HIPK2 Neutralizes MDM2 Inhibition Rescuing P53 Transcriptional Activity and Apoptotic Function

HIPK2 neutralizes MDM2 inhibition rescuing p53 transcriptional activity and apoptotic function

Valeria Di Stefano1, Giovanni Blandino1, Ada Sacchi1, Silvia Soddu1 and Gabriella D’Orazi1,2,*

1Department of Experimental Oncology, Molecular Oncogenesis Laboratory, Regina Elena Cancer Institute, Rome 00158, Italy; 2Department of Oncology and Neurosciences, University ‘G D’Annunzio’, Chieti 66013, Italy

The p53 oncosuppressor protein is subject to negative Oren, 1996). The activation of p53 protein occurs at regulation by MDM2,which efficiently inhibits its activity multiple levels including protein stabilization, nuclear through an autoregulatory loop. In response to stress, accumulation, and post-translational modifications (Ap- however,p53 undergoes post-translational modifications pella and Anderson, 2001). The biological effects of p53 that allow the protein to escape MDM2 control, are largely due to its sequence-specific binding to DNA accumulate,and become active. Recent studies have and transcriptional activation of genes involved in cell- shown that,following DNA damage,the HIPK2 serine/ cycle regulation and apoptosis (Gottlieb and Oren, threonine kinase binds and phosphorylates p53,inducing 1996). The transactivation domain of p53 is located at p53 transcriptional activity and apoptotic function. Here, the N-terminus of the protein, interacts with compo- we investigated the role of HIPK2 in the activation of p53 nents of the basal transcription machinery, and pro- in the presence of MDM2. We found that HIPK2 rescues motes the transcription of genes containing p53-binding p53 transcriptional activity overcoming MDM2 inhibi- sequences (Unger et al., 1992; Lu and Levine, 1995). tion,and that restoration of this p53 function induces How p53 activation and stabilization occur following apoptosis. Recovery of p53-dependent apoptosis is stress is still not completely understood. achieved by preventing p53 nuclear export and ubiquitina- A loss of p53 function, either by mutations or by tion mediated by MDM2 in vitro and in vivo following deregulation of regulatory proteins, is often detrimental, genotoxic stress. These results shed new light on the leading to tumor development and progression mechanisms by which the HIPK2/p53 pathway promotes (Valculescu and El Deiry, 1996; Vogt Sionov and apoptosis and suppression of tumorigenesis. Haupt, 1999). One of the key regulatory elements of Oncogene (2004) 23, 5185–5192. doi:10.1038/sj.onc.1207656 p53 is mdm2, a cellular proto-oncogene (Oliner et al., Published online 3 May 2004 1992). MDM2 restrains p53 function by concealing its transcriptional activation domain, thereby inhibiting Keywords: HIPK2; p53; ubiquitination; MDM2; nucle- p53-mediated transactivation (Momand et al., 1992). ar export; apoptosis MDM2 is also able to induce p53 nuclear export, preventing the access of p53 to DNA (Freedman and Levine, 1998; Roth et al., 1998; Tao and Levine, 1999). Finally, MDM2 is an E3–ubiquitin ligase able to target p53 for ubiquitination and rapid degradation via Introduction proteasome (Haupt et al., 1997; Kubbutat et al., 1997). The expression of the mdm2 gene is regulated by p53 The tumor suppressor p53 is important in the cellular itself in a negative feedback loop (Barak et al., 1993; Wu response to genotoxic damage. In normal cells, p53 is et al., 1993). It is worth noting that the mdm2 gene is present at low concentration and is kept silent. Follow- amplified in a significant proportion of human tumor ing stress stimuli, such as hypoxia or DNA damage, p53 types, thereby contributing to tumor development by accumulates in the nucleus and becomes activated efficiently reducing the availability of a functional p53 inducing several different responses such as growth protein (Oliner et al., 1992; Momand et al., 1998). The arrest, senescence, apoptosis, and DNA repair (Giaccia MDM2-negative regulation of p53 protein can be and Kastan, 1998; Lakin and Jackson, 1999; Prives and neutralized by partner proteins and by specific protein Hall, 1999). These activities contribute to the role of p53 modifications. Several studies have shown that phos- as oncosuppressor by preventing the establishment of phorylation of the p53 protein can regulate its functions mutations in future generation of cells (Gottlieb and in vivo and in vitro (Appella and Anderson, 2001). In particular, DNA-damage-induced phosphorylation of serine and threonine residues at the N-terminus con- *Correspondence: G D’Orazi, Molecular Oncogenesis Laboratory, tributes to p53 stability by preventing MDM2 from Regina Elena Cancer Institute, Via delle Messi d’Oro 156, Rome 00158, Italy; E-mail: [email protected] degrading it (Shieh et al., 1997; Unger et al., 1999). It is Received 27 November 2003; revised 16 February 2004; accepted 16 quite evident that one way to stabilize and activate p53 February 2004; published online 3 May 2004 in response to stress is by interfering either with the HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5186 interaction between MDM2 and p53 or with the ability of MDM2 to target bound p53 for degradation. The mechanisms of p53 stabilization and activation are, however, more complex and might also involve stress- induced modification of MDM2 and post-translational modifications of other interacting proteins (Appella and Anderson, 2001). Several kinases have been identified that detect genotoxic stress and initiate signaling pathways through p53 phosphorylation (Fuchs et al., 1998; Appella and Anderson, 2001). A series of different phosphorylation sites in the p53 protein are presumed to be involved in p53 oncosuppressing functions (Siliciano et al., 1997). A novel serine/threonine kinase, namely HIPK2, has been recently shown to activate the p53 protein through its direct binding and phosphorylation at residue Ser46 (D’Orazi et al., 2002; Hofmann et al., 2002). Phosphor- ylation at this site is a late event after severe DNA damage and specifically regulates p53-induced apoptosis (Bulavin et al., 1999; Oda et al., 2000b). We have recently found that HIPK2 can trigger p53-mediated apoptosis through phosphorylation of p53 at Ser46 after treatment with the chemotherapeutic drug cisplatin; this finding supports a key role for HIPK2 in the stimulation of p53 in response to genotoxic stress (Di Stefano et al., 2004). Recently, exogenous HIPK2 overexpression was shown to correlate with the reduction of MDM2 protein levels (Wang et al., 2001). Since MDM2 is the major player in p53-negative regulation and HIPK2 is a positive regulator of p53, we investigated the possibility that HIPK2 may antagonize the inhibitory effect of MDM2 on p53. Here, we demonstrate that HIPK2 can Figure 1 HIPK2 rescues p53 transcriptional activity overcoming efficiently rescue p53 transcriptional activity over MDM2 inhibition. (a)H1299 cells were transfected with expression MDM2 inhibition. This action of HIPK2 involves the vectors encoding for p53 (1 mg), MDM2 (3 mg), HIPK2 (3, 5, 10 mg), neutralization of MDM2-dependent p53 ubiquitination and K221R (8 mg)proteins, along with the bax-luciferase reporter and nuclear export in vitro and in vivo following DNA vector (1 mg), using the BES method. At 24 h after transfection, damage, with induction of p53 apoptotic function. cells were harvested and luciferase activity determined following normalization to b-gal activity. (b)Same as in ( a), using the PIG3- luc reporter vector. (c)Saos2 cells were transfected as in ( a)with the Noxa-luciferase reporter vector. (d)H1299 cells were transfected with expression vectors encoding for p53S46A mutant, Results MDM2, and HIPK2 proteins. Results are shown as fold of induction of luciferase activity over control sample transfected with HIPK2 abrogates MDM2-mediated repression of p53 empty vector. Standard deviation (s.d.)of triplicate is indicated. ( e) transcriptional activity H1299 cells were transfected with p53 (1 mg), MDM2 (4 mg), and HIPK2 (8 mg)expression vectors. At 24 h after transfection, equal The nuclear serine–threonine kinase HIPK2 is able to amounts of cell proteins were subjected to immunoprecipitation induce p53 transcriptional activity (Wang et al., 2001; analysis with anti-p53 monoclonal antibodies (DO1 and Ab1801). Western blot analysis was performed with anti-p53 polyclonal D’Orazi et al., 2002; Hofmann et al., 2002; Kim et al., (FL393)and anti-MDM2 monoclonal antibodies 2002). To investigate whether HIPK2 has the ability of activate p53 even in the presence of overexpressed MDM2, we used reporter constructs containing the relative to the empty vector, while cotransfection with luciferase gene under the control of the bax, PIG3,or MDM2 reduced the ability of p53 to transactivate the Noxa gene promoters (Miyashita and Reed, 1995; reporter construct, consistent with previous studies Polyak et al., 1997; Oda et al., 2000a). H1299 cells were (Haupt et al., 1997). HIPK2 expression increased p53- cotransfected with various combinations of expression mediated transactivation and strongly rescued the p53 vectors encoding for p53, MDM2, HIPK2, or the transcriptional activity, in a dose-dependent manner, HIPK2 kinase dead mutant-K221R (which is not able counteracting the inhibitory effect of MDM2. In to phosphorylate p53 at Ser46 and also does not induce contrast to this, K221R failed to neutralize the p53-mediated transactivation; Wang et al., 2001; D’Or- inhibitory effect of MDM2 (Figure 1a). Comparable azi et al., 2002; Hofmann et al., 2002)along with the results were obtained with PIG3 reporter vector bax-andPIG3-luciferase reporters. As shown in (Figure 1b)and with Saos2 cells cotransfected with the Figure 1a, p53 expression efficiently induced bax activity luciferase gene under the control of Noxa promoter

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5187 (Figure 1c). Consistent with the HIPK2 kinase activity Localization of p53 was almost entirely nuclear when on p53 at Ser46, the transcriptional activity of a expressed alone, but also became cytoplasmic when p53S46A mutant was not modified by HIPK2 either in coexpressed with MDM2 (Figure 2a). Indeed, under this the absence or in the presence of MDM2 (Figure 1d). condition, the proportion of cells with cytoplasmic We also investigated whether p53 was able to interact staining increased almost four-fold (Figure 2b), consis- with MDM2 in the presence of HIPK2. We transfected tent with previous findings (Freedman and Levine, 1998; H1299 cells with p53, p53 and MDM2, or both together Roth et al., 1998; Tao and Levine, 1999). Interestingly, with HIPK2 expression vectors. At 24 h after transfec- in the presence of both HIPK2 and MDM2 proteins, tion, cells were lysed and equal amounts of total cell p53 was mostly detected in the nuclei (i.e. only 10% of extracts were immunoprecipitated with anti-p53 mono- the cells had cytoplasmic staining), as with p53 alone, clonal antibodies. Figure 1e shows that MDM2 was no while, in the presence of both K221R and MDM2, p53 longer able to interact with p53 in the presence of was cytoplasmic in a larger number of cells (Figure 2a HIPK2. and b), indicating that HIPK2 can block MDM2- Taken together, these results suggest that HIPK2 can mediated p53 nuclear export. rescue p53 transcriptional activity over MDM2 inhibition, and that the HIPK2 kinase domain plays a role in HIPK2 attenuates ubiquitination of p53 mediated this effect. Moreover, HPK2 is able to change the by MDM2 interaction between MDM2 and p53. To investigate whether HIPK2 can interfere with p53 HIPK2 prevents MDM2-mediated p53 nuclear export inhibition mediated by MDM2 through the ubiquitin– proteasome pathway (Haupt et al., 1997; Kubbutat The inhibition of p53 function by MDM2 requires, to a et al., 1997), we transfected H1299 cells with p53, p53 larger extent, the nuclear export of p53 (Freedman and and MDM2, or both together with HIPK2 or K221R Levine, 1998; Roth et al., 1998; Tao and Levine, 1999). expression vectors. At 24 h after transfection, cells were Thus, we asked whether HIPK2 might interfere with the either lysed to analyse p53 protein levels by direct MDM2-mediated nuclear/cytoplasmic shuttling of p53. Western blotting or treated with the proteasome To address this question, H1299 cells were cotransfected inhibitor MG132, to prevent p53 degradation, prior to with vectors expressing p53 in combination with p53 immunoprecipitation from equal amounts of total MDM2, HIPK2, and K221R and, 24 h later, fixed and cell extracts. Figure 3a shows that, in the absence of the stained for p53 immunofluorescence analysis. As shown proteasome inhibitor MG132, MDM2 was able to in Figure 2, the proportion of cells with p53 nuclear degrade p53, as already reported (Haupt et al., 1997), staining, compared with cells with both nuclear and while HIPK2 overexpression recovered p53 stabiliza- cytoplasmic staining, was scored for each combination. tion, even to a larger extent than the p53 overexpressing cells. The catalytically inactive K221R mutant showed an intermediate phenotype, being able to recover partially p53 stabilization. The p53 immunoprecipitation analyses of MG132-treated cells revealed a ladder of p53-specific high molecular weight bands and a smear, indicative of protein ubiquitination, almost exclusively when MDM2 was coexpressed with p53 and, in a lesser extent, when K221R mutant was added (Figure 3b), consistent with the stabilization results shown in Figure 3a. On the contrary, addition of HIPK2 strongly reduced the extent of p53 ubiquitination. Western blot analysis of the transduced proteins shows a downmodulation of MDM2 protein levels when MDM2 was coexpressed with p53 and HIPK2, as suggested (Wang et al., 2001)(Figure 3b, lower panel). These results show that HIPK2 protects p53 stabilization by preventing its in vivo ubiquitination by MDM2 and that HIPK2 activities other than the kinase one can be involved in this regulation. Figure 2 HIPK2 blocks the nuclear export of p53 by MDM2. (a) H1299 cells were transfected with the indicated combinations of expression vectors, using the lipofectaminePlus method. Subcel- HIPK2 rescues the p53 apoptotic function inhibited lular localization of p53, stained using anti-p53 monoclonal by MDM2 antibody DO1, was examined with a fluorescent microscopy. Cells were simultaneously stained for DNA using Hoechst. (b)Summary The clear effect of HIPK2 on MDM2-mediated inhibi- of p53 localization in stained cells (200–500 cells)from three tion of p53 prompted us to evaluate its biological independent experiments. The staining phenotype was categorized in two groups, one with nuclear p53 staining only and one with consequences. One of the most important functions of nuclear and cytoplasmic staining. The graph shows the percentage p53 is induction of apoptosis, and it is well known that of cells with p53 cytoplasmic staining7s.d. MDM2 inhibits it (Chen et al., 1996; Haupt et al., 1996)

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5188 ab 60 40

50 30 40

30 20

dead cells (%) 20 10 10 TUNEL positive cells (%) 0 0 − + ++++ CMV-p53 − + + + + CMV-p53 − − + − ++ CMV-MDM2 − − + − + CMV-MDM2 − − − − HIPK2-Flag − − − ++ HIPK2-Flag ++ K22IR-Flag − −−−−+ Figure 4 HIPK2 relieves the inhibitory effect of MDM2 on p53- mediated apoptosis. (a)H1299 cells were transfected in duplicate, using the lipofectaminePlus method, with p53 (100 ng), MDM2 (500 ng), and HIPK2 (1 mg). To equalize for the transfection efficiencies, soon after transfection, cells were trypsinized and pooled, and approximately equal number of cells were divided among five plates. Cell viability, of both floating and adherent cells, was then analysed by trypan blue exclusion 48 h later. The graph shows the percentage of dead cells. (b)For TUNEL assay cells were transfected as in (a). After 48 h, cells were fixed and stained for TUNEL reaction according the manufacturer’s manual (Boehrin- ger Mannheim, Germany). A summary of TUNEL-positive cells is shown as the percentage of stained cells out of almost 300 counted. Results are the mean of three different experiment7s.d.

HIPK2, and K221R vectors. In agreement with the viability data, we observed a consistent inhibitory effect of MDM2 on p53-induced TUNEL positivity. As shown in Figure 4b, MDM2 reduced the level of p53- mediated apoptosis by 50%, but the inhibition was relieved by coexpression of HIPK2, while the K221R Figure 3 HIPK2 protects p53 from MDM2-mediated ubiquitina- mutant failed to do so. These results suggest that HIPK2 tion. (a)H1299 cells were transfected with p53 (1 mg), MDM2 (4 mg), HIPK2 (8 mg), and K221R (8 mg)expression vectors, using protects p53 apoptotic activity from the inhibitory effect the BES method. At 24 h after transfection equal amounts of cell of MDM2 in a kinase-dependent manner. proteins were subjected to Western blot analysis with anti-p53 polyclonal antibody (FL393). Tubulin was used for protein loading control. (b)Cells were transfected as in ( a)and treated for 12 h with Inhibition of p53 activity mediated by MDM2 is relieved the proteasome inhibitor MG132 (10 mM). Equal amount of cell following cisplatin treatment proteins was subjected to immunoprecipitation analysis with anti- p53 monoclonal antibodies (DO1 and Ab1801)and Western blot To address the physiological relevance of the role played analysis was performed with anti-p53 polyclonal antibody (FL393). by HIPK2 during p53 activation, despite the presence of The position of the p53–ubiquitin (p53–Ubn)conjugates is MDM2, we took advantage of the recent finding that indicated. The lower panel shows a Western blot analysis of the expression of the transfected MDM2, p53, and Flag proteins. Anti HIPK2 plays an essential role in inducing p53-depen- b-gal monoclonal antibody was used for loading control of equally dent apoptosis in response to the antineoplastic drug transfected samples cisplatin (Kim et al., 2002; Di Stefano et al., 2004). To this end, RKO cells were transfected with the MDM2 expression vector (to mimic in vivo mdm2 amplification) whereas HIPK2 induces it (D’Orazi et al., 2002; and 24 h later treated with cisplatin. As shown in Hofmann et al., 2002). To examine whether and to Figure 5a, the expression of HIPK2 increased concomi- what extent HIPK2 interferes with MDM2 in this tantly to total p53 protein and Ser46 phosphorylation, process, cell viability was first analysed in H1299 cells as reported (Di Stefano et al., 2004), and remained high transfected with p53, either alone or together with despite the presence of MDM2. Interestingly, MDM2 MDM2, and HIPK2 vectors. As shown in Figure 4a, was not able to impair the HIPK2 ability to phosphor- induction of cell death by HIPK2 expression rescued the ylate p53 at Ser46. These results were confirmed on p53-mediated death inhibited by MDM2 overexpres- MG132-treated cells, in which p53 is stabilized indepen- sion. Apoptosis was then measured with the TUNEL dent of phosphorylation, indicating that p53 phosphor- assay. H1299 cells were transfected with relatively high ylation at Ser46 depends on cisplatin treatment, rather amount of p53 expression vector, using the efficient than protein accumulation per se (Figure 5b). lipofectaminePlus transfection method in order to To better elucidate whether the cisplatin-mediated induce at least 15% apoptosis, and with MDM2, HIPK2 induction could interfere with p53 ubiquitina-

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5189 shown in Figure 5c, Western blot analysis of total cell extracts revealed that treatment with cisplatin strongly reduced the extent of p53 ubiquitination in the pSuper cells (Figure 5c, in the left panel see the ladder of p53- specific high molecular weight bands detected exclusively in the p53 and MDM2 cotransfected sample). Interestingly, MDM2 overexpression along with treatment with cisplatin was not able to induce p53 ubiquitination. On the contrary, in the HIPK2-interfered cells, p53 was again ubiquitinated by MDM2 as shown by the ladder of p53-specific high molecular weight bands (Figure 5c, right panel). Altogether, these data suggest the requirement of HIPK2 for p53 stabilization after genotoxic damage and MDM2 overexpression. To investigate whether the cisplatin-mediated HIPK2 induction correlated with its ability of activate p53 apoptotic function, despite the presence of MDM2, we performed a luciferase assay. To this end, RKO cells were transfected with the expression vector encoding for MDM2 along with the bax-, PIG3-, and Noxa-luciferase reporters and treated with cisplatin. As shown in Figure 5d, cisplatin efficiently induced bax-, PIG3-, and Noxa-luciferase activity relative to the untreated cells; interestingly, MDM2 overexpression was not able to inhibit p53-mediated transcription of target genes induced by cisplatin treatment. Taken together, these results show that genotoxic stress activates p53-mediated transcription of apoptotic genes likely through HIPK2 kinase activity on p53Ser46 that renders p53 more resistant to in vivo ubiquitination by MDM2. Figure 5 Cisplatin-mediated HIPK2 induction relives p53 inhibition by MDM2. (a)Western blot analyses of HIPK2, MDM2, p53, and p53Ser46 phosphorylation on total cell proteins from RKO cells transfected with MDM2 expression vector, using the lipofectaminePlus method, and treated with cisplatin (5 mg/ml)for Discussion 24 h. Equal amount of proteins was separated on a denaturing SDS–PAGE and immunoblot analysis was performed using specific The tumor suppressor p53 induces cell death by antibodies. Blots were reprobed with antitubulin to ensure equivalence of loading. (b)Western blot analysis of RKO cells apoptosis in response to various stresses, such as treated with cisplatin (5 mg/ml)for 24 h and with the proteasome oncogene activation or DNA damage (Giaccia and inhibitor MG132 (10 mM)for 12 h. Equal amount of proteins was Kastan, 1998; Lakin and Jackson, 1999). The loss of p53 separated on a denaturing SDS—PAGE, and immunoblot analysis tumor-suppressor activity, either by mutation/deletion was performed using specific polyclonal anti-Ser46 and monoclonal of the TP53 gene or by inhibition of the p53 protein, anti-p53 (DO1)antibodies. ( c)Control RKO (pSuper)-or HIPK2- interfered (pSuper-HIPK2)cells were transfected with MDM2 favors the development of cancer (Gottlieb and Oren, vector and treated with cisplatin, as above. Subsequently, cells were 1996). Thus, regulation of p53 stability and activity and treated with the proteasome inhibitor MG132 (10 mm)for 12 h. its control on cell cycle and apoptosis are central to the Equal amount of proteins was subjected to Western blot analysis problem of human tumorigenesis. The recently identi- with anti-p53 polyclonal antibody (FL393). The positions of the fied function of HIPK2 as an activator of p53 (Wang p53–ubiquitin (p53–Ubn)conjugates are indicated. ( d)Induction of luciferase activity on RKO cells transfected with MDM2 expres- et al., 2001; D’Orazi et al., 2002; Hofmann et al., 2002; sion vector, using the lipofectaminePlus method, along with the Kim et al., 2002)raised the possibility that HIPK2 bax-, PIG3-, and Noxa-luciferase reporter vectors, and treated with might play an important role in the regulation of p53 cisplatin (5 mg/ml)for 24 h. All transfections were performed in apoptotic function. Indeed, in the present study, we duplicate. Results are the mean of three different experiment7s.d. demonstrate the involvement of HIPK2 in the rescue of a transcriptionally active p53 from MDM2 inhibition. We show that HIPK2 prevents the MDM2-mediated cytoplasmic shuttling and ubiquitination of p53, in vitro tion by MDM2, we transfected RKO-control (pSuper) and in vivo following genotoxic stress, neutralizing its and RKO-HIPK2-interfered cells (pSuper-HIPK2)(Di destabilization. Moreover, HPK2 is also able to change Stefano et al., 2004)with MDM2 expression vector. the interaction between MDM2 and p53. By these After 24 h, cells were treated with cisplatin for 24 h and, processes, HIPK2 is able to recover the p53 apoptotic subsequently, with the proteasome inhibitor MG132. As function.

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5190 The MDM2 protein is a negative regulator of p53: enhances its transcriptional activity. Since p53 functions indeed both proteins participate in an autoregulatory as a transcription factor, it must be in the nucleus to be feedback loop (Picksley and Lane, 1993; Wu et al., fully active. Indeed, sequestration of p53 in the 1993). After binding to p53, MDM2 inhibits its cytoplasm of some tumors suggests that this localization transcriptional activity (Momand et al., 1992), favors of p53 may represent a nonmutational mechanism of its nuclear export (Freedman and Levine, 1998; Roth inactivation (Moll et al., 1992). The physiological et al., 1998; Tao and Levine, 1999), and stimulates its relevance of the cooperation between p53 and HIPK2 degradation (Haupt et al., 1997; Kubbutat et al., 1997). in the presence of MDM2 is shown after genotoxic The importance of MDM2 in controlling p53 activity is stress. We have recently demonstrated that HIPK2 demonstrated by the mdm2-knockout mice. Their activates p53 apoptotic function, through phosphoryla- embryos die very early during gestation, but additional tion of Ser46, after treatment with cisplatin. The deletion of p53 rescues them from death (Jones et al., relevance of HIPK2 on p53 activity was demonstrated 1995; Montes de Oca Luca et al., 1995). Cell growth or with HIPK2 gene silencing by RNA interference, which cell death depends, to a large extent, on fine-tuning of efficiently reduced p53-mediated apoptosis (Di Stefano the p53–MDM2 interaction. Overexpression of MDM2 et al., 2004). We show here that one possible mechanism, blocks p53-mediated cell-cycle arrest and apoptosis leading to the activation of the HIPK2/p53 pathway, is (Chen et al., 1996; Haupt et al., 1996), and this should the reduction of p53 ubiquitination mediated by have negative consequences for cells, because it MDM2. Thus, only in HIPK2-interfered cells p53 is diminishes their ability to activate the p53 pathway ubiquitinated by MDM2. How HIPK2 prevents the under stress conditions. Indeed, in different types of ubiquitination of p53 by MDM2 is still a matter of tumors, MDM2 overexpression favors uncontrolled cell speculation and further investigations. A p53-mediated proliferation (Leite et al., 2001; Polsky et al., 2001; downregulation of MDM2 expression was observed in Eymin et al., 2002). Saos2 cells overexpressing HIPK2 and p53 (Wang et al., As MDM2 controls p53 activity, inhibition of such 2001, and present paper). This MDM2 downregulation association allows an immediate p53-mediated response could depend on multiple mechanisms including p53 (Prives, 1998). One of the mechanisms that control the post-translational modifications and additional protein/ MDM2/p53 interaction is DNA damage-induced phos- protein interactions. Therefore, a direct interaction phorylation of different p53 residues (e.g. Ser15, Thr18, between HIPK2 and MDM2 cannot be excluded. Ser20), which hinders p53 binding to MDM2. After Further experiments are under evaluation to better stress, modification of the p53 protein prevents or elucidate this hypothetical regulatory mechanism. disrupts the MDM2/p53 interaction, which results in In conclusion, in this study, we have shown that the p53 activation (Craig et al., 1999; Sakaguchi et al., activation of the HIPK2/p53 pathway plays an im- 2000). Hence, protein kinases involved in phosphorylat- portant role in the inhibition of the MDM2/p53 ing p53 sites that regulate its interaction with MDM2 interaction rescuing p53 apoptotic function. Among are particularly useful in order to better understand the the many genes that appear to play a role in the mechanisms of p53 activation. Here we report that regulation of apoptosis, p53 has emerged as one of the HIPK2, a kinase that activates p53 through its specific leading ‘death star’ (Vousden, 2000). In this regard, phosphorylation at Ser46 (D’Orazi et al., 2002; molecules such as HIPK2 that appear to regulate Hofmann et al., 2002), can efficiently rescue p53 specifically the p53 apoptotic pathway, and that we transcriptional activity and apoptosis from MDM2 show to regulate also the MDM2/p53 complex, might be inhibition. One possible explanation for this action useful as attractive new anticancer strategies aimed at might be that HIPK2 phosphorylation at Ser46 renders activating the p53 apoptotic pathway in tumor that p53 more resistant to inhibition by MDM2, maintaining overexpress MDM2 but retain wild-type p53 (Chene, its capacity to transactivate responsive genes and induce 2003). apoptosis, in agreement with the model of Prives and co- workers (Shieh et al., 1997; Prives, 1998), where phosphorylation at the N-terminal sites of p53 (where Materials and methods Ser46 resides)may prevent the inhibitory effect of MDM2 and render p53 active. Cells, transfection, and treatments One other important mechanism by which MDM2 H1299 lung adenocarcinoma (p53 null), Saos2 osteosarcoma negatively regulates p53 is by promoting its degradation (p53 null), and RKO colon cancer cells (wtp53) were grown in through the ubiquitin–proteasome pathway (Haupt RPMI-1640 medium (GIBCO-BRL, Life Technology, Grand et al., 1997; Kubbutat et al., 1997). The nuclear export Island, NY, USA)supplemented with 10% heat-inactivated of p53 is important for this process. In this regard, we fetal bovine serum (GIBCO-BRL). show that HIPK2 protects p53 from nuclear export and Transient transfection assays were performed using the BES ubiquitination by MDM2. We also show that, in the or the lipofectaminePlus (Invitrogen)methods, as described earlier (D’Orazi et al., 2002). The amount of plasmid DNA presence of HIPK2, MDM2 is no longer able to interact was equalized in each sample by supplementing with empty with p53. These findings provide a mechanistic explana- plasmid. Expression vectors used in this study were: tion for the cooperation between p53 and HIPK2 in the pCAG3.1wtp53, p53S46A (kindly provided by E Appella, presence of MDM2. Thus, the maintenance of nuclear NIH, Bethesda, USA), pCMV-MDM2, HIPK2-Flag, K221R- p53 by HIPK2, overcoming MDM2 negative effects, Flag (HIPK2 kinase dead mutant).

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5191 Cells treatments: Cisplatin was used at 5 mg/ml; MG132 inhibitors. Equal amounts of protein were immunoprecipitated proteasome inhibitor (Biomol Research Laboratories, PA, with anti-p53 antibodies (DO1 and Ab1801, Santa Cruz USA)was maintained in DMSO and used at a ﬁnal Biotechnology)preadsorbed to protein G–agarose (Pierce). concentration of 10 mM. Immunocomplexes were collected by centrifugation, separated by 9% SDS–PAGE, and blotted onto nitrocellulose membrane Transactivation assay (Bio-Rad, Hercules, CA, USA). For Western blotting, membranes were incubated with anti-p53 polyclonal antibody Cells were transiently transfected using BES or lipofectamine- (FL393, Santa Cruz Biotechnology), anti-Flag monoclonal Plus (Invitrogen)methods, with the luciferase reporter gene antibody (Upstate Biotechnology), anti-MDM2 monoclonal driven by the p53-dependent promoters bax (Miyashita and antibody (Ab1, Santa Cruz), anti-HIPK2 rabbit antiserum Reed, 1995), PIG-3 (Polyak et al., 1997), or Noxa (Oda et al., (kindly provided by ML Schmitz, University of Bern, Switzer- Science 2000; kindly provided by T Taniguchi, University of land), anti-phospho-p53-Ser46 rabbit polyclonal antibody Tokyo, Japan). Transfection efﬁciencies were normalized with (Cell Signaling), and anti b-gal monoclonal antibody (Sigma, the use of a cotransfected b-galactosidase construct, and Bio-Sciences, St Louis, MO, USA). Immunoreactivity was luciferase activity was assayed as described previously detected by ECL chemiluminescence reaction kit (Amersham (D’Orazi et al., 2002). Corp., Arlington Heights, IL, USA).

Western blotting and immunoprecipitation analyses For ubiquitination assay, cells were transfected with the Acknowledgements indicated vectors and 24 h later treated with the proteasome This work was supported by grants from Associazione Italiana inhibitor MG132 (10 mM)for additional 12 h. Total cell lysates Ricerca sul Cancro (AIRC), Ministero Sanita` , FIRB, MIUR- were prepared by incubating cell pellets for 30 min on ice in CNR, and MIUR Fondi Ateneo. We gratefully acknowledge lysis buffer (2 mM Tris-HCl pH 7.5; 5 mM EDTA; 150 mM constructive criticisms and helpful discussion by Dr F Moretti, NaCl; 1% sodium dehoxycolate; 0.025% SDS)plus protease Dr S Bacchetti, and Dr A Porrello.

References

Appella E and Anderson CW. (2001). Eur. J. Biochem., 268, Leite KR, Franco MF, Srougi M, Nesrallah LJ, Nesrallah A, 2764–2772. Bevilacqua RG, Darini E, Carnalho M, Meirelles MI, Barak Y, Juven T, Haffner R and Oren M. (1993). EMBO J., Santana I and Camara-Lopes LH. (2001). Mod. Pathol., 12, 461–468. 12, 428–436. Bulavin DV, Saito S, Hollander MC, Sakaguchi K, Anderson Lu H and Levine AJ. (1995). Proc. Natl. Acad. Sci. USA, 92, CW, Appella E and Fornace AJ. (1999). EMBO J., 18, 5154–5158. 6845–6854. Miyashita T and Reed JC. (1995). Cell, 80, 293–299. Chen J, Wu X, Lin J and Levine AJ. (1996). Mol. Cell. Biol., Moll UM, Riou G and Levine AJ. (1992). Proc. Natl. Acad. 16, 2445–2452. Sci. USA, 89, 7262–7266. Chene P. (2003). Nat. Rev. Cancer, 3, 102–109. Momand J, Jung D, Wilczynski S and Niland J. (1998). Craig AL, Burch L, Vojtesek B, Mikutowska J, Thompson A Nucleic Acid Res., 26, 3453–3459. and Hupp TR. (1999). Biochem. J., 342, 133–141. Momand J, Zambett GP, Olson DC, Gorge D and Levine AJ. D’Orazi G, Cecchinelli B, Bruno T, Manni I, Higashimoto Y, (1992). Cell, 69, 1237–1245. Saito S, Gostissa M, Coen S, Marchetti A, Del Sal G, Montes de Oca Luca R, Wagner DS and Lozano G. (1995). Piaggio G, Fanciulli M and Soddu S. (2002). Nat. Cell Biol., Nature, 378, 203–206. 4, 11–19. Oda E, Ohki R, Murasawa H, Nemoto J, Shibue T, Yanashita Di Stefano V, Rinaldo C, Sacchi A, Soddu S and D’Orazi G. T, Tokino T, Taniguchi T and Tanaka N. (2000a). Science, (2004). Exp. Cell Res., 293, 311–320. 288, 1053–1058. Eymin B, Gazzeri S, Brambilla C and Brambilla E. (2002). Oda K, Arakawa H, Tanaka T, Matsuda K, Tanikawa C, Oncogene, 21, 2750–2761. Mori T, Nishimori H, Tamai K, Tokino T, Nakamura Y Freedman DA and Levine AJ. (1998). Mol. Cell. Biol., 18, and Taya Y. (2000b). Cell, 102, 849–862. 7288–7293. Oliner JD, Kinzler KW, Meltzer PS, George DL and Fuchs SY, Fried VA and Ronai Z. (1998). Oncogene, 17, Vogelstein B. (1992). Nature, 358, 80–83. 1483–1490. Picksley SM and Lane DP. (1993). BioEssay, 15, 689–690. Giaccia J and Kastan LB. (1998). Genes Dev., 12, 2973–2983. Polsky D, Bastian BC, Hazan C, Melzer K, Pack J, Houghton Gottlieb TM and Oren M. (1996). Biochim. Biophys. Acta, A, Busan K, Cordon-Cardo C and Osman I. (2001). Cancer 1287, 77–102. Res., 61, 7642–7646. Haupt Y, Barak Y and Oren M. (1996). EMBO J., 15, 1596–1606. Polyak K, Xia Y, Zweir JL, Kinzler KW and Vogelstein B. Haupt Y, Maya R, Kazaz A and Oren M. (1997). Nature, 387, (1997). Nature, 389, 300–305. 296–299. Prives C. (1998). Cell, 95, 5–8. Hofmann TG, Moller A, Sirma H, Zentgraf H, Taya Y, Droge Prives C and Hall PA. (1999). J. Pathol., 187, 112–126. W, Will H and Schmitz ML. (2002). Nat. Cell Biol., 4, 1–10. Roth J, Dobbelstein M, Freedman DA, Shenk T and Levine Jones SN, Roe AE, Donehower LA and Bradley A. (1995). AJ. (1998). EMBO J., 17, 554–564. Nature, 378, 206–208. Sakaguchi K, Saito S, Higashimoto Y, Roy S, Anderson CW Kim E-J, Park J-S and Um S-J. (2002). J. Biol. Chem., 277, and Appella E. (2000). J. Biol. Chem., 275, 9278–9283. 32020–32028. Shieh S-Y, Ikeda M, Taya Y and Prives C. (1997). Cell, 9, Kubbutat MH, Jones SN and Vousden KH. (1997). Nature, 325–334. 387, 299–303. Siliciano JD, Canman CE, Taya Y, Sakaguchi K, Appella E Lakin ND and Jackson S. (1999). Oncogene, 1, 7644–7655. and Kastan MB. (1997). Genes Dev., 11, 3471–3481.

Oncogene HIPK2 protects p53 from MDM2 ubiquitination VD Stefano et al 5192 Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA, 96, Valculescu VE and El Deiry WS. (1996). Clin. Chem., 42, 858–868. 3077–3080. Vogt Sionov R and Haupt Y. (1999). Oncogene, 18, 6145–6157. Unger T, Juven-Gershon T, Moallem E, Berger M, Vogt Vousden KH. (2000). Cell, 103, 691–694. Sionov R, Lozano G, Oren M and Haupt Y. (1999). EMBO Wang Y, Bedatin K-M and Hug H. (2001). BMC Mol. Biol., 2, J., 18, 1805–1814. 8–16. Unger T, Nau MM, Segal S and Minna JD. (1992). EMBO J., Wu X, Bayle JH, Olson D and Levine AJ. (1993). Genes Dev., 11, 1383–1390. 7, 1126–1132.

Oncogene