The possible function of Flp1 in homologous recombination repair in Saccharomyces cerevisiae

Saccharomyces cerevisiae Mus81 is a structure-selective endonuclease which constitutes an alternative pathway in parallel with the helicase-topoisomerase Sgs1-Top3-Rmi1 complex to resolve a number of DNA intermediates during DNA replication, repair, and homologous recombination. Previously, it was showed that the N-terminal region of Mus81 was required for its in vivo function in a redundant manner with Sgs1; mus81Δ120N mutant that lacks the first 120 amino acid residues at the N-terminus exhibited synthetic lethality in combination with the loss of SGS1. In this study, the physiologically important role of the N-terminal region of Mus81 in processing toxic intermediates was further investigated. We examined the cellular defect of sgs1Δmus81Δ100N cells and observed that although viable, the cells became very sensitive to DNA damaging agents. A single-copy suppressor screening to seek for a factor(s) that could rescue the drug sensitivity of sgs1Δmus81Δ100N cells was performed and revealed that Flp1, a site-specific recombinase 1 encoded on the 2-micron plasmid was a suppressor. Moreover, Flp1 overexpression could partially suppress the drug sensitivity of mus81Δ cells at 37 °C. Our findings suggest a possible function of Flp1 in coordination with Mus81 and Sgs1 to jointly resolve the branched-DNA structures generated in cells attempting to repair DNA damages.


Introduction
Mus81, a highly conserved DNA structure-selective endonuclease, is related to the XPF/Rad1 family of proteins involved in DNA nucleotide excision repair. Mus81 functions as a heterodimeric protein complex with a partner, namely Eme1 and Eme2 in humans, Eme1 in fission yeast, and Moreover, these physical and functional interactions are significantly important for cellular function of Mus81 in both budding yeast [26,27,30] and human [31]. Taking together all, Mus81-Mms4 is most likely to work with other proteins involved in a variety of DNA metabolisms including DNA repair, replication fork stability, and joint molecule formation/resolution during homologous recombination in order to safeguard the genome integrity.
Here, we further investigated the significance role of Mus81 N-terminus in vivo by performing a single-copy suppressor screening to seek for a factor(s) that can suppress the cellular defect caused by function loss of Mus81 N-terminal region. Unlike the mus81 Δ120N mutant allele which is synthetic lethality when combined with sgs1Δ, the sgs1Δmus81 Δ100N double mutant cells could survive, however, exhibited excessive sensitivity to DNA damaging agents. Through screening, we successfully recovered a candidate that can rescue the HU sensitivity of sgs1Δmus81 Δ100N mutant cells, namely FLP1, a site-specific recombinase 1 which is encoded on the 2-micron plasmid. Accordingly, Flp1 is the possible suppressor of cellular defect caused by the dysfunction of Sgs1 and Mus81 when cells are challenged by DNA damages.

Drop dilution assay
The plasmid pRS325, a yeast episomal vector with a LEU2 marker, harboring wild-type MUS81 or mus81 mutant alleles driven by ADH1 promoter was transformed into NJY1777. The transformants were grown on plates containing selective synthetic defined media and a single colony from each of the transformants was inoculated into liquid media (1 mL) until saturation. Cell densities were adjusted to OD 600 = 1 (~2 × 10 7 cells/mL) by diluting with dH 2 O, followed by spotting of 10-fold serial dilutions onto selective media plates containing with or without DNA damaging agents. The plates were incubated for 4 days at 30 °C. The complementation of sgs1Δmus81Δ synthetic lethality by mus81 mutant alleles was performed in the presence of 5-FOA (uracil analog, 5-fluoroorotic acid). To test the drug sensitivity of mus81 mutant alleles in the sgs1Δ null strain, cells which grew in the presence of 5-FOA were spotted onto plates containing indicated concentrations of DNA damaging agents.

Screening single-copy suppressors of sgs1Δmus81 Δ100N mutant
Yeast genomic DNA library was constructed by Sau3AI-partial digestion of genomic DNA of S. cerevisiae YPH499 strain (MATa ura3-52 lys2-801 ade2-101 trp1-Δ63 his3-Δ200 leu2-Δ1). The fragmented genomic DNA with estimately 5.6 kb in length on average was ligated into BamHI-digested pRS315 plasmid, a yeast centromere vector with a LEU2 marker. Ligation product was then transformed into Escherichia coli competent cells and the library plasmids were extracted and stored at −80 °C for long-term usage. Before screening, the yeast genomic DNA library was amplified using Plasmid Maxi Kit (QIAGEN, Hilden, Germany).
NJY1777 cell containing a plasmid harboring wild-type SGS1 gene with a URA3 marker was transformed with mus81 Δ100N gene consisted in pRS314 plasmid, a yeast centromere vector with a TRP1 marker. Transformants were grown in the selective media and transferred onto plates containing 5-FOA, producing double mutant sgs1Δmus81 Δ100N cells. The double mutant cells were then transformed with yeast genomic DNA library. Transformants were grown in selective media for 24 hours at 30 °C, followed by replica plating onto the same medium supplemented with 20 mM HU. The plates were incubated for additional 3-4 days. Selected colonies that grew on HU plates were examined for HU resistant capability by drop dilution assay. The HU-resistant colonies were transferred to liquid medium, and total plasmids were isolated. To confirm single-copy suppression, recovered plasmids were retransformed into the sgs1Δmus81 Δ100N mutant cells and examined for their ability to support cell growth in the presence of HU. Double-checked plasmids were analyzed by sequencing to identify genomic DNA fragments inserted. One of the analyzed plasmids contained the full length of FLP1 gene.

The significance of the N-terminal region of Mus81
We previously found that the sgs1Δmus81 Δ120N double mutant cell was synthetic lethality similarly to the sgs1Δmus81Δ double mutant, indicating that the deletion of the first 120 amino acids at the N-terminal region of Mus81 affected the cellular function of this enzyme ( [26]; Figure 1A). In order to further define in vivo defects associated with the loss of Mus81 N-terminus, we examined the viability of sgs1Δmus81 Δ80N and sgs1Δmus81 Δ100N mutant cells using sgs1Δmus81Δ double mutant cells that were rendered viable by the presence of a plasmid expressing wild-type Sgs1. With the control empty vector transformed, in the presence of 5-FOA the sgs1Δmus81Δ cells had lost the plasmid containing SGS1 and became lethal ( Figure 1A). Expectedly, these cells developed normally in the presence of episomal copies of wild-type MUS81 ( Figure 1A). In contrast to the synthetic lethality of sgs1Δmus81 Δ120N , we found that the presence of mus81 Δ80N and mus81 Δ100N alleles, despite of poor development, succeeded to support the growth of sgs1Δmus81Δ cells ( Figure 1A). These findings show that a region between the 100 th and 120 th amino acid residues at the N-terminus of Mus81 is crucial for the function of Mus81 in vivo which is relevant to the cell survival in the absence of Sgs1. Here, it was observed that the deficient growth of the sgs1Δmus81 Δ100N and sgs1Δmus81 Δ80N double mutants represented the "bottleneck effect" where a size of population reduced sharply due to a sudden change of environment.
Next, MMS sensitivity of the mus81 Δ80N and mus81 Δ100N single mutant cells and the sgs1Δmus81 Δ80N and sgs1Δmus81 Δ100N double mutant cells were investigated. When wild-type MUS81 was introduced into the mus81Δ deletion cells, they grew normally in the presence of MMS ( Figure 1B). Unlike drug sensitivity phenotype of mus81Δ and mus81 Δ120N [26], the mus81 Δ80N and mus81 Δ100N cells behaved like wild-type ( Figure 1B). However, the sgs1Δmus81 Δ80N and sgs1Δmus81 Δ100N double mutants were sensitive to MMS and HU treatment ( Figure 1C), indicating that the defect causing by the absence of the first 80 and 100 amino acid residues at Mus81 N-terminus, respectively, induced severe troubles to the cell, especially in dealing with DNA-lesion induced by DNA damaging agents. Overall, these results further demonstrated the critical function of Mus81 N-terminal region in homologous recombination repair pathways when cells are faced with accumulated DNA lesions caused by DNA damaging agents.

The single-copy suppressor screening to find out a factor(s) that can suppress the cellular defect causing by the dysfunction of N-terminal region of Mus81
We aim to seek for an alternative pathway that can cope with the loss of function of the important Mus81 N-terminal region. The results in Figure 1A point out that the region between the 100 th to 120 th amino acid residues at the N-terminus of Mus81 is indispensible for the cellular function of Mus81 in the absence of Sgs1 because removal of this small portion in the absence of Sgs1 led to cell death unavoidably. To perform the single-copy suppressor screening to define a suppressor of Mus81 lacking N-terminus mutant, we decided to maintain this small region due to its essential function in vivo. Thus, we chose the defective phenotype of sgs1Δmus81 Δ100N cells ( Figure 1C) to carry out the suppressor screening in order to identify a factor that can rescue this defect.
As shown in Figure 2, the yeast genomic DNA library was transformed into the sgs1Δmus81 Δ100N cells and the transformants were replicated onto plates containing 20 mM HU. Survival colonies were selected and serial-diluted spotted onto HU-plates in order to evaluate the HU resistant ability. Well growing cells were selected and the transformed plasmids were extracted. The plasmids were next transformed back into the sgs1Δmus81 Δ100N cells and transformants were re-examined for HU-resistant capability. Finally, the plasmids producing resistant transformants were sequenced to define the possible genes involving in suppression of the Mus81 Δ100N dysfunction. Collectively, after replica plating step, there were fifty-seven colonies that could grow on HU plates. Choosing those colonies and using drop dilution assay, we were able to examine the HU-resistant ability of fifty-four colonies (Figure 3). Among fifty-four checked colonies, forty-one were capable of suppress HU sensitivity of the sgs1Δmus81 Δ100N mutant. There were twenty-three strong suppressors in comparison to wild-type cells (Figure 3, Table 1). Next, plasmids from forty-one colonies were extracted and re-transformed into the sgs1Δmus81 Δ100N cells. Among forty-one candidate plasmids extracted, thirty-eight successfully created transformants. Then transformants were serial-diluted spotted onto plates containing HU to evaluate their survival. Among thirty-eight obtained transformants, only sixteen were capable of resisting to HU treatment ( Figure 4, Table 1).

Flp1, a site-specific recombinase 1 encoded on the 2-micron plasmid is a suppressor of mus81 mutant lacking the N-terminal region
Sixteen plasmids that transformants created were able to grow well in the presence of HU were sequenced to identify the genomic DNA fragments inserted. Finally, it revealed fourteen plasmids harboring SGS1 sequence including either or both of its upstream and downstream sequence and nearby region on chromosome XIII, jointly forming inserted fragment of approximately 6 kb. The presence of SGS1 in screening results served as a positive control for our suppressor screening approach. The plasmid number 31 contained an upstream sequence (around 250 base pairs) and full length of a gene called FLP1, a site-specific recombinase 1 encoded on the 2-micron plasmid which is a multi-copy selfish extrachromosomal DNA element found in the nucleus in budding yeast. The pRS315 plasmid that we used to construct genomic library is the yeast centromere vector that does not harbor the 2-micron origin of replication and is maintained at a low copy number. Moreover, even yeast episomal plasmids do not contain the full-length Flp1 sequence and its upstream region, indicating that the presence of Flp1 sequence in screening result is definitely not derived from the used plasmid. Here, it is clear that the construct of genomic DNA fragment found in plasmid number 31 could guarantee the expression of functional Flp1 due to the presence of its native promoter in upstream sequence and completed coding sequence.
The 2-micron plasmid is relatively small (approximately 6.3 kb), and resides in most common strains of S. cerevisiae at steady-state copy number of 40-60 molecules per haploid cell [33,34]. The plasmid harbors an origin of bidirectional DNA replication called ARS (autonomous replication sequence), FLP1 gene, three other genes encoding proteins required for regulation of Flp1 expression, namely REP1, REP2, and FAF1, a set of small direct repeats forming a partitioning locus named STB, and two 599-bp inverted repeat sequences called Flp Recombination Targets (FRTs) [33,[35][36][37][38]. Flp1 with a couple of FRT sites residing in the plasmid in a head-to-head orientation belong to the amplification system which is responsible for the 2-micron plasmid self-replication [38]. We questioned that the suppression of HU sensitivity by Flp1 overexpression in sgs1Δmus81Δ cells expressing Mus81 Δ100 is associated with the loss of N-terminal region of Mus81 which is equivalent to loss of Mus81 function, or the loss of Sgs1, or a combination of these defects. Thus, using the plasmid number 31, we examined the ability of Flp1 overexpression in rescuing cellular defects of the mus81Δ and sgs1Δ mutants, respectively and found that Flp1 overexpression was unable to suppress the HU sensitivity of sgs1Δ ( Figure 5). The transformation of plasmid number 31 could not suppress the mus81Δ mutant defect at temperature 30 °C, however, it succeeded to make the mutant partially resistant to HU treatment at 37 °C ( Figure 5). These findings, together with the screening results, indicate that the suppression by Flp1 is related to the loss of Mus81 function, not Sgs1. Figure 5. Transformation of plasmid number 31 can partially suppress the drug sensitivity of mus81Δ at 37 °C. NJY1777 strain was transformed with an empty vector pRS315 or vectors containing MUS81 or screened plasmid number 31. To create sgs1Δ mutant cells, NJY1777 strain was transformed with pRS314 vector containing MUS81 and transformants were grown in the presence of 5-FOA to remove pJM500-URA3-SGS1 plasmid.
Next, FLP1 gene was separately cloned into pRS315 vector and using drop dilution assay, we observed that the Flp1 overexpression could partially rescue the HU sensitivity of the sgs1Δmus81 Δ100N mutant cells ( Figure 6A). This result confirmed that Flp1 is a single-copy suppressor of the sgs1Δmus81 Δ100N mutant. Besides, overexpression of Flp1 failed to suppress the HU sensitivity of rad52Δ and rad52Δsgs1Δmus81 Δ100N cells ( Figure 6B), indicating that Flp1 does not involve in the processing of intermediates at early stage of homologous recombination repair. Also, it was unable to rescue the wild type cells at high amount of HU treatment ( Figure 6C), confirming the specific suppressive effect towards mus81 mutants which lack the full function of Mus81-Mms4 in vivo. Flp1 catalyzes a site-specific recombination reaction via interacting with DNA sequences within its FRT sites, producing a single-stranded break on both DNA strands within the target sites that results in inversion of a segment of the 2-micron plasmid [38,39]. Flp1 was determined to remain binding to the 3' side of each break by an O-phosphotyrosyl residue [40]. Accordingly, Flp1 is a very promising suppressor because all of its activities are potentially required in homologous recombination repair pathway when Sgs1 is absent and Mus81 is partially dysfunctional.

Conclusions
In mitotic cells, the homologous recombination intermediates are preferentially processed by a "dissolution" activity of the Sgs1-Top3-Rmi1 complex to merely generate non-crossover products. When cells are challenged with DNA damaging agents interfering with normal progression of replication forks or producing excessively double strand breaks, the joint work of Sgs1 and Mus81 could rapidly resolving the excess amount of DNA damages, despite of the risk of crossover product formation. Our previous study has shed light on the important role of Mus81 N-terminal region for the full function of Mus81 complex in the absence of Sgs1 during homologous recombination repair pathway [26]. Next, we raised a question for the alternative pathway that can rescue the cellular defect of mus81 mutant isolated in the study. As described in Results section, we have succeeded to recover the FLP1 gene on the 2-micron plasmid as the suppressor. The 2-micron circle independently reproduces itself with chromosome-like stability via the joined activity of a plasmid amplification system and a plasmid partitioning system. There has not been obvious evidence for the advantage or disadvantage of this plasmid existence to its host [33]. However, a very high copy numbers of the plasmid is harmful to the host, leading to cell cycle misregulation and cell lethality [34]. There is a fact that our strain cells should already harbor 2-micron plasmids but were failed to survive when treated with HU without being further transformed with an extra-plasmid containing FLP1 sequence. This observation can be explained by regulation of the endogenous 2-micron plasmid number and gene expression. In details, three proteins Rep1, Rep2, and Faf1 which belong to the partitioning system control the copy number of plasmid by repressing expression of Flp1 protein. The amount of three repressors is in proportion to the 2-micron plasmid copy number, hence, expression of Flp1 is suppressed when the plasmid copy number is high, and if the plasmid copy number is small, Flp1 expression is induced. Therefore, expression of endogenous Flp1 is not relevant to the condition that the cell has to face. However, expression of Flp1 on an extra-plasmid transformed is segregated to the number and regulation of endogenous 2-micron system. The protein is considered as to be overexpressed and can be potentially beneficial when cells are challenged by DNA damaging agents.
We already prove that the lethality of sgs1Δmus81 Δ120N was rescued by further deletion of Rad52, a key homologous recombination mediator in homologous recombination in budding yeast, indicating that the cellular defect related to dysfunction of Mus81 lacking N-terminal region was caused by the accumulation of unprocessed toxic recombination intermediates [26]. Therefore, the suppression of cellular defect of sgs1Δmus81 Δ100N by Flp1 overexpression should be derived from the ability to reduce the late recombination intermediates accumulated. This notion is further supported by the result that overexpression of Flp1 was unable to rescue the cellular defect caused by the absence of Rad52 which functions upstream of Mus81 complex. The possibility that Flp1 can participate in processing toxic intermediates arising when cells are faced with DNA lesions was also supported by the observation that the overexpression of Flp1 can partially rescue the HU sensitivity of mus81Δ mutant at non-permissive temperature (37 °C). At 37 °C, a condition that the cells grow and divide rapidly, HU treatment induces quickly high accumulation of toxic intermediates that only Sgs1 activity could not cope with efficiently, representing by the highly drug-sensitive level of the cells ( Figure 5). This excessive accumulation of recombination intermediates could lead to a condition that the cells may activate Flp1 function in mediating the removal of the toxic intermediates besides Sgs1, showing the partially resistant capability to drug ( Figure 5).
Until now, the reason of 2-micron plasmid presence inside yeast cell has not been clearly defined, hence, identifying Flp1 as the suppressor of Mus81 partial dysfunction raises the possible explanation of advantage of readily containing this plasmid inside the cells as a backup system. As discussed above, we believe that Flp1 overexpression suppressed the cellular defects related to the loss of Mus81 function caused by the loss of N-terminal region via procession of toxic recombination intermediates. Although Flp1 tetramers might bind DNA in a configuration similar from recombination repair intermediates, Flp1 is a site-specific recombinase requiring a specific DNA sequence to induce DNA cleavage and recombination. It has not been completely clear that the Flp1 overexpression suppressor effect depends on its enzymatic activities, pointing to an involvement of DNA cleavage and recombination, or its potential function is perhaps merely structural. Therefore, Flp1 should be further investigated for its possible function in resolving homologous recombination intermediates generated when cells try to repair DNA damages.