Stimulation of Chromosomal Rearrangements by Ribonucleotides | Genetics

Stimulation of Chromosomal Rearrangements by Ribonucleotides

  1. Hailey N. Conover*,†,
  2. Scott A. Lujan‡,
  3. Mary J. Chapman*,
  4. Deborah A. Cornelio*,
  5. Rabab Sharif*,
  6. Jessica S. Williams‡,
  7. Alan B. Clark‡,
  8. Francheska Camilo*,
  9. Thomas A. Kunkel‡,1 and
  10. Juan Lucas Argueso*,‡,1
  1. *Department of Environmental and Radiological Health Sciences and Institute for Genome Architecture and Function, Colorado State University, Fort Collins, Colorado 80523
  2. †Cell and Molecular Biology Graduate Program, Colorado State University, Fort Collins, Colorado 80523
  3. ‡Genome Integrity and Structural Biology Laboratory, National Institute of Environmental Health Sciences, National Institutes of Health, Research Triangle Park, North Carolina 27709
  1. 1Corresponding authors:. 493 MRB, 1618 campus delivery, Colorado State University, Fort Collins, CO 80523. E-mail: lucas.argueso{at}colostate.edu; and Genome Integrity and Structural Biology Laboratory, National Institute of Environmental Health Sciences, NIH, P.O. Box 12233, Mail Drop E3-01, Research Triangle Park, NC 27709. E-mail: kunkel{at}niehs.nih.gov

Abstract

We show by whole genome sequence analysis that loss of RNase H2 activity increases loss of heterozygosity (LOH) in Saccharomyces cerevisiae diploid strains harboring the pol2-M644G allele encoding a mutant version of DNA polymerase ε that increases ribonucleotide incorporation. This led us to analyze the effects of loss of RNase H2 on LOH and on nonallelic homologous recombination (NAHR) in mutant diploid strains with deletions of genes encoding RNase H2 subunits (rnh201Δ, rnh202Δ, and rnh203Δ), topoisomerase 1 (TOP1Δ), and/or carrying mutant alleles of DNA polymerases ε, α, and δ. We observed an ∼7-fold elevation of the LOH rate in RNase H2 mutants encoding wild-type DNA polymerases. Strains carrying the pol2-M644G allele displayed a 7-fold elevation in the LOH rate, and synergistic 23-fold elevation in combination with rnh201Δ. In comparison, strains carrying the pol2-M644L mutation that decreases ribonucleotide incorporation displayed lower LOH rates. The LOH rate was not elevated in strains carrying the pol1-L868M or pol3-L612M alleles that result in increased incorporation of ribonucleotides during DNA synthesis by polymerases α and δ, respectively. A similar trend was observed in an NAHR assay, albeit with smaller phenotypic differentials. The ribonucleotide-mediated increases in the LOH and NAHR rates were strongly dependent on TOP1. These data add to recent reports on the asymmetric mutagenicity of ribonucleotides caused by topoisomerase 1 processing of ribonucleotides incorporated during DNA replication.

  • LOH
  • NAHR
  • genome stability
  • recombination
  • ribonucleotides

THE replicative DNA polymerases of Saccharomyces cerevisiae, DNA polymerases α (Pol α), δ (Pol δ), and ε (Pol ε), frequently incorporate ribonucleotides into DNA both in vitro and during nuclear DNA replication in vivo (Nick McElhinny et al. 2010a,b; Williams and Kunkel 2014; Williams et al. 2015). These ribonucleotides are efficiently removed when RNase H2 incises the DNA backbone containing a ribonucleotide to initiate ribonucleotide excision repair (RER) (Nick McElhinny et al. 2010a; Sparks et al. 2012). When the RNH201 gene that encodes the catalytic subunit of RNase H2 (Cerritelli and Crouch 2009) is deleted, RER is defective and many unrepaired ribonucleotides remain in the genome. A subset of these unrepaired ribonucleotides can be removed when topoisomerase 1 (TOP1) incises a DNA backbone containing a ribonucleotide (Williams et al. 2013). However, TOP1 incision creates nicks with unligatable ends and elicits several RNA–DNA damage phenotypes, including slow growth, activation of the genome integrity checkpoint and altered progression through the cell cycle, sensitivity to the replication inhibitor hydroxyurea (HU), and strongly elevated rates for deletion of 2–5 bp from low-complexity DNA sequences (Nick McElhinny et al. 2010a; Clark et al. 2011; Kim et al. 2011). These effects are elicited primarily by ribonucleotides incorporated by Pol ε, but not by ribonucleotides incorporated by Pol α or Pol δ (Williams et al. 2015). Loss of RNase H2 is also associated with reduced efficiency of mismatch repair (MMR), thereby elevating the rate of single-base mutations (Ghodgaonkar et al. 2013; Lujan et al. 2013). This mutator phenotype is consistent with the hypothesis (Nick McElhinny et al. 2010a) that nicks resulting from RNase H2 incision at ribonucleotides can signal for strand discrimination during removal of DNA replication errors.

In addition to the point mutations mentioned above, larger types of genome instability have also been observed in RNase H2-defective cells. For example, in a study of gross chromosomal rearrangements (GCRs) in haploid yeast cells, RNase H2 defects alone had little effect, but GCR rates were elevated in double mutant strains lacking the noncatalytic Rnh203 subunit in combination with deletions of any of eight other genes affecting DNA metabolism (Allen-Soltero et al. 2014). An earlier GCR study reported that rnh201Δ single mutants displayed a fourfold increase in instability of a nonessential yeast artificial chromosome (YAC loss and terminal deletions) (Wahba et al. 2011). This instability may also be related to the fact that defects in the yeast Rnh202 subunit of RNase H2 increase the rate of gene conversion (also in haploids), an effect that is partially suppressed by deleting TOP1 (Aguilera and Klein 1988; Ii et al. 2011; Potenski et al. 2014). Similarly, mouse embryonic fibroblasts lacking the noncatalytic RNASEH2B subunit of RNase H2 have increased levels of micronuclei and chromosomal rearrangements (Reijns et al. 2012). The mechanisms responsible for these types of large-scale genome instability are not yet fully understood, but could involve DNA strand breaks arising during processing of unrepaired ribonucleotides incorporated during replication, processing of unresolved R-loops formed during transcription, or both.

The present study was designed to answer three questions. First, do ribonucleotides incorporated during nuclear DNA replication in RER-defective yeast strains elevate the rate of two types of large-scale genome instability in diploid cells: mitotic interhomolog allelic homologous recombination leading to LOH and nonallelic homologous recombination (NAHR) leading to chromosomal translocations and copy number alterations? Second, if so, do elevated LOH or NAHR rates depend on ribonucleotides incorporated by Pol ε, Pol α, or Pol δ? Third, do elevated LOH or NAHR rates depend on TOP1? To answer these questions, we studied a series of homozygous diploid yeast strains that vary in RNase H2 and TOP1 status and that also vary in the propensity to incorporate ribonucleotides by Pol ε (pol2-M644G, increased; pol2-M644L, decreased) or by Pols α and δ (increased in both pol1-L868M and pol3-L612M, respectively). These strains, and their parents encoding wild-type replicases, were modified to assay for LOH and NAHR. The results indicate that ribonucleotides incorporated by Pol ε cause TOP1-dependent chromosome instability, while the role of those incorporated by Pols α and δ is not as strong. The results support a model in which asymmetric processing of ribonucleotides in DNA by TOP1 causes both local and large-scale genome destabilization.

Materials and Methods

Yeast strains

The S. cerevisiae strains used for the whole genome sequencing mutation accumulation experiment were diploids descended from Δ|(-2)|-7B-YUNI300 (Pavlov et al. 2001). They were homozygous for his7-2, leu2Δ::kanMX, ura3Δ, trp1-289, ade2-1, lys2ΔGG2899-2900, and agp1::URA3. Sanger sequencing confirmed that the pol2-M644G mutation was homozygous, and rnh201Δ::hphMX was confirmed to be homozygous by PCR fragment-size analysis. The strains used in the main set of LOH assays and all NAHR assays (presented in Supporting Information, Table S1) were isogenic to the CG379 strain background (Morrison et al. 1991; Argueso et al. 2008). We also conducted a limited number of LOH assays using hybrid diploids resulting from the mating of MATa haploids isogenic to CG379 and MATα haploids isogenic with YJM789 (Wei et al. 2007). CG379 is a laboratory strain background closely related to the S288c reference yeast strain, while YJM789 is a diverged background derived from a clinical isolate. There are ∼60,000 single nucleotide polymorphisms (SNPs) between two strains’ genomes, some of which are associated with restriction fragment length polymorphisms (RFLPs) that can be used to monitor recombination between homologous chromosomes. For the isogenic and hybrid LOH assays, the CORE2 cassette containing the Kluyveromyces lactis URA3 gene, the S. cerevisiae URA3 gene, and the kanMX geneticin resistance gene was amplified from plasmid pJA40 (Zhang et al. 2013) and integrated at chromosome 7 (Chr7) downstream of the MAL13 gene (distal side), ∼20 kb from the right telomere. For the NAHR assay, we used a PCR-based approach to delete a 180-bp segment spanning the 3′ end of the open reading frame (ORF) and the immediate downstream sequence of the URA3 gene at its native position on Chr5. This ura3Δ3′ allele corresponds to a clean (marker-less) deletion between Chr5 coordinates 116,823 and 117,003 from the reference S. cerevisiae S288c genome (Cherry et al. 2012). The specific coordinates given are from the release version 64-2-1 of the S288c reference genome. In addition, a cassette containing the kanMX marker and a segment of the URA3 gene missing the promoter sequences and 34 bp from 5′ end of the ORF was integrated on Chr14, downstream of the PEX17 gene (proximal side). This nonallelic PEX17::kanMX-ura3Δ5′ (-RA3) insertion shares 622 bp of perfect homology (RA region) with the ura3Δ3′ allele described above. Both RA sequences are oriented 5′ to 3′ in the top strand of the reference genome (Watson strand), on the left arms of their respective chromosomes, and recombination between them can regenerate a fully functional URA3 gene that can be selected for by cell growth on uracil drop-out medium.

The RNH201, RNH202, RNH203, and TOP1 genes were deleted with the hphMX or natMX drug resistance cassettes using a PCR-based approach (Goldstein and McCusker 1999). The pol2-M644G and pol2-M644L alleles were integrated through a two-step allele replacement procedure (URA3 pop-in, pop-out) using plasmids described earlier (Kirchner et al. 2000). We built custom plasmids to integrate the pol3-L612M and pol1-L868M alleles, pHC1 and pHC2, respectively. We used overlapping PCR approaches to create mutant pol3-L612M and pol1-L868M fragments containing synonymous substitutions that created diagnostic TaqI restriction enzyme polymorphisms immediately adjacent to the codons encoding the Leu to Met mutations of interest in each gene. These PCR fragments were cloned into the pRS306 URA3 integrative vector (Sikorski and Hieter 1989) and Sanger sequenced to validate the constructions. pHC1 was linearized with BamHI and pHC2 was linearized with SalI prior to integration into POL3 and POL1 homozygous diploid strains, respectively. Ura+ clones containing the respective hemizygous plasmid integrations were patched onto 5-FOA medium to select for plasmid pop-outs. PCR and TaqI restriction analysis were used to screen candidate 5-FOA-resistant derivatives and identify diploids heterozygous for POL3/pol3-L612M or POL1/pol1-L868M, from which haploids carrying the respective DNA polymerase δ and α mutations were obtained by sporulation and tetrad dissection. All combinations of LOH and NAHR experimental markers and mutant backgrounds were obtained by intercrossing isogenic haploid strains followed by sporulation, tetrad dissection, and genotyping of haploid spores. The DNA sequences of all plasmids and oligonucleotide primers used in the constructions above are available upon request.

Detection of LOH by whole genome sequencing

RNH201-dependent LOH events were monitored during mutation accumulation experiments performed as previously described (Lujan et al. 2014). LOH rates are lower bound estimates calculated from the number of homozygous mutations accumulated in a particular isolate, divided by both the total number of mutations and by the number of accumulated yeast generations for that isolate. P-values comparing LOH rates between strains were calculated with a one-tailed Welch’s t-test (Welch 1947).

View this table:
Table 1 LOH events detected by whole genome sequencing of mutation accumulation lines

Quantitative recombination assays

Yeast cells were streaked to single colonies on solid YPD medium and incubated on plates at 30° for 3 days. For the LOH assays, fresh whole colonies were picked, resuspended in 1 ml of sterile distilled water, serially diluted, and plated on synthetic complete medium containing 1 g/liter of 5-FOA (selective) and YPD (permissive). For the NAHR assays, fresh single colonies were picked, inoculated in tubes containing 4 ml of liquid YPD, and incubated for 24 hr at 30° in a rotating drum. One-milliliter aliquots from these cultures were transferred to microcentrifuge tubes, pelleted, washed, and resuspended in 1 ml of sterile distilled water. Serial dilutions of cells were plated on uracil drop-out (selective) and YPD (permissive). For both assays, plates were incubated at 30°, and colonies were counted from the permissive plates after 2 days of growth and from the selective plates after 4 days. The colony counts were used to calculate recombination rates and 95% confidence intervals (Table S2) using the Lea and Coulson method of the median within the FALCOR web application (http://www.keshavsingh.org/protocols/FALCOR.html) (Lea and Coulson 1949; Hall et al. 2009). Statistical analyses of comparisons between recombination rates were performed using a two-sided nonparametric Mann–Whitney test in GraphPad Prism software. The calculated P-values for several of the most relevant pairwise recombination rate comparisons are shown in Table S3.

Molecular karyotype analysis

We analyzed the spectra of chromosomal rearrangements present in recombinant Ura+ clones derived from WT and rnh201Δ NAHR diploid strains. Independent Ura+ clones were picked from uracil drop-out plates, purified to single colonies in YPD plates, and then patched on to diagnostic plates containing uracil drop-out, 5-FOA, YPD geneticin, and YPD media. Full-length chromosomal DNA embedded in agarose was prepared and fractionated by pulse field gel electrophoresis (PFGE), and in some cases the DNA was used for comparative genome hybridization microarrays (array-CGH) using the methods and microarray design described previously (Zhang et al. 2013).

Data availability

Yeast strains and plasmids are available upon request. The whole genome sequencing data used to identify genome-wide LOH events (findings in Table 1) have been deposited at the Sequence Read Archive 10 (SRA) database under study no. SRP062900 and experiment nos. SRX1165947, SRX1165952, SRX1165955, and SRX1165953.

Results

Genome-wide ribonucleotide-dependent LOH in a pol2-M644G rnh201Δ strain

We constructed diploid yeast strains carrying the pol2-M644G allele and either RNH201 or rnh201Δ. Multiple clonal isolates were passaged on solid, complete medium, their genomes were sequenced, and mutations were identified by comparison to “zero passage” genomes for each strain (Lujan et al. 2014). In strains proficient in RNase H2, the majority of the mutations that accumulated were single base events whose allelic fractions were between 40 and 60%, as expected for a heterozygous state for new mutations. The identities, distributions in the genome, and rates of formation of these mutations in RNase H2-proficient strains have recently been described in detail (Lujan et al. 2014). In comparison, the pol2-M644G rnh201Δ strain accumulated single base mutations and 2- to 5-bp deletions at higher rates (to be described in detail elsewhere). However, what was not expected was that the allelic fraction of a substantial number of point mutations approached 100%. These LOH mutations suggested that a subset of ribonucleotide-induced mutations quickly became homozygous through interhomolog recombination. The number of these LOH events was larger in the pol2-M644G rnh201Δ strain as compared to the pol2-M644G RNH201 strain (111 vs. 14; Table 1). Given the number of generations over which these events accumulated, we conservatively estimate that a defect in RNase H2 increased the rate of LOH in the pol2-M644G strain by 3.7-fold. This result motivated the more specific genetic analyses described next.

Chromosome instability as measured by LOH

We investigated the role that ribonucleotides play in stimulating structural genomic alterations using assays specifically designed to measure allelic and nonallelic homologous recombination in proliferating yeast diploid cells. The first assay was designed to characterize the LOH effect observed in the mutation accumulation experiment described above. We used isogenic diploids (Figure 1; Table S1) that were homozygous at all positions on Chr7, except that they contained a single hemizygous copy of the CORE2 counter-selectable cassette with two diverged orthologous copies of the URA3 gene and a marker for geneticin resistance (Zhang et al. 2013) inserted near the right end of Chr7. An allelic interhomolog recombination event occurring anywhere within the 575-kb region between CEN7 and the CORE2 insertion can result in LOH that may cause a diploid to become homozygous for the distal regions of the right arm of Chr7. This strain setup allows selection for homozygosity of the homolog lacking the CORE2 insertion, therefore the recombination rates measured by this approach correspond to only half of the LOH events that occur in the region.

Figure 1
Figure 1

Loss-of-heterozygosity (LOH) assay system in isogenic diploids. Schematic representation of the karyotype of the isogenic diploid strains used in the LOH assays in Figure 2A and Table S2. A hemizygous counter-selectable CORE2 cassette was inserted near the right end of one of the Chr7 homologs. Allelic homologous recombination between the two homologs may result in homozygosity for the region lacking the CORE2 insertion. Derived clones carrying this LOH event were resistant to 5-FOA and sensitive to geneticin. Terminal boxes labeled L and R correspond to the left and right telomeres, respectively. The circle corresponds to the centromere.

We also conducted a limited number of experiments using hybrid diploids that resulted from mating between two diverged haploids (Materials and Methods; Figure S1). In this case, the genotype for two heterozygous positions along Chr7 were monitored in clones selected for the loss of the hemizygous CORE2 insertion. All examined 5-FOAR clones derived from wild-type and RNase H2-deficient hybrid diploids remained heterozygous at a position on the left arm of Chr7, and most became homozygous for a marker located on the right arm ∼11 kb proximal to the CORE2 insertion. This result indicated that the majority of 5-FOAR clones selected through this approach formed by interhomolog recombination leading to LOH, and that chromosome loss was not a frequent occurrence. In both the isogenic (Figure 1) and hybrid (Figure S1) diploid backgrounds, the LOH rates measured from diploids carrying the hemizygous CORE2 insertion were in the 10−5 to 10−4 5-FOAR/cell/division range, and all 5-FOAR clones examined became concomitantly sensitive to geneticin. When similar measurements were made from haploid strains, the rates were in the 10−8 range (data not shown). Thus, this assay provides a specific measurement for allelic interhomolog mitotic recombination that happens frequently in diploid genomes, without interference from chromosome loss, or from rare and more complex mutational mechanisms such as gross chromosomal rearrangements or clustered point mutations that simultaneously inactivate the two URA3 genes.

Ribonucleotide-dependent LOH

We analyzed isogenic diploid strains (Figure 1 and Figure 2A) lacking each of the three genes encoding subunits of the RNase H2 enzyme: rnh201Δ, rnh202Δ, and rnh203Δ. All three single mutants showed a similarly significant six- to eightfold elevation of the LOH rate relative to wild type (Table S2 and Table S3; RNH201 vs. rnh201Δ; P < 0.0001). A similar phenotypic differential was observed between the LOH rate measurements made using the hybrid diploid strain background (rate in CG379 × YJM789 RNH201 1.8 × 10−5, and 8.5 × 10−5 in CG379 × YJM789 rnh201Δ). These direct measurements were consistent with the genome-wide result presented in Table 1, confirming that a defect in RER can cause chromosomal instability in the form of LOH.

Figure 2
Figure 2

Quantitative analyses of mitotic recombination. The graphs show the recombination rates determined from the LOH (A) and NAHR (B) assays, described in Figure 1 and Figure 3, respectively. The bars correspond to the median recombination rates and the error bars represent the 95% confidence intervals. The results are grouped according to DNA polymerase genotype, and the bars are color coded according to RNAase H2 and topoisomerase 1 genotypes. The same numerical values presented graphically in A and B are reproduced in Table S2 for reader reference, also including the number of cultures assayed for each genotype. Statistical significance analyses of specific pairwise comparisons between genotypes are shown in Table S3.

The mutagenic effects of unrepaired ribonucleotides have previously been determined to be dependent on inappropriate cleavage by TOP1, including induction of 2- to 5-bp deletions, gene conversions, and chromosomal rearrangements. We found this to also be true for LOH stimulation, as the rnh201Δ TOP1Δ double mutant displayed a rate that was significantly lower than that of the rnh201Δ single (P < 0.0001). The LOH rate in the TOP1Δ single mutant was slightly lower than that of wild type, but not significantly so (P < 0.0853).

A recent study used haploid strains carrying mutant alleles of three different DNA polymerase genes that differentially alter the rate of incorporation of ribonucleotides into DNA (Williams et al. 2015). We pursued a similar approach to investigate whether the asymmetric mutagenic effect of ribonucleotides also applies to LOH stimulation. Diploids homozygous for the pol2-M644G allele that encodes a mutant version of Pol ε that increases ribonucleotide incorporation showed an 8-fold elevation in LOH (P < 0.0001). This rate was further elevated to 23-fold above wild type (P < 0.0001) when the ability to remove these extra ribonucleotides was eliminated in a pol2-M644G rnh201Δ double mutant. We also examined strains carrying the pol2-M644L allele that encodes a mutant version of Pol ε that incorporates fewer ribonucleotides. By itself this mutation did not significantly alter the LOH rate, but in combination with rnh201Δ, the LOH rate was lower than that measured for the wild-type version of Pol ε (POL2 rnh201Δ vs. pol2-M644L rnh201Δ; P = 0.0349). For both alleles of POL2, deletion of TOP1 caused significant decreases in the LOH rate. Together, these results support the existence of a strong direct relationship between the frequency of ribonucleotide incorporation by Pol ε and the rate of TOP1-dependent LOH.

Diploids homozygous for the pol1-L868M or the pol3-L612M alleles that respectively encode mutant versions of Pol α and Pol δ, which increase ribonucleotide incorporation, did not display altered rates of LOH relative to diploids with wild-type polymerases. Neither mutation increased the rate of LOH when combined with loss of RNase H2 activity. The rate of LOH in pol1-L868M rnh201Δ was similar to the rate in POL1 rnh201Δ (P = 0.1821), and surprisingly, the rate of LOH in pol3-L612M rnh201Δ was ∼50% lower than that in POL3 rnh201Δ strains (P = 0.0026). Finally, deletion of TOP1 from strains containing the pol1-L868M or the pol3-L612M alleles in combination with rnh201Δ significantly decreased the LOH rate. We conclude from these results that ribonucleotides incorporated by Pol α or Pol δ do not induce TOP1-dependent LOH as potently as those incorporated by Pol ε.

Chromosome instability as measured by NAHR

In the second assay, we investigated the formation of chromosomal translocations that result from NAHR in diploid cells. We used a classic heteroallele recombination approach in which two incomplete overlapping sequences can recombine to regenerate a functional selectable marker. In this case, a 3′ truncation of the URA3 gene designated URA− was present on Chr5 while a 5′ truncation (-RA3) was located on Chr14 (Figure 3A). Homologous recombination involving the shared central regions (RA; 622 bp, 100% sequence identity) can create a functional copy of the URA3 gene at the breakpoint of a Chr14/Chr5 translocation. This event may occur through a reciprocal mitotic crossover, or nonreciprocal mechanisms such as break-induced replication (BIR) or half crossover (HC) (Malkova and Ira 2013; Symington et al. 2014; Vasan et al. 2014). In the simple recombination scenarios presented in Figure 3, if a reciprocal crossover occurs between the RA substrates, depending on how the recombinant chromatids segregate in the following cell division, the karyotype of the resulting Ura+ cells may contain either two balanced reciprocal translocations (class 1; Figure 3B) or only the nonreciprocal Chr14/Chr5 translocation associated with a terminal deletion on the left arm of Chr14 and a terminal amplification on the left arm of Chr5 (class 2; Figure 3C). If BIR or HC were the mechanisms of recombination between the RA sequences, the same class 2 outcome would be observed. Additionally, since the kanMX marker is distal to the -RA3 sequence on Chr14, the expectation is that Ura+ clones carrying a terminal deletion of Chr14 should lose kanMX, thus becoming sensitive to geneticin. The class 1 and class 2 karyotype configurations predicted from this model were validated using PFGE and array CGH (Figure 3, B and C).

Figure 3
Figure 3

Nonallelic homologous recombination (NAHR) assay and predicted recombination outcomes. (A) Schematic representation of the karyotype of the diploid strain used in the NAHR assay. Chr5 (blue) and Chr14 (red) are drawn to approximate scale. Terminal boxes labeled L and R correspond to the left and right telomeres, respectively, and the numbered circles correspond to the centromeres. Only one of the Chr14 homologs contains a Kan-RA3 insertion, and only one of the Chr5 homologs contains a URA sequence. Both recombination substrates are present in Watson orientation and no other URA3 sequences are present in the genome. Recombination between the RA sequences can regenerate a full-length functional copy of URA3, selectable on uracil drop-out medium. (B) The class 1 reciprocal crossover outcome is shown to the left, with the respective sizes and structure of the associated chromosomal rearrangements. The array-CGH plots for Chr5 and Chr14 and PFGE from a representative class 1 clone are shown. The array-CGH plots y-axis corresponds to copy number [Log2 (Cy5-labeled Ura+ clone DNA/Cy3-labeled parental DNA)]. The x-axis corresponds to the probe coordinates along the respective chromosomes. The white circles indicate the positions of CEN5 and CEN14. The gray dots indicate the Log2 Cy5/Cy3 values and chromosome position of each array probe. The copy number profile of class 1 clones was fully balanced, with no gains or losses relative to the parental diploid. The PFGE was cropped for emphasis, showing only the region from Chr8 (540 kb) to Chr2 (815 kb), with lane quantification traces flanking the image. Chr5 and Chr14 trace peaks are shaded in blue and red, respectively. (C) The class 2 nonreciprocal karyotype outcome is shown to the right. The array-CGH plots for a representative class 2 clone show a deletion (one copy; pink shaded) on the left arm of Chr14 from TEL14L to PEX17, and an amplification (three copies; purple shaded) on the left arm of Chr5 from TEL05L to URA3. The class 2 PFGE profile and its quantitative analysis are also shown. Note that the WT diploid strain used as reference in the PFGEs in B and C, and in Figure S2, has a slightly longer Chr5 band because it is homozygous for the ura3-52 allele (Ty1 insertion) rather than the ura3Δ3′ and ura3Δ0 alleles present in the NAHR parent diploid strain. The overall difference in Chr5 sizes caused by the Ty1 insertion and ura3 deletions is ∼7 kb.

To characterize this NAHR assay system, we analyzed the karyotypes and the presence of the kanMX marker in 67 independent Ura+ clones derived from RNH201 and rnh201Δ diploids (Table 2). We analyzed the PFGE profiles of these clones to determine the number of copies of the parental-sized Chr5 and Chr14, the presence of the 650-kb Chr14/Chr5 translocation with URA3 at the breakpoint, and the reciprocal 700-kb Chr5/Chr14 translocation with kanMX-RA at the breakpoint. In addition, we inspected the PFGE karyotypes for the presence of any other chromosomal rearrangements not predicted by the model in Figure 3. The results showed that the reciprocal crossover outcome (class 1) was infrequent, having only one example detected from each genotype. This was expected, since the reciprocal crossover mechanism requires the spontaneous initiating double-strand break (DSB) lesion to occur within the relatively small RA region of the recipient chromosome. In contrast, nonreciprocal outcomes can be initiated by DSBs within RA or anywhere in the 245-kb distal region of Chr14. This higher probability for the formation of spontaneous initiating lesions was reflected in the higher abundance of nonreciprocal recombination outcomes (classes 2–7). Interestingly, while class 2 clones were the most abundant category recovered, other configurations were also detected at substantial frequency, particularly classes 3 and 4 (Figure S2). Class 3 clones had karyotypes and copy number profiles that were indistinguishable from those of class 2, with the exception that they were resistant to geneticin. Because class 3 clones contained only one parental-sized copy of Chr14 and retained the kanMX marker, they must somehow have lost the copy of Chr14 that did not contain the RA recombination substrate. Class 4 was also resistant to geneticin, but contained two parental-sized copies of Chr14. Finally, the relatively rare nonreciprocal classes 5, 6, and 7 involved loss of one of the parental copies of Chr5. Although these alternative events were quite interesting, the detailed characterization of the recombination mechanisms leading to their formation was beyond the scope of this project. Regardless of the final karyotype configurations of the Ura+ clones, all of them resulted from homologous recombination between nonallelic RA repeats leading to gross chromosomal rearrangements.

View this table:
Table 2 Classes of Ura+ NAHR clones

The overall distribution of Ura+ clones in the various karyotype classes was very similar between RNH201 and rnh201Δ. However, we detected six cases in which rnh201Δ-derived clones each carried one additional chromosomal rearrangement band with size different from those predicted for the URA−/−RA3 translocations between Chr14 and Chr5 (Table 2). No such other chromosomes were detected among the wild-type-derived Ura+ clones. While we did not characterize the structures of these other rearrangements, and the numbers of clones analyzed were relatively small, their presence exclusively in the rnh201Δ background suggested a higher occurrence of complex genome rearrangements in diploids defective for RNase H2.

Ribonucleotide-dependent NAHR

The experiments described above showed that the recombinogenic effect of RNA–DNA damage, first demonstrated for intramolecular gene conversion (Aguilera and Klein 1988; Ii et al. 2011; Potenski et al. 2014), also extend to allelic interhomolog recombination leading to LOH. In the absence of RNase H2, these recombination events are presumably initiated by DNA breaks that accumulate following processing of ribonucleotides or R-loops by pathways other than RER. Therefore, it is to be expected that the same lesions may also increase the formation of more complex outcomes such as gross chromosomal rearrangements. This hypothesis was tested previously using the original version of the haploid URA3-CAN1 GCR assay (Allen-Soltero et al. 2014), and a YAC stability assay (Wahba et al. 2011). Importantly, the majority of the GCR events detected with these two specific strain setups were formed through mechanisms other than homologous recombination, as there were no significant proximal homologous sequence substrates in the regions assayed (Chen and Kolodner 1999). Allen-Soltero et al. (2014). showed that single mutants lacking RNase H2 did not alter the GCR rate, but synergistic stimulation was observed in double mutant combinations with specific suppressors of chromosomal instability. Interestingly, deletion of RAD51 partly rescued this phenotype, suggesting that a substantial fraction of the GCR events that are increased in RNase H2 mutants form through the homologous recombination pathway, possibly involving nonallelic repeats. We specifically investigated this possibility using the diploid NAHR assay described above, which was designed to detect chromosomal translocations formed by recombination between homologous substrates present on Chr5 and Chr14 (Figure 3A).

Our quantitative analyses of chromosomal instability (Figure 2, Table S2, and Table S3) showed that Chr14/Chr5 NAHR events were quite rare compared to Chr7 LOH events (the baseline rate of NAHR was two to three orders of magnitude lower than LOH). Only minor, not significant, stimulation of the NAHR rate was detected in single mutants lacking the RNase H2 catalytic subunit (1.3-fold up in rnh201Δ vs. RNH201; P = 0.2817). A similarly insignificant increase was observed in rnh202Δ, and no alteration at all in rnh203Δ. These small (or no) rate differentials in the single mutants, within a context of rare mutational events, indicated that the NAHR rate measurements were more susceptible than LOH to interference from other confounding pleiotropic effects of the various DNA polymerase genotypes tested, such as differences in cell growth kinetics, mutator phenotypes, roles in replicative repair, or others. Specifically, single mutants carrying either the pol1-L868M or pol2-M644G, which result in higher incorporation of ribonucleotides by Pol α and Pol ε, respectively, showed mild, yet significant decreases in the NAHR rate (each ∼40% reduction relative to wild type, each with P < 0.0001). Despite this complication, the results obtained when we combined RNase H2 deficiency with mutant replicases and TOP1Δ showed a trend analogous to that observed in the LOH experiments, and suggested that ribonucleotides incorporated into DNA also contribute to NAHR.

Within each of the four mutant replicase genetic backgrounds tested, the largest NAHR rate increase between RNH201 and rnh201Δ was observed in strains carrying the pol2-M644G allele encoding a mutant version of Pol ε that increases the incorporation of ribonucleotides (5.4-fold NAHR rate elevation in pol2-M644G rnh201Δ vs. pol2-M644G RNH201; P < 0.0001). pol2-M644G strains also displayed the largest TOP1-dependent NAHR rate reduction (9.4-fold decrease in pol2-M644G rnh201Δ TOP1Δ vs. pol2-M644G rnh201Δ TOP1; P < 0.0001). pol2-M644L strains that have lower Pol ε ribonucleotide incorporation did not significantly increase the NAHR rate in combination with rnh201Δ (only 1.6-fold elevation in pol2-M644L rnh201Δ vs. pol2-M644L RNH201; P = 0.0871). These results were consistent with our observations for LOH and showed that the frequency of ribonucleotide incorporation by Pol ε is an important contributor to the formation of chromosomal rearrangements by NAHR.

The NAHR results obtained within the other two mutant replicase backgrounds that increase ribonucleotide incorporation were not as straightforward as in the LOH experiments. Deficiency of RNase H2 in combination with either pol1-L868M or pol3-L612M resulted in ∼2.3-fold increases in the NAHR rate. These increases were not as robust as the 5.4-fold effect observed for pol2-M644G, but they were both significant (pol1-L868M rnh201Δ vs. pol1-L868M RNH201 and pol3-L612M rnh201Δ vs. pol3-L612M RNH201; each had P < 0.0001). In both cases, the increases in the NAHR rate were dependent on TOP1. A notable result was the significant ∼60% increase in the NAHR rate in pol3-L612M rnh201Δ compared to POL3 rnh201Δ (P < 0.0001), a recombination rate change of approximately the same magnitude but opposite direction as observed in the LOH assay (∼50% decrease; Figure 2A). One possible reason for these contrasting results might be related to the fact that Pol δ is known to participate in the BIR mechanism (Symington et al. 2014), which we showed accounts for the majority of the NAHR events detected in our assay (Table 2). In this scenario, the pol3-L612M mutation might somehow make Pol δ more efficient at initiating or sustaining BIR. This would promote the nonreciprocal recombination mechanism associated with the NAHR assay, and conversely, might disfavor the reciprocal interhomolog mitotic crossover pathway most often associated with LOH.

The results obtained for NAHR with the mutant alleles of POL1 and POL3 can be interpreted as a sign that ribonucleotides incorporated by these polymerases have a larger contribution to structural chromosomal rearrangements than they do to allelic interhomolog recombination (i.e., LOH). However, even in this scenario, the ribonucleotides incorporated by Pol ε remain as the ones with the most destabilizing effect in both recombination assays used in this study. Alternatively, the inconsistencies in pol1-L868M and pol3-L612M behavior between our two assays might be primarily the result of interference in NAHR rate measurements caused by confounding phenotypes associated with these mutations, and not by their higher inherent rate of ribonucleotide incorporation.

Discussion

A number of recent studies have shown that ribonucleotides are incorporated throughout the genome during DNA replication (reviewed in Jinks-Robertson and Klein 2015). Incorrect removal of such abundant ribonucleotides can lead to genomic instability in the form of point mutations, short deletions within low complexity regions, and chromosomal rearrangements. Interestingly, the mutagenic effect of ribonucleotides appears to be asymmetric, according to which DNA polymerase was responsible for their incorporation. Specifically, TOP1-dependent 2- to 5-bp deletions in an RNase H2 mutant background were shown to be the product of ribonucleotides incorporated by Pol ε, but not by Pol α and Pol δ (Williams et al. 2015).

Earlier reports described increased chromosome-scale instability in RNase H2 mutant haploid yeast cells, including TOP1-dependent gene conversion (Potenski et al. 2014) and gross chromosomal rearrangements (Allen-Soltero et al. 2014). Another study measured the rate of LOH at Chr3 in RNase H homozygous mutant diploids (Wahba et al. 2011). Although Wahba et al. (2011). did not observe significant change in LOH in rnh1Δ or rnh201Δ single mutants, they did report an eightfold elevation in rnh1Δ rnh201Δ double mutants lacking both RNase H1 and H2 activities. Finally, a study complementary to our own has been reported recently using hybrid yeast diploids to quantify LOH stimulation genome-wide and characterize the associated recombination tracts in RNase H defective cells. These experiments showed elevated LOH in rnh201Δ and rnh1Δ rnh201Δ, but not in rnh1Δ single mutants (O’Connell et al. 2015). Neither of the diploid studies described above investigated the TOP1 dependency of the reported increases in LOH.

In this study, we show that genome-wide LOH occurs more frequently in diploids lacking RNase H2 activity, and that the rate of LOH on Chr7 is elevated in mutants carrying deletion of RNase H2 subunits. The increased LOH phenotype in rnh201Δ was almost entirely suppressed by deletion of the TOP1 gene. Further, we showed that the rate of LOH is modulated by increasing or decreasing the number of ribonucleotides incorporated by mutant alleles of Pol ε, and this effect was also dependent on TOP1. In contrast, alleles of Pol α or Pol δ that increase ribonucleotide incorporation did not alter the rate of allelic interhomolog recombination.

We also observed a role for ribonucleotides incorporated by Pol ε in stimulating chromosomal translocations formed by NAHR. Although the phenotypic differential provided by this assay system is small, the highest rate of NAHR was measured in the rnh201Δ pol2-M644G strain, and chromosomal size polymorphisms other than the selected translocations were only detected in the rnh201Δ background. Together, these observations were consistent with earlier reports for a role of ribonucleotides in the generation of gross chromosomal rearrangements in yeast (Wahba et al. 2011; Allen-Soltero et al. 2014) and increased cytogenetic abnormalities in mammalian cells (Reijns et al. 2012).

The recombinogenic effects associated with RNase H2 mutants may result from misprocessing of scattered ribonucleotides incorporated into DNA, misprocessing of R-loops, or a combination of these two defects. The observation that increased and decreased ribonucleotide incorporation by Pol ε correlates with the LOH rate leads us to propose that much of the LOH observed here in RNase H2 mutants is triggered by ribonucleotides incorporated by Pol ε during leading strand replication. This idea is consistent with our earlier proposal that ribonucleotides incorporated by Pol ε, but not (or less so) by Pols α and δ, result in 2- to 5-bp deletions that result from incisions by TOP1 (Williams et al. 2015). Thus, it appears that after Pol ε incorporates a ribonucleotide, topoisomerase 1 cleavage of the DNA backbone at a ribonucleotide provides opportunities for multiple types of genome instability, including loss of a short repeat, LOH and NAHR. Possible mechanistic reasons for this observed asymmetry have been proposed (figure 5 in Williams et al. 2015) and are currently being tested. The first centers on the fact that the RER-defective pol2-M644G strain has a higher density of ribonucleotides in the nascent leading strand than is observed in the nascent lagging strand for the pol1-L868M and pol3-L612M mutator strains lacking RNH201, suggesting that ribonucleotide density must exceed a certain threshold to observe TOP1-dependent recombination events. Alternatively, there may be additional pathways involved in removal of nascent lagging strand ribonucleotides that are either specific for this DNA strand or less available for ribonucleotide removal from the nascent leading strand. Finally, we propose that negative supercoils may accumulate in the continuous nascent leading strand in the wake of the replisome and TOP1, through its interaction with the CMG helicase (Gambus et al. 2006), would be in an ideal physical position to relieve this torsional stress. Such helical tension may not build up in the discontinuous nascent lagging strand because of the presence of DNA ends that could allow rotation. As with more localized forms of RNA–DNA damage, the asymmetry of TOP1-dependent recombination events support the concept that failure to remove ribonucleotides incorporated into DNA by Pol ε during leading strand synthesis puts genome stability at risk.

Acknowledgments

We thank Clinton Orebaugh and Marta Garbacz for helpful comments on the manuscript; Adam Burkholder and the National Institute of Environmental Health Sciences (NIEHS) Integrative Bioinformatics Support Group for help with whole genome mutation calling; Matheus A. Pichorim for offering technical support during the recombination rate assays; and Karen O’Connell, Sue Jinks-Robertson, and Tom Petes for helpful discussions and for sharing unpublished data. We are also thankful for the valuable feedback provided by the associate editor and anonymous reviewers. This work was supported by project Z01 ES065070 to T.A.K. from the Division of Intramural Research of the National Institutes of Health, NIEHS, and by a Webb-Waring biomedical research award from the Boettcher Foundation (to J.L.A.). D.A.C. received a postdoctoral fellowship from CNPq/Brazil. F.C. was supported by the National Science Foundation Research Experience for Undergraduates program at Colorado State University. The authors have no conflict of interest.

Note added in proof: See O’Connell et al. 2015 (pp. 963–975) in this issue for a related work.

Footnotes

  • Received August 5, 2015.
  • Accepted September 12, 2015.

Literature Cited

    1. Aguilera A.,
    2. Klein H. L.
    , 1988 Genetic control of intrachromosomal recombination in Saccharomyces cerevisiae. I. Isolation and genetic characterization of hyper-recombination mutations. Genetics 119: 779–790.
    1. Allen-Soltero S.,
    2. Martinez S. L.,
    3. Putnam C. D.,
    4. Kolodner R. D.
    , 2014 A Saccharomyces cerevisiae RNase H2 interaction network functions to suppress genome instability. Mol. Cell. Biol. 34: 1521–1534.
    1. Argueso J. L.,
    2. Westmoreland J.,
    3. Mieczkowski P. A.,
    4. Gawel M.,
    5. Petes T. D.,
    6. et al.
    , 2008 Double-strand breaks associated with repetitive DNA can reshape the genome. Proc. Natl. Acad. Sci. USA 105: 11845–11850.
    1. Cerritelli S. M.,
    2. Crouch R. J.
    , 2009 Ribonuclease H: the enzymes in eukaryotes. FEBS J. 276: 1494–1505.
    1. Chen C.,
    2. Kolodner R. D.
    , 1999 Gross chromosomal rearrangements in Saccharomyces cerevisiae replication and recombination defective mutants. Nat. Genet. 23: 81–85.
    1. Cherry J. M.,
    2. Hong E. L.,
    3. Amundsen C.,
    4. Balakrishnan R.,
    5. Binkley G.,
    6. et al.
    , 2012 Saccharomyces Genome Database: the genomics resource of budding yeast. Nucleic Acids Res. 40: D700–D705.
    1. Clark A. B.,
    2. Lujan S. A.,
    3. Kissling G. E.,
    4. Kunkel T. A.
    , 2011 Mismatch repair-independent tandem repeat sequence instability resulting from ribonucleotide incorporation by DNA polymerase epsilon. DNA Repair (Amst.) 10: 476–482.
    1. Gambus A.,
    2. Jones R. C.,
    3. Sanchez-Diaz A.,
    4. Kanemaki M.,
    5. van Deursen F.,
    6. et al.
    , 2006 GINS maintains association of Cdc45 with MCM in replisome progression complexes at eukaryotic DNA replication forks. Nat. Cell Biol. 8: 358–366.
    1. Ghodgaonkar M. M.,
    2. Lazzaro F.,
    3. Olivera-Pimentel M.,
    4. Artola-Boran M.,
    5. Cejka P.,
    6. et al.
    , 2013 Ribonucleotides misincorporated into DNA act as strand-discrimination signals in eukaryotic mismatch repair. Mol. Cell 50: 323–332.
    1. Goldstein A. L.,
    2. McCusker J. H.
    , 1999 Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15: 1541–1553.
    1. Hall B. M.,
    2. Ma C. X.,
    3. Liang P.,
    4. Singh K. K.
    , 2009 Fluctuation analysis CalculatOR: a web tool for the determination of mutation rate using Luria-Delbruck fluctuation analysis. Bioinformatics 25: 1564–1565.
    1. Ii M.,
    2. Ii T.,
    3. Mironova L. I.,
    4. Brill S. J.
    , 2011 Epistasis analysis between homologous recombination genes in Saccharomyces cerevisiae identifies multiple repair pathways for Sgs1, Mus81-Mms4 and RNase H2. Mutat. Res. 714: 33–43.
    1. Jinks-Robertson S.,
    2. Klein H. L.
    , 2015 Ribonucleotides in DNA: hidden in plain sight. Nat. Struct. Mol. Biol. 22: 176–178.
    1. Kim N.,
    2. Huang S. N.,
    3. Williams J. S.,
    4. Li Y. C.,
    5. Clark A. B.,
    6. et al.
    , 2011 Mutagenic processing of ribonucleotides in DNA by yeast topoisomerase I. Science 332: 1561–1564.
    1. Kirchner J. M.,
    2. Tran H.,
    3. Resnick M. A.
    , 2000 A DNA polymerase epsilon mutant that specifically causes +1 frameshift mutations within homonucleotide runs in yeast. Genetics 155: 1623–1632.
    1. Lea D. E.,
    2. Coulson C. A.
    , 1949 The distribution of the numbers of mutants in bacterial populations. J. Genet. 28: 491–511.
    1. Lujan S. A.,
    2. Williams J. S.,
    3. Clausen A. R.,
    4. Clark A. B.,
    5. Kunkel T. A.
    , 2013 Ribonucleotides are signals for mismatch repair of leading-strand replication errors. Mol. Cell 50: 437–443.
    1. Lujan S. A.,
    2. Clausen A. R.,
    3. Clark A. B.,
    4. MacAlpine H. K.,
    5. MacAlpine D. M.,
    6. et al.
    , 2014 Heterogeneous polymerase fidelity and mismatch repair bias genome variation and composition. Genome Res. 24: 1751–1764.
    1. Malkova A.,
    2. Ira G.
    , 2013 Break-induced replication: functions and molecular mechanism. Curr. Opin. Genet. Dev. 23: 271–279.
    1. Morrison A.,
    2. Bell J. B.,
    3. Kunkel T. A.,
    4. Sugino A.
    , 1991 Eukaryotic DNA polymerase amino acid sequence required for 3′-5′ exonuclease activity. Proc. Natl. Acad. Sci. USA 88: 9473–9477.
    1. Nick McElhinny S. A.,
    2. Kumar D.,
    3. Clark A. B.,
    4. Watt D. L.,
    5. Watts B. E.,
    6. et al.
    , 2010a Genome instability due to ribonucleotide incorporation into DNA. Nat. Chem. Biol. 6: 774–781.
    1. Nick McElhinny S. A.,
    2. Watts B. E.,
    3. Kumar D.,
    4. Watt D. L.,
    5. Lundstrom E. B.,
    6. et al.
    , 2010b Abundant ribonucleotide incorporation into DNA by yeast replicative polymerases. Proc. Natl. Acad. Sci. USA 107: 4949–4954.
  23. O’Connell, K., S. Jinks-Robertson, and T. D. Petes, 2015 Elevated genome-wide instability in yeast mutants lacking RNase H activity. Genetics 201: 959–971.
    1. Pavlov Y. I.,
    2. Nguyen D.,
    3. Kunkel T. A.
    , 2001 Mutator effects of overproducing DNA polymerase eta (Rad30) and its catalytically inactive variant in yeast. Mutat. Res. 478: 129–139.
    1. Potenski C. J.,
    2. Niu H.,
    3. Sung P.,
    4. Klein H. L.
    , 2014 Avoidance of ribonucleotide-induced mutations by RNase H2 and Srs2-Exo1 mechanisms. Nature 511: 251–254.
    1. Reijns M. A.,
    2. Rabe B.,
    3. Rigby R. E.,
    4. Mill P.,
    5. Astell K. R.,
    6. et al.
    , 2012 Enzymatic removal of ribonucleotides from DNA is essential for mammalian genome integrity and development. Cell 149: 1008–1022.
    1. Sikorski R. S.,
    2. Hieter P.
    , 1989 A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122: 19–27.
    1. Sparks J. L.,
    2. Chon H.,
    3. Cerritelli S. M.,
    4. Kunkel T. A.,
    5. Johansson E.,
    6. et al.
    , 2012 RNase H2-initiated ribonucleotide excision repair. Mol. Cell 47: 980–986.
    1. Symington L. S.,
    2. Rothstein R.,
    3. Lisby M.
    , 2014 Mechanisms and regulation of mitotic recombination in Saccharomyces cerevisiae. Genetics 198: 795–835.
    1. Vasan S.,
    2. Deem A.,
    3. Ramakrishnan S.,
    4. Argueso J. L.,
    5. Malkova A.
    , 2014 Cascades of genetic instability resulting from compromised break-induced replication. PLoS Genet. 10: e1004119.
    1. Wahba L.,
    2. Amon J. D.,
    3. Koshland D.,
    4. Vuica-Ross M.
    , 2011 RNase H and multiple RNA biogenesis factors cooperate to prevent RNA:DNA hybrids from generating genome instability. Mol. Cell 44: 978–988.
    1. Wei W.,
    2. McCusker J. H.,
    3. Hyman R. W.,
    4. Jones T.,
    5. Ning Y.,
    6. et al.
    , 2007 Genome sequencing and comparative analysis of Saccharomyces cerevisiae strain YJM789. Proc. Natl. Acad. Sci. USA 104: 12825–12830.
    1. Welch B. L.
    , 1947 The generalisation of student’s problems when several different population variances are involved. Biometrika 34: 28–35.
    1. Williams J. S.,
    2. Kunkel T. A.
    , 2014 Ribonucleotides in DNA: origins, repair and consequences. DNA Repair (Amst.) 19: 27–37.
    1. Williams J. S.,
    2. Smith D. J.,
    3. Marjavaara L.,
    4. Lujan S. A.,
    5. Chabes A.,
    6. et al.
    , 2013 Topoisomerase 1-mediated removal of ribonucleotides from nascent leading-strand DNA. Mol. Cell 49: 1010–1015.
    1. Williams J. S.,
    2. Clausen A. R.,
    3. Lujan S. A.,
    4. Marjavaara L.,
    5. Clark A. B.,
    6. et al.
    , 2015 Evidence that processing of ribonucleotides in DNA by topoisomerase 1 is leading-strand specific. Nat. Struct. Mol. Biol. 22: 291–297.
    1. Zhang H.,
    2. Zeidler A. F.,
    3. Song W.,
    4. Puccia C. M.,
    5. Malc E.,
    6. et al.
    , 2013 Gene copy-number variation in haploid and diploid strains of the yeast Saccharomyces cerevisiae. Genetics 193: 785–801.