Heterozygous inversion breakpoints suppress meiotic crossovers by altering recombination repair outcomes

Abstract

Heterozygous chromosome inversions suppress meiotic crossover (CO) formation within an inversion, potentially because they lead to gross chromosome rearrangements that produce inviable gametes. Interestingly, COs are also severely reduced in regions nearby but outside of inversion breakpoints even though COs in these regions do not result in rearrangements. Our mechanistic understanding of why COs are suppressed outside of inversion breakpoints is limited by a lack of data on the frequency of noncrossover gene conversions (NCOGCs) in these regions. To address this critical gap, we mapped the location and frequency of rare CO and NCOGC events that occurred outside of the dl-49 chrX inversion in D. melanogaster. We created full-sibling wildtype and inversion stocks and recovered COs and NCOGCs in the syntenic regions of both stocks, allowing us to directly compare rates and distributions of recombination events. We show that COs outside of the proximal inversion breakpoint are distributed in a distance-dependent manner, with strongest suppression near the inversion breakpoint. We find that NCOGCs occur evenly throughout the chromosome and, importantly, are not suppressed near inversion breakpoints. We propose a model in which COs are suppressed by inversion breakpoints in a distance-dependent manner through mechanisms that influence DNA double-strand break repair outcome but not double-strand break formation. We suggest that subtle changes in the synaptonemal complex and chromosome pairing might lead to unstable interhomolog interactions during recombination that permits NCOGC formation but not CO formation.

Fig 1

A) Canonical recombination pathways used in meiosis. 1) Meiosis is initiated by an enzymatically mediated DSB. 2) The DSB is resected into single-stranded ends, one of which invades the homologous chromosome and primes DNA synthesis. This forms a displacement loop (D-loop) and during synthesis dependent strand annealing (SDSA), this structure can be unwound by a helicase into a NCO. 3) If the second end of the DSB is captured by the D-loop, it also primes synthesis. 4) The second-end capture intermediate is ligated into a double Holliday Junction, which is cleaved by a meiosis-specific endonuclease to form mostly COs, although some NCOs can occasionally form. B) Predicted pairing arrangement between an inversion and a standard arrangement homologous chromosome. Single COs within the inversion breakpoints lead to acentric and dicentric chromosomes with deletions and duplications. If these COs occur, they are not recovered in the offspring (the so-called transmission distortion effect). COs are also suppressed in the collinear regions outside of the inversion breakpoints, even though COs in these regions would not lead to chromosome rearrangements. This suggests that COs are prevented from forming, as opposed to being suppressed due to transmission distortion.

Fig 2

A) Genomic coordinates of the phenotypic markers and the dl-49 inversion breakpoints. The inversion breakpoints of dl-49 are at positions 4.89 Mb and 13.4 Mb. All numbers are in Mb using the coordinates from genome version dm6 [47]. B) Cross schemes for all three experimental set ups. COs were identified by scoring for new combinations of parental alleles. In cross 2, cv and wy were not scored because the Cy phenotype interfered with scoring for wy; for consistency, we also did not score cv and used only y and f.

Fig 3. Location and frequency of COs recovered in Oregon-RM and dl-49 heterozygotes.

When selecting for COs between y and cv or wy and f, we recovered chromosomes that had two COs, one of which was between cv and wy. These are shown in the figure, but the COs were not included in any analyses. CO counts were normalized to sample size by dividing the number of COs per interval by the total number of COs. 95% confidence intervals were calculated and can be found in S1 File. In Oregon-RM, CO distribution between the dl-49 breakpoint and f are not correlated with distance (Spearman’s rank correlation, r = 0.18, p = 0.35). In dl-49 heterozygotes, CO distribution between the dl-49 breakpoint and f are correlated with distance (Spearman’s rank correlation, r = 0.47, p = 0.01).

Fig 4. Location and frequencies of NCOGCs recovered from Oregon-RM and dl-49 heterozygotes.

Frequencies are raw counts per 150 kb window.

Fig 5. Analysis of phosphorylated H2AV in Oregon-RM and dl-49 heterozygotes.

A) Germaria from Oregon-RM and dl-49 heterozygotes were stained with anti-phospho-H2AV, anti-C(3)G, and DAPI. Meiotic nuclei are identified by the presence of C(3)G, a member of the synaptonemal complex. DSBs–marked by phosphorylated H2AV–begin to form in region 2A and are finished being repaired by the time the cyst enters region 3 in wildtype germaria [47]. Note that H2AV is phosphorylated in non-meiotic cells in response to DNA damage. B) Quantitative analysis of DSB timing and number in full sibling Oregon-RM and dl-49 heterozygotes. Average number of foci with the standard error of the mean are shown. A two-way ANOVA with Tukey’s posthoc correction shows that the number of DSBs and the timing of their repair in dl-49 heterozygotes are statistically the same as in Oregon-RM (see main text for details).