Profiling Tel1 signaling reveals a non-canonical motif targeting DNA repair and telomere control machineries

Abstract

The stability of the genome relies on phosphatidyl inositol 3-kinase-related kinases (PIKKs) that sense DNA damage and trigger elaborate downstream signaling responses. In Saccharomyces cerevisiae, the Tel1 kinase (ortholog of human ATM) is activated at DNA double-strand breaks (DSBs) and short telomeres. Despite the well-established roles of Tel1 in the control of telomere maintenance, suppression of chromosomal rearrangements, activation of cell cycle checkpoints, and repair of DSBs, the substrates through which Tel1 controls these processes remain incompletely understood. Here we performed an in-depth phosphoproteomic screen for Tel1-dependent phosphorylation events. To achieve maximal coverage of the phosphoproteome, we developed a scaled-up approach that accommodates large amounts of protein extracts and chromatographic fractions. Compared to previous reports, we expanded the number of detected Tel1-dependent phosphorylation events by over 10-fold. Surprisingly, in addition to the identification of phosphorylation sites featuring the canonical motif for Tel1 phosphorylation (S/T-Q), the results revealed a novel motif (D/E-S/T) highly prevalent and enriched in the set of Tel1-dependent events. This motif is unique to Tel1 signaling and not shared with the Mec1 kinase, providing clues to how Tel1 plays specialized roles in DNA repair and telomere length control. Overall, these findings define a Tel1-signaling network targeting numerous proteins involved in DNA repair, chromatin regulation, and telomere maintenance that represents a framework for dissecting the molecular mechanisms of Tel1 action.

Keywords: DNA damage signaling; DNA repair; Tel1; kinase signaling; mass spectrometry; phosphoproteomics; phosphorylation; telomere maintenance.

Figure 1

Large-scale phosphoproteomics of Tel1-dependent phosphorylation. A, overview of DNA lesions activating PIKKs, with Mec1 and Tel1 canonically becoming activated by different forms of DNA damage. B, a schematic of the large-scale phosphoproteomic pipeline. Large SILAC yeast cultures are required to obtain sufficient starting material for tryptic digestion and phosphopeptide enrichment. Extensive HILIC prefractionation reduces sample complexity and allows for in-depth LC-MS/MS analysis, followed by database searching and Bowtie replicate filtering. C, Bowtie filtering of forward and reverse replicates performed for each experimental condition, in which experimental conditions are inverted between SILAC channels. The replicates are then plotted against one another to determine which phosphopeptides feature appropriately inverting quantitative ratios, discarding any that do not invert. D, plot depicting averaged ratios for 17,401 phosphopeptides that passed Bowtie filtering. Among these are phosphopeptides on Yku80 and components of the MRX complex, most of which are found to be Tel1-dependent. Asterisk indicates an average of the quantitative ratios of both the forward and reverse replicates.

Figure 2

The S/T-Q motif is enriched among Tel1-dependent phosphorylation events on proteins involved in the DNA damage response and telomere maintenance.A, plot depicting the averaged ratio for 17,401 phosphopeptides that passed Bowtie filtering. S/T-Q sites are highlighted in green. Tel1-dependent phosphorylation events are over 3X enriched for the canonical preferred PIKK S/T-Q motif when compared to unchanged phosphorylation events. Asterisk indicates an average of the quantitative ratios of both the forward and reverse replicates. B, comparison of the number of Tel1-dependent phosphorylation events detected in the current study versus our previous investigation of the Tel1 signaling network. C, the top 10 GO terms (minus terms containing “process”) enriched among proteins featuring Tel1-dependent S/T-Q sites. D, telomere-related GO term enrichment among proteins featuring Tel1-dependent S/T-Q sites.

Figure 3

A novel D/E-S/T motif is enriched among Tel1-dependent phosphorylation events.A, heatmap illustrating enrichment of amino acid residues at loci relative to phosphorylation events found to be Tel1-dependent. Acidic residues, particularly aspartic acid, are found to be enriched in the −1 locus. Enrichment for the S/T-Q motif is also illustrated here. B, plot depicting the averaged ratio for 17,401 phosphopeptides that passed Bowtie filtering with D/E-S/T sites highlighted in pink and the sites with the combined motif (D/E-S/T-Q) highlighted in purple. Asterisk indicates an average of the quantitative ratios of both the forward and reverse replicates. C, prevalence of four phospho-motifs among subsets of the Tel1 signaling dataset. The most enriched motif in phosphopeptides that decreased upon deletion of Tel1 was D/E-S/T, with S/T-Q and D/E-S/T-Q motifs also showing high enrichment among this same subset of the data. D, the top 10 GO terms (minus terms containing “process”) enriched among proteins featuring Tel1-dependent D/E-S/T sites. E, telomere-related GO term enrichment among proteins featuring Tel1-dependent D/E-S/T sites.

Figure 4

The D/E-S/T motif is more enriched in Tel1-dependent phosphorylation events compared to Mec1-dependent phosphorylation events.A, scatter plot with phosphoproteomic data of Tel1 dependency plotted against phosphoproteomic data of Mec1 dependency. Increased abundance of Tel1-dependent phosphorylation events upon loss of Mec1 supports the notion that Tel1 engages in some amount of compensatory signaling when Mec1 activity is compromised. Asterisk indicates an average of the quantitative ratios of both the forward and reverse replicates. B, enrichment of the S/T-Q motif among Tel1- and Mec1-dependent phosphorylation events illustrates a higher enrichment for this motif among Mec1-dependent phosphorylation events. C, enrichment of the D/E-S/T motif among Tel1- and Mec1-dependent phosphorylation events strongly suggests that D/E-S/T motif enrichment is exclusive to Tel1-dependent signaling.

Figure 5

Tel1-dependent D/E-S/T phosphorylation events are not dependent on downstream kinases Hrr25, Dun1, or Rad53.A, plot of phosphoproteomic data showing that S/T-Q sites on kinases Hrr25 and Dun1 were found to be Tel1-dependent. Asterisk indicates an average of the quantitative ratios of both the forward and reverse replicates. B, Sequence map of Hrr25 showing where an analog-sensitivity mutation was made (Ile 82) as well as the location of the Tel1-dependent S/T-Q phosphorylation event (Ser 438). C, sequence map of Dun1 showing the location of the Tel1-dependent S/T-Q phosphorylation event (Ser 345). D, scatter plot for phosphoproteomic data of Tel1 dependency plotted against phosphoproteomic data of Hrr25-dependency. No significant co-dependence of signaling events was observed. The D/E-S/T motif was not found to be enriched among Hrr25-dependent phosphorylation events. Hrr25 kinase inhibition was achieved by treating cultures with 1 micromolar 1NM-PP1. E, scatter plot for phosphoproteomic data of Tel1 dependency plotted against phosphoproteomic data of Dun1/Rad53 dependency. No significant co-dependence of signaling events was observed. The D/E-S/T motif was not found to be enriched among Dun1/Rad53-dependent phosphorylation events.

Figure 6

Substrate and kinase mutations validate Tel1 preference for D/E-S/T motifs. A, diagram illustrating the Tel1-dependent phosphorylation site at serine 469 on Rad50. The aspartic acid previous to this serine was mutated to alanine to assess the importance of a negative residue at the −1 position for Tel1-dependent phosphorylation. B, FLAG-tagged Rad50 with and without the motif mutation was immunoprecipitated and analyzed via mass spectrometry. Peak area quantification reveals a drop in phosphorylation of Rad50 serine 469 after motif mutation. C, structural superimposition of yeast Mec1 and Tel1 (PDB IDs 6S8F and 6Z2X respectively) highlighting the substrate-binding pocket and key residues involved in the recognition of peptide substrates via their positions −1 (Tel1 H²⁵²⁶, R²⁵⁴⁴, and N²⁶¹⁶; Mec1 R²¹³⁹, K²¹⁵⁶ and E²²²⁸) and +1 (Tel1 L²⁶⁴² and Y²⁷⁸⁰; Mec1 L²²⁵⁴ and Y²³⁶¹). Inserts on the right show the electrostatic potential of surfaces around the −1 binding pocket for each kinase, blue indicating positive charge and red indicating negative charge. In all figures, the peptide substrate shown in yellow is superimposed from a structure of ATM bound to p53 (PDB ID 8OXO) with the original Leu at the −1 position is remodeled as Asp (asterisk). ATP loops of kinases are omitted to facilitate visualization. Carbon atoms are tan for Mec1 and green for Tel1. D, phosphoproteomic analysis showing that R2544A and N2616E mutations in the Tel1 substrate binding pocket selectively reduced phosphorylation of Tel1-dependent D/E-S/T phosphorylation sites, exerting no such effect on Tel1-dependent S/T-Q sites. Sites with ratios over 7 were capped at 7 for visualization purposes. E, yeast spot assay in which cells with the Tel1 double mutation show markedly increased sensitivity to DSB-inducing agent camptothecin (CPT) at a concentration of 30 micromolar. 4-fold dilution.

Figure 7

An expanded network of Tel1-dependent phosphorylation targeting S/T-Q and D/E-S/T motifs.A, overview of Tel1-dependent phosphorylation events that were able to be localized and their motifs. ∗D/E-S/T and S/T-Q motif prevalence calculated excluding the combined motif, D/E-S/T-Q. B, Tel1-dependent phosphorylation events with either the S/T-Q or D/E-S/T motif implicated in the DNA damage response, DNA-templated transcription, telomere organization, or noncoding RNA processing. Proteins denoted as having “both” motifs either contain multiple sites with both motifs represented or at least one site with the combined motif, D/E-S/T-Q. C, breakdown of D/E-S/T and/or S/T-Q phosphorylation events implicated in telomere organization as denoted by panelB.

Figure S1

Figure S2

Figure S3

Figure S4