Physical Mapping of Pm57, a Powdery Mildew Resistance Gene Derived from Aegilops searsii

Powdery mildew caused by Blumeria graminis f. sp. tritici (Bgt) is one of many severe diseases that threaten bread wheat (Triticum aestivum L.) yield and quality worldwide. The discovery and deployment of powdery mildew resistance genes (Pm) can prevent this disease epidemic in wheat. In a previous study, we transferred the powdery mildew resistance gene Pm57 from Aegilops searsii into common wheat and cytogenetically mapped the gene in a chromosome region with the fraction length (FL) 0.75–0.87, which represents 12% segment of the long arm of chromosome 2Ss#1. In this study, we performed RNA-seq using RNA extracted from leaf samples of three infected and mock-infected wheat-Ae. searsii 2Ss#1 introgression lines at 0, 12, 24, and 48 h after inoculation with Bgt isolates. Then we designed 79 molecular markers based on transcriptome sequences and physically mapped them to Ae. searsii chromosome 2Ss#1- in seven intervals. We used these markers to identify 46 wheat-Ae. searsii 2Ss#1 recombinants induced by ph1b, a deletion mutant of pairing homologous (Ph) genes. After analyzing the 46 ph1b-induced 2Ss#1L recombinants in the region where Pm57 is located with different Bgt-responses, we physically mapped Pm57 gene on the long arm of 2Ss#1 in a 5.13 Mb genomic region, which was flanked by markers X67593 (773.72 Mb) and X62492 (778.85 Mb). By comparative synteny analysis of the corresponding region on chromosome 2B in Chinese Spring (T. aestivum L.) with other model species, we identified ten genes that are putative plant defense-related (R) genes which includes six coiled-coil nucleotide-binding site-leucine-rich repeat (CNL), three nucleotide-binding site-leucine-rich repeat (NL) and a leucine-rich receptor-like repeat (RLP) encoding proteins. This study will lay a foundation for cloning of Pm57, and benefit the understanding of interactions between resistance genes of wheat and powdery mildew pathogens.


Introduction
Common wheat (Triticum aestivum L., 2n = 6x = 42, AABBDD) is the most widely grown food crop, a cash crop with world trade greater than that of all other crops combined, and also a worldwide staple food, providing valuable plant protein for humans. Wheat is susceptible to more than 30 diseases, of which powdery mildew, caused by the fungus Blumeria graminis f. sp. tritici (Bgt), current study, we have performed RNA-seq using Chinese Spring (CS)-Ae. searsii introgression lines to develop 2S s #1 specific markers and identified ph1b-induced 2S s #1 recombinants to physically map Pm57 into a small genomic region.

RNA-seq Quantity, Sequence Assembly, and Differential Expression Gene (DEGs) Analysis
A total of 94.88-111.83 × 10 6 high-quality reads, which constituted 10. 19-16.35 Gb of the cDNA sequences were generated for each sample in this study. Sequencing quality scores, Q30, which infers a base call accuracy of 99.90%, were more than 93.49% for each sample, signifying that RNA-seq quality in the study is sufficient for subsequent sequence assembly. The high-quality reads were further assembled into 71,313 unigenes (46.87 Gb) of 331 bp median length, ranging from 201 to 63,334 bp in length, using the short-read assembly software Trinity [49].
Pairwise comparison of the unigenes in CS at 0 h post-inoculation (hpi) using Bgt isolates (RNA-seq library referred as A-CS) and at 12 Table S2). Comparing the unigenes of all three Pm57 lines (2011-400, 89-88, and 89-69) against those of CS (Supplementary Tables S1 and S2), 500 candidate DEGs were selected for the design of specific PCR primers for chromosome 2S s #1 and 2S s #1 long-arm segment of FL0.67-0.87.

Validation of DEGs by Quantitative RT-PCR (qRT-PCR)
To validate the DEGs based on RNA-seq analysis, 10 Bgt-infection induced DEGs described as "plant disease resistance-related protein" in the protein family (Pfam) database were further evaluated by qRT-PCR in the Bgt-resistant recombinants 89-88 and 89-69, and susceptible CS. Primer pairs were designed for qRT-PCR for each of the 10 DEGs (Supplementary Table S3). The qRT-PCR results ( Figure 1, left side of each graph) showed that expression levels of seven (comp181267_c0, comp69335_c0, comp45021_c0, comp53009_c0, comp23383_c0, comp57561_c0 and comp70653_c0) of ten DEGs reached to a maximum at 24 hpi. One gene (comp33120_c0) was highest at 12 hpi. The unigene comp61771_c0 caused significant up-regulation at 12 hpi and 48 hpi, whereas comp65225_c0 was up-regulated at three time points (12, 24 and 48 hpi) in resistant lines, but down-regulated in the susceptible CS (Figure 1,  Supplementary Table S4). Although gene expression levels of line 89-69 evaluated by qRT-PCR were not exactly the same as those identified based on RNA-seq analyses, gene comp61771_co of line 89-69, which showed highest expression levels at 24 and 48 hpi based on qRT-PCR, was identified to show maximum expression level at 24 hpi by RNA-seq analysis, but expression patterns of the remaining nine genes were consistent with those of line 89-69 (Figure 1, right side of each graph). These results of RNA-seq and qRT-PCR analyses confirmed the robustness and reproducibility of this study.  Table  S3; gene expression of line 89-69 (in red color) on the right side of each graph was based on analysis of RNA sequencing data in each photo.

Development of CS-Ae. searsii Disomic 2S s #1 Recombinants and Physical Mapping of Pm57
To physically map Pm57 we screened 2280 individuals of an F 2 population of TA3809/recombinant 89-88 segregating for Pm57 in a homozygous ph1b background using only two molecular markers, X23241 and X216815, located at LFLC-0.64 and LFL0.64-0.67, respectively, in the long arm of 2S s #1 (Supplementary Table S6), and Bgt-responsive assay of the individuals. A total of 46 CS-Ae. searsii 2S s #1 recombinants were identified based on the presence of either or both markers X23241 and X216815 (for Bgt-susceptible individuals), and absence of the markers (for Bgt-resistant plants). These recombinants, which included 28 resistant and 18 susceptible recombinant lines, formed a secondary recombinant population for subsequent physical mapping of Pm57, along with lines 89-88, 2011-400, and CS as checks. All recombinants were genotyped using 28 chromosome 2S s #1-specific molecular markers, of which 24 markers were located at LFL0.72-0.87 where Pm57 is mapped, and 4 at Pm57 flanking intervals of LFLC-0.64 (1), LFL0.64-0.67 (1), and LFL0.87-1.00 (2), respectively ( Table 2,  Supplementary Tables S5 and S6). Comparison of molecular marker order at 2S s #1 interval of LFL0.72-0.87 with the corresponding genomic block of chromosome 2B of T. aestivum revealed collinearity except for markers X26866 and X251565 at current loci 7 and 10, respectively. Analyses of these 28 markers and the Bgt-responses of the recombinants classified the 46 CS-Ae. searsii recombinants into 21 types with different recombination breakpoints, including 11 types as resistant and 10 types of susceptible recombinants (Figures 4 and 5). Molecular marker analysis showed that resistant recombinant 88R-3-19-1 had the shortest 2S s #1 genomic region < 9.88 Mb locating Pm57, between markers X67593 at locus 17 (773.72 Mb) and

Comparative Synteny and Genes in Pm57 Candidate Region
For the discovery of putative resistance gene (R-gene) candidates of Pm57, we conducted a synteny comparison with related species using the markers X67593 at 773.72 Mb and X62492 at 778.85 Mb flanking the 5.13 Mb Pm57 genomic region. The 5.13 Mb syntenic region corresponded to genomic regions of 5.13 Mb, 4.48 Mb, 3.64 Mb, 1.79 Mb, and 212.61 kb on chromosomes 2B, 2D, 2H, 2A, and Bd5 from T. aestivum (72 genes), Ae. tauschii (70 genes), Hordeum vulgare (123 genes), T. urartu (42 genes), and Brachypodium distachyon (38 genes), respectively. Though a total of 345 genes were identified in the syntenic blocks in different species where Pm57 mapped only 12 of these genes were annotated as putative R-genes in the PRGdb database (http://prgdb.crg.eu/wiki/Main_Page). Of these 12 R-genes, 8 genes, 4 from T. aestivum (TraesCS2B02G587400, TraesCS2B02G588500, TraesCS2B02G590100 and TraesCS2B02G593900), 2 from Ae. tauschii (AET2Gv21222400 and AET2Gv21227500), and 2 from H. vulgare (HORVU2Hr1G119780 and HORVU2Hr1G119380), are CNL class R-genes encoding proteins with at least a coiled-coil domain, a nucleotide-binding site and a leucine-rich repeat (CC-NB-LRR). The other 2 genes TuG1812G0200005979.01, and TuG1812G0200005980.01 from T. urartu, and one gene TraesCS2B02G593700 from T. aestivum are NL class R-genes encoding proteins consist of nucleotide-binding subdomain at N-terminal and leucine-rich repeat at the C-terminal, but lack of coiled-coil structures. Additionally, one gene AET2Gv21229600 from Ae. tauschii annotated as RLP class R-gene encoding protein contains leucine-rich receptor-like repeat, a transmembrane region of 25AA, and a short cytoplasmic region was also identified in the syntenic region. None of the genes from B. distachyon were annotated as putative R-genes even though up to 38 genes were located in the syntenic blocks ( Figure 6, Supplementary Table S7). Of the 8 CNL class R-genes, 3 genes, gene TraesCS2B02G588500 from T. aestivum, AET2Gv21222400 from Ae. tauschii, and HORVU2Hr1G119780 from H. vulgare are all annotated to a putative R-gene PRGDB00189661 (Hv.31127) in class CNL, decreasing the number of CNL class R-genes from 8 to 6 different genes ( Figure 6). Moreover, chromosome inversion was identified between markers X67593 and X62492 at chromosome 2H of Hordeum vulgare based on synteny comparison with group 2 chromosomes 2A, 2B, 2D of T. aestivum, 2A of T. urartu, and 2D of Ae. tauschii. were annotated as putative R-genes in the PRGdb database (http://prgdb.crg.eu/wiki/Main_Page). Of these 12 R-genes, 8 genes, 4 from T. aestivum (TraesCS2B02G587400, TraesCS2B02G588500, TraesCS2B02G590100 and TraesCS2B02G593900), 2 from Ae. tauschii (AET2Gv21222400 and AET2Gv21227500), and 2 from H. vulgare (HORVU2Hr1G119780 and HORVU2Hr1G119380), are CNL class R-genes encoding proteins with at least a coiled-coil domain, a nucleotide-binding site and a leucine-rich repeat (CC-NB-LRR). The other 2 genes TuG1812G0200005979.01, and TuG1812G0200005980.01 from T. urartu, and one gene TraesCS2B02G593700 from T. aestivum are NL class R-genes encoding proteins consist of nucleotide-binding subdomain at N-terminal and leucine-rich repeat at the C-terminal, but lack of coiled-coil structures. Additionally, one gene AET2Gv21229600 from Ae. tauschii annotated as RLP class R-gene encoding protein contains leucine-rich receptor-like repeat, a transmembrane region of 25AA, and a short cytoplasmic region was also identified in the syntenic region. None of the genes from B. distachyon were annotated as putative R-genes even though up to 38 genes were located in the syntenic blocks ( Figure 6, Supplementary Table S7). Of the 8 CNL class R-genes, 3 genes, gene TraesCS2B02G588500 from T. aestivum, AET2Gv21222400 from Ae. tauschii, and HORVU2Hr1G119780 from H. vulgare are all annotated to a putative R-gene PRGDB00189661 (Hv.31127) in class CNL, decreasing the number of CNL class R-genes from 8 to 6 different genes ( Figure 6). Moreover, chromosome inversion was identified between markers X67593 and X62492 at chromosome 2H of Hordeum vulgare based on synteny comparison with group 2 chromosomes 2A, 2B, 2D of T. aestivum, 2A of T. urartu, and 2D of Ae. tauschii.

Discussion
Powdery mildew causes significant losses to wheat production worldwide. Exploring the gene pools of wheat wild relatives for new and durable Bgt-resistance could lead to the development of wheat cultivars with robust powdery mildew resistance. However, mapping, characterization, and finally cloning R-genes derived from hexaploid wheat, which has three closely related and independent genomes (A, B and D), is challenging due a huge genome size estimated at~17 Gb [32,42,43]. This effort is further complicated if the genes are originating from wheat wild relatives as a result of low wheat-alien homoeologous recombination that is strictly controlled by Ph genes in hexaploid wheat genetic backgrounds. The other limitation is the availability of alien genome-specific molecular markers due to the lack of reference genomic sequences in the wild wheat species.
To deal with these problems of alien gene mapping, we first performed RNA sequencing using three wheat-Ae. searsii chromosome 2S s #1 introgression lines (TA3581, 89-88, and 89-69) carrying Pm57 and a susceptible control CS, which provided abundant transcriptome sequences for the development of Ae. searsii chromosome 2S s #1 specific molecular markers. Here we physically mapped 27 markers specific to chromosome 2S s #1L at the fraction length interval of LFL0.72-0.87 where Pm57 gene is mapped, of which 24 markers were consistent and used for subsequent physical map of Pm57, together with other 4 markers at flanking intervals. Secondly, we constructed a mapping population with a genetic background of homozygous ph1b genes that enhance the homoeologous recombination of chromosome 2S s #1 with its homoeologous counterpart in wheat. This resulted in the identification of 46 CS-Ae. searsii recombinants, which provide useful materials for subsequent physical mapping of Pm57.
By analysis of 46 ph1b-induced 2S s #1L recombinants with different Bgt-responses using 24 2S s #1L molecular markers at the LFL interval where Pm57 is located, and 4 markers at flanking intervals of LFLC-0.64, LFL0.64-0.67, and LFL0.87-1.00, we physically mapped Pm57 to the long arm of 2S s #1 in a 5.13-Mb genomic region flanked by markers X67593 at 773.72 Mb and X62492 at 778.85 Mb. In this 5.13-Mb genomic region, a total of 12 putative plant defense-related genes (R-genes) were discovered based on comparative synteny analysis of T. aestivum, T. urartu, Ae. tauschii, H. vulgare, and B. distachyon.
Of the 90 cataloged Pm genes, at least five, including Pm2, Pm3b, Pm8, Pm21, and Pm60, have been successfully isolated using different gene cloning approaches. Map-based cloning was used for Pm3b, Pm60, and Pm21; homologous cloning for Pm8, which is homologous locus to Pm3b; an R-gene mutant-located chromosome flow sorting and sequencing (MutChromSeq) for Pm2, and resistance gene enrichment sequencing (RenSeq) and Pacific Biosciences single-molecule real-time sequencing for Pm21 isolation [8,15,36,37,39,41]. Although isolated by different strategies, all the Pm genes currently cloned are coiled-coil, nucleotide-binding site, leucine-rich repeat (CNL) class R-genes, which initiate effector-triggered immunity usually accompanied by local cell death, also known as the hypersensitive response [52][53][54]. In this study, a total of 12 putative R-gene candidates were discovered locating in collinear genomic regions, of which 8 R-genes, including 4 genes from T. aestivum (TraesCS2B02G587400, TraesCS2B02G588500, TraesCS2B02G590100 and TraesCS2B02G593900), 2 from Ae. tauschii (AET2Gv21222400 and AET2Gv21227500), and another two from H. vulgare (HORVU2Hr1G119780 and HORVU2Hr1G119380), are CNL type R-genes, to which all cloned Pm genes belong. Furthermore, three of these CNL type R-genes (TraesCS2B02G588500, AET2Gv21222400 and HORVU2Hr1G119780) are likely homologs (PRGDB00189661, Hv.31127). Thus, these six CNL type R-genes in the syntenic region could be potential candidates for Pm57 for further cloning of Bgt-resistance gene derived from Ae. searsii. Functional validation of Bgt-resistance of these six CNL type R-genes using virus-induced gene silencing (VIGS) technology is under progress.

Plant Materials
A total of 9 lines were used in this study, including the wheat landrace CS (TA3808), a CS ph1b mutant stock (TA3809) lacking the Ph1 gene and thereby permitting homoeologous recombination, a CS-Ae. searsii disomic 2S s #1 addition line (2011-400, accession ID TA3581) where a pair of chromosome 2S s #1 were added into CS genetic background (CS-DA 2S s #1), and 6 CS-Ae. searsii disomic 2S s #1 recombinant stocks [7] (Table 1). The #1 designation is used to distinguish between the same Ae. searsii chromosome derived from different Ae. searsii accessions [55]. All materials, except the CS-Ae. searsii disomic 2S s #1 recombinant stocks are kindly provided by the Wheat Genetics Resource Center (WGRC) at Kansas State University, USA and maintained at the Experimental Station of Henan Agricultural University, China.

Construction of cDNA Libraries for Illumina Sequencing
Three wheat-Ae. searsii chromosome 2S s #1 introgression lines with Bgt-resistance, including TA3581, 89-88 and TA5109 (i.e., 89-69) carrying Pm57 and the susceptible control CS, were used for RNA-seq in this study. Pm57 was mapped at the interval of LFL0.72-0.87 on 2S s #1, which was shared by these three introgression lines based on comparison of the recombination breakpoints of the lines.
Seedling growing, fresh leaf collection for RNA sample preparation for cDNA library construction The generated paired-end reads were used for downstream sequence assembly using Trinity software [49], and transcriptome data analyses after removal of sequences containing adapters or poly-N above 5%, reads less than 100 bp in length and those of low quality.

RNA-seq Data Analysis
The assembled sequences were designed as unigenes. Read counts of unigenes were normalized as RPKM. All the unigenes were assigned to chromosomes or chromosome arms based on blastn alignment against wheat reference genomic sequences (Wheat_IWGSC_RefSeq_v1_chromosomes) with a cutoff of expect ≤ 1e-10 and qcov 75% at URGI BLAST (https://urgi.versailles.inra.fr/blast/blast.php).

Validations of RNA-seq Data by Quantitative RT-PCR
cDNA from Bgt-inoculated seedling leaves of CS, 89-88, and 89-69 at time points 0, 12, 24 and 48 hpi were used to validate RNA-seq data by quantitative RT-PCR (qRT-PCR). Total RNA was isolated using Trizol reagent (Cat. No. B511311, Sangon Biotech (Shanghai) Co., Ltd., China). A total of 1-2 µg RNA of each sample was used to synthesize first-strand cDNA using a Thermo Scientific RevertAid First-Strand cDNA Synthesis Kit (Cat. no. K1622, Thermo Fisher Scientific Inc., USA) following the manufacturer's instructions. The reverse transcription product was checked by PCR using GAPDH primer sets.  [58]. Sequences of the primer sets of the selected and actin gene were designed using the software Primer Premier 5 (PREMIER Biosoft, CA, USA) and were listed in Supplementary  Table S3.

Molecular Marker Analysis
Genomic DNA (gDNA) was isolated from 5-10 cm segments of young leaves with a DNeasy Plant Mini Kit (Qiagen, Cat No. 69104) following the instruction guide. PCR primer pairs specific for 2S s #1 were designed based on RNA-seq sequences of 2S s #1 which uniquely presented in 2011-400 but in CS, and sequences shared by three 2S s #1 introgression lines 2011-400, 89-88, and 89-69. The PCR reaction mixture preparation and PCR amplification by "Touch-down 63" followed Li et al. (2019) and Liu et al. (2017) [7,59], respectively. PCR products were resolved in 1.5% agarose gels and visualized by ethidium bromide staining by a Tanon 2500 Gel Imaging System (Tanon Science & Technology Co., Ltd., Shanghai, China).

Powdery Mildew Response Assay
Powdery mildew assays were conducted by using a mixture of Bgt isolates (composite) collected in the Henan Province and were kindly provided by Yuli Song in Henan Academy of Agricultural Sciences. The Bgt composite was inoculated after the first leaf of each seedling had fully unfolded and maintained in a controlled greenhouse with a daily cycle of 14 h light at 22 ± 2 • C and 10 h of darkness at 18 ± 2 • C. Powdery mildew infection types (IT) were scored 7-10 days post-inoculation, when the susceptible controls were heavily infected. A 0-4 IT scale was used, where 0 = no visible symptoms; 0; = hypersensitive necrotic flecks; 1 = small and sparse conidial development; 2 = colonies with moderately developed hyphae, but few conidia; 3 = colonies with well-developed hyphae and abundant conidia, but colonies not joined; and 4 = colonies with well-developed hyphae and abundant conidia with mostly overlapping colonies. Plants with ITs 0-2 were considered resistant, whereas those with ITs 3-4 were susceptible [5,60].

Identification of CS-Ae. searsii 2S s #1 Recombinants and Physical Mapping of Pm57
The CS-ph1b mutant stock TA3809 was crossed with the 89-88 recombinant, which had a pair of recombined chromosomes where the distal parts of the short and long arms of 2S s #1 were replaced by homoeologous segments derived from wheat chromosome 2A (Ti2AS-2S s #1S·2S s #1L-2AL), to generate F 1 hybrids that contained a monosomic recombined chromosome Ti2AS-2S s #1S·2S s #1L-2AL in a homozygous ph1b background. The F 2 segregating population derived from self-pollinated F 1 hybrids were used for selection of 2S s #1 recombinants using 2S s #1-specific molecular markers X23241 (488.87 Mb) and X216815 (658.61 Mb), which located at 2S s #1 intervals of LFLC-0.64 and LFL0.67-0.72 neighboring to the centromere, respectively.
After completing Bgt-responsive assays of the recombinants, Bgt-resistant F 2 individuals lacking either or both of the markers X23241 and X216815, should contain the new interstitially recombined chromosome (Ti2AS.2AL-2S s #1L-2AL), and the Bgt-susceptible, presenting either or both of the markers, were selected to compose a secondary mapping population for further mapping of Pm57 using more markers at the interval of LFL0.72-0.87, where Pm57 mapped based on the RNA-seq sequences.

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

Abbreviations
Bgt Blumeria graminis f. sp. tritici FL fraction length CNL coiled-coil nucleotide-binding site-leucine-rich repeat DEG differential expression gene SFL short-arm fraction length LFL long-arm fraction length RPKM reads per kilobase per million mapped reads IT infection types VIGS virus-induced gene silencing