TAp73 Is a Downstream Target of p53 in Controlling the Cellular Defense against Stress*

TAp73 is a p53 tumor suppressor gene homologue that is known to be mainly involved in apoptosis. We report here that TAp73 is necessary for the cellular response to oxidative stress and that TAp73 functions as a downstream target of p53 in this process. We show that p53 physically interacts with the TAp73 promoter under stress conditions that lead to cell death. Particularly, p53 binds to a palindromic site in the TAp73 promoter, activates the promoter of TAp73, and selectively induces TAp73 transcription. TAp73 expression is highly increased under oxidative stress in a p53-dependent manner. Furthermore, knock-down of TAp73 expression inhibits the cellular apoptotic response to oxidative damage. In contrast, the ectopic expression of TAp73 in p53–/– mouse embryonic fibroblasts induces oxidative cell death. Our findings demonstrate that p53 is a direct transcriptional regulator of TAp73. Our data reveal a new pathway for cellular protection against oxidative damage and provide evidence that TAp73 is a stress-response gene and a downstream effector in the p53 pathway.

p53 is the most frequently mutated gene in human cancers (1). p53 plays a fundamental role in multiple cellular processes, including cell cycle checkpoint control, cell death induction, DNA repair, and genetic stability (2)(3)(4)(5). p53 acts as a sequencespecific DNA-binding protein and activates transcription by binding specific DNA consensus sequences (6). The p53 consensus sequence was initially defined as two copies of the 10-bp motif 5Ј-PuPuPuC(A/T)(T/A)GPyPyPy-3Ј, separated by a 0 -13-bp spacer (7). As a negative regulator of cell cycle progression, p53 functions to control the transition from G 1 to S phase of the cell cycle. p53 does this through transcriptional induction of p21, a universal inhibitor of cyclin/cyclin-dependent kinases, which is required for cell cycle arrest in response to DNA damage caused by ␥-irradiation (8).
It is well established that p53 functions as a mediator of cell death (9). p53 is required for irradiation-induced apoptosis in mouse thymocytes and is associated with cellular sensitivity to drug-mediated cell killing in chemotherapy (10,11). p53 is also required for the apoptotic response to oxidative stress (12). The mechanisms by which p53 regulates this apoptotic pathway have been explored extensively. It is known that p53 transactivation plays an essential role in induction of apoptosis (13). Indeed, many of the known transcriptional targets of p53 are involved in cell death processes, such as Bax (14), Noxa (15), p53AIP1 (16), Apaf-1 (17), PUMA (18,19), Pidd (20), and p53DINP1 (21). Almost all of these are functionally related with caspase signaling and mitochondrial apoptosis pathways (22). Although p53 mediates apoptosis in response to oxidative stress, the downstream effectors of p53 function in this pathway are largely unknown. We previously reported that PAC1 and MKP2, protein phosphatases and inhibitors of the mitogenactivated protein kinase (MAPK) cascade, are two transcriptional targets of p53 in signaling the cellular response to nutritional stress and oxidative stress (23,24). We propose that there are more targets of p53 transactivation involved in cellular defense processes. p73 was discovered due to its considerable homology to p53, and it was originally considered a potential tumor suppressor because of its location at chromosome 1p36, a region frequently deleted in neuroblastoma and other tumors (25). The p73 gene is able to encode transcriptionally active TA73, as well as an N-terminally truncated form, ⌬Np73 (DNp73), lacking the transactivation domain, which is transcribed from a different promoter within intron 3 of the gene (26). TAp73 protein shares structural and functional similarities with p53. The TAp73 gene encodes four polypeptides through alternative splicing of exons 11-13, designated as p73␣, -␤, -␥, and -␦. TAp73 is expressed in either a full-length form, p73␣, or shorter mRNA variants: ␤ (splicing out exon 13), ␥ (splicing out exon 11), and ␦ (splicing out exons 11, 12, and 13) (27). TAp73 has significant sequence identity to the transactivation, DNA binding, and oligomerization domains of p53. In addition, TAp73␣ has a SAM-like domain in its C terminus. TAp73 can bind to a p53 consensus sequence and transactivate some, but not all, p53-regulated genes. A main function of TAp73 is to mediate apoptosis, but unlike p53, it appears to be activated in response to certain types of DNA damage but not others (25,28). In contrast, DNp73 acts in a dominant-negative manner and blocks transactivation of target genes by p53 and TAp73 (29). The role of p73 in tumorigenesis is controversial. Unlike p53, which is the most commonly mutated gene in human cancer, mutations in the p73 gene are extremely rare (30,31), and there are no mutations found in hematopoietic cancers (32), breast cancer (30), and prostate cancer (33). p73 mutations are restricted to certain tumors, such as non-Hodgkin lymphomas (34). The expression of p73, however, is frequently altered in human tumors. For example, the expression of TAp73 is reduced in lymphoma and leukemia (35). In contrast, the overexpression of p73 was found in a variety of cancers, including those from breast, ovary, bladder, colon, and lung (36 -39). These observations suggest that p73 is not a classic tumor suppressor and even an oncogene. Although p53-null mice develop normally and are susceptible to tumorigenesis (40), p73 knockout mice have neurological, pheromonal, and inflammatory defects but do not develop spontaneous tumors (41). Thus, p53 and TAp73 have overlapping and distinct functions; p53 regulates the stress response to suppress tumorigenesis, and TAp73 might regulate both the stress response and development. Because p53 and TAp73 are linked to different upstream pathways, this family of transcription factors might regulate a common set of genes in response to different extracellular signals and developmental cues. On the other hand, p53 and TAp73 seem to be linked functionally in signaling apoptosis in mouse embryonic fibroblast (MEFs) 3 induced by DNA-damaging agents. It is shown that p53deficient E1A mouse embryo fibroblasts are highly resistant to apoptosis (11). However, p53-mediated cell death induced by DNA-damaging agents is inhibited in p73-null E1A mouse embryo fibroblasts, suggesting that p73 is required for p53-dependent apoptosis in these E1A MEFs (42). This is puzzling because if p73 functions as a transcription factor in parallel with p53, loss of p73 should not affect p53 function. Instead, it is possible that p73 is a downstream target of p53. That could explain why p73 overexpression induces apoptosis in the absence of p53, but p53 needs p73 to mediate apoptosis. Our data lend support to this model.
We have previously demonstrated that p53 is required for the cellular response to oxidative stress (12,43). In this study, we report that p73␣, which is referred to as TAp73 or TAp73␣ in the text, is transcriptionally regulated by p53 for signaling apoptosis following oxidative stress. We reveal that p53 transactivates the first promoter of the human p73 gene. We further demonstrate that p53 physically and specifically binds to a novel palindrome in the p73 promoter. It has been reported that combined loss of p63 and p73 inhibits p53-dependent apoptosis in response to DNA damage. We show here that reduction of TAp73 expression by its small interference RNA blocks p53-mediated cell death following oxidative stress. We propose that p53 is a transcriptional regulator of TAp73 and that TAp73 is necessary for the cellular response to oxidative stress.

EXPERIMENTAL PROCEDURES
Cell Culture and DNA Transfection-EB and EB-1 have been described previously (44). MEFs were described elsewhere (12,45). Primary p73 ϩ/ϩ and p73 Ϫ/Ϫ MEFs at passages 2-4 were cultured under normal growth conditions (42). Cancer cell lines used were originally purchased from ATCC. These cell lines were maintained in Earle's minimal essential medium (Invitrogen) supplemented with 10% fetal bovine serum (FBS) at 37°C in a humidified atmosphere with 5% CO 2 . Transfection was performed using Lipofectamine reagent (Invitrogen). For transient transfection, 2 ϫ 10 5 cells were seeded per well of a 6-well culture plate and incubated for 24 h. Plasmid DNA was added into 100 l of serum-free medium, mixed with 50 l of Lipofectamine reagent, and incubated for 15 min. The solution was added directly onto the cells in 2 ml of serum-free medium. The cells were incubated in Earle's minimal essential medium plus 20% FBS overnight. For stable clones, transfected cells were grown for about 2 weeks in a complete growth medium containing 2 g/ml hygromycin B for hygro resistance. Colonies were picked for further studies.
Construction of Luciferase Reporters, TAp73 Expression Plasmid, and siRNA Vector-Human genomic DNA was purified from normal human fibroblasts as template for promoter cloning. The regulatory region of the human TAp73␣ was amplified by PCR. The primers used to amplify the p73 promoter by PCR are as follows: 5Ј-GCGGGACCGTGGCTCCACAGGAGAA-GTG-3Ј (forward); 5Ј-GCGCCAAGTCCCAGGCCGATCCA-ACTCC-3Ј (reverse). PCR reactions were performed under the following conditions: seven cycles of 94°C, 2 s, 72°C, 3 min, 35 cycles of 94°C, 2 s, 67°C, 3 min, and one cycle of 67°C, 4 min. The promoter was ligated into the pGL3-basic reporter (Promega), resulting in pGL3/p73p-luc. pGL3/p73mt-luc was generated by changing the palindromic sequence from 5Ј-TGGACGCGGCCA-3Ј to 5Ј-TTGACGCGGCCA-3Ј as underlined using the QuikChange site-directed mutagenesis kit (Stratagene), according to the manufacturer's protocol. The mouse TAp73 expression vector was constructed from a selectable constitutive expression vector pcDNA3.1 (Invitrogen). The full coding region of mouse TAp73␣ (1.9 kb) was amplified from a mouse cDNA library (Clontech) by PCR using the following primers: mp73-20up, 5Ј-GCAGAATGAGCGGCAGCGTT-3Ј (forward); mp73-20dn, 5Ј-CGTTCCTCAGTGGCTCTCTG-3Ј (reverse). The resulting PCR product (1928 bp) was ligated into pT-Adv and then subcloned into pcDNA3.1, resulting in pcDNA3/mp73. pSilencer 1.0-U6 siRNA expression vector (Ambion) was used to generate a TAp73/siRNA made against a 19-nucleotide-specific mouse TAp73␣ sequence, which is in the starting coding region of the mouse TAp73␣. The U6/p73 siRNA construct was created by ligating the following annealed oligonucleotides into the designed sites of ApaI and EcoRI of the U6 vector as follows: forward, 5Ј-GGGACTAGCGAGGC-ATCAGTTCAAGAGACTGATGCCTCGCTAGTCCCTTT-TTT-3Ј; reverse, 5Ј-AATTAAAAAAGGGACTAGCGAGG-CATCAGTCTCTTGAACTGATGCCTCGCTAGTCCCG-GCC-3Ј. A random sequence, GACCTCCTGTCCTATG-AGG, was ligated into the U6 as a scrambled control. The constructs were sequenced to confirm that the designed fragments were introduced in-frame into vectors.
according to the manufacturer's instructions. Luciferase activity was measured with a Berthold Autolumat LB953 rack luminometer. Luciferase values were normalized against ␤-galactosidase activity. Luciferase readout was always obtained from triplicate transfections and averaged. RNA Isolation, Northern Blot, and Western Blot-Total RNA was isolated from growing cells using TRIzol reagent (Invitrogen). Poly(A) ϩ RNA was purified using a PolyATtract mRNA isolation system (Promega), according to the manufacturer's instructions. For Northern analysis, RNA was separated on a 1.2% formaldehyde gel and transferred to an N-Hybond membrane using a Turboblotter system (Schleicher & Schuell). The probe for TAp73 is based on the N-terminal sequence of TAp73, which is different from that of DNp73. The DNp73 probe is a unique sequence corresponding to the exon 3Ј of DNp73␣. DNA probes were labeled with [␣-32 P]dCTP (Amersham Biosciences) using the Prime-It RmT random primer labeling Kit (Stratagene). The membrane was hybridized with labeled DNA probes in the QuikHyb hybridization (Stratagene) at 65°C for 2 h and developed for autoradiography. For Western blotting, growing cells at 60 -70% confluence were lysed in cold Nonidet P-40 buffer with protease/phosphatase inhibitors. The samples were resolved by 7.5% SDS-polyacrylamide gels and then transferred onto nitrocellulose filters after protein separation. For the detection of TAp73, a monoclonal antibody specific for TAp73␣ (Calbiochem, ER-13) was used to recognize the ␣ form of TAp73. We used an anti-DNp73 monoclonal antibody that recognizes the ␦ form of p73 (Calbiochem, 38C674). Immunoblots were incubated with primary antibodies and then incubated with peroxidase-conjugated rabbit antimouse IgG as secondary antibody. The signals were detected with enhanced chemiluminescence (Amersham Biosciences).
DNase I Footprinting Analysis-Fragments of the human p73 promoter were end-labeled with [␥-32 P]ATP by T4 polynucleotide kinase, digested with BamHI, and purified as the sense strand 5Ј-end-labeled probe. Recombinant p53 protein was produced in insect cells infected with a baculovirus vector expressing human WT p53 and partially purified through affinity chromatography. p53 protein was added to bind to radiolabeled probe fragment (1-2 10 4 cpm) at 37°C for 20 min followed by the addition of DNase I from the Core footprinting system (Promega) according to the manufacturer's protocol. The reaction products were subjected to polyacrylamide gel electrophoresis under denaturing conditions, and the gel was dried and exposed for autoradiography.
Electrophoretic Mobility Shift Assay-Oligonucleotides (pairs of sense and antisense) were synthesized, annealed, and labeled with 32 P by using T4 polynucleotide kinase and [␥-32 P]ATP as described elsewhere (46). Recombinant p53 protein was produced in insect cells infected with a baculovirus vector expressing human WT p53 and partially purified through affinity chromatography. 32 P-labeled probes (2 ϫ 10 4 cpm) were mixed with 0.5 g of purified recombinant p53 in a 20-l DNA binding reaction buffer consisting of 20 mM Tris-HCl (pH 7.5), 4% Ficoll-400, 2 mM EDTA, 0.5 mM dithiothreitol, 0.2 mg of poly(dI-dC). For specificity or competition controls, a labeled random oligonucleotide and excess of unlabeled corresponding oligonucleotide were added together in the reactions.

Transcriptional Regulation of TAp73 by p53 in Signaling
Apoptotic Stress Response-To reveal the mechanism of the cellular response to genotoxic stress, we performed microarray analysis of gene expression profiles under stress conditions. We observed that TAp73 expression is increased following oxidative damage (data not shown). Because p53 is a controller of this process, we proposed that p53 is a regulator of TAp73. To determine whether p53 regulates the transcription of TAp73, we examined the expression of TAp73␣, the full-length form of TAp73, in the p53-inducible system, EB-1. The EB cell line was originally derived from a human colon cancer with a mutant p53 gene. EB-1 is a stable cell clone of EB that expresses ectopic wild-type p53 (WT p53) under the control of the metallothionein promoter. p53 is induced by ZnCl 2 (47). EB-1 cells undergo cell cycle arrest following WT p53 induction alone and typical apoptosis following serum deprivation in the presence of p53 (23,47). As shown in Fig. 1A, the transcriptional level of p21, a p53 target gene involved in cell cycle regulation (48,49), is low in uninduced EB-1 cells (middle panel, lane 1) but elevated following WT p53 induction by ZnCl 2 (middle panel, lanes 2 and 3). TAp73␣ transcripts are undetectable in EB-1 without p53 (upper panel, lane 1). Interestingly, TAp73␣ transcription is only slightly induced in EB-1 cells with ZnCl 2 (lane 1 versus lane 2), although p21 levels are highly increased (middle panel, lane 2), and cell cycle progression is blocked (8,23). However, TAp73␣ expression is greatly increased in EB-1 cells treated with both ZnCl 2 and serum starvation (upper panel, lane 3) or H 2 O 2 (lane 6), both of which increase the levels of p53 protein (lanes 3 and 6) and induce apoptosis (23,47). ZnCl 2 has no effect on TAp73␣ in parental EB cells (data not shown). We further examined the protein levels of TAp73␣ under these conditions. Although p53 protein is induced in the presence of ZnCl 2 (Fig. 1B, lanes 2, 3, 5, and 6), the expression of TAp73␣ protein is increased only when p53 is induced by ZnCl 2 plus serum starvation or oxidative damage (Fig. 1B, upper panel, lanes 3 and 6). Our results suggest that TAp73␣ is regulated by p53 and that stress conditions are necessary for TAp73␣ induction.
To determine whether p53 is required for the induction of TAp73␣ transcription, we tested normal and p53-null mouse embryo fibroblasts (p53 ϩ/ϩ and p53 Ϫ/Ϫ MEFs) for TAp73␣ expression in response to oxidative damage by H 2 O 2 , which induces apoptosis (43). We used a ribonuclease protection assay (RPA) for precise detection and quantitation of TAp73␣ mRNA following oxidative stress. As shown in Fig. 1C, the level of TAp73␣ mRNA is extremely low in p53 ϩ/ϩ MEFs (lane 1). However, TAp73␣ transcription is increased 11-12-fold following oxidative damage by H 2 O 2 (lanes 2-4). In contrast, the level of TAp73␣ transcription is only increased by 1-1.5-fold in p53 Ϫ/Ϫ MEFs following oxidative damage (lanes 5-8). Similarly, the expression of TAp73␣ protein is greatly increased following oxidative stress (Fig. 1D, lane 1 versus lane 3) in p53 ϩ/ϩ MEFs but not in p53 Ϫ/Ϫ MEFs (Fig.  1D, lane 5 versus lane 7). TAp73␣ protein is not induced during cell cycle arrest caused by either ␥-irradiation or serum starvation in these p53 ϩ/ϩ MEFs (Fig. 1D, lanes 2 and  4), although p21 is induced under these conditions. p21 is up-regulated by p53 following ␥-irradiation or serum starvation in these MEFs at the translational and transcriptional levels (Fig. 1, D and E, lanes 2  and 4). These observations suggest that TAp73 is specifically inducible by oxidative stress in a p53-dependent manner and that p53 can regulate TAp73 in response to oxidative damage. To determine whether p53 regulates TAp73 in vivo, we monitored the levels of TAp73 protein in various mouse tissues from p53 knock-out mice. As shown in Fig.  1F, TAp73 is expressed in all the normal mouse tissues examined, including brain, lung, thymus, liver, kidney, and bladder. However, TAp73 is low or almost undetectable in thymus (lane 5 versus lane 6), liver (lane 7 versus lane 8), and kidney (lane 9 versus lane 10) in the absence of p53. TAp73 expression is not significantly reduced in brain, lung, and bladder, where p53 is disrupted. These results indicate that p53 regulates TAp73 in a tissue-specific manner.
Transactivation of the Human p73 Gene Promoter by p53-To determine the molecular basis for transcriptional regulation of TAp73 by p53, we cloned the first promoter of the human p73 gene by PCR. Sequence analysis of the entire promoter revealed no known p53 consensus binding site defined as 5Ј-PuPuPuC(A/T)(T/A)GPyPyPy-3Ј (7). However, there is a 12-bp palindromic sequence (5Ј-TGGACGCGGCCA-3Ј) from Ϫ219 to Ϫ230 in the human p73 promoter, as indicated in Fig. 2A in bold and underlined letters. It was reported that p53 physically binds to a palindromic site in the promoter of its target gene PAC1 and activates the transcription of this gene (23). These two palindromes are different in sequence, but both are GC-rich elements. Thus, it is possible that p53 can also transactivate the p73 promoter by binding this GCrich palindromic site. The cells were harvested after 3 h, and mRNA was fractionated on a 1.2% formaldehyde agarose gel and transferred for Northern blotting using [␣-32 P]dCTPlabeled TAp73␣, p21, or GAPDH cDNA probes, respectively, following standard procedures. B, induction of TAp73 protein by p53. Exponentially growing EB-1 cells were treated as indicated. Cell extracts were resolved by 10% SDS-PAGE and transferred for Western blotting using a monoclonal anti-TAp73 antibody, anti-p53 antibody (Calbiochem), or anti-actin antibody (Santa Cruz Biotechnology) followed by appropriate secondary antibodies detected by chemiluminescence. C, RPA for TAp73 transcription. Exponentially growing MEFs were treated with H 2 O 2 (140 M) for the indicated time points, and RNA was isolated for RPA. The antisense TAp73 fragment was synthesized by in vitro transcription from the TAp73␣ under the control of the T7 promoter and was labeled with [␥-32 P]ATP as a probe. The 32 P-labeled cRNA probe was hybridized with an equal amount of poly(A)RNA from each group of cells and was subject to RNase A digestion using an RPA III kit (Ambion), according to the manufacturer's protocol. RPA for the detection of ␤-actin mRNA from the same groups was performed as a control. Yeast RNA (lane 9) was incubated with RNase as a control for RNase activity. Lane 10, free probe only. The mixture was separated with a 5% denaturing polyacrylamide, 7 M urea gel. Full-length TAp73␣ probe (upper bands) and protected TAp73␣ fragments (lower bands) are indicated as TAp73 probe and TAp73 mRNA, respectively. Undigested TAp73␣ and actin probes were also loaded on the gel for corresponding sizes. D, the expression of TAp73␣ protein in MEFs under stress conditions. MEFs were exposed to ␥-irradiation (6 Gy), H 2 O 2 (140 M), or serum starvation (0.1% FBS). The cells were harvested after 3 h, and total protein was resolved by 10% SDS-PAGE and then transferred for immunoblotting using the anti-TAp73 monoclonal antibody, which was developed by enhanced chemiluminescence. The levels of p53 and p21 protein are also detected using the corresponding antibodies. E, transcriptional regulation of p21 in MEFs under various conditions. MEFs were exposed to ␥-irradiation (6 Gy), H 2 O 2 (140 M), or serum starvation (0.1% FBS). The cells were harvested after 2 h for Northern analysis using a labeled mouse p21 cDNA fragment as a probe. F, the expression of TAp73 in tissues from p53 ϩ/ϩ and p53 Ϫ/Ϫ mice. Cell lysates were from mouse tissues with the indicated p53 statuses. The levels of TAp73 and p53 protein were examined by Western blotting using the anti-TAp73 antibody or anti-p53 antibody. Tubulin was used as a loading control. KO, knockout.
TAp73 Is a Stress-response Gene in the p53 Pathway OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40

JOURNAL OF BIOLOGICAL CHEMISTRY 29155
To determine whether the activity of the p73 promoter is regulated by p53, the putative human p73 promoter sequence was amplified from Ϫ726 to Ϫ15. The PCR fragment was ligated into a luciferase reporter vector pGL3-basic (Promega) adjacent to a luciferase reporter gene, resulting in the reporter pGL3/hp73-luc. The p73 luciferase reporter was then transfected into p53-null H1299 cells, along with a wild-type p53 expression vector (pCMV/wtp53) (50) or with a pCMV vector as a control. As shown in Fig. 2B, the pGL3/hp73-luc reporter has a basal activity in the absence of functional p53. However, its activity is greatly induced by a wild-type p53 expression plasmid but not by a mutant p53 plasmid (pC53-248) containing a point mutation at codon 248. To test whether the palindromic sequence is important for p53-mediated transactivity, we used a PCR-based site-directed mutagenesis to create a single point mutation by changing G to T in the palindromic sequence, resulting in a mutant reporter, referred to as pGL3/ hp73mt-luc. We observed that luciferase activity was reduced significantly in the luciferase reporter containing the mutated palindromic sequence in the presence of p53. These results suggest that p53 regulates the p73 promoter activity likely through this palindromic site. However, there is still a 2-fold increase of the pGL3/p73mt-luc in the presence of p53, indicating that p53 may regulate the promoter through other unknown mechanisms.
To determine whether p53 can also regulate the promoter of mouse p73, we cloned the first promoter of mouse p73. We found that there is no consensus binding site for p53. However, there is a 12-bp palindromic sequence (5Ј-GGACTG-CAGGCC-3Ј) located from Ϫ485 to Ϫ474 in the mouse p73 promoter (supplemental Fig. 1). Although it is different from the palindrome (5Ј-TGGACGCGGCCA-3Ј) in the human TAp73 promoter, these two palindromes are both GC-rich. We created a luciferase reporter for the mouse p73 promoter (pGL3/mp73luc) by ligating the mouse p73 promoter sequence (from Ϫ758 to ϩ21) into the luciferase reporter vector pGL3-basic (Promega) adjacent to a luciferase reporter gene. To examine luciferase activity, the luciferase reporters were transfected into p53 Ϫ/Ϫ MEFs along with a pCMV vector (pcDNA3) as a control or with a pCMV/wild-type p53 expression vector (pCMV/wtp53). As shown in Fig. 2B, there is a basal activity of the mouse p73 promoter in the absence of p53. The activity of the mouse p73 promoter. Nucleotide sequence upstream of the transcription start site is annotated from a PCR-amplified regulatory region of human p73 gene. The TATA box-like sequence is bolded. The 12-bp palindromic site is bold and underlined. Numbering is with respect to the transcriptional initiation site. B, activation of the p73 promoters by wild-type p53. Luciferase assays were carried out for induction of promoter activities of p73. H1299 cells (groups 1-6) or p53 Ϫ/Ϫ MEFs (groups 7 and 8) were transiently transfected with respective plasmids described below using Lipofectamine in OPTI-MEM medium (Invitrogen). Cell extracts were assayed for luciferase activity on a Berthold Autolumat LB953 rack luminometer. The luciferase activity readout is expressed as means Ϯ S.D. of triplicate cultures and transfections. The transfection groups are as follows: 1) empty pCMV vector plus pGL3-Basic; 2) pCMV/wtp53 plus pGL3-Basic; 3) pCMV vector plus human p73 promoter reporter, pGL3/hp73luc; 4) pCMV/wtp53 plus pGL3/hp73-luc; 5) pC53-248 plus pGL3/hp73-luc; 6) pCMV/wtp53 plus mutated reporter, pGL3/hp73mt-luc; 7) pcDNA3 empty vector plus mouse p73 promoter reporter, pGL3/mp73-luc; and 8) pCMV/wtp53 plus pGL3/mp73-luc.

TAp73 Is a Stress-response Gene in the p53 Pathway
promoter is significantly increased in the presence of pCMV/ wtp53. This result indicates that p53 can also transactivate the mouse p73 promoter.
Identification of Novel Palindromic Motifs as p53-binding Sites in the Promoters of Human and Mouse p73 Gene-As we discussed above, there is no conventional p53-binding consensus sequence in the p73 promoter. Nevertheless, since p53 transactivates the p73 promoter, it is possible that p53 may physically bind to the promoter. To test this possibility, we performed a chromatin immunoprecipitation (ChIP) assay using an anti-p53 antibody to detect the physical interaction between p53 and the p73 promoter. We chose the p53-inducible system, EB-1, for induction of p53. As shown in Fig. 3A, the p73 promoter was PCR-amplified from the ChIP products pulled out by the p53 antibody in group 3 where EB-1 cells were treated by 100 M ZnCl 2 plus serum deprivation (lane 3) or in group 6 where EB-1 were treated with H 2 O 2 in the presence of 100 M ZnCl 2 (upper panel, lane 6). Interestingly, under these conditions, the level of wtp53 protein was increased, and the EB-1 cells undergo apoptosis (23,47). These results suggest that p53 can bind to the p73 promoter in vivo and that p53 does so only under conditions that cause apoptosis. To identify the minimum p53-binding sequence, we conducted DNase I footprint analysis using the Core footprinting system (Promega), according to manufacturer's protocol. A human p73 promoter fragment was end-labeled with [␥-32 P]ATP by T4 polynucleotide kinase and digested with DNase I in the absence or presence of recombinant p53 protein. If p53 can bind a sequence in the promoter, this p53-bound sequence should be protected from DNase I digestion, leaving an undigested region, the so-called footprint, that can be revealed by sequencing gel analysis. Indeed, a clear footprint was observed between Ϫ190 and Ϫ230 of the p73 promoter in the presence of p53 (Fig. 3B, lane 2  versus lanes 3 and 4), corresponding to the palindromic site described above. These results indicate that p53 interacts specifically with the palindromic site of the p73 promoter.
To further demonstrate that p53 physically binds the palindromic site, we synthesized a 36-bp oligonucleotide encompassing the palindromic sequence (p73-36W; for sequence information, see the oligonucleotide list under "Experimental Procedures") and used it as a probe to perform an EMSA. As shown in Fig. 3C, one prominent shifted band was formed when this radiolabeled probe was incubated with recombinant human wild-type p53 (lane 2). This shifted band was blocked by adding an excess of the same unlabeled p73-36w oligonucleotide (lane

3) but not by an unrelated oligonucleotide NS30w (lane 4).
To confirm binding specificity, we included a p53 monoclonal antibody (pAb 1801) in the reaction. The addition of the p53 antibody resulted in a supershifted band (lane 5). This supershifted band is specific because the antibody does not form a complex with the probe in the absence of p53 protein (lane 6). These results demonstrate that p53 protein binds to the element containing the palindromic sequence. To better define the binding sequence, we synthesized oligonucleotides containing point mutations either within the palindromic sequence or in the flanking sequence and tested the interactions of these probes with p53. As shown in Fig. 3C, no shifted band shift was generated from reactions with oligonucleotides containing mutations within the palindromic sequence (lanes 7 and 8). In contrast, the probe containing a mutation outside the palindromic site still forms a shifted band with p53 (lane 9). As  , lane 6). The cells were harvested after 2 h and processed for ChIP by an anti-p53 antibody (pAb 1801). The ChIP was processed using a ChIP assay kit (Upstate Biotechnology), according to the manufacturer's instructions. The p73 promoter associated with p53 was amplified by PCR using the corresponding primers. Genomic DNA from each group was used as positive control for the PCR reaction (Input). B, footprinting of the p53-binding site in the promoter of human p73. The end-labeled promoter fragment was incubated with DNase I in the absence or presence of recombinant human wild-type p53 protein at the indicated concentrations. The reaction was electrophoresed in sequencing gels. The A/G ladders were prepared by using the Maxam-Gilbert sequencing method as size markers. The region protected from DNase I digestion between Ϫ190 and Ϫ230 was indicated. C, electrophoretic mobility shift detection for specific binding of p53 protein to a palindromic motif in the p73 promoter. Purified recombinant wild-type p53 protein was incubated with ␥-32 P-labeled double-stranded oligonucleotide containing the palindromic sequence from the human p73 promoter. For competition, a 50-fold excess of the unlabeled oligonucleotides was included in the reactions. The anti-p53 antibody was included in the indicated reactions for detecting a supershift. The reactions were separated by 4% native polyacrylamide gel electrophoresis and visualized by autoradiography. OCTOBER 5, 2007 • VOLUME 282 • NUMBER 40 a control, incubation of a labeled 30-bp oligonucleotide encompassing the p53 consensus binding site in the p21 promoter (p21-30W) resulted in a shift band (lane 11), which was supershifted by the anti-p53 antibody pAb 1801 (lane 12). These results clearly demonstrate that p53 can physically bind the palindromic sequence in the p73 promoter.

TAp73 Is a Stress-response Gene in the p53 Pathway
As shown in Fig. 2B, p53 is also able to transactivate the promoter of mouse p73. Because p53 can interact with the promoter of human p73, we set up the ChIP assay to determine whether endogenous p53 can also interact with the promoter of mouse p73. We chose p53 ϩ/ϩ MEFs for the ChIP assay and detected whether p53 is associated with the mouse p73 promoter under different conditions. As shown in Fig. 4A, p53 is associated with the p73 promoter following oxidative stress by 1 versus lane 3). Interestingly, there is no evidence for p53 association with the mouse p73 promoter when MEFs are exposed to ␥-rays (lane 1 versus lane 2). Our data indicate that p53 selectively interacts with the p73 promoter under certain stresses. This is interesting because oxidative stress causes cell death in MEFs, whereas ␥-irradiation induces cell cycle arrest in MEFs (23,51). To elucidate the basis for selective interaction, we examined the phosphorylation status of p53 in MEFs under stress conditions. As shown in Fig. 4B, lower panel, 1 versus lane 3). p53 is also phosphorylated at serine 18 under oxidative stress, whereas it is slightly phosphorylated at this serine 18 following ␥-irradiation (lane 1 versus lane  2). These results suggest that p53 is phosphorylated at different sites under different conditions, which may influence the choice of p53 binding to the promoter under these conditions. As we discussed above, there is no classic consensus site for p53 binding, but there is a 12-bp palindromic sequence (5Ј-GGACTGCAGGCC-3Ј) located from Ϫ485 to Ϫ474 in the first promoter of the mouse p73 gene. This is also a GC-rich element, although it is different from the human p73 palindrome in sequence. To determine whether p53 can physically interact with this palindrome, EMSA was performed using a 32-mer oligonucleotide corresponding to the mouse p73 promoter containing the palindromic sequence (mp73-32W; see the oligonucleotide list under "Experimental Procedures"). As shown in Fig. 4C, there is a shifted band as a complex of p53 and labeled mp73-32W (lane 2). This band can be competed out by excess of cold mp73-32W (lane 3) but not by the unrelated oligonucleotide NS30W (lane 4), indicating that this binding is specific. The shifted band can be further shifted by the addition of an anti-p53 antibody (lane 5). As a control, p53 forms a shift band with p21-30W that contains a conventional p53 consensus binding site (lane 8). As expected, this shifted band can be supershifted by the anti-p53 antibody (lane 9). These results suggest that p53 can directly bind to the palindrome in the promoter of mouse p73 gene.

Induction of Cell Death by TAp73 in Response to Oxidative Stress-If
TAp73 is a key downstream mediator of p53 function, the ectopic expression of TAp73 should be able to mimic p53 by inducing the cellular response to oxidative stress in the absence of p53. To define the function of TAp73 in the p53 pathway, we constructed a mouse TAp73␣ expression vector by cloning the 1.9-kb full-length coding sequence of mouse TAp73␣ cDNA into a mammalian expression vector driven by the cytomegalovirus promoter (pcDNA3/hygro, Invitrogen). The resulting TAp73␣ expression vector, pcDNA3/mp73, or an empty vector, pcDNA3, was transfected into p53 Ϫ/Ϫ MEFs, which were then selected with hygromycin B to isolate stable clones of p53 Ϫ/Ϫ MEF/p73. In addition to an empty vector-transformed FIGURE 4. Association of p53 with the mouse p73 promoter in vivo and in vitro. A, stress-induced interaction between endogenous p53 and the mouse promoter of p73. Growing MEFs were subject to ␥-irradiation (6 Gy) or H 2 O 2 (140 M) for 2 h and processed for ChIP with an anti-p53 antibody (pAb 421) plus an anti-phospho-p53 (Ser-15). The p53-associated mouse p73 promoter was amplified by PCR using the corresponding primers. Genomic DNA from each group was used as input. IP, immunoprecipitation. B, phosphorylation of p53 following oxidative damage. MEFs were either exposed to ␥-irradiation (6 Gy) or treated with H 2 O 2 (140 M) for 2 h. Total protein from these groups was resolved by 10% SDS-PAGE and then transferred for immunoblotting using a set of anti-phospho p53 antibodies (Cell Signaling). Actin was probed by the anti-actin antibody as a loading control. C, examination of interaction between p53 and a palindrome in the promoter of mouse p73. EMSA was performed using purified recombinant wild-type p53 protein. Protein (500 ng/reaction) was incubated with 32 P-labeled oligonucleotide containing the palindromic sequence in the mouse p73 promoter or the p53 consensus sequence in the p21 promoter. A 50-fold excess of the unlabeled oligonucleotides was included in the reactions for competition. To produce supershift bands, the anti-p53 antibody was included in the reaction. The reactions were separated by 4% native polyacrylamide gel electrophoresis and visualized by autoradiography. The oligonucleotide sequences are listed as one strand (5Ј-3Ј) under "Experimental Procedures." TAp73 Is a Stress-response Gene in the p53 Pathway clone (p53 Ϫ/Ϫ MEF/pcDNA), we obtained two stable clones that have moderate to high levels of ectopic TAp73␣ expression, named p53 Ϫ/Ϫ MEF/p73-c6 and p53 Ϫ/Ϫ MEF/p73-c11 (Fig. 5A). We then tested the susceptibility of these p53-null MEF cells with or without ectopic TAp73␣ expression to oxidative stress. As shown in Fig. 5B, whereas control p53 Ϫ/Ϫ MEFs are highly resistant to oxidative stress, the same cells expressing ectopic TAp73␣ undergo cell death following H 2 O 2 treatment. As control, three clones exhibit similar viability without any treatment. These results suggest that TAp73␣ can cooperate with a collateral signal induced by H 2 O 2 for induc-tion of cell death in the absence of p53. Therefore, TAp73 is a downstream effector of p53 function during the cellular response to oxidative stress.
We previously reported that p53 is required for the apoptotic response to oxidative stress (12). As a transcriptional target of p53, TAp73 is specifically up-regulated by p53 in response to oxidative stress. The fact that TAp73 is only responsive to oxidative stress but not to ␥-irradiation indicates that TAp73 may be necessary in this former process. To demonstrate the importance of TAp73 in signaling cell death, we used RNA interference technology to reduce TAp73 expression. Using the pSilencer 1.0-U6 vector for siRNA expression (52), we constructed a mouse TAp73/siRNA vector under the control of the U6 RNA polymerase III promoter. The mouse TAp73siRNA vector and a scrambled control U6 vector containing a random sequence (U6) were then introduced into MEFs, and stable clones selected by hygromycin B were tested for TAp73 expression by Western blotting (Fig.  5C). MEFs with TAp73 siRNA contain either low or undetectable TAp73 protein levels when compared with their normal counterparts under oxidative stress (Fig. 5C,  lanes 2 and 3 versus lanes 4 -7). This reduction is specific because the expression of DNp73 is still inducible in these cells by oxidative stress (Fig. 5D). As a control, p53 is highly induced by H 2 O 2 in these MEF/p73 siRNA cells (Fig. 5C), suggesting that p53 is still functional in these clones. We then determined the susceptibility of MEF/U6 and MEF/ p73 siRNA to oxidative stress. As shown in Fig. 5E, the MEF/p73 siRNA cells with undetectable or lower levels of TAp73 were highly resistant to cell killing by H 2 O 2 . We also performed a TUNEL assay to determine the nature of cell death and to measure the kinetics of cell death. As shown in Fig. 5F, MEFs gradually underwent apoptosis following treatment with H 2 O 2 , whereas the majority of MEF/p73 siRNA cells did not become apoptotic by H 2 O 2 . These results demonstrate that TAp73 plays an essential role in the p53-mediated apoptotic response to oxidative damage. Our results are consistent with the previous report that loss of p73 inhibits p53-dependent apoptosis induced by E1A following DNA damage (42). It was recently reported that p73 is not required for p53-dependent apoptosis in T cells (53). These findings indicate that p73 plays different roles in different cell types and that it is differentially responsive to various stresses.

DISCUSSION
p53 plays a fundamental role in controlling apoptosis. TAp73 also mediates cell death in a similar manner, and interestingly, it is required for p53-dependent apoptosis. Therefore, it is important to explore the relationship between these two homologous tumor suppressor genes. In this study, we demonstrate that TAp73, or TAp73␣, is a transcriptional target for p53 in mediating apoptosis. As a homologue of p53, TAp73 has been widely considered to function in parallel with p53. Indeed, TAp73 can regulate some p53 target genes and mediate apoptosis in response to either DNA damage or chemotherapeutic agents (54). However, it is unclear whether and how TAp73 is regulated at the transcriptional level. Regulation of p53 function occurs mostly at the protein level, whereas p53 transcription is rarely influenced by environmental stress. However, TAp73 transcription is regulated by transcription factor E2F-1, and TAp73 is involved in E2F-1-mediated apoptosis as a downstream effector (55,56). It has been shown that p73 is required for the p53-mediated apoptotic response to DNA damage, raising the question of whether p73 is a downstream component in the p53 pathway (42). It was previously reported that p73 is transcriptionally regulated by DNA damage, p53, and p73 (57). We found that TAp73 is significantly up-regulated by p53 under conditions of serum starvation or oxidative damage that causes apoptosis but not cell cycle arrest. Our results demonstrate that p53 directly regulates the expression of TAp73 during the apoptotic response to nutritional stress or oxidative damage.
One fundamental question is how p53 regulates TAp73 in response to oxidative stress. The well characterized p53 consensus binding sequence encompasses two copies of the 10-bp motif 5Ј-PuPuPuC(A/T)(T/A) GPyPyPy-3Ј, separated by a 0 -13-bp spacer (7). This consensus site has been found in the regulatory regions of a number of genes controlled by p53, such as p21, Bax, and PIG3 (14,48,58). However, there are many cases where no known p53-responsive consensus sequence can be identified in p53-inducible genes. This is true in our study of the regulation of TAp73. Although p53 can transactivate the p73 promoter, there is no known p53 consensus binding site in the first promoter of the human p73 gene. We have previously identified a novel mechanism for p53 to regulate its target gene PAC1 through binding a palindromic motif in the promoter (23). Interestingly, there is also a similar palindromic site in the promoter of the human p73 gene. We speculated that p53 is able to bind to this palindromic motif to activate TAp73 transcription in response to apoptotic stimuli. As shown above, we have generated sufficient data to classify TAp73 as a stressinducible gene in the p53 pathway. Systematic analysis of p53 target genes using DNA microarray technology has demonstrated that p53 regulates more than 60 genes with defined functions in many systems, resulting in a variety of biological effects (44). The remaining challenge is to elucidate how p53 regulates so many genes and what is the molecular basis for p53 function in response to various factors and environmental stresses. We propose that p53 utilizes this palindromic binding motif to regulate a distinct class of target genes, which differ functionally from the genes regulated through the conventional p53-binding sequence. Interestingly, endogenous p53 only binds to the p73 promoter in vivo when it is activated under oxidative stress but not following ␥-irradiation. Thus, we speculate that p53 may be modified under oxidative stress and that this modified form of p53 may have access to the promoter of p73 containing the palindrome. As supporting evidence, we observed that mouse p53 is phosphorylated after incurring oxidative damage. Because p53 can bind to the p21 promoter containing a perfect p53 consensus site and induces the expression of p21 following ␥-irradiation, it would be conceivable that under different conditions, p53 may have options to bind the promoters of a subset of its target genes containing different types of binding sites: conventional consensus sites versus palindrome motifs, leading to distinct transcriptional and biological outcomes. We have searched the human genome data base for other genes with palindromic sequences and found that there are a number of genes containing either identical or similar palindromic motifs in their promoters. Therefore, the identification of the palindromic sequences as a second type of p53-binding site will pave the way for a better understanding of how p53 plays multiple roles and how other factors influence p53 binding and transactivation of select target genes.
The importance of TAp73 in signaling oxidative cell death was evaluated using newly developed RNA interference technology. Because reduction of TAp73 by RNA interference significantly blocks cell death triggered by oxidative stress, we propose that TAp73 is required for the p53-mediated apoptotic response to oxidative stress. Our data reveal a novel mechanism for tumor suppression by p53 and may explain why p53 has broad and powerful effects on genetic defense and why most cancer-associated mutations occur in p53 instead of its p73 homolog. Because TAp73 is an essential downstream effector of p53 and it is not mutated in human cancers, TAp73 may serve as an excellent target for inducing apoptosis in tumor cells using chemical compounds, providing more effective treatments for cancer patients.
Ever since p73 was cloned, it has been speculated that p73 and p53 have similar functions. However, the observations that rare mutations and the overexpression of p73 are found in some human cancers imply that p73 is at least not a classic tumor suppressor and even an oncogene. Consistently, an early report on a p73 knock-out mouse model implicates p73 in epithelial stem cells and development neurogenic responses but not in tumor suppression (41). The complex pattern of p73 transcription resulting in isoforms with a diversity of functions may be an explanation for these contradictory reports. For example, TAp73 seems to function in a manner much like p53, whereas DNp73 is oncogenic. It has been shown that p53 induces the expression of its antagonist DNp73, establishing an autoregulatory feedback loop (59). Since DNp73 is also induced by oxidative stress, the fate of the stressed cells should be determined by the balance between TAp73 and DNp73, both of which are induced by p53. Obviously, these cells are killed following oxi-dative stress. Thus, TAp73 seems to be dominant over DNp73 in this event. Recent studies (60) show that mice heterozygous for p53 and p73 display a higher incidence of spontaneous tumor and metastasis when compared with p53 ϩ/Ϫ mice. p63 ϩ/Ϫ and p73 ϩ/Ϫ mice also developed spontaneous tumors. Strikingly, aging p73 Ϫ/Ϫ mice develop lung adenocarcinomas at a very high frequency (60%) (60). Interestingly, as shown in Fig. 1F, TAp73 is only regulated by p53 in a tissue-specific manner, and TAp73 is not regulated by p53 in lung. This implies that p53 and TAp73 function differently in lung, although they both contribute to suppression of lung carcinoma in mouse models. Consistently, p53-deficient mice mainly develop sarcomas and lymphomas (40), whereas p73 knock-out mice develop lung cancers (60). The reported tumor spectrum in aging p73 Ϫ/Ϫ mice is in sharp contrast to the multiple tumor development in p53-null mice as young as 6 months of age (40), which may reflect the tumor suppressive role of p73 in a tissuespecific manner. The difference of tumor development in age may also point to developmental defects caused by loss of p73 in the immune system that leads to chronic inflammation (41), which often is the culprit of tumorigenesis. Our results establish a novel link between p53 and TAp73 and provide a mechanism by which TAp73 functions as a tumor suppressor. There are several isoforms of p73 that have similar or opposite roles in tumor development. Our siRNA approach allowed us to specifically demonstrate the role of TAp73 in controlling cell death. The high frequency of tumor development in p73-null mice may be related to the accumulation of genetic damage caused by endogenous reactive oxygen species in the absence of TAp73. Our findings that TAp73 eliminates oxidatively damaged cells provide insights into the mechanism by which p73 functions as a tumor suppressor gene.
In summary, we have demonstrated that TAp73 is a direct transcription target of p53 and that p53 selectively regulates TAp73 expression by binding to the p73 promoter. We suggest that p53 may be modified following different damage signals, and modified p53 may specifically and selectively regulate its target genes through different binding sites to mediate proper responses to environmental stresses. TAp73 functions to mediate cell death induced by nutritional and oxidative stresses.