MOLECULAR AND CELLULAR BIOLOGY, May 2007, p. 3542–3555 0270-7306/07/$08.00⫹0 doi:10.1128/MCB.01595-06 Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 27, No. 9
The MDM2 Ubiquitination Signal in the DNA-Binding Domain of p53 Forms a Docking Site for Calcium Calmodulin Kinase Superfamily Members䌤 Ashley L. Craig,1 Jennifer A. Chrystal,1 Jennifer A. Fraser,1 Nathalie Sphyris,1 Yao Lin,1 Ben J. Harrison,1 Mary T. Scott,1† Irena Dornreiter,2 and Ted R. Hupp1* University of Edinburgh, Cancer Research Centre, CRUK p53 Signal Transduction Group, South Crewe Road, Edinburgh, United Kingdom EH4 2XR,1 and Heinrich-Pette-Institut fur Experimentelle Virologie und Immunologie, Universitat Hamburg, Hamburg, Germany2 Received 26 August 2006/Returned for modification 16 October 2006/Accepted 31 January 2007
Genetic and biochemical studies have shown that Ser20 phosphorylation in the transactivation domain of p53 mediates p300-catalyzed DNA-dependent p53 acetylation and B-cell tumor suppression. However, the protein kinases that mediate this modification are not well defined. A cell-free Ser20 phosphorylation site assay was used to identify a broad range of calcium calmodulin kinase superfamily members, including CHK2, CHK1, DAPK-1, DAPK-3, DRAK-1, and AMPK, as Ser20 kinases. Phosphorylation of a p53 transactivation domain fragment at Ser20 by these enzymes in vitro can be mediated in trans by a docking site peptide derived from the BOX-V domain of p53, which also harbors the ubiquitin signal for MDM2. Evaluation of these calcium calmodulin kinase superfamily members as candidate Ser20 kinases in vivo has shown that only CHK1 or DAPK-1 can stimulate p53 transactivation and induce Ser20 phosphorylation of p53. Using CHK1 as a prototypical in vivo Ser20 kinase, we demonstrate that (i) CHK1 protein depletion using small interfering RNA can attenuate p53 phosphorylation at Ser20, (ii) an enhanced green fluorescent protein (EGFP)–BOX-V fusion peptide can attenuate Ser20 phosphorylation of p53 in vivo, (iii) the EGFP–BOX-V fusion peptide can selectively bind to CHK1 in vivo, and (iv) the ⌬p53 spliced variant lacking the BOX-V motif is refractory to Ser20 phosphorylation by CHK1. These data indicate that the BOX-V motif of p53 has evolved the capacity to bind to enzymes that mediate either p53 phosphorylation or ubiquitination, thus controlling the specific activity of p53 as a transcription factor. domain containing a binding site for the cofactor p300 (27). The BOX-I domain also contains the FXXWXXXL consensus MDM2 binding site that is required for MDM2-mediated inhibition of p53. DNA damage-activated protein kinases like CHK1/2 modify the BOX-I domain of p53 at Thr18 and Ser20 (46) by an allosteric mechanism (10). Phosphorylation of p53 at Thr18 and Ser20 can in turn differentially modulate the binding of the coactivator p300 or the inhibitor MDM2. Phosphorylation at Thr18 has the most striking effect on blocking MDM2 binding (13), while phosphorylation at Ser20 has no effect on MDM2 binding (45). Rather, phosphorylation at Ser20 creates a p300-phospho-consensus LXXLL binding site in the activation domain of p53 (16), within a novel phosphoLXXLL peptide-binding module in p300 (18, 21). Thus, these phosphorylation events can stimulate the p53 transcriptional response. Mutation of Ser20 to Ala20 can reduce the specific activity of p53 in vitro (52) and in vivo, increasing cancer incidence in mice (36), suggesting that Ser20 phosphorylation is a key modifier of the p53 response in some cell types, including B cells. Consistent with this, phosphomimetic amino acid substitutions at Ser20 or Thr18 can increase the specific activity of p53 as a transcription factor (28). As such, identification of the stressactivated Ser20 kinases is an important goal in understanding the mechanisms underlying p53 activation as a tumor suppressor. The enzymes that modify p53 within the transactivation domain at Thr18 and Ser20 are beginning to be defined, although the data are relatively controversial. CHK2 and CHK1 were the original enzymes reported to modify both Thr18 and
The tumor suppressor protein p53 is activated as a transcription factor in response to a variety of genotoxic and metabolic stresses, resulting in alterations in gene expression along with either transient cell cycle arrest or apoptosis, depending on cell type and the severity of damage (24). Several functional domains of p53 are involved in cooperating in its transactivation function, which include (i) an LXXLL-type transactivation domain that interacts with p300 (3), (ii) a proline repeat transactivation domain that binds directly to p300 and mediates DNA-dependent acetylation (17); (iii) a central sequence-specific DNA-binding domain that harbors most of the mutations found in p53 (39), (iv) a tetramerization domain (37), and (v) a multifunctional C-terminal regulatory domain whose phosphorylation stimulates DNA binding (33). This combined action of phosphorylation and acetylation clamps the p300 coactivator to a promoter and recruits chromatin-remodeling enzymes that cooperate in transcription activation (4, 15, 19). The N-terminal conserved BOX-I domain of p53 encompasses approximately 15 amino acids and has evolved as a multiprotein binding domain resulting in an interaction with a set of acetyltransferases, ubiquitin ligases, and protein kinases. The BOX-I domain harbors the LXXLL-type transactivation
* Corresponding author. Mailing address: University of Edinburgh, Cancer Research Centre, South Crewe Road, Edinburgh, United Kingdom, EH4 2XR. Phone: 44-131-777-3500. Fax: 44-131-777-3520. E-mail:
[email protected]. † Present address: CRUK Beatson Laboratories, Glasgow, Scotland. 䌤 Published ahead of print on 5 March 2007. 3542
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
Ser20 (46), and CHK2 deletion reduces the specific activity of p53 in irradiated cells (35, 50). Furthermore, one study indicates that CHK2 phosphorylates p53 at only Thr18 and Ser20 in the activation domain (10), while another report has indicated that CHK2 modifies multiple sites in the C-terminal domain of p53 (41); thus, an understanding of where CHK2 modifies p53 is still unclear at a mechanistic level. Furthermore, additional studies indicate that neither CHK2 nor CHK1 may be a radiation-induced Ser20 kinase in vivo (2, 30); thus, it is not evident what the actual Ser20 kinase(s) for p53 protein might be. The Ser20 site in the transactivation domain of p53 will presumably be the target of distinct stress-activated signaling pathways containing kinases recruited to phosphorylate and stimulate p53 transcriptional activity. In order to begin to identify the class of Ser20 kinases that exist for p53, we have set up a cell-free Ser20 kinase assay which demonstrated that a subset of the calcium calmodulin kinase superfamily can function as Ser20 kinases. These enzymes have the property of being activated in vitro by a docking site peptide derived from the BOX-V region in the DNA-binding domain of p53. This docking site motif overlaps with the more recently defined MDM2binding site in the BOX-V domain of p53 that plays an additional role as a ubiquitination signal (47, 54). Furthermore, using an in vivo Ser20 kinase assay and CHK1 as a prototypical member of the calcium calmodulin kinase superfamily, we show how deletion of the region containing the BOX-V domain in the DNA-binding domain of p53 prevents Ser20 phosphorylation in vivo and that an enhanced green fluorescent protein (EGFP)–BOX-V fusion peptide can bind to and attenuate CHK1-mediated Ser20 phosphorylation in vivo. These data demonstrate a role for the conserved BOX-V domain docking site in mediating not only p53 ubiquitination, but also p53 phosphorylation. Furthermore, the data provide a framework from which to evaluate the role of a subset of calcium calmodulin kinase superfamily members as stress-induced p53 activators in vivo. MATERIALS AND METHODS Chemicals, reagents, and protein purifications. Chemicals were acquired from Sigma-Aldrich unless indicated otherwise. Overlapping synthetic 15-mer biotinylated peptides from the human p53 coding sequence were described previously (10) (as in Fig. 3I) and were obtained from Chiron Mimitopes. Peptides were resuspended in dimethyl sulfoxide to a concentration of 5 mg/ml, and peptides were added to kinase reaction mixtures to a final concentration of 50 M. Peptide stimulation of kinase function or peptide inhibition towards p53 tetramers was monitored as described previously for CHK1 and CHK2 (10). Glutathione S-transferase (GST)- or His-tagged CHK2, CHK1, DAPK-1, and DRAK-1 were purified using glutathione or nickel affinity chromatography. p53 was purified as described previously (25). DNA-activated protein kinase (DNAPK) was purchased from Promega. AMPK was acquired from Upstate Biotechnologies. Full-length and core domain human CHK1, CHK2, DRAK-1, DAPK-3, and DAPK-1 homologues were cloned from total RNA (A375 human cell line) into the Invitrogen Gateway system, and Escherichia coli expression vectors were generated by recombining the respective alleles into the His-tagged or GST-tagged destination vectors according to the manufacturers’ recommendations. A kinase-inactive form of DAPK-1 was generated by mutagenesis of codon 42 from Lys to Ala. Protein kinase reactions. Preparation of high-salt extracts from A375 cells for chromatography was done essentially according to Achari and Lees-Miller (1). Fractionation of kinase activity from high-salt lysates was performed using QHiTrap (Pharmacia) with a 40-column volume 0 to 0.5 M KCl gradient in buffer A (25 mM HEPES [pH 8.0]), 10% glycerol, 1 mM benzamidine, 1 mM dithiothreitol, 0.02% Triton X-100). Equal units of native or recombinant kinase were added to phosphorylation reaction mixtures shown in Fig. 1, as determined by
3543
the SAMS peptide kinase assay (as in reference 14; data not shown). The full-length p53 and GST-p53N1-66 kinase assay was described previously (10). Briefly, recombinant or native enzyme (⬃50 ng) was incubated with p53 substrate (150 ng) and with a 50 M final concentration of peptides unless indicated otherwise. Reaction buffer contained 50 mM HEPES (pH 7.5), 10 mM MgCl2, 1 mM ATP, 1 mM dithiothreitol, and 0.8 mM EDTA with 0.1 mM -glycerol phosphate in a final volume of 20 l. Reaction mixtures were incubated at 30°C for 30 min unless indicated otherwise. Reactions were terminated by addition of Laemmli sample buffer and resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. When radioactive kinase assays were set up, [␥-32P]ATP incorporation into reaction products was visualized by autoradiography after electrophoresis. Cell biology and immunochemical assays. H1299 cells were cultured in RPMI 1640 medium (Invitrogen) supplemented with 5% fetal bovine serum, and other cell types were grown in Dulbecco’s modified Eagle’s medium with 10% fetal calf serum at 37°C with 5% CO2. Cells were harvested 24 h after transfection by being scraped into phosphate-buffered saline, and cells were collected by low-speed centrifugation at 4°C. Cell pellets were lysed with 3 volumes of urea lysis buffer (6 M urea, 20 mM HEPES [pH 8.0], 25 mM NaCl, 100 mM dithiothreitol, 0.5% Triton X) for 30 min on ice and clarified by cold centrifugation at 13,000 rpm in a bench-top centrifuge. When active enzyme was required, the lysis buffer replaced urea with 1%Triton X-100 and 0.15 M KCl unless indicated otherwise. The protein concentration of the lysates was determined by Bradford analysis, and samples were prepared using Laemmli sample buffer to a final concentration of 1 mg/ml. Transfection into the indicated cells was carried out using Lipofectamine from Invitrogen, as reported previously (15, 54). A mixed pool of small interfering RNA (siRNA) oligonucleotides towards CHK1, CHK2, DAPK-1, DAPK-2, and DAPK-3 was obtained from Dharmacon. Transfection of p53 and p53 reporter vectors; EGFP–BOX-V, EGFP–BOX-I, and control EGFP-peptide fusions; and CHK1, CHK2, and DAPK genes into A375, SAOS-2, H1299, and HCT116 was performed as described previously (17, 54). The p53⌬ gene cloning was described previously (44). MDM2-mediated ubiquitin of p53 in vivo was described previously (54). Protein G beads were obtained from Pharmacia. DAPK-1, CHK1, CHK2, DAPK-3, and GST antibodies were obtained from Sigma, Santa Cruz, and BD Biosciences. GFP antibodies were obtained from Clontech. Monoclonal antibodies to phospho-BOX-I epitope peptides including FPS20, FPT18, and FPS15 (as used in Fig. 1 and 2) were developed by standard methods as described previously (12, 13). Immunoblotting for Ser20 or Thr18 site phosphorylation of p53 after kinase incubation at 30°C for 30 min was performed by immunoblotting with anti-phospho-site immunoglobulin G (IgG) at a concentration of 1 g/ml essentially as described previously (10).
RESULTS Calcium calmodulin kinase superfamily members are stimulated in vitro by a BOX-V domain peptide derived from p53. In order to begin to define the class of enzyme that targets the Ser20 site of p53 in vivo, we initially examined the homology between the p53 transactivation domain (i.e., BOX-I domain [Fig. 1A and B]) and consensus phosphorylation sites of the key stress-activated protein kinases with homology to CHK2 (Fig. 1D). We reasoned that enzymes with homology to CHK2 (i.e., members of the calcium calmodulin kinase superfamily) might have evolved a similar specificity for p53 at Ser20. Members of the calcium calmodulin kinase superfamily that have a previously defined consensus peptide phosphorylation site (Fig. 1D) were examined and include CHK1, CHK2, calcium calmodulin II and IV, AMP-activated protein kinase, and DAPK-1. Of these consensus sites, the DAPK-1 phosphorylation site peptide consensus has the most obvious homology to the BOX-I domain of p53: the amino acids of homology form the consensus QXX(Y/F)SD(L/V)(W/F) (53). An independent approach using the DAPK-1 homologue DAPK-3 as a scaffold for selecting peptides from a combinatorial library also surprisingly yielded peptides with a striking degree of homology to the p53 activation domain forming a consensus sequence of SD(L/W)WXXP (Fig. 1E) (7).
3544
CRAIG ET AL.
MOL. CELL. BIOL.
FIG. 1. The BOX-V domain peptide stimulates p53 transactivation domain phosphorylation in vitro by calcium calmodulin kinase superfamily members. (A) Summary of the domain structure of p53 with conserved BOX domains I through V, the proline repeat domain (PRO) that binds p300, the tetramerization domain (TET), and the C-terminal sites of ubiquitination and acetylation (UB/AC) highlighted. (B) Summary of phospho-acceptor sites in the BOX-I transactivation domain (fish versus human p53 homology alignment) with sites modified by ATM (Ser15), CHK-1/2 (Thr18 and Ser20), and DAPK-1 (Ser20) highlighted. (C) Specificity of phospho-specific monoclonal antibodies (MAB) for the BOX-I transactivation domain. The indicated biotinylated peptides (non-phospho; P-Ser15, P-Thr18, or P-Ser20) were bound to streptavidin and incubated with the DO-1, P-Thr18, or P-Ser20 monoclonal antibodies, and monoclonal antibody binding was quantified using a tetramethylbenzidine-based spectrophotometric assay at 450 nm, as described previously (13). (D) Homology of the p53 transactivation domain to consensus phosphorylation peptides from members of the calcium calmodulin kinase superfamily. Shown is an alignment of the consensus phosphorylation site peptides for CHK1 (26), CHK2 (40), AMPK (14), calcium calmodulin kinases (55), and DAPK-1 (53) relative to the p53 phosphorylation sites in the BOX-I transactivation domain. Sites of homology with p53 are highlighted in gray. (E) Homology of the p53 transactivation domain to consensus peptides acquired from the peptide library display of DAPK-3. An alignment of the phage-peptides (-1 and -2) isolated from a DAPK-3 screen (7), relative to the p53 phosphorylation sites in the BOX-I transactivation domain is indicated in gray. (F and G) p53 transactivation domain kinase function by calcium calmodulin kinases is stimulated by the BOX-V peptide from the DNA-binding domain of p53. The p53 transactivation domain fragment (GST-p53N1-66 was assembled in kinase reactions with the indicated kinase (DAPK-1, CHK2, CHK1, DRAK-1, or AMPK), with crude lysate (as indicated in lanes 13 and 14), and without or with the BOX-V docking site peptide, as indicated. Reaction products were immunoblotted and analyzed for (F) Ser20 or (G) Thr18 phosphorylation of p53. In the case of AMPK, AMP was also included (lane 11 versus 10) to evaluate AMP dependence in the assay as a reflection of AMPK integrity.
These biochemical properties prompted us to evaluate whether key members of this kinase superfamily can function as p53-Ser20 kinases. The kinases evaluated include CHK2, CHK1, DAPK-1, DAPK-3, the DAPK-1/3 homologue DRAK-1, and AMPK. The kinase assay utilized phospho-specific monoclonal antibodies with validated specificity for phospho-sites in the p53 activation domain (Fig. 1C). When fragments of p53 harboring the activation domain alone were used as a kinase substrate, most of these enzymes exhibited little detectable activity towards p53 (data not shown). One exception was DAPK-1, which had detectable basal activity to the p53 transactivation domain fragment (as in Fig. 3G, lane 5 versus 1), which can be explained by the high degree of
hydrophobic amino acid homology between p53 and the DAPK-1 consensus site peptides (Fig. 1D). The second exception was AMPK (Fig. 1F, lane 10 versus 1), which exhibits AMP-stimulated phosphorylation of p53 at Ser20 (Fig. 1F, lane 11 versus 10). The other enzymes do not have an intrinsic capacity to modify the minimal transactivation domain of p53 (Fig. 1F and G). Although the information required for a protein kinase to phosphorylate a target is generally thought to occur within the primary phosphorylation site, growing evidence points to a second determinant of specificity involving enzyme docking (5). The role of small linear interaction motifs in driving protein-protein interactions may be more widespread than real-
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
3545
FIG. 2. Chromatographic fractionation of Ser20 kinases from human cell lines. (A) Cell-free Ser20 kinase assay development. Full-length native p53 protein was incubated in kinase buffer either alone (lane 1), with lysate from proliferating A375 cells (lane 2), or with lysate from cells treated with 2 J/m2 (lane 3), 5 J/m2 (lane 4), 10 J/m2 UV-C (lane 5), or 20 J/m2 UV-C (lane 6). Total p53 protein in each assay was monitored by immunoblotting (top panel, “P53 levels”), and Ser20 site-specific phosphorylation of p53 was detected following immunoblotting reaction products with the Ser20-phospho-specific monoclonal antibody, P-S20 (bottom panel, “Ser20-P”). (B) p53 activity from A375 cells exposed to progressively higher levels of UV irradiation. A375 cells were transfected with the p21-luc reporter and pCMV--gal control and exposed to increasing amounts of UV-C irradiation followed by measurement of p53-dependent reporter activity (expressed in relative light units [RLU] for luciferase/galactosidase activity). (C to F) Chromatographic fractionation of A375 lysates from undamaged cells. Cell lysates from undamaged A375 cells were fractionated using a linear gradient on a Q-Trap ion-exchange resin, and Ser20 kinase activity (C) was measured as a function of elution of DAPK-1 (D), CHK2 (E), and CHK1 (F). (G to I) Chromatographic fractionation of A375 lysates from irradiated cells. Cell lysates from A375 cells irradiated with 20 J/m2 were fractionated using a linear gradient on a Q-Trap ion-exchange resin, and Ser20 kinase activity (G) was measured as a function of elution of CHK2 (H) and CHK1 (I). (J to L) DAPK-1 protein levels in distinct tumor cell lines. Lysates from the indicated cells (A549, A375, and HCT116 [wild-type p53, p53 null, and p21 null, respectively]) were immunoblotted with antibodies to (J) DAPK-1, (K) p53, and (L) DAPK-3. (M to P) Chromatographic fractionation of A549 lysates from undamaged cells. Cell lysates from undamaged A549 cells were fractionated using a linear gradient on a Q-Trap ion-exchange resin, and BOX-V-stimulated Ser20 kinase activity (P) was measured as a function of elution of DAPK-1 (M), CHK2 (N), and DAPK-3 (O). (Q to R). The DAPK-1 antibody attenuates Ser20 kinase activity. The peak of fractionated DAPK-1 from A549 cells (1 l of fraction 26 [Fract 26]) was incubated in kinase buffer containing the p53 tetramer without (lane 1) or with DAPK-1 monoclonal antibody (MAb) and incubated at 30°C for 30 min. The extent of (R) Ser20 site phosphorylation was measured using the immunochemical blot assay and normalized to total p53 protein in the fraction (Q).
3546
CRAIG ET AL.
ized (38). In the case of p53, the protein kinases CHK2 and CHK1 have been shown in vitro to modify p53 at Ser20 by a docking-dependent interaction with the BOX-V domain of p53 (10). When DAPK-1, CHK2, CHK1, DRAK, and AMPK were assayed for p53 activation domain Ser20 kinase activity in the presence of the peptide fragment from the BOX-V domain of p53, significant stimulation of Ser20 kinase activity was detected (Fig. 1F, lanes 3, 5, 7, 9, and 12 versus lanes 2, 4, 6, 8, and 10, respectively). Furthermore, in a crude cell-free system, the same docking site peptide induced Ser20 site phosphorylation, highlighting an enzyme(s) in cell lysates with similar docking-dependent Ser20 kinase activity (Fig. 1F, lane 14 versus 13). We also examined DAPK-1, CHK2, CHK1, DRAK, and AMPK for p53 transactivation domain Thr18 kinase activity in the presence of the docking site peptide fragment from the BOX-V domain of p53. Of these enzymes, CHK1, DRAK-1, and AMPK exhibited the most pronounced Thr18 kinase activity in a docking-dependent manner (Fig. 1G, lanes 7, 9, and 12 versus lanes 6, 8, and 10, respectively). DAPK-1 did not exhibit same dual-site kinase activity (Fig. 1G, lane 3) and maintained an absolute specificity in targeting the Ser20 site of p53 (see Fig. 3). Ser20 kinase activity from cell lysates coelutes with CHK1, CHK2, and/or DAPK-1. Together, these data affirm that members of the calcium calmodulin kinase superfamily have evolved a docking-dependent capacity to target the Ser20 and/or Thr18 site of p53, but whether these enzymes mediate this phosphorylation in cells is not clear. In order to first identify the major Ser20 kinases from cells, we had to develop a highly specific cell-free immunochemical assay that uses tetrameric p53 as a substrate. The identification of a cell line that produces high levels of Ser20-modified p53 in the absence of DNA damage would increase the likelihood of yielding a source of activity distinct from CHK2. Undamaged cycling A375 cells containing wild-type p53 produce an active pool of p53 that is constitutively modified at Ser20 (6) (11). Kinase activity in a cell-free system was assayed with native p53 tetramers as substrates using three phospho-specific monoclonal antibodies that bind to the phospho-Ser15, phospho-Thr18, or phospho-Ser20 p53 activation domain, respectively. Significant levels of Ser20 kinase activity can be detected in lysates from undamaged cells (Fig. 2A, lane 2 versus 1), consistent with the existence of an enzyme that can modify the Ser20 site constitutively in cycling A375 cells. When p53 activity was measured using a reporter assay and compared to cell-free Ser20 kinase activity, a correlation was observed between p53 activity and Ser20 kinase function. Low levels of UV irradiation (1 to 2 J/m2) attenuated basal p53 activity (Fig. 2B) that correlated with loss of Ser20 kinase function (Fig. 2A, lane 3 versus 2). Progressively elevated levels of UV radiation that gave rise to increasing p53 activity also gave rise to progressively higher levels of cell-free Ser20 kinase activity (Fig. 2B versus A). This cell system was used to determine whether the major Ser20 kinase was coincident with or distinct from CHK1, CHK2, or DAPK-1. AMPK and DRAK-1 were not evaluated since there was no detectable Thr18 site kinase activity detectable in A375 cell lysates (Fig. 1G). Lysates from damaged and unirradiated cells were fractionated to determine whether the Ser20 kinase activity copurified with CHK1/2 or DAPK-1/3. The application of lysates from cycling
MOL. CELL. BIOL.
A375 cells to an ion-exchange column results in the very broad elution of Ser20 kinase activity (Fig. 2C, fractions 12 to 20 and 21 to 26 [data not shown]). The bulk of the broadly eluting Ser20 kinase activity is separate from CHK1 (Fig. 2F, two peaks over fractions 5 to 13) and CHK2 (Fig. 2E, fractions 14 to 15). However, high levels of UV damage that results in progressively elevated levels of p53 activity (Fig. 2B) results in increasing levels of cell-free Ser20 kinase activity (Fig. 2A, lanes 4 to 6). This UV-inducible Ser20 kinase activity (Fig. 2G, lanes 9 to 11) copurifies with CHK2 (Fig. 2H, lanes 9 to 11), not CHK1, and is consistent with CHK2 being a major DNA damageinducible p53-Ser20 kinase (46). Given the ability of recombinant DAPK-1 to function as Ser20 kinase, we next wanted to determine whether DAPK-1 protein coeluted with the Ser20 kinase activity from A375 cells. The total Ser20 kinase activity eluting from fractions 13 to 19 (Fig. 2C) is plotted as a function of the elution of DAPK-1 (Fig. 2D, fractions 15 to 19). DAPK-1 eluted as a broad peak distinct from CHK1 and CHK2, and the broad elution of DAPK-1 correlated with the broad elution of the Ser20 kinase activity, thus indicating that one possible identity of Ser20 kinase was DAPK-1. It has not been possible to purify the Ser20 kinase for identification from monolayer cycling A375 cells, as its abundance in hindsight was too low (data not shown; see Fig. 2J). Additional cell lines were screened for those with higher levels of endogenous DAPK-1 protein to determine whether DAPK-1 elution correlated with cell-free Ser20 kinase activity. A set of cells with known targeting of the Ser20 site on p53 were evaluated, including A375, A549, and HCT116 cells (Fig. 2J). Of these, A549 cells exhibited the highest levels of DAPK-1 protein, and lysates were fractionated and analyzed for BOX-V-stimulated kinase activity. The majority of DAPK-1 protein (Fig. 2M) correlated with BOX-V stimulated kinase activity (Fig. 2P; broad elution from fractions 24 to 32) and was distinct from CHK2 (Fig. 2N, fraction 22) and DAPK-3 (Fig. 2O, fraction 30). Furthermore, the peak fraction of the Ser20 kinase activity (fraction 26) could be neutralized as a Ser20 kinase using an anti-DAPK-1 IgG, thus indicating there was a DAPK-1 component to the Ser20 kinase (Fig. 2R, lane 2 versus 1). Recombinant DAPK-1 was characterized further in relation to the established Ser20 kinases, CHK1 and -2. The DAPK-1 purified from E. coli had an intrinsic in vitro Ser20 kinase activity (Fig. 3A, lane 4 versus 1). DAPK-1 has no intrinsic Thr18 or Ser15 kinase activity (Fig. 3B and C, lane 4), making it distinct from CHK2 or DRAK-1, which are dual-site kinases modifying both Ser20 and Thr18 (Fig. 3D and E). A kinasedead form of DAPK-1 is not active as a Ser20 kinase (Fig. 3D, lane 3 versus lane 2). DAPK-3 was also examined in p53 kinase assays, and it also modifies only Ser20 and not Thr18 (Fig. 3A, lane 5), consistent with the -peptide consensus defined for DAPK-3, where the Ser20 amino acid equivalent plays a role in kinase recognition (Fig. 1E). As positive controls for the immunochemical kinase assay, DNA-PK has no detectable Ser20 kinase activity (Fig. 3A, lane 3 versus lane 1), but has relatively high levels of intrinsic Ser15- and Thr18-associated kinase activity (Fig. 3B and C, lane 3). Similar to CHK2, the Ser20 kinase activity associated with DAPK-1 was not inhibited when the BOX-I phosphorylation site peptide was included in a Ser20 kinase assay containing tetrameric p53 as a substrate (Fig. 3E, lane 3 versus lane 2).
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
3547
FIG. 3. The BOX-V domain peptide modifies Ser20 kinase function in vitro. (A to D) DAPK-1/3 modifies only Ser20. The indicated protein kinases were assembled in kinase assays with p53 tetramers as substrates: no kinase (lane 1), Ser20 kinase from A375 cells (lane 2), DNA-activated protein kinase (lane 3), recombinant DAPK-1 (lane 4), and recombinant DAPK-3 (lane 5). After incubation, the reaction products were analyzed for phosphorylation at (A) Ser20, (B) Thr18, or (C) Ser15 using the appropriate phospho-specific monoclonal antibodies (MAb) to each site. (D) Recombinant DAPK-1 and the DAPK-1 kinase-dead form (DAPkd) were purified and incubated in reaction mixtures containing p53 tetramers as substrates: no kinase (lane 1), DAPK-1 (lane 2), and the kinase-dead form of DAPK-1 (lane 3). Reaction products were immunoblotted to quantify phosphorylation at Ser20. (E and F) Ser20 kinase (E) and recombinant DAPK-1 (F) were analyzed for a CHK1/2-like docking site motif. (E) Ser20 kinase from cells (DAPK-1 was assembled in kinase assays with p53 tetramers: no kinase (No Kin; lane 1), Ser20 kinase (lane 2), and Ser20 kinase and peptides 1 to 20 (lanes 3 to 24, as in panel I). The key inhibition reactions include Ser20 kinase activity in the presence of the BOX-I peptide (lane 3), BOX-II peptide (lane 7), and the BOX-V peptide (lane 22). (F) Recombinant DAPK-1 was assembled in kinase assays with p53 tetramers: no kinase (lane 1); Ser20 kinase (lane 2); and Ser20 kinase and BOX-I peptides 4, 5, 6, 17, 18, and 19 (lanes 3 to 9). The key inhibition reactions include Ser20 kinase activity in the presence of the BOX-I peptide (lane 3), BOX-II peptide (lane 5), and the BOX-V peptide (lane 8). (G) The BOX-II and BOX-V domain peptides stimulate CHK2 and DAPK-1 in trans. Kinase reaction mixtures containing [␥-32P]ATP, p53N1-66, and CHK2 (lanes 1 to 4) or DAPK-1 (lanes 5 to 8) were assembled without peptide (lanes 1 and 5) or with the indicated peptides (lanes 2 to 4 and 6 to 8, respectively). Reaction products were analyzed for p53N1-66 phosphorylation after electrophoresis by autoradiography. (H) The p53-DNA complex (dark gray) is shown with the BOX-II, BOX-IV/V linker, and BOX-V domain peptide regions (dark gray) residing in the same plane forming the basis for an anchoring site for DAPK-1. (I) The relative locations of the peptides from p53 analyzed for activity in competitive kinase assays are as summarized. The BOX-I peptide contains amino acids 12 to 27, the BOX-II peptide (peptide 5) contains amino acids 117 to 131, and the BOX-V peptide (peptide 18) contains amino acids 270 to 284 of p53.
The same BOX-I peptide can bind MDM2 (16, 54), indicating that the peptide is bioactive. The inability of the p53 transactivation domain peptide (BOX-I) to inhibit DAPK-1 suggests that DAPK-1 might have higher-affinity docking sites that mediate p53 tetramer phosphorylation. In fact, overlapping peptides from the entire coding region of p53 (Fig. 3I) were used to show that two major regions within the core domain of p53 have apparent contact sites for DAPK-1: BOX-II- and BOX-
V-derived peptides (peptides 5 and 18) attenuated phosphorylation of p53 (Fig. 3E, lane 7 versus 2 and lane 22 versus 14; and Fig. 3F, peptides 5 and 18). The highly conserved BOX-II, BOX-IV/V linker, and BOX-V domain peptides reside in the same plane that may form the basis for an anchoring site for DAPK-1 (Fig. 3H), as is the case for CHK1 and CHK2 (10). The addition in trans of the BOX-II, BOX-IV/V linker, and BOX-V domain peptides to CHK2 allos-
3548
CRAIG ET AL.
MOL. CELL. BIOL.
FIG. 4. In vivo analysis of DAPK-1–p53 interactions. (A) Immunoprecipitation (IP) of p53 with DAPK-1 in A375 cells. The indicated genes (1 g of vector, DAPK-1, and p53) were transfected into A375 cells (containing both p53 and low levels of DAPK-1, as in Fig. 2J and K) for 24 h, cells were lysed and subject to immunoprecipitation with anti-DAPK antibodies, and the amounts of p53 protein in the DAPK-1 immunoprecipitate (bound) and unbound fractions were analyzed after immunoblotting with anti-p53 antibodies. (B) Immunoprecipitation of GST–DAPK-1 with p53 in A375 cells. The indicated genes (1 g of vector or GST–DAPK-1) were transfected into A375 cells for 24 h, cells were lysed and subject to immunoprecipitation with anti-p53 antibodies, and the amounts of GST–DAPK-1 protein in the p53 immunoprecipitate (bound) and unbound fractions were analyzed after immunoblotting with anti-DAPK antibodies. (C to G) DAPK-1 stabilizes p21WAF1 protein in HCT116 p53⫹/⫹ cells. HCT116 p53 wild-type cells (lanes 4 to 6) or HCT116-p53-null cells (lanes 1 to 3) were transfected with the indicated genes: lanes 1 and 4 for pCMV-DAPK-1, lanes 2 and 5 for pCMV-DAPK-3, and lanes 3 and 6 for pCMV-vector control. Twenty-four hours after transfection, the cells were lysed and blotted for p21 (C), p53 (D), DAPK-1 (E), DAPK-3 (F), and the actin control for protein loading (G). (H to O) siRNA to DAPK-1 depletes p53 protein steady-state levels in A375 cells. A375 cells, which contain the low levels of DAPK-1 originally identified as a Ser-20 kinase (Fig. 2J), were treated with siRNA to various DAPK homologues (control, DAPK-1, DAPK-2, and DAPK-3) for the indicated times (0, 8, or 16 h; lanes 1 to 3), and the steady-state levels of either (H to K) p53 protein or (L to O) DAPK-1 protein were analyzed by immunoblotting and quantified in panel P.
terically activates the enzyme towards a small transactivation domain fragment of p53 (p53N1-66; Fig. 3G, lanes 2 to 4 versus 1). DAPK-1 has a significantly higher basal kinase activity towards p53 than CHK2 (Fig. 3G, lane 5 versus 1), which is consistent with the fact that DAPK-1 consensus site peptides have more homology to p53 than the CHK2 consensus peptides (7, 53) (Fig. 1D). The inclusion of BOX-II, BOX-IV/V linker, and BOX-V peptides stimulated p53N1-66 substrate phosphorylation by DAPK-1 (Fig. 3G, lanes 6 to 8 versus 5). These biochemical characterizations have established that recombinant DAPK-1 is a bona fide Ser20 kinase and that its docking dependence towards p53 is similar to that of the established Ser20 kinase, CHK2. However, DAPK-1 is quite distinct in that it cannot function as a dual-site kinase like CHK1 or CHK2. Evaluation of DAPK-1, CHK1, and CHK2 as p53-activating kinases in vivo. CHK1 and CHK2 enzymes have been vali-
dated previously in cells as p53-interacting kinases (46), and we next evaluated whether DAPK-1 similarly forms stable complexes with p53 in cells after transfection. The immunoprecipitation of DAPK-1 after transfection with the indicated DAPK-1 or p53 expression vectors shows that a relatively small proportion of transfected p53 protein is bound to endogenous DAPK-1 (Fig. 4A, lane 5 versus 1). In contrast, the immunoprecipitation of endogenous p53 after transfection with the DAPK-1 gene shows that a relatively large proportion of transfected DAPK-1 is bound to endogenous p53 (Fig. 4B, lane 3 versus 4). These data suggest that a relatively large pool of ectopically expressed DAPK-1 can interact stably with p53 in cells, while the majority of endogenous DAPK-1 in cells cannot be easily dissociated from its anchor to bind transfected p53. In order to examine whether this binding of DAPK-1 to p53 correlated with changes in p53 activity, endogenous p21 protein changes were evaluated after transfection with the indi-
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
cated DAPK-1 and DAPK-3 genes into the HCT116 p53 wildtype and p53-null cell lines. DAPK-1 transfection (Fig. 4E, lanes 1 and 4) can elevate the basal levels of p21 protein in p53-containing HCT116 cells, relative to the isogenic p53-negative HCT116 cells (Fig. 4C, lane 4 versus 1). In contrast, DAPK-3 transfection (Fig. 4F, lanes 5 and 2) is unable to induce a similar p53-dependent induction of p21 (Fig. 4C, lane 5 versus 2). This is consistent with the p21-luc reporter template response (see Fig. 5K below) and indicates that p53dependent induction of p21 is DAPK-1 specific. We further evaluated, using siRNA, whether DAPK-1 interacts genetically with p53, as described previously with murine knockout cells (42). The treatment of A375 cells with siRNA to DAPK-1 depletes p53 protein within 8 h (Fig. 4I, lanes 2 and 3 versus 1, and Fig. 7D, lane 4 versus 1), suggesting that DAPK-1 can stabilize steady-state levels of p53 protein in this cell line. As controls, siRNA to DAPK-1 reduces DAPK-1 protein levels (Fig. 4M, lanes 2 and 3 versus 1, and as quantified in panel P). The more pronounced depletion of p53 protein relative to DAPK-1 may be due to the fact the enzymatic activity of DAPK-1 is required for p53 protein stabilization and that a small reduction in DAPK-1 enzyme activity by siRNA can relatively quickly destabilize p53 protein. The use of siRNA in A375 cells towards DAPK-2 or DAPK-3 did not deplete p53 protein (Fig. 4J and K, lanes 1 to 3), nor did depletion of CHK1 or CHK2 using siRNA alter p53 protein levels in A375 cells (Fig. 7D, lanes 1 to 3). A375 cells therefore appear to utilize DAPK-1 as the predominant kinase for maintaining high levels of endogenous p53 protein. We continued DAPK-1, CHK1, and CHK2 characterization in order to determine which of these three enzymes can be the most potent Ser20 kinase towards p53 in cells. In order to determine whether CHK2, CHK1, or DAPK-1 can mediate in vivo phosphorylation of p53 at Ser20, a cell-based transfection assay was set up to evaluate Ser20 phosphorylation of p53 protein. The use of the ectopically expressed p53 transfection system is required, since the Ser20 phospho-specific antibodies we use do not have a high enough affinity to detect unambiguously endogenous p53 phosphorylation at Ser20 (12). The transfection of increasing amounts of the p53 expression vector into H1299 cells (null for p53) gave rise to elevated p53 protein production (Fig. 5A) and Ser20 site phosphorylation (Fig. 5B) in a dose-dependent manner. These data indicate that H1299 cells have an endogenous Ser20 kinase capable of modifying a fraction of the ectopically expressed p53 protein. Using a level of p53 gene transfection that gave dose-dependent Ser20 site phosphorylation in vivo, we cotransfected CHK2, CHK1, fulllength DAPK-1, and DAPK-1 core domain (Fig. 5C to G) to determine whether transcription stimulation correlated with increased Ser20 site phosphorylation. Only full-length DAPK-1 transfection (Fig. 5D, lane 8 versus 2) and CHK1 transfection (Fig. 5D, lane 12 versus 2) gave rise to elevated Ser20 site phosphorylation of p53. CHK1 transfection (from 1 to 5 g of DNA) also gave rise to a dose-dependent increase in p53-dependent transcription (Fig. 5J). Similar to CHK1, transfection of DAPK-1 gave rise to elevated p53-dependent transcription (Fig. 5K), while the DAPK-1 core or full-length DAPK-1 kinase-dead expression vectors attenuated the basal p53-dependent transcription (Fig. 5L). In contrast, CHK2 transfection actually attenuated the
3549
basal Ser20 site phosphorylation (Fig. 5D, lane 6 versus 2), which is consistent with the dose-dependent attenuation of p53-dependent transcription by CHK2 (Fig. 5J). Furthermore, compared to DAPK-1 transfection, which was able to elevate basal Ser20 site phosphorylation (Fig. 5I, lanes 3 to 5 versus 1), DAPK-3 transfection also attenuated basal Ser20 site phosphorylation (Fig. 5I, lanes 6 to 8 versus 1). The latter data are also consistent with the ability of DAPK-3 to attenuate p53-dependent transcription function (Fig. 5K and 4C). Thus, by two distinct transfection-based assays (p53-dependent transcription reporter activity and Ser20 site phosphorylation), only CHK1 and DAPK-1 are able to positively affect the p53 pathway, while DAPK-3 and CHK2 attenuate p53 function. Mutation of the Ser20 site to Ala20can attenuate p53 activity in apoptotic and transgene-based systems (36, 52), suggesting a role for phosphorylation in p53 regulation. However, in a transfectionbased system, the mutation of the Ser20 residue to Asp20 only marginally elevated basal p53-dependent transcription (Fig. 5M). The kinase-dead form of DAPK-1 is unable to stimulate p53 (Fig. 5O), suggesting that the kinase activity is required to stimulate p53 function. Nevertheless, the Asp20 mutant p53 is further stimulated more strikingly by transfected DAPK-1 (Fig. 5N), suggesting that the transfection system reflects a greater contribution of kinase docking to p53 over kinase phosphorylation of p53. In fact, the Ala20 mutant version of p53 is stimulated as well as the wild-type p53 or Asp20 mutant p53 from the bax promoter (Fig. 5O), highlighting the importance of kinase docking compared to phosphorylation in this transfection assay. As such, we next evaluated whether deletion of the BOX-V region on p53 prevented kinase stimulation of p53 activity and associated Ser20 phosphorylation in cells. Calcium calmodulin kinase interaction with p53 requires the BOX-V domain of p53. A recent report (44) has highlighted the existence of a naturally occurring spliced form of p53 that has removed approximately 55 internal amino acids harboring the BOX-V docking region of p53 (Fig. 6A). The BOX-V docking site of p53 that is required for the calcium calmodulin kinase superfamily to modify the Ser20 site of p53 (Fig. 1) also contains the “ubiquitination signal” that is required for dualsite stimulation of the E3 ubiquitin ligase function of MDM2 (47, 54). The removal of this region within p53⌬ in fact impairs MDM2-mediated ubiquitination of the transfected p53⌬ isoform (Fig. 6B, lane 4 versus 3). However, the p53⌬ protein is able to drive p53-dependent activity from the p21 reporter template (Fig. 6C), indicating that this BOX-V domain deletion does not simply unfold and inactivate p53. That this p53 spliced variant is active as a transcription factor is relatively surprising given that so many tumor-derived point mutations within this region of p53 can inactivate p53 function. Similar to the inability of p53⌬ to be ubiquitinated by MDM2, p53⌬ basal activity is also not stimulated by CHK1 relative to wild-type full-length p53 (Fig. 6C). Mutant p53 encoded by the HIS175 allele induced no activity from the p21 reporter, as expected (Fig. 6C). However, the p53 HIS175 protein is reactivated after CHK1 transfection (Fig. 6C). Under these conditions, the p53 HIS175 protein is hyperubiquitinated in vivo due to exposure of the BOX-V ubiquitination signal (48). The ability of CHK1 to drive the p53 HIS175 protein into an active state suggests that CHK1 and MDM2 can compete for binding to the BOX-V
3550
CRAIG ET AL.
MOL. CELL. BIOL.
FIG. 5. In vivo induction of Ser20 phosphorylation of p53 by the calcium calmodulin kinase superfamily members CHK1 and DAPK-1. (A and B) Analysis of p53 phosphorylation by an endogenous Ser20 site kinase in p53-null H1299 cells. Cells were transfected with increasing amounts of p53 vector DNA (from 30 ng to 1 g), and lysates were immunoblotted for (A) p53 and (B) Ser20 site phosphorylation of p53. (C to G) Effects of kinase gene transfection on Ser20 site phosphorylation of p53. H1299 cells were transfected with the p53 expression vector (100 ng) and the indicated kinase expression vector (CHK2, DAPK-1 full length, DAPK-1 core kinase domain, and CHK1), and lysates were immunoblotted for (C) p53, (D) Ser20 site phosphorylation of p53, (E) CHK2, (F) DAPK-1, and (G) CHK1. (H and I) Effects of DAPK-1/3 kinase gene transfection on Ser20 site phosphorylation of p53. H1299 cells were transfected with the p53 expression vector (100 ng) and the indicated kinase expression vector (DAPK-1 full length or ZIPK/DAPK-3), and lysates were immunoblotted for (H) p53 and (I) Ser20 site phosphorylation of p53. (J to L) Effects of calcium calmodulin superfamily gene transfection on p53-dependent transcription reporter activity. (J). The indicated kinase-encoding expression vectors (1 to 5 g) were transfected into p53-null cells along with vector controls or p53, and with p21-luc and --gal reporters, as indicated. After 24 h, cell lysates were analyzed for luciferase and -galactosidase activity as reported previously (16). The analyses include (J) CHK1 versus CHK2, (K) DAPK-1 versus DAPK-3, and (L) DAPK-1 mutant series on p21 reporter vectors. In panel J, “⫺” and “⫹” indicate the omission and inclusion of p53 expression vector and the numbers 1.0 and 2.0 indicate the amount of DNA added in micrograms. (M to O) Effects of p53 phospho-acceptor site mutation on p53-dependent transcription reporter activity. (M) Activity of p53S20D. The indicated p53encoding genes (ng of wild-type p53 and p53S20D) were transfected into p53-null cells along with vector controls or p53, along with p21-luc and --gal reporters, as indicated. After 24 h, cell lysates were analyzed for luciferase and -galactosidase activity. (N). Stimulation of p53S20D by DAPK-1 from a p21 reporter. The indicated p53-encoding genes (ng of wild-type p53 and p53S20D) were cotransfected into p53-null cells along with vector controls or DAPK-1, along with p21-luc and --gal reporters, as indicated. After 24 h, cell lysates were analyzed for luciferase and -galactosidase activity. (O) Stimulation of p53S20A by DAPK-1 from a bax reporter. The indicated p53-encoding genes (ng of wild-type p53, p53S20A, and p53S20D) were cotransfected into p53-null cells along with vector controls or DAPK-1, together with bax, luc, or -gal reporters, as indicated. After 24 h, cell lysates were analyzed for luciferase and -galactosidase activity as reported previously (16).
domain and drive p53 into either a ubiquitinated or transactivation-competent state. Although Ser20-site phosphorylation of full-length p53 is stimulated by CHK1 (Fig. 6F, lane 3 versus 4) or DAPK-1 (Fig.
6H, lanes 3 and 4), the transfected p53⌬ has no basal Ser20 site phosphorylation (Fig. 6F and H, lane 5 versus 3, and K and J, lane 2 versus 1). Transfected CHK1 and DAPK-1 are unable to induce Ser20 site phosphorylation on p53⌬ (Fig. 6F and H, lane
FIG. 6. Deletion of the central region of p53 containing the BOX-V domain prevents both Ser20 site phosphorylation and ubiquitination of p53. (A) General structure and conserved domains of full-length p53 and the alternatively spliced p53⌬. Highlighted is the BOX-I transactivation containing the phospho-acceptor sites and the region deleted in p53⌬ that removes the BOX-V region (44). CAL-Cam, calcium comodulin family. (B) MDM2mediated ubiquitination is attenuated using p53⌬ as a substrate. H1299 cells were cotransfected with p53 (lanes 1 and 3) or p53⌬ (lanes 2 and 4) and MDM2 (lanes 3 and 4), as indicated. After 24 h, cells were immunoblotted with an anti-p53 antibody to identify the high-molecular-mass adducts reflective of MDM2-mediated ubiquitination (bracket). (C) p53⌬ activity from the p21 reporter is not stimulated by protein kinase transfection. The CHK1-encoding gene (5 g) was cotransfected into p53-null cells along with either p53 or p53⌬ and with p21-luc and --gal reporters, as indicated. After 24 h, cell lysates were analyzed for luciferase and -galactosidase activity as reported previously (16). (D to I) p53⌬ is not modified at Ser20 after protein kinase transfection. H1299 cells were cotransfected with the p53 or p53⌬ expression vector (100 ng) and the indicated CHK1 or DAPK-1 kinase expression vectors (as indicated by “⫹” in lanes 4 and 6), and lysates were immunoblotted for (D) p53 protein, (E) evidence for deletion of the p53 central region using the DO-12 antibody (which binds in the BOX-V region), (F) Ser20 site phosphorylation of p53 without or with CHK1, (G) CHK1, (H) Ser20 site phosphorylation of p53 without or with DAPK-1, and (I) DAPK-1. (J and K) p53⌬ does not compete with wild-type p53 phosphorylation at Ser20. p53 or p53⌬ was transfected into H1299 cells, and using fixed p53, increasing amounts of p53⌬ were transfected into cells and immunoblotted for both p53 and Ser20 phospho-p53 levels. (L to N) Expression levels of transfected CHK1. (L) Hemagglutinin (HA)-tagged CHK-1 (0, 0.1, 0.2, 0.5, 1.0, and 2.0 g in lanes 1 to 6, respectively) was transfected into H1299 cells and immunoblotted for CHK1 (lower band is endogenous CHK1, and the upper band is tagged CHK1). (M) FLAG-tagged CHK1 (0, 0.1, 0.2, 0.5, 1.0, and 2.0 g in lanes 1 to 6, respectively) was transfected into H1299 cells and immunoblotted with an anti-FLAG IgG. (N) Loading control for the FLAG-CHK-1 transfection experiment in panel M. 3551
3552
CRAIG ET AL.
6 versus 4). Controls measuring transfected tagged or untagged CHK1 protein levels are depicted in Fig. 6L to N. These data indicate that the BOX-V docking region of p53 is required for basal Ser20 site phosphorylation in vivo by an endogenous Ser20 site kinase, is required for both CHK1- and DAPK-1mediated Ser20 phosphorylation in vivo, and is required for MDM2-mediated ubiquitination of p53. Thus, the BOX-V region appears to be a multiprotein docking site rivaling the N-terminal transactivation domain of p53, which also has both an MDM2 binding site, a p300 binding site, and the calcium calmodulin kinase superfamily phospho-acceptor site. Having established that CHK1 and DAPK-1 are the most potent inducers of the Ser20 site phosphorylation in H1299 cells and that this reaction is BOX-V docking dependent, we focused on determining which of the three enzymes under study—CHK2, CHK1, or DAPK-1—was the major genetic mediator of Ser20 phosphorylation of ectopically expressed p53 in H1299 cells. The use of siRNA to DAPK-1, CHK1, or CHK2 can reliably attenuate the levels of each kinase in A375 cells without affecting the levels of the others (Fig. 7A to C), and of these three manipulations, only DAPK-1 depletion reduced p53 protein levels (Fig. 7D). Using these siRNA oligonucleotides in H1299 cells, we found that only CHK1 protein depletion (Fig. 7G, lane 3 versus 1 and 2) gave rise to the most pronounced attenuation of Ser20 site phosphorylation and reduction in p53 protein levels of ectopically expressed p53 (Fig. 7E and F, lane 3 versus 1, 2, and 4). As a control, the depletion of CHK2 had no effect on the endogenous Ser20 site phosphorylation and p53 protein levels (Fig. 7E and F, lane 4 versus 1 to 3). Thus, endogenous CHK1 is required for the majority of Ser20 site phosphorylation of ectopically expressed p53 in H1299 cells. We evaluated whether a GFP–BOX-V fusion peptide could attenuate full-length p53 phosphorylation at Ser20. This would affirm that the Ser20 site kinase in cells is similar to CHK1 in that it binds to the BOX-V domain of p53. H1299 cells were transfected with expression vectors for p53 and CHK1 along with either the EGFP control (Fig. 7H to K, lanes 2, 4, 6, and 8) or EGFP–BOX-V fusion peptide (Fig. 7H to K, lanes 3, 5, 7, and 9). This EGFP–BOX-V fusion peptide contains the ubiquitination signal for MDM2 and can inhibit MDM2-mediated ubiquitination of p53 in vitro and in cells (54). Lysates were evaluated for Ser20 site phosphorylation of p53 after transfection of the EGFP constructs (Fig. 7H), and the data demonstrate that the BOX-V peptide can attenuate the endogenous Ser20 site phosphorylation of p53 (Fig. 7H, lanes 3, 5, 7, and 9 versus 2, 4, 6, and 8), which is consistent with the inability of the p53⌬ to function as a CHK1 substrate in cells. The p53⌬ protein cannot be reactivated for Ser20 site phosphorylation in vivo using the GFP–BOX-V peptide fusion (Fig. 7N, lanes 6 versus 2 and 3), under conditions where full-length p53 Ser20 site phosphorylation is attenuated (Fig. 7N, lane 3 versus 2). The attenuation of p53 phosphorylation by the EGFP–BOX-V fusion peptide correlated with the stable binding of CHK1 to the EGFP–BOX-V fusion peptide in cell lysates (Fig. 7T, lane 4 versus 7R, lane 4). Together, these data indicate that the BOX-V region in p53 binds to members of the calcium calmodulin kinase superfamily and that this binding is required to stimulate p53 function. However, only CHK-1 and DAPK-1 exhibit a pronounced ability to mediate p53 protein
MOL. CELL. BIOL.
stabilization and/or p53 activation in experimentally tractable systems. DISCUSSION Given that p53 activity is switched on by a variety of environmental stresses, definition of the enzymes that phosphorylate the p53 transactivation domain opens the door to mapping out distinct signaling transduction pathways that control p53dependent transcription and associated tumor suppression. Although mutation of the Ser20 phospho-acceptor site in the p53 transactivation domain has been reported to either attenuate p53 activity (52) or have no effect on p53 function (29), mutation of the site in murine transgenes reduces in vitro apoptotic responses and increases B-cell lymphoma (36). The latter data highlight the importance of transactivation domain integrity to p53 activity and suggest that the kinases that modify this Ser20 site will play an important role in controlling p53 function, perhaps in a tissue-specific or cell-type-specific manner. CHK2 was the original biochemically and genetically defined Ser20 kinase for p53 (46, 50). Other protein kinases presumably have evolved to target the p53 transactivation domain and link distinct stresses to p53 activation. We have used bioinformatic and biochemical fractionation approaches to establish that members of the calcium calmodulin kinase superfamily have the common feature of being able to target the p53 activation domain at Thr18 and/or Ser20 and that they all also harbor a high-affinity interaction with the BOX-V domain of p53. The enzymes have included CHK1, CHK2, DAPK-1, DAPK-3, DRAK-1, and AMPK. Using this benchmark, we fractionated cell lysates to determine what the major Ser20 kinases are in a tumor cell used to study p53 and found that CHK1, CHK2, and DAPK-1 were the major Ser20 kinases. As such, we focused on the study of these three enzymes to determine, using transfection technologies, which of these might be the most potent mediator of Ser20 site phosphorylation of p53 in cells and also to determine whether the BOX-V docking site plays a dominant role in calcium calmodulin kinase-p53 interactions. Five observations are notable: (i) CHK1 and DAPK-1 transfections into cells mediate the best stimulation of p53-dependent transcription reporter activity; (ii) CHK1 and DAPK-1 transfections into cells induce the most significant Ser20 site phosphorylation of p53; (iii) endogenous CHK1 is the most evident mediator of Ser20 site phosphorylation in H1299 cells, while endogenous DAPK-1 is the most effective p53 protein stabilizer in A375 cells, highlighting a possible cell-specific function for the Ser20 site kinases; (iv) CHK1 is attenuated as an inducer of Ser20 site phosphorylation by transfection of the EGFP–BOX-V peptide fusion protein; and (v) deletion of the BOX-V region of p53 prevents Ser20 site phosphorylation in vivo and prevents stimulation of p53 function by transfected kinase. Furthermore, CHK2 and DAPK-3 were distinct from CHK1 and DAPK-1 in that ectopic expression of the former two was unable to induce either Ser20 site phosphorylation or p53 transactivation. These data highlight the calcium calmodulin kinase superfamily as the most likely type of Ser20 site kinase, establish the BOX-V docking dependence in Ser20 site phosphorylation, and open the door to evaluate how distinct stresses might recruit different members of the calcium calmodulin kinase superfamily to mediate genetic activation of
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
3553
FIG. 7. The EGFP–BOX-V fusion peptide binds CHK1 and attenuates Ser20-site phosphorylation of p53 in vivo. (A to D) Depletion of DAPK-1 in A375 cells selectively depletes endogenous p53. Validation of siRNA specificity for CHK1, CHK2, and DAPK-1. A375 cells, which express all three calcium calmodulin kinase superfamily members CHK1, CHK2, and DAPK-1 (as in Fig. 2), were transfected with siRNA oligonucleotides for 48 h and immunoblotted for changes in (A) CHK1, (B) CHK2, (C) DAPK-1, and (D) p53 protein. (E to G) Depletion of endogenous CHK1 protein in H1299 cells using siRNA attenuates in vivo Ser20 site phosphorylation of p53. The p53-null H1299 cell line was cotransfected with the expression vector p53 and siRNA oligonucleotides to CHK1 or CHK2, as indicated. Cell lysates were immunoblotted for (E) p53 protein, (F) Ser20 site phosphorylation of p53, and (G) CHK1 protein. (H to K) Transfection of EGFP–BOX-V peptide fusion attenuates Ser20 site phosphorylation of p53 in vivo. Vectors encoding CHK1 and p53 were cotransfected into cells with EGFP control (lanes 2, 4, 6, and 8) or EGFP–BOX-V vector (lanes 3, 5, 7, and 9), and lysates were immunoblotted for (H) Ser20 site phosphorylation of p53, (I) p53 protein, (J) CHK1, and (K) EGFP levels. (L to O) the EGFP–BOX-V peptide fusion cannot activate p53⌬ phosphorylation in vivo. Vectors encoding p53 (lanes 1 to 3) and p53⌬ (lanes 4 to 6) were cotransfected into cells with EGFP control (C; lanes 2 and 5) or EGFP–BOX-V vector (V; lanes 3 and 6), and lysates were immunoblotted for (L) p53 protein, (M) evidence for deletion of the p53 central region using the DO-12 antibody (which binds in the BOX-V region), (N) Ser20 site phosphorylation of p53, and (O) EGFP protein levels. (P to T) The EGFP–BOX-V peptide fusion can form a stable immune complex with CHK1. Vectors encoding FLAG-CHK1 and EGFP variants were cotransfected into H1299 cells, lysates (lane 1) were precleared (lane 2), and FLAG-CHK1 was subjected to immunoprecipitation with FLAG-agarose beads followed by quantitation of the EGFP variants in the unbound fraction (lane 3) or EGFP in the bound fraction (lane 4). Samples after the FLAG pull-down were immunoblotted with an anti-EGFP IgG for the respective binding reactions including (P) control, (Q) vector, (R) EGFP, (S) EGFP–BOX-I, and (T) EGFP–BOX-V. (U) Multiprotein binding sites in the BOX-I and BOX-V domains of p53. The BOX-I domain was the first key protein docking site identified in p53; this linear motif has evolved interactions with p300 and MDM2 and is a phospho-acceptor site. MDM2 was the first protein found to have a defined contact site in the BOX-V domain of p53 (47, 56). This ubiquitination signal is driven by dual-site binding of MDM2 to both the BOX-I domain and the BOX-V domain to mediate C-terminal ubiquitination of p53 (54). A set of protein kinases including CHK1/2 (10) or DAPK-1 (this study) have evolved a docking-dependent interaction with the BOX-V domain of p53 that mediates N-terminal phosphorylation of p53 in the BOX-I motif. Competition for binding to the BOX-V domain or deletion of the BOX-V domain in p53 can alter E3-ubiquitin ligase or protein kinase modification of the p53 tetramer.
p53. For example, given that mutation of the Ser20-equivalent phospho-acceptor site in mice accelerates B-cell lymphoma, we are now examining in B cells which calcium calmodulin kinase superfamily plays the most dominant role in activating p53-
dependent apoptosis in response to DNA damage. Such a model provides a framework for evaluating calcium calmodulin kinase superfamily members as stress-regulated activators of p53 in a cell-specific manner.
3554
CRAIG ET AL.
Apart from the ability of CHK1 to function as the most potent docking-dependent kinase for p53, we also report that DAPK-1 can stimulate Ser20 site phosphorylation of p53. DAPK-1 is defined as a tumor suppressor and was shown to be part of a p53-dependent cell death pathway that operates via p19ARF (42). No physiological cell cycle substrate that mediates DAPK-1-dependent activation of p53 is known, although it is involved in intriguing biological pathways, including autophagic vesicle formation (23). In our biological analysis of the DAPK family, we find that the p53-specific modification in vitro at Ser20 involves primarily DAPK-1/3 rather than DRAK-1 homologues. A p53 specificity for DAPK-1 was also observed in (i) p53-dependent reporter transcription assays; (ii) p21 protein stabilization; (iii) siRNA analysis, where p53 can be depleted upon DAPK-1 protein attenuation in A375 cells; and (iv) the inability of DAPK-3 to simulate the p53 response in vivo. Although CHK2 is held to be the radiation damage-induced p53 kinase, a recent report has shown that irradiation-induced p53 protein stabilization in vivo is attenuated in kinase domain mutants of DAPK-1 (32), suggesting that DAPK-1 can be a component of a genetic program to activate p53 in irradiated cells. It remains to be determined how DAPK-1 coordinately stimulates p53-dependent checkpoint control. The mechanism appears to be distinct from CHK2, since we have been unable to show that DAPK-1 can modify E2F-1 (data not shown); thus, DAPK-1 cannot replace CHK2 as a p53- and E2F-1-activating kinase pathway (49). A high-resolution structure of the p53 tetramer has not been defined, but biochemical studies have shown that some of the functions of the tetramer are allosterically controlled. One pronounced allosteric effect of p53 protein is the DNA dependence in acetylation (17, 18), suggesting conformational changes mediate both binding stability of the acetyltransferase p300 and release of the acetylation motif from a cryptic state (8, 43). Kinase modification of p53 has evolved a dockingdependent interaction similar to that of the p300 acetyltransferase or the E3 ubiquitin ligase MDM2, though p300, MDM2, and protein kinases dock at distinct sites on p53: (i) P300 binds primarily at the BOX-I domain and the proline-repeat domain, (ii) MDM2 binds at the BOX-I and BOX-V domains, (iii) CDK2 modification at Ser315 requires a docking interaction within the tetramerization domain between residues 321 and 348 (20), (iv) PKC-mediated modification at Ser371 also requires distal docking motifs to mediate substrate phosphorylation that are distinct from the CDK2 interface (34), and (v) CK2 docking to the tetramerization domain of p53 tethers the kinase to the low-affinity phospho-acceptor site at Ser392 (34). The BOX-V domain of p53 forms a docking site for the calcium calmodulin kinase superfamily, as described here, but MDM2 was the first BOX-V domain binding protein (47). The BOX-V region is now defined as a key ubiquitination signal for MDM2, which is driven by a dual-site docking mechanism by MDM2 on the p53 tetramer (54). The unfolding of BOX-V region in mutant p53 protein also enhances p53 ubiquitination in vivo (48). Thus, the BOX-V interface appears to be a conformationally flexible multiprotein binding domain that controls covalent modification of the p53 tetramer. Studies using nuclear magnetic resonance have confirmed that the binding of MDM2 to the core domain of p53 can induce chemical shifts within the BOX-V motif (56). There is potential, therefore, for
MOL. CELL. BIOL.
cross talk through competition for the same BOX-V interface: CHK1/DAPK-1 and MDM2 could potentially antagonize each other through competition for the same multiprotein docking site. Although the exposure of this BOX-V region on mutant p53 can enhance MDM2-dependent ubiquitination of p53 in vitro and in cells (48), it is striking that CHK1 can restore partial transcriptional activity to the same mutant p53 protein in vivo (Fig. 7C). These data suggest CHK1 could overcome MDM2-mediated ubiquitination by competition for the same interface in the DNA-binding domain of p53. Furthermore, physiological deletion of this BOX-V region on p53, as exemplified in the p53⌬ S-phase alternatively spliced variant (44), would provide a transcriptionally active form of p53 (as in Fig. 7C) that cannot be degraded well by MDM2 (Fig. 7B) and that cannot be modified by enzymes with the specificity of CHK1 (Fig. 7F). The p53⌬ mutant may have evolved the capacity to evade these regulators during S phase, to alter cell-cycle-responsive promoters, and possibly to directly alter DNA replication (9). As an example of how different regulators of p53 may direct the protein to distinct compartments, in UV-irradiated human skin stem cells, Ser15 and Ser392 phosphorylation of p53 can occur in the same nucleus but within distinct regions of the nucleus (22), suggesting promoter-specific functions for these kinase modifications. Together, these data highlight the potential of the BOX-V domain of p53 to function as a calcium calmodulin kinase docking site, an MDM2 ubiquitination signal, and a region whose removal by splicing offers the cell a range of mechanisms to coordinate p53-dependent responses to distinct cellular stresses. ACKNOWLEDGMENTS A.C. is supported by a Royal Society of Edinburgh Research Fellowship, B.H. is supported by a CRUK Ph.D. studentship, and T.H. is funded by a CRUK Programme grant. REFERENCES 1. Achari, Y., and S. P. Lees-Miller. 2000. Detection of DNA-dependent protein kinase in extracts from human and rodent cells. Methods Mol. Biol 99:85–97. 2. Ahn, J., M. Urist, and C. Prives. 2003. Questioning the role of checkpoint kinase 2 in the p53 DNA damage response. J. Biol. Chem. 278:20480–20489. 3. Avantaggiati, M. L., V. Ogryzko, K. Gardner, A. Giordano, A. S. Levine, and K. Kelly. 1997. Recruitment of p300/CBP in p53-dependent signal pathways. Cell 89:1175–1184. 4. Barlev, N. A., L. Liu, N. H. Chehab, K. Mansfield, K. G. Harris, T. D. Halazonetis, and S. L. Berger. 2001. Acetylation of p53 activates transcription through recruitment of coactivators/histone acetyltransferases. Mol. Cell 8:1243–1254. 5. Biondi, R. M., and A. R. Nebreda. 2003. Signalling specificity of Ser/Thr protein kinases through docking-site-mediated interactions. Biochem. J. 372: 1–13. 6. Bond, J. A., K. Webley, F. S. Wyllie, C. J. Jones, A. Craig, T. Hupp, and D. Wynford-Thomas. 1999. p53-dependent growth arrest and altered p53-immunoreactivity following metabolic labelling with 32P ortho-phosphate in human fibroblasts. Oncogene 18:3788–3792. 7. Burch, L. R., M. Scott, E. Pohler, D. Meek, and T. Hupp. 2004. Phagepeptide display identifies the interferon-responsive, death-activated protein kinase family as a novel modifier of MDM2 and p21WAF1. J. Mol. Biol. 337:115–128. 8. Ceskova, P., H. Chichger, M. Wallace, B. Vojtesek, and T. R. Hupp. 2006. On the mechanism of sequence-specific DNA-dependent acetylation of p53: the acetylation motif is exposed upon DNA binding. J. Mol. Biol. 357:442–456. 9. Cox, L. S., T. Hupp, C. A. Midgley, and D. P. Lane. 1995. A direct effect of activated human p53 on nuclear DNA replication. EMBO J. 14:2099–2105. 10. Craig, A., M. Scott, L. Burch, G. Smith, K. Ball, and T. Hupp. 2003. Allosteric effects mediate CHK2 phosphorylation of the p53 transactivation domain. EMBO Rep. 4:787–792. 11. Craig, A. L., J. P. Blaydes, L. R. Burch, A. M. Thompson, and T. R. Hupp. 1999. Dephosphorylation of p53 at Ser20 after cellular exposure to low levels of non-ionizing radiation. Oncogene 18:6305–6312.
VOL. 27, 2007
KINASE DOCKING WITHIN THE MDM2 UBIQUITINATION SIGNAL
12. Craig, A. L., S. E. Bray, L. E. Finlan, N. M. Kernohan, and T. R. Hupp. 2003. Signaling to p53: the use of phospho-specific antibodies to probe for in vivo kinase activation. Methods Mol. Biol. 234:171–202. 13. Craig, A. L., L. Burch, B. Vojtesek, J. Mikutowska, A. Thompson, and T. R. Hupp. 1999. Novel phosphorylation sites of human tumour suppressor protein p53 at Ser20 and Thr18 that disrupt the binding of mdm2 (mouse double minute 2) protein are modified in human cancers. Biochem. J. 342:133–141. 14. Dale, S., W. A. Wilson, A. M. Edelman, and D. G. Hardie. 1995. Similar substrate recognition motifs for mammalian AMP-activated protein kinase, higher plant HMG-CoA reductase kinase-A, yeast SNF1, and mammalian calmodulin-dependent protein kinase I. FEBS Lett. 361:191–195. 15. Dornan, D., M. Eckert, M. Wallace, H. Shimizu, E. Ramsay, T. R. Hupp, and K. L. Ball. 2004. Interferon regulatory factor 1 binding to p300 stimulates DNA-dependent acetylation of p53. Mol. Cell. Biol. 24:10083–10098. 16. Dornan, D., and T. R. Hupp. 2001. Inhibition of p53-dependent transcription by BOX-I phospho-peptide mimetics that bind to p300. EMBO Rep. 2:139– 144. 17. Dornan, D., H. Shimizu, L. Burch, A. J. Smith, and T. R. Hupp. 2003. The proline repeat domain of p53 binds directly to the transcriptional coactivator p300 and allosterically controls DNA-dependent acetylation of p53. Mol. Cell. Biol. 23:8846–8861. 18. Dornan, D., H. Shimizu, N. D. Perkins, and T. R. Hupp. 2003. DNAdependent acetylation of p53 by the transcription coactivator p300. J. Biol. Chem. 278:13431–13441. 19. Espinosa, J. M., and B. M. Emerson. 2001. Transcriptional regulation by p53 through intrinsic DNA/chromatin binding and site-directed cofactor recruitment. Mol. Cell 8:57–69. 20. Ferguson, M., M. G. Luciani, L. Finlan, E. M. Rankin, S. Ibbotson, A. Fersht, and T. R. Hupp. 2004. The development of a CDK2-docking site peptide that inhibits p53 and sensitizes cells to death. Cell Cycle 3:80–89. 21. Finlan, L., and T. R. Hupp. 2004. The N-terminal interferon-binding domain (IBiD) homology domain of p300 binds to peptides with homology to the p53 transactivation domain. J. Biol. Chem. 279:49395–49405. 22. Finlan, L., R. Nenutil, S. Ibbotson, B. Vojtesek, and T. Hupp. 2006. CK2-site phosphorylation of p53 is induced in ⌬Np63 expressing basal cells in UVB irradiated human skin. Cell Cycle 5:2489–2494. 23. Gozuacik, D., and A. Kimchi. 2004. Autophagy as a cell death and tumor suppressor mechanism. Oncogene 23:2891–2906. 24. Harris, S. L., and A. J. Levine. 2005. The p53 pathway: positive and negative feedback loops. Oncogene 24:2899–2908. 25. Hupp, T. R., and D. P. Lane. 1994. Allosteric activation of latent p53 tetramers. Curr. Biol. 4:865–875. 26. Hutchins, J. R., M. Hughes, and P. R. Clarke. 2000. Substrate specificity determinants of the checkpoint protein kinase Chk1. FEBS Lett. 466:91–95. 27. Iyer, N. G., S. F. Chin, H. Ozdag, Y. Daigo, D. E. Hu, M. Cariati, K. Brindle, S. Aparicio, and C. Caldas. 2004. p300 regulates p53-dependent apoptosis after DNA damage in colorectal cancer cells by modulation of PUMA/p21 levels. Proc. Natl. Acad. Sci. USA 101:7386–7391. 28. Jabbur, J. R., and W. Zhang. 2002. p53 antiproliferative function is enhanced by aspartate substitution at threonine 18 and serine 20. Cancer Biol. Ther. 1:277–283. 29. Jackson, M. W., M. K. Agarwal, M. L. Agarwal, A. Agarwal, P. StanhopeBaker, B. R. Williams, and G. R. Stark. 2004. Limited role of N-terminal phosphoserine residues in the activation of transcription by p53. Oncogene 23:4477–4487. 30. Jallepalli, P. V., C. Lengauer, B. Vogelstein, and F. Bunz. 2003. The Chk2 tumor suppressor is not required for p53 responses in human cancer cells. J. Biol. Chem. 278:20475–20479. 31. Jones, R. G., D. R. Plas, S. Kubek, M. Buzzai, J. Mu, Y. Xu, M. J. Birnbaum, and C. B. Thompson. 2005. AMP-activated protein kinase induces a p53dependent metabolic checkpoint. Mol. Cell 18:283–293. 32. Kishino, M., K. Yukawa, K. Hoshino, A. Kimura, N. Shirasawa, H. Otani, T. Tanaka, K. Owada-Makabe, Y. Tsubota, M. Maeda, M. Ichinose, K. Takeda, S. Akira, and M. Mune. 2004. Deletion of the kinase domain in deathassociated protein kinase attenuates tubular cell apoptosis in renal ischemiareperfusion injury. J. Am. Soc. Nephrol. 15:1826–1834. 33. Lane, D. P., and T. R. Hupp. 2003. Drug discovery and p53. Drug Discov. Today 8:347–355. 34. Luciani, M. G., J. R. Hutchins, D. Zheleva, and T. R. Hupp. 2000. The C-terminal regulatory domain of p53 contains a functional docking site for cyclin A. J. Mol. Biol. 300:503–518.
3555
35. Lukas, C., J. Falck, J. Bartkova, J. Bartek, and J. Lukas. 2003. Distinct spatiotemporal dynamics of mammalian checkpoint regulators induced by DNA damage. Nat. Cell Biol. 5:255–260. 36. MacPherson, D., J. Kim, T. Kim, B. K. Rhee, C. T. Van Oostrom, R. A. DiTullio, M. Venere, T. D. Halazonetis, R. Bronson, A. De Vries, M. Fleming, and T. Jacks. 2004. Defective apoptosis and B-cell lymphomas in mice with p53 point mutation at Ser 23. EMBO J. 23:3689–3699. 37. Mateu, M. G., M. M. Sanchez Del Pino, and A. R. Fersht. 1999. Mechanism of folding and assembly of a small tetrameric protein domain from tumor suppressor p53. Nat. Struct. Biol 6:191–198. 38. McEntyre, J., and T. Gibson. 2004. Patterns and clusters within the PSM column in TiBS, 1992–2004. Trends Biochem. Sci. 29:627–633. 39. Olivier, M., R. Eeles, M. Hollstein, M. A. Khan, C. C. Harris, and P. Hainaut. 2002. The IARC TP53 database: new online mutation analysis and recommendations to users. Hum. Mutat. 19:607–614. 40. O’Neill, T., L. Giarratani, P. Chen, L. Iyer, C. H. Lee, M. Bobiak, F. Kanai, B. B. Zhou, J. H. Chung, and G. A. Rathbun. 2002. Determination of substrate motifs for human Chk1 and hCds1/Chk2 by the oriented peptide library approach. J. Biol. Chem. 277:16102–16115. 41. Ou, Y. H., P. H. Chung, T. P. Sun, and S. Y. Shieh. 2005. p53 C-terminal phosphorylation by CHK1 and CHK2 participates in the regulation of DNAdamage-induced C-terminal acetylation. Mol. Biol. Cell 16:1684–1695. 42. Raveh, T., G. Droguett, M. S. Horwitz, R. A. DePinho, and A. Kimchi. 2001. DAP kinase activates a p19ARF/p53-mediated apoptotic checkpoint to suppress oncogenic transformation. Nat. Cell Biol. 3:1–7. 43. Rippin, T. M., S. M. Freund, D. B. Veprintsev, and A. R. Fersht. 2002. Recognition of DNA by p53 core domain and location of intermolecular contacts of cooperative binding. J. Mol. Biol. 319:351–358. 44. Rohaly, G., J. Chemnitz, S. Dehde, A. Nunez, J. Heukeshoven, W. Deppert, and I. Dornreiter. 2005. A novel p53 isoform is an essential element of the ATR-intra-S phase checkpoint. Cell 122:21–32. 45. Schon, O., A. Friedler, M. Bycroft, S. M. Freund, and A. R. Fersht. 2002. Molecular mechanism of the interaction between MDM2 and p53. J. Mol. Biol. 323:491–501. 46. Shieh, S. Y., J. Ahn, K. Tamai, Y. Taya, and C. Prives. 2000. The human homologs of checkpoint kinases Chk1 and Cds1 (Chk2) phosphorylate p53 at multiple DNA damage-inducible sites. Genes Dev. 14:289–300. 47. Shimizu, H., L. R. Burch, A. J. Smith, D. Dornan, M. Wallace, K. L. Ball, and T. R. Hupp. 2002. The conformationally flexible S9-S10 linker region in the core domain of p53 contains a novel MDM2 binding site whose mutation increases ubiquitination of p53 in vivo. J. Biol. Chem. 277:28446–28458. 48. Shimizu, H., D. Saliba, M. Wallace, L. Finlan, P. R. Langridge-Smith, and T. R. Hupp. 2006. Destabilizing missense mutations in the tumour suppressor protein p53 enhance its ubiquitination in vitro and in vivo. Biochem. J. 397:355–367. 49. Stevens, C., L. Smith, and N. B. La Thangue. 2003. Chk2 activates E2F-1 in response to DNA damage. Nat. Cell Biol. 5:401–409. 50. Takai, H., K. Naka, Y. Okada, M. Watanabe, N. Harada, S. Saito, C. W. Anderson, E. Appella, M. Nakanishi, H. Suzuki, K. Nagashima, H. Sawa, K. Ikeda, and N. Motoyama. 2002. Chk2-deficient mice exhibit radioresistance and defective p53-mediated transcription. EMBO J. 21:5195–5205. 51. Tereshko, V., M. Teplova, J. Brunzelle, D. M. Watterson, and M. Egli. 2001. Crystal structures of the catalytic domain of human protein kinase associated with apoptosis and tumor suppression. Nat. Struct. Biol 8:899–907. 52. Unger, T., T. Juven-Gershon, E. Moallem, M. Berger, R. Vogt Sionov, G. Lozano, M. Oren, and Y. Haupt. 1999. Critical role for Ser20 of human p53 in the negative regulation of p53 by Mdm2. EMBO J. 18:1805–1814. 53. Velentza, A. V., A. M. Schumacher, C. Weiss, M. Egli, and D. M. Watterson. 2001. A protein kinase associated with apoptosis and tumor suppression: structure, activity, and discovery of peptide substrates. J. Biol. Chem. 276: 38956–38965. 54. Wallace, M., E. Worrall, S. Pettersson, T. Hupp, and K. Ball. 2006. Dual site regulation of the E3 ligase activity of MDM2. Mol. Cell 23:251–263. 55. White, R. R., Y. G. Kwon, M. Taing, D. S. Lawrence, and A. M. Edelman. 1998. Definition of optimal substrate recognition motifs of Ca2⫹-calmodulin-dependent protein kinases IV and II reveals shared and distinctive features. J. Biol. Chem. 273:3166–3172. 56. Yu, G. W., S. Rudiger, D. Veprintsev, S. Freund, M. R. Fernandez-Fernandez, and A. R. Fersht. 2006. The central region of HDM2 provides a second binding site for p53. Proc. Natl. Acad. Sci. USA 103:1227–1232.
