Condensins regulate meiotic DNA break distribution, thus crossover frequency, by controlling chromosome structure

Abstract

Meiotic crossover (CO) recombination facilitates evolution and accurate chromosome segregation. CO distribution is tightly regulated: homolog pairs receive at least one CO, CO spacing is nonrandom, and COs occur preferentially in short genomic intervals called hotspots. We show that CO number and distribution are controlled on a chromosome-wide basis at the level of DNA double-strand break (DSB) formation by a condensin complex composed of subunits from two known condensins: the C. elegans dosage compensation complex and mitotic condensin II. Disruption of any subunit of the CO-controlling condensin dominantly changes DSB distribution, and thereby COs, and extends meiotic chromosome axes. These phenotypes are cosuppressed by disruption of a chromosome axis element. Our data implicate higher-order chromosome structure in the regulation of CO recombination, provide a model for the rapid evolution of CO hotspots, and show that reshuffling of interchangeable molecular parts can create independent machines with similar architectures but distinct biological functions.

Figure 1. Identification of a Condensin Complex that Controls CO Distribution

(A) Colloidal blue stained proteins from SDS-PAGE-fractionated IP reactions using DPY-26 antibodies and protein extracts from mixed-stage worms. Proteins were identified by mass spectrometry (Table S1). Red, condensin II subunit; blue, condensin I^DC subunit; red-blue, subunit in both complexes. (B–D) Reciprocal IPs and Western blot analysis using L4 extracts. The color scheme is as in (A). (B) Reciprocal IPs confirm association of DPY-26 with condensin II subunit SMC-4, condensin I^DC subunits DPY-27, CAPG-1, and DPY-28, and subunit MIX-1, common to condensin II and condensin I^DC. Antibodies for IPs are above the blots and antibodies for probes below. (C) IP for cohesin subunit SMC-3 (black) failed to recover DPY-26, showing that protein associations in (A) and (B) are not mediated by DNA. An IP for condensin II subunit HCP-6 failed to recover DPY-26, showing that DPY-26 does not associate with all condensin II subunits. DPY-27, SMC-4, MIX-1, and CAPG-1 IPs are positive controls. (D) Reciprocal IPs verify the association of SMC-4 with condensin I^DC subunits CAPG-1, DPY-26, DPY-28 and with shared condensin subunit MIX-1. DPY-27 IPs identify DPY-26 but not SMC-4. HCP-6 IPs identify SMC-4 but not DPY-26, indicating that HCP-6 and DPY-27 are not part of condensin I. (E) Subunit composition of three condensin complexes in C. elegans: condensin I^DC, condensin II, and condensin I. Condensin I, inferred from data in (A)–(D), includes the two SMC proteins MIX-1 and SMC-4 from mitotic condensin II and the three non-SMC proteins DPY-28, DPY-26, and CAPG-1 from condensin I^DC.

Figure 2. Mutation of any Gene Encoding a Condensin I Subunit Increases CO Frequency and Shifts CO Distribution to the Right End of X

CO analysis of X in heterozygous condensin I mutants using snip-SNPs. The relative physical and genetic positions of SNPs (red) used to map CO sites are above the chart. For each genotype (left), the CO frequencies (numbers in the colored boxes) were calculated by the formula (number of COs in the interval)/(total meiotic products assayed). Box colors represent the relative recombination frequencies in each interval between mutant and wild-type animals; the key is at the bottom. Shown to the right are the number of triple (3-CO), double (2-CO), single (1-CO), and non- (0-CO) crossover chromatids and the total number (n) of chromatids scored. (%), Percentage of 0-COs was calculated by the formula 100(0-CO/n). Asterisks mark CO intervals or frequencies statistically different (p < 0.01, Fisher’s exact test) from those in wild-type animals. (A) Heterozygous mutations in condensin I genes shift COs to the right end of X, but mutation of DC-specific gene dpy-27 does not. dpy-28(y283)/+ data are from Tsai et al. (2008). (B) γ-irradiation increases the number of COs on the left end of X and has an additive effect on CO frequency when combined with dpy-28 or dpy-26 mutations.

Figure 3. The Shift in CO Distribution Correlates Directly with the Shift in RAD-51 Distribution in dpy-28(y283) Mutants

(A) Pachytene chromosomes from wild-type and dpy-28(y283) animals labeled with X chromosome FISH probes from the center (red) and right end (blue) of X and antibodies to axial element HTP-3 (green) and RAD-51 (purple). X chromosome traces (yellow) are used to straighten each X and permit assessment of RAD-51 positions relative to FISH probes. (B) The relative genetic maps of dpy-28(y283) and wild-type animals show that interval A–D is reduced in CO frequency in y283 and interval D–F is increased. Individual SNPs scored are shown as red circles. Distances between SNPs reflect the frequency of COs between SNPs. Boundaries between red and blue intervals (red arrow) or blue and green intervals (blue arrow) are the approximate genetic positions of middle (red) or right-end (blue) FISH probes, respectively. dpy-28(y283) CO data are from Tsai et al. (2008). (C) dpy-28(y283) mutants show a dramatic decrease in RAD-51 foci in the left interval (A–D, red) of X, which has map compression, and a dramatic increase in RAD-51 foci in the center interval (D–F, blue), which has map extension. Values shown in white are the percent of total RAD-51 foci in each interval (left, red; middle, blue; right, green) as demarcated by FISH probes. The number of X chromosomes and RAD-51 foci scored for each genotype are shown below each graph. The number of foci in red and blue intervals are statistically different between wild-type and dpy-28(y283) animals (p < 0.002, Fisher’s exact test). (D and E) An obligate DSB. (D) Three-dimensional traces of chromosomes in a rad-54(ok615) pachytene nucleus (P2) permit quantification of RAD-51 foci per bivalent. Chromosomal axes are stained with HTP-3 antibodies (turquoise). Each chromosome trace is matched in color to its RAD-51 foci. (E) An obligate DSB. Quantification of RAD-51 foci (purple) on each of 198 bivalents from 33 rad-54(ok615) pachytene (P2) nuclei is plotted relative to the expected number of foci (turquoise) in each category based on the Poisson distribution. The y axis shows percentage of bivalents having the number of RAD-51 foci given on the x axis. The number of bivalents with zero foci (1%) is significantly less than expected, and the number with one focus (38%) is significantly more (p < 0.0001, binomial test), revealing a mechanism to guarantee at least one DSB per bivalent. Scale bars represent 1 μm.

Figure 4. TUNEL Assay Shows that Twice as Many DSBs Occur as COs in C. elegans

(A and B) TUNEL assay detects SPO-11-dependent DSBs (green) on pachytene chromosomes (red). The scale bar represents 4 μm. (C) Most TUNEL foci (green) colocalize with RAD-51 foci (red) in pachytene. The scale bar represents 1 μm. (D–I) Histograms show quantification of either RAD-51 or TUNEL foci in wild-type or rad-54(ok615) germlines. Each column color represents a class of nuclei with the indicated number of foci. A color key is at the bottom. The y axis shows the percentage of foci in each class. The x -axis shows the position along the germline: premeiotic region (M), transition zone (TZ), the first third of pachytene (P1), the second third of pachytene (P2), and the last third of pachytene (P3). The number of nuclei (n) scored, the average number of foci (avg), and standard error of the mean (SEM) are shown beneath each stage. (D and E) DSB number, as measured by TUNEL, correlates well with RAD-51 foci in wild-type germlines. (F and G) The plateau value of DSBs and RAD-51 foci in pachytene nuclei of rad-54(ok615) germlines shows an average value of ~12 DSBs in each meiocyte, twice as many DSBs as COs. (H) Elimination of germline cell death by ced-4(RNAi) in the rad-54(ok615) mutants reduces the average number of RAD-51 foci only in P3, where apoptosis occurs. (I) γ-irradiation (7.5 Gy) of rad-54(RNAi) animals increases the plateau value of RAD-51 foci, indicating that RAD-51 and the machinery to make RAD-51 foci are not limiting in the rad-54(RNAi) animals.

Figure 5. Condensin I Mutants Have More DSBs than Wild-Type Animals

(A) Shown are high-resolution images of early- to mid-pachytene nuclei from wild-type and mutant animals labeled with antibodies to RAD-51 (green) and the axis protein HTP-3 (red). Pachytene nuclei from mutants defective in the DCC-specific gene dpy-27 have a similar number of RAD-51 foci as wild-type animals, while animals heterozygous for a mutation disrupting any condensin I subunit show an increase in RAD-51 foci. Fields of nuclei are shown in Figure S2. The scale bar represents 1 μm. (B and C) Histograms show quantification of TUNEL foci in rad-54(RNAi); dpy-28(s939) or rad-54(RNAi) germlines. Histograms are labeled as in Figure 4. rad-54(RNAi); dpy-28(s939) mutants have a higher plateau value of TUNEL foci than rad-54(RNAi) animals (~15.4 versus ~12), consistent with the increase in COs and RAD-51 foci in the mutants. The average number of DSBs per nucleus in P1–P3 is statistically different between (B) and (C) (p < 0.001, two-tailed t test). (D–I) Histograms show quantification of RAD-51 foci in mutant germlines, as labeled above. (D and E) The average number of RAD-51 foci per nucleus in P1-P3 of rad-54(RNAi); dpy-28(s939) germlines (~14) is statistically different from that in rad-54(ok615) (Figure 4G) or rad-54(RNAi) germlines (~11) (p < 0.001, two-tailed t test), consistent with the s939-induced increase in COs. (F and G) The plateau value of RAD-51 foci is similar in rad-54(RNAi); dpy-28(y283) and rad-54(RNAi) germlines, consistent with y283 not increasing COs. (H and I) dpy-28(s939); unc-22(RNAi) germlines have increased RAD-51 foci compared to unc-22(RNAi) controls, which show the RNAi process does not affect RAD-51 foci (compare to Figure 4E).

Figure 6. Meiotic Chromosome Axis Length Is Expanded in Condensin I Mutants

(A) Shown are high-resolution images of nuclei from the first third of pachytene in wild-type and dpy-28(s939) germlines labeled for the axis protein HTP-3 (green) and a right-end X FISH probe (blue). A 3D X chromosome trace (yellow) was used to straighten each chromosome. (B–E) Computationally straightened chromosomes are displayed horizontally. Genotypes, average total axis length, and SEM are shown below each axis. (B) Disruption of dpy-28 causes an increase in X chromosome axis length that is independent of programmed DSBs made by SPO-11. Induction of extra DSBs by γ-irradiation does not increase axis length. (C) Mutation of any gene encoding a condensin I subunit causes a haploinsufficient extension of x axis length. In contrast, mutation of the condensin I^DC-specific dpy-27 gene does not. (D) The axis expansion caused by disrupting condensin I requires axis protein HIM-3. (E) The chromosome I axis is expanded in dpy-28(s939) mutants compared to wild-type animals. The scale bar represents 1 μm.

Figure 7. Condensin II Disruption Expands Chromosomal Axes but Alters CO Distribution Differently from Condensin I Disruption

(A) Condensin II complex. (B) Straightened X chromosomes from pachytene nuclei. Genotypes, average axis length, and SEM are below each axis. X axis lengths in kle-2 (null)/+ and dpy-28(null)/+ mutants are similar but longer than in wild-type animals. Axis length in dpy-28/+; kle-2/+ double mutants is greater than in either single mutant, but axis length in dpy-28/+; dpy-26/+ double mutants is similar to that in either single mutant, showing independent action of condensin II and I. (C) Axis expansion in kle-2/+ pachytene chromosomes correlates with a DSB increase. (D) kle-2/+ mutants show an increase in 2-COs and a shift in CO distribution to the left end of X, the opposite end of condensin I mutants. CO analysis and presentation are as in Figure 2. Asterisks mark CO intervals or frequencies statistically different (p < 0.01, Fisher’s exact test) from those in wild-type animals. The scale bar represents 1 μm.