Global analysis of genomic instability caused by DNA replication stress in Saccharomyces cerevisiae

Abstract

DNA replication stress (DRS)-induced genomic instability is an important factor driving cancer development. To understand the mechanisms of DRS-associated genomic instability, we measured the rates of genomic alterations throughout the genome in a yeast strain with lowered expression of the replicative DNA polymerase δ. By a genetic test, we showed that most recombinogenic DNA lesions were introduced during S or G₂ phase, presumably as a consequence of broken replication forks. We observed a high rate of chromosome loss, likely reflecting a reduced capacity of the low-polymerase strains to repair double-stranded DNA breaks (DSBs). We also observed a high frequency of deletion events within tandemly repeated genes such as the ribosomal RNA genes. By whole-genome sequencing, we found that low levels of DNA polymerase δ elevated mutation rates, both single-base mutations and small insertions/deletions. Finally, we showed that cells with low levels of DNA polymerase δ tended to accumulate small promoter mutations that increased the expression of this polymerase. These deletions conferred a selective growth advantage to cells, demonstrating that DRS can be one factor driving phenotypic evolution.

Keywords: DNA polymerase; DNA replication stress; genome instability.

Fig. 1.

Examples of LOH and other genomic rearrangements detected by microarray. (A) Microarray analysis of a gene conversion event on chromosome X in isolate DZ12-3. The y axis shows the normalized hybridization ratio between W303-1A–specific SNPs (red) and YJM789-specific SNPs (blue); the x axis shows SGD coordinates numbered from the left telomere. A moving window that includes the hybridization values of nine SNPs is shown. On the right side of the figure, the chromatids are shown as horizontal lines connected by ovals (centromeres). The daughter cell with the chromosomes represented by the microarray is outlined. (B) Microarray analysis of a crossover (or break-induced replication event) on chromosome XI in isolate DZ12-2. (C) A heterozygous deletion on chromosome IX in isolate DZ12-7. Because the break points of the deletion occur at repetitive δ elements (indicated as arrows on the right side of the figure), one possible mechanism of deletion formation is unequal crossing over. (D) Microarray analysis of an isolate monosomic for chromosome XIV (DZ12-2).

Fig. S1.

Effects of lowered expression of POL3 on cell cycle and growth rate. (A) Cell cycle progression of synchronized DZ12 cells in high-galactose– [0.05% galactose, 3% (grams/100 milliliters) raffinose; high gal] or low-galactose–containing [0.005% galactose, 3% (grams/100 milliliters) raffinose; low gal] media. Yeast cells were grown to midlog phase in high-gal and low-gal media, respectively. The cells were then arrested in G₁ phase using α-factor. The α-factor was removed after 2.5 h and the cells were incubated in media containing either high or low galactose. A total of 5 × 10⁶ cells was collected and stained with 10 μg/mL propidium iodide at each time point. The cell cycle progression of the cells was monitored by flow cytometry machine (BD FACSCanto System). The blue and red lines indicate the location of cells with G₁ and G₂ DNA content, respectively. (B) Growth of wild-type strains (haploids JSC54-1 and JSC20-1; diploid DZ3) and GAL-POL3 strains (haploids DZ1 and DZ2; diploid DZ12) on high-gal, low-gal, and YPD plates. Yeast cells were precultured in 10 mL of liquid high-gal medium overnight. Cells were then collected and serially diluted with sterile water at an initial density of OD₆₀₀ of 1. Cell suspensions (4 μL) were spotted on plates and then incubated at 30 °C for 1–3 d. Experiments were done three times.

Fig. 2.

Chromosomal alterations observed in 35 DZ12 derivatives. (A) The distribution of LOH and duplication/deletion events across yeast genome. CO/BIR, CON, TER DEL/DUP, and INT DUP/DEL represent crossover/BIR, gene conversion, terminal deletions/duplications, and interstitial duplications/deletions, respectively. Centromeres are shown as black ovals, and ORFs are shown as short vertical lines. (B) Two-way hierarchical clustering analysis of aneuploidy events. The red, orange, blue, and yellow colors indicate trisomy, uniparental disomy, monosomy, and no changes from euploidy, respectively.

Fig. S2.

Paired duplication and deletion events in the isolate DZ12-11 suggests a BIR event between Ty elements located on nonhomologs. (A) Microarray analysis indicated a terminal duplication event on chromosome IV. The green line indicates the YDRCTy2-1 element near the break point. (B) A terminal deletion event on chromosome VI with a break point near YFLWTy2-1. (C) Generation of a translocation by BIR. A DSB within a Ty element on chromosome VI can be repaired by a BIR event involving a Ty element on chromosome IV as shown. Ty elements are indicated by arrows. The isolate shown in the rectangle would have a deletion of sequences from chromosome VI and a duplication of sequences from chromosome IV. (D) Contour-clamped homogeneous electric field gel analysis of the chromosome rearrangement in DZ12-11 (ethidium bromide-stained gel on Left; Southern analysis on Right). The strains JSC54-1 and JSC20-1 are isogenic with the haploids W303-1A and YJM789, respectively. DZ12 is the progenitor diploid before growth in low-galactose medium, and DZ12-11 is the strain with the deletion/duplication. The mechanism shown in Fig. S2C would be expected to result in a novel chromosome of about 650 kb that hybridizes to a probe derived from the left arm of IV; the expected band was observed using a probe from this region (POL3).

Fig. S3.

Comparison of patterns of chromosomal alterations caused by lowered expression of POL1 and POL3. Data from the low–DNA polymerase-α strains are from Song et al. (10), and data from the low–DNA polymerase-δ strains are from Dataset S5 of the current study. (A) The ratios of loss of heterozygosity (LOH), duplications (DUP), deletions (DEL), and changes in ploidy (PLOIDY) event compared with the total events in DZ12-derived (low Pol3; shown in gray) and WS84-derived (low Pol1; shown in purple). The asterisks indicate significance at a level of P < 0.05 by the Fisher exact test. (B) The ratios of crossovers/BIR (CO/BIR), gene conversions unassociated with crossover (CON), terminal deletions (TER DEL), terminal duplications (TER DUP), interstitial deletions (INT DEL), interstitial duplications (INT DUP), monosomy (MON), trisomy (TRI), and uniparental disomy (UPD) observed in strains with low levels of Pol3 and Pol1.

Fig. S4.

Correlation analysis of the enrichment of genomic elements near the break points of LOH events or other rearrangements caused by lowered levels of Pol1 (y axis) or Pol3 (x axis). POL1 (y axis) and lowered POL3 (x axis). The numbers on each axis are the ratios of the observed number of elements in the break points divided by the expected numbers. Data for the strains with low levels of DNA polymerase δ were derived from Datasets S1 and S6, and Table S3; data for strains with low levels of DNA polymerase α were from Song et al. (10).

Fig. S5.

Two types of gene conversion events associated with crossovers that give rise to red/white sectored colonies in DZ12. (A) Crossover associated with the repair of a single broken chromatid, generated in S or G₂ and resulting in a 3:1 conversion event. The region outlined in black indicates the conversion tract. The red and blue chromatids are derived from W303-1A and YJM789, respectively. (B) Crossover associated with the repair of two broken chromatids. The replication of a chromosome broken in G₀ or G₁ results in two broken chromatids (3). Repair of these breaks results in a 4:0 conversion.

Fig. 3.

Spontaneous recombination events and events induced by low levels of DNA polymerase δ on the right arm of chromosome IV. These plots summarize the number of times a SNP is included in conversion tracts associated with a crossover on chromosome IV. (A) Conversion tract distribution of spontaneous events (19). The labels HS1–HS7 are hot spots for crossovers. (B) Conversion tracts associated with low levels of DNA polymerase δ in DZ12. The highest peak contains the tandem array of the closely related HXT7/6/3 genes.

Fig. S6.

Deletion analysis of the HS5 hot spot. (A) The yeast strains used to monitor the activity of the wild-type HS5 hot spot (DZ67) and various deletion derivatives. The HIS3 and URA3 genes were inserted centromere-proximal and -distal to HS5, respectively, on the W303-1A-derived homolog (blue). The YJM789-derived homolog (red) was unaltered. (B) Modifications of the HS5 sequences. In DZ67, the unmodified hot spot consists of HXT7, HXT6, ARS432, and HXT3. The hphMX4 gene (resulting in hygromycin resistance) was used to generate the deletion derivatives DZ67d7, DZ67dA, DZ67d6, and DZ67d3, respectively. (C) Frequency of recombination between HIS3 and URA3 in the strains with wild-type HS5 activity and in various deletion derivatives.

Fig. S7.

Example of a gene conversion event on chromosome X (isolated DZ12-3) that was undetectable by SNP microarrays, but detectable by DNA sequencing. Red and blue lines/points indicate the relative hybridization ratio (HR) or relative sequence coverage (RC) of SNPs of W303-1A and YJM789, respectively. (A) Analysis of region located between 390 and 420 kb on chromosome X by SNP microarrays. (B) High-resolution microarray analysis of the region between 403 and 407 kb on chromosome X. (C) Whole-genome sequencing of the same region as in Fig. S7A indicates an LOH event. (D) High-resolution sequencing analysis of the region between 403 and 407 kb indicates an LOH event involving a single SNP. This SNP was not represented on the microarray.

Fig. S8.

Analysis of the same recombination event by both SNP microarrays and sequencing. This figure shows the analysis of an unselected recombination event on chromosome X in isolate DZ12-17. (A and B) Low- and high-resolution depiction of results from SNP microarray analysis showing a single transition between heterozygosity and homozygosity at about SGD coordinate 224,000. (C and D) Low- and high-resolution depictions of sequencing data for the same sample and the same region. This analysis shows there is a short LOH region (boxed in Fig. S8D) located near SGD coordinate 223,000 that was not detected by the SNP arrays. This region was not detected because oligonucleotides containing these SNPs were not present on the array.

Fig. 4.

CNV of repeated genes in strains with low levels of DNA polymerase δ. The copy numbers of each class of repeat were determined by DNA sequencing, using the number of “reads” for each repeat relative to the number of reads for single-copy sequences. (A) Depictions of ribosomal RNA and ASP3 genes on chromosome XII; only 2 of the 150 copies of the rRNA genes are shown. The indicated SNPs in the nontranscribed spacer region were used to measure the copy numbers of W303-1A– and YJM789-derived rRNA genes. (B) Numbers of rRNA genes in 15 sequenced isolates compared with the parental strain DZ12. The gray and purple colors indicate YJM789- and W303-1A–derived repeats, respectively. In isolates labeled “C,” there was a crossover within the rDNA. (C) The number of ASP3 repeats in DZ12 and 15 isolates; this repeat is found only in the W303-1A–derived homolog. (D) The numbers of CUP1 repeats. In the strain labeled “M,” chromosome VIII (which contains the CUP1 array) became monosomic. (E) An example of deletion between the HXT7 and HXT6 genes. This deletion was detected by a reduction in the sequencing coverage of the SNPs located at 1,156,234 and 1,157,008.

Fig. S9.

Comparison of point mutations in DZ12-derived isolates and in a wild-type yeast strain (25). (A) The ratio of mutations that occur on G or C bases relative to A or T bases. (B) The ratios of various types of transition and transversion mutations in strains with low DNA polymerase δ compared with wild type. In total, 867 point mutations were examined by Zhu et al. (24), and we analyzed 71.

Fig. 5.

DNA replication stress drives phenotypic changes in DZ12-derived isolates. (A) Comparison of the growth rates of a wild-type diploid (DZ3), DZ12, and 35 DZ12-derived isolates grown on medium containing low galactose. Approximately equal numbers of cells were placed on medium containing high [0.05% galactose, 3% (grams/100 milliliters) raffinose], low [0.005% galactose, 3% (grams/100 milliliters) raffinose], and YPD [2% (grams/100 milliliters) glucose, no galactose]. Numbers 1–35 represent the 35 DZ12-derived isolates (DZ12-1 to DZ12-35). (B) Deletions within the GAL1 promoter in three DZ12 isolates capable of growth on YPD. When genomic samples of DZ12-12, DZ12-14, and DZ12-9 were amplified using primers P2 and P3 (Table S4), the resulting products were shorter than the original fragment derived from DZ12. By sequence analysis, we found that all three strains contained deletions that were flanked by microhomologies [shown in green (DZ12-14), orange (DZ12-12), and blue (DZ12-9)].

Fig. S10.

Certain deletions of the GAL1 promoter allow GAL-POL3 strains to grow in the absence of galactose. The three mutant GAL1 promoters (GAL1mut1, GAL1mut2, and GAL1mut3 described in Fig. 5) were amplified using the genomic DNA from isolates DZ12-9, DZ12-12, and DZ12-14, respectively. The resulting fragments were used to transform DZ1 (a wild-type derivative of W303-1A), resulting in the strains WPM1, WPM2, and WPM3, respectively. Similarly, PCR fragments containing the GAL1mut1, GAL1mut2, and GAL1mut3 promoters were used to transform JSC20-1 (a wild-type derivative of YJM789) to produce the strains YPM1, YPM2, and YPM3, respectively. We compared the growth rates of the wild-type haploid strains, haploid strains with the undeleted GAL1 promoter (DZ1 and DZ2), and the strains with deletions within the GAL1 promoter on three types of medium.

Fig. S11.

Transcription factor binding motifs at the GAL1 promoter and adjacent sequences. The deletions that occurred in the GAL1 promoters in DZ12-derived isolates DZ12-9, DZ12-12, and DZ12-14 cover either the UAS_GAL and/or the Mig1 binding motif. Other binding motifs were located in the upstream region using Yeastract (www.yeastract.com/).

Fig. S12.

Strains with three copies of GAL-POL3 grow faster on medium containing low levels of galactose or no galactose, relative to strains that have two copies of GAL-POL3. As described in the text, DZ12-derived isolates DZ12-10, DZ12-11, and DZ12-30 have three copies of GAL-POL3. Strains DZ12-10dp, DZ12-11dp, and DZ12-30dp were deleted for one of these three copies, resulting in strains DZ12-10, DZ12-11, and DZ12-30, respectively. We tested the three-copy and two-copy strains for their abilities to grow on high-galactose, low-galactose, and glucose-containing plates. Strains with three copies of GAL-POL3 had a growth advantage on low-galactose and YPD plates, but not on the high-galactose plates.

Fig. S13.

Models depicting deletions generated by homologous recombination between repetitive DNA sequences. The blue and red lines are the two homologs, and the black arrows on the red homolog are repeated genes. (A) Unequal recombination between repeats on sister chromatids. (B) Intrachromosomal recombination between direct repeats.

Fig. 6.

Different patterns of chromosomal alterations caused by lowered expression of POL1 and POL3. Both types of polymerase-depleted strains have elevated rates of DSBs during S and/or G₂. We suggest that these breaks are efficiently repaired in strains with low levels of DNA polymerase α, but inefficiently repaired in strains with low levels of DNA polymerase δ.