MOLECULAR AND CELLULAR BIOLOGY, Sept. 2004, p. 7987–7997 0270-7306/04/$08.00⫹0 DOI: 10.1128/MCB.24.18.7987–7997.2004 Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Vol. 24, No. 18
Regulation of Human p53 Activity and Cell Localization by Alternative Splicing Anirban Ghosh,† Deborah Stewart,† and Greg Matlashewski* Department of Microbiology and Immunology, McGill University, Montreal, Canada Received 6 May 2004/Accepted 14 June 2004
The development of cancer is a multistep process involving mutations in proto-oncogenes, tumor suppressor genes, and other genes which control cell proliferation, telomere stability, angiogenesis, and other complex traits. Despite this complexity, the cellular pathways controlled by the p53 tumor suppressor protein are compromised in most, if not all, cancers. In normal cells, p53 controls cell proliferation, senescence, and/or mediates apoptosis in response to stress, cell damage, or ectopic oncogene expression, properties which make p53 the prototype tumor suppressor gene. Defining the mechanisms of regulation of p53 activity in normal and tumor cells has therefore been a major priority in cell biology and cancer research. The present study reveals a novel and potent mechanism of p53 regulation originating through alternative splicing of the human p53 gene resulting in the expression of a novel p53 mRNA. This novel p53 mRNA encodes an N-terminally deleted isoform of p53 termed p47. As demonstrated within, p47 was able to effectively suppress p53-mediated transcriptional activity and impair p53-mediated growth suppression. It was possible to select for p53-null cells expressing p47 alone or coexpressing p53 in the presence of p47 but not cells expressing p53 alone. This showed that p47 itself does not suppress cell viability but could control p53-mediated growth suppression. Interestingly, p47 was monoubiquitinated in an Mdm2-independent manner, and this was associated with its export out of the nucleus. In the presence of p47, there was a reduction in Mdm2-mediated polyubiquitination and degradation of p53, and this was also associated with increased monoubiquitination and nuclear export of p53. The expression of p47 through alternative splicing of the p53 gene thus has a major influence over p53 activity at least in part through controlling p53 ubiquitination and cell localization. what physiological role p47 may play. It has recently been reported, with a transgenic mouse model, that overexpression of p47 (mouse p44) resulted in p53-dependent cellular senescence and reduced life span in these mice (14). Taken together, the N-terminally truncated version of p53 (p47) has emerged as a potentially significant p53 regulatory protein, and it is therefore important to define the mechanisms of p47 expression and regulation of p53 activity, as addressed in the present study. During the original cloning of the human p53 gene (16, 18) a partial cDNA clone, terminating at the 5⬘ end within the intron 2 sequence, was isolated from a cDNA library constructed from primary human foreskin fibroblast mRNA (17). Since this cDNA was incomplete and did not contain an inframe start methionine codon at the 5⬘ end, no further work was carried out on this cDNA clone until the present investigation. Because of the growing interest in N-terminally truncated p53 family members, and because intron 2 is downstream from the p53 start codon, we resumed an investigation of this novel p53 transcript. As detailed within, an intron 2-containing p53 transcript has been identified in mature polysomal mRNA, which is capable of expressing an N-terminally truncated isoform of p53 termed p47. The alternative splice-derived p47 product did not suppress cell viability but was able to control p53 ubiquitination, cell localization, and activity. These observations argue that alternative splicing of the p53 gene results in potent p53 regulatory activity.
The p53 tumor suppressor protein inhibits malignant cell transformation by mediating cell cycle arrest and apoptosis following cellular stress, including ectopic oncogene expression (1, 11). Mutations in the p53 gene or disruptions of the pathways involved in the activation of p53 appear to be a common feature of all cancers. Moreover, p53-deficient mice are rendered highly susceptible to sporadic cancers (4), and germ line mutations in p53 result in Li-Fraumeni syndrome, which predisposes individuals to a variety of cancer types (15). p53 is considered the prototype tumor suppressor gene, and defining the mechanisms that regulate p53 function is important for understanding the development of cancer. The p53 protein belongs to a family of analogous proteins, including p63 and p73, which share substantial sequence identity, structure, and are sequence specific transcription factors capable of mediating apoptosis (9, 19). Both p63 and p73 genes undergo alternative splicing, giving rise to the expression of a variety of isoforms, including the ⌬N isoforms, which lack the N-terminal transactivation domain. ⌬Np73 is capable of inhibiting both p73 and p53 activity (19, 27). A ⌬N isoform of human p53, termed p47, which lacks the N-terminal transactivation domain, has also been identified (3, 27). These studies reported that p47 arises through the use of different sites for translation initiation on the same p53 mRNA. However, it remains poorly understood how p47 regulates p53 activity and * Corresponding author. Mailing address: Department of Microbiology and Immunology, McGill University, 3775 University St., Room 511, Montreal, Quebec, Canada H3A 2B4. Phone: (514) 398-3914. Fax: (514) 398-7052. E-mail:
[email protected]. † A.G. and D.S. contributed equally to this research.
MATERIALS AND METHODS Cell lines and reagents. Human H1299 cells were kindly provided by Sam Benchimol. Human cell lines Saos-2, HeLa, MCF-7, HT1080, and SIHA and
7987
7988
GHOSH ET AL.
murine p53-null 10(1) cells were originally obtained from the American Type Culture Collection (Rockville, Md.). The cells were grown at 37°C in 5% CO2 in Dulbecco’s modified Eagle’s medium (Invitrogen) supplemented with 10% fetal bovine serum and 100 ng/ml each of streptomycin and penicillin. Monoclonal antibodies DO1 and 1801 were obtained from Oncogene Research Products; anti-mouse immunoglobulin G–fluorescein isothiocyanate was from Santa Cruz Biotechnology Inc.; anti-mouse immunoglobulin G–horseradish peroxidase was from Amersham Bioscience Ltd.; SYBR Green I was from Sigma; luciferase assay kit was from Stratagene; protease inhibitor cocktail was from Roche Diagnostics; RNase inhibitor was from Sigma; protein A-Sepharose was from Sigma; transfection reagent Lipofectamine Plus was from Invitrogen; proteasome inhibitor MG-132 was from Calbiochem; and Trizol reagent was from Invitrogen. Molecular and cellular biological techniques. All the RNA preparations were made DNA free before subsequent experiments. A human p53 intron 2-specific reverse primer was used in a 5⬘ rapid amplification of cDNA ends on total HeLa RNA with a GeneRacer kit (Invitrogen). The p53(EII) cDNA was constructed with PCR to combine the intron 2-specific 5⬘ rapid amplification of cDNA ends fragment and full-length p53 sequences, followed by cloning into pCDNA3. The final p53(EII) cDNA was sequenced to confirm the absence of mutations during cloning. The prototype p53 cDNA was previously described (18). Plasmid pCDNA3-p47 was created by deleting a BamHI fragment containing exon 1, exon 2, and intron 2 from pCDNA3-p53(EII). The oligomerization mutants were generated with PCR to insert a stop codon corresponding to amino acid 338 in the p53, p53(EII), and p47 cDNA sequences. For first-strand cDNA synthesis and real-time PCR, either oligo(dT) or human p53 exon 11-specific reverse primers were used on 5 g of DNA-free polysomal or total RNA for reverse transcription with the Superscript 1st Strand synthesis system (Invitrogen) according to the manufacturer’s protocol. Out of 20 l of first-strand cDNA, 1 l was subjected to PCR amplification with primers specific for each isoform of the p53 transcript. Real-time PCR was conducted with an ABI Prism 7700 in the presence of SYBR Green I with Advantage2 polymerase (Clontech). For the colony formation assay and the generation of stable cell lines, the above p53/p47 cDNAs were subcloned into the pCIN4 vector, which contains the encephalomyocarditis virus promoter and an internal ribosome entry site (IRES) downstream from the cloned sequences, followed by the neomycin resistance gene. Polyribosome purification. Following two washes with cold phosphate-buffered saline, 108 cells were lysed in 1 ml of polysomal buffer (25 mM Tris [pH 7.5], 50 mM NaCl, 5 mM Mgcl2, 0.25 M sucrose, 200 U of RNase inhibitor per ml) with 1% Triton X-100 on ice for 20 min. After removing the nuclei and cell debris by centrifugation at 16,000 ⫻ g for 15 min at 4°C, the polyribosomes were purified by pelleting through a 2 M sucrose cushion (in polysomal buffer) by ultracentrifugation at 600,000 ⫻ g for 2 h in a 75 Ti rotor (Beckman) at 4°C. The RNA was isolated from polyribosomes with Trizol reagent. Immunofluorescence microscopy. For immunofluorescence experiments, cells grown on glass coverslips were transfected or treated as indicated and fixed after 24 h or the indicated time in 80% methanol–20% acetone for 20 min at ⫺20°C. Cells were then rehydrated three times with cold phosphate-buffered saline and blocked with 10% milk in phosphate-buffered saline with 0.02% bovine skin gelatin. Cells were incubated with monoclonal antibody DO1 at 1:500 or monoclonal antibody 1801 at 1:50 for 2 h. Following three washes with phosphatebuffered saline, cells were incubated with secondary antibody at 1:500 (antimouse immunoglobulin G–fluorescein isothiocyanate) in blocking solution. After three washes, coverslips were mounted with PermaFluor aqueous mountant (Pierce). The slides were observed with either a Zeiss LSM510 laser scanning microscope or an inverted fluorescence microscope. Western blot and immunoprecipitation analysis. Western blot analysis of p53 was performed as previously detailed (25) with monoclonal antibody DO1 to detect only full-length p53 and monoclonal antibody 1801 to detect both fulllength p53 and p47. For immunoprecipitations, cells were washed twice with cold phosphate-buffered saline and lysed on ice in lysis buffer (50 mM Tris [pH 8.0], 150 mM NaCl, 1% NP-40, 1 tablet of protease inhibitor cocktail/10 ml). Cell debris was discarded after centrifugation at 16,000 ⫻ g for 10 min. Cleared lysate was incubated at 4°C for 2 h with monoclonal antibody DO1 (1:1,000 dilution), followed by 1/10 volume of protein A-Sepharose for 30 min. Immunoprecipitates were washed four times with cold lysis buffer and subjected to sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) followed by Western blot analysis with monoclonal antibody 1801. Luciferase transcription assay. Luciferase activity was determined in cells transfected with p53/p47 cDNAs and a p53-responsive p21 promoter luciferase gene construct, as previously described (25). Cells were cotransfected with a
MOL. CELL. BIOL. -galactosidase expression plasmid and galactosidase activity was measured to ensure equal transfection efficiencies, as previously described (25). Cell viability and colony formation assays. For the colony formation assay, 5 ⫻ 105 Saos-2 or H1299 cells were seeded onto six-well plates and transfected with the various p53/p47 cDNAs cloned into the pCIN4 plasmids described above. The pCIN4 vector was used for this assay because it contains an internal ribosome entry site and neomycin resistance gene downstream from the inserted gene. In this manner, cells expressing the transfected gene can be selected in the presence of G418. After 24 h, transfected cells were placed in medium supplemented with 500 g of G418 per ml to select cells taking up the transfected plasmids. Cells were grown for 14 days, at which time surviving colonies were apparent and stained with Giemsa. Ubiquitination assays. H1299 cells were cotransfected with the various p53/ p47-expressing cDNAs cloned into the pCDNA3 expression plasmid, a hemagglutinin (HA)-tagged ubiquitin expression plasmid, a -galactosidase expression plasmid, and either control pcDNA3 or an Mdm2 expression plasmid. Cells were harvested as described above, with the inclusion of 2 mM N-ethylmaleimide (Sigma) to prevent deubiquitination. Lysates were precleared with normal mouse serum, followed by immunoprecipitation with either monoclonal antibody DO1 (1:1,000), which immunoprecipitates only full-length p53, or monoclonal antibody 1801 (1:3 hybridoma supernatant), which immunoprecipitates both p53 and p47. Loading volumes of the washed immunoprecipitates were standardized for transfection efficiency according to -galactosidase activity, and samples were analyzed by Western blot with anti-HA-labeled immunoglobulin G-horseradish peroxidase conjugate (Roche). Oligomerization assays. H1299 cells were transfected with the various p53/p47 pCDNA3 expression plasmids, and cell lysates were prepared as described above. Equal amounts of protein standardized for transfection efficiency were treated with 0, 0.01, or 0.1% glutaraldehyde (ICN) for 5 min on ice. Following addition of SDS sample buffer, samples were resolved by SDS-PAGE on a 4 to 15% gradient gel (Bio-Rad). Western blot analysis was performed with monoclonal antibody 1801, which detects both p53 and p47. Membranes were then stripped and reprobed with monoclonal antibody DO1 to detect p53-specific bands.
RESULTS Identification of alternatively spliced p53 transcripts and their products. We initially set out to determine whether the p53 transcripts containing intron 2 sequences identified previously (17) [termed p53(EII) in our study] were present in various human cell lines and could encode p53 proteins. A reverse primer specific to p53 intron 2 was used in a 5⬘ rapid amplification of cDNA ends PCR on HeLa cell mRNA, and the resulting cDNA sequence was compared to the p53 genomic sequence (10) and the human p53 sequence in the National Center for Biotechnology Information database. Figure 1 highlights the differences between the novel intron 2-containing p53 transcript [p53(EII)] and the prototype p53 transcript. Comparison of these p53 transcripts raised the possibility that alternative splicing generates two distinct p53 transcripts: the prototype p53 transcript, in which intron 2 is removed by splicing and encodes p53, and the novel p53 transcript, which has retained intron 2 as a novel exon (shown in blue and referred to as the EII exon). Within the EII exon sequence are three stop codons that are in-frame with the first start codon (M1 in green) in exon 2. The next start codon (M2 in green) downstream from the EII exon in this novel transcript is the methionine codon in exon 4, which also contains a consensus Kozak sequence for initiation of translation and thus potentially encodes an N-terminally deleted p53 isoform termed p47 (Fig. 1). It was important first to confirm that the novel p53(EII) transcript was present in mature p53 mRNA in different human cell lines. Polyribosomes contain cytoplasmic mRNA that is undergoing translation into protein and thus represent mature mRNA. Cytoplasmic polyribosomes were therefore puri-
VOL. 24, 2004
REGULATION OF HUMAN p53 BY ALTERNATIVE SPLICING
7989
FIG. 1. Comparison of prototype p53 mRNA and p53(EII) mRNA. The top p53 gene diagram shows the novel EII exon (shown in blue) in the context of p53 exons 1 through 11. The two start codon methionines (M1 and M2) present in exons 2 and 4, respectively, are shown in green, and the stop codons (T) present in the novel EII exon and exon 11 are indicated in red. The middle diagram compares the p53(EII) mRNA to the prototype p53 mRNA, expressing p47 and p53, respectively, and the primers used to characterize these mRNAs, as detailed within. The bottom diagram shows the corresponding protein translation products derived from the p53(EII) mRNA and the locations of the epitopes for monoclonal antibodies DO1 and 1801.
fied from several human cell lines, and the mRNA derived from the polysomes was analyzed by reverse transcription-PCR and restriction enzyme digestion to specifically identify the EII exon in mature p53 mRNA. Forward primers specific to exon 2 (E2F) or the novel EII exon (3⬘EIIF) were used in combination with a reverse primer specific to exon 5 (5⬘E5R), and the amplified products were analyzed by size and for the presence of an NcoI site present in the EII exon (see Fig. 1 for location of primers and the NcoI site). As shown in Fig. 2A, the p53(EII) transcript was present in polyribosome-associated mRNA from all the human cell lines examined. Likewise, the EII exon forward primer (3⬘EIIF) and either exon 8 (5⬘E8R) or exon 11 (E11R) reverse primers, followed by NcoI digestion, also confirmed the presence of the EII exon in mature p53 mRNA (data not shown). Size analysis of these reverse transcription-PCR products revealed that alternative splicing only involved intron 2 (EII exon). These data confirmed that the EII exon is present in mature p53 mRNA. The same reverse transcription-PCR and NcoI digestion were also performed on mRNA isolated from normal human lymphocytes and compared to several wild-type p53-containing cell lines. As shown in Fig. 2A (right panel), the p53(EII) transcript was present in normal lymphocyte RNA and all the human cell lines examined. Furthermore, this is consistent with the original observation that an EII exon-containing truncated p53 cDNA was isolated from an oligo(dT)-primed cDNA li-
brary made from normal human foreskin fibroblast mRNA (17). Real-time PCR was performed to compare the level of the p53(EII) mRNA relative to the prototype p53 mRNA (Fig. 2B). The threshold cycles (CT) for p53(EII) mRNA and prototype p53 plus p53(EII) mRNA were 37.03 and 21.38, respectively, in normal human lymphocytes, 40.95 and 20.33, respectively, in HeLa, 38.98 and 19.53, respectively, in HT1080, and 38.06 and 18.33, respectively, in MCF-7 cells. These results further confirm the presence of the p53(EII) mRNA and also show that it is much less abundant than the prototype p53 mRNA in these cells. It was next important to compare the proteins derived from the p53(EII) transcript to the prototype p53 transcript. To this end, the cDNAs were inserted into the pCDNA3 expression vector and transfected into p53-null human H1299 and murine 10(1) cells, followed by Western blot analysis. A truncated p53 cDNA deleted in exons 1, 2, and the novel EII exon defined above, in which translation initiation would take place only at the second methionine (the M2 site in exon 4 as shown in Fig. 1), was also included for comparison. Western blot analysis was carried out with monoclonal antibody DO1 recognizing amino acids 20 to 25 just downstream from the first start methionine (M1), and monoclonal antibody 1801, recognizing amino acids 46 to 55 just downstream from the second start methionine
7990
GHOSH ET AL.
MOL. CELL. BIOL.
FIG. 2. Detection of p53(EII) mRNA in human cell lines and normal lymphocytes and analysis of its protein products. (A) (Left panel) Purified polyribosomal RNA from different human cell lines was subjected to reverse transcription-PCR analysis to identify the prototype p53 and p53(EII) mRNAs with the E2F/5⬘E5R and 3⬘EIIF/5⬘E5R primer pairs, respectively (see Fig. 1 for locations of primers). The resulting PCR products were digested with NcoI, which generates an 86-bp fragment from the p53(EII) PCR product but not from the prototype p53 PCR product (see Fig. 1 for location of the NcoI site in the EII exon). (Right panel) The same diagnostic reverse transcription-PCR followed by NcoI digestion as above was performed on total RNA isolated from normal human lymphocytes and different human cell lines. (B) Comparison of prototype p53 and p53(EII) mRNA levels. RNA was extracted from the indicated normal lymphocytes or established cell lines, and cDNA synthesis and kinetic quantitative PCR was carried out. To quantify p53 prototype plus EII exon-containing transcripts, primer set E2F/5⬘E5R was used (beaded lines). To quantify EII exon-containing transcripts, primer set 3⬘EIIF/5⬘E5R was used (solid lines). Data are representative of at least five independent experiments. (C) Western blot (WB) analysis of p53 and p47 in H1299 and 10(1) cells transfected with the indicated expression plasmids (see Fig. 1 for the locations of monoclonal antibody 1801 and DO1 epitopes). pCDNA3 represents the control plasmid-transfected cells. Note that the p53(EII) cDNA expressed both p53 and p47 proteins, as detected with monoclonal antibody 1801. (D) Western blot analysis of p53 and p47 in 10(1) cells transfected with p53(EII) cDNAs isolated from MCF-7 cells (lanes 3 and 4). For comparison, the first lane contains the p53(EII) cDNA derived from HeLa cells, and the second lane contains the p53 cDNA truncated in exons 1, 2, and EII as detailed in the text.
(M2) (see Fig. 1 for locations of M1 and M2 start methionines and monoclonal antibody DO1 and 1801 epitopes). As shown in Fig. 2C, the prototype p53 cDNA expressed the predicted 53-kDa protein containing both the monoclonal antibody DO1 and 1801 epitopes, and the truncated cDNA expressed the N-terminally deleted 47-kDa protein which retained the monoclonal antibody 1801 epitope but lost the monoclonal antibody DO1 epitope, as expected. This confirmed that translation initiation for the truncated 47-kDa protein occurred at M2. Interestingly, transfection of the p53(EII) cDNA resulted in the expression of both the 53-kDa and 47kDa proteins. This observation can be explained by the expression of two different transcripts from the p53(EII) plasmid. One transcript, which retained the EII exon, expresses the 47-kDa protein due to initiation of translation at the second methionine (M2). The other transcript, in which the EII exon was removed by splicing, results in the prototype p53 mRNA expressing full-length p53 protein. This is consistent with the epitope analysis of the p47 protein in cells transfected with the EII exon-containing p53 cDNA, which revealed the presence of the monoclonal antibody 1801 epitope downstream from the second start methionine (M2) but not the monoclonal antibody DO1 epitope downstream from the first start methionine (M1). Reverse transcription-PCR analysis confirmed that both types
of transcripts were present in the p53(EII) cDNA-transfected 10(1) cells (data not shown). This result is also consistent with the data presented in Fig. 2A, confirming the presence of both p53 mRNAs in the various human cells examined. Based on these observations, we reasoned that it would be possible to clone the p53(EII) cDNAs by screening a library made by reverse transcription-PCR with forward and reverse primers specific to exon 1 and exon 11, which border the EII exon. This library was constructed from polysomal RNA derived from MCF-7 cells. Out of a total of 21 p53 cDNA clones screened by diagnostic restriction enzyme digestion, two were shown to contain the EII exon. This confirmed that it was possible to clone p53(EII) cDNAs from human cells without with an EII exon-specific primer and further established the authenticity of this novel p53(EII) transcript. Transfection of the p53(EII) cDNAs derived from MCF-7 cells into 10(1) cells confirmed that it likewise expressed the 47-kDa and 53-kDa proteins (Fig. 2D, lanes 3 and 4). Taken together, the data from Fig. 2A to D demonstrate that EII exon-containing transcripts are present in mature p53 mRNA from transformed and normal human cells, confirming alternative splicing of the p53 gene transcripts, and that expression of this novel p53 transcript results in an N-terminally truncated protein of 47
VOL. 24, 2004
REGULATION OF HUMAN p53 BY ALTERNATIVE SPLICING
kDa, in which initiation of translation occurs at the second methionine in exon 4. Control of p53 activity by p47 derived from alternative splicing. It was previously reported that a large excess of p47 relative to p53 was required to suppress p53 transcriptional activity when cotransfecting separate plasmids expressing p53 and p47 (3), and we have made similar observations (data not shown). However, as shown in Fig. 2, transfection of the p53(EII) cDNA into cells produces comparable or lower levels of p47 relative to p53. It was therefore necessary to determine whether p53 activity was suppressed in cells transfected with the p53(EII) cDNA. Human p53-null H1299 cells and p53-null murine 10(1) cells were cotransfected with control and p53(EII) cDNAs described above, and a p53-responsive p21 promoter luciferase reporter plasmid. As shown in Fig. 3A, p53-mediated transcriptional transactivation was effectively suppressed in cells transfected with the p53(EII) cDNA, where similar or lower expression levels of p47 relative to p53 were obtained by differential splicing. Likewise, p47 alone (expressed from the truncated p53 cDNA with deleted exons 1, 2, and EII) was unable to induce transcription in this assay. The same observations were made with p53(EII) cDNAs derived from HeLa or MCF-7 cells (data not shown). Since p53 is wild type in HeLa and MCF-7 cells, this confirms that the absence of p53 activity in the p53(EII) cDNA-transfected cells was not due to mutations in p53. Sequence analysis of the p53(EII) cDNA confirmed that the full sequence was wild type, eliminating mutation in the p53(EII) cDNA as an explanation for the lack of transcriptional transactivation activity of the p53 protein expressed from this plasmid. This revealed that p47 derived from the p53(EII) cDNA was a potent inhibitor of p53 transcriptional transactivation activity even in the presence of equal or higher levels of p53. It is currently unclear whether p47 is growth suppressive and whether p47 can rescue cells from p53-mediated suppression of cell viability. Although it was reported that p47 induced apoptosis in p53-null H1299 cells and therefore was growth suppressive (26), it has also been argued that p47 was not growth suppressive in p53-null cells and that a 10-fold excess of p47 relative to p53 could counteract the growth suppressive effect of p53 (3). It was therefore necessary in the present study to determine whether p47 was growth suppressive and whether p47 derived from the p53(EII) cDNA expression plasmid could repress p53-mediated growth suppression. To address this issue, we performed a colony formation assay with p53-null H1299 cells, which were transfected with plasmids expressing p53, p47, or the p53(EII) cDNA sequences and placed in G418 to select surviving cells expressing the transfected plasmids. As shown in Fig. 3B, colony formation was dramatically reduced in p53-transfected cells compared with vector- or p47-transfected cells. In comparison, coexpression of p47 with p53 from the p53(EII) cDNA expression plasmid significantly enhanced colony formation relative to the p53-transfected cells. Similar observations were made with p53-null Saos-2 cells. Moreover, we also confirmed that p47 derived from the p53(EII) cDNA impaired the growth suppressive effect of p53 in transient transfections assays of Saos-2 cells transfected with a farsenylated green fluorescent protein expression plasmid, which permits quantitation of live
7991
FIG. 3. p47 inhibition of p53-mediated transcriptional transactivation and growth suppression. (A) (Top panel) p53-mediated transcriptional transactivation as determined by luciferase activity in H1299 and 10(1) cells cotransfected with a p53-responsive p21 promoter luciferase plasmid, the indicated p53 and p47 expression plasmids and a -galactosidase expression plasmid to control for transfection efficiency. The p53-mediated transcriptional transactivation activity was expressed as relative light units (RLU) over pCDNA3, the control plasmid. (Bottom panel) The same cell lysates used for the luciferase assay were analyzed by Western blot with monoclonal antibody 1801 to detect p53 and p47 protein levels. This experiment was repeated four times with the same result. (B) p53-mediated suppression of colony formation is reversed by p47. H1299 cells were transfected with the empty vector (pCIN4) or with the indicated cDNAs as detailed in the text. The vector (pCIN4) used to express p53 and/or p47 contained the neomycin resistance gene, and therefore surviving colonies were selected for 2 weeks by addition of G418 to the culture medium, and drug-resistant colonies were counterstained with Giemsa. (C) Western blot analysis of H1299 cells stably expressing p47 or coexpressing p53 in the presence of p47. Cells as in panel B were pooled and subjected to Western blot analysis with monoclonal antibodies 1801 and DO1 as indicated. Note that no cell lines survived following selection for p53 expression alone with the pCIN4-p53 plasmid.
cells by fluorescence-activated cell sorting analysis (data not shown). It was necessary to confirm whether the cells stably transfected with the p53(EII) cDNA or the p47 expression plasmids and selected with G418 continued to stably express both p53 and p47. As shown in Fig. 3C, cells stably transfected with the p47 expression plasmid contained p47 protein, and cells stably transfected with the p53(EII) cDNA expression plasmid continued to express both p53 and p47. These observations con-
7992
GHOSH ET AL.
MOL. CELL. BIOL.
FIG. 4. Association and oligomerization of p47 and p53. (A) Complexing of p53 with p47 in H1299 cells transfected with the p53(EII) cDNA. Control cells were also transfected with p53 or p47 expression plasmids. (Left panel) Transfected cell lysates were first subjected to immunoprecipitation with monoclonal antibody DO1, specific for p53, and the resultant precipitates were subjected to Western blot analysis with monoclonal antibody 1801 to detect both p53 and the coimmunoprecipitated p47. (Right panel) Western blot analysis with monoclonal antibody 1801 to determine p53 and p47 levels in the whole-cell lysates of the transfected cells used for the immunoprecipitation analysis. (B) Oligomerization analysis of p47 with p53. Cell lysates prepared from H1299 cells transfected with the indicated cDNAs were treated with 0, 0.01, or 0.1% glutaraldehyde for 5 min on ice. Treated lysates were resolved on a 4 to 15% gradient SDS-PAGE gel to differentiate monomers from oligomers. Western blot (WB) analysis was performed with monoclonal antibody 1801 (left panel), and the blot was stripped and reprobed with monoclonal antibody DO1 to detect p53-specific bands (right panel). Note that control p53-338 is a C-terminal mutant of p53 that is unable to oligomerize.
firmed that p53-null tumor cells could tolerate p47 expression or p53 expression in the presence of p47 but were unable to tolerate p53 expression in the absence of p47. Taken together, these results argue that p47 does not suppress cell viability but controls p53-mediated suppression of cell viability. p47 oligomerization with p53. Because p47 effectively impaired p53 activity, it was necessary to examine the mechanism of inhibition. Our initial approach involved characterizing the association between p53 and p47. To this end, we determined whether the majority of p53 was complexed with p47 and whether p47 could disrupt p53 tetramer formation, a characteristic required for p53-mediated transcriptional transactivation. First, to examine the p53-p47 association, we performed immunoprecipitation specifically for full-length p53, followed by Western blot analysis with monoclonal antibody 1801 to detect p53 and p47. This approach would permit an estimation of how much of the p47 was coimmunoprecipitated with p53. We reasoned that if the majority of p53 was not complexed with p47 in H1299 cells transfected with the p53(EII) cDNA, then selective immunoprecipitation of full-length p53, followed by Western blot analysis to detect p53 and coimmunoprecipitated p47, would reveal an excess of p53 relative to p47. However, as shown in Fig. 4A (left panel), there was no excess of p53 relative to p47 in this assay. Western blot analysis of the total cell lysate prior to immunoprecipitation with monoclonal antibody 1801 confirmed that there were similar levels of ex-
pression of p53 and p47 in the transfected cells (Fig. 4A, right panel). This argues that the majority of the p53 was present in a complex with p47 in these transfected cells. To determine whether p47 could interfere with the ability of p53 to form tetramers, which could in turn alter p53 activity, we performed cross-linking experiments to visualize the different oligomeric complexes formed by p53 and p47. Cell lysates were treated with glutaraldehyde to cross-link any dimers and tetramers formed. As a negative control, we used a C-terminal mutant of p53 (p53-338), which was previously shown to be unable to form dimers or tetramers (24). As shown in Fig. 4B, both p53 and p47 expressed alone were able to form homodimers and homotetramers (Fig. 4B, left panel, lanes 3 and 9, respectively). However, a lower concentration of glutaraldehyde permitted visualization of p47 homodimers by Western blot analysis with 1801, while p53 homodimers were only readily detected with the higher concentration of glutaraldehyde (Fig. 4B, left panel, compare lanes 2 and 8). This suggests that p47-p47 interactions are more stable than p53-p53 interactions. A similar stabilization effect of p47 was observed in cells transfected with the p53(EII) cDNA. In these cells, which coexpress p53 and p47, there appeared to be a stronger association between p53-p47 heterodimers (Fig. 4B, left panel, lane 5) than between p53-p53 homodimers (Fig. 4B, left panel, lane 2). In order to determine if the majority of p53 was complexed
VOL. 24, 2004
REGULATION OF HUMAN p53 BY ALTERNATIVE SPLICING
7993
FIG. 5. Control of p53 cell localization by p47 in transiently transfected and stable cell lines. (A) Percentage of cells with p53 and p47 localized predominantly in the nucleus or the cytoplasm in transiently transfected cells. The indicated expression plasmids were transfected into H1299 or 10(1) cells and the cell localization of p53 and p47 was determined by immunofluorescence (IF) microscopy with monoclonal antibody DO1 to detect p53 and monoclonal antibody 1801 to detect p47 in cells expressing only p47. Note that expression of p47 from the p53(EII) cDNA shifted the localization of p53 from the nucleus to the cytoplasm. The data are representative of three independent experiments. (B) Representative immunofluorescence images of 10(1) cells transiently transfected with the indicated expression cDNAs with monoclonal antibody DO1 specific for p53 and monoclonal antibody 1801 to detect p47. Note that, as expected, DO1 did not show any signal when used on p47-transfected cells because p47 lacks the DO1 N-terminal epitope. (C) Localization of p53 and p47 at day 1 and day 7 posttransfection with the indicated plasmids, where p53 was localized with monoclonal antibody DO1 and p47 was localized with monoclonal antibody 1801, as indicated. Note that on day 7, no surviving cells expressing p53 were detected. (D) Localization of p47 in stably transfected H1299 cells expressing p47 alone in untreated control cells and cells exposed to stress by transfecting empty plasmid pCDNA3 or treatment with adriamycin. (E) Localization of p53 in stably transfected H1299 cells coexpressing p53 and p47 in untreated control cells and cells exposed to stress by transfecting empty plasmid pCDNA3 or treatment with adriamycin.
with p47, the Western blot filter was stripped and reprobed with monoclonal antibody DO1 (Fig. 4B, right panel). As monoclonal antibody DO1 detects only p53, the molecular weights of the major bands representing p53-p53 and p53-p47 dimers may be compared. If the majority of p53 was complexed with p47 in cells transfected with the p53(EII) cDNA, the major bands representing p53-p47 heterodimers will be lower molecular weight than bands representing p53-p53 homodimers. By comparing the bands observed in lanes 2 and 5 (Fig. 4B, right panel), the majority of p53-specific bands were downshifted when coexpressed with p47 compared to cells expressing p53 alone. This supports the immunoprecipitation results shown in Fig. 4A, arguing that the majority of p53
expressed is complexed with p47. Note that glutaraldehyde treatment of the control p53-338 mutant failed to cross-link the monomer, as oligomeric forms of the monomeric mutant were not detected. This confirmed that the glutaraldehyde-mediated visualization of oligomerization shown in Fig. 4B is specific to the C-terminal oligomerization domain of p53. Taken together, these data show that in cells transfected with the p53(EII) cDNA, the majority of p53 was complexed with p47 and that p47 appears to enhance p53-p47 oligomerization. Control of p53 cell localization by p47. Since p53 transcriptional transactivation activity requires p53 localization to the nucleus, we next examined whether p47 altered the cell localization of p53 in transfected p53-null H1299 and 10(1) cells.
7994
GHOSH ET AL.
Localization of p53 and p47 was determined by immunofluorescence 24 h after transfection with monoclonal antibodies DO1 and 1801, respectively. As expected, p53 localized predominantly to the nucleus in the transfected H1299 and 10(1) cells (Fig. 5A). Figure 5B depicts the typical immunofluorescence cell localization patterns observed. Surprisingly however, p47 derived from the truncated p53 cDNA (exons 1, 2, and EII deleted) localized predominantly to the cytoplasm in the majority of the cells, as determined with monoclonal antibody 1801. As shown in Fig. 5A and B, when p53 and p47 were coexpressed from the p53(EII) cDNA, p53 detection by monoclonal antibody DO1 followed the same cell localization pattern as p47, localizing predominantly to the cytoplasm. These data demonstrate that following transient transfection, p47 localized predominantly to the cytoplasm and was able to shift the localization of p53 from the nucleus to the cytoplasm. In contrast, p53 remained predominantly nuclear when it was expressed in the absence of p47. The marked difference in cell localization between p47 and p53 was unexpected, and we therefore carried out a time course experiment to determine whether the cytoplasmic localization of p47 and of p53 in the presence of p47 was maintained beyond the 24 h following transient transfection. As shown in Fig. 5C, p53 was localized predominantly in the nucleus 24 h after transfection but became undetectable by day 7 following transfection, consistent with p53 suppression of cell viability. In comparison, p47 was present predominantly in the cytoplasm 24 h after transfection but was relocated to the nucleus by day 7. Likewise, in the presence of p47, p53 was predominantly cytoplasmic 24 h following transfection but, by day 7, had shifted into the nucleus with kinetics similar to that of p47. These data reveal that p47 localized to the cytoplasm following transient transfection and subsequently was able to accumulate in the nucleus and that p53 followed the same cellular localization pattern in the presence of p47. Furthermore, these observations are consistent with p47’s impairing p53-mediated suppression of cell viability, since in the presence of p47, p53-containing cells were evident 7 days following transfection, whereas in the absence of p47, p53-containing cells were not observed. Based on the preceding localization results, we predicted that p53 and p47 would be located predominantly in the nucleus in the stably transfected H1299 cells described above (Fig. 3). Moreover, following a cellular stress, such as DNA transfection, p47 and p53 in the presence of p47 would relocate to the cytoplasm in these cells. To test this possibility, the cellular localization of p47 and p53 in cells stably coexpressing p53 and p47 was determined before and 24 h after inducing cell stress by empty plasmid (pCDNA3) transfection and treatment with adriamycin. As shown in Fig. 5D and E, p47 and p53 (in the cells coexpressing p47) were localized predominantly in the nucleus in resting cells. There was a substantial redistribution of p47 and p53 (in the cells coexpressing p47) from the nucleus to the cytoplasm following cellular stress by DNA transfection or adriamycin treatment. Taken together, cell stress induced by transfection or adriamycin treatment triggered both the relocation of p47 from the nucleus to the cytoplasm and the associated nuclear export of p53 to the cytoplasm in cells coexpressing p53 and p47.
MOL. CELL. BIOL.
FIG. 6. Association and cell localization of p53 and p47 C-terminal oligomerization mutants. (A) Association analysis of p53 and p47 in H1299 cells transfected with the indicated plasmids expressing either wild-type or oligomerization-deficient p53, p53(EII), or p47 cDNAs. Oligomerization mutants p53-338, p53(EII)-338, and p47-338 lack the C terminus (amino acids 338 to 393), which includes the oligomerization domain. (Top panel) Transfected cell lysates were first subjected to immunoprecipitation with monoclonal antibody DO1 specific for p53, and the resultant immunoprecipitates were subjected to Western blot (WB) analysis with monoclonal antibody 1801 to detect both p53 and the coimmunoprecipitated p47. (Bottom panel) Western blot analysis with monoclonal antibody 1801 to determine p53 and p47 levels in the whole cell lysates of the transfected cells used for the immunoprecipitation analysis. (B) Cell localization of wild-type and oligomerization-deficient p53 and p47 proteins in H1299 cells 24 h posttransfection with the indicated plasmids. Cell localization of p53 and p47 was determined by immunofluorescence (IF) microscopy with monoclonal antibody DO1 to detect p53 and monoclonal antibody 1801 to detect p47 in cells expressing only p47. Note that in cells transfected with the p53(EII)-338 cDNA, in which p53 and p47 do not associate due to deletion of the oligomerization domain, p47 is no longer able to shift the localization of p53 to a more cytoplasmic distribution compared to cells transfected with the wild-type p53(EII) cDNA. The data are representative of three independent experiments.
Involvement of the C-terminal oligomerization domain in p47-mediated p53 nuclear export. To determine whether p47mediated p53 nuclear export required a physical association involving the C-terminal tetramerization domain, p47- and p53(EII)-encoding oligomerization mutants were generated as per the control p53-338 oligomerization mutant (24) used in the glutaraldehyde cross-linking experiments shown previously. These oligomerization mutants lack amino acids 338 to 393. As shown in the top panel of Fig. 6A, immunoprecipitation with the p53-specific DO1 antibody, followed by Western blot analysis with 1801, demonstrated that p47 does not associate with p53 in cells transfected with the p53(EII)-338 cDNA (Fig. 6A, top panel, lane 5) which expressed the p53 and p47
VOL. 24, 2004
REGULATION OF HUMAN p53 BY ALTERNATIVE SPLICING
mutants lacking the C-terminal oligomerization domain. In contrast, p47 was coimmunoprecipitated by the DO1 antibody in the control wild-type p53(EII) cDNA transfected cells (Fig. 6A, top panel, lane 2). Note that p47 was expressed at levels similar to that of p53 in both the control wild-type p53(EII) and the p53(EII)-338 oligomerization mutant-transfected cells, as demonstrated in the bottom panel of Fig. 6A (lanes 2 and 5). These results demonstrate that p47 association with p53 requires the C-terminal oligomerization domain. To confirm that a physical association facilitated by the Cterminal oligomerization domain was necessary for p47-mediated nuclear export of p53, localization studies were conducted. H1299 cells were transfected with the different mutant and wild-type expression plasmids, and the localization of p53 and p47 was determined by immunofluorescence 24 h after transfection with monoclonal antibodies DO1 and 1801, respectively. As shown in Fig. 6B, p47 effectively redistributed p53 to the cytoplasm in cells transfected with the wild-type p53(EII) cDNA, as expected. Mutation of the C-terminal oligomerization domain altered the localization of p53, as shown in cells expressing the p53-338 mutant. In these cells, approximately 35% of cells exhibited a predominantly nuclear localization pattern. However, in cells transfected with the p53(EII)-338 cDNA lacking the oligomerization domain, p47 had no effect on p53 localization. This is consistent with the observation that p47 is unable to physically associate with p53 if the C-terminal oligomerization domain is deleted (Fig. 6A) and argues that this oligomerization domain is required for p47-mediated p53 nuclear export. Localization of the p47-338 oligomerization mutant showed a distribution similar to the wild-type p47 protein, in which only 20% of the cells showed a predominantly nuclear localization. Thus, in contrast to the importance of the C-terminal oligomerization domain for wildtype p53 localization to the nucleus, this domain does not appear to be critical for wild-type p47 cytoplasmic localization following transfection. Mono- and polyubiquitination of p47 and p53 in the presence of Mdm2. As described above, it was of considerable interest that p47 was predominantly cytoplasmic and mediated a cytoplasmic localization of p53 following transient transfection. To investigate the mechanism of this nuclear export, we examined the ubiquitination of p47 and p53 in the presence of Mdm2, a cellular ubiquitin ligase. It has recently been shown that monoubiquitination of p53 results in nuclear export and that polyubiquitination of p53 results in nuclear export and proteasome-mediated degradation (12). We therefore examined the ubiquitination of p47 and p53 in both the presence and absence of Mdm2. Mdm2 complexes with the N-terminal region of p53 in the nucleus prior to export into the cytoplasm, where p53 is degraded by the ubiquitin-mediated proteasome pathway (2, 5, 6). To detect ubiquitination of p47 and p53, H1299 cells were cotransfected with a plasmid expressing HAtagged ubiquitin and the various p47 and p53 expression plasmids. Immunoprecipitation of p47 and p53 was followed by Western blot analysis with anti-HA antibodies to detect ubiquitinated p47 and p53 proteins, as detailed in Materials and Methods. As shown in the first two lanes of Fig. 7A, Mdm2 efficiently mediated both monoubiquitination and polyubiquitination of p53. Mdm2-mediated polyubiquitination of p53 was associated
7995
FIG. 7. Mdm2-mediated ubiquitination and degradation of p53 and p47. (A) H1299 cells were cotransfected with the indicated plasmids expressing p53, p53/p47 (p53[EII]), or p47 together with an HA-tagged ubiquitin expression plasmid and either control pCDNA3 plasmid or an Mdm2 expression plasmid as indicated. A -galactosidase expression plasmid was also included to control for transfection efficiency. Cell lysates were prepared 24 h posttransfection and immunoprecipitated with either DO1 or 1801 followed by Western blot analysis with anti-HA antibody to detect ubiquitinated p53 and/or p47. Control lanes included cells coexpressing p53 or p47 with Mdm2 in the absence of HA-tagged ubiquitin (lanes 10 and 11) and cells expressing only HA-tagged ubiquitin (lane 12) revealed no ubiquitinated proteins, as expected. Note that in lane 1, no ubiquitinated p53 species were detected in the absence of Mdm2. Note also the predominance of monoubiquitinated species when p53 was coexpressed with p47 (lanes 3 and 4) or when p47 was expressed alone (lane 5) or in the presence of Mdm2 (lane 6). (B) Western blot analysis (monoclonal antibody 1801) of p53 and p47 protein levels in the whole-cell lysates (5% input) which were used to carry out the ubiquitination analysis shown in panel A. Note that the presence of p47 protects p53 from Mdm2-mediated degradation when comparing lanes 1 and 2 to lanes 3 and 4, and that Mdm2 was unable to mediate p47 degradation.
with efficient Mdm2-mediated p53 degradation (Fig. 7B, lanes 1 and 2). In contrast to p53, p47 was ubiquitinated in the absence of Mdm2 and p47 ubiquitination represented predominantly monoubiquitinated species (Fig. 7A, lane 5). Coexpression of p47 in the presence of Mdm2 did not significantly increase polyubiquitination of p47 (Fig. 7A, lanes 6), which is consistent with the absence of the N-terminal region containing the Mdm2 binding site on p47 and the inability of Mdm2 to mediate the degradation of p47 (Fig. 7B, lanes 5 and 6). Treatment with the proteasome inhibitor MG-132 confirmed that in the absence of Mdm2, p47 was substantially ubiquitinated (Fig. 7A, compare lanes 5 and 9). In contrast, ubiquitination of p53, which was not detected in the absence of transfected Mdm2 (Fig. 7A, lane 1) even in the presence of the proteasome inhibitor, MG-132 (Fig. 7A, lane 7). In the presence of p47,
7996
GHOSH ET AL.
there was also a shift from polyubiquitination to monoubiquitination of p53 either in the presence or absence of Mdm2 (Fig. 7A, lanes 3 and 4) compared to Mdm2-mediated ubiquitination of p53 in the absence of p47 (Fig. 7A, lanes 1 and 2). Interestingly, this is also consistent with the observation that in the presence of p47, there was reduced Mdm2-mediated degradation of p53 (Fig. 7B, lanes 3 and 4). Control lanes in which the HA-tagged ubiquitin expressing plasmid was excluded (lanes 10 and 11) or when control empty plasmid was cotransfected with the HA-tagged ubiquitin expression plasmid (lane 12), confirmed that the bands present in lanes 1 through 9 represent ubiquitinated p47 and p53 proteins. Taken together, the monoubiquitination and cytoplasmic localization of p47, independent of Mdm2 expression, is consistent with the observation that monoubiquitination induces nuclear export (12). Moreover, p47-mediated monoubiquitination of p53 is consistent with the nuclear export of p53 observed in Fig. 5 and protection against Mdm2-mediated degradation shown in Fig. 7B. DISCUSSION Stringent regulation of p53 is essential because of its firm authority over cell cycle, apoptosis, and senescence, properties which make p53 the prototype tumor suppressor gene (1, 7). The present study reveals a novel mechanism of p53 regulation originating through alternative splicing of p53 transcripts, resulting in the expression of an N-terminally truncated version of p53, termed p47, which is able to control p53 activity. We have further demonstrated that p47 strongly associates with p53 through its C-terminal oligomerization domain, is monoubiquitinated in an Mdm2-independent manner, and is able to mediate the export of p53 out of the nucleus. In this manner, p47 also impaired Mdm2-mediated p53 polyubiquitination and degradation. Taken together, these observations reveal that the p47 product from the alternatively spliced p53 gene has potent p53 regulatory activity through its ability to control p53 ubiquitination, degradation, and cell localization. This also represents the first study to reveal the regulation of human p53 activity through an alternatively spliced p53 gene product. It has been reported that the p47 N-terminally deleted isoform of p53 can also arise by use of different sites for translation initiation on the same p53 mRNA (3, 26) and that Mdm2 enhances translation initiation from the second start methionine codon, giving rise to p47 (26). Therefore, p47 can arise through two distinct mechanisms in human cells, including alternative splicing as demonstrated within and initiation of translation at different start sites (3, 26). The observation that p47 can arise by different mechanisms supports the argument that it plays an important physiological role. This is consistent with a recent study with a transgenic mouse model, where it was shown that that overexpression of p47 (mouse p44) resulted in p53-dependent cellular senescence and reduced life span (14). It was suggested that the balance between p44 and p53 was important to control cellular senescence and apoptosis (14). However, earlier studies have also shown that the truncated mouse p44 can be tumorigenic in mice (20, 21), presumably through its ability to impair p53 activity. Additional studies are needed to establish the physiological role of p47 and whether it plays similar roles in murine and human cells.
MOL. CELL. BIOL.
It has also been argued that human p47 can induce apoptosis (26), but a subsequent study reported that p47 was unable to mediate apoptosis and could negatively regulate p53-mediated growth suppression (3). Our observations strongly argue that p47 alone does not induce apoptosis in p53-null H1299 or Saos-2 cells, since it was possible to select for these cells expressing p47 alone but not p53. Moreover, H1299 and Saos-2 cells tolerated p53 expression in the presence of p47, further arguing that p47 controlled p53-mediated growth suppression. These observations do not, however, rule out the possibility that p47 can modulate distinct p53-mediated physiological roles, such as aging (14). Several studies have provided evidence that the export of p53 out of the nucleus is dependent on Mdm2 activity (2, 5, 6), although there is also evidence that p53 can be exported from the nucleus independently of Mdm2 (23). The p47 protein retains the nuclear localization signal in its C-terminal region (22). However, this study revealed that p47 could be exported from the nucleus in an Mdm2-independent manner, since p47 lacks the N-terminal region Mdm2 binding site. Moreover, p47 nuclear export was associated with p47-mediated p53 nuclear export. A recent study revealed that monoubiquitination of p53 resulted in p53 nuclear export and stabilization (12). We observed that p47 was likewise monoubiquitinated and this was associated with export of p47 out of the nucleus in an Mdm2independent manner. Moreover, in the presence of p47, there appeared to be an increase in the monoubiquitination of p53 and an impairment of Mdm2-mediated polyubiquitination of p53, and this was associated with reduced Mdm2-mediated p53 degradation. This suggests that p47-mediated nuclear export of p53 resulted from an increased monoubiquitination of both p47 and p53, and this is consistent with p47-mediated impairment of p53 activity. Additional studies are needed to investigate how monoubiquitinated p53 is further processed and whether it is a substrate for the herpesvirus-associated ubiquitin-specific protease ubiquitin hydrolase (13). In summary, this study demonstrates a novel splicing mechanism of the p53 gene and how this controls p53 cell localization and activity. Various types of tumors, including breast, colon, hepatocellular carcinomas, undifferentiated neuroblastomas, and retinoblastomas, have demonstrated abnormal p53 cytoplasmic localization, and this is associated with tumor metastasis and poor prognosis (reviewed in reference 8). It will be important to examine the role of p47 in such tumors and whether p47 may influence tumor response to chemotherapy. ACKNOWLEDGMENTS We are grateful to D. Bohmann for the HA-ubiquitin expression plasmid and to S. Wing for the gift of N-ethylmaleimide. This work was supported by the National Sciences and Engineering Research Council of Canada (NSERC). G.M. holds a Canadian Institutes of Health Research Senior Investigator Award, and D.S. has been supported by NSERC and FCAR student fellowships. REFERENCES 1. Balint, E., and K. Vousden. 2001. Activation and activities of the p53 tumor suppressor protein. Br. J. Cancer 85:1813–1823. 2. Boyd, S., Tsai, K., and Jacks, T. 2000. An intact HDM2 RING-finger domain is required for nuclear exclusion of p53. Nat. Cell Biol. 2:563–568. 3. Courtois, S., G. Verhaegh, S. North, M. G. Luciani, P. Lassus, U. Hibner, M. Oren, and P. Hainaut. 2002. DeltaN-p53, a natural isoform of p53 lacking the first transactivation domain, counters growth suppression by wild-type p53. Oncogene 21:6722–6728.
VOL. 24, 2004
REGULATION OF HUMAN p53 BY ALTERNATIVE SPLICING
4. Donehower, L. 1996. The p53-deficient mouse: a model for basic and applied cancer studies. Semin. Cancer Biol. 7:269–278. 5. Freedman, D., and A. Levine. 1998. Nuclear export is required for degradation of endogenous p53 by Mdm2 and human papillomavirus E6. Mol. Cell. Biol. 18:7288–7293. 6. Geyer, R., Z. Yu, and C. Maki. 2000. The MDM2 Ring-finger domain is required to promote p53 nuclear export. Nat. Cell Biol. 2:569–573. 7. Hahn, W., and R. Weinberg. 2002. Rules for making human tumor cells. N. Engl. J. Med. 347:1593–1603. 8. Jimenez, G., S. Khan, J. Stommel, and G. Wahl. 1999. p53 regulation by post-translational modification and nuclear retention in response to diverse stresses. Oncogene 18:7656–7665. 9. Kaelin, W. 1999. The p53 gene family. Oncogene 18:7701–7705. 10. Lamb, P., and L. Crawford. 1986. Characterization of the human p53 gene. Mol. Cell. Biol. 5:1379–1385. 11. Levine, A. J. 1997. p53, the cellular gatekeeper for growth and division. Cell 88:323–331. 12. Li, M,. Brooks, C., Wu-Baer, F., D. Chen, R. Baer, and W. Gu. 2003. Monoversus polyubiquitination: differential control of p53 fate by Mdm2. Science 302:1972–1975. 13. Li, M., C. Brooks, N. Kon, and W. Gu. 2004. A dynamic role for HAUSP in the p53-Mdm2 pathway. Mol. Cell 13:879–886. 14. Maier, B., W. Gluba, B. Bernier, T. Turner, K. Mohammad, T. Guise, A., Sutherland, M. Thorner, and H. Scrable. 2003. Modulation of mammalian life span by the short isoform of p53. Genes Dev. 18:306–319. 15. Malkin, D., F. Li, L. Strong, J. Fraumini, C. Nelson, D. Kim, J. Kassel, M. Gryka, F. Bischoff, M. Tainsky, and S. Friend. 1990. Germ line p53 mutations in a familial syndrome of breast cancer, sarcomas, and other neoplasias. Science 250:1233–1238. 16. Matlashewski, G., P. Lamb, D. Pim, J. Peacock, L. Crawford, and S. Benchimol. 1984. Isolation and characterization of a human p53 cDNA clone: expression of the human p53 gene. EMBO J. 3:3257–3262. 17. Matlashewski, G., D. Pim, L. Banks, and L. Crawford. 1987. Alternative splicing of human p53 transcripts. Oncogene Res. 1:79–88.
7997
18. Matlashewski, G. J., S. Tuck, D. Pim, P. Lamb, J. Schneider, and L. V. Crawford. 1987. Primary structure polymorphism at amino acid residue 72 of human p53. Mol. Cell. Biol. 7:961–963. 19. Melino, G., V. DeLaurenzi, and K. Vousden. 2002. p73: Friend or foe in tumorigenesis. Nat. Rev. Cancer 2:1–11. 20. Mowat, M., A. Cheng, A. Kimura, A. Bernstein, and S. Benchimol. 1985. Rearrangement of the cellular p53 gene in erythroleukemia cells transformed by Friend virus. Nature 314:633–636. 21. Rovinski, B., D. Munroe, J. Peacock, M. Mowat, A. Berstein, and S. Benchimol. 1987. Delection of 5⬘-coding sequences of the cellular p53 gene in mouse erythroleukemia: a novel mechanism of oncogene regulation. Mol. Cell. Biol. 7:847–853. 22. Shaulsky, G., N. Goldfinger, Ben-Ze’ev, A., and V. Rotter. 1990. Nuclear accumulation of p53 is mediated by several nuclear localization signals and plays a role in tumorigenesis. Mol. Cell. Biol. 10:6565–6577. 23. Stommel, J., N. Marchenko, G. Jimenez, U. Moll, T. Hope, and G. Wahl. 1999. A leucine-rich nuclear export signal in the p53 tetramerization domain: regulation of subcellular localization and p53 activity by NES masking. EMBO J. 18:1660–1672. 24. Sturzbecher, H. W., R. Brain, C. Addison, K. Rudge, M. Remm, M. Grimaldi, E. Keenan, and J. R. Jenkins. 1992. C-terminal alpha-helix plus basic region motif is the major structural determinant of p53 tetramerization. Oncogene 7:1513–1523. 25. Thomas, M., A. Kalita, S. Labrecque, D. Pim, L. Banks, and G. Matlashewski. 1999. Two polymorphic variants of wild-type p53 differ biochemically and biologically. Mol. Cell. Biol. 19:1092–1101. 26. Yin, Y., M.-G. Luciana, and R. Fahraeus. 2002. p53 stability and activity is regulated by Mdm2-mediated induction of alternative p53 translation products. Nat. Cell Biol. 4:462–467. 27. Zaika, A. I., N. Slade, S. H. Erster, C. Sansome, T. W. Joseph, M. Pearl, E. Chalas, and U. M. Moll. 2002. DeltaNp73, a dominant-negative inhibitor of wild-type p53 and TAp73, is up-regulated in human tumors. J. Exp. Med. 196:765–780.
