Regulation of Mec1 (ATR) signaling in budding yeast

Cells are continuously challenged by various sources of DNA damage that can contribute to cancer formation if not appropriately repaired. To cope with this threat, cells have conserved mechanisms called the DNA damage checkpoints that sense damaged DNA, stop the cell cycle, and upregulate DNA repair. Central players in these checkpoints are the PI3K-like kinases ATM and ATR (S.c. Tel1 and Mec1). Mec1 senses single stranded DNA (ssDNA) that is exposed at stalled replication forks and activates the S phase checkpoint. However, ssDNA, which is generated at the lagging strand during normal replication, does not cause detectable checkpoint activation. It is unknown how Mec1 is regulated in S phase. To study this, we took advantage of a mutant allele of MEC1, mec1-100, which is proficient for the G2 DNA damage checkpoint, but is compromised in G1-S and intra-S-phase checkpoints.
In the first part of this thesis we aimed at identifying regulatory factors. We screened for spontaneous survivors on a lethal dose of the replication fork-stalling agent hydroxyurea (HU) for mec1-100 cells. We mapped additional mutations in mec1-100 or mutations in either PPH3 or PSY2, which form a highly conserved phosphatase (PP4) complex. In a second, more unbiased, high-throughput screen we combined mec1-100 with a collection of 1525 gene deletions involved in chromatin processes and scored double mutants for HU sensitivity. pph3Δ and psy2Δ were among the top mec1-100 suppressor hits, confirming a strong genetic interaction. Suppression by pph3Δ was correlated with the phosphorylation of the downstream kinase Rad53. However, it did not depend exclusively on Rad53, because residual suppression of mec1-100 by pph3Δ could also be observed in rad53Δ cells. We tested whether Psy2-Pph3 might regulate Mec1 directly, and found a physical interaction between Psy2 and Ddc2-Mec1. Moreover, we found that a phosphorylation site within Mec1 (S1991) is downregulated in mec1-100 cells and restored when Pph3 is also lost. However, we were unable to demonstrate that Pph3 dephosphorylates Mec1 directly in vitro. Phosphorylation required both Mec1 kinase activity and Rad53. Thus, we speculate that Mec1 phosphorylation is achieved through Rad53, which is in turn regulated by Pph3, indicating the existence of a feedback loop from Rad53 back to Mec1. Mutation of the phosphorylation site renders cells sensitive to the radiomimetic drug Zeocin, indicating an important role in the survival of DNA damage. Finally, we applied quantitative phosphoproteomics to identify Mec1 and Pph3 targets. We found that the levels of the majority of the phosphopeptides that are affected by a tel1Δ mec1-100 mutation but not by rad53Δ can be rescued due to additional deletion of PPH3. The data presented here support a model in which Pph3 is a major regulator of Mec1 signaling.
In a second part mec1-100 was further characterized in order to understand the mechanism by which its two point mutations outside of the catalytic domain (F1179S, N1700S) cause defects in the replication checkpoint. We find that the mutations leave kinase activity in vitro, oligomerization and Ddc2-Mec1 interaction intact. Genetic analysis shows that mec1-100 is additive, rather than epistatic with mutation or deletion of any of the canonical checkpoint activating proteins Ddc1, Dna2, Dpb11, Rad24, Mrc1, Rad9, Tel1 or Chk1. Thus, we conclude that mec1-100 does not impair function of any of these proteins. We hypothesized that the mutated region might constitute a regulatory domain that is bound by a yet unknown factor. IP experiments followed by mass spectrometry analysis did not show reproducibly decreased interaction of any protein. Additional detailed biochemical analysis is needed to fully understand the mechanism of the two mec1-100 mutations.
We further characterize intragenic mec1-100 suppressor mutations by mapping them to a homology model. While some mutations reside within the kinase domain, and could influence catalytic activity, others might as well be involved protein-protein interactions. We asked whether suppression would involve Rad24 dependent Mec1 activation. Interestingly, we find that suppression by mutations in residues that might make protein-protein contacts completely requires Rad24. Other suppressor mutations relied less on Rad24. Thus, we conclude that intragenic suppression of mec1-100 HU sensitivity employs at least two different mechanisms: one that is Rad24-dependent and a second that is Rad24–independent. These unpublished results will help in understanding Mec1 function and regulation once structural data is available.
The third experimental part resolves the role of the RecQ helicase Sgs1 in replication checkpoint signaling. It was shown before that Sgs1 and Mec1 synergistically contribute to replication fork stabilization under replication stress. Both interact with the ssDNA binding protein RPA. Here, we created a mutant, sgs1-r1, which lacks the RPA interaction domain. While sgs1-r1 is proficient to stabilize stalled forks under replication stress, it is synthetic lethal with mus81Δ, slx4Δ, slx5Δ and slx8Δ. These could provide alternative means to recover stalled forks by resolving crossover structures, DNA repair or break induced replication. . Sgs1 was previously shown to promote Rad53 activation in a manner independent of its helicase activity. We show here that Sgs1 checkpoint function requires the R1 domain. Mec1 phosphorylates Sgs1 in this domain and Sgs1 phosphorylation allows its binding to Rad53 in vitro and in vivo. We thus propose that Sgs1 serves as a mediator in checkpoint signaling by recruiting Rad53 to stalled replication forks for activation.
This work provides new insights into Mec1 signaling by elucidating the checkpoint function of Sgs1 and defining Psy2-Pph3 as a major regulator of this pathway.