 cDNA-Much of the 3'-untranslated region of hIGFBP-5 cDNA clone BP-5.12 was not originally sequenced (4). This sequence is shown now in Fig. 1;the first nucleotide presented here immediately follows the last thymidineresidue of the publishedsequence (4). Two A m A motifs, which may promote mRNA degradation when located in the 3"untranslated region of an mRNA transcript (35), are identified and complement two additional ATITA motifs presentinthe previously reported 3'-untranslated sequence (4). Isolation and Characterization of the hIGFBP-5 Gene-The A clone hgBP5-17 was isolated by screening 6 x lo5plaques from

10894

A

hIGFBP-5 Gene and Promoter

A B C D E F G H I J K L

--

-4.0 3.0

- 1.6

2.0

1.0

I3 ll

c,

FIG.4. Localization of the hIGFBP-6 gene to chromosome 2. A, PCR amplifications from a human-rodent cell hybrid; representative analysis of PCR products generated by amplification of total genomic DNA (A), total hamster DNA ( K ) ,and somatic human-rodent cell hybrids containing human chromosomes 1-9 (BJ). The length in kilobases of the PCR products is given tothe left, and of the size marker (L) to the right. B, fluorescence hybridizationof genomic exon-containing subclones to human metaphase chromosomes. Fluorescent signals for both chromatids of each chromosome 2 are indicated by arrows.

a human leukocyte genomic library using thehIGFBP-5 cDNA clone, BP-5.1, (4) as probe. When analyzedby Southern blotting after cleavage with BamHI, BglI, HindIII, SacII, and/or SalI, the hgBP5-17 clone hybridized only to a BgZII fragment which spans the 5'-most 455 nucleotides of the BP-5.1 cDNA. Two A fragments, a 6.6-kb BamHI-Sal1 fragment and a 1.1-kb HindIII-BglI fragment, were subcloned in pGEM-3 and pSP73 to create phBP5-BS6.6 and phBP5-HB1l.l, respectively (Fig. 2). The phBP5-BS6.6 clone was used as hybridization probe to screen 8 x lo5 colonies of a human cosmid library. One plaquepurified positive clone, chBP5-9.2, was recognized by each of the threeBgZII fragments from the BP5.12 cDNA, EcoRI cleavage of chBP5-9.2 released nine fragments, of which the 4.6-, 2.5, and 5.6-kb fragments were shown to contain exon sequence by hybridizing with one or more of the BP-5.12 BgZII fragments. These threeEcoRI fragments from chBP5-9.2 were subcloned in pGEM3Zf(+) togeneratephBP5E4.6, phBP5E2.5, and phBP5E5.6, respectively (Fig. 2) and were used, along with theA fragment phBP5-BS6.6, to sequence the coding and 3"untranslated regions and exodintron borders of the hIGFBP-5 gene(Fig. 3). The 3'-untranslated region of genomic hIGFBP-5 was notcompletely sequenced, leaving a gap equivalent to3095 bpof the hIGFBP-5 cDNA clone BP-5.12 (Fig. 1). However, the 3"untranslated region was studied by PCR amplifications using three primerpairs: BP5-9/BP5-18, BP5-19/ BP5-20, and BP5-21/BP5-16 (see Figs. 1 and 3 legends). The PCR product from each primer pair was of identical length when either total human DNA, the cosmid subclone phBP5E5.6, or the hIGFBP-5 cDNA BP-5.12 served as template (data not shown), suggesting strongly that no additional introns are

present in the 3'-untranslatedregion of the hIGFBP-5 gene. The order of the nine EcoRI fragments from the chBP5-9.2 cosmid insert wasmapped by the method of Wahl et aZ. (36) to further characterize the introdexon organization of the hIGFBP-5 gene(Fig. 2). Briefly, Southern analysisused chBP5-9.2 DNA that wascompletely cleaved with SfiI and then partially cleaved with EcoRI; SfiI was chosen since this enzyme cleaves the cosmid insert at a single site found in the 4.6-kb (exon 1-containing) EcoRI fragment. Hybridization with oligonucleotide probes 5735 and5:PE, located 5' and 3' to theSfiI cleavage site,respectively (see Fig. 6), determined the relative order of the nineEcoRI fragments of the cosmid insert. The orderof these EcoRI fragments suggested by the above studies was confirmed by additional PCR and hybridization experiments (data not shown). The exact location of each exon within subclones and the cosmid clone was determinedby PCR amplifications with exonspecific oligonucleotides and specific vector primers. The sizes of introns 2 and 3, determined by PCR amplifications, are 0.6 and 1.3 kb, respectively. Based on the EcoRI restriction map and additional PCR amplifications within subclones, intron 1 has a length of -25 kb. The authenticity of the cosmid fragments was confirmed by hybridizing the 32P-labeledcosmid clone to EcoRI-cleaved total human DNA. Of the nine genomic EcoRI fragments detected, seven had lengths corresponding to the seven internal cosmid fragments shown inFig. 2, while the others measured-12 and -16 kb, much longer than the outertwo fragments of the cosmid insert. The32P-labeledphBP5-BS6.6 clone, which contains the 5"flanking region of the hIGFBP-5 gene, hybridized with -16-, 1.0-, and 4.6-kb fragments of EcoRI-cleaved total human DNA, while 32P-labeled phBP5-E5.6, which represents the 3' end of the cosmid clone, hybridized only with a -12-kb fragment ofEcoRI-cleaved total human DNA(data not shown). This analysis indicates that thehaploid human genome contains a single copy of the hIGFBP-5 geqe.and that this gene is organized as presented in Fig. 2. Chromosomal Localization of the hIGFBP-5 Gene-The hIGFBP-5 genewas localized using a panel of somatic cell hybrids, each carryingone human chromosome on a rodent background. PCR reactions, using primersfrom the coding and the adjacent 3'-untranslated regions of exon 4 amplified an expected 1.7-kb fragment from total human DNA and from the cell hybrid carrying human chromosome 2, but not from any otherhybrid, nor from hamster or mouse DNA (Fig. 4A). Chromosomal localization was also investigated by fluorescent in situ hybridization which showed that, for all analyzed metaphases, only chromosome 2 hybridized with exon-containing probes(Fig. 4B ). QFQbanding placed the IGFBP-5 gene on the long arm at q33-q34 (data not shown),the sameregion where theIGFBP-2 gene was previously localized (26). Human DNA was cleaved with different rare cuttingrestriction enzymes, separated by PFGE, transferred to a filter, and then hybridized successively with four DNA probes derived from the 5' and 3' regions of the IGFBP-2 and -5 genes, respectively. As shown in Fig. 5A, the 3' probes for IGFBP-5 and -2 hybridized to BssHII, EagI, NaeI, NarI, SacII, and SfiI fragments of the same length;most common fragments were -90 kb or less in size, indicating a close linkage between the two genes. In contrast to the 3' probes, the 5' probes for IGFBP-5 and -2 hybridized to fragments of the same length only when DNA samples were cleaved by NarI and SfiI; additionally, for DNA samples cleaved with either BssHII or NaeI, both 5' probes hybridized to different PFGE fragments than did their respective 3' probes. The results of the Southern blots, presented inFig. 5B, indicate that thehIGFBP-2 and -5 genes are transcribed convergently and are separated by -20-40 kb of

hIGFBP-5 Gene and Promoter 3' ICFBP-5

A FIG. 5. Physical linkageof the hIGFBP-5 and -2 genes. A, Southern analysis of DNA fragments separated PFGE by and hybridized with probes taken from the 3' regions of the IGFBP-5 and -2 DNA,prepared genes. Each lane contains from lymphoblastoidcell lines, that was cleaved with the restriction endonuclease indicated below. The position of 90-kb the DNA size marker is indicatedon the left. B,table summarizing the size, in kb, of restriction fragments separated PFGE by and then recognized on Southern analysis by IGFBP probes. Enzymes used are shown at the top and the 5'- or 3'-IGFBP-5 or -2 probe usedfor the Southern blot is shown on the left.The shortest DNA size marker was 90 kb; thus, a fragment presented as