This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Guillemain, G.
Right arrow Articles by Marsolier-Kergoat, M.-C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Guillemain, G.
Right arrow Articles by Marsolier-Kergoat, M.-C.

 Previous Article  |  Next Article 

Molecular and Cellular Biology, May 2007, p. 3378-3389, Vol. 27, No. 9
0270-7306/07/$08.00+0     doi:10.1128/MCB.00863-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Mechanisms of Checkpoint Kinase Rad53 Inactivation after a Double-Strand Break in Saccharomyces cerevisiae{triangledown}

Ghislaine Guillemain,1,{dagger},{ddagger} Emilie Ma,1,{dagger} Sarah Mauger,1 Simona Miron,2 Robert Thai,1 Raphaël Guérois,1 Françoise Ochsenbein,1 and Marie-Claude Marsolier-Kergoat1*

CEA, Direction des Sciences du Vivant, Institut de Biologie et de Technologies de Saclay, 91191 Gif-sur-Yvette Cedex, France,1 Institut Curie, INSERM U759, 91405 Orsay, France2

Received 15 May 2006/ Returned for modification 5 July 2006/ Accepted 9 February 2007


arrow
ABSTRACT
 
In Saccharomyces cerevisiae, double-strand breaks (DSBs) activate DNA checkpoint pathways that trigger several responses including a strong G2/M arrest. We have previously provided evidence that the phosphatases Ptc2 and Ptc3 of the protein phosphatase 2C type are required for DNA checkpoint inactivation after a DSB and probably dephosphorylate the checkpoint kinase Rad53. In this article we have investigated further the interactions between Ptc2 and Rad53. We showed that forkhead-associated domain 1 (FHA1) of Rad53 interacts with a specific threonine of Ptc2, T376, located outside its catalytic domain in a TXXD motif which constitutes an optimal FHA1 binding sequence in vitro. Mutating T376 abolishes Ptc2 interaction with the Rad53 FHA1 domain and results in adaptation and recovery defects following a DSB. We found that Ckb1 and Ckb2, the regulatory subunits of the protein kinase CK2, are necessary for the in vivo interaction between Ptc2 and the Rad53 FHA1 domain, that Ckb1 binds Ptc2 in vitro and that ckb1{Delta} and ckb2{Delta} mutants are defective in adaptation and recovery after a DSB. Our data thus strongly suggest that CK2 is the kinase responsible for the in vivo phosphorylation of Ptc2 T376.


arrow
INTRODUCTION
 
The DNA checkpoint is a surveillance mechanism that detects DNA lesions or replication blocks and coordinates various responses such as cell cycle arrests and transcriptional or posttranscriptional modifications. This mechanism is present in all eukaryotes and has been particularly analyzed in the yeast Saccharomyces cerevisiae, where it was originally identified (14, 53). In S. cerevisiae, activation of the DNA checkpoint by DNA lesions depends essentially on two sets of proteins, Rad24 and the PCNA-like trimer Rad17-Mec3-Ddc1, on the one hand, and the ATR homolog, the phosphatidylinositol 3-kinase-like Mec1 (in complex with an auxiliary subunit Ddc2), on the other hand (reviewed in references 28 and 58). Both the Rad17-Mec3-Ddc1 and the Mec1-Ddc2 complexes have been shown to be simultaneously and independently recruited to a double-strand break (DSB) artificially induced by the HO endonuclease (15, 29). Once activated, Mec1 induces the phosphorylation and the activation of two central transducers, the Rad53 and Chk1 kinases, which subsequently phosphorylate downstream effectors. The phosphorylation of Rad53 and Chk1 also depends on so-called "adaptors," Rad9 in the case of DNA damage and Mrc1 in the case of replication blocks and DNA lesions during S phase (for a review on Rad53 activation, see reference 33).

Rad53 plays a central part in S. cerevisiae DNA checkpoint: it controls the majority of the DNA damage responses and rad53{Delta} cells are strongly hypersensitive to all genotoxic stresses. Rad53 is the founding member of the conserved family of FHA (forkhead associated) domain-containing checkpoint kinases, which also includes mammalian Chk2 and Schizosaccharomyces pombe Cds1. It contains two FHA domains, FHA1 and FHA2, flanking the protein catalytic domain. FHA domains are protein-protein interaction domains that specifically bind phosphothreonine residues (for reviews, see references 9 and 22). Both Rad53 FHA domains are required for DNA damage-dependent Rad53 activation (they both bind the Rad9 adaptor) and probably for transducing the checkpoint signal to some downstream targets (7, 8, 35, 36, 41, 46).

Like all stress responses, the responses elicited by the DNA checkpoint have to be inactivated for cells to regain physiological growth conditions. Cell cycle arrests, in particular, have to be suppressed since their maintenance brings about cell lysis and corresponds to reproductive death (49). Inactivation of the DNA checkpoint-induced responses (in particular, resumption of the cell division cycle) was shown to occur in two situations: once the DNA lesions are repaired (a process called "recovery") or in the continued presence of a limited amount of irreparable DNA damage (a process called "adaptation"). Adaptation to DNA damage has been studied so far mostly in S. cerevisiae and after the occurrence of HO-induced DSBs that cannot be repaired by homologous recombination (17, 39, 49). A single, irreparable DSB is sufficient to trigger a prolonged DNA checkpoint-dependent G2/M arrest, but wild-type cells can ultimately adapt and resume cell cycle progression even if they still harbor a broken chromosome. Similarly, the process of recovery from DNA damage in S. cerevisiae has been studied primarily after formation of an HO-induced DSB whose repair takes place in 6 to 10 h (51).

Adaptation and recovery are brought about by inactivating upstream elements of the DNA checkpoint rather than downstream effectors that block cell cycle progression. Several studies have demonstrated that Rad53 inactivation is a key event of both adaptation and recovery (20, 34, 51). Thus, adaptation and recovery are always accompanied by the disappearance of Rad53 phosphorylated forms, and conversely, all adaptation- and recovery-defective mutants exhibit a persistent phosphorylation of Rad53 (19, 20, 34, 51). Moreover, degradation of thermosensitive forms of Rad53 or Mec1 is sufficient to suppress the cell cycle arrest of adaptation-defective mutants following HO-induced DSB (34).

Adaptation to irreparable DSBs seems to be influenced by many parameters. Thus, whereas wild-type cells can resume cell cycle progression after the occurrence of a single irreparable DSB, they are unable to do so in the presence of two irreparable DSBs, which illustrates the strong sensitivity of adaptation to the amount of DNA damage (17). Adaptation to a single DSB in S. cerevisiae was also shown to be altered by mutations in genes encoding proteins involved in homologous recombination and the processing of DSBs, like Rad51, Tid1, Yku70, Yku80, Srs2, or Rfa1 (17-19, 51). These observations suggested either that these proteins could be implied in the assessment of the amount of DNA lesions or that their absence or mutation could modify the processing of the DSBs and, consequently, the nature of the DNA lesions and the resulting activation of the DNA checkpoint (with the underlying assumption that the stronger the DNA checkpoint activation, the more difficult the adaptation). Other mutants, affected in the polo kinase Cdc5 and in the regulatory subunits Ckb1 and Ckb2 of the protein kinase CK2, were also isolated as adaptation defective, but Cdc5 and CK2 functions in adaptation have remained elusive (49).

We have demonstrated that the protein phosphatase 2C (PP2C) phosphatases Ptc2 and Ptc3 are required for DNA checkpoint inactivation after a DSB. We have shown that Ptc2 specifically binds the Rad53 FHA1 domain, which strongly suggests that Ptc2 and Ptc3 inactivate Rad53-dependent pathways and contribute to adaptation and recovery by dephosphorylating Rad53. In this article we have investigated further the interactions between Ptc2 and Rad53. We show that the FHA1 domain of Rad53 interacts with a specific threonine of Ptc2, T376, located in a TXXD motif which constitutes an optimal FHA1 binding sequence in vitro. Mutating T376 abolishes Ptc2 interaction with the Rad53 FHA1 domain and results in adaptation and recovery defects. We found that Ckb1 and Ckb2 are necessary for the in vivo interaction between Ptc2 and Rad53 FHA1, that Ckb1 binds Ptc2 in vitro, and that ckb1{Delta} and ckb2{Delta} mutants are defective in adaptation and recovery after a DSB. Taken together, our data strongly suggest that CK2 is the kinase responsible for the in vivo phosphorylation of Ptc2 T376, which could have some implications for other Rad53 FHA1 ligands.


arrow
MATERIALS AND METHODS
 
Strains and plasmids. The yeast strains used in this study are listed in Table 1. Genes were disrupted or tagged by PCR targeting using the Kluyveromyces lactis URA3 gene (KlURA3) (16) or the kanMX cassette (52). JKM139 ckb1{Delta} cells with a mutated, uncuttable HO site were selected as mutants able to form viable colonies on galactose-containing plates. pGBT9/GAL11 was a kind gift from B. Guglielmi (CEA/Saclay). The pGEX1{lambda}T/PTC2 construct encoding the glutathione S-transferase (GST)-Ptc2 fusion has been previously described (54) and was a generous gift from A. A. Welihinda and R. J. Kaufman. The two-hybrid plasmids pACTIIst/RAD53(7-164) and pGBT9/PTC2(1-464) have been described previously (20). The pGBT9/PTC2(1-428) construct was fortuitously recovered as a mutated pGBT9/PTC2(1-464) construct containing a stop codon at position 429. The sequences encoding Ptc2 residues 1 to 314 [Ptc2(1-314)], Ptc2(174-355), and Ptc2(295-428) were amplified by PCR and cloned between the EcoRI and PstI sites of pGBT9 (Clontech), so as to create pGBT9/PTC2(1-314), pGBT9/PTC2(174-355), and pGBT9/PTC2(295-428), respectively. The plasmid used to complement PTC2 deletion, pRS314/PTC2M3-11, contains a fragment of PTC2 starting 293 bp before the translation initiation codon and ending about 100 bp after the stop codon. This fragment had been previously isolated as a suppressor of a hyperactive allele of RAD53 (26) and was subcloned into pRS314 (42) SalI and NotI sites. The T363A and T376A mutations in pGBT9/PTC2(1-428), pGEX1{lambda}T/PTC2, and pRS314/PTC2M3-11 were realized using a QuikChange site-directed mutagenesis system (Stratagene). In order to express a hemagglutinin (HA)-tagged version of Ckb1, the full-length coding sequence of CKB1 was amplified by PCR and first cloned into pACTIIst NcoI and EcoRI sites. The BglII/XhoI fragment encoding HA-Ckb1 was then extracted from the pACTIIst plasmid and subcloned into p424GAL1 BamHI and XhoI sites (behind the GAL1 promoter) to construct p424GAL1/HA-CKB1. The sequence encoding the Rad53 FHA1 domain containing residues 1 to 164 [FHA1(1-164)] was cloned in the pETM-30 vector (a gift from G. Stier, EMBL, Heidelberg, Germany). All constructs were verified by sequencing.


View this table:
[in this window]
[in a new window]

  		TABLE 1. Yeast strains

Other materials. Phosphorylated and nonphosphorylated Ptc2 peptides [373IDD(pT)DADTDAENL385, 372DIDD(pT)DADTDAE383, 373IDDTDADTDAENL385, and 372DIDDTDADTDAE383; in the sequences pT is phosphothreonine] were produced by chemical synthesis (CovalAb and Asynt). 373IDD(pT)DADTDAENL385 was used for the generation of rabbit polyclonal antibodies (CovalAb). Enzyme-linked immunosorbent assays demonstrated that the serum tested had similar affinity for the phosphorylated and the nonphosphorylated Ptc2 peptides. The antibodies were affinity purified using the GST-Ptc2 fusion protein prepared from Escherichia coli extracts as described by Leroy et al. (20).

Analysis of adaptation and recovery after HO cutting. JKM139 and YMV2 derivatives were grown overnight in yeast extract-peptone (YP) medium containing 2% (wt/vol) raffinose (YPRaffinose). HO cutting was induced by the addition of 2% galactose at time zero. After 6 h, when cells were synchronized in G2/M, an aliquot was spotted on YP plates containing 2% galactose for the microcolony assay. Plates were incubated at 30°C and inspected by microscopic observation after 24 h (200 items, comprised of either cells in G2/M, microcolonies, or viable colonies, were examined for each strain). The number of cells or buds per microcolony was determined to monitor adaptation. Recovery was evaluated by counting the numbers of microcolonies, viable colonies, and cells arrested in G2/M.

Western blotting analysis used to monitor Rad53 phosphorylation or to quantify the expression of the PTC2 constructs was performed as described previously (26), using either the Rad53 (yC-19) antibody (Santa Cruz Biotechnology) or the affinity-purified antibody directed against Ptc2 (CovalAb) and described above.

Analysis of DSB repair was carried out as described previously (51), except that genomic DNA was digested by KpnI and StuI and that the blots were probed with the 0.5-kb EcoRV-KpnI fragment of LEU2 coding sequence and with the 1.5-kb XhoI fragment of RAD53 (to have an internal control for normalizing the repair signal). Blots were analyzed by using a Molecular Dynamics PhosphorImager.

Protein purification, nuclear magnetic resonance (NMR) spectroscopy, isothermal titration calorimetry (ITC), and liquid chromatography-electrospray ionization (LC-ESI) ion trap mass spectrometry (MS) analysis. E. coli strain BL21-Gold(DE3) (Stratagene) carrying the pETM-30/FHA1(1-164) plasmid was grown in M9 minimal medium supplemented with (15NH4)2SO4 (0.5 g liter–1; Eurisotop) as the sole nitrogen source. After IPTG (isopropyl-1-thio-ß-D-galactopyranoside) induction at an optical density of 0.8, cells were lysed by freeze-thaw and disrupted by sonication. The His6-GST-TEV site-FHA1(1-164) (where TEV is tobacco etch virus) fusion proteins were immobilized on glutathione (GSH)-coated agarose beads (Sigma), eluted with an excess of GSH, and cleaved using a His6-tagged TEV protease (1% [wt/wt] protease/fusion protein). A Ni-nitrilotriacetic acid agarose column (QIAGEN) was used to trap the His6-tagged TEV protease and the His6-tagged GST.

NMR samples were prepared in 10 mM sodium phosphate buffer, pH 6.5, containing 1 mM dithiothreitol, 0.1% NaN3, 1 mM EDTA, 0.1 mM disuccinimidyl suberate, and 10% D2O. The amounts of FHA1 and Ptc2 peptides were precisely measured by amino acid analysis. The concentration of the FHA1 domain was 163 µM, and the molar ratio FHA1:peptide at the end of the titration was 1:4. NMR experiments were carried out on a Bruker Advance-700 spectrometer equipped with a cryoprobe at 293 K. Experimental conditions were identical to those used in Yuan et al. (56).

ITC experiments were performed using a MicroCal MCS instrument (Northampton, MA) at 30°C. A 15 µM solution of the FHA1 domain in the calorimeter cell (1.34 ml) was titrated by a 200 µM solution of the Ptc2 peptide 372DIDD(pT)DADTDAE383 using automatic injections of 6 µl. Integration of the peaks corresponding to each injection and correction for the baseline were done using Origin-based software provided by the manufacturer. Fitting of the data to a single-site binding model yielded the value of the equilibrium dissociation constant (Kd).

LC coupled to tandem MS (LC-MS/MS) of the Ptc2 peptide 372DIDDTDADTDAE383 before and after treatment with CK2 was performed by high-performance LC (Agilent 1100; Santa Clara, CA) coupled online through a flow splitter to an ESI ion trap Esquire HCT mass spectrometer (Bruker, Germany). Samples were acidified by concentrated formic acid prior to desalting and separation using a reverse-phase column (Atlantis dC18; 150-mm by 4.6-mm internal diameter; 5-µm particle size; 10-nm porosity; Waters) with a linear gradient of acetonitrile-water in 0.1% formic acid. The mass spectrometer was optimized for detection of compounds in the mass range of 200 to 1,600, and mass isolation width (m/z) was set to 1.0 for MS/MS sequencing.

GST pull-down assays and CK2 protein kinase assay. GST and the GST-Ptc2 and GST-Ptc2T376A fusion proteins were purified from E. coli extracts as described previously (20). JKM139 ckb1{Delta} cells with either an intact or an uncuttable HO site (strain MCM449 or MCM714, respectively) containing the p424GAL1/HA-CKB1 plasmid were grown overnight in YPRaffinose, and the expression of both the HA-Ckb1 fusion and the HO endonuclease was induced by galactose for 6 h. Native extracts (0.5 mg of protein) were incubated with GST, GST-Ptc2, or GST-Ptc2T376A proteins bound to GSH-coated agarose beads in 400 µl of HEPES buffer (20 mM HEPES [pH 7.5], 1 mM dithiothreitol, 0.5% NP-40, plus complete protease inhibitor mix) containing various concentrations of NaCl for 2 h at 4°C on a rotating wheel. Beads were then washed in HEPES buffer, and the protein complexes were eluted by boiling in sample buffer prior to sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Western blots were revealed with the anti-HA 16B12 antibody (Promega).

For the CK2 kinase assay, GST-Ptc2 and GST-Ptc2T376A fusion proteins purified as described in Leroy et al. (20) and immobilized on GSH-coated agarose beads were cleaved by thrombin (MP Biomedicals), and the Ptc2 and Ptc2T376A proteins were recovered by elution. A total of 0.05 µg of each protein was treated with the human CK2 kinase (New England BioLabs), according to the manufacturer's instructions.


arrow
RESULTS
 
Ptc2 threonine T376 is required for the interaction between Ptc2 and the Rad53 FHA1 domain. We had previously shown by two-hybrid analysis and by GST pull-down assay that Ptc2 and Rad53 interact in vivo and in vitro, and we had characterized more specifically a strong interaction (as monitored by the activation of the reporter genes in the two-hybrid system) between Ptc2 and the FHA1 domain of Rad53, which required FHA1 phosphopeptide binding function (20). We fortuitously recovered a mutant construct encoding a fusion between the Gal4 DNA binding domain (Gal4BD) and a fragment of Ptc2 encompassing amino acids 1 to 428 [Ptc2(1-428)] which interacted with the Rad53 FHA1 domain as strongly as the full-length Gal4BD-Ptc2 protein (Fig. 1B). We therefore used this construct in subsequent experiments to characterize Ptc2 interaction with FHA1. We first tested three Ptc2 fragments for their ability to interact with FHA1 (Fig. 1A): (i) the Ptc2 catalytic domain, characteristic of the members of the PP2C family [Ptc2(1-314)], (ii) part of the Ptc2 C-terminal domain which shows no homology with any other yeast protein but Ptc3 [Ptc2(295-428)], and (iii) Ptc2(174-355), a fragment which was shown to interact with the Ire1 kinase (54). As shown in Fig. 1B (lines 1 to 6), only the Ptc2 C-terminal domain interacted with FHA1 in the two-hybrid assay.


Figure 1
View larger version (30K):
[in this window]
[in a new window]

  		FIG. 1. Rad53 FHA1 interacts with Ptc2 C-terminal domain and Ptc2 T376 is required for this interaction. (A) A schematic showing Ptc2 domains. (B) The empty vector pACTIIst or the plasmid pACTIIst/RAD53(7-164) expressing a protein fusion between Gal4 activating domain (Gal4AD) and Rad53 FHA1 were introduced into the Y190 tester strain along with plasmids harboring different parts of PTC2 fused to the sequence encoding Gal4BD [pGBT9/PTC2(1-428), pGBT9/PTC2(1-314), pGBT9/PTC2(174-355), and pGBT9/PTC2(295-428)]. The two-hybrid interaction was revealed by growth on plates lacking histidine (–His) complemented with 25 mM 3-amino-triazole (3AT) and by X-Gal (5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside) staining. (C) Expression levels of the Gal4BD-Ptc2(1-428) and Gal4BD-Ptc2(1-428)T376A fusion proteins. Total protein extracts prepared from Y190 cells carrying the pGBT9/PTC2(1-428), the pGBT9(1-428)T376A, or the pGBT9 plasmid [along with the pACTIIst/RAD53(7-164) construct] were analyzed by Western blotting using antibodies directed against Ptc2. (D) Sequence alignment of part of the C-terminal domains of Ptc2 and Ptc3. The TXXD motifs of Ptc2 and Ptc3 sequences are shaded in gray. The acidic regions bordering Ptc2 T376 and Ptc3 T366 are boxed.

Two groups of investigators, using different phosphopeptide libraries, had found that Rad53 FHA1 specifically bound phosphothreonines inside pTXXD motifs in vitro (10, 23). We therefore considered two threonines of the Ptc2 C-terminal domain, T363 and T376, located within TXXD sequences, as candidates for binding by FHA1. Mutating T376 into alanine abolished Ptc2 interaction with FHA1, whereas mutating T363 did not affect the interaction (Fig. 1B, compare lines 1 and 7 to 8). Of importance, the point mutation T376A is unlikely to affect Ptc2 catalytic activity since the truncation mutant protein Ptc2(1-355) has a wild-type phosphatase activity in vitro (55). We also verified that the T376A mutation did not reduce the expression level of the Gal4BD-Ptc2(1-428) construct (Fig. 1C).

These observations strongly suggest that FHA1 binds Ptc2 pT376. This hypothesis correlates well with the facts that the 376TXXD motif is also present in the Ptc3 sequence (in contrast to 363TXXD) and is surrounded by an acidic sequence which could favor the formation of a protruding loop (Fig. 1D). Liao et al. had also found that FHA1 exhibited a strong preference toward pTXXD peptides containing an alanine at the phosphothreonine +2 position (23), and it can be noted that the 376TDAD sequence presents this feature.

FHA domains have little affinity for unphosphorylated threonines in vitro (8, 10). We had also found that Rad53 in vitro interaction with purified GST-Ptc2 protein produced in E. coli was strictly dependent on prior in vitro phosphorylation of GST-Ptc2 (20). The fact that Ptc2 and FHA1 are found to interact in the two-hybrid assay in the absence of genotoxic stress suggests that at least a fraction of Ptc2 proteins is permanently phosphorylated and interacts constitutively with Rad53 in vivo, independently of the occurrence of DNA damage.

Rad53 FHA1 binds in vitro to a phosphopeptide encompassing Ptc2 pT376. The direct binding of FHA1 to the phosphothreonine peptide 372DIDD(pT)DADTDAE383 derived from Ptc2 sequence [pT(Ptc2)] was probed using NMR spectroscopy. The 1H-15N heteronuclear single-quantum coherence spectrum of the free FHA1 domain was nearly identical to the one published by Yuan et al. (56), thus allowing the assignment of five signals corresponding to R35, R70, R83, S85, and N86 (Fig. 2). As observed for the Rad9-derived 188SLEV(pT)EADATFVQ200 phosphopeptide [pT(Rad9)] which binds FHA1 with high affinity (56), the exchange rate between FHA1 and pT(Ptc2) was slow compared to the chemical shift time scale. The chemical shift variations observed after addition of pT(Ptc2) were virtually identical in magnitude and direction to those observed for pT(Rad9) (56) (Fig. 2A). In particular, S85, N86, and the side chain proton of R70 (involved in the binding of the phosphothreonine phosphate group) showed large chemical shift variations, whereas the signal corresponding to NH{varepsilon}/R35 remained unchanged. A large variation in the signal corresponding to the R83 side chain was also observed, suggesting that Ptc2 D379 (at the phosphothreonine +3 position) formed a salt bridge with FHA1 R83. These data indicated a similar binding mode for pT(Ptc2) and pT(Rad9), different from that of the 301NDPD(pT)LEIYS310 peptide from the Mdt1 protein [pT(Mdt1)], another FHA1 ligand (25, 56). We verified as a control that addition of the nonphosphorylated Ptc2 peptide led to almost no signal variation (Fig. 2B).


Figure 2
View larger version (32K):
[in this window]
[in a new window]

  		FIG. 2. 1H-15N heteronuclear single-quantum coherence spectra showing Ptc2 peptide binding. (A) Titration with the phosphorylated pT(Ptc2) peptide 372DIDD(pT)DADTDAE383. (B) Titration with the nonphosphorylated Ptc2 peptide. The spectra of the FHA1 domains in free and complexed forms (after addition of a 1:4 peptide ratio) are represented in blue and red, respectively. NH signals of S85 and N86 are indicated. Circled peaks are folded NH{varepsilon} resonance signals of arginine residues R35, R70, and R83. (C) Thermogram (upper panel) and the binding isotherm (lower panel) of the titration of FHA1 (15 µM) with the phosphorylated pT(Ptc2) peptide (200 µM) at 30°C. The thermogram and the isotherm were fitted to a single-site interaction model with a stoichiometry of 1:1.

The thermodynamic parameters of the molecular interactions between FHA1 and pT(Ptc2) were investigated by ITC. Figure 2C shows the thermogram and the binding isotherm curves. pT(Ptc2) was found to bind tightly FHA1 with a Kd of 2.3 µM, intermediate between the Kd values observed for binding of FHA1 to pT(Rad9) (0.36 µM [23]) and to pT(Mdt1) (15 µM [25]).

These data clearly indicate that Rad53 FHA1 interacts directly with a peptide encompassing Ptc2 pT376 in vitro and support our hypothesis that a similar, direct interaction occurs in vivo between Rad53 FHA1 and Ptc2.

The ptc2T376A allele is defective for adaptation and recovery. If FHA1 binds Ptc2 pT376 and if this binding is required for Rad53 inactivation by Ptc2, we should expect strains harboring the ptc2T376A allele to be defective in adaptation and recovery. We studied adaptation and recovery as previously described (17, 51), using strains that carry a galactose-inducible HO gene and that have a deletion of both HML and HMR donor sequences. An irreparable HO cut is produced at the MAT locus in the adaptation-tester strain JKM139. The MAT locus is deleted in the recovery-tester strain YMV2, and HO cutting is induced at a unique HO cleavage site artificially inserted at the LEU2 locus. A fragment with part of the LEU2 sequence is inserted about 30 kb centromere-distal to the HO site so as to allow repair of the HO break by single-strand annealing (SSA) between the two homologous LEU2 sequences.

Adaptation- and recovery-tester strains JKM139 and YMV2 were deleted of the PTC2 gene and transformed with vectors harboring either the wild-type PTC2 or mutant ptc2T376A alleles. We confirmed by Western blotting that the PTC2 and ptc2T376A alleles were expressed at the same levels (Fig. 3H). Strains were grown on raffinose medium overnight, synchronized in G2/M phase by induction of HO expression with galactose for 6 h, and plated on galactose-containing plates. Cells were monitored for adaptation or recovery after 24 h (Fig. 3). Adaptation-tester cells with a wild-type PTC2 allele were able to adapt to the HO cut and to resume mitosis, producing microcolonies of three to more than seven cells, which eventually die because essential genes of the broken chromosome III are lost (17). As previously shown, about 60% of ptc2{Delta} cells remained arrested as large-budded cells after HO cutting (20). We found that ptc2T376A cells were almost as defective in adaptation as ptc2{Delta} cells since about 50% remained blocked in G2/M (Fig. 3A). Kinetic analyses of DNA content and of Rad53 phosphorylation after HO cutting confirmed the adaptation defect of the ptc2T376A mutant (Fig. 3B and C). Previous studies of wild-type cells have shown that Rad53 phosphorylation and kinase activity start increasing about 2 h after formation of the HO-induced DSB, remain elevated for 8 to 12 h, and decrease at the time cells adapt (34). As shown in Fig. 3B, Rad53 phosphorylated forms practically disappeared 12 h after galactose induction in PTC2 cells whereas they were still predominant after 15 h in the ptc2{Delta} and ptc2T376A mutants. The fluorescence-activated cell sorter (FACS) profiles of ptc2{Delta} and ptc2T376A cells were also comparable: whereas PTC2 cells were arrested for 5 to 8 h and resumed their division cycle about 12 h after HO cutting, the ptc2{Delta} and ptc2T376A mutants remained permanently blocked in G2/M (Fig. 3C).


Figure 3
View larger version (56K):
[in this window]
[in a new window]

  		FIG. 3. ptc2T376A strains are defective for adaptation and recovery after a DSB. (A, B, and C) The ptc2T376A allele is defective for adaptation. Either the empty vector pRS314 or plasmids harboring the wild-type PTC2 gene (pRS314/PTC2M3-11) or the ptc2T376A allele (pRS314/PTC2M3-11T376A) were introduced into the adaptation-tester strain JKM139 with a deletion of PTC2 (strain MCM413). Cells were grown overnight in raffinose medium, and HO expression was induced by galactose at time zero. Aliquots were taken at the indicated times after addition of galactose and analyzed by FACS (C) and by Western blotting to monitor Rad53 phosphorylation (B). Adaptation was monitored after 24 h by counting the number of buds and cells of microcolonies (A). (D, E, F, and G) The ptc2T376A allele is defective for recovery. Either the empty vector pRS314 or plasmids harboring PTC2 or ptc2T376A (pRS314/PTC2M3-11 and pRS314/PTC2M3-11T376A, respectively) were introduced into the recovery-tester strain YMV2 with a deletion of PTC2 (YMV56). Cells were grown overnight in raffinose medium, and HO expression was induced by galactose at time zero. Recovery (D), Rad53 phosphorylation (E), and cell cycle progression (F) were analyzed as described above. (G) The efficiency of HO break repair by SSA was analyzed by Southern blotting in YMV56 cells carrying either the empty vector pRS314 or the pRS314/PTC2M3-11T376A plasmid expressing ptc2T376A. The RAD53 fragment was used as an internal control for DNA loading. (H) Expression levels of the PTC2 and ptc2T376A alleles. Total protein extracts prepared from JKM139 ptc2{Delta} (strain MCM413) or YMV2 ptc2{Delta} (strain YMV56) cells carrying either the empty vector pRS314 or plasmids harboring PTC2 or ptc2T376A (pRS314/PTC2M3-11 and pRS314/PTC2M3-11T376A, respectively) were analyzed by Western blotting using antibodies directed against Ptc2.

The ptc2T376A mutant also proved defective for recovery after HO cutting (Fig. 3D to F). About 25% of ptc2T376A cells gave rise to viable colonies after HO cutting, whereas 55% and 12% of PTC2 and ptc2{Delta} cells, respectively, were able to do so (Fig. 3D). The phenotype of the ptc2T376A mutant was thus intermediate between the phenotypes of PTC2 and ptc2{Delta} cells, suggesting a residual activity of Ptc2 toward Rad53. The kinetic analyses of Rad53 phosphorylation and DNA content confirmed these observations. It has been previously shown that upon HO cutting in a wild-type strain, Rad53 phosphorylation is maintained for 7 or 8 h and disappears about 12 h after HO induction, whereas Rad53 phosphorylation persists much longer in ptc2{Delta} cells (20, 51). Here, we found that the Rad53 phosphorylation state in ptc2T376A cells was intermediate between the states of PTC2 and of ptc2{Delta} cells (Fig. 3E). Similarly, FACS analysis indicated that whereas PTC2 cells exited the G2/M arrest between 6 and 9 h after HO induction, ptc2{Delta} and ptc2T376A cells were still predominantly blocked in G2/M 9 h after HO cutting (Fig. 3F). Mutant ptc2{Delta}, ptc3{Delta}, and ptc2{Delta} ptc3{Delta} YMV2 cells were previously shown to exhibit wild-type SSA repair of the HO-induced DSB (20). We verified by Southern blotting that the introduction of the ptc2T376A allele in ptc2{Delta} YMV2 cells did not modify the extent or the kinetics of HO break repair (Fig. 3G), thus ruling out the hypothesis that the recovery defect of ptc2T376A mutants could be due to a defect in DNA repair.

In summary, we have demonstrated that Ptc2 T376 is required for adaptation and recovery after HO cutting, which substantiates the hypotheses that pT376 is the phosphothreonine bound by FHA1 and that the Ptc2-FHA1 interaction is instrumental in Rad53 inactivation.

Wild-type protein kinase CK2 is required for Ptc2-FHA1 interaction. We then sought to identify the kinase responsible for the phosphorylation of Ptc2 T376. Protein kinase CK2 appeared as a good candidate for two reasons: (i) the CK2 major consensus site corresponds to the core motif (S/T)(D/E)X(D/E) surrounded by several acidic residues (27), which 376TDAD matches perfectly (Fig. 1D); and (ii) we had found that the in vitro interaction between Rad53 and bacterially expressed GST-Ptc2 was strictly dependent on prior incubation of GST-Ptc2 with human CK2 (20). CK2 is a serine/threonine/tyrosine-directed enzyme expressed in all eukaryotes (for reviews, see references 24 and 27). The holoenzyme is a heterotetramer composed of two catalytic subunits (encoded in S. cerevisiae by the CKA1 and CKA2 genes) separately bound to a dimer of regulatory subunits (encoded by CKB1 and CKB2). The free catalytic subunits of CK2 also display high catalytic activity in the absence of the regulatory subunits, which seem to operate as a docking platform for binding substrates and/or substrate-directed effectors, thus modulating CK2 substrate specificity rather than its activity. If CK2 were responsible for phosphorylating Ptc2 T376, we would expect mutations in CK2 encoding genes to abolish specifically the interaction between Ptc2 and FHA1. As shown in Fig. 4A, the Ptc2-FHA1 interaction was suppressed in two-hybrid tester strains deleted of either CKB1 or CKB2, whereas the expression levels of the PTC2 constructs were similar in the wild type and in the ckb1{Delta} and ckb2{Delta} mutants (Fig. 4B). In contrast, the ckb1{Delta} and ckb2{Delta} mutations did not affect the activation of the reporter genes induced by the binding of the mediator subunit Gal11.


Figure 4
View larger version (52K):
[in this window]
[in a new window]

  		FIG. 4. (A) Wild-type CK2 is required for Ptc2/FHA1 interaction. The empty vector pACTIIst or the plasmid pACTIIst/RAD53(7-164) was introduced either into the wild-type (WT) tester strain Y190 or into ckb1{Delta} or ckb2{Delta} mutants (strains MCM630 and MCM460) along with the empty vector pGBT9 or with plasmids harboring sequences encoding either the full-length Ptc2, part of the Ptc2 C-terminal domain [Ptc2(295-428)], or the Gal11 mediator subunit [pGBT9/PTC2(1-464), pGBT9/PTC2(295-428), and pGBT9/GAL11, respectively]. Transformants were spotted onto plates lacking histidine complemented with 25 mM 3-amino-triazole and stained with X-Gal (5-bromo-4-chloro-3-indolyl-ß-D-galactopyranoside). (B) Expression levels of the Gal4BD-Ptc2(1-464) and Gal4BD-Ptc2(295-428) fusion proteins. Total protein extracts prepared from either Y190, Y190 ckb1{Delta} (strain MCM630; 1{Delta}) or Y190 ckb2{Delta} (strain MCM460; 2{Delta}) cells carrying either the pGBT9/PTC2(1-464) or the pGBT9/PTC2(295-428) plasmid along with the pACTIIst/RAD53(7-164) construct were analyzed by Western blotting using antibodies directed against Ptc2. (C and D) Ptc2 and CK2 interact in GST pull-down assay independently of Ptc2 T376 and of DNA damage. JKM139 ckb1{Delta} cells containing either an intact HO site (strain MCM449; HO cutting+) or an uncuttable HO site (strain MCM714; HO cutting–) were transformed with the p424GAL1/HA-CKB1 plasmid expressing HA-Ckb1 under the control of a galactose-inducible promoter. Cells were grown on raffinose medium overnight, and the expression of HO and HA-Ckb1 was induced by galactose for 6 h. Cell extracts were incubated with immobilized GST, GST-Ptc2, or GST-Ptc2T376A proteins in the presence of various concentrations of NaCl as indicated. Bound proteins were analyzed by Western blotting using anti-HA antibodies. (E) JKM139 ckb1{Delta} ptc2{Delta} (strain MCM800) carrying the p424GAL1/HA-CKB1 plasmid was grown on raffinose medium overnight, and the expression of HO and HA-Ckb1 was induced by galactose for 6 h. Cell extracts were incubated with GST, GST-Ptc2, or GST-Ptc2T376A proteins in the absence of NaCl, and bound proteins were analyzed as described above. (F) Coomassie staining showing the amounts of GST, GST-Ptc2, and GST-Ptc2T376A proteins used in the GST pull-down assays in panels C, D, and E. (G) Ptc2T376A is phosphorylated by CK2 in vitro to a lesser degree than Ptc2. Ptc2 and Ptc2T376A proteins (0.05 µg) were treated with CK2 in the presence of [{gamma}-32P]ATP. Phosphorylated proteins were revealed by PhosphorImager analysis. Coomassie staining shows equal quantities (1 µg) of the input proteins (only 1/20 of this amount was used for the phosphorylation tests). The relative phosphorylation value for Ptc2T376A represents the mean of two independent experiments done in duplicate or triplicate and the standard deviation. (H) The ESI-MS/MS spectrum of the monophosphorylated Ptc2 peptide 372DIDD(pT)DADTDAE383 detected after treatment of the nonphosphorylated peptide with CK2 was obtained from collision-induced dissociation fragmentation of the precursor ion at m/z 1416.5 Da (indicated with a black diamond). The amino acid sequence of the peptide was identified by the production in the ion trap of y-type (C terminus sequence, continuous lines) and b-type (N terminus sequence, broken lines) ions. The position of the phosphorylated threonine pT376 (threonine with an excess mass of 80 due to the phosphate group) is boxed.

Ptc2 interacts with Ckb1, one of the CK2 regulatory subunits. We then sought to demonstrate a direct interaction between Ptc2 and CK2. JKM139 ckb1{Delta} cells containing either an intact or a mutated uncuttable HO site were transformed with a plasmid expressing an HA-tagged version of Ckb1 under the control of the GAL1 promoter. The cells were grown overnight in YPRaffinose, and the expression of both HO and HA-Ckb1 was induced by galactose for 6 h. Native protein extracts were incubated with GST, GST-Ptc2, or GST-Ptc2T376A proteins immobilized on GSH-coated beads (Fig. 4F), and the bound proteins were analyzed by Western blotting. As shown in Fig. 4C, HA-Ckb1 remained specifically and stably bound to the GST-Ptc2 and GST-Ptc2T376A fusion proteins even in the presence of 1 M NaCl. This binding was observed with extracts from cells containing either the intact or the mutated HO sites, indicating that it was independent of the occurrence of a DSB (Fig. 4D). We also observed the interaction between GST-Ptc2T376A and HA-Ckb1 with extracts from ckb1{Delta} ptc2{Delta} cells (Fig. 4E), which ruled out the possibility that the interaction between Ptc2T376A and Ckb1 could be indirectly due to a homotypic interaction between Ptc2T376A and the endogenous Ptc2. These results suggest that in vivo CK2 can bind and phosphorylate Ptc2 constitutively, irrespective of the presence of DNA damage.

We had shown that the interaction between Rad53 and GST-Ptc2 was dependent on prior incubation of GST-Ptc2 with human CK2 (20). We verified that Ptc2 is, indeed, an in vitro substrate of CK2 (Fig. 4G). Interestingly, although Ptc2T376A was also phosphorylated by CK2, the amount of radiolabeled phosphate incorporated into Ptc2T376A was about 60% of that incorporated into the wild-type Ptc2, which indicated that T376 could be one of several CK2 phosphorylation sites present in Ptc2. To unambiguously demonstrate that Ptc2 T376 was a substrate for CK2 phosphorylation, the Ptc2 peptide 372DIDDTDADTDAE383 encompassing T376 was incubated with CK2, and the reaction products were analyzed by MS. Monophosphorylated, but not diphosphorylated, peptides were detected. Collision-induced dissociation fragmentation of the monophosphorylated peptides allowed their sequencing from both C-terminal and N-terminal ends and demonstrated the presence of the phosphothreonine pT376 whereas threonine T380 remained unphosphorylated (Fig. 4H).

CK2 regulatory subunits Ckb1 and Ckb2 are required for adaptation and recovery after a DSB. We expected that the deletions of CKB1 and CKB2, which abrogate the interaction between Ptc2 and FHA1, would also bring about a defect in adaptation and recovery. Indeed, as shown in Fig. 5A and C, tester strains with deletions of the CKB1 or CKB2 genes were found strongly defective in adaptation and recovery. More than 90% of ckb1{Delta} and ckb2{Delta} cells remained permanently blocked in G2/M after induction of an irreparable HO cut (Fig. 5A) and less than 20% of YMV2 ckb1{Delta} and ckb2{Delta} mutants were able to form viable colonies after HO cutting (Fig. 5C). These results confirmed the observations by Toczyski et al., who had demonstrated that ckb1{Delta} and ckb2{Delta} mutants are adaptation-defective after a DSB by using a different tester strain (49). The adaptation and recovery defects of the ckb1{Delta} and ckb2{Delta} mutants are also evident from an analysis of Rad53 phosphorylation. The decrease in Rad53 phosphorylated forms was much delayed in ckb1{Delta} and ckb2{Delta} mutants compared to wild-type cells (Fig. 5B and D). We tested HO break repair by Southern blotting, and we observed little difference between the wild type and CK2-defective cells (Fig. 5E), which suggests that the recovery defects of ckb1{Delta} and ckb2{Delta} mutants should result mostly from the defective dephosphorylation of Rad53 by Ptc2 and is in keeping with the observation that the recovery defects of the ptc2{Delta} ptc3{Delta}, ckb1{Delta}, and ckb2{Delta} mutants are comparable (as monitored by the percentages of viable colonies formed after HO cutting) (Fig. 5C). In contrast, we observed that the adaptation defects of the ckb1{Delta} and ckb2{Delta} mutants (assessed by the percentages of cells permanently arrested in G2/M) (Fig. 5A) were reproducibly more severe than the defect of ptc2{Delta} ptc3{Delta} cells, thus hinting at Ptc2-independent functions of CK2 in adaptation.


Figure 5
View larger version (54K):
[in this window]
[in a new window]

  		FIG. 5. Ckb1 and Ckb2 are required for adaptation and recovery after a DSB. (A and B) Mutants of the adaptation-tester strain JKM139 affected in CK2 regulatory subunits (ckb1{Delta} [strain MCM449] and ckb2{Delta} [MCM458]) are defective in adaptation. Wild-type, ckb1{Delta}, ckb2{Delta}, and ptc2{Delta} ptc3{Delta} cells were grown overnight in raffinose medium, and HO expression was induced by galactose at time zero. (A) Adaptation was monitored by the microcolony assay as described in the legend of Fig. 3. (B) Aliquots were taken at the indicated times after addition of galactose and analyzed by Western blotting. (C and D) Mutants of the recovery-tester strain YMV2 affected in CK2 regulatory subunits (ckb1{Delta} [strain MCM755] and ckb2{Delta} [MCM485]) are defective for recovery. Wild-type, ckb1{Delta}, ckb2{Delta}, and ptc2{Delta} ptc3{Delta} cells were grown overnight in raffinose medium, and HO expression was induced by galactose at time zero. Recovery (C) and Rad53 phosphorylation (D) were analyzed as described in the legend of Fig. 3. (E) The efficiency of HO break repair by SSA was analyzed by Southern blotting in YMV2 and YMV2 ckb2{Delta} cells.


arrow
DISCUSSION
 
Molecular mechanisms of Rad53 regulation by Ptc2. A model for Rad53 activation proposes that, following DNA damage, Rad53 is first phosphorylated by Mec1 on critical residues in a process stimulated by Rad9, which is also phosphorylated by Mec1 (47). This priming phosphorylation of Rad53 would then facilitate its binding to Rad9 and increase its local concentration to a level that allows its in trans autophosphorylation (12). Two groups have recently analyzed Rad53 phosphorylation sites. Smolka et al. have shown that Rad53 is constitutively phosphorylated on seven sites in the absence of DNA damage. After treatment with methyl methane sulfonate, the phosphorylation of these sites increases by a factor 1.7 to 14.3, and eight additional sites get phosphorylated (43). Durocher and collaborators have also determined the phosphorylation sites of a wild-type and a kinase-dead form of Rad53 after treatment with 4-nitroquinoline oxide (4-NQO) (47). Together these studies have characterized 32 phosphorylation sites on Rad53, only 10 of which have been found by both groups.

A first comment regarding these analyses concerns the fact that Rad53 is phosphorylated on several residues in the absence of DNA damage (43). Three of these phosphorylation sites conform to the (S/T)Q consensus for phosphatidylinositol 3-kinase-related protein kinases like Mec1 and Tel1, whereas three others appear to be proline-directed sites, hinting at cyclin-dependent kinase (CDK) phosphorylation. These observations thus suggest that at least a fraction of Rad53 could be constitutively phosphorylated by the Mec1/Tel1 kinases and the cell cycle CDK Cdc28 and constitute a substrate for regulatory phosphatases like Ptc2 and Ptc3. This hypothesis is substantiated by the fact that Ptc2 and Ptc3 interact with Rad53 FHA1 in the absence of DNA damage. The fact that Ptc2 and Ptc3 would regulate Rad53 basal activity by dephosphorylating it would account for the observations that deleting PTC2 suppresses the lethality of MEC1 deletion and that overexpression of PTC2 is lethal to mec1-21, rad53-21, and dun1{Delta} mutants (21). Alternatively, Ptc2 could operate by competing with other partners of Rad53 for binding to FHA1 and could reduce constitutively Rad53 interactions with activators like Rad9 (8) or potential substrates like Dbf4 (7) or Asf1 (41). It has to be noted that ptc2{Delta} ptc3{Delta} mutants do not show any hyperphosphorylation of Rad53 in the absence of genotoxic stress, which indicates that Rad53 hyperphosphorylation is strictly dependent on the activation of the upstream components of the DNA checkpoint and is not brought about by the inactivation of Ptc2 and Ptc3.

A second comment regards the observation that Rad53 phosphorylation sites are quite different after treatment with methyl methane sulfonate and with 4-NQO. This indicates that Rad53 phosphorylation may vary considerably according to the nature and the number of the DNA lesions (and probably other parameters like the cell cycle phase). Consequently, the phosphatases regulating Rad53 phosphorylation could also change according to these parameters. The variability of Rad53 phosphorylation sites according to genotoxic stresses could explain the a priori puzzling observation that ptc2{Delta} ptc3{Delta} cells are hypersensitive to only a subset of genotoxic stresses: they are deficient in recovery and adaptation following HO-induced DSB and are hypersensitive to camptothecin and 4-NQO (C. Clémenson and A. Peyroche, unpublished results) but are not hypersensitive to UV or ionizing radiations or to phleomycin or hydroxyurea (26).

Determining the residues targeted by Ptc2 and Ptc3 among the 32 phosphorylation sites that have already been identified in Rad53 is a difficult task as PP2C phosphatases, in contrast to most protein kinases which show a marked specificity for the residues surrounding the phosphoamino acid, do not exhibit a significant site specificity (see for reviews references 6 and 38). Given that several PP2C phosphatases were found to dephosphorylate mitogen-activated protein kinases and CDKs on threonine residues within their activation loops (4, 5, 31, 48), Ptc2 and Ptc3 could inactivate Rad53 by dephosphorylating T354 in its activation loop (T354 is one of the Rad53 phosphorylation sites determined by Sweeney et al. [47]).

Rad53 phosphorylation upon genotoxic stress results from a balance between the kinase activity of the upstream components of the DNA checkpoint and the phosphatase activity of Ptc2 and Ptc3 (plus unknown positive and negative regulators). The major effect of Ptc2 and Ptc3 on Rad53 phosphorylation is seen during Rad53 inactivation following a DSB. The decrease in Rad53 phosphorylated forms after a DSB could theoretically be brought about by a decrease in kinase activity and/or an increase in phosphatase activity toward Rad53. A decrease in kinase activity toward Rad53 when DSBs are repaired (recovery) can be easily understood as the upstream checkpoint components are no longer recruited onto DNA lesions and activated. Even in the case of adaptation, when DNA damage persists, Toczyski and collaborators have observed a loss of intensity of Ddc2-green fluorescent protein (GFP) foci about 12 h after induction of the DSB, correlating with the time at which cells undergo adaptation (29). The disappearance of Ddc2-GFP foci presumably reflects a decrease in Mec1 activity. Since Pelliciolli et al. have shown that Mec1 activity is needed to maintain Rad53 phosphorylation after a DSB (34), a mere decrease in Mec1 activity could have accounted for the inactivation of Rad53, which is not the case since Ptc2 and Ptc3 are required for adaptation and recovery. One could imagine that Ptc2 would be recruited specifically at the time of Rad53 inactivation via a regulatory phosphorylation inducing its binding by Rad53 FHA1. However, the facts that FHA1 binds Ptc2 in the absence of genotoxic stress and that Ptc2 is active as a monomere, without any regulatory subunit, suggest that Ptc2 activity toward Rad53 is constitutive. The events triggering Rad53 inactivation in cases of adaptation and recovery, therefore, seem to be directly linked to the inactivation of Mec1/Ddc2.

Specificity of Ptc2 and Ptc3 activity toward Rad53. Besides Ptc2 and Ptc3, four other S. cerevisiae proteins (Ptc1, Ptc4, Ptc5, and Ycr079w) belong to the PP2C family and could conceivably be involved in adaptation and recovery. However, overexpression of neither Ptc1, Ptc4, nor Ptc5 was able to rescue the adaptation defect of ptc2{Delta} ptc3{Delta} mutants, and deleting PTC4 affected neither adaptation nor recovery after HO cleavage (unpublished results). This specificity of Ptc2 and Ptc3 activity toward Rad53 can be explained by the importance of Rad53 FHA1 binding to the noncatalytic C-terminal moiety of Ptc2 and Ptc3 which is missing in all other PP2C phosphatases of S. cerevisiae. Substrate recognition by phosphatases has now been found in many cases to be mediated by targeting subunits and/or structural modules outside the catalytic core of the phosphatase (45, 50, 57), and Rad53 recognition by Ptc2 and Ptc3 appears to follow this general rule.

Functions and binding specificity of Rad53 FHA1 domain. Rad53 FHA1 is among the first and most studied FHA domains. Two groups of investigators independently found by screening peptide libraries that FHA1 specifically bound pTXXD motifs in vitro (10, 23). However, evidence suggesting that Rad9 390TXXV and Mdt1 305TXXI are not only in vitro but also in vivo recognition sites of FHA1 recently led to questions concerning FHA1 binding specificity in vivo (25, 36, 40). To our knowledge, our study provides the first evidence that the pTXXD motif defined in vitro is indeed a Rad53 FHA1 binding site in vivo. In vitro, threonine-phosphorylated peptides following the optimal sequence derived from library screening typically bind FHA domains with an affinity 10 to 100 times higher than peptides with more divergent sequences (3, 10, 23). However, it is not clear how the cellular context further tunes the selectivity of FHA1 binding for its targets, and whether proteins that harbor the optimal motif will be preferentially bound in vivo and overcome the effect of other binding partners is still an unanswered issue. Moreover, as was shown for the Cdc4 binding sites in Sic1 (30), multiple suboptimal FHA1 binding sites could act in concert and allow the setting of thresholds of protein kinase activity for the regulation of protein interactions. As the list of Rad53 FHA1 interactors keeps growing (Rad9 [8], Dbf4 [7], Asf1 [41], Ptc2 [20], Mdt1 [36], Sgs1 [2], and the septins Shs1, Cdc10, and Cdc11 [44]), competition and interplay between the various phosphorylated patterns will have to be further elucidated to fully understand the molecular logic underlying Rad53 interaction network.

We had previously demonstrated that the mutated FHA1-R70A and FHA1-S85A domains, which are deficient in phosphopeptide binding ability (10, 23), were unable to bind Ptc2 (20). We then sought to assess the function of FHA1 in adaptation and recovery. However, we found (our unpublished results) that the FHA1-defective rad53 mutants were partially deficient in checkpoint activation after HO-induced DSBs (as they are after other kinds of DNA damage [35, 41]), which prevented an estimation of FHA1 function specifically in Rad53 inactivation.

CK2 functions in adaptation and recovery after a DSB. We have shown that deletion of the CKB1 and CKB2 genes encoding CK2 regulatory subunits abrogates the interaction between Ptc2 and Rad53 FHA1 and that Ckb1 interacts with Ptc2, which suggests that CK2 is the kinase responsible for the in vivo phosphorylation of Ptc2 T376. The fact that Ptc2 interacts with both CK2 and FHA1 in the absence of genotoxic stress is in keeping with the observation that CK2 seems constitutively active, independently of second messengers or phosphorylation events (24, 37). Since TXXD, the FHA1 optimal binding motif in vitro, is included in the minimal consensus sequence for phosphorylation by CK2, CK2 could phosphorylate constitutively other ligands of FHA1, including Rad9, Asf1, Sgs1, or Dbf4, which possess several TXXD motifs and were found to bind FHA1 in the absence of genotoxic stress (2, 7, 8, 41).

CK2 has been shown to phosphorylate a large number of proteins (a recent inventory counts 307 CK2 substrates [27]) including topoisomerases I and II and the yeast cell cycle regulators Cdc28 or Cdc37 (27). It would therefore not be surprising if CK2 were also involved in adaptation or recovery not only through Ptc2 phosphorylation but also via other substrates. Indeed, we found that whereas the ptc2{Delta} ptc3{Delta}, ckb1{Delta}, and ckb2{Delta} mutants exhibit similar recovery defects after DSB, the ckb1{Delta} and ckb2{Delta} mutants were reproducibly more defective in adaptation than ptc2{Delta} ptc3{Delta} cells, hinting at additional functions of CK2 during adaptation.

Evidence has recently emerged for a crucial role of CK2 in cell survival, including cancer cell survival (1). In addition, the mechanisms of DNA checkpoint inactivation seem relatively conserved among eukaryotes. For example, the PP2C phosphatase Wip1, a human homolog of Ptc2 and Ptc3, was recently found to be involved in the attenuation of Chk2 activation (Chk2 is the mammalian homolog of Rad53) (11, 32). CK2 in higher eukaryotes could also promote cell survival after DNA damage in a way similar to what we have described in S. cerevisiae, which remains the subject of future studies.


arrow
ACKNOWLEDGMENTS
 
We thank A. A. Welihinda, R. J. Kaufman, and B. Guglielmi for the gift of plasmids; A. Peyroche for useful comments on the manuscript; and A. Lautrette and S. Dubois for assistance in the NMR and the mass spectrometry experiments, respectively.

This work was financed in part by the Association pour la Recherche sur le Cancer.


arrow
FOOTNOTES
 
* Corresponding author. Mailing address: Service de Biologie Intégrative et de Génétique Moléculaire, P.O. Box 22, CEA/Saclay, 91191 Gif-sur-Yvette Cedex, France. Phone: 33 1 69 08 83 54. Fax: 33 1 69 08 47 12. E-mail: mcmk{at}cea.fr Back

{triangledown} Published ahead of print on 26 February 2007. Back

{dagger} G.G. and E.M. contributed equally to this work. Back

{ddagger} Present address: Faculté de Médecine Necker, INSERM E363, 75730 Paris, France. Back


arrow
REFERENCES
 
    1
  1. Ahmed, K., D. A. Gerber, and C. Cochet. 2002. Joining the cell survival squad: an emerging role for protein kinase CK2. Trends Cell Biol. 12:226-230.[CrossRef][Medline]
    2. 2
  2. Bjergbaek, L., J. A. Cobb, M. Tsai-Pflugfelder, and S. M. Gasser. 2005. Mechanistically distinct roles for Sgs1p in checkpoint activation and replication fork maintenance. EMBO J. 24:405-417.[CrossRef][Medline]
    3. 3
  3. Byeon, I. J., S. Yongkiettrakul, and M. D. Tsai. 2001. 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. 314:577-588.[CrossRef][Medline]
    4. 4
  4. Cheng, A., P. Kaldis, and M. J. Solomon. 2000. Dephosphorylation of human cyclin-dependent kinases by protein phosphatase type 2C alpha and beta 2 isoforms. J. Biol. Chem. 275:34744-34749.[Abstract/Free Full Text]
    5. 5
  5. Cheng, A., K. E. Ross, P. Kaldis, and M. J. Solomon. 1999. Dephosphorylation of cyclin-dependent kinases by type 2C protein phosphatases. Genes Dev. 13:2946-2957.[Abstract/Free Full Text]
    6. 6
  6. Donella-Deana, A., M. Boschetti, and L. A. Pinna. 2003. Monitoring of PP2A and PP2C by phosphothreonyl peptide substrates. Methods Enzymol. 366:3-17.[Medline]
    7. 7
  7. Duncker, B. P., K. Shimada, M. Tsai-Pflugfelder, P. Pasero, and S. M. Gasser. 2002. 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 99:16087-16092.[Abstract/Free Full Text]
    8. 8
  8. Durocher, D., J. Henckel, A. R. Fersht, and S. P. Jackson. 1999. The FHA domain is a modular phosphopeptide recognition motif. Mol. Cell 4:387-394.[CrossRef][Medline]
    9. 9
  9. Durocher, D., and S. P. Jackson. 2002. The FHA domain. FEBS Lett. 513:58-66.[CrossRef][Medline]
    10. 10
  10. Durocher, D., I. A. Taylor, D. Sarbassova, L. F. Haire, S. L. Westcott, S. P. Jackson, S. J. Smerdon, and M. B. Yaffe. 2000. The molecular basis of FHA domain:phosphopeptide binding specificity and implications for phospho-dependent signaling mechanisms. Mol. Cell 6:1169-1182.[CrossRef][Medline]
    11. 11
  11. Fujimoto, H., N. Onishi, N. Kato, M. Takekawa, X. Z. Xu, A. Kosugi, T. Kondo, M. Imamura, I. Oishi, A. Yoda, and Y. Minami. 2006. Regulation of the antioncogenic Chk2 kinase by the oncogenic Wip1 phosphatase. Cell Death Differ. 13:1170-1180.[CrossRef][Medline]
    12. 12
  12. Gilbert, C. S., C. M. Green, and N. F. Lowndes. 2001. Budding yeast Rad9 is an ATP-dependent Rad53 activating machine. Mol. Cell 8:129-136.[CrossRef][Medline]
    13. 13
  13. Harper, J. W., G. R. Adami, N. Wei, K. Keyomarsi, and S. J. Elledge. 1993. The p21 Cdk-interacting protein Cip1 is a potent inhibitor of G1 cyclin-dependent kinases. Cell 75:805-816.[CrossRef][Medline]
    14. 14
  14. Hartwell, L. H., and T. A. Weinert. 1989. Checkpoints: controls that ensure the order of cell cycle events. Science 246:629-634.[Abstract/Free Full Text]
    15. 15
  15. Kondo, T., T. Wakayama, T. Naiki, K. Matsumoto, and K. Sugimoto. 2001. Recruitment of Mec1 and Ddc1 checkpoint proteins to double-strand breaks through distinct mechanisms. Science 294:867-870.[Abstract/Free Full Text]
    16. 16
  16. Langle-Rouault, F., and E. Jacobs. 1995. A method for performing precise alterations in the yeast genome using a recycable selectable marker. Nucleic Acids Res. 23:3079-3081.[Free Full Text]
    17. 17
  17. Lee, S. E., J. K. Moore, A. Holmes, K. Umezu, R. D. Kolodner, and J. E. Haber. 1998. Saccharomyces Ku70, mre11/rad50 and RPA proteins regulate adaptation to G2/M arrest after DNA damage. Cell 94:399-409.[CrossRef][Medline]
    18. 18
  18. Lee, S. E., A. Pellicioli, A. Malkova, M. Foiani, and J. E. Haber. 2001. The Saccharomyces recombination protein Tid1p is required for adaptation from G2/M arrest induced by a double-strand break. Curr. Biol. 11:1053-1057.[CrossRef][Medline]
    19. 19
  19. Lee, S. E., A. Pellicioli, M. B. Vaze, N. Sugawara, A. Malkova, M. Foiani, and J. E. Haber. 2003. Yeast Rad52 and Rad51 recombination proteins define a second pathway of DNA damage assessment in response to a single double-strand break. Mol. Cell. Biol. 23:8913-8923.[Abstract/Free Full Text]
    20. 20
  20. Leroy, C., S. E. Lee, M. B. Vaze, F. Ochsenbein, R. Guerois, J. E. Haber, and M. C. Marsolier-Kergoat. 2003. PP2C phosphatases Ptc2 and Ptc3 are required for DNA checkpoint inactivation after a double-strand break. Mol. Cell 11:827-835.[CrossRef][Medline]
    21. 21
  21. Leroy, C., C. Mann, and M. C. Marsolier. 2001. Silent repair accounts for cell cycle specificity in the signaling of oxidative DNA lesions. EMBO J. 20:2896-2906.[CrossRef][Medline]
    22. 22
  22. Li, J., G. I. Lee, S. R. Van Doren, and J. C. Walker. 2000. The FHA domain mediates phosphoprotein interactions. J. Cell Sci. 113:4143-4149.[Abstract]
    23. 23
  23. Liao, H., C. Yuan, M. I. Su, S. Yongkiettrakul, D. Qin, H. Li, I. J. Byeon, D. Pei, and M. D. Tsai. 2000. Structure of the FHA1 domain of yeast Rad53 and identification of binding sites for both FHA1 and its target protein Rad9. J. Mol. Biol. 304:941-951.[CrossRef][Medline]
    24. 24
  24. Litchfield, D. W. 2003. Protein kinase CK2: structure, regulation and role in cellular decisions of life and death. Biochem. J. 369:1-15.[CrossRef][Medline]
    25. 25
  25. Mahajan, A., C. Yuan, B. L. Pike, J. Heierhorst, C. F. Chang, and M. D. Tsai. 2005. FHA domain-ligand interactions: importance of integrating chemical and biological approaches. J. Am. Chem. Soc. 127:14572-14573.[CrossRef][Medline]
    26. 26
  26. Marsolier, M. C., P. Roussel, C. Leroy, and C. Mann. 2000. Involvement of the PP2C-like phosphatase Ptc2p in the DNA checkpoint pathways of Saccharomyces cerevisiae. Genetics 154:1523-1532.[Abstract/Free Full Text]
    27. 27
  27. Meggio, F., and L. A. Pinna. 2003. One-thousand-and-one substrates of protein kinase CK2? FASEB J. 17:349-368.[Abstract/Free Full Text]
    28. 28
  28. Melo, J., and D. Toczyski. 2002. A unified view of the DNA-damage checkpoint. Curr. Opin. Cell Biol. 14:237-245.[CrossRef][Medline]
    29. 29
  29. Melo, J. A., J. Cohen, and D. P. Toczyski. 2001. Two checkpoint complexes are independently recruited to sites of DNA damage in vivo. Genes Dev. 15:2809-2821.[Abstract/Free Full Text]
    30. 30
  30. Nash, P., X. Tang, S. Orlicky, Q. Chen, F. B. Gertler, M. D. Mendenhall, F. Sicheri, T. Pawson, and M. Tyers. 2001. Multisite phosphorylation of a CDK inhibitor sets a threshold for the onset of DNA replication. Nature 414:514-521.[CrossRef][Medline]
    31. 31
  31. Nguyen, A. N., and K. Shiozaki. 1999. Heat-shock-induced activation of stress MAP kinase is regulated by threonine- and tyrosine-specific phosphatases. Genes Dev. 13:1653-1663.[Abstract/Free Full Text]
    32. 32
  32. Oliva-Trastoy, M., V. Berthonaud, A. Chevalier, C. Ducrot, M. C. Marsolier-Kergoat, C. Mann, and F. Leteurtre. 2007. The Wip1 phosphatase (PPM1D) antagonizes activation of the Chk2 tumour suppressor kinase. Oncogene 26:1449-1458.[CrossRef][Medline]
    33. 33
  33. Pellicioli, A., and M. Foiani. 2005. Signal transduction: how Rad53 kinase is activated. Curr. Biol. 15:R769—R771.
    34. 34
  34. Pellicioli, A., S. E. Lee, C. Lucca, M. Foiani, and J. E. Haber. 2001. Regulation of Saccharomyces Rad53 checkpoint kinase during adaptation from DNA damage-induced G2/M arrest. Mol. Cell 7:293-300.[CrossRef][Medline]
    35. 35
  35. Pike, B. L., S. Yongkiettrakul, M. D. Tsai, and J. Heierhorst. 2003. Diverse but overlapping functions of the two forkhead-associated (FHA) domains in Rad53 checkpoint kinase activation. J. Biol. Chem. 278:30421-30424.[Abstract/Free Full Text]
    36. 36
  36. Pike, B. L., S. Yongkiettrakul, M. D. Tsai, and J. Heierhorst. 2004. Mdt1, a novel Rad53 FHA1 domain-interacting protein, modulates DNA damage tolerance and G2/M cell cycle progression in Saccharomyces cerevisiae. Mol. Cell. Biol. 24:2779-2788.[Abstract/Free Full Text]
    37. 37
  37. Pinna, L. A. 2003. The raison d'etre of constitutively active protein kinases: the lesson of CK2. Acc. Chem. Res. 36:378-384.[CrossRef][Medline]
    38. 38
  38. Pinna, L. A., and A. Donella-Deana. 1994. Phosphorylated synthetic peptides as tools for studying protein phosphatases. Biochim. Biophys. Acta 1222:415-431.[Medline]
    39. 39
  39. Sandell, L. L., and V. A. Zakian. 1993. Loss of a yeast telomere: arrest, recovery, and chromosome loss. Cell 75:729-739.[CrossRef][Medline]
    40. 40
  40. Schwartz, M. F., J. K. Duong, Z. Sun, J. S. Morrow, D. Pradhan, and D. F. Stern. 2002. Rad9 phosphorylation sites couple Rad53 to the Saccharomyces cerevisiae DNA damage checkpoint. Mol. Cell 9:1055-1065.[CrossRef][Medline]
    41. 41
  41. Schwartz, M. F., S. J. Lee, J. K. Duong, S. Eminaga, and D. F. Stern. 2003. FHA domain-mediated DNA checkpoint regulation of Rad53. Cell Cycle 2:384-396.[Medline]
    42. 42
  42. Sikorski, R. S., and P. Hieter. 1989. A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19-27.[Abstract/Free Full Text]
    43. 43
  43. Smolka, M. B., C. P. Albuquerque, S. H. Chen, K. H. Schmidt, X. X. Wei, R. D. Kolodner, and H. Zhou. 2005. Dynamic changes in protein-protein interaction and protein phosphorylation probed with amine-reactive isotope tag. Mol. Cell Proteomics 4:1358-1369.[Abstract/Free Full Text]
    44. 44
  44. Smolka, M. B., S. H. Chen, P. S. Maddox, J. M. Enserink, C. P. Albuquerque, X. X. Wei, A. Desai, R. D. Kolodner, and H. Zhou. 2006. An FHA domain-mediated protein interaction network of Rad53 reveals its role in polarized cell growth. J. Cell Biol. 175:743-753.[Abstract/Free Full Text]
    45. 45
  45. Song, H., N. Hanlon, N. R. Brown, M. E. Noble, L. N. Johnson, and D. Barford. 2001. Phosphoprotein-protein interactions revealed by the crystal structure of kinase-associated phosphatase in complex with phosphoCDK2. Mol. Cell 7:615-626.[CrossRef][Medline]
    46. 46
  46. Sun, Z., J. Hsiao, D. S. Fay, and D. F. Stern. 1998. Rad53 FHA domain associated with phosphorylated Rad9 in the DNA damage checkpoint. Science 281:272-274.[Abstract/Free Full Text]
    47. 47
  47. Sweeney, F. D., F. Yang, A. Chi, J. Shabanowitz, D. F. Hunt, and D. Durocher. 2005. Saccharomyces cerevisiae Rad9 acts as a Mec1 adaptor to allow Rad53 activation. Curr. Biol. 15:1364-1375.[CrossRef][Medline]
    48. 48
  48. Takekawa, M., T. Maeda, and H. Saito. 1998. Protein phosphatase 2Calpha inhibits the human stress-responsive p38 and JNK MAPK pathways. EMBO J. 17:4744-4752.[CrossRef][Medline]
    49. 49
  49. Toczyski, D. P., D. J. Galgoczy, and L. H. Hartwell. 1997. CDC5 and CKII control adaptation to the yeast DNA damage checkpoint. Cell 90:1097-1106.[CrossRef][Medline]
    50. 50
  50. Tonks, N. K., and B. G. Neel. 2001. Combinatorial control of the specificity of protein tyrosine phosphatases. Curr. Opin. Cell Biol. 13:182-195.[CrossRef][Medline]
    51. 51
  51. Vaze, M. B., A. Pellicioli, G. Ira, S. E. Lee, G. Liberi, A. Arbel-Eden, M. Foiani, and J. E. Haber. 2002. Recovery from checkpoint-mediated arrest after repair of a double-strand break requires SRS2 helicase. Mol. Cell 10:373-385.[CrossRef][Medline]
    52. 52
  52. Wach, A., A. Brachat, R. Pohlmann, and P. Philippsen. 1994. New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae. Yeast 10:1793-1808.[CrossRef][Medline]
    53. 53
  53. Weinert, T. A., and L. H. Hartwell. 1988. The RAD9 gene controls the cell cycle response to DNA damage in Saccharomyces cerevisiae. Science 241:317-322.[Abstract/Free Full Text]
    54. 54
  54. Welihinda, A. A., W. Tirasophon, S. R. Green, and R. J. Kaufman. 1998. Protein serine/threonine phosphatase Ptc2p negatively regulates the unfolded-protein response by dephosphorylating Ire1p kinase. Mol. Cell. Biol. 18:1967-1977.[Abstract/Free Full Text]
    55. 55
  55. Young, C., J. Mapes, J. Hanneman, S. Al-Zarban, and I. Ota. 2002. Role of Ptc2 type 2C Ser/Thr phosphatase in yeast high-osmolarity glycerol pathway inactivation. Eukaryot. Cell 1:1032-1040.[Abstract/Free Full Text]
    56. 56
  56. Yuan, C., S. Yongkiettrakul, I. J. Byeon, S. Zhou, and M. D. Tsai. 2001. Solution structures of two FHA1-phosphothreonine peptide complexes provide insight into the structural basis of the ligand specificity of FHA1 from yeast Rad53. J. Mol. Biol. 314:563-575.[CrossRef][Medline]
    57. 57
  57. Zhan, X. L., and K. L. Guan. 1999. A specific protein-protein interaction accounts for the in vivo substrate selectivity of Ptp3 towards the Fus3 MAP kinase. Genes Dev. 13:2811-2827.[Abstract/Free Full Text]
    58. 58
  58. Zhou, B. B., and S. J. Elledge. 2000. The DNA damage response: putting checkpoints in perspective. Nature 408:433-439.[CrossRef][Medline]

Molecular and Cellular Biology, May 2007, p. 3378-3389, Vol. 27, No. 9
0270-7306/07/$08.00+0     doi:10.1128/MCB.00863-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.



This article has been cited by other articles:

  • Kim, J.-A., Haber, J. E. (2009). Chromatin assembly factors Asf1 and CAF-1 have overlapping roles in deactivating the DNA damage checkpoint when DNA repair is complete. Proc. Natl. Acad. Sci. USA 106: 1151-1156 [Abstract] [Full Text]  
  • Mahajan, A., Yuan, C., Lee, H., Chen, E. S.-W., Wu, P.-Y., Tsai, M.-D. (2008). Structure and Function of the Phosphothreonine-Specific FHA Domain. Sci Signal 1: re12-re12 [Abstract] [Full Text]  
  • Luo, H., Li, Y., Mu, J.-J., Zhang, J., Tonaka, T., Hamamori, Y., Jung, S. Y., Wang, Y., Qin, J. (2008). Regulation of Intra-S Phase Checkpoint by Ionizing Radiation (IR)-dependent and IR-independent Phosphorylation of SMC3. J. Biol. Chem. 283: 19176-19183 [Abstract] [Full Text]  
  • Travesa, A., Duch, A., Quintana, D. G. (2008). Distinct Phosphatases Mediate the Deactivation of the DNA Damage Checkpoint Kinase Rad53. J. Biol. Chem. 283: 17123-17130 [Abstract] [Full Text]  
  • Nyswaner, K. M., Checkley, M. A., Yi, M., Stephens, R. M., Garfinkel, D. J. (2008). Chromatin-Associated Genes Protect the Yeast Genome From Ty1 Insertional Mutagenesis. Genetics 178: 197-214 [Abstract] [Full Text]  

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrowReprints and Permissions
Right arrow Copyright Information
Right arrow Books from ASM Press
Right arrow MicrobeWorld
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Guillemain, G.
Right arrow Articles by Marsolier-Kergoat, M.-C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Guillemain, G.
Right arrow Articles by Marsolier-Kergoat, M.-C.

Copyright © 2009 by the American Society for Microbiology.   For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»