MOLECULAR AND CELLULAR BIOLOGY, May 2007, p. 3367–3377
Copyright © 2007, American Society for Microbiology. All Rights Reserved.
Vol. 27, No. 9
Function of a Conserved Checkpoint Recruitment Domain
in ATRIP Proteins?
Heather L. Ball,1Mark R. Ehrhardt,2Daniel A. Mordes,1Gloria G. Glick,1
Walter J. Chazin,2and David Cortez1*
Department of Biochemistry1and Departments of Biochemistry and Chemistry and Center for Structural Biology,2
Vanderbilt University, Nashville, Tennessee 37232
Received 29 November 2006/Returned for modification 29 December 2006/Accepted 16 February 2007
The ATR (ATM and Rad3-related) kinase is essential to maintain genomic integrity. ATR is recruited to
DNA lesions in part through its association with ATR-interacting protein (ATRIP), which in turn interacts with
the single-stranded DNA binding protein RPA (replication protein A). In this study, a conserved checkpoint
protein recruitment domain (CRD) in ATRIP orthologs was identified by biochemical mapping of the RPA
binding site in combination with nuclear magnetic resonance, mutagenesis, and computational modeling.
Mutations in the CRD of the Saccharomyces cerevisiae ATRIP ortholog Ddc2 disrupt the Ddc2-RPA interaction,
prevent proper localization of Ddc2 to DNA breaks, sensitize yeast to DNA-damaging agents, and partially
compromise checkpoint signaling. These data demonstrate that the CRD is critical for localization and optimal
DNA damage responses. However, the stimulation of ATR kinase activity by binding of topoisomerase binding
protein 1 (TopBP1) to ATRIP-ATR can occur independently of the interaction of ATRIP with RPA. Our results
support the idea of a multistep model for ATR activation that requires separable localization and activation
functions of ATRIP.
ATR (ATM and Rad3-related) kinase is a protein kinase
that coordinates cellular responses to genotoxic stress. ATR
activation occurs primarily in S phase due to replication stress
induced by DNA-damaging agents or replication inhibitors.
More specifically, ATR activation is stimulated when the rep-
lication machinery encounters a DNA lesion and becomes
uncoupled (the helicase continues to unwind DNA while the
polymerase becomes stalled at the site of DNA damage) (9).
The critical factor that promotes ATR activation is believed
to be the accumulation of RPA (replication protein A)-coated
single-stranded DNA (ssDNA) (11, 33, 43). At least two sep-
arate checkpoint complexes accumulate in distinct foci that
colocalize with RPA. Rad17, a PCNA-like clamp loader pro-
tein, is recruited to RPA-ssDNA and loads the Rad9-Rad1-
Hus1 checkpoint clamp at the junction of double-stranded and
single-stranded DNA (4, 14, 53). Independently, ATR is re-
cruited by ATR-interacting protein (ATRIP), which binds the
RPA-ssDNA that accumulates at DNA lesions (3, 15, 37, 52).
ATRIP is required for ATR function, and mutation of either
ATR or ATRIP causes the same phenotypes (3, 12). The strict
requirement for ATRIP is conserved in Schizosaccharomyces
pombe (Rad3 and Rad26), Saccharomyces cerevisiae (Mec1 and
Ddc2/Lcd1/Pie1), and Xenopus laevis (xATR and xATRIP)
(13, 38, 41, 51). An N-terminal domain of ATRIP binds RPA-
ssDNA and is necessary for stable ATR-ATRIP localization to
damage-induced nuclear foci (3, 25).
The ATR signaling pathway is currently viewed as an im-
portant target for the development of cancer therapies (10, 22,
24, 32, 34). However, the mechanism by which ATR is acti-
vated remains unclear. Localization to sites of DNA damage or
replication stress has been suggested to be essential and per-
haps sufficient to promote ATR signaling. However, mutations
in ATRIP that disrupt the stable RPA-ATRIP interaction and
impair the accumulation of ATR-ATRIP complexes in DNA-
damage-induced foci have minimal effects on ATR activation
and signaling (3, 25). Furthermore, topoisomerase binding
protein 1 (TopBP1) was recently discovered to stimulate ATR
kinase activity, suggesting regulation by a means other than
localization (28). To clarify the functions of ATRIP, RPA, and
TopBP1 in mediating ATR-dependent checkpoint response we
have performed a series of biochemical and genetic experi-
ments in human and yeast systems. We report structural and
functional data that support a model for ATR activation in
which two separable ATRIP activities—localization and acti-
vation—cooperate to promote ATR signaling.
MATERIALS AND METHODS
Yeast strains. All strains used in this study are described in Table 1. Myc-Ddc2
and Myc-Ddc2 mutant strains were generated by expressing mutants from a
centromeric plasmid under the control of the endogenous DDC2 promoter in
strain DMP2995/1B (MATa sml1D::KanMX4 ddc2D::KanMX4) (38). GFP-
Ddc2?N was generated in JK8-1 (36) by use of the delitto perfetto system (47).
Strain yHB244 was generated by expressing RNR3 by use of pBAD79 and
deleting DDC2 by use of pGEM499 in the JKM179 strain (30). Myc-DDC2 and
Myc-Ddc2?N were expressed in strain yHB244 from the pNML1 centromeric
RPA-ssDNA and RPA binding. The 14 kDa and 70 kDa RPA subunits were
tagged with an His6epitope tag (45). RPA was purified from Escherichia coli by
use of nickel affinity chromatography followed by Superdex fractionation. A
20-pmol volume of biotin-labeled 69-bp single-stranded oligonucleotide was
bound to streptavidin beads and incubated with binding buffer (10 mM Tris [pH
7.5], 100 mM NaCl, 10% glycerol, 0.02% Igepal CA-630, 10 ?g/ml bovine serum
albumin) alone or with a 4 M excess of RPA in binding buffer. The RPA-ssDNA-
streptavidin beads were washed three times with binding buffer prior to use.
Hemagglutinin-ATRIP (HA-ATRIP) fragments were generated using in vitro
* Corresponding author. Mailing address: Department of Biochem-
istry, Vanderbilt University, 613 Light Hall, 23rd @ Pierce Ave., Nash-
ville, TN 37232. Phone: (615) 322-8547. Fax: (615) 343-0704. E-mail:
?Published ahead of print on 5 March 2007.
transcription/translation (Promega) and added to recombinant His-RPA or His-
RPA-ssDNA beads in binding buffer, and RPA was isolated using His-Select
(Sigma Aldrich) or ssDNA-Sepharose beads. Proteins bound to beads were
washed with binding buffer three times, eluted, and separated by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) prior to blotting.
Kinase assays. ATR kinase assays were performed essentially as described
previously (28) with the following alterations. HA-ATRIP and Flag-ATR ex-
pression vectors were transfected into 293T cells and the ATR-ATRIP com-
plexes purified by immunoprecipitation. Kinase reactions were performed with
the antibody-linked ATR-ATRIP complex. Recombinant TopBP1 and RPA
heterotrimer was purified from E. coli. The Phas1 substrate was purchased from
A.G. Scientific. The in situ Rad53 autophosphorylation assay following denatur-
ation/renaturation was performed as previously described (39).
Cell culture. All cells were grown in Dulbecco modified Eagle medium
(DMEM)–7.5% fetal bovine serum. Plasmid transfections were performed with
Lipofectamine 2000 (Invitrogen). The ATRIP wild-type (WT) and ATRIP?N
293T cells were generated by retroviral infection and selection (3). ATRIP small
interfering RNA (siRNA) transfections and immunofluorescence methods were
also performed as described previously (3).
Protein purification and antibodies. Tandem affinity purification (TAP)-
Rfa1 was purified from soluble yeast extracts as described previously (1, 17).
RPA70N was kindly provided by Cheryl Arrowsmith in a pET15b vector
(Novagen) and was grown on defined M9 medium supplemented with15N-
NH4Cl and D-glucose and purified over nickel-nitrilotriacetic acid. ?H2AX
(Upstate), Myc9E10 (Covance), and HA.11 (Covance) were purchased from
the indicated companies. Rad53 antibody was a gift from Stephen Elledge.
Rfa1 antibody was a gift from Steven Brill.
NMR analysis and homology modeling. Nuclear magnetic resonance (NMR)
spectra were collected on ?100 ?M15N-RPA70N in a buffer containing 2 mM
?-mercaptoethanol, 50 mM NaCl, and 20 mM Tris–d11at pH 7.4. ATRIP1-107
and ATRIP54-70 were added at a four- to sixfold molar excess to maximize the
bound-state population of the observed component, RPA70N. NMR experi-
ments were performed at 25°C using Bruker AVANCE 500 MHz or 600 MHz
NMR spectrometers equipped with a 5 mm single axis z-gradient Cryoprobe.
Two-dimensional, gradient-enhanced15N-1H heteronuclear single-quantum cor-
relation (HSQC) spectra were recorded with 1,024 complex datum points in the
1H dimension and 96 complex points in15N dimension.1H and15N backbone
NMR assignments for RPA70N were kindly provided by Cheryl Arrowsmith.
Chromatin immunoprecipitation (ChIP) analysis. HO expression in strains
yHB245 (vector), yHB246 (Myc-DDC2), and yHB247 (Myc-Ddc2?N) containing
galactose-inducible expression of HO endonuclease was performed as described
previously (30). Cells were cross-linked with 1% formaldehyde, lysed, and son-
icated to generate DNAs with an average size of 500 bp. Myc-Ddc2 protein-DNA
complexes were isolated using Myc9E10 antibody and protein G-Sepharose
beads, washed extensively, and eluted from beads. Cross-links were reversed by
overnight incubation at 65°C. DNA was precipitated and amplified using the
following primers specific to a region adjacent to the HO break site (HO-A or
HO-B) or to a region of SMC2: HO-A1 (5?-CTCATCTGTGATTTGTGG-3?),
HO-A2 (5?-AGAGGGTCACAGCACTAATACA-3?), HO-B1 (5?-CCAGAT
TTGTATTAGACGAGGGACGGAGTGA-3?), HO-B2 (5?-AGAGGGTCA
CAGCACTAAATACAGCTCGAAT-3?), SMC2-1 (5?-AAGAGAAACTTTA
GTCAAAACATGGG-3?), and SMC2-2 (5?-CCATCACATTATACTAACT
Characterization of ATRIP and RPA binding domains. Our
analysis started with an examination of the binding of ATRIP
fragments to RPA in the absence and presence of ssDNA.
Previous analysis of ATRIP identified at least three domains:
an N-terminal RPA-ssDNA binding domain, a dimerization
domain predicted to fold into a coiled-coil structure, and a
C-terminal ATR-interaction domain (2, 3). HA-tagged, intact
ATRIP and ATRIP fragments spanning the various domains
were generated using a coupled transcription/translation sys-
tem. These ATRIP proteins were added to purified His-tagged
RPA heterotrimer bound to ssDNA displayed on Sepharose
beads or His-tagged RPA heterotrimer bound to nickel beads.
After incubation and washing, the bound ATRIP proteins were
detected by Western blot analysis. We found that all ATRIP
fragments containing the N-terminal 107 amino acids bound
well to RPA-ssDNA and His-RPA in the absence of DNA
(Fig. 1A and B). Thus, the first 107 amino acids of ATRIP
contain a protein-protein interaction domain that mediates
binding to the RPA heterotrimer.
HA-ATRIP fragments lacking this N-terminal RPA binding
domain of ATRIP (N-RBD) were deficient in binding RPA-
ssDNA and His-RPA (Fig. 1A and B). Long exposures of
Western blots did show a small degree of association of these
ATRIP fragments with RPA. In fact, all of the protein frag-
ments that we tested, including a fragment of Brca1, bound
weakly to RPA-ssDNA (Fig. 1A). These interactions may re-
flect additional RPA-ssDNA binding domains on ATRIP, as
has been previously reported (37).
To determine which subunit of the RPA heterotrimer inter-
TABLE 1. Yeast strains used in this study
Strain Description-genotypeReference or source
MATa ade2-1 can1-100 his3-11,15 leu2-3,112 trp1-1 ura3 sml1::KanMX4
DMP2995/1B ?p1220: Myc-URA3-CEN?
DMP2995/1B ?pNML1: Myc-DDC2-URA3-CEN?
DMP2995/1B ?pHB126: Myc-ddc2?N-URA3-CEN?
DMP2995/1B ?pHB157: Myc-ddc2D12KD13K-URA3-CEN?
DMP2995/1B ?pHB155: Myc-ddc2NAAIRS14-19-URA3-CEN?
MATa his3-1 leu2-0 met15-0 ura3-0 Rfa1-TAP
Rfa1-TAP ?pNML1: Myc-DDC2-URA-CEN?
Rfa1-TAP ?pHB126: Myc-ddc2?N-URA-CEN?
Rfa1-TAP ?pHB157: Myc-ddc2D12KD13K-URA3-CEN?
Rfa1-TAP ?pHB155: Myc-ddc2NAAIRS14-19-URA3-CEN?
ho? MATa hml?::ADE1 hmr?::ADE1 ade1-100 leu2-3,112 lys5 trp1::hisG?
JKM179 ddc2?::LEU2 ?pBAD79: RNR3-TRP1-CEN?
yHB244 (GAL-HO, ?ddc2) ?p1220: Myc-URA3-CEN?
yHB244 (GAL-HO, ?ddc2) ?pNML1: Myc-DDC2-URA3-CEN?
yHB244 (GAL-HO, ?ddc2) ?pHB126: Myc-ddc2?N-URA3-CEN?
mat? can1 ade2 trp1 his3 ura3 leu2 lys5 cyh2 ade3::GalHO adh4::HO site::HIS3
yJK8-1 ddc2?N-GFP ?pBAD79: RNR3-TRP1-CEN?
3368 BALL ET AL.MOL. CELL. BIOL.
acts with ATRIP, we purified recombinant RPA domains in-
dividually or in combination as His-tagged proteins (Fig. 1C).
Using pull-down assays with in vitro-translated ATRIP pro-
teins, we found that full-length HA-ATRIP or the isolated
N-RBD bound only to RPA fragments containing the N-ter-
minal RPA70 oligonucleotide/oligosaccharide (OB) fold do-
main (RPA70N) (Fig. 1D). No significant binding to other
RPA domains was detectable, and no binding of ATRIP?N
(ATRIP108-791) protein lacking the N-RBD to any RPA frag-
ment (Fig. 1D) was detectable in this assay. Taken together,
these data suggest that the ATRIP N-RBD interacts directly
with the 70N domain of RPA.
A conserved acidic domain in the ATRIP N terminus inter-
acts with the basic cleft of the RPA70N OB fold. The specific
residues involved in the interaction of RPA and ATRIP were
identified using an NMR chemical-shift-mapping approach.
This strategy involves monitoring NMR chemical shifts of one
protein over the course of a titration with a binding partner.
Measurement of the RPA15N-1H-HSQC NMR spectrum of
15N-enriched RPA70N as ATRIP N-RBD after titration in
solution showed that only a subset of the RPA70N signals was
affected (Fig. 2A). The observation of effects in the fast-to-
intermediate-exchange regimen on the NMR timescale sug-
gests that binding was occurring with a dissociation constant in
the low micromolar range. When the chemical shifts are
mapped onto the crystal structure of RPA70N (7), it is appar-
ent that ATRIP N-RBD interacts within the basic cleft of
RPA70N (Fig. 2B).
Initial insight into characteristics of the RPA70N binding
site of the ATRIP N-RBD was obtained from sequence anal-
ysis. When the N termini of five ATRIP orthologs were
aligned, minimal sequence similarity was observed, with the
notable exception of a small, acidic region spanning approxi-
mately 15 amino acids (Fig. 2C). On the basis of putative
electrostatic complementarity, we hypothesized that this small
acidic region made contact with the basic surface in the cleft of
the OB fold of RPA70N. To test this hypothesis, an ATRIP
peptide spanning the conserved acidic region (ATRIP54-70)
was synthesized, and the RPA70N titration was repeated. The
titration with the peptide perturbed most of the same residues
FIG. 1. ATRIP N terminus binds RPA70N in vitro. HA-tagged, full-length ATRIP or ATRIP fragments generated using a coupled transcrip-
tion/translation system were incubated with single-stranded DNA bound to Sepharose beads in the presence (?) or absence (?) of purified RPA
(A) or His-RPA bound to nickel beads (B). After washing, the bound proteins were eluted, separated by SDS-PAGE, blotted, and probed with
HA antibody. Input (In) (10%) data are included for comparison. (C and D) Purified, recombinant His-tagged RPA domains were added to in
vitro translation reaction mixtures containing HA-ATRIP, HA-ATRIP?N (C), or HA-ATRIP1-107 (D). Protein complexes were isolated using
nickel beads, separated by SDS-PAGE, blotted, and probed using an HA antibody. (E) Diagram of RPA heterotrimer. Black bars above protein
segments indicate protein interaction domains.
VOL. 27, 2007 FUNCTIONS OF RPA AND TOPBP1 IN ATR-ATRIP SIGNALING 3369
as the titration with ATRIP N-RBD, indicating that the
ATRIP peptide binds in the same manner within the basic cleft
of RPA70N (Fig. 2D). In addition, we analyzed the binding of
ATRIP1-107 containing charge-reversal mutations at positions
D58 and D59 to RPA70N by use of NMR. This mutant binds
much more weakly than wild-type ATRIP.
The basic cleft of RPA70N has been shown to bind peptides
that can mimic DNA in a manner similar to the binding of
ssDNA to the A and B domains of RPA70 (5, 7). RPA70N
binds an acidic helical peptide of p53, and the crystal structure
of the p53 peptide bound in the cleft was determined by a
method previously reported (7). Alignment of this p53 peptide
with ATRIP54-70 indicates significant homology between the
two peptides (Fig. 2E). Therefore, the crystal structure of
RPA70N bound to the p53 peptide was used to generate a
homology model for the ATRIP peptide-RPA70N interaction.
The strategy involved using the backbone coordinates of the
RPA70N and the p53 peptide along with the side chains of
FIG. 2. A conserved acidic domain in the ATRIP N terminus interacts with the basic cleft of RPA70N. (A) The15N-1H-HSQC NMR spectrum
of15N-enriched-RPA70N in the absence (blue) and presence (red) of ATRIP1-107. (B) RPA70N residues perturbed (blue) upon addition of
ATRIP1-107 mapped onto the crystal structure of RPA70N (PDB accession 2B3G). (C) Sequence alignment of the conserved acidic region in the
N terminus of five ATRIP orthologues. (D) The15N-1H-HSQC NMR spectrum of15N-enriched RPA70N acquired in the absence (blue) and
presence (red) of ATRIP54-70. (E) Alignment of the p53 and ATRIP peptides used in homology modeling. (F) Space-filling diagram of RPA70
and ATRIP55-66 (red), with the residues of RPA70N in the ATRIP binding pocket colored blue. (G) Predicted electrostatic interactions between
basic RPA70N basic residues K88 and R41 with ATRIP acidic residues D58 and D59.
3370 BALL ET AL.MOL. CELL. BIOL.
RPA70N from the crystal structure. The p53 amino acid side
chains were replaced with the ATRIP amino acid 55 to 66 side
chains, and the best fit of the ATRIP peptide into the con-
strained RPA70N was determined using ROSETTA (Fig. 2F)
The model predicts that there are several specific electro-
static interactions between the acidic residues on ATRIP and
the basic residues on RPA. In particular, the absolutely con-
served aspartic acid residues D58 and D59 of ATRIP are likely
to make contact with R41 and K88 of RPA70N (Fig. 2G). Basic
residues at these positions in RPA are highly conserved. The
NMR data and molecular modeling are fully consistent with
the previously described pull-down experiment results, indicat-
ing that the N terminus of ATRIP binds directly to RPA70N.
Importantly, these data create a structural framework within
which specific ATRIP-RPA binding mutants can be designed
and used for functional analysis.
The N-RBD of human ATRIP is conserved in the S. cerevi-
siae ATRIP ortholog Ddc2. The functional consequences of
disrupting the ATRIP-RPA interaction in human cells were
previously characterized using an ATRIP mutant lacking
the entire N-RBD (ATRIP?N). Unlike wild-type ATRIP,
ATRIP?N has a severe defect in localizing to damage- or
replication stress-induced nuclear foci (3). Despite this local-
ization defect, cells depleted of endogenous ATRIP and com-
plemented with ATRIP?N exhibit normal ATR-dependent
signaling following DNA damage (3). The only checkpoint
defect that we have uncovered in the ATRIP?N-expressing
cells is a slight delay in recovery from hydroxyurea (HU)-
induced stalling of replication (H. L. Ball, unpublished data).
The use of RNA interference in endogenous ATRIP synthesis
is not 100% effective, and the results are variable from cell to
cell. In addition, the level of retrovirally expressed ATRIP or
ATRIP?N after integration of the retroviral vector is variable
and not equivalent to the endogenous protein levels (3). These
technical limitations to performing genetic analysis of human
cell cultures may confound our ability to detect the phenotypic
consequences of abrogating the ATRIP-RPA interaction.
Therefore, we sought to examine the physiological role of the
ATRIP-RPA interaction in the genetic system of another or-
The RPA binding domain of ATRIP is N terminal to the
predicted coiled-coil domain. To determine whether the equiv-
alent region of yeast ATRIP (Ddc2) mediates binding to yeast
RPA70 (Rfa1), we deleted 42 amino acids N terminal to the
predicted coiled-coil domain of Ddc2 (Ddc2?N) (Fig. 3A).
Binding of Ddc2 and Ddc2?N to Rfa1 was assayed using co-
immunoprecipitation. Myc-tagged Ddc2 (WT) or Ddc2?N
(?N) was expressed from plasmids under the control of the
endogenous DDC2 promoter in ?ddc2 cells containing HA-
tagged Mec1. Myc-Ddc2 and Myc-Ddc2?N were immunopre-
cipitated using a Myc antibody, and the coassociated Rfa1 and
Mec1 were assayed by Western blotting. As expected, both
Mec1 and Rfa1 were coimmunoprecipitated with Ddc2 (Fig.
3B). In comparison, Rfa1 association with Ddc2?N was greatly
reduced, although Ddc2?N continued to bind Mec1 (Fig. 3B).
These data suggest that the N terminus of Ddc2 is required for
a stable Ddc2-Rfa1 interaction. The amount of Rfa1-associ-
ated Ddc2 was not altered by exposing cells to UV damage,
suggesting that the Ddc2-Rfa1 interaction may not be regu-
lated by DNA damage (Fig. 3B). However, these experiments
utilized soluble extracts, so it is possible that the interaction
with DNA-bound RPA is regulated.
Sequence alignment of human and yeast ATRIP indicates
that the small acidic region in the ATRIP N-RBD is conserved
(Fig. 2C). The homology model generated from NMR data
predicts that the absolutely conserved aspartic acid residues in
this region (D12 and D13 in Ddc2) could make contacts with
conserved basic amino acids on yeast Rfa1. Therefore, we
hypothesized that mutating these residues would disrupt Ddc2-
Rfa1 binding. To test this hypothesis a Ddc2 mutant was gen-
erated with aspartic acid-to-lysine charge-reversal mutations in
these two aspartic acids (Ddc2DK) (Fig. 3C). In addition, a
mutant (Ddc2N14) was generated replacing Ddc2 residues 14
to 19 with a peptide (NAAIRS) that is known to adopt a
helical conformation (Fig. 3C). Myc-Ddc2, Myc-Ddc2?N, Myc-
Ddc2DK, or Myc-Ddc2N14 was expressed in a yeast strain
containing TAP-tagged Rfa1. TAP purification of Rfa1 protein
complexes indicated they contained Myc-Ddc2 whether or not
cells were pretreated with methyl methanesulfonate (MMS)
FIG. 3. Ddc2-Rfa1 interaction requires the conserved acidic region
of Ddc2. (A) Schematic diagram of ATRIP, RPA-binding mutant
ATRIP (ATRIP?N), Ddc2, and Ddc2?N. Locations of predicted
coiled-coil domains (gray) are indicated. (B) ?ddc2 HA-MEC1 yeast
containing myc-vector (Vec), myc-Ddc2 (WT), or myc-Ddc2?N (?N)
were exposed to 0 (?) or 60 (?) J/m2UV and harvested 1 h later. Myc
immunoprecipitates from soluble extracts were separated by SDS-
PAGE, blotted, and probed with Myc, Rfa1, and HA antibodies.
(C) Alignment of conserved acidic region in the N terminus of ATRIP
and Ddc2 and schematic of mutations generated in this region of Ddc2.
(D) Yeasts containing TAP-Rfa1 and Ddc2 (WT), Ddc2?N (?N),
Ddc2DK (DK), or Ddc2N14 (N14) were damaged with 0.01% MMS
(?) or left untreated (?) and harvested 1 h later. Cells were lysed, and
TAP-Rfa1 was isolated using immunoglobulin G beads. TAP protein
complexes were separated by SDS-PAGE and Western blotted using
Myc (Ddc2) and Rfa1 antibodies.
VOL. 27, 2007 FUNCTIONS OF RPA AND TOPBP1 IN ATR-ATRIP SIGNALING3371