PMC

Representative time courses are shown.

Representative time courses are shown.

Representative pulsed-field gel Southern blots probed for Chr IX are shown, labeled as in .

a, b, Spo11-oligo complex labeling from representative time courses is in a and quantification from ≥ 3 cultures (mean ± SD) is in b. c, Meiotic progression (percent of cells completing the first division).

Representative Southern blots of DNA separated on a conventional agarose gel and probed for GAT1 (a), CCT6 (b, c), and ERG1 (d). JMs, joint molecules; P, parental-length DNA; W, wells. The arrowhead in a indicates signal from the CCT6 parental band that remained after stripping and reprobing for GAT1. e, Quantifications for b, c, d (mean ± SD for 3 cultures).

a, b, Recombination reporters at the ERG1 (a) and GAT1 (b) hotspots. c, d, e, Representative Southern blots of parental and recombinant DNA molecules at CCT6 (c), ERG1 (d), and GAT1 (e). The arrowhead in e indicates a non-reproducible radiolabeled species. f, Local distribution of DSBs around recombination reporter locations is not altered in zip3 mutants. Spo11-oligo profiles (averages for wild type and zip3) are smoothed with 201-bp Hann window; zip3 values are offset to separate profiles.

a, b, ZMM status is irrelevant in a dmc1 background. Broken chromosomes accumulate to similar levels in a dmc1 single mutant and dmc1 zmm double mutants. Representative pulsed-field gel Southern blots probed for Chr IX are in a and Poisson-corrected quantification of DSBs is in b (mean ± SD, 3 cultures). P, parental; W, wells. c, Reducing Spo11 activity in a zip3 mutant partially alleviates the prophase I delay/arrest. Meiotic progression was assessed by staining with DAPI and measuring the percentage of cells that had completed MI (±MII). Data are means ± SD for 3 cultures, except wild type and spo11-HA, each analyzed once.

a, Schematic of arg4 heteroalleles, showing ORFs and mutated restriction sites. Below, Spo11-oligo profile shows DSB distribution (RPM, reads per million mapped; smoothed with 201-bp Hann window). b, Heteroallele recombination frequencies (mean ± SD). *, significantly different from wild type (p<0.02, t test); **, significantly different from ndt80 (p<0.006). c, Recombination reporter at the CCT6 hotspot. d, Representative Southern blots of parental and recombinant DNA molecules (crossovers (COs) and noncrossover gene conversions (NCOs)) at CCT6 resolved by two-dimensional gel electrophoresis. e, Recombination frequencies (mean ± SD). Crossover frequencies were halved to convert to per-DSB equivalent because each crossover yields two recombinant molecules. *, total recombination significantly different from wild type (p<0.003); **, crossing over significantly different from wild type (p<0.04). f, g, DSBs at CCT6 and GAT1. A representative Southern blot probed for CCT6 is in f and quantifications for CCT6 and GAT1 are in g (mean ± SD for 3 cultures, except 8 hr for zip3 at GAT1, analyzed twice). JMs, joint molecules; P, parental; W, wells.

a, Southern blots probed for Chr III. High molecular weight chromosomal DNA was purified 6 hr after transfer to sporulation medium from meiotic rad50S cultures carrying the indicated SPO11 alleles (in spo11-yf the catalytic tyrosine 135 is mutated to phenylalanine), then separated on pulsed-field electrophoresis gels. Samples from a rad50S spo11-HA strain are shown for comparison; HA-tagged Spo11 has reduced DSB frequency. Each lane represents an independent culture (SPO11⁺ samples from the same cultures were run on both gels). P, parental-length DNA; W, wells. b, Quantification of blots in panel a and separate blots (not shown) probed for Chr VII or VIII. Break frequencies are percent of DNA in lane (mean ± SD of 3–4 cultures). Numbers in parentheses indicate values from each tagged strain relative to SPO11⁺ for the same chromosome. Relative DSB frequencies at the bottom are averages across the three chromosomes assayed.

a, Spo11 generates a covalent protein-linked DSB; endonucleolytic cleavage releases Spo11 bound to a short oligo (detection method at left). Resection is followed by strand invasion and ZMM-dependent stabilization of intermediates fated to become crossovers. b, c, Representative pulsed-field gel Southern blots probed for Chr IX are in b and Poisson-corrected DSB quantification in c (mean ± SD, 3 cultures). P, parental; W, wells. d, e, Representative Spo11-oligo complex time courses are in d and quantification in e (mean ± SD for 3 cultures, except at 10 hr for msh5 and zip3 analyses (1 culture)). Radiolabeled Spo11-oligo complexes were detected by autoradiography (top panels) and total Spo11 was detected by anti-flag western blot (WB, middle). The main labeled species differ in oligo size. Nearly all of the WB signal is Spo11 that has not made a DSB. Asterisk, species co-migrating with upper Spo11-oligo complexes; arrowhead, proteolytic product. Extract samples run separately and stained with Coomassie control for input to the IPs (bottom). In panels c and e, mutants are plotted with wild-type data collected in parallel.

a, b, Quantitative reproducibility of Spo11-oligo maps. In a, comparisons are shown for individual wild type (WT) or zip3 datasets from the present study, or the previously published spo11-HA data (from ref ). Uniquely mapped Spo11 oligos were summed in non-overlapping 5-kb bins and expressed as RPM per kb (plotted on a log scale). In b, pairwise correlation coefficients for the datasets from the current study are shown (Pearson's r; box colors scaled from blue to red proportional to strength of correlation). For the comparison of this study's wild-type average with data from Pan et al., r = 0.949. Note that Pan et al. used a different strain background with different auxotrophies, which may alter DSB distributions^61,62, and a hypomorphic spo11 allele (spo11-HA), which may affect DSBs to different extents at different locations⁵⁶. Note that biological replicates (WT-1 vs. WT-2 or zip3-1 vs. zip3-2) agreed better than comparisons between cultures of different genotype. c, DSBs form at the same hotspots and with similar distribution within and between hotspots in wild type and zip3. Unsmoothed Spo11 oligo maps are shown in the vicinity of the well-characterized ARE1 (YCR048w) hotspot.

a, Upper, reproducibility of Spo11-oligo maps. Lower, DSBs form at the same hotspots in zip3 as wild type. Smoothed with 201-bp Hann window. b, Zip3 is required for chromosome size-dependent variation in Spo11-oligo density. Lines, least squares fits (dashed = not significant). c, Larger chromosomes experience greater increase in Spo11 oligos. Fold change is the per-chromosome Spo11-oligo density in zip3 over wild type (WT). Open circles, Chr XII (“12”, omitting rDNA length) and the portions of Chr XII left or right of the rDNA (“12L”, “12R”). Regression line treats 12L and 12R as separate chromosomes. d, Regional variation in response to zip3 mutation. Each point is the change at a hotspot (plotted on log scale). Red lines, local regression (loess); green circles, centromeres. e, Local domains of correlated behavior. Each point compares hotspots to their neighbors in 5-kb-wide windows the indicated distance away. Nearby hotspots show correlated behavior for fold change in zip3 (red), but not heat (Spo11-oligo frequency) in wild type (black). Shaded areas denote 95% CI estimates for hotspots randomized within-chromosome (randomized r>0 for zip3-fold-change because of the chromosome size effect). f, Correlation between log-fold-change in zip3 and binding of indicated proteins, binned in non-overlapping windows of varying size. For clarity, other proteins are in . Pericentric, sub-telomeric, and rDNA-proximal regions were censored. Closed symbols, p<0.05. g, Fit of multiple regression model predicting changes (log scale) in Spo11-oligo density in 35-kb windows from ChIP data, G+C content, and chromosome size (). Dashed lines, observed mean fold change. h, Network of feedback circuits controlling DSB formation. Circuit 1: DSBs activate Tel1 (ATM in mouse), which inhibits further DSB formation. Circuit 2: ZMM-dependent interactions between homologous chromosomes inhibit Spo11. Circuit 3: Ndt80 shuts down DSB formation and drives pachytene exit; Mec1 kinase delays or blocks Ndt80 activation when DSBs are present.

a, Change in Spo11-oligo counts in hotspots grouped by chromosomal context. Tel, within 20 kb of telomeres; Cen, within ±10 kb of centromeres; rDNA, from 60 kb leftward to 30 kb rightward of rDNA; Interstitial, all others. Dashed lines mark values assumed as no change and average change (1.8-fold). Boxes indicate median and interquartile range; whiskers indicate the most extreme data points which are ≤1.5 times the interquartile range from the box; individual points are outliers. Sub-telomeric and pericentric zones show less increase in zip3 on average, thus, ZMM-dependent feedback contributes less than other, unknown factors to suppressing DSBs in these regions. The zone near the rDNA showed no increase or was even decreased; thus, zip3 mutants are competent for this region's DSB suppression, which is dependent on the ATPase Pch2 and the replication factor Orc1 (ref 63). Note that the remaining interstitial hotspots showed highly variable response to zip3 mutation (>20 fold). b, Correlation between log-fold change in Spo11-oligo counts in zip3 and the binding of the indicated proteins, binned in non-overlapping windows of varying size. Closed symbols, p<0.05. ChIP data are from ref . c, Average ChIP profiles around interstitial hotspots divided into three equal-sized groups according to the average fold change in zip3. Top: the box-and-whisker plot (as described for panel a) shows the distribution of fold changes for the three groups. Below: ChIP profiles for each of the indicated proteins. Note that the profiles lie atop one another for Rec102 and Rec104. Dashed arrows indicate direction of the change in the average profiles with increasing fold change in zip3. ChIP data are from refs and . d, High degree of colinearity of log₂-transformed ChIP data for Rec114, Mei4, and Mer2 (which are essential for DSB formation) and Hop1 and Red1 (axis proteins that promote normal DSB formation). More than 90% of the variance for this combination of ChIP data is captured in the first principal component (PC1). The high degree of correlation between these proteins was described previously. e, Correlations between the fold change in zip3 (zip3 FC, log₂ and assuming 1.8-fold increase genome-wide) and various chromosomal features: principal component 1 (PC1) for Rec114, Mei4, Mer2, Hop1, and Red1 ChIP data (same as in panel d); chromosome size (log_e(bp)); G+C content (%); and ChIP data for the indicated proteins (log₂). In d and e, upper right panels show pair-wise scatter plots and lower left panels show corresponding correlation coefficients (Pearson's r) for data for interstitial regions binned in 35-kb non-overlapping windows. Essentially identical results were obtained with different window sizes (20–40 kb) or with varying placement of windows (data not shown). f, Essentially no correlation between DSB activity in wild type and change in zip3, whether considering interstitial regions divided into non-overlapping 35-kb bins (upper panel) or interstitial hotspots (lower panel). A 1.8-fold increase genome-wide in zip3 is assumed. Note: Fold change is labeled according to a linear scale but plotted in a log scale in panels a, c, f.