SCFβTrCP1 Activates and Ubiquitylates TAp63γ*

p63 is a member of the p53 tumor suppressor family that is critical for epithelial differentiation and also has an important role in cancer progression. Currently, the molecular mechanisms governing regulation of p63 function remain largely unclear. This study identifies a unique E3 ubiquitin ligase for p63, SCFβTrCP1. SCFβTrCP1 is able to bind p63γ isoforms, with a higher affinity for the TAp63γ isoform. Strikingly, co-expression of TAp63γ and βTrCP1 leads to the stabilization of TAp63γ. This stabilization of TAp63γ leads to up-regulation of p21 at the mRNA and protein level by increased binding of TAp63γ at the p21 promoter. The up-regulation of p21 causes a subsequent increase in G1 phase cell cycle arrest. Last, SCFβTrCP1 is able to ubiquitylate TAp63γ, and this ubiquitylation, as well as the increased activity of TAp63γ, is ablated with the expression of a ubiquitin-deficient mutant of βTrCP1 (ΔFβTrCP1). Therefore, our study reveals that SCFβTrCP1 is an E3 ligase that activates p63 through ubiquitylation.

ther, the C-terminal regions of these isoforms are alternatively spliced. These variants are termed ␣, ␤, and ␥ (1, 3). The p63␣ isoforms are the longest and most commonly expressed in adult tissue (2,3), whereas the p63␥ isoforms are the shortest and the most potent transcriptional activators of the six isoforms (3,4). Because of its strong transactivation and its role in the stress response, TAp63␥ is considered the p63 isoform "most like" p53.
The biological role of p63 is more complex than that of a classical tumor suppressor. p63 knock-out mice have a severe defect in epithelial stratification and fail to form associated tissues, such as teeth, hair, and mammary glands (2,5). Also, it is debated whether p63 might act as a tumor suppressor or an oncogene (1). Recent data suggest that the ⌬Np63 isoforms, when overexpressed, may be tumorigenic, whereas the TA isoforms may play a more important role in tumor suppression (1). Several papers have suggested a reciprocal relationship between ⌬Np63␣ and TAp63␥, since, under a variety of cellular stimuli, the levels of one are often high when those of the other are low, and vice versa (1). What is clear is that a tightly controlled balance of p63 isoform levels in the cell is critical to homeostasis.
This tight balance is maintained through both direct binding of p53 family member transcription sites and heterodimerization among the p53 family members. Although commanding a unique set of target genes, p63 also shares several target genes with the other p53 family members, such as p21 (6 -8). In a p53-deficient background, TAp63␥ may rescue the growthinhibitory function of p53 by activating some of these common target genes and causing growth arrest in response to various forms of stress (6 -8). The elucidation of both common and isoform-specific signaling pathways will be critical to our understanding of the role of p63 in development and oncogenesis.
In an effort to understand one of the biochemical mechanisms that regulate p63, we searched for a p63-specific regulator that could modulate p63 ubiquitylation and, therefore, its activity. A search of the literature revealed that the inhibitor of IB kinase ␣ knock-out mouse displayed an epidermal phenotype very similar to that of the p63 knock-out mouse (9). The inhibitor of IB kinase kinase complex works in conjunction with the E3 ubiquitin ligase complex SCF ␤TrCP1 to degrade the IB and allow subsequent up-regulation of NFB (10). ␤TrCP1 (␤-transducin repeats-containing protein 1, Fwd1, etc.) is one of the many substrate recognition components of the SCF complex (10). Humans have two ␤TrCP proteins, which act in homo-or heterodimers as part of the SCF complex: ␤TrCP1 and ␤TrCP2. When exogenously expressed, ␤TrCP1 is nuclear, whereas ␤TrCP2 is cytoplasmic (10). ␤TrCP has two major protein motifs: an F-box motif that connects ␤TrCP to the rest of the SCF complex though its interaction with Skp1 and a WD domain, which binds the many substrates of ␤TrCP (10). ␤TrCP proteins can recognize a canonical DSX 2-3 S motif, where the Ser residues are usually phosphorylated (10). However, several noncanonical sequences and regions have also been found. These regions usually depend on a cluster of highly charged or phosphorylated residues and are found in proteins, such as Emi1 and Cdc25a (10 -12).
Therefore, it is very plausible that SCF ␤TrCP1 may also regulate the p63 pathway. Our study, as described here, tests this idea. We have chosen to focus on the p63␥ isoforms, because they are the most potent regulators of transcription and may be important in modulating the activity of the other family members during cell growth. Our data demonstrate that SCF ␤TrCP1 binds TAp63␥ and, to a lesser degree, ⌬Np63␥. This interaction takes place at sites on both the N and the C termini of TAp63␥. Further, this binding leads to subsequent stabilization of TAp63␥, leading to increased activation of p21, because of increased binding of TAp63␥ to the p21 promoter. In turn, the up-regulation of p21 leads to an increase in G 1 phase cells. Further, ␤TrCP1 is able to ubiquitylate TAp63␥ in vitro and in cells, demonstrating that although it regulates the activity of TAp63␥, ␤TrCP1 also has a dual role as an E3 ligase to TAp63␥. This E3 ligase activity is critical to the increased activation of TAp63␥ by ␤TrCP1, since a ␤TrCP1 mutant that can bind but not ubiquitylate TAp63␥ (⌬F␤TrCP1) is unable to mediate this activity. Thus, this study provides insight into a novel ubiquitindependent activation mechanism for TAp63␥.
Transient Transfection/siRNA Transfection-Transient transfection used Transfectin (Bio-Rad) reagent according to the manufacturer's directions at ϳ70% confluence for all cell lines. For each experiment, a 1:2 g ratio of pEGFP-TAp63␥ to pCDNA3-FLAG␤TrCP1 vector or a 1:2.5 g ratio of pEGFP-TAp63␥ to pCDNA3-FLAG⌬F␤TrCP1 vector was used and scaled to the appropriate size of the plate (for ChIP, 0.2 g of pEGFP-TAp63␥: 2 g of pCDNA3-FLAG␤TrCP1 or 2.5 g of pCDNA3-FLAG⌬F␤TrCP1 was used per 10-cm plate). siRNA oligonucleotide infection was performed using cells at ϳ50% confluence. For each experiment, a ␤TrCP1/2 siRNA oligonucleotide pool (Santa Cruz Biotechnology) or scramble siRNA oligonucleotide mix was added to each plate at a final concentration of 100 nM. The cells were harvested 48 h post-transfection.
Western Blot Analysis-Clarified whole cell lysates were loaded directly onto an SDS gel and probed with antibodies, as noted in the figure legends, according to the previous protocol (7,16). Band density was calculated using Adobe Photoshop or Optiquant software and normalized to tubulin. For Western blot analysis (WB) indicating protein levels after co-expression, cells were harvested 48 h post-transfection.
Co-immunoprecipitation-GFP-TAp63␥, GFP-⌬Np63␥, and FLAG-␤TrCP1 vectors were co-expressed in HEK293 or H1299 cells, according to protocols above and the figure legends. Clarified whole cell lysate (concentration indicated per figure) was prebound for 2 h using 2 g of antibody, monoclonal IgG, or polyclonal IgG at 4°C, and then 30 l of protein G-agarose bead slurry (Santa Cruz Biotechnology) was added, and the samples continued to rotate at 4°C overnight. The samples were then washed once with general lysis buffer (50 mM Tris-HCl, pH 7.5, 5 mM EDTA, 150 mM NaCl, 0.5% Nonidet P-40), once with SNNTE buffer (50 mM Tris-HCl, pH 7.4, 5 mM EDTA, 500 mM NaCl, 1% Nonidet P-40, 50% sucrose), and a second time with lysis buffer. The samples were then prepared using SDS sample buffer (50 mM Tris-Cl, pH 6.8, 100 mM dithiothreitol, 2% SDS, 0.1% bromphenol blue, 10% glycerol) and analyzed by WB according to the above protocols. For endogenous assays, the co-IP was performed using the noted endoge-nous antibodies under the same conditions as above with 800 g of protein lysate harvested from keratinocytes.
Stability Assay-Cells were plated onto 60-mm culture dishes and transfected, as noted above, for regular transfection or siRNA depletion. 48 h post-transfection, cells were treated with 100 g/ml cycloheximide and harvested at the indicated time points before preparation for WB.
Semiquantitative and Real Time PCR-For RT-PCR, cells were plated according to the protocols above. If transfection was used, cells were harvested 24 h post-transfection. The RNA was isolated using TRIZOL reagent (Invitrogen), according to the manufacturer's directions. cDNA was produced, and RT-PCR was performed using our previous protocol and the same primers for p63, p21, ␤-actin, and bax, as noted previously (7,16). ␤TrCP1 primers were as follows: 5Ј-ATTGTGCCCAAG-CAGCGGAAAC-3Ј and 3Ј-TGTTCTCAGCGATGTGGTC-CAG-5Ј. Endogenous mouse TAp63 primers were kindly provided by Dr. M. Koster in the laboratory of Dr. D. Roop. Image J software was used to estimate band density. qPCR was performed with 2 l of cDNA as described previously using the listed primers against p21 and glyceraldehyde-3-phosphate dehydrogenase (18). Levels were normalized to input and glyceraldehyde-3-phosphate dehydrogenase and then graphed using Microsoft Excel. Data were analyzed for statistical significance using Student's t test.
Flow Cytometry-H1299 cells were transfected with vectors as indicated. After 1.5 days, the cells were treated with 150 ng/ml nocodazole to arrest them in G 2 , fixed, permeabilized, and stained with propidium iodide (PI), as noted previously (7,16). Cells were analyzed for DNA content with a BD Biosciences FACScan flow cytometer. Data were analyzed using a multicycle software program using a polynomial S-phase algorithm.
Chromatin Immunoprecipitation-H1299 cells were transfected according to the protocol above. 24 h post-transfection, the cells were fixed for ChIP and processed according to the previous protocol (18), except that the IP was carried out with 2 g of ChIP grade Pab4A4 (Abcam) or 2 g of monoclonal IgG (Sigma). 2 l of the purification elutant was loaded onto a 96-well plate, and qPCR was performed as noted above using primers against regions of the p21 promoter or gene taken from a previous publication (19). Each sample was normalized to input using Microsoft Excel.
In Vitro Ubiquitylation Assay-The SCF ␤TrCP1 -mediated in vitro ubiquitylation assay was performed similarly to previous assays (20), with some modifications. The p63 in vitro ubiquitylation assay was carried out in 40 l of a reaction mixture containing 50 mM Tris-HCl (pH 8.0), 5 mM MgCl 2 , 0.5 mM dithiothreitol, 2 mM NaF, 3 M okadaic acid, and the noted concentrations of His-TAp63␥, 10 l of SCF ␤TrCP1 , 400 ng of UBA-1 (Boston Biochem), 400 ng of UbcH3 (Boston Biochem), 1000 ng of HA-ubiquitin, 5 mM ATP (Amersham Biosciences), and 1.5 mM ATPS (Fisher ICN). The mixture was incubated at 37°C for 1.5 h and analyzed afterward by SDS-PAGE followed by WB, as noted above.
In Vivo Ubiquitylation Assay-10-cm plates were transfected as noted above. 24 h post-transfection, the cells were treated with 10 M MG132 (a protease inhibitor) for 6 h. The cells were harvested, and the pellet was split in half. Half underwent standard WB sample preparation (above), and half was prepared using our previous protocol (21).

RESULTS
␤TrCP1 Expression Stabilizes TAp63␥-To determine if ␤TrCP1 could act as a regulator for TAp63␥, we wished to determine if ␤TrCP1 would have an effect on the stability of TAp63␥, since ␤TrCP1 is a well known regulator of protein stability. To answer this question, we used primary MEFs at passage 2 derived from ␤TrCP1 ϩ/Ϫ and ␤TrCP1 Ϫ/Ϫ littermates (13) and examined the endogenous levels of TAp63␥ mRNA and protein. Strikingly, there was a significantly lower amount of TAp63␥ protein in the ␤TrCP1 Ϫ/Ϫ MEFs compared with the ␤TrCP1 ϩ/Ϫ MEFs (ϳ30%), although the TAp63␤ protein level is not affected (Fig. 1A, top). Since the p63 antibody is pan-specific, O3C keratinocytes were used as positive control for the p63 isoforms (23,24). Further, when the mRNA levels were examined using semiquantitative RT-PCR with TA-specific p63 primers, little change was observed (less than 0.5-fold) in TAp63 between the ␤TrCP1 ϩ/Ϫ and the ␤TrCP1 Ϫ/Ϫ MEFs (Fig. 1A, bottom). Since the protein level of TAp63␤ is minor compared with the level of TAp63␥, by WB, this RT-PCR measurement is representative of the TAp63␥ RNA levels. Next, to determine if this observation was specific to ␤TrCP1 expression, we repeated this experiment in H1299 cells, a cell line that is p53 null and has undetectable endogenous expression of p63 and p73 by WB. In fact, in the presence of increased ␤TrCP1, the protein level of TAp63␥ increased significantly above the level of TAp63␥ alone (Fig. 1B, lane 4 versus lane 2). Therefore, TAp63␥ protein levels are also increased by a subsequent increase of ␤TrCP1. These results suggest that ␤TrCP1 must mediate the endogenous protein stability of TAp63␥ through a post-translational mechanism.
In order to verify these intriguing observations, we measured the half-life of exogenous TAp63␥ in the presence ␤TrCP1. Again, to our surprise, although ␤TrCP1 has been shown to be involved in the degradation of its substrates, our data demonstrate that it stabilizes TAp63␥. As shown in Fig. 1C, although the exogenous GFP-TAp63␥ alone has a half-life of ϳ2-3 h, comparable with others' published data (25), the GFP-TAp63␥ co-expressed with the FLAG-␤TrCP1 was highly stabilized over 6 h. We also confirmed this result using a labeled pulsechase (data not shown).
Further, to examine the effect of ␤TrCP on endogenous p63 stability, we used a human keratinocyte line with several p63 isoforms and ␤TrCP1 (HaCaT). The cells were treated with siRNA oligonucleotides against ␤TrCP1/2 (since ␤TrCP2 often rescues ␤TrCP1 if knocked down alone). An initial experiment was performed to determine the efficiency of knockdown at both 24 and 48 h post-knockdown to find the optimal time to start the degradation assay. Knockdown was confirmed, at 0.47fold versus the scramble control at 48 h (Fig. 1D). This degree of reduction is biologically significant, since several groups have shown that ␤TrCP1 is affected by both haploinsufficiency and dominant-negative mutants in vivo (10), compared with the control, by WB (Fig. 1D). Therefore, for our half-life, time 0 for the cycloheximide exposure in Fig. 1E is 48 h after oligonucleo-␤TrCP1 Activation and Stabilization of p63 tide transfection. As predicted from the data above, the half-life of p63 in the ␤TrCP1/2 knockdown cells was shorter (about 1.5 h) versus the control (about 3.5 h) (Fig. 1E). The ␤TrCP1/2 siRNA treatment also reduced the level of endogenous steadystate p63 (data not shown), supporting the result in Fig. 1A. Therefore, ␤TrCP1 co-expression stabilizes TAp63␥ exogenously and endogenously.

B.
A. GFP-TAp63␥, GFP-⌬Np63␥, and FLAG-␤TrCP1 were expressed in HEK293 cells. 448 g of clarified whole cell lysate co-immunoprecipitated per sample, with 15.6% of the prebound lysate loaded as input. WB was performed using the indicated antibodies. B, TAp63␥ binds ␤TrCP1 in vitro. GST-␤TrCP1 was bound to GSH-agarose bead. 60 ng of purified His-TAp63␥ was bound to bead, 10% input, and binding samples were analyzed by WB. C, TAp63␥ and TAp63␣ bind ␤TrCP1 endogenously. HEKn-E6/E7 cells were harvested, and 850 g of total lysate was immunoprecipitated using the noted antibodies (left). The panel on the right shows the IgG control taken from the same gel, same exposure.  ␤TrCP1 Activation and Stabilization of p63 weaker interaction shown between ⌬NP63␥ and ␤TrCP1 in Fig. 2A. TAp63␥ Activation of p21, but Not Bax, Is Increased by ␤TrCP1 Expression-Since TAp63␥ was stabilized by ␤TrCP1, and they interacted directly, we next asked whether ␤TrCP1 could affect TAp63␥ function as a transcription factor. Two target genes of TAp63␥ are p21 and bax (1,4,8), so we examined the effect of TAp63␥ and ␤TrCP1 co-expression on the p63 transcriptional targets p21 and bax. Consistent with our previous reports (7), TAp63␥ alone activated p21 at both the mRNA and protein levels significantly over the vector alone control and also had some effect on Bax (Fig. 4, A and B, lane 2). However, when TAp63␥ was co-expressed with ␤TrCP1, p21 mRNA levels increased 55% (Fig. 4A, lane 4 and graph), and protein levels increased ϳ50% over the TAp63␥ control (Fig. 4B, lane  4). In contrast, Bax expression did not increase over the TAp63␥ control at either the mRNA or protein level (Fig. 4, A and B, lane 4). Additionally, the TAp63␥ mRNA expression levels do not change in the presence of ␤TrCP1, again suggesting that the up-regulation of the protein level is occurring through a post-translational mechanism, supporting our endogenous result in Fig. 1A. Therefore, the co-expression of TAp63␥ with ␤TrCP1 causes apparent up-regulation of p21, but not bax, over the level of either TAp63␥ or ␤TrCP1 alone.
TAp63␥ Binding at the p21 Promoter Increases in the Presence of ␤TrCP1 in Cells-Since ␤TrCP1 stabilized TAp63␥, and ␤TrCP1 also augmented TAp63␥ activation of p21 levels, we next asked if ␤TrCP1 could enhance the presence of TAp63␥ at the p21 promoter. The regions on the promoter that were used for analysis began at amino acids Ϫ2283, Ϫ20, and ϩ7878 and spanned 50 bp downstream in each case. These regions are called P1, P2, and P3, respectively (Fig. 5A). The P1 primers include a high affinity site for the p53 family members, whereas the P3 primers are a negative control. Last, sequence analysis of the p21 promoter revealed both CATG sequences and CGTG sequences that may be favorable for p63 binding near the P2 primer set (4). Thus, we also tested this region for TAp63␥ binding. Consistent with our earlier results, TAp63␥ alone resulted in a 20-fold increase of promoter binding over the vector alone control, whereas co-expression of TAp63␥ with ␤TrCP1 resulted in an increase of over 50-fold of TAp63␥ bound to the p21 promoter (Fig. 5B). The P3 primers in the gene were negative. Interestingly, P2 primers also could detect  . TAp63␥ activation of p21, but not Bax, is increased by ␤TrCP1 expression. A, TAp63␥ activation of p21, but not Bax, is increased by ␤TrCP1 expression at the mRNA level. RT-PCR for p21, bax, ␤TrCP1, p63, and ␤-actin was performed in H1299 cells as shown (right). qPCR was also performed, and p21 was normalized to glyceraldehyde-3-phosphate dehydrogenase (right) Representative data are shown. *, p Ͻ 0.0002. B, TAp63␥ activation of p21, but not Bax, is increased by ␤TrCP1 expression at the protein level. 70 g of whole cell lysate was analyzed by WB.
A.  an active site of TAp63␥ recruitment on the p21 promoter (Fig.  5B), demonstrating the existence of a novel site for p63 binding at the p21 promoter. Therefore, the increased stability of TAp63␥ in the presence of ␤TrCP1 leads to increased TAp63␥ bound to the promoter, using both a canonical p53 family binding region and a novel p63 binding region, accounting for the activation of p21 noted above. TAp63␥ Activation of Growth Arrest Is Increased by ␤TrCP1 Expression-Since our data clearly pointed to a role of ␤TrCP1 in TAp63␥ expression of p21, and p21 is an established regulator of cell cycle arrest, we asked whether the up-regulation of p21 protein would translate to a functional role in G 1 cell cycle arrest. Consistent with our previous data, although TAp63␥ alone increases the proportion of cells in G 1 (26%), the cells co-expressing TAp63␥ and ␤TrCP1 showed a significant increase (42%) (Fig. 6). Therefore, the increased activity of p21 upon TAp63␥ co-expression with ␤TrCP1 translates to an equally striking up-regulation of G 1 cell cycle arrest.
p63␥ Is Ubiquitylated by SCF ␤TrCP1 -Our data above identify ␤TrCP1 as a regulator of TAp63␥, which acts by increasing TAp63␥ through a post-translational mechanism. Since SCF ␤TrCP1 is an active member of the SCF complex, we next examined whether SCF ␤TrCP1 could ubiquitylate TAp63␥. First, we wished to examine the general ubiquitylation of p63 in H1299 cells. Using a Ni 2ϩ -nitrilotriacetic acid binding assay, GFP-TAp63␥, Myc-TAp63␣, and Myc-⌬Np63␥ showed a shift to higher weight moieties in the presence of ubiquitin (Fig. 7A,  left, lanes 4, 6, and 8), but not when transfected alone, demonstrating that all of the isoforms in this case are ubiquitylated. Interestingly, the ⌬N isoform, although expressed at equivalent levels, has less ubiquitylation than the TA isoforms, as shown by the much lighter laddering. Therefore, several isoforms of p63 are ubiquitylated, with the TA isoforms ubiquitylated to a greater degree, suggesting that ubiquitin may be one way to account for the differences in p63 isoform activity in the cell. Next, to ask if the ubiquitylation we observed was ␤TrCP1-dependent, we performed an in vitro ubiquitylation assay using a purified SCF complex, p63, E1, E2, and ubiquitin. In fact, in the presence of all of the ubiquitylation components, TAp63␥ shifted to several ubiquitylated forms of the protein, which ␤TrCP1 Activation and Stabilization of p63 were absent in the control lanes (Fig. 7B, lane 7). We then tested whether FLAG-␤TrCP1 expression would increase the ubiquitylation of TAp63␥ in H1299 cells. In this assay, we used FLAG-␤TrCP1 or the F-box mutant of ␤TrCP1, FLAG-⌬F␤TrCP1. ⌬F␤TrCP1 is a dominant negative truncation mutant of ␤TrCP1 that lacks the F-box. Therefore, the ⌬F␤TrCP1 protein can bind the substrates of ␤TrCP1, but those substrates are not presented to the SCF complex for ubiquitylation (data not shown) (10,26). Consequently, the level of generally ubiquitylated TAp63␥ precipitated in cells (Fig. 7C, lane 3) increased specifically in the presence of FLAG-␤TrCP1 (lane 5) and decreased in the presence of the FLAG-⌬F␤TrCP1 (lane 7). Further, as an additional control to ensure that our result was not due to the function of another SCF complex, we also performed experiments co-expressing TAp63␥ with another SCF substrate recognition component, such as Skp2, and these experiments were negative (data not shown). Thus, this finding demonstrates that SCF ␤TrCP1 could act as a specific E3 ubiquitin ligase for TAp63␥ and that the ubiquitylation of TAp63␥ requires the presence of the rest of the SCF complex. Therefore, while also acting functionally as a regulator of TAp63␥, ␤TrCP1 acts as an E3 ubiquitin ligase for TAp63␥. p63␥ Activation by SCF ␤TrCP1 Is Mediated by Its Ubiquitylation-Since we found that ␤TrCP1 could both activate and ubiquitylate TAp63␥, we next wished to determine if the activation was related to the ubiquitin modification. Therefore, we repeated some of the activation assays using the FLAG-⌬F␤TrCP1 mutant. First, we examined TAp63␥ half-life in H1299 cells in a similar manner to Fig. 1C, with the addition of the ⌬F␤TrCP1 mutant. Where co-expression of GFP-TAp63␥ and FLAG-␤TrCP1 stabilized the TAp63␥ (Fig. 8A, lanes  7-12), co-expression of GFP-TAp63␥ with FLAG-⌬F␤TrCP1 at the same expression ratio as in Fig. 7C did not stabilize the TAp63␥ when compared with the GFP-TAp63␥ alone (lanes [13][14][15][16][17][18]. In order to observe if this loss of TAp63␥ stabilization also correlated to a functional decrease, we continued by examining the level of p21 expression by qPCR. Again, when GFP-TAp63␥ and the FLAG-⌬F␤TrCP1 were co-expressed, there was no significantly increased expression of p21 compared with the p21 induction of GFP-TAp63␥ alone, whereas p21 expression is significantly increased again in the presence of both GFP-TAp63␥ and FLAG-␤TrCP1 (Fig. 8B). Therefore, we wanted to see if this loss of activation was also true at the level of promoter binding. Again, we used H1299 cells and performed ChIP. As shown in Fig. 8C, although the GFP-TAp63␥ has the expected increase of binding to the p21 P1 and P2 promoter regions (top and bottom, respectively), and GFP-TAp63␥ and FLAG-␤TrCP1 co-expression leads to a significant increase in promoter binding, the level of GFP-TAp63␥ bound to the p21 promoter in the presence of FLAG-⌬F␤TrCP1 significantly decreases compared with GFP-TAp63␥ alone on both the P1 and P2 regions. The P3 results were similarly negative, as in Fig.  5 (data not shown). Concomitantly, these data demonstrate that the increased stabilization and activation of TAp63␥ by ␤TrCP1 are direct results of the ubiquitylation itself, since the ⌬F␤TrCP1 mutant is still able to bind TAp63␥ but not ubiquitylate it. A, ⌬F␤TrCP1 co-expression does not stabilize TAp63␥. H1299 cells were transfected with GFP-TAp63␥, FLAG-␤TrCP1, and FLAG-⌬F␤TrCP1. 48 h later, they were treated with cycloheximide (CHX) and harvested over the indicated time course. B, ⌬F␤TrCP1 co-expression does not activate p21. qPCR was performed on p21, as in Fig. 4. Representative data are shown. *, p ϭ 0.022 increase from TAp63␥ alone control; **, p Ͻ 0.003 decrease from TAp63␥ ϩ ␤TrCP1 sample. C, ⌬F␤TrCP1 expression decreases the amount of TAp63␥ bound to the p21 promoter. H1299 cells were transfected and subjected to ChIP using the indicated ChIP antibody/qPCR primer combinations. Representative data are shown; the top panel shows data from the P1 primer set, and the bottom panel shows data from the P2 primer set on the same experiment. For the P1 graph, *, p ϭ 0.006 increase from TAp63␥ alone control; **, p Ͻ 0.002 decrease from TAp63␥ ϩ ␤TrCP1 sample. For the P2 graph, *, p ϭ 0.005 increase from TAp63␥ alone control; **, p ϭ 0.013 decrease from TAp63␥ ϩ ␤TrCP1 sample.

DISCUSSION
In this study, we provide evidence for a novel activation pathway between TAp63␥ and ␤TrCP1, beginning with the observation that endogenous and exogenous steady-state levels and the half-life of TAp63␥ increase dramatically in the presence of ␤TrCP1. ␤TrCP1 binds to TAp63␥ directly and also to ⌬NP63␥ to a lesser degree, consistent with our data mapping ␤TrCP1 binding to the N and C termini of TAp63␥. This interaction is also present endogenously. The result of this interaction is the subsequent stabilization of TAp63␥ and increased activation of p21 because of increased binding of TAp63␥ to the p21 promoter, in turn leading to an increase in G 1 phase cells. Further, ␤TrCP1 is able to ubiquitylate TAp63␥ in vitro and in cells, and this ubiquitylation is needed for the functional activation of TAp63␥.
The literature on various E3 ligases in signaling, in terms not only of degradation but also of activation, is increasing at an incredible rate. Although our report is the first to demonstrate a direct role for SCF ␤TrCP1 in stabilization and activation of its substrate, ␤TrCP1 is involved in activation pathways, such as its well established role in aiding in the activation of the NF-B pathway by degrading IB␣ (27). Additionally, the association between transcriptional activators and E3 ligases is also well trodden experimental ground. In fact, another E3 ligase that forms a complex with the SCF core is Skp2, which co-activates the oncogene c-myc (28,29). Our data clearly demonstrate that TAp63␥ is stabilized exogenously at subsaturating levels of both TAp63␥ and ␤TrCP1 and destabilized endogenously by knockdown of ␤TrCP1 and ␤TrCP2 (Fig. 1, C-E). Our endogenous data in primary MEFs shows that the TAp63␥ protein level is markedly reduced with the loss of ␤TrCP1 (Fig. 1A). Consistent with this result, increased levels of ␤TrCP1 raise the steady-state levels of TAp63␥ (Figs. 1B and 4B), again while the mRNA levels remain largely even (Figs. 1A and 4A). Taken together, these data demonstrate that the stabilization and activation of TAp63␥ by SCF ␤TrCP1 are present endogenously.
An intriguing observation is the clear effect of the ubiquitin modification itself on TAp63␥ stability. Our data show that the p63 isoforms have a distinctly differing degree of ubiquitylation, demonstrating that ubiquitylation could be used as one mechanism to distinguish specific activity of each p63 isoform (Fig. 7A). Further, TAp63␥ is ubiquitylated in the presence of ␤TrCP1, but not ⌬F␤TrCP1 (Fig. 7, B and C), suggesting that this binding includes the entire SCF complex. Our data clearly show that significantly more TAp63␥ is recruited to the p21 promoter in the presence of ␤TrCP1 (Figs. 5B and 8C) after its stabilization, accounting for the significant activation of p21 at both the mRNA and protein level (Figs. 4, A and B, and 8B). However, strong evidence in Fig. 8 shows that in each case when the ubiquitylation activity is lost through expression of the ⌬F␤TrCP1 mutant, which can bind the substrates of ␤TrCP1 but not carry them to the SCF complex for ubiquitylation, TAp63␥ stability (Fig. 8A), its up-regulation of p21 (Fig. 8B), and its binding to the p21 promoter (Fig. 8C) are lost as well. Adding further detail to our model are the facts that the stability of TAp63␥ is initiated off chromatin, since our IP data are taken from a whole cell lysate that would not contain chromatin-associated proteins, and that this signaling event most likely occurs in the nucleus, since ␤TrCP1 and TAp63␥ both are nuclear proteins and are therefore aptly placed for activation (7,10). Although ubiquitin is traditionally thought of as a degradation signal, the list of proteins that use ubiquitin as an activating stimulus is growing. Some ubiquitin linkages have been linked to activation, and a recent study shows that forked polyubiquitin chains created by some E2s, including UbcH5c, may also be a signal used for a function other than proteasomal degradation (27,30). Further, p53 is targeted for ubiquitylation by E4F1, and this modification stabilizes it for a p21-specific activation cascade very similar to what we observe with p63 (31). It is also unsurprising that p63 is targeted by an SCF complex, since p53 is also recognized by a cullin-containing complex during viral infection (22). The characterization of ␤TrCP1 ubiquitylation of TAp63␥ would be an interesting future study, especially with the conservation of this stabilizing role of ubiquitin among two of the three p53 family members (1). For now, what is clear in light of our data and earlier studies is that the relationship among p63 DNA binding, subsequent transactivation activity, and stability is most likely the result of subtle modifications and interactions tailored to each isoform.
Specific isoform interaction would be provided by binding affinity and presence of ␤TrCP1 in the cell. ␤TrCP1 binds TAp63␥ and ⌬Np63␥ isoforms in vitro and in cells at both the N-and C-terminal segments of the protein (Figs. 2 and 3), regions known to be important for moderating stability among the p53 family members. It is quite intriguing that ␤TrCP1 makes both N-and C-terminal contacts with varying interaction at those regions, because the difference in binding strength could easily provide a mechanism for the modulation of isoform-specific activity. Other than conferring a selective TA isoform-specific advantage in binding to and subsequent activation by ␤TrCP1, this dual binding site may also serve three distinct purposes. First, the placement of the N-terminal site may occlude the binding of other substrates, allowing for selective activation of specific downstream targets involved in growth arrest, but not apoptosis, as our later mRNA/protein activation and cell cycle profile suggests (Figs. 4 and 6). In this respect, the lack of binding in the case of endogenous TAp63␤ may be explained in that other interacting proteins may have a higher affinity for TAp63␤ at those sites (Fig. 2C). Second, it could also prevent export to the cytoplasm, in the case of MDM2 occlusion, or prevent immediate degradation by ligases, such as Itch. Last, it is also possible that ␤TrCP1 binding facilitates the placement of other post-translational modifications on p63. Further, in each of these cases, not only the ␤TrCP1 affinity for the individual p63 isoforms, but also ␤TrCP1 affinity for its other substrates would play a critical role in its regulation of p63 under our model, since SCF ␤TrCP1 protein levels have been demonstrated as a limiting factor in endogenous signaling (10,27). Our data show that ␤TrCP1 also binds TAp63␣ and ⌬NP63␣; this binding could provide another level of regulation through isoform competition. Likewise, future studies will probably elucidate the effect of the specific balance of p63 isoforms, as well as the other p53 family members, on this pathway in vivo, since these studies focus largely on TAp63␥ and our binding and ubiquitylation data suggest that differences in ␤TrCP1 Activation and Stabilization of p63 binding and ubiquitylation would be important for the regulation of p63 signaling through this mechanism. Therefore, most likely a fine balance exists between the availability of both ␤TrCP1 and p63 impacting the ability of ␤TrCP1 to act as a regulator of p63.
Thus, in light of our data, we propose a model ( Fig. 9) in which SCF ␤TrCP1 associates with and stabilizes TAp63␥ in the nucleus through the covalent addition of ubiquitin. The increased load of ubiquitylated TAp63␥ at the p21 promoter then leads to an increased level of transcript and G 1 cell cycle arrest. It will be an interesting future question to explore what upstream signal might begin this cascade endogenously and how this modification would interact with others as more regulators of p63 are found. Clearly, discerning more about such subtle regulations of the p63 isoforms will lead to a greater understanding of the role of p63 in differentiation and oncogenesis.  TAp63␥ binds SCF ␤TrCP1 (top). SCF ␤TrCP1 then ubiquitylates TAp63␥ on unknown sites, stabilizing it; thus, more TAp63␥ binds the endogenous p21 promoter (bottom). p21 transcription is increased due to the increased binding of TAp63␥ to the promoter.