HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 18, 16102-16115, May 3, 2002

This Article Abstract Full Text (PDF) An addition or correction has been published All Versions of this Article: 277/18/16102    most recent M111705200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by O'Neill, T. Articles by Rathbun, G. A. Search for Related Content PubMed PubMed Citation Articles by O'Neill, T. Articles by Rathbun, G. A. Social Bookmarking                      What's this?

Determination of Substrate Motifs for Human Chk1 and hCds1/Chk2 by the Oriented Peptide Library Approach^*

Ted O'Neill, Lauren Giarratani, Ping Chen^§, Lakshmanan Iyer^¶, Chang-Hun Lee^**, Matthew Bobiak, Fumihiko Kanai^§§, Bin-Bing Zhou, Jay H. Chung, and Gary A. Rathbun^¶¶

From the  Center for Blood Research, Department of Pediatrics, Children's Hospital, Harvard Medical School, Boston, Massachusetts 02115, ^§ Agouron Pharmaceuticals, Inc., San Diego, California 92121, ^¶ Research Computing Center, Harvard Medical School, Boston, Massachusetts 02115,  Laboratory of Biochemical Genetics, HLBI, National Institutes of Health, Bethesda, Maryland 20892, the  Department of Oncology Research, GlaxoSmithKline, King of Prussia, Pennsylavania 19406, and the ^§§ Department of Medicine, Beth Israel Deaconess Hospital, Harvard Medical School, Boston, Massachusetts 02115

Received for publication, December 7, 2001, and in revised form, January 24, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mammalian Chk1 and Chk2 are two Ser/Thr effector kinases that play critical roles in DNA damage-activated cell cycle checkpoint signaling pathways downstream of ataxia telangiectasia-mutated and ataxia telangiectasia-related. Endogenous substrates have been identified for human hCds1/Chk2 and Chk1; however, the sequences surrounding the substrate residues appear unrelated, and consensus substrate motifs for the two Ser/Thr kinases remain unknown. We have utilized peptide library analyses to develop specific, highly preferred substrate motifs for hCds1/Chk2 and Chk1. The optimal motifs are similar for both kinases and most closely resemble the previously identified Chk1 and hCds1/Chk2 substrate target sequences in Cdc25C and Cdc25A, the regulation of which plays an important role in S and G₂M arrest. Essential residues required for the definition of the optimal motifs were also identified. Utilization of the peptides to assay the substrate specificities and catalytic activities of Chk1 and hCds1/Chk2 revealed substantial differences between the two Ser/Thr kinases. Structural modeling analyses of the peptides into the Chk1 catalytic cleft were consistent with Chk1 kinase assays defining substrate suitability. The library-derived substrate preferences were applied in a genome-wide search program, revealing novel targets that might serve as substrates for hCds1/Chk2 or Chk1 kinase activity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

In the presence of DNA damage or incomplete DNA replication, eukaryotic cells activate cell cycle checkpoints that temporarily halt the cell cycle to permit DNA repair or completion of DNA replication to take place. In the presence of extensive damage or absence of timely repair, these checkpoint signaling pathways may also trigger a pathway that effects programmed cell death or apoptosis (reviewed in Refs. 1 and 2). DNA damage-activated cell cycle checkpoints are regulated in part by the phosphoinositide kinase family of checkpoint components, including the yeast Rad3 in Schizosaccharomyces pombe (3-5), Mec1/Tel1 in Saccharomyces cerevisiae (6, 7), mammalian ataxia telangiectasia-mutated (ATM)¹ (8), ATM/Rad3-related (ATR) (9), MEI-41 in Drosophila (10, 11), and X-ATM and X-ATR in Xenopus (12). These checkpoint kinases regulate the activities of two downstream effector serine/threonine kinases, Cds1 and Chk1, that are evolutionarily conserved. The Cds1 family includes conserved representatives from the yeasts (Cds1 in fission yeast and Rad53 in budding yeast) to man (hCds1/Chk2, hereafter termed Chk2) (13-21). This family of kinases is characterized by the presence of an N-terminal SQ-TQ cluster, followed by a Forkhead-associated domain and a C-terminal kinase domain. The Chk1 kinase family (called grp in Drosophila) contains an N-terminal kinase domain, C-terminal SQ cluster, and a TRF "domain" (see Ref. 22 and reviewed in Ref. 23). ATM and ATR phosphorylation of certain residues within the TQ or SQ sequences modulates mammalian Chk2 and Chk1 kinase activity, respectively (23).

Analyses utilizing the yeast systems have been critical in understanding the roles of Chk1 and Chk2 in DNA damage responses. In S. pombe, Chk1 is the effector of the G₂M DNA damage checkpoint pathway and is downstream of the Rad3 kinase (24), whereas in S. cerevisiae, Chk1 is downstream of Mec1 (25). One of the best defined and well characterized targets of Chk1 is Cdc25C, the dual specificity tyrosine/threonine phosphatase that dephosphorylates the mitotic inducer Cdc2 on Thr¹⁴ and Tyr¹⁵, resulting in Cdc2 Ser/Thr kinase activity that drives mitosis (23). Chk1 phosphorylation of Cdc25C promotes the binding of a 14-3-3 protein, which sequesters Cdc25C in the cytoplasm (26-30). Cds1 is the effector of the replication checkpoint pathway and is required for cells to survive treatments that block replication, such as hydroxyurea (31, 32). Cds1 appears to arrest cells in G₂ by phosphorylating Cdc25C on amino acid residues also targeted by Chk1 (29, 30). In the absence of Cds1, Chk1 can sometimes act to impose a checkpoint delay (26, 31, 33) indicating the existence of a partial functional overlap between the two kinases. In addition to inactivating Cdc25C, Cds1 is essential for transcriptional up-regulation of Mik1, a tyrosine kinase that phosphorylates and inhibits Cdc2 in replication checkpoint-arrested cells (31, 34).

The Chk1 and Chk2 Ser/Thr kinases also play important roles in cell cycle checkpoint signaling pathways in higher organisms (20, 35). In Xenopus and Drosophila, Chk1 functions not only in a checkpoint triggered by UV-damaged DNA but also in an S phase checkpoint triggered by a replication blockade (36-38). Cds1/Rad53/Chk2 homologues, termed CeCds1, CeCds2, and Dmnk/Chk2, respectively, have been identified in Caenorhabditis elegans and Drosophila (20, 39, 40). In C. elegans, CeCds1 and CeCds2 are required for meiotic recombination; loss of Dmnk/Chk2 in Drosophila results in ionizing radiation sensitivity, resistance to ionizing radiation-induced apoptosis, and genetic instability (41).

Mammalian Chk2 is activated by ATM phosphorylation of Thr⁶⁸ in response to DNA double strand breaks (42-44), and this phosphorylation event is required for Chk2 activation loop autophosphorylation necessary for Chk2 kinase activity (45). Vertebrate Chk1 is activated by ATR-mediated phosphorylation of several Ser residues in a C-terminal SQ cluster including Ser³¹⁷ and Ser³⁴⁵ (35, 36, 46). Targeted mutation Chk2 and Chk1 in mice have demonstrated profound differences in cellular requirements for the two effector kinases. Both Chk1/ and Chk2/ ES cells exhibit defective G₂/M DNA damage checkpoint function (35, 47, 48). However, Chk1 is an essential gene as targeted mutation of Chk1 results in an early embryonic lethal phenotype (35, 48). Chk2 loss of function has not been characterized with regard to the effects of germ line loss, rather only in the context of certain effects on ES cells and partial effects on T cell development in a Rag1/ blastocyst complementation analysis. In this context, Chk2 functional loss is not ES cell or T cell lethal (47). Chk2/ T cells fail to stabilize p53 after ionizing radiation and appear resistant to radiation-induced apoptosis (47). Loss of Chk1 function is ES cell lethal, and during development, Chk1 null cells appear to die of spontaneous apoptosis.

Chk2 phosphorylates Ser⁹⁸⁸ of Brca-1 (49) and Ser²⁰ of p53 (50, 51) in response to double strand DNA breaks. Chk1 also phosphorylates Ser²⁰ of p53 in vitro (47, 50, 51). Phosphorylation of Brca-1 by Chk2 in response to DNA damage is required for survival after DNA damage. Phosphorylation of p53 Ser²⁰ blocks the ability of Mdm2 to complex with p53 and shunt the latter into a degradation pathway, allowing the G₁ checkpoint to be activated by p53 (52). In vitro, Chk1 and Chk2 both phosphorylate Ser²¹⁶ of Cdc25C. Chk2 also phosphorylates Ser¹²³ in Cdc25A after double-stranded DNA breaks in an ATM-dependent manner resulting in a replication stage S phase cell cycle checkpoint (53). Chk1 has been shown to phosphorylate Cdc25A, Cdc25B, and Cdc25C in vitro (54).

Whereas the substrate motifs in Cdc25C and Cdc25A are similar, the motifs of other endogenous substrates targeted by Chk2 and Chk1 appear unrelated (see Table I). Thus, the two Ser/Thr kinases are considered "versatile" protein kinases with a potentially wide range of acceptable substrate targets (51). We have utilized an oriented, degenerate peptide library approach to determine clear, unambiguous substrate specificities for human Chk2 and Chk1. Library-derived optimal peptides were compared with peptides representing the endogenous substrates in terms of Chk2 and Chk1 phosphorylation activities and utilized as probes to show both strong similarities as well as striking differences between the two Ser/Thr kinases. The substrate motifs established from peptide library analyses were also used in a genome-wide search program in an attempt to identify additional potential targets that might serve as novel substrates for Chk2 or Chk1.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mammalian C1 and Chk1 Recombinant Protein-- Full-length Chk1 was subcloned into baculovirus expression vector pFASTBAC with glutathione S-transferase (GST) fused to N terminus of Chk1 via a linker containing a thrombin cleavage site. Spodoptera frugiperda (Sf9) cells expressing GST-Chk1 were scaled up, harvested, and then frozen until purification. To purify Chk1, a frozen cell pellet was resuspended on ice in lysis buffer containing 50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 1 mM DTT, 0.1% Brij, a protease inhibitor mixture (2 µg/ml E64, 1 mM 4-(2-aminoethyl)benzenesulfonyl fluoride, and 1 µg/ml pepstatin A), and 1 mM sodium orthovanadate, homogenized with a Tekmar tissuemizer, and disrupted through Microfluid fluidizer (M110-Y). The extracts were centrifuged at 100,000 × g for 30 min. The supernatant was added to glutathione-Sepharose 4B (Amersham Biosciences AB) beads equilibrated in wash buffer (20 mM Tris-HCl, pH 7.0, 10 mM MgCl₂, 100 mM NaCl, 1 mM DTT, protease mixture) and rocked at 4 °C for 30 min. The suspension was transferred into a column and allowed to pack; the wash buffer was then applied extensively. The GST-Chk1 was eluted from the column with 10 mM glutathione in 50 mM Tris-HCl, pH 8.0, collected, and dialyzed into 25 mM HEPES, pH 7.5, 50 mM KCl, and 5 mM DTT. GST-Chk1 sample was further purified by Superdex 200 (equilibrated with the dialysis buffer), at 4 °C, and more than 90% purity was estimated based on SDS-PAGE and silver staining analysis (Silver Stain Plus; Bio-Rad).

Purification of Strep-tagged Chk2 (ST-Chk2)-- Recombinant human Chk2 was produced in baculovirus using a MaxBac2.0 transfection kit (Invitrogen) as described previously (18). ST-Chk2 protein from the lysate was then purified using a StrepTactin-Sepharose column (Genosys Biotechnologies, Inc.). Briefly, the lysate (~15 ml) was applied to the column (1 ml) in buffer containing 100 mM Tris-HCl, pH 8.0, 1 mM EDTA. After washing the column five times with the same buffer, Strep-tagged Cds1 was eluted with a buffer containing 2.5 mM desthiobiotin, 100 mM Tris-HCl, pH 8.0, 1 mM EDTA. Purity of the eluate was assessed by silver staining using Silver Stain Plus (Bio-Rad).

Peptide Libraries-- Peptide libraries utilized in this analysis to derive consensus substrate motifs for Cds1 and Chk1 were designated as follows: RXXS, MAXXXXRXXSXXXXAKKK; SI, MAXXXXXSIAKKK; SF, MAXXXXXSFXXXXAKKK; SQ, MAXXXXXXSQXXXXAKKK; SP, MAXXXXSPXXXXAKKK; 4S4+ or 4S4, MAXXXXSAKKK; 4T4+ or 4T4, MAXXXXTXXXXAKKK; 4Y4+ or 4Y4, MAXXXXYXXXXAKKK. Degenerate positions in the sequences are represented by X in which all amino acids were represented with the exception of Cys; + or  indicates the presence or absence of Tyr, Ser, and Thr at the degenerate positions. To assay peptide libraries with maximum degeneracy at non-fixed positions, and thus provide the truest preference values, we analyzed libraries that included Ser, Thr, and Tyr and compared the results to those with only the fixed target residue for phosphorylation by Chk1 and Chk2. We saw no striking differences between the two sets of libraries with fixed central Ser or Thr (see for example, Fig. 1B, and data not shown). Inclusion of Ser and Thr in the fixed Tyr library enhanced the phosphorylation of the 4Y4 library by Chk1 and Chk2 (data not shown) indicating preferences for aromatic hydrophobic residues N- or C-terminal to phosphorylated Ser or Thr.

Chk2 and Chk1 Kinase Assays-- The reaction buffer for Chk2 consisted of a final concentration of 10 mM HEPES, pH 7.5, 75 mM KCl, 10 mM MgCl₂ 0.5 mM EDTA, 1.25 mM DTT, 100 µM cold ATP, 10 µCi of [-³²P]ATP, 100 ng of recombinant human Chk2 in a 30-µl total reaction volume at 30 °C. Small scale peptide library analyses were conducted utilizing 100 µg of each library. To assay Chk1 phosphorylation activity, the reaction buffer utilized consisted of 20 mM HEPES, pH 7.5, 50 mM KCl,10 mM MgCl₂, 0.1 mM EGTA, 100 µM cold ATP, 10 µCi of [-³²P]ATP, 50-100 ng of Chk1 in a total volume of 30 µl at 30 °C. To monitor peptide phosphorylation by Chk2 or Chk1, 2 µl of the reaction mix was transferred to an Eppendorf tube containing an equal volume of 30% acetic acid to halt the reaction, and 2.5-3.0 µl were spotted onto to P-81 phosphocellulose paper squares and allowed to air dry. These samples were washed three times in 1% o-phosphoric acid (5 min per wash), placed in scintillation vials, and counted. Peptide phosphorylation experiments including kinetic assays and relative velocities of Chk2 and Chk1 were reproduced multiple times at several concentrations.

Large Scale Peptide Library Analyses-- Large scale analyses to determine preferred substrate motifs for Chk2 and Chk1 were performed using the RXXS, SF, SI, and SQ peptide libraries. Large scale peptide library assays were scaled up 10-fold and repeated at least twice for each library using the methods described previously (55, 56). Approximately 1% of each library was phosphorylated in the appropriate reaction conditions in the presence of 300 µM ATP and 0.33 µCi of [-³²P]ATP; the kinase reaction was terminated by the addition of acetic acid, and the reaction was desalted and partially purified using a 1 ml of DEAE-Sephacel (Sigma) column. Radiolabeled fractions were pooled, lyophilized, reconstituted in distilled H₂O, and applied to a ferric chelation nitrilotriacetic acid-agarose column as described previously (56, 57). Eluted phosphopeptides were pooled, lyophilized, reconstituted in distilled H₂O, and sequenced. To determine amino acid preferences at each position in the sequenced peptides, each cycle was first internally normalized for all amino acids present at that cycle using a program designed to analyze relative amino acid abundance at each cycle (56).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Mammalian Chk1 and Chk2 Preferentially Phosphorylate the RXXS Fixed Degenerate Peptide Library-- To establish whether Chk1 and Chk2 exhibit preferences for distinct substrate motifs, we assayed phosphorylation of several different fixed, degenerate peptide libraries by the two Ser/Thr kinases. We included the most degenerate libraries fixed at only a central Ser, Thr, or Tyr (i.e. 4S4, 4T4 and 4Y4, respectively) as part of an unbiased survey of Chk1 and Chk2 target phosphorylation preferences. Fig. 1A shows that Chk2 prefers the RXXS peptide library that contains a fixed Arg at the 3-position (N-terminal) relative to the fixed Ser targeted by Chk2 kinase activity (Fig. 1A). Ser appears more highly selected than Thr as a target for phosphorylation by Chk2 in the context of the peptide libraries. A second tier of strongly selected libraries is that with hydrophobic residues (Phe and Ile) fixed at the +1-position (C-terminal) relative to Ser, as well as a degenerate library with a central SQ sequence (Fig. 1A). A library with Pro at +1 relative to Ser was not significantly phosphorylated by Chk2 (Fig. 1A).

View larger version (30K): [in this window] [in a new window]   Fig. 1.   Phosphorylation of several fixed, degenerate peptide libraries reveal similarities and differences in library substrate preferences of Chk2 (A) and Chk1 (B). Peptide library compositions are listed under "Experimental Procedures." Depicted are representative experiments of small scale analyses in which 100 µg of each library was added to a Chk2 or Chk1 kinase reaction in the presence of radioisotope and assayed for phosphorylation at the indicated time points. Both Chk2 and Chk1 prefer the Ser and RXXS libraries but differ in their comparative phosphorylation of other peptide libraries.

Chk1 Exhibits Similarities and Significant Differences to Chk2 Regarding Peptide Library Preferences-- Consistent with Chk2, Chk1 also selected the RXXS library but diverged from Chk2 in that the SI peptide library appeared equivalently phosphorylated compared with the RXXS library (Fig. 1B). The SF peptide library is also highly preferred; therefore, similar to Chk2, Chk1 also selects a hydrophobic residue at the +1-position. The SQ library (Fig. 1B) is modestly selected by Chk1; similar to Chk2, the SP library is less preferred and Chk1 also appeared to prefer Ser over Thr as the residue targeted for phosphorylation. Our results indicate that the 3- and +1-positions relative to a Ser targeted for phosphorylation appear to constitute a minimum selected motif for both Chk1 and Chk2.

Fig. 2, A and B, shows that the kinase-inactive (KI) versions of Chk2 and Chk1 failed to phosphorylate the RXXS peptide library, indicating that phosphorylation was carried out by the catalytic site of Chk1 and Chk2 and not that of an associated enzyme. Chk2 KI continued to express an apparent low level autophosphorylation activity (data not shown). There was no detectable Ser/Thr autophosphorylation of Chk1 KI.

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Lack of KI Chk2 K249R (A) and Chk1 D130A (B) phosphorylation activity for preferred peptide libraries indicates specificity of catalytic activity by kinase-active (KA) versions of the two Ser/Thr kinases.

Selection of a Distinct Core Substrate Motif by Chk2-- The results shown in Fig. 1, A and B, indicated that we could utilize data from several unrelated peptide libraries in large scale analyses to develop optimal substrate motifs for Chk2 and Chk1. We assayed the most highly selected RXXS library to determine whether there was a strong selection for residues at positions in addition to the fixed Arg at position 3 as well as to confirm the predilection for hydrophobic residues at the C-terminal +1-position adjacent to Ser. To verify the strong preference of Chk1 and Chk2 for Arg or a basic residue at position 3, we employed peptide libraries fixed at the +1-position but degenerate at 3. For these latter analyses we assayed the SI, SF, and SQ libraries. An amino acid was considered strongly selected or preferred at a given cycle if its normalized value was greater than 1.00 (base line) (shown in Tables II and III).

Results derived from large scale peptide library analyses for Chk2 established consistently strong preferences at the 7-, 5-, 3-, and +1-positions relative to the phosphorylated serine. At position 7, both hydrophobic and basic residues (Phe, Met, and Arg) were highly selected indicating a predisposition for charged residues with bulky side chains at that position. All four peptide libraries showed a strong selection for hydrophobic aliphatic residues, in particular, Leu and Ile, at the 5-position. A strong preference for arginine at position 3 was clearly evident in the SI, SF, and SQ libraries (Table II) thus confirming initial peptide library screening.

An examination of the amino acid preferences C-terminal to the fixed Ser in the RXXS library reveals a strong selection for hydrophobic amino acids, particularly Phe and Ile at the +1-position when +1 was degenerate (Table II and data not shown). These preferences are consistent with the high level of phosphorylation observed in small scale SF and SI peptide library assays (Fig. 1A). The preferences observed at the +3- and +4-position are characterized by bulky amino acids (Table II). With the exception of the +1-position, the N-terminal positions appeared to demonstrate the strongest predisposition for consistent selection of a given residue. We conclude that the minimal consensus substrate of Chk2 as defined by the peptide library analyses consists of 7(Arg/Phe) 6(Arg/hydrophobic) 5(Leu/Ile) 4(Lys/Arg) 3(Arg) 2(Xaa) 1(Xaa) Ser +1(Phe/Ile) +2(Phe/Ile/Arg).

Definition of a Consensus Substrate Motif for Chk1-- Chk1 substrate motif preferences exhibited strong similarities to those of Chk2 as well clear differences (Table III). Similar to Chk2, a basic amino acid is strongly selected at the 3-position, with Arg demonstrating the highest preference values. Aliphatic hydrophobic amino acids, with Leu predominating, are highly preferred by Chk1 at position 5 as was the case for Chk2. We also found a preference for Arg at position 5 in addition to selection for hydrophobic residues at this site when the 3-position is fixed for Arg. Leucine and arginine appeared at 7 also consistent preferences at this position by Chk2. Analysis of the RXXS library revealed that optimal preferences for Chk1 at position +1 are hydrophobic residues (Phe and Met; Table III). Similar to Chk2, the +2-, +3-, and +4-positions were unremarkable for specific preferences. Taken together, the simplest consensus substrate for Chk1 is 7(Leu/Arg) 6(Xaa) 5(Leu/hydrophobic/Arg) 4(basic/Val) 3(Arg/Lys) 2(Tyr/Xaa) 1(Xaa) Ser +1(Phe/Met/hydrophobic).

Phosphorylation of Optimal Peptide Substrates and Peptides Representing Endogenous Targets by Chk1 and Chk2-- Based upon the results of amino acid substrate preferences summarized in Tables II and III, we generated predicted optimal peptide substrates (Table IV) for Chk1 and Chk2, designated Chk1-tide and Chk2-tide, respectively. Fig. 3A shows that the Chk-tides and Cdc25C are highly suited as substrates for Chk1 catalytic activity followed by modest phosphorylation of Brca-1. As substrates for Chk2 kinase activity, only the Chk-tides were optimal substrates. The Brca-1 and Cdc25C peptides were much less suited as substrates for Chk2 (Fig. 3B).

View this table: [in this window] [in a new window]   Table IV Peptide substrates Peptides represent specific endogenous substrate targets (Cdc25CSer216, Brca-1Ser988, and p53Ser20) of Chk2 and Chk1, consensus optimal peptides (termed Chk2-tide and Chk1-tide, respectively) derived from large scale library analyses, and the Chk2-tide peptide with positions L(5)A or R(3) substitutions. Single letter codes are used to represent amino acids. Lysines at the C-terminal ends of several peptides were added to ensure solubility. Composition of the Chk2-tide and Chk1-tide peptides was obtained from two sources. 1) The strongest preferences (e.g. positions 5, 3, and +1) were acquired from analysis of several peptide libraries (see Tables II and III). 2) For positions within the Chk1-tide and Chk2-tide peptides that exhibited only mild or modest preferences in the peptide libraries analyses, selection of a residue at a given position depended upon relative abundance of the amino acid as demonstrated by its increase from the previous cycle (data not shown).

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Chk2 and Chk1 differentially phosphorylate specific peptide substrates. Shown are representative phosphorylation assays of 50 µM peptide substrate by Chk2 (A) or 25 µM peptide substrate by Chk1 (B). p53 S20* indicates synthetic phospho-Ser15 version of the p53Ser20 peptide (see Table IV).

Ser²⁰ in p53 has been identified as a substrate target of Chk2 and Chk1. However, we found that the p53Ser20 peptide, which contains the amino acid sequence surrounding Ser²⁰ in p53, was not phosphorylated by either Chk kinase (Table IV; Fig. 3, A and B). Because Ser¹⁵ within the Leu-Ser¹⁵-Gln-Glu motif within p53 has been identified as a target for ATM kinase activity (56, 58, 59), we asked whether a phospho-Ser¹⁵ version of the p53Ser20 peptide might serve to prime the C-terminal Ser²⁰ site for subsequent Chk2/Chk1 phosphorylation (see Table IV). We found that presenting either kinase with the phospho-Ser¹⁵ p53Ser20 peptide resulted in no increase in phosphorylation of the peptide by either Chk2 or Chk1 (Fig. 3, A and B).

Because the 5- and 3-positions were highly selected as key residues in substrate preferences for Chk1 and Chk2 (Tables II and III), these sites were mutated to alanine in the Chk2-tide peptide to assess their importance in the definition of an optimal substrate. Fig. 4, A and B, shows that the L(5)A and R(3)A substitutions substantially diminished the suitability of the Chk2-tide peptide as a substrate target for both Ser/Thr kinases. The 3-position appeared the most critical of the two positions because the presence of Ala at this position abrogated the ability of Chk2-tide to act as a substrate for the two kinases. In contrast to Chk2, the 5-position is significantly more important as a key residue in the Chk2-tide sequence for Chk1 as substitution of Ala for Leu at this position drastically reduced the ability of the L(5)A peptide to serve as a substrate for Chk1 (Fig. 4B). As little or no phosphorylation of the R(3)A and p53Ser20 peptides was observed, the two were not included in subsequent analyses.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Representative experiment showing that alanine substitutions at the highly selected 5- and 3-positions vary in their effects on Chk2-tide peptide substrate suitability for Chk2 (A) and Chk1 (B).

Analysis of Chk2 and Chk1 Catalytic Activities in Terms of Peptide Substrates-- Although there were substantial similarities in the requirements for a minimal optimal core substrate motif for Chk2 and Chk1 (Table IV; also see above), a more detailed comparison of Chk2 and Chk1 catalytic activities revealed striking differences between the two Ser/Thr kinases (Table V). Peptide substrate K_m values determined for Chk2 were relatively high, suggesting that K_m appears to play a much less important role in the overall definition of peptide substrate suitability. As was the case in previous studies of peptide substrates for basophilic kinases (i.e. kinases preferring substrates with a basic residue at position 3) such as AKT and several of the PKC isotypes (60), the V_max/K_m ratios appeared to reflect the most informative values regarding peptide substrate suitability for Chk2. The V_max/K_m values of Chk2-tide and Chk1-tide indicated that the library-generated peptides were equivalent as the two most highly suited substrates for Chk2 kinase activity. The L(5)A substitution resulted in a markedly poorer substrate as indicated by the 4.6-fold reduction in V_max/K_m value compared with "wild type" Chk2-tide (Table V). The peptide representing mammalian Cdc25C Ser²¹⁶ exhibited the lowest V_max and V_max/K_m values; this peptide showed a 16-fold loss when compared with Chk2-tide. Surprisingly, V_max/K_m ratios indicated that the Brca-1 peptide, with Pro at position 3, was ~3-fold better suited than Cdc25C as a substrate for Chk2. However, V_max/K_m values of Brca-1 indicated that this peptide was 5-6-fold less suitable compared with the library-derived peptides, probably because of the presence of Pro instead of Arg at the critical 3-position. In the context of the Chk2-tide substrate, Ala substituted for Arg at 3 appears to be a more severe substitution for both Chk2 and Chk1 as it effectively abolishes the ability of this peptide to serve as a substrate target. These results suggest that the Chk2-tide sequence, in contrast to that of the Brca-1 peptide, lacks the capability to compensate for the R(3)A substitution.

Chk1 exhibits significant differences to Chk2 in that the K_m values for the peptides are low and generally more reflective of the differences in peptide substrate suitability (Table V). We found that distinct from Chk2 in which Chk2-tide and Chk1-tide demonstrate comparable V_max/K_m values, the Chk1-tide peptide is 3-fold better suited as a substrate for Chk1. A major difference is that the Cdc25C peptide is much better suited as a substrate for Chk1 than Brca-1, as reflected in the equivalent V_max/K_m values for the Cdc25C and Chk2-tide peptide substrates for Chk1 (Table V). Similar to Chk2, the Brca-1 peptide is a better substrate for Chk1 than the L(5)A-substituted peptide. Our results indicate that for both Chk kinases, the 5-position and 3-positions together played critical roles as a driving force in optimal substrate selection from the degenerate peptide libraries. We found that consistent with previous studies (55, 61), peptides with high V_max/K_m ratios are selected from the peptide libraries by both Chk1 and Chk2.

For all tested peptides except one, Chk2 demonstrated higher relative velocities than Chk1 in terms of ATP turnover (Table VI) ranging from ~3.0- to 8.0-fold greater activity, depending on the peptide. The one interesting exception to the overall greater catalytic activity of Chk2 versus Chk1 occurred when the Cdc25C peptide was analyzed, providing further support for the better suitability of this sequence for Chk1. Together, the data suggest that although Chk2 and Chk1 select extremely similar consensus substrate motifs, the two Ser/Thr kinases vary substantially in both single peptide preferences as well in terms of catalytic activity.

Distinct from Chk2 but similar to the AKT peptide library analyses (61), Chk1, in addition to a selection for aliphatic hydrophobic amino acids at the 5-position, also demonstrated a preference for Arg at this position in the RXXS library (Table III). At position 7, Chk1 and Chk2 selected Leu, Phe, Met, and Arg. The potential motif of Arg-Xxx-Arg-Xxx-Arg-Xxx-Xxx-Ser (hydrophobic) as a substrate for Chk1 leaves open the intriguing possibility for AKT, PKC, and Chk1 to target shared downstream substrates and potentially convergent signaling pathways. Interestingly, a comparison of Chk1 and Chk2 phosphorylation of AKTide, the peptide sequence identified as an optimal substrate for AKT kinase activity (61), shows a striking specificity of this peptide to serve as a substrate for Chk1 versus minimal AKTide phosphorylation by Chk2 (Fig. 5). We are currently investigating whether a potential AKT-Chk1 link exists in vivo. AKTide joins the Cdc25C peptide and Chk1-tide as substrates that clearly distinguish Chk1 and Chk2 kinase activities.

View larger version (10K): [in this window] [in a new window]   Fig. 5.   The AKTide peptide substrate (25 µM) is specifically phosphorylated by Chk1 but not Chk2.

Utilization of the Peptide Library-derived Substrate Motifs for Chk2 and Chk1 to Identify Novel Targets in Data Base Searches-- The data in Tables II and III for Chk2 and Chk1 was used to develop preference matrices that were then applied to the SCANSITE genomic search program (62) to identify potential novel substrates of Chk2 and Chk1. Multiple possible substrates were selected in the data base searches for mouse, human (Table VII), and the yeasts (Table VIII). Some of the substrate targets appearing to have a role in regulation of developmental processes, DNA expression, or DNA repair are listed. Potential targets are organized from the highest match percentiles (i.e. candidate molecules containing sequences that theoretically best matched the peptide library-derived motifs) to the lowest, ranging from 0.002% for Cdc25C to 0.089% for Rad51 for Chk1 (Table VII). For Chk2, values ranged from 0.014% for the DMR-N9 protein and 0.067% for DCAMKL1 to less highly matched possible substrates (0.154% for Rag1). For Chk1, the Rag1 preference value was 0.054%, which is theoretically a better match.

The highest match percentile for Chk1 was Cdc25C at Ser²¹⁶ (0.002%) and Ser³⁰⁹ (0.020%) in Cdc25B. For Chk2, only Ser³⁰⁹ in Cdc25B was selected in the search analysis, and the match percentile for this sequence indicated that it was modest compared with other more highly selected Chk2 targets (data not shown). Lack of a strong match for Cdc25C Ser²¹⁶ in the SCANSITE data base analysis is in keeping with our experimental results showing that the Cdc25C peptide is a much stronger target substrate for Chk1 than Chk2. Noteworthy, however, is the recent finding that targeted mutation of Cdc25C creating a functionally null gene resulted in no abnormalities in mice indicating that the other Cdc25 members are apparently able to substitute for the roles occupied by Cdc25C (63). Cdc25A at Ser¹²³ has been recently identified as a downstream target of ATM and Chk2 in a replication checkpoint pathway (53). The preference values generated in the SCANSITE search indicate that Cdc25A is a better substrate for Chk2 than Chk1 (data not shown).

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The oriented peptide library approach has been utilized to derive consensus substrate motifs for Chk1 and Chk2, two Ser/Thr effector kinases that play critical but different downstream roles in DNA damage-activated cell cycle checkpoint signaling pathways. Because several independent peptide libraries, each containing over 10⁸ independent potential substrate targets, were analyzed for each kinase, we conclude that the substrate motifs selected in these analyses, which are similar for both, are compelling. For both Chk2 and Chk1, we observed striking preferences for certain amino acids at the 7-, 5-, 3-, and +1-positions relative to the phosphorylated Ser. The peptide substrate motif selected by both Ser/Thr kinases most strongly resembles the sequences containing Ser²¹⁶ in mammalian Cdc25C, Ser³⁰⁹ in Cdc25B, and Ser¹²³ in Cdc25A. The peptide library-derived motifs versus previously identified endogenous substrates and the differences in the catalytic activities of the two kinases raise interesting questions in terms of Chk1 and Chk2 regulation and function (discussed in more detail below). The selected substrate motifs define Chk2 and Chk1 as additional members within a growing list of basophilic Ser/Thr kinases important in various stress-related responses, including C-TAK1, the protein kinase C (PKC) family of isoenzymes, mammalian AMP-activated protein kinase, SNF1, calcium/calmodulin-dependent kinase, phosphorylase kinase, and AKT (protein kinase B). All have in common a strong preference for basic residues at the 3-position, particularly arginine, and hydrophobic residues at the +1-position relative to the phosphorylated Ser (60, 61, 64-66).

Chk1 and Chk2 are similar to PKC I and II, and PKCµ (60) in exhibiting a strong selection for Leu at 5. Substitution of Ala for Leu at position 5 resulted in a moderate loss of the ability of the optimal Chk2-tide peptide to serve as a substrate for Chk2 and a much more severe impact with regard to Chk1, reflecting a difference between Chk1 and Chk2. The impairment of substrate suitability resulting from the L(5)A substitution is supported by the high preference values for aliphatic hydrophobic residues observed in the peptide library analyses for both Ser/Thr kinases. Modeling analysis of a linear Cdc25C peptide into the Chk1 catalytic pocket indicates a strong potential for an important anchoring function served by hydrophobic residues at positions 5 and +1 (67) which extend toward the hydrophobic pockets H1 and H3, respectively (see Fig. 6A). The Cdc25C and peptide library-derived Chk1-tide and Chk2-tide peptides contain the appropriate residues at these crucial positions (Fig. 6, A-C). Fig. 6E shows that Ala at position 5 in the Chk2-tide peptide clearly appears insufficient as an anchor substitute for the H1 pocket.

View larger version (114K): [in this window] [in a new window]   Fig. 6.   Peptides are modeled into the catalytic cleft of the Chk1 kinase domain molecular surface using INSIGHT II (65). Peptides substrates are as follows: Cdc25C (A), Chk1-tide (B), Chk2-tide (C), Brca-1 (D), and the L(5)A/R(3)A peptide (representing a Chk2-tide peptide containing Ala substitutions at both the 5- and 3-positions) (E). The peptides, represented from the 5- to the +3-positions, are shown as stick models with non-polar, acidic, and basic side chains represented in violet, red, and blue, respectively. Atom type in the peptides are denoted by red, oxygen; green, carbon; blue, nitrogen; and yellow, sulfur. The peptides were docked into the Chk1 catalytic cleft based upon the structure of the ternary complex of PHK, AMP-PNP, and the MC-peptide substrate (67, 90, 91). Using this template as a model, Arg3 of the substrate may play a critical role in its interaction with Glu91 (depicted in green) of Chk1 by providing stabilization of the nearby ribose sugar of adenine, resulting in alignment of the -phosphoryl for catalytic transfer (67, 91). The catalytic Asp130 is indicated in orange. Surface hydrophobic pockets are indicated as follows (64): H1 (Phe93, Ile96, and Pro98), dark blue; H2 (Leu206, Thr170, and Pro172), red; and H3 (Leu171, Val174, Leu1178, Leu179, and Met167), yellow. Leu206 (magenta) is shared between H1 and H2. H1 and H3 are predicted as important regions that anchor substrate positions 5 and +1. Residues Ile4 and Phe2 in the Brca-1 peptide may provide a critical additional anchor by fitting into the H2 cluster.

We were surprised to find that the Brca-1 peptide appears 3-fold better suited than Cdc25C as a substrate for Chk2 because it lacked the critical basic residue at position 3. In our analyses, the 3-position is least forgiving in terms of a substitution in the context of the Chk2-tide peptide; the R(3)A mutation effectively abolished the ability of the peptide to serve as a substrate for either Chk kinase. Crystallographic analysis of the Chk1 kinase domain suggests that Arg³ in a Cdc25C substrate peptide is the only residue that can interact directly with Glu⁹¹ within the glycine-rich stretch of the N-terminal loop of the Chk1 kinase domain (67). Substitution of Ala at position 3 in the otherwise highly suitable Chk2-tide sequence results in loss of the essential role of the 3-position and a profound loss of the ability of the peptide to serve as a substrate for either Chk1 and Chk2 (Fig. 6D, Fig. 4, A and B).

At position 3 in the Brca-1 peptide, Pro appears better tolerated by both Chk1 and Chk2 suggesting that this residue may be a less severe replacement. However, although Brca-1 lacks Arg at position 3, it contains the strongly preferred Leu and Phe anchors at positions 5 and +1, respectively (Table IV and Fig. 6D). The 7-position in the Brca-1 sequence is also a Pro residue, consistent with peptide library preferences of Chk1 and Chk2 for a basic or an aliphatic hydrophobic residue at this position. Additionally, when the Brca-1 peptide was docked into the Chk1 catalytic pocket, we noted that Ile and Phe at positions 4 and 2, respectively, may provide an important added anchoring capacity in terms of the H2 cluster of hydrophobic residues represented by Thr¹⁷⁰, Pro¹⁷², and Leu²⁰⁶ (Fig. 6D). This extra anchor may provide sufficient added stability of the Brca-1 peptide to enable it to serve as a modest substrate. Notably, however, both Chk-tides and the Cdc25C peptide, all with Arg at position 3, fare significantly better as substrates than the Brca-1 peptide for Chk-1.

It is puzzling, however, that in the context of Chk2 catalytic activity, the Brca-1 peptide, with Pro at 3, serves as a better substrate than Cdc25C, the sequence of which appears to fulfill all the necessary requirements for a good substrate. Our results suggest that within the context of Chk2-tide, there is another position within a substrate sequence important for optimal Chk2 phosphorylation activity. One possible candidate is Glu at the +3-position in Cdc25C which may not be an ideal amino acid at this position for Chk2 kinase activity. In this regard, Chk2 may differ from Chk1 (the latter of which lacks an activation loop Thr motif) and resemble cAMP-dependent protein kinase, PKC, Cdc2, and CK1 in a strong preference for a hydrophobic or basic residue at the +3-position in a substrate (see Tables II-IV) (60, 66). Notably, the two Chk-tides, both optimal substrates for Chk2, both contain an aliphatic hydrophobic amino acid at the +3-position, and the Brca-1 peptide contains a lysine at this position. The +3-position may be important for substrate stabilization within the Chk2 catalytic cleft through interaction with phosphothreonine in the Chk2 activation loop (60, 66, 68). Alternatively, the ability of Cdc25C to serve as a highly preferred substrate for Chk2 phosphorylation may require modulation of Chk2 kinase activity by as yet undefined interacting components. Another possibility may be a required stabilization or structural modification of a seemingly optimal Cdc25C substrate sequence within the catalytic cleft of Chk2 that occurs in the context of the entire Cdc25C protein structure.

Critical residues for human Chk1 substrate specificity were mapped using a Xenopus Cdc25C Ser²⁸⁷ template peptide (the sequence of which is highly conserved with human Cdc25C Ser²¹⁶ sequence) to derive a "minimal" core Cdc25C peptide substrate sequence (65, 69, 70). In this study, aliphatic hydrophobic residues were preferred at the 5-position, whereas arginine and lysine were selected at position 3. Mutation of these positions to Ala substantially diminished the substrate suitability of the peptides for Chk1 kinase activity (70). These results are very similar to our findings using the peptide libraries and our analysis of the Leu⁵ and Arg³ positions in Chk2-tide. Interestingly, substrate efficiency of the Xenopus minimal Cdc25C peptide was diminished but not abrogated following an M(+1)A substitution (70). Our results showing that the SQ peptide library provided a modest substrate may likewise indicate that the requirement for a hydrophobic residue at +1 is not absolute (Fig. 2B and Table III). We conclude that in terms of relative importance for Chk1 and Chk2 substrate preferences, positions 3 > 5 +1.

The peptide library-derived substrate motifs and peptide modeling analyses provide rationales for Chk1 and Chk2 phosphorylation of the endogenous substrates Cdc25C, Cdc25A (and Cdc25B), and Brca-1. Our study fails to support phosphorylation of p53 Ser²⁰ by either Ser/Thr kinase. The disparity in our results versus other studies (50, 51) may be due to our utilization of peptide substrates incapable of forming higher order structures that result in the dimerization or tetramerization of p53. Such a configuration is apparently required to generate an adequate target substrate motif of the Ser²⁰ sequence in p53 for Chk2 and Chk1 (50, 51) coupled with a potential intrinsic versatility in the ability of the two effector kinases to phosphorylate a broad range of substrates (51). This wide ranging capacity was not immediately obvious in our analyses of Chk1 and Chk2 kinase activity although some modest flexibility was clearly evident as shown by phosphorylation of the Brca-1 peptide and SQ peptide library phosphorylation.

It is uncertain from our study how the structure of p53 at Ser²⁰ might be altered through oligomerization such that it develops a functional substrate for Chk2 and Chk1. Thus, other explanations for apparent Chk2- or Chk1-dependent p53 Ser²⁰ phosphorylation may also be relevant. For example, Chk2 and Chk1 may function upstream to activate a yet unidentified Ser/Thr kinase that specifically targets and phosphorylates Ser²⁰ in p53. S. pombe Chk1 has been identified in a complex with several other components (71, 72), and Xenopus Chk1 has been isolated in association with a component termed Claspin (73). It is conceivable that a given complex containing either Chk kinase might serve to alter substrate specificity and enhance the capability of Chk1 or Chk2 to target a motif that is normally much less than ideal. These and other complexes may also participate in regulating localization of Chk1 and Chk2, thus altering accessibility of two highly active Ser/Thr kinases and promoting an extended association with a less ideal substrate or dissociation from a substrate containing a more optimal motif.

Although the substrate motifs selected by Chk1 and Chk2 appear similar, a recent study determined that the anti-cancer agent UCN-01 blocked Chk1 activity resulting in loss of Cdc25C phosphorylation yet left the Chk2 pathway intact (74). This observation is consistent with our results showing that the single exception in which Chk1 catalytic activity is more robust than Chk2 involved phosphorylation of the Cdc25C peptide (Tables V and VI). Together, these results underscore the potentially essential positioning of the Cdc25 phosphatase family downstream of Chk1 kinase activity in replication phase or G₂M checkpoint activation.

The Li Fraumeni syndrome in humans, which presents with tumor development previously attributed to p53 functional loss, has in certain cases been recently assigned to a Chk2 haploinsufficiency (75). To date there are no reports of lymphoid cell tumor development in Rag1-complemented Chk2/ mice nor have there been studies yet published that characterize germ line functional loss of Chk2 in mice. Chk1, an essential gene for embryonic stem cells and critical for early embryonic viability (35, 48), may be able to partially compensate for loss of Chk2 in the context of T cell development and maintain differentiated cell viability on the basis of common recognition of overlapping substrate motifs. However, that tumor progression in humans occurs as a result of a Chk2 loss or diminishment of function indicates that not all roles of Chk2 can be rescued by Chk1.

The potential targets for mammalian Chk2 and Chk1 listed in Table VII consist of several potentially relevant proteins in the checkpoint response, including transcription factors, DNA repair molecules, and in one case, a V(D)J recombinase component. Phosphatases and kinases generally appearing to be important in development were also selected. There are intriguing possibilities among these potential substrates. For example, NEK2, a NIMA-related Ser/Thr kinase, may play an important role in G₂M transition in mitosis and may participate in meiosis as well. A second interesting Ser/Thr kinase selected from the Chk2 data base search is DCAMKL1 which contains homology to Doublecortin and is thought to play a role in neuronal migration and nervous system maturation (76). A third appealing candidate is PCTAIRE-1, a member of an enigmatic family of Cdc2-like Ser/Thr kinases consisting of two other related family members, which in some studies have been shown to interact with 14-3-3 proteins upon phosphorylation (77-80). PCTAIRE-1 is highly expressed in brain, and of particular interest are reports (81-83) of PCTAIRE-1 expression in terminally differentiated neuronal cells including Purkinje cells in rodents.

A striking and mysterious feature of ATM functional loss in humans is the ataxic phenotype arising from a neurodegenerative process involving not only Purkinje cell loss but also ectopic placement of these cells. Granular and basket cell loss has also been described in patients with ataxia telangiectasia (84). These cells may accumulate unrepaired DNA damage in the absence of ATM during a proliferative developmental phase, the consequences of which are only evident in a subsequent post-mitotic phase (85). The possibility of DCAMKL1 and PCTAIRE as potential downstream substrate targets of human Chk2 (or Chk1) again raises the simple question whether the neurodegenerative effects that result from loss of a phosphoinositide kinase family member primarily remarkable for its DNA damage-activated checkpoint function are also indirect. Substantial cytoplasmic ATM and Chk2 have been observed in neuronal cells (86-89). Loss of ATM kinase activity may at least in part be responsible for lack of appropriate activation of critical downstream effector kinase(s) whose primary role in nervous system development is not necessarily only related to a DNA damage-activated cell cycle checkpoint function.

A high degree of conservation exists between yeast and mammalian DNA damage-induced cell cycle checkpoint signaling pathways (23). In both, an ATM-related phosphoinositide kinase activates a homologous downstream effector kinase which then phosphorylates a conserved site in both yeast and mammalian Cdc25 homologues. In S. pombe, Cd25, Ser⁹⁹ within the sequence Arg-Thr-Leu-Phe-Arg-Ser-Leu-Ser⁹⁹-Cys-Thr-Val-Glu-Thr-Pro is specifically targeted for in vitro phosphorylation by both Cds1 and Chk1 (30); this sequence was among the top four percentile matches when the mammalian Chk1 and Ch2 substrate motifs were tested against fission yeast Cdc25 (data not shown). Therefore, we also probed yeast data bank sequences with the motifs determined for human Chk1 and Chk2. As was the case in the mammalian search, high match percentile matches in S. cerevisiae and S. pombe belonged to interacting factors with DNA and DNA repair components important in replication and meiosis (Table VIII). Collectively, therefore, the results underscore the evolutionary conservation and importance of the Chk1 and Chk2 substrate target motif: 5(hydrophobic/basic) 4(Xaa/basic) 3(Arg/basic) 2(Xaa) 1(Xaa) Ser/Thr +1(hydrophobic), as a critical module for communication in multiple stress response signaling cascades related to the basophilic family of Ser/Thr kinases.

	ACKNOWLEDGEMENTS

We are grateful to Dr. Lewis C. Cantley for critically reading the manuscript, thoughtful discussions, and assistance with the SCANSITE data base search program. We also thank Dr. Anna Russell for insights. We thank Patrick Lazorchak for assistance.

	FOOTNOTES

^* The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^** Present address: Dept. of Pathology, Yomsei University, College of Medicine, Seoul 120-752, Korea.

^¶¶ Supported by National Institutes of Health Grant GM57018. To whom correspondence should be addressed: Center for Blood Research, Dept. of Pediatrics, Children's Hospital, 200 Longwood Ave., Warren Alpert Bldg., Rm. 135, Harvard Medical School, Boston, MA 02115. Tel.: 617-278-3226, Fax: 617-278-3131; E-mail: grathbun@cbr.med.harvard.edu.

Published, JBC Papers in Press, January 30, 2002, DOI 10.1074/jbc.M111705200

	ABBREVIATIONS

The abbreviations used are: ATM, ataxia telangiectasia-mutated; ATR, ataxia telangiectasia-related; GST, glutathione S-transferase; DTT, dithiothreitol; KI, kinase-inactive; PKC, protein kinase C; ES, embryonic stem; AMP-PNP, adenosine 5'-(,-imino)triphosphate.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES