Chaperone-Dependent Stabilization and Degradation of P53 Mutants

Home , Hsp90

Oncogene (2008) 27, 3371–3383 & 2008 Nature Publishing Group All rights reserved 0950-9232/08 $30.00 www.nature.com/onc ORIGINAL ARTICLE Chaperone-dependent stabilization and degradation of p53 mutants

P Muller1,2, R Hrstka1, D Coomber2, DP Lane2 and B Vojtesek1

1Masaryk Memorial Cancer Institute, Brno, Czech Republic and 2Institute of Molecular and Cell Biology, Proteos, Singapore

p53 missense mutant proteins commonly show increased inactivated either by mutations leading to absence of stability compared to wild-type p53, which is thought to protein or by production of truncated protein product, depend largely on the inability of mutant p53 to induce the the majority of mutations found in TP53 gene ubiquitin ligase MDM2. However, recent work using (Blagosklonny et al., 1995) result in single amino-acid mouse models has shown that the accumulation of mutant substitutions. These missense mutations take place in p53 occurs only in tumour cells, indicating that stabiliza- the DNA-binding core domain of p53, producing a full- tion requires additional factors. To clarify the stabiliza- length protein that is incapable of binding DNA and is tion of p53 mutants in tumours, we analysed factors that therefore nonfunctional as a transcriptional activator/ affect their folding and degradation. Although all missense repressor. One of the most interesting aspects of mutant mutants that we studied are more stable than wild-type p53 proteins is their structural, biochemical and p53, the levels correlate with individual structural biological heterogeneity. The mutations in the core characteristics, which may be reflected in different gain- region, found in human tumours, have been classified of-function properties. In the absence of Hsp90 activity, based on thermodynamic stability and DNA-binding the less stable unfolded p53 mutants preferentially properties. The contact mutants with changes in residues associate in a complex with Hsp70 and CHIP (carboxy that are directly involved in binding DNA (such as 248R terminus of Hsp70-interacting protein), and we show that or 273R) have little effect on protein folding or stability. CHIP is responsible for ubiquitination and degradation of However, the majority of mutations cause local distor- these mutants. The demonstration of a complex interplay tion, mainly in proximity to the DNA-binding site (for between Hsp90, Hsp70 and CHIP that regulate the example, R249S) or can cause global unfolding (for stability of different p53 mutant proteins improves our example, mutation in the b-sandwich, for example, understanding of the pro-tumorigenic effects of increased V157F) (Bullock and Fersht, 2001). Hsp90 activity during multi-stage carcinogenesis. Under- The level of p53 protein is mainly regulated at the standing the roles of Hsp90, Hsp70 and CHIP in cancers post-translational level by the MDM2 protein, while may also provide an important avenue through which to expression of MDM2 is activated by p53 at the target p53 to enhance treatment of human cancers. transcription level, forming a negative feedback loop Oncogene (2008) 27, 3371–3383; doi:10.1038/sj.onc.1211010; to maintain p53 protein at low levels under normal published online 28 January 2008 conditions (Barak et al., 1993; Wu et al., 1993). This feedback loop does not necessarily involve only MDM2, Keywords: chaperones; p53 mutant; stabilization; as other ligases were found to be responsible for degradation; CHIP degradation of p53 protein. Furthermore, different mechanisms were described to stabilize wild-type p53 protein, including inhibitors of transcription inducing stabilization of wild-type p53 via initial inhibition of MDM2 and p21WAF1 (Blagosklonny, 2004). Stabilization Introduction of mutated p53 in tumours is mostly ascribed to lack of induction of MDM2 by mutant p53, although recent Mutations of p53 generally result in the inactivation of genetic models have suggested that this is an over- its tumour-suppressor functions, and the role of simplification because mice engineered to express different p53 mutants in cancer progression is associated mutant p53 protein either with or without the wild-type either with wild-type independent oncogenic gain of allele show high levels of protein expression only in function or with trans-dominant suppression of wild- tumour tissue and not in normal tissue (Lang et al., type p53. While most tumour suppressor genes are 2004). An additional factor responsible for regulating the stability of p53 is interaction with molecular chaperones, Correspondence: Dr B Vojtesek, Department of Clinical and Experi- including Hsp70 and Hsp90 (Hinds et al., 1987; Selkirk mental Pathology, Masaryk Memorial Cancer Institute, Zluty kopec 7, et al., 1994). It was also shown that binding of Hsp90 to Brno 656 53, Czech Republic. p53 inhibits the ability of MDM2 to promote p53 E-mail: [email protected] Received 13 June 2007; revised 9 November 2007; accepted 4 December ubiquitination and degradation, resulting in the stabili- 2007; published online 28 January 2008 zation of both mutant p53 and MDM2 (Nagata et al., Stabilization and degradation of p53 mutants P Muller et al 3372 1999; Peng et al., 2001a, b). Although MDM2 plays a tion of p53 (PAb240-positive) in both untreated and central role in proteasomal targeting of p53, additional 17AAG-treated cells, while cells with mutation at ubiquitin ligases participate in normal cells, including codons R273H, R280K, R249S and L194F contain p300, Pirh2, COP1 and CHIP (carboxy terminus of both native (PAb1620-positive) and denatured confor- Hsp70-interacting protein) (Grossman et al., 2003; Leng mation of p53 (PAb240-positive). Inhibition of Hsp90 et al., 2003; Dornan et al., 2004; Esser et al., 2005). Esser by 17AAG leads to reduced levels of folded (PAb1620- et al. (2004) have identified a pathway for protein positive) p53 in these cells, indicating the essential role degradation that involves close cooperation of the of Hsp90 in folding these mutant proteins to adopt a molecular chaperones Hsp70 and Hsp90 with the partially native conformation. ubiquitin–proteasome system. The major influence on To correlate the stabilization of mutant p53 protein this degradation pathway is mediated by the chaperone- with conformation, the level of p53 was measured using associated ubiquitin ligase CHIP (Ballinger et al., 1999). two-site ELISA (Figure 2a) and western blotting In our study, we have extended these initial observations (Figure 2b) of the same lysate. Figure 2a represents and investigated the role of these chaperones in folding the relative level of p53 proteins, captured with DO-1 and stabilization of p53 mutants with different proper- monoclonal antibody and detected with CM1 polyclonal ties. We demonstrate a major role for CHIP ubiquitin sera, and was normalized to the total protein concentra- ligase in the degradation of unfolded p53 mutants, with tion in cell lysates. All mutants are present at higher little or no roles for CHIP in degrading wild-type or levels than the wild-type protein in MCF-7cells. mutant p53 proteins that adopt a folded conformation. However, the protein levels of mutants that display predominantly well-folded proteins (PAb1620-positive) are particularly high compared with proteins that adopt a predominantly denatured conformation. These results Results indicate that the unfolded mutant p53 proteins that do not rely on Hsp90 activity are less stable than folded Correlation of p53 level with conformation and impact of mutant p53 proteins that require Hsp90 activity, to Hsp90 achieve their final conformation. We analysed the folding of different p53 mutants using conformational-specific monoclonal antibodies. We simultaneously analysed the role of Hsp90 in mutant Negligible role of MDM2in degradation of p53 mutants p53 folding using 17AAG to inhibit Hsp90 activity To evaluate the importance of MDM2 in mutant p53 (Figure 1). In cells bearing the mutations E285K, degradation, cells were treated with Nutlin-3 that R175H and R156P, we detect only denatured conforma- specifically inhibits the interaction between p53 and

Figure 1 Conformation of mutated p53. Immunoprecipitation of p53 with PAb1620 and PAb240 was used to assess the folding status of p53. Cells were treated with Hsp90 inhibitor 17AAG (4 mM) for 12 h. Immunoprecipitated p53 was detected by CM1 antibody and the band intensity was measured by densitometry. Bar graph shows the ratio of folded (PAb1620-positive, black) and unfolded (PAb240-positive, grey) p53. 100% is the sum of band densities of p53 precipitated by PAb1620 and PAb240 antibody. This graph approximately reﬂects the ratio of folded and unfolded protein.

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3373

Figure 2 Determination of p53 mutant level using two-site ELISA. (a) Ninety-six-well plates were coated by DO-1 antibody and p53 protein was detected in equal amounts of whole protein lysates using p53 polyclonal antibody CM1. (b) Western blotting analysis of p53 protein level in the cell extract used for ELISA experiments. (* mostly folded mutants)

MDM2. Figure 3 shows the results for cell lines that (Figures 4a and b). Inhibiting Hsp90 with 17AAG contain mixtures of folded and denatured p53 (MDA- causes variable degrees of degradation of p53 mutants MB-231 and T47D) or only denatured p53 (HOS). and a minimal effect on wild-type p53 in MCF-7cells. Complexes were demonstrated in all cells using im- Nutlin-3 caused a progressive stabilization of wild-type munoprecipitation for p53 followed by western blotting p53 after Hsp90 inhibition. In contrast, mutant p53 was for MDM2. The p53–MDM2 interaction was disrupted not progressively stabilized but remained susceptible to by treatment with Nutlin-3 in all cell lines, as expected degradation after Nutlin-3 and 17AAG treatment (Figure 3). However, Nutlin-3 has no effect on p53 levels (Figure 4a). Thus, the decrease in mutant p53 stability in these cells, indicating the inability of MDM2 to seen after inhibition of Hsp90 activity is not dependent degrade both native and denatured conformations on MDM2. of mutant p53. To further deﬁne the role of MDM2 This result also identiﬁed the lower stability of in ubiquitination and degradation of p53, we used unfolded mutants in comparison with folded mutants Nutlin-3 treatment after Hsp90 inhibition with 17AAG (folded mutants are marked by *). To quantify the

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3374 and MDM2, whereas naturally folded mutants R273H and R280K require 17AAG treatment for ubiquitination and L194F ubiquitination is enhanced by 17AAG. As with the denatured mutants, MDM2 is not involved in ubiquitination of naturally folded mutant p53 proteins. To exclude a possible role of different genetic backgrounds in our panel of cell lines that might reflect intrinsic differences in MDM2, Hsp90 or other activities, p53-null H1299 cells were used to analyse the conformational status and degradation of two major p53 mutants, R175H and R273H, that are expressed in SKBR3 and MDA-MB-468 cells, respectively. Identical results were obtained with the mutants expressed in H1299 cells as were seen in the original cell lines, indicating that it is the intrinsic nature of the mutation that regulates stability and ubiquitination, not differences in the cellular background (data not shown). Thus, ubiquitination of mutant p53 molecules that adopt a native conformation is highly sensitive to inhibition of Hsp90 activity, whereas ubiquitination of Figure 3 Detection of p53–MDM2 complexes. Cells were treated with 10 mM Nutlin-3 for 3 h and after lysis in native conditions the unfolded p53 mutants occurs in the absence of Hsp90 p53–MDM2 complexes were detected by immunoprecipitation using activity. monoclonal anti-p53 antibody Bp53-10. (a) Detection of p53 in We also used p53À/À MDM2À/À mouse embryonic whole-cell lysate (b) and in immunoprecipitated fractions was fibroblasts (MEFs) to prove that p53 mutant degrada- performed using anti-p53 polyclonal antibody CM1. (c) MDM2 was detected in lysate or (d) immunoprecipitate using antibody 2A9. tion is not dependent on MDM2 but is related to Hsp90 activity. p53À/À MDM2À/À MEFs were transfected with human mutant p53 R175H and treated by chemically unrelated inhibitors of Hsp90 17AAG, degradation rate, we employed densitometric analysis of radicicol and novobiocin. Figure 6b shows that inhibi- immunoblots showing degradation of p53 mutant after tion of Hsp90 causes a decrease in the levels of this treatment by 17AAG. The curve representing the mutant protein and the additional inhibition of protea- decrease of p53 level was fitted by exponential function. some activity by MG-132 confirms that this mutant is This regression function was used to calculate the rate of ubiquitinated in the absence of active Hsp90. decay of p53 after treatment by 17AAG (Supplementary data 1). The graph in Figure 4b shows the time required Interaction of unfolded p53 and Hsp70/Hsc70 for the p53 level to decay to half of its initial value. In Cell line MDA-MB-468 bearing p53 R273H mutation addition, we used the inducible stable transfected hot- was employed to test the interaction of Hsp70 with p53. spot mutant p53 R175H (typical unfolded mutant) and This mutant exhibits structural stability comparable to mutant p53 R273H (fully-folded mutant) to show that wild-type p53 protein and its degradation is not the unfolded mutants exhibit distinctively longer half- influenced by MDM2 (Figure 6a). Combined treatment life in the same cellular background (Figure 5 and with 17AAG and MG-132 at either 37 or 39 1C was Supplementary data 2). applied to investigate the link between structural stability of p53, its folding and degradation. Immuno- Effect of 17AAG on ubiquitination of p53 mutants precipitation with PAb1620 shows that mutant p53 The above data are in agreement with the findings of R273H protein is folded and highly expressed at both Blagosklonny et al. (1995) that geldanamycin specifi- temperatures (Figure 7). 17AAG treatment, responsible cally affects mutated but not wild-type p53, indicating for elevation of unfolded p53 protein fraction, is that mutant p53 is degraded when Hsp90 is inhibited. associated with a slight degradation at 39 1C. Treatment To investigate the role of ubiquitination in Hsp90- with MG-132 shows that mutant p53 R273H is mediated degradation of different p53 mutants, cells ubiquitinated to a higher degree at both 37and 39 1C were treated with 17AAG and/or proteasome inhibitor after inhibition of Hsp90 with 17AAG. These results MG-132. Simultaneous treatment by Nutlin-3 was also support the idea that unfolded p53 protein is destined used to investigate MDM2 independent ubiquitination. for rapid degradation. Altogether, inhibition of Hsp90, Ubiquitination of wild-type p53 protein in MCF-7cells inhibition of proteasome and higher temperature requires MDM2 to interact with p53, seen by the lack of increase the ratio between the unfolded and folded p53 ubiquitination in the presence of Nutlin-3 (Figure 6a). fraction. Hsp90 inhibition has only a marginal effect on We were unable to detect complex formation between ubiquitination of wild-type p53. On the other hand, p53 and Hsp70 under physiological conditions at 37 1C. denatured mutants E285K, R249S, R175H and R156P Stable complex formation was detected either after are ubiquitinated independently of both Hsp90 inactivation of Hsp90 activity with 17AAG or during

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3375

Figure 4 Time-dependent degradation of p53 mutants after 17AAG and Nutlin-3 treatment. Equal numbers of cells were treated with 4 mM 17AAG or with 4 mM 17AAG and 10 mM Nutlin-3 and harvested at various times. The level of p53 was detected using western blotting and antibody DO-1. Folded mutants are labelled by ‘*’ (a). The immunoblots were analysed by densitometry; the column graph (b) shows the half-life of p53 after 17AAG treatment. Full regression analysis is presented in Supplementary data 1A–H. (X) No data were obtained, because the level of wild-type p53 is increased after Nutlin-3 treatment.

inhibition of proteasomal degradation with MG-132. CHIP is responsible for degradation of unfolded p53 The interaction between Hsp70 and mutant p53 was The above data show that Hsp70 speciﬁcally interacts further analysed after temperature shift to 39 1C. We with unfolded mutant p53 and targets it for degradation, found that this interaction is characterized by recruit- suggesting the possible role of Hsp70-associated ubiqui- ment of inducible form of Hsp70 (Hsc70) not only tin ligase CHIP in this process. To analyse the ability of following higher temperature exposure but also after CHIP to degrade unfolded p53, full-length or truncated 17AAG treatment. isoforms of CHIP lacking the U-Box domain were

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3376

Figure 5 Comparison of half-life of folded mutant p53 R273H and unfolded mutant p53 R175H. H1299-PERV3 lacZeo3 cells containing either mutant p53 R273H or mutant p53 R175H recombined in the same locus were used. The cells were continuously treated with 5 mM Ponasterone A to induce the expression of p53 protein. The Ponasterone A treatment was disrupted in several time intervals to measure the degradation rate. Cells were treated with 10 mM Nutlin-3 or 4 mM 17AAG after removal of Ponasterone A (c). The level of p53 was detected by DO-1 antibody (a). The immunoblot was analysed by densitometry (Supplementary data 2C) and column graph (b) shows computed half-life of p53 mutants R175H and R273H in cells treated with Nutlin-3 or 17AAG.

transiently transfected in to H1299-175 cells expressing character of the less stable complex formed by full- unfolded p53 R175H mutant (Figure 8). Our data length CHIP, which results in ubiquitination and confirmed the ability of full-length CHIP to degrade degradation of denatured p53. The fact that the protein mutant p53 R175H, with higher degradation rate in the complex between p53, Hsp70 and CHIP is assembled presence of 17AAG (Figure 8a). Transfection of especially after Hsp90 inhibition emphasizes the role of truncated 200AA CHIP lacking U-Box region did not CHIP in the degradation of unfolded p53. destabilize unfolded p53 mutant, but attenuated the effect of 17AAG on p53 degradation. These data demonstrate the requirement of both TRP and U-Box Discussion domain for degradation of unfolded p53 mutants. Similar data were obtained using H1299-273 cells. The aim of our study has been to elucidate the role of Flow cytometry was additionally used to quantify the Hsp90 and other chaperones in folding and stabilization effect of CHIP on the level of mutant p53 R175H of different p53 mutants. Regulation of stability of (Figure 8b and Supplementary data 3). Full-length unfolded p53 protein is mediated through Hsp90, so it CHIP markedly decreased the level of p53 in transfected precedes negative feedback loop regulation and results in cells, whereas the truncated 200AA CHIP had the either folding or degradation of p53. Our work shows that opposite effect, indicating that this form of CHIP is increased Hsp90 activity and the preference of folding is responsible for the increase of p53 level but is also another important condition for expression of high levels capable in blocking the effect of inhibition of Hsp90 by of mutant p53. This finding can explain why mouse 17AAG. models that inherit homozygous p53 mutations express the mutant p53 protein at low levels in differentiated tissues (Lang et al., 2004). Therefore, alterations of the CHIP forms a complex with Hsp70 and ubiquitinates balance of folding/degradation can be responsible for bound p53 tumour-specific stabilization of mutant p53. The immunoprecipitation of a complex between Hsp70, We show that p53 mutants adopting mainly unfolded p53 and CHIP was performed to clearly demonstrate the conformations are associated with lower protein levels, mechanism of p53 ubiquitination by CHIP. Figure 9 implying that these p53 mutants exhibit higher levels of shows that full-length CHIP induces ubiquitination of degradation. These mutants require inhibition of Hsp90 p53 R273H and the effect is enhanced if the folding is for ubiquitination, whereas ubiquitination of other hot- blocked by 17AAG. The dominant negative 200AA spot mutants proceeds in the absence of Hsp90 activity. CHIP is unable to induce ubiquitination of p53 even We also showed that unfolded p53 mutants interact with though it forms a stable complex with Hsp70 and Hsp70, and that CHIP can induce degradation of these p53. The relatively higher stability of the complex mutants. The data suggest that ubiquitination of mutant with truncated 200AA CHIP underlines the temporal p53 is a consequence of its denaturation and interaction

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3377

Figure 6 (a) MDM2-independent ubiquitination of p53 mutants. Cells were treated with 4 mM 17AAG to inhibit Hsp90 and to increase p53 unfolded fraction. Proteasome inhibitor MG-132 (20 mM) was used to inhibit degradation of ubiquitinated p53. Nutlin-3 (10 mM) was used to disclose the role of MDM2 in the ubiquitination process. (b) p53À/À MDMÀ/À MEFs were transfected with human p53 R175H and treated with 4 mM 17AAG, 4 mM radicicol, 10 mM novobiocin and MG-132 (20 mM). (* mostly folded mutants). with Hsp70, leading to CHIP-mediated ubiquitination of denatured p53 in tumour cells. Our data also show and proteasomal degradation. that Hsp90 is responsible for proper folding of p53 or Many p53 mutants responsible for transformation forms a stable complex with mutants that adopt display structural defects and associate with the unfolded conformations. Therefore, the activity of molecular chaperones Hsc70 and Hsp90 in tissue culture Hsp90 and its relationship with Hsp70 and other co- cells and human tumours (Soussi and Lozano, 2005). chaperones are crucial factors that determine stabiliza- We showed that increased temperature and/or inhibition tion or degradation of newly synthesized proteins, of Hsp90 by 17AAG lead to an elevation in the fraction depending on the precise mutation and the effect on

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3378

Figure 7 Analysis of p53 R273 H conformation and its interaction with Hsp70. MDA-MB-468 cells containing folded p53 R273H were treated for 8 h at 37or 39 1C. Cells were treated with 4 mM 17AAG (inhibition of Hsp90) and with 20 mM MG-132 (proteasome inhibition). The level of Hsp90, Hsp70 and p53 was detected using western blotting. Conformational changes of p53 were assessed using immunoprecipitation with conformation-speciﬁc monoclonal antibodies PAb1620 and PAb240. Complexes between p53 and Hsp70 were analysed using immunoprecipitation of p53 with rabbit polyclonal antibody CM1.

p53 conformation. In particular, conformationally tumour-specific stabilization (Lang et al., 2004; Olive unfolded p53 mutants form a complex with Hsc70 and et al., 2004; Soussi and Lozano, 2005). One possibility is are addressed to degradation by CHIP if Hsp90 activity that p53 modifications such as acetylation or phospho- is low. Temperature shift to 39 1C further augments this rylation are responsible for this tumour-specific stabili- effect, both by induction of CHIP and Hsp70 homo- zation (Lang et al., 2004; Olive et al., 2004). However, logues and also by further destabilization of p53 because Hsp90 exists in cancer cells in a highly activated structure. Although all p53 missense mutants studied state, susceptible to bind to its inhibitor 17AAG here exhibit increased stability compared to wild-type, a approximately 100 times more tightly compared with quantitative analysis and comparison with different Hsp90 expressed in normal cells (Kamal et al., 2003; conformation-specific antibodies show for the first time Workman, 2004), our data provide an additional that proteins that adopt a native conformation are explanation for tumour-specific stabilization by protect- particularly stable. Experiments performed in trans- ing against interaction with Hsp70 and CHIP. The level fected H1299 cells indicate that these differences are not of this unfolded protein can then reflect differences in due to different cellular backgrounds. In human chaperone constitution in normal and cancer cells, tumours, the level of mutant p53 detected by immuno- where assembly of multichaperone complexes and high histochemistry shows a broad range (Nenutil et al., Hsp90 activity will stabilize mutated p53 in cancer. The 2005), suggesting that specific mutation site may be one relationship between p53 mutants and Hsp90 activity is factor that influences the overall level of p53 in primary further supported by the recent demonstration that human tumours. The model explaining the longer half- Hsp90b coding gene repression is mediated by wild-type life and higher expression of folded mutants is described p53 (Zhang et al., 2004), suggesting that the develop- in Supplementary data 4. ment of p53 mutation and increased Hsp90 activity can The inherent stability and dominant gain-of-function be coincident processes in tumorigenesis. These factors properties of mutant p53 protein in tumours have been may also influence the ability of mutant p53 to exhibit questioned by the analysis of p53 in knock-in p53 mice dominant effects. In general, proteins that adopt only (Lang et al., 2004; Olive et al., 2004). Levels of p53 the denatured conformation are less stable and, there- R172H and R270H are very low in normal tissues and fore, may be less effective in inhibiting wild-type p53 are stabilized only in some cancers in these mice, functions, which may also help to explain the different indicating that other cellular factors are involved in tumorigenic properties observed for different mutations

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3379

Figure 8 Degradation of mutant R175H mediated by CHIP. (a) Full-length CHIP contains N-terminal Flag, tetratricorepeat (TPR) domain and U-Box domain with ubiquitin ligase activity. Truncated CHIP lacking U-Box domain was tested as a dominant negative mutant. H1299-175 cells were transiently transfected with empty pcDNA3 or pcDNA3 vectors expressing full-length CHIP and 200AA CHIP, respectively. These cells were also treated with 4 mM 17AAG to evaluate the role of Hsp90 in this process. CHIP was detected by anti-Flag antibody. (b) The level of mutant p53 R175H in cells was analysed using ﬂow cytometry. Cells transfected with either Flag-200AA CHIP or Flag-CHIP were gated as nontransfected Flag-negative (grey colour column) and as transfected Flag-positive (black column). The statistical analysis of the p53 level in the individual cell population is displayed as 25–75 percentile (column) and 5–95 percentile (bar), as a median (*) and as an arithmetic mean (O). Details of ﬂow cytometry are described in Supplementary data 3A–F.

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3380

Figure 9 Analysis of complex between mutant p53 R273H, Hsp70 and CHIP. H1299 cells were cotransfected with p53 R273H with either GFP or Flag-CHIP or Flag-200AA CHIP. Cells were treated 24 h after transfection with 4 mM 17AAG for 6 h. Mutated p53 protein, actin, Hsp90, Hsp70 and CHIP were detected in protein lysate (left panels). Protein complexes between mutant p53, Hsp70 and CHIP were detected by immunoprecipitation (right panels). Antibodies DO-1 anti-p53 and 9H10 anti-Flag were used for immunoprecipitation. Antibodies H70.1, DO-1-HRP and anti-Flag-HRP were used to detect Hsp70, p53 and transfected Flag-CHIP, respectively.

expressed in heterozygous mutant mice (Olive et al., histone deacetylase (HDAC) inhibitors (such as FK228, 2004). An alternative proposal that post-translational LAQ824 or LBH589) were also found to be able to modifications may be required for mutant p53 to inhibit inhibit Hsp90 and alter the function of Hsp70 (Bali wild-type p53 (Lang et al., 2004) is not supported by the et al., 2005; Wang et al., 2007). Accordingly, the findings that p53 phosphorylation does not correlate inhibition of Hsp90 by HDAC inhibitors may be with transcriptional activity in primary human breast, involved in depletion of mutant p53 caused by ovary and cervix tumours (Nenutil et al., 2005). trichostatin A or FR901228 treatment (Blagosklonny It was shown earlier that the loss of p53 trans- et al., 2005). Furthermore, it was shown that HDAC function is sufficient for its stabilization without any inhibitors potentiate their anti-tumour effect with change in conformation (Blagosklonny, 2004). More- 17AAG and promote proteasomal degradation of over, we found that mutant p53 proteins exhibiting Hsp90 clients such as Bcr-Abl (Nimmanapalli et al., wild-type conformation are expressed at higher levels. 2003; George et al., 2005). Our data indicate that unfolded p53 mutant proteins In summary, we found that Hsp90 plays a key role in interact with Hsp70 and are hence available for CHIP- the folding of p53 mutants with native conformation. mediated degradation. This finding is also in agreement The levels of denatured p53 mutants are generally lower with the theoretical prediction that degradation of p53 than natively folded p53 mutants and the Hsp90 mutant is not associated with MDM2 feedback loop inhibitor 17AAG decreases the level of denatured p53 (Blagosklonny, 1997). Higher expression of folded p53 mutants, indicating that Hsp90 prevents degradation of mutants was also observed in knock-in mouse models denatured p53 mutants. Surprisingly, MDM2 is unable bearing either folded or unfolded p53 mutants (Jackson to ubiquitinate p53 mutants even though it forms stable et al., 2005). Chaperone-mediated degradation of p53 complexes, leaving space for the function of other may also be involved in other drug-induced degradation ubiquitin ligases such as CHIP. Our model showing the of p53 mutants. Besides geldanamycin analogues, balance between folding and degradation of p53 by several groups of drugs were found to influence the CHIP is described in Figure 10. The relevancy of ability of Hsp90 to fold its client proteins. Different cooperation between heat-shock proteins and p53 for

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3381

Figure 10 Folding–degradation balance of the p53 protein. Native or denatured p53 protein is bound by Hsp70 (I). In collaboration with other co-chaperones, the unfolded p53 is sent to the next processing. Especially HOP (II) helps to link Hsp70 substrate to Hsp90 and enables the folding of p53 by Hsp90 (King et al., 2001; Zylicz et al., 2001). Naturally unfolded mutants, such as p53 R175H, form stable complex with chaperones and Hsp90 activity is crucial to prevent their degradation (III). If the Hsp90 activity is lower or blocked by 17AAG (IV), the denatured mutants bind complex Hsp70 and CHIP (V). Tetratricorepeat (TPR) domain of CHIP binds the C terminus of Hsp70 and the U-box domain is responsible for ubiquitination of unfolded p53. The ubiqutinated p53 protein is degraded in proteasome (VI). cancer therapy was recently shown by O’Shea et al. fetal bovine serum, 300 mg lÀ1 L-glutamine and penicillin/ (2005). The authors demonstrate that the induction of streptomycin. Cells were grown to 80% confluence before heat-shock or treatment with 17AAG selectively sensi- experimental treatments. 17AAG (17-(allylamino)-17-de- tizes resistant tumour cells to ONYX-015 oncolytic methoxygeldanamycin), Nutlin-3 and MG-132 were added to replication (O’Shea et al., 2005). a final concentration of 2–20 mM. CHIP was recently found to be a regulator of Hsp70 as well as p53 level. Thus, lower expression or activity Antibodies of CHIP might be one reason for overexpression of We used the following anti-p53 antibodies for immuno- À1 chaperone complexes in cancer and stabilization of p53 precipitation (1 mg per reaction) or western blotting (1 mgml ): protein. The ability of CHIP to degrade both Hsp70 and monoclonal antibody DO-1 recognizing the N-terminus of the protein (Vojtesek et al., 1992), Bp53-10 monoclonal antibody its clients makes CHIP a candidate tumour suppressor recognizing the C-terminus of the protein (Bartek et al., 1993), protein. In addition, our data indicate that manipulating PAb1620 monoclonal antibody directed towards the wild-type CHIP activity in cells may prove useful for the treatment conformation of human p53 protein (Milner, 1997; Wang of tumours that bear specific types of p53 mutation. In et al., 2001), PAb240 monoclonal antibody directed towards this respect, careful characterization of the ability of the mutant conformations of p53 (Gannon et al., 1990) and Hsp90, Hsp70 and CHIP to regulate p53 stability and CM1 rabbit polyclonal antibody. MDM2 was detected by conformation will be necessary for the development of monoclonal antibody 2A9 (Chen et al., 1993), Hsp90 by therapies that are designed to interfere with mutant p53 monoclonal antibody AC88 (StressGen, Canada), and Hsp70 activities in tumours. and Flag epitope by in-house mouse monoclonal antibodies.

ELISA assay A sandwich immunoassay to measure the level of p53 protein Materials and methods in total cell extracts was performed as described (Vojtesek et al., 1993). The immunoassay was performed using mono- Cell culture and treatment clonal antibody DO-1 as the solid-phase reagent and Human cancer cell lines HOS, BT474, MDA-MB-468, MDA- polyclonal rabbit antiserum CM-1 to detect the captured MB-231, T47D, BT549 and SKBR3 contain different p53 protein. ELISA was quantiﬁed using linear regression of the mutations (R156P, E285K, R273H, R280K, L194F, R249S relationship between optical density and total protein con- and R175H, respectively), whereas the wild-type allele was lost centration (quantifying speciﬁc antibody concentrations by in all these cell lines. MCF-7cells contain wild-type p53, and enzyme-linked immunosorbent assay using slope correction) H1299 cells are p53-null. Cell lines H1299-neo, H1299-175 and (Barrette et al., 2006). H1299-273 were developed by stable transfection of empty pcDNA3 vector or pcDNA3 containing the p53 hot-spot mutants R175H and R273H, respectively. MEFs from p53À/À Immunoblotting and immunoprecipitation MDM2À/À mice were generously provided by Dr G Lozano Treated and control cells were collected by centrifugation, (Lang et al., 2004). Cells were cultured in D-MEM, 10% washed three times with ice-cold PBS and lysed in 150 mM

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3382 NaCl; 50 mM TrisCl, pH 8.0; 5 mM EDTA; 1% NP-40; 1 mM cell line H1299-PERV3 lacZeo3. The plasmid pOG44 PMSF buffer supplemented with protease inhibitor cocktail (Invitrogen Corp., USA) containing Flp recombinase was (Sigma-Aldrich, Inc., USA). A 20 mg portion of total proteins used for recombination. was separated by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) on 10 or 12.5% gels and transferred onto Transfection and FLP recombination nitrocellulose membranes. Membranes were blocked in 5% 6 milk and 0.1% Tween 20 in PBS and probed overnight with 10 H1299 cells were transfected with 2 mg of plasmid DNA in specific monoclonal antibodies or rabbit polyclonal sera. To nucleofector Kit-C solution using program X-005 (Amaxa GmbH, Germamy). Transfection efficiency was evaluated confirm equal protein loading, immunodetection was performed with the anti-actin AC-40 monoclonal antibody. using pmaxGFP. Cells were treated and harvested 36 h after transfection. H1299 cells intended for stable integration of p53 Peroxidase-conjugated rabbit anti-mouse immunoglobulin or À1 swine anti-rabbit immunoglobulin antisera (DAKO Ltd., mutants were treated with geneticin G418 (3 mg ml ). Stable Denmark) were used as the secondary antibodies and clones forming colonies were selected within 4 weeks and visualized with ECL reagents from Amersham-Pharmacia tested for p53 expression. The Flp-In system (Invitrogen (Little Chalfont, England). Corp., USA) was used to make integration of different p53 Immunoprecipitation of conformationally different forms of mutants in H1299 cells at a specific genomic location. H1299 p53 protein was performed by using 1 mg of mouse monoclonal cells were transfected by pFRT/lacZeo construct to generate clones containing single FRT site for Flp-mediated recombi- antibodies DO-1, PAb240 or PAb1620, as described previously À1 (Gannon et al., 1990; Milner, 1997). For co-immunoprecipita- nation. Clones were selected by Zeocin (100 mgml ) and were tested for lacZ activity. The correct integration of single copy tion of p53 complexes, cells were lysed in 150 mM NaCl, 50 mM HEPES (pH 7.2) and 0.5% Tween 20 and sonicated. Cell lysate of the pFRT/lacZeo construct was tested by Southern blotting (200 ml) containing 200–800 mg of total protein was incubated and real-time PCR of genomic DNA. Selected clone H1299 lac with 2 mg of appropriate antibody. A 15 ml volume of protein Zeo was than stably transfected with PERV3 construct G-conjugated magnetic beads (Dynal, Norway) was used to (Stratagene Inc., USA) to produce clones expressing modified precipitate the antibody. After three washes in lysis buffer, ecdysone receptor VgEcR and transcription co-activator RXR. The clones were selected by geneticin G418 (3 mg mlÀ1) beads were resuspended in 60 mlof2Â concentrated Laemmli buffer. For electrophoresis, 8 ml of samples were used. and were tested for expression of VgEcR and RXR by anti-VP-16 anti RXR monoclonal antibody (Santa Cruz Biotechnology, USA; sc-7545 and sc-46659). The resultant Constructs clone, called H1229-PERV3 laczeo3, was used for integration CHIP cDNA was amplified using total RNA and one-step of the construct pEGSH FRT containing ecdysone-inducible RT-PCR with primers 50-CTTAAGAGCGCGGGGACAGG-30 0 0 promoter and either mutant p53 R175H or mutant p53 and 5 -ACAGATAGTGGCGGCAATGG-3 . Nested forward R273H. The recombination of the p53 mutant constructs primer 50-gacaGGTACCATGGATTACAAGGATGACGAC 0 was facilitated by transient transfection of pOG44 expressing GATAAGaagggcaaggaggagaaggag-3 containing N-terminal Flp recombinase. Successfully recombined clones containing Flag (italics) and KpnI restriction site (underlined) and 0 0 correctly integrated pEGSH p53 R175H or R273H were reverse primers 5 -tctgGAATTCtcagtagtcctccacccagcc-3 or À1 0 0 selected with 500 mgml Hygromycin (recombination is 5 -tctgGAATTCctagtacttgtcgtgcttggcctc-3 containing EcoRI schematically presented in Supplementary data 2A). restriction site (underlined) and stop codon (italics) were used for amplification and cloning of full-length and truncated CHIP into pcDNA3 vector. p53 mutants R175H and R273H Acknowledgements were expressed in pCDNA3 plasmid. Inducible p53 mutants R175H and R273H were cloned using gateway (Invitrogen) We are thankful to Dr G Lozano for generous gift of MEFs into pEGSH FRT GW, a newly constructed plasmid and Dr PJ Coates and E Michalova for critical reading of the containing ecdysone-inducible promoter and FRT recombina- manuscript. DC, DPL and PM were supported by IMCB tion site (Supplementary data 2B). The plasmids PERV3 Singapore. This work was supported by a grant from IGA MZ (Stratagene Inc., USA) and pFRT/lacZeo (Invitrogen Corp., CR (NR8338-3/2005, GA CR (301/05/0416) and Grant USA) were used to make stable isogenic, ecdysone-inducible MSMT LC06035.

References

Bali P, Pranpat M, Bradner J, Balasis M, Fiskus W, Guo F et al. paraffin sections using a series of novel monoclonal antibodies. (2005). Inhibition of histone deacetylase 6 acetylates and disrupts the J Pathol 169: 27–34. chaperone function of heat shock protein 90: a novel basis for Blagosklonny MV. (1997). Loss of function and p53 protein antileukemia activity of histone deacetylase inhibitors. J Biol Chem stabilization. Oncogene 15: 1889–1893. 280: 26729–26734. Blagosklonny MV. (2004). Flavopiridol, an inhibitor of transcription: Ballinger CA, Connell P, Wu Y, Hu Z, Thompson LJ, Yin LY implications, problems and solutions. Cell Cycle 3: 1537–1542. et al. (1999). Identification of CHIP, a novel tetratricopeptide Blagosklonny MV, Toretsky J, Neckers L. (1995). Geldanamycin repeat-containing protein that interacts with heat shock proteins selectively destabilizes and conformationally alters mutated p53. and negatively regulates chaperone functions. Mol Cell Biol 19: Oncogene 11: 933–939. 4535–4545. Blagosklonny MV, Trostel S, Kayastha G, Demidenko ZN, Barak Y, Juven T, Haffner R, Oren M. (1993). mdm2 expression is Vassilev LT, Romanova LY et al. (2005). Depletion of mutant induced by wild type p53 activity. EMBO J 12: 461–468. p53 and cytotoxicity of histone deacetylase inhibitors. Cancer Res Barrette RW, Urbonas J, Silbart LK. (2006). Quantifying specific 65: 7386–7392. antibody concentrations by enzyme-linked immunosorbent assay Bullock AN, Fersht AR. (2001). Rescuing the function of mutant p53. using slope correction. Clin Vaccine Immunol 13: 802–805. Nat Rev Cancer 1: 68–76. Bartek J, Bartkova J, Lukas J, Staskova Z, Vojtesek B, Lane DP. Chen J, Marechal V, Levine AJ. (1993). Mapping of the p53 and mdm- (1993). Immunohistochemical analysis of the p53 oncoprotein on 2 interaction domains. Mol Cell Biol 13: 4107–4114.

Oncogene Stabilization and degradation of p53 mutants P Muller et al 3383 Dornan D, Wertz I, Shimizu H, Arnott D, Frantz GD, Dowd P et al. Nimmanapalli R, Fuino L, Bali P, Gasparetto M, Glozak M, Tao J (2004). The ubiquitin ligase COP1 is a critical negative regulator of et al. (2003). Histone deacetylase inhibitor LAQ824 both lowers p53. Nature 429: 86–92. expression and promotes proteasomal degradation of Bcr-Abl and Esser C, Alberti S, Hohfeld J. (2004). Cooperation of molecular induces apoptosis of imatinib mesylate-sensitive or -refractory chaperones with the ubiquitin/proteasome system. Biochim Biophys chronic myelogenous leukemia-blast crisis cells. Cancer Res 63: Acta 1695: 171–188. 5126–5135. Esser C, Scheffner M, Hohfeld J. (2005). The chaperone-associated Olive KP, Tuveson DA, Ruhe ZC, Yin B, Willis NA, Bronson RT ubiquitin ligase CHIP is able to target p53 for proteasomal et al. (2004). Mutant p53 gain of function in two mouse models of degradation. J Biol Chem 280: 27443–27448. Li–Fraumeni syndrome. Cell 119: 847–860. Gannon JV, Greaves R, Iggo R, Lane DP. (1990). Activating O’Shea CC, Soria C, Bagus B, McCormick F. (2005). Heat shock mutations in p53 produce a common conformational effect. A phenocopies E1B-55K late functions and selectively sensitizes monoclonal antibody speciﬁc for the mutant form. EMBO J 9: refractory tumor cells to ONYX-015 oncolytic viral therapy. Cancer 1595–1602. Cell 8: 61–74. George P, Bali P, Annavarapu S, Scuto A, Fiskus W, Guo F et al. Peng Y, Chen L, Li C, Lu W, Agrawal S, Chen J. (2001a). Stabilization (2005). Combination of the histone deacetylase inhibitor LBH589 of the MDM2 oncoprotein by mutant p53. J Biol Chem 276: and the hsp90 inhibitor 17AAG is highly active against human 6874–6878. CML-BC cells and AML cells with activating mutation of FLT-3. Peng Y, Chen L, Li C, Lu W, Chen J. (2001b). Inhibition of MDM2 by Blood 105: 1768–1776. hsp90 contributes to mutant p53 stabilization. J Biol Chem 276: Grossman SR, Deato ME, Brignone C, Chan HM, Kung AL, 40583–40590. Tagami H et al. (2003). Polyubiquitination of p53 by a ubiquitin Selkirk JK, Merrick BA, Stackhouse BL, He C. (1994). Multiple p53 ligase activity of p300. Science 300: 342–344. protein isoforms and formation of oligomeric complexes with heat Hinds PW, Finlay CA, Frey AB, Levine AJ. (1987). Immunological shock proteins Hsp70 and Hsp90 in the human mammary tumor, evidence for the association of p53 with a heat shock protein, hsc70, T47D, cell line. Appl Theor Electrophor 4: 11–18. in p53-plus-ras-transformed cell lines. Mol Cell Biol 7: 2863–2869. Soussi T, Lozano G. (2005). p53 mutation heterogeneity in cancer. Jackson EL, Olive KP, Tuveson DA, Bronson R, Crowley D, Biochem Biophys Res Commun 331: 834–842. Brown M et al. (2005). The differential effects of mutant p53 alleles Wang PL, Sait F, Winter G. (2001). The ‘wildtype’ conformation on advanced murine lung cancer. Cancer Res 65: 10280–10288. of p53: epitope mapping using hybrid proteins. Oncogene 20: Kamal A, Thao L, Sensintaffar J, Zhang L, Boehm MF, Fritz LC et al. 2318–2324. (2003). A high-afﬁnity conformation of Hsp90 confers tumour Wang Y, Wang SY, Zhang XH, Zhao M, Hou CM, Xu YJ et al. selectivity on Hsp90 inhibitors. Nature 425: 407–410. (2007). FK228 inhibits Hsp90 chaperone function in K562 cells King FW, Wawrzynow A, Hohfeld J, Zylicz M. (2001). Co-chaperones via hyperacetylation of Hsp70. Biochem Biophys Res Commun 356: Bag-1, Hop and Hsp40 regulate Hsc70 and Hsp90 interactions with 998–1003. wild-type or mutant p53. EMBO J 20: 6297–6305. Vojtesek B, Bartek J, Midgley CA, Lane DP. (1992). An immuno- Lang GA, Iwakuma T, Suh YA, Liu G, Rao VA, Parant JM et al. chemical analysis of the human nuclear phosphoprotein p53. New (2004). Gain of function of a p53 hot spot mutation in a mouse monoclonal antibodies and epitope mapping using recombinant p53. model of Li-Fraumeni syndrome. Cell 119: 861–872. J Immunol Methods 151: 237–244. Leng RP, Lin Y, Ma W, Wu H, Lemmers B, Chung S et al. (2003). Vojtesek B, Fisher CJ, Barnes DM, Lane DP. (1993). Comparison Pirh2, a p53-induced ubiquitin-protein ligase, promotes p53 between p53 staining in tissue sections and p53 proteins levels degradation. Cell 112: 779–791. measured by an ELISA technique. Br J Cancer 67: 1254–1258. Milner J. (1997). Structures and functions of the tumor suppressor p53. Workman P. (2004). Altered states: selectively drugging the Hsp90 Pathol Biol (Paris) 45: 797–803. cancer chaperone. Trends Mol Med 10: 47–51. Nagata Y, Anan T, Yoshida T, Mizukami T, Taya Y, Fujiwara T et al. Wu X, Bayle JH, Olson D, Levine AJ. (1993). The p53–mdm-2 (1999). The stabilization mechanism of mutant-type p53 by impaired autoregulatory feedback loop. Genes Dev 7: 1126–1132. ubiquitination: the loss of wild-type p53 function and the hsp90 Zhang Y, Wang JS, Chen LL, Cheng XK, Heng FY, Wu NH association. Oncogene 18: 6037–6049. et al. (2004). Repression of hsp90beta gene by p53 in UV Nenutil R, Smardova J, Pavlova S, Hanzelkova Z, Muller P, Fabian P irradiation-induced apoptosis of Jurkat cells. J Biol Chem 279: et al. (2005). Discriminating functional and non-functional p53 42545–42551. in human tumours by p53 and MDM2 immunohistochemistry. Zylicz M, King FW, Wawrzynow A. (2001). Hsp70 interactions with J Pathol 207: 251–259. the p53 tumour suppressor protein. EMBO J 20: 4634–4638.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Oncogene