The 3′-Untranslated Region of p21WAF1 mRNA Is a Composite cis-Acting Sequence Bound by RNA-binding Proteins from Breast Cancer Cells, Including HuR and Poly(C)-binding Protein*

Despite promoting growth in many cell types, epidermal growth factor (EGF) induces growth inhibition in a variety of cancer cells that overexpress its receptor. The cyclin-dependent kinase inhibitor p21WAF1 is a central component of this pathway. We found in human MDA-468 breast cancer cells that EGF up-regulates p21WAF1 mRNA and protein, through a combination of increased mRNA stability and transcription. The decay rate of a hybrid luciferase reporter full-length p21WAF13′-untranslated region (UTR) mRNA was significantly faster than that of a control mRNA. Transfections with a variety of p21WAF1 3′-UTR constructs identified multiplecis-acting elements capable of reducing basal reporter activity. Short wavelength ultraviolet light induced reporter activity in constructs containing the 5′ region of the p21WAF1 3′-UTR, whereas EGF induced reporter activity in constructs containing sequences 3′ of the UVC-responsive region. Thesecis-elements bound multiple proteins from MDA-468 cells, including HuR and poly(C)-binding protein 1 (CP1). Immunoprecipitation studies confirmed that HuR and CP1 associate with p21WAF1mRNA in MDA-468 cells. Over- and underexpression of HuR in MDA-468 cells did not affect EGF-induced p21WAF1 protein expression or growth inhibition. However, binding of HuR to its target 3′-UTRcis-element was regulated by UVC but not by EGF, suggesting that these stimuli modulate the stability of p21WAF1mRNA via different mechanisms. We conclude that EGF-induced p21WAF1 protein expression is mediated largely by stabilization of p21WAF1 mRNA elicited via multiple 3′-UTR cis-elements. Although HuR binds at least one of these elements, it does not appear to be a major modulator of p21WAF1 expression or growth inhibition in this system. CP1 is a novel p21WAF1 mRNA-binding protein that may function cooperatively with other mRNA-binding proteins to regulate p21WAF1 mRNA stability.

Despite promoting growth in many cell types, epidermal growth factor (EGF) induces growth inhibition in a variety of cancer cells that overexpress its receptor. The cyclin-dependent kinase inhibitor p21 WAF1 is a central component of this pathway. We found in human MDA-468 breast cancer cells that EGF up-regulates p21 WAF1 mRNA and protein, through a combination of increased mRNA stability and transcription. The decay rate of a hybrid luciferase reporter full-length p21 WAF1 3-untranslated region (UTR) mRNA was significantly faster than that of a control mRNA. Transfections with a variety of p21 WAF1 3-UTR constructs identified multiple cisacting elements capable of reducing basal reporter activity. Short wavelength ultraviolet light induced reporter activity in constructs containing the 5 region of the p21 WAF1 3-UTR, whereas EGF induced reporter activity in constructs containing sequences 3 of the UVC-responsive region. These cis-elements bound multiple proteins from MDA-468 cells, including HuR and poly(C)-binding protein 1 (CP1). Immunoprecipitation studies confirmed that HuR and CP1 associate with p21 WAF1 mRNA in MDA-468 cells. Over-and underexpression of HuR in MDA-468 cells did not affect EGFinduced p21 WAF1 protein expression or growth inhibition. However, binding of HuR to its target 3-UTR ciselement was regulated by UVC but not by EGF, suggesting that these stimuli modulate the stability of p21 WAF1 mRNA via different mechanisms. We conclude that EGF-induced p21 WAF1 protein expression is mediated largely by stabilization of p21 WAF1 mRNA elicited via multiple 3-UTR cis-elements. Although HuR binds at least one of these elements, it does not appear to be a major modulator of p21 WAF1 expression or growth inhibition in this system. CP1 is a novel p21 WAF1 mRNA-binding protein that may function cooperatively with other mRNA-binding proteins to regulate p21 WAF1 mRNA stability.
Inhibition of human tumor cell growth is mediated by a variety of cell cycle-related proteins and tumor suppressors. p53, a well characterized tumor suppressor, activates transcription of a number of target genes, including p21 WAF1 (wildtype p53 activated fragment-1) (1, 2), which encodes a protein of M r 21,000 (p21), also known as cyclin-dependent kinaseinteracting protein 1. p21 WAF1 inhibits cyclin-cyclin-dependent kinase activity, preventing phosphorylation of critical cyclindependent kinase substrates, blocking transition from G 1 to S phase of the cell cycle (3), as well as inducing apoptosis (4). Recent evidence suggests that factors other than p53, such as EGF 1 (16), can induce p21 WAF1 expression in various cell types (p53-independent pathways). Because most human tumors lack p53 function (5), investigation of the mechanisms that regulate p21 WAF1 expression through alternative growth factor-induced pathways has become an important focus in cancer research. In particular, a major goal is to devise approaches that would increase expression of p21 WAF1 in tumors to reduce proliferation and tumor growth.
Although EGF is typically growth-proliferative in breast cancer cells (6), some cancer cells are growth-inhibited by EGF (e.g. MDA-468 breast (7,8), A431 epidermoid (9,10)). EGF-induced growth inhibition of these cells is associated with EGF receptor (EGFR) overexpression (11) and appears to be mediated by induction of p21 WAF1 mRNA and protein (8). Multiple reports show conclusively that the regulation of p21 WAF1 expression by growth factors and other ligands occurs predominantly at the level of mRNA stability (12)(13)(14)(15)(16)(17). However, there is little understanding of the specific RNA-protein interactions involved in this process, particularly in breast cancer cells. Thus, the EGFinduced up-regulation of p21 WAF1 mRNA provides an ideal system to investigate the mechanisms governing p21 WAF1 mRNA decay.
The regulation of mRNA decay is a critical mechanism in the control of gene expression (reviewed in Hollams et al. (18)) that involves interactions between cis-acting sequences that confer instability to mRNA and the trans-acting protein factors that bind them. Many cis-acting sequences consist of AU-rich ele-ments (AREs), most often located in the 3Ј-untranslated region (3Ј-UTR) of labile mRNAs. However, cis-acting elements are also found within the coding regions and 5Ј-UTRs of various mRNAs (e.g. c-fos, c-myc) (19). AREs often contain single or multiple repeats of pentamer (AUUUA) sequences, and inclusion of the AUUUA pentamer motif often targets the mRNA for rapid cytoplasmic degradation (20). One well characterized cis-acting element is the AU-rich sequence found within the 3Ј-UTR of granulocyte monocyte colony-stimulating factor mRNA, which is able to reduce the half-life of ␤-globin mRNA from many hours to less than 30 min (20). Other studies have shown the UUAUUUA(A/)U(A/U) nonamer sequence to be more predictive of rapid mRNA decay than the AUUUA pentamer motif (21,22). Several other RNA binding motifs have been identified, including the C-rich motif that is the target for the poly(C)-binding proteins (CPs) (23). To date, the cis-activity of the 3Ј-UTR of p21 WAF1 has not been characterized extensively.
HuR, a ubiquitously expressed member of the Hu/ELAV family, is involved in the shuttling of transcripts from the nucleus into the cytoplasm (33)(34)(35), as well as in the regulation of mRNA stability (17,(35)(36)(37)(38). In RKO colorectal carcinoma cells, HuR mediates UVC-induced stabilization of p21 WAF1 mRNA (17), and of interest, HuD, a neuron-specific member of the Hu/ELAV family (30), has been shown to bind to a 42-nt sequence within the 3Ј-UTR of p21 WAF1 mRNA (39). It seemed possible, therefore, that HuR would play an important role in the regulation of p21 WAF1 mRNA stability in breast cancer cells.
Here, we show that the 3Ј-UTR of p21 WAF1 mRNA contains multiple cis-acting regions that reduce basal reporter activity and confer EGF-and UVC-induced changes to reporter constructs in a region-specific manner. These 3Ј-UTR elements are the target for a number of RNA-binding proteins, including HuR and CP1, from MDA-468 breast cancer cells. Despite its role in the mediation of p21 WAF1 mRNA stabilization and p21 WAF1 expression by UVC in other cell systems, HuR does not appear to have a major role in EGF-induced p21 WAF1 expression in MDA-468 breast cancer cells (EGFR overexpressed, mutant p53).

MATERIALS AND METHODS
Cell Culture-The MDA-468 (HTB 132) human breast cancer cell line was obtained from ATCC (Manassas, VA). Cells were routinely cultured in Dulbecco's modified Eagle's medium/F-12 medium supplemented with 10% fetal calf serum (Invitrogen). BING cells (40) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. All cell lines were cultured in the presence of penicillin (50 units/ml) and streptomycin (50 g/ml). Cells were utilized within 12 passages of the original stock received from ATCC for all experiments.
Cell Cycle Analysis-For cell cycle analysis, EGF-treated (25 ng/ml, 4 nM) and control MDA-468 cells were harvested by trypsinization and then permeabilized and stained for DNA in phosphate-buffered saline (PBS) containing 0.1% Nonidet P-40, 5 mM EDTA, 5 mM EGTA, 5 g/ml propidium iodide, and 100 g/ml RNase A. Flow cytometry was performed on a Coulter EPICS XL-MCL (Coulter Corp., Hialeah, FL), and cell cycle analysis was performed with MultiPlus AV MultiParameter data analysis software (Phoenix Flow Systems, San Diego, CA).
RNA Isolation and Northern Analysis-MDA-468 cells were solubilized in 4 M guanidinium isothiocyanate, and total RNA was isolated using the method of Chomczynski and Sacchi (43). RNA (10 -15 g per sample) was size-fractionated on a 1% agarose-formaldehyde gel and transferred to Hybond-Nϩ membrane (Amersham Biosciences). RNA was UV cross-linked to the membrane, which was prehybridized for 4 h at 42°C in a buffer containing 50% formamide, 0.75 M NaCl, 0.075 M sodium citrate, pH 7.0, 5ϫ Denhardt's solution, 1% SDS, and 200 g/ml salmon sperm DNA and then hybridized in the same buffer overnight at 42°C with 32 P-labeled p21 WAF1 cDNA probe at 10 6 cpm/ml. The membrane was washed sequentially in 2ϫ SSC/0.1% SDS for 20 min at 22°C, 0.2ϫ SSC/0.1% SDS for 20 min at 22°C, and finally in 0.2ϫ SSC/0.1% SDS at 65°C for 5 min. Membranes were imaged with a PhosphorImager (Molecular Dynamics, Sunnyvale, CA) and quantified using ImageQuant software (Molecular Dynamics, Sunnyvale, CA). In all experiments an 18 S rRNA cDNA probe was used for normalization.
mRNA Turnover Studies-MDA-468 cells (70 -80% confluent) were treated with EGF (25 ng/ml) (Promega) or cycloheximide (10 g/ml) for 2 h followed by the addition of the transcription inhibitor actinomycin D (ActD) at 7.5 g/ml (Sigma). Total RNA was isolated from the cells at 0-, 2-, 4-, and 8-h time intervals after addition of ActD and subjected to Northern analysis as described earlier. p21 WAF1 mRNA half-life was determined using linear regression analysis.
Nuclear Run-on Transcription Assay-MDA-468 cells (70 -80% confluent) were treated with EGF (25 ng/ml) for 2 h. Nuclei were isolated as described previously (44), rapidly frozen, and stored at Ϫ85°C. The transcription assay was performed as described previously (45). Briefly, the nuclei were thawed on ice, resuspended in 100 l of reaction buffer (10 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 300 mM KCl, 5 mM dithiothreitol (DTT), 0.5 mM each of ATP, CTP, and GTP, and 100 Ci of [ 32 P]UTP (3000 Ci/mmol; Amersham Biosciences)), and incubated at 30°C for 30 min. Labeled RNA was isolated and hybridized to nitrocellulose filters onto which 5 g of p21 WAF1 and 18 S rRNA cDNAs had been blotted. Filters were washed and then analyzed by PhosphorImager and Image-Quant software.
Transfection and Luciferase Assays-MDA-468 cells (50% confluent) were transiently transfected with 8 g of pGL3 Ϯ various p21 WAF1 3Ј-UTR regions (see Fig. 1A) and 100 ng of pRL-SV40 as a control, using FuGENE 6 as above. Some cells were cultured following treatment with EGF (25 ng/ml) or UVC (254 nM, 20 J/m 2 ), for 8 or 6 h respectively, prior to lysate extraction. Cells were washed in PBS, harvested by trypsinization, and lysed, and supernatant luciferase activity was measured using the dual luciferase reporter assay kit (Promega) and a Wallac Victor 1420 multilabel counter (Wallac Oy; Turku, Finland), according to the manufacturer's instructions. Firefly luciferase (pGL3) activity was normalized against Renilla luciferase (pRL-SV40) activity to yield the relative luciferase activity.
Preparation of Cytoplasmic Extracts for RNA Gel Shift Assays-MDA-468 cells were grown to 70 -80% confluence in 10-cm culture dishes. Cytoplasmic extracts were prepared as described previously (41). Briefly, cells were scraped from the culture dishes in chilled PBS,

FIG. 1. Schematic of p21 WAF1 3-UTR and clones used in vitro.
A, schematic representation of the p21 WAF1 mRNA 3Ј-UTR sequence and clones used for transfection (pGL3) and REMSA (pBluescript). The 42-nt HuD binding sequence (39) and three AU-rich regions (A, B, and C) are delineated, together with a number of C-rich motifs (shown beneath the 3Ј-UTR sequence, denoted C). The numbers refer to nucleotide positions within the p21 WAF1 mRNA sequence (GenBank TM accession number U03106). B, sequences of c-fos HuD (nt 3399 -3425 of GenBank TM accession number V01512), WAF1-HuD (nt 657-698 of GenBank TM accession number U03106), and WAF1-1/6 (nt 571-829 of GenBank TM accession number U03106). Within WAF1-1/6, WAF1-HuD is underlined and italicized, a CCUCC consensus motif for CP1 is shown in bold italics, and the three AU-rich sequences, A, B, and C, are shaded. C, schematic representation of the GST fusion proteins used for in vitro assays. HuR contains three RNA recognition motif domains (RRMI, RRMII, RRMIII). CP1 contains three K-homology domains (KHI, KHII, KHIII). RNA-binding domains are shown to scale, and the amino acid boundaries of each are defined. centrifuged at 450 ϫ g for 4 min at 4°C, washed again with PBS, and then incubated for 20 min with cold cytoplasmic extract buffer (CEB; 10 mM HEPES, 3 M MgCl 2 , 40 mM KCl, 5% glycerol, 0.2% Nonidet P-40, 1 mM DTT), containing freshly added protease inhibitors (0.5 mM PMSF, 10 g/ml leupeptin, 2 g/ml aprotinin). Lysates were cleared by centrifugation at 4°C for 10 min at 12,100 ϫ g, and the supernatant was snap-frozen in liquid nitrogen and stored at Ϫ80°C. Protein concentrations were determined using the Bio-Rad protein assay kit.
Preparation of Whole Cell Extracts for RNA Gel Shift Assays-MDA-468 cells were grown to 70 -80% confluence in 10-cm culture dishes. Medium was removed, the cell monolayer was washed twice in ice-cold PBS, and the cells were lysed in 0.5 ml of chilled lysis buffer (containing 50 mM Tris, pH 7.5, 5 mM EDTA, pH 8.5, 150 mM NaCl, 1% Triton X-100, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM PMSF, 2 mM NaVO 4 , 50 mM NaF, and 10 mM Na 2 MoO 4 ⅐2H 2 O) on ice for 10 min. Cells were scraped and transferred to Eppendorf tubes, and lysates were then cleared by centrifugation at 4°C for 10 min at 12,100 ϫ g and stored at Ϫ80°C. Protein concentrations were determined by Bio-Rad protein assay kit.
RNA Electrophoretic Mobility Shift Assay (REMSA)-Binding reactions were performed as described previously (41) with 5 g of cytoplasmic extract or 200 ng of recombinant protein and 10 5 cpm of 32 P-labeled RNA (ϳ2-5 pg). Briefly, binding reactions were incubated at 22°C for 30 min, after which 0.3 units of RNase T1 (Roche Molecular Biochemicals) was added for 10 min, followed by the addition of heparin (final concentration 5 mg/ml) (Sigma) for 10 min. Complexes were separated by 4% PAGE, visualized by PhosphorImager, and analyzed by Image-Quant software (Molecular Dynamics, Sunnyvale, CA). In competition assays, extracts were incubated with an unlabeled competitor RNA (ϳ100-fold excess) for 30 min prior to addition of the 32 P riboprobe. For supershift assays with the HuR antibody, the method used was as described previously (46).
UV Cross-linking (UVXL) of RNA-Protein Complexes-RNA-protein binding reactions were carried out as described above, using 20 g of whole cell extract or 100 -500 ng of recombinant protein, 1.5 ϫ 10 5 cpm (10 -15 pg) of 32 P riboprobe (41), and 2-5 g of tRNA. Following the addition of heparin, samples were placed on ice in a microtiter tray and UV-irradiated 1 cm below the Stratalinker UV light source (240 nm UV-bulb; Stratagene) for 10 min. After UVXL, samples were incubated with RNase A (final concentration 100 g/ml) (Roche Molecular Biochemicals) at 37°C for 15 min. The samples were boiled for 3 min in SDS sample buffer (50% glycerol, 0.25 M Tris, pH 6.8, 10% SDS, 4% ␤-mercaptoethanol) and subjected to SDS-PAGE (gels ranging from 8.5% to 12%), and RNA-protein complexes were detected by PhosphorImager. In some competition experiments, recombinant proteins were incubated with ribohomopolymers (poly(C), poly(A)) for 20 min, prior to addition of riboprobe. For UVXL immunoprecipitation assays, UVXL was performed as described above, and the reactions were then incubated for 1 h with HuR or GST antibody at 4°C, followed by incubation with protein A and G beads (Sigma) for 45 min at 4°C. After washing, RNA-protein complexes were resolved by SDS-PAGE (gels ranging from 8.5 to 12%) and detected by PhosphorImager. In all UVXL experiments, 14 C Rainbow molecular mass markers (Amersham Biosciences) were used.
IP-RT-PCR Assay-MDA-468 cells were grown to 50% confluence in 10-cm dishes, and cytoplasmic extracts were harvested as described above. Cytoplasmic extract (200 g) was added to 10 g of HuR, CP1, or EGFR antibody. After incubation on ice for 45 min, 5 g each of protein A (Amersham Biosciences) and protein G (Sigma) beads was added to all tubes, which were mixed for a further 45 min at 4°C. The tubes were centrifuged at 2,000 ϫ g for 2 min, and the supernatants were removed for RNA extraction. The pelleted beads were washed with cold CEB (10 ϫ 1 ml), and RNA was extracted using TRIzol reagent. Reverse transcription was performed using random hexamers (Promega) and standard procedures. PCR was performed for 33 cycles, comprising five cycles of denaturation at 94°C/30 s, annealing at 66°C/30 s, and extension at 72°C/1 min, followed by 28 cycles of denaturation at 94°C/30 s, annealing at 55°C/30 s, and extension at 72°C/1 min, with the primers 481F (5Ј-GAC TCT CAG GGT CGA AAA CG-3Ј) and 585R (5Ј-CTT CCT GTG GGC GGA TTA G-3Ј) from within the coding sequence of p21 WAF1 . These primers produce an amplicon that spans an intron, allowing the discrimination between cDNA-and genomic DNArelated products in these PCR experiments. PCR products were resolved on an ethidium bromide-stained 3% agarose gel.
Retroviral Expression of HuR-Full-length HuR cDNA was cloned into the pBabe puro vector (42) in the sense and antisense orientation and then transiently transfected into the retroviral packaging cell line BING (40) using FuGENE according to the manufacturer's protocol. Retroviral-containing conditioned medium was collected from the BING cells at ϳ48 h after transfection. Following filtration (0.45 m) and the addition of 4 g/ml polybrene (hexadimethrine bromide; Sigma), the retroviral-containing medium was incubated overnight with the target cells (MDA-468). Cells were selected in 1 g/ml puromycin (Sigma) starting 48 h after infection. Pools of puromycin-resistant cells were analyzed by Western blotting to confirm transgene expression. All subsequent experiments were performed using pools of infected cells.
Colony Formation Assays-Colony formation assays were performed as described previously (58). Briefly, MDA-468 sublines were plated in triplicate at a density of 5000 cells per 10-cm plate and then incubated overnight. A single dose of EGF (25 ng/ml) or PBS control was added at 24 h, and the cells were allowed to grow for a further 10 days. After fixation in methanol:acetic acid (3:1), colonies were stained with Giemsa (Fluka) and counted using a Quantimet 520 image analyzer (Leica).
Statistical Analysis-Transfection and luciferase mRNA decay data are shown as mean Ϯ S.D. Statistical analysis was performed using Student's t test, with a p value of Ͻ0.05 regarded as significant.

EGF Up-regulates p21 WAF1 Expression in MDA-468 Cells-
The MDA-468 breast cancer cell line, which contains mutant p53 and overexpresses the EGFR (7), provides an excellent model system to investigate the mechanisms underlying p21 WAF1 gene expression and its regulation by EGF. To establish the validity of this cell line as a model of EGF-induced cell cycle arrest and growth inhibition, the cells were treated with EGF (25 ng/ml) for 8 h, and the proportion of cells in S-phase of the cell cycle was determined using flow cytometry (see "Materials and Methods"). EGF treatment reduced the proportion of cells in S phase from 30 to 12.5% (data not shown). Furthermore, in a colony formation assay (see "Materials and Methods"), EGF treatment led to Ͼ98% reduction in the number of detectable colonies (data not shown).
We next examined the effect of EGF on endogenous p21 WAF1 mRNA and protein levels in MDA-468 cells. EGF rapidly upregulated p21 WAF1 protein ( Fig. 2A) and p21 WAF1 mRNA (Fig.  2B) within 2 h. To determine whether this up-regulation of p21 WAF1 mRNA occurred at the post-transcriptional level, we treated MDA-468 cells with EGF for 2 h and performed ActD chase studies. In the absence of EGF, the basal half-life of p21 WAF1 mRNA was relatively short (2.7 h) (Fig. 2C). However, in the presence of EGF the half-life was increased to 11.5 h (Fig. 2C). Treatment of the cells with cycloheximide, a translational inhibitor, also stabilized p21 WAF1 mRNA (Fig. 2C), suggesting that (i) ongoing translation is required for maintenance of the short basal half-life, and/or (ii) existing cellular proteins mediate the stabilization of p21 WAF1 mRNA.
Nuclear run-on assays were employed to evaluate the effect of EGF on p21 WAF1 transcriptional activity. After treatment of MDA-468 cells with EGF, p21 WAF1 transcription increased by ϳ2-fold (Fig. 2D). Taken together, these data suggest that the EGF-induced growth arrest in MDA-468 cells is associated with a rapid increase in p21 WAF1 mRNA and protein, which results from a combination of increased mRNA stability and increased transcription.
Identification of cis-Acting Elements in the 3Ј-UTR of p21 WAF1 mRNA-Based on the observation that the stabilization of p21 WAF1 mRNA is a major contributor to the overall up-regulation of p21 WAF1 protein expression in EGF-treated MDA-468 cells, we next sought to elucidate some of the mechanisms underlying this effect. In breast cancer cells, little is known of the cis-acting elements or trans-acting factors involved in the regulation of p21 WAF1 mRNA stability. Transfection studies with ActD chase and real-time PCR demonstrated that WAF1-1/7 (entire p21 WAF1 3Ј-UTR; see Fig. 1A) destabilized luciferase mRNA significantly, in a manner equivalent to that of the highly unstable c-fos ARE (Fig. 3). These data provided strong evidence that the p21 WAF1 3Ј-UTR contains one or more cis-elements that confer basal mRNA instability and potentially contribute to the regulation of p21 WAF1 mRNA stability.
The 3Ј-UTR of p21 WAF1 contains an AU-rich region at the 5Ј end that spans ϳ250 nt, termed WAF1-1/6 (see Fig. 1, A and  B), which contains at least one known HuR binding site (17). Within WAF1-1/6 is a 42-nt sequence, termed WAF1-HuD, which contains an imperfect consensus nonamer and is the target for HuD binding (39). The WAF1-1/6 region also contains several smaller stretches of AU-rich sequence, denoted A, B, and C (see Fig. 1, A and B, and see "Materials and Methods") MDA-468 cells were transfected with 8 g of pGL3-control, pGL3-WAF1-1/7 (see Fig. 1), or pGL3-c-fos ARE, and 38 h after transfection they were treated with ActD (7.5 g/ml) for 0, 2, and 4 h. Total MDA-468 RNA was harvested, and cDNA was generated by RT and used in real-time PCR with luciferase-and ␤-actin-specific primers. Data were normalized using results obtained for ␤-actin, and the ratio of luciferase RNA remaining for pGL3-WAF1-1/7 and pGL3-c-fos-ARE was expressed relative to pGL3-control. The data are representative of three separate experiments performed in triplicate. **, p Ͻ 0.001.

FIG. 2. EGF increases p21 WAF1 mRNA and protein levels and p21 WAF1 mRNA stability and transcription in MDA-468 cells. A,
Western blot analysis showing p21 WAF1 protein levels in MDA-468 cells following treatment with EGF (25 ng/ml). Cell lysates (10 g) were subjected to SDS-PAGE, transferred to polyvinylidene difluoride membranes, and probed with anti-p21 WAF1 and actin antibodies. ECL-generated images were quantified using ImageQuant and p21 WAF1 protein levels graphed over time. B, Northern blot analysis of total RNA extracted from MDA-468 cells after treatment with EGF (25 ng/ml) for the times indicated. Following hybridization with a 32 P-labeled p21 WAF1 cDNA probe, each blot was normalized using a 32 P-labeled 18 S ribosomal RNA cDNA probe. Quantification was performed using a PhosphorImager and ImageQuant software, and p21 WAF1 mRNA levels were graphed over time. C, ActD chase studies in MDA-468 cells. Cells were grown to 50% confluence and treated with 25 ng/ml EGF or 10 g/ml cycloheximide (CHX) for 2 h and then 7.5 g/ml ActD. Total RNA was extracted from the cells at 0, 2, 4, or 8 h after ActD treatment and analyzed by Northern blot as in B. p21 WAF1 mRNA was normalized against 18 S rRNA (image not shown). Half-life of p21 WAF1 mRNA was 11.5 h (EGF-treated cells) and 2.7 h (control cells). **, p Ͻ 0.01. D, transcription run-on analysis of MDA-468 cells after treatment with EGF (25 ng/ml) for 2 h. p21 WAF1 transcription rates were measured in isolated nuclei by run-on transcription assays, and the results were analyzed by PhosphorImager and ImageQuant and shown in this figure relative to 18 S rRNA transcription levels. CON, control; EGF, 2-h EGF treatment. and was therefore a candidate cis-acting sequence.
To further dissect the cis-activity of the p21 WAF1 3Ј-UTR, we generated several reporter constructs containing portions of the 3Ј-UTR of p21 WAF1 (Fig. 1A). In transfection assays using MDA-468 cells, the full-length 3Ј-UTR, WAF1-1/7, reduced basal reporter activity by ϳ85% (Fig. 4, A and B), supporting our ActD-luciferase mRNA data (Fig. 3). Subsequent analysis of the three major components of the 3Ј-UTR (WAF1-1/6, WAF1-879, WAF1-1512) showed that each reduced reporter activity but that the major effect was 3Ј of the previously identified AU-rich region contained within WAF1-1/6 ( Fig.  4B). In support of this observation, clones WAF1-2/7 and WAF1-6/7, which harbored deletions of the AU-rich regions, reduced reporter activity similarly to WAF1-1/7. Taken together, these results suggest that the WAF1-1/6 region is not the sole determinant of basal p21 WAF1 mRNA stability in MDA-468 cells and that the four AU-rich regions (HuD; see Fig. 1, A-C) are not major contributors to basal turnover of p21 WAF1 mRNA in MDA-468 cells. Furthermore, although WAF1-879 and WAF1-1512 each decrease reporter activity, neither is as effective as the combined sequence (WAF1-6/7) (Fig. 4B). This suggests the presence of multiple cis-elements within the 3Ј-UTR of p21 WAF1 mRNA.
We next examined the effect of EGF on the reporter activity of each of these constructs in MDA-468 cells. EGF increased reporter activity by ϳ60% in the case of WAF1-1/7, with most of this effect being contained within the WAF1-6/7 region (Fig.  4C). Both the WAF1-879 and WAF1-1512 constructs conferred EGF-induced up-regulation of luciferase reporter activity, whereas the WAF1-1/6 construct appeared to be relatively EGF-unresponsive (Fig. 4C). Taken together, these results suggest that the predominant cis-element(s) within the 3Ј-UTR of p21 WAF1 that are responsible for both basal mRNA instability and EGF-inducibility in MDA-468 cells reside downstream of the WAF1-1/6 sequence.
To compare these results with another regulator of p21 WAF1 mRNA stability, we tested the effect of UVC treatment on reporter activity using the same constructs in transfection experiments with MDA-468 cells. The full-length 3Ј-UTR (WAF1-1/7) increased reporter activity by ϳ70% after UVC treatment (Fig. 4D). Further analysis revealed that the UVCmediated up-regulation of reporter activity occurred predominantly through sequences contained within the WAF1-1/6 construct, with a lesser contribution from sequences downstream of WAF1-1/6 (Fig. 4D). This result suggests that the WAF1-1/6 region is the major 3Ј-UTR determinant of UVC-induced stabilization of p21 WAF1 mRNA in MDA-468 cells. These data illustrate that even within the one cell type, different stimuli may lead to the preferential regulation of different and specific cis-elements within the p21 WAF1 3Ј-UTR, presumably via different sets of RNA-protein interactions.
HuR and CP1 Bind to p21 WAF1 mRNA-Previous studies in other cell types have demonstrated that the WAF1-HuD region within p21 WAF1 mRNA is a target for members of the ELAV RNA-binding protein family (e.g. HuD) (39). Wang et al. (17) showed that increased binding of HuR to the WAF1-1/6 element mediated the UVC-induced stabilization of p21 WAF1 mRNA (17). This suggested that the ϳ36-kDa band observed in UVXL studies with the WAF1-HuD and WAF1-1/6 probes and using MDA-468 breast cancer extracts (Fig. 5C, lanes 1 and 3) contained HuR. In addition, a preponderance of potential CP1 binding sites within the p21 WAF1 3Ј-UTR (see Fig. 1A), together with the observed RPCs at ϳ42 kDa using the p21 WAF1 3Ј-UTR riboprobes (Fig. 5C, lanes 1 and 3; Fig. 5D, lanes 2 and 3), suggested that CP1 protein might target p21 WAF1 mRNA.
To investigate the association among HuR, CP1, and p21 WAF1 mRNA in MDA-468 cells, we utilized an immunoprecipitation-RT-PCR assay with primers that target the p21 WAF1 coding region (see "Materials and Methods"). Using HuR and CP1 antibodies, we were able to co-immunoprecipitate p21 WAF1 mRNA from MDA-468 cells (Fig. 6A, lanes 2 and 3). However, no p21 WAF1 -specific PCR product was identified when using an unrelated antibody (EGFR) (Fig. 6A, lane 4) or with no antibody (beads alone) (Fig. 6A, lane 1). Controls were used routinely in these assays: positive (assay of supernatant following immunoprecipitation; see . These data provide definitive evidence that HuR and CP1 interact closely with p21 WAF1 mRNA in MDA-468 cells. To definitively map the binding site of HuR to WAF1-HuD we performed a UVXL assay using MDA-468 extracts and immunoprecipitated the resultant RPCs with HuR antibody. We identified a single major RPC with a molecular mass of ϳ36 kDa that was not present when the RPCs were precipitated with GST antibody or with beads alone (Fig. 6B, lanes 3-5). Similar results were obtained with the WAF1-1/6 probe (Fig.  6B, lanes 8-10). These findings provided strong evidence that the 36-kDa RPC detected in UVXL in Fig. 5C represents HuR bound to the WAF1-HuD and WAF1-1/6 probes. When tested, thrombin-cleaved recombinant GST-HuR bound to the WAF1-HuD and c-fos HuD probes (Fig. 6C, lanes 1 and 3), and in each case, the RPC could be supershifted with HuR antibody (Fig.  6C, lanes 2 and 4).
We next used MDA-468 whole cell extracts treated with EGF or UVC in UVXL experiments with WAF1-HuD and WAF1-1/6 probes. In each case, UVC up-regulated binding of HuR (ϳ36 kDa) to the probe, whereas EGF did not (WAF1-HuD; see Fig.  6D, lanes 1-5 and WAF1-1/6; see Fig. 6D, lanes 6 -10). A similar UVC-induced increase in binding of HuR to p21 WAF1 FIG. 5. Proteins from MDA-468 cells bind specifically to cis-acting elements from the 3-UTR of p21 WAF1 mRNA. MDA-468 extracts were incubated with a panel of 32 P-labeled p21 WAF1 or vector control (pBluescript) riboprobes, and REMSA (4% PAGE) or UVXL (10 min of UVXL exposure, SDS-PAGE 8.5 to 12%) was performed as described under "Materials and Methods." A, REMSA utilizing various p21 WAF1 (lanes 1, 2, and  5-10), vector control (pBluescript) ( lanes  11 and 12), or c-fos HuD (lanes 3 and 4)  mRNA has been observed in RKO colorectal carcinoma cells (17). However, no significant change was seen in the pattern of binding for any of the other p21 WAF1 RNA-binding proteins. These data suggest an important role for HuR in the UVCinduced up-regulation of p21 WAF1 mRNA stability and implicate a lesser role for HuR in the EGF-mediated changes in p21 WAF1 mRNA turnover. These results support our transfection data (Fig. 4, C and D) in which EGF up-regulated luciferase reporter activity through sequences downstream of WAF1-HuD and WAF-1/6, whereas UVC regulated luciferase activity primarily through sequences within WAF1-1/6, presumably those which bind HuR.
To determine the functional role of HuR in the regulation of p21 WAF1 expression and control of cell cycle in MDA-468 cells, we used retroviral vectors to generate stable pools of MDA-468 cells expressing antisense or sense HuR. HuR protein levels varied significantly between the antisense and sense MDA-468 sublines. Despite this, no significant difference was observed in either p21 WAF1 or actin protein levels following EGF treatment (Fig. 7A). We also used flow cytometry to examine the effect of HuR levels on the cell cycle profile. However, no difference was seen between the sublines (Fig. 7B), where the proportion of cells in S phase was unaffected by increasing (sense) or decreasing (antisense) HuR levels. The EGF-induced reduction in S phase content was also unaffected by HuR levels (Fig. 7B). Similarly, in a colony formation assay (not shown), EGF induced Ͼ98% reduction in the number of colonies with each subline. Taken together, these data suggest that although HuR binds to the WAF1-HuD sequence of p21 WAF1 mRNA (within the context of the larger WAF1-1/6 region) in MDA-468 cells, modification of cellular HuR levels has little or no effect on the regulation of p21 WAF1 protein levels or progression of cells through the cell cycle. DISCUSSION p21 WAF1 plays a central role in various models of growth inhibition, although the molecular mechanisms that regulate p21 WAF1 expression in breast cancer cells are not well understood. Here we have shown that EGF-induced growth inhibition in MDA-468 breast cancer cells (mutant p53) is associated with a rapid increase in p21 WAF1 protein and mRNA expression, which results from the combination of increased transcription and stabilization of p21 WAF1 mRNA. Significantly, we established that the 3Ј-UTR of p21 WAF1 contains cis-acting  4 and 9), and with beads alone (lanes 5 and 10) prior to 8.5% SDS-PAGE and PhosphorImager analysis compared with 14 C molecular mass markers (lanes 1 and 6). C, REMSA-supershift assay with cleaved recombinant HuR protein and WAF1-HuD (lanes 1 and 2) or c-fos HuD (lanes 3 and 4) riboprobes. HuR antibody was used to shift RPCs (lanes 2 and 4). UVXL assay (lanes 6 -12) using WAF1-1/6 riboprobe and recombinant GST (lane 7) or cleaved CP1 protein (lanes 8 -12), in the presence of either poly(C) (lanes 9 and 11) or poly(A) (lanes 10 and 12) is shown compared with 14 C molecular mass markers (lane 5). D, UVXL assay with either WAF1-HuD or WAF1-1/6 riboprobes and control (lanes 2 and 8), EGF (lanes 4 and 9)-, or UVC (lanes 5 and 10)-treated MDA-468 cell extracts compared with 14 C molecular mass markers (lanes 3 and 6).
elements that modulate the stability of a heterologous reporter mRNA. Through transfection studies, distinct regions have been defined within the p21 WAF1 3Ј-UTR that basally regulate, and confer UVC-and EGF-inducible changes to, heterologous reporter activity. Furthermore, these cis-elements are the target for several RNA-binding proteins, including HuR and CP1. Despite its documented role in mediating the stabilization of a number of transcripts, including p21 WAF1 mRNA in UVCtreated RKO colorectal carcinoma cells (17), and in promoting cyclooxygenase-2 expression in colon carcinoma cells (59), our binding assays, together with sense and antisense expression studies, suggest that HuR does not play a major role in the EGF-induced expression of p21 WAF1 in MDA-468 breast cancer cells.
The data presented herein support the notion that posttranscriptional pathways are a major regulator of p21 WAF1 gene expression. Several compounds have been shown to increase the stability of p21 WAF1 mRNA in a variety of cells. These include phorbol ester (phorbol 12-myristate 13-acetate; ϳ8-fold increase in p21 WAF1 mRNA stability in human ovarian carcinoma SKOV-3 cells) (12), tumor necrosis factor ␣ (ϳ5-fold increase in human myeloid leukemic KG-1 cells) (13), a novel retinoid CD437 (ϳ3-fold increase in MDA-468 and MCF-7 breast cancer cells) (14), UV light (ϳ4-fold increase in mouse embryonal fibroblasts) (15), phenylephrine (ϳ3-fold increase in transfected HepG2 cells, mediated by p42/44 MAP kinase) (48), and of direct relevance to this work, EGF (ϳ2-fold in A431 human epidermoid carcinoma cells) (16). This response to EGF may only be apparent in cells that overexpress EGFRs, such as MDA-468 cells and A431 cells, as EGF did not modulate p21 WAF1 mRNA stability in MCF-7 cells, which express lower levels of EGFR (16). In each of these reports, where measured, the transcriptional increase in p21 WAF1 contributed less than the increase in mRNA stability to the total mRNA level.
Our data provides definitive mRNA decay evidence implicating the p21 WAF1 3Ј-UTR in cis, supporting the findings of Liu et al. (48). Previous studies into the effect of shorter elements, such as WAF1-HuD or WAF1-1/6, demonstrated that the WAF1-HuD region did not destabilize a CAT reporter RNA in transfected HepG2 cells significantly (48). Similarly, Li et al. (48) found that the ARE motifs present within the p21 WAF1 3Ј-UTR (HuD; A, B, and C in Fig. 1, A and B) did not contribute substantially toward message instability in breast cancer cells (14). Instead, they found that the predominant basal instability sequences were contained downstream of the WAF1-1/6 region thought responsible for UVC-induced stabilization of p21 WAF1 mRNA. Together with our data, these findings suggest that the p21 WAF1 3Ј-UTR is a composite cis-acting sequence, with contributions to basal turnover from each of the WAF1-1/6, WAF1-879, and WAF1-1512 regions, but that the majority of the effect is because of sequences downstream of WAF1-1/6 (cf. with Li et al. (14)). Interestingly, EGF and UVC augmented reporter activity preferentially via different components of the 3Ј-UTR. In particular, EGF-induced up-regulation of reporter activity occurred predominantly via a combination of WAF1-879 and WAF1-1512, with little effect via WAF1-1/6. In contrast, UVC augmented reporter activity predominantly through WAF1-1/6, consistent with the findings of Wang et al. (17), together with smaller, yet significant, up-regulation via WAF1-879 and WAF1-1512 sequences. Further analysis of each of these three components of the 3Ј-UTR will be required in different cell types and with different stimuli to develop a definitive understanding of the mechanisms underlying p21 WAF1 mRNA turnover.
We have produced several lines of evidence in support of the association of HuR and CP1 with p21 WAF1 mRNA in MDA-468 breast cancer cells. These include the immunoprecipitation of endogenous HuR bound to the WAF1-HuD and WAF1-1/6 riboprobes and the immunoco-purification of p21 WAF1 mRNA from MDA-468 cell extracts using HuR and CP1 antibodies. These assays (UVXL-IP and IP-RT-PCR) provide the first definitive evidence for a close association of HuR and CP1 with p21 WAF1 mRNA in MDA-468 cells.
Based on these observations and the findings of others, we presumed there would be a significant role for HuR in the EGF-induced regulation of p21 WAF1 mRNA stability in breast cancer cells. For example, HuR and other members of the ELAV protein family have been shown to stabilize AU-rich mRNAs in several other cell systems (35). These include the stabilization of VEGF mRNA (36), GLUT-1 mRNA (49), and p21 WAF1 mRNA in UVC-treated colorectal carcinoma cells (17). In the latter report, the shortest riboprobe that the authors used was equivalent to our WAF1-1/6 probe (see Fig. 1A). HuR was one of only two proteins detected by UVXL, and HuR antibody produced a partial supershift in cell extracts. Furthermore, antisense HuR-expressing clones (with a 4 -6-fold reduction in HuR levels) demonstrated a decrease in both basal and UV-induced p21 WAF1 expression (mRNA stability and protein levels). However, we observed that the WAF1-1/6 region did not have a major role in modulating EGF-induced reporter activity in MDA-468 cells. Moreover, we found that treatment of MDA-468 cells with UVC, but not EGF, regulated the binding of HuR to the WAF1-HuD and WAF1-1/6 riboprobes. Thus, we were not surprised to find that modification of HuR expression in MDA-468 cells had no detectable effect on EGF-induced p21 WAF1 protein levels, cell cycle, or growth. Taken together, these observations suggest that in these cells, HuR plays a relatively minor role in the regulation of basal and EGF-induced expression of p21 WAF1 .
In this context, Liu et al. (48) found that non-HuR RNAbinding proteins (24 and 52 kDa) mediated the induction of p21 WAF1 mRNA stability by the ␣ 1 adrenergic agonist, phenylephrine (48). They did not observe binding of cellular HuR to their riboprobe, which was identical to our WAF1-HuD ribo- probe. We therefore presume that the role of HuR in regulating p21 WAF1 expression varies according to the mode of stimulus and is dependent upon cell type. It also emphasizes that RNAbinding proteins other than HuR can regulate p21 WAF1 mRNA turnover.
We have identified a U-and C-rich cis-element in the 3Ј-UTR of the human androgen receptor mRNA that is the target for simultaneous, co-operative binding of HuR and CP1/CP2 (47). The UVXL assays presented herein with recombinant CP1 suggest that CP1 may bind to the UVC-responsive WAF1-1/6 element. The close proximity of HuR and CP1 binding sites within WAF1-1/6 might allow both proteins to participate in coordinated mRNA decay. It also emphasizes the need to examine the functional role of CP1 in p21 WAF1 mRNA turnover in MDA-468 cells.
In summary, EGF increases p21 WAF1 mRNA expression in p53 mutant breast cancer cells through a combination of mRNA stabilization and transcriptional up-regulation. We have identified cis-elements within the p21 WAF1 3Ј-UTR that are distinctly EGF-or UVC-inducible in MDA-468 cells. This implies that different stimuli can regulate p21 WAF1 mRNA stability via independent cis-elements. HuR binding modulates p21 WAF1 expression in UVC-treated RKO cells but not in EGFtreated MDA-468 cells. This indicates that there is an HuRindependent, cell type-specific mechanism through which EGF induces p21 WAF1 expression via stabilization of p21 WAF1 mRNA. CP1 and other RNA-binding proteins associate with p21 WAF1 mRNA in MDA-468 cells and may direct its turnover. The cloning and characterization of these proteins are the subject of further investigation.