help button home button Endocrine Society Endocrinology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, X.
Right arrow Articles by Adamo, M. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wang, X.
Right arrow Articles by Adamo, M. L.
Endocrinology Vol. 139, No. 3 1054-1066
Copyright © 1998 by The Endocrine Society

ARTICLES

A CACCC Box in the Proximal Exon 2 Promoter of the Rat Insulin-Like Growth Factor I Gene Is Required for Basal Promoter Activity1

Xia Wang, Jose L. Talamantez and Martin L. Adamo

Department of Biochemistry, University of Texas Health Science Center, San Antonio, Texas 78284-7760

Address all correspondence and requests for reprints to: Martin L. Adamo, Ph.D., Department of Biochemistry, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284-7760. E-mail: adamo{at}bioc02.uthscsa.edu


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
The insulin-like growth factor I gene is transcribed from two promoter regions, resulting in alternative first exons in insulin-like growth factor I messenger RNAs. A previous study showed that the sequence from -73 to +44 (where +1 is the first nucleotide in the exon 2 transcription initiation cluster) contained an active exon 2 promoter, and that sequences between -73 and -36 were required for promoter activity. In the current study, the roles of two putative cis-acting elements within the -73 to +44 region in basal exon 2 promoter activity were evaluated using mutagenesis and nuclear protein-DNA binding assays. Mutation of the CCCCACCC sequence at position -53 to GAAATCCC resulted in a complete loss of promoter activity in transient transfection assays in GH3, OVCAR-3, C6, and Chinese hamster ovary (CHO) cells. A -73/+24 exon 2 promoter-luciferase construct had partial promoter activity. Mutation of a putative initiator motif surrounding the major exon 2 start site did not alter the activity of this construct. In electrophoretic mobility shift assays, a 32P-labeled oligomer extending from -73 to +44 in the exon 2 promoter was specifically bound by GH3 cell nuclear extracts. A 32P-labeled oligomer which extended from -63 to -37 in the exon 2 promoter was specifically bound by GH3 and OVCAR-3 cell nuclear extracts. These unlabeled oligomers inhibited the binding of a labeled -236/+44 exon 2 promoter fragment to OVCAR-3 nuclear extracts. Mutation of the CCCCACCC sequence prevented the unlabeled -73/+44 oligomer from inhibiting the binding of the -236/+44 fragment. An unlabeled oligomer containing a consensus activating protein-2 (AP-2)-binding site inhibited labeled -236/+44, -73/+44, and -63/-37 exon 2 promoter binding with a much lower affinity than did the respective unlabeled oligomers. Purified AP-2 protein did not bind to the exon 2 promoter fragment, nor did anti-AP-2 antibody alter the binding. Cotransfection of AP-2 expression vector did not significantly increase exon 2 promoter activity. On the other hand, an oligomer containing a consensus Sp1-binding site inhibited labeled -63/-37 exon 2 promoter binding by GH3 cell nuclear extracts with an affinity similar to that of the unlabeled -63/-37 oligomer. A mutation in the Sp1-binding site in this same oligomer resulted in a complete loss of binding affinity. Purified Sp1 bound to the -63/-37 exon 2 promoter oligomer. Addition of polyclonal antibody to Sp1 resulted in a partial supershift of the complex formed between GH3 cell and OVCAR-3 cell nuclear extracts and the labeled -63/-37 oligomer. However, in Drosophila Schneider cells, which are an experimental model for studying the ability of Sp1 to activate transcription, the -73/+44 exon 2 promoter construct was not stimulated by cotransfection with an Sp1 expression plasmid. UV cross-linking studies indicated that proteins of approximate molecular mass 125, 76, 47, and 38 kDa are bound to the proximal (-236/+44) exon 2 promoter region. It is concluded that the CCCCACCC sequence at -53 is required for exon 2 promoter activity. Moreover, specific binding of nuclear proteins to the proximal exon 2 promoter region requires the CCCCACCC sequence. Sequences downstream of the exon 2 initiation site from +24 to +44 are required for full promoter activity. However, the putative initiator surrounding the major transcription start site at +1 does not appear to be important for the strength of the proximal promoter. The CCCCACCC sequence at -53 appears to be a CACCC box, which binds zinc finger transcription factors of the Kruppel family such as Sp1, although protein factors in addition to Sp1 are required to activate exon 2 transcription through the -73/+44 proximal promoter region.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
IN MAMMALS, insulin-like growth factor I (IGF-I) messenger RNAs (mRNAs) contain alternative first exons. Either exon 1 or exon 2 is spliced to a common block of exons that contains the mature peptide-coding sequence. Exons 1 and 2 encode alternative 5'-untranslated regions, and distinct amino-terminal amino acids of the prepro-IGF-I signal peptide (reviewed in Ref.1). The alternative 5'-untranslated regions dramatically affect the translational efficiency of IGF-I mRNAs (2). Exon 1-containing mRNAs are ubiquitous and predominant, and in some tissues are the only form of IGF-I mRNA (3, 4). However, exon 2 mRNAs comprise about 30% of the total IGF-I mRNA in the liver (3, 4, 5). A few other tissues express exon 2 mRNA (3, 4), as do some cell lines (6, 7, 8). OVCAR-3 cells are unique in that exon 2 mRNAs appear to be predominant over exon 1 mRNAs in this cell line (7, 8). In the liver, exon 2 mRNA levels are regulated in a coordinate manner with exon 1 mRNA levels in response to diabetes, altered nutritional states, and acute GH treatment (3, 5). During development and possibly during chronic GH administration, there are tissue-specific differences in the regulation of exon 1 and exon 2 mRNAs (4, 5, 9, 10, 11, 12).

Exon 2 transcription initiation occurs from two clusters located about 1.8 kilobases (kb) downstream from the 3'-end of exon 1 (3, 4, 13, 14, 15). The major initiation cluster is located 62–68 bp upstream of the 3'-end of exon 2, and a minor cluster is located at 52–53 bp upstream of the 3'-end of exon 2. In those rat tissues in which exon 2 mRNAs are expressed, the location of the transcription start sites appears to be invariant (3, 4). In studies using the human IGF-I gene, the entire 1.8 kb of sequence 5' to exon 2 and 58 bp of exon 2 function as a promoter when fused to a luciferase reporter gene and tested in transient transfection assays (16). The promoter activity was 4-fold higher in exon 2-expressing OVCAR-3 cells than in exon 1-expressing SK-N-MC cells and was 10- to 15-fold higher than that in HepG2 cells, which do not express detectable IGF-I mRNA. When 600 bp of the 5'-flanking sequence were removed, there was increase in promoter activity in all three cell lines, suggesting that negative regulatory elements were present in the upstream sequence between -1800 and -1200 (16). However, there was a greater difference in exon 2 promoter activity between the SK-N-MC cells and the OVCAR-3 cells when about 1200 bp of 5'-flanking sequence were present, suggesting that elements important for cell type-specific promoter activity are contained in the region downstream of -1200.

This laboratory has recently characterized the activity of the rat exon 2 promoter (17). We found that a promoter construct containing 1.5 kb of 5'-flanking sequence and 44 bp of exon 2 sequence was essentially inactive in exon 1-producing C6 cells and had slight activity in GH3 cells, which express exon 1 and exon 2 mRNAs. Activity was highest in OVCAR-3 cells and to a lesser extent in Chinese hamster ovary (CHO) cells. When deletions of 5'-flanking sequence were performed, promoter activity was increased, especially in the C6 and GH3 cells, to a lesser extent in the CHO cells, and to a minor extent in OVCAR-3 cells. These results suggest that negative regulatory elements between -1200 and approximately -400 contribute to cell type-specific transcription of exon 2. A construct that contained 73 bp of 5'-flanking sequence was fully active. Thus, the CCAAAT sequence at -80, which we had previously hypothesized to be a CCAAT box (13), was not essential for exon 2 promoter activity. However, when only 36 bp of 5'-flanking sequence were present, promoter activity was completely lost. These results indicated that the sequence between -73 and -36 contained an element(s) essential for exon 2 transcription (17). This region contains two in vitro footprints that are produced by rat liver nuclear extracts (18). A computer-aided search indicated to others (18) and to us that the sequence CCCCACCC at -53 was a potential activating protein-2 (AP-2)-binding site. In this study, we have demonstrated that the CCCCACCC sequence is indeed essential for exon 2 promoter activity. However, it does not appear to be a high affinity AP-2-binding site. This sequence may be a CACCC box, as it is specifically bound by the zinc finger transcription factor Sp1 with high affinity.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
IGF-I exon 2 promoter-luciferase constructs
All constructs (diagrammed in Fig. 1Go) were prepared by ligation of PCR-generated DNA fragments into the pGL2-Basic expression vector (Promega, Madison, WI). As described previously (17), the -73/+44 and the -36/+44 exon 2 promoter fragments were generated by PCR using the -73/+44 CAC-W and -36/+44 Inr-W sense primers (Table 1Go), respectively, and the -73/+44 CAC-W antisense primer. The PCR products were ligated into the SmaI and BglII sites of pGL2-Basic. For construction of the -73/+24, the -45/+24, the -73/+44 CAC-M, the -73/+24 CAC-M, and the -73/+24 Inr-M exon 2 promoter fragments, PCR primers (Table 1Go) included a GCGC clamp, and either a KpnI restriction site in the sense primers or a HindIII site in the antisense primers. The -73/+24 fragment was amplified using the -73/+24 Inr-W sense primer and the -73/+24 Inr-W antisense primer. The -45/+24 fragment was amplified with the -45/+24 CAC-D sense primer and the -73/+24 Inr-W antisense primer. The -73/+44 CAC-M fragment, in which the CCCCACCC sequence at -53 was mutated to GAAATCCC, was amplified using the -73/+44 CAC-M sense primer and the -73/+44 CAC-W antisense primer. This latter primer differed from that shown in Table 1Go in that it contained a HindIII restriction site rather than a BglII site. The -73/+24 CAC-M fragment was amplified using the -73/+44 CAC-M sense primer and the -73/+24 Inr-W antisense primer. Finally, the -73/+24 Inr-M fragment, in which the putative initiator sequence GGCCTCATAAT was mutated to GGGCAGATAAT, was amplified using the -73/+24 Inr-W sense primer and the -73/+24 Inr-M antisense primer. Primers were synthesized at the Center for Advanced DNA Technologies at University of Texas Health Science Center (San Antonio, TX). A SmaI-SmaI fragment of the exon 2 promoter region extending from -1500 to +44 (where +1 is the first transcription initiation site in the major initiation cluster) was used as a template (17). PCR reactions contained 50 ng template DNA; 0.5 µM each of sense and antisense primers; 200 µM deoxy (d)-ATP, dCTP, dGTP, and dTTP; 2.5 U native Taq DNA polymerase; and 10 µl 10 x buffer [500 mM KCl, 100 mM Tris-HCl (pH 8.3), 15 mM MgCl2, and 0.01% (wt/vol) gelatin] in a final volume of 100 µl. PCR reagents other than template and primers were obtained from Perkin-Elmer/Cetus (Norwalk, CT). PCR reactions were conducted using a Crocodile II thermocycler (Appligene, Pleasonton, CA) as follows. Reactions were hot started at 94 C for 5 min (initial denaturation), and then 15 of the following cycles were performed: denaturation at 94 C for 1 min, annealing at either 45 or 55 C for 1 min, and elongation at 72 C for 1 min. After the last cycle, a 7-min elongation at 72 C was performed. The PCR reactions were subjected to ethanol precipitation, and the precipitates were digested with HindIII and KpnI at 37 C overnight. Restriction enzymes were heat inactivated at 75 C for 15 min, and the digested PCR products were extracted with phenol-chloroform, precipitated with ethanol, and quantified by ethidium bromide staining of aliquots run on 4% agarose gels.The digested PCR products were ligated into the KpnI and HindIII sites of pGL2-Basic. Plasmids were amplified in either HB101 or DH5-{alpha} Escherichia coli cells and were checked by restriction mapping of minipreps. The plasmid DNA was sequenced to confirm the identity of the inserts. For transfection, plasmid DNAs were purified using the Qiagen column system (Santa Clarita, CA) and quantified by measuring the absorbance at 260 nm. Quantitation was verified by ethidium bromide staining of linearized DNAs on 0.8% agarose gels.



View larger version (26K):
[in this window]
[in a new window]
  		Figure 1. Structure and sequence of the proximal exon 2 promoter, putative cis-acting elements and determination of proximal promoter limits by transfection assays. The locations within the proximal exon 2 promoter of the first transcription start site in the major initiation cluster (arrow at +1) (13), a putative TdT gene initiator-like motif (Inr) (32) surrounding the major start sites, and a putative CACCC (28–30) box located at -53 are all shown at the top of A. The sequence of the proximal exon 2 promoter region is shown in B, with the sequences of the cis-acting elements underlined. The nucleotides in italics represent the exon 2 transcription start sites mapped in rat liver (13). To the left in A are shown the exon 2 promoter fragments with the indicated length of sequence upstream (negative numbers) and downstream (positive numbers) of the first major start site, which were fused upstream of the luciferase reporter gene in the pGL2-Basic expression vector (large empty arrowhead). The x’s show the positions and numbers of mutated nucleotides in the mutated CACCC box and putative initiator (CAC-M and Inr-M, respectively). Equimolar amounts of DNAs (~8 µg from different plasmid preparations) were introduced into OVCAR-3 cells, C6 cells, and CHO cells by the calcium phosphate method and into GH3 cells by lipofectin reagent. Luciferase activities and protein concentrations of cell lysates were measured 24 h after transfection. Relative luciferase activity (light units normalized to A595, as determined by the Bradford method) (19) is presented in the right part of A as the fold stimulation ± SEM over pGL-2-Basic vector for 4–10 independent transfections performed in duplicate. For the -73/+44 and -36/+44 promoter constructs, 4 of the 10 replicate experiments reported in the figure were taken from Ref. 17.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Oligonucleotides and PCR primers used for EMSA and promoter constructs

 
Transient transfection
Rat C6 glial tumor cells, rat GH3 pituitary tumor cells, human OVCAR-3 ovarian adenocarcinoma cells, and CHO cells were obtained from American Type Culture Collection (Rockville, MD). C6 and GH3 cells were grown in DMEM (4.5 g/liter glucose) containing 10% newborn calf serum or FBS. OVCAR-3 cells were grown in RPMI 1640 medium containing 10% FBS. CHO cells were grown in Ham’s F-12 medium with 10% FBS. All sera were heat inactivated at 56 C for 30 min. All media were supplemented to final concentrations of 4 mM glutamine, 50 IU/ml penicillin, and 50 µg/ml streptomycin. The cells were maintained in a humidified 5% CO2-95% air incubator at 37 C. All cells were used between passages 5–15 (the frozen stock obtained from American Type Culture Collection was arbitrarily referred to as passage 1). Before transfection, 6 x 105 C6, CHO, or OVCAR-3 cells or 2 x 106 GH3 cells were plated in 60-mm plates. Forty-eight hours later, equimolar amounts of promoter-luciferase constructs (~8 µg) or promoterless pGL2-Basic plasmid were introduced into C6, OVCAR-3, and CHO cells by the calcium phosphate precipitation procedure, using reagents supplied by 5 Prime-3 Prime (Boulder, CO). GH3 cells were transfected with lipofectin reagent (Life Technologies, Gaithersburg, MD). Twenty-four hours after the transfection, the cells were lysed with Reporter lysis buffer (Promega). Lysates were clarified by centrifugation (12,000 x g for 4–5 min). The luciferase activity of aliquots of the lysate was measured with the luciferase assay system reagents (Promega) in a model ILA911 semiautomatic luminometer (Tropix, Bedford, MA), over a 10-sec interval. The protein concentrations of the lysates were measured using the Bradford method (19). Light units of luciferase activity were normalized to the A595. Specific luciferase enzyme activity was then divided by that seen in the promoterless luciferase plasmid pGL2-Basic to quantify promoter activity.

For AP-2 cotransfection studies, 3 x 106 HepG2 cells (obtained from American Type Culture Collection) were plated onto 60-mm plates and grown for 48 h in DMEM (1 g/liter glucose), 10% FBS, glutamine, and antibiotics as described above. The cells were then transfected with 10 µg pGL2-Basic, 10 µg IGF-I exon 2 -73/+44 promoter in pGL2-Basic or 10 µg IGF-binding protein-5 (IGFBP-5) promoter luciferase plasmid pBP5P/luc (-503/+775) (20) and 1, 2, or 4 µg of either SPRSV-AP2, which is an AP-2 expression vector under control of the Rous sarcoma virus (RSV) enhancer/promoter, or SPRSV, which is the same vector without the AP-2-coding sequence. Plasmid pBP5P/luc was provided by Drs. Cunming Duan and David Clemmons, University of North Carolina (Chapel Hill, NC), and served as a positive control for AP-2 cotransfection experiments. Plasmids SPRSV-AP2 and SPRSV were provided by Dr. Trevor Williams, Yale University (New Haven, CT). The cotransfections that received 1 and 2 µg AP-2 vector also received 3 and 2 µg, respectively, of an inactive pGL2-Basic vector (from which the luciferase-coding region had been partially removed) so that equal amounts of DNA were used. Transfections were performed using the calcium phosphate method, and luciferase and protein were measured, all as described above. The ratio of relative luciferase activity in the presence of SPRSV-AP2 to that in the presence of SPRSV was calculated for each construct. The experiment was repeated three times in duplicate.

For Sp1 cotransfection studies, Schneider’s Drosophila cell line 2 (SL-2) was obtained from American Type Culture Collection and grown in Schneider’s Drosophila medium (Life Technologies) containing 10% FBS, antibiotics, and 4 mM glutamine as described above. Approximately 8 x 105 cells were plated onto 60-mm plates and grown for 48 h. Five micrograms of pGL2-Basic, exon 2 -73/+44 promoter fragment in pGL2-Basic, p0Luc (provided by Dr. Alan Brasier), or IGF-I receptor promoter (-476/+640) in p0Luc (provided by Drs. Haim Werner, Charles T. Roberts, Jr., and Derek LeRoith) plasmids were cotransfected with 15 µg of either pAdhSp1, which is the human Sp1-coding sequence under control of the Drosophila alcohol dehydrogenase (Adh) promoter, or pAdh, which is the same vector without the Sp1-coding sequence, by the calcium phosphate method (21). The IGF-I receptor promoter was used as a positive control, as it contains multiple Sp1-binding sites, which mediate Sp1 activation of the IGF-I receptor promoter in Drosophila cells. Plasmids were provided by Dr. Charles T. Roberts of Oregon Health Science University, with permission also obtained from Dr. Robert Tjian of University of California-Berkeley, who originally supplied the Sp1 plasmids to Drs. Werner, Roberts, and LeRoith at the NIH. Forty-eight hours after transfection, luciferase enzyme activities and total cellular protein concentration were assayed in cellular lysates. Data were calculated as the fold increase in luciferase in the presence of pAdhSp1 over that in the presence of pAdh and are reported as the mean ± SEM for four separate experiments, performed in duplicate.

For both the AP-2 and Sp1 cotransfection studies, statistical differences in fold stimulation between the promoterless luciferase plasmids and the corresponding promoter luciferase plasmids were determined using one-way ANOVA in the SIMSTAT 3 package (Normand Peladeau, Provalis Research, Montreal, Canada).

Electrophoresis mobility shift assay (EMSA)
Nuclear extracts from GH3 cells and OVCAR-3 cells were prepared by the high salt method (22), as follows. Two grams of packed cells were suspended in 10 ml 10 mM HEPES-KOH (pH 7.9), 10 mM KCl, 1.5 mM MgCl2, 0.1 mM EGTA, 0.5 mM dithiothreitol (DTT), 2 µg/ml of leupeptin and pepstatin A, and 0.5 mM phenylmethylsulfonyl fluoride (PMSF). The cells were homogenized with a glass Ten-Broeck homogenizer (Pyrex), and the homogenates were centrifuged at 1200 x g for 10 min. The nuclear pellet was washed twice with 10 ml homogenization buffer, and then resuspended in 5 ml 10 mM HEPES-KOH (pH 7.9), 0.4 M NaCl, 1.5 mM MgCl2, 0.1 mM EGTA, 0.5 mM DTT, 5% glycerol, and 0.5 mM PMSF. The nuclei were homogenized using the Ten-Broeck device, and the homogenate was stirred at 4 C for 30 min. The nuclear homogenate was centrifuged at 100,000 x g for 60 min. The supernatant was dialyzed in 1000 dalton molecular mass cut-off dialysis tubing against 20 mM HEPES-KOH (pH 7.9), 75 mM NaCl, 0.1 mM EDTA, 0.5 mM DTT, 0.5 mM ZnCl2, 0.5 mM PMSF, and 20% glycerol at 4 C for 4 h. The nuclear extract was removed from the tubing and centrifuged at 25,000 x g for 15 min. The nuclear extract in the supernatant was stored in frozen aliquots at -80 C. Protein concentrations were determined by the method of Bradford (19).

Table 1Go shows the sequences of the double stranded oligomers that were used in EMSA. The exon 2 -63/-37 oligomer was prepared by annealing the single-stranded oligonucleotides that were synthesized at the Center for Advanced DNA Technologies at University of Texas Health Science Center. Annealing was performed in 0.1 M NaCl by heating the oligonucleotides to 90 C for 20 min, and then allowing the annealing reaction to cool at room temperature. The annealed oligomer was visualized on an ethidium bromide-stained 4% agarose gel and then stored frozen at -20 C. The -236/+44 fragment was isolated from a -485/+44 exon 2 promoter region by digestion with Sau3A and SmaI. The -73/+44 wild-type (CAC-W), the -73/+44 CCCCACCC mutation (CAC-M), the -45/+44 CCCCACCC deletion (CAC-D), and the -73/+44 initiator mutation (Inr-M) exon 2 promoter fragments were prepared by PCR, as described above, using the primers listed in Table 1Go. All of these DNA fragments were quantified by ethidium bromide staining of aliquots on agarose gels and comparison to DNA standard ladders. The oligomers containing the consensus AP-2- and Oct-1-binding sites were purchased from Promega. The oligomers containing the wild-type and mutant Sp1 sites were purchased from Santa Cruz Biotechnologies (Santa Cruz, CA). Recombinant human AP-2 and Sp1 protein were purchased from Promega. Rabbit polyclonal antibodies to human AP-2 and Sp1 were obtained from Santa Cruz Biotechnologies.

The -236/+44 exon 2 promoter fragment was 32P labeled using a Klenow fill-in reaction (23). The labeled probe was separated from unincorporated nucleotides on an Elutip-D column (Schleicher and Schuell, Keene, NH), followed by ethanol precipitation. Specific radioactivity was estimated by dividing the amount of radioactivity recovered by the amount of DNA recovered. This latter parameter was estimated as 65% of the input DNA, based on control experiments in which large amounts of unlabeled input DNA were visualized on 4% ethidium bromide-stained agarose gels after Elutip-D recovery. The -73/+44 exon 2 promoter fragment was uniformly labeled with 32P using a PCR procedure (24). Reactions (100 µl) contained 50 ng template DNA; 0.5 µM sense and antisense primers; 50 µM dATP; 200 µM of dCTP, dTTP, and dGTP; 5 µl [{alpha}-32P]dATP (3000 Ci/mmol); 10 µl 10 x reaction buffer (defined above), and 2.5 U native Taq DNA polymerase. Fifteen to 20 cycles of hot start PCR were performed as described above. The labeled PCR products were purified using Elutip-D columns. The amount of radiolabeled DNA probe generated by the PCR procedure was determined by comparison with {phi}x174 DNA markers (HinF1 digest) in an ethidium bromide-stained 4% agarose gel. The -63/-37 exon 2 promoter double stranded oligomer was 32P labeled using either the Klenow fill-in reaction (23), or 5' end labeled using [{gamma}-32P]ATP and T4 polynucleotide kinase (25). Specific radioactivity was calculated on the basis of estimated recovered radioactivity and mass of oligomer, determined as described above. The double stranded AP-2 oligomer was end labeled with T4 polynucleotide kinase (25). EMSAs were performed using the Bandshift kit (Pharmacia, Piscataway, NJ). 32P-labeled DNA probes in the amounts indicated in the figure legends were incubated with 1 µg nuclear proteins, 400 ng poly-(dI-dC)·poly-(dI-dC), and unlabeled competitor DNA at the molar excess concentrations indicated in the figures in a buffer containing 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 0.5 mM DTT, 0.05% Nonidet P-40, and 10% glycerol in a final volume of 20 µl. In some cases, purified Sp1, AP-2, or antibodies to Sp1 or AP-2 were added as described. After a 20-min preincubation, the probe was added, and the reactions were incubated for an additional 40 min at room temperature. The DNA-protein complexes were separated by electrophoresis (10 V/cm) on 5% native polyacrylamide gels. Dried gels were autoradiographed.

UV cross-linking of nuclear proteins to DNA probes
DNA probes for use in UV cross-linking experiments were labeled by the PCR procedure described above (24). For UV cross-linking, 1–10 µg nuclear protein were preincubated for 20 min at room temperature with 5 µg poly-(dI-dC) · poly-(dI-dC) or sonicated calf thymus DNA and in some cases unlabeled competitor oligomers in the same binding buffer as that described above. Then, 10 ng 32P-labeled probe (260,000 cpm) were added, and reactions were incubated at room temperature for an additional 40 min in a 40-µl final volume. The reaction tubes were then irradiated from a distance of 5 cm with a UV transilluminator (Spectronics Co., Westbury, NY), emitting light at a wavelength of 312 nm, with an intensity of 7000 µW/cm2, for 30 min (26). Eight-tenths of a microliter of 0.5 M CaCl2, 4 U deoxyribonuclease I (DNase I), and 1 U micrococcal nuclease (Worthington Diagnostics, Freehold, NJ) were added, either without or with 2 U proteinase K. Samples were incubated at 37 C for 30 min to digest unbound DNA. Eight microliters of 5 x SDS sample buffer [0.1 M Tris-HCl (pH 6.8), 20% glycerol, 3.5% SDS, 0.1 M DTT, and 1 mg/ml bromophenol blue] were then added to each tube. The tubes were boiled for 5 min. Samples were electrophoresed on 10% SDS-PAGE minigels along with prestained high mol wt protein markers from Life Technologies. The dried gel was autoradiographed at -80 C for 3–4 days.


   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Mapping of the 5' and 3' limits of the exon 2 promoter
Our previous studies indicated that significant exon 2 promoter activity was observed in a -73/+44 construct, but the activity was completely lost in a -36/+44 construct (17) (Fig. 1AGo). This result suggested that the 5' limit of the promoter was between -73 and -36. A 3'-deletion that produced the -73/+24 construct resulted in luciferase activity that was 2- to 5-fold higher than that of the promoterless control plasmid (Fig. 1AGo). However, luciferase activity of the -73/+24 promoter construct was 2.5-, 2.1-, 1.5-, and 1.9-fold lower than that of the -73/+44 construct in OVCAR-3, GH3, C6, and CHO cells, respectively (Fig. 1AGo). When the partially active -73/+24 construct was deleted at the 5'-end to produce the -45/+24, promoter activity was completely abolished (Fig. 1AGo). These results indicated that the 5' limit of the exon 2 promoter is between -73 and -45. The 3' limit is upstream of +24, although sequences between +24 and +44 are required for maximal exon 2 promoter activity. An evaluation of multiple transfection experiments indicated that the highest luciferase activity resulting from the IGF-I promoter was never greater than 10% of the activity of the pGL2-Control vector, which uses the simian virus 40 promoter/enhancer to drive luciferase expression.

The sequence 5'-CCCCACCC-3', beginning at -53, (underlinedin Fig. 1BGo) was identified as a potential AP-2 transcription factor-binding site after searches by us [using the program FINDPATTERNS in the GCG program package (WI Package version 9.1, Genetic Computer Group) to search the Transcription Factor Database, release 7.4, 1995] and others (18). However, this sequence could also be a CACCC box, which binds Kruppel-like transcription factors (27, 28, 29, 30, 31). To examine the contribution of the CCCCACCC sequence to basal promoter activity, the sequence was mutated to 5'-GAAATCCC-3' (-73/+44 CAC-M). Promoter activity was abolished in all four cell lines tested (Fig. 1AGo). In the partially active -73/+24 construct, mutation of the CACCC sequence (-73/+24 CAC-M) also completely abolished promoter activity.

A putative initiator (Inr)-like motif surrounds the major exon 2 transcription initiation sites (Fig. 1Go, A and B). This putative initiator has the sequence 5'-GGCCTCATAAT-3'. The nucleotides in boldface are the transcription start sites (13), and the core initiator motif (32, 33) is underlined. To determine whether this putative initiator contributed to basal promoter activity, it was mutated to 5'-GGGCAGATAAT-3' in the -73/+24 construct. This mutation did not alter the promoter activity of the -73/+24 construct in any of the four cell lines studied (Fig. 1AGo).

Characterization of nuclear protein binding to the CCCCACCC sequence
To further demonstrate the importance of the CCCCACCC sequence at -53, its contribution to the binding of nuclear proteins to the exon 2 promoter was evaluated in EMSA. When the 32P-labeled -236/+44 fragment of the exon 2 promoter was incubated with OVCAR-3 cell nuclear extracts, a complex was formed with a slower mobility than that of the free probe (Fig. 2AGo, lanes 1 and 2). The intensity of the DNA-protein complex was not altered by inclusion of a 100-fold molar excess of an oligomer containing an Oct-1 transcription factor-binding site (see Table 1Go for sequence; lane 3). The active -73/+44 exon 2 promoter fragment inhibited the binding of the labeled -236/+44 fragment when added at a 100-fold molar excess (lane 4). However, when the CCCCACCC sequence was mutated to GAAATCCC in the -73/+44 fragment (CACCC-M) or when it was deleted in the -45/+44 fragment (CACCC-D), these unlabeled oligomers (at a 100-fold molar excess) were not able to inhibit the binding of the -236/+44 promoter fragment (lanes 5 and 6). However, the -73/+44 exon 2 promoter fragment with the mutation in the putative initiator sequence was able to inhibit binding (lane 7). The wild-type CCCCACCC sequence is retained in this fragment. As previously demonstrated (17), the intensity of the band(s) was reduced when unlabeled -236/+44 DNA was included as a competitor at 50-, 100-, or 200-fold molar excess (Fig. 2AGo, lanes 8–10). A double stranded exon 2 promoter oligomer extending from -63 to -37, which contained the CCCCACCC sequence (see Table 1Go for sequence), inhibited binding of the labeled -236/+44 exon 2 promoter probe in a concentration-dependent manner (lanes 11–13). A 26-bp oligomer containing a canonical AP-2-binding site (see Table 1Go for sequence) was also able to inhibit binding of the -236/+44 fragment, although with a lower affinity than that of the -63/-37 oligomer (Fig. 2AGo, compare lanes 11–13 to lanes 14–16).



View larger version (80K):
[in this window]
[in a new window]
  		Figure 2. Specific binding of nuclear proteins to exon 2 promoter regions determined by EMSA. A, IGF-I exon 2 promoter fragment -236/+44 was labeled with 32P using the Klenow fill-in reaction. Approximately 1 ng (~10,000 cpm) was incubated with nuclear extract (1 µg protein) from OVCAR-3 cells. B, IGF-I exon 2 promoter fragment -73/+44 was labeled with 32P by the PCR procedure. Approximately 1 ng (~5000 cpm) was incubated with nuclear extract (1 µg protein) from GH3 cells. For both panels, some reactions contained unlabeled competitor DNA at the indicated molar excess. Where not indicated, unlabeled DNAs were used at a 100-fold molar excess. Binding reactions were conducted as described in Materials and Methods. The DNA-protein complexes (whose migration is indicated by the bar to the left in A and by the arrowhead in B) were separated from free DNA (bands at the bottom of the gel) on 5% native polyacrylamide gels. Dried gels were autoradiographed overnight at -80 C.

 
Similar results were observed when GH3 cell nuclear extracts were incubated with a 32P-labeled -73/+44 exon 2 promoter fragment (Fig. 2BGo). In this case, a complex was formed with slower mobility than that of the free probe (compare lanes 1 and 2), although it was not clear that multiple bands appeared within the shifted complex. The Oct-1 site DNA did not inhibit binding at a 100-fold molar excess (lane 3), whereas the unlabeled -73/+44 double stranded oligomer (probe oligo) inhibited binding at 50-, 100-, 250-, and 500-fold molar excesses (lanes 4–7). These results indicated that the binding was specific. The -63/-37 (CACCC box) oligomer inhibited binding at a 50- to 500-fold molar excess, whereas over the same range of concentrations, the 26-mer containing the consensus AP-2-binding site was much less effective (compare lanes 8–11 to lanes 12–15).

AP-2 does not appear to bind to or to activate the exon 2 promoter
To further characterize the protein(s) that binds to the CCCCACCC sequence, the -63/-37 double stranded exon 2 promoter oligomer was end labeled with 32P and used as a probe in the EMSA. Incubation of this probe with GH3 and OVCAR-3 cell nuclear extracts resulted in formation of a complex with slower mobility than that of the free probe (Fig. 3Go, compare lane 1 to lanes 2 and 9). When using GH3 cell nuclear extracts, a 100-fold molar excess of unlabeled Oct-1 DNA did not inhibit binding (lane 3), whereas a 100-fold molar excess of the -63/-37 oligomer almost completely inhibited binding (lane 4). To determine whether the binding was due to AP-2 protein, a polyclonal antibody against AP-2 (100 ng) was added. As shown in lane 5, there was neither inhibition of binding of the -63/-37 oligomer to GH3 cell nuclear extracts, nor any apparent supershift in the DNA-protein complex. When 10 ng purified recombinant human AP-2 were incubated with the labeled -63/-37 probe under the same EMSA conditions, there was no apparent complex formed, regardless of whether unlabeled -63/-37 or unlabeled Oct-1 DNA was added (lanes 6–8). When the 26-mer containing the consensus AP-2-binding site was also end labeled and used as a probe in the EMSA, no binding to this probe by GH3 or OVCAR-3 cell nuclear extracts could be demonstrated (Fig. 3Go, compare lane 10, free probe, to lanes 11–13 and 18). When 10 ng purified recombinant AP-2 protein were used, a smeared gel shift was produced, whose intensity was not reduced by excess unlabeled Oct-1 DNA (lanes 14 and 15). However, excess unlabeled AP-2 oligomer reduced the intensity (lane 16). A clearer indication that AP-2 was able to bind the AP-2 consensus sequence oligomer was seen when incubation of the labeled AP-2 oligomer with purified AP-2 protein and a polyclonal antibody to AP-2 resulted in a supershifted band, indicative of the formation of a ternary complex (lane 17).



View larger version (95K):
[in this window]
[in a new window]
  		Figure 3. Purified AP-2 protein binds to a consensus AP-2 oligomer, but not to the exon 2 CACCC box oligomer. An exon 2 promoter oligomer extending from -63 to -37 (sequence shown in Table 1Go; lanes 1–9) or a 26-bp oligomer containing a consensus AP-2-binding site (sequence shown in Table 1Go; lanes 10–18) were end labeled with 32P using T4 polynucleotide kinase. Approximately 0.03 ng of the -67/-37 probe or 0.04 ng of the AP-2 probe (~10,000 cpm) were incubated with GH3 cell nuclear extract (1 µg; lanes 2–5 and 11–13), OVCAR-3 cell nuclear extract (1 µg; lanes 9 and 18), or purified AP-2 protein (10 ng; lanes 6–8 and 14–17). Unlabeled competitor DNAs and/or a polyclonal anti-AP-2 antibody (100 ng) were added in the indicated lanes. DNA-protein complexes were separated from free DNA on a 5% polyacrylamide gel, which was dried and autoradiographed for 3 days at -80 C.

 
To further determine whether the CCCCACCC sequence at -53 could be a functional AP-2 site, cotransfection experiments were performed using an AP-2 expression plasmid in HepG2 cells (Fig. 4Go). These cells do not produce endogenous AP-2 protein and are used to study the function of AP-2 (20). The pBP5P/luc construct contains a functional AP-2 site (20) and was used as a positive control. Ten micrograms of pGL2-Basic, the exon 2 -73/+44 promoter-luciferase construct, or pBP5P/luc plasmid DNA were cotransfected with 1, 2, or 4 µg of either AP-2 expression vector (SPRSV-AP2) or the same plasmid without the AP-2-coding sequence (SPRSV). The fold stimulation was computed as the relative luciferase activity observed with SPRSV-AP2 divided by that observed with SPRSV. One microgram of AP-2 plasmid stimulated all three test plasmids by 5-fold. However, when 2 or 4 µg SPRSV-AP2 were used, pBP5P/luc was significantly stimulated about 3.4-fold more than was pGL2-Basic (P < 0.05). Using 2 µg SPRSV-AP2, the exon 2 promoter was not stimulated more than pGL2-Basic. At 4 µg, the 2-fold increase in exon 2 -73/+44 promoter activity compared with that in pGL2-Basic was not significant (P = 0.25).



View larger version (32K):
[in this window]
[in a new window]
  		Figure 4. Effect of AP-2 transcription factor on exon 2 promoter activity. HepG2 cells were cotransfected with 10 µg pGL2-Basic, 10 µg IGF-I exon 2 -73/+44 promoter-luciferase construct in pGL2-Basic, or 10 µg pBP5P/luc (an IGFBP-5 promoter construct from -503 to +775 in pGL2-Basic) and with 1, 2, or 4 µg of an AP-2 expression vector (SPRSV-AP2) or the same vector without the AP-2-coding sequence (SPRSV). The total amount of transfected DNA was made to 14 µg using a luciferase-null pGL2-Basic vector. Luciferase activity and protein levels (i.e. A595) were assayed 24 h after transfection. The ratio of relative luciferase activity in the presence of SPRSV-AP2 to that in the presence of the SPRSV is shown for all three test constructs and at all doses of AP-2 vector. The data are presented as the mean ± SEM for three separate experiments performed in duplicate. One-way ANOVA indicated that AP-2 significantly increased IGFBP-5 promoter activity compared with pGL2-Basic vector when 2 and 4 µg AP-2 expression vector were used (P < 0.05).

 
The zinc finger transcription factor Sp1 can bind to the exon 2 promoter CACCC box, but does not stimulate proximal exon 2 promoter activity
Recent reports from the literature also identify the sequence CCCCACCC as a CACCC box, which is bound by members of a larger family of zinc finger transcription factors, such as erythroid (28), lung (29), and Basic (30) Kruppel-like factors (EKLF, BKLF, and LKLF); Sp1 (30, 31); and Wilms’ tumor gene product (WT-1) (28, 34). We evaluated the ability of Sp1 to bind to the exon 2 promoter CACCC box. As expected, GH3 cell nuclear extracts formed a gel shift complex with the 32P-labeled exon 2 -63/-37 double stranded oligomer (Fig. 5AGo, lanes 1 and 2). The binding was not altered by excess unlabeled Oct-1 DNA (lane 3), but was inhibited by excess unlabeled -63/-37 oligomer (lanes 4–6). A 22-bp oligomer containing a consensus Sp1-binding site (see Table 1Go for sequence) inhibited the binding of the labeled -63/-37 exon 2 promoter region by GH3 cell nuclear extracts (lanes 7–9). The magnitude of inhibition of binding was very similar to that produced by the unlabeled -63/-37 oligomer itself (compare lanes 4–6 with lanes 7–9). In confirmation of the data shown in Fig. 2Go, the oligomer containing the consensus AP-2-binding site showed less inhibition of binding than did the unlabeled -63/-37 oligomer or the Sp1 oligomer (Fig. 5AGo, compare lanes 10–12 with lanes 4–9). A double stranded 22-bp oligomer (250-fold molar excess) identical to the Sp1 consensus oligomer except for a 2-bp mutation that prevents Sp1 binding (see Table 1Go for sequence) was unable to inhibit binding of the exon 2 -63/-37 oligomer to GH3 cell nuclear extracts (lane 13).



View larger version (59K):
[in this window]
[in a new window]
  		Figure 5. Sp1 transcription factor binds to the exon 2 promoter CACCC box. A, The oligomer extending from -63 to -37 of the exon 2 promoter (sequence shown in Table 1Go) was end labeled with 32P, and about 0.1 ng (~24,000 cpm) was incubated without (lane 1) or with (lane 2) 1 µg GH3 cell nuclear extract (GH3 N.E.). The indicated unlabeled DNAs were added as competitors in lanes 3–13. Oct-1 is a 22-mer fragment containing an Oct-1 transcription factor-binding site (sequence shown in Table 1Go) and was used at a 100-fold molar excess (lane 3). The -63/-37 oligo (lanes 4–6) is the same oligomer as that labeled for use as the probe. The Sp1 (consensus) is a 22-bp oligomer containing a consensus Sp1 transcription factor-binding site (sequence shown in Table 1Go). The AP-2 (consensus) is the oligomer described in Fig. 3Go. The Sp1 M (250-fold molar excess) is identical to the Sp1 (consensus) except for a 2-bp mutation (shown in boldface in Table 1Go) that prevents Sp1 binding. B, The labeled -63/-37 probe (~0.1 ng; ~30,000 cpm) was incubated without (lane 1) or with GH3 nuclear extract (N.E.; lanes 2–4), OVCAR-3 nuclear extract (N.E.; lane 5), or purified Sp1 protein (1 fpu; lane 6). The indicated unlabeled competitor DNAs at 200-fold molar excess were used in lanes 3 and 4. The labeled oligomer containing the AP-2 consensus binding site was incubated without (lane 7) or with purified AP-2 protein (10 ng; lane 8), purified AP-2 protein and polyclonal anti-AP-2 antibody (10 and 100 ng, respectively; lane 9), GH3 cell and OVCAR-3 cell nuclear extracts (N.E.; 1 µg each; lanes 10 and 11, respectively), or purified Sp1 protein (1 fpu; lane 12). The free and protein-bound probes were separated on 5% native gels, which were dried and autoradiographed overnight at -80 C.

 
The data shown in Fig. 5BGo again showed that GH3 and OVCAR-3 cell nuclear extracts specifically bound the -63/-37 exon 2 probe (lanes 1–5). Incubation of this probe with 1 footprint unit (fpu) of Sp1 protein resulted in formation of a gel shift complex (lane 6). When the labeled AP-2 site probe was used, there was no binding by GH3 or OVCAR-3 cell nuclear extracts (lanes 7, 10, and 11). Purified AP-2 protein produced a somewhat smeared gel shift pattern with the AP-2 probe (lane 8), which could be supershifted by anti-AP-2 antibody (lane 9). Purified Sp1 protein (1 fpu) did not bind to the labeled AP-2 oligomer (lane 12).

To determine further whether Sp1 was a component of the protein-DNA complexes formed between cell nuclear extracts and the -63/-37 oligomer, supershift assays were performed using a polyclonal antibody to Sp1. As shown in Fig. 6AGo, a gel-shifted complex was formed between GH3 cell nuclear extract and the labeled -63/-37 exon 2 promoter oligomer (lane 2) that was not competed by excess unlabeled Oct-1 oligomer (lane 3), but was completely inhibited by excess unlabeled -63/-37 oligomer (lane 4). Addition of Sp1 antibody resulted in formation of a supershifted complex (lane 5) with slower mobility than that produced by GH3 nuclear extracts (lane 2). A portion of the complex was not affected by Sp1 antibody. Purified Sp1 was bound specifically to the labeled -63/-37 oligomer (lanes 6–8). The mobility of the complex formed with purified Sp1 was slower than but overlapped with that formed with nuclear extracts. The binding was competed by the unlabeled Sp1 oligomer, but not by the mutant Sp1 oligomer (lanes 9 and 10). In control experiments, Sp1 antibody was able to supershift the complex formed between different amounts of purified Sp1 protein and the labeled -63/-37 oligomer (lanes 11–13 compared with lanes 14–16).



View larger version (62K):
[in this window]
[in a new window]
  		Figure 6. Effect of anti-Sp1 antibody on mobility shift of protein-DNA complexes formed by GH3 and OVCAR-3 cell nuclear extracts. A, The -63/-37 double stranded exon 2 promoter fragment was end labeled by a Klenow fill-in reaction, and 30,000 cpm (~0.5 ng) were incubated without (lane 1) or with 1 µg nuclear proteins from GH3 cells (lanes 2–5) or 0.25–1 fpu (U) purified recombinant Sp1 protein (lanes 6–16). Unlabeled competitor oligomers (oligo) defined in previous figures were added in the indicated lanes at 100-fold molar excess. One tenth microgram of purified rabbit polyclonal antibody (IgG) against human Sp1 was incubated with GH3 cell nuclear extracts (lane 5) or with purified Sp1 protein (lanes 14–16) before adding the probe. The dried gel was exposed for 12 h at -80 C. The small arrowheads to the left and right indicate the migration of the GH3 nuclear extract-probe and purified Sp1-probe complexes, respectively. The triangles represent the corresponding supershifted complexes. B, An experiment similar to that described in A was performed using OVCAR-3 cell nuclear extracts (lanes 1–7). Lane 8 shows the result obtained when the Sp1 antibody alone was incubated with the labeled probe. Lanes 9–11 show the effect of a 100-fold molar excess of an oligomer containing a consensus C/EBP transcription factor-binding site (lane 11) on the binding of OVCAR-3 cell nuclear extracts (lane 10) to the -63/-37 labeled exon 2 promoter probe. The triangle and the arrowhead to the left show the migration of the gel shift and supershift complexes, respectively.

 
Similar results were obtained using OVCAR-3 cell nuclear extracts (Fig. 6BGo). The -63/-37 oligomer was bound by OVCAR-3 cell nuclear extracts (lane 2). The binding was not altered by excess unlabeled Oct-1 oligomer (lane 3), but was inhibited by excess unlabeled -63/-37 oligomer (lane 4). The Sp1 oligomer completely inhibited binding (lane 5), whereas the mutant Sp1 oligomer was unable to inhibit binding (lane 6). Antibody to Sp1 partially supershifted the complex formed between OVCAR-3 cell nuclear extracts and the labeled -63/-37 oligomer (lane 7). Sp1 antibody itself did not form any complex with the -63/-37 oligomer (lane 8).

The sequence from -63 to -55 in the bottom strand (GAAAGTGTT) could be a binding site for NF-IL6/C/EBP, as a portion of the consensus C/EBP binding site sequence is GAAAGATTG (35). To test this possibility, an unlabeled oligomer containing a consensus C/EBP-binding site (100-fold molar excess) was added to OVCAR-3 cell nuclear extracts and the labeled -63/-37 oligomer. No inhibition of binding was apparent (Fig. 6BGo, lanes 9–11).

To determine whether Sp1 stimulates exon 2 transcription, an Sp1 expression vector under control of the Drosophila Adh promoter was cotransfected with pGL2-Basic, exon 2 -73/+44 construct, p0Luc, or IGF-I receptor promoter construct, in Drosophila SL-2 cells (Fig. 7Go). The IGF-I receptor promoter construct, cloned into p0Luc, contains multiple binding sites for Sp1 and was used as a positive control. The SL-2 cells do not produce endogenous Sp1 and are classically used to study the function of Sp1 (36). Cotransfection of 15 µg Sp1 expression vector (pAdhSp1) with IGF-I receptor promoter resulted in a 33-fold increase in luciferase activity compared with cotransfection of this plasmid with pAdh. This increase was significantly greater (P < 0.01) than the 1.7-fold increase in p0Luc (Fig. 7Go). In contrast, the luciferase activity of the exon 2 -73/+44 promoter construct (3.9-fold) was not significantly increased by Sp1 compared with pGL2-Basic (2.5-fold).



View larger version (17K):
[in this window]
[in a new window]
  		Figure 7. Effect of Sp1 cotransfection on exon 2 promoter activity. Drosophila Schneider line 2 (SL-2) cells were transfected with 5 µg pGL2-Basic, exon 2 -73/+44 in pGL2-Basic, p0Luc, or IGF-I receptor promoter (-476/+640) in p0Luc. Each of the these constructs was cotransfected with 15 µg of either a vector containing the Sp1-coding sequence under control of the Drosophila Adh promoter (pAdhSp1) or the same vector without the Sp1-coding sequence (pAdh). Forty-eight hours after transfection, luciferase activities and protein levels (i.e. A595) in the cellular lysates were assayed, and relative luciferase enzyme activity was calculated. The data are presented as the fold increase in luciferase activity in the presence of pAdhSp1 over that in the presence of pAdh and are shown as the mean ± SEM for four determinations performed in duplicate. One-way ANOVA indicated that Sp1 significantly increased IGF-I receptor promoter activity compared with the promoterless p0Luc plasmid (P < 0.01).

 
Identification of the molecular size of exon 2 promoter-binding proteins
We wished to determine what proteins in nuclear extracts were bound to the exon 2 promoter region. To characterize these proteins by their molecular size, the -236/+44 exon 2 promoter fragment was 32P labeled and incubated without or with GH3 cell nuclear extracts. This was followed by UV cross-linking, DNase digestion, and resolution on SDS-PAGE minigels. In the absence of nuclear extract, no bands were observed (Fig. 8Go, lane 1). In the presence of GH3 cell nuclear extract (5 µg protein), the labeled -236/+44 exon 2 promoter fragment was cross-linked to proteins with molecular sizes of approximately 125 (p125), 76 (p76), 47 (p47), and 38 (p38) kDa (lanes 2 and 6). The intensity of the bands was not reduced by coincubation with a 100-fold molar excess of unlabeled Oct-1 DNA (lane 3) or by a 75-fold molar excess of unlabeled -63/-37 DNA (lane 4). The use of 1 µg GH3 cell nuclear extract barely revealed the presence of cross-linked bands (lane 5). The use of 10 µg nuclear extract (lane 7) did not yield more intense cross-linking than did the use of 5 µg (lane 6). When the unlabeled -63/-37 oligomer was used at a 750-fold molar excess, the intensity of the bands was reduced when 5 µg nuclear protein were cross-linked to the labeled -236/+44 exon 2 promoter fragment (compare lanes 6 and 8). The major cross-linked bands were p38 and p47. The bands corresponding to p125 and p76 were more easily discerned upon longer exposure of the gels. No cross-linked bands were observed when proteinase K was coincubated in the binding reactions (data not shown).



View larger version (54K):
[in this window]
[in a new window]
  		Figure 8. Identification of exon 2 promoter binding proteins by UV cross-linking. The exon 2 -236/+44 promoter fragment was uniformly labeled with 32P by a modified PCR protocol as described in Materials and Methods. Approximately 10 ng labeled probe (~260,000 cpm) were incubated without (lane 1) or with 5 µg (lanes 2 and 6), 1 µg (lane 5), or 10 µg (lane 7) GH3 cell nuclear extract (N.E.) protein. Unlabeled competitor DNA containing an Oct-1-binding site (lane 3) or the unlabeled -63/-37 exon 2 promoter oligomer (lanes 4 and 8) were added at the indicated excess molar concentrations. The DNA-protein complexes formed in the binding reactions were cross-linked with UV irradiation and digested with DNases as described in Materials and Methods. The bound proteins were separated on a 10% SDS-PAGE minigel. The gel was dried and autoradiographed at -80 C for 4 days. The lines and numbers to the left indicate the positions and mol wt, respectively, of prestained protein mol wt markers (high mol wt range from Life Technologies) that were electrophoresed along with samples. The arrowheads to the right indicate the positions of the cross-linked proteins. The radioactive bands at the bottom of the autoradiograph represent digested free probe.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
The sequence CCCCACCC (CACCC box) is an essential exon 2 promoter element
Deletion analysis indicated that the sequence between -73 and +44 (where +1 is the first major transcription start site) of the proximal exon 2 promoter is required for full promoter activity (17). Within this region is a CACCC box at -53. Deletion or point mutations of the CACCC box lead to complete loss of promoter activity in transfection assays and also the loss of binding affinity for proteins in nuclear extracts from IGF-I-producing OVCAR-3 and GH3 cells. Thus, the CACCC box at -53 is the first cis-acting element within the exon 2 promoter that has been demonstrated to be required for basal promoter activity and nuclear protein binding. It will be important to determine whether rat liver exon 2 footprint Pr2F, which partially overlaps the CACCC sequence (18), is dependent on the intact CACCC sequence. Of interest, GH did not alter any of the exon 2 promoter footprints produced by rat liver nuclear extracts, although GH increased the levels of nuclear exon 2 transcripts (18).

The CACCC box was originally defined as an essential element of the ß-globin promoter that was bound by EKLF (28). A naturally occurring point mutation in the CACCC box led to reduced ß-globin expression in ß-thalassemia patients (37). Inactivation of the EKLF gene in knock-out mice results in lethal ß-thalassemia (NOREF>38). It is now recognized that many genes, including the GATA-1 transcription factor gene, the pyruvate kinase gene, and the simian virus 40 enhancer region, contain a CACCC box that is important for transcription (30). The nearness of the exon 2 CACCC box to the transcription initiation cluster suggests that it functions as part of the proximal exon 2 promoter.

The exon 2 promoter CACCC box is bound, but not activated, by the zinc finger transcription factor Sp1
EMSAs and UV cross-linking experiments demonstrated that the exon 2 promoter region was bound by multiple nuclear proteins from GH3 and OVCAR-3 nuclear extracts. Mutation of the CACCC box at -53 abolished the binding affinity of GH3 and OVCAR-3 cell nuclear extracts for exon 2 promoter fragments. Thus, binding of nuclear proteins to the proximal exon 2 promoter region reflects direct binding to the CACCC box and/or binding to sites that is dependent on the CACCC box at -53. As the addition of anti-AP-2 antibody did not produce either inhibition or a supershift of binding of nuclear extracts to the exon 2 promoter oligomer, it is possible that the DNA-protein complexes did not contain any detectable AP-2 transcription factor. Western immunoblot experiments using the AP-2 antibody revealed multiple bands (data not shown) and thus did not allow us to conclude definitively that there is no AP-2 in GH3 or OVCAR-3 cell nuclear extracts. However, purified AP-2 protein did not bind to the exon 2 promoter CACCC box probe, but bound to an oligomer that contained a consensus AP-2-binding site under the same conditions. Expression of exogenous AP-2 did not stimulate exon 2 promoter activity in transient cotransfection assays more than it did a promoterless luciferase expression vector. Thus, the CCCCACCC sequence at -53 in the exon 2 promoter is probably not a functional AP-2-binding site. A similar conclusion was reached regarding the CCCCACCC sequence in the IGFBP-5 gene promoter (20). In this case, two distal overlapping CACCC-like putative AP-2-binding sites were able to bind AP-2 protein, but were not activated by expression of AP-2 in cotransfection assays. The proximal GCCAGGGGC AP-2-like sequence was identified as a functional AP-2 site in the IGFBP-5 promoter (20). The observation that an oligomer containing a consensus AP-2-binding site was able to inhibit exon 2 promoter binding by nuclear extracts with relatively low affinity, suggests that the high affinity CACCC box-binding protein(s) in the nuclear extracts can bind to a canonical AP-2-binding site with low affinity. However, it is also possible that there are very low levels of AP-2 protein in the nuclear extracts from GH3 and OVCAR-3 cells that could bind to the CACCC box with low affinity (39), but not activate transcription (20).

CACCC boxes can potentially be bound by a large family of Kruppel-like zinc finger transcription factors (27, 28, 29, 30, 31, 34, 40). All of these transcription factors contain three zinc finger motifs involved in DNA binding, which are similar to the zinc finger motifs of the Drosophila gap gene Kruppel (27). In the current studies of the exon 2 promoter, an oligomer containing a consensus Sp1-binding site competed for nuclear extract binding to the exon 2 promoter containing the CACCC box with an affinity as great as the promoter oligomer itself. Mutation in the Sp1-binding site prevented the Sp1 oligomer from competing for binding. The exon 2 promoter oligomer was able to directly bind purified Sp1 protein. The diffuse nature of the Sp1-shifted band may be due to the presence of isoforms of Sp1 in the preparation (41). The exon 2 promoter CACCC box appears to be a binding site for at least one member of the Kruppel family of transcription factors, namely Sp1. The general transcription machinery can interact with Sp1, resulting in transcriptional activation (42). The fact that we were unable to observe activation of the proximal IGF-I promoter by Sp1 in cotransfection assays indicates that Sp1 binding alone to this sequence cannot stimulate the general transcription machinery in Drosophila cells. Strong stimulation was observed using the IGF-I receptor promoter construct, which contains multiple GC-rich Sp1-binding sites (21). Thus, we cannot yet conclude whether in the case of the exon 2 promoter, other transcription factors that are not present in SL-2 cells participate in transcriptional activation through the CACCC box or whether the binding of Sp1 to this site is not involved in transcriptional activation. It is possible that a single CACCC box requires other factors that act in a synergistic manner with Sp1 to activate the transcription machinery at the core promoter (43). The observation that several proteins were cross-linked to the proximal exon 2 promoter region supports this view.

The putative exon 2 initiator is not functional in transient transfection assays
A terminal deoxynucleotidyl transferase (TdT) gene initiator-like motif (32) was found surrounding the major exon 2 transcription initiation site (17). Initiators function to determine the transcription start site in TATA-less promoters and to support basal transcription (44, 45). Several lines of evidence indicate that the putative IGF-I exon 2 promoter initiator motif may not be a functional initiator. First, although the putative exon 2 initiator contains the core sequence CTCATA, exon 2 mRNA transcription initiation occurs from multiple nucleotide positions, which is uncommon for a typical initiator (32, 45). Secondly, the mutation in the putative exon 2 motif that was shown to reduce initiator activity in the TdT initiator (32, 45), did not reduce the activity of the partially active -73/+24 exon 2 promoter fragment. In addition, the -73/+44 exon 2 promoter fragment with the mutated initiator sequence was still able to compete with the labeled -236/+44 exon 2 promoter fragment for nuclear protein binding in competition EMSAs. Moreover, the putative exon 2 initiator was not footprinted by rat liver nuclear extracts (18). Thus, nuclear protein(s) binding to the exon 2 promoter was probobly not related to the putative initiator sequence.

Recently, a single GC-rich Sp1 site located in the proximal IGF-II promoter (that is active in adult human liver) was found to be essential for basal promoter activity and for the activation of transcription by liver-enriched transcription factors (46). Whether this site in the IGF-II promoter could be activated by cotransfection with Sp1 in SL-2 cells has not been reported. However, Sp1 alone cannot activate IGF-I exon 2 transcription through the proximal promoter region that contains an Sp1-binding site, the CACCC box, in Drosophila cells. Our future direction will focus on characterization of other potential exon 2 CACCC box-binding proteins, and the mechanisms by which these putative proteins and Sp1 communicate with the general transcription machinery to activate exon 2 transcription initiation.


   Acknowledgments
 
The authors thank Dr. Trevor Williams (Yale University, New Haven, CT) for supplying the AP-2 expression vectors; Dr. Robert Tjian for supplying the Sp1 expression vectors; Drs. Haim Werner, Charles Roberts, and Derek LeRoith for supplying the IGF-I receptor promoter construct; Dr. Allan Brasier for supplying p0Luc; and Dr. Cunming Duan and David Clemmons for supplying the IGFBP-5 promoter construct. We are grateful to Dr. Stephen Hardies of the Department of Biochemistry, University of Texas Health Science Center (San Antonio, TX), for enabling us to search the rat IGF-I proximal promoter sequence using the FINDPATTERNS program. The authors thank Melissa Loyd for providing general technical assistance.


   Footnotes
 
1 This work was supported by NIH Grant DK-47357 and the South Texas Health Research Center at University of Texas Health Science Center (San Antonio, TX). Portions of this work were presented at the 10th International Congress of Endocrinology, San Francisco, CA, June 1996 (Abstract P1–523), and at the 79th Annual Meeting of The Endocrine Society, Minneapolis, MN, June 1997 (Abstract P2–276). Back

Received May 22, 1997.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Adamo ML 1995 Regulation of IGF-I gene expression. Implications for normal and pathological growth. Diabetes Rev 3:2–27
  2. Yang H, Adamo ML, Koval AP, McGuinness MC, Ben-Hur H, Yang Y, LeRoith D, Roberts Jr CT 1995 Alternative leader sequences in insulin-like growth factor I mRNAs modulate translational efficiency and encode multiple signal peptides. Mol Endocrinol 9:1380–1395[Abstract]
  3. Hall LJ, Kajimoto Y, Bichell D, Kim S-W, James PL, Counts D, Nixon LJ, Tobin G, Rotwein P 1992 Functional analysis of the rat insulin-like growth factor-I gene and identification of an IGF-I gene promoter. DNA Cell Biol 11:301–313[Medline]
  4. Shemer J, Adamo ML, Roberts Jr CT, LeRoith D 1992 Tissue-specific transcription start site usage in the leader exons of the rat insulin-like growth factor-I gene: evidence for differential regulation in the developing kidney. Endocrinology 131:2793–2799[Abstract]
  5. Adamo ML, Ben-Hur H, Roberts Jr CT, LeRoith D 1991 Regulation of start site usage in the leader exons of the rat insulin-like growth factor-I gene by development, fasting and diabetes. Mol Endocrinol 5:1677–1686[CrossRef][Medline]
  6. Lowe Jr WL, Adamo M, LeRoith D, Roberts CT Jr 1989 Expression and stability of insulin-like growth factor-I (IGF-I) mRNA splicing variants in the GH3 rat pituitary cell line. Biochem Biophys Res Commun 162:1174–1179[CrossRef][Medline]
  7. Tobin G, Yee D, Brunner N, Rotwein P 1990 A novel human insulin-like growth factor I messenger RNA is expressed in normal and tumor cells. Mol Endocrinol 4:1914–1920[CrossRef][Medline]
  8. Yee D, Morales FR, Hamilton TC, Von Hoff DD 1991 Expression of insulin-like growth factor-I, its binding proteins and its receptor in ovarian cancer. Cancer Res 51:5107–5112[Abstract/Free Full Text]
  9. Lowe Jr WL, Roberts Jr CT, Lasky SR, LeRoith D 1987 Differential expression of alternative 5'-untranslated regions in mRNAs encoding rat insulin-like growth factor-I. Proc Natl Acad Sci USA 84:8946–8950[Abstract/Free Full Text]
  10. Hoyt EC, Van Wyk JJ, Lund PK 1988 Tissue and development specific regulation of a complex family of rat insulin-like growth factor-I messenger ribonucleic acids. Mol Endocrinol 2:1077–1086[CrossRef][Medline]
  11. Adamo M, Lowe WL Jr, LeRoith D, Roberts Jr CT 1989 Insulin-like growth factor-I messenger ribonucleic acids with alternative 5'-untranslated regions are differentially expressed during development of the rat. Endocrinology 124:2737–2744[Abstract]
  12. Kikuchi K, Bichell DP, Rotwein P 1992 Chromatin changes accompany the developmental activation of insulin-like growth factor-I gene transcription. J Biol Chem 267:21505–21511[Abstract/Free Full Text]
  13. Adamo ML, Ben-Hur H, LeRoith D, Roberts Jr CT 1991 Transcription initiation in the two leader exons of the rat IGF-I gene occurs from dispersed vs. localized sites. Biochem Biophys Res Commun 176:887–893[CrossRef][Medline]
  14. Jansen E, Steenbergh PH, LeRoith D, Roberts Jr CT, Sussenbach JS 1991 Identification of multiple transcription start sites in the human insulin-like growth factor-I gene. Mol Cell Endocrinol 78:115–125[CrossRef][Medline]
  15. Simmons JG, Van Wyk JJ, Hoyt EC, Lund PK 1993 Multiple transcription start sites in the rat insulin-like growth factor-I gene give rise to IGF-I mRNAs that encode different IGF-I precursors and are processed differently in vitro. Growth Factors 9:205–223[Medline]
  16. Jansen E, Steenbergh PH, van Schaik FMA, Sussenbach JS 1992 The human IGF-I gene contains two cell type-specifically regulated promoters. Biochem Biophys Res Commun 187:1219–1226[CrossRef][Medline]
  17. Wang X, Yang Y, Adamo ML 1997 Characterization of the rat insulin-like growth factor I gene promoters and identification of a minimal exon 2 promoter. Endocrinology 138:1528–1536[Abstract/Free Full Text]
  18. LeStunff C, Thomas MJ, Rotwein P 1995 Rapid activation of rat insulin-like growth factor-I gene transcription by growth hormone reveals no changes in deoxyribonucleic acid-protein interactions within the second promoter. Endocrinology 136:2230–2237[Abstract]
  19. Bradford MM 1976 A rapid and sensitive method for the quantitation of microgram quantities of protein using the principle of protein-dye binding. Anal Biochem 72:248–254[CrossRef][Medline]
  20. Duan C, Clemmons DR 1995 Transcription factor AP-2 regulates human insulin-like growth factor binding protein-5 gene expression. J Biol Chem 270:24844–24851[Abstract/Free Full Text]
  21. Werner H, Bach MA, Stannard B, Roberts Jr CT, LeRoith D 1992 Structural and functional analysis of the insulin-like growth factor I receptor gene promoter. Mol Endocrinol 6:1545–1558[Abstract]
  22. Henninghausen L, Lubon H 1987 Interaction of protein with DNA in vitro. In: Berger ST, Kimmel AR (eds) Guide to Molecular Cloning Techniques. Academic Press, New York, pp 721–735
  23. Tabor S, Struhl K, Scharf SJ, Gelfand DH 1995 Unit 3.5: DNA-dependent DNA polymerases. Klenow fragment of Escherichia coli DNA polymerase I. In: Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K (eds) Short Protocols in Molecular Biology, ed 3. Wiley and Sons, New York, pp 3–19
  24. Schowalter DB, Sommer SS 1989 The generation of radiolabeled DNA and RNA probes with polymerase chain reaction. Anal Biochem 177:90–94[CrossRef][Medline]
  25. Adamo ML, Stannard B, LeRoith D, Roberts Jr CT 1993 Approaches for the purification, quantitation, and analysis of hormone and receptor mRNAs. In: de Pablo F, Scanes CG, Weintraub BD (eds) Handbook of Endocrine Research Techniques. Academic Press, San Diego, pp 421–455
  26. Chodosh LA 1995 Unit 12.5: UV cross-linking of proteins to nucleic acids. UV cross-linking using a non-BrdU-substituted probe. In: Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith JA, Struhl K (eds) Short Protocols in Molecular Biology, ed 3. Wiley and Sons, New York, pp 12–22
  27. Rosenberg UB, Schroder C, Preiss A, Kienlin A, Cote S, Riede I, Jackle H 1986 Structural homology of the product of the Drosophila Krupple gene with the Xenopus transcription factor IIIA. Nature 319:336–339[CrossRef]
  28. Miller IJ, Bieker JJ 1993 A novel, erythroid cell-specific murine transcription factor that binds to the CACCC element and is related to the Kruppel family of nuclear proteins. Mol Cell Biol 13:2776–2786[Abstract/Free Full Text]
  29. Anderson KP, Kern CB, Crable SC, Lingrell 1995 Isolation of a gene encoding a functional zinc finger protein homologous to erythroid Kruppel-like factor: identification of a new multigene family. Mol Cell Biol 15:5957–5965[Abstract]
  30. Crossley M, Whitelaw E, Perkins A, Williams G, Fujiwara Y, Orkin SH 1996 Isolation and characterization of the cDNA encoding BKLF/TEF-2, a major CACCC-box binding protein in erythroid cells and selected other cells. Mol Cell Biol 16:1695–1705[Abstract]
  31. Gregory RC, Taxman DJ, Seshasayee D, Kensinger MH, Bieker JJ, Wojchowski DM 1996 Functional interaction of GATA1 with erythroid Kruppel-like factor and SP1 at defined erythroid promoters. Blood 87:1793–1801[Abstract/Free Full Text]
  32. Smale ST, Baltimore D 1989 The "initiator" as a transcription control element. Cell 57:103–113[CrossRef][Medline]
  33. Weis L, Reinberg D 1992 Transcription by RNA polymerase II: initiator-directed formation of transcription-competent complexes. FASEB J 6:3300–3309[Abstract]
  34. Call CM, Glaser T, Ito CY, Buckler AJ, Pelletier J, Haber DA, Rose EA, Kral A, Yeger H, Lewis WH, Jones C, Housman DE 1990 Isolation and characterization of a zinc finger polypeptide gene at the human chromosome 11 Wilm’s tumor locus. Cell 60:509–520[CrossRef][Medline]
  35. Faisst S, Meyer S 1992 Compilation of vertebrate-encoded transcription factors. Nucleic Acids Res 20:3–26[Free Full Text]
  36. Courey AJ, Tjian R 1988 Analysis of Sp1 in vivo reveals multiple transcriptional domains, including a novel glutamine-rich activation motif. Cell 55:887–898[CrossRef][Medline]
  37. Feng WC, Southwood CM, Bieker JJ 1994 Analyses of