Analysis of Mec1- and Tel1-Dependent Phosphorylation Sites.

To determine which phosphorylation sites depend on Mec1 and Tel1, we next examined the relative abundance of each phosphopeptide in WT and mec1Δtel1Δ cells. Most of the phosphopeptides were present at similar abundances in WT and mec1Δtel1Δ cells, as revealed by an abundance ratio within 2-fold (exemplified in Fig. 2A Left). To be considered Mec1/Tel1-dependent, a phosphorylation must be identified in a phosphopeptide that was either absent or considerably reduced in abundance (>4-fold) in mec1Δ tel1Δ cells (exemplified in Fig. 2A Right). As shown in Fig. 2B, only 62 phosphopeptides (≈2% of the total identified phosphopeptides) from 55 different proteins were found to be Mec1/Tel1-dependent. We then examined whether direct targets of Mec1 and Tel1 were found. Because Mec1 and Tel1 phosphorylate S/T-Q sites (30), we calculated the frequency of the amino acid residue at the +1 position of all of the phosphorylation sites identified (see Fig. 2C). Interestingly, although phosphorylation of the S/T-Q motif accounts for ≈3% of all phosphorylation sites identified, it represents 39% of the Mec1/Tel1-dependent phosphorylation. In contrast, proline-directed phosphorylation is very common among all of the identified phosphopeptides, yet only two such phosphorylation sites were found to be Mec1/Tel1-dependent. Therefore, the list of Mec1/Tel1-dependent phosphorylation is likely enriched for direct targets of Mec1 and Tel1.

Open in a separate window

Fig. 2.

Identification and analysis of Mec1/Tel1-dependent phosphorylation. (A) Relative ion abundances of selected phosphopeptides (with the indicated phosphorylation site) in WT (filled circles) and mec1Δ tel1Δ (open circles) cells. (Left) Example of phosphopeptides present at similar abundance in both cells, thereby containing a Mec1/Tel1-independent phosphorylation. (Right) Example of phosphopeptides specifically present in wild-type cells, thereby containing a Mec1/Tel1-dependent phosphorylation. For these Mec1/Tel-dependent phosphorylations, the open circles indicate the expected position of the phosphopeptide ions from mec1Δ tel1Δ cells (labeled with d10-Nisotag). (B) Number and percentage of Mec1/Tel1-dependent phosphorylation identified relative to the total number of phosphorylation sites identified (2,689 phosphorylation sites and 1,109 proteins; see Methods for detailed description of criteria used). (C) Comparison of the frequency of amino acids at the +1 position of phosphorylated serine or threonine for all identified phosphorylation (gray bars) and Mec1/Tel1-dependent phosphorylation (black bars).

Analysis of Rad53-Dependent Phosphorylation Sites.

Because Mec1 and Tel1 control the activities of Rad53, Dun1 and possibly other kinases (12, 13, 31, 32), Mec1- and Tel1- dependent phosphorylation should include both direct targets of Mec1 and Tel1 and those of their downstream kinases. Phosphorylation of the S/T-ψ motif (ψ denotes a hydrophobic residue: F, I, L, or V) was also found to be a relatively common Mec1/Tel1-dependent phosphorylation, accounting for 23% of these phosphopeptides (see Fig. 2C). We recently showed that Rad53 prefers to phosphorylate Dun1 in its S/T-ψ motif (32). This led us to reason that Mec1/Tel1-dependent phosphorylation of the S/T-ψ motif may be directly phosphorylated by Rad53. To test this hypothesis, we performed a similar experiment as shown in Fig. 1A, using rad53Δ cells instead of mec1Δ tel1Δ cells. Of the 2,689 phosphorylation sites consistently found, only 32 were found to be Rad53-dependent and 13 of them are located in a S/T-ψ motif (see Fig. 3 A and B). Importantly, all of them were also identified as Mec1/Tel1-dependent, consistent with Rad53 activity being Mec1/Tel1-dependent.

Open in a separate window

Fig. 3.

Identification and analysis of Rad53-dependent phosphorylation. (A) Number and percentage (related to total number of Mec1/Tel1-dependent phosphorylation identified) of Rad53-dependent phosphorylation identified. All Rad53-dependent phosphorylation was also Mec1/Tel1-dependent (n = 62). (B) Frequency of amino acids at the +1 position of phosphorylated serine or threonine for Rad53-independent and Mec1/Tel1-dependent phosphorylation (black bars) and Rad53-dependent phosphorylation (gray bars). (C) Simplified representation of the Mec1, Tel1, and Rad53 kinases and their targets in the DNA damage response.

Because phosphorylation of the majority of the non S/T-Q sites depends on Rad53 (Fig. 3B), Rad53 appears to be the major kinase that functions downstream of Mec1 and Tel1 in yeast samples analyzed 3 h after MMS exposure. Because Rad53 controls the activity of Dun1 and possibly other kinases, it is not surprising that phosphorylation motifs other than S/T-ψ were also found to depend on Rad53 (Table 1). Four phosphorylation sites in a non S/T-ψ motif were detected in Rad53 (S424 and S560) and Dun1 (S10 and S139). These sites were previously found to be autophosphorylation sites through the analysis of purified kinases (6, 32). Their detection here further validated our approach.

Direct Targets of Mec1 and Tel1.

To identify direct Mec1/Tel1 targets, we only considered those Mec1/Tel1-dependent and Rad53-independent phosphorylation sites (see rationale in Fig. 3C). As shown in Table 2, 24 of the 30 phosphorylation sites are located in an S/T-Q motif. This is remarkable because S/T-Q is a relatively rare phosphorylation motif in cells (see Fig. 2C). Clearly, our phosphoproteomic approach has detected the consensus motif preferentially phosphorylated by Mec1 and Tel1. This analysis strongly suggests that most proteins listed in Table 2 are likely direct substrates of Mec1 and/or Tel1. It is interesting to note that most of them are nuclear proteins involved in DNA replication and repair, DNA damage checkpoint, and RNA metabolism and transcription. Among them, Rpa1 and Mec3 were reported to be directly phosphorylated by Mec1 and/or Tel1 (24, 33) and Rad9 was known to undergo a Mec1/Tel1-dependent phosphorylation (34, 35). The other targets are previously undescribed findings.

We next addressed whether any new Mec1/Tel1 targets could be validated by other methods and chose to examine Cbf1 and Trm8 further. Both Cbf1 and Trm8 underwent Mec1/Tel1 dependent and Rad53 independent phosphorylation as shown in Fig. 4 A and D. Substrates of Mec1 or Tel1 should be phosphorylated in an MMS-inducible manner. To test whether this is the case for Cbf1 and Trm8, we purified these proteins from untreated and MMS treated cells and detected their phosphorylation using an antibody that specifically recognizes phosphorylated S/T-Q motif. As shown in Fig. 4 B and E, this antibody detected phosphorylation in Cbf1 and Trm8 that is strongly induced by MMS treatment. This signal was abolished by prior phosphatase treatment. Further, the detected phosphorylation of both Cbf1 and Trm8 was found to be mostly dependent on Mec1, less on Tel1 (see Fig. 4 C and F, respectively). These results confirmed that Cbf1 and Trm8 undergo an MMS-induced phosphorylation on an S/T-Q motif.

Open in a separate window

Fig. 4.

Cbf1 and Trm8 are phosphorylated in an S/T-Q motif, in an MMS-induced manner. (A and D) Relative ion abundance of the Cbf1 and Trm8 phosphopeptides, with phosphorylation at S45 and S7, respectively, in WT (filled circles) and the indicated kinase-null (open circles) cells. (B and E) Western blot analysis of Cbf1 and Trm8 immunoprecipitated (IP) from control or MMS-treated cells and eluted by using TEV protease. Phosphorylated S/T-Q motifs were detected with an antiphospho-S/T-Q antibody. For loading control, the blot was later detected with an anti-CBP antibody. (C and F) Western blot was performed as described in B and E, and proteins were purified from MMS-treated cells lacking the indicated kinases.

Cell Growth and Protein Extraction.

Two liters of yeast cells (WT or kinase-null) were grown in YPD medium to an OD₆₀₀ of 0.5 and cells were treated with 0.05% MMS for 3 h. Cells were broken in an ice-cooled bead beater with 40 ml of lysis buffer containing 50 mM Tris·HCl (pH 8.0), 150 mM NaCl, 0.2% Nonidet P-40, 0.5 mM DTT, 5 mM NaF, 10 mM β-glycerolphosphate, 1 mM sodium vanadate, 5 mM EDTA, 1 mM PMSF, 0.2 mM benzamidine, 1 μM leupeptin, and 1.5 μM pepstatin. Cell debris was removed by centrifugation at 30,000 × g for 30 min. For each experiment, ≈50 mg of proteins were then denatured by boiling in the presence of 2% SDS and 10 mM DTT for 5 min. Proteins were alkylated with 50 mM iodoacetamide, precipitated with 3 volumes of cold ethanol:acetone (1:1, vol/vol) and then resuspended with buffer containing 2 M urea and 50 mM Tris·HCl, pH 8.0. One milligram of trypsin (Worthington, Lakewood, NJ) was added for overnight digestion, and then the tryptic peptides were desalted by using a 2-g C18 column (Waters, Milford, MA).

Phosphopeptide Purification.

N-isotag labeling and SCX fractionation. Phosphopeptides were purified by using immobilized metal affinity column, as described (6). Eluted phosphopeptides were labeled with either d0-N-isotag (WT cells) or d10-N-isotag (kinase-null cells), as described (44). The labeled phosphopeptides from WT cells were then combined with the labeled phosphopeptides from either mec1Δ tel1Δ or rad53Δ cells and then fractionated using a 2 × 50 mm PolySulfoethyl-based Strong Cation Exchange column (Polymicro Technologies, Phoenix, AZ), as described (6).

Mass Spectrometry and Data Analysis.

Ten SCX fractions were collected and analyzed by a Thermo Finnigan (Waltham, MA) LTQ quadrupole ion trap mass spectrometer. For data analysis, COMET was used for peptide identification (45), and XPRESS and INTERACT softwares were used for quantitation as described (6, 44). The complete yeast database was used to analyze MS/MS spectra with no restriction on the enzyme used. Variable modifications of 80 Da (phosphorylation) on serine and threonine residues and 10 Da (light and heavy N-isotag mass difference) on lysine residues and N terminus were included in the database search. A static modification of 113 Da was included on lysine and N termini of all peptides. Because of oxidation of methionine and tryptophan residues during N-isotag labeling, a static modification of 16 Da was added to these resides. Each large-scale phosphoproteomic experiment (WT vs. mec1Δ tel1Δ cells and WT vs. rad53Δ cells) was repeated five times.

To generate a list of the identified phosphopeptides (SI Table 4), we considered only the top-matched and doubly tryptic phosphopeptides that were identified at least once in each of the experiments (WT vs. mec1Δ tel1Δ cells and WT vs. rad53Δ cells). Using these criteria, we obtained a nonredundant list of 2,457 phosphopeptides, containing 2,689 phosphorylation sites (see SI Table 4). This list of phosphopeptides was then subjected to further more stringent analysis, as described below.

Identification of Mec1/Tel1- or Rad53-Dependent Phosphorylation Sites.

To identify Mec1/Tel1- or Rad53-dependent phosphorylation, we searched for phosphopeptides that fitted three additional criteria (see SI Tables 5 and 6 for detailed information of each peptide): (i) abundance is reduced at least 4-fold in the kinase-null cells, compared with WT cells; (ii) signal-to-noise ratio of the identified parent ion is at least 5-to-1; (iii) the phosphopeptide was identified more times as labeled with d0-N-isotag (WT) than with d10-N-isotag (kinase-null mutant). MS/MS spectra of all of the phosphopeptides that fitted the above three criteria were further manually inspected to confirm the identification of the phosphopeptide and the assignment of the phosphorylation sites.

Western Blot Analysis of Cbf1 and Trm8 Phosphorylation.

One hundred milliliters of Cbf1-TAP and Trm8-TAP cells was grown in YPD medium to an OD₆₀₀ of 0.5, and cells were either treated with 0.05% MMS for 3 h or mock-treated for 2 h. Cells were then harvested and broken as described above but using 1 ml of lysis buffer instead. The protein extracts were incubated with 20 μl of IgG resin overnight. The IgG resins were then washed extensively by TBS-T and subjected to TEV cleavage (1 unit, Invitrogen, Carlsbad, CA) to elute proteins. Part of the elution derived from MMS-treated cells was further treated by CIP (5 units, New England Biolabs, Ipswich, MA) for 1 h. For Western blot analysis, proteins were detected by using antiphospho-S/T-Q antibody (Cell Signaling Technology, Boston, MA) or anti-CBP antibody (Upstate, Charlottesville, VA).

We thank Dr. Raymond Deshaies and Dr. Richard Kolodner for critical reading and suggestions on this work. This work was supported by the Ludwig Institute for Cancer Research and National Human Genome Research Institute (Grant K22 HG002604, to H.Z.).