Structured DNA promotes phosphorylation of p53 by DNA-dependent protein kinase at serine 9 and threonine 18 Se ´ bastien Soubeyrand 1 , Caroline Schild-Poulter 1 and Robert J. G. Hache ´ 1,2 Departments of 1 Medicine and 2 Biochemistry, Microbiology and Immunology, University of Ottawa, the Ottawa Health Research Institute, Ottawa, Ontario, Canada Phosphorylation at multiple sites within the N-terminus of p53 promotes its dissociation from hdm2/mdm2 and sti- mulates its trans criptional r egulatory potential. T he lar ge phosphoinositide 3-kinase-like kinases ataxia telangiectasia mutated gene p roduct and t he ataxia t elangectasia and RAD-3-related kinase promote phosphorylation o f human p53 at Ser15 and Ser20, and are required for the activation of p53 f ollowing DNA damage. DNA-dependent protein kin- ase (DNA-PK) is another large phosphoinositide 3-kinase- like kinase with the potential to phosphorylate p53 at Ser15, and has been proposed to enhance phosphorylation of these sites in vivo. Moreover, recent studies support a role for DNA-PK in the regulation of p53-mediated apoptosis. We have shown previously that colocalization of p53 and DNA-PK to structured single-stranded DNA dramatically enhances the potential for p53 phosphorylation by DNA- PK. We r eport here the identification of p53 phosphoryla- tion at two novel sites f or DNA-PK , Thr18 and Ser9. Colocalization o f p53 and DNA-PK o n s tructured DNA was required for efficient phosphorylation of p53 at multiple sites, while specific recognition of Ser9 and Th r18 appeared to be dependent upon additional determinants of p53 beyond the N-terminal 65 amino acids. Our results suggest a role for DNA-PK in the modulation of p53 activity resultant from the convergence of p53 and DNA-PK on structured DNA. Keywords: DNA-dependent protein kinase; p53; structured single-stranded DNA; phosphorylation. The large phosphatidylinositide 3-kinase (PI3K)-like kinases are broad specificity serine/threonine kinases w ith essential roles in regulating DNA metabolism and responses to DNA damage. Three of these kinases, DNA-dependent protein kinase (DNA-PK), the ataxia telangiectasia mutated gene product (ATM) and the ataxia telangectasia and RAD-3-related kinase (ATR) [1,2] show a redundant specificity for accessible SQ and TQ motifs in vitro that has hindered definition of their individual roles in DNA repair an d metabolism. In particular, while DNA-PK and its associated kinase activity have been shown to be required for double-stranded DNA (dsDNA) break repair through the nonhomologous end-joining pathway, for V(D)J recombination, and to play at least some role in the regulation o f other processes including transcription, DNA replication and viral integration, demonstration of a role for DNA-PK in specific protein phosphorylation in vivo has remained elusive [1]. We a nd others have hypothesized that substrate phosphorylation by DNA-PK in vivo depends to a large extent on mechanisms that promote the recruitment of substrates to DNA-bound, acti ve, DNA-PK [3–6]. p53 is a key regulatory protein that has the potential to be phosphorylated by DNA-PK, ATM and ATR as Ser15 of human p53 is efficiently phosphorylated by all three kinases in vitro [7]. Phosphorylation a s well as ubiquitylation a nd acetylation control the activation status of p53 [8]. A majority of the phosphorylation sites on p53 are clustered within the N-terminal 40 amino acids (see Fig. 1) and modification at some of these sites, particularly Ser20 and Thr18, promotes the accumulation of active p53 by destabilizing the interaction of p 53 with hdm2/mdm2 [9,10]. Phosphorylation of other sites, such as Ser15, appear to stimulate the transcriptional activation potential of p53, while the exact influence of phosphorylation at other sites remains to b e determined [11]. While p53 phosphorylation in response to DNA damage has long made it an attractive in vivo candidate target of DNA-PK, ATM and ATR are now believed to constitute the main effectors leading, directly as well as indirectly, to p53 phosphorylation in response to DNA damage [12,13]. Nonetheless, it has been reported that in cells lacking ATM, accumulation of p53 and phosphorylation within the N-terminus of p53 in response to treatment with agents that induce dsDNA breaks still occurs, albeit at a lower levels or with delayed kinetics [14,15]. Fu rther, in certain situations DNA-PK is essential for p53-dependent DNA damage- mediated apopto sis [16,17]. In addition, DNA-PK and p53 have both been implicated in controlling the integrity of DNA replication and repair [18–24]. D NA-PK r eaches a maximum level during G 1 /early S phase, suggesting that DNA replicative structures can activate DNA-PK [25]. Correspondence to S. Soubeyrand, Departments of Medicine, Uni- versity of Ottawa, the Otta wa Hea lth Research Institute, 725 Parkdale Avenue, Ottawa, Ontario, Canada K1Y 4E9. Fax: +613 7615036; Tel.: +613 7985555 ext 13705; E-mail: ssoubeyrand@ohri.ca Abbreviations: ATM, ataxia telangiectasia mutated gene product; ATR, ataxia telangectasia and RAD-3-related kinase; dsDNA, double-stranded DNA; DNA-PK, DNA-dependent protein kinase; PI3K, phosphoinositide 3-kinase; ssDNA, single-stranded DNA. (Received 10 March 2004, revised 6 July 2004, accepted 2 A ugust 2 004) Eur. J. Biochem. 271, 3776–3784 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04319.x In vitro, p53 and DNA-PK both interact with single- stranded, st ructured and damaged DNA [26–30]. The sequence-independent DNA binding ability o f p53, which depends on its C-terminus as well as the core domain, has been proposed to play an important part in the initiation o f cellular responses to DNA damage [31–34]. Recently, we have shown that DNA-PK is activated from structured single-stranded DNA (ssDNA) and hairpin DNA ends resembling r eplication and recombination intermediates [30,35]. Preliminary studies indicated the phosphorylation of p53 by DNA-PK was dramatically enhanced by the colocalization of p53 to the ssDNA [30]. In the present study we have performed a detailed analysis of the phosphorylation of p53 in the presence of ssDNA. We report the identification of two sites of DNA-PK phos- phorylation in the N-terminus of p53, Thr18 and Ser9, which are preferentially phosphorylated by DNA-PK when p53 and DNA-PK are colocalized to ssDNA. These results reinforce the importance of colocalization for substrate phosphorylation b y DNA-PK and emphasize that DNA- PK is a kinase with a broad specificity. They also suggest a specific role for DNA-PK in the phosphorylation of p53 from structured DNA in vivo. Materials and methods Chemicals and recombinant proteins Purified DNA-PK and the p53-derived peptide were obtained from Promega (Madison, WI, USA). Wortmannin was from Sigma (St. Louis, MO, USA) and LY294002 from Calbiochem (San Diego, CA, USA). The p53 wt as well as the Ser15 and Ser37 variants have been described elsewhere [11], while the other point mutants were generated by QuikChange mutagenesis (Stratagene, La Jolla, CA, USA). Mutations were confirmed by dideoxynucleotide sequen- cing. The truncated p53 forms, p53 1)65 and p5 3 D30C ,were generated by introducing nonsense mutations at positions 66 (Met) and 363 (Arg), respectively. The recombinant p53s were expressed as fusion proteins from pGEX-6P1 and purified on Glutathione Sepharose 4B (Amersham Phar- macia Biotech; Piscataway, NJ, USA) and then cleaved from the GST with PreScission protease. The purity of all p53 p reparations was monitored b y SDS/PAGE a nalysis. Single-stranded M13 DNA (ssM13) was obtained from New England Biolabs (Beverly, MA, USA) while linearized pBluescript DNA was prepared by extraction of Hin dIII digested plasmid (Qiagen) from agarose gels. The p53 (FL- 393) and p53 pSer9 polyclonal antibodies were obtained from Santa Cruz (Santa Cruz, CA, USA) and Cell Signaling Technology (Beverly, MA, USA), respectively. DNA-PK kinase assays Assays were performed with 0.2 l M of p53 peptides or 100 n g of recombinant p53 at 30 °Cfor15mininthe presence of 4.2 n M of [ 32 P]ATP[cP] (3000 CiÆmmol )1 ), 10 ng of DNA and 10 units of DNA-PK in 20 lLof reaction buffer (50 m M HEPES, 100 m M KCl, 10 m M MgCl 2 ,0.2m M EGTA, pH 7.5). The final ATP concentra- tion was adjusted to 50 l M where indicated. Following completion of the reaction, the substrates were resolved by SDS/PAGE (8–12%) and visualized by autoradiography. Quantification was performed by Phosphorimager analysis (Typhoon 8600, Molecular Dynamics) in the presence of a series of [ 32 P]ATP[cP] standards. Inhibition experiments with LY294002 (0.3–300 l M ) and wortmannin (3–3000 n M ) dissolved in dimethyl sulfoxide w ere performed as above except that t he inhibitor was incubated for 5 min at 30 °C in the presence of either ssM13 DNA (for p53 wt and p53 S15A/S37A phosphorylation) or dsDNA (for p53 peptide phosphorylation) prior to the additio n of substrate. Phos- phorylation was quantified by Phosphorimager analysis of the polyacrylamide gel. Phosphorylation was expressed relative to the mock-treated kinase and the resulting ratios fit to a sigmoidal curve used to derive the IC 50 .The inhibition of p53 phosphorylation subsequent to DNA-PK autoinactivation was assessed as previously described [30]. DNA-PK was preincubated in the presence of ssM13 with or without 50 l M ATP for 10 min at 30 °C. Subsequently [ 32 P]ATP[cP] in a final ATP concentration of 5 0 l M was added and incubation continued f or 15 min. Dissection of p53 phosphorylation by DNA-PK in vitro Radiolabeled p53 was resolved by SDS/PAGE, excised and digested at 37 °C in situ in 400 lL of digestion buffer (50 m M NH 4 HCO 3 , pH 8.0) containing 0.050 lgÆlL )1 TPCK-treated trypsin (Worthington Biochemical Corpora- tion; Freehold, NJ, USA) for 16 h . This was followed by the addition of fresh trypsin (0.025 lgÆlL )1 ) and redigestion for 3 h . The supernatant was evaporated under vacuum and the Fig. 1. Phosp horylation of p53 by DNA-PK from ssDNA at sites beyond Ser15 a nd Ser37. (A) Schematic presentatio n of p53 h igh- lighting Ser (S) and Thr (T) residues in the N-terminal 60 amino acids. The filled a rrowh eads indic ate the position of trypsin cleavage, whereas the open arrowhead indicates the location of a CNBr cleavage site. (B,C) Phosphorylation of recombinantp53.(B)Purifiedrecombinant p53 wt and the indicated variants were phosphorylated by DNA-PK in the presence of 40 n M [ 32 P]ATP[cP] (3 CiÆlmol )1 )intheabsenceof DNA (–), or in the p resence o f 10 ng of ssM13 DNA (ss) or SmaI- linearized pBluescript (ds) DNA. The phosphorylated p53s were resolved b y SDS/PAGE and quantified by Phosphorimage r analysis. Phosphate incorporation is indicated (values below the lanes; fmol P). (C) P hospho rylation of p53 wt and p 53 S15A/S37A as in (B) but in the presence of 50 l M [ 32 P]ATP[cP]. Phosphate incorporation is indicated (values below the lanes; pmol P). Ó FEBS 2004 Specificity of DNA-PK towards p53 (Eur. J. Biochem. 271) 3777 pellet resuspended in a denaturation buffer (6 M urea, 25 m M Tris/HCl, pH 8.0). The sample was then loaded onto a 40% alkaline acrylamide gel as described previously [35] and resolved for 6000 VÆh )1 at 3 W. For CNBr digests, the protein was trypsinized as above, evaporated to dryness and incubated in 100 lL of CNBr i n formic aci d (100 mgÆmL )1 )for90minat20°C. The samples were then vacuum-dried and the pellets resuspended in denaturation buffer and submitted to electrophoresis as above. Each analysis was confirmed by obtaining two to five repetitions with reproducible results. For phosphoamino acid mapping, the trypsinized p53 fragments were resolved by 40% PAGE and eluted in H 2 O. An aliquot was t hen evaporated to dryness a nd 5.5 M HCl wasaddedfor1hat110°C. The hydrolysate was evaporated, mixed with unlabeled pSer, pThr and pTyr standards and then applied onto 10 · 10 cm plastic backed cellulose thin layer chromatography plates (Merck, Darms- tadt, Germany). Phosphomanino acids were resolved by two consecutive ascending chromatographies in ethanol/ acetic acid/H 2 O (1 : 1 : 1, v/v/v; 80 min) and 2 -propanol/ HCl/H 2 O (7 : 1.5 : 1.5, v/v/v; 180 min). The phospho- amino acids were then visualized by spraying the plates with 0.25% (v/v) ninhydrin/acetone. Results Phosphorylation of p53 by DNA-PK from ssDNA at novel sites for DNA-PK While p53 has been reported t o be phosphorylated exclu- sively on Ser15 and Ser37 by DNA-PK in the presence of double-stranded linear DNA, to d ate no study has evalu- ated the impact of structured DNA or DNA colocalization on the kinase specificity [11,36]. To begin detailed analysis of the phosphorylation of p53 colocalized to ssDNA with DNA-PK, we compared the phosphorylation of recombin- ant p53 with the phosphorylation of S15A and S37A substituted p53 (p53 S15A ,p53 S37A ) in the presence of ssM13 and linearized double-stranded plasmid DNA (Fig. 1). In the presence of ssDNA, substrate phosphorylation occurs in competition with DNA-PK autophosphorylation and autoinactivation [30]. Previously we demonstrated that this potent autoinactivation of DNA-PK linked in cis can be minimized when assessing phosphorylation of heterologous DNA-PK substrates by performing the kinase reactions at the limiting ATP concentration of 40 n M [30]. At 40 n M ATP, p53 w as phosphorylated 11 ± 2.5 ( n ¼ 5) times more efficiently by DNA-PK in t he presence of the optimal amount of ssM13 than in the presence of an equimolar amount of linearized double-stranded plasmid DNA (Fig. 1B). Unexpectedly, p53 substituted at Ser15 and Ser37 remained a strong substrate for DNA-PK in the presence of ssM13, with 20 ± 5% (n ¼ 4) of the phosphate incorporation of p53 wt . Indeed p53 S15A/S37A was phosphor- ylated three times more efficiently in the presence of ssM13 than was p53 wt in the presence of linear plasmid DNA (Fig. 1 B, lanes 3, 11). To determine whether this additional phosphorylation of p53 a rose due to the limiting concentration o f ATP in the assay, we repeated the experiment at the usual ATP concentration employed for DNA-PK, 50 l M (Fig. 1 C). Phosphorylation of p53 in the presence of ssM13 was reduced to 3.0 ± 0.4 (n ¼ 4) times the efficiency of p53 phosphorylation in the presence of linear double-stranded DNA. This high remaining level of p53 phosphorylation by DNA-PK i n the presence of ssM13 DNA has previously been shown to be dire ctly a ttributable to colocalization of p53 to the ssM13 with DNA-PK, wh ich allows for rapid p53 phosphorylation prior to a DNA-PK autoinactivation [30]. Before pursuing the sites of this n ew phosphorylation it was important to ensure that the phosphorylation observed was mediated directly by the DNA-PK rather than a minor contaminant of the DNA-PK p reparation. Although our SDS/PAGE analysis indicated that the DNA-PK was about 90% pure, the potential contribution of contaminating kinases had to be excluded. Consequently, we titrated the sensitivity o f phosphorylation of p53 S15A/S37A and the classical p53-derived DNA-PK peptide substrate containing only Ser15 to the DNA-PK inhibitors wortmannin and LY294002. The IC 50 values for p53 S15A/S37A closely matched the values obtained for the p53 peptide, confirming that both activities were due to a single enzyme species, namely DNA-PK (Fig. 2A). Notably, these values exclude Fig. 2. DNA-PK directly mediates phosphorylation of p53 S15A/S37A . (A) IC 50 values for the in hibition of p53 phosphorylation. Kinase reactions with p53 S15A/S37A (100 n g) or the synthetic p53-derived peptide (0.5 lg) w ere performed with DNA-PK in the presen ce of increasing amounts of Wortmannin (3–3000 n M ), LY294002 (0.3–300 l M )oran equivalent amount of dimethyl sulfoxide. The reaction products were quantified by P hosp horimage analysis of polyacrylamide gels. Th e IC 50 values are the mean of two interpolations from two independent inhibition profiles. (B) Autoinactivation of DNA-PK prevents r p53 S15A/S37A phosphorylation. DNA-PK was preincubated for 10 min either with ( lan es 1, 3) or without ( lan es 2, 4 ) 50 l M ATP i n the presence of ssM13. [ 32 P]ATP[cP]wasthenaddedandtheATPcon- centration raised to 50 l M in all the samples and kinasing of p53 wt (lanes 1, 2) or p53 S15A/S37A (lanes 3, 4) was performed by standard assay. On the left is a Phosphorimager analysis of a representative gel and o n the right i s a graphical display of the Phosphorimager quan- tification of two independent determinations (± SD ) expressed as the ratio of p53 phosphorylation following a preincubation with ATP over a control preincubation without ATP. 3778 S. Soubeyrand et al. (Eur. J. Biochem. 271) Ó FEBS 2004 phosphorylation of p53 S15A/S37A by minor contaminating amounts of ATM or ATR i n the kinase preparation, as these kinases are not inhibited by the concentrations of wortmannin and LY294002 employed [37–40]. To further ascertain that DNA-PK is directly involved in p53 S15A/S37A phosphorylation, we took advantage of the rapid autoinactivation of DNA-PK that occurs on ssM13 in the presence of 50 l M ATP [30]. It was reasoned that a contaminating kinase should remain unaffected by this rapid, DNA-dependent, inactivation and that p53 phos- phorylation should then proceed normally. On the contrary, preincubation of the DNA-PK for 10 min in the presence of 50 l M ATP and ssM13 prior to addition of p53 and [ 32 P]ATP[cP], led to phosphorylation of both p53 wt and p53 S15A/S37A by 80% (Fig. 2B). Hence DNA-PK directly targets p53 S15A/S37A . To begin analysis of p53 phosphorylation by DNA-PK in the p resence o f ssM13 in greater d etail, tryptic digests of p53 wt phosphorylated in the presence o f 40 n M and 50 l M ATP were resolved on a 40% alkaline polyacrylamide gels (Fig. 3 ). Alkaline PAGE allows separation of peptides according to a combination of charge and size; the presence of additional negative charges, such as those induced by phosphorylation or by substitution of Ser with Asp or Glu, enhances peptide migration. Trypsin digestion of p53 is expected to lead to the separation of Ser15 and Ser37 onto two peptides containing amino acids 1–24 and 25–65, respectively (Fig. 1A). p53 phosphorylation a t 40 n M ATP in the presence of ssM13 resulted in the resolution of two major tryptic phosphopep- tides ( A a nd B) on alkaline g els (Fig. 3A). Two peptides with the same corresponding migrations were also observed following trypsin digestion of an N-terminal p53 peptide (aa 1–65) phosphorylated by DNA-PK (Fig. 3B). A third peptide w hose appearance varied in intensity through the course of the s tudy, designated A ¢, was observed in both instances. This peptide likely reflects an alternative cleavage product of peptide A as both bands were abrogated by the Ala37 substitution (Fig. 4B). In summary, these data suggested that the additional phosphorylation of p53 detected at 40 n M ATP occurred i n the N-terminus of p53 within the two peptides containing Ser15 and Ser37. Interestingly, at 50 l M ATP, two additional phospho- peptides, with intensity approaching t hat o f p eptide B as well as a somewhat weaker band were detected within full- length p53 (Fig. 3C, peptides C, D and E, respectively). Additionally, t he signal yielded by peptides A broadened and decreased in resolution. These results suggested that the activity of DNA-PK a t the higher ATP concen tration was increased to include additional sites within p53. Import- antly, although weaker in intensity, highly similar t ryptic profiles were obtained in the presence of dsDNA ends (data not shown), indicating that although colocalization stimu- lated phosphorylation of p53 it did not appear to induce the exposure of new sites on p53. DNA-PK phosphorylates p53 at Thr18 and Ser9 The relative s implicity of tryptic peptide digestion pattern of p53 phosphorylated at 40 n M ATP prompted us t o first characterize the additional p53 phosphorylation under these conditions. T o identify peptides A and B , p53 phosphorylated by DNA-PK from ssM13 DNA at 40 n M ATP was treated with CNBr which cleaves p53 tryptic peptide 25–65, but not 1–24 (Fig. 1A). CNBr treatment of the tryptic digest converted peptide A to a higher mobility peptide, without affecting the intensity or m obility of peptide B (Fig. 4A). This identified peptide A as contain- ing amino acids 25–65 of p53 and peptide B as containing amino acids 1–24. Substitution of Ser37 with Ala in full-length p53 elimin- ated the signal from peptides A/A ¢ while conversion of Ser15 to Ala strongly interfered with, but did not abrogate, fragment B phosphorylation (Fig. 4B). Together these results identify the presence of a new DNA-PK phosphory- lation site in amino acids 1–24 of p53. The presence of additional phosphorylation site(s) within tryptic peptide B was also observed in t he context of a polypeptide spanning aa 1–65 following Ser15 and Ser37 substitutions, despite a > 95% reduction in total phosphorylation (Fig. 4C). In addition to Ser15, peptide B contains Ser6, Ser9, Thr18 and Ser20 as well as an additional Ser (at position 1) that comigrates upon cleavage of the GST tag (Fig. 1A). Phosphoamino acid analysis of peptide B from Ala15/37- substituted p53 revealed the predominant presence of phosphothreonine (Fig. 4D, top), thereby establishing Thr18 (the only threonine residue in amino acids 1–24) as a third major DNA-PK phosphorylation site within the N-terminus of p53. Interestingly a similar analy sis of the wild-type protein showed proportionally less but significant Thr18 phosphorylation demonstrating that phosphoryla- tion does indeed occur at this site in the Wt context (Fig. 4 D, bottom). While at limiting ATP concentrations p53 was almost exclusively phosphorylated on Ser15, Thr18 or Ser37, at the saturating and physiologically relevant ATP concentration of 50 l M , additional radiolabeled tryptic p53 peptides were observed (Fig. 3C, bands C and D). The in troduction of T18E or S15D mutations shifted the migration of these phosphopeptides indicating that they were phosphopeptide B-derived (data not shown). Wt 1-65 Wt ABC A B A B A B C D E 40 nM ATP 50 µM ATP 40 nM ATP Fig. 3. Tryptic analysis of p53 phosphorylation. Alkaline PAGE ana- lysis of t he ph osphorylation of tryptic peptides of p53 wt (A,C) and p53 1)65 (B) phosphorylated by DNA-PK in the presence of ssM13 (A,C) or linearized pBluescript DNA (B) and 40 n M (A,B) o r 50 l M [ 32 P]ATP[cP] (C). Aliquots of 2000 cpm from the tryptic digests of phosphorylation reaction were resolved through 40% alkaline PAGE. Tryptic phosphopeptides were labeled A–E on the basis of increasing mobility. Peptide A¢ is a subordinate cleavage product of peptide A as discussed in the text. Ó FEBS 2004 Specificity of DNA-PK towards p53 (Eur. J. Biochem. 271) 3779 To identify the r emaining three bands originating f rom p53 tryptic peptide 1–24, we assessed the effect of additional substitutions on the phosphorylation of p53 at 50 l M ATP (Fig. 4 E). A s m entioned above, the recombinant p53 used in the mapping contained a serine residue at position 1 as which was replaced with Ala. Substitution of this Ala reduced the peptides migrating in the range B-E from 3 to 2 indicating that it was indeed phosphorylated (Fig. 4E, top, lanes 1 and 2). Within that context, substitution of Ser9, but not Ser6 nor Ser20, with Ala resulted in the loss of the remaining higher mobility peptide, leaving a single peptide, presumably phosphorylated at Thr18 (Fig. 4E, lane 4). Finally, mutation of both Thr18 and Ser9 to Ala in the context of the Ser1/15 mutation abrogated fragment B (data not shown), consistent with phosphorylation o f both Thr18 and Ser 9 . Introduction of the single Thr18 an d S er9 mutations in the wild-type protein background resulted in the a brogation o f o ne band, further i ndicating that these sites are genuine targets in the wild-type protein (Fig. 4E, lanes 6–8) and not artifacts due to Ser15 mutation. Finally, to confirm the presence of the nonconsensus p53 phosphorylation in the context of a wild-type protein, western blot analysis of p53 wt phosphorylated by DNA-PK was performed. Because of a lack of a suitable pThr18 antibody, we focused on Ser9 phosphorylation. Ser9 phosphorylation was observed only in the presence of both DNA-PK and p53 wt (and not in the alanine-substituted control p53), indicating that Ser9 was targeted by DNA-PK in the context of t he wild-type protein (Fig. 4F). Perhaps not surprisingly, in view of the lack of effect on total phosphorylation by the S15A and S37A single mutations (Fig. 1 B), initial attempts at comparing total phosphorylation of p53 wt and p53 S9A/T18V revealed no significant difference (data not shown). Consequently, the proportional significance of these sites on total phosphory- lation was rather e stimated in the co ntext of the wild-type protein by quantifiying the tryptic profiles of phosphoryl- ated p53 wt ; this a pproach had the additional advantage of circumventing potential artifacts arising from the introduc- tion of mutations. Taking into consideration that the fastest Fig. 4. Phosphorylation of p53 on Thr18 and Ser9. (A) Alkaline PAGE analysis of CNBr cleavage of tryptic phosphopeptides derived from p53 wt phosphorylated by DNA-PK in the presence of ssM13 and 40 n M [ 32 P]ATP[cP]. The migration o f tryptic phosphopeptides A and B are indicated by arrows. (B) Tryptic phosphopeptide profiles of p53 wt ,p53 S15A ,p53 S37A and p53 S15A/S37A phosphorylated by DNA-PK in the presence of 40 n M [ 32 P]ATP[cP] and ssM13 DNA. (C) Tryptic phosphopeptide profiles of p53 1)65 (Wt 1–65, 2 l M ) and S15A/S37A substituted p53 1)65 (S15A/S37A 1–65, 2 l M ) phosphorylated by DNA-PK as in (B). Phosphate incorporation (pmol) is indicated at the bottom below the e xposure. (D) Tryptic phosphopeptide B o f DNA-PK p hosphorylated p53 Wt or p53 S15A/S37A waselutedfroma40%alkalinePAGEgelandhydrolyzedinHCl. Phosphoamino acids were resolved by TLC in the presence of phosphoserine and phosphothreonine markers. Assignment of phosphorylation was made by su perimpo sition of autoradiographs and ninhydrin staining, with the po sition o f phosp hoserine ( pSer) and phosphothreonine (pThr) migration indicated to the left of the phosphorimage. (E) Alkalin e PAG E a nalysis of tryptic phosphopeptides derived from recombinant p53s following incubation with DNA-PK in the presence of ssM13 and 50 l M [ 32 P]ATP[cP]. (F) Western blot analysis of p53 Wt or p53 S9A/T18V phosphorylation by DNA-PK . DNA-PK was incubated with the indicated p53 species in the presence of 50 l M ATP and assessed for total phosphorylation (top) or pSer9 phosphorylation (bottom) by Western blotting. The amino acid substitutions within full-length recombinant p53 are listed at the top of each lane in the panels. In panel (E) and (F), p53 substituted at Ser1 with Ala is highlighted by asterisks. 3780 S. Soubeyrand et al. (Eur. J. Biochem. 271) Ó FEBS 2004 band (Fig. 3C or Fig. 5B, band E) has four phosphate groups, with a progressive reduction of one band per phosphate removed, one can e stimate the contri bution of Ser9 and Thr18 on total phosphorylation. Conservatively assuming that all of t he B-E bands are phosphorylated at Ser15 and that C-E are also phosphorylated on pS er1, leaving D and E as containing phosphorylation on Ser9 and Thr18, phosphorylation at the latter sites would account for 10% ± 2% of total phosphorylation. Taking the least conservative ap proach, i.e. inferring that pSer15/p Ser1 phosphorylation correspond to the two lowest intensity fragments, would increase this value to 18% ± 2%. Thus phosphorylation at t hese two sites probably accounts for 8–20% of total p53 phosphorylation. Phosphorylation of p53 at Ser9 andThr18 is preferentially enhanced within full-length p53 Initial experiments comparing p53 phosphorylation by DNA-PK with the phosphorylation of a p53 peptide containing only the N-terminal 65 amino acids (p53 1)65 ) indicated that p53 1)65 was a noticeably poor substrate for DNA-PK in the presence of ssDNA at 40 n M ATP (Fig. 3 B). Similarly, at 50 l M ATP, p53 1)65 phosphoryla- tion occurred with an effic iency less than 1% that of p53 wt (Fig. 5 A). By contrast, phosphorylation by DNA-PK in the presence of dsDNA increased the absolute phosphorylation of the p53 peptide 30-fold, while decreasing phosphate incorporated into p53 wt by 2.5-fo ld. T ogether, these d ata suggested that p53 phosphorylation was affected by the nature of the DNA cofactor and by t he remainder of p53 beyond amino acid 65. Previously, we have demonstrated that the efficiency of phosphorylation o f r ecombinant p53 by DNA-PK in the presence of ssDNA correlated directly with the ssDNA binding ability of p53 [30]. In the present experiments, however, the reduction i n th e efficiency of phosphorylation of the p53 peptide could not entirely be accounted for by the loss of ssDNA binding (Fig. 5 A). Phosphorylation of a mutated version o f p53 l acking the C -terminal 30 amino acids (p53 D30 C ) that is unable to interact with ssDNA [27], occurred w ith an effic iency that was o nly eightfold lower than p5 3 wt inthepresenceofssM13,leavingthelevelof phosphorylation of p53 D30 C 30-fold higher than that of p53 1)65 (0.37/0.011 pmol). To investigate whether DNA binding and the presen ta- tion of full-length p53 also influenced the recognition of individual phosphorylation sites by DNA-PK, we com- pared the pattern of tryptic phosphopeptides obtained from p53 wt ,p53 D30C , and the amino acid 1–65 p eptide phosphor- ylated by DNA-PK in the presence of ssM13 DNA (Fig. 5 B). Interestingly, while the ratio between peptides A/A¢ an d B showed little variation between substrates, the level of phosphorylation of peptides C-D was markedly decreased for p53 D30 C and was undetectable for the p53 peptide (Fig. 5B), even upon prolonged exposure of the gels. In order to b etter discriminate the contribution o f structural elements within p53 that may promote its phosphorylation at S er9 and Thr18 from the direct contri- bution of p53 DNA binding to structured DNA, we quantified the absolute levels of phosphorylation of p53 wt with p53 S15A/S37A in the presence of dsDNA. Utilization of dsDNA minimizes DN A binding by p53 and resulted in more similar total phosphorylation levels (Fig. 5A). Substi- tution of Ser15 and Ser37 with Ala in full-length recombin- ant p53-reduced phosphate incorporation to 35% of the level of both p53 wt and p53 D30 C at 50 l M ATP, confirming that colocalization to DNA was not required for the phosphorylation of Ser9 and Thr18 by DNA-PK. In contrast, DNA-PK w as essent ially unable to e ffect phos- phorylation of p53 1)65, S15A/S37A . Thus, t hese data indicate that phosphorylation of p53 at Ser9 and Thr18 by DNA- PK is dependent upon specific determinants within the remainder of the p53 protein that are not directly related to its ability to bind DNA structures. Discussion Our results demonstrate the phosphorylation of p53 at two sites, Ser9 and Thr18, which have not previously been appreciated as potential targets for DNA-PK in vitro. Importantly, phosphorylation at Ser9 and Thr18 showed a strong preference for the colocalization of p53 and DNA- PK on ssDNA. This may explain why these sites have not been previously recognized as bona fide DNA-PK targets. Indeed, typical DNA-PK activity assays involve dsDNA ends in combination with p eptides or polypeptides s pan- ning the N-terminal portion of p53. Another ancillary Fig. 5. Phosphorylation of p53 within the novel N-terminal sites is dependent on binding to ssDNA and full-length p53. The phosphoryla- tion o f p53 wt ,p53 D30C and p 53 1)65 by DNA-PK in the presence of 50 l M [ 32 P]ATP[cP] is compared. (A) Comparison of total phosphate incorporation (pmol) in the presence o f ssM13 and linear pbluescript dsDNA. Data shown is representative of the results of three inde- pendent experiments. (B) A lkaline PAGE analysis of tryptic phos- phopeptide l abeling o f t he three forms of p53 phosphorylated by DNA-PK in the presence o f ssM13. The position o f migration of phosphopeptides A–E is indicated to the left o f t he panel. (C) The contribution of phosphorylation of p53 at Ser15 and Ser37 to the total phosphorylation of p53 by DNA-PK in the presence of dsD NA was determined by comparing 32 P incorporation int o wild-type an d Ala- substituted recombinant p53s. Fo llowing in cubation with DNA-PK, the p 53 polypeptides we re resolved by S DS/PAGE and phosphate incorporation was quantified by Phosphorimager. The results are expressed as a ratio of the phosphorylation of the alanine-substituted p53 variant (hatched bars) over its serine equivalent (100%, s olid bars). Data represent the mean ± SD of three determinations performed in duplicates. Ó FEBS 2004 Specificity of DNA-PK towards p53 (Eur. J. Biochem. 271) 3781 explanation resides in the relatively low phosphorylation level of these sites. We have estimated that phosphorylation at these sites may account for 10-20% of total phosphory- lation of the wild-type protein at 50 l M ATP. Clea rly th is does not account for the 35% phosphorylation remaining observed in the absence of both Ser15 and Ser37. This discrepancy suggests that Ala mutations may either intro- duce potential novel sites elsewhere in p53 or somehow facilitate phosphorylation of Ser9 and Thr18. The a bsence o f Ser9/T hr18 phosphorylation in p53 1)65 suggests that the overall conformation of p53 or determi- nants beyond the N-terminal 65 amino acids are important for phosphorylation at Ser9 and Thr18. These results also suggest that the conformation change induced by the binding of p53 to ssDNA and DNA ends facilitates the presentation of Ser9 and T hr18 in a m anner th at makes them attractive substrates for DNA-PK. This may be mediated in part by the core domain of p53 which although insufficient, has been shown to b e r equired for sequence- independent binding [33]. Alternatively, a second possibility is that full-length p53 becomes involved in a protein–protein interaction with DNA-PK that promotes p53 [41]. While p53 has been known t o i nteract w ith linear and ssDNA for several years, the functional implications of th is activity have been uncertain. Binding of p53 to ssDNA is known to stimulate sequence-specific DNA binding and may play a role i n promoting tetramerization of the protein [27]. Our present and previous results [30] show that colocaliza- tion of p53 and DNA-PK to such DNAs promote a close to 10-fold enhancement of p53 phosphorylation. Thus colocal- ization of DNA-PK and p53 to DNA would likely be important for regulation of p53 by DNA-PK in vivo. Phosphorylation was highly specific as several other sites in the N-terminus of p53, including Ser6 and Ser20 were not recognized by DNA-PK. Phosphorylation of Thr18 appeared to be preferred to phosphorylation at Ser9 in vitro, as it was the only additional site detected at limiting ATP concentrations. It is interesting that in human p53 wt ,Ser9 follows a P ro residue, a s it has been suggested p reviously that such Pro-Ser/Thr might in fact form a variant consensus site for DNA-PK [42]. By contrast the two additional Ser-Gln dipeptides in p53, at amino acids 99–100 and 166–167 were not recognized in full-length p53, as assessed by a lack of a shift in fragment A migration (data not shown), even though a p eptide containing Ser99 is recognized by DNA-PK [43]. This apparent discrepancy reiterates how important the molecular environment of the target site is in determining th e specificity of the k inase. DNA-PK is not the only candidate kinase for phos- phorylation of p53 at Ser9 and Thr18. Previous work has shown that casein kinase 1 also has the potential to phosphorylate p53 at Ser9 and Thr18 [44,45]. For casein kinase 1, phosphorylation o f p 53 at Ser9 and Thr18 was dependent on prior phosphorylation of Ser6 and Ser15. Phosphorylation was also readily observed with N-terminal peptides of p53. By contrast, phosphorylation of Ser9 and Thr18 by DNA-PK was depend ent on full-length p53 but was independent of phosphorylation at other sites in p53. It was also independent of the addition of IC 261, a specific casein kinase 1 inhibitor [46]. The checkpoint kinases Chk1 and Chk2 have also been associated with phosphorylation at several N - and C-terminal sites of p53 in vitro including Ser20 [47]. Here again DNA-PK differs as no phosphory- lation of Ser20 was detected in our assay. Of significant interest, Chk1 directly stimulates the ability of DNA-PK to phosphorylate p53 [48]. While th e authors focused most of their study on a truncated version of p53 and did not evaluate the stimulation in the presence of ssDNA, it will be interesting to evaluate the impact of Chk1 on the specificity of DNA-PK toward full-length p53. The gatekeeper function of p53 depends principally on its ability to monitor progression of cells through the cell cycle, and to induce cell cycle arrest or direct a cell towards apoptosis in re sponse to a variety of stresses [12]. Numer- ous reports have demonstrated that phosphorylation of N-terminal domain of p53 is essential to the accumulation of p53 and potentiates p53 acetylatio n and its transactiva- tion function [49]. Identification of the kinases involved in vivo has been challenging and it has become obvious that there is currently no simple Ôone site-one kinaseÕ model to fit the experimental evidence. Rather, p53 phosphorylation probably involves a complex network of kinases whose interactions between themselves and p53 depend upon the exact n ature o f t he st ress and the cell type involved. For instance while Chk1 and C hk2 were long held as kinases acting immediately upstream of p53, two recent reports have questioned t heir implications in p53 stabilization, at least in certain cancer cells, and it has been su ggested that a yet-to-be identified kinase(s) is(are) involved instead [50,51]. Currently, several lines of evidence point to a role of DNA-PK in the apoptotic branch of the p53 pathway. Indeed, activation of DNA-PK in response to ionizing radiation is d irectly linked t o t he activation of the latent cellular population of p53 that directs cells towards DNA damage-induced apoptosis [16]. Further, the presence of shortened telomeres that result from telomerase deficiency fail to induce apoptosis in the absence of DNA-PKcs [52]. Thus despite the overlap between the l arge PI3K-like kinases i n their ability to phosphorylate p53 in vitro,p53 phosphorylation by DNA-PK might occur under appro- priate circumstances. A chanta et al . has provided evidence that DNA-PK may also play an important role in the p53- dependent induction of apoptosis that follows nucleoside- induced arrest of DN A s ynthesis [41]. They a lso s howed that p53 and DNA-PK colocalize in the nuclei of nucleo- side-treated cells and could be coimmunoprecipitated. Our results offer the intriguing possibility that the accumulation of stalled replication intermediates, which contain ssDNA regions, may directly facilitate the phosphorylation of p53 by DNA-PK [53]. In conclusion, our results broaden the previously recog- nized specificity o f DNA-PK t owards p53 t o include two new sites, Ser9 and Thr18. It will be important to next determine whether DNA-PK plays a role in mediating the phosphorylation of these sites in response to dsDNA breaks and to explore whether the action of DNA-PK on p53 occurs in response to other forms of cellular stress, such as replication blocks induced by nucleoside analogues or topoisomerase poisons. Given the similarities in substrate selection by DNA-PK, ATM and ATR, it will also be interestingtoassesswhetherSer9andThr18canalsobe targeted by these kinases, particularly because Ser15 phos- phorylation in vivo is not required to mediate cell cycle regulation following ionizing radiation [54]. 3782 S. Soubeyrand et al. (Eur. J. Biochem. 271) Ó FEBS 2004 Acknowledgements We are grateful to Dr Lambert (National Institutes of Health, Bethesda, Maryland) for providing plasmids encoding p53 wt ,p53 S15A , p53 S37A and p53 S15A/S37A mutants as GST fusion p roteins. This work was supported by a grant from the Canadian I nstitutes for Health Research to RJGH. SS was supported by a fellowship from Canadian Institutes for Health R esearch while RJGH is a n Investigator of the Canadian Institutes for Health Research. References 1. Smith, G.C. & Jackson, S.P. (1999) The DNA-dependent protein kinase. Genes Dev. 13, 916–934. 2. Abraham, R.T. (1996) Phosphatidylinositol 3-kinase related kin- ases. Curr. Opin. Immunol. 8, 412–418. 3. Gottlieb, T.M. & Jackson, S.P. (1993) The DNA-dependent protein kinase: requirement for DNA ends and association with Ku antigen. Cell 72, 131–142. 4. Finnie, N.J., Gottlieb, T.M., Blunt, T., Jeggo, P .A. & Jackson, S.P. (1995) DNA-dependent protein kinase activity is a bsent i n x rs-6 cells: implications for site-specific recombination and DNA dou- ble-strand break repair. Proc. Natl Acad. Sci. USA 92, 320–324. 5. Giffin, W., Torrance, H., Rodda, D.J., Pre ´ fontaine, G.G., Pope, L. & Hac he ´ , R.J. (1996) Sequence-specific DNA binding by Ku autoantigen and its effects on transcription. Nature 380, 265–268. 6. Anderson, C .W. (1993) DNA dam age and the DNA-activated protein kinase. Trends Biochem. Sci. 18, 433–437. 7.Kim,S.T.,Lim,D.S.,Canman,C.E.&Kastan,M.B.(1999) Substrate specificities and identification of putative substrates of ATM kinase family members. J. Biol. Chem. 274, 37538–37543. 8. Brooks, C.L. & Gu, W. (2003) Ubiquitination, phosphorylatio n and acetylation: the molecular basis for p53 regulation. Curr. Opin. Cell. Biol. 15, 164–171. 9. Craig,A.L.,Burch,L.,Vojtesek,B.,Mikutowska,J.,Thompson, A. & Hupp, T.R. (1999) Novel phospho rylation sites of human tumour suppressor protein p53 at Ser20 and Thr18 that disrupt the binding of mdm2 (mouse double min ute 2) protein are modified in human cancers. Biochem. J. 342, 133–141. 10.Dumaz,N.,Milne,D.M.,Jardine,L.J.&Meek,D.W.(2001) Critical roles for the serine 20, but not th e serine 15, phosphor- ylation s ite and for the po lyproline domain i n r egulating p53 turnover. Biochem. J. 359, 459–464. 11. Lambert, P.F., Kashanchi, F., Radonovich, M.F., Shiekhattar, R. & Brady, J.N. (1998) Phosphorylation of p53 serine 15 increases interaction with CBP. J. Biol. Chem. 273, 33048–33053. 12. Ryan, K.M., Phillips, A.C. & Vousden, K.H. (2001) Regulation and function of the p53 tumor su ppre ssor protein. Curr. Opin. Cell Biol. 13, 332–337. 13. Khanna, K.K. & Jackson, S.P. (2001) DNA double-strand breaks: signaling, repair and the cancer connection. Nat. Genet. 27 , 247–254. 14. Canman, C.E., Lim, D.S., Cimprich, K.A., Taya, Y., Tamai, K., Sakaguchi, K., Appella, E., Kastan, M.B. & Siliciano, J.D. (1998) Activation of the A TM kinase by ionizing radiat ion and ph os- phorylation of p53. Science 281, 1677–1679. 15. Gottifredi, V., Shieh, S., Taya, Y. & Prives, C. (2001) From the cover: p53 accumulates but is functionally imp aire d whe n D NA synthesis is blocked. Proc. Natl Acad. Sci. USA 98, 1036–1041. 16. Woo, R.A., Jack, M.T., Xu, Y., Burma, S., Chen, D.J. & Le e, P.W. (2002) DNA damage-induced apoptosis requires the DNA- dependent protein kinase, and i s me diated by the latent population of p53. EMBO J. 21, 3000–3008. 17. Wang, S., Guo, M., Ouyang, H., Li, X., Cordon-Cardo, C., Kurimasa, A., Chen, D.J., Fuks, Z., Ling, C.C. & Li, G.C. (2000) The cat alytic su bunit o f DNA-dependent protein kinase selectively regulates p53-dependent apoptosis but not cell-cycle arrest. Proc. NatlAcad.Sci.USA97, 1584–1588. 18. Saintigny, Y. & Lopez, B.S. (2002) Homologous recombination induced by r eplication inhibition, is stimulated b y expression of mutant p53. Oncogene 21, 488–492. 19. Lee, S., Cavallo, L. & Griffith, J. (1997) Human p53 binds Hol- liday junctions strongly and facilitates their cleavage. J. Biol. Chem. 272, 7532–7539. 20. Willers,H.,McCarthy,E.E.,Wu,B.,Wunsch,H.,Tang,W., Taghian, D.G., Xia, F. & Powell, S.N. (2000) Dissociation of p53- mediated suppression of homologous recombination from G1/S cell cycle checkpoint control. Oncogene 19, 632–639. 21. Sengupta, S., Linke, S.P., Pedeux, R., Yang, Q., Farnsworth, J., Garfield, S.H., Valerie, K., Shay, J.W., Ellis, N.A., Wasylyk, B. & Harris, C.C. (2003) BLM helicase-dependent transport of p53 to sites of s talle d DN A r ep lication forks modulates homologous recombination. EMBO J. 22, 1210–1222. 22. Allen, C., Kurimasa, A., Brenneman, M.A., Chen, D.J. & Nickoloff, J.A. (2002) DNA-dependent protein kinase suppresses double-strand break-indu ced and spontaneo us homologo us recombination. Proc. N atl Acad. Sci. USA 99, 3758–3763. 23. Shao, R.G., Cao, C.X., Zhang, H., Kohn, K.W., Wold, M.S. & Pommier, Y. (1999) Replication-mediated DNA damage by camptothecin induces phosphorylation o f RPA by DNA-depen- dent pro tein kinase and dissociates RPA: DNA-PK complexes. EMBO J. 18, 1397–1406. 24. Akyuz, N., Boehden, G.S., Susse, S., Rimek, A., Preuss, U., Scheidtmann, K.H. & Wiesmuller, L. (2002) DNA substrate dependence of p53-mediated regulation of double-strand break repair. Mol. Cell Biol. 22, 6306–6317. 25. Lee,S.E.,Mitchell,R.A.,Cheng,A.&Hendrickson,E.A.(1997) Evidence for DNA-PK-depende nt and -indepen dent DNA dou- ble-strandbreakrepairpathwaysinmammaliancellsasafunction of the cell cycle. Mol. Cell Biol. 17, 1425–1433. 26. Bakalkin, G., Selivanova, G., Yakovleva, T., Kiseleva, E., Kashuba, E., Magnusson, K.P., Szekely, L., Klein, G., Terenius, L. & Wiman, K.G. (1995) p53 binds single-stranded DNA ends through the C-terminal d omain and internal DNA segments via t he m iddle doma in. N ucleic A cids Res. 23, 362– 369. 27. Selivanova, G., Iotsova, V., Kiseleva, E., Strom, M., Bakalkin, G., Grafstrom, R.C. & Wiman, K.G. (1996) The single-stranded DNA end binding site of p53 coincides with the C-terminal reg- ulatory region. Nucleic Acids Res. 24, 3560–3567. 28. Plumb, M.A., Smith, G.C., Cunniffe, S.M., Jackson, S.P. & O’Neill, P. (1999) DNA-PK activation by i onizing radiation- induced DNA single-strand breaks. Int. J. Radiat. Biol. 75, 553– 561. 29. Hammarsten, O., DeFazio, L.G. & Chu, G. (2000) Activation of DNA-dependent protein kinase b y single-stranded D NA ends. J. Biol. Chem. 275, 1541–1550. 30. Soubeyrand, S., Torrance, H., Giffin, W., Gong, W., Schild- Poulter, C. & Hache, R.J. (2001) Activation and autoregulation of DNA-PK from structured single-stranded DNA an d coding end hairpins. Proc. Natl Acad. Sci. USA 98, 9605–9610. 31. Liu, Y. & Kulesz-Martin, M. (2001) p53 protein at the hub of cellular DNA damage response pathways through sequence-spe- cific and non-seque nce-specific DNA binding. Carcinogenesis 22, 851–860. 32. Zotchev, S.B., P rotopopova, M. & Selivanova, G . (2000) p53 C-terminal interaction with DNA ends and gaps has opp osing effect on specific DNA binding by the core. Nucleic Acids Res. 28, 4005–4012. 33. Wolcke, J., Reimann, M., Klumpp,M.,Gohler,T.,Kim,E.& Deppert, W. (2003) Analysis of p53 Ôla te ncyÕ and ÔactivationÕ by fluorescence correlation spectroscopy: evidence for d ifferent Ó FEBS 2004 Specificity of DNA-PK towards p53 (Eur. J. Biochem. 271) 3783 modes of high a ffinity DNA binding. J. Biol. Chem. 278, 32587– 32595. 34.Dudenhoffer,C.,Kurth,M.,Janus,F.,Deppert,W.&Wies- muller, L. (1999) Dissociation of the recombination control and the sequence-sp ecific tra nsactivation fun ction of P53. Oncogene 18, 5773–5784. 35. Soubeyrand, S., Pope, L., Pakuts, B. & Hache, R.J. (2003) Threonines 2638/2647 in DNA-PK are essential for cellular resistance to ionizing radiation. Cancer Res. 63, 1198–1201. 36. Lees-Miller, S.P., Sakaguchi, K., Ullrich, S.J., Appella, E. & Anderson, C .W. (1992) Human DNA-activated protein kinase phosphorylates serine s 1 5 and 37 in the amino-terminal transac- tivation domain of human p53. Mol. Cell Biol. 12, 5041–5049. 37. Griffin, R., Calvert, H., Curtin, N., Durkacz, B., Golding, B ., Hardcastle, I., Leahy, J., Martin, N., Newell, D., Rigoreau, L., Smith, G., Stockley, M., Veuger, S. & Hickson, I. (2002) Struc- ture–activity relationships and cellular activity of chromenone and pyrimidoisoquinolinone inhibitors o f DNA-dependent protein kinase (DNA-PK). Proc. Am. Assoc. Cancer Res. 43, 849. 38. Sarkaria, J.N., Tibbetts, R.S., Busby, E.C., Kennedy, A.P., Hill, D.E. & A braham, R .T . (1998) Inh i bition of phosphoinositide 3-kinase related kinases by the radiosensitizing agent wortmannin. Cancer Res. 58, 4375–4382. 39. Hall-Jackson, C.A., Cross, D.A., Morrice, N. & Smythe, C. (1999) ATR is a caffeine-sensitive, DNA-activated protein kinase with a substrate s pecificity distinct from DNA-PK. Oncogene 18, 6707– 6713. 40. Banin, S., Moyal, L., Shieh, S., Taya, Y., Anderson, C.W., Chessa, L., Smorodinsky, N.I., Prives, C., Reiss, Y., Shiloh, Y. & Ziv, Y. (1998) Enhanced phosphorylation of p53 by ATM in response to DNA damage . Science 281, 1674–1677. 41. Achanta, G., Pelicano, H., Feng, L., Plunkett, W. & Huang, P. (2001)Interactionofp53andDNA-PKinresponsetonucleoside analogues: potential role as a sensor complex for DNA damage. Cancer Res. 61, 8723–8729. 42. Watanabe, F., Teraoka, H., Iijima, S., Mimori, T. & Tsukada, K. (1994) Molecular properties, substrate specificity and regulation of DNA-dependent protein kinase from Raji Burkitt’s lymphoma cells. Biochim. Biophys. Acta 1223, 255–260. 43. Anderson, C.W. & Lees-Miller, S.P. (1992) The nuclear serine/ threonine pro tein kinase DNA-PK. Cr it. Rev. Eukar yot. Gene Expr. 2, 283–314. 44. Sakaguchi, K., Saito, S.I., Higashimoto, Y., Roy, S., Anderson, C.W. & Appella, E. (2000) Damage-mediated phosphorylation of human p53 threonine 18 through a cascade mediated by a casein 1-like kinase: effect on Mdm2 binding. J. Biol. Chem. 27 5, 9278– 9283. 45. Dumaz, N., Milne, D.M. & Meek, D.W. (1999) Protein k inase CK1 is a p53-threonine 18 k inase whic h req uires prior phos- phorylation of serine 15. FEB S Lett. 463, 312–316. 46. Mashhoon, N., DeMaggio, A.J., Tereshko, V., Bergmeier, S.C., Egli, M., Hoekstra, M.F. & Kuret, J. (2000) Crystal structure of a conformation-selective casein kinase-1 inhibitor. J. Biol. Chem. 275, 20052–20060. 47. Shieh, S.Y., Ahn, J., Tamai, K., Taya, Y. & Prives, C. (2000) The human homologs of checkpoin t kinases Chk1 and Cds1 (Chk2) phosphorylate p53 at multiple DNA damage-inducible sites. Genes Dev. 14, 289–300. 48. Goudelock, D.M., Jiang, K., Pereira, E., Russell, B. & Sanchez, Y. (2003) Regulatory interactions between the checkpoint kinase Chk1 and t he proteins of t he D NA-dependent protein kinase complex . J. Biol. Chem. 278, 29940–29947. 49. Bean, L .J. & Stark, G.R. (2002) Regulation of the accumulation and function of p53 by phosphorylation of two residues within the domain that binds to Mdm2. J. Biol. Chem. 277, 1864– 1871. 50. Jallepalli, P.V., Lengauer, C., V ogelstein, B. & Bunz, F. (2003) The Chk2tumorsuppressorisnotrequiredforp53responsesin human cancer cells. J. Biol. Chem. 278, 20475–20479. 51. Ahn,J.,Urist,M.&Prives,C.(2003)Questioningtheroleof checkpoint kinase 2 in the p53 DN A damage response. J. Biol. Chem. 278, 20480–20489. 52. Espejel, S., Franco, S., Sgura, A., Gae, D., Bailey, S.M., Taccioli, G.E. & Blasco, M.A. (2002) Functional interaction between DNA-PKcs and telomerase in telomere length maintenance. EMBO J. 21, 6275–6287. 53. Reckmann,B.,Grosse,F.,Urbanke,C.,Frank,R.,Blocker,H.& Krauss, G. (1985) Analysis of secondary structures in M13m p8 (+) single-strande d DNA by the pausing of DNA polymerase alpha. Eur. J. Biochem. 152, 633–643. 54. Sluss, H.K., Armata, H., Gallant, J. & Jones, S.N. (2004) Phos- phorylation o f serine 18 r egulates distinct p53 fu nctions in mice. Mol. Cell Biol. 24 , 976–984. 3784 S. Soubeyrand et al. (Eur. J. Biochem. 271) Ó FEBS 2004 . Structured DNA promotes phosphorylation of p53 by DNA- dependent protein kinase at serine 9 and threonine 18 Se ´ bastien Soubeyrand 1 , Caroline Schild-Poulter 1 and Robert J. G of p53 Wt or p53 S9A/T18V phosphorylation by DNA- PK . DNA- PK was incubated with the indicated p53 species in the presence of 50 l M ATP and assessed for total phosphorylation (top) or pSer9 phosphorylation. for DNA- PK in the modulation of p53 activity resultant from the convergence of p53 and DNA- PK on structured DNA. Keywords: DNA- dependent protein kinase; p53; structured single-stranded DNA; phosphorylation. The