FISH landmarks reflecting meiotic recombination and structural alterations of chromosomes in wheat (Triticum aestivum L.)

DNA sequence composition affects meiotic recombination. However, the correlation between tandem repeat composition and meiotic recombination in common wheat (Triticum aestivum L.) is unclear. Non-denaturing fluorescent in situ hybridization (ND-FISH) with oligonucleotide (oligo) probes derived from tandem repeats and single-copy FISH were used to investigate recombination in three kinds of the long arm of wheat 5A chromosome (5AL). 5AL535–18/275 arm carries the tandem repeats pTa-535, Oligo-18, and pTa-275, 5AL119.2–18/275 arm carries the tandem repeats pSc119.2, Oligo-18 and pTa-275, and 5AL119.2 arm carries the tandem repeats pSc119.2. In the progeny of 5AL535–18/275 × 5AL119.2, double recombination occurred between pSc119.2 and pTa-535 clusters (119–535 interval), and between pTa-535 and Oligo-18/pTa-275 clusters (535–18 interval). The recombination rate in the 119–535 interval in the progeny of 5AL535–18/275 × 5AL119.2–18/275 was higher than that in the progeny of 5AL535–18/275 × 5AL119.2. Recombination in the 119–535 interval produced 5AL119 + 535 segments with pTa-535 and pSc119.2 tandem repeats and 5ALNo segments without these repeats. The 5AL119 + 535 and 5ALNo segments were localized between the signal sites of the single-copy probes SC5A-479 and SC5A-527. The segment between SC5A-479 and SC5A-527 in the metaphase 5ALNo was significantly longer than that in the metaphase 5AL119 + 535. The structural variations caused by tandem repeats might be one of the factors affecting meiotic recombination in wheat. Meiotic recombination aggregated two kinds of tandemly repeated clusters into the same chromosome, making the metaphase chromosome more condensed. To conclude, our study provides a robust tool to measure meiotic recombination and select parents for wheat breeding programs.


Background
Increasing the chromosomal recombination rate can accelerate new cultivars' development in common wheat (Triticum aestivum L.). Researchers have extensively studied the molecular mechanisms of meiotic recombination in plants and reported many genes involved in the process [1], which provide an opportunity to manipulate the mechanisms for crop improvement [2]. Mutation in the anticrossover gene FANCM (Fanconi anemia complementation group M) resulted in a two-fold increase in meiotic recombination in hybrid rice and pea [3]. However, the fancm mutation has almost no effect on recombination in hybrid Arabidopsis but resulted in a three-fold increase in inbreds [3][4][5]. Combining the anticrossover mutants recq4 (RecQ like helicase 4) and figl1 (FIDGETIN-like-1) resulted in a 7.8-fold increase in crossover frequency, while the fancm, recq4 and figl1 triple mutant displayed less recombination [5]. These findings indicated that factors other than genes help control meiotic recombination. Both genes and chromatin structure control the recombination rate. In humans and animals, the recombination rate is related to cytogenetic structures that were displayed by staining intensity of G bands and sequence compositions including repetitive elements, GC content, CpG density and poly(A)/poly(T) stretches [6,7]. In Arabidopsis thaliana, A-rich, CCN-repeat and CTT-repeat motifs are enriched in the crossover regions [8]. Reduced DNA methylation at the CG sites increases recombination rate in the euchromatic, but not in the pericentric heterochromatin regions [9]. Heterochromatin plays a role in meiotic recombination. It represses centromeric meiotic recombination in fission yeast [10]. The role of pericentric heterochromatin in suppressing meiotic recombination has been widely studied in eukaryotes [11]. The disruption of H3K9me2 and non-CG DNA methylation pathways via gene mutations increased pericentromeric crossovers in hybrid and inbred Arabidopsis [12]. Meanwhile, in mice, a tandem array of mo-2 minisatellite conferred higher-order structures crucial for recombination in the pseudoautosomal region of sex chromosomes [13]. Therefore, the effect of tandem repeats on chromosomal recombination should be investigated. Common wheat can be used as a reference model to study the role of tandem repeats in chromosomal recombination. Previous studies have indicated a correlation between the wheat chromosomal crossover and gene-rich regions [14][15][16][17][18]. However, different aspects of recombination should be considered for common wheat because of allopolyploidy and repetitive sequences [19].
Some new tandem repeats were discovered from common wheat [20,21]. In our previous work, oligonucleotide (oligo) probes derived from these tandem repeats displayed large structural variations in the 5AL arms of common wheat [22,23]. However, the correlation between the composition of tandem repeats and the rate of recombination in wheat is still unclear. This study investigates the effect of genomic structural variations reflected by tandem repeats on recombination frequency in the wheat chromosome 5A.

FISH karyotypes of 5A chromosomes
In wheat, the 5A chromosomes can be distinguished from the other chromosomes based on the signal patterns of Oligo-713, Oligo-pSc119.2-1, and Oligo-pTa535-1 [22,24]. The Oligo-pSc119.2-1 signals on the telomeric region of 5BS (the short arm of 5B chromosome) are stronger than that on 5AS (the short arm of 5A chromosome), and the signal patterns of Oligo-pTa535-1 on 5D chromosomes are different from that on 5A chromosomes [22,24]. However, when Oligo-pSc119.2-1 signals occur on both 5AS and 5AL, it is challenging to distinguish 5A and 2B chromosomes. Then, the 5A chromosomes are identified based on the submetacentric feature. The pericentromeric region of 7AS (the short arm of the 7A chromosome) and 7DS (the short arm of the 7D chromosome) contains Oligo-713 signals, however, the 7A and 7D chromosomes can be distinguished from 5A chromosomes based on the signal patterns of Oligo-pSc119.2-1 and Oligo-pTa535-1 [22].
Researchers have observed a higher rate of motifs of DNA transposons in the recombination intervals. In potato, the Stowaway family of miniature invertedrepeat transposable elements spanned the crossover regions [25]. Darrier et al. observed a higher frequency of a DNA motif specific to the TIR-Mariner DNA transposon in common wheat recombinant intervals [16]. These studies found differences in DNA transposon composition between recombination and nonrecombination hotspots along the same chromosome, a phenomenon observed in diverse populations [16,25].  Putative sequence divergence or insertions/deletions in wheat led to significant differences in crossover frequency along 3B chromosomes between two different F 2 segregating populations [15]. Meanwhile, a nested association mapping (NAM) population used to map the QTL affecting the crossover distribution and frequency indicated a lower recombination rate in the regions with more single nucleotide polymorphisms (SNPs) than with fewer SNPs in wheat [17]. Besides, similar crossover patterns were detected in populations derived from closely related parents [18]. Even in the mutants of anticrossover genes, recombination was prevented in regions with the highest sequence divergence, displayed by SNP [3].
In Arabidopsis, the msh2 (MutS-related heterodimers) mutant displayed significantly reduced SNP enrichment around crossovers compared with the wild type [26]. These studies indicated that SNP polymorphisms and indels are important factors affecting meiotic recombination. Although a significant difference was observed in the recombination rate in the 119-535 interval between   (Fig. 5a), other sequence differences among the 5AL arms, especially in 119-535 and 535-18 intervals, that result in different genetic distances cannot be ignored. In addition to the tandem repeats, these sequence components may also affect the recombination rate between the different 5AL arms. Meanwhile, the recombination rate in the 119-535 interval in the progeny derived from 5AL 535-18/275 × 5AL 119.2-18/275 was higher than that in the progeny derived from 5AL 535-18/275 × 5AL 119.2 . Therefore, the effects of tandem repeats on meiotic recombination should be further investigated.

Effects of juxtaposed heterozygous and homozygous regions on recombination
Although it cannot be concluded that the composition of tandem repeats alone resulted in a comparatively low recombination rate in the 119-535 interval in the progeny of CM39  [27]. Thus, our findings, together with these previous reports, indicate the need to investigate the effects of juxtaposition of heterozygous and homozygous intervals on meiotic recombination.

Effects of tandem repeats on metaphase chromosome condensation
In this study, recombination formed the 5AL 119 + 535 and 5AL No metaphase segments. A significant difference in length was observed only between 5AL 119 + 535 and 5AL No segments derived from the F 2 populations; 5AL 119 + 535 of KCM2 was significantly shorter than that of XKM8 (5AL 119 ), CM36 (5AL 535 ), and CM90 (5AL No ) (Fig. 6b, c). These observations indicate other sequence differences in the 119-535 interval of the 5AL arms, affecting the length of 5AL 119 , 5AL 535 , 5AL 119 + 535 and 5AL No segments. However, for both F 2 populations and wheat cultivars, the 5AL No segments were significantly longer than the 5AL 119 + 535 segments. These findings indicate that the metaphase chromosome segments containing two tandemly repeated clusters close to each other in the same chromosome will be more condensed. Conversely, the metaphase chromosome segments with fewer tandem repeats will be more relaxed. Moreover, the cellSens Dimension software-based method used in this study to measure the length is reliable. Tandemly repeated satellite DNA sequences can drive population and species divergence by inducing alterations in heterochromatin and/or centromere [28]. Satellite DNA sequences are the major heterochromatin components that play an essential role in heterochromatin formation and regulation [29]. In mouse, the heterochromatin alters the loop size of the chromatin of mitotic chromosome [30]. Therefore, the 5AL 119 + 535 segment of this study might be highly heterochromatinized by the aggregating pSc119.2 and pTa-535 making the segment more condensed.

Application of ND-FISH in wheat breeding programs
Understanding the factors affecting the variations in recombination rate and recombination point is important in wheat breeding programs. Generally, crossover frequency decreases from telomere to centromere in eukaryotes [31], including wheat [15,16,32]. Mutations in the anticrossover genes do not increase the recombination in regions close to centromere [3]. However, increasing the interstitial recombination rate reduces deleterious genetic load, which will benefit crop improvement [17]. Therefore, we should understand the mechanisms and methods to increase the recombination rate in the proximal regions [3]. The differences in recombination rate detected in this study reflect differences in sequence composition in 119-535 and 535-18 intervals derived from different crosses. ND-FISH assay displays the differences in tandem repeat composition in wheat chromosomes and can be used to predict the recombination rate. Earlier, Derrier et al. used chromosome 3B pseudomolecule and high-throughput SNP detection to map 252 crossover events at intervals of < 26 kb, a high-resolution crossover location in common wheat [16]. Although high-throughput sequencing and DNA markers have great advantages in studying chromosomal recombination [15][16][17][18]32], large number of individual recombinants need to be sequenced. Cytological markers have advantages of visualization and intuition in studying chromosomal recombination [33][34][35]. ND-FISH technology based on oligo probes can be used to establish abundant FISH karyotypes of wheat chromosomes conveniently [21,36], and oligo probes derived from tandem repeats can display structural variations in wheat chromosomes [22,23,36,37]. Besides, chromosomal recombination indicated by FISH karyotypes is visual comprehend and can be used conveniently in wheat breeding programs [34,35]. Therefore, high-throughput sequencing and DNA markers should be combined with FISH karyotypes to study the wheat chromosomal recombination rules.

Conclusion
The present study confirms the effects of tandem repeats on meiotic recombination in wheat, however, studies should be carried out to identify other factors determining the recombination rate in 5AL arms. Our study provides a robust visual tool based on ND-FISH to measure meiotic recombination and crossover interference in wheat.

Non-denaturing FISH (ND-FISH)
Oligo-pSc119.2-1, Oligo-pTa535-1 [24], Oligo-713, Oligo-275.1 [22], and Oligo-18 [21] were used as oligo probes for ND-FISH. The information on these oligo probes is listed in Additional file 5. These oligo probes can replace their original sequences' roles in identifying wheat chromosomes [22,24]. The sequence of the Oligo-18 probe is identical to the original one, both are 18 bp long [21]. The metaphase chromosomes were prepared from the root tips according to the method described by Han et al. [38]. ND-FISH was performed according to the method described by Tang et al. [21]. An epifluorescence microscope (BX51, Olympus Corporation, Tokyo, Japan) with cellSens Dimension software (Olympus Corporation, Tokyo, Japan) was used to capture images.

Single-copy FISH
The Tandem Repeat Finder (TRF, Version 4.09) [39] program was used to filter the tandem repeats in the 5A chromosome (IWGSC RefSeq Version 2.0). An in-house R package was then used to filter other repeated sequences, including retrotransposon and transposon elements. The remaining sequences were used as the query to align with the full-length sequence of chromosome 5A of the bread wheat variety Chinese Spring (IWGSC RefSeq Version 2.0) using the in-house R package. The query sequences aligned to themselves at a single site were kept as preparatory probes. The sequences of Oligo-pSc119.2-1, Oligo-pTa535-1 [24], and Oligo-18 [21] were also used as the query to align with the fulllength sequence of 5A chromosome using BLAST in the B2DSC web server (http://mcgb.uestc.edu.cn/b2dsc) [40]. The sequences of Oligo-pTa535-1 and Oligo-pSc119.2-1 hit their targets with high copy numbers at 431-432 Mbp and 507-508 Mbp sites on the 5AL arm, respectively. These were used to narrow down the options for single copy probes. Finally, two single-copy sequences, SC5A-479 (479272790-479,275,822 bp) and SC5A-527 (527288389-527,289,202 bp), were selected as probes. The sequence of Oligo-18 hit the target with a high copy number at one site (584-585 Mbp) on the 5AL arm. Then a single copy sequence SC5A-586 (586379936-586,380,903 bp) was selected as the probe. The primer pairs of SC5A-479 (5'TCGTTGACTAGA AAGACGTG TGT3', 5'ACGCCTGTGTTAAGTTAA GTGAC3'), SC5A-527 (5'TGCGTACATAGGGTGAGT GTATG3', 5'GGCCTCTGGAAGAACGTTTTAT3'), and SC5A-586 (5'TTGCTCGTGTCCACCATTGA3', 5'TGTGGAATACTTACCGCGCA3') were used to amplify the three single-copy probe sequences. The target sequences were cloned into the TSINGKE pClone007 vector (TSINGKE, Chengdu, China) and labeled with Texas-Red-5-dUTP (PerkinElmer, USA) according to the method described by Han et al. [38]. The root tip metaphase chromosomes were prepared, and the hybridization was performed as described by Han et al. [38] with slight modifications. The probe mixture contained 5-6 ng/μL of each probe and 1 × ENZO buffer (ENZO Life Science Inc. USA). Slides were washed in 2 × SSC buffer containing 0.1% NP-40 detergent (Solarbio Life Sciences Ltd. China) for 3-5 min at 45-50°C. An epifluorescence microscope (BX51, Olympus Corporation, Tokyo, Japan) with cellSens Dimension software (Olympus Corporation, Tokyo, Japan) was used to capture the images.
Measurement of length between SC5A-479 and SC5A-527 probes on metaphase chromosomes The length of metaphase 5AL segment (MSL) between the signals of the two single-copy probes SC5A-479 and SC5A-527 was measured. Four kinds of metaphase 5AL segments were measured: segments with Oligo-pSc119.2-1 signal (named 5AL 119 ), segments with Oligo-pTa535-1 signal (named 5AL 535 ), segments with signals of the two probes (named 5AL 119 + 535 ), and segments without the signals of the two probes (named 5AL No ). These four kinds of segments were derived from the four F 2 generations, wheat cultivars CM36, CM90, XKM8, and KCM2. The length of the metaphase 5AL segment between SC5A-479 and SC5A-527 (MSL) was determined as relative metaphase length (RML) to avoid errors caused by chromosome condensation. RML = MSL/WL where WL represents the whole length of the 5AL arm from the central point of the centromere to the distal end. The MSL and WL were measured using cellSens Dimension software (Olympus Corporation, Tokyo, Japan). A one-way ANOVA was carried out using GraphPad Prism software (Version 5) for pairwise comparisons among the groups. Graphs were also plotted using GraphPad Prism software.