The use of the ph1b mutant to induce recombination between the chromosomes of wheat and barley

Intensive breeding has led to a narrowing in the genetic base of our major crops. In wheat, access to the extensive gene pool residing in its many and varied relatives (some cultivated, others wild) is hampered by the block on recombination imposed by the Ph1 (Pairing homoeologous 1) gene. Here, the ph1b mutant has been exploited to induced allosyndesis between wheat chromosomes and those of both Hordeum vulgare (cultivated barley) and H. chilense (a wild barley). A number of single chromosome Hordeum sp. substitution and addition lines in wheat were crossed and backcrossed to the ph1b mutant to produce plants in which pairing between the wheat and the non-wheat chromosomes was not suppressed by the presence of Ph1. Genomic in situ hybridization was applied to almost 500 BC1F2 progeny as a screen for allosyndetic recombinants. Chromosome rearrangements were detected affecting H. chilense chromosomes 4Hch, 5Hch, 6Hch, and 7Hch and H. vulgare chromosomes 4Hv, 6Hv, and 7Hv. Two of these were clearly the product of a recombination event involving chromosome 4Hch and a wheat chromosome.


Introduction
Bread wheat (Triticum aestivum) is one of the most important food crops of the world, and continuous improvement in its productivity will be required to keep pace with global population growth. The genetic base of the species is rather narrow, as its speciation was very recent (Salamini et al., 2002;Riehl et al., 2013). However, a large number of sexually compatible species (some wild and some cultivated) are known, and these represent a much needed reservoir of potentially exploitable genetic variation.
The genome of an interspecific or (intergeneric) hybrid combines the haploid complements of each of its sexual parents. Even though their genomes are closely related to one another, in most cases, the chromosomes of wheat and those of its relatives fail to pair with one another and thus allosyndetic recombination is rare. The failure of homoeologs (chromosomes from related genomes but not completely homologous) to pair at meiosis is ensured by the wild type allele at the Ph1 locus (Riley and Chapman, 1958;Sears and Okamoto, 1958;Sears, 1976). This gene imposes diploid-like chromosome behavior during meiosis, even though the constituent sub-genomes of this hexaploid species are known to be very closely related to one another. Deletion of the Ph1 locus allows homoeologs to pair relatively freely with one another (Moore, 2014), a situation which  has been exploited for introgression purposes through the use of the ph1b mutant (Riley et al., 1968b;Sears, 1977Sears, , 1981Sears, , 1982Khan, 1999;Lukaszewski, 2000;Qi et al., 2008;Liu et al., 2011;Zhao et al., 2013). Hordeum chilense, a species which is readily crossable with wheat, is a diploid relative of cultivated barley. It has been identified as a potential donor to wheat for a number of traits of agronomic interest (Martín et al., 1998. The bread wheat × H. chilense hybrid has been the source of a collection of single (Hordeum) chromosome addition lines and chromosome substitution lines in a bread wheat genetic background (Miller et al., 1982), and similar cytogenetic stocks have been developed involving the cultivated barley (H. vulgare) chromosomes (Islam et al., 1978(Islam et al., , 1981. The self-fertile amphidiploid Tritordeum represents the product of chromosome doubling of the hybrid T. turgidum × H. chilense (Martín and Sanchez-Mongelaguna, 1982). The presence of Ph1 maintains the integrity of Hordeum sp. chromosome(s) in all of this germplasm, meaning that the introgression of favorable non-wheat genes is inevitably accompanied by the inheritance of a large number of unwanted ones. The experience with introgression into wheat from other related species suggests that this linkage drag can best be overcome by employing a ph1b-based strategy. Here, we describe progress made with an introgression program using the ph1b mutant to induce chromosome pairing and recombination between the chromosomes of H. chilense or H. vulgare, and those of wheat.      (Islam et al., 1978(Islam et al., , 1981Miller et al., 1982) used as the female parent in crosses with the ph1b mutant (Sears, 1977). Grains were germinated on wet filter paper in the dark for 5 days at 4 • C, followed by a period of 24 h at 25 • C. Emerging seedling roots were excised, incubated for 4 h in 0.05% w/v colchicine at 25 • C, fixed in Carnoy's solution (three parts 100% ethanol plus one part glacial acetic acid), and finally stored at 4 C for at least 1 month. The plants were subsequently raised in a greenhouse held at 26 C during the day and 22 • C during the night (16 h photoperiod). Immature spikes were fixed in Carnoy's solution and used to characterize chromosome pairing at meiosis metaphase I.

DNA Marker Characterization
Genomic DNA was extracted from frozen seedling leaves following Murray and Thompson (1980), as modified by Hernández et al. (2001). The absence of Ph1 was verified using a PCR assay described by Wang et al. (2002). Each 30 µL PCR contained 1x PCR buffer with MgCl 2 (Bioline USA, Taunton, MA, USA), 0.25 mM dNTP, 0.17 µM primers, 0.02 U/µL Taq DNA polymerase (Bioline USA), and 20 ng template. The reaction was first denatured (94 • C/5 min), and then subjected to 35 cycles of 94 • C/60 s, 51 • C/60 s, and 72 • C/60 s, followed by a final extension (72 • C/7 min). The PCR products were electrophoretically separated through a 1% agarose gel and visualized by EtBr staining. The presence of each Hordeum sp. chromosome was based on PCR assays described by Liu et al. (1996) and Hagras et al. (2005) as detailed in Table 2. The composition of these PCR reactions was as above, while the amplification regime comprised an initial denaturing step (94 • C/5 min), followed by 35 cycles of 94 • C/15 s, 50-65 • C (primer dependent, see Table 2) /30 s, 72 • C/60 s, and completed by a final extension (72 • C/6 min). The amplicons were separated as described above.

Cytogenetic Analysis
Chromosome spreads were prepared from both pollen mother cells (PMCs) at meiotic metaphase I and from root tip cells.   The material was macerated in a drop of 45% glacial acetic acid, squashed under a cover slip, and dipped in liquid nitrogen in order to remove the cover slip. The preparations were then airdried and either processed directly for in situ hybridization, or stored at 4 • C until required. The probe used for genomic in situ hybridization was genomic DNA extracted from H. chilense (or H. vulgare) seedling leaves. The DNA was labeled with either biotin-11-dUTP (H. vulgare) or digoxigenin-11-dUTP (H. chilense; both from Roche Corporate, Basel, Switzerland) by nick-translation. The in situ hybridization protocol followed that described by Prieto et al. (2004). The GAA-satellite sequence (Pedersen et al., 1996) and the pAs1 probe (Rayburn and Gill, 1986) were used to identify chromosomes involved in homoeologous pairing, chromosomal translocations, or chromosomal rearrangements. The GAA-satellite sequence identifies all the A and B wheat chromosomes (Pedersen and Langridge, 1997), whereas the pAs1 identifies the D wheat and the H. chilense chromosomes (Cabrera et al., 1995). The GAA-satellite sequence and the pAs1 probes were also labeled by nick translation with biotin-11-dUTP and digoxigenin-11-dUTP, respectively. Biotin-or digoxigenin-labeled DNA were detected using, respectively, streptavidin-Cy3 (Sigma, St. Louis, MO, USA) and antidigoxigenin-FITC (Roche Applied Science, Indianapolis, IN, USA). After counter-staining with DAPI (4 , 6-diamidino-2-phenylindole), the preparations were mounted in Vectashield (Vector Laboratories, Burlingame, CA, USA). Hybridization signals were visualized using a Nikon Eclipse 80i epifluorescence microscopy, and the images captured with a CCD camera (Nikon Instruments Europe BV, Amstelveen, The Netherlands).

Statistical Methods
Statistical analyses were performed using the STATISTIX v9.0 software (Analytical Software, Tallahassee, FL, USA). Wilcoxon (or U of Mann-Whitney) test was used to determine the statistical significance of differences between means.

Converting the Substitution and Addition Lines into a ph1b Mutant Background
The crossing scheme used is illustrated in Figure 1, and the details of the crossing outcomes from the F1 to the BC 1 F 2 generation are given in Table 1. The F 1 hybrid progeny were genotyped by PCR to ensure that they had retained the expected Hordeum sp. chromosome (Table 2; Figure 2A), then crossed again to the ph1b mutant in order to establish individuals in which the Hordeum sp. chromosome was now present in a ph1bph1b background. Zygosity at the Ph1 locus was predicted using a PCR assay ( Figure 2B). The meiotic behavior of the selected individuals was characterized by GISH analysis of metaphase I in PMCs, and the plants were allowed to self-pollinate.

Allosyndetic Pairing in BC 1 F 1 Selections Lacking Ph1
Meiosis was characterized in 63 BC 1 F 1 segregants carrying a Hordeum chromosome in the absence of Ph1 and compared to those carrying the Hordeum chromosome in its presence ( Table 3). No wheat/Hordeum chromosome pairing occurred in plants of genotype Ph1Ph1 (Table 3; Figures 3A,D). In contrast, in the absence of the Ph1 locus, although the Hordeum chromosomes remained unpaired in most metaphase I PMCs (Figures 3B,E), pairing was observed in 1.77% of the PMCs in H. chilense (Table 3; Figure 3C). The equivalent frequency with respect to H. vulgare chromosomes was 1.84% (Table 3; Figure 3F). The frequency of plants displaying wheat/Hordeum chromosome associations was lower in H. chilense than in H. vulgare (45.23% and 61.90%, respectively), although variability depending on the specific Hordeum sp. chromosome introgressed was found. Most of the associations between a Hordeum and a wheat chromosome involved the formation of a rod bivalent harboring a single sub-terminal chiasma (Figures 4A-A"), although in some cases the chiasma occurred more proximally (Figures 4B-B"). In a few PMCs, the Hordeum sp. chromosome formed part of a multivalent (Figures 4C-C") as the result of chiasmata between homoeologous chromosomes, or reflecting the re-arrangement of the wheat genome induced by successive meiosis during the generations of selfing used to maintain the ph1b mutant stock. Wilcoxon test showed that the frequency of allosyndesis was not Hordeum sp. chromosome specific, since there was no significant difference in pairing frequency between either chromosomes 4H ch , 6H ch , and 7H ch or between chromosomes 4H v , 6H v , and 7H v (Table 4A). In addition, using the same statistical test, no significance differences where found when compared the effect of the genome (H. chilense or H. vulgare) for the same homoeologous group (p = 0.39, 0.41, and 0.70 for chromosomes 4, 6, and 7, respectively; Table 4A). A statistical comparison of chromosome pairing frequency involving a H. chilense chromosome and each of its wheat homoeologs was also carried out and showed no evidence for any preferential pairing (Table 4B).

Genetic Evidence for Hordeum sp. Introgression Induced by the Absence of Ph1
A total of 473 BC 1 F 2 progeny were analyzed by GISH analysis to detect and characterize Hordeum sp. chromosome rearrangements in the background of the ph1b mutant. About 60% of the progeny lacked any Hordeum sp. chromatin. Overall, with respect to the Hordeum sp. chromosome, about 3% of the progeny were disomic and about 33% were monosomic. The highest transmission rate of a Hordeum chromosome was observed among the progeny derived from the (4B) 4H ch substitution line. Two recombinants were identified, both involving chromosomes 4H ch and 4D (

Discussion
Interspecific hybridization retains its potential to widen the gene pool available to the wheat breeder. Combining in situ hybridization with DNA-based genotyping has eased the process considerably since the initial efforts which followed the recognition that recombination could be induced by the deletion of Ph1 (Koebner and Shepherd, 1986;Qi et al., 2007). An in situ hybridizationbased screening strategy has previously been applied to characterize introgressions from both H. chilense and H. vulgare, resulting in the recognition of a number of wheat/Hordeum sp. translocations . Here, the intention was to exploit the abolition of strict homologous pairing induced by the absence of Ph1 to generate material where recombination had shortened the length of the introgressed segment. Chromosome 4H ch is of particular interest as it harbors a gene (or possibly genes) encoding resistance against the fungal pathogen Septoria tritici . Two recombinants involving chromosome 4H ch were obtained in this work as the results of the same recombination event between 4DL and 4H ch L chromosome arms, and can help to locate those resistance genes on chromosome 4H ch L. Similarly, chromosome 7H ch has been targeted for its positive effect on grain carotenoid content (Alvarez et al., 1999), and chromosome 5H ch for its contribution to enhancing salinity tolerance (Forster et al., 1990). Although inter-chromosome translocations are known to occur spontaneously (Mettin et al., 1973;Zeller, 1973;Prieto et al., 2001), and can be induced by ionizing radiation and the action of certain gametocidal genes (Sears, 1956(Sears, , 1993Endo, 1988Endo, , 1990Endo and Gill, 1996), the particular advantage of exploiting the ph1b mutant to promote allosyndesis is that the translocations are non-random: rather, they tend to involve the exchange of genetically related material. Its disadvantage is that the frequency of allosyndesis (and hence of recombination) is rather low, especially between chromosomes of more distantly related genomes such as Triticum and Hordeum.
The level of ph1b-induced pairing between wheat and cereal rye (Secale cereale) chromosomes has been estimated to be around 4% (Miller et al., 1994), which is about double the level noted here between the chromosomes of wheat and either of the two Hordeum sp. Moreover, the frequency of recombination was correlated with the frequency of wheat-rye pairing in metaphase I in ABDR hybrids in the absence of the Ph1 locus (Naranjo and Fernández-Rueda, 1996). However, an extensive ph1b-based attempt to reduce the length of the rye chromosome segment present in the widely used wheat/rye Robertsonian translocation 1BL.1RS resulted in an estimated recombination frequency of only around 0.7% (Koebner and Shepherd, 1986;Lukaszewski, 2000). The levels achievable in more closely related species, notably in the genus Aegilops (Riley et al., 1968a;Gill and Raupp, 1987;Koebner and Shepherd, 1987;Farooq et al., 1990;Ceoloni et al., 1992), are much higher than this. Our results showed that homoeologous recombination between Hordeum sp. and wheat chromosomes did only depend on the absence of the Ph1 locus as no differences in the frequency of pairing were found when chromosome association in different homoeologous groups was studied. Most of chromosome associations between Hordeum sp. and wheat chromosomes were end-to-end extremely distal associations as described previously (Werner et al., 1992;Benavente et al., 1996;Calderón et al., 2014).
In summary, the use of the ph1b mutant does induce a low, but significant level of chromosome pairing and recombination between wheat and Hordeum sp. chromosomes. The translocation and introgression chromosomes detected in the present work will serve as potential donor material for the breeding of cultivars having a higher grain carotenoid content, stronger resistance against S. tritici and improved salinity tolerance.

Author Contributions
M-DR, MC, and PP carried out the experiments and analyzed the data. M-DR and PP planned the study and wrote the manuscript. All authors read and approved the final manuscript.