Ataxia telangiectasia-mutated phosphorylates Chk2 in vivo and in vitro.

Abstract

The protein kinase Chk2, the mammalian homolog of the budding yeast Rad53 and fission yeast Cds1 checkpoint kinases, is phosphorylated and activated in response to DNA damage by ionizing radiation (IR), UV irradiation, and replication blocks by hydroxyurea (HU). Phosphorylation and activation of Chk2 are ataxia telangiectasia-mutated (ATM) dependent in response to IR, whereas Chk2 phosphorylation is ATM-independent when cells are exposed to UV or HU. Here we show that in vitro, ATM phosphorylates the Ser-Gln/Thr-Gln (SQ/TQ) cluster domain (SCD) on Chk2, which contains seven SQ/TQ motifs, and Thr68 is the major in vitro phosphorylation site by ATM. ATM- and Rad3-related also phosphorylates Thr68 in addition to Thr26 and Ser50, which are not phosphorylated to a significant extent by ATM in vitro. In vivo, Thr68 is phosphorylated in an ATM-dependent manner in response to IR, but not in response to UV or HU. Substitution of Thr68 with Ala reduced the extent of phosphorylation and activation of Chk2 in response to IR, and mutation of all seven SQ/TQ motifs blocked all phosphorylation and activation of Chk2 after IR. These results suggest that in vivo, Chk2 is directly phosphorylated by ATM in response to IR and that Chk2 is regulated by phosphorylation of the SCD.

Full text

Page 1

Ataxia telangiectasia-mutated phosphorylates Chk2
in vivo and in vitro
Shuhei Matsuoka*, Galit Rotman†, Akira Ogawa‡, Yosef Shiloh†, Katsuyuki Tamai‡, and Stephen J. Elledge*§
*Howard Hughes Medical Institute, Verna and Marrs McLean Department of Biochemistry and Molecular Biology, Department of Molecular and Human
Genetics, Baylor College of Medicine, Houston, TX 77030;†Department of Human Genetics and Molecular Medicine, Sackler School of Medicine,
Tel Aviv University, Ramat Aviv 69978, Israel; and‡Medical and Biological Laboratories Co., Ina, Nagano 396-0002, Japan
Edited by Bert Vogelstein, Johns Hopkins Oncology Center, Baltimore, MD, and approved July 11, 2000 (received for review January 24, 2000)
The protein kinase Chk2, the mammalian homolog of the budding
yeast Rad53 and fission yeast Cds1 checkpoint kinases, is phos-
phorylated and activated in response to DNA damage by ionizing
radiation (IR), UV irradiation, and replication blocks by hydroxyu-
rea (HU). Phosphorylation and activation of Chk2 are ataxia telan-
giectasia-mutated (ATM) dependent in response to IR, whereas
Chk2 phosphorylation is ATM-independent when cells are exposed
to UV or HU. Here we show that in vitro, ATM phosphorylates the
Ser-Gln?Thr-Gln (SQ?TQ) cluster domain (SCD) on Chk2, which
contains seven SQ?TQ motifs, and Thr68 is the major in vitro
phosphorylation site by ATM. ATM- and Rad3-related also phos-
phorylates Thr68 in addition to Thr26 and Ser50, which are not
phosphorylated to a significant extent by ATM in vitro. In vivo,
Thr68 is phosphorylated in an ATM-dependent manner in response
to IR, but not in response to UV or HU. Substitution of Thr68 with
Ala reduced the extent of phosphorylation and activation of Chk2
in response to IR, and mutation of all seven SQ?TQ motifs blocked
all phosphorylation and activation of Chk2 after IR. These results
suggest that in vivo, Chk2 is directly phosphorylated by ATM in
responsetoIRandthatChk2isregulatedbyphosphorylationofthe
SCD.
D
the transcription of genes that facilitate repair (1). Defects in this
DNAdamageresponsepathwaycanresultingenomicinstability,
a mutagenic condition leading to cancer predisposition (2). Key
regulators of the DNA damage response pathway are the phos-
phoinositide kinase homologs ataxia telangiectasia-mutated
(ATM) and ATM- and Rad3-related (ATR), the protein kinases
Chk1 and Chk2, and the tumor suppressor protein p53 (1, 3).
ATM controls cell cycle arrest in G1and G2. Control of G2arrest
by ATM is thought to involve maintenance of Cdc2 in a
tyrosine-phosphorylated state (1), but precisely how ATM reg-
ulates Cdc2 has been unclear. Recently, Chk2, the mammalian
homolog of the budding yeast Rad53 and fission yeast Cds1
checkpointkinases(4–6),wasclonedandfoundtolinkATMand
Cdc2 (7, 8). In response to ionizing radiation (IR), Chk2 is
rapidly phosphorylated and activated in an ATM-dependent
manner, and in vitro, Chk2 phosphorylates Cdc25C on Ser216 (7,
8), which interferes with Cdc25C’s ability to activate Cdc2 (9). In
Chk2?/?mouse embryonic stem cells, maintenance of G2 arrest
and reduced Cdc2 kinase activity in response to IR is defective
(10). These results suggested a model in which ATM controls G2
arrest in part by activating Chk2.
The functional role of ATR and Chk1 in the checkpoint
pathway is not fully understood. Overexpression of ATR par-
tially suppresses cells lacking ATM, and overexpression of a
kinase-defective ATR mutant in wild-type cells abrogates G2?M
arrest after exposure to IR (11) and increases sensitivity of cells
to IR and UV (12), suggesting an overlapping function between
ATM and ATR. Chk1 is modified in response to DNA damage
and can phosphorylate Ser216 on Cdc25C in vitro (13). Recently,
ATR?/?blastocyst cells were found to die with chromosome
fragmentation, implying an essential role for ATR in the DNA
NA damage activates a response pathway that arrests the
cell cycle to provide time for repair of damage and induces
replication checkpoint (14), and mouse embryonic stem cells
lacking Chk1 have a defective G2?M DNA damage checkpoint
(15). Furthermore, in response to UV, HU, and IR, Chk1 is
found to be phosphorylated on Ser345, and the phosphorylation
in response to UV is ATR dependent (15). ATR and Chk1
disruptions both lead to peri-implantation embryonic lethality in
mice (14–16). Thus, Chk1 is regulated by ATR (15).
ATM-dependent G1 arrest in response to IR is achieved by
activation of p53 (17), which acts as a sequence-specific DNA-
binding transcription factor to induce transcription of the Cdk
inhibitor p21CIP1/WAF1(18, 19). Activation of p53 in response to
DNA damage is accomplished by both increasing its protein
levels and by increasing its transcriptional activity. The increase
in p53 levels in response to DNA damage is caused by decreased
degradation of p53, and phosphorylation of Ser20 on p53 is
required for the decrease in p53 degradation (20–22). Recently,
Chk2 was found to phosphorylate Ser20 on p53 in vitro (10, 23,
24), and we have found that Chk2 is required for increased
stability of p53 in response to IR (10), providing a possible
mechanism for regulation of p53 stability by Chk2.
Transcriptional activation of p53 is mediated by acetylation
and phosphorylation of p53. Acetylation of p53 by histone
acetyltransferases CBP?p300, which function as a coactivator
for p53-mediated transcription (25, 26), is shown to enhance
sequence-specific DNA binding of p53 (27). Phosphorylation
of Ser15 on p53 was found to increase the interaction with
CBP?p300 and acetylation of p53 (28). ATM phosphorylates
Ser15 in vitro, and in response to IR, Ser15 is phosphorylated
in an ATM-dependent manner, suggesting that Ser15 is phos-
phorylated by ATM in vivo (29, 30). However, Ser15 is
phosphorylated in an ATM-independent manner when cells
are exposed to UV, indicating that a kinase distinct from ATM
phosphorylates Ser15. A candidate kinase is ATR, which
phosphorylates Ser15 in vitro, and overexpression of kinase
dead ATR interferes with phosphorylation of Ser15 after
exposure of cells to IR or UV (31).
ATM-dependent phosphorylation of Chk2 is seen only when
cells are damaged by IR. Chk2 is phosphorylated in an ATM-
independent manner in response to UV or HU (7, 8, 32–34).
Thus, Chk2 phosphorylation parallels p53 phosphorylation on
Ser15. These results prompted us to examine Chk2 phosphory-
lation by ATM and ATR. In this study we report that Chk2 is
phosphorylated on Thr68 by both ATM and ATR in vitro, and
that Thr68 is phosphorylated in an ATM-dependent manner in
This paper was submitted directly (Track II) to the PNAS office.
Abbreviations: IR, ionizing radiation; HU, hydroxyurea; GST, glutathione S-transferase;
SQ?TQ, Ser-Gln?Thr-Gln; SCD, SQ?TQ cluster domain; MEF cells, mouse embryonic fibro-
blastcells;WT,wildtype;KD,kinasedead;ATM,ataxiatelangiectasia-mutated;ATR,ATM-
and Rad3-related.
§To whom reprint requests should be addressed. E-mail: selledge@bcm.tmc.edu.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
Article published online before print: Proc. Natl. Acad. Sci. USA, 10.1073?pnas.190030497.
Article and publication date are at www.pnas.org?cgi?doi?10.1073?pnas.190030497
PNAS ?
September 12, 2000 ?
vol. 97 ?
no. 19 ?
10389–10394
CELL BIOLOGY

Page 2

vivo when cells are exposed to IR but in an ATM-independent
manner when cells are treated with UV or HU. Furthermore,
Chk2 was found to be regulated by phosphorylation of the
N-terminal Ser-Gln?Thr-Gln (SQ?TQ) cluster domain (SCD).
Materials and Methods
Plasmids. Full-length Chk2 cDNA with NdeI site at its first ATG
was constructed in pBluscript by replacing a KpnI–NcoI frag-
ment of a Chk2 cDNA clone with a double-stranded oligonu-
cleotide with a NdeI site. pBSChk2-KD, a plasmid having kinase
dead Chk2 with substitution of Asp347 by Ala, was created by
PCR. pGST-Chk2 and pGST-Chk2-KD were respectively con-
structed by ligating a NdeI-EcoRI fragment of full-length Chk2
from pBSChk2 and pBSChk2-KD into pGEX2TKcs. To make
pGST-Chk2-KD2, Ala-substitution at Lys249 was introduced by
PCR. pGST-SQ?TQ was made by ligating NdeI–ApaI (blunt-
ended) fragment (1–96 amino acids) from pBSChk2 into NdeI–
EcoRI (blunt-ended) sites of pGEX2TKcs. pGST-?SQ?TQ was
constructed by ligating a ApaI (blunt-ended)–EcoRI fragment
(99–543 amino acids) from pBSChk2-KD into NcoI(blunt-
ended)–EcoRI sites on pGEX2TKcs. pRetFLAG-Chk2,
retrovirus vector encoding FLAG-tagged human Chk2, was
constructed in pBabe-puro (35). Site-directed mutagenesis for
replacement of SQ?TQ with Ala-Gln (AQ) was performed by
PCR, and all mutations were confirmed by DNA sequencing.
Plasmids expressing FLAG-ATM-WT and FLAG-ATM-KD
(30) were from M. Kastan (St. Judes Children’s Hospital,
Memphis, TN), and FLAG-ATR-WT and FLAG-ATR-KD (31)
were from R. Abraham (Duke University, Durham, NC).
Production of Anti-Phospho-Chk2 Antibodies. Rabbit polyclonal
anti-phospho-Thr68 antibodies were raised against a peptide
containing phospho-Thr68 [ETVST(-PO4)QELYS] and were
affinity-purified using a column of the antigen peptide. The
affinity-purified anti-phospho-Thr68 antibodies were further
purified by passage through a column of a non-phospho-Chk2-
peptide (ETVSTQELYS) to deplete contaminated antibodies
reactive with unphosphorylated antigen.
Western Blotting. Cells were lysed in 20 mM Tris?HCl buffer (pH
8.0) containing 100 mM NaCl, 0.5% Triton X-100, 10 mM NaF,
1 mM Na3VO4, 2 ?g?ml aprotinin, 2 ?g?ml leupeptin, 2 ?g?ml
antipain, and 1 mM phenylmethylsulfonyl fluoride (PMSF) for
15 min at 4°C. After centrifugation at 10,000 ? g for 10 min, cell
extracts were incubated with anti-Chk2 antibodies and protein
A–Sepharose (Amersham Pharmacia) for 1 h at 4°C. The
precipitated Sepharose beads were washed four times with 20
mM Tris?HCl buffer (pH 8.0) containing 100 mM NaCl, 0.1%
Triton X-100, 10 mM NaF, 1 mM Na3VO4, 2 ?g?ml aprotinin,
2 ?g?ml leupeptin, 2 ?g?ml antipain, and 1 mM PMSF and then
applied to SDS-PAGE and transferred to nitrocellulose mem-
branes. After blocking with 5% nonfat-milk in 50 mM Tris?HCl
(pH 7.5), 100 mM NaCl, and 0.05% Tween 20, the membranes
were incubated with anti-phospho-Thr68 antibodies and 100
?g?ml non-phospho-Chk2-peptides (ETVSTQELYS), and then
horseradish peroxidase-conjugated anti-rabbit IgG antibodies
(Promega), followed by detection by enhanced chemilumines-
cence (Amersham Pharmacia). To reprobe with anti-Chk2 an-
tibodies, membranes were washed twice with 50 mM glycine (pH
2.0) and then twice with 50 mM Tris?HCl (pH 7.5), 100 mM
NaCl, 0.05% Tween 20 before blocking with 5% nonfat-milk in
50 mM Tris?HCl (pH 7.5), 100 mM NaCl, 0.05% Tween 20.
In Vitro ATM?ATR Kinase Assay. ATM kinase assay with immuno-
precipitated ATM from normal and AT lymphoblast cells was
carried out as described (29). ATM-WT, ATM-KD, ATR-WT,
and ATR-KD proteins were expressed as FLAG-tagged protein
in 293T cells by transfection of the expression vector for ATM
(28) or ATR (29). DNA transfection into 293T cells was per-
formed with lipofectamine and Opti-MEM (GIBCO?BRL).
Forty-eight hours after transfection, cells were lysed by sonica-
tion in 50 mM Tris?HCl buffer (pH 7.5) containing 50 mM
?-glycerophosphate, 150 mM NaCl, 10% glycerol, 1% Tween 20,
1 mM NaF, 1 mM Na3VO4, 1 mM PMSF, 2 ?g?ml aprotinin, 2
?g?ml leupeptin, 2 ?g?ml antipain, and 1 mM DTT. After
centrifugation at 10,000 ? g for 10 min, FLAG-tagged proteins
were immunoprecipitated with anti-FLAG M2 antibodies (Sig-
ma) and protein A?G-Sepharose (Santa Cruz Biotechnology)
and washed four times with the lysis buffer, twice with 50 mM
Tris?HCl buffer (pH 7.5) containing 0.5 M LiCl, and then three
times with 20 mM Hepes (pH 7.4), 10 mM MgCl2, and 10 mM
MnCl2. Kinase reactions contained immunoprecipitated FLAG-
tagged proteins bound to protein A?G–Sepharose and GST-
Chk2substratesin20mMHepes(pH7.4),10mMMgCl2,10mM
MnCl2, 25 ?M ATP, and 15 ?Ci [?-32P]ATP for 30 min at
30°C. Proteins were separated by SDS-PAGE and visualized
by autoradiography.
Retrovirus-Mediated Expression of Human Chk2 in Chk2?/?Mouse
Embryonic Fibroblast (MEF) Cells. pRetFLAG-Chk2 or the empty
vector was transfected into Phoenix-eco cells (36), and, after
48 h, viral supernatants were harvested and used to infect
immortalized Chk2?/?MEF cells (10). Forty-eight hours after
infection, cells were treated with or without 20 Gy of ionizing
Fig. 1.
from normal lymphoid cells. Immunoprecipitates with anti-ATM antibodies
from normal lymphoblastoid cell lines (L40 and C3ABR) or AT lymphoblastoid
cell lines (AT59 and L6) were incubated with [?-32P]ATP and KD Chk2 fused to
GST (GST-Chk2-KD) or p53-N47 (29). The control reaction was carried out
without immunoprecipitate (No kinase). Proteins were resolved by SDS-PAGE
and visualized by autoradiography. (B) Chk2 phosphorylation by transfected
ATM.Immunoprecipitateswithanti-FLAGfrom293Tcellstransfectedwiththe
vectorexpressingFLAG-taggedwild-typeATM(ATM-WT)orkinasedeadATM
(ATM-KD) were incubated with GST-Chk2-KD and [?-32P]ATP. Proteins were
resolvedbySDS-PAGEandvisualizedbyautoradiographyorWesternblotting
with anti-FLAG antibodies.
Phosphorylation of Chk2 by ATM. (A) Chk2 phosphorylation by ATM
10390 ?
www.pnas.orgMatsuoka et al.

Page 3

radiation (IR), and cell extracts were prepared as described
previously (7).
Results and Discussion
To examine Chk2 phosphorylation by ATM, an in vitro ATM
kinase assay was performed, with GST-Chk2-KD as substrate.
GST-Chk2-KD is defective for kinase activity because of sub-
stitution of the catalytically essential Asp347 residue with Ala.
This mutation prevents autophosphorylation of Chk2, which
might complicate detection of phosphorylation by ATM. ATM
immunoprecipitates from the normal lymphoblastoid cell lines
L40 and C3ABR phosphorylated GST-Chk2 as well as p53-N47
(29), whereas immunoprecipitates obtained from AT lympho-
blastoid cell lines AT59 and L6, which lack the ATM protein, did
not phosphorylate GST-Chk2 and p53-N47 (Fig. 1A). To elim-
inate the possibility that an ATM-associated kinase phosphor-
ylatedChk2,transfectedFLAG-taggedwildtype(WT)orkinase
dead (KD) ATM was immunoprecipitated and used in a Chk2
phosphorylation assay. Immunoprecipitates from ATM-WT ex-
pressing cells phosphorylated GST-Chk2, but immunoprecipi-
tates from cells expressing ATM-KD did not (Fig. 1B). These
results demonstrate that Chk2 is phosphorylated by ATM in
vitro.
ATM is known to phosphorylate Ser15 on p53 (29, 30), and
this Ser residue is followed by Gln. Recently, ATM was found to
phosphorylate multiple Ser residues in the SCD on BRCA1 (37).
Chk2 has nine SQ?TQ motifs; seven of them are located in its
N terminus, and two are in its kinase domain (Fig. 2A). This
N-terminal SCD, in addition to the kinase and fork head-
associated domains, is conserved in Chk2, Rad53, and Cds1. To
determine the ATM phosphorylation sites on Chk2, we made
two deletion mutants of Chk2: GST-SQ?TQ, which contains
seven SQ?TQ motifs in the N terminus, and GST-?SQ?TQ,
which lacks the N-terminal SQ?TQ motifs but contains the fork
head-associated domain and the kinase domain, which is inac-
tivated by the replacement of Asp347 with Ala (Fig. 2A). As
shown in Fig. 2B, GST-SQ?TQ was readily phosphorylated by
ATM, whereas GST-?SQ?TQ was very poorly phosphorylated,
suggesting that ATM primarily phosphorylates Ser and?or Thr
residues in the N-terminal SQ?TQ motifs of Chk2.
To further map the ATM phosphorylation sites on Chk2, a
SQ?TQ to AQ substitution was introduced on each SQ?TQ
motif by site-directed mutagenesis. Whereas substitution of
Ser19, Thr26, Ser28, Ser33, Ser35, or Ser50 by Ala did not show
significant reduction in32P incorporation into GST-SQ?TQ,
substitution of Thr68 by Ala largely reduced32P incorporation
into GST-SQ?TQ (Fig. 3A). Quantitative analysis of radioactiv-
ity in GST-SQ?TQ proteins revealed that GST-SQ?TQ(T68A)
was phosphorylated to an extent that was less than 20% of the
level observed for WT GST-SQ?TQ. Furthermore, introduction
of Ala substitution at Thr68 into full-length KD Chk2 fused to
GST (GST-Chk2-KD(T68A)) resulted in a large reduction of32P
incorporation into GST-Chk2, compared with ‘‘wild type’’ KD
GST-Chk2 (WT) (Fig. 3B), indicating that Thr68 is the major
phosphorylation site by ATM. It should be noted that
incorporation into GST-Chk2-KD(T68A) is more than GST-
Chk2-KD(7A), in which all Ser and Thr in the N-terminal
SQ?TQ motifs were changed to Ala. Therefore, other Ser or Thr
residues in addition to Thr68 in the N-terminal SQ?TQ motifs
are phosphorylated by ATM (Fig. 3B), although to a significantly
lesser degree than Thr68. Finally, Ala substitutions were intro-
ducedintoSQ?TQmotifsinthekinasedomainofChk2.Because
32P
Fig. 2.
of human Chk2. SQ?TQ, Ser-Gln and Thr-Gln cluster domain; FHA, fork-head
associateddomain;Kinase,kinasedomain;O,SQandTQmotifs.(B)Phosphor-
ylation of the SQ?TQ-cluster domain on Chk2 by ATM. Immunoprecipitates
with anti-FLAG from 293T cells transfected with FLAG-ATM were incubated
with [?-32P]ATP and GST-SQ?TQ, GST-?SQ?TQ, or GST. Regions of SQ?TQ and
?SQ?TQfusedtoGSTareshowninA.ProteinswereresolvedbySDS-PAGEand
visualized by autoradiography.
Mapping ATM phosphorylation sites on Chk2. (A) Domain structure
Fig. 3.
with anti-FLAG from 293T cells transfected with the vector expressing FLAG-
ATMwereincubatedwith[?-32P]ATPandwild-typeGST-SQ?TQorGST-SQ?TQ
withAlasubstitutionatSer19(S19A),Thr26(T26A),Ser28(S28A),Ser33(S33A),
Ser35 (S35A), Ser50 (S50A), or Thr68 (T68A). Proteins were resolved by SDS-
PAGE and visualized by autoradiography. (B) FLAG-ATM was incubated with
[?-32P]ATP and full-length ‘‘wild type’’ kinase dead Chk2 fused to GST (WT),
GST-Chk2-KD with Ala substitution at Thr68 (T68A), or Ala substitution at all
sevenThrandSerintheN-terminalSQ?TQmotifs(7A).Proteinswereresolved
by SDS-PAGE and visualized by autoradiography. (C) FLAG-ATM was incu-
bated with [?-32P]ATP and GST-Chk2-KD2 (WT), a full-length ‘‘wild-type’’
kinase dead Chk2 fused to GST, or GST-Chk2-KD2 with Ala substitution at
Ser357 (S357A), Thr432 (T432A), or at both Thr68 and Ser357 (T68A?S357A).
Proteins were resolved by SDS-PAGE and visualized by autoradiography.
Phosphorylation of Thr68 on Chk2 by ATM. (A) Immunoprecipitates
Matsuoka et al.
PNAS ?
September 12, 2000 ?
vol. 97 ?
no. 19 ?
10391
CELL BIOLOGY

Page 4

Asp347 is close to Ser357, substitution at Asp347 might con-
ceivably affect the phosphorylation of Ser357. Therefore, we
made a second kinase defective mutant, GST-Chk2-KD2, in
which Lys249 was replaced by Ala, and three SQ?TQ mutants of
GST-Chk2-KD2, which have Ala substitution at Ser357 (S357A),
Thr432 (T432A), or at both Thr68 and Ser357 (T68A?S357A).
Fig. 3C shows that a similar amount of32P was incorporated into
‘‘wild type’’ GST-Chk2-KD2 (WT) and mutants of Ser357
(S357A) and Thr432 (T432A). However,32P incorporation was
largelyreducedinthemutantwithAlasubstitutionatbothThr68
and Ser375 (T68A?S357A), compared with a mutant with single
Ala substitution at Ser357 (S357A) (Fig. 3C). From these
findings we concluded that Thr68 is the major in vitro site of
phosphorylation by ATM on Chk2.
To determine whether Thr68 is a phosphorylation site for
Chk2 in vivo, polyclonal rabbit anti-phospho-Thr68 antibodies
were raised against a Chk2 peptide containing phosphorylated
Thr68. Results obtained by ELISA indicated that the anti-
phospho-Thr68 antibodies recognize a Chk2 peptide containing
phosphorylated Thr68, but not unphosphorylated Thr68, and
that the antibodies do not cross-react with Chk2 peptides
containing other SQ?TQ motifs in the phosphorylated state
(data not shown). Chk2 protein immunoprecipitated with anti-
Chk2antibodiesfrom293TcellstreatedwithIR,UV,orHUand
untreated 293T cells was immunoblotted with anti-phospho-
Thr68 antibodies. Fig. 4A shows that Thr68 on Chk2 is phos-
phorylated in response to IR, UV, or HU in vivo. The anti-
phospho-Thr68 antibody did not react with the T68A Chk2
mutantthatwasexpressedincellsdamagedbyIR,indicatingthat
the antibody is highly specific to phosphorylated Thr68 on Chk2
(Fig. 4B). Although phosphorylation and activation of Chk2 in
response to IR is ATM dependent, Chk2 is phosphorylated in an
ATM-independent manner when cells are treated with UV or
HU (7, 8, 32–34). To investigate whether phosphorylation of
Thr68 is ATM dependent or not, Chk2 was immunoprecipitated
from AT cells into which a functional ATM gene or the vector
was reintroduced (7, 38) and subjected to Western blotting with
anti-phospho-Thr68 antibodies. In response to IR, Thr68 is
phosphorylated in cells expressing wild-type ATM cDNA, but
not in cells lacking ATM, and phosphorylation of Thr68 is seen
after UV or HU treatment, regardless of the status of ATM (Fig.
4C).TheseresultsindicatethatThr68isphosphorylatedbyATM
in response to IR in vivo and suggest the existence of kinase(s)
other than ATM, which phosphorylates Thr68 on Chk2 in
response to UV or HU treatment.
Because Ser15 on p53 is phosphorylated by not only ATM but
also by ATR in vitro (29–31), we examined in vitro Chk2
phosphorylation by ATR. ATR-WT or ATR-KD was expressed
and immunoprecipitated from 293T cells and used in a Chk2
phosphorylation assay. GST-SQ?TQ was phosphorylated by
ATR-WT but not by ATR-KD (Fig. 5A). Whereas Ala substi-
tution at Thr68 resulted in a significant reduction of32P incor-
porationintoGST-SQ?TQbyATM(Fig.3A),substitutionofany
Fig. 4.
responsetoDNAdamageinvivo.293Tcellswereuntreatedortreatedwith20
Gy IR or 50 J?m2UV radiation and harvested after 2 h, or grown in 1 mM HU
for 24 h before harvesting. Chk2 protein immunoprecipitated with anti-Chk2
antibodiesfromthesecellswasimmunoblottedwithanti-phospho-Thr68and
anti-Chk2antibodies.(B)293Tcellsweretransfectedwithplasmidsexpressing
FLAG-taggedwild-typeChk2(WT),ortheT68AmutantChk2(T68A)andChk2
proteins were immunoprecipitated with anti-FLAG antibody from untreated
cells (?) or cells treated (?) with 20 Gy IR. Chk2 was immunoblotted with
anti-phospho-T68 and anti-Chk2 antibodies. (C) ATM-dependent phosphory-
lation of Thr68 in response to IR. AT cells containing the vector alone (?ATM)
and the same cells containing the vector expressing ATM (?ATM) were
untreatedortreatedwith10GyIRor50J?m2UVradiationandharvestedafter
1 h or were grown in 1 mM HU for 24 h before harvesting. Chk2 immunopre-
cipitated with anti-Chk2 antibodies from cells was immunoblotted with anti-
phospho- and anti-Chk2 antibodies.
Phosphorylation of Thr68 in vivo. (A) Phosphorylation of Thr68 in
Fig. 5.
(ATR-WT) or kinase dead ATR (ATR-KD) was incubated with [?-32P]ATP and
wild-type GST-SQ?TQ (WT) or GST-SQ?TQ with Ala-substitution at Ser19
(S19A), Thr26 (T26A), Ser28 (S28A), Ser33 (S33A), Ser35 (S35A), Ser50 (S50A),
or Thr68 (T68A). Proteins were resolved by SDS-PAGE and visualized by auto-
radiography. (B) FLAG-ATM or FLAG-ATR was incubated with [?-32P]ATP and
wild-typeGST-SQ?TQ(WT),GST-SQ?TQwithAlasubstitutionatsixSerandThr
residues in SQ?TQ motifs except Ser19 (S19), Thr26 (T26), Ser28 (S28), Ser33
(S33), Ser50 (S50), or Thr68 (T68); or GST-SQ?TQ with Ala substitution at all
seven Ser and Thr residues in SQ?TQ motifs (7A). Proteins were resolved by
SDS-PAGE and visualized by autoradiography.
Phosphorylation of Chk2 by ATR. (A) FLAG-tagged wild-type ATR
10392 ?
www.pnas.org Matsuoka et al.

Page 5

single SQ or TQ with AQ did not show a significant reduction
of phosphorylation by ATR (Fig. 5A). This finding suggests that
ATR phosphorylates multiple Ser and Thr residues in the
N-terminal SCD. To determine the sites of in vitro phosphory-
lation by ATR, kinase assays were carried out with GST-SQ?TQ
mutants, which have an Ala substitution at six of the seven Ser
and Thr residues in the SQ?TQ motif, so that each mutant has
only a single SQ or TQ motif. ATM efficiently phosphorylated
wild-type GST-SQ?TQ (WT) and GST-SQ?TQ(T68), a mutant
that has intact Thr68 and six Ala substitutions at other SQ?TQ
motifs, but32P incorporation into GST-SQ?TQ mutants that
have intact Ser or Thr at Ser19 (S19), Thr26 (T26), Ser28 (S28),
Ser33 (S33), or Ser50 (S50) was as poor as GST-SQ?TQ(7A), in
which all seven SQ?TQ motifs are changed to AQ (Fig. 5B).
These results are completely consistent with the notion that
Thr68 is the major in vitro phosphorylation site by ATM.
ATR phosphorylated GST-SQ?TQ mutants that have intact
Thr26 (T26), Ser50 (S50), or Thr68 (T68), at levels that were as
much as 20–40% of the phosphorylation of wild-type GST-
SQ?TQ (WT), but other mutants that have intact Ser19 (S19),
Ser28 (S28), and Ser33 (S33) were phosphorylated as poorly as
GST-SQ?TQ(7A) (Fig. 5B). Because Ala substitution at Thr26,
Ser28, Ser33, or Ser35 could possibly affect phosphorylation of
the adjacent Ser and?or Thr residues in the SQ?TQ motif by
altering the recognition sequence for ATR, we are unable to
exclude the possibility of phosphorylation of Ser28, Ser33, and
Ser35byATR.However,theseresultsindicatethatThr26,Ser50,
and Thr68 could be phosphorylated by ATR in vitro and
demonstrate that ATM and ATR have overlapping but distinct
substrate specificities in vitro.
Finally, to explore the function of SQ?TQ phosphorylation,
retroviruses expressing wild-type human Chk2, the T68A Chk2
mutant, or a mutant in which all seven SQ?TQ motifs in the SCD
were changed to AQ, were introduced into Chk2?/?MEF cells
(10). In these retroviruses human Chk2 is expressed at a level
similar to that of endogenous Chk2 in human cell lines (data not
shown). Wild-type human Chk2 expressed in Chk2?/?MEF cells
shows the same mobility shift on SDS-PAGE as the endogenous
Chk2 does in human cells (7), suggesting that it is correctly
phosphorylated in response to IR (Fig. 6A). Substitution of
Thr68 with Ala reduced the amount of fully phosphorylated
Chk2 by 50%, and Ala substitution at all seven SQ?TQ motifs
abolished the mobility shift altogether, indicating that phosphor-
ylation of Thr68 is not solely responsible for the shift in
electrophoretic mobility after IR (Fig. 6A). Kinase assays with
Chk2 immunoprecipitated from these cells revealed that T68A
Chk2 mutants displayed a significantly reduced IR induction of
kinase activity (75% reduction relative to wild-type activation).
The mutant Chk2 lacking all SQ?TQ motifs in the SCD did not
show any activation by IR (Fig. 6B). These results suggest that
phosphorylation of the SCD regulates Chk2 activation in re-
sponse to IR.
The results shown here indicate that Chk2 is a direct substrate
of ATM both in vitro and in vivo, and that Thr68 is the major site
of phosphorylation in vitro and is phosphorylated in response to
DNA damage in vivo. Although substitution at Thr68 largely
reduced32P incorporation into GST-Chk2-KD, further reduc-
tion in32P incorporation into GST-Chk2-KD was observed when
all seven SQ?TQ motifs in the N terminus were changed to AQ
(Fig. 3B). Thus, there must be additional ‘‘minor’’ in vitro
phosphorylation sites in the N-terminal SQ?TQ motifs. These
results are consistent with those obtained from the in vivo
experiments on the expression of wild-type and mutant Chk2 in
Chk2?/?MEF cells, in which the Chk2 T68A mutant showed
partial phosphorylation and activation in response to IR that was
dependent upon the remaining six SQ?TQ motifs in the SCD
(Fig.6).Althoughwewereunabletoidentifythese‘‘minor’’sites,
by the use of additional anti-phospho antibodies, Ser28 was also
foundtobephosphorylatedinanATM-dependentmannerwhen
cells were damaged by IR (data not shown). This finding
indicates that sites other than Thr68 could be phosphorylated by
ATM or an unidentified ATM-dependent kinase in vivo.
We have previously shown that phosphatase treatment of
Chk2 from IR-treated cells reduces its kinase activity, which is
activated upon IR (7). Thus, phosphorylation of Chk2 can
activate Chk2 kinase activity. However, we were not able to
detect activation of Chk2 when the wild-type GST-Chk2 was
phosphorylated by ATM in vitro (data not shown). We may have
failed to detect activation of Chk2 in this instance because only
a small fraction of the Chk2 was phosphorylated by ATM in this
reaction. It is also possible that an additional protein that binds
and activates phosphorylated Chk2 in vivo is absent from the in
vitro reaction. Alternatively, it is possible that all of the phos-
phorylation sites necessary for Chk2 activation are not accessible
to ATM and therefore are not phosphorylated in vitro when
Chk2 is fused to GST and?or ATM is bound to Sepharose beads.
In response to UV or HU treatment, Thr68 is phosphorylated
in an ATM-independent manner, indicating phosphorylation of
Thr68 by a kinase other than ATM. ATR phosphorylates Chk2
on Thr68 in vitro, and phosphorylation of Thr68 in response to
IR occurs later in cells lacking ATM (data not shown). Phos-
phorylation of Ser15 on p53 in response to IR is also delayed in
ATM mutant cells (39), and ATR is suggested to phosphorylate
Ser15 in late phase in response to IR (31). Therefore, we believe
ATR is likely to phosphorylate Thr68 on Chk2 in vivo. ATR also
phosphorylates Thr26 and Ser50 in vitro. Further experiments
will be required to determine whether these sites are phosphor-
ylated by ATR in vivo.
This paper is dedicated to Dr. Akio Matsukage on the occasion of
his 60th birthday. We thank B. Zhou and C. McGowan for sharing
Fig. 6.
(A) Chk2?/?MEF cells infected with retroviruses encoding empty vector (Vec-
tor),wild-typehumanChk2(WT)ortheChk2T68Amutant(T68A),oramutant
with Ala substitution at all seven N-terminal SQ?TQ motifs (7A) were not
treated (?) or were treated (?) with 20 Gy ionizing radiation (IR) and har-
vested after 1 h. Protein from these cells was fractionated by SDS-PAGE and
immunoblotted with anti-human Chk2 antibody. (B) Chk2 was immunopre-
cipitated with anti-Chk2 antibodies from cells in A. Immunoprecipitates were
incubated with GST-Cdc25C (200–256) and [?-32P]ATP. Proteins were resolved
by SDS-PAGE and visualized by autoradiography or Western blotting with
anti-human Chk2 antibodies.
Regulation of Chk2 by phosphorylation of N-terminal SQ?TQ motifs.
Matsuoka et al.
PNAS ?
September 12, 2000 ?
vol. 97 ?
no. 19 ?
10393
CELL BIOLOGY