Research Article 1245 The neurogene BTG2TIS21/PC3 is transactivated by ∆Np73α via p53 specifically in neuroblastoma cells

David Goldschneider1, Karine Million1, Anne Meiller3, Hedi Haddada1, Alain Puisieux2, Jean Bénard1, Evelyne May3,* and Sétha Douc-Rasy1,‡ 1CNRS UMR 8126, Institut Gustave Roussy, 39 rue Camille Desmoulins, 94800 Villejuif, France 2INSERM U590, Centre Leon Bérard, 28 rue Laënnec, 69008 Lyon, France 3CEA-CNRS UMR217, PB6, 92265 Fontenay-aux-Roses, France *Present address: CNRS UPR9045, Institut André Lwoff, 7 rue Guy Moquet, 94801 Villejuif, France ‡Author for correspondence (e-mail: [email protected])

Accepted 6 January 2005 Journal of Cell Science 118, 1245-1253 Published by The Company of Biologists 2005 doi:10.1242/jcs.01704

Summary The p53 gene and its homologue p73 are rarely mutated it occurs in a p53+/+ neuroblastoma cell line but not in a in neuroblastoma. In recent studies, we showed that p53+/+ breast tumor cell line). The fact that ∆Np73α may overexpression of ∆Np73α, an isoform lacking the N- either inhibit or stimulate wt-p53 transcriptional activity, terminal transactivation (TA) domain, surprisingly induces depending on both the p53 target gene and the cellular p53 protein accumulation in the wild-type (wt) p53 human context, was confirmed by real-time quantitative PCR. neuroblastoma line SH-SY5Y. As can be expected owing to Moreover, transactivation of the BTG2TIS21/PC3 promoter its dominant-negative effect, ∆Np73α inhibits Waf1/p21 requires a complete ∆Np73α C-terminus sequence as it is gene expression, but equally importantly, it upregulates not observed with ∆Np73β, which lacks most of the C- BTG2TIS21/PC3, another p53 target gene. This effect is terminal domain. We have previously shown that ∆Np73α not observed in neuroblastoma cells that express a is the only p73 isoform expressed in undifferentiated mutated p53. To better understand the ∆Np73-mediated neuroblastoma tumors. In light of all these findings, we transactivation of the BTG2TIS21/PC3 gene we performed propose that ∆Np73α not only acts as an inhibitor of luciferase assays with two reporter plasmids harboring p53/TAp73 functions in neuroblastoma tumors, but also long and short BTG2 promoter sequences in three human cooperates with wt-p53 in playing a physiological role neuroblastoma cell lines and one breast cancer cell line. through the activation of BTG2TIS21/PC3 gene expression. Our results demonstrate that BTG2TIS21/PC3 transactivation

Journal of Cell Science by ∆Np73α depends on both p53 status (as it is not observed Key words: p73, p53, BTG2TIS21/PC3, Transactivation, in a p53–/– neuroblastoma cell line) and cellular context (as Neuroblastoma, Apoptosis, Differentiation, Development

Introduction Structural homology between p53 and p73 leads to The p73 gene and the p53 tumor suppressor gene share functional homology. TAp73 mimics p53 in its ability to homology in terms of sequence and the structural organization transactivate many p53-responsive genes (including the Mdm2, of the three canonical functional domains, namely, the acidic N- an E3 ubiquitin ligase and the cyclin-dependent kinase terminal transactivating domain (TA), the core DNA binding inhibitor p21) and to induce apoptosis (Jost et al., 1997). domain (DBD) and the C-terminal oligomerization domain (OD) However, in sharp contrast with p53, Mdm2 cannot target p73 (Kaghad et al., 1997). The p73 gene encodes 636 amino acids for ubiquitin-proteasome degradation, although it can block its generated from 14 exons and specifies two main N-terminal transcriptional activity (Zeng et al., 1999). Initial studies on isoforms: the full-length TAp73 (TAp73α) initiated from the P1 DNA-damaging agents report a lack of p73 induction upon promoter and the N-terminally truncated p73 isoforms. These actinomycin D treatment or UV irradiation. However, several TA-deficient p73 isoforms include isoforms initiated from P1 reports show that endogenous p73 is a target of the non- but that are lacking exon 2 as a result of alternate splicing receptor tyrosine kinase c-Abl in p53-deficient mouse embryo (p73∆exon2) and isoforms initiated from a second internal fibroblast cells in response to genotoxic stress such as cisplatin promoter, P2 (∆Np73α) (Yang et al., 2002). Another particular (Gong et al., 1999). Furthermore, p73 and c-Abl can associate feature of p73 may be the fact that its C-terminus contains three with each other via the p73 PxxP motif and the c-Abl SH3 additional domains: a second transactivation domain TA2 domain in γ-irradiated cells to induce apoptosis or cell cycle (Takada et al., 1999), a SAM (‘sterile alpha motif’) protein- arrest in the G1 phase (Agami et al., 1999). protein interaction domain (Chi et al., 1999) and a basic C- In neuroblastoma cells, ∆Np73 expression is stimulated by terminus sequence known as a transcriptional negative regulator overexpression of TAp73 or p53; moreover, induction or (Ozaki, et al., 1999). Moreover, it is generally thought that the overexpression of ∆Np73 promotes cell survival by C-terminus domain negatively regulates the transcriptional competition with TAp73 or wild-type (wt-) p53 (Nakagawa et ability of TAp73α (Yang et al., 2002). al., 2002; Stiewe et al., 2002). In vitro, the spliced variant 1246 Journal of Cell Science 118 (6)

p73∆exon2 also functions as a transdominant-negative ∆Np73α-dependent upregulation of the BTG2TIS21/PC3 gene inhibitor of TAp73 and wt-p53 as it interferes with the initially observed in neuroblastoma cells. apoptotic function of p53 and is therefore one of the key regulatory elements involved in p53 signaling pathways (Fillippovich et al., 2001; Zaika et al., 2002). Materials and Methods ∆ ∆ In human tumors, Np73 and/or p73 exon2 transcript levels Plasmids and Recombinant p73-adenovirus constructs are found to be upregulated in about 80% of human cancers, pcDNA-TAp73α/pcDNAp73β and pcDNA-∆Np73α/pcDNAp73β including cancers of the ovary, endometrium, cervix, vulva, plasmid constructs were donated by Mourad Kaghad (Sanofi vagina, breast, kidney and colon (Zaika et al., 2002). In human Recherche, Labège). The recombinant adenoviral vector expressing neuroblastic tumors, p53 and p73 are very rarely mutated full-length p73α (Ad-p73α) and ∆N-p73α (Ad-∆Np73α) were (Bénard et al., 2003) but p73 protein expression is clearly produced by pcDNA-p73α and pcDNA-∆N-p73α in vivo homologous deregulated (strong accumulation in undifferentiated recombination in 293 cells as previously described (Goldschneider et neuroblastic tumors and a total lack in Schwann cells). In al., 2004). pDDm-TO were obtained by inserting the sequence that addition, the ∆Np73α protein is the only isoform observed by encodes the mouse p53DD truncated protein (amino acid residues 1- western blotting in undifferentiated neuroblastic tumors 14 and 302-390) into the Invitrogen vector pcDNA/TO/B (Drané et . al., 2002). pE1B-hWAF1 contains the p53RE of the Waf1/p21 (Douc-Rasy et al , 2002), which coincides with a poor outcome promoter cloned ahead of the luciferase gene (Munsch et al., 2000). in these neuroblastoma patients (Casciano et al., 2002). TIS21/PC3 The pBTG2-Luc reporter constructs were obtained as previously BTG2 , the antiproliferative human homologue gene described (Duriez et al., 2002). of PC3 (rat) or TIS-21 (mouse) belongs to the B-cell translocation gene family, initially found to be rapidly and transiently induced by nerve growth factor (NGF) and Cell culture and adeno-recombinant infection depolarization (Bradbury et al., 1991). It has been reported that The parental human neuroblastoma cell lines, SH-SY5Y and IMR-32, BTG2TIS21/PC3 expression is induced through p53-dependent were purchased from the European Collection of Cell Cultures mechanisms and that the BGT2 function may play a role in cell (ECACC, Wiltshire, UK) and express wt-p53. The neuroblastoma cycle control and cellular response to DNA damage (Rouault LAN-1 cell line (gift of Nicole Gross, Pediatric Oncology Research, . BTG2TIS21/PC3 Lausanne, Switzerland) does not express p53. The MCF-7 cell et al , 1996). has been found to be involved in line, purchased from American Type Culture Collection (ATCC, cell growth, differentiation and DNA repair (Tirone, 2001). Manassas, VA) expresses wt-p53. The human IGR-N-91 cell line was PC3 mRNA expression coincides both spatially and temporally established in our laboratory from the bone marrow of a patient with with the pattern of CNS neurogenesis in the developing rat metastatic neuroblastoma after unsuccessful adriamycin-vincristine (Iacopetti et al., 1994). The human BTG2TIS21/PC3 was recently chemotherapy (Ferrandis et al., 1994). The IGR-N-91 p53 gene is evaluated as a candidate tumor suppressor gene; the gene was mutated by an in-frame duplication of exons 7-8-9 (Goldschneider et cloned and its structural organization was determined (Duriez al., 2004). The human breast cancer MCF-7 and the neuroblastoma et al., 2002). In vitro studies have shown that BTG2TIS21/PC3 cell lines were grown under standard conditions in Dulbecco’s promotes neuronal differentiation and prevents apoptosis of modified Eagle’s medium (DMEM) supplemented with 2 mM L- terminally differentiated cells (el-Ghissassi et al., 2002). In glutamine and 10% calf fetal serum. Cells were infected with the Journal of Cell Science BTG2TIS21/PC3 recombinant adenovirus vector expressing either p73 full-length (Ad- normal human tissues, is differentially expressed p73α or β) or N-terminal truncated p73 (Ad-∆Np73α or β) at an MOI in several differentiated epithelial cells including the breast, of 15. The cells were harvested after 48 hours’ infection. lung, intestine and pancreas (Melamed et al., 2002). Physical interactions between BTG2TIS21/PC3 and other cell cycle regulator proteins, such as CCR4-CAF, have recently been Western blotting demonstrated in HeLa cells (Morel et al., 2003). The Western blotting was performed as previously described BTG2TIS21/PC3, CAF1 and CCR4 complex has been shown to (Goldschneider et al., 2004). Briefly, the cells were directly lysed in regulate the transcription activity mediated through the Laemmli buffer and the total protein extracts (50 µg to detect estrogen receptor α. endogenous protein or 1 µg to detect protein in infected cells) were In recent studies involving p73 target genes in loaded onto SDS-PAGE (7.5%). After electrophoresis, the proteins were electrotransferred to nitrocellulose filters and the filters were neuroblastoma cells using the adenoviral (Ad)-p73 probed with either polyclonal p73 antibody to reveal p73, or DO-7 recombinant approach we showed that both Ad-∆Np73α and α monoclonal anti-p53 antibody (DAKO) to reveal p53, or with anti- Ad-TAp73 induce accumulation and activation of the p73β (clone GC15 from Upstate Biotechnology) to reveal p73β or anti- endogenous wt-p53 in wt-p53 expressing cells (Goldschneider p63 (clone 4A4 from Oncogene Research) to reveal p63. The dilution et al., 2003; Miro-Mur et al., 2003; Goldschneider et al., for each antibody was 1:1000. Protein detection was carried out using 2004). Strikingly, ∆Np73 selectively regulates p53 target an ECL kit (Amersham Pharmacia Biotech, France). Controls were genes as the BTG2TIS21/PC3 transcript level was shown to be performed using β-actin antibody MAB 1501 (Chemicon). upregulated both in Ad-∆Np73α and Ad-TAp73-infected cells, thus contrasting with other p53 target genes such as Real-time quantitative PCR (RTQ-PCR) Waf1/p21Wa f , which was completely repressed by ∆Np73 BTG2 Waf1/p21 . . and mRNA were quantified using the Applied (Miro-Mur et al , 2003; Goldschneider et al , 2004). The Biosystem apparatus, Abi Prism 7000 Sequence Detection system. present study was performed to determine whether: (1) Briefly, total RNA was extracted from non-infected or infected cells TIS21/PC3 ∆ BTG2 activation by Np73 requires both using RNeasy kit (Qiagen, Courtaboeuf, France) according to the TIS21/PC3 BTG2 promoter sequences and p53; (2) the C-terminus manufacturer’s instructions. Total RNA (1 µg) was reversed plays a role in the ∆Np73-dependent transactivation of transcribed using SuperScript II (Invitrogen). RTQ-PCR was BTG2TIS21/PC3; and (3) the cellular context contributes to the performed in a final volume of 25 µl containing 25 ng of each total ∆Np73α transactivates BTG2TIS21/PC3 via p53 1247

A p53BS -2700 luciferase

-266 luciferase -119

B D 100 400 + empty vector SH-SY5Y (wt-p53) 90 LAN-1 (null-p53) + empty vector 350 + TAp73α 80 + TAp73α + ∆Np73α 300 (A. U.) (A. U.) 70 + ∆Np73α + p53DD 250 60 + p53 50 200 + TAp73α + p53 40 150 + ∆Np73α + p53 30 100 20 Luciferase activity

Luciferase activity 50 10

0 0 BTG2 2700 BTG2 266 BTG2 2700 BTG2 266 C E 200 IMR-32 (wt-p53) + empty vector 500 MCF-7 (wt-p53) + empty vector 180 + TAp73α 450 + TAp73α 160 + ∆Np73α 400 + ∆Np73α (A. U.) 140 + p53DD (A. U.) 350 + p53DD 120 300

100 250

80 200 60 150 40 100 Luciferase activity

20 Luciferase activity 50 0 0 BTG2 2700 BTG2 266 BTG2 2700 BTG2 266

Fig. 1. BTG2TIS21/PC3 promoter activation by TA- and ∆N-p73α in wt-p53 neuroblastoma cell lines but not in wt-p53 breast cancer line MCF-7.

Journal of Cell Science (A) Schematic representation of pGL3-reporter constructs containing p53 binding site (BS) at –119, as previously described (Duriez et al., 2002); long BTG2TIS21/PC3 promoter sequence fragments of 2700 bp (–2700) and short promoter fragment (–266). (B) SH-SY5Y, (C) IMR-32, (D) LAN-1 and (E) MCF-7 cells were transiently co-transfected with pBTG2-Luc plasmid carrying long (2700) or short (266) promoter sequences in the presence of either TA- or ∆Np73 expressing vectors or empty control plasmids. Transfection techniques and luciferase activity measurements were carried out as described in the Materials and Methods. Error bars represent standard deviation calculated from three independent experiments. Cells were transfected with empty vector, TAp73α, ∆Np73α, p53DD, TAp73α + p53, ∆Np73α + p53 as indicated.

RNA template, 10 pmoles of each primer and 12.5 µl of a SYBR®- 2×104 cells per cm2 and transfected with 0.5 µg (2.5 µg/ml) of either Green master mix. The primers were designed using the Oligo 6 pGL3 firefly luciferase reporter gene plasmids containing a 2700 bp Primer Analysis Software (Molecular Biology Insights). Three sets of fragment (BTG2 2700) or a 266 bp fragment of the BTG2 (BTG2 primers were used: GAPDH-F, 5′-AGCTCACTGGCATGGCCTTC- 266) or the Waf1/p21 luciferase reporter plasmid, pE1B-hWAF1 3′; GAPDH-R, 5′-ACGCCTGCTTCACCACCTTC-3′; p21-F, 5′-GG- (WAF-1) using the calcium phosphate method as previously described ACCTGTCACTGTCTTGTA-3′; p21-R, 5′-GGCTTCCTCTTGGA- (Drané et al., 2001). To evaluate the transactivation effect of p73, cells GAAGAT-3′; BTG2-F, 5′-CGAGCAGAGGCTTAAGGTCTTC-3′; were co-transfected with 1 µg (5 µg/ml) of either TAp73α- or ∆Np73 BTG2-R, 5′-CTGGCTGAGTCCGATCTGG-3′. Quantification was (α or β)-expressing vectors or control plasmid. At 24 hours after carried out using the comparative CT method and water was used as transfection, cells were lysed with 200 µl/well of passive lysis buffer the negative control. An arbitrary threshold was chosen on the basis provided with the Luciferase assay kit (Promega). Luciferase activity of the variability of the baseline. Threshold cycle (CT) values were was measured according to the manufacturer’s protocol with the calculated by determining the point at which the fluorescence Microlumat LB96P luminometer (EG&G Berthold). exceeded the threshold limit. CT was reported as the cycle number at this point. The average of target gene was normalized to GAPDH as endogenous housekeeping gene and relative to the non-infected Results and Discussion ∆∆ – CT ∆∆ ∆ condition as control and was given by 2 where CT = CT TIS21/PC3 ∆ α ∆ ∆ p53-dependent activation of BTG2 by N-p73 (sample) – CT (control); CT = CT (target gene) – CT (GAPDH). through the BTG2TIS21/PC3 promoter sequence In previous studies we showed that in wt-p53 SH-SY5Y cells Luciferase reporter assays TAp73 proteins were able to upregulate the p53 target gene Cells were seeded in triplicate onto six-well plates at a density of Waf1/p21 involved in cell cycle arrest, the proapoptotic genes 1248 Journal of Cell Science 118 (6)

PUMA and BAX, and the BTG2TIS21/PC3 gene. On the other worthwhile to note that the canonical p53 responsive element of hand, and as expected, ∆Np73α, a dominant-negative isoform Waf1/p21 gene is upregulated by an ectopic expression of p53 of both p53 and TAp73, efficiently inhibited the conventional in these cells (D. G., unpublished data). p53 responsive genes such as MDM2 BAX or Waf1/p21. Surprisingly, however, the expression of BTG2TIS21/PC3 was upregulated by ∆Np73 and this effect occurred specifically in The cellular context may interfere with the ∆Np73α- wt-p53 SH-SY5Y cells but not in mutated-p53 IGR-N-91 cells mediated BTG2 transactivation (Goldschneider et al., 2004). We next addressed the question of whether BTG2TIS21/PC3 can To further define the mechanism by which ∆Np73 activates be activated by the ∆Np73 isoform in wt-p53-expressing cells BTG2TIS21/PC3 transcription, we performed parallel co- derived from other tissue. The same experiments were thus transfection assays with recombinants expressing the various performed in MCF-7, a human breast carcinoma cell line that p73 isoforms and reporter constructs containing either long expresses wt-p53. Like SH-SY5Y and LAN-1, the MCF-7 cells (2700 bp) or short (266 bp) BTG2TIS21/PC3 promoter sequences expressed an endogenous p73α protein as detected by western (Duriez et al., 2002) linked to the luciferase gene. The BTG2- blot analysis (Fig. 2). P73α was also expressed in IMR-32 p53 binding site (p53BS) is located at position –119 from the cells, but to a lesser extent, as the protein could only be initiation site (Fig. 1A). detected when the blot was overexposed (30 minutes compared Results obtained with SH-SY5Y are presented in Fig. 1B. to 0.5 minutes). In contrast, it was not possible to detect either Luciferase activity was significantly higher when the reporter the p73β or the p63α protein by western blotting in any of these gene was driven by short rather than long BTG2TIS21/PC3 cell lines. promoter sequences (Fig. 1B, empty vector). Promoter activity Surprisingly, in the MCF-7 cellular context neither TA- nor was inhibited by the p53 dominant-negative form, p53DD, ∆Np73α ectopic expression was able to stimulate luciferase in both cases, demonstrating that expression through the expression from the BTG2TIS21/PC3 promoter sequences (Fig. BTG2TIS21/PC3 promoter depends on endogenous p53. More 1E). In contrast, the activity of these sequences was diminished. interestingly, co-transfection of ∆Np73 increased transcriptional These results indicate that the stimulatory effect of both p73α activity to the same level as TAp73, strongly suggesting that an and ∆Np73α on the p53-dependent BTG2TIS21/PC3 transcription upregulation of BTG2 promoter activity may account for our is cell specific. As neuroblastoma is a peripheral primitive previous published data showing an increase of endogenous neuroectodermal tumor, derived from neural crests cells, an BTG2 mRNA levels in SH-SY5Y overexpressing ∆Np73 important issue will be to determine whether or not (Goldschneider et al., 2004). BTG2TIS21/PC3 transactivation by ∆Np73α also occurs in other To determine whether BTG2TIS21/PC3 activation by ∆Np73 wt-p53-expressing cells originating from the same precursor. occurs in other neuroblastoma cell lines expressing wt-p53, the same experiments were performed with IMR-32 cells (Fig. 1C). TIS21/PC3 Again, the BTG2TIS21/PC3 promoter was consistently activated by Transactivation by ∆Np73 is restricted to BTG2 ∆Np73 as well as by TAp73 and again, the short promoter It is well established that TAp73 proteins like p53 are able to sequence was more efficient than the long sequence. As with upregulate the Cdk inhibitor Waf1/p21 (De Laurenzi et al., Journal of Cell Science SH-SY5Y, p53DD interfered negatively with the activity of the 1998), whereas ∆Np73 exerts a dominant-negative effect on p53 BTG2TIS21/PC3 promoter in IMR-32 cells. and TA transcriptional activity (Kartasheva et al., 2002; Vossio To ascertain the involvement of wt-p53, luciferase assays were et al., 2002). This coincides with our previous results, which repeated using LAN-1, a p53-deficient (null-p53) neuroblastoma showed that an overexpression of ∆Np73α in SH-SY5Y cell line. Cells were co-transfected with pluc-BTG2 and a vector neuroblastoma cells downregulated Bax, PUMA and Waf1/p21 encoding a p73 isoform and/or a vector encoding wt-p53. In the genes although p53 was accumulated (Goldschneider et al., absence of exogenous p53 expression, neither TA- nor ∆Np73α 2004). However, surprisingly, we found that an overexpression was able to activate the BTG2TIS21/PC3 promoter (Fig. 1D). of ∆Np73α upregulated BTG2TIS21/PC3 expression. This Interestingly, in these cells, an ectopic expression of wt-p53 had unexpected result has now been confirmed at the promoter level. no effect on BTG2TIS21/PC3 promoter activity, but the co- To ascertain that BTG2TIS21/PC3 promoter activation was not an expression of wt-p53 and TA- or ∆Np73α significantly increased artifact of the system analysis, the same experiment was the luciferase activity, supporting a synergic effect of p53 and performed using a plasmid encoding the luciferase reporter gene p73 at the level of BTG2TIS21/PC3 promoter activation. It is under the control of the Waf1/p21 p53-responsive element. The co-transfection of ∆Np73α dramatically decreased luciferase activity in both SH-SY5Y and MCF7 cells, two cell lines expressing wt-p53 (Fig. 3A,B). These results led us to postulate that ∆Np73 is able to cooperate with p53 to activate some specific p53 target genes (BTG2TIS21/PC3 in the present study) depending on both the promoter sequences and the cellular context.

Fig. 2. Western blot showing p73, p63 and p53 endogenous protein expression in SH-SY5Y, IMR-32, LAN-1 and MCF-7 cell lines. Total protein extracts were obtained as previously described (Goldschneider et al., 2004) and 30 µg total lysate was loaded for western blotting. The blot on the right was overexposed for 30 minutes rather than 0.5 minutes as for the blots on the left. ∆Np73α transactivates BTG2TIS21/PC3 via p53 1249 cells (more than tenfold when A B 100 120 SH-SY5Y MCF-7 compared to empty vector

100 80 infected cells) (Fig. 4B). These + empty vector + empty vector results, combined with those

80 (A. U.) + ∆Np73α 60 + ∆Np73α obtained by transfection, + ∆Np73β + ∆Np73β 60 strongly suggest that the + p53DD 40 + p53DD upregulation of the 40 BTG2TIS21/PC3 promoter could 20 TIS21/PC3 20 account for the BTG2 Luciferase activity Luciferase activity (A. U.) mRNA increase in Ad- 0 0 ∆Np73α-infected SH- pE1-hWAF1 pE1-hWAF1 SY5Ycells. Parallel experiments were Fig. 3. Transcriptional activity of Waf1/p21 p53BS is inhibited by ∆Np73. The mean relative luciferase performed in the MCF-7 breast activity±s.e.m. is shown for wt-p53 SH-SY5Y (A) and MCF-7 (B) cells co-transfected with 0.5 µg pE1- cancer cell line. Comparable hWAF1 and 1 µg of plasmid encoding either ∆Np73α or ∆Np73β. The p53DD effect is also shown. results were obtained for Waf1/p21 with a marked increase in Waf1/p21 mRNA ∆Np73α overexpression induces an increase of levels in response to Ad-p73α infection but not to Ad-∆Np73α endogenous BTG2TIS21/PC3 mRNA transcript in SH- infection (Fig. 4C). However, consistent with the luciferase SY5Y but not MCF-7 cells results shown in Fig. 1E, a relatively low increase in the level TIS21/PC3 To confirm the results obtained with reporter plasmids, the of BTG2 mRNA was found in Ad-p73α-infected MCF- endogenous BTG2TIS21/PC3 mRNA levels were compared to that 7 cells and no significant increase was found in parallel cells of Waf1/p21 in SH-SY5Y cells infected with either Ad- infected with Ad-∆Np73α (Fig. 4D). TAp73α or Ad-∆Np73α. The results obtained using SYBR®- Green RTQ-PCR are presented in Fig. 4. As expected, Waf1/p21 expression was strongly upregulated by TAp73α and ∆Np73-mediated stimulation of p53-dependent TIS21/PC3 significantly repressed by the dominant-negative mutant BTG2 transcription required the p73α C- ∆Np73α (Fig. 4A). Very interestingly, however, unlike terminus Waf1/p21, BTG2TIS21/PC3 endogenous transcript levels were It is well known that p73β is more active in transcription than comparably higher in both the TAp73α and ∆Np73α-infected p73α owing to a repressive effect of the C-terminus. We

A 140 SH-SY5Y/Waf1 B 40 SH-SY5Y/BTG2

Journal of Cell Science 120 35

100 30 Uninfected 25 Uninfected 80 TAp73α TAp73α ression/GAPDH 20 60 ∆Np73α ∆Np73α

A expression/GAPDH 15 Empty vector Empty vector 40 10 20 5 p21 mRNAexp 0 BTG2 mRN 0

C 300 MCF-7/Waf1 D MCF-7/BTG2 10 250

200 Uninfected Uninfected TAp73α TAp73α ression/GAPDH 150 ∆Np73α 5 ∆Np73α

100 Empty vector A expression/GAPDH Empty vector

50 p21 mRNAexp 0 BTG2 mRN 0

Fig. 4. Effect of ectopic expression of either TA- or ∆Np73 on the endogenous expression of Waf1/p21 and BTG2TIS21/PC3 in wt-p53-expressing SH-SY5Y and MCF-7 cell lines. The transcript levels of Waf1/p21 (A,C) and BTG2TIS21/PC3 (B,D) were estimated by SYBR®-Green RTQ-PCR from total RNA extracted from SH-SY5Y (A,B) or MCF-7 (C,D) cells infected with either TA- or ∆Np73α-recombinant adenovirus. Results are mean expression levels±s.e.m. compared with those obtained from either uninfected cells or cells infected with an empty vector 1250 Journal of Cell Science 118 (6)

A 180 B 450 SH-SY5Y + empty vector IMR-32 + empty vector 160 400 p53+/+ + ∆Np73β p53+/+ + ∆Np73β 140 350 120 + p53DD 300 + p53DD

100 250 80 200 60 150 40 100

Luciferase (A. activity U.) 20 50 Luciferase (A. activity U.) 0 0 BTG2 2700 BTG2 266 BTG2 2700 BTG2 266

C 500 MCF-7 + empty vector 450 p53+/+ ∆ β 400 + Np73 350 + p53DD 300 250 200 150 100 Luciferase (A. activity U.) 50 0 BTG2 2700 BTG2 266

Fig. 5. BTG2 transactivation is not induced by p73β isoforms. Transcriptional activity of BTG2TIS21/PC3 was analyzed by luciferase assays in wt-p53 SH-SY5Y (A), IMR-32 (B) and MCF-7 (C) cells co-transfected with 0.5 µg of luciferase reporter plasmid (BTG2 2700 or BTG2 266) and 1 µg of either ∆Np73β or p53DD expression plasmids. The values represent mean relative luciferase activity±s.e.m.

wondered whether the C-terminal domain had any effect on luciferase results. Nevertheless, p53 accumulation in SH- BTG2TIS21/PC3 activation by ∆Np73. To this end, we performed SY5Y was observed in both the Ad-∆Np73β and Ad-∆Np73α- luciferase assays using long and short promoter sequences in infected cells (Fig. 7), indicating that the presence of wt-p53 cells transfected by ∆Np73β lacking the SAM domain, and is somewhat necessary but not sufficient for BTG2TIS2/PC3 compared these results to results obtained from the co- upregulation, and suggesting that p73 C-terminus sequences transfection of the p53 dominant-negative mutant, p53DD. might enhance p53 binding with the BTG2 p53-responsive Journal of Cell Science Unlike ∆Np73α, ∆Np73β isoforms did not stimulate the element. When combined, our findings show that BTG2TIS2/PC3 activity of the BTG2TIS21/PC3 promoter sequences in two wt- expression is upregulated in wt-p53 neuroblastoma cells by p53-expressing neuroblastoma cell lines, SH-SY5Y and IMR- ∆Np73α but not by ∆Np73β. The role of p73β, which is known 32 (Fig. 5A,B). Interestingly, the decrease in luciferase activity to be transcriptionally more active than the α isoform, may not is much more pronounced in cells expressing p53DD than in be systematic and should perhaps be revisited. ∆Np73α, thus, those expressing ∆Np73β. One hypothesis to explain these may play a role in upregulating certain genes like BTG2TIS2/PC3 results would be that, in these particular cells, the affinity of in addition to its known dominant-negative role. p53 for the BTG2 p53-responsive element is higher than that A second transactivating domain called (TA2), which is of ∆Np73β. This is consistent with the fact that ∆Np73α does different from the well-known N-terminus TAD domain, has not act as a dominant-negative mutant of p53 in either of these been identified in the C-terminus of the p73 protein (Takada two cell lines. In contrast, the decrease in luciferase activity in et al., 1999). However, this TA2 domain is present both in MCF-7 cells is even more pronounced with ∆Np73β than with p73α and β, and cannot, therefore, account for the stimulation p53DD (Fig. 5C), suggesting once again that the BTG2TIS21/PC3 of BTG2 promoter activity by the ∆Np73α protein. It has promoter response to p53 and p73 expression is cell specific. been demonstrated more recently that a stable expression of The same observation was noted when BTG2TIS2/PC3 mRNA inducible ∆Np73β was able to suppress cell growth and levels were analyzed by RTQ-PCR in cells infected with induce apoptosis, even at a physiological level (Liu et al., ∆Np73β. As already shown in Fig. 4, BTG2TIS2/PC3 2004). The authors attributed this transactivating activity to upregulation was observed in SH-SY5Y cells infected with the N-terminal part of the ∆Np73β protein. Interestingly, Ad-∆Np73α but not in MCF-7 cells (Fig. 4B and 4D ∆Np73α harbors the same N-terminus as the β isoform but, respectively). To confirm that the presence of C-terminus according to their experiments, it was unable to transactivate sequences are necessary for BTG2TIS2/PC3 activation, BTG2 promoters, possibly owing to the established repressor role of mRNA levels were quantified by RTQ-PCR in SH-SY5Y and the p73α C-terminus. So, once again, such a model fails to MCF-7 infected with either Ad-∆Np73β or Ad-∆Np73α. There explain our results. Furthermore, the ∆Np73β transactivating was six times less mRNA in SH-SY5Y cells infected with the potential reported in Liu’s study (Liu et al., 2004), was found β isoform than in those infected with the α isoform, which is to be p53-independent as it can be induced in p53-null H1299 in contrast to MCF-7 (Fig. 6A,B). This is consistent with the cells as well as in wt-p53-expressing MCF-7 cells. Our results ∆Np73α transactivates BTG2TIS21/PC3 via p53 1251

A 25 SH-SY5Y

20 Uninfected TAp73α 15 ∆Np73α TAp73β 10 ∆Np73β Empty vector 5 BTG2 mRNAexpression/GAPDH BTG2

0

Fig. 7. Western blot showing an accumulation of p53 protein in α β B 10 response to an ectopic expression of either the or isoforms of MCF-7 TA- and ∆Np73 proteins.

Uninfected a TAp73α 5 ’PuPuPuC(A/T)(A/T)GPyPyPy N(0 - 13) PuPuPuC(A/T)(A/T)GPyPyPy 3 ’ ∆Np73α

pression/GAPDH 5 TAp73β b ∆Np73β 5 ’ PuPuPuC(A/T)GPuPuPuC(A/T) N(12) PuPuPuC(A/T)PuPuPuC(A/T) 3 ’ Empty vector c 5 ’ (A/T)GPuPuPuPyPyPyC(A/T) N (A/T)GPuPuPuPyPyPyC(A/T) 3 ’ BTG2 mRNAex 0 Fig. 8. Schematic representation of the p53BS of three p53-target genes showing the pentamer orientation in each decamer. (a) Head- Fig. 6. No accumulation of BTG2 endogenous mRNA in Ad-p73β- to-head (HH) for the canonical 20-bp p53-binding site, as in the infected SH-SY5Y cells. Comparative RTQ-PCR analyses of Waf1/p21 promoter; (b) in tandem, head-to-tail (HT), as in the MDR- TIS21/PC3 BTG2 mRNA levels in SH-SY5Y (A) and MCF-7 cells (B) 1 promoter; and (c) tail-to-tail (TT) as in the BTG2TIS21/PC3 promoter. infected with either p73α or p73β isoforms. Pu, purine; Py, pyrimidine.

showed that BTG2TIS21/PC3 activation by ∆Np73α is p53- ∆Np73α and p53 form heterotetramers which are dependent. transcriptionally inactive on the canonical p53BS, and active Journal of Cell Science (through the adoption of a different conformation) on the BTG2TIS21/PC3 promoter p53BS. Such a conformational change TIS21/PC3 BTG2 transactivation may result from the could possibly be due to the p73 C-terminus domain. However, unusual orientation of the pentamers in the p53 such a model is unlikely according to Davison and colleagues responsive element (Davison et al., 1999), who demonstrated that p73 and p53 This study shows that the ∆Np73α isoform activates proteins only interact weakly. A second possibility, therefore, BTG2TIS21/PC3 through its promoter in a p53-dependent manner could be that p73α is unable to bind the non-canonical p53BS in neuroblastoma wt-p53 cells. Transcriptional activation by of BTG2TIS21/PC3, which means that ∆Np73α would not be able p53 requires the fixation of p53 to its consensus binding site to compete with p53 in terms of binding to its responsive (p53BS). The p53BS consists of a canonical sequence element and could, therefore, not repress the p53-dependent comprising two decamers ‘PuPuPuC(A/T)(A/T)GpyPyPy’ transactivation of the BTG2TIS21/PC3 promoter. separated by 0 to 13 base pairs. The two pentamers in each Earlier studies of ours showed that both ∆Np73α and decamer are arranged head to head (HH) (Fig. 8a). The TAp73α were able to induce accumulation and activation of Waf1/p21 promoter is the paradigm of this kind of base the endogenous wt-p53 in various cell lines (Miro-Mur et al., arrangement (El-Deiry, 1998). Alternately, p53 can repress 2003; Goldschneider et al., 2004). The present data indicate the transcription of other genes such as MDR1. The p53BS that ∆Np73α stimulates the p53-dependent activation of harbored by the MDR1 promoter presents a particular BTG2TIS21/PC3 expression at the promoter level. The molecular orientation, with a head-to-tail (HT) pentamer arrangement mechanism by which BTG2TIS21/PC3 sequences might be (Fig. 8b) (Johnson et al., 2001). The authors proposed that this transactivated by ∆Np73α remains to be elucidated. Two orientation could affect p53 protein conformation and lead to possible mechanisms can be proposed. The first model involves an inactive p53. For BTG2TIS21/PC3, the pentamer organization a cooperative effect of ∆Np73α and p53 acting on BTG2- of the p53BS, with a tail-to-tail (TT) arrangement, is again p53BS but not on a canonical p53BS such as that of Waf1/p21. different from that of the canonical p53BS (HH) and MDR1 Such a model could be instrumental in explaining that ∆Np73α promoter (HT) (Fig. 8c). This specific arrangement could not only does not inhibit BTG2TIS21/PC3 but that it also activates account for the particular response of the BTG2TIS21/PC3 it. The second model excludes ∆Np73α-BTG2 promoter promoter to the ∆Np73α protein. One possibility could be that binding because of the BTG2-p53BS TT orientation, which 1252 Journal of Cell Science 118 (6)

means that it could not inhibit BTG2TIS21/PC3 activation as it Courtois, S., Wang, Q., Soussi, T. and Puisieux, A. (2002). The human could for Waf1/p21, where the p53BS is HT orientated. In this BTG2/TIS21/PC3 gene: genomic structure, transcriptional regulation and alternative, the stimulation of BTG2TIS21/PC3 gene expression evaluation as a candidate tumor suppressor gene. Gene 282, 207-214. α El-Deiry, W. (1998). Regulation of p53 downstream genes. Cancer Biology 8, could result from the p73 -dependent activation of p53. 345-357. Indeed, we have already published that an ectopic expression el Ghissassi, F., Valsesia-Wittmann, S., Falette, N., Duriez, C., Walden, P. of p73α activates the transcriptional activity of the endogenous D. and Puisieux, A. (2002). BTG2(TIS21/PC3) induces neuronal wt-p53 expressed in several transformed cell lines, differentiation and prevents apoptosis of terminally differentiated PC12 independently of its own transcriptional activity (Miro-Mur et cells. Oncogene 21, 6772-6778. Ferrandis, E., da Silva, J., Riou, G. and Bénard, J. (1994). Coactivation of al., 2003). We are currently looking for the mechanism that the MDR 1 and MYCN genes in human neuroblastoma cells during the could account for p53 activation by p73α. metastatic process in the nude mouse. Cancer Res. 15, 2256-2261. Fillippovich, I., Sorokina, N., Gatei, M., Haupt, Y., Hobson, K., Moallem, E., Spring, K., Mould, M., McGuckin, M. A., Lavin, M. F. et al. (2001). Conclusion Transactivation-deficient p73alpha (p7∆exon2) inhibits apoptosis and ∆ α competes with p53. Oncogene 20, 514-522. Our current findings, which demonstrate that it is the Np73 Goldschneider, D., Blanc, E., Raguenez, G., Haddada, H., Benard, J. and isoform and not the ∆Np73β isoform that activates the Douc-Rasy, S. (2003). When p53 needs p73 to be functional – forced p73 BTG2TIS21/PC3, are particularly interesting in terms of gene expression induces nuclear accumulation of endogenous p53 protein. regulation and the physiology of neuroblastoma cells. They Cancer Lett. 197, 99-103. suggest that, by activating the BTG2TIS21/PC3, the ∆Np73α Goldschneider, D., Blanc, E., Raguenez, G., Barrois, M., Legrand, A., le Roux, G., Haddada, H., Benard, J. and Douc-Rasy, S. (2004). Differential isoform might function, not only as an anti-apoptotic mediator response of p53 target genes to p73 overexpression in SH-SY5Y in the neuronal system in vivo or as a transdominant negative neuroblastoma cell line. J. Cell Sci. 117, 293-301. in human tumors as previously reported, but may also play Gong, J. G., Costanzo, A., Yang, H. Q., Melino, G., Kaelin, W. G., Jr., an antiproliferative and/or differentiating role in neural Levrero, M. and Wang, J. Y. (1999). The tyrosine kinase c-Abl regulates development and in neuroblastoma. p73 in apoptotic response to cisplatin-induced DNA damage. Nature 399, 806-809. Iacopetti, P., Barsacchi, G., Tirone, F., Maffei, L. and Cremisi, F. (1994). We would like to express our sincerest gratitude to Mourad Kaghad Developmental expression of PC3 gene is correlated with neuronal cell for his kind donation of the pcDNA-p73 plasmid constructs and to birthday. Mech. Dev. 47, 127-137. Eric Le Cam for his helpful contributions. This work was supported Johnson, R. A., Ince, T. A. and Scotto, K. W. (2001). Transcriptional by Ligue Contre le Cancer, Comité du Cher, l’Association pour la repression by p53 through direct binding to a novel DNA element. J. Biol. Recherche sur le Cancer (ARC) and Bonus Qualité Recherche (BQR) Chem. 276, 27716-27720. from University Paris-Sud. Edited by English Booster. Jost, C. A., Marin, M. C. and Kaelin, W. G., Jr. (1997). p73 is a simian [correction of human] p53-related protein that can induce apoptosis. Nature 389, 191-194. References Kaghad, M., Bonnet, H., Yang, A., Creancier, L., Biscan, J. C., Valent, A., Minty, A., Chalon, P., Lelias, J. M., Dumont, X. et al. (1997). Agami, R., Blandino, G., Oren, M. and Shaul, Y. (1999). Interaction of c- Monoallelically expressed gene related to p53 at 1p36, a region frequently Abl and p73alpha and their collaboration to induce apoptosis. Nature 399, deleted in neuroblastoma and other human cancers. Cell 90, 809-819. 809-813.

Journal of Cell Science Kartasheva, N. N., Contente, A., Lenz-Stoppler, C., Roth, J. and Bénard, J., Douc-Rasy, S. and Ahomadegbe, J. C. (2003). TP53 family Dobbelstein, M. (2002). p53 induces the expression of its antagonist members and human cancers. Hum. Mutat. 21, 182-191. p73∆N, establishing an autoregulatory feedback loop. Oncogene 21, 4715- Bradbury, A., Possenti, R., Shooter, E. M. and Tirone, F. (1991). Molecular 4727. cloning of PC3, a putatively secreted protein whose mRNA is induced by Liu, G., Nozell, S., Xiao, H. and Chen, X. (2004). DeltaNp73beta is active nerve growth factor and depolarization. Proc. Natl. Acad. Sci. USA 88, 3353- in transactivation and growth suppression. Mol. Cell. Biol. 24, 487-501. 3357. Melamed, J., Kernizan, S. and Walden, P. D. (2002). Expression of B-cell Casciano, I., Mazzocco, K., Boni, L., Pagnan, G., Banelli, B., Allemanni, translocation gene 2 protein in normal human tissues. Tissue Cell 34, 28- G., Ponzoni, M., Tonini, G. P. and Romani, M. (2002). Expression of DeltaNp73 is a molecular marker for adverse outcome in neuroblastoma 32. patients. Cell Death Differ. 9, 246-251. Miro-Mur, F., Meiller, A., Haddada, H. and May, E. (2003). p73alpha Chi, S. W., Ayed, A. and Arrowsmith, C. H. (1999). Solution structure of a expression induces both accumulation and activation of wt-p53 independent conserved C-terminal domain of p73 with structural homology to the SAM of the p73alpha transcriptional activity. Oncogene 22, 5451-5456. domain. EMBO J. 18, 4438-4445. Munsch, D., Watanabe-Fukunaga, R., Bourdon, J. C., Nagata, S., May, E., Davison, T. S., Vagner, C., Kaghad, M., Ayed, A., Caput, D. and Yonish-Rouach, E., and Reisdorf, P. (2000). Human and mouse Fas (APO- Arrowsmith, C. H. (1999). p73 and p63 are homotetramers capable of weak 1/CD95) death receptor genes each contain a p53-responsive element that heterotypic interactions with each other but not with p53. J. Biol. Chem. is activated by p53 mutants unable to induce apoptosis. J. Biol. Chem. 275, 274, 18709-18714. 3867-3872. De Laurenzi, V., Costanzo, A., Barcaroli, D., Terrinoni, A., Falco, M., Morel, A. P., Sentis, S., Bianchin, C., le Romancer, M., Jonard, L., Rostan, Annicchiarico-Petruzzelli, M., Levrero, M. and Melino, G. (1998). Two M. C., Rimokh, R. and Corbo, L. (2003). BTG2 antiproliferative protein new p73 splice variants, gamma and Delta, with different transcriptional interacts with the human CCR4 complex existing in vivo in three cell-cycle- activity. J. Exp. Med. 188, 1763-1768. regulated forms. J. Cell Sci. 116, 2929-2936. Douc-Rasy, S., Barrois, M., Echeynne, M., Kaghad, M., Blanc, E., Nakagawa, T., Takahashi, M., Ozaki, T., Watanabe, K. K., Todo, S., Raguenez, G., Goldschneider, D., Terrier-Lacombe, M. J., Hartmann, Mizuguchi, H., Hayakawa, T. and Nakagawara, A. (2002). O., Moll, U. et al. (2002). ∆N-p73alpha accumulates in human neuroblastic Autoinhibitory regulation of p73 by ∆Np73 to modulate cell survival and tumors. Am. J. Pathol. 160, 631-639. death through a p73-specific target element within the DeltaN-p73 promoter. Drané, P., Bravard, A., Bouvard, V. and May, E. (2001). Reciprocal down- Mol. Cell. Biol. 22, 2575-2585. regulation of p53 and SOD2 gene expression-implication in p53 mediated Ozaki, T., Naka, M., Takada, N., Tada, M., Sakiyama, S. and Nakagawara, apoptosis. Oncogene 20, 430-439. A. (1999). Deletion of the COOH-terminal region of p73alpha enhances Drané, P., Leblanc, V., Miro-Mur, F., Saffroy, R., Debuire, B. and May, E. both its transactivation function and DNA-binding activity but inhibits (2002). Accumulation of an inactive form of p53 protein in cells treated with induction of apoptosis in mammalian cells. Cancer Res. 59, 5902-5907. TNF alpha. Cell Death Differ. 9, 527-537. Rouault, J. P., Falette, N., Guehenneux, F., Guillot, C., Rimokh, R., Wang, Duriez, C., Falette, N., Audoynaud, C., Moyret-Lalle, C., Bensaad, K., Q., Berthet, C., Moyret-Lalle, C., Savatier, P., Pain, B. et al. (1996). ∆Np73α transactivates BTG2TIS21/PC3 via p53 1253

Identification of BTG2, an antiproliferative p53-dependent component of M. and Costanzo, A. (2002). ∆N-p73 is activated after DNA damage in a the DNA damage cellular response pathway. Nat. Genet. 14, 482-486. p53-dependent manner to regulate p53-induced cell cycle arrest. Oncogene Stiewe, T., Theseling, C. C. and Putzer, B. M. (2002). Transactivation- 21, 3796-3803. deficient Delta TA-p73 inhibits p53 by direct competition for DNA binding: Yang, A., Kaghad, M., Caput, D. and McKeon, F. (2002). On the shoulders implications for tumorigenesis. J. Biol. Chem. 277, 14177-14185. of giants: p63, p73 and the rise of p53. Trends Genet. 18, 90-95. Takada, N., Ozaki, T., Ichimiya, S., Todo, S. and Nakagawara, A. (1999). Zaika, A. I., Slade, N., Erster, S. H., Sansome, C., Joseph, T. W., Pearl, M., Identification of a transactivation activity in the COOH-terminal region of Chalas, E. and Moll, U. M. (2002). DeltaNp73, a dominant-negative p73 which is impaired in the naturally occurring mutants found in human inhibitor of wild-type p53 and TAp73, is upregulated in human tumors. J. neuroblastomas. Cancer Res. 59, 2810-2814. Exp. Med. 196, 765-780. Tirone, F. (2001). The gene PC3(TIS21/BTG2), prototype member of the Zeng, X., Chen, L., Jost, C. A., Maya, R., Keller, D., Wang, X., Kaelin, W. PC3/BTG/TOB family: regulator in control of cell growth, differentiation, G., Jr., Oren, M., Chen, J. and Lu, H. (1999). MDM2 suppresses p73 and DNA repair? J. Cell Physiol 187, 155-165. function without promoting p73 degradation. Mol. Cell. Biol. 19, 3257- Vossio, S., Palescandolo, E., Pediconi, N., Moretti, F., Balsano, C., Levrero, 3266. Journal of Cell Science