Two collections of chromosomally tagged yeast Saccharomyces cerevisiae strains were previously designed to detect protein-protein interactions via the Cross-and-Capture system. Here, we used these strain collections in a different application to screen for proteins that are phosphorylated in response to DNA damage by electrophoretic shift analysis. Modification of a number of proteins that are known targets for checkpoint kinases was confirmed this way. Furthermore, we identified the mismatch repair protein Pms1 as a novel target for phosphorylation in the response to DNA damage and replication fork arrest. Genetic analysis revealed that this phosphorylation is dependent on checkpoint activation by ATM and ATR (yeast Mec1p and Tel1p) kinase. Hence, we demonstrated that the Cross-and-Capture system could be efficiently used to detect posttranslational modifications that modulate and control protein function in eukaryotic cells.
Posttranslational modifications (PTMs) are highly regulated and control the activation of signaling pathways, protein-protein and protein-DNA interactions, and protein trafficking and degradation. Consequently, the comprehensive analysis of PTMs remains one of the major challenges in proteomics today. A well-characterized PTM mechanism is protein phosphorylation by kinases in response to DNA damage. The phosphatidyl-inositol kinase (PIK)–like kinases ATM (ataxia telangiectasia mutated) and ATR (ATM and Rad3 related) are a pair of related protein kinases at the core of the DNA damage signaling cascade (1). ATM and ATR are highly conserved through evolution with the homologs Mec1p/Tel1p in yeast Saccharomyces cerevisiae. ATM/ATR-like kinases preferentially phosphorylate serine or threonine residues that precede glutamines (so called S/T-Q motifs) (2). Whereas S/T-Q motifs occur in some ATM/ATR substrates in high densities or clusters (3), single or non-clustered S/T-Q sites occur in other important substrates. Known substrates of ATM/ATR mostly include proteins that are involved in the DNA damage response and DNA repair mechanisms. Recently, a large-scale proteomic analysis of human proteins phosphorylated in response to DNA damage on ATM/ATR consensus sites identified over 700 proteins that also include components of modules and pathways not previously linked to the DNA damage response (4).
Previously, our lab constructed two arrays of differentially tagged yeast strains that were used for rapid assessment of protein-protein interactions in the so-called Cross-and-Capture procedure (5). In total, ∼500 proteins with roles in DNA repair, recombination, replication, and DNA damage response—and a number of nuclear proteins with unknown functions—were C-terminally tagged in their chromosomal location with a six histidine (6xHIS, bait) or a triple VSV (3xVSV, prey) tag. In addition, both bait and prey tags contained a V5 epitope (5). Here, we surveyed the Cross-and-Capture arrays for proteins that comprise potential targets for PIK-like kinases, based on the total number and the clustering of S/T-Q phosphorylation motifs. In total, 38 proteins were selected (Table 1). These include a number of established targets for PIK-mediated phosphorylation that serve as positive controls. We used the 3xVSV-tagged prey array for rapid assessment of the candidate proteins by monitoring electrophoretic mobility (Figure 1A). Protein phosphorylation in cells treated with methyl methanesulfonate (MMS) is expected to result in altered (slower) electrophoretic mobility of the protein compared with the untreated sample.
3xVSV-tagged prey strains were grown in a small volume (5–50 mL) of yeast extract/peptone/dextrose (YPD) to ∼1.0 × 107 cells/mL and then treated either with or without 0.02% MMS for 2 h before harvest. Protein extraction, Western blotting, and probing with anti-VSV antibody were done as described in the Cross-and-Capture protocol (5), except that extracts were prepared in 100–150 µ1 of lysis buffer. The proteins were separated on 7.5% SDS-PAGE (37.5:1 acrylamide/bis-acrylamide) at 60–80 V for optimal resolution. Band shifts were observed for a number of proteins that were already known as phosphorylation substrates in the presence of DNA damage (Hpr5p, Mrc1p, and Rtt107p; see Figure 1B). Other known phosphorylation targets included Nej1p, Rad53p, Rad55p, Rad9p, and Slx4p (see Table 1 and references therein). Surprisingly, we observed a somewhat slower migration upon MMS treatment for the mismatch repair (MMR) protein Pms1p (Figure 1B), the yeast MutL homolog that forms a heterodimer with Mlh1p (6,7). Phosphorylation of Pms1p has not been reported previously.
In a second step, putative phosphorylation events were confirmed by treatment of the proteins with λ-phosphatase, which is expected to reverse the phosphorylation mobility shift in the gel run (Figure 2A). For this assay, we used the 6xHIS-tagged strain collection of the Cross-and-Capture system, which allows isolation of the tagged proteins by binding to nickel beads. Cultures of Pms1–6xHIS bait strains were grown in 50–250 mL YPD and treated with 0.02% MMS, 5 µg/mL camptothecin (CPT), 10 µg/mL phleomycin (PHL), 0.2 M hydroxyurea (HU), or not treated for 2 h before harvest. Proteins were extracted and bound to nickel beads (ProBond nickel-chelating resin beads; Invitrogen, Carlsbad, CA, USA), according to the Cross-and-Capture protocol (5). The nickel beads were washed twice with lysis buffer with 5 mM β-glycerophosphate and twice without β-glycerophosphate (3 mL volume in each wash step). Individual samples were then split in three. Phosphate groups were removed with 200 U of λ-phosphatase in 100 µl 1×λ-phosphatase buffer for 30 min at 30°C (400,000 U/mL; New England Biolabs, Ipswitch, MA, USA). λ-phosphatase in one control sample was inhibited by the addition of 20 mM sodium orthova-nadate, whereas no λ-phosphatase was added to the mock-treated control. After digestion, λ-phosphatase buffer was spun out and proteins were eluted with 50 µl of 200 mM imidazole. Proteins were separated on 6.5% or 7.5% SDS-PAGE. Immunoblotting with anti-V5 antibody was done as described in the Cross-and-Capture protocol (5). The Pms1–6xHIS protein from the cultures grown in the presence of DNA damaging agents (MMS, CPT, PHL) and replicative stress (HU) migrated faster in the gel upon treatment with λ-phosphatase, indicating that Pms1p is phosphorylated in the presence of DNA damage (Figure 2B). Inhibition of λ-phosphatase by sodium orthovanadate shows that the mobility shift in Pms1p-6xHIS is indeed a consequence of protein phosphorylation. The λ-phosphatase assay also suggests a low level of Pms1p phosphorylation in untreated cells.
To determine the genetic requirements of DNA damage–induced phosphorylation of Pms1p, the Pms1–6xHIS tag was introduced in strain backgrounds with deletions in the checkpoint-kinase genes MEC1, TEL1, and RAD53 (see Supplemental Table 1, available online at www.BioTechniques.com). DNA damage–dependent phosphorylation of Pms1p is dependent on the redundant functions of Tellp and Mec1p, whereas single mutations have no significant effect (Figure 2C). Interestingly, previous studies linked MMR with the activation of the S-phase checkpoint response in human cells (8,9), and the ATM kinase (human homolog of Tellp) was found to associate with the human MLH1 mismatch repair protein (9). A large-scale proteomic analysis identified three human mismatch repair proteins (MSH2, MSH3, and MSH6), but not the human homolog of yeast Pms1p, PMS2, as bona fide targets of ATM/ATR-mediated phosphorylation (4). Hence, whereas a connection between MMR- and ATM/ ATR-dependent signaling is established, further biochemical studies are warranted to show that yeast Pms1p is indeed a conserved direct target of Tellp and Mec1p.
In this article, we demonstrated that the two differentially tagged arrays that were developed for the Cross-and-Capture procedure could be successfully used to screen for and identify PTMs, such as protein phosphorylation. Since the genes are tagged endogenously, possible overexpression artifacts are avoided. Identification of protein phosphorylation is straightforward and the use of the 6xHIS tag not only allows an efficient λ-phosphatase assay but also confirms the band shifts observed with the 3xVSV tag. The screening and detection principle presented here can be readily applied for other PTMs, including ubiquitination and sumoylation. The 6xHIS tag is well suited for the denaturing extraction conditions that are commonly used to avoid degradation of these instable PTMs.
The Stagljar lab is supported by grants from the Canadian Foundation for Innovation (CFI), the Canadian Institute for Health Research (CIHR), the National Cancer Institute of Canada (NCIC), the Gebert Rüf Foundation, Genentech, and Novartis.
The authors declare no competing interests.