[Cell Cycle 2:4, 384-396, July/August 2003]; ©2003 Landes Bioscience

Report

FHA Domain-Mediated DNA Checkpoint Regulation of Rad53

*Correspondence to: David F. Stern; Dept. of Pathology; Yale University School of Medicine; P.O. Box 208023; New Haven, Connecticut 06510 USA; Tel.: 203.785.4832; Fax: 203.785.7467; Email: [email protected]

is t

rd

N

Received 04/22/03; Accepted 05/23/03

fo

address: The Wistar Institute, Philadelphia, PA, 19104

Previously published online as a Cell Cycle Paper in Press at: http://www.landesbioscience.com/journals/cc/toc.php?volume=2&issue=4

e.

†Present

Saccharomyces cerevisiae Rad53 is a protein kinase central to the DNA damage and DNA replication checkpoint signaling pathways. In addition to its catalytic domain, Rad53 contains two forkhead homology-associated (FHA) domains (FHA1 and FHA2), which are phosphopeptide binding domains. The Rad53 FHA domains are proposed to mediate the interaction of Rad53 with both upstream and downstream branches of the DNA checkpoint signaling pathways. Here we show that concurrent mutation of Rad53 FHA1 and FHA2 causes DNA checkpoint defects approaching that of inactivation or loss of RAD53 itself. Both FHA1 and FHA2 are required for the robust activation of Rad53 by the RAD9-dependent DNA damage checkpoint pathway, while an intact FHA1 or FHA2 allows the activation of Rad53 in response to replication block. Mutation of Rad53 FHA1 causes the persistent activation of the RAD9-dependent DNA damage checkpoint pathway in response to replicational stress, suggesting that the RAD53-dependent stabilization of stalled replication forks functions through FHA1. Rad53 FHA1 is also required for the phosphorylation-dependent association of Rad53 with the chromatin assembly factor Asf1, although Asf1 itself is apparently not required for the prevention of DNA damage in response to replication block.

ot

Department of Pathology; Yale University School of Medicine; New Haven, Connecticut

rib ut io n.

ABSTRACT

Marc F. Schwartz† Soo-Jung Lee Jimmy K. Duong Seda Eminaga David F. Stern*

INTRODUCTION

cell cycle checkpoint genes, DNA damage, replication, Rad53 gene product, Rad9 protein, phosphopeptide, FHA domain

Rad53, a budding yeast homolog of the mammalian tumor suppressor Chk2, is a protein kinase that is phosphorylated and activated in response to genotoxic stress.1-3 This process is mediated by at least two signaling networks, the DNA damage checkpoint pathway (DDC), and the DNA replication checkpoint pathway (DRC) (reviewed in refs. 4,5). The DDC functions throughout the cell cycle, including S-phase, whereas the DRC is restricted to cells undergoing active replication. Activation of the DDC during G1 or G2 causes the characteristic arrest of cell cycle progression at the G1/S or G2/M transitions.6-8 Activation of the DDC and/or DRC during S phase causes a delay of progression through S, and prevents elongation of the mitotic spindle.9-13 DNA checkpoint activation also leads to the transcriptional upregulation of a number of DNA repair and replication genes, as well as some of the components of the DNA checkpoint pathways themselves.11,14-18 In addition, components of the DNA checkpoint pathways may themselves be involved in DNA repair, or may directly regulate this process.19-22 Several proteins have been implicated in the initiation of the DDC signal, but their precise function in the recognition of DNA damage is poorly understood. The prototypical checkpoint protein, Rad9, integrates DNA damage signals from at least two incoming pathways that are thought to be involved in the recognition and/or initial processing of the DNA lesion itself. The first incoming DDC sensor pathway, active throughout the cell cycle, requires the RFC-like Rad24/Rfc2-4 and PCNA-like Rad17/Mec3/Ddc1 complexes in parallel with the phosphoinositol-(3’) kinase-like protein kinase (PIKK) Mec1 and its functional partner Ddc2 (reviewed in ref. 5). The second incoming DDC sensor pathway, active in late S and G2, involves the Mre11/Rad50/Xrs2 complex in conjunction with the PIKK Tel1.23 These two pathways converge on Rad9, resulting in the heavy MEC1/TEL1dependent [S/T]Q phosphorylation of Rad9 in response to DNA damage.24-27 Rad9, in turn, mediates the PIKK-dependent activation of the parallel checkpoint effector pathways involving the protein kinases Rad53 and Chk1.13,28 The major Rad9 phosphorylation sites induced by DNA damage are required for both the direct interaction of phosphorylated Rad9 with Rad53, and the subsequent activation of Rad53.27 These Rad9 phosphorylation sites are not essential for the DNA damage-induced phosphorylation of Chk1, suggesting that Rad9 regulates these two downstream pathways through different means.27

©

20 03

La

nd

es

ie

sc

io

B

This work was supported by US Public Health Service grant R01CA82257, and USAMRMC DAMD17-98-1-8272. S-JL and SE were supported by the USAMRMC Predoctoral Training Program in Breast Cancer Research DAMD17-99-1-9461. MFS was additionally supported by USAMRMC DAMD17-99-1-9460.

nc

KEY WORDS

384

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

A

B

Figure 1. Genotoxin-induced phosphorylation of FHA-mutant forms of Rad53 in cycling cells. (A) Diagram of Rad53, with FHA1 (amino-terminal) and FHA2 (carboxyl-terminal) flanking the catalytic kinase domain. Positions of the alanine substitutions that generated a defective FHA1 (R70 and N107), FHA2 (NVS655-657), or kinase domain (K227 D339) are marked (*). (B) Western blot analysis of Rad53 in TCA lysates prepared from asynchronous cultures of U960-5C bearing the indicated forms of Rad53-expressing CEN plasmids. The cultures were either mock treated (0), or treated with 200 mM HU (H) for 90 minutes or 0.1% MMS (M) for 30 minutes.

During S phase, DNA damage can also lead to activation of the RAD9-independent DRC, which also leads to the MEC1-dependent phosphorylation and activation of Rad53.13 In addition to DNA damage, the DRC generally responds to conditions that interfere with DNA replication, such as exposure to the ribonucleotide reductase (RNR) inhibitor hydroxyurea (HU).9,11 Although the precise order or mechanism of function of this pathway (or pathways) is not known, several proteins or protein complexes have also been implicated in the initiation of the DRC signal, including DNA Polε, Dpb11/Sld2, the RFC-like Rad24/Rfc2-5 and Chl12/Rfc2-5 complexes, Tof1, Mrc1, and the PIKK Mec1 complexed with Ddc2 (reviewed in ref. 5) Of these, Mrc1 is suggested to fulfill a role similar to Rad9 in the DRC as it is phosphorylated in response to DNA checkpoint activation in a MEC1-dependent manner, and is required for the DNA checkpoint-induced RAD9-independent phosphorylation of Rad53.29 However, unlike Rad53 and Rad9, a direct interaction between Mrc1 and Rad53 has yet to be demonstrated. In addition to its central protein kinase catalytic domain, Rad53 contains two copies of another conserved sequence motif, the Forkhead associated (FHA) domain (Fig. 1A). The FHA domain was first recognized as a loosely conserved sequence motif found in a subset of the Forkhead-family of transcription factors, as well as other proteins involved in DNA-related functions.30 A physiological function as a protein binding domain was first suggested for the FHA domain by the observation that a portion of the Arabidopsis kinase-associated protein phosphatase (KAPP) containing a FHA domain is required for its interaction with the receptor-like kinase (RLK). Furthermore, kinase-inactivating mutations of RLK also abolished the KAPP-RLK interaction, implicating phosphorylation in this interaction.31 The FHA domain was directly implicated as a phosphorylationdependent protein binding domain with the observation that Rad53 FHA2 is sufficient for the interaction of Rad53 with phosphorylated Rad9.24 This hypothesis was confirmed by the demonstration of direct FHA domain-phosphopeptide binding in vitro.32 FHA domains interact with phosphothreonine (pT) containing peptides,32-34 but a www.landesbioscience.com

very weak phosphotyrosine-binding activity has also been reported for Rad53 FHA2.35 Phosphopeptide library binding experiments determined that FHA domain binding affinity is influenced by the sequence surrounding the pT, especially by the third following residue, and have suggested consensus binding motifs for some of the FHA domains.33,34,36 For example, Rad53 FHA1 displays an in vitro binding preference for peptides containing pTxxD, whereas the Rad53 FHA2 prefers peptides containing pTxx[I/L].33,34,36 However, DNA damage-induced Rad9 phosphorylation sites required for the interaction of Rad9 with Rad53 in vivo do not completely conform to the consensus sequences established in vitro,27suggesting that physiological FHA domain binding sites may exhibit previously unappreciated characteristics. Alternatively, and by analogy to the phosphorylation-dependent regulated degradation of the CDK inhibitor, Sic1,37 the use of suboptimal FHA domain binding sites in vivo may sharpen the signaling threshold and allow a more finely tuned initiation of Rad53 activation. As Rad53 is phosphorylated and activated by both the RAD9-dependent DDC and the RAD9-independent DRC, Rad53 is able to integrate afferent signals from these two pathways. Rad53 FHA2 is required in vivo for RAD9-dependent Rad53 phosphorylation and function, but not for the phosphorylation of Rad53 in response to replicational stress.24 Thus, we had proposed that the Rad53 FHA domains provide a modular basis for the ability of Rad53 to integrate the DDC and DRC signals, wherein FHA1 couples Rad53 to the DRC, presumably through interaction with a phosphoprotein, and FHA2 couples Rad53 to the DDC through interactions with phosphorylated Rad9.24 However, recombinant Rad53 FHA1 is also sufficient for binding phosphorylated Rad9 in fusion protein pull down experiments,32 suggesting that FHA1 may also participate in the regulation of Rad53 by the DDC. Therefore, to determine the physiological contribution of the Rad53 FHA domains to the regulation of Rad53 by both the DDC and DRC signaling pathways, we examined the effects of FHA1 and/or FHA2 mutation in vivo.

Cell Cycle

385

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

MATERIALS AND METHODS Plasmids. The plasmids used are described in Table 1. Proper construct structure was confirmed by restriction endonuclease digest and sequence analysis. The low copy (CEN) plasmid pRS316 RAD53, which allows the expression of RAD53 driven by its own promoter at levels indistinguishable from genomic RAD53, was described previously.14 Kinase defective rad53K227A D339A was described previously.38 To generate rad53R70A N107A, the mutations were introduced by two-step PCR into the PacI-BglII fragment of pRS316 RAD53. Appropriate fragments of the rad53 alleles were recombined through restriction endonuclease digests and ligations to yield the plasmids listed in Table 1. pRS316 RAD533xFLAG is described elsewhere (S-JL, MFS, JKD, DFS, in press). pRS316 RAD93xHA was described previously.27 To create pRS314 DUN113xmyc, the BclI-Pst1 region of the pRS314 DUN1 plasmid pZZ48 was replaced with a BclI-SmaI cut PCR amplified fragment encoding the carboxyl-terminus of Dun1 with a SmaI site introduced before the stop codon, the SmaI-SpeI fragment encoding the 13xmyc tag from pFA6a-13myc-His3MX6,39 and a Spel-Pst1 cut PCR amplified fragment containing DUN1 3’ UTR sequence with a SpeI site introduced immediately following the DUN1 stop codon. For the GST-FHA1 expression construct, sequence encoding Rad53 amino acids 1-197 were PCR amplified from a RAD53 and a rad53R70A N107A construct, and cloned into the BamHI and EcoRI sites of pGEX4T3. GST-Rad53 FHA2 fusion protein expression constructs, containing sequence encoding Rad53 amino acids 450-730, were described previously.24 Strains. The strains used are described in Table 2. These strains are in the W303 background.40 The essential requirement for RAD53 in rad53 strains was suppressed by either mutation or deletion of SML1.41,42 The rad53∆XB::HIS3 allele used to generate U960-5C41 does not delete the entire RAD53 ORF, and strains harboring this allele retain some coding sequences flanking the kinase domain. Rad53 antibodies, raised against a carboxyl-terminal fragment of Rad53,14 detect a truncated version of Rad53 in these strains (MFS, unpublished data). For this reason, key experiments documenting specific Rad53 FHA domain interactions were also performed in rad53∆ strains lacking the entire ORF, or confirmed with GST-Rad53 FHA domain fusion protein pull down experiments. Table 1

PLASMID LIST

Name

Source

yJKD103 was described previously.27 RAD9 was epitope-tagged by two-step allele replacement with the plasmid pRS306 RAD93xHA.27 Genomic sequences encoding the complete ORFs of SML1 and RAD53 were then deleted as described perviously.27 pRS306 rad9∆27 was used to delete RAD9 in U960-5C by two-step allele replacement. Where appropriate, genomic sequences encoding the entire DUN1 and ASF1 ORFs were replaced with kanR cassettes PCR amplified from the corresponding yeast deletion strains (Research Genetics). To epitope-tag Asf1, the tagging cassette from pFA6a-3HA-kanMX639 was PCR amplified with additional primer sequence that targeted the cassette to the 3’ end of the ASF1 ORF. Growth Conditions. Drop-out media was purchased from Bufferad. Synthetic and rich media were typically supplemented with adenine to 47.5 mg/L. For biochemical analyses, cultures were grown to early to mid-log phase. For cdc15-2 synchronization, early-log cultures grown at 23˚C were shifted to 37°C for three hours. As indicated in the figure legends, cultures were mock treated, treated with 0.1% MMS (Sigma) for 30–60 minutes, or treated with 200 mM HU (Sigma) for 90–120 minutes, washed, and used immediately or frozen in liquid nitrogen and stored at -80˚C. Genotoxin Sensitivity Assays. These assays were performed as described previously.27 Briefly, cultures of U960-5C bearing the appropriate plasmids were brought to similar densities, and serially diluted five-fold in a 96-well plate. A 48-pin inoculator was used to spot the diluted cultures onto rich media plates (YPAD), or YPAD plates containing HU or MMS. Sets of inoculated YPAD plates were UV irradiated in a prewarmed Stratalinker (Stratagene). The plate cultures were grown at 30˚C for 2–3 days. Assay for the S/M DNA Replication Checkpoint Arrest. This assay was performed as described previously.11 Briefly, cultures of U960-5C bearing the appropriate plasmids were synchronized in G1 with α-factor, washed, and released into rich media containing 200 mM HU. Spindles in samples taken at the indicated timepoints were visualized by indirect immunofluorescence using the rat anti-tubulin monoclonal YOL1/34 and a FITCconjugated anti-rat secondary antibody (Jackson Immunoresearch), and scored for spindle elongation. Assay for the G2/M DNA Damage Checkpoint Arrest. This assay was performed as described previously.27,43 Briefly, cultures of yJKD103 bearing the appropriate plasmids were grown to early-log phase at 23˚C in synthetic media, shifted into rich media, and synchronized with α-factor. Cells were sonicated briefly to reduce clumping, washed into prewarmed rich media,

pRS316 RAD53

Our laboratory

pRS316 RAD533xFLAG

Our laboratory

pRS316 rad53K227A D339A

Our laboratory

Name

Relevant Genotype

Source

This study

W303-1a

MATa ade2-1 can1-100 his3-11, 15 leu2-3, 112 trp1-1 ura3-1 rad5-535

R. Rothstein

pRS316

rad53R70A N107A

Table 2

STRAIN LIST

pRS316 rad53NVS655-657AAA

Our laboratory

pRS316 rad53R70A N107A NVS655-657AAA

This study

pRS315 RAD53

This study

DLY408

as W303-1a, cdc13-1 cdc15-2

T. Weinert

pRS315 rad53K227A D339A

This study

yJKD103

as DLY408, sml1∆::TRP1 rad53D::HIS3

Our laboratory

DLY418

as W303-1a, cdc15-2

T. Weinert

yMS509

as DLY418, RAD9-3xHA sml1∆::TRP1 rad53∆::HIS3

This study

pRS315

rad53R70A N107A

This study

pRS315 rad53NVS655-657AAA

Our laboratory

pRS315 rad53R70A N107A NVS655-657AAA

This study

pRS315 rad53K227A D339A R70A N107A

This study

1588-4C

W303-1a, RAD5

R. Rothstein

pRS315 rad53K227A D339A NVS655-657AAA

This study

U960-5C

as 1588-4C, sml1-1 rad53∆XB::HIS3

R. Rothstein

yMS665

as U960-5C, RAD93xHA

This study

pRS315 rad53K227A D339A R70A N107A NVS655-657AAA This study pRS316 RAD93xHA

Our laboratory

pRS314 DUN1

S. Elledge

yJKD429

as U960-5C, rad9∆

This study

pRS314 DUN113xmyc

This study

yJKD415

as U960-5C, dun1∆::kanR

This study

pGEX 4T Rad53 1-197

This study

yJKD409

as U960-5C, asf1∆::kanR

This study

pGEX 4T Rad53 1-197 R70A N107A

This study

yJKD303

as 1588-4C, asf1∆::kanR

This study

ASF1-3xHA::kanR

This study This study

pGEX Rad53 450-730

Our laboratory

yJKD403

as U960-5C,

pGEX Rad53 450-730 NVS655-657AAA

Our laboratory

yJKD301

as 1588-4C, ASF1-3xHA::kanR

386

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

and grown at 37˚C. Samples taken at the indicated times were fixed in 70% ethanol, and spotted onto poly-L-lysine (Sigma) coated slides with DAPI mounting media (Vector Laboratories). For every genotype, samples from two independently derived strains were scored in triplicate, with at least 100 cells in each count. Large-budded, mononucleate cells were counted as pre-anaphase cells, and large-budded, binucleate cells were counted as post-anaphase. Cells with gross nuclear or morphological defects, and cells terminally arrested in G1, as determined by size and morphology, were not scored. Immunoprecipitations and Immunoblotting. Immunoprecipitations and Western blotting were as described previously,27 with minor modifications. Briefly, lysates were prepared by mechanical disruption in TG (PBS + 1% Triton X-100, 10% glycerol, and phosphatase and protease inhibitors [Roche and Sigma]). 0.5–2 mg of lysate was used per immunoprecipitation with rabbit anti-Rad53 serum, rabbit anti-Rad9 serum, mouse anti-HA monoclonal 16B12 (Covance), or control antibodies. Immunoprecipitates were captured with protein A sepharose (Amersham). TCA lysates were prepared as described,44 with minor volume adjustments. Samples were resolved by SDS-PAGE prior to transfer to PVDF (Millipore). HA-tagged Rad9 and Asf1 were detected in immunoblot analysis with HRP-conjugated rat anti-HA monoclonal 3F10 (Roche), Rad53 with rabbit anti-Rad53 (serum or affinity purified) or goat anti-Rad53 antibody (Santa Cruz), and myc-tagged Dun1 with HRP-conjugated mouse monoclonal 9E10 (Santa Cruz). GST fusion proteins were detected with HRP-conjugated rabbit anti-GST antibody (Sigma). The rabbit anti-phospho[S/T]Q antibody was used as per manufacturer’s suggestions (Cell Signaling Technology). To detect non-tagged Rad9, rabbit polyclonal antibodies against Rad9, demonstrated previously,27 were raised against a GST fusion protein expressing Rad9 amino acids 1–505, and affinity purified against a similar CBP-tagged Rad9 fusion protein. Fusion Protein Purification. Bacterial cultures bearing pGEX GST-Rad53 FHA1 (1–197) or FHA2 (450–730) were treated with 1.5 mM IPTG for three hours to induce the expression of the fusion protein. Cells were collected by centrifugation, and lysed by sonication in TG with 5mM DTT and protease inhibitor cocktail (Roche). Clarified lysates were rotated with glutathione-sepharose beads (Amersham) overnight at 4˚C. Beads were then washed in batch format ≥ 3 times with > 100x bead bed volume of lysis buffer, resuspended in lysis buffer, and stored at -20°C with the addition of glycerol to ~50%. GST Fusion Protein Precipitations and Phosphatase Treatments. For GST fusion protein precipitation from yeast cell lysates, approximately equal amounts of GST fusion proteins (10–20 µg) bound to glutathione sepharose beads were combined with 0.2–2 mg of yeast lysates, and processed as described for immunoprecipitations. For the precipitations involving phosphatase treatment, 0.25–0.5 mg of nondenaturing lysates prepared in 50 mM Tris, pH 7.5, 100 mM NaCl, 0.05% Triton X-100, and protease inhibitor lacking EDTA (Roche) were diluted to ~2 mg/ml in 1x lambda phosphatase buffer (New England Biolabs) with 2 mM MnCl2, combined with 4000 units of lambda phosphatase (New England Biolabs) per 250 µg, and incubated at 30˚C. Where appropriate, 50 mM EDTA, 10 mM NaF, 10 mM β-glycerophosphate, and 2 mM NaVO3 were included to inhibit the phosphatase reaction. After 90 minutes, phosphatase reactions were stopped with the addition of EDTA, NaF, β-glycerophosphate, and NaVO3. Rad53 Kinase Assays. Immunoprecipitated Rad53 was resolved by SDS-PAGE and transferred to PVDF. Rad53 was then subjected to an in situ autophosphorylation kinase assay as described.45 Similar results were observed with in situ autophosphorylation kinase assays performed on TCA-extracted lysates, and in immune-complex kinase assays performed with material from nondenaturing lysates (data not shown).

RESULTS Effects of Rad53 FHA Domain Mutations on Rad53 Phosphorylation in Cycling Cells. To establish the requirements for the dual Rad53 FHA domains in the physiological function of Rad53, we compared the effects of alanine substitution of conserved residues in either or both Rad53 FHA

www.landesbioscience.com

domains. To disrupt the phosphopeptide binding activity of Rad53 FHA1, we substituted alanine for residues R70 and N107 (rad53R70A N107A) (Fig. 1A). Structural analysis of recombinant FHA1 binding a phosphopeptide demonstrated that R70 and N107 interact with the phosphopeptide backbone, and R70 additionally interacts with the phosphothreonine residue.33 As the single substitution of R70 has previously been shown to reduce, but not eliminate, the affinity of recombinant FHA1 for synthetic phosphopeptides in vitro,33 we used the double substitution to disrupt phosphopeptide binding more effectively. To inactivate the phosphopeptide binding of Rad53 FHA2, we used the triple alanine substitution of N655 (equivalent position to FHA1 N107), V656, and S657 (rad53NVS655-657AAA). This triple substitution was previously shown to abrogate the ability of recombinant FHA2 to either precipitate phosphorylated Rad9 from yeast lysates, or bind Rad9 phosphopeptides in vitro.24,27 These mutations were combined in a single construct to generate the double FHA domain mutant version of Rad53 (rad53R70A N107A NVS655-657AAA). In response to activation of either the DDC or the DRC, Rad53 is phosphorylated and activated in a PIKK-dependent manner.2,3 As the DNA checkpoint-induced phosphorylation of kinase defective Rad53 is reduced but not eliminated,2 this phosphorylation likely involves both transphosphorylation by a PIKK, and Rad53 autophosphorylation. To confirm that the alanine substitutions in the Rad53 FHA domains did not alter Rad53 expression, and to compare the involvement of Rad53 FHA domains in the phosphorylation of Rad53, we evaluated the phosphorylation and activation of Rad53 FHA domain mutants in asynchronously growing cells treated with HU or the alkylating agent methyl methane sulfonate (MMS). Importantly, the expression of the FHA domain-mutant forms of Rad53 was not impaired, relative to wildtype Rad53 (Fig. 1B), suggesting that neither the FHA1 nor the FHA2 mutations alter the basal expression or apparent stability of Rad53 in vivo. However, Rad53 FHA domain mutants did display altered levels of checkpoint-induced phosphorylation. Mutation of FHA1 resulted in an increase of intermediate phosphorylation forms of Rad53, especially in cells activated by replication block (Fig. 1B, lanes 8–9). Kinase defective Rad53 displayed a similar increase of intermediately phosphorylated Rad53 in response to genotoxin exposure (Fig. 1B, lanes 17–18). By contrast, the Rad53 FHA2 mutant displayed nearly normal phosphorylation in response to HU, and was partially impaired for phosphorylation and activation in response to MMS (Fig. 1B, lanes 11–12). Double mutation of the Rad53 FHA domains completely prevented its DNA checkpoint-induced phosphorylation (Fig. 1B, lanes 14–15). The Rad53 FHA Domains Are Required for Genotoxin Survival. RAD53 is an essential gene that is required for the survival of genotoxic stress.10,11,14,46 In a plasmid shuffle assay, individual deletion of either Rad53 FHA domain does not prevent the growth of a rad53 strain.38 Similarly, inactivation of either or both Rad53 FHA domains by alanine substitution also did not prevent growth of a rad53 strain (data not shown), suggesting that fully functional FHA domains are not crucial for the essential function of RAD53. This result is not surprising since only a very low level of Rad53 activity is required to fulfill its essential function, as indicated by the growth of rad53 strains with severe kinase-disabling missense mutations A208P or K227A.38 Rad53 FHA1 is known to contribute to the survival of DNA damage.47 To better understand the relative contribution of Rad53 FHA1 and FHA2 to this aspect of Rad53 function, we assessed the genotoxin sensitivity of Rad53 FHA domain-mutant strains. Mutation of either Rad53 FHA1 or FHA2 similarly and only slightly increased the sensitivity to DNA damage from UV irradiation or MMS (Fig. 2A), suggesting that Rad53 FHA domain function in the survival of DNA damage is largely redundant. Intriguingly, the growth of cells containing mutated Rad53 FHA2 on media containing HU was virtually indistinguishable from that of RAD53 cells, whereas the mutation of Rad53 FHA1 caused an intermediate loss of viability on HU (Fig. 2A). This result suggests that Rad53 FHA1 has DNA replicationspecific activity for genotoxin survival. Finally, the simultaneous mutation of both FHA domains reduced the DNA damage and replication block survival to levels approaching those of rad53 kinase-defective or null strains (Fig. 2A). These results indicate that intact FHA domains are essential for the function of Rad53 in the survival of genotoxic stress.

Cell Cycle

387

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

pathways. Therefore, we examined the FHA domain requirements for Rad53 in its function in the RAD9-dependent DDC. RAD53 is required for the delay of cell cycle progression in response to DNA damage throughout the cell cycle.9-12,48 Cdc13 is a telomere-associated protein that is required for the maintenance of chromosome ends.49-51 At elevated temperatures, cdc13-1 cells accumulate single stranded telomeric DNA, leading to the activation of the RAD9-dependent DNA damage checkpoint pathway and a strong arrest of the cell cycle at the G2/M transition.49,52 Downstream of Rad9, Rad53 and Chk1 function in parallel to enforce this cell cycle arrest, such that loss of RAD53 allows cells to eventually escape from the G2/M arrest and proceed into mitosis.24,28,53,54 We previously reported that the mutation of FHA2 significantly impairs the ability of Rad53 to rescue the G2/M cell cycle checkpoint defects of a cdc13-1 cdc15-2 rad53-11 strain.24 To better determine the effects of FHA domain mutations on this aspect of Rad53 function, we generated a cdc13-1 cdc15-2 strain lacking the complete RAD53 open reading frame, and reintroduced RAD53 or rad53 alleles on a low-copy plasmid. Within 180 minutes following shift to the restrictive temperature, nearly half of the cdc13-1 rad53∆ or cdc13-1 rad53K227A D339A cells had failed to maintain the G2/M arrest, whereas most of the cdc13-1 RAD53 cells were prevented from entering mitosis (Fig. 2B). Mutation of either FHA1 or FHA2 had little apparent effect on the ability of cells to maintain the G2/M arrest (Fig. 2B). However, by 360 minutes after temperature shift, both the single and double FHA domain mutants escaped into mitosis with numbers similar to that of rad53∆ or kinase defective rad53K227A D339A cells (Fig. 2B), indicating that both FHA domains are required for the sustained function of Rad53 in the DNA damage-induced G2/M checkpoint arrest. Among its activities in preventing the progression of the cell cycle, the RAD9-independent DRC delays elongation of the mitotic spindle and subsequent nuclear separation in response to HU exposure (the S/M checkpoint arrest).9,11 Since the genotoxin sensitivity assays indicated that Rad53 FHA1 is especially important for the survival of replication block, we assessed the integrity of the DRC-dependent prevention of spindle elongation after HU treatment in rad53 FHA domain mutant cells. Loss of Rad53 FHA1, or both FHA1 and FHA2, caused spindle elongation Figure 2. Impact of FHA substitutions on DNA checkpoint functions. (A) Survival of genotoxic defects similar to those observed in rad53∆ or stress. Cultures of U960-5C carrying the indicated forms of RAD53 or rad53 were serially dilutrad53K227A D339A strains (Fig. 2C, closed triangles, ed five-fold and spotted onto plates either lacking (untreated) or containing the indicated conclosed diamonds, open circles, and open squares). centrations of HU or MMS, or subjected to UV irradiation. The deletion of RAD9 increased the Mutation of Rad53 FHA2 caused an increase of spindle overall genotoxin sensitivity of these strains, but did not alter the relative impact of FHA1 or elongation after HU treatment intermediate between FHA2 mutations (data not shown). (B) G2/M cell cycle arrest in response to cdc13-1 DNA damwildtype and rad53∆ or rad53K227A D339A strains age. The indicated cultures derived from yJKD103 were synchronized in G1 with α-factor, (Fig. 2C, open triangles, open circles, and open washed, and released into prewarmed media at the restrictive temperature to inactivate squares). Hence, Rad53 FHA1 is primarily required, cdc13-1. Samples taken at the indicated times points were fixed, stained with DAPI, and scored progression through anaphase. (C) S/M cell cycle arrest in response to replication block. and FHA2 partially required, for the function of Cultures derived from U960-5C were synchronized in G1 with α-factor, washed, and released Rad53 in the S/M checkpoint arrest in response to into media containing HU. Samples taken at 30 minute intervals were fixed, stained by immunoreplication block. fluorescence for tubulin and scored for mitotic spindle elongation. FHA1 and FHA2 are Jointly Required to Couple Rad53 to Phosphorylated Rad9. DNA damage leads to the MEC1/TEL1-dependent oligomerization and phosphorylation of Dysfunction of Cell Cycle Checkpoints in Rad53 FHA Domain Rad9.24-26,55,56 RAD9 is essential for the replication-independent phosphoMutants. The difference between the sensitivity to DNA damage versus rylation and activation of Rad53 in response to DNA damage, and phosreplication block of the single FHA1 and FHA2 mutants suggests that the phorylated Rad9 interacts with Rad53.13,24 This interaction is required for Rad53 FHA domains differentially couple Rad53 to the DDC and DRC

A

B

C

388

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

the subsequent phosphorylation and activation of Rad53.24,27,56 To clarify the contribution of the Rad53 FHA domains to the in vivo interaction of Rad53 with phosphorylated Rad9, we determined the ability of Rad53 FHA domain mutants to coimmunoprecipitate with Rad9 in cells synchronized in telophase at the cdc15-2 arrest point, where the DNA damage-dependent activation of Rad53 is entirely RAD9-dependent.27 The mutation of Rad53 FHA1, and especially FHA2, substantially reduced the ability of Rad53 to coimmunoprecipitate with the phosphorylated form of Rad9 induced by MMS exposure (Fig. 3, lanes 3–6). The loss of Rad9 interaction in these FHA domain mutants correlated with strong reductions of the damageinduced phosphorylation of these forms of Rad53 (Fig. 3, lanes 3–6, bottom panel). Notably, we did not observe increased amounts of the intermediate phosphorylation form of FHA1-mutant Rad53 after MMS treatment, suggesting that this alteration of Rad53 phosphorylation is replication-dependent. Mutation of both FHA domains completely eliminated the coimmunoprecipitation of Rad9 with Rad53, and prevented Figure 3. FHA domain-mediated interaction of Rad53 with Rad9. Nondenaturing lysates the RAD9-dependent phosphorylation of Rad53 (Fig. 3, prepared from telophase-synchronized yMS509-derived cultures expressing the indicated forms lanes 7–8). Thus, both Rad53 FHA domains contribute of Rad53 were either mock treated (0) or treated with MMS (M) for 30 minutes, and used for to the stable interaction of Rad53 with phosphorylated the immunoprecipitation (IP) of Rad9 (top panel) and Rad53 (middle panel). Lysates (bottom Rad9 in vivo. panel) were loaded at 25 µg per lane. Lysate and IPs were subjected to Western blot analysis Rad53 kinase activity is required for its release from for Rad9 (top and middle panels) or Rad53 (bottom panel). phosphorylated Rad9 in vitro.56 Therefore, we examined whether Rad53 kinase activity modulates the interaction of the Rad53 FHA domain mutants with phosphorylated Rad9 Rad53 from lysates of cells treated with HU, and this precipitation was in vivo. The kinase-defective Rad53 FHA1 mutant coimmunoprecipitated slightly increased from lysates treated with MMS (Fig. 4, lanes 14–15). The amounts of phosphorylated Rad9 similar to its kinase-active counterpart lesser ability of GST-FHA2 to isolate Rad53 is noteworthy, since this FHA (Fig. 3, lane 12 v. lane 4). Disabling its catalytic domain caused domain pulls down more Rad9 than FHA1 (Fig. 4, upper panel, lane 15 v. FHA2-mutant Rad53 to coimmunoprecipitate with an increased amount of phosphorylated Rad9 (Fig. 3, lane 14 v. lane 6), suggesting that an intact Rad53 kinase domain reduces the association of Rad53 FHA1 with Rad9. Alternatively, the increased endogenous DNA damage arising from the inactivation of the Rad53 kinase domain may have augmented the RAD9-dependent checkpoint signal, thereby increasing the Rad53-Rad9 interaction. Rad53 FHA1 Preferentially Interacts with Rad53 after DNA Checkpoint Activation. One explanation for the stronger FHA1-mediated interaction between kinase-defective Rad53 and Rad9 could be that FHA1 is involved in releasing wildtype Rad53 from its activation complex, possibly via recognition of Rad53 autophosphorylation sites. To investigate this hypothesis, we measured the ability of bacterially produced GSTRad53 FHA1 and FHA2 fusion proteins to pull down Rad53 and Rad9 from lysates of cells treated with either HU or MMS. Both FHA1 and FHA2 fusion proteins precipitate approximately equal amounts of phosphorylated Rad9 from lysates of cells with DNA damage (Fig. 4, lanes 7–9 and 13–15, top panel). As HU does not strongly induce phosphorylated Rad9 in wildtype Figure 4. Differential interaction of the Rad53 FHA domains with Rad9 and Rad53. Cultures of cells, Rad9 was less readily precipitated from cells treated yMFS665 expressing FLAG-tagged Rad53 were either mock treated (0), or treated with 200 mM with HU. Recombinant FHA1 preferentially precipi- HU (H) for two hours or 0.1% MMS (M) for 1 hour, and used to prepare nondenaturing lysates. tates Rad53 from lysates of cells treated with either HU Lysates were incubated with bead-bound bacterially produced GST-Rad53 FHA1 or FHA2 fusion or MMS, but not mock treated cells (Fig. 4, lanes 7–9, proteins, or GST alone, pelleted, and resolved by SDS-PAGE. The portion of the gel containing middle panel), supporting the hypothesis that these the fusion proteins was stained with Coomassie Brilliant Blue to monitor fusion protein loading treatments enable Rad53-Rad53 interactions mediated (lanes 4–18, bottom panel), while the upper portion of the gel was analyzed by Western blot for by FHA1. Similar interactions have been proposed to Rad9 (top panel) or FLAG-tagged Rad53 (middle panel). Lanes 1–3 contain anti-Rad9 immunofacilitate signal amplification by mammalian precipitations (top panel) or lysates (middle panel, 100 µg per lane) to confirm the position of Chk2.57,58 Recombinant FHA2 weakly precipitated Rad53 and Rad9.

www.landesbioscience.com

Cell Cycle

389

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

Apparently, Rad53 FHA2 binds the HU-induced Rad53 phosphorylation sites more weakly than FH1. Mutation of Rad53 FHA1 Increases the Activation of Rad53 in Cycling Cells. Phosphorylation of Rad53 correlates with an increase of its intrinsic kinase activity.2,3 To determine whether mutation of the Rad53 FHA domains alters its activity, we performed in situ kinase assays on Rad53 that was immunoprecipitated, resolved by SDS-PAGE, transferred to PVDF, and renatured.45 Mutation of the Rad53 FHA domains did not alter its basal activity in untreated cells (Fig. 5, lanes 4, 7, 10, and 13), suggesting that the Rad53 FHA domains do not function in cis to regulate basal Figure 5. FHA domain requirements for kinase activation of Rad53. Asynchronous cultures of Rad53 autophosphorylation activity. Mutation of U960-5C bearing the indicated forms of Rad53 were either mock treated (0), or treated with 200 FHA1 caused the hyperactivation of Rad53 in mM HU (H) for 90 minutes or 0.1% MMS (M) for 45 minutes, and nondenaturing lysates were preresponse to DNA checkpoint activation (Fig. 5, pared. Immunoprecipitated Rad53 was resolved by SDS-PAGE, transferred to PVDF, and subjectlanes 8–9), which correlated with the increase of ed to an in situ kinase assay (32P ISA, bottom panel). For the immunoblot of Rad53 (top panel), intermediately phosphorylated forms of Rad53 (Fig. the ISA membrane was washed extensively, and probed with anti-Rad53 antibodies. Similar results 1B, lanes 8–9). Mutation of Rad53 FHA2 caused were observed with conventional Rad53 immune complex kinase assays, in which immunopreonly a partial reduction in the activation of Rad53 cipitated Rad53 complexes were directly incubated with [γ32P]-ATP prior to analysis (data not (Fig. 5, lanes 11–12), while the mutation of both shown). FHA domains prevented any increase of Rad53 kinase activity in response to genotoxin exposure lane 9). This difference of Rad53 FHA2 versus FHA1 suggests that HU and (Fig. 5, lanes 14–15). These results indicate that intact Rad53 FHA domains MMS may induce different sets of Rad53 (or Rad9) phosphorylation sites, are essential for the phosphorylation and activation of Rad53 by the DNA and that the two Rad53 FHA domains differentially recognize these sites. checkpoint pathways. In addition, the increase of the intermediate phosphorylation forms of FHA1-mutant Rad53 in asynchronous cells (Fig. 1B, lanes 8–9) versus non-S phase cells (Fig. 3, lane 4) suggests that the DDC and DRC may activate Rad53 through intrinsically different mechanisms. Rad53 FHA1 is Required to Prevent Activation of the DDC in HU. Replicational stress slows or stalls progression of replication forks.59-62 In wildtype cells, HU-induced S-phase delay is reversible,11 as the stalled replication forks are maintained, and are able to resume replication after removal of HU.62 In mec1 or rad53 cells, HU treatment causes the accumulation of aberrant DNA structures thought to be caused by the collapse of stalled replication forks,62-64 and the DNA-damage like phosphorylation of Rad9.26 Indeed, Rad9 was heavily [S/T]Q phosphorylated in response to HU treatment in rad53K227A D339A and rad53∆ cells (Fig. 6A, lane 14 v. lane 2, and Fig. 7A, lane 8 v. lane 2). Like Rad53 phosphorylation in HU-treated rad53K227A cells,62,63 this Rad9 phosphorylation persisted even after the removal of HU (data not shown). In otherwise wildtype cells, RAD9 is neither required for the survival of HU,9,10 nor the activation of Rad53 in response to HU,13 and HU treatment does not induce the heavy phosphorylation of Rad9.24-26 Thus, these results suggest that DNA damage-like phosphorylation of Rad9 in response to HU is an indicator of the collapse of destabilized replication forks. Figure 6. Interdependency of Rad9 and Rad53 phosphorylation. (A) Cultures of yJKD415 expressThe HU-induced hyperactivation of FHA1ing wildtype Dun1 and the indicated forms of Rad53 were either mock treated (0), or treated with mutant Rad53 in asynchronous cells could arise 200 mM HU (H) for two hours or 0.1% MMS (M) for 1 hour. TCA lysates were prepared, and Rad9 from a requirement for Rad53 FHA1 in the stabiphosphorylation analyzed by Western blot. The immunoblot was then stripped and reprobed with lization of stalled replication forks. If so, mutation of a broad spectrum anti-phospho-[S/T]Q antibody (bottom panel) that was previously shown to Rad53 FHA1 would cause the HU-induced collapse strongly recognize the heavily phosphorylated form of Rad9 induced by DNA damage.27 (B) of replication forks, leading to the activation of the Cultures of U960-5C (lanes 1-3) or yJKD429 (lanes 4-19) bearing the indicated forms of Rad53 RAD9-dependent DDC. Thus, the joint activation were either mock treated (0), or treated with 200 mM HU (H) for two hours or 0.1% MMS for 1 of both the DDC and the DRC pathways may then hour (M). TCA lysates were prepared, and Rad53 phosphorylation analyzed by Western blot.

A

B

390

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

account for the hyperactivation of FHA1-mutant Rad53. To test this hypothesis, we first determined the effect of Rad53 FHA domain mutations on Rad9 phosphorylation in genotoxin-treated asynchronous cells. Compared to cells with wildtype Rad53, HU treatment of FHA1-mutant cells induced the heavy, DNA damage-like phosphorylation of Rad9 (Fig. 6A, lane 2 v. lanes 5 and 11). This phosphorylation persisted even after cells were washed into fresh media (data not shown). By contrast, mutation of Rad53 FHA2 only slightly increased HU-induced Rad9 phosphorylation (Fig. 6A, lane 8). Together, these results suggest that Rad53 FHA1 is required to prevent the activation of the DDC in response to HU, presumably through the stabilization of replication forks. To determine whether the HU-induced activation of the RAD9-dependent DDC in Rad53 FHA1-mutant cells is sufficient to cause the hyperactivation of Rad53, we determined whether inactivation of the DDC pathway by deletion of RAD9 alleviates the accumulation of intermediately phosphorylated Rad53. Although deletion of RAD9 lessened the DNA damage-induced phosphorylation of Rad53 in asynchronous cells (Fig. 6B, lane 3 v. lane 7), the loss of Rad9 did not prevent the accumulation of intermediate phosphorylated Rad53 in rad53R70N N107A or rad53K227A D339A in response to HU exposure (Fig. 6B, lanes 9 and 18). These data suggest that the aberrant activation of the RAD9-dependent DDC in response to HU Figure 7. Rad53 FHA domain requirements for regulation of Dun1. (A) DUN1 is not essential for the treatment does not account for the hyperactivation prevention of DNA damage in response to HU. yJKD415-derived cultures bearing RAD53, DUN1, or empty vector plasmids were mock treated (0), or treated with 200 mM HU for 2 hours (H) or 0.1% of FHA1-mutant Rad53. Dun1 Is Not Required for the Rad53 MMS (M) for 1 hour. TCA lysates were analyzed by Western blot for Rad9 (top panel) and FHA1-Dependent Stabilization of Replication myc-tagged Dun1 (bottom panel). The anti-Rad9 immunoblot was stripped and reprobed with an Forks. Rad53 FHA1 could mediate replication fork anti-phospho-[S/T]Q antibody (middle panel). (B) Redundant function of Rad53 FHA domains in stability either directly through the interaction of Dun1 phosphorylation. Cultures of yJKD415 carrying myc-tagged Dun1 and the indicated forms of FHA1 with a replication fork component, or indi- Rad53 were either mock treated (0), treated with 200 mM HU (H) for 2 hours, or treated with 0.1% MMS (M) for 1 hour, lysed in TCA, and analyzed for Dun1 phosphorylation by Western blot with rectly through a downstream pathway that couples anti-myc antibodies. to Rad53 via FHA1. Dun1 is a FHA domaincontaining protein kinase that is phosphorylated and activated in a MEC1 and RAD53-dependent manner11,15 (Fig. 7A, the phosphorylation of Rad53 in response to replication-block, but not bottom panel, lanes 1–3 v. 7–9), and has been implicated in a number of DNA damage,71,72 implicating Asf1 in the regulation of Rad53 by the 11,15,15,53,65-67 To deterthe RAD53-dependent DNA checkpoint responses. DRC. To determine whether the HU response defects of the Rad53 FHA1 mine whether the HU-induced activation of the DDC caused by mutation mutant involve Asf1, we examined the basal interaction of Asf1 with the of Rad53 FHA1 is due to the disruption of signaling through Dun1, we FHA domain-mutant forms of Rad53. Mutation of Rad53 FHA1, and not assessed Rad9 and Dun1 phosphorylation in dun1∆ and rad53 cells. FHA2, abrogates the coimmunoprecipitation of Asf1 with Rad53, suggesting Although Rad9 phosphorylation was slightly enhanced, the deletion of that Rad53 FHA1 is required for the interaction of Rad53 with Asf1 (Fig. DUN1 did not cause the HU-induced DNA damage-like strong [S/T]Q 8A, middle panel, lanes 3 and 5 v. lane 4). Surprisingly, kinase-defective phosphorylation of Rad9 present in rad53∆ or FHA1-mutant cells (Fig. 7A, Rad53, which contains an intact FHA1, was also severely impaired for coimlane 5 v. lane 8, and Fig. 6A, lane 5). Consistent with this result, the DNA munoprecipitation of Asf1 (Fig. 8A, lane 6), suggesting that the interaction checkpoint-induced phosphorylation of Dun1 was not abolished by the of Asf1 with Rad53 is sensitive to a FHA1-mediated Rad53 kinase activity individual mutation of either Rad53 FHA1 or FHA2 (Fig. 7B, lanes 4–9). in untreated cells. Rather, Dun1 phosphorylation directly correlated with Rad53 phosphoryTo confirm that the interaction of Rad53 with Asf1 is mediated by lation and activation, such that only cells with the non-activatable Rad53 FHA1, we performed GST fusion protein pull down experiments using FHA1/FHA2 double mutant, kinase-defective Rad53, or rad53∆ cells wildtype and mutant versions of Rad53 FHA1. Bacterially produced lacked the DNA checkpoint- induced Dun1 phosphorylation (Fig. 7B, GST-Rad53 FHA1 specifically precipitates Asf1 from yeast cell lysates, and lanes 10–15, and Fig. 7A, bottom panel, lanes 7–9). Together, these results this interaction is severely impaired by the R70A and N107A substitutions indicate that the defects observed within FHA1-mutant cells are not due to that disrupt phosphopeptide binding activity in vitro (Fig. 8B, lanes 2–4). the dysregulation of RAD53-dependent signaling through Dun1. Recombinant GST-FHA2 precipitated a barely detectable amount of Asf1 Phosphorylation-Dependent Interaction of Asf1 with Rad53 FHA1. (Fig. 8B, lanes 5–6), suggesting that the Asf1-binding characteristics of Asf1 is a chromatin assembly factor thought to be involved in the deposition FHA1 are not well conserved in FHA2. of histones H3 and H4 onto DNA after DNA replication and repair.68-71 The finding that Rad53 FHA1 specifically interacts with Asf1 was In addition to interacting with histones H3 and H4, Asf1 interacts with unexpected, as Asf1 is not thought to be a phosphoprotein.71,72 Based on Rad53 in undamaged cells, and MEC1-dependent DNA checkpoint activation parallels with the MH2 domain, it was proposed that FHA domains may reverses or prevents this interaction.71,72 Deletion of ASF1 can interfere with also bind proteins in a phosphorylation-independent manner.73 To assess

A

B

www.landesbioscience.com

Cell Cycle

391

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

A

B

C

Figure 8. Asf1 interacts with Rad53 mainly through FHA1. (A) Rad53 FHA1 is required for the coimmunoprecipitation of Asf1 with Rad53. Immunoprecipitates of Rad53 (top and middle panels) or Asf1 (bottom panel) were prepared from nondenaturing lysates of yJKD403-derived cells bearing the indicated forms of Rad53, and analyzed by Western blot for Rad53 (top panel) or Asf1 (middle and bottom panels). (B) Bacterially produced Rad53 FHA domain precipitations of Asf1. Nondenaturing lysates prepared from untreated yJKD403 cells carrying RAD53 were combined with bead-bound GST, GST-FHA1, or FHA2 fusion proteins. 1% of the lysate (5 µg) used for each pull down (L), and the pull downs were analyzed by Western blot for Asf1 (top panel) and GST for fusion protein loading (bottom panel). Approximately 1/6th of the each pull down was loaded. (C) Nondenaturing lysates prepared from yJKD301 were either untreated (u), mock treated in phosphatase buffer alone (m), treated with lambda phosphatase (+), or treated with lambda phosphatase and phosphatase inhibitors (+i). Each lysate was then split and combined with bead-bound GST, or GST-FHA1 or FHA2 fusion proteins as indicated. Pull downs were analyzed by Western blot for Asf1 (top panel) and GST for fusion protein loading (bottom panel). 1% (2 µg) of the lysate (L) used for the precipitations from each of the treatment conditions were also loaded. Approximately 1/5th of each pull down was loaded.

the possible requirement for phosphorylation in the interaction of Rad53 FHA1 with Asf1, we determined whether phosphatase treatment prevents the pull down of Asf1 by recombinant GST-Rad53 FHA1. As previously reported,71 phosphatase treatment did not alter the electrophoretic profile of Asf1 (Fig. 8C, lane 7 v. lanes 1 and 5). However, treatment of lysates with lambda phosphatase (Fig. 8C, lanes 7–8), but neither a mock incubation nor lambda phosphatase treatment in the presence of phosphatase inhibitors

392

(Fig. 8C, lanes 5–6 and 9–11), reduced the precipitation of Asf1 by recombinant FHA1 (Fig. 8C, lane 8 v. lanes 6 and 10), suggesting that the interaction between Asf1 and Rad53 is indeed phosphorylation-dependent. Checkpoint- and RAD53-Independent Interaction of Rad53 FHA1 and Asf1 In Vitro. Through an unknown mechanism, DNA checkpoint activation reduces the interaction of Asf1 with Rad53.71,72 As the interaction between Asf1 and Rad53 requires the phosphorylation of Asf1, the loss

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

of Asf1 phosphorylation in response to DNA checkpoint activation could account for the loss of this interaction. Therefore, we investigated the ability of Rad53 FHA1 to interact with Asf1 in lysates from cells treated with HU. GST-Rad53 FHA1 readily pulled down Asf1 from cell lysates regardless of DNA checkpoint activation (Fig. 9A), suggesting that the reduction of Asf1 interaction with Rad53 is not due to the modification of Asf1 itself. In addition, recombinant FHA1 is able to precipitate Asf1 from cell lysates regardless of the presence, absence, or mutation of Rad53 in the lysates (Fig. 9B), suggesting that Asf1 is competent to interact with Rad53 in vitro independent of RAD53-dependent in vivo processes. Loss of ASF1 Only Partially Mimics Mutation of Rad53 FHA1. As ASF1 overexpression partially suppresses the HU sensitivity of rad53-21 cells,72 and Rad53 interacts with Asf1 through FHA1, the requirement of Rad53 FHA1 for the prevention of DDC activation in HU may also involve Asf1. To determine whether ASF1 is required for the normal DNA checkpoint response to HU, we deleted the coding sequence for ASF1 in wildtype and rad53 cells. As reported previously,72 untreated asf1∆ cells have increased levels of Rad53 phosphorylation (Fig. 10, bottom panel, lane 4 v. lane 1), suggesting that the lack of Asf1 may cause basal DNA damage that leads to the activation of the DNA checkpoint pathways. Consistent with this hypothesis, we also observed a weak increase of Rad9 phosphorylation in untreated asf1∆ cells (Fig. 10, top panel, lane 4 v. lane 1). However, HU treatment of asf1∆ cells did not cause the heavy phosphorylation of Rad9 seen in rad53∆ cells (Fig. 10, top panel, lanes 5–6 v. 8-9), whereas asf1∆ rad53∆ cells did heavily phosphorylate Rad9 in response to HU, indicating that loss of ASF1 does not contribute to HU-induced DDC activation in either wildtype or rad53 cells. Together, these results suggest that the inability of FHA1-mutant Rad53 to interact with Asf1 does not account for the activation of the DDC in response to HU in rad53 cells.

DISCUSSION

A

B

Figure 9. Interaction of Rad53 FHA1 with Asf1 from cells with a quiescent or active DRC. (A) Rad53 FHA1 can interact with Asf1 regardless of DNA checkpoint activation. Cultures of yJKD403 expressing wildtype Rad53 were either mock treated, or treated with 200 mM HU for 2 hours, and used to generate nondenaturing lysates. 500 µg of each lysate was then combined with bead-bound GST, or GST-FHA1 fusion proteins as indicated. Pull downs were analyzed by Western blot for Asf1 (top panel) and GST for fusion protein loading (bottom panel). Approximately 1/5th of the each pull down was loaded. (B) Rad53 FHA1 can interact with Asf1 independent of cellular Rad53. Cultures of yJKD403 expressing the indicated forms of Rad53 were used to generate nondenaturing lysates. 500 µg of each lysate was then combined with bead-bound GST, or GST-FHA1 fusion proteins as indicated. Pull downs were analyzed by Western blot for Asf1 (top panel) and GST for fusion protein loading (bottom panel). Approximately 1/5th of the each pull down was loaded.

We introduced alanine substitutions in place of conserved residues of Rad53 FHA1 and FHA2 that are required for their phosphopeptide binding activities. The loss of FHA1 approximates the loss of FHA2 in reducing the survival of DNA damage and coupling Rad53 to the RAD9-dependent DNA damage checkpoint pathway. By contrast, Rad53 FHA1, not FHA2 is specifically required for the proper modulation of Rad53 kinase activity and the prevention of DNA damage in response to the DNA replication inhibitor, HU. The simultaneous mutation of both FHA1 and FHA2 results in a rad53 allele that was nearly as defective for genotoxin survival and cell cycle checkpoint function as kinase-defective or null rad53 alleles. Thus, these two FHA domains have both redundant and unique functions that are crucial for the protective DNA checkpoint functions of Rad53. Rad53 is unique among Chk2/Cds1-family checkpoint kinases in that Rad53 contains not one but two FHA domains. Sequence comparison of the Rad53 FHA domains to other FHA domains within the Chk2-family reveals that FHA1 is more similar to other www.landesbioscience.com

FHA domains than it is to FHA2.30 Intriguingly, Cds1, the Schizosaccharomyces pombe homolog of Rad53, normally participates only in the DRC, and S. pombe Chk1 is the checkpoint kinase that is activated by a PIKK in the DDC response.74-77 As budding yeast Rad53 acts in both the DRC and, in parallel with Chk1, the DDC,2,3,10,11,28 it is interesting to speculate that Rad53 acquired its second FHA domain as a means to couple to the DDC. Alternatively, the second FHA domain of Rad53 may serve to increase the affinity of Rad53 for phosphopeptide ligands by increasing the valence of multiprotein complexes, or may increase the potential repertoire of interacting proteins. The Rad53 FHA domains act at several levels of the DNA checkpoint pathways. First, the FHA domains mediate Rad53 activation by regulating interactions of Rad53 with phosphorylated checkpoint components in upstream signaling pathways. Second, the FHA domains within a Rad53 molecule can modulate its activation

Cell Cycle

393

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

Figure 10. ASF1 and Rad9 phosphorylation. 200 mM HU was added to early log-phase cultures of 1588-4C, yJKD303, U960-5C, and yJKD409. Cell samples taken at the indicated time points were lysed in TCA, and analyzed by Western blot for Rad9 (top panel) and Rad53 (bottom panel).

through intramolecular interactions or intermolecular interactions with other Rad53 molecules. Finally, the Rad53 FHA domains may act to target Rad53 activity to effector components of downstream DNA checkpoint pathways. The major replicational stress- and DNA damage-protective responses of the pathway probably require all of these levels of interaction. Activation of Rad53 by the DRC and DDC. Of the DNA checkpoint pathways that regulate Rad53 activation, the RAD9-dependent DDC is the best understood. Upstream of Rad53, Rad9 is multiply phosphorylated in response to DNA damage in a PIKK-dependent manner.24-26 The Rad9 phosphorylation sites bind the Rad53 FHA domains,24,27,32 thereby mediating Rad53 activation through transphosphorylation by PIKKs and autophosphorylation.2,3,56 Mutation of either Rad53 FHA domain results in a strong reduction of the interaction of Rad53 with phosphorylated Rad9, and subsequent RAD9-dependent phosphorylation of Rad53. Thus, both Rad53 FHA domains are required for the robust interaction of Rad53 with Rad9 in vivo. Furthermore, bacterially produced Rad53 FHA1 and FHA2 precipitate similar amounts of phosphorylated Rad9, supporting the hypothesis that the two Rad53 FHA domains contribute similarly to coupling of Rad53 to the DDC through Rad9. The mechanism (or mechanisms) of regulation of Rad53 by the RAD9-independent DRC is less well understood, despite the identification of several proteins required for the activation of Rad53 by replicational stress. In contrast to the DNA damage response, Rad53 FHA domain-dependent differences are seen for the response to replication block. Alanine substitutions within either Rad53 FHA domain individually do not reduce Rad53 phosphorylation and activation in response to HU. Thus, either Rad53 FHA domain is competent for coupling Rad53 to phosphorylation and activation by the DRC. However, the mutation of FHA1, and not FHA2, leads to the genotoxin-induced accumulation of intermediately phosphorylated, hyperactivated Rad53. These effects are RAD9-independent, and do not occur in cells synchronized outside of S phase, suggesting that they arise from the direct action of the DRC, rather than the additive activation of DRC and DDC pathways. As Rad53 can interact with itself in the two-hybrid assay (Z Sun and DFS, unpublished data), and FHA1 interacts with phosphorylated Rad53 more efficiently than FHA2, Rad53 FHA1 may be involved in the modulation of Rad53 activation. Alternatively, genomic defects 394

caused by the disruption of Rad53 FHA1-dependent functions may cause increased signaling through the DRC, leading to increased Rad53 activation. FHA Domain-Mediated Targeting of Rad53 Effector Pathways. Although FHA1- or FHA2-mutant Rad53 retains at least partial genotoxin-induced kinase activity, cells containing these forms of Rad53 are compromised for function of their S/M and G2/M DNA checkpoints. Disruption of either FHA1 or FHA2 also causes essentially similar increases in sensitivity to DNA damaging agents, and loss of FHA1 impairs the survival of HU. Thus, it is possible that mutation of the Rad53 FHA1 and FHA2 domains affects Rad53 signaling output, as well as Rad53 activation. Activation of the checkpoint kinase Dun1 is an output of Rad53,11 and the elimination of either Rad53 phosphorylation (FHA1/FHA2 double mutant) or Rad53 kinase activity disrupts genotoxin-induced Dun1 phosphorylation. However, loss of either FHA1 or FHA2 does not specifically decouple Dun1 from DNA checkpoint regulation. Instead, Dun1 couples to DNA checkpoint activation through the recognition of Rad53 PIKK consensus phosphorylation sites by the Dun1 FHA domain, enabling activation of Dun1 through transphosphorylation by Rad53 (S-JL, MFS, JKD, DFS, in press; and ref. 22). Not all of the DNA checkpoint functions attributed to Rad53 are accounted for by the activation of Dun1. For example, in wildtype cells, the RAD53-dependent maintenance of stalled replication forks is important for the timely recovery from HU treatment.11,59,62 This protective DNA checkpoint effect appears to require Rad53 FHA1, but not Rad53 FHA2 or DUN1. Rad53 relies directly on protein interactions mediated by the Rad53 FHA domains to regulate another DNA checkpoint effector pathway. The protein kinase Cdc7 and its regulatory subunit Dbf4 are required for replication origin firing.78-80 In response to DNA checkpoint activation, Dbf4 becomes phosphorylated and Cdc7 activity is reduced in a RAD53-dependent manner.81 Rad53 can interact with Dbf4,82 and it now appears that both Rad53 FHA domains are competent to interact with Dbf4.83 Thus, the Rad53 FHA domains provide a direct link between Rad53 and DNA checkpoint effector proteins. The Rad53 FHA domains may directly mediate interactions with other downstream signaling pathways, thereby amplifying and distributing the DNA checkpoint signal. In the absence of DNA checkpoint activation, the chromatin assembly factor Asf1 interacts with Rad53.71,72 The Asf1/Rad53 complex does not contain histones H3 and H4, and Rad53 can inhibit the ability of Asf1 to facilitate histone deposition in vitro.71 Thus, RAD53-dependent DNA checkpoint responses may regulate chromatin structure through Asf1. Rad53 FHA1 is specifically required to couple Rad53 to Asf1 in vivo and in vitro. Although this interaction is phosphorylation-dependent, the ability of Asf1 to be pulled down by GST-FHA1 does not vary with either DNA checkpoint activation or cellular Rad53. Perhaps the DNA checkpointinduced loss of Rad53 interaction with Asf1 is due to the differential compartmentalization of Asf1 and Rad53, rather than the loss of Asf1 phosphorylation. Alternatively, DNA checkpoint activation may induce phosphoepitopes on other proteins, including possibly Rad53 itself, which may compete with Asf1 for occupancy of Rad53 FHA1. Multivalency of DNA Checkpoint Signaling Complexes. Like other protein/phosphopeptide interaction domains, FHA domains are involved in a number of critical recognition steps in signal transduction (reviewed in refs. 84,85). The diverse functionality of

Cell Cycle

2003; Vol. 2 Issue 4

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

Rad53 FHA domains reveals the flexibility of FHA domain/phosphopeptide interactions. Rad53 FHA domains are involved in binding Rad53 to an upstream coupling protein, Rad9; in Rad53/Rad53 oligomerization that may be an important component of Rad53 activation or Rad53 dissociation from signaling complexes; and in linking Rad53 to the Dbf4/Cdc7 effector pathway. The finding that Rad53 FHA1 binds Asf1 identifies the fourth in vivo Rad53 FHA domain binding partner (including Rad53, Rad9, and the Dbf4/Cdc7 complex), and it can be anticipated that this list will continue to grow. Rad53 phosphopeptide recognition sites on Rad9 and Rad53 are grouped in clusters (ref. 27; and S-JL, MFS, JKD, and DFS, in press), with sufficient spacing so that multiple Rad53 molecules may be recruited to a single phosphorylated cluster. Cooperative binding and regulatory effects of polyvalent binding may contribute to sharp regulatory transitions that will both ensure rapid responses to DNA damage, and maintain stringent control, preventing baseline activity of the DNA checkpoint signaling networks. The availability of two FHA domains on Rad53 may also facilitate the assembly of polymerized networks containing large numbers of activated Rad53 molecules and other checkpoint proteins. Such exponential interactions can readily lead to assembly of the massive DNA damage response complexes that accumulate focally at sites of DNA damage. Acknowledgements We thank M Kroetz for fusion protein production, G Liu for assistance with Rad9 antibody production, and J Falato for clerical assistance. We also thank the Rothstein, Weinert, Elledge, and Longtine laboratories for strains and/or plasmids, and members of the Stern laboratory for helpful discussions. References 1. Stern DF, Zheng P, Beidler DR, Zerillo C. Spk1, a new kinase from Saccharomyces cerevisiae, phosphorylates proteins on serine, threonine, and tyrosine. Mol Cell Biol 1991; 11:987-1001. 2. Sun Z, Fay DS, Marini F, Foiani M, Stern DF. Spk1/Rad53 is regulated by Mec1-dependent protein phosphorylation in DNA replication and damage checkpoint pathways. Genes Dev 1996; 10:395-406. 3. Sanchez Y, Desany BA, Jones WJ, Liu Q, Wang B, Elledge SJ. Regulation of RAD53 by the ATM-like kinases MEC1 and TEL1 in yeast cell cycle checkpoint pathways. Science 1996; 271:357-60. 4. Elledge SJ. Cell cycle checkpoints: preventing an identity crisis. Science 1996; 274:1664-72. 5. Nyberg KA, Michelson RJ, Putnam CW, Weinert TA. Toward Maintaining The Genome: DNA Damage and Replication Checkpoints. Annu Rev Genet 2002; 36:617-56. 6. Weinert TA, Hartwell LH. The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae. Science 1988; 241:317-22. 7. Siede W, Friedberg AS, Friedberg EC. RAD9-dependent G1 arrest defines a second checkpoint for damaged DNA in the cell cycle of Saccharomyces cerevisiae. Proc Natl Acad Sci USA 1993; 90:7985-9. 8. Siede W, Friedberg AS, Dianova I, Friedberg EC. Characterization of G1 checkpoint control in the yeast Saccharomyces cerevisiae following exposure to DNA-damaging agents. Genetics 1994; 138:271-81. 9. Weinert TA. Dual cell cycle checkpoints sensitive to chromosome replication and DNA damage in the budding yeast Saccharomyces cerevisiae. Radiat Res 1992; 132:141-3. 10. Weinert TA, Kiser GL, Hartwell LH. Mitotic checkpoint genes in budding yeast and the dependence of mitosis on DNA replication and repair. Genes Dev 1994; 8:652-65. 11. Allen JB, Zhou Z, Siede W, Friedberg EC, Elledge SJ. The SAD1/RAD53 protein kinase controls multiple checkpoints and DNA damage-induced transcription in yeast. Genes Dev 1994; 8:2401-15. 12. Paulovich AG, Hartwell LH. A checkpoint regulates the rate of progression through S phase in S. cerevisiae in response to DNA damage. Cell 1995; 82:841-7. 13. Navas TA, Sanchez Y, Elledge SJ. RAD9 and DNA polymerase epsilon form parallel sensory branches for transducing the DNA damage checkpoint signal in Saccharomyces cerevisiae. Genes Dev 1996; 10:2632-43. 14. Zheng P, Fay DS, Burton J, Xiao H, Pinkham JL, Stern DF. SPK1 is an essential S-phase-specific gene of Saccharomyces cerevisiae that encodes a nuclear serine/threonine/tyrosine kinase. Mol Cell Biol 1993; 13:5829-42. 15. Zhou Z, Elledge SJ. DUN1 encodes a protein kinase that controls the DNA damage response in yeast. Cell 1993; 75:1119-27.

www.landesbioscience.com

16. Kiser GL, Weinert TA. Distinct roles of yeast MEC and RAD checkpoint genes in transcriptional induction after DNA damage and implications for function. Mol Biol Cell 1996; 7:703-18. 17. Aboussekhra A, Vialard JE, Morrison DE, de la Torre-Ruiz MA, Cernakova L, Fabre F, et al. A novel role for the budding yeast RAD9 checkpoint gene in DNA damage- dependent transcription. EMBO J 1996; 15:3912-22. 18. Jelinsky SA, Samson LD. Global response of Saccharomyces cerevisiae to an alkylating agent. Proc Natl Acad Sci USA 1999; 96:1486-91. 19. Terleth C, Schenk P, Poot R, Brouwer J, van de Putte P. Differential repair of UV damage in rad mutants of Saccharomyces cerevisiae: a possible function of G2 arrest upon UV irradiation. Mol Cell Biol 1990; 10:4678-84. 20. McAinsh AD, Scott-Drew S, Murray JA, Jackson SP. DNA damage triggers disruption of telomeric silencing and Mec1p- dependent relocation of Sir3p. Curr Biol 1999; 9:963-6. 21. de la Torre-Ruiz M, Lowndes NF. The Saccharomyces cerevisiae DNA damage checkpoint is required for efficient repair of double strand breaks by non-homologous end joining. FEBS Lett 2000; 467:311-5. 22. Bashkirov VI, Bashkirova EV, Haghnazari E, Heyer WD. Direct Kinase-to-Kinase Signaling Mediated by the FHA Phosphoprotein Recognition Domain of the Dun1 DNA Damage Checkpoint Kinase. Mol Cell Biol 2003; 23:1441-52. 23. Usui T, Ogawa H, Petrini JH. A DNA damage response pathway controlled by Tel1 and the Mre11 complex. Mol Cell 2001;7:1255-66. 24. Sun Z, Hsiao J, Fay DS, Stern DF. Rad53 FHA domain associated with phosphorylated Rad9 in the DNA damage checkpoint. Science 1998; 281:272-4. 25. Emili A. MEC1-dependent phosphorylation of Rad9p in response to DNA damage. Mol Cell 1998; 2:183-9. 26. Vialard JE, Gilbert CS, Green CM, Lowndes NF. The budding yeast Rad9 checkpoint protein is subjected to Mec1/Tel1- dependent hyperphosphorylation and interacts with Rad53 after DNA damage. EMBO J 1998; 17:5679-88. 27. Schwartz MF, Duong JK, Sun Z, Morrow JS, Pradhan D, Stern DF. Rad9 phosphorylation sites couple Rad53 to the Saccharomyces cerevisiae DNA damage checkpoint. Mol Cell 2002; 9:1055-65. 28. Sanchez Y, Bachant J, Wang H, Hu F, Liu D, Tetzlaff M, et al. Control of the DNA damage checkpoint by chk1 and rad53 protein kinases through distinct mechanisms. Science 1999; 286:1166-71. 29. Alcasabas AA, Osborn AJ, Bachant J, Hu F, Werler PJ, Bousset K, et al. Mrc1 transduces signals of DNA replication stress to activate Rad53. Nat Cell Biol 2001; 3:958-65. 30. Hofmann K, Bucher P. The FHA domain: a putative nuclear signalling domain found in protein kinases and transcription factors. Trends Biochem Sci 1995; 20:347-9. 31. Stone JM, Collinge MA, Smith RD, Horn MA, Walker JC. Interaction of a protein phosphatase with an Arabidopsis serine- threonine receptor kinase. Science 1994; 266:793-5. 32. Durocher D, Henckel J, Fersht AR, Jackson SP. The FHA domain is a modular phosphopeptide recognition motif. Mol Cell 1999; 4:387-94. 33. Durocher D, Taylor IA, Sarbassova D, Haire LF, Westcott SL, Jackson SP, et al. The molecular basis of FHA Domain:Phosphopeptide binding specificity and implications for phospho-dependent signaling mechanisms. Mol Cell 2000; 6:1169-82. 34. Liao H, Yuan C, Su MI, Yongkiettrakul S, Qin D, Li H, et al. Structure of the FHA1 domain of yeast Rad53 and identification of binding sites for both FHA1 and its target protein Rad9. J Mol Biol 2000; 304:941-51. 35. Liao H, Byeon IJ, Tsai MD. Structure and function of a new phosphopeptide-binding domain containing the FHA2 of Rad53. J Mol Biol 1999; 294:1041-9. 36. Byeon IJ, Yongkiettrakul S, Tsai MD. Solution structure of the yeast Rad53 FHA2 complexed with a phosphothreonine peptide pTXXL: comparison with the structures of FHA2- pYXL and FHA1-pTXXD complexes. J Mol Biol 2001; 314:577-88. 37. Nash P, Tang X, Orlicky S, Chen Q, Gertler FB, Mendenhall MD, et al. Multisite phosphorylation of a CDK inhibitor sets a threshold for the onset of DNA replication. Nature 2001; 414:514-21. 38. Fay DS, Sun Z, Stern DF. Mutations in SPK1/RAD53 that specifically abolish checkpoint but not growth-related functions. Curr Genet 1997; 31:97-105. 39. Longtine MS, McKenzie A, 3rd, Demarini DJ, Shah NG, Wach A, Brachat A, et al. Additional modules for versatile and economical PCR-based gene deletion and modification in Saccharomyces cerevisiae. Yeast 1998; 14:953-61. 40. Thomas BJ, Rothstein R. Elevated recombination rates in transcriptionally active DNA. Cell 1989; 56:619-30. 41. Zhao X, Muller EG, Rothstein R. A suppressor of two essential checkpoint genes identifies a novel protein that negatively affects dNTP pools. Mol Cell 1998; 2:329-40. 42. Zhao X, Georgieva B, Chabes A, Domkin V, Ippel JH, Schleucher J, et al. Mutational and structural analyses of the ribonucleotide reductase inhibitor Sml1 define its Rnr1 interaction domain whose inactivation allows suppression of mec1 and rad53 lethality. Mol Cell Biol 2000; 20:9076-83. 43. Lydall D, Weinert T. Use of cdc13-1-induced DNA damage to study effects of checkpoint genes on DNA damage processing. Methods Enzymol 1997; 283:410-24. 44. Foiani M, Liberi G, Piatti S, Plevani P. Saccharomyces cerevisiae as a model system to study DNA replication. In: Cotterill S, ed. Eukaryotic DNA replication : a practical approach. Oxford, UK: Oxford University Press; 1999:185-200. 45. Pellicioli A, Lucca C, Liberi G, Marini F, Lopes M, Plevani P, et al. Activation of Rad53 kinase in response to DNA damage and its effect in modulating phosphorylation of the lagging strand DNA polymerase. EMBO J 1999; 18:6561-72.

Cell Cycle

395

FHA DOMAIN-MEDIATED DNA CHECKPOINT REGULATION OF RAD53

46. Moore CW. Responses of radiation-sensitive mutants of Saccharomyces cerevisiae to lethal effects of bleomycin. Mutat Res 1978; 51:165-80. 47. Pike BL, Hammet A, Heierhorst J. Role of the N-terminal FHA domain in the cell cycle checkpoint function of the Rad53 kinase. J Biol Chem 2001; 276:14019-26. 48. Sidorova JM, Breeden LL. Rad53-dependent phosphorylation of Swi6 and down-regulation of CLN1 and CLN2 transcription occur in response to DNA damage in Saccharomyces cerevisiae. Genes Dev 1997; 11:3032-45. 49. Garvik B, Carson M, Hartwell L. Single-stranded DNA arising at telomeres in cdc13 mutants may constitute a specific signal for the RAD9 checkpoint. Mol Cell Biol 1995; 15:6128-38. 50. Lin JJ, Zakian VA. The Saccharomyces CDC13 protein is a single-strand TG1-3 telomeric DNA- binding protein in vitro that affects telomere behavior in vivo. Proc Natl Acad Sci USA 1996; 93:13760-5. 51. Nugent CI, Hughes TR, Lue NF, Lundblad V. Cdc13p: a single-strand telomeric DNA-binding protein with a dual role in yeast telomere maintenance. Science 1996; 274:249-52. 52. Lydall D, Weinert T. Yeast checkpoint genes in DNA damage processing: implications for repair and arrest. Science 1995; 270:1488-91. 53. Gardner R, Putnam CW, Weinert T. RAD53, DUN1 and PDS1 define two parallel G2/M checkpoint pathways in budding yeast. EMBO J 1999; 18:3173-85. 54. Liu Y, Vidanes G, Lin YC, Mori S, Siede W. Characterization of a Saccharomyces cerevisiae homologue of Schizosaccharomyces pombe Chk1 involved in DNA-damage-induced M-phase arrest. Mol Gen Genet 2000; 262:1132-46. 55. Soulier J, Lowndes NF. The BRCT domain of the S. cerevisiae checkpoint protein Rad9 mediates a Rad9-Rad9 interaction after DNA damage. Curr Biol 1999; 9(:551-4. 56. Gilbert CS, Green CM, Lowndes NF. Budding yeast Rad9 is an ATP-dependent Rad53 activating machine. Mol Cell 2001; 8:129-36. 57. Xu X, Tsvetkov LM, Stern DF. Chk2 activation and phosphorylation-dependent oligomerization. Mol Cell Biol 2002; 22:4419-32. 58. Ahn JY, Li X, Davis HL, Canman CE. Phosphorylation of threonine 68 promotes oligomerization and autophosphorylation of the Chk2 protein kinase via the forkheadassociated domain. J Biol Chem 2002; 277:19389-95. 59. Santocanale C, Diffley JF. A Mec1- and Rad53-dependent checkpoint controls late-firing origins of DNA replication. Nature 1998; 395:615-8. 60. Shirahige K, Hori Y, Shiraishi K, Yamashita M, Takahashi K, Obuse C, et al. Regulation of DNA-replication origins during cell-cycle progression. Nature 1998; 395:618-21. 61. Tercero JA, Diffley JF. Regulation of DNA replication fork progression through damaged DNA by the Mec1/Rad53 checkpoint. Nature 2001; 412:553-7. 62. Lopes M, Cotta-Ramusino C, Pellicioli A, Liberi G, Plevani P, Muzi-Falconi M, et al. The DNA replication checkpoint response stabilizes stalled replication forks. Nature 2001; 412:557-61. 63. Sogo JM, Lopes M, Foiani M. Fork reversal and ssDNA accumulation at stalled replication forks owing to checkpoint defects. Science 2002; 297:599-602. 64. Cha RS, Kleckner N. ATR homolog Mec1 promotes fork progression, thus averting breaks in replication slow zones. Science 2002; 297:602-6. 65. Huang M, Zhou Z, Elledge SJ. The DNA replication and damage checkpoint pathways induce transcription by inhibition of the Crt1 repressor. Cell 1998; 94:595-605. 66. Zhao X, Rothstein R. The Dun1 checkpoint kinase phosphorylates and regulates the ribonucleotide reductase inhibitor Sml1. Proc Natl Acad Sci USA 2002; 99:3746-51. 67. Hammet A, Pike BL, Heierhorst J. Posttranscriptional regulation of the RAD5 DNA repair gene by the Dun1 kinase and the Pan2-Pan3 poly(A)-nuclease complex contributes to survival of replication blocks. J Biol Chem 2002; 277:22469-74. 68. Tyler JK, Adams CR, Chen SR, Kobayashi R, Kamakaka RT, Kadonaga JT. The RCAF complex mediates chromatin assembly during DNA replication and repair. Nature 1999; 402:555-60. 69. Munakata T, Adachi N, Yokoyama N, Kuzuhara T, Horikoshi M. A human homologue of yeast anti-silencing factor has histone chaperone activity. Genes Cells 2000; 5:221-33. 70. Sharp JA, Fouts ET, Krawitz DC, Kaufman PD. Yeast histone deposition protein Asf1p requires Hir proteins and PCNA for heterochromatic silencing. Curr Biol 2001; 11:463-73. 71. Emili A, Schieltz DM, Yates JR, 3rd, Hartwell LH. Dynamic interaction of DNA damage checkpoint protein Rad53 with chromatin assembly factor Asf1. Mol Cell 2001; 7:13-20. 72. Hu F, Alcasabas AA, Elledge SJ. Asf1 links Rad53 to control of chromatin assembly. Genes Dev 2001; 15:1061-6. 73. Li J, Williams BL, Haire LF, Goldberg M, Wilker E, Durocher D, et al. Structural and functional versatility of the FHA domain in DNA-damage signaling by the tumor suppressor kinase Chk2. Mol Cell 2002; 9:1045-54. 74. Walworth N, Davey S, Beach D. Fission yeast chk1 protein kinase links the rad checkpoint pathway to cdc2. Nature 1993; 363:368-71. 75. al-Khodairy F, Fotou E, Sheldrick KS, Griffiths DJ, Lehmann AR, Carr AM. Identification and characterization of new elements involved in checkpoint and feedback controls in fission yeast. Mol Biol Cell 1994; 5:147-60. 76. Bentley NJ, Holtzman DA, Flaggs G, Keegan KS, DeMaggio A, Ford JC, et al. The Schizosaccharomyces pombe rad3 checkpoint gene. EMBO J 1996; 15(:6641-51. 77. Lindsay HD, Griffiths DJ, Edwards RJ, Christensen PU, Murray JM, Osman F, et al. S-phase-specific activation of Cds1 kinase defines a subpathway of the checkpoint response in Schizosaccharomyces pombe. Genes Dev 1998; 12:382-95.

396

78. Jackson AL, Pahl PM, Harrison K, Rosamond J, Sclafani RA. Cell cycle regulation of the yeast Cdc7 protein kinase by association with the Dbf4 protein. Mol Cell Biol 1993; 13:2899-08. 79. Dowell SJ, Romanowski P, Diffley JF. Interaction of Dbf4, the Cdc7 protein kinase regulatory subunit, with yeast replication origins in vivo. Science 1994; 265:1243-6. 80. Pasero P, Duncker BP, Schwob E, Gasser SM. A role for the Cdc7 kinase regulatory subunit Dbf4p in the formation of initiation-competent origins of replication. Genes Dev 1999; 13:2159-76. 81. Weinreich M, Stillman B. Cdc7p-Dbf4p kinase binds to chromatin during S phase and is regulated by both the APC and the RAD53 checkpoint pathway. EMBO J 1999; 18:5334-46. 82. Dohrmann PR, Oshiro G, Tecklenburg M, Sclafani RA. RAD53 regulates DBF4 independently of checkpoint function in Saccharomyces cerevisiae. Genetics 1999; 151:965-77. 83. Duncker BP, Shimada K, Tsai-Pflugfelder M, Pasero P, Gasser SM. An N-terminal domain of Dbf4p mediates interaction with both origin recognition complex (ORC) and Rad53p and can deregulate late origin firing. Proc Natl Acad Sci USA 2002; 99:16087-92. 84. Li J, Lee GI, Van Doren SR, Walker JC. The FHA domain mediates phosphoprotein interactions. J Cell Sci 2000; 113 Pt 23:4143-9. 85. Durocher D, Jackson SP. The FHA domain. FEBS Lett 2002; 513:58-66.

Cell Cycle

2003; Vol. 2 Issue 4