Dbf4‐dependent kinase and the Rtt107 scaffold promote Mus81‐Mms4 resolvase activation during mitosis
2017; Springer Nature; Volume: 36; Issue: 5 Linguagem: Inglês
10.15252/embj.201694831
ISSN1460-2075
AutoresLissa Nicola Princz, Philipp S. Wild, Julia Bittmann, F. Javier Aguado, Miguel G. Blanco, Joao Matos, Boris Pfander,
Tópico(s)Microtubule and mitosis dynamics
ResumoArticle17 January 2017Open Access Transparent process Dbf4-dependent kinase and the Rtt107 scaffold promote Mus81-Mms4 resolvase activation during mitosis Lissa N Princz Lissa N Princz Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author Philipp Wild Philipp Wild Institute of Biochemistry, Eidgenössische Technische Hochschule, Zürich, Switzerland Search for more papers by this author Julia Bittmann Julia Bittmann Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author F Javier Aguado F Javier Aguado Department of Biochemistry and Molecular Biology, Center for Research in Molecular Medicine and Chronic Diseases, Universidade de Santiago de Compostela, Santiago de Compostela, Spain Search for more papers by this author Miguel G Blanco Miguel G Blanco orcid.org/0000-0002-2883-7326 Department of Biochemistry and Molecular Biology, Center for Research in Molecular Medicine and Chronic Diseases, Universidade de Santiago de Compostela, Santiago de Compostela, Spain Search for more papers by this author Joao Matos Joao Matos Institute of Biochemistry, Eidgenössische Technische Hochschule, Zürich, Switzerland Search for more papers by this author Boris Pfander Corresponding Author Boris Pfander [email protected] orcid.org/0000-0003-2180-5054 Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author Lissa N Princz Lissa N Princz Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author Philipp Wild Philipp Wild Institute of Biochemistry, Eidgenössische Technische Hochschule, Zürich, Switzerland Search for more papers by this author Julia Bittmann Julia Bittmann Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author F Javier Aguado F Javier Aguado Department of Biochemistry and Molecular Biology, Center for Research in Molecular Medicine and Chronic Diseases, Universidade de Santiago de Compostela, Santiago de Compostela, Spain Search for more papers by this author Miguel G Blanco Miguel G Blanco orcid.org/0000-0002-2883-7326 Department of Biochemistry and Molecular Biology, Center for Research in Molecular Medicine and Chronic Diseases, Universidade de Santiago de Compostela, Santiago de Compostela, Spain Search for more papers by this author Joao Matos Joao Matos Institute of Biochemistry, Eidgenössische Technische Hochschule, Zürich, Switzerland Search for more papers by this author Boris Pfander Corresponding Author Boris Pfander [email protected] orcid.org/0000-0003-2180-5054 Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany Search for more papers by this author Author Information Lissa N Princz1, Philipp Wild2, Julia Bittmann1, F Javier Aguado3, Miguel G Blanco3, Joao Matos2 and Boris Pfander *,1 1Max Planck Institute of Biochemistry, DNA Replication and Genome Integrity, Martinsried, Germany 2Institute of Biochemistry, Eidgenössische Technische Hochschule, Zürich, Switzerland 3Department of Biochemistry and Molecular Biology, Center for Research in Molecular Medicine and Chronic Diseases, Universidade de Santiago de Compostela, Santiago de Compostela, Spain *Corresponding author. Tel: +49 89 85783050; Fax: +49 89 85783022; E-mail: [email protected] The EMBO Journal (2017)36:664-678https://doi.org/10.15252/embj.201694831 PDFDownload PDF of article text and main figures. Peer ReviewDownload a summary of the editorial decision process including editorial decision letters, reviewer comments and author responses to feedback. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info Abstract DNA repair by homologous recombination is under stringent cell cycle control. This includes the last step of the reaction, disentanglement of DNA joint molecules (JMs). Previous work has established that JM resolving nucleases are activated specifically at the onset of mitosis. In case of budding yeast Mus81-Mms4, this cell cycle stage-specific activation is known to depend on phosphorylation by CDK and Cdc5 kinases. Here, we show that a third cell cycle kinase, Cdc7-Dbf4 (DDK), targets Mus81-Mms4 in conjunction with Cdc5—both kinases bind to as well as phosphorylate Mus81-Mms4 in an interdependent manner. Moreover, DDK-mediated phosphorylation of Mms4 is strictly required for Mus81 activation in mitosis, establishing DDK as a novel regulator of homologous recombination. The scaffold protein Rtt107, which binds the Mus81-Mms4 complex, interacts with Cdc7 and thereby targets DDK and Cdc5 to the complex enabling full Mus81 activation. Therefore, Mus81 activation in mitosis involves at least three cell cycle kinases, CDK, Cdc5 and DDK. Furthermore, tethering of the kinases in a stable complex with Mus81 is critical for efficient JM resolution. Synopsis DNA joint molecule resolution by the Mus81-Mms4 nucleases is tightly controlled during mitosis. Dbf4-dependent kinase (DDK), best known for its replication roles, now joins previously implicated cell cycle kinases CDK and Cdc5/Plk in this regulation. The budding yeast Cdc7-Dbf4 kinase regulates the Mus81-Mms4 resolvase in mitosis. DDK and Cdc5 form a complex that targets Mms4 and leads to mitotic Mus81 activation. Scaffold protein Rtt107 promotes DDK and Cdc5 activity towards Mus81-Mms4 and is required for full Mus81 activation. DDK, Cdc5 and Rtt107 target Mms4 interdependently, suggesting that they form a signal amplification loop. Introduction Many DNA transactions are under cell cycle control to adjust them to cell cycle phase-specific features of chromosomes (Branzei & Foiani, 2008). Homologous recombination (HR) is cell cycle-regulated at several steps including the first, DNA end resection, and the last, JM removal (Heyer et al, 2010; Ferretti et al, 2013; Mathiasen & Lisby, 2014; Matos & West, 2014). Given that JMs provide stable linkages between sister chromatids, they will interfere with chromosome segregation and therefore need to be disentangled before sister chromatid separation during mitosis. Accordingly, JM resolvases, such as budding yeast Mus81-Mms4 (Interthal & Heyer, 2000; Schwartz et al, 2012) or Yen1 (Ip et al, 2008), become activated during mitosis (Matos et al, 2011, 2013; Gallo-Fernández et al, 2012; Szakal & Branzei, 2013; Blanco et al, 2014; Eissler et al, 2014). In contrast, the alternative JM removal pathway, JM dissolution by the Sgs1-Top3-Rmi1 complex, is thought to be constantly active throughout the cell cycle (Mankouri et al, 2013; Bizard & Hickson, 2014). The activation of JM resolvases in mitosis therefore leads to a shift in the balance between JM removal pathways, with dissolution being preferred outside of mitosis, but JM resolution becoming increasingly important in mitosis (Matos et al, 2011, 2013; Gallo-Fernández et al, 2012; Dehé et al, 2013; Saugar et al, 2013; Szakal & Branzei, 2013; Wyatt et al, 2013). It has been hypothesized that JM resolvases are downregulated at cell cycle stages other than mitosis in order to counteract crossover-induced loss of heterozygosity or to prevent over-active resolvases from interfering with S phase by, for example, cleaving stalled replication forks (Gallo-Fernández et al, 2012; Szakal & Branzei, 2013; Blanco et al, 2014). Budding yeast Mus81-Mms4 has previously been shown to be targeted by two cell cycle kinases, cyclin-dependent kinase Cdc28 (CDK) and the yeast polo-kinase Cdc5 (Matos et al, 2011, 2013; Gallo-Fernández et al, 2012; Szakal & Branzei, 2013). The corresponding Mms4 phosphorylation events were shown to correlate with and to be required for activation of Mus81-Mms4 in mitosis. In 2014, we showed that in mitosis Mus81-Mms4 also forms a complex with Slx4-Slx1 and the scaffold proteins Dpb11 and Rtt107 (Gritenaite et al, 2014). Interestingly, mass spectrometric analysis of this complex (Gritenaite et al, 2014) revealed that Cdc5 and a third cell cycle kinase Dbf4-Cdc7 (Dbf4-dependent kinase, DDK) are also a stable part of this protein assembly (see Appendix Fig S1A). Here, we investigate the role of DDK in Mus81-Mms4 regulation and find that DDK can phosphorylate Mms4 and that DDK and Cdc5 target Mus81-Mms4 in an interdependent manner. Moreover, we show that Rtt107 promotes the association of both kinases with the Mus81-Mms4 complex. The DDK-dependent regulation of Mus81-Mms4 is critical for Mus81 activity thus revealing DDK as a novel regulator of homologous recombination. Results Mus81-Mms4 is a DDK phosphorylation target The cell cycle regulation of JM resolution by Mus81-Mms4 is intricate and involves phosphorylation by the cell cycle kinases CDK and Cdc5 (Matos et al, 2011, 2013; Gallo-Fernández et al, 2012; Szakal & Branzei, 2013) as well as complex formation with the scaffold proteins Dpb11, Slx4 and Rtt107 (Gritenaite et al, 2014). To study this protein complex, we performed an analysis of Mms43FLAG interactors in mitosis by SILAC-based quantitative mass spectrometry (Gritenaite et al, 2014) and found in addition to Dpb11, Slx4, Rtt107 and Cdc5, also Cdc7 and Dbf4 as specific interactors of Mms4 (Appendix Fig S1A). We verified that Cdc7 binds to Mus81-Mms4 in an Mms43FLAG pull down from mitotic cells analysed by Western blots (Fig 1A). The fact that Mus81-Mms4 binds to DDK suggested that it might be involved in the phosphorylation cascade that occurs on Mms4 and controls Mus81 activity in mitosis. Accordingly, we found that purified DDK was able to phosphorylate both subunits of purified Mus81-Mms4 in vitro (Fig 1B, lane 3). When we furthermore compared the DDK-dependent phosphorylation signal to Mms4 phosphorylation by CDK and Cdc5 (Fig 1B, lanes 2–4), we observed different degrees of phosphorylation shifts indicating that the three kinases phosphorylate Mms4 at distinct sites and/or to different degrees. DDK target sites on other proteins have been studied in detail, and in several cases, DDK was found to target (S/T)(S/T) motifs, where phosphorylation was stimulated by a priming phosphorylation usually on the second (S/T) (Masai et al, 2006; Montagnoli et al, 2006; Randell et al, 2010; Lyons et al, 2013). Intriguingly, Mms4 contains 15 of these motifs and we therefore tested whether these could be targeted by DDK and would depend on priming phosphorylation. We therefore turned to a peptide-based assay where Mms4 phosphorylation states are precisely defined. To this end, we synthesized peptides corresponding to two (S/T)(S/T) motifs of Mms4. We chose two representative motifs: S222, as it harbours a minimal CDK consensus motif (S/T)P, and S134, as it contains a non-(S/T)P consensus for CDK [(S/T)X(K/R)(K/R) (Suzuki et al, 2015)]. For each of these motifs, we generated peptides in three different phosphorylation states: non-phosphorylated, phosphorylated at the second serine and doubly phosphorylated (Fig 1C and Appendix Fig S1B). When using such peptides as substrates in in vitro kinase reactions, we saw that CDK targeted specifically only the second serine in each peptide, although much stronger for S222 than for S134, consistent with these residues matching CDK consensus motifs (Fig 1C). In contrast, DDK showed only little activity towards the non-phosphorylated peptides, but was strongly stimulated when the second residue in the (S/T)(S/T) motif was in a phosphorylated state (Fig 1C). DDK may thus be stimulated by priming phosphorylation in order to efficiently phosphorylate Mms4 on (S/T)(S/T) sites. However, using the full-length protein as a phosphorylation substrate, we did not obtain evidence for a stimulatory effect on DDK by prior CDK phosphorylation (Fig 1B and Appendix Fig S1C), perhaps because over the whole 15 (S/T)(S/T) motifs CDK phosphorylation plays a minor role. We also did not reveal any priming activity of either CDK or DDK for Mms4 phosphorylation by Cdc5 (Fig 1B and Appendix Fig S1D). Overall, the data in Fig 1 thus identify Mus81-Mms4 as an interaction partner and potential substrate of DDK. Figure 1. Dbf4-dependent kinase (DDK) binds to the Mus81-Mms4 complex in mitosis and can phosphorylate Mms4 at (S/T)(S/T) motifs Cdc7 and Cdc5 are specifically enriched in Mms43FLAG co-IPs from cells arrested in mitosis (with nocodazole). Under the same conditions, Mus81-Mms4 associates with scaffold proteins such as Dpb11 and Slx4 (Appendix Fig S1A and Gritenaite et al, 2014). DDK can phosphorylate Mus81-Mms4 in vitro. Purified, immobilized Mus81-Mms4 is incubated in an in vitro kinase assay with purified CDK2/cycAN170 (a model CDK), DDK or Cdc5 (lanes 1–4). Additionally, Mus81-Mms4 is incubated with respective kinases after a non-radioactive priming step with CDK (lanes 5–8). DDK phosphorylates Mms4 peptides at (S/T)(S/T) motifs and is enhanced by priming phosphorylation. Mms4 peptides including (S/T)(S/T) motifs (221/222; 133/134) were synthesized in different phosphorylation states (depicted in left panel) and incubated in an in vitro kinase assay with either CDK or DDK. CDK targets unphosphorylated Mms4 peptides 1 and (to a weaker extent) 4 consistent with its substrate specificity (Mok et al, 2010), while DDK primarily targets Mms4 peptides 2 and 5, which harbour a priming phosphorylation at the C-terminal (S/T) site (see Appendix Fig S1B for in-gel running behaviour of peptides). Download figure Download PowerPoint Mus81-Mms4 is phosphorylated by a mitotic Cdc5-DDK complex DDK is present and active throughout S phase and mitosis until anaphase when the Dbf4 subunit is degraded by APC/CCdc20 (Cheng et al, 1999; Weinreich & Stillman, 1999; Ferreira et al, 2000). We therefore tested at which cell cycle stage DDK would associate with Mus81-Mms4 using cells synchronously progressing through the cell cycle. Figure 2A shows that DDK did not associate with Mus81-Mms4 in S phase, but only once cells had reached mitosis. Strikingly, DDK binding therefore coincided with binding of Cdc5, Slx4 and Dpb11 and most notably the appearance of the hyperphosphorylated form of Mms43FLAG (Fig 2A). Figure 2. DDK and Cdc5 target Mus81-Mms4 in an interdependent manner A. DDK stably associates with Mus81-Mms4 in mitosis, but not in S phase or G1. Mms43FLAG pull down experiment (left panel, as in Fig 1A) from cells arrested in G1 (with alpha-factor) or in cells progressing synchronously through S phase until mitosis (arrest with nocodazole) reveals that DDK binds specifically in mitosis concomitant with the raise in Cdc5 levels and Cdc5 binding to Mus81-Mms4. A nocodazole-arrested untagged strain was used as a control. Right panel shows measurements of DNA content by FACS from the respective samples. B. CDK activity is required for DDK and Cdc5 association with Mus81-Mms4. Mms43FLAG pull down as in (A), but in mitotic WT or cdc28-as1 mutant cells treated with 5 μM 1NM-PP1 for 1 h. Additional Western blots of this experiment are shown in Appendix Fig S5B, including as a control the identical anti-FLAG Western blot. C. Cdc5 activity is required for DDK association with Mus81-Mms4. Mms43FLAG pull down as in (A), but with mitotically arrested WT or cdc5-as1 mutant cells treated with 10 μM CMK for 1 h. D, E. DDK is required for Cdc5 binding to Mus81-Mms4 in mitosis and the mitotic Mms4 phospho-shift. (D) Mms43FLAG pull down using mitotically arrested cells as in (A), but using a bob1-1 background (all samples), where the DDK subunit Cdc7 could be deleted. (E) SILAC-based quantification of Mms43FLAG pull downs in mitotically arrested bob1-1 vs. bob1-1 cdc7Δ cells. Plotted are the H/L ratios of two independent experiments including label switch. F. The Cdc5 binding region on Dbf4 is required for interaction of DDK and Cdc5 with Mus81-Mms4 and for efficient Mms4 phosphorylation. Mms43FLAG pull down as in (A), but using mitotically arrested cells expressing N-terminal truncation mutants of Dbf4 lacking aa2–66 (including a D-box motif) or 2–109 [additionally including the Cdc5 binding site (Miller et al, 2009)]. Download figure Download PowerPoint Given this late timing of the association, we tested in co-immunoprecipitation (co-IP) experiments whether DDK binding to Mus81-Mms4 would depend on CDK or Cdc5 activity. Using analog-sensitive mutant yeast strains for CDK [cdc28-as1 (Bishop et al, 2000)] and for Cdc5 [cdc5-as1 (Snead et al, 2007)], we observed that inhibition of these kinases in mitotically arrested cells strongly reduced the hyperphosphorylation shift of Mms4 (see also Matos et al, 2013) and compromised the association with DDK (Fig 2B and C, and Appendix Fig S2A–C). Notably, both conditions also interfered with Cdc5 binding (Fig 2B and C, and Appendix Fig S2A), suggesting that the association of DDK may follow a similar regulation as Cdc5. Next, we tested whether conversely DDK is involved in Mms4 phosphorylation. To bypass the essential function of DDK in DNA replication, we used the mcm5bob1-1 allele (Hardy et al, 1997), which allowed us to test a cdc7Δ mutant. Using Western blot and SILAC-based mass spectrometry as a read-out of Mms43FLAG co-IPs from cells arrested in mitosis, we found that Cdc5 association with Mus81-Mms4 was strongly reduced in the cdc7Δ mutant strain (Fig 2D and E). Moreover, we observed that Mms43FLAG phosphorylation as indicated by mobility shift was decreased in the absence of DDK, although not to the same extent as upon CDK or Cdc5 inhibition (Fig 2D and Appendix Fig S2C). Additionally, as an alternative way to deregulate DDK, we used the cdc7-1 temperature-sensitive mutant. Even with WT cells, we observed that elevated temperature (38°C) leads to a slight reduction in Cdc5 binding to Mus81-Mms4. However, in cdc7-1 mutant cells, incubation at 38°C leads to the complete disappearance of Cdc5 binding to Mus81-Mms4 (Appendix Fig S2D). Therefore, we conclude from these data that DDK and Cdc5 bind to Mus81-Mms4 in an interdependent fashion. Interestingly, Cdc5 was previously shown to interact with DDK via a non-consensus polo-box binding site within Dbf4 (Miller et al, 2009; Chen & Weinreich, 2010). The proposed model based on genetic experiments suggested that DDK binding antagonizes mitotic functions of Cdc5. However, the catalytic activity of Cdc5 was not inhibited in this complex (Miller et al, 2009) and we reason that DDK may simply target Cdc5 to a specific set of substrates. Since the Cdc5 binding site was mapped to the N-terminal portion of Dbf4 (Miller et al, 2009), we tested whether N-terminal truncations of Dbf4 would affect DDK or Cdc5 association with Mus81-Mms4. While the dbf4-ΔN66 truncation lacking the first 66 amino acids (including a D-box motif) did not influence DDK or Cdc5 binding to Mms43FLAG, the dbf4-ΔN109 truncation, which additionally lacks the Cdc5 binding motif (Miller et al, 2009), showed strongly decreased DDK and Cdc5 binding to Mus81-Mms4 (Fig 2F). Additionally, also mitotic hyperphosphorylation of Mms4 was diminished when DDK and Cdc5 could not interact with each other (Fig 2F). Overall, these data strongly suggest that Cdc5 and DDK interact with and target Mus81-Mms4 in an interdependent manner. Furthermore, it is currently unclear whether collaboration of DDK and Cdc5 is a widespread phenomenon that may affect other Cdc5 substrates as well, given that mitotic phosphorylation of two candidate Cdc5 substrates, Ulp2 and Scc1 (Alexandru et al, 2001), was affected to varying degree by the cdc7Δ mutation (Appendix Fig S2E). Given the known cell cycle regulation of Cdc5 and DDK (Shirayama et al, 1998; Cheng et al, 1999; Weinreich & Stillman, 1999; Ferreira et al, 2000; Mortensen et al, 2005), the limiting factor for the temporal regulation of this complex and its restriction to mitosis is expected to be Cdc5 and not DDK, which is present already throughout S phase. Consistently, we observed that forced expression of Cdc5 (using the galactose-inducible GAL promoter) in cells that were arrested in S phase by hydroxyurea (HU) led to the premature occurrence of Mms4 hyperphosphorylation (Fig EV1A; Matos et al, 2013), suggesting that S-phase DDK is in principle competent for Cdc5 binding and joint substrate phosphorylation. Click here to expand this figure. Figure EV1. Cdc5 restricts Mms4 hyperphosphorylation to mitosis Overexpression of CDC5 in S phase results in premature Mms4 hyperphosphorylation. Western blot analysis of Mms49myc, Cdc5 and Dbf4 from whole-cell extracts (upper panel) and FACS data (lower panel). Cells were arrested in G1 (with alpha-factor), S phase (with HU) or G2/M phase (with nocodazole). After arrest, CDC5GFP overexpression was induced by addition of 2% galactose for the indicated time to cells harbouring an additional copy of GFP-tagged CDC5 under the GAL1 promoter. Samples were run in 7% Tris-acetate gels. Mms4 hyperphosphorylation by CDC5 overexpression in S phase is reduced in HU-treated cells. Western blot analysis of Mms49myc and Cdc5 from precipitated whole-cell extracts (upper panel) and FACS data (lower panel) of cells arrested in G1 (with alpha factor) or G2/M phase (with nocodazole), or released to S phase (with or without HU). CDC5GFP overexpression was induced for 30 min by addition of 2% galactose to cells harbouring an additional copy of GFP-tagged CDC5 under the GAL1 promoter. Note that upon CDC5 overexpression cells are partially defective in bulk replication. Samples were run in 7% Tris-acetate gels. Download figure Download PowerPoint Furthermore, we performed additional experiments that addressed the regulation of Mus81-Mms4 by the DNA damage response. In M-phase-arrested cells, association of DDK and Cdc5 with Mus81-Mms4 was reduced after induction of DNA damage with phleomycin (Appendix Fig S2F), but this treatment was not sufficient to induce a significant reduction in the Mms4 phosphorylation shift. Interestingly, when we forced Cdc5 expression in S-phase cells and compared normal S-phase cells to cells treated with hydroxyurea (HU), we observed that the Mms4 phosphorylation shift was less pronounced in the presence of hydroxyurea (HU) (Fig EV1B). These data are therefore consistent with the current view that DNA damage, specifically the DNA damage checkpoint, negatively influences Mus81 resolution activity (Szakal & Branzei, 2013; Gritenaite et al, 2014). Since DDK is known to be targeted and inhibited by the DNA damage checkpoint (Weinreich & Stillman, 1999; Lopez-Mosqueda et al, 2010; Zegerman & Diffley, 2010), it could become particularly critical to regulate Mms4 phosphorylation after DNA damage. Even though DDK and Cdc5 seem to target Mus81-Mms4 in unison, we tested whether it was possible to resolve differences on the level of individual phosphorylation sites. Therefore, we analysed Mms4 phosphorylation sites in M-phase cells after Cdc5 inhibition (Fig 3A and C) or CDC7 deletion (Fig 3B and D) by SILAC-based mass spectrometry. We also applied two different experimental set-ups that used either endogenously expressed Mus81-Mms4 (Fig 3A and B) or overexpressed Mus81-Mms4 (Fig 3C and D), as the latter set-up allowed much better coverage of Mms4 phosphopeptides in higher order phosphorylation states (peptides harbouring > 1 phosphorylated site). Cdc5 inhibition or lack of DDK led to overlapping, but distinct changes in Mms4 phosphorylation sites, suggesting that each kinase phosphorylates specific sites on Mms4. After Cdc5 inhibition, phosphorylation of many sites was reduced and among those were sites that match to a putative Cdc5 consensus [(D/E/N)X(S/T), blue, Fig 3A and C; Mok et al, 2010]. Overall, CDC7 affected Mms4 phosphorylation less than Cdc5 inhibition, but nonetheless, we found widespread changes in the phosphorylation of (S/T)(S/T) motifs (Fig 3B and D). (S/T)(S/T) motifs were found less abundantly in the doubly phosphorylated state (Fig 3D, red), while conversely these motifs were found more abundantly in the state where only the second (S/T) was singly phosphorylated (Fig 3B and D, yellow), as expected for a substrate–product relation. These data are thus consistent with phosphorylation of the second (S/T) priming for phosphorylation at the preceding (S/T) (Appendix Table S1 and Appendix Fig S3). Figure 3. Analysis of Mms4 phosphorylation sites reveals Cdc5 and DDK target sites, as well as the interdependence between the twoChanges of the abundance of phosphorylated Mms4 peptides after Cdc5 inhibition (as in Fig 2C) (A, C) or in the absence of Cdc7 (B, D) in mitotically arrested cells. A, B. Depicted are SILAC-based intensity ratios of individual MS evidences for peptides of endogenously expressed Mms4. Evidences of non-phosphorylated Mms4 peptides are shown in grey; evidences of phosphorylated peptides are shown in black, yellow, orange or blue. Blue colour indicates putative Cdc5 phosphorylation as defined by the (D/E/N)X(S/T) consensus (and additionally S268, which was also very strongly deregulated upon Cdc5 inhibition). Yellow or orange colours mark singly phosphorylated (S/T)(S/T) motifs, with orange marking p(S/T)(S/T) and yellow marking (S/T)p(S/T). Numbers indicate the phosphorylated residue in the depicted peptide. An asterisk marks peptide evidences that contained measured intensity values exclusively in the heavy or light sample. For doubly phosphorylated peptides, the two phospho-sites are separated by a comma. For singly phosphorylated (S/T)(S/T) motifs, peptide ion fragmentation was in some cases unable to unambiguously identify the phosphorylated residue. In these cases, possible phosphorylation sites are indicated as “a/b”. Note that doubly phosphorylated (S/T)(S/T) sites were not reproducibly identified under conditions of endogenous Mus81-Mms4 expression. C, D. As in panels (A, B) but using Mus81-Mms4 expressed from a high-copy promoter. Depicted are SILAC-based H/L ratios of individual MS evidences for phosphorylated peptides only. Peptides were sorted into categories according to their phosphorylation status: putative DDK target sites ((S/T)(S/T) motifs) were differentiated into the categories p(S/T)p(S/T) (red), p(S/T)(S/T) (orange) or (S/T)p(S/T) (yellow). Phosphorylated peptides matching the Cdc5 consensus site are coloured in blue. All other phosphorylated peptides are marked in grey. Bars depict the mean of the ratios of the respective category. Overall, Mms4 H/L ratio is shown on top. Download figure Download PowerPoint DDK phosphorylation is required for activation of Mus81-Mms4 during mitosis Phosphorylation of Mms4 by CDK and Cdc5 has previously been shown to be required for the upregulation of Mus81-Mms4 activity during mitosis (Matos et al, 2011, 2013; Gallo-Fernández et al, 2012; Szakal & Branzei, 2013). Based on our finding that hyperphosphorylation of Mms4 was impaired in the absence of DDK (Fig 2D and Appendix Fig S2C), we predicted that also Mus81-Mms4 activity would be influenced. Therefore, we tested the activity of endogenous Mus819myc-Mms43FLAG immunopurified from G2/M arrested cells (approx. 5 fmol) on a nicked Holliday junction (nHJ) substrate (500 fmol) using an assay related to those in (Matos et al, 2011, 2013; Gritenaite et al, 2014). Notably, the activity of the endogenous purified Mus81-Mms4 from G2/M cells exceeded the activity of recombinant Mus81-Mms4 (subjected to a dephosphorylation step during the purification), indicating that it is the mitotically activated form (Appendix Fig S4A). Moreover, the activity of endogenous purified Mus81-Mms4 was not influenced by 350 mM NaCl salt washes. This indicates that the presence of accessory, salt-labile factors such as Rtt107 or Cdc5 in the reaction is unlikely to contribute to Mus81 activity (Appendix Fig S4B and C). Importantly, when we used this assay to test immunopurified Mus819myc-Mms43FLAG from mitotic cells lacking DDK (cdc7Δ or dbf4Δ), we observed a reduced activity compared to Mus819myc-Mms43FLAG from WT cells (Fig 4A and Appendix Fig S4D; also observed with an RF substrate, Appendix Fig S4E). In order to exclude that indirect effects of the CDC7 deletion may cause the reduction in Mus81 activity, we furthermore created an Mms4 mutant that specifically lacks candidate DDK phosphorylation sites. We chose to mutate (S/T)(S/T) motifs (SS motifs in particular) and created an mms4-8A mutant that harboured eight S to A exchanges at the N-terminal (S/T) of the motifs (see Appendix Fig S3A). This mutant appeared less phosphorylated in mitosis as judged by a less pronounced phosphorylation shift (Fig 4B). Furthermore, we observed a reduction in the association of DDK and Cdc5 with the Mus81-Mms4-8A complex in pull-down experiments (Fig 4B), suggesting that phosphorylation of Mms4 also plays a role in tethering these kinases. Notably, Mus819myc-Mms43FLAG-8A from mitotic cells showed a moderate but reproducible reduction in resolution activity on nHJ substrates compared to WT Mus819myc-Mms43FLAG (Fig 4C and Appendix Fig S4F). These data thus indicate that DDK targets Mus81-Mms4 and that (S/T)(S/T) phosphorylation events are essential for full activation of Mus81 in mitosis. Figure 4. DDK phosphorylation controls activation of Mus81-Mms4 resolvase activity in mitosis A. DDK is required for mitotic activation of Mus81-Mms4. Resolution assay using a nicked Holliday junction (nHJ) substrate and Mus819myc-Mms43FLAG purified from mitotically arrested bob1-1 (DDK-WT+), bob1-1 dbf4Δ and bob1-1 cdc7Δ strains or untagged control cells. Right panel: quantification of cleavage products. See Appendix Fig S4D for Western blots samples of anti-myc IPs. Left panel: representative gel image. B. A defect in the phosphorylation of Mms4 (S/T)(
Referência(s)