p18^Hamlet protein levels are regulated by p38 MAPK

The mRNA levels of p18^Hamlet varied significantly among different human tissues and cell lines in contrast with the p18^Hamlet protein, which was usually difficult to detect (Supplementary Figure 3 and data not shown). We therefore investigated the possibility that p18^Hamlet protein stability could be regulated. As shown in Figure 3A, incubation with the proteasome inhibitor MG132 resulted in the accumulation of endogenous p18^Hamlet protein in both mouse embryonic fibroblasts (MEFs) and human osteosarcoma U2OS cells. To demonstrate that p18^Hamlet was a target of the ubiquitin–proteasome system, we transfected U2OS cells with Myc-p18^Hamlet alone or in combination with HA-ubiquitin and then analyzed the Myc-p18^Hamlet immunoprecipitates by immunoblotting with anti-p18^Hamlet antibody. In this experiment, we detected a smear of slowly migrating p18^Hamlet forms that were not observed in extracts from cells transfected with either HA-ubiquitin or p18^Hamlet alone, suggesting that they probably correspond to p18^Hamlet–ubiquitin conjugates. This was further supported by the recognition of the slowly migrating forms of p18^Hamlet with an anti-HA antibody (Figure 3B). MG132 concentrations that inhibit the proteasome have been also reported to activate p38 MAPK (Wu et al, 2004). We confirmed that MG132-induced p18^Hamlet accumulation correlated with the phosphorylation of p38 MAPK in MEFs, but this effect was abolished when the MG132 treatment was performed in the presence of the p38 MAPK inhibitor SB203580 (Supplementary Figure 4). These results strongly suggest that p18^Hamlet accumulation requires the activation of p38 MAPK.

Open in a separate window

Figure 3

Accumulation of p18^Hamlet protein in response to DNA damage-inducing agents. (A) MEFs and U2OS cells were treated with the proteasome inhibitor MG132 (25 μM) for 2 h and then lysed. Expression of endogenous p18^Hamlet was analyzed by Western blotting. (B) U2OS cells were transfected with HA-ubiquitin and Myc-p18^Hamlet, as indicated, and 16 h after transfection, were treated with MG132 (25 μM) for 5 h. Myc immunoprecipitates were analyzed by Western blotting using p18^Hamlet and HA antibodies. (C) SK-Mel-103 and U2OS cells were treated with UV and cell lysates were analyzed by Western blotting using the indicated antibodies. (D) SK-Mel-103 cell lysates were prepared 3 h after UV treatment, either in the presence or absence of SB203580 (SB, 10 μM), and analyzed by Western blotting. (E) SK-Mel-103 cells were treated with cisplatin for the indicated times and p18^Hamlet accumulation was analyzed by Western blotting. (F) U2OS cells were cotransfected with MKK6DD (600 ng), p18^Hamlet (1 μg) and increasing amounts of p38α, as indicated. Twenty-four hours after transfection, lysates were prepared from both attached and floating cells (that express higher levels of p18^Hamlet) and analyzed by Western blotting.

Next, we investigated the effect of genotoxic stresses that activate the p38 MAPK pathway, such as UV (Kyriakis and Avruch, 1996), on the endogenous p18^Hamlet protein levels. In agreement with the above results, UV-induced p38 MAPK phosphorylation correlated with a small (about two-fold) but reproducible increase in endogenous p18^Hamlet protein levels in different cell lines (Figure 3C). The amount of p18^Hamlet protein typically peaked between 1 and 6 h after the treatment and later decreased to levels lower than those in untreated cells. Importantly, the UV-induced accumulation of p18^Hamlet was prevented by pretreatment with SB203580 (Figure 3D).

We then investigated whether p18^Hamlet accumulation was specific for UV-induced p38 MAPK activation or if it was a more general response to stress. Treatment of cells with the chemical DNA damage-inducing agent cisplatin induced p38 MAPK activation, which was accompanied by a significant increase in endogenous p18^Hamlet levels (Figure 3E).

To further strengthen the connection between p38 MAPK activation and accumulation of p18^Hamlet, we transfected U2OS cells with a low amount of p18^Hamlet (to limit apoptosis induction, see below) together with increasing amounts of p38α and a constant amount of its activator MKK6DD (Figure 3F). In this experiment, we observed a direct correlation between p38α activation levels and the amount of p18^Hamlet protein expressed. We also detected maximal p18^Hamlet phosphorylation on Thr103 with the lowest concentration of p38α used, suggesting that this residue might be efficiently phosphorylated in situations of poor p38 MAPK activation. However, the T103A mutant protein was still able to accumulate upon p38α activation (Supplementary Figure 5), suggesting that Thr103 phosphorylation may contribute to, but is not essential for, p18^Hamlet accumulation.

Finally, we confirmed that p18^Hamlet was an unstable protein with a half-life of less than 3 h in cycloheximide-treated U2OS cells (Figure 4A). Interestingly, specific activation of p38α was sufficient to significantly increase the half-life of p18^Hamlet (Figure 4B and C). In contrast, we could detect no changes in p18^Hamlet mRNA levels when cells were treated with cisplatin or UV (Figure 4D). These results indicated that stress-induced accumulation of p18^Hamlet was mainly regulated at the level of protein stability. To determine the importance of p38 MAPK-mediated phosphorylation in p18^Hamlet protein stability, we performed cycloheximide-chase experiments in U2OS cells transfected with p18^Hamlet wt, T103A or the quadruple mutant 4 × T/A and then UV irradiated the cells (Figure 4E). Whereas the T103A and wt proteins behaved similarly and were both significantly accumulated in response to UV, 4 × T/A mutant expression levels were not affected by this treatment. Moreover, the 4 × T/A mutant was expressed at levels lower than wt p18^Hamlet, supporting the idea that phosphorylation of these sites could be important for the regulation of p18^Hamlet protein stability.

Open in a separate window

Figure 4

Stabilization of p18^Hamlet protein in response to p38α activation. (A) U2OS cells were cotransfected with p18^Hamlet (1 μg) and GFP (500 ng) and incubated with cycloheximide (CHX, 30 μg/ml) for the indicated times. Total cell lysates were analyzed by Western blotting. (B, C) U2OS cells were transfected with p18^Hamlet, either alone or together with p38α and MKK6DD, and 24 h later were incubated with CHX for up to 4 h. Expression of p18^Hamlet protein was determined by Western blotting. The blots corresponding to cells untreated or treated with CHX for 4 h, in the presence or absence of active p38α, are shown in (C). (D) Total RNAs were obtained from UV- or cisplatin-treated SK-Mel-103 cells and were analyzed by Northern blotting with a p18^Hamlet probe. GAPDH was used to confirm equal RNA loading. (E) U2OS cells were transfected with 6 μg of p18^Hamlet wt, T103A or 4 × T/A, as indicated. Twenty-four hours after transfection, cells were treated with CHX alone or in combination with MG132 for 4 h. Where indicated, cells were also UV irradiated 1 h before collection. A GFP expression vector (200 ng) was cotransfected to ensure equal efficiency of transfection.

Accumulation of p18^Hamlet induces apoptosis

Once we established that p18^Hamlet levels increased in response to DNA damage, we investigated the biological significance of this accumulation. U2OS cells were transfected with either GFP-p18^Hamlet or GFP alone and then sorted to analyze the cell-cycle profile in the fluorescent cell population. We found that around 6% of the GFP-expressing cells were apoptotic (sub-G0/G1 population), whereas this amount increased to 18% in the GFP-p18^Hamlet-positive cells, indicating that p18^Hamlet overexpression was sufficient to induce apoptosis. Interestingly, GFP-p18^Hamlet overexpression did not affect apoptosis levels in SAOS cells, a p53-deficient human osteosarcoma cell line (Figure 5A).

Open in a separate window

Figure 5

Induction of apoptosis by p18^Hamlet overexpression. (A) U2OS and SAOS cells were transfected with GFP or GFP-p18^Hamlet and analyzed by flow cytometry. The percentage of cells with a sub-G0/G1 DNA content in a representative experiment is shown. (B) U2OS and SAOS cells were transfected as indicated in (A) and the expression of the indicated proteins was analyzed by Western blotting. (C) U2OS cells expressing inducible p18^Hamlet were treated with tetracycline for 24 h before RNA and protein extraction. Samples were analyzed by quantitative RT–PCRs (left) and Western blotting (right). (D) U2OS cells expressing tetracycline-inducible p18^Hamlet were incubated with tet for 24 h and then treated with cisplatin for another 24 h before analyzing apoptosis by annexin V staining. Numbers on top of the bars indicate total percentage of early and late apoptotic events as well as dead cells. The lower panel shows the protein levels of overexpressed Myc-p18^Hamlet, endogenous (End.) p18^Hamlet (marked with an asterisk) and phospho-Ser15-p53. (E) Primary MEFs were infected with pBABE puro or p18^Hamlet-expressing pBABE puro retroviruses. After puromycin selection, cells were treated for 24 h with cisplatin or UV as indicated, and the percentage of sub-G0/G1 cell population, as a measure of apoptosis levels, was analyzed by FACS.

Given the key role of p53 in the apoptotic response induced by DNA damage, we investigated if p18^Hamlet could regulate p53. In U2OS cells, GFP-p18^Hamlet overexpression did not affect the total levels of p53 (Figure 5B), but resulted in higher levels of the proapoptotic p53 target gene NOXA, whereas it had no effect on other p53-dependent proapoptotic genes such as PUMA or Bax. In contrast, NOXA levels were not affected by GFP-p18^Hamlet overexpression in SAOS cells. To study the induction of apoptosis by p18^Hamlet in more detail, we generated a tetracycline-inducible system in which p18^Hamlet protein expression peaked at about 16 h after addition of tetracycline to U2OS cells (Supplementary Figure 6). Consistent with the above results, tetracycline-induced p18^Hamlet expression was accompanied by an increase in both NOXA mRNA and protein levels (Figure 5C), whereas we could detect no significant changes in the expression of other p53-dependent targets, such as PUMA, Bax, p21 or Hdm2 (Figure 5C and data not shown). We also confirmed that p18^Hamlet induction was sufficient on its own to significantly increase the early and late apoptotic populations. In addition, p18^Hamlet cooperates with cisplatin treatment in apoptosis induction (Figure 5D). To confirm the proapoptotic function of p18^Hamlet in a more physiological system, we overexpressed p18^Hamlet in primary MEFs. As shown in Figure 5E, p18^Hamlet overexpression was sufficient to increase apoptosis levels in non-stressed cells, and it also strongly promoted apoptosis induced by cisplatin or UV. Taken together, these results support a role for p18^Hamlet in p53-mediated apoptosis induction.

To confirm the role of endogenous p18^Hamlet as an apoptosis mediator, we designed siRNA oligonucleotides that efficiently downregulated p18^Hamlet (Figure 6A and Supplementary Figures 7 and 8). Previous work has shown that p53 contributes to apoptosis induced by UV or cisplatin in both U2OS and MCF7 cells (Bergamaschi et al, 2003). Using a highly sensitive, quantitative method that detects apoptotic nucleosomes (see Materials and methods), we found that p18^Hamlet downregulation did not affect basal apoptosis levels, but significantly impaired apoptosis induced by either UV or cisplatin in U2OS cells (Figure 6A). The same effect was observed in MCF7 cells treated with cisplatin (Supplementary Figure 7). We then analyzed the effect of p18^Hamlet downregulation on the expression of p53-dependent target genes upon cisplatin and UV treatment. Interestingly, only certain p53 target genes were induced in response to these two types of stress. In particular, NOXA and Hdm2 responded to cisplatin, whereas p21, Bax and PUMA remained unaffected. In the case of UV, only NOXA, and also slightly PUMA, protein levels were upregulated, whereas p21, Hdm2 and Bax levels did not increase (Figure 6B). This result suggests that specific programs of gene expression account for the p53-dependent apoptosis in response to each particular type of stress. Downregulation of p18^Hamlet prevented NOXA and, in the case of UV irradiation, also PUMA protein accumulation, but had no effect on Hdm2 protein induction. Taken together, the results support an important role for p18^Hamlet in stress-induced apoptosis.

Open in a separate window

Figure 6

p18^Hamlet is required for apoptosis induction in response to DNA damage. (A) U2OS cells were transfected with p18^Hamlet or control siRNAs and the levels of endogenous p18^Hamlet were analyzed by Western blotting (upper). Forty-eight hours after transfection, cells were treated with UV or cisplatin for 24 h and apoptosis was quantified by measuring DNA fragmentation in a colorimetric assay. Means±standard deviations of three independent experiments are represented. Statistical significance was evaluated with the Student's t-test (P-values are shown). (B) U2OS cells were treated with siRNAs and UV or cisplatin as described in (A). Expression of the indicated proteins was detected by Western blotting.

p18^Hamlet activates p53-dependent gene promoters

The ability of p18^Hamlet to upregulate p53-dependent proapoptotic genes prompted us to investigate whether both proteins could interact. We found that tetracycline-induced p18^Hamlet co-immunoprecipitated with endogenous p53 from U2OS cells (Figure 7A, upper panel), and we also managed to co-immunoprecipitate the two endogenous proteins (Figure 7A, lower panel). The interaction was also observed in pull-down assays using recombinant GST-p53 and ³⁵S-labelled p18^Hamlet and it involves the zinc-finger domain of p18^Hamlet and the C-terminal region of p53 (see below, Figure 9D and Supplementary Figure 9). In contrast, overexpression of p18^Hamlet did not affect phosphorylation of p53 on Ser15 and Ser46, two common p53 phosphorylation events in response to stress (Figure 7B).

Open in a separate window

Figure 7

p18^Hamlet stimulates some p53-regulated genes. (A) Upper: U2OS cells with inducible p18^Hamlet were treated with tetracycline (tet) for 24 h and total cell lysates were immunoprecipitated with p53 or control antibodies and then blotted with p18^Hamlet antibodies. Lower: total lysates of MG132-treated U2OS cells were immunoprecipitated and blotted as above to detect interaction between to endogenous p18^Hamlet and p53. (B) U2OS cells expressing tet-inducible p18^Hamlet were treated with tet or cisplatin and then analyzed by Western blotting. (C) U2OS cells were transfected with empty vector (control) and either p18^Hamlet wt or p18^{Hamlet(1–117)}, together with reporter constructs containing different p53-responsive promoters upstream of the luciferase gene, as indicated. Luciferase activity was analyzed 16 h later and transfection efficiency was normalized to Renilla activity. Means±s.d. of three independent experiments are represented. (D) Transfections and luciferase assays were performed exactly as in (C) to test the wt and mutant PUMA 4 × BS2 reporters. (E) U2OS cells were transfected with the indicated reporter constructs. Twenty-four hours later, cells were treated with cisplatin and after 16 h, luciferase activity was measured and normalized to Renilla. (F) U2OS cells were transfected with p18^Hamlet or control siRNAs and 24 h later cotransfected with the indicated luciferase reporters. Twenty-four hours after transfection, cells were incubated with cisplatin for 16 h and luciferase activity was measured. (G) U2OS cells were transfected with p18^Hamlet or empty vector (control) in combination with Hdm2 and NOXA promoter reporters. Twenty four hours after transfection, cells were mock treated or treated with cisplatin and luciferase activity was measured 10 h later. Means±s.d. of three independent experiments are represented. (H) U2OS cells were transfected with GFP alone (control) or the indicated GFP-tagged p18^Hamlet proteins and 48 h later the sub G0/G1 percentage in the fluorescent population was determined by FACS (white bars). U2OS cells were transfected with empty vector (control) or Myc-tagged wt and mutant p18^Hamlet proteins, together with the PUMA 4 × BS2 reporter, and luciferase activity was measured 16 h later (black bars). Expression levels of Myc- and GFP-tagged p18^Hamlet proteins are shown in Supplementary Figure 12.

Open in a separate window

Figure 9

The p53 target gene cyclin G1 controls p18^Hamlet protein levels under normal growing conditions. (A) MEFs (wt and p53^−/−) were immunostained with p18^Hamlet antibodies. Nuclear localization was confirmed by DAPI staining. (B) Expression of the indicated proteins was analyzed by Western blotting in wt and p53^−/− MEFs. (C) U2OS cells with inducible p18^Hamlet were treated with tetracycline (tet) and 24 h later were transfected with p53 and analyzed by Western blotting. (D) GST pull-down assay was performed by incubation of ³⁵S-labelled p18^Hamlet wt and 1–117 with GST and GST-fused p38α, cyclin G1 or p53, as indicated. (E) HEK-293 cells were cotransfected with Myc-p18^Hamlet (5 μg) and increasing amounts of HA-cyclin G1 (0–10 μg). Twenty-four hours after transfection, the expression levels of the indicated proteins were analyzed by Western blotting using HA and Myc antibodies. Transfection efficiency was evaluated by cotransfection with GFP (500 ng). (F) HEK-293 cells were cotransfected with p18^Hamlet (1 μg) and YFP-cyclin G1 (9 μg) and 16 h after transfection, cells were treated for 5 h with MG132. The expression levels of the indicated proteins were analyzed by Western blotting. (G) U2OS cells expressing tet-inducible p18^Hamlet were transfected with HA-cyclin G1. Cellular localization was analyzed by immunostaining with HA and p18^Hamlet antibodies. Colocalization areas are indicated in yellow (merge).

We then analyzed whether p18^Hamlet could regulate the transactivation function of p53. For these experiments, we expressed the p53-regulated promoters of Hdm2, Bax, NOXA and PUMA in U2OS cells, which contain wt p53 protein. As shown in Figure 7B, p18^Hamlet stimulated the transcription of the three p53-regulated proapoptotic genes (only a minor effect was observed in the case of Bax), but had no effect on the Hdm2 promoter. We confirmed that both Bax and Hdm2 promoters were indeed able to respond to p53 (Supplementary Figure 10). Importantly, stimulation of p53 transcriptional activity by p18^Hamlet was dependent upon the integrity of its C-terminal zinc-finger HIT domain, as a p18^Hamlet derivative lacking the last 37 amino acids (p18^{Hamlet(1–117)}) had no significant effect on any of the p53-regulated promoters (Figure 7C). The C-terminally truncated p18^{Hamlet(1–117)} protein was still able to bind to p38α and localized to the nucleus, as the full-length p18^Hamlet, but failed to interact with p53 (Figure 9D and Supplementary Figure 9). It therefore appears that p38α and p53 interact with different parts of the p18^Hamlet protein and that the C-terminal domain of p18^Hamlet is required for p53 transactivation.

To confirm that the effect of p18^Hamlet was indeed p53 dependent, we used a reporter plasmid with the PUMA minimal promoter (PUMA 4 × BS2) consisting of four tandem repeats of the p53 binding site or a mutant version of this reporter, which does not bind to p53 (Yu et al, 2001). Full-length p18^Hamlet increased PUMA 4 × BS2 transcription by about four-fold, whereas the mutant p18^{Hamlet(1–117)} had no effect. In contrast, neither full-length nor truncated p18^Hamlet proteins were able to stimulate transcription of the mutated PUMA 4 × BS2 promoter (Figure 7D). Based on these results, we conclude that p18^Hamlet can activate, through its C-terminal domain, the transcription of several p53-dependent genes.

Consistent with the results shown in Figure 6B, cisplatin was able to differentially transactivate several p53-dependent promoters, being more effective in the cases of NOXA and PIG-3 (Figure 7E) when compared with other genes such as Bax or PUMA. The downregulation of p18^Hamlet inhibited by about 50% the ability of cisplatin to transactivate the NOXA and PUMA promoters, whereas it had a more moderate effect on p21 and Bax promoters (Figure 7F). In addition, when p18^Hamlet overexpression was combined with cisplatin treatment, we observed an additive effect on the NOXA promoter, without affecting the Hdm2 promoter, leading to an imbalance that was clearly favorable to the transcription of the proapoptotic gene NOXA (Figure 7G). Of note, overexpression of p18^Hamlet in p53-deficient SAOS cells had no effect on the PUMA 4 × BS2-reporter construct, unless the cells were cotransfected with p53 (Supplementary Figure 11).

To determine the importance of p38 MAPK-mediated phosphorylation in p18^Hamlet function, we analyzed the ability of different p18^Hamlet mutants to transactivate the PUMA 4 × BS2 promoter and to induce apoptosis. The three p18^Hamlet mutants were able to bind to p38α to a similar extent, but phosphorylation was significantly reduced in p18^Hamlet-T103A and specially in the quadruple mutant p18^{Hamlet-4 × T/A} (Supplementary Figure 2). However, only the p18^Hamlet-T103A mutant recapitulated the activity of the wt protein regarding PUMA promoter transactivation and apoptosis induction, consistent with the idea that phosphorylation of this residue was not essential for p18^Hamlet function and accumulation (Figure 7H). In contrast, truncated p18^{Hamlet(1–117)} was not able to induce cell death, supporting the importance of p53 interaction for the induction of apoptosis by p18^Hamlet. Finally, p18^{Hamlet-4 × T/A} showed a significant decrease in both PUMA transactivation and apoptosis induction, suggesting that p18^Hamlet phosphorylation is essential for its activity (Figure 7H).

Our results indicated that the ability of p18^Hamlet to stimulate p53-induced transcription was not related to p53 phosphorylation or stabilization. We therefore performed chromatin IP (ChIP) experiments to investigate the recruitment of p53 to the promoter of its target genes. We found that p18^Hamlet overexpression was sufficient to enhance p53 binding to both the PUMA and NOXA promoters, whereas it had no effect on the ability of p53 to bind to the Hdm2 promoter (Figure 8A). In addition, p18^Hamlet itself was also bound to the NOXA promoter, even in unstimulated U2OS cells, and the binding was increased in response to cisplatin treatment (Figure 8B). Interestingly, p18^Hamlet knockdown had a profound effect on p53 loading onto the NOXA promoter (Figure 8C), supporting a key role for p18^Hamlet in the recruitment of p53 to certain target promoters.

Open in a separate window

Figure 8

Recruitment of p53 and p18^Hamlet to p53-regulated promoters. (A) U2OS cells expressing tetracycline (tet)-inducible p18^Hamlet were treated with tet for 24 h or UV irradiated for 8 h and then subjected to ChIP analysis. The DNA associated with the p53 immunoprecipitates was subjected to PCR with primers specific for the Hdm2, PUMA and NOXA promoters. (B) U2OS cells were treated with cisplatin for 6 h and then subjected to ChIP analysis using both p53 and p18^Hamlet antibodies and NOXA primers. (C) p53 recruitment to NOXA promoter was analyzed by ChIP assay in U2OS cells 72 h after incubation with control and p18^Hamlet siRNAs.

p18^Hamlet links the p38 MAPK and p53 pathways

Our results indicate that p38 MAPK plays an important role in the regulation of p18^Hamlet half-life. In particular, p38 MAPK activation is required for the accumulation of p18^Hamlet induced by DNA damage-inducing agents such as UV. We also showed that several sites that are phosphorylated by p38α in vitro are also important for p18^Hamlet protein stability in cells. However, the exact contribution of specific phosphorylation sites to p18^Hamlet protein stability remains to be elucidated. It is also possible that p38 MAPK might regulate the stabilization of the p18^Hamlet protein by other mechanisms, in addition to direct phosphorylation.

Previous studies have documented that activation of the p38 MAPK pathway may lead to p53-induced apoptosis (Bulavin and Fornace, 2004). The connection between p38 MAPK activation and increased transcription from p53-regulated promoters has been classically attributed to the ability of p38 MAPK to directly phosphorylate p53 on Ser33 and Ser46 (Bulavin et al, 1999; Sanchez-Prieto et al, 2000). Our findings provide a new mechanism by which p38 MAPK can contribute to p53-induced apoptosis, namely by contributing to the stabilization of p18^Hamlet, a p53 coactivator that stimulates p53-dependent transcription.

We have shown that the stimulatory effect of p18^Hamlet on p53-regulated genes is mediated by the p53 binding sites in the promoters. Nevertheless, we cannot rule out that p18^Hamlet could also have p53-independent functions. The biological significance of p18^Hamlet homologues in yeast, which lack p53-related proteins, is as yet unclear but might support this possibility.

p18^Hamlet as a determinant for p53 response specificity

There are several mechanisms by which p18^Hamlet could provide specificity to p53-regulated stress responses. First, p18^Hamlet does not seem to be a ubiquitously expressed protein, in contrast with p38 MAPK and p53, and might therefore provide tissue-specific responses. Second, p18^Hamlet accumulates only in response to certain types of stresses such as UV and cisplatin but not in response to γ irradiation (not shown). Finally, p18^Hamlet can specifically stimulate the transcription of certain p53-dependent promoters. In particular, in response to cisplatin, p18^Hamlet contributes to the p53-dependent transcriptional activation of NOXA but not Hdm2, suggesting that the effects of p18^Hamlet on p53-mediated transcription are promoter specific. This property of p18^Hamlet is shared by other p53 coactivators. For example, ASPP proteins can stimulate transcription of the proapoptotic genes Bax and PIG3, but not Mdm2 or p21^Cip1 (Samuels-Lev et al, 2001), whereas hDaxx specifically represses p53-mediated induction of genes involved in cell-cycle arrest such as p21^Cip1 (Gostissa et al, 2004). Other p53 coactivators, such as the p300/CBP cofactor JMY, can efficiently upregulate Bax but not p21^Cip1 (Shikama et al, 1999).

It has been clearly established that p21^Cip1 has a key role in p53-induced cell-cycle arrest, but the molecular pathways involved in p53-mediated apoptosis are not fully understood. Several proapoptotic molecules can be transcriptionally induced by p53, but the contribution of each factor to the apoptotic response depends on both the cell type and nature of the stress. We have found that p18^Hamlet can stimulate the recruitment of p53 to the p21^Cip1 promoter (not shown), but we could not observe p21^Cip1 protein induction in response to UV or cisplatin, most likely due to post-translational downregulation (Fotedar et al, 2004). Thus, the contribution of p18^Hamlet to the regulation of the p21^Cip1 promoter needs to be further investigated.

The function of p18^Hamlet as a transcriptional coactivator is further supported by recent work, which has identified this protein as a potential subunit of the SRCAP (SNF2-related CBP-activating protein) complex (Cai et al, 2005). Interestingly, SRCAP may contribute to the recruitment of histone acetyltransferase CBP to certain promoters (Eissenberg et al, 2005). The p300/CBP proteins are well-established regulators of the p53 response that control p53 acetylation and its DNA binding activity (Barlev et al, 2001; Espinosa and Emerson, 2001). It is therefore tempting to speculate that p53 modulation by p18^Hamlet could involve the regulation of p300/CBP.

Cell culture

HEK-293, HeLa, SAOS, MCF7, U2OS and melanoma SK-Mel-103 cells as well as wt and p53^−/− MEFs were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum. FuGene reagent (Roche Applied Science) was used for cell transfection according to the manufacturer's protocol. Cells were treated with UV (50–100 J/m²) and cisplatin (5–10 μg/ml), as indicated.

The generation of stable cell lines expressing inducible p18^Hamlet and the retroviral infections were performed as indicated in Supplementary data.

Transfection of siRNA and apoptosis measurement

The siRNA oligonucleotide for p18^Hamlet (UGCGGACACUGGAAAGAAAUU) was obtained from Dharmacon (Lafayette, CO). U2OS and MCF7 cells were grown to 50% of confluency and transfected with Dharmafect reagent 1 (Dharmacon) according to the manufacturer's protocol. Cells were treated with UV or cisplatin 48 h after siRNA transfection. Human Lamin A siRNA (siGLO™, Dharmacon) was used as a control. Apoptosis was analyzed using the cell death detection ELISA^PLUS kit (Roche Applied Science).

Flow cytometry analysis

Cells were trypsinized, washed with PBS, fixed with chilled 70% ethanol for 30 min at 4°C and then incubated in PBS containing 30 mg/ml of RNAse and stained for 30 min at 37°C with propidium iodide (25 μg/ml). Apoptotic cells were determined by their hypochromic subdiploid staining profiles. To estimate early apoptotic cells, Alexa 488-conjugated annexin V was used together with propidium iodide counterstain following the manufacturer's recommendations (Molecular Probes Inc.).

Luciferase expression analysis

U2OS and SAOS cells (2 × 10⁵) were plated 24 h before transfection in six multiwell dishes. Transactivation assays contained 30 ng of the Renilla expression construct phRL-TK (Promega), as a transfection control, 10 ng of p53, 300 ng of promoter reporter and 700 ng of full-length or mutated p18^Hamlet, as indicated. Cells were lysed in reporter lysis buffer 24 h after transfection. In the case of cisplatin treatments, cells were treated with the drug 24 h after transfection and collected 10–16 h later. Luciferase and Renilla activities were measured using the Dual-Luciferase^® Reporter kit (Promega).

Antibodies, Western blotting, IP, pull-down and kinase assays

Western blot analysis was performed using 40–60 μg of cell lysates prepared in ice-cold IP lysis buffer. Buffers and antibodies are described in Supplementary data.

For IPs, 20 μl of anti-Myc or anti-HA agarose conjugates were incubated with 250–500 μg of protein lysates for 14 h at 4°C. The beads were then washed three times in IP buffer and analyzed by immunoblotting or further washed in kinase buffer and used for kinase assays (Alonso et al, 2000).

GST pull-downs and in vitro kinase assays were performed as described in Supplementary data.

This work was initiated while AC, ID and ARN were working at the EMBL (Heidelberg, Germany). We thank Manuel Serrano, Marisol Soengas, Ivan Dikic and Pedro Lazo for providing advice and reagents and Esther Seco and Sara Sánchez for technical support. AC acknowledges a post-doctoral fellowship from Fundación Ramón Areces (Spain). ARN is supported by grants from MEC and the Fundación Científica de la AECC (Spain) and PC by the UK MRC and the Royal Society.