The Unique NH2-terminally Deleted (ΔN) Residues, the PXXP Motif, and the PPXY Motif Are Required for the Transcriptional Activity of the ΔN Variant of p63*

p63, a member of the p53 family of transcription factors, is known to be involved in epithelial development. However, its role in tumorigenesis is unclear. Contributing to this uncertainty, the TP63 locus can express multiple gene products from two different promoters. Utilization of the upstream promoter results in expression of the TAp63 variant with an activation domain similar to p53. In contrast, the NH2-terminally deleted (ΔN) p63 variant, transcribed from a cryptic promoter in intron 3, lacks such an activation domain. Thus, the TAp63 and ΔNp63 variants possess a wide ranging ability to up-regulate p53 target genes. Consequentially, the disparity in transactivation potential between p63 variants has given rise to the hypothesis that the ΔNp63 variant can serve as oncoprotein by opposing the activity of the TAp63 variant and p53. However, recent studies have revealed a transcriptional activity for ΔNp63. This study was undertaken to address the transcriptional activity of the ΔNp63 variant. Here, we showed that all NH2-terminally deleted p63 isoforms retain a potential in transactivation and growth suppression. Interestingly, ΔNp63β possesses a remarkable ability to suppress cell proliferation and transactivate target genes, which is consistently higher than that seen with ΔNp63α. In contrast, ΔNp63γ has a weak or undetectable activity dependent upon the cell lines used. We also demonstrate that an intact DNA-binding domain is required for ΔNp63 function. In addition, we found that the novel activation domain for the ΔNp63 variant is composed of the 14 unique ΔN residues along with the adjacent region, including a PXXP motif. Finally, we demonstrated that a PPXY motif shared by ΔNp63α and ΔNp63β is required for optimal transactivation of target gene promoters, suggesting that the PPXY motif is requisite for ΔNp63 function.

p53 functions as a tumor suppressor by transactivating target genes that mediate cell cycle arrest, apoptosis, and other p53-dependent activities. In response to DNA damage, oncogene activation, or other forms of cellular stress, p53 is stabilized by a complex series of post-translational modifications and protein-protein interactions that allow p53 to carry out its tumor suppression activity. Structurally, p53 is organized into several domains, each of which contributes to p53 transcriptional activity. p53 contains two amino-terminal activation domains, AD1 within residues 1-42 and AD2 within residues 43-92, including the proline-rich domain, which enable p53 to form direct associations with transcriptional coactivators. The DNA-binding domain allows for sequence-specific recognition of response elements in p53 target gene promoters. The tetramerization domain directs formation of the p53 tetramer required for DNA binding. Finally, the carboxyl-terminal basic domain has a regulatory function by controlling p53 stability and transcriptional activity (reviewed in Ref. 1).
p63, a member of the p53 family of transcription factors, was identified in 1998 (2,3). Like p53, p63 contains a proline-rich domain, a DNA-binding domain, and a tetramerization domain, all of which share strong sequence similarity with p53 (reviewed in Ref. 4). Accordingly, p63 is able to regulate many p53 target genes. However, due to two transcriptional start sites, the TP63 locus is capable of expressing two major variants, called TA 2 and ⌬N, both of which have multiple isoforms as a result of alternative splicing at the COOH terminus. Transcribed from the upstream promoter, the TA variant contains an acidic, transcriptional activation domain (AD) similar to p53 and is able to transactivate p53 target genes as well as inhibit cell proliferation and induce apoptosis. In contrast, the ⌬Np63 variant, transcribed from the cryptic promoter in intron 3, lacks the activation domain encoded by exons 2 and 3 but gains 14 unique residues in its NH 2 terminus. Alternative splicing of p63 transcripts results in expression of at least three COOH-terminal isoforms, ␣, ␤, and ␥. Thus, the TP63 gene is capable of producing at least six transcripts: p63␣, p63␤, p63␥, ⌬Np63␣, ⌬Np63␤, and ⌬Np63␥ (Fig. 1A).
The discovery of p53 family members, p73 in 1997 (5) and p63 in 1998 (2,3), was first thought to introduce two new tumor suppressors in the fight against cancer. Although enthusiasm was initially tempered by findings that p63 and p73 do not function as classic tumor suppressors, recent evidence suggests that p63 might serve an important role in tumor suppression. Amplification and overexpression of p63 has been linked to increased survival rates in lung cancer patients (6), and loss of p63 expression has been linked to an increase in metastasis in bladder (7)(8)(9) and breast cancers (10). Likewise, fibroblasts derived from p63 knock-out mice were shown to be defective in the p53-mediated apoptotic response (11). In addition, a follow up study has demonstrated that mice heterozygous for p63 developed an increased tumor burden and metastasis rate, which was compounded in mice harboring heterozygous alleles of p53 and/or p73 (12).
Among all p63 isoforms, ⌬Np63␣ is the most predominantly expressed. Our laboratory and others have demonstrated that ⌬Np63␣ retains transcriptional activity and is competent to induce growth suppression (13). When expressed in H1299 cells, ⌬Np63␣ is able to inhibit cell proliferation, induce cell death, and up-regulate GADD45 (growth arrest and DNA damage-45) and p21. Importantly, these activities are dependent on the presence of the NH 2 terminus, since the ⌬⌬Np63␣ construct lacking the first 26 amino acids was nonfunctional. These * This work is supported in part by National Institutes of Health Grant RO1 CA102188.
findings have forced the scientific community to reconsider the role of ⌬Np63␣ in tumor formation and development. Evidence supporting a transcriptional activity for ⌬Np63␣ has been demonstrated by others. For example, ⌬Np63␣ was shown to function both as a positive regulator of p53-responsive promoters and as a negative regulator of maturation-specific genes during Ca 2ϩ -induced keratinocyte differentiation (14). Similarly, small interfering RNA knockdown of ⌬Np63␣ in immortalized mammary epithelial cells revealed that ⌬Np63␣ promotes transcription of PUMA and NOXA but negatively regulates the P21 and P53 promoters (15). Finally, ⌬Np63␣ activates the HSP70 promoter, as well as induces hsp70 protein expression (16).
Whereas a tumor suppressor function has recently been ascribed to p63, previous studies have suggested that p63 might function as an oncogene. For example, ⌬Np63␣ was characterized as transcriptionally incompetent and shown to be a dominant negative to TAp63 and p53 in luciferase reporter assays (3,17). In addition, loricrin promoter-driven expression of ⌬Np63␣ in the epidermis of transgenic mice hindered apoptosis induced by UV-B but did not alter epithelial cell proliferation or stratification (18). Furthermore, chromosomal arm 3q, which contains the TP63 locus, is amplified, and the ⌬Np63␣ transcript is overexpressed in lung cancer and squamous cell carcinomas of the head and neck (19). Likewise, TAp63 isoforms are overexpressed in certain lymphomas (20).
Debate on the function of ⌬Np63␣ will continue until regulation of p63-mediated target gene expression is elucidated. To gain a better understanding of the diverse activities found in the p63 isoforms, we mutated or deleted specific regions in p63. By comparing the transcriptional activity of wild-type and mutant p63 proteins in stable, inducible cell lines and in transient expression studies, we have characterized several functional domains that regulate ⌬Np63 transactivation potential. Our results demonstrated that ⌬Np63␣ and ⌬Np63␤ are capable of inducing target gene expression and inhibiting cell proliferation. We showed that these activities are dependent upon DNA binding, since DNA-binding domain mutants were inactive. In addition, we have defined the activation domain in ⌬Np63, which includes the 14 unique ⌬N residues and the adjacent proline-rich domain. Finally, our results demonstrate that a PPXY motif, present in ⌬Np63␣ and ⌬Np63␤, modulates ⌬Np63 function.

EXPERIMENTAL PROCEDURES
Reagents-HF-Taq polymerase and tetracycline-free fetal bovine serum were obtained from Clontech. The dual luciferase kit was obtained from Promega. The site-directed mutagenesis kit was obtained from Stratagene. Primary anti-Myc antibody was collected from the 9e10.2 hybridoma cell line. The p21(C-19) antibody was obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). The rabbit anti-␤actin antibody and horseradish peroxidase-conjugated secondary antibodies were purchased from Sigma. Unless otherwise indicated, all remaining reagents were purchased from Sigma.
Cell Culture and Stable Transfections-Stable cell lines were generated using the tetracycline-off system as previously described (24). Transfections were carried out using the calcium phosphate precipitation method (25).
Growth Curve Analysis-The overall viability of each cell line was addressed by growth curve analysis, as previously described (26). Briefly, 5 ϫ 10 4 H1299 and 1 ϫ 10 5 MCF-7 cells were plated per 6-cm dish and either induced (without tetracycline) or uninduced (with tetracycline). Adherent cells were counted at 24-h intervals over 5 days in culture using a Coulter cell counter. Medium was changed at day 3 for both control and experimental cell groups.
DNA Histogram Analysis-The assay was carried out using a FACS-Caliber cell sorter (BD Biosciences) as previously described (24). Briefly, 3 ϫ 10 5 cells were plated per 10-cm dish. H1299 and MCF-7 cells were induced or uninduced to express various ⌬Np63 proteins for 3 days. Trypsinized, adherent cells and floating cells were combined, washed with phosphate-buffered saline, pH 7.4 (PBS), and fixed in 100% ethanol. When ready to stain, cells were washed with PBS and resuspended in staining buffer with 100 g/ml RNase A and 50 g/ml propidium iodine (Molecular Probes, Inc., Eugene, OR). RNA was digested, and cells were stained for 30 min at room temperature. The percentage of cells in each phase of the cell cycle (G 1 , S, and G 2 -M) was analyzed using Cell Quest software, with sub-G 1 accumulation being used as a marker for apoptosis.
Trypan Blue Dye Exclusion-H1299 and MCF-7 cells were induced or uninduced to express various ⌬Np63 proteins for 3 days. Trypsinized, adherent cells and floating cells were collected and stained with trypan blue dye for 5 min. Both live (unstained) and dead (stained) cells were counted two times using a hemocytometer. Results are expressed as a percentage of dead cells over total cells counted.
Western Blot Analysis-Cells were plated at a density of 2.5 ϫ 10 6 cells/10-cm dish and incubated for 24 h. Following induction, plates were washed with cold PBS (pH 7.4) and collected in 500 l of 2ϫ SDS sample buffer and heated for 7 min at 95°C. SDS-PAGE and Western blots were carried out as previously described (27). Nitrocellulose membranes were blocked with 5% dry milk in PBS with 0.1% Tween 20 (PBST) for 30 min, incubated in primary antibody overnight at 4°C or for 2 h at room temperature, and washed three times for 10 min in PBST, and then the membrane was incubated in horseradish peroxidase-conjugated secondary antibody for 1 h at room temperature and washed four times in PBST. Blots were developed using Pierce Pico-West reagents, followed by exposure to x-ray film. All antibody dilutions were carried out with 2% dry milk in PBST.
Northern Blot Analysis-Cells were plated at a density of 2.5 ϫ 10 6 cells/10-cm dish and incubated for 24 h. Following induction, plates were washed with cold PBS (pH 7.4) and collected in 1.0 ml of Trizol reagent (Invitrogen). Total RNA was isolated following the manufacturer's instructions. Northern blots were prepared using 10 g of total RNA. The P21, DKK1 (DICKKOPF-1), PIG3 (p53-induced gene-3), GADD45, FDXR, and GAPDH probes were prepared as previously described (27,28). Blots were exposed to x-ray film and quantified by densitometry.
Luciferase Assays-Transient transfections were carried out using the calcium phosphate precipitation method (25). Briefly, p53-null H1299 cells were plated at 5 ϫ 10 4 cells/well in 12-well plates and allowed to recover overnight. Using the calcium phosphate method, cells were co-transfected with 100 ng of wild-type or mutant p63 construct in pcDNA3 and 100 ng of pGL2 reporter vector/well. As an internal control, 5 ng of pRL-CMV, a Renilla luciferase vector (Promega), was also co-transfected per well. 36 h post-transfection, cells were washed with cold PBS and lysed in 150 l of passive lysis buffer. Luciferase activity was measured using the Promega dual luciferase kit and a Turner Designs luminometer. The -fold increase in relative luciferase activity was the product of the luciferase activity induced by pcDNA3 constructs expressing wild-type or mutant p63 divided by that induced by a pcDNA3 empty control vector.

All ⌬Np63 Isoforms Have the Potential to Inhibit Cell Proliferation-
To determine the role of ⌬Np63 isoforms in cell cycle progression and cell death, stable cell lines were created using the tetracycline-off system. H1299, a p53-null lung cancer cell line, was utilized to determine the p53-independent function of ⌬Np63 isoforms. Previously published data from our laboratory demonstrate that tetracycline does not alter H1299 cell proliferation (29). Multiple cell lines were generated, and two representative cell lines for each ⌬Np63 isoform were selected for further study. As shown in Fig. 1B, these cell lines were able to inducibly express ⌬Np63␣, ⌬Np63␤, or ⌬Np63␥. In addition, the protein level of each isoform was comparable (Fig. 1B). To determine the transcriptional activity of each isoform, we monitored the level of p21 WAF1 , a known p53 target gene. Similar to our previous finding (13), ⌬Np63␣ was able to induce a modest increase in p21 expression. Whereas p21 expression was weakly induced by ⌬Np63␥, ⌬Np63␤ was highly active in inducing this p53 target gene. To determine the ability of ⌬Np63 isoforms to inhibit cell proliferation, growth rate analysis was performed ( Fig. 1C). Consistent with our previous report, ⌬Np63␣ reduced the proliferation potential of H1299 cells by day 3 (13). Furthermore, we showed that ⌬Np63␤ and ⌬Np63␥ were capable of inhibiting cell proliferation, with ⌬Np63␤ displaying the greatest ability to induce growth suppression, followed by ⌬Np63␣ and then ⌬Np63␥ (Fig. 1C).
To characterize the decreased proliferative potential of H1299 cells seen in our growth curve assay, trypan blue dye exclusion studies were performed as an indicator of cell viability (Fig. 1D). Following induction of ⌬Np63␤, we observed a significant increase in the number of trypan blue-positive, nonviable cells over that seen for uninduced controls. Likewise, we observed a moderate increase in nonviable cells upon induction of ⌬Np63␣. In contrast, there was no difference in the viability of control cells and those induced to express ⌬Np63␥. In addition, DNA histogram analysis was performed to determine the cell cycle profile of these cells in the presence or absence of each ⌬Np63 isoform (Fig. 1E). We found that sub-G 1 accumulation, associated with apoptotic cell death, was increased upon induction of ⌬Np63␣ and ⌬Np63␤. We also found that ⌬Np63␣ and ⌬Np63␤ induced cell cycle arrest, primarily in G 1 phase. Upon induction of ⌬Np63␥, cell cycle arrest in G 1 phase, but not significant apoptosis, was detected in both ⌬Np63␥expressing cell lines. Taken together, these results demonstrate that ⌬Np63␣ and ⌬Np63␤ can induce cell death.
To rule out the possibility that these ⌬Np63 activities are cell typespecific, we used MCF-7 breast adenocarcinoma cells to produce additional stable cell lines. Unpublished results from our laboratory have demonstrated that tetracycline does not alter MCF-7 cell proliferation. Multiple cell lines inducibly expressing ⌬Np63␣, ⌬Np63␤, or ⌬Np63␥ were generated, and two representative clones were chosen for further studies. Western blot analysis was performed to determine the level of protein expressed ( Fig. 2A). We found that these ⌬Np63 isoforms were expressed at a comparable level and that p21 was up-regulated by ⌬Np63␤ but not by ⌬Np63␣ and ⌬Np63␥. Similar to our results observed in H1299 cells, growth curve analysis demonstrated that ⌬Np63␣ and ⌬Np63␤ inhibited cell proliferation, whereas ⌬Np63␥ was almost inert (Fig. 2B). It is noteworthy that p21 induction was low in ⌬Np63␣, suggesting that a target gene other than p21 was responsible for mediating ⌬Np63␣-dependent growth suppression.
To further demonstrate the transcriptional activity of p63 isoforms, Northern blot analysis was performed on MCF-7 cell lines induced to express TAp63␣, ⌬Np63␣, TAp63␥, and ⌬Np63␥ for 1 and 7 days (Fig.  3). As expected, TAp63␥ possessed the greatest potential to up-regulate PIG3, a proapoptotic gene, and P21, a cyclin-dependent kinase inhibi-tor, whereas TAp63␣ had reduced transactivation potential due to its inhibitory COOH terminus. In addition, ⌬Np63␣ was able to induce expression of GADD45 in MCF-7 cells similar to our previously published results seen in H1299 cells (13). Interestingly, ⌬Np63␥ was able to induce DKK1, an inhibitor of Wnt signaling. These data effectively demonstrate the differential transcriptional activity of the p63 isoforms and highlight the need for additional studies involving isoform-specific gene activation.
DNA Binding Is Essential for p63 Function-Most hotspot mutations found in TP53 are located in the DNA-binding domain. These mutations, characterized as either contact site or conformational mutants, are often associated with loss of tumor suppressor function. Similar mutations, associated with EEC syndrome (reviewed in Ref. 30  . p21 was detected with anti-p21 antibody (C19). Actin was detected with an anti-actin polyclonal antibody. C, growth curve analysis was performed over a 5-day period as described under "Experimental Procedures" (numerals in the graph titles refer to clonal isolate number). D, cell viability was determined by counting live (unstained) and dead (stained) cells following staining with trypan blue dye. E, DNA content was quantified following propidium iodide staining of fixed cells that were cultured for 3 days in the presence (Ϫp63) or absence (ϩp63) of tetracycline. 175 in p53 (Fig. 4A). Multiple cell lines were generated, and two representative clones were used for further study. Western blot analysis was performed and revealed that the level of ⌬Np63␣(R149W) was equal to or greater than wild-type ⌬Np63␣ (Fig. 4B). Consistent with our earlier published reports (13), Northern blot analysis demonstrated that ⌬Np63␣ was transcriptionally active in MCF-7 cells and induced expression of the FDXR, GADD45, and P21 genes (Fig. 4C). However, contrary to wild-type ⌬Np63␣, the transcriptional activity of the R149W mutant was abrogated (Fig. 4C). Consequentially, ⌬Np63␣(R149W) were unable to suppress cell proliferation (Fig. 4D). These results suggest that an intact DNA-binding domain is required for ⌬Np63␣ function.
N6H Mutation Modestly Compromises ⌬Np63␣ Function-A patient diagnosed with ADULT syndrome was found to have a mutation in exon 3Ј of the TP63 gene resulting in the substitution of an asparagine with a histidine at codon 6 in all ⌬Np63 isoforms (31). To date, the N6H mutation is the only naturally occurring mutation discovered in the unique NH 2 -terminal region of ⌬Np63, but its effect on ⌬Np63 function is unclear. To address this, multiple stable cell lines expressing ⌬Np63␣(N6H) were generated in MCF-7 cells. Western blot analysis demonstrated that the level of ⌬Np63␣(N6H) expressed was comparable with that of the wild-type (Fig. 5A). To examine the effect of this mutation on p63 function, growth curve analysis was performed. Fol-lowing induction of ⌬Np63␣(N6H), cell proliferation was inhibited (Fig.  5B), but to a lesser extent compared with wild-type ⌬Np63␣ in MCF-7 cells (Fig. 2B).
The Proline-rich Domain Contributes to ⌬Np63 Transcriptional Activity-Src homology 3 domains have been shown to function as protein-protein interaction motifs in various signaling pathways (reviewed in Ref. 32). PXXP motifs, identified in the proline-rich regions of p53 and p73, serve as ligand binding sites for Src homology 3 domaincontaining proteins and regulate the transcriptional activity of these p53 family members (reviewed in Ref. 33). To explore the effect of the proline-rich domain on ⌬Np63 function, we mutated the PXXP motif at residues 69 -72 to PXXA. Since ⌬Np63␣ exhibited a strong growth suppression in MCF-7 cells (Fig. 2C), cell lines expressing ⌬Np63␣(P72A) were generated, and two representative clones were selected for further study. As demonstrated in Fig. 6A, Western blot analysis revealed that the mutant ⌬Np63␣(P72A) protein was expressed at levels equal to or greater than that of the wild-type. Growth curve analysis was performed and showed that, in comparison with ⌬Np63␣, ⌬Np63␣(P72A) had a diminished ability to inhibit cell proliferation (Fig. 6B).
⌬Np63 Isoforms Differ in Their Ability to Activate the p53 Target Gene Promoters-To better understand the contribution of specific amino acid sequences to p63 function, we characterized the ability of wild-type and mutant p63 constructs to activate p53-responsive promoters. As determined by Western blot analysis, ⌬Np63␤ and ⌬Np63␥ were expressed at equivalent levels, whereas the level of ⌬Np63␣ was slightly higher following transient transfection in H1299 (Fig. 7A). To measure the transactivation potential of the wild-type ⌬Np63 isoform, each ⌬Np63-expressing vector was co-transfected with a luciferase reporter construct into the p53-null H1299 cells (Fig. 7B). We showed that ⌬Np63␣, ⌬Np63␤, and ⌬Np63␥ differentially regulated the P21, GADD45, and FDXR reporters (Fig. 7B). Similar to the results obtained in the stable cell line studies, luciferase reporter assays demonstrated that ⌬Np63␤ was highly potent, whereas ⌬Np63␥ was limited, to transactivate P21, GADD45, and FDXR promoters. However, transiently expressed ⌬Np63␣ was unable to transactivate these promoters.
⌬Np63␤ DNA-binding Domain Mutants Exhibit Loss of Function-Among ⌬Np63 isoforms, ⌬Np63␤ is the most potent in transactivation FIGURE 3. Unique sequences in the NH 2 terminus and COOH terminus of p63 isoforms differentially regulate their ability to induce endogenous target genes. Northern blots were prepared using 10 g of RNA purified from uninduced (Ϫp63) and induced cells (ϩp63) cultured for 1 day or 7 days post-induction. Blots were analyzed with cDNA probes for P21, DKK1, PIG3, and GADD45, as indicated to the right of the blot. GAPDH was utilized as a loading control. Lanes 1-16 are labeled at the bottom. and in inducing cell cycle arrest and apoptosis. Thus, the remaining studies to analyze the domain and function of ⌬Np63 were carried out with the ⌬Np63␤ isoform. Since mutations in the DNA-binding domain of p63 have been identified in several developmental disorders, we utilized these mutants to demonstrate a requirement for DNA binding for p63 function (reviewed in Refs. 30 and 34). For example, codons corresponding to arginine 149 and arginine 249 were found to be mutated in EEC syndrome, whereas arginine 243 was found to be mutated in ADULT syndrome (Fig. 4A). These mutants were generated and used for luciferase assay. As revealed in Fig. 8A, both wild-type and mutant ⌬Np63␤ constructs were expressed at comparable levels. To determine their transactivation potential, these DNA-binding domain mutants were analyzed for their ability to activate the P21, GADD45, and FDXR reporter constructs (Fig. 8B). Consistent with data obtained from ⌬Np63␣(R149W)-expressing cell lines (Fig. 4), ⌬Np63␤(R149W) showed a complete lack of transcriptional activity for all reporter constructs (Fig. 8B). Likewise, ⌬Np63␤(R243Q) was inactive. Surprisingly, we found that ⌬Np63␤(R249W) was able to activate the P21, GADD45, and FDXR reporters, albeit less than the wild-type protein (Fig. 8B). These data suggest that the DNA-binding domain is necessary for ⌬Np63 activity and that some critical residues in the p53 DNA-binding domain are conserved in p63, whereas others are not.
The 14 Unique ⌬N Residues and Adjacent Proline-rich Domain Constitute the Activation Domain for ⌬Np63-Previously, we showed that a deletion construct lacking the first 26 residues in the NH 2 terminus of ⌬Np63␣, termed ⌬⌬Np63␣, lost the ability to induce P21 WAF1 and GADD45 (13). In order to extend this observation, additional NH 2terminal mutants were constructed to uncover the residues required for transactivation by ⌬Np63 (Fig. 9A). Following transfection into H1299 cells, Western blot analysis revealed that all NH 2 -terminal constructs were expressed at similar levels (Fig. 9B). The P21 reporter construct was used to measure the transactivation potential for these mutant constructs. As shown in Fig. 9C, we found that the luciferase activity induced by ⌬Np63␤ (⌬2-14) was substantially reduced. In addition, subsequent deletion of additional NH 2 -terminal residues, as in ⌬Np63␤ (⌬2-19), led to a further reduction in its transactivation potential. Consistent with our previous finding that ⌬⌬Np63␣ is inert (13), deletion of the first 26 residues of ⌬Np63␤ completely abolished its transcriptional activity. Western blot analysis of Myc-tagged p63 and actin was performed on extracts collected 24 h postinduction from ⌬Np63␣(R149W) control cells (Ϫp63) and induced cells (ϩp63). C, Northern blots were prepared using 10 g of RNA purified from uninduced (Ϫp63) and induced cells (ϩp63) cultured for 24 h postinduction. Blots were analyzed with cDNA probes for FDXR, GADD45, and P21 genes as indicated to the right of the blot. GAPDH was utilized as a loading control. D, growth curve analysis was performed over a 5-day period as described under "Experimental Procedures" (numerals in graph titles refer to clonal isolate number).  To further analyze the ⌬N activation domain, we generated point mutations at residues deemed likely to contribute to ⌬Np63 function. First, we generated ⌬Np63␤ carrying the N6H mutation associated with ADULT syndrome (31). Although the N6H mutation consistently decreased the ability of ⌬Np63␤ to transactivate the P21 reporter, the data failed to rise to the level of statistical significance (Fig. 9C). However, when taken together with the ability of the ⌬Np63␣(N6H) mutation to attenuate growth suppression (Fig. 4B), it is likely that the N6H mutation can alter p63 function, thus accounting for the ADULT syndrome phenotype (31). Mutation of hydrophobic residues in the NH 2terminal activation domain (AD1) has been shown to reduce the tran-scriptional activity of p53 (35). Therefore, the presence of two analogous hydrophobic residues in the NH 2 terminus of ⌬Np63 led us to construct the ⌬Np63␤(L22Q,L23S) mutant. As measured in a luciferase reporter gene assay, the (L22Q,L23S) mutation resulted in a small, but reproducible, decrease in promoter activation (Fig. 9C).
Several studies have demonstrated that the proline-rich domain in p53 and p73 contributes to their transcriptional activity (reviewed in Ref. 33). Likewise, we showed that the PXXP motif in ⌬Np63␣ is necessary for inducing growth suppression (Fig. 6B). To determine the effect of the PXXP motif on the transcriptional activity of ⌬Np63␤, we constructed ⌬Np63␤(P72A), in which the PXXP motif was mutated to PXXA, and ⌬Np63␤(⌬69 -72), in which the PXXP motif was deleted (Fig. 9A). Although ⌬Np63␤(P72A) and ⌬Np63␤(⌬69 -72) were expressed at a higher level than wild-type ⌬Np63␤ (Fig. 9B), we found that the ability of these mutants to activate the P21 promoter was greatly reduced compared with wild-type ⌬Np63␤ (Fig. 9C). To further analyze the function of the PXXP motif, cell lines expressing the ⌬Np63␤(P72A) FIGURE 7. ⌬Np63 isoforms differ in their ability to activate p53 target gene promoters following transient transfection. A, the levels of ⌬Np63 isoforms and actin were assayed by Western blot analysis of extracts collected 24 h following transient transfection utilizing anti-Myc antibody for Myc-tagged p63 and anti-actin antibody. B, transcriptional activity of ⌬Np63␣, ⌬Np63␤, and ⌬Np63␥ was determined by luciferase assay following cotransfection with pGL2-p21, pGL2-GADD45, or pGL2-FDXR. The -fold increase in relative luciferase activity was calculated using an empty pcDNA3 vector as a control. FIGURE 8. DNA-binding domain mutants greatly reduce the transactivation potential of ⌬Np63␤. A, the levels of wild type, DNA-binding domain mutants, and actin were assayed by Western blot analysis of extracts collected 24 h following transient transfection utilizing anti-Myc antibody for Myc-tagged p63 and anti-actin antibody. B, transcriptional activity of DNA-binding domain mutants for ⌬Np63␤ was determined by luciferase assay following cotransfection with pGL2-p21, pGL2-GADD45, or pGL2-FDXR. mutant were created in the H1299 background. Multiple clones were collected, and two clones were selected for further study. Western blot analysis revealed that the mutant was expressed at levels comparable with that of wild-type ⌬Np63␤ (Fig. 9D). To determine the transcriptional activity of ⌬Np63␤(P72A), endogenous p21 protein levels were monitored. Consistent with the transient transfection assay, p21 was not up-regulated in cells induced to express ⌬Np63␤(P72A) (Fig. 9D,  p21 panel). Coincidentally, a commonly used p63 monoclonal antibody (p63 Ab-1; Oncogene) only partially recognized the P72A mutant and did not recognize the ⌬69 -72 mutant. Taken together, these data suggest that the 14 unique residues in the NH 2 terminus and adjacent residues, including the PXXP motif, are required for transactivation by all ⌬Np63 isoforms.
A Region Present in ⌬Np63␣ and ⌬Np63␤, but Not ⌬Np63␥, Contributes to Their Transcriptional Activity-The data above clearly indicate that ⌬Np63␤ is the most potent transcription factor among the ⌬Np63 isoforms, suggesting that a positive regulatory domain exists in the COOH terminus of ⌬Np63␤. Therefore, five COOH-terminal truncation mutants were generated to map this domain in ⌬Np63␤ (Fig. 10A). Western blot analysis showed that when transfected into H1299 cells, these constructs were expressed at equivalent levels (Fig. 10B). Of importance, the transcriptional activity of ⌬Np63(⌬457-ter), which lacks the unique residues associated with the ␤ isoform, and ⌬Np63(⌬356-ter), which lacks the unique residues associated with the ␥ isoform, was not altered compared with wild-type ⌬Np63␤ and ⌬Np63␥, respectively (Fig. 10C). Therefore, the five unique ␤ and 39 unique ␥ residues present in their COOH termini do not appear to have any effect on ⌬Np63 activity. However, the ability of ⌬Np63(⌬431-ter), ⌬Np63(⌬406-ter), and ⌬Np63(⌬381-ter) to activate the P21 promoter was diminished compared with that of ⌬Np63␤. In sum, these data suggest that residues 431-455 are responsible for the differential transcriptional activities exhibited by the three ⌬Np63 isoforms.
A COOH-terminal PPXY Motif Confers Increased Transactivation Potential to ⌬Np63␤-PPXY motifs serve as ligands for WW domaincontaining proteins. Recently, studies showed that p73, another p53 family member, contains a PPXY motif that can be regulated by WW domain-containing proteins (36 -39). Thus, we searched for such a motif between residues 431 and 455 and found a PPXY motif present in the ␣ and ␤ isoforms (Fig. 11A). To explore the role of the PPXY motif in FIGURE 9. The 14 unique ⌬N residues and adjacent region form an activation domain for ⌬Np63 isoforms. A, schematic diagram of NH 2 -terminal deletion and point mutation constructs for ⌬Np63␤. B, levels of p63 and actin were assayed by Western blot analysis using anti-Myc antibody for Myc-tagged p63 and anti-actin antibody. C, transcriptional activity of ⌬Np63␤ carrying mutations in NH 2 -terminal residues was determined by luciferase assay following cotransfection with pGL2-p21. The -fold increase in relative luciferase activity was calculated using an empty pcDNA3 vector as a control. D, the levels of p63, p21, and actin expressed in wild-type ⌬Np63␤ and ⌬Np63␤(P72A) were assayed by Western blot analysis. p63 function, three mutants, which disrupt the PPXY motif, were generated (Fig. 11B). Western blot showed that both wild-type and mutant ⌬Np63␤ constructs were expressed at equivalent levels (Fig. 11C). Interestingly, these mutants exhibited a reduced ability to activate the P21 promoter (Fig. 11D). Nevertheless, the PPXY mutants of ⌬Np63␤ were still more potent than ⌬Np63␥ to activate the P21 promoter (Fig.  11D). These data suggest that the PPXY motif and additional amino acids between residues 431 and 455 are capable of enhancing the transcriptional activity of ⌬Np63␤.

DISCUSSION
The role of ⌬Np63 in cancer is highly debatable. Previous studies have suggested that ⌬Np63␣ is dominant negative toward p53 and TAp63, since it is capable of inhibiting these proteins in reporter assays (3). However, we and others have demonstrated a role for ⌬Np63␣ in the induction of gene expression (13)(14)(15)(16). In addition, some studies have suggested that ⌬Np63␣ is a potential oncoprotein, because it is overexpressed in lung cancer and squamous cell carcinomas of the head and neck (19), whereas others have observed a correlation between loss of p63 expression and an increase in tumorigenesis and metastasis (9,12). To clarify the ambiguous role of p63 in cancer, we further characterized the ⌬Np63 isoforms. We found that all three ⌬Np63 isoforms, although lacking an activation domain similar to p53 or TAp63, retain the ability to induce target genes and inhibit cell proliferation. Of significance, the ability of ⌬Np63␤ to induce target gene expression is the strongest among ⌬Np63 isoforms, followed by ⌬Np63␣ and then ⌬Np63␥.
Having demonstrated a transcriptional activity for the ⌬N variants, we searched for the activation domain. Previous work in our laboratory demonstrated that ⌬Np63␣ requires residues 1-26 to be transcriptionally active (13). Here, we found that the first 26 residues in ⌬Np63␤ were also required for transactivation of P21. Interestingly, deletion of the 14 unique ⌬N residues (⌬1-14) or of additional residues reaching into the proline-rich domain (⌬1-19) diminished, but did not abolish, the activity of ⌬Np63␤. Of interest, we observed that an intact PXXP motif is required for ⌬Np63␣ to suppress cell proliferation and for ⌬Np63␤ to transactivate the P21 gene. Thus, the 14 residues unique to the ⌬N variant and the adjacent region, including the PXXP motif, comprise an activation domain for the ⌬N variant.
Our data indicate that a COOH-terminal proline-rich domain within residues 431-455, including a PPXY motif found at residues 446 -449, is to some extent responsible for the differential biochemical and biological activities associated with the ⌬Np63 isoforms. Supporting our hypothesis, deletion of the COOH-terminal proline-rich domain or mutation of the PPXY motif revealed that this domain is required for optimal transactivation of the P21 promoter. Additionally, ⌬Np63␣ and ⌬Np63␤, which contain the COOH-terminal proline-rich domain, are effective in suppressing cell proliferation. In contrast, ⌬Np63␥, which lacks the COOH-terminal proline-rich domain, possesses a much weaker potential in growth suppression. Of relevance to these latter results, fusion of TP63 exons 11 or 12 to the DNA-binding domain of GAL4 revealed a weak transcriptional activity associated with the COOH-terminal proline-rich domain (40). These findings suggest that the COOH-terminal proline-rich domain augments the transcriptional activity of the ⌬N variant. However, it is unlikely that this COOHterminal region constitutes an independent activation domain, since deletion of the NH 2 -terminal activation domain makes ⌬Np63 transcriptionally inactive. Thus, alternative splicing confers an enhanced transactivation potential to ⌬Np63␣ and ⌬Np63␤, which is not extended to ⌬Np63␥.
The question remains as to how the COOH-terminal PPXY motif modulates the transcriptional activity of ⌬Np63. Instead of functioning as an activation domain, we hypothesize that the PPXY motif within the COOH-terminal proline-rich domain is likely to alter p63 function by promoting interaction with WW domain-containing proteins. Several studies have demonstrated that WW domain-containing proteins can alter the function of transcription factors. For example, Strano et al. (41) observed that yes-associated protein binds to a PPXY motif in TAp73␣ and enhances p73-mediated Bax promoter activation. In addition, WW-containing ubiquitin ligases, including Itch and NEDL2, can interact with the PPXY motif in p73 and regulate p73 protein stability (38,39). Since mutation of the PPXY motif in ⌬Np63␤ reduced its ability to transactivate the P21 reporter, it is likely that WW domain-containing proteins modulate ⌬Np63␤ function in a manner similar to that seen with p73.
We have established that following inducible expression in stably transfected cells, ⌬Np63␣ is able to induce growth suppression and transactivate target genes. In contrast, following transient transfection, we and others demonstrated that ⌬Np63␣ is unable to transactivate known p53 target gene reporters (3,17) but can transactivate the HSP70 reporter, a non-p53 target gene (16). In addition, several recent studies have described a positive and negative role for ⌬Np63␣ in the transcriptional regulation of endogenous target genes (14,15). Given these obser- FIGURE 11. Point mutation or deletion of the PPXY motif attenuates the ability of ⌬Np63␤ to transactivate the p21 promoter. A, alignment of p63 and p73 COOHterminal residues including a conserved PPXY motif. B, schematic presentation of the ⌬Np63␤ COOH terminus proline-rich domain. The PPXY motif is disruped by point mutation in ⌬Np63␤(Y449A) and ⌬Np63␤(Y449D) or by deletion in ⌬Np63␤(⌬445-449). Dashes represent deleted residues. C, Western blot analysis of p63 and actin in H1299 cells transfected with wild-type ⌬Np63␤ and PPXY mutants. D, transcriptional activity of ⌬Np63␤ and mutants carrying mutations in the PPXY motif was determined by luciferase reporter assay following cotransfection with pGL2-p21.
vations, the transcriptional activity of ⌬Np63␣ is likely to require complex regulation found at the level of chromatin DNA, which is not extended to the naked DNA utilized in most reporter assays. Furthermore, it is apparent that the effect of the ␣ COOH terminus on transcription is context-specific, since some genes are induced whereas others are repressed in the presence of the ␣ COOH terminus. However, it is still unclear how the ␣ COOH terminus modulates p63 function. Previous studies have demonstrated that the transcriptional activity of TAp63␣ is inhibited by its COOH terminus (40,42). One hypothesis suggests that transactivation by TAp63␣ is inhibited through an interaction between the ␣ COOH terminus and the TA variant activation domain (42). Importantly, our data indicate that ⌬Np63␣ is less active than ⌬Np63␤ to induce growth suppression, suggesting that the ␣ COOH terminus can also attenuate the activity of the ⌬N variant ( Figs.  1 and 2). Thus, the currently accepted model cannot be applied to ⌬Np63␣, since the ⌬N variant does not encode the residues required for this inhibitory interaction. Therefore, at least for the ⌬N variant, the inhibitory effect of the ␣ COOH terminus must function through an unknown mechanism. These observations point to a complex regulation of ⌬Np63 transcriptional activity, which may account for the differential activities attributed to ⌬Np63␣ (12)(13)(14)(15)19).
In conclusion, this study demonstrates a transcriptional activity for all three ⌬Np63 isoforms and defines the activation domain for the ⌬N variant to include the 14 unique ⌬N residues and the adjacent region, including the proline-rich domain. We found that the DNA-binding domain is necessary for ⌬Np63 activity and that some critical residues in the p53 DNA-binding domain are conserved in p63, whereas others are not. Interestingly, our results suggest that regulation of and by the PPXY motif is likely to confer isoform-specific function to the ⌬N variant. Furthermore, it is probable that regulation of proteins interacting with the PXXP and PPXY motifs accounts for the paradoxical functions attributed to p63 in cancer.