Vol. 189, No. 2, 1992 December 15, 1992
OF A RAT INSULIN-LIKE GENE PROMOTER
RESEARCH COMMUNICATIONS Pages 972-978
William L. Lowe, Jr.,and Rebecca M. Teasdale
Dept. of Internal Medicine, Room 3E-17, VA Medical Center and University of Iowa College of Medicine, Iowa City, IA 52246 Received
SUMMARY Rat IGF-I mRNAs contain one of two alternative 5’untranslated regions which are encoded by alternative exons (exons 1 and 2) and whose expression is controlled by alternative promoter elements. We investigated the ability of fra ments of DNA which contain exon 1 and its 5’-flanking region to regulate transcription o f a luciferase reporter gene in transient transfection assays. Maximal promoter activity was obtained with a construct which contained 412 bp of 5’flanking re ion, while constructs which contained 1120 and 1690 bp of 5’-flanking region induced -5 %% less enzymatic activity. Mapping of transcription start sites by RNase protection assay demonstrated that native start sites were used by these constructs, although the relative use of different start sites was different from start site usage by the endogenous gene. These data demonstrate that the 5’-flanking B 1992 Academic region of exon 1 is capable of regulating transcription of IGF-I mRNAs. Press.
Insulin-like growth factor I (IGF-I) is a polypeptide with both growth-promoting and metabolic effects which mediates many of the growth-promoting effects of growth hormone (1). Growth hormone is one of the primary regulators of IGF-I gene expression, although metabolic alterations, including fasting and diabetes, are also important in the regulation of tissue IGF-I mRNA levels (2-4). These changes in IGF-I mRNA levels in response to growth hormone and metabolic alterations occur, at least in part, at a transcriptional level (5,6), although the molecular mechanisms responsible for this regulation are unknown. Progress has been made in characterizing the IGF-I gene from various species (4,79). IGF-I mRNAs in rats and humans contain one of two 5’-untranslated regions, which are encoded by separate exons (either exon 1 or exon 2) and whose expression is thought to be regulated by alternative promoter elements. In rats, IGF-I mRNAs which contain exon 1 are present in greatest abundance in all tissues and are the only IGF-I mRNAs expressed in some tissues (10). Mapping of the transcriptional start sites in exon 1 of the human and rat IGF-I gene has demonstrated at least four clusters of transcription initiation sites which are separated by as many as 350 base pairs (bp), which is consistent with the absence of classic TATA and CCAAT motifs in the region 20 to 80 bp upstream of the start sites (8,9, 11,12). In this study, we have now examined the promoter activity of the 5’-flanking region of exon 1 of the IGF-I gene by examining its ability to regulate transcription of a reporter gene in transient transfection assays. 0006-291X/92
Copyright All rights
1992 by Academic Press, reproduction in any form
Isolation. orenaration. and characterization of genomic clones. A rat liver genomic library, which was prepared using a partial &zu3A digest of enomic DNA, was obtained (Promega Biotech, Madison, WI). Five hundred thousan % independent plaques were screened with a “P-labelled IGF-I exon 1 cDNA using standard techniques. The single laque was plaque-purified, and phage DNA was prepared using a Qiagen Lambda aration (Chatsworth, CA) according to the manufacturer’s mstructions. Exon 1 and 5’region of the IGF-I gene were isolated from the 15 kilobase pair genomic clone using tfl e polymerase chain reaction (PCR) with a vector-based S-primer and an oli onucleotide complementary to the terminal 22 nucleotides of exon 1. The resulting 7 ! 4 bp product was purified, digested with BgZII (which removed the terminal 20 bp of exon l), subcloned 5’ to the luciferase structural ene in the promoterless plasmid pGL2 Basic (Promega Biotech) to generate the ‘i GF-I-luciferase fusion plasmid pGL2IGF0.8, and subjected to DNA sequence analysis using an A plied Biosystems 370A DNA se uencer. A second genomic clone was prepared by P 8 R amplification of rat genomic B NA using an oligonucleotide identical to nucleotides 36-57 of the previously described se uence of the Y-flanking region of the rat IGF-I gene (13) and the oligonucleoti 2 e complementary to the terminal 22 nucleotides of exon 1 as primers. The resulting 2050 bp product was purified, digested with BgZII, subcloned into GL2 Basic to generate pGIZGF2.0, and subjected to DNA sequence analysis as describe B above. Assavs for transient exoression of IGF-I-luciferase fusion Dlasmids. Rat dermal fibroblasts were repared from skin of fetal rats at 18-20 days gestation, as described previously (14). -pn e cells were grown in Eagle’s Minimal Essential Medium with 10% fetal calf serum at 37 C in a humtdified atmosphere containing 5% C&. Upon reaching confhtence, cells were replated at a dilution of 1:3 and used at passages 4 and 5 for experiments. For transfection assays, cells lates at a density of 1 x 106 cells per well. C6 cells Eagle’s Medium with 5% fetal calf serum as described assays, cells were split at a density of 1:7 and plated onto for fibroblasts and 3 pg for C6 cells; purified by two two independent reparations of plasmid DNA were used) was transfected into cells using the cationic lipid I5 OTAP according to the manufacturer’s instructions (Boehringer Marmheun, Indianapolis, IN). After incubation for 18 hrs, medium was removed and fresh medium was added. Following an additional 30 hr incubation, cells were washed twice with phosphate buffered salme and incubated on ice in 150 ~1 of lysis buffer (15 mM MgSO4, 1 mM DTT, 0.4 PM EGTA, 1% Triton X-100, 25 mM glycyl- lycine, H 7.8). Protein content of the cell lysate was determined using the method o f Lowry P16). To measure luciferase activity, 65 ~1 of cell lysate from fibroblasts or 40 ~1 from C6 cells were added to 350 ~1 of assay buffer (15 mM MgS04, 1 mM DTT, 0.4 /IM EGTA, 14.5 mM KHzPO4,4.8 mM ATP, 25 mM glycyl-glycine, pH 7.8). The reaction mix was placed in a Monolight 2001 luminometer (Analytical Luminescence Laboratories., San Diego, CA). The light reaction was initiated by the in’ection of 100 ~1 of luciferin mtx ( 0.46 mM luciferin, 15 mM MgSO4, 0.4 PM EGTA, 1 m !I4 DTT, 25 mM glycyl-glycine, pH 7.8), and light emission was measured using a 10 set integration mode. The luciferase activity present in each sample was normalized using the protein content of the sample. RNA isolation and analvsis. Total cellular RNA was repared using the guanidine thiocyanate-cesium chloride method and analyzed 1 y denaturin agarose gel electrophoresis, as described previously (14). For preparation of RNA &rom transfected cells, cells were harvested 48 hrs after transfection, and followin purification of total cellular RNA as described above, RNA was treated with 10 rg/ml %Nase I (Worthington Biochemical Corp., Freehold, NJ) for 15 min at 37 C. Transcription initiation sites were ma ped using a solution hybridtzation/RNase protection assay as described previously (105: Ant’ lsense RNAs for this assay were generated from a SacI-XbaI fragment of pGL2IGF0.8 (see Fig. 2 for a schematic diagram of this fragment) which contained 412 bp of 5’-flanking region, 362 bp of exon 1, and 86 bp of the luciferase structural gene and which was subcloned into pGEM 42 (Promega Biotech).
189, No. 2, 1992
RESULTS Genomic clones were isolated by screening a rat genomic library with an IGF-I exon 1 cDNA and by PCR amplification of rat genomic DNA and subcloned 5’ to the luciferase structural gene to generate IGF-I-luciferase fusion plasmids. The IGF-I-luciferase fusion plasmid pGIJIGFO.8 contained 412 bp of S-flanking region and the first 362 bp of exon 1 (which does not contain an ATG which is in-frame with luciferase), relative to the most 5’ transcription initiation site described by Adam0 et al. (ll), while the IGF-I-luciferase fusion plasmid pGL2IGF2.0 contained 1690 bp of 5’-flanking region and the first 362 bp of exon 1. DNA sequence analysis of the initial 800 bp at the 5’ end of the insert in pGL2IGF2.0 and initial 346 bp at the 5’ end of the insert in pGL2IGF0.8 demonstrated 4 and 3 single base pair differences from the sequence described previously by Shimatsu and Rotwein (13), respectively. DNA sequence analysis of the final 289 bp at the 3’ end of the inserts in pGIZIGF0.8 and pGL2IGF2.0 demonstrated 3 single base pair differences between the two clones. The sequence of this region of pGL2IGF0.8 was identical to the previously published sequence (13). A final clone, pGL2IGF1.5, was prepared using a SmaI restriction digest of pGL2IGF2.0 to generate a construct which contained 1120 bp of 5’-flanking region and the first 362 bp of exon 1. The identity of the 5’ and 3’ ends of this clone were confirmed by DNA sequence analysis. To analyze the activity of the DNA 5’ to exon 1 of the IGF-I gene as a promoter, the IGF-I-luciferase fusion plasmids were used in transient transfection assays. Background activity in all experiments was determined by transfecting cells with the promoterless plasmid pGL2. The fusion plasmids were transfected into a cell line which we have previously used as a model system to study the regulation of IGF-I mRNA levels, rat dermal fibroblasts in primary culture, which express IGF-I mRNAs containing almost exclusively (> 95%) exon 1 (14). All of the constructs increased luciferase activity as compared to the activity present in cells transfected with pGL2 (Fig. 1). Interestingly, pGL2IGF0.8, which contained 412 bp of 5’flanking region, had the greatest activity and increased luciferase activity 8.3-fold as compared to background. Extracts from cells transfected with pGL2IGF2.0 and pGL2IGFl.S increased enzymatic activity 4.1- and 5.2fold over background, respectively. For comparsion, the fusion plasmids were also transfected into an established cell line, rat glioma C6 cells (Fig. 1). These cells have been shown previously to express IGF-I mRNAs which contain only exon 1 (15). Similar results were obtained with these cells. Again, pGL2IGF0.8 increased luciferase activity 8.8-fold as compared to the activity present in cells transfected with pGL2, while pGIXGF2.0 and pGL2IGF1.5 increased luciferase activity 4.3- and 5.2-fold, respectively. As noted, multiple transcription initiation sites which span -350 bp of exon 1 have been described previously (9, 11, 12). To determine which sites were being used by pGL21GF0.8, transcription initiation sites were mapped using a solution hybridization ribonuclease protection assay. For this assay, an 864 base antisense RNA which contained 412 bp of 5’flanking region, 362 bp of exon 1, and 86 bp of the luciferase gene was used (Fig. 2). As demonstrated in Fig. 2, transcription initiation sites used in nontransfected 974
189, No. 2, 1992
4.1 * 0.7
5.2 f 0.4
6.3 t 0.9
C6 Glioma Cells PG~ pGL2lGF2.0
4.3 2 0.6
5.2 f 0.6
6.6 2 1.1
Fig. 1. Identification of romoter activity in the S-flanking region of the rat IGF-I gene. IGF-I-luciferase fusion p Pasmids containing fragments of the IGF-I S-flanking region plus the first 362 bp of exon 1 were transfected into rat dermal fibroblasts in primary culture or rat C6 glioma cells. Cells were harvested 48 hr after transfection, and luciferase activity was measured as described in the Materials and Methoak. All transfections were performed in triplicate. Values represent the relative luciferase activity as compared to the activity in cells transfected with the promoterless plasmid pGL.2 which was defined as 1.0 and are the mean + SEM of 9 to 12 inde endent experiments for the fibroblasts and 8 to 11 independent experiments for the 8 6 glioma cells.
fibroblasts (lane b) are similar to those used in liver (lane c), although subtle differences in the relative use of different start sites are apparent. The transcription start sites which are present in fibroblasts and liver are similar to the previously described transcription initiation sites (indicated by the arrows in Fig. 2) (9, 11, 12), but other protected bands, suggesting the presence of additional minor start sites, are also present. In transfected fibroblasts (lane a), the prominent 152-base protected band (indicated by an arrowhead in Fig. 2) is derived from IGF-I-luciferase mRNAs which utilize a transcription initiation site on the IGF-I-luciferase
fusion plasmid which is 86 bases upstream of the 3’ end of exon 1. A 66-base protected band corresponding to endogenous IGF-I mRNAs with this start site
was also present in nontransfected fibroblasts and liver (data not shown), although this start site has not been previously described. Additional faint protected bands (in lane a of Fig. 2 as indicated by the other arrowheads) which correspond to IGF-I-luciferase mRNAs transcribed from pGL2IGF0.8 which initiate in the vicinity of the previously described native transcription start sites of the endogenous IGF-I gene are also present. In C6 cells, the vast majority of transcription from pGL2IGF0.8 also initiated from the downstream start site seen in transfected fibroblasts (data not shown). The specificity of the protected bands is demonstrated both by the absence of protected bands when RNA prepared from bovine pulmonary artery endothelial cells was used as the target RNA in the assay (lane d) and by the absence of the protected bands corresponding to the transcripts which initiate 975
189, No. 2, 1992
Fig. 2. Identification of transcription initiation sites in transfected fibroblasts. Top, Schematic dia ram of the IGF-I-luciferase fusion lasmid. Depicted is the Sac1 - X!xzI fragment whtc 3 was subcloned into pGEM 42 and tRe antisense RNA which was generated from this template. Also depicted is the primer used to generate the DNA sequencing ladder present on the autoradiogram. Borrotom,Autoradiogram demonstrating transcription initiation sites as determined by solution hybridization/RNase protection assay mapping. The target RNA which was used in the assay was isolated from fibroblasts 48 hr after transfection with pGL2IGF0.8 (lane a), nontransfected fibroblasts (lane b), adult rat liver (lane c), or bovine pulmonary artery endothelial cells (lane d). Protected bands in lanes b and c which were generated by IGF-I mRNAs transcribed from the endogenous IGF-I gene and which utilize the previously described transcription initiation sites (9, 11, 12) are indicated by the arrows. Protected bands in lane a which correspond to RNAs transcribed from the fusion lasmid which initiate in the region of the transcription start sites of the endogenous IG If -1 gene are indicated by arrowheads. ‘Ihe locatton of these protected bands was determined from the DNA sequencing ladder seen on the left which was derived from the plasmid template.
from pGLZGF0.8 when RNA prepared from fibroblasts and C6 cells transfected with pGL2 was used in the assay (data not shown). These data demonstrate that transcription from the IGF-I-luciferase fusion plasmid initiated in the region of the native start sites, 976
189, No. 2, 1992
although there were differences in the relative use of different start sites between the endogenous gene and the fusion plasmid. Relative differences in start site usage between human IGF-I-luciferase fusion plasmids and the endogenous human IGF-I gene (as well as the use of start sites on the fusion plasmid which were not used by the endogenous gene) were also present in a previous study (8). The origin of the two pairs of more slowly migrating protected bands which are 5’ to the upstream transcription initiation site in lane a is unclear, although similar, but fainter, bands were also present in samples from nontransfected cells upon prolonged exposure of the gel.
DISCUSSION In this study, we have characterized the functional role of DNA sequences 5’ to exon 1 of the IGF-I gene as a promoter for the IGF-I gene. Our data demonstrate that IGF-Iluciferase fusion plasmids which contain 412 to 1690 bp of 5’-flanking region are capable of regulating transcription of a reporter gene, although the IGF-I-luciferase fusion plasmid which contained 412 bp of S-flanking region induced twice as much enzymatic activity as fusion plasmids containing 1690 and 1120 bp of 5’flanking region. These data are consistent with the presence of negative regulatory elements between -1690 and -412 of the 5’-flanking region of the IGF-I gene and with basal promoter activity for the IGF-I gene being located in the first 412 bp of 5’-flanking region. Further studies will be required to confirm this and to specifically localize the minimal promoter region and regulatory elements within that region. Previous studies have characterized the rat and human IGF-I promoter (8, 9). In contradistinction to our results, the results of the previous studies suggested the presence of positive regulatory elements between -1711 and -395 of the 5’-flanking region of the IGF-I gene in that these studies demonstrated a maximal 9- to 1Zfold increase in luciferase activity using rat and human fusion plasmids which contained 823 to 1711 bp of S-flanking region, while fusion plasmids which contained 385 to 395 bp of 5’-flanking region stimulated only a 2.6- to 4.0-fold increase in luciferase activity (8,9). A similar pattern was seen with fusion plasmids which contained fragments of the chicken IGF-I gene 5’-flanking region (7). In contrast, in our studies, maximal activity was achieved with a fusion plasmid which contained 412 bp of 5’-flanking region, whereas fusion plasmids which contained 1120 or 1690 bp of 5’-flanking region induced -50% less enzyme activity. Of note, the previous studies which characterized the rat, chicken, and human IGF-I gene promoters were performed with the same cell type, SK-N-MC cells, a human cell line which was not used in these studies. Taken together with the results of these previous studies, our data suggest that there is either cell type specificity or species specificity in the activity of the IGF-I gene promoter which regulates expression of mRNAs which contain exon 1. In summary, these data demonstrate that fusion plasmids which contain fragments of exon 1 of the IGF-I gene and its 5’-flanking region are able to regulate expression of a reporter gene in cells which are known to express the IGF-I gene, demonstrating that IGF-I 977
189, No. 2, 1992
promoter activity is present in this region. Further studies utilizing these constructs should facilitate discerning the molecular mechanisms for the regulation of IGF-I gene expression by growth hormone and different metabolic states. Acknowlednments: The authors would like to thank Drs. Greg Tennyson and Jim Flanagan for many helpful discussions and critical reading of the manuscript and Ms. Elizabeth Hancock for her secretarial assistance. This material is based upon work supported by the Office of Research and Development, Medical Research Service, Department of Veterans Affairs and NIH Diabetes-Endocrinology Research Center Grant DK-25295.
REFERENCES 1. 2. 3.
5: 4: i: 10. 11. 12. 13. 14. 15. 16.
Lowe, Jr., W.L. (1991) In Insulin-Like Growth Factors: Molecular and Cellular Aspects (D. LeRoith, Ed.). pp. 49-85. CRC Press, Inc., Boca Raton, FL. Bomfeldt, K.E., Arn vist, H.J., Enberg, B., Mathews, L.S., and Norstedt, G. (1989) J. Endocrinol. 122,6 9 l-656. Lowe, Jr., W.L. Adamo, M., Werner, H., Roberts, Jr., C.T., and LeRoith D. (1989) J. Clin. Invest. 84,619-626. Rotwein, P. (1991) Growth Factors 5,3-18. Pao, C.-I., Farmer, P.K., Be ovic, S., Goldstein, S., Wu, G.-j., and Phillips, L.S. (1992) Mol. Endocrinol. 6,96 s-977. Straus, D.S., and Takemoto, C.D. (1990) Mol. Endocrinol. 4,91-100. Kajimoto, Y., and Rotwein, P. (1991) J. Biol. Chem. 266,9724-9731. Kim, S.-W., Lajara, R., and Rotwein, P. (1991) Mol. Endocrinol. 5, 1964-1972. Hall,, L.J., Kajimoto, Y., Bichell, D., Kim, S.-W., James, P.L., Counts, D., Nixon, L.J., Tobm, G., and Rotwein, P. (1992) DNA and Cell Biol. 11,301-313. Lowe, Jr., W.L., Roberts, Jr., C.T., L&y, S.R., and LeRoith, D. (1987) Proc. Natl. Acad. Sci. USA. 84,8946-8950. Adamo, ML., Ben-Hur, H., LeRoith D., and Roberts, Jr., C.T. (1991) Biochem. Biophys. Res. Commun. 176,887-893. Adamo, ML., Ben-Hur, H., Roberts, Jr., C.T., and LeRoith D. (1991) Mol. Endocrinol. 5, 1677-1686. Shimatsu, A., and Rotwein, P. (1987) J. Biol. Chem. 262,7894-7900. Lowe, Jr., W.L., Kummer, M., Karpen, C.W., and Wu, X.-D. (1990) Endocrinology 127.2854-2861. Lowe, Jr., W.L., Meyer, T., Karpen, C.W., and Lorentzen, L.R. (1992) Endocrinology 130,2683-2691. Lo O.H., Rosebrough, N.J., Farr, A.L., and Randall, R.J. (1951) J. Biol. Chem. F! 65-272. 193,