Roos-Mattjus et al 1 Phosphorylation of Human Rad9 Is Required ...
Transcript of Roos-Mattjus et al 1 Phosphorylation of Human Rad9 Is Required ...
Roos-Mattjus et al
1
Phosphorylation of Human Rad9 Is Required for Genotoxin-Activated Checkpoint
Signaling
Pia Roos-Mattjus1,9, Kevin M. Hopkins7, Andrea J. Oestreich3, Benjamin T. Vroman4,
Kenneth L. Johnson6, Stephen Naylor1,2,6,10, Howard B. Lieberman7, and Larry M.
Karnitz2,3,4,5,8
Department of Biochemistry and Molecular Biology1, the Department of
Molecular Pharmacology and Experimental Therapeutics2, the Program in Tumor
Biology3, the Divisions of Developmental Oncology Research4 and Radiation Oncology5,
and the Biomedical Mass Spectrometry and Functional Proteomics Facility6
Mayo Clinic and Foundation
Rochester, MN 55905, USA
and
The Center for Radiological Research, Columbia University7
630 West 168th Street
New York, NY 10032, USA.
8Corresponding author: Division of Developmental Oncology Research, Guggenheim 13,Mayo Clinic and Foundation, 200 First Street Southwest, Rochester, MN 55905, USATelephone: 507-284-3124FAX: 507-284-3906email: [email protected]
9Present Address:Turku Centre for BiotechnologyBiocity, 5th floorTykistokatu 6BFin-20520 TurkuFinland
10Present Address:Beyond Genomics40 Bear Hill RoadWaltham, MA 02451
Running title: Rad9 tail phosphorylation affects 9-1-1 function
Copyright 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
JBC Papers in Press. Published on April 21, 2003 as Manuscript M301544200 by guest on M
arch 17, 2018http://w
ww
.jbc.org/D
ownloaded from
Roos-Mattjus et al
2
Abstract
Rad9, a key component of genotoxin-activated checkpoint signaling pathways,
associates with Hus1 and Rad1 in a heterotrimeric complex (the 9-1-1 complex). Rad9 is
inducibly and constitutively phosphorylated. However, the role of Rad9 phosphorylation
is unknown. Here we identified nine phosphorylation sites, all of which lie in the carboxyl-
terminal 119-amino acid Rad9 "tail," and examined the role of phosphorylation in
genotoxin-triggered checkpoint activation. Rad9 mutants lacking a Ser272
phosphorylation site, which is phosphorylated in response to genotoxins, had no effect
on survival or checkpoint activation in Mrad9-/- mouse ES cells treated with hydroxyurea
(HU), ionizing radiation (IR), or ultraviolet radiation (UV). In contrast, additional Rad9 tail
phosphorylation sites were essential for Chk1 activation following HU, IR, and UV
treatment. Consistent with a role for Chk1 in S-phase arrest, HU- and UV-induced S-
phase arrest was abrogated in the Rad9 phosphorylation mutants. In contrast, however,
Rad9 did not play a role in IR-induced S-phase arrest. Clonogenic assays revealed that
cells expressing a Rad9 mutant lacking phosphorylation sites were as sensitive as Rad9-
/- cells to UV and HU. Although Rad9 contributed to survival of IR-treated cells, the
identified phosphorylation sites only minimally contributed to survival following IR
treatment. Collectively, these results demonstrate that the Rad9 phospho-tail is a key
participant in the Chk1 activation pathway and point to additional roles for Rad9 in
cellular responses to IR.
Keywords: checkpoint/DNA damage/phosphorylation/Rad9
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
3
Introduction
Cells respond to genotoxic stress by activating checkpoint signaling pathways
that delay cell cycle progression, providing time to repair DNA damage, activate
transcriptional responses and, if the damage is too severe, induce apoptotic pathways
(1-6). The phosphatidylinositol-3 kinase-related kinases (PIKK) ATM and ATR are
central components of the checkpoint signaling pathways (7). ATM and ATR are
activated by genotoxins and phosphorylate downstream signaling proteins, including
Chk1 and Chk2, two protein kinases that phosphorylate proteins that regulate checkpoint
responses (8-10).
In addition to the PIKKs, Rad1, Hus1, Rad9, and Rad17 (using
Schizosaccharomyces pombe nomenclature) are evolutionarily conserved proteins
essential for Chk1 activation (11-13). Rad9, Hus1 and Rad1 form a stable heterotrimeric
complex, called the 9-1-1 complex (14-17). Biochemical, biophysical, and molecular
modeling studies predict that the 9-1-1 complex resembles PCNA (14,18-21), a
homotrimeric clamp that is loaded around DNA at primer-template junctions (22,23).
Loading of PCNA is carried out by the clamp loader replication factor C (p140-RFC), a
heteropentameric complex consisting of 1 large subunit (p140) and four small subunits
(24). The 9-1-1 complex also interacts with a putative clamp loader, the Rad17-RFC
complex (25,26), which is composed of Rad17, an RFC-like protein, and the four small
RFC subunits (15,19). In vitro, recombinant Rad17-RFC complex loads the 9-1-1
complex onto DNA (27), and in intact cells, multiple genotoxins induce the association of
the 9-1-1 clamp with chromatin (28) in a process that requires Rad17 (12). Collectively,
these observations support a model in which the Rad17-RFC clamp loader is a DNA
damage sensor that loads the 9-1-1 clamp onto sites of DNA damage in the checkpoint
signaling pathway (1,3,29,30).
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
4
Recent studies demonstrated that ATR and the 9-1-1 complex collaboratively
induce Chk1 activation. In both humans and S. cerevisiae, the 9-1-1 complex associates
with chromatin in a PIKK-independent manner following DNA damage (12,31,32).
Likewise, ATR forms foci in the absence of Rad17 (12). Taken together, these results
suggest that the 9-1-1 complex and ATR associate with DNA lesions independently of
one another. Nonetheless, both the 9-1-1 complex and ATR are required for Chk1
activation (11-13,33), suggesting that these chromatin-bound complexes coordinately
transduce a Chk1-activating signal in response to genotoxins; it remains unclear,
however, how these protein complexes are regulated or how they cooperatively couple
with the downstream signaling pathways.
One possible 9-1-1 regulatory mechanism is phosphorylation. Both Rad9 and
Rad1 are phosphorylated in response to DNA damage (28,34-38). Because Rad9 binds
DNA in the absence of ATR and pharmacologic inhibition of PIKKs does not affect Rad9
chromatin binding, it is unlikely that Rad9 phosphorylation regulates 9-1-1 chromatin
binding (12,39). Instead, phosphorylation of 9-1-1 complex members may regulate
downstream signals that are activated by the clamp complex.
In the present study we examined the role of Rad9 phosphorylation in checkpoint
signaling. We identified 9 phosphorylation sites in the carboxyl-terminal 119-amino acid
Rad9 "tail," which extends beyond the amino-terminal PCNA homology domains and is
not required for interaction with Hus1 and Rad1 (14,40). One of the sites we found was
the previously identified Rad9 damage-inducible phosphorylation site Ser272. We
assessed the effects of Rad9 phosphorylation by expressing Rad9 mutants in Mrad9-/-
embryonic stem (ES). These studies showed that cells expressing a Rad9 mutant in
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
5
which Ser272 was converted Ala were indistinguishable from cells expressing wild-type
Rad9. In contrast, the remaining Rad9 phosphorylation sites were essential for some but
not all Rad9-dependent cellular responses. Collectively, these results demonstrate that
the phospho-Rad9 tail plays a critical role in the transduction of downstream checkpoint
signals.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
6
Experimental Procedures
Antibodies
Antibodies generated to human Rad9 (monoclonal and polyclonal), Rad1
(polyclonal), Hus1 (monoclonal) and Rad17 (polyclonal) have been described previously
(28,34). Phospho-Chk1 (P-Ser345) and Chk1 (G-4) antibodies were purchased from Cell
Signaling Technology (Beverly, MA) and Santa Cruz Biotechnology (Santa Cruz, CA),
respectively, and were used according to the manufacturer’s instructions. Anti-Chk2
antibodies (41) were a generous gift from J. Chen, (Mayo Clinic, Rochester, MN) and
anti-phospho-Rad9 (P-Ser272) (37) were a generous gift from E. Lee (University of
Texas, San Antonio Texas). The p53 antibody (CM5) was purchased from Novocastra
Laboratories, Ltd. (Newcastle upon Tyne, UK). The JNK1 (C-17) antibody was
purchased from Santa Cruz Biotechnology.
Cell culture and cell transfection
Human K562 erythroleukemia cells were cultured in RPMI-1640 (BioWhittaker,
Walkersville, MD) containing 10% fetal bovine serum (Biofluids, Rockville, MD). K562
cells (1x107) were transiently transfected as previously described (34). Mouse ES cells
were maintained on gelatinized tissue culture plates in knock-out DMEM (Invitrogen,
Carlsbad, CA) containing 15% ES cell-qualified fetal bovine serum (Cell & Molecular
Technologies, Inc., Phillipsburg, NJ), 0.1 mM non-essential amino acids, 2 mM L-
glutamine (Invitrogen), 10-4 M 2-mercaptoethanol (Sigma) and 103 units/ml ESGRO
(Chemicon International, Temecula, CA). Cells were grown at 37°C and 5% CO2. ES
cells were transfected with Lipofectamine 2000 according to manufacturer’s instructions
(Invitrogen). Stable clones were selected in medium containing G418 (0.2 mg/ml) and
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
7
characterized for expression levels by immunoblotting. Two independent clones of each
mutant expressing similar levels of protein were used for further studies.
Derivation of Mrad9-/- ES cells.
The full derivation of the Mrad9-/- ES cells will be described elsewhere. Briefly,
mouse ES cells were electroporated with a targeting vector (Mrad9neo/loxP) that, when
integrated into the MRad9 locus, inserts loxP sites in the 5' untranslated region and in
the second intron. Following selection in 200 µg/ml G418 and gancyclovir to enrich for
homologous targeting events, Southern blots and PCR were used to identify clones in
which a single Mrad9 allele was targeted. One Mrad9neo/loxP ES cell line was cultured in
300 µg/ml G418 to select for cells that targeted the second Mrad9 allele, and a cell line
with two targeted MRad9 alleles was identified by Southern blotting and PCR analyses.
To delete the first and second Mrad9 exons via Cre-mediated recombination (42), the
cells were transiently transfected with an HS-cre expression vector (43) and pPur
(Clontech, Palo Alto, CA), and selected in puromycin. The clones were then examined
by Southern blotting and PCR analyses to identify one in which the first and second
exons of both Mrad9 alleles were deleted.
Drug and radiation treatments
Hydroxyurea (Sigma, St. Louis, MO) was prepared fresh in phosphate-buffered
saline (PBS). Cells were irradiated with a 137Cs source at a dose rate of approximately
10 Gy/min (Mark I, JLS Shepherd and Associates, San Fernando, CA). For UV
treatment, cells were washed with PBS to remove medium. Residual PBS was removed
prior to irradiation with a UV-C irradiator (Spectrolinker XL-1000, Spectronics Corp.,
Westbury, NY). Fresh, warmed medium (37°C) was added after UV irradiation.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
8
Cell extract preparation and analysis of Rad9 chromatin binding
To analyze protein expression levels, cells were lysed in 50 mM HEPES, pH 7.4,
1% Triton X-100, 10 mM NaF, 30 mM Na4P207, 150 mM NaCl, 1 mM EDTA, containing
freshly added 10 mM ß-glycerophosphate, 1 mM Na3VO4, 10 µg/ml pepstatin A, 5 µg/ml
aprotinin, 10 µg/ml leupeptin and 20 nM microcystin-LR for 10 min at 4°C. For protein-
protein interaction studies, cells were lysed in 150 mM KCl, 10 mM MgCl2 and 10 mM
HEPES, pH 7.4, (supplemented with 100 µg/ml digitonin, 10 µg/ml aprotinin, 5 µg/ml
pepstatin, 5 µg/ml leupeptin, 20 nM microcystin-LR, 1 mM Na3VO4, and 10 mM
ß–glycerophosphate) for 10 min at 4ºC. All lysates were either immunoprecipitated as
indicated or boiled with sodium dodecyl sulfate-polyacrylamide gel electrophoresis
(SDS-PAGE) sample buffer. Proteins were separated by SDS-PAGE, transferred to
Immobilon P membranes (Millipore, Bedford, MA) and immunoblotted as described (34).
Fractionation of the 9-1-1 complex into released and chromatin-bound fractions was
done as described previously (28).
Rad9 expression vectors and Rad9 mutagenesis
S-tag-Rad9-pIRES2-EGFP was generated by adding an in-frame S-tag™
(Novagen Inc.) to the amino terminus of human Rad9 using a PCR strategy. The
resulting PCR product was cloned into pIRES2-EGFP (BD Sciences Clontech, Palo Alto,
CA) using XhoI and BamHI. To generate stable ES cells expressing wild-type and
mutant Rad9s, untagged human Rad9 was cloned into pIRES2-EGFP. Human Rad9
phospho-deficient mutants were generated by mutating the mapped phosphorylation
sites using either the GeneEditor kit (Promega, Madison, WI) according to the
manufacturer’s instructions or by sequence-overlap extension (44). All mutants were
sequenced to confirm mutations.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
9
Purification of S-tagged Rad9
A total of 2x109 K562 cells were transiently transfected by electroporation with S-
tag-Rad9-pIRES2-EGFP (50 pooled transfections). For Edman radiosequencing
analysis, cells were recovered by centrifugation 24 hours after transfection, washed
twice with phosphate-free media (ICN Biomedicals, Inc., Aurora, OH), and resuspended
in 80 ml of phosphate-free media containing 5% fetal bovine serum and 5 mCi
[32P]orthophosphate (ICN Biomedicals, Inc.). Cells were cultured an additional two
hours, harvested, washed in ice-cold PBS, and lysed in buffer containing 10 mM
HEPES, pH 7.4, 150 mM KCl, and 10 mM MgCl2 (containing freshly added 100 µg/ml
digitonin, 10 µg/ml aprotinin, 5 µg/ml pepstatin, 5 µg/ml leupeptin, 20 nM microcystin-LR,
1 mM Na3VO4, and 10 mM ß–glycerophosphate) for 10 min at 4ºC. Following
centrifugation, the supernatant was added to S-protein agarose beads (Novagen Inc.),
and urea was added to a final concentration of 2 M. The mixture was incubated with
rotation for 2 h at 4°C. The beads were washed with cold lysis buffer (50 mM HEPES,
pH 7.6, 1% Triton-X100, 10 mM NaF, 30 mM NaP2O7, 150 mM NaCl, 1 mM EDTA)
supplemented with 2 M urea, 1 mM Na3VO4, and 10 mM ß–glycerophosphate. Proteins
were released from the beads by heating in SDS-PAGE sample buffer and were
resolved by SDS-PAGE. The gel was stained with Coomassie Blue, and the slowest-
migrating (most highly phosphorylated) Rad9 band was excised.
Radiosequencing of Rad9
To identify the amino acid sequence of the radiolabeled peptides, the Coomassie
Blue-stained band corresponding to fully phosphorylated, radiolabeled Rad9 was
excised from the gel and cut into pieces no larger than 2 mm. Prior to digestion, the gel
pieces were destained until clear, then reduced with dithiothreitol, and alkylated with
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
10
iodoacetamide. In-situ enzymatic digestion was performed overnight at 37°C with
modified sequencing-grade trypsin (Promega). Peptides were extracted from the gel,
and a portion of the sample was separated on an Applied Biosystems 173A Microblotter
Capillary HPLC System (Applied Biosystems, Foster City, CA) using a 0.5-mm x 150-
mm C18 column with an acetonitrile gradient in the presence of 0.1% trifluoroacetic acid.
The eluting peptides were spotted directly onto a strip of polyvinylidinedifluoride
membrane. The strip was exposed to film to identify 32P-labeled peptides. Spots
containing 32P-labeled peptides were excised, treated with Biobrene in methanol, and
applied to either an Applied Biosystems Procise 492 HT or the Procise cLC Protein
Sequencing System and run using pulsed liquid chemistry. These data were analyzed
with the ABI Model 610A data analysis software (Applied Biosystems). To identify the
positions of 32P-labeled amino acids within the radioactive peptides, the remainder of the
sample was separated on the capillary HPLC exactly as described above. However, the
radiolabeled fractions were instead collected manually and covalently bound to aryl
amine-derivatized polyvinylidinedifluoride disks using the Sequelon-AA kit (Applied
Biosystems), following the suggested instructions. The disks were chromatographed on
an Applied Biosystem’s Procise cLC modified to collect the derivatized amino acids
released by each cycle directly into a 96-well microtiter plate. The extraction solvent of
90% MeOH/10% water/0.01% H3PO4 was substituted for n-butyl chloride in the S3 bottle
on the sequencer. Each sequencer fraction, representing one Edman degradation cycle,
was spotted onto a custom-made polystyrene plate and dried. The plate containing the
released amino acids was counted either by direct counting using a Packard Matrix 96
counter (Packard Biosciences, Boston, MA) or by exposing the plate to a phosphor
screen that was scanned with a Storm Phosphorimager (Molecular Dynamics,
Piscataway, NJ).
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
11
Nano-scale liquid chromatography-mass spectrometry (nLC/MS) and tandem mass
spectrometry (nLC/MS/MS) of Rad9
For mass spectrometric analysis, S-tag-Rad9 was purified as described above
with the exception that the cells were not labeled with [32P]orthophosphate. After S-
protein purification, one portion of the sample was left untreated and the other portion
was treated with l-protein phosphatase according to the manufacturer’s instruction (New
England Biolabs, Beverly, MA). The proteins were released from the beads by heating in
SDS-PAGE sample buffer and were resolved by SDS-PAGE. The gel was stained with
Coomassie Blue. Protein bands corresponding to dephosphorylated and fully
phosphorylated Rad9 were excised. Gel slices containing phosphorylated and
dephosphorylated Rad9 were destained, reduced with dithiothreitol, and alkylated with
iodoacetamide. The gel was dried under vacuum and rehydrated in buffer containing 200
mM ammonium bicarbonate containing 0.2 mg/ml Zwittergent 3-16 and 40 ng/ml porcine
trypsin (Promega) for 20 min. The gel slice was rinsed in 50 mM ammonium bicarbonate
and incubated overnight at 37°C in 50 mM ammonium bicarbonate. The next day the
sample was sonicated, acidified with 10% TFA, and sonicated again.
nLC/MS and nLC/MS/MS measurements were performed on aliquots of the
tryptic digest using a Micromass Q-Tof-II mass spectrometer (Micromass, Manchester,
UK). Reversed-phase nLC separations were performed on a 75-mm i.d. x 5-cm long
PicoFit column packed with Aquasil C18 stationary phase (NewObjective, Inc.,
Cambridge, MA). Mobile phase A consisted of water/acetonitrile/n-propanol (98:1:1)
containing 0.2% formic acid. Mobile phase B consisted of acetonitrile/n-propanol/water
(80:10:10) containing 0.2% formic acid. A linear gradient from 0-50% B over 30 minutes
was performed. The liquid chromatography pumping system (Michrom UMA, Michrom
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
12
BioResources Inc., Auburn, CA) was operated at 50 ml/min and split to a column flow of
0.2 mL/min just prior to the sample introduction valve.
Peptides from both the phosphorylated and dephosphorylated samples were
compared using Mass Lynx software (Micromass, Manchester, UK). Novel peptides
were considered putative phospho-peptides if the mass of the of the tryptic peptide
corresponded to the mass of Rad9 tryptic peptide plus 80 mass units per phosphate
group. Peptides of the appropriate masses were subjected to MS/MS to identify
phosphorylation sites. MS/MS experiments were performed using argon as the collision
gas, and the collision energy was determined as a function of mass/charge (m/z) and
charge (z). Sites of phosphorylation were determined by a combination of Sequest
database searching (ThermoFinnigan, San Jose, CA) and manual spectrum
interpretation.
S-phase checkpoint assay
One day prior to the experiment, 1-2 x 104 cells were plated onto gelatinized 96-
well plates. Cells were either left untreated or treated with IR (30 Gy) or UV irradiated
(20 J/m2), and incubated for 40 minutes at 37°C following treatment. [Methyl-
3H]thymidine (TRA120, Amersham Pharmacia) was then added at 2 µCi/well and the
cells were incubated for an additional 20 minutes. Cells were released by trypsinization
and harvested by transferring to glass filters. The filter-bound cells were lysed with
distilled water (Skatron semiautomatic harvester, Skatron As., Lier, Norway). Filter-
bound radioactivity was determined by liquid scintillation counting. 3H-thymidine
incorporation was calculated as the ratio of treated to control samples.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
13
Cell survival and clonogenic assays
Varying numbers of ES cells were plated on gelatinized 60-mm dishes in
triplicate and 4 h later treated with the indicated doses of ionizing radiation or HU. HU
was washed off with PBS 24 h later, and fresh medium was added. Cells were incubated
for 14 days, stained with Coomassie Blue, and colonies were counted. Survival was
calculated as the percentage of colonies in the untreated dishes compared to the treated
dishes, taking into account the plating efficiency, using the equation: colonies
counted/(cells plated)(plating efficiency).
For UV survival assays 1-2 x 104 cells were plated onto gelatinized 6-well plates.
Plates were washed with PBS 24 h later. After removal of residual PBS, the cells were
UV irradiated (15 J/m2) and fresh media was added. Plates were washed 72 h later and
stained with Coomassie Blue.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
14
Results
Mapping of Rad9 phosphorylation sites
Rad9 is extensively and constitutively phosphorylated even in cells that have not
been exposed to exogenous genotoxins (34-36). Rad9 is further inducibly
phosphorylated in response to DNA damage (34-38). Despite the extensive
phosphorylation, the role of this Rad9 modification remains unclear. To identify Rad9
phosphorylation sites, we subjected Rad9 tryptic peptides to both Edman
radiosequencing and mass spectrometry. To isolate sufficient quantities of Rad9 for
these analyses, we transiently over-expressed S-tag-Rad9 in K562 cells. S-tag-Rad9
interacts with Hus1 and Rad1 and is converted to a chromatin-bound form following
treatment with ionizing and UV radiation (39). Moreover, S-tag-Rad9 restores Chk1
activation in Rad9-deficient cells (data not shown). Collectively, these results
demonstrate that S-tag-Rad9 has biochemical properties similar to endogenous Rad9.
Using Edman radiosequencing, we identified four radiolabeled peptides, and
within these peptides we identified six phosphorylation sites in S-tag Rad9 (Fig. 1A and
1B, Ser272, Ser277, Ser 328, Ser375, Ser 380 and Ser387). Using mass spectrometry,
we identified five phosphorylated sites (Fig. 1A and 1B). Two of these sites (Ser277 and
Ser328) were also identified by radiosequencing. The three additional sites (Ser336,
Ser341 and Thr355) were new. All the phosphorylation sites were located within the
carboxyl-terminal 119-amino acid Rad9 tail. This portion of Rad9 has no homology with
PCNA, is not required for Rad9 to interact with Hus1 and Rad1 (14,40), but is involved in
nuclear transport of the protein due to the presence of a nuclear localization sequence
(40). Collectively, the radiosequencing and mass spectrometric analyses mapped nine
phosphorylation sites. Mutation of all nine sites to alanine blocked the Rad9 mobility shift
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
15
observed by SDS-PAGE (data not shown and Fig. 2), indicating that these studies
identified all the phosphorylation sites responsible for the mobility shift.
Expression of wild-type Rad9 and mutant Rad9 in Mrad9-/- ES cells.
To determine the role of Rad9 phosphorylation in cellular responses to
genotoxins, we created Mrad9-/- mouse ES cells stably expressing human wild-type
Rad9 or Rad9 mutants that could not be phosphorylated. (The full derivation of the
Mrad9-/- cells will be described elsewhere.) To insure that epitope tags would not affect
the function of Rad9, untagged Rad9 was used for the remaining studies presented
here. We generated the following untagged Rad9 mutants (Fig. 1C): (1) Rad9-S272A,
which converted Ser272, a previously identified DNA damage-induced phosphorylation
site, to Ala; (2) Rad9-9A, which converted all the phosphorylation sites identified in this
study (including Ser272) to Ala residues; and (3) Rad9-8A, which retained Ser272 but
converted all the remaining phosphorylation sites to Ala residues. The Mrad9-/- ES cells
did not express immunologically detectable full-length mouse Rad9, whereas the
corresponding wild-type ES cells did (Fig. 2). We also did not detect any amino-
terminally truncated Rad9 proteins by immunoblotting with antisera that recognize the
carboxyl terminus of Rad9, indicating that the targeted Mrad9 loci are not producing
truncated protein (data not shown). We transfected Mrad9-/- ES cells with expression
vectors encoding wild-type human Rad9 and the three mutant Rad9s (Rad9-S272A,
Rad9-8A, Rad9-9A). We identified two independently derived clonal cell lines expressing
each form of Rad9 at approximately equivalent levels (Fig. 2). It is worth noting that
although the transfected cell lines (expressing human Rad9s) appeared to express
slightly higher levels of Rad9 than did wild-type ES cells (expressing mouse Rad9), this
may result from different immunoreactivities of human Rad9 compared to mouse Rad9.
(The antibodies were generated to human Rad9 antigen.) Consistent with the role of
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
16
phosphorylation in Rad9's extensive mobility shift, the Rad9-8A and Rad9-9A mutants
exhibited dramatically increased mobility that was identical to the mobility of
phosphatase-treated Rad9 (data not shown).
Rad9 phosphorylation is not required for interaction with Hus1, Rad1, and Rad17.
Rad9 tail phosphorylation could affect Rad9 function at numerous points.
Because Rad9 associates with Hus1 and Rad1 as part of the 9-1-1 complex, and with
Rad17, which is the putative clamp loader for the 9-1-1 complex, we first assessed
whether tail phosphorylation was required for interactions with these binding partners.
We immunoprecipitated wild-type human Rad9 and the phospho-mutant Rad9 proteins
and immunoblotted the precipitates to detect interacting endogenous mouse Rad1,
Hus1, and Rad17 in the reconstituted ES cell clones. All the Rad9 mutants interacted
with Hus1, Rad1, and Rad17 as well as wild-type human and wild-type endogenous
mouse Rad9 did (Fig. 2), demonstrating that Rad9 phosphorylation is not required for 9-
1-1 complex formation or for the 9-1-1 complex to interact with the Rad17-containing
clamp loader.
Rad9 phosphorylation regulates survival following treatment with HU and UV light
Once we established that the Rad9 mutants associated with Hus1, Rad1, and
Rad17, we then asked whether the mutants affected cellular responses to genotoxins.
Mhus1-/- mouse embryonic fibroblasts (MEF) are sensitive to UV light and the replication
inhibitor hydroxyurea (HU) (13). We therefore examined the role of Rad9 and Rad9
phosphorylation in ES cell clonogenicity following treatment with these agents. Mrad9-/-
cells, wild-type ES cells (Rad9+/+), and two independently derived clones of Mrad9-/- cells
expressing human wild-type Rad9 were exposed to increasing concentrations of HU for
24 hr. Like Mhus1-/- MEFs, the Mrad9-/- ES cells were exquisitely sensitive to HU (Fig.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
17
3A). Expression of wild-type human Rad9 in the Mrad9-/- mouse ES cells restored the
HU resistance to near normal levels, demonstrating that human Rad9 complemented the
defect in Mrad9-/- cells (Fig. 3A).
We then assessed clonogenic survival of the Mrad9-/- cells expressing the three
Rad9 phosphorylation-deficient mutants treated with HU. Rad9 Ser272 was originally
reported to be phosphorylated in response to IR in an ATM-dependent manner (37). To
determine whether other genotoxins also induce Ser272 phosphorylation, we treated
Mrad9-/- ES cells expressing wild-type human Rad9 with IR, HU, or UV. Immunoblotting
of the Rad9 immunoprecipitates with anti-phospho-Rad9 (P-Ser272) antibodies revealed
that all three genotoxins induced Rad9 Ser272 phosphorylation (Fig. 3E). As a control,
we also treated Mrad9-/- ES cells expressing Rad9-S272A and found that Rad9-S272A
did not react with the anti-phospho-Rad9 (P-Ser272) antibodies after genotoxin
exposure. Thus, we examined whether Mrad9-/- ES cells expressing Rad9-S272A
exhibited increased sensitivity to HU. Surprisingly, these cells were indistinguishable
from Mrad9-/- cells expressing wild-type Rad9 (Fig. 3B). In contrast, Mrad9-/- cells
expressing Rad9-8A and Rad9-9A were nearly as sensitive to HU as the parental mutant
cells (Figs. 3C and 3D).
We also examined the role of Rad9 phosphorylation in cells treated with UV light.
Wild-type and Mrad9-/-ES cells, and Mrad9-/- ES cells expressing wild-type human Rad9
and the three Rad9 phosphorylation mutants were exposed to 15 J/m2 of UV light. Three
days later the plates were stained with Coomassie Blue. As expected, cells lacking Rad9
were very sensitive to UV damage (Fig. 3F, left panel). Introduction of wild-type or Rad9-
S272A restored the UV resistance to essentially wild-type ES cell levels (Fig. 3F, middle
panel). However, as with the HU experiments, cells expressing Rad9-8A and Rad9-9A
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
18
were as sensitive to UV as the Mrad9 null cells (Fig. 3F, right panel). Collectively, these
results demonstrated that Rad9 S272 phosphorylation is not required for cellular
responses that affect clonogenicity following replication inhibition or UV irradiation. In
contrast, one or more of the phosphorylation sites mutated in the 8A and 9A mutants is
required for normal responses to these genotoxic stresses.
Rad9 promotes cell survival following IR, but Rad9 phosphorylation plays only a minor
role
Mhus1-/- MEFs are sensitive to both UV light and HU, but have near wild-type
sensitivity to IR (45). Because Hus1 and Rad9, along with Rad1, form a heterotrimeric
complex (14-16), we anticipated that Mrad9-/- ES cells would be only modestly sensitive
to IR. Surprisingly, when assessed by clonogenic assays, the Mrad9-/- ES cells were far
more sensitive to IR than wild-type ES cells (Fig. 4A), and expression of wild-type Rad9
in the Mrad9-/- ES cells fully restored the wild-type levels of resistance (Fig. 4A). Like the
results found with HU and UV, cells expressing Rad9-S272A were no more sensitive to
IR than were cells expressing wild-type Rad9, suggesting that this phosphorylation site
is not required for any Rad9 function that contributes to ES cell survival (Fig. 4B). In
sharp contrast, cell lines expressing Rad9-8A and Rad9-9A exhibited near wild-type
sensitivity to lower doses (≤4 Gy) of IR and modestly increased sensitivity at 8 Gy IR
(Figs. 4C and 4D). Taken together, these results indicate that although Rad9 plays an
important role in the cellular response to IR, the Rad9 phosphorylation sites identified
here are not central to this function.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
19
Rad9 phosphorylation is not required for chromatin binding in response to genotoxic
stress
We then assessed which aspect of Rad9 function required phosphorylation. In
the current model for Rad9 function, the Rad17-RFC clamp loader loads the 9-1-1
complex onto DNA at sites of damage (1,3,29,30). In Fig. 2, we demonstrated that Rad9
phosphorylation did not affect interaction with Hus1 and Rad1, the other members of the
9-1-1 complex, or with Rad17. Because these interactions remained intact, we then
asked whether Rad9 chromatin binding, a Rad17-dependent event (12), required Rad9
phosphorylation. To streamline further biochemical characterizations for the studies that
follow, we selected a representative clone from each transfected cell line. We treated
wild-type cells, Mrad9-/- cells, and Mrad9-/- cells expressing Rad9 and the three Rad9
phosphorylation mutants with UV or ionizing radiation. The cells were sequentially
extracted with a low salt buffer to first remove unbound Rad9 and then with a high salt
buffer that dislodges chromatin-bound Rad9 (28). Endogenous Rad9 inducibly bound to
chromatin after UV and IR treatment (Fig. 5). Likewise, wild-type Rad9 and Rad9-S272A
also bound chromatin after genotoxic stress, showing that Ser272 phosphorylation is not
required for chromatin binding. Similarly, Rad9-8A and Rad9-9A also inducibly
associated with chromatin after genotoxic stress. However, we did note a slightly higher
level of chromatin-bound Rad9 in the 8A and 9A mutants even in the absence of DNA
damage, suggesting that there may a slightly higher level of basal genomic stress in
these cells. Taken together, these data indicate that none of the phosphorylation sites
targeted in this study is required for genotoxin-induced chromatin binding.
Rad9 is not required for IR-induced Chk2 or S-phase checkpoint activation
Rad9, Hus1 and Rad1 function early in a checkpoint signaling cascade. To
determine what checkpoint signaling pathways lie downstream of Rad9 and are
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
20
dependent upon Rad9 phosphorylation, we examined genotoxin-induced Chk1 and
Chk2 activation, and downstream events that require Chk1 and Chk2. Chk2, which is
phosphorylated and activated by PIKKs in response to genotoxins (8,9), undergoes a
phosphorylation-dependent mobility shift detected by SDS-PAGE. Wild-type and Mrad9-
/- ES cells were exposed to increasing doses of IR, and the cell lysates were
immunoblotted for Chk2 (Fig. 6A). At both low and high doses we did not see any effect
of Rad9 status on IR-induced Chk2 phosphorylation. Therefore, even though the Mrad9-/-
cells were more sensitive to IR, they did not have a Chk2 activation defect.
IR activates an S-phase checkpoint that slows DNA synthesis and blocks firing of
late DNA replication origins. An ATM-activated signaling pathway regulates this IR-
induced S-phase response, and mutations in ATM disrupt this checkpoint and cause a
characteristic phenotype of radio-resistant DNA synthesis (RDS), in which cells
synthesize DNA even in the presence of DNA damage (46-48). We treated cells with
increasing doses of IR or with UV light and examined their ability to synthesize DNA
(Fig. 6B). As expected, the UV-induced S-phase checkpoint was disrupted in cells
lacking Mrad9. In contrast, the IR-induced S-phase checkpoint did not require Rad9 and
was fully reversed by caffeine, a known inhibitor of ATM (49-51). Collectively, these
results demonstrate that Rad9 is not involved in the RDS response.
Rad9 is not required for IR-, UV-, or HU-induced p53 accumulation
We also examined whether Rad9 was required for accumulation of the DNA
damage-responsive transcription factor p53. After genotoxic stress, p53 undergoes
posttranslational modifications, including phosphorylation by Chk2 and ATM, which
participate in p53 stabilization and activation of p53 target genes (52). Basal p53 levels
were low in untreated wild-type cells and slightly higher in untreated Mrad9 null cells
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
21
(Fig. 7), consistent with the results observed in Mhus1-/- MEFs (13). When treated with
UV or ionizing radiation, wild-type and Mrad9-/- ES cells accumulated similar levels of
p53 after 2 and 8 hr (Fig. 7), demonstrating that Rad9 is not required for p53 stabilization
in response to these genotoxic stresses.
Rad9 and its phosphorylation are required for Chk1 activation
Replication inhibition and UV irradiation trigger ATR- and Hus1-dependent Chk1
phosphorylation on Ser345 (13,33), which is essential for Chk1 activation (2,53). To
assess whether Rad9 and its phosphorylation also participate in Chk1 activation, wild-
type, Mrad9-/- ES cells, and Mrad9-/- ES cells expressing wild-type human Rad9 or the
three Rad9 phosphorylation mutants were treated with HU, UV, or increasing amounts of
IR. All three stimuli triggered robust Chk1 Ser345 phosphorylation in wild-type ES cells
(Fig. 8A and 8B). In contrast, UV- and IR-induced Chk1 phosphorylation was eliminated,
and HU-induced phosphorylation was dramatically reduced in Mrad9-/- cells. We also
noted that the mobility of the residual phospho-Chk1 in HU-treated Mrad9-/- cells was
altered compared to wild-type cells (Fig. 8 and data not shown). We then asked whether
Rad9 phosphorylation was required for Chk1 activation. Wild-type Rad9 restored Chk1
phosphorylation after genotoxic damage, as did the S272A mutant (Fig. 8C and 8D).
Consistent with the defects in survival, Rad9-8A and Rad9-9A did not restore Chk1
phosphorylation, indicating that Rad9 tail phosphorylation is required for downstream
signaling from the 9-1-1 complex to Chk1 in response to genotoxic stress.
Rad9 and its phosphorylation are required for UV-induced S-phase checkpoint activation
Chk1 participates in the genotoxin-activated S-phase checkpoint (54).
Consequently, we asked whether Rad9 phosphorylation was required for S-phase arrest
following DNA damage. We UV irradiated wild-type ES cells, Mrad9-/- ES cells, and
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
22
Mrad9-/- ES cells expressing Rad9-S272A, Rad9-8A, and Rad9-9A. Forty minutes after
UV irradiation 3H-thymidine was added to detect DNA synthesis, and the cells were
incubated for an additional 20 minutes prior to harvest. The level of 3H-thymidine
incorporation decreased in wild-type cells to approximately 60% of control, whereas the
3H-thymidine incorporation in Mrad9-/- cells was only slightly inhibited by UV irradiation
(Fig. 9), indicating that the S-phase checkpoint requires Rad9. Expression of wild-type
Rad9 in the Mrad9-/- ES cells reconstituted the checkpoint, as did the S272A mutant. In
contrast, both Rad9-8A and Rad9-9A incorporated 3H-thymidine at levels similar to
Mrad9 null cells (Fig. 9). Taken together, these results demonstrate that Rad9
phosphorylation at Ser272 is not required for the S-phase checkpoint; however, the
additional phosphorylation sites in the Rad9 tail are required for this cellular response.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
23
Discussion
Mapping Rad9 phosphorylation sites
Several Rad9 phosphorylation sites have been identified previously (28,34-38).
Our analysis confirmed these sites (except Tyr28 and Thr292) and found four additional
sites (Ser341, Ser375, Ser380 and Ser387). Even though Ser272 is a DNA damage-
inducible site, we identified this site in the present radiosequencing analysis, possibly
because the cells were radiolabeled with large amounts of radiolabel, a procedure that
activates p53 and generates DNA damage (55,56). Alternatively, Ser272 may also be
phosphorylated by ATR. ATR is activated during unperturbed S-phase progression in
Xenopus egg extracts (57), raising the possibility that this site is phosphorylated at low
levels during S-phase. In contrast, based on Rad9 mobility studies, the remaining sites
that we identified are phosphorylated in the absence of exogenous genotoxic stress,
suggesting that they are constitutively phosphorylated; however, we cannot rule out the
possibility that one or more may be inducibly phosphorylated. Nonetheless, our results
clearly demonstrate that these phosphorylation sites play essential role(s) in Chk1
activation and highlight the role of the Rad9 tail in the Chk1 activation process.
Rad9 is also phosphorylated on Thr292, a site that is preferentially
phosphorylated in mitotic cells (35). Even though our radiosequencing analysis extended
through the peptide containing Thr292, we did not observe phosphorylation of this site.
Because we used an exponentially growing cell population, which contains few mitotic
cells, this site may have escaped our analysis and was not included in our studies of
Rad9 function in intact cells.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
24
Additionally, Rad9 is inducibly phosphorylated on Tyr28 in a c-Abl-dependent
manner following DNA damage (38), although neither of our phosphorylation site
analyses detected Tyr28 phosphorylation. Possible explanations for this discrepancy are
that the radiolabeling procedure did not cause sufficient damage to induce this
phosphorylation or that the techniques used here were not sufficiently sensitive to detect
this phosphorylation event.
The Rad9 tail
All Rad9 orthologs identified to date possess a carboxyl-terminal tail. The tail is
the least conserved portion of Rad9, and yet all Rad9 proteins identified have tails of
varying length, suggesting that it has a role in Rad9 function. Despite the fact that all
Rad9s possess a tail, its role in Rad9 function is unknown. A recent study in S. pombe
demonstrated that the S. pombe Rad9 tail was required for checkpoint signaling (18).
However, this study did not address how loss of the tail affected interaction of Rad9 with
its interacting partners or Rad9 chromatin binding. Another analysis revealed that the
human Rad9 tail contains a nuclear localization sequence, which is required for Rad9 to
enter the nucleus (40). This observation raises the possibility that mutation of the tail
phosphorylation sites might impair nuclear localization. However, both Rad9-8A and
Rad9-9A inducibly associated with chromatin following UV and IR treatment, indicating
that they were present in the nucleus at the time of damage. Thus, the present studies
allowed us to examine the role of the tail in checkpoint activation without deleting the tail,
a manipulation that would affect other aspects of Rad9 function.
Rad9 and Rad9 phosphorylation in checkpoint activation
In many respects the Mrad9-/- ES cells and Mhus1-/- MEFs have nearly identical
genotoxin- triggered cellular responses. First, both Hus1- and Rad9-deficient cells were
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
25
sensitive to UV and HU (45). Second, the Mhus1-/- MEFs and the Mrad9-/- ES cells were
defective in Chk1 activation in response to UV, HU, and IR (13). Third, neither Hus1 nor
Rad9 was required for Chk2 activation or p53 accumulation induced by these genotoxins
(13). Fourth, the Hus1- and Rad9-deficient cells have defects in the S-phase checkpoint
triggered by bulky DNA lesions but they do not have defects in IR-induced S-phase
checkpoint (58). Collectively, these results further support the model that Hus1 and
Rad9 function together in the 9-1-1 signaling complex.
In contrast, survival following treatment with IR differed significantly between the
cell lines. The Hus1 null MEFs were not substantially more sensitive to IR than paired
control cells (45), and we anticipated that Mrad9-/- ES cells would also be relatively
insensitive to IR. Surprisingly, our studies demonstrated that Mrad9-/- ES cells were very
sensitive to IR, much like S. pombe lacking Rad9, suggesting that Rad9 either has roles
outside the 9-1-1 complex or that in some cell types the 9-1-1 complex may participate in
cellular responses to IR. Because IR-induced Chk2 and S-phase checkpoint activation
were normal in the Mrad9-/- ES cells, the increased sensitivity to IR could not be
explained by loss of these checkpoint signaling events. In contrast, IR-induced Chk1
activation was disrupted in the Mrad9 null ES cells, raising the possibility that the Chk1
signaling deficiency might contribute to the survival defect. However, cells expressing
Rad9-8A and Rad9-9A also exhibited the Chk1-activation defect. Yet these cells
exhibited near-wild-type IR sensitivity at lower doses of IR, demonstrating that, although
Rad9 is important for survival following IR, Chk1 activation does not contribute to
survival at lower doses of IR (≤4 Gy). Because the cells expressing Rad9-8A and Rad9-
9A were more sensitive to a higher dose of IR (8 Gy) than cells expressing wild-type
Rad9, it is possible that IR-induced Chk1 may contribute to cell survival under these
conditions. These results demonstrate that Rad9 plays a key role in IR-treated cells that
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
26
depends only modestly on Rad9 tail phosphorylation. However, this role does not
depend on Chk1 or Chk2 activation, suggesting that Rad9 must participate in other IR-
activated cellular responses that are independent of the Rad9 phospho-tail.
Rad9 is inducibly phosphorylated on Ser272 in response to IR (37) as well as HU
and UV radiation (Fig. 3). In the case of IR, this phosphorylation has been implicated in
the regulation of G1 to S phase progression (37). Because ES cells lack a strong DNA
damage-induced G1 checkpoint, we were unable to examine this phenotype. However,
we did show that Ser272 phosphorylation was not required for survival following
treatment with HU, UV, or ionizing radiation. We also showed that HU- and UV-induced
Chk1 activation and UV-induced activation of the S-phase checkpoint did not require
Ser272 phosphorylation. Collectively, these results suggest that Rad9 Ser272
phosphorylation is not required for the cellular and biochemical responses assayed in
our ES cell studies; however, they do not rule out a role for this phosphorylation in other
cell types or in the G1 checkpoint.
In contrast, one or more of the additional Rad9 tail phosphorylation sites was
essential for Chk1 activation. Because Rad9 mutants lacking tail phosphorylation sites
form a 9-1-1 complex that associates with chromatin in response to genotoxins, these
results demonstrated that 9-1-1 clamp binding was not sufficient for Chk1 activation.
Moreover, they also demonstrate that the mutation of the phosphorylation sites did not
disrupt the overall structure of Rad9; it is possible, however, that the defects in Rad9-8A
and Rad9-9A are due to conformational alterations rather than changes in
phosphorylation status. The precise roles of the non-SQ phosphorylation sites in Chk1
activation are not yet understood. Although many of these sites are likely
phosphorylated constitutively (based on their retarded mobility when analyzed by SDS-
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
27
PAGE (Fig. 2)), we cannot exclude the possibility that some of the sites may also be
inducibly phosphorylated. The phosphorylation site sequences provide few clues as to
potential kinases that might phosphorylate Rad9. Six of the sites are followed by pro
residues, suggesting that pro-directed kinases such as MAP kinase or cyclin-dependent
family members might target these sites, whereas two sites are embedded in regions
rich in negatively charged amino acids, raising the possibility that they may be
phosphorylated by CKII. Interestingly, while this work was under review, a study was
published implicating protein kinase C (PKC) d in Rad9 phosphorylation (59). However,
because the published study did not identify Rad9 phosphorylation sites and none of the
sites reported here is a consensus PKC d site, it is unclear whether PKC d targets the
sites identified in the present work. Studies to identify the roles of specific
phosphorylation sites and the kinases that phosphorylate them are presently underway.
A checkpoint signaling model
The 9-1-1 complex and ATR, as a complex with its binding partner ATRIP (60),
are independently recruited to DNA lesions (12,31,32), where these complexes
cooperatively activate Chk1. Importantly, both the 9-1-1 complex and ATR are required
for Chk1 activation. The present studies provide additional insight into this requirement.
Our results demonstrate that formation of the 9-1-1 complex and genotoxin-induced
chromatin binding of that complex are not sufficient for Chk1 activation; the Rad9
phospho-tail plays a critical role in Rad9-dependent Chk1 activation. Because neither
ATR nor Chk1 co-precipitates with Rad9 (P.R.-M., unpublished observations), the role of
the Rad9 phospho-tail may be to facilitate the transient assembly of a multisubunit
complex that participates in Chk1 activation (Fig. 10). Although additional studies are
required to determine the precise role of Rad9 phosphorylation, our investigations
highlight for the first time the key role played by the Rad9 phospho-tail in genotoxin-
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
28
induced checkpoint activation.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
29
Acknowledgements
We wish to thank Drs. J. Chen, S. Kaufmann, and J. Sarkaria for thoughtful
discussions, insights, and critical reading of the manuscript. We thank B. Madden and
Dr. D. McCormick of the Mayo Protein Core Facility for performing the Edman
radiosequencing analyses. We also thank Dr. J. Chen for providing the anti-Chk2
antiserum and Dr. E. Lee for sharing the anti-phospho-Rad9 (P-Ser272) antibodies. This
work was funded by NIH grants CA84321 (L.M.K.), GM52493 (H.B.L.), CA89816
(H.B.L.), the Mayo Clinic Foundation (L.M.K.) and the Magnus Ehrnrooth and Oskar
Öflund foundations (P.R.-M.).
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
30
References
1. Lowndes, N.F., and Murguia, J.R. (2000) Curr. Opin. Genet. Dev. 10, 17-25
2. Zhou, B.B., and Elledge, S.J. (2000) Nature 408, 433-439
3. Melo, J., and Toczyski, D. (2002) Curr. Opin. Cell Biol. 14, 237-245
4. Bartek, J., and Lukas, J. (2001) Curr. Opin. Cell Biol. 13, 738-747
5. O'Connell, M.J., Walworth, N.C., and Carr, A.M. (2000) Trends Cell Biol. 10, 296-
303
6. Dasika, G.K., Lin, S.-C.J., Zhao, S., Sung, P., Tomkinson, A., and Lee, E.Y.-H.P.
(1999) Oncogene 18, 7883-7899
7. Abraham, R.T. (2001) Genes Dev. 15, 2177-2196
8. Bartek, J., Falck, J., and Lukas, J. (2001) Nat. Rev. Mol. Cell Biol. 2, 877-886
9. McGowan, C.H. (2002) Bioessays 24, 502-511
10. Rhind, N., and Russell, P. (2000) J. Cell Sci. 113, 3889-3896.
11. Walworth, N.C., and Bernards, R. (1996) Science 271, 353-356
12. Zou, L., Cortez, D., and Elledge, S.J. (2002) Genes Dev. 16, 198-208.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
31
13. Weiss, R.S., Matsuoka, S., Elledge, S.J., and Leder, P. (2002) Curr. Biol. 12, 73-
77.
14. Burtelow, M.A., Roos-Mattjus, P.M., Rauen, M., Babendure, J.R., and Karnitz,
L.M. (2001) J. Biol. Chem. 276, 25903-25909.
15. Lindsey-Boltz, L.A., Bermudez, V.P., Hurwitz, J., and Sancar, A. (2001) Proc.
Natl. Acad. Sci. USA 98, 11236-11241
16. Kaur, R., Kostrub, C.F., and Enoch, T. (2001) Mol. Cell. Biol. 12, 3744-3758
17. Hang, H., and Lieberman, H.B. (2000) Genomics 65, 24-33
18. Kaur, R., Kostrub, C.F., and Enoch, T. (2001) Mol. Biol. Cell 12, 3744-3758
19. Griffith, J.D., Lindsey-Boltz, L.A., and Sancar, A. (2002) J. Biol. Chem. 277,
15233-15236
20. Thelen, M.P., and Fidelis, K. (1999) Cell 96, 769-770
21. Caspari, T., Dahlen, M., Kanter-Smoler, G., Lindsay, H.D., Hofmann, K.,
Papadimitriou, K., Sunnerhagen, P., and Carr, A.M. (2000) Mol. Cell. Biol. 74,
1254-1262
22. Kelman, Z. (1997) Oncogene 14, 629-640.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
32
23. Tsurimoto, T. (1998) Biochim. Biophys. Acta 1443, 23-39
24. Waga, S., and Stillman, B. (1998) Annu. Rev. Biochem. 67, 721-751
25. Rauen, M., Burtelow, M.A., Dufault, V.M., and Karnitz, L.M. (2000) J. Biol. Chem.
275, 29767-29771
26. Parker, A.E., Van de Weyer, I., Laus, M.C., Verhasselt, P., and Luyten, W.H.
(1998) J. Biol. Chem. 273, 18340-18346
27. Bermudez, V.P., Lindsey-Boltz, L.A., Cesare, A.J., Maniwa, Y., Griffith, J.D.,
Hurwitz, J., and Sancar, A. (2003) Proc. Natl. Acad. Sci. USA 100, 1633-1638
28. Burtelow, M.A., Kaufmann, S.H., and Karnitz, L.M. (2000) J. Biol. Chem. 275,
26343-26348
29. Venclovas, C., and Thelen, M.P. (2000) Nucleic Acids Res. 28, 2481-2493
30. Caspari, T., and Carr, A.M. (2002) Curr. Biol. 12, R105-107
31. Melo, J.A., Cohen, J., and Toczyski, D.P. (2001) Genes Dev. 15, 2809-2821
32. Kondo, T., Wakayama, T., Naiki, T., Matsumoto, K., and Sugimoto, K. (2001)
Science 294, 867-870.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
33
33. Liu, Q., Guntuku, S., Cui, X.S., Matsuoka, S., Cortez, D., Tamai, K., Luo, G.,
Carattini-Rivera, S., DeMayo, F., Bradley, A., Donehower, L.A., and Elledge, S.J.
(2000) Genes Dev. 14, 1448-1459
34. Volkmer, E., and Karnitz, L.M. (1999) J. Biol. Chem. 274, 567-570
35. St. Onge, R.P., Besley, B.D., Park, M., Casselman, R., and Davey, S. (2001) J.
Biol. Chem. 276, 41898-41905.
36. St. Onge, R.P., Udell, C.M., Casselman, R., and Davey, S. (1999) Mol. Biol. Cell
10, 1985-1995
37. Chen, M.J., Lin, Y.T., Lieberman, H.B., Chen, G., and Lee, E.Y. (2001) J. Biol.
Chem. 276, 16580-16586.
38. Yoshida, K., Komatsu, K., Wang, H.G., and Kufe, D. (2002) Mol. Cell. Biol. 22,
3292-3300
39. Roos-Mattjus, P., Vroman, B.T., Burtelow, M.A., Rauen, M., Eapen, A.K., and
Karnitz, L.M. (2002) J. Biol. Chem. 277, 43809-43812
40. Hirai, I., and Wang, H.G. (2002) J. Biol. Chem. 277, 25722-25727
41. Ward, I.M., Wu, X., and Chen, J. (2001) J. Biol. Chem. 276, 47755-47758.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
34
42. Duyao, M.P., Auerbach, A.B., Ryan, A., Perischetti, F., Barnes, G.T., McNeil,
S.M., Ge, P., Vonsattel, J.P., Gusella, J.F., Joyner, A.L., and MacDonald, M.E.
(1995) Science 269, 407-410
43. Dietrich, P., Dragatsis, I., Xuan, S., Zeitlin, S., and Efstradiadis, A. (2000) Mamm.
Genome 11, 196-205
44. Horton, R.M., Cai, Z.L., Ho, S.N., and Pease, L.R. (1990) Biotechniques 8, 528-
535
45. Weiss, R.S., Enoch, T., and Leder, P. (2000) Genes Dev. 14, 1886-1898
46. Rotman, G., and Shiloh, Y. (1999) Oncogene 18, 6135-6144
47. Kastan, M.B., and Lim, D.S. (2000) Nat. Rev. Mol. Cell Biol. 1, 179-186
48. Falck, J., Mailand, N., Syljuasen, R.G., Bartek, J., and Lukas, J. (2001) Nature
410, 842-847
49. Sarkaria, J.N., Busby, E.C., Tibbetts, R.S., Roos, P., Taya, Y., Karnitz, L.M., and
Abraham, R.T. (1999) Cancer Res. 59, 4375-4382
50. Blasina, A., Price, B.D., Turenne, G.A., and McGowan, C.H. (1999) Curr. Biol. 9,
1135-1138
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
35
51. Zhou, B.B., Chaturvedi, P., Spring, K., Scott, S.P., Johanson, R.A., Mishra, R.,
Mattern, M.R., Winkler, J.D., and Khanna, K.K. (2000) J. Biol. Chem. 275, 10342-
10348.
52. Wahl, G.M., and Carr, A.M. (2001) Nat. Cell Biol. 3, E277-286
53. Zhao, H., and Piwnica-Worms, H. (2001) Mol. Cell. Biol. 21, 4129-4139
54. Guo, N., Faller, D.V., and Vaziri, C. (2002) Cell Growth Differ. 13, 77-86
55. Yeargin, J., and Haas, M. (1995) Curr. Biol. 5, 423-431
56. Hu, V.W., Heikka, D.S., Dieffenbach, P.B., and Ha, L. (2001) FASEB J. 15, 1562-
1568
57. Hekmat-Nejad, M., You, Z., Yee, M.C., Newport, J.W., and Cimprich, K.A. (2000)
Curr. Biol. 10, 1565-1573
58. Wiess, R.S., Leder, P., and Vaziri, C. (2003) Mol. Cell. Biol. 23, 791-803
59. Yoshida, K., Wang, H.G., Miki, Y., and Kufe, D. (2003) EMBO J. 22, 1431-1441
60. Cortez, D., Guntuku, S., Qin, J., and Elledge, S.J. (2001) Science 294, 1713-
1716
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
36
Figure legends
Figure 1
Mapping of Rad9 phosphorylation sites. (A) The amino terminal two-thirds of Rad9
contains two predicted PCNA folds that form a PCNA-like structure. The carboxyl
terminus, dubbed the "tail," contains all the mapped Rad9 phosphorylation sites. (B)
Rad9 phosphorylation sites, their surrounding sequences, and the methods used to
identify the site. (C) Three different mutants were generated: Ser272 mutated to alanine
(S272A), all mapped phosphorylation sites except for Ser272 mutated to alanine (8A)
and all mapped phosphorylation sites mutated to alanine (9A).
Figure 2
Characterization of Rad9-expressing mouse ES cell lines. Mrad9-/- ES cells were
transfected with plasmids encoding wild-type Rad9, Rad9-S272A, Rad9-8A, and Rad9-
9A. Stable cell lines expressing approximately equivalent amounts of wild-type and
mutant Rad9 proteins were identified. Two independent clones of each Rad9-expressing
cell line were characterized alongside wild-type and Mrad9-/- ES cells. The clone names
are indicated below the panels. Rad9 was immunoprecipitated from equivalent amounts
of total protein and sequentially immunoblotted for Rad1, Hus1, Rad17, and Rad9. The
lower band in the Rad1 blot is a nonspecific band.
Figure 3
Rad9 and Rad9 phosphorylation are required for resistance to HU and UV. (A-D) ES
cells were plated, and four hours later treated with the indicated concentrations of HU.
After 24 h, the cells were washed to remove HU and fresh media was added. The cells
were cultured for 14 days, and colonies were stained and counted. The experiment has
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
37
been repeated three times with identical results. The results shown are from a
representative experiment. Each point shows the mean of three samples, with error bars
showing standard deviation. Two independent clones expressing wild-type or Rad9
mutants were compared. The clone names are indicated in parentheses. (E) ES cells
expressing wild-type Rad9 or Rad9-S272A were treated with 50 Gy IR, 10 mM HU, or 40
J/m2 UV, and cells were lysed 1 h later. Rad9 immunoprecipitates were sequentially
immunoblotted for phospho-Rad9 (P-Ser272) and total Rad9. (F) ES cells were treated
with 15 J/m2 the following day. The cells were stained 72 h later. One representative cell
line per clone is shown in this experiment. The clone names are indicated in
parentheses.
Figure 4
Rad9 but not Rad9 phosphorylation is required for resistance to IR. (A-D) ES cell lines
were plated and treated with the indicated doses of IR four hours later. The experiment
has been repeated three times with identical results. The results shown are from a
representative experiment. Error bars indicate standard deviation. Two independent
clones expressing wild-type or Rad9 mutants were compared. Clone names are
indicated in parentheses.
Figure 5
Rad9 phosphorylation is not required for chromatin binding in response to IR or UV.
Wild-type, Mrad9-/- and Mrad9-/- cells expressing wild-type and Rad9 mutants were left
untreated or treated with either 50 Gy IR (panel A) or 40 J/m2 UV radiation (panel B).
One hour later the cells were harvested and fractionated into low and high salt
(chromatin-bound) fractions to separate free and chromatin-bound Rad9. Rad9 was
immunoprecipitated from equivalent amounts of total protein, except for wild-type ES
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
38
cells, where Rad9 was immunoprecipitated from twice the amount of protein to facilitate
detection of Rad9. The immunoprecipitates were immunoblotted for Rad9. The clones
used in both experiments were wild-type Rad9 (B3-24), Rad9-S272A (3-9-15), Rad9-8A
(D3-8) and Rad9-9A (E2-2).
Figure 6
Rad9 is not required for IR-induced Chk2 or S-phase checkpoint activation. (A) Wild-type
or Mrad9-/- ES cells were treated with the indicated doses of IR. One hour later the cells
were harvested and lysed. Chk2 was immunoblotted with an anti-Chk2 antiserum. (B)
Wild-type or Mrad9-/- cells were pretreated with nothing or 3 mM caffeine and treated
with 2, 4, or 10 Gy IR or with 30 J/m2 UV light. [methyl-3H]thymidine was added 40 min
later, and the cultures were incubated for an additional 20 min. 3H-thymidine
incorporation is plotted as percent of incorporation compared to untreated cells. Values
are the mean from three independent experiments. Error bars indicate standard error.
Figure 7
Rad9 is not required for genotoxin-induced p53 accumulation. Wild-type and Mrad9-/-
cells were treated with 20 Gy IR or 20 J/m2 UV radiation and harvested 2 or 8 hours
later. Equal amounts of total protein were separated by SDS-PAGE and sequentially
immunoblotted for p53 and JNK (as a loading control).
Figure 8
Rad9 and Rad9 phosphorylation are required for genotoxin-induced Chk1 activation.
Wild-type or Mrad9-/- ES cells were treated with 10 mM HU or 50 J/m2 UV radiation
(panel A) or with increasing amounts of IR (panel B). Mrad9-/- and Mrad9-/- cells
expressing wild-type Rad9 and Rad9 mutants were treated with 10 mM HU or 50 J/m2
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Roos-Mattjus et al
39
UV radiation (panel C) or with 50 Gy IR (panel D). One hour later the cells were
harvested and lysed, and equal amounts of total protein were sequentially
immunoblotted with an antibody recognizing phosphorylated Chk1 (P-Ser345) and an
antibody recognizing total levels of Chk1. The clones used were wild-type Rad9 (B3-24),
Rad9-S272A (3-9-15), Rad9-8A (D3-8) and Rad9-9A (E2-2).
Figure 9
The UV irradiation-induced S-phase checkpoint requires Rad9 and Rad9
phosphorylation. Wild-type, Mrad9-/- and reconstituted ES cells were plated in 96-well
plates. The clones used were wild-type Rad9 (B3-24), S272A (3-9-15), 8A (D3-8) and 9A
(E2-2). The cells were treated with 20 J/m2 and [methyl-3H]thymidine was added 40 min
later and the cells were incubated for an additional 20 min. [methyl-3H]thymidine
incorporation is plotted as percent of incorporation relative to untreated cells. Values are
the mean of three independent experiments. Error bars indicate standard deviation.
Figure 10
Model for the role of the Rad9 phospho-tail in Chk1 activation. See text for discussion.
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
���
� �������� � ��������
�����
�����
�����
����
����
� �� ������� ���
����
����
�����
��
� �������� � ��������
�����
� �� ������� ���
���
�����
���
� �������� � ��������
�����
�����
�����
�����
����
����
� �� ������� ���
����
����
�����
��
���
� �������� � ��������
����
����
����
����
���
���
� �� ������� ���
���
���
�����
�
amino acid sequence identified by
������ ��S�� �� ������� �S�� �� ����������� �S��� �� ����������� ��S��� �������� ��S �������� ��T��� �������� ��S��� �� ������� ��S�� �� ������� �S� �� �
�
Figure 1
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Rad9
Rad1
Hus1Rad17
Mra
d9-/-
wt E
S w
t Rad
9
S272
A
8A 9A
B3-24
B7-3
3-3-7 3-9-15D
3-6D
3-8E
2-2
E6-1Clone names:
Figure 2
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Figure 3A B
µM Hydroxurea
0 50 100 150
% S
urvi
val
1
10
100
Mrad9-/-
wt ES wt Rad9 (B3-24) wt Rad9 (B7-3)
µM Hydroxurea
0 50 100 150
% S
urvi
val
1
10
100
Mrad9-/-
wt Rad9 (B3-24) S272A (3-3-7)S272A (3-9-15)
C D
µM Hydroxurea
0 50 100 150
% S
urvi
val
1
10
100
Mrad9-/-
wt Rad9 (B3-24) 8A (D3-6)8A (D3-8)
µM Hydroxurea
0 50 100 150
% S
urvi
val
1
10
100
Mrad9-/-
wt Rad9 (B3-24) 9A (E2-2) 9A (E6-1)
by guest on March 17, 2018 http://www.jbc.org/ Downloaded from
control 15 J/m2
Mrad9 -/-
wt ES
control 15 J/m2
wt Rad9 (B3-24)
S272A(3-9-15)
control 15 J/m2
8A (D3-8)
9A (E2-2)
�
- IR UVHU
total Rad9
P-Rad9 (S272)
wt Rad9 S272A
- IR UVHU
�
Figure 3
by guest on March 17, 2018 http://www.jbc.org/ Downloaded from
Figure 4A B
Gy
0 2 4 6 8
% S
urvi
val
0.001
0.01
0.1
1
10
100
Mrad9-/-
wt ESwt Rad9 (B3-24)wt Rad9 (B7-3)
Gy
0 2 4 6 8
% S
urvi
val
0.001
0.01
0.1
1
10
100
Mrad9-/-
wt Rad9 (B3-24) S272A (3-3-7) S272A (3-9-15)
C D
Gy
0 2 4 6 8
% S
urvi
val
0.001
0.01
0.1
1
10
100
Mrad9-/-
wt Rad9 (B3-24) 8A (D3-6) 8A (D3-8)
Gy
0 2 4 6 8
% S
urvi
val
0.001
0.01
0.1
1
10
100
Mrad9-/-
wt Rad9 (B3-24) 9A (E2-2) 9A (E6-1)
by guest on March 17, 2018 http://www.jbc.org/ Downloaded from
- + - + - + - + - + - + IR
Rad9
chromatin-bound fraction
Mrad9-
/-
wt ES
wt Rad
9
S272A
8A 9A
Mrad9-
/-
wt ES
wt Rad
9
S272A
8A 9A
- + - + - + - + - + - + UV
Rad9
chromatin-bound fraction
B
A
Figure 5
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
- 2 5 20 - 2 5 20IR, GyWild-type Mrad9-/-
Chk2
0
50
100
3H-T
hym
idin
be
inco
rpo
rati
on
(
% o
f co
ntr
ol)
IR
Mrad9-/-
Wild type
2 Gy 4 Gy 10 Gy 2 Gy UV+Caff
A
B
Figure 6
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
Mrad9-/- wt ES
- 2 8 2 8 - 2 8 2 8 h IR UV IR UV
p53
JNK
Figure 7
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
- HUUV - HUUVwt ES Mrad9-/-
P-Chk1 (S345)
total Chk1
- 10 25 50 - 10 2550
wt ES Mrad9 -/-
Gy
P-Chk1 (S345)
total Chk1
- HU UV- HU UV- HU UV- HU UV - HU UV M
rad9-
/-
wt Rad
9
S272A
8A 9A
P-Chk1 (S345)
total Chk1
- IR - IR- IR- IR- IR 50 GyMra
d9-/-
wt Rad
9
S272A
8A 9A
P-Chk1 (S345)
total Chk1
A
B
C
D
Figure 8
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
3 H-th
ymid
ine
inco
rpor
atio
n (%
of c
ontr
ol)
40
60
80
100
Mrad9-/- wt ES wt R9 S272A 8A 9A
UV 20 J/m2
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from
L. Johnson, Steve Naylor, Howard B. Lieberman and Larry M. KarnitzPia Roos-Mattjus, Kevin M. Hopkins, Andrea J. Oestreich, Benjamin T. Vroman, Kenneth
signalingPhosphorylation of human Rad9 is required for genotoxin-activated checkpoint
published online April 21, 2003J. Biol. Chem.
10.1074/jbc.M301544200Access the most updated version of this article at doi:
Alerts:
When a correction for this article is posted•
When this article is cited•
to choose from all of JBC's e-mail alertsClick here
by guest on March 17, 2018
http://ww
w.jbc.org/
Dow
nloaded from