- Email: [email protected]
Regulation of Insulin-like Growth Factor Binding Protein-5 mRNA Abundance in Rat Intestinal Smooth Muscle
Biochemical and Biophysical Research Communications 275, 422– 427 (2000) doi:10.1006/bbrc.2000.3283, available online at http://www.idealibrary.com on
Regulation of Insulin-like Growth Factor Binding Protein-5 mRNA Abundance in Rat Intestinal Smooth Muscle Y. T. Hou,* X. P. Xin,† L. Li,† and E. M. Zimmerman† ,1 *Department of Anatomy and Cell Biology and †Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan
Received July 18, 2000
IGF-I increases abundance of IGFBP-5 mRNA in rat intestinal smooth muscle cells (RISM), and IGFBP-5 protein in RISM conditioned media. The translational blocker, cycloheximide, decreased the abundance of IGFBP-5 mRNA to undetectable levels, suggesting that IGFBP-5 mRNA integrity is linked to protein synthesis. We studied the mechanism of IGF-I’s effect on IGFBP-5 mRNA, and the role of cytoplasmic proteins in modulating IGFBP-5 mRNA abundance. Anisomycin, emetine, and puromycin abolished IGFBP-5 mRNA as seen with cycloheximide. Cycloheximide had a dose- and time-dependent effect on IGFBP-5 mRNA. IGF-I increased IGFBP-5 nuclear transcripts by reverse transcription-polymerase chain reaction (RTPCR), suggesting that IGF-I acts at least partially by increasing IGFBP-5 mRNA transcription. Protein synthesis inhibitors did not affect IGFBP-5 nuclear transcripts, therefore, they affect only mature mRNA. The IGFBP-5 mRNA 3ⴕ and 5ⴕ UTRs were cloned and their sequences searched for adenosine-uridine rich elements (AUREs), elements shown to regulate RNA stability. RNA mobility gel shift assay showed two protein activities that bind to nt 922 to 2076 of the 3ⴕ UTR, a region that contains an AURE. One protein activity (BA2) was decreased in cytoplasmic extracts from cycloheximide-treated RISM. These data demonstrate that IGFBP-5 mRNA integrity is dependent on protein synthesis. The 3ⴕ UTR of IGFBP-5 contains elements shown to bind proteins important for RNA stability regulation. This region binds RISM cytoplasmic proteins, and may mediate the dramatic effect of cycloheximide on IGFBP-5 abundance. RNA–protein interactions may be important to IGFBP-5 mRNA stability and ultimately, to IGFBP-5 actions. © 2000 Academic Press Key Words: protein synthesis inhibitor; cycloheximide; cytoplasmic extract; IGF-I; IGFBP-5; mRNA binding proteins.
1 To whom correspondence should be addressed at 4410 Kresge III, University of Michigan, Ann Arbor, MI 48109-0589. Fax: 734-7637280. E-mail: [email protected].
0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Insulin-like growth factor binding proteins (IGFBPs) are a family of secreted proteins that can bind IGF-I (insulin-like growth factor I) and modulate its functions (1, 2). One of the IGFBPs, IGFBP-5, is increased in inflamed and fibrotic intestine from patients with Crohn’s disease and rats with experimental enterocolitis (3). IGFBP-5 is unique in that it enhances the actions of IGF-I possibly by associating with extracellular matrix (ECM) proteins near the target cell and accumulating IGF-I near its receptor or protecting it from proteolysis (4, 5). Intestinal smooth muscle cells, the primary site of collagen synthesis in fibrotic intestine in Crohn’s disease, synthesize IGFBP-5. In vitro, IGF-I increased collagen synthesis and synthesis of IGFBP-5 by rat intestinal smooth muscle cells (RISM) (3). Up-regulation of IGFBP-5 by IGF-I may represent a positive feedback loop whereby IGF-I increases synthesis of IGFBP-5 which, in turn, enhances the fibrogenic actions of IGF-I. Regulation of IGF-I and IGFBP-5 in inflamed intestine may be key to the development of intestinal fibrosis in Crohn’s disease. IGF-I increased IGFBP-5 abundance in several in vitro systems including RISM (3, 6, 7). IGF-I increased IGFBP-5 transcription; effects on mRNA stability have not been well studied (3, 6, 7). In RISM, basal and IGF-I stimulated levels of IGFBP-5 mRNA were reduced dramatically by cycloheximide suggesting that de novo protein synthesis is essential for IGFBP-5 mRNA integrity. This observation is often seen in systems where proteins interact with the mRNA untranslated regions (UTRs) that are important for maintaining RNA integrity or regulating stability. The best described of these regulatory sequences are adenosineuridine rich elements (AUREs; 8, 9). AUREs bind cytoplasmic proteins that can modulate the half-life of mRNAs and ultimately, determine protein abundance. IGFBP-5 mRNA size and sequence are highly conserved among Xenopus, chicken, mouse, rat, and human species. The size of IGFBP-5 mRNA is 6 kb, compared with other known IGFBP transcripts that are smaller than 2.5 kb (10). The large size of IGFBP-5
422
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
mRNA results from an extremely long 3⬘ untranslated region (UTR) of 4.4 kb. Given the long 3⬘ UTR of IGFBP-5 mRNA and the dramatic effect of protein synthesis inhibitors on mRNA abundance, we investigated the possibility that cytoplasmic protein(s) stabilize IGFBP-5 mRNA by binding to its 3⬘ UTR. The effects of protein synthesis inhibitors on regulation of IGFBP-5 mRNA and protein abundance were studied. The 3⬘ and 5⬘ UTRs of IGFBP-5 were cloned and 3 AUREs were identified. RISM cytoplasmic extracts were studied using RNA gel mobility shift assay to determine if cytoplasmic proteins bind the 3⬘ UTR of IGFBP-5. MATERIALS AND METHODS Smooth muscle cell culture. RISM from the muscularis externa were isolated from the distal colon of Lewis strain rats by a standard collagenase digestion method (11) and cultured in DMEM media (GibcoBRL, Gaithersburg, MD). Cells were grown to 80% confluence in 100 mm dishes, washed with serum-free DMEM media and then cultured in the serum-free media for 24 h to minimize the effect of serum-binding proteins. IGF-I (UBI, Inc, Lake Placid, NY), cycloheximide, anisomycin, emetine, puromycin, or puromycin aminonucleoside (Sigma Chemical, St. Louis, MO) were added to the culture dishes. Conditioned media was collected for Western analysis and cells were collected for RNA analysis and preparation of cytoplasmic extracts. IGFBP-5 RNA and protein analysis. Total RNA was isolated from RISM by the acid-guanidine thiocyanate phenol-chloroform method (12). For Northern blots (3), 10 g total RNA was loaded in each lane and electrophoresized on 1% agarose-formaldehyde gels and transferred to Nytran membrane (Amsham, Arlington, IL). For slot blots, 5 g RNA was loaded onto the slot blot apparatus (Schleicher & Schuell, Keene, NH) and transferred to the membrane. Blots were probed for IGFBP-5 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using random primed cDNA probes (3, 13). Specific RNA was quantitated by densitometric analysis of exposed autoradiograms. Western ligand blot and immunoblot for IGFBP-5 detection were performed as previously detected (3). RT-PCR nuclear transcript analysis. Nuclear transcript RT-PCR was performed as described previously (14). Briefly, nuclear transcripts were reverse-transcribed from total RNA. The 50 l reaction volume contained 2 g of total RNA, 0.4 mM each of dGTP, dCTP, dATP, dTTP, 2 mM DTT, 10 units of RNA inhibitor (Boehringer Mannheim, Indianapolis, IN), 2 M random primers and 200 units of M-MLV (Moloney Murine Leukemia Virus) reverse transcriptase (Gibco-BRL). Fifty microliters of water was added to the reaction mixtures after 30 min incubation at 42°C, and the samples were heated at 95°C for 5 min. Five microliters of cDNA were used for each 50 l PCR reaction which also contained 2 mM MgCl 2, 0.1 mM each of dGTP, dCTP, dATP, dTTP, 0.4 mM each of two rat IGFBP-5 intron 2 primers (GCACCATCCCAATCTGAT, CTTAGGATGCACGTGGTT) (6, 15), 0.1 mM of each 2 rat -actin intron D primers (ACCACAGCTGAGAGGGAAATGGT, AGAGGTCTTTACGG-ATGTCAACGT), 0.25 l of 10 Ci/l 32P-dCTP (Amsham), and 2.5 units of AmpliTaq DNA polymerase (Perkin-Elmer, Norwalk, CT). The amplification was performed in a Perkin-Elmer thermocycler (Model 480) for 25 cycles of 95°C for 30 s, 58°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 7 min. Co-amplification of IGFBP-5 and increasing amounts of the DNA mimic demonstrated linearity of IGFBP-5 abundance using the experimental conditions described. The mimic was constructed using core DNA flanked by identical PCR primers that were used to amplify IGFBP-5 intron 2 in RT-PCR (6). Five microliters of each PCR mixture were loaded onto 6% polyacrylamide gels and electrophoresized at 15 V per cm for 2 h. The gels
TABLE 1
Primer sequence Riboprobe 1 Riboprobe 2 Riboprobe 3 Riboprobe 4 Riboprobe 5
5⬘-T...GGCTTTGACGCGTCCCT-3⬘ 5⬘-TCTCCGCCTATCCTCCTC-3⬘ 5⬘-T...GGAGGAGGATAGGCGGAA-3⬘ 5⬘-TCTCCTGGCAGGCAATCT-3⬘ 5⬘-T...GGCTTAGGACTGTCTT-3⬘ 5⬘-ATTGC-AAGCCTGCCTTCT-3⬘ 5⬘-T...GGATGCCCTTGTCC-AT-3⬘ 5⬘-AGCCTGTGAAGTGGTGAAAG-3⬘ 5⬘-T...GGCAGCAGTCCTTCTG-3⬘ 5⬘-CTATTCAG-CGGCCGCTTT-3⬘
Primer positions (nt) ⫺6 to 940 922 to 2076 2031 to 3190 3144 to 4032 3704 to 4434
Note. Twenty-three nucleotides of the T7 promoter were synthesized at 5⬘-end of the primers (underlined T...G ⫽ TGTAATACGACTCACTATAGG) for in vitro transcription. The number 1 in primer positions starts at the first nucleotide of the 3⬘ UTR.
were then dried and exposed to X-ray film at ⫺80°C for 5 h with an intensifying screen. Cloning and sequencing the 5⬘- and 3⬘-UTRs. The 0.8 kb 5⬘ UTR and 4.4 kb 3⬘ UTR were cloned using the Marathon cDNA Amplification kit (Clontech Laboratories, Palo Alto, CA). Briefly, two 25-mer DNA oligonucleotides specific for 5⬘ and 3⬘ UTR (5⬘-GCACAGGCGGCCAGCAGCAGGAGGA-3⬘ and 5⬘-TGACGCGT-CCCCTCCCTTCCTCC-3⬘) were synthesized based on published sequence data (13). One microgram of Poly (A) ⫹ RNA from IGF-I treated cells was reverse-transcribed to generate cDNA. Rapid amplification of cDNA ends (RACE) PCR was performed with 5⬘ and 3⬘ UTR specific primers and poly (A) ⫹ primers using a modification of the manufacturer’s recommendations. DNA fragments of correct sizes were isolated and subcloned into pCR2.1 vector (Invitrogen, Carlsbad, CA). Sequencing of the cloned fragments was performed at the Sequencing Core of The University of Michigan. Sequences of the cloned UTRs were submitted to GenBank (Accession No. AF139830). Riboprobes and RNA gel mobility shift assays. For preparation of cytoplasmic extracts, cells were removed from dishes by trypsinization, pelleted, and washed twice with phosphate-buffered saline (PBS) by centrifugation. Cell pellets were resuspended in hypotonic buffer (25 mM Tris–HCl, pH 7.9, 0.5 mM EDTA) and lysed by repetitive cycles of freeze-thaw (6). The lysates were centrifuged at 15,000g at 4°C for 15 min, and the supernatant was removed and stored at ⫺70°C. The protein concentration of the cytoplasmic extracts was determined by the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA). Five PCR primers specific for the (⫹) strand of the 3⬘ UTR were designed (Table 1). Twenty-three nucleotides of the T7 promoter were synthesized at the 5⬘ end of the (⫹) strand primers for in vitro transcription. These primers and their corresponding (⫺) strand primers were used in PCR to generate five 730 –1100 bp DNA fragments. Five riboprobes were transcribed from the DNA fragments using T7 RNA polymerase (Gibco-BRL) and 32P-UTP (Amsham). 2 ⫻ 10 5 cpm of 32P-labeled riboprobes were mixed with 20 or 80 g of cytoplasmic extracts in 10% glycerol, 12 mM Hepes (pH 7.9), 15 mM KCl, 0.25 mM EDTA, 5 mM MgCl 2, and 1.5 ml of E. coli tRNA (200 ng/ml) in a total volume of 15 l at 30°C for 10 min (17). Ten units of RNase T1 (GIBCOBRL) was added and the mixtures were incubated at 37°C for 30 min, followed by 10 min incubation at room temperature with addition of 1.5 l of 50 mg/ml heparin. The mixtures were loaded onto 5% native polyacrylamide gels and electrophoresized at 35 mA for 2 h. Gels were dried and exposed to X-ray films at ⫺80°C. Statistical analysis. Densitometry was performed on scanned autoradiograms using NIH image software. Differences between
423
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Protein synthesis inhibitor-induced IGFBP-5 mRNA degradation. RISM were treated with Ani: anisomycin; Eme: emetine; Puro: puromycin; PuroA: puromycin aminonucleoside at 0, 2, 10, or 20 M for 24 h. Some cells were treated with IGF-I (100 ng/ml) which was added at the same time as the protein synthesis inhibitors. Each lane represents 10 g of RISM total RNA. Blots were simultaneously probed with radiolabled probes for IGFBP-5 (6.0 kb) and GAPDH (2.4 kb).
groups of unpaired samples and slopes of lines were determined by the Mann Whitney U test and analysis of covarience, respectively. Significance was defined as P ⬍ 0.05.
RESULTS Effect of Protein Synthesis Inhibitors on IGFBP-5 mRNA and Protein Abundance Previous experiments showed that cycloheximide decreased the abundance of IGFBP-5 mRNA in RISM to undetectable levels (3). To confirm that this effect was due to protein synthesis inhibition, rather than toxicity of cycloheximide, three additional protein synthesis inhibitors, anisomycin, emetine, and puromycin, were studied. Anisomycin (2 M), emetine (2 M), and puromycin (10 M) decreased the abundance of IGFBP-5 mRNA to undetectable levels (Fig. 1). Puromycin at 5 M dose (data not shown) had the same effect as the 10 M. In addition, puromycin aminonucleoside, an analogue of puromycin that has no effect on protein synthesis, had no effect on IGFBP-5 mRNA. The effect of cycloheximide on IGFBP-5 mRNA was dose- and time-dependent. Cycloheximide (0.1 g/ml) decreased IGFBP-5 mRNA abundance. IGFBP-5 mRNA abundance was decreased to undetectable levels by 2 g/ml cycloheximide (Fig. 2A). RISM treated with cycloheximide (5 g/ml for 2 h demonstrated decreased IGFBP-5 mRNA abundance, at 8 h the mRNA was nearly undetectable (Fig. 2B). Cycloheximide decreased IGFBP-5 protein abundance in RISM conditioned medium by both ligand blot and immunoblot (Fig. 3A). Densitometric data were plotted for immunoblot analyses (Fig. 3B).
tem, an RT-PCR procedure to selectively co-amplify target (IGFBP-5) and control (-actin) nuclear transcripts under conditions shown to give linear increases in signal with increasing target was used (14). In this experiment, regions of IGFBP-5 intron 2 and -actin intron D (364 bp) were co-amplified from total RNA extracted from RISM (Fig. 4A). These samples were previously used in Northern blots and demonstrated over a two fold IGF-I induction in IGFBP-5 mRNA level (3). IGF-I-treated samples demonstrated 24% increase (1.24 ⫾ 0.09; n ⫽ 7, p ⫽ 0.01) in IGFBP-5 nuclear transcript abundance (Fig. 4B) indicating a modest increase in IGFBP-5 transcription in RISM in the presence of IGF-I. Effect of Protein Synthesis Inhibitors on IGFBP-5 Nuclear Transcript To further investigate the effect of protein synthesis inhibitors, RT-PCR for nuclear transcripts was performed using samples treated with anisomycin, emetine, puromycin, and cycloheximide (Fig. 5, lanes 2–7). Abundance of IGFBP-5 intron 2 transcript was no different in cells treated with protein synthesis inhibitors compared with controls, suggesting that protein synthesis inhibitors only affect mature IGFBP-5 mRNA. Effect of IGF-I on Cycloheximide Action The effect of IGF-I on IGFBP-5 mRNA in the presence and absence of cycloheximide was determined by Slot-blot hybridization. The mean densitometric value for IGFBP-5 mRNA abundance at 0, 2, 4, and 8 h after
Effect of IGF-I on IGFBP-5 Nuclear Transcript IGF-I increased the steady state level of IGFBP-5 mRNA in RISM (3; Fig. 1). To investigate whether the IGF-I effect on IGFBP-5 mRNA is due to increased transcription or increased mRNA stability in our sys-
FIG. 2. The effect of cycloheximide on IGFBP-5 mRNA was doseand time-dependent. Each lane represents 10 g of total RNA from RISM treated with increasing concentrations of cycloheximide for 24 h (A) or with 5 g/ml cycloheximide for 0, 2, 4, and 8 h (B). IGFBP-5 and GAPDH mRNA transcripts are indicated.
424
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. The effect of cycloheximide on IGFBP-5 protein abundance. After RISM were treated with IGF-I (100 ng/ml) for 24 h, cycloheximide (5 g/ml) was added to the conditioned media. Media was collected at 0, 2, 4, 8, and 24 h and Western ligand blot using radiolabeled IGF-I (A, top panel; exposed to film for 2 days) and immunoblot using IGFBP-5 specific antibody (30) (A, bottom panel; exposed to film for 10 min) were performed. Each lane represents concentrated conditioned media (30 g protein/lane) from a different cell culture dish. IGFBP-5 (31 kDa) abundance was determined by densitometry and values for immunoblot data were plotted (3B). The mean ⫾ standard error is given from four different experiments.
addition of cycloheximide was plotted, IGF-I decreased the decay rate of IGFBP-5 mRNA in the presence of cycloheximide shown by the change of the slope between the two curves (p ⫽ 0.001; Fig. 6). This may reflect increased transcription of IGFBP-5 by IGF-I or an effect of IGF-I to enhance IGFBP-5 mRNA stability.
FIG. 4. Effect of IGF-I on IGFBP-5 nuclear transcript. IGFBP-5 intron 2 and -actin intron D were co-amplified in RT-PCR performed with DNA derived from 2 g total RNA from untreated RISM and RISM treated with 100 ng/ml IGF-I. IGFBP-5 nuclear transcript abundance was determined by densitometry and normalized with the signal for -actin. The mean ⫾ standard error is given from two different experiments (n ⫽ 7).
UTR of IGFBP-5 and is not present, or present in decreased abundance, in the presence of cycloheximide. A cycloheximide-sensitive RNA binding protein would be a candidate protein for increasing the stability of IGFBP-5 mRNA, and a candidate protein for mediating the profound effect of cycloheximide on the abundance of IGFBP-5 mRNA in RISM. The cloned 3⬘ UTR was transcribed into radiolabeled RNAs that spanned the 3⬘ UTR (Table 1). Each RNA probe was mixed with RISM cytoplasmic extracts. RISM cytoplasmic extracts were prepared from un-
Sequence Analysis of 5⬘ and 3⬘ UTRs of IGFBP-5 5⬘ and 3⬘ UTRs of rat IGFBP-5 were cloned and sequenced. In the 4.4 kb 3⬘ UTR, there were three AUREs: element 1 (963–1068 nt) is 49% U and 19% A; element 2 (3883– 4016 nt) is 40% U and 16% A; element 3 (4242–7378 nt) is 52% U and 15% A. RNA Binding Protein(s) To investigate the possibility that one or more proteins bind to the 3⬘ UTR of IGFBP-5 mRNA, RNA gel mobility shift assays were performed. The aim was to identify a cycloheximide-sensitive RNA binding protein, that is, a cytoplasmic protein that binds the 3⬘
FIG. 5. Effect of protein synthesis inhibitors on IGFBP-5 nuclear transcript abundance. IGFBP-5 intron 2 and -actin were coamplified in RT-PCR performed with DNA derived from 2 g total RNA from untreated RISM (lanes 1, 8, and 9) and RISM treated with anisomycin (lane 2, 5 M; lane 3, 50 M), emetine (lane 4, 5 M; lane 5, 50 M), puromycin (lane 6, 5 M; lane 7, 50 M) and 1, 2, 5, and 10 g/ml of cycloheximide (lanes 10, 11, 12, and 13). DNA standard was the 32P-end-labeled pBR322/MSP1.
425
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 6. The effect of IGF-I on IGFBP-5 mRNA abundance. RISM were treated with either cycloheximide (5 g/ml) or cycloheximide and IGF-I (100 ng/ml) for 0 to 8 h. Total RNA was isolated and three slot blots were performed with 5 g per sample in doublet. Mean densitometric value ⫾ SE for normalized blots is shown for each time point in cultures treated with cycloheximide or cycloheximide and IGF-I.
treated cells, cells treated for 24 h with IGF-I (100 ng/ml) or cells treated with cycloheximide (5 g/ml for 24 h). Samples containing the RNA probe and the cytoplasmic extract were treated with RNase T1 to degrade RNA that was not protected by cytoplasmic proteins. Samples were size separated by polyacrylamide gel electrophoresis. All riboprobes tested showed protein binding activities, however, only riboprobe #2 showed a binding activity (BA2) that was present in untreated RISM cytoplasmic extracts (Fig. 7, lane 4), but not in cycloheximide-treated RISM cytoplasmic extracts (Fig. 7, lane 8). A second binding activity was identified when RISM cytoplasmic extracts were mixed with riboprobe #2 (BA1) however, BA1 abundance did not change when cycloheximide-treated cytoplasmic extracts were used (Fig. 7, lane 4, 17) suggesting that not all binding activities are cycloheximide sensitive. As negative controls, a sample containing all reagents but without cytoplasmic extract did not show any binding activities (Fig. 7, lane 1), and a sample with RISM cytoplasmic extract identical to lane 4 but with proteinase K to degrade all proteins did not show any binding activities (Fig. 7, lane 2). The use of cytoplasmic extracts from RISM treated with 100 ng/ml IGF-I for 24 h did not significantly change the abundance of BA1 or BA2 (Fig. 7, lane 6). DISCUSSION Our data suggest that IGF-I increases abundance of IGFBP-5 mRNA in RISM by transcriptional and posttranscriptional mechanisms. Data using the transcriptional inhibitor 5, 6-dichloro-1--D-ribofuranosylbenzimidazole (DRB) (3) and data presented here using
RT-PCR on nuclear RNA suggest an effect on transcription. In our study, IGF-I increased IGFBP-5 nuclear transcript abundance by 24%. The modest but consistent effect in nuclear transcript abundance is consistent with the twofold increase in RNA abundance seen by Northern analysis. The profound effect of translational inhibitors on IGFBP-5 mRNA abundance suggest that protein synthesis is important for IGFBP-5 mRNA integrity. Our data demonstrate that the relevant proteins are cytoplasmic, and that they affect integrity of mature RNA rather than nuclear RNA. This is consistent with a post-transcriptional mechanism of IGFBP-5 mRNA regulation. Cycloheximide and other protein synthesis inhibitors dramatically decreased basal and IGF-Istimulated IGFBP-5 mRNA in RISM; the effect was time- and dose-dependent. In RISM, 5 g/ml of cycloheximide reduced IGFBP-5 mRNA by 100% and protein level by approximately 70%. In the present study, none of the four protein synthesis inhibitors changed IGFBP-5 nuclear transcript levels, suggesting that their effects on IGFBP-5 mRNA took place in the cyto-
FIG. 7. RNA gel mobility shift assay. Radiolabeled riboprobes were generated by PCR as shown in Table 1. Riboprobes were mixed with cytoplasmic extracts from untreated RISM, RISM treated with IGF-I (100 ng/ml for 24 h) or RISM treated with cycloheximide (5 g/ml for 24 h). RNase T1 “protected” proteins were size separated by polyacrylamide gel electrophoresis. Representative experiment shown using riboprobe #2 (922 to 2076 nt). Lane 1: negative control lane without cytoplasmic extract (cyto extract), Lane 2: negative control lane with cytoplasmic extract from untreated RISM and proteinase K, Lane 3: 20 g cytoplasmic extract from untreated RISM, Lane 4: 80 g cytoplasmic extract from untreated RISM, Lane 5: 20 g cytoplasmic extract from IGF-I treated RISM, Lane 6: 80 g cytoplasmic extract from IGF-I treated RISM, Lane 7: 20 g cytoplasmic extract from cycloheximide treated RISM, Lane 8: 80 g cytoplasmic extract from cycloheximide treated RISM.
426
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
plasm, and that an intermediate protein(s) was involved in regulating IGFBP-5 mRNA stability. The effect of cycloheximide on IGFBP-5 mRNA and protein abundance has been studied in other systems. In aortic smooth muscle cells, cycloheximide abrogated the IGFI-induced increase in IGFBP-5 mRNA but not basal IGFBP-5 mRNA (7). In human fibroblasts, cycloheximide (25 g/ml) was shown to completely inhibit the secretion of all IGFBPs; the effect of cycloheximide on RNA abundance was not studied (18). Many mRNA binding proteins and their cognate binding sites that stabilize or destabilize mRNAs have been identified and characterized (9, 19). One important set of binding sites are the adenosine and uridinerich elements (AUREs) that residue in 3⬘ UTRs of many short-lived mRNAs coding for growth factor, transcription activators, proto-oncogenes, cytokines, and neuropeptides (8, 9). The observation that many unstable mRNAs contain AUREs implies that these sites function as mRNA destabilizing elements. A family of AURE-binding proteins (AUBPs) has been described (8). These protein factors bind to AUREs with different affinities and can either protect mRNA from degradation or recruit RNases and enhance decay. In the present study, a region of the 3⬘ UTR (922 to 2076 nt) of IGFBP-5 mRNA was identified that bound a cycloheximide-sensitive protein activity. This region contains AURE #1. However, it cannot be determined from our studies whether RNA binding proteins bind to the AURE #1, or if the RNA binding proteins were responsible for cycloheximide-induced IGFBP-5 mRNA degradation; future studies will clarify this issue. Interestingly, IGF-I decreased the rate of cycloheximideinduced IGFBP-5 mRNA degradation (Fig. 3), suggesting that IGF-I may regulate the RNA stabilizing protein factors. Our hypothesis is that, after IGFBP-5 gene is transcribed, its nuclear transcript is spliced and exported into cytoplasm where one or more RNA binding proteins attaches to the mRNA. Inhibition of the translation of these RNA stabilizing factors by protein synthesis inhibitors leads to rapid IGFBP-5 mRNA degradation. Since IGF-I is a potent growth factor, it could stimulate the production of the RNA binding protein(s) and therefore, more RNA binding proteins would be available to stabilize IGFBP-5 mRNA. This may be an important mechanism in determining IGFBP-5 abundance and therefore, may be important in determining the biological effects of IGF-I in the intestine and other tissues. REFERENCES 1. Shimasaki, S., Shimonaka, M., Zhang, H-P., and Ling, N. (1991) Identification of five different insulin-like growth factor binding proteins (IGFBPs) from adult rat serum and molecular cloning of a novel IGFBP-5 in rat and human. J. Biol. Chem. 266, 10646 – 10653.
2. Clemmons, D. R. (1997) Insulin-like growth factor binding proteins and their role in controlling IGF actions. Cytokine Growth Factor Rev. 8, 45– 62. 3. Zimmermann, E. M., Li, L., Hou, Y. T., Cannon, M., Christman, G. M., Lund, P. K., and Bitar, K. N. (1997) Insulin-like growth factor I induces collagen and insulin-like growth factor binding protein 5 synthesis in cultured rat intestinal smooth muscle. Am. J. Physiol. 273(36), G875–G882. 4. Nam, T. J., Busby, W., Jr., and Clemmons, D. R. (1997) Insulinlike growth factor binding protein-5 binds to plasminogen activator inhibitor-I. Endocrinology 138, 2972–2978. 5. Zheng, B., Duan, C., and Clemmons, D. R. (1998) The effect of extracellular matrix proteins on porcine smooth muscle cell insulin-like growth factor (IGF) binding protein-5 synthesis and responsiveness to IGF-I. J. Biol. Chem. 273, 8994 –9000. 6. Yu, D., and Canalis, E. (1995) Insulin-like growth factor (IGF) I and retinoic acid induce the synthesis of IGF-binding protein 5 in rat osteoblastic cells. Endocrinology 136, 2000 –2006. 7. Duan, C., Hawes, S. B., Prevette, T., and Clemmons, D. R. (1996) Insulin-like growth factor-I (IGF-I) regulated IGF-binding protein-5 synthesis through transcriptional activation of the gene in aortic smooth muscle cells. J. Biol. Chem. 271, 4280 – 4288. 8. Chen, C-Y. A., and Shyu, A-B. (1995) AU-rich elements: Characterization and importance in mRNA degradation. Trends Biochem. Sci. 20, 465– 470. 9. Ross, J. (1995) mRNA stability in mammalian cells. Microbiol. Rev. 59, 423– 450. 10. James, P. L., Jones, S. B., Busby, W. H., Clemmons, D. R., and Rotwein, P. (1993) A highly conserved insulin-like growth factorbinding protein (IGFBP-5) is expressed during myoblast differentiation. J. Biol. Chem. 268, 22305–22312. 11. Graham, M. F., Diegelmann, R. F., Elson, C. O., Bitar, K. N., and Ehrlich, H. P. (1984) Isolation and culture of human intestinal smooth muscle cells. Proc. Soc. Exp. Biol. Med. 176, 503–506. 12. Chomczynski, P., and Sacchi, N. (1987) Single step method of RNA isolation by acid guanidinium thiocipunate-phenolchloroform extraction. Anal. Biochem. 162, 156 –159. 13. Zhu, X., Ling, N., and Shimasaki, S. (1993) Cloning the rat insulin-like growth factor binding protein-5 gene and DNA sequence analysis of its promoter region. Biochem. Biophysical. Res. Commun. 190, 1045–1052. 14. Hayden, J. M., Strong, D. D., Baylink, D. J., Powell, D. R., Sampath, T. K., and Mohan, S. (1997) Osteogeneic protein-1 stimulates production of insulin-like growth factor binding protein-3 nuclear transcripts in human osteosarcoma cells. Endocrinology 138, 4240 – 4247. 15. Canalis, E., and Gabbitas, B. (1995) Skeletal growth factors regulate the synthesis of insulin-like growth factor binding protein-5 in bone cell cultures. J. Biol. Chem. 270, 10771–10776. 16. Malter, J. S. (1989) Identification of an AUUUA-specific messenger RNA binding protein. Science 246, 664 – 666. 17. Rondon, I. J., MacMillan, L. A., Beckman, B. S., Goldberg, M. A., Schneider, T., Bunn, H. F., and Malter, J. S. (1991) Hypoxia up-regulates the activity of a novel erythropoietin mRNA binding protein. J. Biol. Chem. 266, 16594 –16598. 18. Camacho-Hubner, C., Busby, W. H., Jr., McCusker, R. H., Wright, G., and Clemmons, D. R. (1992) Identification of the forms of insulin-like growth factor binding proteins produced by human fibroblasts and the mechanisms that regulate their secretion. J. Biol. Chem. 267, 11949 –11956. 19. Caponicro, G., and Parker, R. (1996) Mechanisms and Control of mRNA turnover in Saccharomyces cerevisae. Microbiol. Rev. 60, 23–249.
427
Regulation of Insulin-like Growth Factor Binding Protein-5 mRNA Abundance in Rat Intestinal Smooth Muscle Y. T. Hou,* X. P. Xin,† L. Li,† and E. M. Zimmerman† ,1 *Department of Anatomy and Cell Biology and †Department of Internal Medicine, University of Michigan, Ann Arbor, Michigan
Received July 18, 2000
IGF-I increases abundance of IGFBP-5 mRNA in rat intestinal smooth muscle cells (RISM), and IGFBP-5 protein in RISM conditioned media. The translational blocker, cycloheximide, decreased the abundance of IGFBP-5 mRNA to undetectable levels, suggesting that IGFBP-5 mRNA integrity is linked to protein synthesis. We studied the mechanism of IGF-I’s effect on IGFBP-5 mRNA, and the role of cytoplasmic proteins in modulating IGFBP-5 mRNA abundance. Anisomycin, emetine, and puromycin abolished IGFBP-5 mRNA as seen with cycloheximide. Cycloheximide had a dose- and time-dependent effect on IGFBP-5 mRNA. IGF-I increased IGFBP-5 nuclear transcripts by reverse transcription-polymerase chain reaction (RTPCR), suggesting that IGF-I acts at least partially by increasing IGFBP-5 mRNA transcription. Protein synthesis inhibitors did not affect IGFBP-5 nuclear transcripts, therefore, they affect only mature mRNA. The IGFBP-5 mRNA 3ⴕ and 5ⴕ UTRs were cloned and their sequences searched for adenosine-uridine rich elements (AUREs), elements shown to regulate RNA stability. RNA mobility gel shift assay showed two protein activities that bind to nt 922 to 2076 of the 3ⴕ UTR, a region that contains an AURE. One protein activity (BA2) was decreased in cytoplasmic extracts from cycloheximide-treated RISM. These data demonstrate that IGFBP-5 mRNA integrity is dependent on protein synthesis. The 3ⴕ UTR of IGFBP-5 contains elements shown to bind proteins important for RNA stability regulation. This region binds RISM cytoplasmic proteins, and may mediate the dramatic effect of cycloheximide on IGFBP-5 abundance. RNA–protein interactions may be important to IGFBP-5 mRNA stability and ultimately, to IGFBP-5 actions. © 2000 Academic Press Key Words: protein synthesis inhibitor; cycloheximide; cytoplasmic extract; IGF-I; IGFBP-5; mRNA binding proteins.
1 To whom correspondence should be addressed at 4410 Kresge III, University of Michigan, Ann Arbor, MI 48109-0589. Fax: 734-7637280. E-mail: [email protected].
0006-291X/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Insulin-like growth factor binding proteins (IGFBPs) are a family of secreted proteins that can bind IGF-I (insulin-like growth factor I) and modulate its functions (1, 2). One of the IGFBPs, IGFBP-5, is increased in inflamed and fibrotic intestine from patients with Crohn’s disease and rats with experimental enterocolitis (3). IGFBP-5 is unique in that it enhances the actions of IGF-I possibly by associating with extracellular matrix (ECM) proteins near the target cell and accumulating IGF-I near its receptor or protecting it from proteolysis (4, 5). Intestinal smooth muscle cells, the primary site of collagen synthesis in fibrotic intestine in Crohn’s disease, synthesize IGFBP-5. In vitro, IGF-I increased collagen synthesis and synthesis of IGFBP-5 by rat intestinal smooth muscle cells (RISM) (3). Up-regulation of IGFBP-5 by IGF-I may represent a positive feedback loop whereby IGF-I increases synthesis of IGFBP-5 which, in turn, enhances the fibrogenic actions of IGF-I. Regulation of IGF-I and IGFBP-5 in inflamed intestine may be key to the development of intestinal fibrosis in Crohn’s disease. IGF-I increased IGFBP-5 abundance in several in vitro systems including RISM (3, 6, 7). IGF-I increased IGFBP-5 transcription; effects on mRNA stability have not been well studied (3, 6, 7). In RISM, basal and IGF-I stimulated levels of IGFBP-5 mRNA were reduced dramatically by cycloheximide suggesting that de novo protein synthesis is essential for IGFBP-5 mRNA integrity. This observation is often seen in systems where proteins interact with the mRNA untranslated regions (UTRs) that are important for maintaining RNA integrity or regulating stability. The best described of these regulatory sequences are adenosineuridine rich elements (AUREs; 8, 9). AUREs bind cytoplasmic proteins that can modulate the half-life of mRNAs and ultimately, determine protein abundance. IGFBP-5 mRNA size and sequence are highly conserved among Xenopus, chicken, mouse, rat, and human species. The size of IGFBP-5 mRNA is 6 kb, compared with other known IGFBP transcripts that are smaller than 2.5 kb (10). The large size of IGFBP-5
422
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
mRNA results from an extremely long 3⬘ untranslated region (UTR) of 4.4 kb. Given the long 3⬘ UTR of IGFBP-5 mRNA and the dramatic effect of protein synthesis inhibitors on mRNA abundance, we investigated the possibility that cytoplasmic protein(s) stabilize IGFBP-5 mRNA by binding to its 3⬘ UTR. The effects of protein synthesis inhibitors on regulation of IGFBP-5 mRNA and protein abundance were studied. The 3⬘ and 5⬘ UTRs of IGFBP-5 were cloned and 3 AUREs were identified. RISM cytoplasmic extracts were studied using RNA gel mobility shift assay to determine if cytoplasmic proteins bind the 3⬘ UTR of IGFBP-5. MATERIALS AND METHODS Smooth muscle cell culture. RISM from the muscularis externa were isolated from the distal colon of Lewis strain rats by a standard collagenase digestion method (11) and cultured in DMEM media (GibcoBRL, Gaithersburg, MD). Cells were grown to 80% confluence in 100 mm dishes, washed with serum-free DMEM media and then cultured in the serum-free media for 24 h to minimize the effect of serum-binding proteins. IGF-I (UBI, Inc, Lake Placid, NY), cycloheximide, anisomycin, emetine, puromycin, or puromycin aminonucleoside (Sigma Chemical, St. Louis, MO) were added to the culture dishes. Conditioned media was collected for Western analysis and cells were collected for RNA analysis and preparation of cytoplasmic extracts. IGFBP-5 RNA and protein analysis. Total RNA was isolated from RISM by the acid-guanidine thiocyanate phenol-chloroform method (12). For Northern blots (3), 10 g total RNA was loaded in each lane and electrophoresized on 1% agarose-formaldehyde gels and transferred to Nytran membrane (Amsham, Arlington, IL). For slot blots, 5 g RNA was loaded onto the slot blot apparatus (Schleicher & Schuell, Keene, NH) and transferred to the membrane. Blots were probed for IGFBP-5 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) using random primed cDNA probes (3, 13). Specific RNA was quantitated by densitometric analysis of exposed autoradiograms. Western ligand blot and immunoblot for IGFBP-5 detection were performed as previously detected (3). RT-PCR nuclear transcript analysis. Nuclear transcript RT-PCR was performed as described previously (14). Briefly, nuclear transcripts were reverse-transcribed from total RNA. The 50 l reaction volume contained 2 g of total RNA, 0.4 mM each of dGTP, dCTP, dATP, dTTP, 2 mM DTT, 10 units of RNA inhibitor (Boehringer Mannheim, Indianapolis, IN), 2 M random primers and 200 units of M-MLV (Moloney Murine Leukemia Virus) reverse transcriptase (Gibco-BRL). Fifty microliters of water was added to the reaction mixtures after 30 min incubation at 42°C, and the samples were heated at 95°C for 5 min. Five microliters of cDNA were used for each 50 l PCR reaction which also contained 2 mM MgCl 2, 0.1 mM each of dGTP, dCTP, dATP, dTTP, 0.4 mM each of two rat IGFBP-5 intron 2 primers (GCACCATCCCAATCTGAT, CTTAGGATGCACGTGGTT) (6, 15), 0.1 mM of each 2 rat -actin intron D primers (ACCACAGCTGAGAGGGAAATGGT, AGAGGTCTTTACGG-ATGTCAACGT), 0.25 l of 10 Ci/l 32P-dCTP (Amsham), and 2.5 units of AmpliTaq DNA polymerase (Perkin-Elmer, Norwalk, CT). The amplification was performed in a Perkin-Elmer thermocycler (Model 480) for 25 cycles of 95°C for 30 s, 58°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 7 min. Co-amplification of IGFBP-5 and increasing amounts of the DNA mimic demonstrated linearity of IGFBP-5 abundance using the experimental conditions described. The mimic was constructed using core DNA flanked by identical PCR primers that were used to amplify IGFBP-5 intron 2 in RT-PCR (6). Five microliters of each PCR mixture were loaded onto 6% polyacrylamide gels and electrophoresized at 15 V per cm for 2 h. The gels
TABLE 1
Primer sequence Riboprobe 1 Riboprobe 2 Riboprobe 3 Riboprobe 4 Riboprobe 5
5⬘-T...GGCTTTGACGCGTCCCT-3⬘ 5⬘-TCTCCGCCTATCCTCCTC-3⬘ 5⬘-T...GGAGGAGGATAGGCGGAA-3⬘ 5⬘-TCTCCTGGCAGGCAATCT-3⬘ 5⬘-T...GGCTTAGGACTGTCTT-3⬘ 5⬘-ATTGC-AAGCCTGCCTTCT-3⬘ 5⬘-T...GGATGCCCTTGTCC-AT-3⬘ 5⬘-AGCCTGTGAAGTGGTGAAAG-3⬘ 5⬘-T...GGCAGCAGTCCTTCTG-3⬘ 5⬘-CTATTCAG-CGGCCGCTTT-3⬘
Primer positions (nt) ⫺6 to 940 922 to 2076 2031 to 3190 3144 to 4032 3704 to 4434
Note. Twenty-three nucleotides of the T7 promoter were synthesized at 5⬘-end of the primers (underlined T...G ⫽ TGTAATACGACTCACTATAGG) for in vitro transcription. The number 1 in primer positions starts at the first nucleotide of the 3⬘ UTR.
were then dried and exposed to X-ray film at ⫺80°C for 5 h with an intensifying screen. Cloning and sequencing the 5⬘- and 3⬘-UTRs. The 0.8 kb 5⬘ UTR and 4.4 kb 3⬘ UTR were cloned using the Marathon cDNA Amplification kit (Clontech Laboratories, Palo Alto, CA). Briefly, two 25-mer DNA oligonucleotides specific for 5⬘ and 3⬘ UTR (5⬘-GCACAGGCGGCCAGCAGCAGGAGGA-3⬘ and 5⬘-TGACGCGT-CCCCTCCCTTCCTCC-3⬘) were synthesized based on published sequence data (13). One microgram of Poly (A) ⫹ RNA from IGF-I treated cells was reverse-transcribed to generate cDNA. Rapid amplification of cDNA ends (RACE) PCR was performed with 5⬘ and 3⬘ UTR specific primers and poly (A) ⫹ primers using a modification of the manufacturer’s recommendations. DNA fragments of correct sizes were isolated and subcloned into pCR2.1 vector (Invitrogen, Carlsbad, CA). Sequencing of the cloned fragments was performed at the Sequencing Core of The University of Michigan. Sequences of the cloned UTRs were submitted to GenBank (Accession No. AF139830). Riboprobes and RNA gel mobility shift assays. For preparation of cytoplasmic extracts, cells were removed from dishes by trypsinization, pelleted, and washed twice with phosphate-buffered saline (PBS) by centrifugation. Cell pellets were resuspended in hypotonic buffer (25 mM Tris–HCl, pH 7.9, 0.5 mM EDTA) and lysed by repetitive cycles of freeze-thaw (6). The lysates were centrifuged at 15,000g at 4°C for 15 min, and the supernatant was removed and stored at ⫺70°C. The protein concentration of the cytoplasmic extracts was determined by the Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA). Five PCR primers specific for the (⫹) strand of the 3⬘ UTR were designed (Table 1). Twenty-three nucleotides of the T7 promoter were synthesized at the 5⬘ end of the (⫹) strand primers for in vitro transcription. These primers and their corresponding (⫺) strand primers were used in PCR to generate five 730 –1100 bp DNA fragments. Five riboprobes were transcribed from the DNA fragments using T7 RNA polymerase (Gibco-BRL) and 32P-UTP (Amsham). 2 ⫻ 10 5 cpm of 32P-labeled riboprobes were mixed with 20 or 80 g of cytoplasmic extracts in 10% glycerol, 12 mM Hepes (pH 7.9), 15 mM KCl, 0.25 mM EDTA, 5 mM MgCl 2, and 1.5 ml of E. coli tRNA (200 ng/ml) in a total volume of 15 l at 30°C for 10 min (17). Ten units of RNase T1 (GIBCOBRL) was added and the mixtures were incubated at 37°C for 30 min, followed by 10 min incubation at room temperature with addition of 1.5 l of 50 mg/ml heparin. The mixtures were loaded onto 5% native polyacrylamide gels and electrophoresized at 35 mA for 2 h. Gels were dried and exposed to X-ray films at ⫺80°C. Statistical analysis. Densitometry was performed on scanned autoradiograms using NIH image software. Differences between
423
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 1. Protein synthesis inhibitor-induced IGFBP-5 mRNA degradation. RISM were treated with Ani: anisomycin; Eme: emetine; Puro: puromycin; PuroA: puromycin aminonucleoside at 0, 2, 10, or 20 M for 24 h. Some cells were treated with IGF-I (100 ng/ml) which was added at the same time as the protein synthesis inhibitors. Each lane represents 10 g of RISM total RNA. Blots were simultaneously probed with radiolabled probes for IGFBP-5 (6.0 kb) and GAPDH (2.4 kb).
groups of unpaired samples and slopes of lines were determined by the Mann Whitney U test and analysis of covarience, respectively. Significance was defined as P ⬍ 0.05.
RESULTS Effect of Protein Synthesis Inhibitors on IGFBP-5 mRNA and Protein Abundance Previous experiments showed that cycloheximide decreased the abundance of IGFBP-5 mRNA in RISM to undetectable levels (3). To confirm that this effect was due to protein synthesis inhibition, rather than toxicity of cycloheximide, three additional protein synthesis inhibitors, anisomycin, emetine, and puromycin, were studied. Anisomycin (2 M), emetine (2 M), and puromycin (10 M) decreased the abundance of IGFBP-5 mRNA to undetectable levels (Fig. 1). Puromycin at 5 M dose (data not shown) had the same effect as the 10 M. In addition, puromycin aminonucleoside, an analogue of puromycin that has no effect on protein synthesis, had no effect on IGFBP-5 mRNA. The effect of cycloheximide on IGFBP-5 mRNA was dose- and time-dependent. Cycloheximide (0.1 g/ml) decreased IGFBP-5 mRNA abundance. IGFBP-5 mRNA abundance was decreased to undetectable levels by 2 g/ml cycloheximide (Fig. 2A). RISM treated with cycloheximide (5 g/ml for 2 h demonstrated decreased IGFBP-5 mRNA abundance, at 8 h the mRNA was nearly undetectable (Fig. 2B). Cycloheximide decreased IGFBP-5 protein abundance in RISM conditioned medium by both ligand blot and immunoblot (Fig. 3A). Densitometric data were plotted for immunoblot analyses (Fig. 3B).
tem, an RT-PCR procedure to selectively co-amplify target (IGFBP-5) and control (-actin) nuclear transcripts under conditions shown to give linear increases in signal with increasing target was used (14). In this experiment, regions of IGFBP-5 intron 2 and -actin intron D (364 bp) were co-amplified from total RNA extracted from RISM (Fig. 4A). These samples were previously used in Northern blots and demonstrated over a two fold IGF-I induction in IGFBP-5 mRNA level (3). IGF-I-treated samples demonstrated 24% increase (1.24 ⫾ 0.09; n ⫽ 7, p ⫽ 0.01) in IGFBP-5 nuclear transcript abundance (Fig. 4B) indicating a modest increase in IGFBP-5 transcription in RISM in the presence of IGF-I. Effect of Protein Synthesis Inhibitors on IGFBP-5 Nuclear Transcript To further investigate the effect of protein synthesis inhibitors, RT-PCR for nuclear transcripts was performed using samples treated with anisomycin, emetine, puromycin, and cycloheximide (Fig. 5, lanes 2–7). Abundance of IGFBP-5 intron 2 transcript was no different in cells treated with protein synthesis inhibitors compared with controls, suggesting that protein synthesis inhibitors only affect mature IGFBP-5 mRNA. Effect of IGF-I on Cycloheximide Action The effect of IGF-I on IGFBP-5 mRNA in the presence and absence of cycloheximide was determined by Slot-blot hybridization. The mean densitometric value for IGFBP-5 mRNA abundance at 0, 2, 4, and 8 h after
Effect of IGF-I on IGFBP-5 Nuclear Transcript IGF-I increased the steady state level of IGFBP-5 mRNA in RISM (3; Fig. 1). To investigate whether the IGF-I effect on IGFBP-5 mRNA is due to increased transcription or increased mRNA stability in our sys-
FIG. 2. The effect of cycloheximide on IGFBP-5 mRNA was doseand time-dependent. Each lane represents 10 g of total RNA from RISM treated with increasing concentrations of cycloheximide for 24 h (A) or with 5 g/ml cycloheximide for 0, 2, 4, and 8 h (B). IGFBP-5 and GAPDH mRNA transcripts are indicated.
424
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 3. The effect of cycloheximide on IGFBP-5 protein abundance. After RISM were treated with IGF-I (100 ng/ml) for 24 h, cycloheximide (5 g/ml) was added to the conditioned media. Media was collected at 0, 2, 4, 8, and 24 h and Western ligand blot using radiolabeled IGF-I (A, top panel; exposed to film for 2 days) and immunoblot using IGFBP-5 specific antibody (30) (A, bottom panel; exposed to film for 10 min) were performed. Each lane represents concentrated conditioned media (30 g protein/lane) from a different cell culture dish. IGFBP-5 (31 kDa) abundance was determined by densitometry and values for immunoblot data were plotted (3B). The mean ⫾ standard error is given from four different experiments.
addition of cycloheximide was plotted, IGF-I decreased the decay rate of IGFBP-5 mRNA in the presence of cycloheximide shown by the change of the slope between the two curves (p ⫽ 0.001; Fig. 6). This may reflect increased transcription of IGFBP-5 by IGF-I or an effect of IGF-I to enhance IGFBP-5 mRNA stability.
FIG. 4. Effect of IGF-I on IGFBP-5 nuclear transcript. IGFBP-5 intron 2 and -actin intron D were co-amplified in RT-PCR performed with DNA derived from 2 g total RNA from untreated RISM and RISM treated with 100 ng/ml IGF-I. IGFBP-5 nuclear transcript abundance was determined by densitometry and normalized with the signal for -actin. The mean ⫾ standard error is given from two different experiments (n ⫽ 7).
UTR of IGFBP-5 and is not present, or present in decreased abundance, in the presence of cycloheximide. A cycloheximide-sensitive RNA binding protein would be a candidate protein for increasing the stability of IGFBP-5 mRNA, and a candidate protein for mediating the profound effect of cycloheximide on the abundance of IGFBP-5 mRNA in RISM. The cloned 3⬘ UTR was transcribed into radiolabeled RNAs that spanned the 3⬘ UTR (Table 1). Each RNA probe was mixed with RISM cytoplasmic extracts. RISM cytoplasmic extracts were prepared from un-
Sequence Analysis of 5⬘ and 3⬘ UTRs of IGFBP-5 5⬘ and 3⬘ UTRs of rat IGFBP-5 were cloned and sequenced. In the 4.4 kb 3⬘ UTR, there were three AUREs: element 1 (963–1068 nt) is 49% U and 19% A; element 2 (3883– 4016 nt) is 40% U and 16% A; element 3 (4242–7378 nt) is 52% U and 15% A. RNA Binding Protein(s) To investigate the possibility that one or more proteins bind to the 3⬘ UTR of IGFBP-5 mRNA, RNA gel mobility shift assays were performed. The aim was to identify a cycloheximide-sensitive RNA binding protein, that is, a cytoplasmic protein that binds the 3⬘
FIG. 5. Effect of protein synthesis inhibitors on IGFBP-5 nuclear transcript abundance. IGFBP-5 intron 2 and -actin were coamplified in RT-PCR performed with DNA derived from 2 g total RNA from untreated RISM (lanes 1, 8, and 9) and RISM treated with anisomycin (lane 2, 5 M; lane 3, 50 M), emetine (lane 4, 5 M; lane 5, 50 M), puromycin (lane 6, 5 M; lane 7, 50 M) and 1, 2, 5, and 10 g/ml of cycloheximide (lanes 10, 11, 12, and 13). DNA standard was the 32P-end-labeled pBR322/MSP1.
425
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
FIG. 6. The effect of IGF-I on IGFBP-5 mRNA abundance. RISM were treated with either cycloheximide (5 g/ml) or cycloheximide and IGF-I (100 ng/ml) for 0 to 8 h. Total RNA was isolated and three slot blots were performed with 5 g per sample in doublet. Mean densitometric value ⫾ SE for normalized blots is shown for each time point in cultures treated with cycloheximide or cycloheximide and IGF-I.
treated cells, cells treated for 24 h with IGF-I (100 ng/ml) or cells treated with cycloheximide (5 g/ml for 24 h). Samples containing the RNA probe and the cytoplasmic extract were treated with RNase T1 to degrade RNA that was not protected by cytoplasmic proteins. Samples were size separated by polyacrylamide gel electrophoresis. All riboprobes tested showed protein binding activities, however, only riboprobe #2 showed a binding activity (BA2) that was present in untreated RISM cytoplasmic extracts (Fig. 7, lane 4), but not in cycloheximide-treated RISM cytoplasmic extracts (Fig. 7, lane 8). A second binding activity was identified when RISM cytoplasmic extracts were mixed with riboprobe #2 (BA1) however, BA1 abundance did not change when cycloheximide-treated cytoplasmic extracts were used (Fig. 7, lane 4, 17) suggesting that not all binding activities are cycloheximide sensitive. As negative controls, a sample containing all reagents but without cytoplasmic extract did not show any binding activities (Fig. 7, lane 1), and a sample with RISM cytoplasmic extract identical to lane 4 but with proteinase K to degrade all proteins did not show any binding activities (Fig. 7, lane 2). The use of cytoplasmic extracts from RISM treated with 100 ng/ml IGF-I for 24 h did not significantly change the abundance of BA1 or BA2 (Fig. 7, lane 6). DISCUSSION Our data suggest that IGF-I increases abundance of IGFBP-5 mRNA in RISM by transcriptional and posttranscriptional mechanisms. Data using the transcriptional inhibitor 5, 6-dichloro-1--D-ribofuranosylbenzimidazole (DRB) (3) and data presented here using
RT-PCR on nuclear RNA suggest an effect on transcription. In our study, IGF-I increased IGFBP-5 nuclear transcript abundance by 24%. The modest but consistent effect in nuclear transcript abundance is consistent with the twofold increase in RNA abundance seen by Northern analysis. The profound effect of translational inhibitors on IGFBP-5 mRNA abundance suggest that protein synthesis is important for IGFBP-5 mRNA integrity. Our data demonstrate that the relevant proteins are cytoplasmic, and that they affect integrity of mature RNA rather than nuclear RNA. This is consistent with a post-transcriptional mechanism of IGFBP-5 mRNA regulation. Cycloheximide and other protein synthesis inhibitors dramatically decreased basal and IGF-Istimulated IGFBP-5 mRNA in RISM; the effect was time- and dose-dependent. In RISM, 5 g/ml of cycloheximide reduced IGFBP-5 mRNA by 100% and protein level by approximately 70%. In the present study, none of the four protein synthesis inhibitors changed IGFBP-5 nuclear transcript levels, suggesting that their effects on IGFBP-5 mRNA took place in the cyto-
FIG. 7. RNA gel mobility shift assay. Radiolabeled riboprobes were generated by PCR as shown in Table 1. Riboprobes were mixed with cytoplasmic extracts from untreated RISM, RISM treated with IGF-I (100 ng/ml for 24 h) or RISM treated with cycloheximide (5 g/ml for 24 h). RNase T1 “protected” proteins were size separated by polyacrylamide gel electrophoresis. Representative experiment shown using riboprobe #2 (922 to 2076 nt). Lane 1: negative control lane without cytoplasmic extract (cyto extract), Lane 2: negative control lane with cytoplasmic extract from untreated RISM and proteinase K, Lane 3: 20 g cytoplasmic extract from untreated RISM, Lane 4: 80 g cytoplasmic extract from untreated RISM, Lane 5: 20 g cytoplasmic extract from IGF-I treated RISM, Lane 6: 80 g cytoplasmic extract from IGF-I treated RISM, Lane 7: 20 g cytoplasmic extract from cycloheximide treated RISM, Lane 8: 80 g cytoplasmic extract from cycloheximide treated RISM.
426
Vol. 275, No. 2, 2000
BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS
plasm, and that an intermediate protein(s) was involved in regulating IGFBP-5 mRNA stability. The effect of cycloheximide on IGFBP-5 mRNA and protein abundance has been studied in other systems. In aortic smooth muscle cells, cycloheximide abrogated the IGFI-induced increase in IGFBP-5 mRNA but not basal IGFBP-5 mRNA (7). In human fibroblasts, cycloheximide (25 g/ml) was shown to completely inhibit the secretion of all IGFBPs; the effect of cycloheximide on RNA abundance was not studied (18). Many mRNA binding proteins and their cognate binding sites that stabilize or destabilize mRNAs have been identified and characterized (9, 19). One important set of binding sites are the adenosine and uridinerich elements (AUREs) that residue in 3⬘ UTRs of many short-lived mRNAs coding for growth factor, transcription activators, proto-oncogenes, cytokines, and neuropeptides (8, 9). The observation that many unstable mRNAs contain AUREs implies that these sites function as mRNA destabilizing elements. A family of AURE-binding proteins (AUBPs) has been described (8). These protein factors bind to AUREs with different affinities and can either protect mRNA from degradation or recruit RNases and enhance decay. In the present study, a region of the 3⬘ UTR (922 to 2076 nt) of IGFBP-5 mRNA was identified that bound a cycloheximide-sensitive protein activity. This region contains AURE #1. However, it cannot be determined from our studies whether RNA binding proteins bind to the AURE #1, or if the RNA binding proteins were responsible for cycloheximide-induced IGFBP-5 mRNA degradation; future studies will clarify this issue. Interestingly, IGF-I decreased the rate of cycloheximideinduced IGFBP-5 mRNA degradation (Fig. 3), suggesting that IGF-I may regulate the RNA stabilizing protein factors. Our hypothesis is that, after IGFBP-5 gene is transcribed, its nuclear transcript is spliced and exported into cytoplasm where one or more RNA binding proteins attaches to the mRNA. Inhibition of the translation of these RNA stabilizing factors by protein synthesis inhibitors leads to rapid IGFBP-5 mRNA degradation. Since IGF-I is a potent growth factor, it could stimulate the production of the RNA binding protein(s) and therefore, more RNA binding proteins would be available to stabilize IGFBP-5 mRNA. This may be an important mechanism in determining IGFBP-5 abundance and therefore, may be important in determining the biological effects of IGF-I in the intestine and other tissues. REFERENCES 1. Shimasaki, S., Shimonaka, M., Zhang, H-P., and Ling, N. (1991) Identification of five different insulin-like growth factor binding proteins (IGFBPs) from adult rat serum and molecular cloning of a novel IGFBP-5 in rat and human. J. Biol. Chem. 266, 10646 – 10653.
2. Clemmons, D. R. (1997) Insulin-like growth factor binding proteins and their role in controlling IGF actions. Cytokine Growth Factor Rev. 8, 45– 62. 3. Zimmermann, E. M., Li, L., Hou, Y. T., Cannon, M., Christman, G. M., Lund, P. K., and Bitar, K. N. (1997) Insulin-like growth factor I induces collagen and insulin-like growth factor binding protein 5 synthesis in cultured rat intestinal smooth muscle. Am. J. Physiol. 273(36), G875–G882. 4. Nam, T. J., Busby, W., Jr., and Clemmons, D. R. (1997) Insulinlike growth factor binding protein-5 binds to plasminogen activator inhibitor-I. Endocrinology 138, 2972–2978. 5. Zheng, B., Duan, C., and Clemmons, D. R. (1998) The effect of extracellular matrix proteins on porcine smooth muscle cell insulin-like growth factor (IGF) binding protein-5 synthesis and responsiveness to IGF-I. J. Biol. Chem. 273, 8994 –9000. 6. Yu, D., and Canalis, E. (1995) Insulin-like growth factor (IGF) I and retinoic acid induce the synthesis of IGF-binding protein 5 in rat osteoblastic cells. Endocrinology 136, 2000 –2006. 7. Duan, C., Hawes, S. B., Prevette, T., and Clemmons, D. R. (1996) Insulin-like growth factor-I (IGF-I) regulated IGF-binding protein-5 synthesis through transcriptional activation of the gene in aortic smooth muscle cells. J. Biol. Chem. 271, 4280 – 4288. 8. Chen, C-Y. A., and Shyu, A-B. (1995) AU-rich elements: Characterization and importance in mRNA degradation. Trends Biochem. Sci. 20, 465– 470. 9. Ross, J. (1995) mRNA stability in mammalian cells. Microbiol. Rev. 59, 423– 450. 10. James, P. L., Jones, S. B., Busby, W. H., Clemmons, D. R., and Rotwein, P. (1993) A highly conserved insulin-like growth factorbinding protein (IGFBP-5) is expressed during myoblast differentiation. J. Biol. Chem. 268, 22305–22312. 11. Graham, M. F., Diegelmann, R. F., Elson, C. O., Bitar, K. N., and Ehrlich, H. P. (1984) Isolation and culture of human intestinal smooth muscle cells. Proc. Soc. Exp. Biol. Med. 176, 503–506. 12. Chomczynski, P., and Sacchi, N. (1987) Single step method of RNA isolation by acid guanidinium thiocipunate-phenolchloroform extraction. Anal. Biochem. 162, 156 –159. 13. Zhu, X., Ling, N., and Shimasaki, S. (1993) Cloning the rat insulin-like growth factor binding protein-5 gene and DNA sequence analysis of its promoter region. Biochem. Biophysical. Res. Commun. 190, 1045–1052. 14. Hayden, J. M., Strong, D. D., Baylink, D. J., Powell, D. R., Sampath, T. K., and Mohan, S. (1997) Osteogeneic protein-1 stimulates production of insulin-like growth factor binding protein-3 nuclear transcripts in human osteosarcoma cells. Endocrinology 138, 4240 – 4247. 15. Canalis, E., and Gabbitas, B. (1995) Skeletal growth factors regulate the synthesis of insulin-like growth factor binding protein-5 in bone cell cultures. J. Biol. Chem. 270, 10771–10776. 16. Malter, J. S. (1989) Identification of an AUUUA-specific messenger RNA binding protein. Science 246, 664 – 666. 17. Rondon, I. J., MacMillan, L. A., Beckman, B. S., Goldberg, M. A., Schneider, T., Bunn, H. F., and Malter, J. S. (1991) Hypoxia up-regulates the activity of a novel erythropoietin mRNA binding protein. J. Biol. Chem. 266, 16594 –16598. 18. Camacho-Hubner, C., Busby, W. H., Jr., McCusker, R. H., Wright, G., and Clemmons, D. R. (1992) Identification of the forms of insulin-like growth factor binding proteins produced by human fibroblasts and the mechanisms that regulate their secretion. J. Biol. Chem. 267, 11949 –11956. 19. Caponicro, G., and Parker, R. (1996) Mechanisms and Control of mRNA turnover in Saccharomyces cerevisae. Microbiol. Rev. 60, 23–249.
427