We identified a number of proteins with a specific role in SCR that include Wss1, a SUMO or Ub-SUMO protease, and several proteins involved in chromatin remodeling, as Ahc1 (structural subunit of ADA histone acetyltransferase complex) and Hst3 or Rtt109, involved in acetylation/deacetylation of histone H3 lysine K56 (H3K56). These functions are necessary for the repair of replication-born DSBs by SCR. Mutations in the histone H3K56 residue to A, R and Q reveal that H3K56 acetylation/deacetylation is critical to promote SCR. This is the first evidence that chromatin marks can be used for the choice of repair template as a mechanism to warrant genome integrity, uncovering new functions for chromatin remodeling in genome dynamics. In addition, our study shows that Rad52 has specific residues with a key role in SCR but little or no impact on DSB repair via HR between homolog chromosomes, as deduced from the analysis of the rad52-C180A mutant.
The role for the SUMO or Ub-SUMO protease Wss1 in SCR 
is particularly intriguing, given the known relevance of SUMOylation in different DSB repair pathways in yeast and mammals 
; reviewed in 
, suggesting that SUMOylation may act at various steps and via different protein targets. Interestingly, a role for the Smc5-Smc6 complex containinig the Mms21/Nse2 SUMO ligase activity has been reported in SCR, even though the role of SUMOylation in this particular case has not been defined 
. It is also worth noting that Rad52 is indeed a target of SUMOylation that affects its DNA repair ability 
Irc4, Irc9 and Irc19 are new proteins involved in SCR, as well as Bud27 and Pdr10, two proteins involved in stress response. The biochemical function of these proteins is yet unknown and further investigation of them is required to define their role in SCR. Our study also revealed that proteins involved in chromatin remodeling, such as the Ahc1 and Ada2 subunits of the ADA histone acetyl-transferase of the SAGA complex is important for SCR 
. One of the functions of SAGA is transcriptional, in particular in transcription of RNA polII genes 
. Perhaps these results imply a possible interconnection between transcription and DNA metabolism via a transcription-dependent chromatin remodeling, which is an interesting possibility.
A major focus of this work has been on the role of histone H3K56 acetylation/deacetylation in SCR. In S. cerevisiae
acetylation of H3K56 (H3K56ac) occurs on newly synthesized histone H3 molecules by Rtt109 acetyl-transferase, facilitating their deposition onto newly replicated DNA during S phase, but disappears rapidly by the action of sirtuins Hst3 and Hst4 when cells enter G2/M 
. Their deposition also increases in response to DNA damage in S phase 
. Strains lacking an acetylatable histone H3K56 show genetic instability and sensitivity to a subset of genotoxic agents including camptothecin (CPT) 
. This phenotype is possibly due to a key role of this modification in nucleosome assembly following DNA replication and DNA repair 
. Indeed, in agreement with our results implying a function in SCR, it has been recently suggested that H3K56 acetylation in nascent chromatin is important to complete the repair of DNA lesions and/or DNA replication 
. As with other mutations affecting chromatin assembly, hyper-recombination can be explained by defective replication fork progression that would lead to DNA breaks (see 
We show that hst3
mutations specifically impair SCR. Given the redundancy of the two deacetylases, the synergistic effect of the mutations in the accumulation of Rad52 foci and the defect in SCR demonstrates that histone H3 deacetylation is critical in SCR and genome stability. Furthermore, the analysis of specific A, R and Q mutations of H3K56 that mimic either hyper-acetylation or deacetylation strengthens the notion that this mark is important for efficient SCR and for preventing genome instability. The relevance of histone H3K56 acetylation/deacetylation dynamics in genome instability has also been reported in mammalian cells for p300/CBP H3K56 acetyl-transferase and SIRT1 deacetylase 
. We propose that the histone H3K56 acetylation/deacetylation profile serves as a cell marker to favor SCR versus
other mechanisms of repair of replication-born DSBs. It is worth noting that the effect of asf1Δ
, which also impairs H3K56 acetylation 
, may be different as asf1Δ
mutants are weakly affected in SCR at the early time points of the reaction 
, likely due to its function in other processes such as the DNA damage checkpoint 
One of the known functions of histone H3K56 acetylation/deacetylation in chromatin dynamics during replication 
is that acetylated histone H3 is incorporated into newly synthesized chromatin behind the replication fork, whereas deacetylated “old” histones are ahead of the fork. Here, we propose a model, depicted in , to explain its role in favoring the choice for the sister as the preferential repair template for replication-generated DSBs. The preference for the sister for DSB repair is lost if H3K56 is deacetylated on both sides of the fork or hyper-acetylated. Deacetylated chromatin is involved in silencing and chromatin condensation 
, which may also explain the decreased efficiency of repair observed here due to limited accessibility of DNA repair proteins. It could also be that absence of H3K56 acetylation causes a defect in nucleosome assembly responsible for an impairment of SCR or negatively affects loading of cohesins, which has been shown to be required for SCR 
. Nevertheless, the fact that H3K56 acetylation causes similar effects on SCR than H3K56 deacetylation or the rad52-C180A
mutation (see below) makes rather unlikely that cohesin loading is the major cause of the SCR impairment. Therefore, the asymmetry of the acetylation state around the fork may facilitate the repair of a broken chromatid with its sister.
Model to explain how the state of acetylation/deacetylation of H3K56 influences SCR.
The intimate link of repair with chromatin modifications suggests that particular recombination proteins may have a differential capacity to interact with differently modified histones. In this sense, the existence of specific rad52
alleles, known as class C mutants, which are defective in the repair of DSBs but proficient in spontaneous recombination () 
, is particularly intriguing. In this work we demonstrate, using the rad52-C180A
allele, that this phenotype is explained by a defect in SCR ( and ). It would be interesting to see whether specific mutations in the early acting HR protein Rad52 might impair its ability to recognize different states of acetylated/deacetylated histone H3, therefore randomizing the template choice. Nevertheless, this is just one possibility as it is also plausible that a number of Rad52 residues, likely those identified in the class C alleles, play a role in favoring the sister as the main repair template choice by either facilitating interaction with some components of the sister such as particular histone residues, cohesins, etc. which would be worth investigating in the future.
Finally, our work provides genetic evidence for two HR pathways to reconstitute replication forks via SCR. We find that hst3Δ hst4Δ
mutants are lethal with rad52Δ
but not with rad51Δ
unless the Pol32 subunit of Pol
is ablated (
; ). The same is observed in rad3-102
mutants that accumulate single-strand DNA nicks that precede DSBs occurring by replication fork breakage 
. These observations support a model of two mitotic Rad52/MRX-dependent mechanisms of SCR for the repair of replication-born DSBs, one being Rad51-dependent and the other Pol32-dependent 
, even though a synergistic effect caused by a masked role of Po32 in replication cannot be discarded.
In summary, our work provides new insights into SCR as a major mechanism of repair of replication-born DNA breaks. It shows the existence of factors and specific protein residues that play a role in the choice of the sister chromatid as the DNA repair template. These functions include the state of histone H3 K56 acetylation/deacetylation or specific DSB repair proteins acting at the early steps of homologous recombination such as Rad52. Importantly, our study demonstrates that failure to repair a replication-born DSB with the sister can lead to genome instability, raising new questions about the mechanisms by which DSB repair proteins and chromatin interact to favor one DSB repair pathway versus another.