Physical Localization of a Locus from Agropyron cristatum Conferring Resistance to Stripe Rust in Common Wheat

Stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is one of the most destructive diseases of wheat (Triticum aestivum L.) worldwide. Agropyron cristatum (L.) Gaertn. (2n = 28, PPPP), one of the wild relatives of wheat, exhibits resistance to stripe rust. In this study, wheat-A. cristatum 6P disomic addition line 4844-12 also exhibited resistance to stripe rust. To identify the stripe rust resistance locus from A. cristatum 6P, ten translocation lines, five deletion lines and the BC2F2 and BC3F2 populations of two wheat-A. cristatum 6P whole-arm translocation lines were tested with a mixture of two races of Pst in two sites during 2015–2016 and 2016–2017, being genotyped with genomic in situ hybridization (GISH) and molecular markers. The result indicated that the locus conferring stripe rust resistance was located on the terminal 20% of 6P short arm’s length. Twenty-nine 6P-specific sequence-tagged-site (STS) markers mapped on the resistance locus have been acquired, which will be helpful for the fine mapping of the stripe rust resistance locus. The stripe rust-resistant translocation lines were found to carry some favorable agronomic traits, which could facilitate their use in wheat improvement. Collectively, the stripe rust resistance locus from A. cristatum 6P could be a novel resistance source and the screened stripe rust-resistant materials will be valuable for wheat disease breeding.


Introduction
Stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is one of the most devastating and widespread diseases of wheat (Triticum aestivum L.) around the world [1][2][3]. Stripe rust has become a major threat to wheat production, causing yield losses of 5-25% [4,5]. In recent years, new virulent Pst races appeared in a short period of time so that many wheat varieties were ineffective against prevalent races [6]. Therefore, there is a need to screen new stripe rust resistance genes for broadening the wheat gene pool and providing new potential resistance genes for the wheat breeding of stripe rust resistance.
Wide hybridization is an efficient way of transferring beneficial resistance genes to common wheat. Different resistance genes from distant genera, such as Yr9 from Secale cereal [7,8], Pm21 from Haynaldia villosa [9] and YrH9020 from Psathyrostachys huashanica [10], have been transferred to common wheat, and were effective in enhancing the disease resistance of wheat. The disease-resistant wheat-alien species derivative lines will be used as new wheat resources for breeding new resistant varieties.
To map alien desirable genes, a series of wheat-alien translocation lines and deletion lines were produced to construct the physical map. Qi et al. [24] constructed the physical mapping of Haynaldia villosa 6V using one spontaneous deletion line and two deletion lines induced by irradiation. Ashida et al. [25] created chromosomal breaks in the offspring of a 5H addition line induced by the gametocidal chromosome and constructed the physical map of barley chromosome 5H. Wheat-A. cristatum translocation lines and deletion lines have been developed through the irradiation of wheat-A. cristatum disomic addition lines, and high-resolution physical maps of A. cristatum 2P, 6P and 7P have been constructed [21,22,[26][27][28][29]. The physical maps and the introgression lines can provide a solid foundation for the exploration and utilization of elite genes from the wide relatives of wheat.
In this study, wheat-A. cristatum 6P derivative lines and the populations were used to map the stripe rust resistance locus from A. cristatum 6P to the specific chromosomal region. The resistance locus could be a novel stripe rust-resistant source and the screened resistant materials will be valuable for wheat disease breeding.

Molecular Cytogenetic Analysis and Chromosomal Arm Localization of the Locus Conferring Resistance to Stripe Rust
A series of wheat-A. cristatum 6P translocation lines and deletion lines have been developed through irradiation of the addition line 4844-12 [26,29]. Ten translocation lines and five deletion lines were used to map the stripe rust resistance locus. All of these lines contained a pair of segments of A. cristatum 6P detected by GISH ( Figure 2). A. cristatum 6P-specific STS molecular markers were used to trace different chromosomal segments of A. cristatum 6P to confirm the constitutions of P chromatin in each line ( Figure 3).

Molecular Cytogenetic Analysis and Chromosomal Arm Localization of the Locus Conferring Resistance to Stripe Rust
A series of wheat-A. cristatum 6P translocation lines and deletion lines have been developed through irradiation of the addition line 4844-12 [26,29]. Ten translocation lines and five deletion lines were used to map the stripe rust resistance locus. All of these lines contained a pair of segments of A. cristatum 6P detected by GISH ( Figure 2). A. cristatum 6P-specific STS molecular markers were used to trace different chromosomal segments of A. cristatum 6P to confirm the constitutions of P chromatin in each line ( Figure 3).     To identify the arm carrying the resistance locus, four wheat-A. cristatum 6P whole-arm translocation lines (WAT639b, WAT638a, WAT639a and WAT638b) and two 6P ditelosomic addition lines (del21 and del20) were tested using a mixture of the Pst races in two sites. Among them, all lines carrying chromosome arm 6PS were resistant; all lines carrying chromosome arm 6PL were susceptible (Table 1, Figure 1). This identifies the short arm of 6P as the carrier of the resistance locus. To identify the arm carrying the resistance locus, four wheat-A. cristatum 6P whole-arm translocation lines (WAT639b, WAT638a, WAT639a and WAT638b) and two 6P ditelosomic addition lines (del21 and del20) were tested using a mixture of the Pst races in two sites. Among them, all lines carrying chromosome arm 6PS were resistant; all lines carrying chromosome arm 6PL were susceptible (Table 1, Figure 1). This identifies the short arm of 6P as the carrier of the resistance locus. The BC 2 F 2 and BC 3 F 2 populations of WAT639a and WAT639b were constructed to confirm the resistance locus on chromosome 6PS. P genome-specific molecular markers [30] were used to detect the populations (Figure 4). In the BC 2 F 2 population (2015-2016 season) of the line WAT639b, 22 plants carried 6PS arm, which were resistant; while 32 did not carried 6PS, which were susceptible. In the BC 3 F 2 population (2016-2017 season) of the line WAT639b, 97 plants carried 6PS arm, which were resistant; while 52 did not carry 6PS, which were susceptible (Table 2). Chi squared (χ 2 ) tests for independence showed that stripe rust resistance was significantly affected by the 6PS arm. In the populations of the line WAT639a, no matter whether the plants contained the 6PL arm or not, all were susceptible, suggesting that there was no correlation between stripe rust resistance and the 6PL arm. Therefore, the 6PS arm surely carried the resistance locus.
centromere was considered as 0, while the terminal end of the 6PS/6PL arm was considered as 1. The letters R and S in "Stripe rust response" column indicated plants were resistant and susceptible, respectively.
The BC2F2 and BC3F2 populations of WAT639a and WAT639b were constructed to confirm the resistance locus on chromosome 6PS. P genome-specific molecular markers [30] were used to detect the populations (Figure 4). In the BC2F2 population (2015-2016 season) of the line WAT639b, 22 plants carried 6PS arm, which were resistant; while 32 did not carried 6PS, which were susceptible. In the BC3F2 population (2016-2017 season) of the line WAT639b, 97 plants carried 6PS arm, which were resistant; while 52 did not carry 6PS, which were susceptible (Table 2). Chi squared (χ 2 ) tests for independence showed that stripe rust resistance was significantly affected by the 6PS arm. In the populations of the line WAT639a, no matter whether the plants contained the 6PL arm or not, all were susceptible, suggesting that there was no correlation between stripe rust resistance and the 6PL arm. Therefore, the 6PS arm surely carried the resistance locus.

Evaluation of Agronomic Traits of the Stripe Rust-Resistant Translocation Lines
The spike agronomic traits of six wheat-A. cristatum 6P translocation lines (WAT638a, WAT639b, WAT641a, WAT648, WAT655 and WAT646) conferring stripe rust resistance were evaluated at BC2F3 progeny, including spike length, spikelet number per spike, kernel number per spikelet, grain number per spike and thousand-grain weight in this study ( Table 3). The seeds of the lines WAT638a, WAT639b, WAT648, WAT655 and WAT646 displayed longer and wider size than that of Fukuhokomugi (Figure 7), so that these translocation lines exhibited higher thousand-grain weights. Compared to Fukuhokomugi, the lines WAT655 and WAT646 showed higher grain number per spike, contributed by the higher spikelet number per spike based on Duncan's multiple-range test.

Evaluation of Agronomic Traits of the Stripe Rust-Resistant Translocation Lines
The spike agronomic traits of six wheat-A. cristatum 6P translocation lines (WAT638a, WAT639b, WAT641a, WAT648, WAT655 and WAT646) conferring stripe rust resistance were evaluated at BC 2 F 3 progeny, including spike length, spikelet number per spike, kernel number per spikelet, grain number per spike and thousand-grain weight in this study ( Table 3). The seeds of the lines WAT638a, WAT639b, WAT648, WAT655 and WAT646 displayed longer and wider size than that of Fukuhokomugi (Figure 7), so that these translocation lines exhibited higher thousand-grain weights. Compared to Fukuhokomugi, the lines WAT655 and WAT646 showed higher grain number per spike, contributed by the higher spikelet number per spike based on Duncan's multiple-range test.  The left map showed that the diagram of A. cristatum chromosome 6P, which was modified in the physical map of A. cristatum 6P as described by Song et al. [29]. Pink and blue colors represented A. cristatum and wheat chromosomal segments, respectively. Dotted line boxes indicated the missing segments of 6P chromosome. The letters R and S indicated materials were resistant and susceptible, respectively. Figure 6. Chromosomal localization of the novel stripe rust resistance locus from A. cristatum 6P and chromosome diagrams of the deleted and the translocated 6P chromosomes. The left map showed that the diagram of A. cristatum chromosome 6P, which was modified in the physical map of A. cristatum 6P as described by Song et al. [29]. Pink and blue colors represented A. cristatum and wheat chromosomal segments, respectively. Dotted line boxes indicated the missing segments of 6P chromosome. The letters R and S indicated materials were resistant and susceptible, respectively.

Discussion
Broadening the genetic base of common wheat by transferring resistance genes from wide relatives may enhance the chance of achieving adequate resistance against stripe rust. Wheat-A. cristatum 6P disomic addition line 4844-12 was immune to the infection with stripe rust (CYR32 and CYR33). Six translocation lines WAT638a, WAT639b, WAT641a, WAT648, WAT655 and WAT646 were highly resistant to stripe rust. Among these lines, WAT646 was slightly different from others in the response to stripe rust ( Figure 5). The leaves of WAT646 carried few spores of Pst. There may be two reasons for this difference. Firstly, WAT655 was a homoeologous translocation line, because the 6P segment was translocated to 6D, while the 6P segment of WAT646 was translocated to 1B [29]. Compensating translocations between homoeologous wheat and alien segments are favorable and beneficial for wheat improvement [31], so that WAT655 may display better complementary than WAT646. Secondly, the 6PS (0.81-1.00) of WAT655 was larger than the 6PS (0.86-1.00) of WAT646, so the 6P segment of WAT655 may contain more resistance genes than that of WAT646. Therefore, we localized the resistance locus on the 6P segment of WAT655. Collectively, the six translocation lines exhibit high resistance to stripe rust and can be the new stripe rust-resistant resources for resistance breeding.
A. cristatum, as a favorable genetic resource, carried various disease resistance genes [16,22,23]. The leaf rust resistance locus from A. cristatum has been transferred to common wheat [16]. Copete et al. [23] utilized the addition lines to confirm A. cristatum 2PL and 6PL carried powdery mildew resistance gene(s). Song et al. [26] mapped a leaf rust resistance locus of A. cristatum on 6PS (0.81-1.00) using A. cristatum 6P deletion lines. In this study, the stripe rust resistance locus was found to be located on 6PS (0.81-1.00). The leaf rust resistance locus and the stripe rust resistance locus from A. cristatum 6P were coincidently located on the same chromosomal segment. The translocation lines (WAT639b, WAT638a, WAT648 and WAT655) were resistant to stripe rust in this research, and they also were resistant to leaf rust. Therefore, the resistance locus mapped on 6PS (0.81-1.00) may be a broad-spectrum resistance locus, and we speculate that there may be one favorable disease resistance gene cluster on 6PS (0.81-1.00).
Molecular marker technique has been a considerably efficient and convenient method to detect the alien genomic component. A series of P genome-specific STS markers were designed through EST sequences of A. cristatum transcriptome sequences [32]. Song et al. [20] mapped 255 STS markers on the physical map of A. cristatum 6P. These markers were used to trace the 6P segments of translocation lines and deletion lines in this research. Twenty-nine STS markers were mapped on the region of the resistance locus. These 6P-specific STS markers will be useful for screening disease-resistant materials in wheat breeding, which will provide a basis for fine mapping of the wheat rust resistance locus in future work.
In recent years, many wheat varieties have lost resistance against stripe rust due to the deficiency of resistance genes and variation of Pst races. The emergence of two prevalent physiological races (CYR32 and CYR33) have resulted in the loss of stripe rust resistance of more varieties. However, only a few resistance genes conferred still resistance to stripe rust in common wheat, such as Yr5, Yr10, Yr15, Yr24/Yr26 and Yr50 [33][34][35][36]. Therefore, continuous exploration of new stripe rust resistance genes, particularly wide-spectrum genes, will be imperative, which will enhance the diversity of stripe rust resistance gene. Transferring stripe rust resistance genes from the wild relatives to common wheat has been an effective approach to enhance the agronomic performance of wheat [37]. Wheat-rye 1BL·1RS translocation line possesses the stripe rust resistance gene Yr9 [7,8]. The stripe rust resistance genes from emmer wheat, such as Yr15 [38], Yr35/Lr52 [39] and Yr30/Sr2 [40], have been transferred to common wheat. In this research, the stripe rust resistance locus of A. cristatum 6P has been transferred into common wheat in translocation line form. The stripe rust resistance locus is a broad-spectrum resistance locus, which can be used for genetic improvement of wheat as a new resistance source. The stripe rust-resistant lines with favorable agronomic traits can be utilized in wheat breeding as new disease-resistant wheat germplasms. The small segmental lines (WAT648, WAT655 and WAT646) will be as the basis for further exploring the stripe rust resistance locus through RNA-Sequence in future research.
In summary, the novel stripe rust resistance locus from A. cristatum 6P was located on the region 6PS (0.81-1.00) using homozygous strains and the populations. The stripe rust-resistant translocation lines will be used for wheat disease-resistant breeding as new germplasms and for fine mapping of the novel stripe rust resistance locus as the basic materials.
anti-digoxigenin-rhodamine (red) were purchased from Roche, Mannheim, Germany. Signals were observed using an OLYMPUS AX80 (Olympus Corporation, Tokyo, Japan) fluorescence microscope. Images were captured with a CCD camera (Diagnostic Institute, Inc., Sterling Height, MI, USA) and processed with Photoshop CS 3.0. P genome-specific markers and 6P-specific STS markers [30,32] were used to detect translocation lines and deletion lines. P genome-specific markers can trace the P genomic component in wheat background, which were designed through specific DNA sequences distributing the entire P genome [30]. Zhang et al. [32] designed the 6P-specific STS markers according to the EST sequences from A. cristatum transcriptome. Song et al. [30] mapped 255 6P-specific STS markers on the physical map using the deletion lines and translocation lines. We used 6P-specific STS markers to trace the different 6P segments of translocation lines and deletion lines.

Evaluation of Stripe Rust Resistance at Adult Stage
Homozygous materials and the BC 2 F 2 and BC 3 F 2 populations of WAT639a and WAT639b were planted in a random complete block design with three replicates in the fields of Yangling (34 •  Host responses to infection were recorded when leaves of Huixianhong were fully rusted [45]. The infection type (IT) of each plant was recorded based on 0-9 rating scale, with 0 as immune (no visible signs), 1-2 as high resistance (no or few sporulation), 3-4 as intermediate resistance (trace sporulation), 5-6 as intermediate susceptibility (intermediate sporulation), 7-8 as susceptibility (abundant sporulation) and 9 as high susceptibility (no necrosis or chlorosis; abundant sporulation). Plants with IT 0-4 were considered resistant, while plants with IT 5-9 were considered susceptible.

Statistical Analysis of Field Experiment
Statistical Analysis System (Version 9.2, SAS Institute, Cary, NC, USA) was used for statistical analysis in this study. Plants in each population were classified to two types according to molecular marker analysis: plants with P genome-specific markers and plants without P genome-specific markers. Chi-squared (χ 2 ) tests for independence were used to determine the association between 6P segments and response to stripe rust.
The translocation lines were manually harvested at the maturity stage. We measured and counted the spike agronomic traits including spike length, spikelet number per spike, kernel number per spikelet, grain number per spike and thousand-grain weight. The analysis of variance was performed to test the difference between the translocation lines and the parent Fukuhokomugi in the agronomic traits.

Conflicts of Interest:
The authors declare no conflict of interest.