DNA AND CELL BIOLOGY Volume 11, Number 4, 1992 Mary Ann Liebert, Inc., Publishers
Functional Analysis of the Rat Insulin-Like Growth Factor I Gene and Identification of an IGF-I Gene Promoter LISA J. HALL,* YOSHITAKA KAJIMOTO,* DAVID BICHELL,§ SUNG-WOON KIM,* PAYTON L. JAMES,* DEBRA COUNTS.t LINDA J. NIXON,* GARRY TOBIN,* and PETER ROTWEIN*>t
ABSTRACT Insulin-like growth factor I (IGF-I) mediates many of the systemic growth-promoting effects of growth hormone and also functions as a locally acting growth stimulator. In mammals, IGF-I gene expression is complicated, as the gene is transcribed and processed into multiple mRNAs (ranging in length from less than 1 to nearly 7.5 kb) that encode at least two protein precursors. As a step toward understanding the regulation of IGF-I, we report the complete organization of the rat IGF-I gene, including identification of the structural determinants for all IGF-I mRNA species, and an initial functional analysis of its promoters. The gene is composed of 6 exons distributed over nearly 80 kb of chromosomal DNA and is structurally heterogeneous. Several transcription start sites were identified within IGF-I exons 1 and 2, adjacent to presumptive promoters 1 and 2, respectively, and at least three polyadenylation sites were mapped to exon 6. To test promoter function, fusion genes were constructed linking fragments of IGF-I DNA to a reporter plasmid. Chimeric genes containing at least 395 bp of DNA from the 5 -flanking region of exon 1 enhanced luciferase activity after transfection into the IGF-I-producing SK-N-MC cell line, while fusion plasmids containing up to 1,300 bp of DNA from the 5 -flanking region of exon 2 were inactive. Relative levels of IGF-I mRNAs containing exons 1 or 2 varied among different rat tissues, although in response to acute or chronic growth hormone treatment both classes of transcripts were induced coordinately in rat liver. These observations represent the first thorough characterization of a mammalian IGF-I gene, and provide a starting point for defining the mechanisms by which growth hormone and other trophic factors regulate IGF-I gene expression.
(IGF-I), 70-residue, sinIandgle-chain peptide that is structurally related insulin IGF-II fundamental nsulin-like growth factor
role in (Humbel, 1989), plays a growth and development as both a mediator of many of the actions of growth hormone (GH) and as a locally acting stimulator of tissue growth and differentiation (Daughaday and Rotwein, 1989). Although the structure of IGF-I is simple (Blundell and Humbel, 1980), the organization and pattern of expression of its gene is complicated, and remains at best incompletely defined. In human beings and rats, the six known IGF-I exons are distributed over more than 70 kb of chromosomal DNA, and in both species multiple large and small mRNAs are generated that
encode at least two precursor proteins (Bell et ai, 1985; Rotwein, 1986; Rotwein et ai, 1986; dePagter-Holthuizen er ai, 1986; Roberts et ai, 1987a,b; Shimatsu and Rotwein, 1987a,b; Bucci et ai, 1989; Lund et ai, 1989; Tobin et ai, 1990; Jansen et ai, 1991; Steenburgh et ai, 1991). Investigation into the mechanisms involved in the regulation of expression of IGF-I by GH and other trophic agents has been hampered by a lack of complete understanding of the IGF-I gene. Although it has been shown that several days of GH treatment enhances IGF-I gene
ai, 1986; Doglio
Metabolism Division, Departments of »Internal Medicine and tpediatrics, ^Department of Biochemistry and Molecular and §Department of Surgery, Washington University School of Medicine, St. Louis, MO 63110.
and also may preferentially stimulate the accumulation of a subset of IGF-I mRNA species (Lowe et ai, 1987), the molecular basis of each of these effects is unknown. The ways in which other local and systemic modifiers of IGF-I
HALL ET AL.
expression act are similarly unknown. As a prelude to an- different levels of activity in rat tissues but are induced coswering these questions, we have sought to characterize ordinately in response to both acute and chronic GH treatIGF-I genes in their entirety and have recently determined ment. the structure of the relatively simple chicken gene (Kajimoto and Rotwein, 1991). We now report our analysis of the rat IGF-I gene, including the identification of its hetMATERIALS AND METHODS
erogeneous 5' and 3' ends and an initial functional assessment of its two promoters. Our results indicate that the generation of multiple IGF-I mRNAs is a consequence of
several mechanisms, including the use of tandem promoters, each with multiple transcription initiation sites, alternative RNA processing, and variable RNA polyadenylation. We also show that the two promoters appear to have
Materials Restriction endonucleases,
polymerases, and other
zymes were purchased from commercial suppliers (United States Biochemical Corp., Cleveland, OH; Perkin-Elmer Cetus, Norwalk, CT; Stratagene Cloning Systems, San Diego, CA; Promega Biotec, Madison, WI; New England
B. Gene 3 4 12 5 6 7 probes I—I-I-1-1-1-1-1
5--1 t t
C. cDNAs pi A9 A7
A8 A21 A3
FIG. 1. Analysis of the 3' end of the rat IGF-I gene. A. Results of Northern blotting experiments using the singlestranded genomic probes depicted in B. The length of each hybridizing band was determined by comparison with RNA size markers. Autoradiographs were performed at -80°C for 16-36 hr, using Kodak XAR 5 X-ray film and a DuPont Lightening Plus intensifying screen. B. Map of exon 6 and its flanking DNA. Sites for Eco RI and Hind III are indicated. The exon is represented by a box, with the coding region solid black and the 3' noncoding region cross-hatched. The 5' intron and 3'-flanking DNA are depicted by thin lines. Polyadenylation sites are indicated by vertical arrows. The two 5' sites were identified previously (Shimatsu and Rotwein, 1987a). C. Length and location of several cDNA clones. Clone pi contains sequences from exon 4, as signified by the dotted line. The presence of a poly(A) tract is indicated at the 3' end of X3.
RAT INSULIN-LIKE GROWTH FACTOR I GENE
acagaaacactcagactaatcatttat ca tagagat at agaact tacaggttaaagt a ttcacaatgt atgtaagagagt t a tga.
FIG. 2. Nucleotide sequence of the 3' end of the rat IGF-I gene. DNA sequences of cDNA clone X3 and of exon 6 were determined over the 309 nucleotides from an Eco RI site to their point of divergence. The cDNA terminates with a poly(A) tail and the gene contains a series of guanine-thymidine oligomers (underlined).
Biolabs, Beverly, MA; Bethesda Research Laboratories, Gaithersburg, MD). Deoxy-, dideoxy-, and ribonucleotide triphosphates were obtained from Pharmacia-LKB Biotechnology Inc. (Piscataway, NJ), and radionuclides ([ct-"P]dATP, [a-32P]dCTP, [a-32P]CTP, and [a-"S]dATP) from Dupont-New England Nuclear (Wilmington, DE) and Amersham Corp. (Arlington Heights, IL). Nitrocellulose filters (BA85) were purchased from Schleicher & Schuell (Keene, NH), and plasmids pGEM3Z and Bluescript were from Promega Biotec and Stratagene, respectively. Tissue culture media and serum were from BRL-GIBCO (Grand Island, NY). Oligonucleotides were prepared on an Applied Biosystems Model 380B DNA synthesizer at the Washington University Protein Chemistry Laboratory.
for 10-14 days under a 12-hr light-dark cycle with free acfood and drinking water before experimental study. Completeness of pituitary ablation was assessed by lack of weight gain during the observation period (age-matched control rats gained 3 grams/day). Female Wistar rats (age 12-15 weeks) were housed under similar conditions. Human recombinant GH in 0.14 M sterile NaCl was administered as a single intraperitoneal injection of 100 ¡ig to hypophysectomized male rats or as 14 daily injections to normally growing female rats. Rats were anesthetized with diethyl ether before sacrifice. All protocols were approved by the Washington University Animal Welfare Committee. cess to
RNA isolation and RNA
Characterization of genomic clones DNA from previously isolated X clones (Shimatsu and Rotwein, 1987a) was mapped after digestion with restriction endonucleases by Southern blot hybridization (Southern, 1975) with "P-labeled IGF-I cDNA fragments or "Plabeled synthetic oligonucleotides. Selected hybridizing fragments were subjected to additional mapping and DNA sequence analysis after being subcloned into bacterial plasmids pGEM3Z or Bluescript. DNA sequence DNA
plasmid DNA templates using dideoxy chain-terminating inhibitors and [a-35S]dATP (Chen and Seeburg, 1985). All sequence information was obtained from both DNA strands. Data were analyzed with computer programs from the University of Wisconsin Genetics Computing Group (Devereux
Experimental animals Male Sprague-Dawley
6-7 weeks of age at Sasco (Indianapolis, IN), and were housed at the Washington University Animal Care Facility
isolated from tissues of 7- to 8-week-old male
Sprague-Dawley rats and 12- to 15-week-old female Wistar rats using guanidinium thiocyanate and guanidine hydrochloride (Chirgwin et ai, 1979). Northern blots were performed following standard procedures, using 5 fig of total cellular RNA that had been separated by size on 0.9% agarose, 2.2 M formaldehyde gels and transferred to nitrocellulose filters by blotting. Hybridizations were performed at 60°C in buffer containing 50% formamide (Melton et ai, 1984) and 1 x 10' cpm of [o>32P]CTP-labeled antisense RNA probes, prepared from linearized plasmid templates using T3, T7, or SP6 RNA polymerase (Melton et ai, 1984). Filters were washed for 1 hr at 65°C in buffer containing 15 mM sodium chloride, 1.5 mM sodium citrate pH 7.0, and 0.1% NaDodS04, and were exposed to Kodak XAR5 film using intensifying screens at -80°C. Solutionhybridization nuclease protection assays followed the protocol of Zinn et al. (1983), using total RNA and "P-labeled IGF-I-specific antisense RNA probes. Protected fragments were separated on 6% polyacrylamide/8.3 M urea gels, and were seen by autoradiography, as outlined above. RNA abundance was quantitated by direct counting of dried gels using a Betascope 603 (Betagen). cDNA
the 3' end of the rat IGF-I gene
A -1000 -800
G A T C Liver tRNA
_2Q— + l-
Sfct + 20-