close
Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Oct 26;193(3):1913-1932.
doi: 10.1093/plphys/kiad432.

Comprehensive dissection of meiotic DNA double-strand breaks and crossovers in cucumber

Yanling Wang  1 , Zhaonian Dong  1 , Yalin Ma  2 , Yi Zheng  1 , Sanwen Huang  1   3 , Xueyong Yang  2
Affiliations

Comprehensive dissection of meiotic DNA double-strand breaks and crossovers in cucumber

Yanling Wang et al. Plant Physiol. .

Abstract

Meiotic recombination drives genetic diversity and crop genome optimization. In plant breeding, parents with favorable traits are crossed to create elite varieties. Different hybridizations produce diverse types of segment reshuffling between homologous chromosomes. However, little is known about the factors that cause hybrid-specific changes in crossovers (COs). Here, we constructed 2 F2 populations from crosses between a semiwild and 2 domesticated cucumber (Cucumis sativus) accessions and examined CO events. COs mainly occurred around genes and differed unevenly along chromosomes between the 2 hybrids. Fine-scale CO distributions were suppressed in regions of heterozygous structural variations (SVs) and were accelerated by high sequence polymorphism. C. sativus RADiation sensitive 51A (CsRAD51A) binding, histone H3 lysine 4 trimethylation (H3K4me3) modification, chromatin accessibility, and hypomethylation were positively associated with global CO landscapes and in local DNA double-strand break (DSB) hotspots and genes. The frequency and suppression of COs could be roughly predicted based on multiomic information. Differences in CO events between hybrids could be partially traced to distinct genetic and epigenetic features and were significantly associated with specific DSB hotspots and heterozygous SVs. Our findings identify the genomic and epigenetic features that contribute to CO formation and hybrid-specific divergence in cucumber and provide theoretical support for selecting parental combinations and manipulating recombination events at target genomic regions during plant breeding.

PubMed Disclaimer

Conflict of interest statement

Conflict of interest statement. None declared.

Figures

Figure 1.
Figure 1.
Experimental design for multiomic analysis of meiotic recombination formation in cucumber. A) Construction of cucumber F2 populations. The semiwild Xishuangbanna cucumber accession Cuc80 was chosen as a shared female parent to cross with cultivated accessions 9930 (East Asia type) and Cuc37 (Eurasian type), respectively. The obtained F1 hybrids were self-pollinated to produce F2 populations. B) Multiomic integration of cucumber recombination. DSB sites were measured by CsRAD51A ChIP-seq. Epigenetic modifications were carried out by H3K4me3 ChIP-seq, FAIRE-seq, and WGBS. Transcription levels were presented with RNA-seq. 5-Methylcytosine (5-meC) represents methylation of the fifth position of cytosine.
Figure 2.
Figure 2.
Enrichment of COs within various genomic components in cucumber. A) Relative abundance of COs enriched within the genic region and flanking intergenic region in cucumber. Stack bar charts show frequencies of COs overlapping with gene body and intergenic region with increasing distances to the genes in Cuc80 × 9930 (left) and Cuc80 × Cuc37 (right) populations. B, C) Enrichment proportion of COs in different genomic components. Pie charts show the proportion of COs overlapping with the CDS, UTR, intron, promoter (2 kb upstream of TSS), and distal intergenic region in B) Cuc80 × 9930 and C) Cuc80 × Cuc37, respectively. D, E) Relative enrichment and significance of COs in different genomic components. Histogram charts represent the relative enrichments of COs in corresponding genomic components in D) Cuc80 × 9930 and E) Cuc80 × Cuc37 normalized by the expected CO numbers. Compared to the expected number, deposition and depletion are marked by positive and negative log values, respectively. Significances are assessed by 10,000 times permutation test. **P < 0.01, and ***P < 0.001.
Figure 3.
Figure 3.
Comparison of CO landscapes between cucumber populations. The green shadows represent the predicted positions of centromeres and pericentromeres (1 Mb flanking regions of centromeres). Blue and orange lines indicate chromosomal patterns in Cuc80 × 9930 and Cuc80 × Cuc37, respectively. The windows in Cuc80 × 9930 with a significantly higher CO rate than Cuc80 × Cuc37 is marked with blue hollow circles, and the opposite is orange. Unique CO hotspots in Cuc80 × 9930 and Cuc80 × Cuc37 are marked with blue and orange six-pointed stars, respectively. The bottom stacks represent gene density. Bin size = 50 kb. The significance of difference in recombination rates is calculated with 10,000 times permutation test, P < 0.05.
Figure 4.
Figure 4.
COs are suppressed around SV but positively associated with SNP density on the fine scale. A, B) Relative enrichment and significance of COs in different types of SVs in A) Cuc80 × 9930 and B) Cuc80 × Cuc37. Histogram charts represent for relative enrichments of COs in corresponding genomic components normalized by the expected CO number. Compared to expected number, deposition and depletion are shown in red and dark gray, respectively. C, D) Distances of nearest COs to SVs. DEL, deletion; INS, insertion; INV, inversion; TRANS, translocation (n = 2,540 for total COs with intervals <100 kb, 396 for SV-overlapping COs, 186 for DEL-overlapping COs, 196 for INS-overlapping COs, 1 for INV-overlapping COs, 13 for TRANS-overlapping COs in Cuc80 × 9930; n = 2,582 for total COs with intervals <100 kb, 632 for SV-overlapping COs, 265 for DEL-overlapping COs, 342 for INS-overlapping COs, 1 for INV-overlapping CO, 24 for INS-overlapping COs, INV-overlapping COs, TRANS-overlapping COs in Cuc80 × Cuc37). Boxplots show the median with center line, upper and lower quartiles with box limits, outliers with points, and the whiskers represent 1.5× interquartile range limits. E, F) SNP density around centers of CO intervals in cucumber. CO center indicates the midpoint of the CO interval. E) Hollow circles stand for SNP density (count/kb) in Cuc80 × 9930 within increasing distance away from CO centers of total COs, COs with intervals <5 kb, COs with intervals <10 kb, COs with intervals <15 kb, COs with intervals <20 kb, and random genomic region. The green dash line represents for average SNP density in genome-wide. F) Distributions in Cuc80 × Cuc37 (n = 1,413 for CO intervals <5 kb, 1,729 for CO intervals <10 kb, 1,912 for CO intervals <15 kb, 2,080 for CO intervals <20 kb, 2,585 for random and total COs in Cuc80 × 9930; n = 1,596 for CO intervals <5 kb, 1,879 for CO intervals <10 kb, 2,036 for CO intervals <15 kb, 2,152 for CO intervals <20 kb, 2,663 for random and total COs in Cuc80 × Cuc37). Significances are assessed by A, B) 10,000 times permutation test and C, D) Mann–Whitney U test. ***P < 0.001.
Figure 5.
Figure 5.
Genome-wide associations between meiotic CO and multiomic signals in cucumber meiosis. A) Genomic landscapes of CO, genomic and epigenetic features during meiosis in C. sativus cv. 9930. Circos plot showing coverage profiles along (a) physical coordinate (black contig for predicted centromere and pericentromere) of cucumber chromosome for (b) gene density, (c) repeat density, (d) DSB density, (e) recombination rate (centimorgan per megabase [cM/Mb]), (f) CsRAD51A binding (standardized log2[ChIP/Input] ratio), (g) H3K4 trimethylation level (standardized log2[ChIP/Input] ratio), (h) chromatin accessibility measured by FAIRE-seq (standardized log2[ChIP/Input] ratio), (i) transcription level (TPM), and (j) cytosine methylation level (mC/C ratio). B) Genomic landscapes of CO, genomic and epigenetic features during meiosis in C. sativus cv. Cuc37. C, D) The heatmaps represent pairwise Spearman's rank correlation matrix for CsRAD51A binding, chromatin accessibility, H3K4me3, transcription, gene density, repeat density, SNP density between homologs, and CO rate (cM/Mb). Color intensity indicates the strength of correlations.
Figure 6.
Figure 6.
Genomic location annotation and motif analysis of meiotic DSB hotspots in cucumber. A, B) Enrichment proportion of DSB hotspots in different genomic components. Pie charts show the proportion of DSB hotspots overlapping with the CDS, UTR, intron, promoter (2-kb upstream of TSS), and distal intergenic region in A) 9930 and B) Cuc37, respectively. C, D) Relative enrichment and significance of DSB hotspots in different genomic components. Histogram charts represent the relative enrichment of DSB in corresponding genomic components in C) 9930 and D) Cuc37 normalized by the expected DSB numbers. Compared to expected number, deposition and depletion are marked by postive and negative log values, respectively. E) Most abundant DNA sequence motifs identified in 9930 and Cuc37 DSB hotspots. The alignment quality (P-value) and enrichment proportions are displayed. PWM, position weight matrix. Significances are assessed by 10,000 times permutation test. *P < 0.05, **P < 0.01, and ***P < 0.001.
Figure 7.
Figure 7.
Aggregated plots of CsRAD51A, H3K4 trimethylation, chromatin accessibility, and DNA methylation around meiotic DSB hotspots in cucumber. A) DSB hotspots in 9930 were divided into 2 groups based on decreasing recombination rate (Quantile 1 in red, n = 749; Quantile 2 in blue, n = 2,731) or random grouping (Random 1 in red, n = 749; Random 2 in blue, n = 2,731), respectively. For comparison, the random genomic region with the same length and chromosome distribution was taken as control (Random 1 in red, n = 962; Random 2 in blue, n = 2,518). Quantile 1 for DSBs with recombination rate higher than genomic average Quantile 2 for DSBs with recombination rate lower than genomic average. Random 1 and Random 2 for randomly dividing evenly into 2 groups. Lines show the average epigenetic signals and corresponding shadows for a 95% confidence interval. H3K4me3, histone H3 lysine 4 trimethylation; FAIRE, formaldehyde-assisted isolation of regulatory elements; cM/Mb, centimorgan per megabase. B) DSB hotspots in Cuc37. Quantile 1 and Random 1 for DSBs, n = 1,079; Quantile 2 and Random 2 for DSBs, n = 4,526. Random 1 for genomic regions, n = 1,627; Random 2 for genomic regions, n = 3,978.
Figure 8.
Figure 8.
Aggregated plots of CsRAD51A, H3K4 trimethylation, chromatin accessibility, and DNA methylation around gene regions in cucumber. A) Gene regions in 9930 were divided into 2 groups based on decreasing recombination rate (Quantile 1 in red, n = 10,045; Quantile 2 in blue, n = 13,477) and randomly grouping (Random 1 in red, n = 10,045; Random 2 in blue, n = 13,477), respectively. For comparison, the random genomic region with the same length and chromosome distribution was taken as control (Random 1 in red, n = 11,760; Random 2 in blue, n = 11,762). Quantile 1 for genes with recombination rate higher than genomic average and Quantile 2 for genes with recombination rate lower than genomic average. Random 1 and Random 2 for randomly dividing evenly into 2 groups. Lines show the average epigenetic signals and corresponding shadows for a 95% confidence interval. H3K4me3, histone H3 lysine 4 trimethylation; FAIRE, formaldehyde-assisted isolation of regulatory elements; cM/Mb, centimorgan per megabase. B) Gene regions in Cuc37. Quantile 1 and Random 1 for genes, n = 9,906; Quantile 2 and Random 2 for genes, n = 13,616. Random 1 for genomic regions, n = 11,760; Random 2 for genomic regions, n = 11,762.
Figure 9.
Figure 9.
Comparison of observed and predicted CO frequency along cucumber chromosomes. The predictions of CO landscapes were performed by a random forest regression. Line plots strand for observed and predicted CO rates in 2.5-Mb sliding windows in A) Cuc80 × 9930 and B) Cuc80 × Cuc37.
Figure 10.
Figure 10.
Enrichment of distinct epigenetic modifications and SVs in CO alterations between cucumber hybrids. Upset diagrams show the number of unique and overlapping with distinct meiotic DSB hotspots, open chromatin (FAIRE), H3K4me3, hypo-DMR, absence of a heterozygous SV (SV escape) and interaction of features in A) Cuc80 × 9930 and B) Cuc80 × Cuc37. Pie charts show the total proportion of unique COs overlapping with distinct features and interactions. Red means features significantly overlapped with unique COs (P < 0.05 for 10,000 permutation test).
See this image and copyright information in PMC

References

    1. Bai SL, Peng YB, Cui JX, Gu HT, Xu LY, Li YQ, Xu ZH, Bai SN. Developmental analyses reveal early arrests of the spore-bearing parts of reproductive organs in unisexual flowers of cucumber (Cucumis sativus L). Planta 2004:220(2):230–240. 10.1007/s00425-004-1342-2 - DOI - PubMed
    1. Blackwell AR, Dluzewska J, Szymanska-Lejman M, Desjardins S, Tock AJ, Kbiri N, Lambing C, Lawrence EJ, Bieluszewski T, Rowan B. MSH2 shapes the meiotic crossover landscape in relation to interhomolog polymorphism in Arabidopsis. EMBO J. 2020:39(21):e104858. 10.15252/embj.2020104858 - DOI - PMC - PubMed
    1. Brick K, Thibault-Sennett S, Smagulova F, Lam KG, Pu Y, Pratto F, Camerini-Otero RD, Petukhova GV. Extensive sex differences at the initiation of genetic recombination. Nature 2018:561(7723):338–342. 10.1038/s41586-018-0492-5 - DOI - PMC - PubMed
    1. Browning SR, Browning BL. Rapid and accurate haplotype phasing and missing-data inference for whole-genome association studies by use of localized haplotype clustering. Am J Hum Genet. 2007:81(5):1084–1097. 10.1086/521987 - DOI - PMC - PubMed
    1. Cai C, Pelé A, Bucher J, Finkers R, Bonnema G. Fine mapping of meiotic crossovers in Brassica oleracea reveals patterns and variations depending on direction and combination of crosses. Plant J. 2023:113(6):1192–1210. 10.1111/tpj.16104 - DOI - PubMed

Publication types