An Assay to Study Intra-Chromosomal Deletions in Yeast

An accurate DNA damage response pathway is critical for the repair of DNA double-strand breaks. Repair may occur by homologous recombination, of which many different sub-pathways have been identified. Some recombination pathways are conservative, meaning that the chromosome sequences are preserved, and others are non-conservative, leading to some alteration of the DNA sequence. We describe an in vivo genetic assay to study non-conservative intra-chromosomal deletions at regions of non-tandem direct repeats in Schizosaccharomyces pombe. This assay can be used to study both spontaneous breaks arising during DNA replication and induced double-strand breaks created with the S. cerevisiae homothallic endonuclease (HO). The preliminary genetic validation of this assay shows that spontaneous breaks require rad52+ but not rad51+, while induced breaks require both genes, in agreement with previous studies. This assay will be useful in the field of DNA damage repair for studying mechanisms of intra-chromosomal deletions.


Introduction
The inappropriate repair of DNA double-strand breaks (DSBs) can cause different forms of structural chromosomal instability such as translocations, deletions, duplications and inversions [1,2]. These aberrations are the types of genomic instability seen in cancer cell karyotypes [3].
There are two main sources of DNA double-strand breaks-endogenous and exogenous. Both sources can produce similar types of lesions and are repaired by the same mechanisms [4]. Replication stress is the greatest producer of endogenous breaks. Replication forks may stall as they run into hard-to-replicate heterochromatin structures [5,6] or during collisions with RNA polymerases [7]. Such replication stalls are rescued by the recombination machinery which has evolved precisely to deal with stress arising from the replication of long genomes [8]. Exogenous breaks are caused by environmental chemicals or radiation. DNA DSB repair occurs by two distinct genetic pathways: homologous recombination (HR) or non-homologous end joining (NHEJ) [4]. HR uses an intact template sequence to copy the missing/broken region and requires the RAD52 epistatic group (RAD52, RAD51, RAD54, RAD55/57 as well as several accessory nucleases and helicases) [9,10]. NHEJ requires Ku70/80 and Lig4, and involves the localized repair of breaks with no major sequence rearrangements [11]. In yeast, as in humans, the two repair pathways are cell-cycle regulated under the control of the cyclin-dependent 2 of 9 kinase with NHEJ, acting primarily in G1 and HR in S-phase and mitosis [12][13][14][15][16][17]. Furthermore, NHEJ antagonizes HR as Ku70/80 inhibits the HR pathway in G1 [15].
Several in vivo assays have been designed in various model systems, as well as in human cells, to study the genetic requirements for DSB repair (we only reference a few) [18][19][20][21][22][23][24][25][26]. Such studies have contributed vastly to our understanding of how chromosomal instability arises. We also previously briefly described an assay to study chromosomal instability at non-tandem direct repeats arising from spontaneous damage [27]. However, this assay was only briefly characterized and not sufficiently validated. Here, we present an improved assay that can be used to study both random and induced DSBs and describe the protocol for this method. We also provide a preliminary characterization of the genetic requirements for repair of these different types of breaks. Our results show that induced breaks are likely to be repaired by different mechanisms than spontaneous breaks, in agreement with what has been previously shown. This assay should become a valuable tool in the field of yeast genetics to study intra-chromosomal deletions.

Experimental Design, Methods and Materials
We previously reported an assay that monitors chromosomal instability [27]. In this assay, two non-functional ura4 fragments were placed on either side of a functional his3 + gene ( Figure 1A). The ura4 + fragments contain 200 bp of overlapping identical sequences (gray region), creating two non-tandem repeats (referred to from here on as the ura-his-ura cassette). Here, we improved on this assay by introducing the S. cerevisiae homothallic endonuclease (HO) restriction site next to the his3 + gene. The endonuclease restriction site is identical to that described in [28]. Primers 5 -ggaattcggccaggtacctttcagctttccgcaacagtataaagtactctgca-3 and 5 gagtactttatactgttgcggaaagctgaaaggtacctggccgaatcctgca-3 with PstI restriction site overhangs were used to amplify the HO endonuclease restriction site. The PCR was cloned into the PstI site of plasmid pRCP16 described previously [27] to generate pRCP20. The ura-his-ura cassette was released with SacI and KpnI and transformed into FY1828 to create RCP24. Unlike in our previous report where we studied recombination in the centromere heterochromatin, here we introduced the ura-his-ura cassette at the endogenous ura4 + locus (Chr.III 116575-115781). The ura4 + was replaced by the cassette. Thus, this strain is ura auxotrophic and his prototrophic. This assay can monitor both random breaks that may arise during DNA replication and HO-induced breaks. The HO endonuclease is expressed from the pREP81X-HO plasmid [29]. pREP81X was used as vector control.

5.
Count colonies on both the YES control and EMM-Uracil plates and record the numbers. Although Phloxin B allows for easier differentiation of ura4+ prototrophic colonies, to ensure that all colonies on the EMM-Uracil plates are in fact Ura + , this plate can be replica plated onto 5-FOA. All ura4+ colonies that grow on EMM-Uracil should die on 5-FOA.   Incubate all plates at 32 • C until colonies appear. 5.

Induced Break Recombination Protocol
Count colonies on both YES and EMM-Uracil plates and record the numbers.

Characterization of the Assay
Spontaneous ura4 + his3 − recombinants arise at an average frequency of approximately 1 in 10 4 cells ( Figure 1B). As expected, when the break is made by the endonuclease, the frequency is much higher (2.5 in 10 colonies) ( Figure 1B). This assay does not appear to report conversion (e.g., ura4 + his3 + colonies). PCR analysis of several recombinants with primers flanking the ura4 + ORF showed that the ura-his-ura cassette has been converted to ura4 + ( Figure 1C). Sporadically, we did find some colonies that were ura4 + his3 + , which appeared at a much lower frequency and only when we induced the break (Supplemental Figure S1A). To understand what these ura4 + his3 + colonies were, we used PCR to check the size of the locus in the HO-induced recombinant colonies (Supplementary Figure S1B). When primers flanking the ura4 + ORF are used, we found that both the ura4 + and the ura4 + his3 + are the same size-indicating that both are deletion outcomes. Next, we checked whether the ura4 + his3 + colonies arose as a result of gene conversion between the his3 + in our assay and the his3-D1 locus [30]. PCR across the his3 + locus showed that the his3-D1 deletion is present in both the ura-his-ura (pre) and the recombinant ura4 + his3 + colonies, suggesting that the his3-D1 allele has not been converted to his3 + (WT) (Supplementary Figure S1C). However, PCR with primers within the ORF his3 + detected the presence of the his3 + ORF in the ura4 + his3 + strains. We concluded from this PCR analysis and the very low frequency of the ura4 + his3 + recombinants that the his3 + must arise due to some spurious integration of the ORF elsewhere in the genome. Thus, this assay can be primarily used to test deletions.

Strains
The strains used in this study are listed in Supplementary Table S1. The construction of the ura-his-ura recombination cassette has been described previously [27]. This cassette has been slightly modified by cloning the S. cerevisiae HO endonuclease restriction site immediately upstream of the his3 + gene, as described above. The HO restriction site sequence is identical to that described in [28]. Standard yeast genetics have been used to cross the recombination mutants with the ura-his-ura cassette.

Data Analysis
For all assays, the data were adjusted for viability and error in plating using the numbers on the YES plates (# colonies EMM-Uracil/(# colonies on YES/1000). The number "1000" represents the colonies intended to be plated on the YES plate. For example, if only 900 colonies appeared on the YES plate, then 900/1000 = 0.9 efficiency of plating. This means that a 10% error was made and the division of EMM-Uracil colonies by 0.9 normalizes the numbers to 100% plating efficiency. Thus, the YES plate serves as a control for both viability and plating errors and normalizes all experiments. This normalization was also important in order to control for systematic errors that might have Methods Protoc. 2019, 2, 74 5 of 9 been introduced as different people did the experiments, and is not unlike normalizations used previously [27,31]. The resulting value was then multiplied by the dilution factor so that the results were normalized to a recombination frequency of 10 5 for spontaneous breaks and 10 4 for induced breaks. Descriptive statistics and graphs were generated using SPSS.

Analysis of Spontaneous Breaks
We next carried out some preliminary characterization of the genetic requirements for these deletions (Figure 2). We found that rad52 + is required for spontaneous breaks but rad51 + and pku70 + are not (Figure 2, Supplementary Tables S2 and S3). In fact, both rad51 + and pku70 + appeared to inhibit deletion outcomes. These results suggest that spontaneous break repair relies on a pathway that requires both rad51 + and pku70 + . rad51 + has been previously shown to suppress chromosomal rearrangements in S. pombe arising from improperly repaired spontaneous breaks [25,32]. rad51 + is also not required for deletion outcomes, but is essential for gene conversion [33]. This indicates that spontaneous break repair is initially channeled through a homologous recombination pathway that requires rad51 + , presumably by attempting to initiate a crossover. However, this pathway is not very efficient, most likely because the break occurs between direct repeats-a process which favors repair by single-strand annealing. Mechanisms of repair of spontaneous breaks by single-strand annealing that do not rely on rad51 + have been proposed in S. pombe [34]. colonies intended to be plated on the YES plate. For example, if only 900 colonies appeared on the YES plate, then 900/1000 = 0.9 efficiency of plating. This means that a 10% error was made and the division of EMM-Uracil colonies by 0.9 normalizes the numbers to 100% plating efficiency. Thus, the YES plate serves as a control for both viability and plating errors and normalizes all experiments. This normalization was also important in order to control for systematic errors that might have been introduced as different people did the experiments, and is not unlike normalizations used previously [27,31]. The resulting value was then multiplied by the dilution factor so that the results were normalized to a recombination frequency of 10 5 for spontaneous breaks and 10 4 for induced breaks. Descriptive statistics and graphs were generated using SPSS.

Analysis of Spontaneous Breaks
We next carried out some preliminary characterization of the genetic requirements for these deletions (Figure 2). We found that rad52 + is required for spontaneous breaks but rad51 + and pku70 + are not (Figure 2, Supplementary Tables S2, S3). In fact, both rad51 + and pku70 + appeared to inhibit deletion outcomes. These results suggest that spontaneous break repair relies on a pathway that requires both rad51 + and pku70 + . rad51 + has been previously shown to suppress chromosomal rearrangements in S. pombe arising from improperly repaired spontaneous breaks [25,32]. rad51 + is also not required for deletion outcomes, but is essential for gene conversion [33]. This indicates that spontaneous break repair is initially channeled through a homologous recombination pathway that requires rad51 + , presumably by attempting to initiate a crossover. However, this pathway is not very efficient, most likely because the break occurs between direct repeats-a process which favors repair by single-strand annealing. Mechanisms of repair of spontaneous breaks by single-strand annealing that do not rely on rad51 + have been proposed in S. pombe [34]. The loss of pku70 + also increases recombination outcomes arising from spontaneous breaks, indicating that pku70 + suppresses these deletions as well. Recent evidence in S. pombe shows that pku70 + controls resection at stalled replication forks [35]. The loss of pku70 + leads to an increase in the resection tract but a decrease in Rad51 binding. These two events combined suggest that pku70 + and The loss of pku70 + also increases recombination outcomes arising from spontaneous breaks, indicating that pku70 + suppresses these deletions as well. Recent evidence in S. pombe shows that pku70 + controls resection at stalled replication forks [35]. The loss of pku70 + leads to an increase in the resection tract but a decrease in Rad51 binding. These two events combined suggest that pku70 + and rad51 + work in the same pathway and in the repair of direct repeats in our assay. This repair pathway most likely occurs through single-strand annealing when either pku70 + or rad51 + is lost.
In higher eukaryotes and fission yeast, rad52 + is not essential for all forms of homologous recombination repair [36,37]. The fact that some repair still occurs in the absence of rad52 + indicates that, at a low percentage, ura4 + may be reconstituted by some other form of repair that does not rely on rad52 + . These genetic results validate that this assay reports deletions only.
Sometimes the deletion of rad52 + in S. pombe acquires a suppressor that attenuates the sensitivity of the strains to DNA-damaging drugs. However, these results are not due to the effect of such a suppressor, because all strains used here are still sensitive to methyl-methanosulfonate (MMS) (Supplementary Figure S2).

Analysis of Induced Breaks
Next, we validated the genetic requirements for induced breaks. Because the HO endonuclease is expressed from a plasmid behind the nmt1 promoter, we used the following controls: 1) a control with thiamine to check whether the promoter is leaky and 2) a vector control which should give a frequency comparable to the spontaneous frequency (Figure 3, Supplementary Table S4). We found that recombinants appeared at about three orders of magnitude higher when the break was induced. The addition of thiamine drastically reduced the recombination frequency-albeit not to the spontaneous levels, suggesting that there is some promoter leakage. The vector control gives spontaneous level recombinants.
Methods Protoc. 2019, 1, x FOR PEER REVIEW 6 of 9 rad51 + work in the same pathway and in the repair of direct repeats in our assay. This repair pathway most likely occurs through single-strand annealing when either pku70 + or rad51 + is lost.
In higher eukaryotes and fission yeast, rad52 + is not essential for all forms of homologous recombination repair [36,37]. The fact that some repair still occurs in the absence of rad52 + indicates that, at a low percentage, ura4 + may be reconstituted by some other form of repair that does not rely on rad52 + . These genetic results validate that this assay reports deletions only.
Sometimes the deletion of rad52 + in S. pombe acquires a suppressor that attenuates the sensitivity of the strains to DNA-damaging drugs. However, these results are not due to the effect of such a suppressor, because all strains used here are still sensitive to methyl-methanosulfonate (MMS) (Supplementary Figure S2).

Analysis of Induced Breaks
Next, we validated the genetic requirements for induced breaks. Because the HO endonuclease is expressed from a plasmid behind the nmt1 promoter, we used the following controls: 1) a control with thiamine to check whether the promoter is leaky and 2) a vector control which should give a frequency comparable to the spontaneous frequency ( Figure 3, Supplementary Table S4). We found that recombinants appeared at about three orders of magnitude higher when the break was induced. The addition of thiamine drastically reduced the recombination frequency-albeit not to the spontaneous levels, suggesting that there is some promoter leakage. The vector control gives spontaneous level recombinants. Both rad52 + and rad51 + are required for recombinants arising from induced breaks, suggesting that induced breaks are likely repaired through a crossover-either intrachromosomal or unequal sister chromatid exchange. As previously shown [38,39], pku70 + antagonizes recombination and, not unexpectedly, we also show that the deletion of pku70 + increases recombination outcomes. However, the function of pku70 + here is distinct from its function at spontaneous breaks, because it does not have the same phenotype as rad51 + . The small increase in the frequency of recombinants in the absence of pku70 + suggests that the cell attempts to repair some of the generated two-ended breaks through NHEJ, which does not result in a ura4 + phenotype. Non-homologous end-joining and pku70 + Both rad52 + and rad51 + are required for recombinants arising from induced breaks, suggesting that induced breaks are likely repaired through a crossover-either intrachromosomal or unequal sister chromatid exchange. As previously shown [38,39], pku70 + antagonizes recombination and, not unexpectedly, we also show that the deletion of pku70 + increases recombination outcomes. However, the function of pku70 + here is distinct from its function at spontaneous breaks, because it does not have the same phenotype as rad51 + . The small increase in the frequency of recombinants in the absence of pku70 + suggests that the cell attempts to repair some of the generated two-ended breaks through NHEJ, which does not result in a ura4 + phenotype. Non-homologous end-joining and pku70 + have been previously shown to compete with recombination [29,38,39]. Because we select for ura4 + , these cells will die. The requirement of pku70 + for the induced breaks, but not rad51 + , may be explained by the stage of the cell cycle where the breaks are repaired. We presume that spontaneous breaks are generated as a consequence of DNA replication and may be repaired in S-phase. HO-induced breaks cause a cell cycle delay due to activation of the rad3 + -dependent checkpoint [29]. It is therefore possible that induced breaks pile up in G2, while spontaneous breaks are repaired in S-phase. These results suggest that the mechanism of repair of induced breaks is distinct from that of spontaneous breaks.

Conclusions
Here we describe an assay to study intra-chromosomal deletions arising at regions of non-tandem repeats and provide preliminary data that show that induced breaks have different genetic requirements than spontaneous breaks-in agreement with what has been previously shown. We believe that this assay will be an important tool in the field of DNA damage repair to study deletions.
It is also worth noting that some cell cycle regulators, such as CDKN2A, are inactivated in cancer cells by deletion of the entire gene rather than by point mutations [40,41]. This shows that these intra-chromosomal deletions could introduce enough genetic change in human cells that may cause cancer. Since there is enough conservation in repair genes between yeast and human cells, this assay could be used to leverage the powerful yeast genetics to identify the mechanisms of these intra-chromosomal deletions.
Supplementary Materials: The following are available online at http://www.mdpi.com/2409-9279/2/3/74/s1, Figure S1: Analysis of Ura+His+ recombinants, Figure S2: Sensitivity of strains to MMS, Table S1: Strains used in this study, Table S2: Descriptive statistics for spontaneous breaks when cells were released in Edinburgh Minimal Media and platted on EMM-Uracil with Phloxin B, Table S3: Descriptive statistics for spontaneous breaks when cells were released in Edinburgh Minimal Media and platted on EMM-Uracil without Phloxin B, Table S4: Descriptive statistics for induced breaks when cells were released in Edinburgh Minimal Media.