Distinct Targets of the Eco1 Acetyltransferase Modulate Cohesion in S Phase and in Response to DNA Damage

doi:10.1016/j.molcel.2009.04.008

Mol Cell. Author manuscript; available in PMC 2009 November 15.

Published in final edited form as:

Mol Cell. 2009 May 15; 34(3): 311–321.

doi: 10.1016/j.molcel.2009.04.008

PMCID: PMC2737744

HHMIMSID: HHMIMS119231

Distinct Targets of the Eco1 Acetyltransferase Modulate Cohesion in S Phase and in Response to DNA Damage

Jill M. Heidinger-Pauli,^1,² Elçin Ünal,^1,²^* and Douglas Koshland¹^#

¹Carnegie Institution, Howard Hughes Medical Institute, Department of Embryology, 3520 San Martin Dr, Baltimore, Maryland 21218, USA

²Johns Hopkins University, Department of Biology, 3400 N Charles St, Baltimore, Maryland 21218, USA

^# Correspondence: Email: koshland/at/ciwemb.edu, phone: (410) 246-3016, fax: (410) 467-1147

^*Current address: Massachusetts Institute of Technology, Howard Hughes Medical Institute, David H. Koch Institute for Integrative Cancer Research, 77 Massachusetts Ave, Cambridge 02139, USA

The publisher's final edited version of this article is available at Mol Cell

See other articles in PMC that cite the published article.

Publisher's Disclaimer

Summary

Chromosome segregation and the repair of DNA double-strand breaks (DSBs) require cohesin, the protein complex that mediates sister chromatid cohesion. Cohesion requires both a chromatin binding step and a subsequent tethering step called cohesion generation. Here we provide insight into how cohesion generation is restricted to S phase but can be activated in G2/M by a DSB in budding yeast. We show that Wpl1p inhibits cohesion in G2/M. A DSB counteracts Wpl1p and stimulates cohesion generation by first inducing the phosphorylation of the Mcd1p subunit of cohesin. This phosphorylation activates Eco1p-dependent acetylation of Mcd1p, which in turn antagonizes Wpl1p. Previous studies show that Eco1p antagonizes Wpl1p in S phase by acetylating the Smc3p subunit of cohesin. We show that Mcd1p and Smc3p acetylation antagonize Wpl1p only in their proper context. Thus, Eco1p antagonizes Wpl1p in distinct ways to modulate cohesion generation during the cell cycle and after DNA damage.

Introduction

Cohesin physically tethers sister chromatids together to mediate sister chromatid cohesion. Cohesion ensures accurate chromosome segregation at the metaphase to anaphase transition (reviewed in (Onn et al., 2008)), and efficient repair of DNA double-strand breaks (DSBs) (Kim et al., 2002; Sjogren and Nasmyth, 2001; Strom et al., 2004; Unal et al., 2004). To achieve proper chromosome segregation and efficient DNA repair, cohesin binding to chromatin is regulated both temporally and spatially. From G1/S through G2/M in budding yeast, cohesins are targeted to pericentric regions and to distinct AT-rich intergenic regions along chromosome arms, called cohesin-associated regions (CARs) (reviewed in (Onn et al., 2008)). Upon a DSB, cohesins are targeted to a large domain flanking the break site (Strom et al., 2004; Unal et al., 2004).

Mere binding of cohesin to chromatin is not sufficient to tether sister chromatids together. Rather a second step, henceforth referred to as cohesion generation, is needed to convert chromatin-bound cohesin to a stable cohesive state (Skibbens et al., 1999; Toth et al., 1999). Cohesion generation is restricted to S phase, but is induced to occur in G2/M when a DSB is present (Haering et al., 2004; Lengronne et al., 2006; Strom et al., 2007; Strom et al., 2004; Uhlmann and Nasmyth, 1998; Unal et al., 2007). This reactivation of cohesion generation in response to a DSB is required for efficient post-replicative repair of DSBs (Strom et al., 2007; Unal et al., 2007). How the reactivation of cohesion generation upon DSBs is regulated is an important but poorly understood aspect of cohesion biology.

The key promoter of cohesion generation in S and G2/M phase is the evolutionarily conserved Eco1 (Ctf7) acetyltransferase (Skibbens et al., 1999; Strom et al., 2007; Toth et al., 1999; Unal et al., 2007). Recombinant Eco1 protein (Eco1p) can acetylate a number of cohesin subunits in vitro: Mcd1p (Scc1p), Scc3p, Smc3p, and Pds5p (Bellows et al., 2003; Ivanov et al., 2002; Williams et al., 2003; Unal et al., 2008). Eco1p mediated acetylation of the Smc3p cohesin subunit at Smc3-K112, K113 is essential for the establishment of cohesion in S phase in budding yeast (Ben-Shahar et al., 2008; Unal et al., 2008; Zhang et al., 2008). Acetylation of these residues inhibits Wpllp (Rad61p), an evolutionarily conserved antagonist of cohesion (Ben-Shahar et al., 2008; Bernard et al., 2008; Sutani et al., 2009; Dobie et al., 2001; Gandhi et al., 2006; Kueng et al., 2006; Verni et al., 2000).

In undamaged cells, cohesion generation in G2/M is prevented because Eco1p activity becomes limiting after S phase (Strom et al., 2007; Unal et al., 2007). G2/M cells behave as though they lack Eco1p; cohesins load at CARs and pericentric regions but do not become cohesive (Haering et al., 2004; Lengronne et al., 2006; Strom et al., 2004; Uhlmann and Nasmyth, 1998). Over-expression of Eco1p drives cohesins loaded in G2/M to become cohesive (Unal et al., 2007), further supporting the idea that cohesion generation in G2/M is prevented due to limiting Eco1p activity.

This situation changes when a G2/M cell suffers a DSB. Eco1p acts on the cohesins that are loaded both in the vicinity of the break site and genome-wide to convert them to their cohesive state (Strom et al., 2007; Unal et al., 2007). Thus, the normal inhibition of Eco1p in G2/M is relieved by a DSB. This change in Eco1p activity is mediated through a DNA damage signal transduction pathway requiring two key damage induced kinases, the master checkpoint kinase Mec1p/ATR, and its downstream target Chk1p (Heidinger-Pauli et al., 2008; Strom et al., 2007; Unal et al., 2007). Genetic evidence supports a model where the Chk1p kinase stimulates cohesion generation by phosphorylating the S83 residue of Mcd1p (Scc1p), the kleisin subunit of cohesin (Heidinger-Pauli et al., 2008). This kinase cascade acts upstream of Eco1p in DSB-induced cohesion generation. However, how this kinase cascade activates Eco1p is unknown.

Since Eco1p is required for cohesion generation both in S phase and after a DSB, it seemed reasonable it would activate cohesion generation by the same mechanism in these two contexts. However, cells expressing Eco1p with the eco1-R222G, K223G mutations in its acetyltransferase domain are defective only in DSB-induced cohesion but not S phase cohesion (Brands and Skibbens, 2005; Unal et al., 2007). Previous work has shown that this mutant is only partially compromised for its acetyl-transferase activity in vivo (Unal et al., 2007; Zhang et al., 2008). This separation-of-function allele of ECO1 suggests that some aspect of Eco1p function must differ between S phase and DNA damage. One possibility is the acetylation targets of Eco1p are different. Eco1p acetylates the Smc3p subunit to generate cohesion in S phase (Ben-Shahar et al., 2008; Unal et al., 2008; Zhang et al., 2008). However, our studies point to the Mcd1p kleisin subunit as the key to controlling cohesion generation after a DSB (Heidinger-Pauli et al., 2008). Therefore Eco1p might generate cohesion by acetylating Smc3p in S phase and Mcd1p after a DSB.

In the current study, we address two important questions about the regulation of cohesion generation: how does Eco1p stimulate cohesion generation in G2/M in response to DSBs and how does this function of Eco1p compare with its role in cohesion generation during S phase. Our data suggest that Eco1p promote cohesion generation in S phase and in G2/M after a DSB by acetylating different cohesin subunits, which in turn antagonize Wpl1p by distinct mechanisms. We discuss the implications of these observations for the mechanism of cohesion generation.

Results

The Eco1p Acetyltransferase Potentially Regulates DSB-induced Cohesion Generation by Targeting Mcd1p

Since precedent exists for the phosphorylation of a protein enhancing its subsequent acetylation (Lo et al., 2000; Sakaguchi et al., 1998), we hypothesized that DSB-induced phosphorylation of Mcd1-S83 enhances its acetylation by Eco1p to activate DSB-induced cohesion generation. To test for a functional interaction between Mcd1p phosphorylation and Eco1p acetyltransferase activity, we exploited an assay that we developed previously to specifically monitor cohesion generated in G2/M ((Unal et al., 2007), Figure 1A). In this assay, S phase cohesion is established by cohesins composed of a thermo-sensitive mcd1 allele at the permissive temperature. Cells are subsequently arrested in G2/M and variants of Mcd1p (wild-type or mutant) are expressed either alone or concomitant with the induction of two DSBs on chromosome III. Finally, S phase established cohesion is inactivated by a temperature shift, which allows the measurement of cohesion generated solely in G2/M through the use of a GFP marking system on chromosome IV (Figure 1A, B).

Figure 1

Mcd1 Phosphomimic Rescues the DSB-induced Cohesion Defect of the eco1-R222G, K223G Mutant

If Chk1p dependent phosphorylation of Mcd1p stimulates Mcd1p acetylation by Eco1p, then overexpression of Eco1p might bypass the need for Mcd1p phosphorylation to generate cohesion in G2/M. Indeed, over-expression of Eco1p generates cohesion in G2/M in the absence of a DSB, a condition where Chk1p kinase activity is repressed (Unal et al., 2007). Furthermore, Eco1p over-expression generates cohesion in G2/M even when phosphorylation of Mcd1p at Mcd1-S83 is prevented by mutation of the residue to alanine. (Supplemental Figure 1). Thus, elevating the levels of Eco1p can bypass the need for Mcd1-S83 phoshorylation in cohesion generation in G2/M.

Cells expressing eco1p-R222G, K223G are defective for DSB-induced cohesion but not S phase cohesion ((Unal et al., 2007), Figure 1C), indicating eco1p-R222G, K223G’s acetyltransferase activity is altered but not eliminated. If eco1p-R222G, K223G is less efficient at acetylating Mcd1p and if Mcd1p phosphorylation augments its subsequent acetylation, then a phospho-mimetic mutant of Mcd1-S83 (S83D) might enable eco1p-R222G, K223G to acetylate Mcd1p. Indeed, the cohesion defect of cells containing eco1p-R222G, K2223G is suppressed by expression of mcd1p-S83D (Figure 1C). In summary, these genetic interactions between Mcd1p phosphorylation and Eco1p acetylatransferase activity in DSB-induced cohesion are consistent with Mcd1p phosphorylation augmenting its acetylation by Eco1p.

Mcd1p-K84, K210 are Required for DSB-induced Cohesion and are Likely Targets of Eco1p

To test the hypothesis that Mcd1p might be acetylated in response to DSBs, we mutated Mcd1-K210, which is acetylated by Eco1p in vitro, to the non-acetylatable residue arginine (Ivanov et al., 2002). Cohesin containing mcd1p-K210R is partially deficient for DSB-induced cohesion generation (Figure 2A). Because the defect is only partial, we reasoned that additional Eco1p acetylation site(s) on Mcd1p may be required for DSB-induced cohesion. An attractive candidate site was Mcd1-K84, a conserved lysine that was immediately adjacent to Mcd1-S83. Mcd1-K84 was attractive not only because of its proximity to Mcd1-S83, but also because it shared some similarities in adjacent amino acid content to Mcd1-K210, and the lysines acetylated by Eco1p on Smc3p (Supplemental Figure 2). Cohesin containing mcd1p-K84R is also partially defective for generating DSB-induced genome-wide cohesion in response to a DSB (Figure 2A). However, the double mutant is unable to activate cohesion generation in response to DSBs despite being proficient for S phase cohesion (Figure 2A, Supplemental Figure 4). Mcd1-K84, K210 residues are therefore likely targets of Eco1p acetylation in response to DSBs.

Figure 2

Mcd1-K84, K210 Residues are Required for DSB-induced Cohesion

To test whether mcd1p-K84R, K210R causes a cohesion defect in G2/M because it blocks Eco1p acetylation, we asked whether mcd1-K84R, K210R mutants phenocopy the defect of eco1 cells in DSB-induced cohesion. That is, the mutant cohesin should be proficient to bind chromatin but unable to generate cohesion (Skibbens et al., 1999; Toth et al., 1999). This is a rare cohesin mutant phenotype, seen only in eco1 cells and cells with the mcd1-S83A allele (Heidinger-Pauli et al., 2008; Unal et al., 2007). We examined the chromatin binding of cohesin containing mcd1p-K84R, K210R by chromatin spreads and chromatin immunoprecipitation. By these two assays, chromatin binding of cohesin containing wild-type Mcd1p or mcd1p-K84R, K210R is indistinguishable genome-wide (Figure 2B, Supplemental Figure 3). Furthermore, cohesin containing wild-type Mcd1p or mcd1p-K84R, K210R is enriched around the site of a DSB (Figure 2C). Thus mcd1-K84R, K210R mutant cells, like eco1 cells and cells with the mcd1-S83A allele, are specifically defective in converting chromatin-bound cohesin into a cohesive state in G2/M. Mcd1-K84, K210 are thus strong candidates for in vivo acetylation targets of Eco1p.

If Mcd1-K84, K210 are the critical Mcd1p residues acetylated by Eco1p in response to a DSB, then cells expressing Mcd1p with acetyl-mimics of these residues (mcd1p-K84Q, K210Q) should generate cohesion in G2/M under conditions when Eco1p activity is limiting. Cells containing mcd1p-K84Q, K210Q should thus be able to generate cohesion in G2/M in the absence of a DSB as well as in the presence of either eco1p-R222G, K223G or eco1-203, a thermo-sensitive allele, at the non-permissive temperature. Under all of these conditions, cells expressing wild-type Mcd1p fail to generate cohesion while cells expressing mcd1p-K84Q, K210Q can generate cohesion to nearly the level that occurs when wild-type Eco1p is fully activated by a DSB (Figure 3A-3D, Supplemental Figure 5). While biochemical evidence for modification of these residues would provide the final proof of their acetylation, we have so far been unable to obtain such evidence in vivo (see experimental procedures). Nonetheless, our genetic data strongly argues that these two Mcd1p residues are critical acetylation targets of Eco1p: the acetyl-mimic mcd1p-K84Q, K210Q bypasses requirement for Eco1p in DSB-induced cohesion, and the acetyl-null mcd1p-K84R, K210R blocks G2/M cohesion generation after chromatin binding identical to inactivation of Eco1p.

Figure 3

Mcd1p-K84Q, K210Q Bypass the Requirement for a Break and Eco1p to Generate Cohesion in G2/M

Mcd1-K84, K210 residues act Downstream of Chk1p Kinase Activity to Generate DSB-induced Cohesion

Our modification null and mimic alleles enable us to order the Chk1p kinase and Eco1p acetyltransferase in the signal transduction pathway for DSB-induced cohesion generation. With the modification null and mimic alleles of mcd1-S83 and mcd1-K84, K210, two outcomes are predicted. First, a signal blocked by inactivation of an upstream regulator in the pathway will be restored by constitutive activation of a downstream regulator. Conversely the signal generated by constitutively active upstream regulator will be blocked by inactivation of a downstream regulator. Indeed, cells blocked for Mcd1p phosphorylation by mcd1p-S83A are unable to induce DSB-induced cohesion ((Heidinger-Pauli et al., 2008) and Figure 4A). However, over-expression of Eco1p or introduction of mcd1p-K84Q, K210Q (constitutive-acetyl mimic) into these mutants restores cohesion (Figure 4A and Supplemental Figure 1). Furthermore, induction of mcd1p-K84Q, K210Q rescues the defect of cells lacking Chk1p, the kinase likely responsible for Mcd1-S83 phosphorylation (Figure 4B). These results suggest that Eco1p mediated acetylation is downstream of Chk1p kinase in DSB-induced cohesion generation. Furthermore, constitutive cohesion generation caused by expression of the constitutive phosphomimic, mcd1p-S83D, or over-expression of Eco1p is blocked by introducing the Mcd1-K84R, K210R mutations (Figure 4C,D). These results also place the Chk1p kinase and Mcd1p phosphorylation upstream of Eco1p acetylation. The ordering of these activities suggests a model where Chk1p mediated phosphorylation of Mcd1p directly stimulates its subsequent acetylation by Eco1p.

Figure 4

Mcd1p-K84QK210Q Bypasses the Requirement for Chk1p and Mcd1-S83 to Generate Cohesion in G2/M

Mcd1-K84 and Smc3-K112, K113 Residues are Proximal in the Cohesin Complex and Mediate Their Function through Wpl1p

While no cohesin structure contains Mcd1-K84 or Mcd1-K210, we could approximate the spatial position of Mcd1-K84 by creating a model of the Smc3p head bound to the Mcd1p N-terminus. This model was obtained by superimposing the sequences of the Smc3p head and Mcd1p N terminus on the highly homologous crystal structure of the Smc1p head and Mcd1p C terminus (Figure 5A) (Haering et al., 2004). Mcd1-K84 maps proximal to the ATP binding pocket of Smc3p and the Smc3-K112, K113 acetylation sites. The proximity of Mcd1-K84 and Smc3- K112, K113 to each other is consistent with Eco1p acetylation modulating cohesion generation in S phase and after a DSB by similar mechanism, perhaps by modulating Smc3p’s ATPase activity.

Figure 5

wpl1Δ allows Cohesion Generation in G2/M Without a Break and Bypasses the Requirement for the K84, K210 Residues on Mcd1p

A previous study indicated that Eco1p dependent acetylation of Smc3p promotes cohesion generation by antagonizing Wpl1p, a cohesion inhibitor (Ben-Shahar et al., 2008, Sutani et al., 2009). The spatial proximity of K84 to the Smc3p acetylation sites suggests that Eco1p mediated acetylation could promote cohesion generation in the context of DNA damage also by antagonizing Wpl1p. If this is true, then deleting Wpl1p (wpl1Δ) should bypass the need for Eco1p mediated acetylation and allow cohesion in G2/M without a DSB.

To test this, we had to modify our assay for cohesion generation because we were unable to introduce a wpl1Δ into our assay strains, which contain mcd1-1, due to extreme synthetic sickness. Therefore, we engineered a new strain built upon the same principles but using a thermo-sensitive smc3 allele and inducible Smc3p. Similar to the mcd1 based assay, the smc3 based assay reveals that cohesion generation does not occur in G2/M in the absence of a DSB or without G2/M expression of wild-type Smc3p (Figure 5B). We then asked whether deletion of Wpl1p (wpl1Δ) bypasses the need for a DSB to generate cohesion in G2/M. Indeed its elimination allows cohesion generation in G2/M both in the presence and absence of a DSB (Figure 5C).

We next constructed a strain where mcd1p-K84R, K210R was the sole source of Mcd1p with or without the wpl1Δ. Consistent with our previous results using our mcd1 based assay, cells expressing mcd1p-K84R, K210R are unable to generate DSB-induced cohesion in the smc3 based assay (Figure 5D). In contrast, cohesin containing mcd1p-K84R, K210R and lacking Wpl1p were able to generate DSB-induced cohesion as robust as cells containing Mcd1p and Wpl1p cells (Figure 5D). Therefore, Eco1p mediated acetylation promotes cohesion generation during S phase and after a DSB by antagonizing the Wpl1p inhibitor.

Mcd1-K84, K210 and Smc3-K112, K113 Residues Antagonize Wpl1 by Distinct Pathways

Given that Mcd1-K84 and Smc3-K112, K113 map to a similar location within the cohesin complex and commonly function by antagonizing Wpl1p, it seemed reasonable that these modifications might function by a common mechanism. To address the functional interplay between Mcd1-K84, K210 and Smc3-K112, K113, we began by asking whether Mcd1-K84, K210 were also required for S phase cohesion. Cells expressing only mcd1p-K84R, K210R are viable, and completely proficient for S phase cohesion (Supplemental Figure 4). Thus acetylation of Mcd1-K84, K210 are not necessary for S phase cohesion. This result is consistent with the fact that phosphorylation of Mcd1-S83 is also required for DSB-induced cohesion but not S phase cohesion. However, it was still possible that the acetylation of Mcd1p-K84, K210 might be able to substitute for acetylation of Smc3p-K112, K113 to antagonize Wpl1p in S phase. If so, then mcd1p-K84Q, K210Q would bypass the requirement of Eco1p in S phase cohesion. However, mcd1p-K84Q, K210Q fails to rescue the inviability of a thermo-sensitive eco1 allele or the smc3-K112R, K113R (Figure 6A, 6B). Therefore, Mcd1p acetylation cannot substitute for Smc3p acetylation in S phase. Together these results suggest that Eco1p-dependent acetylation of Mcd1p antagonizes Wpl1p only in G2/M.

Figure 6

Smc3 Acetylation Events are Uniquely Required for S phase Cohesion and not G2/M Cohesion and Mcd1-K84, K210 Residues are Uniquely Required for DSB-induced Cohesion

We next asked whether Smc3p acetylation was required for DSB-induced cohesion generation and whether the Smc3p acetyl-mimic was capable of antagonizing Wpl1p in G2/M. To address this question we used our modified G2/M cohesion assay based upon the thermo-sensitive smc3 allele. Cells harboring the thermo-sensitive smc3p are allowed to pass through S phase at the permissive temperature to establish cohesion and then arrested in G2/M. Then a double strand break is induced and Smc3p or smc3p-K112R, K113R is expressed. S phase cohesin is inactivated by raising the temperature, revealing the cohesion generated in G2/M by cohesin containing smc3p-K112R, K113R. Cells expressing smc3p-K112R, K113R in G2/M can mediate DSB-induced cohesion almost as well as cells expressing Smc3p (Figure 6C). Thus Smc3 acetylation is not obligatory for DSB-induced cohesion generation despite being essential for S phase cohesion. We then used this assay to ask whether cells expressing the acetyl-mimic smc3p-K112Q, K113Q could bypass the need for a DSB and allow cohesion generation in unperturbed G2/M cells. Cells expressing smc3p-K112Q, K113Q can generate cohesion in G2/M in the presence of a DSB but not in the absence of a DSB (Figure 6D). Therefore, Smc3p acetylation is unable to antagonize Wpl1p in G2/M phase even though it is proficient to do so in S phase. Thus Eco1p antagonizes Wpl1p in S phase and G2/M, but only by acetylating distinct subunits of cohesin.

Discussion

In the current study, we provide significant insights into how Eco1p stimulates cohesion generation in G2/M in response to DSBs, how the DNA damage signal transduction pathway coordinate this response, and finally how DSB-induced cohesion compares to the cohesion generation that occurs during S phase. We provide multiple lines of evidence that Eco1p regulates DSB-induced cohesion by acetylating Mcd1-K84, K210 upon DNA damage-induced phosphorylation of Mcd1-S83 ((Heidinger-Pauli et al., 2008), Figure 7). Nonacetylatable mutation of mcd1-K84, K210 phenocopies eco1 mutations. Both block DSB-induced cohesion at a step after cohesin binding to chromatin and both block constitutive cohesion of Mcd1-S83D, placing them downstream of Chk1p in the DNA damage kinase cascade. The acetyl-mimics of Mcd1 (Mcd1-K84Q, K210Q) also phenocopy Eco1p over-expression: both allow cohesion generation in G2/M in the absence of a DSB and Chk1p. Furthermore, the acetyl-mimics bypass the requirement for Eco1p in DSB-induced cohesion. These data provide strong support for the model that Mcd1-K84, K210 are Eco1p acetylation targets, although future biochemical studies will be required to confirm this conclusion in vivo.

Figure 7

Model Illustrating the Regulation of Cohesion Generation in S Phase and G2/M Phase

Here we provide evidence that Eco1p’s ability to acetylate Mcd1p is enhanced by Mcd1p phosphorylation. Preventing Mcd1-S83 phosporylation, or the deleting its likely kinase, Chk1p, results in an identical phenotype to preventing acetylation of Mcd1p at residues of K84 and K210. Namely, DSB-induced cohesion, but not S phase cohesion is blocked at a step after cohesin binding ((Heidinger-Pauli et al., 2008), this study). Acetyl-mimic and phospho-mimic mutants phenocopy elevated expression of Eco1p; they all generate cohesion in G2/M in the absence of a DSB. Finally epistasis analysis with the modification-null and modification–mimic alleles indicate that DSB-induced phosphorylation of Mcd1-S83 enhances Eco1p’s ability to acetylate Mcd1p in G2/M, thereby increasing the relative activity of Eco1p.

Based upon these observations, we speculate that Eco1p has a poor affinity for its substrates not only in G2/M, but also in S phase. Indeed, Eco1p interacts poorly with soluble cohesin and requires Scc2p/Scc4p (loading factors for cohesin) to acetylate Smc3p during S phase, leading to the suggestion that Eco1p preferentially acetylates cohesin bound to chromatin (Noble et al., 2006; Unal et al., 2008). During S phase, Eco1p is likely recruited to chromatin through PCNA associated with DNA replication, thereby ensuring the proper timing of cohesion generation(reviewed in(Skibbens et al., 2007)). In G2/M, phosphorylation of Mcd1p may serve as a replication independent surrogate to facilitate Eco1p function. These results set the precedent that other unique facilitators of Eco1p may exist to modulate cohesin function in other contexts such as transcription.

We propose that Eco1p promotes DSB-induced cohesion by antagonizing Wpl1p. The putative targets of Eco1p for DNA damage, Mcd1-K84, K210, are no longer required for DSB-induced cohesion when Wpl1p is absent. Furthermore, wpl1Δ allows cohesion generation in G2/M in the absence of a DSB. These results strongly suggest that Wpl1p antagonizes cohesion generation in G2/M, explaining why cohesion generation is normally limited to S phase, and that the function of Mcd1-K84, K210 acetylation is to counteract Wpl1p (Figure 7).

The role of Wpl1p as an antagonist in mitosis in budding yeast agrees with previous studies of its mammalian homolog, Wapl. During prophase, Wapl is required for the dissolution of S phase established cohesion from chromosome arms (Gandhi et al., 2006; Kueng et al., 2006). Wapl could also act immediately on any cohesion formed during G2/M thereby blocking cohesion generation. Indeed, depletion of Wapl in results in a larger amount of cohesin on chromosomes in prophase than would be expected by just inhibiting the normal removal of cohesin loaded on chromosome arms during S phase. Thus Wapl inactivation may allow new cohesion generation in G2/M in mammalian cells similar to what we have shown for Wpl1p in budding yeast (this study). Interestingly, Wapl-stimulated cohesion dissolution in mammalian cells leads to the removal of cohesin from chromosome arms as well as loss of cohesion. One intriguing possibility is that Wapl counteracts cohesion at a step independent from cohesin binding as Wpl1p does in budding yeast, but these non-cohesive cohesins are then more susceptible to being removed from chromosomes in mammalian cells.

Here we provide a model where Eco1p and Wpl1p act as a simple positive and negative regulatory switch controlling cohesion generation (Figure 7). However, both the regulatory switch and the function of the Wpl1p and Eco1p are likely to be more complicated. The regulatory switch may contain additional components like Pds5p, a cohesion accessory factor that associates with Eco1p and possibly Wpl1p as well (Gandhi et al., 2006; Noble et al., 2006; Sutani et al., 2009). Furthermore, Wpl1p may have additional functions beyond controlling cohesion generation. For example, wpl1Δ cells are mildly sensitive to irradiation, a phenotype that cannot be explained by constitutive cohesion generation in G2/M (Game et al., 2003). Similarly Eco1p may regulate other cohesin functions. In fact, Eco1p acetylates other residues on Smc3p and other cohesin subunits (Ivanov et al., 2002; Unal et al., 2008). The physiological relevance of these acetylation events will be determined when they are analyzed in their proper context.

Eco1p antagonizes Wpl1p likely by acetylating Mcd1-K84, K210 for DSB-induced cohesion and Smc3-K112, K113 for S phase cohesion in budding yeast (This study, (Ben-Shahar et al., 2008)). These pathways are distinct and each modification is functional only in its specific context. These observations reveal a surprising complexity to the regulation of cohesion generation. To explain this complexity we envision at least two possibilities. Wpl1p may always inhibit the same step in cohesion generation, for example by binding to the Smc3p-Mcd1p subdomain and modulating Smc3p ATPase activity. However, the interaction between Wpl1p and this subdomain might change during the cell cycle such that Eco1p must acetylate different targets to antagonize the interaction. Alternatively, Wpl1p blocks steps in cohesion generation that are unique to the stages of the cell cycle, for example steps unique to the presence or absence of DNA replication. Our data provide a framework for future studies to build upon and distinguish these two models.

Experimental Procedures

Yeast Strains

Yeast strains and plasmids used in this study are listed in Supplementary Table 1.

Cell Synchronization

Exponentially dividing cultures were arrested in S or G2/M using 130mM hydroxyurea (HU) (Sigma) or 15 μg/ml nocodazole (NZ) (Sigma) for 4-4.5 hours in YEP media with 3% glycerol (EMD, 30% v/v stock), 2% lactic acid (Fisher, 40% v/v stock ph 5.7) and 0.01 mg/ml adenine.

mcd1 based DSB-induced Cohesion Assay

The assay was preformed as described (Unal et al., 2007). All error bars are SD from multiple individual experiments. All Mcd1 mutants were tested for robust galactose-induced protein expression. For each 30° C experimental condition, n≥3 experiments, and at least 300 cells were counted. For each 37.5° C experimental condition, n≥3 experiments, and at least 600 cells were counted.

smc3 based DSB-induced Cohesion Assay

Cells were grown to log phase in YEP media with 3% glycerol (EMD, 30% v/v stock), 2% lactic acid (Fisher, 40% v/v stock ph 5.7) and arrested in S phase using hydroxurea at 30° C. Cells were then centrifuged and washed three times with media and released into media containing 15 μg/ml nocodazole for 1 hour 45 minutes. At this point cultures were split in half and wildtype Smc3p was induced to one half by adding galactose (to 2% final) to the media for 1 hour. Half of each flask was then shifted to 37.5° C to inactivate the thermo sensitive allele of smc3 for 45 minutes. 1.5 mls of cells were then harvested and prepared for microscopy as described (Unal et al., 2007).

Attempts for Detecting Mcd1p Acetylation

We tried to identify K84 and K210 acetylation by MALDI-MS both from in vivo samples and from samples where bacterially purfied Mcd1p was acetylated by Eco1p in vitro (collaboration with Dr. Steve Gygi). We tried digesting Mcd1p using a regular trypsin digest, a trypsin partial digest, Glu-C, Asp-N, Clu-C and Asp-N combination, and Lys-C. Unfortunately the peptide containing Mcd1-K84 eludes detection, which precludes identification of acetylation of Mcd1p at K84 by MALDI-MS either in vitro or in vivo. We were able to repeat the previously published results showing that K210 is acetylated by Eco1p in vitro (Ivanov et al., 2002). We also attempted to identify Mcd1 acetylation by western blotting with four purchased antibodies Calbiochem anti acetyl-Lysine ab (cat# ST1027), cell signaling AC-K-103 #9681S, chemicon AB3879, and Abcam anti acetyl-Lysine. We could not reproducibly detect Mcd1p acetylation by western blotting with these antibodies.

Microscopy

Fluorescence was observed using a Zeiss Axioplans 2 microscope (100X objective, NA=1.40) with a Quantix CCD camera (Photometrics).

Chromatin Immunoprecipitation

Chromatin immunoprecipitation was preformed as described (Heidinger-Pauli et al., 2008).

Chromosome Spreads

Chromosome spreads were preformed as described (Hartman et al., 2000).

Supplementary Material

Click here to view.^{(452K, pdf)}

Acknowledgments

We would like to thank Chen-ming Fan, Alex Bortvin, Dean Calahan, Vinny Guacci, Margaret Hoang, Itay Onn, Fred Tan, Lamia Wahba, and Aaron Welch for constructive comments on the manuscript; Ellen Camon, Patricia Camon, and Carol Davenport for excellent technical support; and Cynthia Wagner, Judith Yanowitz and all members of our laboratory for advice and helpful discussions. Furthermore, we would like to thank Vinny Guacci for kindly providing antibodies against Mcd1p, and Dan Gottschling for providing the pGAL:HO LEU2 integrating plasmid. D.K. is supported by grants from the Howard Hughes Medical Institutes.

Footnotes

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

References

Bellows AM, Kenna MA, Cassimeris L, Skibbens RV. Human EFO1p exhibits acetyltransferase activity and is a unique combination of linker histone and Ctf7p/Eco1p chromatid cohesion establishment domains. Nucleic Acids Res. 2003;31:6334–6343. [PMC free article] [PubMed]
Ben-Shahar TR, Heeger S, Lehane C, East P, Flynn H, Skehel M, Uhlmann F. Eco1-dependent cohesin acetylation during establishment of sister chromatid cohesion. Science. 2008;321:563–566. [PubMed]
Bernard P, Schmidt CK, Vaur S, Dheur S, Drogat J, Genier S, Ekwall K, Uhlmann F, Javerzat JP. Cell-cycle regulation of cohesin stability along fission yeast chromosomes. Embo J. 2008;27:111–121. [PubMed]
Brands A, Skibbens RV. Ctf7p/Eco1p exhibits acetyltransferase activity--but does it matter? Curr Biol. 2005;15:R50–51. [PubMed]
Dobie KW, Kennedy CD, Velasco VM, McGrath TL, Weko J, Patterson RW, Karpen GH. Identification of chromosome inheritance modifiers in Drosophila melanogaster. Genetics. 2001;157:1623–1637. [PubMed]
Game JC, Birrell GW, Brown JA, Shibata T, Baccari C, Chu AM, Williamson MS, Brown JM. Use of a genome-wide approach to identify new genes that control resistance of Saccharomyces cerevisiae to ionizing radiation. Radiat Res. 2003;160:14–24. [PubMed]
Gandhi R, Gillespie PJ, Hirano T. Human Wapl is a cohesin-binding protein that promotes sister-chromatid resolution in mitotic prophase. Curr Biol. 2006;16:2406–2417. [PMC free article] [PubMed]
Haering CH, Schoffnegger D, Nishino T, Helmhart W, Nasmyth K, Lowe J. Structure and stability of cohesin’s Smc1-kleisin interaction. Mol Cell. 2004;15:951–964. [PubMed]
Hartman T, Stead K, Koshland D, Guacci V. Pds5p is an essential chromosomal protein required for both sister chromatid cohesion and condensation in Saccharomyces cerevisiae. J Cell Biol. 2000;151:613–626. [PMC free article] [PubMed]
Heidinger-Pauli JM, Unal E, Guacci V, Koshland D. The kleisin subunit of cohesin dictates damage-induced cohesion. Mol Cell. 2008;31:47–56. [PubMed]
Ivanov D, Schleiffer A, Eisenhaber F, Mechtler K, Haering CH, Nasmyth K. Eco1 is a novel acetyltransferase that can acetylate proteins involved in cohesion. Curr Biol. 2002;12:323–328. [PubMed]
Kim ST, Xu B, Kastan MB. Involvement of the cohesin protein, Smc1, in Atm-dependent and independent responses to DNA damage. Genes Dev. 2002;16:560–570. [PubMed]
Kueng S, Hegemann B, Peters BH, Lipp JJ, Schleiffer A, Mechtler K, Peters JM. Wapl controls the dynamic association of cohesin with chromatin. Cell. 2006;127:955–967. [PubMed]
Lengronne A, McIntyre J, Katou Y, Kanoh Y, Hopfner KP, Shirahige K, Uhlmann F. Establishment of sister chromatid cohesion at the S. cerevisiae replication fork. Mol Cell. 2006;23:787–799. [PubMed]
Lo WS, Trievel RC, Rojas JR, Duggan L, Hsu JY, Allis CD, Marmorstein R, Berger SL. Phosphorylation of serine 10 in histone H3 is functionally linked in vitro and in vivo to Gcn5-mediated acetylation at lysine 14. Mol Cell. 2000;5:917–926. [PubMed]
Noble D, Kenna MA, Dix M, Skibbens RV, Unal E, Guacci V. Intersection Between the Regulators of Sister Chromatid Cohesion Establishment and Maintenance in Budding Yeast Indicates a Multi-Step Mechanism. Cell Cycle. 2006;5 [PubMed]
Onn I, Heidinger-Pauli JM, Guacci V, Unal E, Koshland DE. Sister chromatid cohesion: a simple concept with a complex reality. Annu Rev Cell Dev Biol. 2008;24:105–129. [PubMed]
Sakaguchi K, Herrera JE, Saito S, Miki T, Bustin M, Vassilev A, Anderson CW, Appella E. DNA damage activates p53 through a phosphorylation-acetylation cascade. Genes Dev. 1998;12:2831–2841. [PubMed]
Sjogren C, Nasmyth K. Sister chromatid cohesion is required for postreplicative double-strand break repair in Saccharomyces cerevisiae. Curr Biol. 2001;11:991–995. [PubMed]
Skibbens RV, Corson LB, Koshland D, Hieter P. Ctf7p is essential for sister chromatid cohesion and links mitotic chromosome structure to the DNA replication machinery. Genes Dev. 1999;13:307–319. [PubMed]
Skibbens RV, Maradeo M, Eastman L. Fork it over: the cohesion establishment factor Ctf7p and DNA replication. J Cell Sci. 2007;120:2471–2477. [PubMed]
Strom L, Karlsson C, Lindroos HB, Wedahl S, Katou Y, Shirahige K, Sjogren C. Postreplicative formation of cohesion is required for repair and induced by a single DNA break. Science. 2007;317:242–245. [PubMed]
Strom L, Lindroos HB, Shirahige K, Sjogren C. Postreplicative recruitment of cohesin to double-strand breaks is required for DNA repair. Mol Cell. 2004;16:1003–1015. [PubMed]
Sutani T, Kawaguchi T, Kanno R, Itoh T, Shirahige K. Budding Yeast Wpl1(Rad61)-Pds5 Complex Counteracts Sister Chromatid Cohesion-Establishing Reaction. Curr Biol. 2009;19 [PubMed]
Toth A, Ciosk R, Uhlmann F, Galova M, Schleiffer A, Nasmyth K. Yeast cohesin complex requires a conserved protein, Eco1p(Ctf7), to establish cohesion between sister chromatids during DNA replication. Genes Dev. 1999;13:320–333. [PubMed]
Uhlmann F, Nasmyth K. Cohesion between sister chromatids must be established during DNA replication. Curr Biol. 1998;8:1095–1101. [PubMed]
Unal E, Arbel-Eden A, Sattler U, Shroff R, Lichten M, Haber JE, Koshland D. DNA damage response pathway uses histone modification to assemble a double-strand break-specific cohesin domain. Mol Cell. 2004;16:991–1002. [PubMed]
Unal E, Heidinger-Pauli JM, Kim W, Guacci V, Onn I, Gygi SP, Koshland DE. A molecular determinant for the establishment of sister chromatid cohesion. Science. 2008;321:566–569. [PubMed]
Unal E, Heidinger-Pauli JM, Koshland D. DNA double-strand breaks trigger genome-wide sister-chromatid cohesion through Eco1 (Ctf7) Science. 2007;317:245–248. [PubMed]
Verni F, Gandhi R, Goldberg ML, Gatti M. Genetic and molecular analysis of wings apart-like (wapl), a gene controlling heterochromatin organization in Drosophila melanogaster. Genetics. 2000;154:1693–1710. [PubMed]
Williams BC, Garrett-Engele CM, Li Z, Williams EV, Rosenman ED, Goldberg ML. Two putative acetyltransferases, san and deco, are required for establishing sister chromatid cohesion in Drosophila. Curr Biol. 2003;13:2025–2036. [PubMed]
Zhang J, Shi X, Li Y, Kim BJ, Jia J, Huang Z, Yang T, Fu X, Jung SY, Wang Y, et al. Acetylation of Smc3 by Eco1 is required for S phase sister chromatid cohesion in both human and yeast. Mol Cell. 2008;31:143–151. [PubMed]