Characterization of a Novel Positive Transcription Regulatory Element That Differentially Regulates the Insulin-like Growth Factor Binding Protein-3 (IGFBP-3) Gene in Senescent Cells*

Insulin-like growth factor binding protein-3 (IGFBP-3) is a well documented growth inhibitor and pro-apoptotic factor. IGFBP-3 mRNA and its protein are overexpressed by senescent human diploid fibroblasts. However, the mechanism responsible for the up-regulation of its expression is still unclear. This report describes a novel transcriptional regulatory element, IGFBP-3 enhancer element (IEE), identified in the 5′ untranslated region of the IGFBP-3 gene. This element differentially activates IGFBP-3 expression in senescent versus young fibroblasts. Electrophoretic mobility shift assays revealed abundant complexes in senescent cell nuclear extracts compared with young cell nuclear extracts. Similar to young proliferative cells, young quiescent cells showed reduced binding activity; enhancement of this activity was specific to senescent cells and not an effect of cell cycle arrest. The DNase I footprint revealed the protein-binding core sequence within the IEE through which the protein binds the IEE. Site-directed mutagenesis within IEE abolished binding activity and selectively decreased IGFBP-3 promoter activity in senescent (but not young) cells. Furthermore, introduction of an IEE decoy suppressed the endogenous IGFBP-3 gene expression specifically in senescent cells. These results point to the IEE as being a positive transcription regulatory element that contributes to the up-regulation of IGFBP-3 during cellular senescence.

Insulin-like growth factor binding protein-3 (IG-FBP-3) is a well documented growth inhibitor and proapoptotic factor. IGFBP-3 mRNA and its protein are overexpressed by senescent human diploid fibroblasts. However, the mechanism responsible for the up-regulation of its expression is still unclear. This report describes a novel transcriptional regulatory element, IG-FBP-3 enhancer element (IEE), identified in the 5 untranslated region of the IGFBP-3 gene. This element differentially activates IGFBP-3 expression in senescent versus young fibroblasts. Electrophoretic mobility shift assays revealed abundant complexes in senescent cell nuclear extracts compared with young cell nuclear extracts. Similar to young proliferative cells, young quiescent cells showed reduced binding activity; enhancement of this activity was specific to senescent cells and not an effect of cell cycle arrest. The DNase I footprint revealed the protein-binding core sequence within the IEE through which the protein binds the IEE. Site-directed mutagenesis within IEE abolished binding activity and selectively decreased IGFBP-3 promoter activity in senescent (but not young) cells. Furthermore, introduction of an IEE decoy suppressed the endogenous IGFBP-3 gene expression specifically in senescent cells. These results point to the IEE as being a positive transcription regulatory element that contributes to the upregulation of IGFBP-3 during cellular senescence.
Insulin-like growth factor binding protein-3 (IGFBP-3), 1 a 40 -50-kDa glycoprotein, is by far the most abundant IGFBP in the circulation. More than 90% of the circulating IGFs are bound to IGFBP-3 and the acid-labile subunit to form a 150-kDa complex (1). Circulating IGFBP-3, produced primarily by hepatic endothelial and Kupffer cells (2,3), modulates the amount of bioavailable free IGF and inhibits its transfer from circulation to tissue sites of action (4). At the cellular level, IGFBP-3 inhibits IGF action by competitively binding IGFs and thereby preventing their binding to the IGF receptor I (IGF-IR). This action has been demonstrated by numerous experiments in various cell types using destripeptide-(1,3)-IGF-I, an IGF-I analog that binds IGF-IR and stimulates DNA synthesis but cannot bind IGFBP-3 (5-7). By preventing IGFs from stimulating the IGF-IR, IGFBP-3 inhibits the IGF signaling pathway. However, the effects of IGFBP-3 on IGF-dependent cellular functions are complex, with both stimulatory and inhibitory actions reported, even within the one cell type (8).
Evidence from several studies has shown that, in addition to its ability to regulate access of IGF to its receptor, IGFBP-3 may have important IGF-independent antiproliferative effects. For example, transfection of human IGFBP-3 cDNA inhibits growth in various cell types, including human breast cancer cells (9) and murine 3T3 fibroblasts (10). In addition, Valentinis et al. (11) showed that transfection of human IGFBP-3 cDNA inhibits proliferation of fibroblasts with a targeted disruption of the IGF-IR. Further evidence has come from studies showing that IGFBP-3 gene expression is induced by potent growth-inhibitory proteins and agents, including tumor suppressor protein p53 (12), transforming growth factor (TGF)-␤ (13,14), retinoic acid (15), tumor necrosis factor ␣ (16,17), and anti-estrogen (18), which suggests that these agents may mediate their inhibitory effects through IGFBP-3. More recently, it has been suggested that the growth-inhibitory effects of IGFBP-3 may be mediated via an induction of apoptosis. Indirect evidence has come from reports that an increase in IG-FBP-3 expression is associated with the induction of apoptosis (19,20). More directly, Nickerson et al. (21) demonstrated that the addition of recombinant IGFBP-3 induces apoptosis in breast carcinoma cells. This effect was abrogated in the presence of an IGF-I analog that binds IGF-IR (but not IGFBP), which suggests that, in this system, IGFBP-3 induces apoptosis indirectly by reducing the bioavailability of ligands for IGF-IR. However, direct pro-apoptotic effects of IGFBP-3 have been reported in cells lacking IGF-IR (22) and under conditions where IGF-I could not elicit a survival effect (23), pointing to the existence of an IGF-independent mode of IGFBP-3 action.
Compared with young cells, senescent fibroblasts show substantially up-regulated IGFBP-3 mRNA and protein (24 -26), and the protein also accumulates in the growth medium of young cells to levels directly correlated with the chronological age of the donor (25)(26)(27). Furthermore, Grigoriev et al. (28) found that IGFBP-3 accumulation in the medium of senescent human diploid fibroblasts can bind and sequester IGFs and thereby attenuate the response of senescent cells to IGF-I. Whether IGFBP-3 exerts its influence on senescent cells through an IGF-independent pathway is still unclear. In-creased IGFBP-3 expression appears to suppress growth (but not induce) apoptosis in senescent fibroblasts, because senescent fibroblasts are more resistant to apoptotic stimuli than young cells (29).
Buckbinder et al. (12) first identified IGFBP-3 as one of the p53-inducible genes by employing EB1 colon carcinoma cells carrying an inducible wild-type p53 transgene and saos-2-D4H cells containing a temperature-sensitive p53 mutant. Furthermore, ultraviolet radiation and doxorubicin induced IG-FBP-3 expression in fibroblasts containing the wild-type (but not mutant) p53. As well, on the basis of homology to the p53-binding consensus sequence (30), two p53-binding sites within the IGFBP-3 gene were identified and confirmed by electrophoretic mobility shift analyses (EMSAs) (12). Thus, IGFBP-3 expression is positively regulated by p53. However, IGFBP-3 expression could also be induced by a p53independent pathway (31).
During cellular senescence, p53 activity increases greatly (32). Thus, we wondered whether the increased p53 activity contributes to overexpression of IGFBP-3 in senescent cells. However, treating senescent cells with the HPV-E6-expressing vector or another p53 inhibitor, PFT-␣, did not reduce IGFBP-3 expression. This report describes our subsequent examination of potential 5Ј regulatory regions and mechanisms that may be responsible for the up-regulation of IGFBP-3 in senescent human diploid fibroblasts. Analysis of the IGFBP-3 promoter via transient transfection of nested promoter deletions into young and old fibroblasts has identified an enhancer element. This element is located in the 5Ј-UTR of IGFBP-3 and binds a protein of 27 kDa more avidly in senescent cells than in young cells, which suggests that enhanced binding to this element in senescent cells contributes to the increased expression of IG-FBP-3 during senescence.

MATERIALS AND METHODS
Cells and Cell Culture-Human embryonic lung diploid fibroblast 2BS cells (obtained from the National Institute of Biological Products, Beijing, China) were previously isolated from female fetal lung fibroblast tissue and have been fully characterized (33). The current expected life span is ϳ70 population doublings (PD). 2BS cells were considered to be young at PD30 or below and fully senescent at PD55 or above. The cells were maintained in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% (v/v) fetal bovine serum and 1% penicillin/streptomycin. Senescent cells used in this study did not reach confluence within 2 weeks after subculturing at a 1:2 split ratio, demonstrated a characteristic large flattened senescent cell morphology, and were positive for senescenceassociated ␤-galactosidase activity.
Northern Blot Analysis-Total RNA was isolated from cells using an RNeasy total RNA kit (Qiagen, Valencia, CA). DNA fragments from IGFBP-3 cDNA and GAPDH cDNA were labeled with [␣-32 P]dCTP with use of a random priming kit (Promega, Madison, WI). A quantity of 30 g of total RNA from each sample was electrophoresed on 1% agrose/ 2.2 M formaldehyde gels. The RNA was then transferred to nylon membranes and cross-linked by UV light exposure. Blots were hybridized using Expresshyb TM hybridization solution (Clontech, Palo Alto, CA) and autoradiographed.
Deletion Construct of 5Ј-Flanking Region of IGFBP-3-The 2-kb human IGFBP-3 promoter containing the 5Ј-UTR was generated by PCR with plaque-forming unit polymerase. The upper primer was 5Ј-GCT AGC ACC TGG GAC CTC AAG AAT TGC ATT T-3Ј (the underlined sequence is the NheI recognition site), and the lower primer was 5Ј-AGA TCT CGG AGC AGC ACC AGC AGA GTC AG-3Ј (the underlined sequence is the BglII recognition site). The PCR fragment was ligated into pGEM-T Easy Vector (Promega). The resulting plasmid was digested with NheI/BglII to release the 2-kb promoter. Thereafter, the fragment was subcloned into the NheI/BglII site of the pGL3 basic vector (designated pIGFBP-3). For 5Ј-nested deletion, the restriction endonuclease SacII at position Ϫ1303, PmacI at Ϫ778, XhoI at Ϫ305, SmaI at Ϫ141 in the promoter sequence, and NheI in the multicloning site of the pGL3 vector were chosen. pIGFBP-3 was digested with SacII/NheI, PmacI/NheI, XhoI/NheI, and SmaI/NheI. The fragments containing the vector sequence were end-blunted with T4 DNA poly-merase and self-ligated with T4 DNA ligase. The deleted constructs were analyzed by agarose gel electrophoresis and transformed to obtain clones that were confirmed by sequencing. These clones were designated as Ϫ2031 (full-length), Ϫ1303, Ϫ778, Ϫ305, and Ϫ141 relative to the translational start site.
For 3Ј deletion of the 5Ј-UTR, the Ϫ305 clone in the pGL3 vector was cut with HindIII, filled in with thionucleotides, and redigested with BglII at the 3Ј end. After exonuclease III/S1 nuclease treatment, the deleted clones were ligated, screened, and sequenced. The clones with the 3Ј-end deletion were designated as Ϫ58 to ϩ59 and Ϫ192 to ϩ59.
Transient Transfection-Plasmid DNA was introduced into young and senescent cells with use of Effectene transfection reagent (Qiagen). To standardize the transfection efficiency, 0.1 g of pRL-CMV vector (pRL Renilla reniformis luciferase control reporter vector; Promega) was cotransfected in all experiments. Empty pGL3-Basic vector was used as a negative control. At 36 -48 h after transfection, cells were harvested and lysed in 200 l of reporter lysis buffer (Promega). A luciferase assay was performed using a dual luciferase assay kit (Promega), and activity was measured with an Optocomp luminometer (MGM Instruments, Inc., Hamden, CT). Promoter activity of each construct is represented by relative light output normalized to pRL-CMV control.
Electrophoretic Mobility Shift Assays-Nuclear extracts from young and senescent 2BS cells were prepared as described previously (34). The oligonucleotides were synthesized, and double-stranded oligonucleotides were generated by annealing equimolar complementary oligonucleotides in 50 mM Tris-HCl (pH 7.4), 1 mM EDTA, 100 mM NaCl, and 13 mM MgCl 2 as follows: 88°C for 2 min, 65°C for 10 min, 37°C for 10 min, and 25°C for 5 min. The double-stranded oligonucleotides were endlabeled with [␥-32 P]ATP (3000 mCi/mmol; FuRui Co., Beijing, China) and T4 polynucleotide kinase, and the labeled probes were subsequently purified using Sephadex G-50. Binding assays involved the use of a mixture containing 32 P-labeled oligonucleotide (0.3 ng), 5 g of nuclear protein, 1 g of poly(dI-dC) (Amersham Biosciences) adjusted to 20 l with binding buffer (20 mM HEPES, pH 7.9, 1.5 mM MgCl 2, 0.2 mM EDTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, 7.5% glycerol). Binding reactions were carried out for 30 min at room temperature. A 15-l aliquot of each reaction was loaded onto a 5% nondenaturing polyacrylamide gel and run in 0.5ϫ TBE buffer at 120 V. Following electrophoresis, the gels were dried and autoradiographed. In the competition assays, unlabeled competitor oligonucleotides were added at 100-fold excess before the addition of the 32 P-labeled probe.
DNase I Footprinting-Nuclear extracts were prepared as described previously (34), DNase I footprinting experiments were performed with a fragment corresponding to Ϫ63 to ϩ59 of the IGFBP-3 gene transcription starting site. The fragment was labeled with [␣-32 P]dATP and incubated with nuclear extracts at room temperature for 30 min in a reaction mixture containing 4% glycerol, 10 mM Tris-HCl, pH 7.5, 50 mM NaCl, 2.5 mM MgCl 2 , 0.5 mM dithiothreitol, and 2 g of poly(dI: dC). After incubation, DNase I was added and incubation was continued for exactly 1 min. Footprinting was done using a Core footprinting system kit from Promega, according to the manufacturer's instructions. The samples were analyzed on a 6% acrylamide/7 M urea denaturing gel.
Site-directed Mutagenesis-Bases were mutated using the Quik-Change site-directed mutagenesis kit (Stratagene, La Jolla, CA). A mutagenic primer (45mer) was synthesized and annealed to the doublestranded Ϫ305 construct vector DNA. Plaque-forming unit DNA polymerase was used to synthesize the mutagenic promoter, followed by digestion of the parental plasmid by DpnI according to the manufacturer's instructions. The mutated plasmid was transformed into XL1-Blue competent cells, and the resulting plasmid was isolated and sequenced to confirm the mutations.
Decoy Approach-0 -14 g of IGFBP-3 enhancer element (IEE) decoy was introduced into senescent cells cultured with Dulbecco's modified Eagle's medium with 10% fetal calf serum in a 10-cm diameter dish by the lipofection method. At 24 h after the administration of DNA solution, total RNA was extracted from the cells, and Northern blot analysis was performed. For control reasons, the mutant IEE oligonucleotides (shown in Fig. 10A) were used.
UV Cross-linking of DNA-Protein Complexes-To estimate the relative molecular mass of the DNA-binding proteins, binding reactions with oligonucleotide 5Ј-UTR-2 were performed. Twice the amount of the labeled probe (4 ϫ 10 4 counts/min) and nuclear extracts (10 g) were used. After 30 min of incubation at room temperature, the binding reaction was subjected to UV light, and unprotected DNA was digested with DNase I. The samples were irradiated by a 305-nm inverted UV transilluminator at 7 milliwatts/cm 2 for 5 min. The cross-linked reactions were electrophoresed through 10% SDS-polyacrylamide gels, dried, and visualized by autoradiograph.
Dual Luciferase Reporter Assays-Constructs containing the wildtype and mutant IEE were made using pGL3 promoter plasmids containing the SV40 promoter (Promega). A total of 30 bp of oligonucleotides was cloned into the MluI/BglII in the multicloning site region upstream of the SV40 promoter and the HindIII/NcoI between the SV40 promoter and the lucϩ reporter gene. HindIII/NcoI was chosen because of its proximity to the functional gene (i.e. the in vivo IEE is located in the 5Ј-UTR of the IGFBP-3 gene). MluI/BglII was chosen to determine whether a changed IEE location affects its transcriptional regulation. These constructs were then co-transfected into young and senescent 2BS cells with pRL-CMV as a transfection efficiency control. Luciferase activity was measured using an Optocomp luminometer and normalized on the basis of pRL-CMV expression.

Increased p53 Activity Does Not Contribute to IGFBP-3 Overexpression in Senescent 2BS
Cells-Northern blot analysis with IGFBP-3 cDNA as a probe confirmed whether IGFBP-3 expression is increased in senescent 2BS cells. As shown in Fig. 1A, IGFBP-3 expression increased substantially in senescent cells as compared with young cells (24 -26). Many studies have shown that the induction of the IGFBP-3 gene is associated with p53 activity. Coincidentally, p53 activity greatly increases in senescent cells. However, we wondered whether increased p53 activity caused IGFBP-3 overexpression. The expression vector expressing HPV-E6 protein, which is a well known p53 inhibitor, was used to examine whether or not p53 activity is associated with IGFBP-3 expression in senescent 2BS cells. For the positive control, we used a pp53-TA-Luc vector (Clontech) containing a p53 Cis-acting enhancer element upstream of the luciferase reporter gene, because p53 target genes have not been definitely documented in senescent cells. We co-transfected senescent cells with both vectors; after 48 h, luciferase activity was assayed. The result from Fig. 1B showed that luciferase activity was decreased in HPV-E6 vector-treated senescent cells in a dose-dependent manner, suggesting HPV-E6 vector is functional in antagonizing p53 function in this system. Subsequently, senescent cells were treated with the same method as above; RNA were isolated and analyzed by Northern blotting with IGFBP-3 cDNA as a probe and GAPDH as an internal control. As shown in Fig. 1C, no change in IGFBP-3 expression was observed in HPV-E6 vector-transfected senescent cells, suggesting that elevated p53 activity does not contribute to IGFBP-3 gene overexpression in senescent cells. Such a view was further supported by the results with another p53 inhibitor, PFT-␣ (as shown in Fig. 1, D and  E). Thus, we subsequently cloned the IGFBP-3 promoter and analyzed potential regulatory regions that may be responsible for the up-regulation of IGFBP-3 seen in the senescent 2BS cells.

Identification of the Functional Region in the 5Ј-UTR of IG-FBP-3 That Contributes to Overexpression of IGFBP-3 in Senescent Fibroblasts by Deletion
Analysis-To determine the basal activity of the IGFBP-3 promoter in young and senescent cells, a 2-kb IGFBP-3 promoter fused to the luciferase reporter gene ( Fig. 2A) was used in transient transfection studies. Various 5Ј-deleted pIGFBP-3 constructs were generated (Fig. 2B) and transiently co-transfected with a cytomegalovirus-driven Renilla luciferase vector (as an internal control) into young and senescent cells. After 36 h of incubation, the cells were harvested and analyzed for luciferase activity. As shown in Fig.  2B, luciferase activity caused by the full-length IGFBP-3 promoter was ϳ3-fold greater in senescent than in young cells. Deletion of segments from the 5Ј end, from Ϫ2031 to Ϫ778, slightly reduced the difference in expression between young and senescent cells. However, further deletion of the IGFBP-3 promoter, to 305 bp upstream of the ATG translation initiation site, did not reduce but rather increased the difference in expression by 4 -5-fold, which indicates that the region from Ϫ305 to ϩ59 strongly contributes to increased expression in senescent cells.
We then generated a 3Ј-nested deletion of the Ϫ305 to ϩ59 region to further define the region involved in the differential regulation of IGFBP-3. Fig. 3A shows that deletion of the region from ϩ59 to Ϫ58 substantially attenuated luciferase expression in senescent cells, but the luciferase activity in young cells remained relatively unchanged, thus reducing the difference in expression between senescent and young cells (Fig. 3B). Further deletion from ϩ59 to Ϫ192 completely abolished luciferase expression in both cells, probably because the transcription initiation site had been removed (Fig. 3A). Thus, a potential binding site for a transcriptional repressor in young cells and/or a transcriptional activator in senescent cells exists in the 100-bp region between Ϫ58 and ϩ59.
Identification of a 30-bp Enhancer Element with Overlapping Oligonucleotides-To define the region involved in protein binding, we used electrophoretic mobility gel shift assays (but not DNaseI) footprinting assays, because the former can semiquantitate protein-DNA binding. As shown in Fig. 4A, five doubled-stranded 30-bp oligonucleotides, designated 5Ј-UTR-1 to 5Ј-UTR-5, were synthesized and used in EMSA of young and senescent cell nuclear extracts. No detectable activity was observed with the 5Ј-UTR-1, -3, -4, and -5 oligonucleotides (Fig.  4B, lanes 6, 8, 14, 16,18,20,22,24). A specific complex was formed when 5Ј-UTR-2 was used as a probe with senescent cell extracts (Fig. 4B, lane 12), whereas the level was dramatically reduced with young cell extracts (Fig. 4B, lane 10). Specificity of the complex was confirmed by incubation with a 100-fold excess of unlabeled oligonucleotide, which competed with the labeled probe (Fig. 4B, lanes 11 and 13). Thus, the 30-bp region was involved in protein binding and termed IEE. To rule out the possibility that the nuclear extracts from young 2BS fibroblasts were deficient in DNA binding ability, we performed gel shift studies using the same extracts with an oligonucleotide containing the Sp1-binding site. As shown in Fig. 4C, Sp1-binding activity actually was slightly higher in young cell extracts (lane 2) than in senescent extracts (lane 4). As expected, 100-fold competition with unlabeled oligonucleotide bearing the same Sp1 sequence competed with the complex formed on the labeled Sp1 (Fig. 4C, lanes 3 and 5). Given previous reports that Sp1-binding activity is about equal in young and senescent cells (35), the specific binding activity of the complex shown in Fig. 4B would actually be reduced ϳ4fold in extracts from young as compared with senescent cells (Fig. 4D).
Increased expression of IGFBP-3 is also seen upon serum withdrawal or contact inhibition-induced cell cycle arrest (25). To test whether the increased binding activity was specific to senescence, we incubated the 30-bp element with extracts from young proliferating cells, young quiescent cells, and subconfluent senescent cells. As shown in Fig. 4E, increased binding was not seen in young quiescent cells (lane 6), which indicates that elevated binding depends on senescence rather than growth.
To help determine the specific transcription factor producing the band shift shown in Fig. 4B, we first performed a detailed computer analysis using the MatInspector program but did not find highly homologous consensus with known transcription factors. To further confirm the bases within the IEE responsible for protein binding, oligonucleotides of 5Ј-UTR-2 containing various mutations (m1-m15) were generated as shown in Fig.  5A. In each case, purines and pyrimidines were exchanged for noncomplementary pyrimidines and purines, respectively. Surprisingly, EMSA of senescent cell nuclear extracts with these oligonucleotides (Fig. 5B) revealed that these mutations greatly reduced (lanes 3 and 14) or completely abolished (other lanes) the binding activity, suggesting that these bases within the IEE were necessary for transcription factor binding.
DNase I Footprint Identified the Protein-binding Core Sequence within the IEE-That seemingly every base pair within the oligonucleotide is important for the protein-DNA complex is a surprising phenomenon. To define the protein-binding sequence within the IEE, in vitro DNase I footprinting was performed with nuclear extracts from young and senescent 2BS cells and single end-labeled probe (-63 to ϩ59) covering the IEE region. As seen in Fig. 6, two DNase I-protected regions, designated as sites A and B, were detected in this region. Coincidentally, the two sites are located in the 30-bp IEE region, with site A covering the sequence CTGCCA and site B covering the sequence GCGTGCCCCG. Regions A and B were separated only by four nucleotides. Because a single protein-DNA complex was detected in EMSA, the protein (or protein complex) binds the probe likely through sites A and B, two contacting sites within the DNA probe. Besides the bases within the regions A and B, other base mutations in the 30-bp oligonucleotide also affected protein-DNA interaction, perhaps reflecting stabilization of the complexes by these bases.
Mutation of the IEE Decreases IGFBP-3 Promoter Activity in Senescent Cells-Three sets of 2-bp bases (randomly selected within the IEE) changed in the Ϫ305 construct, and corresponding to the mutation in 5Ј-UTR-m6, -m8, and -m10 (Fig.  7A), were introduced by site-directed mutagenesis. We trans- fected the promoter Ϫ305 construct and the mutated constructs into young and senescent cells and measured the luciferase activity resulting from each construct after normalizing for pRL-CMV control. As shown in Fig. 7B, all three mutations within the IEE resulted in a nearly 3-fold decrease in luciferase activity in senescent cells as compared with the wild-type construct. In contrast, the luciferase activity from the mutant construct was slightly decreased as compared with that in the wild-type construct in young cells. Fig. 7C shows the results of the same experiment plotted as -fold difference in senescent FIG. 4. Identification of the IEE with overlapping oligonucleotides. A, a portion of the IGFBP-3 promoter 5Ј-UTR showing a 117-bp region divided into five overlapping 30-bp oligonucleotides (5Ј-UTR-1 to 5Ј-UTR-5). B, electrophoretic mobility shift assay of the overlapping oligonucleotides shown in A with young (Y) and senescent (S) cell nuclear extracts. Lanes 1-5 are control reactions in the absence of nuclear extracts. The formation of a complex was evident in senescent cell nuclear extracts (lane 12) and weaker in young cell nuclear extracts with 5Ј-UTR-2 used as a probe (lane 10), but the signal was not detected with 5Ј-UTR-1, -3, -4, and -5 used as probes (lanes 6, 8, 14, 16, 18, 20, 22, and 24). Unlabeled 5Ј-UTR-1, -2, -3, -4, and -5 oligonucleotides were also used as probes for a competition experiment at a 100-fold excess as indicated in lanes 7,9,11,13,15,17,19,21,23, and 25 to determine the specificity of binding. C, control gel shift with use of a consensus Sp1-binding site with the same young (Y) and senescent (S) nuclear extracts described in B. Lane 1 shows the probe alone in the absence of nuclear extracts. Lanes 2 and 4 show the formation of a complex with the addition of young senescent nuclear extracts. Lanes 3 and 5 show the reaction in the presence of 100-fold of a cold Sp1 oligonucleotide. D, the binding activity of senescent and young cells with 5Ј-UTR-2 in B was quantified as determined against the Sp1 control in C. E, electrophoretic mobility shift assay with independently isolated nuclear extracts from senescent subconfluent cells (S), young growing cells (Y), and young quiescent cells (Q), with 5Ј-UTR-2 used as a probe. Lane 1 shows 5Ј-UTR-2 in the absence of the nuclear extract, and lanes 2, 4, and 6 show the degree of binding activity of three nuclear extracts to 5Ј-UTR-2. Lanes 3, 5, and 7 show the reaction in the presence of 100-fold of a cold 5Ј-UTR-2 oligonucleotide. versus young cells. Mutations of the IEE resulted in a nearly 60% decrease in the difference in luciferase activity between mutated and unmutated controls. These results suggest that the IEE within the 5Ј-UTR of IGFBP-3 constitutes a binding site for a potential transcriptional activator in senescent cells, the activity and/or levels of which are reduced in young cells.
Inhibition of the IGFBP-3 Gene Expression by Introducing IEE Decoy into Senescent 2BS Cells-To demonstrate the feasibility of suppressing the endogenous IGFBP-3 gene expression by the decoy strategy, we examined the effect of the IEE decoy on the IGFBP-3 mRNA expression in senescent cells. Either 0 -14 g of IEE decoy oligonucleotides or control mutant IEE oligonucleotides (shown in Fig. 10A) was introduced into senescent cells cultured in 10-cm diameter dishes by the lipofection method. When IEE decoy was introduced, the IGFBP-3 mRNA level was decreased in a dose-dependent manner compared with the level in cells transfected with a control (mutant) IEE ( Fig. 8A and B); on the other hand, GAPDH mRNA expression in IEE decoy-introduced senescent cells and mutated IEEtransfected ones revealed no difference, suggesting that the introduction of the IEE decoy suppressed the endogenous IG-FBP-3 gene expression specifically.
Estimation of the Molecular Weight of the Protein Binding IEE-To estimate the molecular mass of any protein(s) bound specifically to IEE, nuclear extracts from young and senescent cells were incubated with the 30-bp oligonucleotide (5Ј-UTR-2) and subjected to UV cross-linking and then underwent SDS-PAGE. To demonstrate the sequence-specific nature of the protein, we added a 500-fold excess of cold competitors to the binding reaction mixture. As shown in Fig. 9, a complex was detected in both kinds of cells and was slightly more apparent in senescent than in young cells. Because the size of the complex was 45 kDa, the molecular mass of the protein is expected to be ϳ27 kDa.
The IEE Preferentially Activates Gene Expression in Senescent Fibroblasts-To test the activity of the 30-bp wild-type and mutant IEE, both forms were ligated into the HindIII/NcoI site between the luciferase reporter gene and the SV40 promoter and the MluI/BglII multicloning site upstream of the SV40 promoter in the pGL3 promoter vector. These recombinant plasmids were transfected into young and senescent 2BS fibroblasts. Transfection efficiencies were normalized on the basis of Renilla luciferase (pRL-CMV) expression. As shown in Fig.  10B, the level of reporter was increased by Ͼ6.5-fold in senescent cells, with the wild-type IEE located in its natural site (HindIII/NcoI between the luciferase reporter gene and the SV40 promoter (but only slightly in young cells). No change was found with the ligation of the mutant IEE form in senescent or young cells. Therefore, although the IEE enhanced transcription in both young and senescent cells, the effect was markedly greater in senescent cells. The same phenomenon was observed in cells with the IEE cloned into its non-natural site (MluI/BglII) (Fig. 10B), although the transcription effect was weaker than with cloning into the natural site.

DISCUSSION
In this study, we cloned the IGFBP-3 promoter. Deletion analysis showed that the 117-bp region from Ϫ58 to ϩ59 strongly contributes to increased IGFBP-3 gene expression in senescent cells. Use of overlapping oligonucleotides corresponding to this region narrowed the protein-binding region further to a 30-bp sequence in the 5Ј-UTR of IGFBP-3, which we have termed the IEE (from Ϫ37 to Ϫ8). Further EMSA with young quiescent and senescent nuclear extracts showed that the elevation of the protein-binding activity depends on senescence. Although mutation experiments showed that every base within the IEE appears to be important for forming protein-DNA complexes, the DNaseI footprint revealed the proteinbinding core sequence within the IEE, through which the pro- tein bind the IEE. Site-directed mutagenesis within IEE may markedly reduce IGFBP-3 promoter activity in senescent (but not young) cells; furthermore, introduction of the IEE decoy suppressed the endogenous IGFBP-3 gene expression specifically in senescent cells, indicating that the IEE is a positive transcription regulatory element. Luciferase reporter con- structs further demonstrated the ability of this element to promote gene transcription in senescent cells. Finally, results of UV cross-linking assays suggested that the molecular mass of protein binding the IEE was ϳ27 kDa.
It has been well documented that the IGFBP-3 gene is activated by p53. However, we showed that increased p53 activity during cellular senescence was not required for induction of IGFBP-3, which indicates a p53-independent regulation of IG-FBP-3 expression. A similar phenomenon was observed for the resistance of senescent cells to apoptosis. Although p53 is an important apoptotic factor in other cells, increased p53 activity does not sensitize senescent fibroblasts to apoptotic stimuli. In fact, senescent fibroblast cells are refractory to apoptosis (29). This difference in function probably results from the differential post-translational modification of p53 in senescent cells. It is well known that p53 is activated mainly through chemical modification, such as phosphorylation, acetylation, and ADPribosylation. A detailed analysis of p53 phosphorylation in senescent fibroblasts showed that senescence is associated with a major change at a putative regulatory site in the N and C termini of p53, with increased phosphorylation at Ser-15, Thr-18, and Ser-376 and decreased phosphorylation at Ser-392. Ionizing and UV radiation generated an overlapping but distinct profile of response, with increased serine-15 phosphorylation being the only common change (36). This observation suggests that the difference in modification likely makes p53 function slightly differently in senescent than in other cells. Fig. 1A shows that IGFBP-3 expression in senescent cells was Ͼ20-fold higher than that in young cells, but subsequent luciferase assay (Fig. 2B) showed only an ϳ4 -5-fold difference FIG. 9. Determination of the approximate molecular mass of protein(s) binding to the IEE. Binding reactions of the indicated probes with young (Y) and senescent (S) cell nuclear extracts were subjected to UV cross-linking and analyzed by SDS-PAGE. The lanes without UV cross-linking served as negative controls (lanes 2 and 5). The arrow indicates the DNA-protein complex.
FIG. 10. Dual luciferase reporter assays of the wild-type and mutant IEE sequence. A, a mutant IEE (mut-IEE) with a 6-bp change in the IEE region was synthesized and cloned into the pGL3 promoter vector. B, IEE (wild-type) or mutant IEE sequences were ligated into the HindIII/NcoI site between the luciferase (LUC) reporter gene and SV40 promoter and MluI/BglII multicloning site upstream of the SV40 promoter in the pGL3 promoter firefly luciferase reporter plasmids in the sense orientation (left graph). The results in senescent (black bars) and young (white bars) cells were normalized to pRL-CMV activity and represent the mean Ϯ S.D. for three independent experiments (right graph).
in expression. This observation indicates that other elements exist and contribute to the differential expression of IGFBP-3 in young and senescent cells.
Although the expression of a number of genes changes during senescence (37,38), the mechanism by which senescence dependently regulates this expression has not been extensively studied. However, a few reports have suggested that regulation of some gene expression in senescent cells could differ from that in other cells. For example, Wang et al. (39) showed that a novel negative regulatory element at Ϫ495 to Ϫ485 bp of the p16 INK4a promoter contributes to overexpression of p16 in senescent fibroblasts. As well, Berardi et al. (40) reported that a novel transcriptional inhibitory element in the 5Ј-UTR of cyclin D1 was related to the up-regulation of cyclin D1 in senescent fibroblasts. In this paper, another novel transcription enhancer element in the 5Ј-UTR of IGFBP-3 was identified and shown to contribute to the increased IGFBP-3 expression in senescent cells. Thus, inactivation or activation of some unidentified transcription factors during cellular senescence likely contributes to differential regulation of some gene expression.
Although the regulation of gene expression via the 5Ј-UTR has not been widely reported, such a mechanism appears to repress transcription. For example, p53 suppresses the expression of bcl-2, at least in part through a p53 response element located in the 5Ј-UTR of the bcl-2 gene (41). Similarly, a suppressor element has been identified in the 5Ј-UTR of the androgen receptor gene (42). Recently, a transcription inhibitory element in the 5Ј-UTR of cyclin D1 has been found to suppress the cyclin D1 gene in young fibroblasts (40). In this report, however, a positive transcription regulatory element was identified in the 5Ј-UTR of IGFBP-3 in senescent fibroblasts. Thus, the regulation of gene expression via the 5Ј-UTR may also activate gene transcription.
These data indicate that a novel transcription enhancer contributes to selectively increasing the expression of IGFBP-3 during cellular senescence. Use of the IEE to search transcription factor databases revealed no homologies to known human transcription factor binding sites, which suggests that the protein(s) binding the IEE are novel. Experiments to identify the putative enhancer protein by a one-hybrid assay are under way and should reveal important insights regarding its role(s) in cellular aging.