Fine mapping of Pi57(t) conferring broad spectrum resistance against Magnaporthe oryzae in introgression line IL-E1454 derived from Oryza longistaminata

Wild species of the genus Oryza are excellent gene pools for improvement of agronomic traits of Asian cultivated rice. The blast resistance gene Pi57(t) in the introgression line IL-E1454 derived from Oryza longistaminata was previously mapped on rice chromosome 12. Inoculation with 322 Magnaporthe oryzae isolates collected from 6 countries indicated that Pi57(t) conferred broad spectrum resistance against M. oryzae. Two mapping populations consisting of 29070 and 10375 F2 plants derived from the crosses of resistant donor IL-E1454 with susceptible parents RD23 and Lijiangxintuanheigu respectively, were used for fine mapping of Pi57(t) locus. Based on genotyping and phenotyping results of recombinants screened from the two crosses, Pi57(t) was finally mapped to a 51.7-kb region flanked by two molecular markers (STS57-320 and STS57-372) on the short arm and close to the centromere of chromosome 12. Six candidate resistance genes were predicted in the target region according to the reference sequence of Nipponbare. These results could facilitate both marker-assisted selection for disease-resistant breeding and gene cloning.


Introduction
Rice blast, caused by the filamentous ascomycete Magnaporthe oryzae [1], is one of the most destructive diseases for rice (Oryza sativa L.), and is responsible for significant yield losses under favorable environmental conditions worldwide [2]. Rice-M. oryzae pathosystem follows the gene-for-gene relationship during the host-pathogen interaction [3][4]. The use of resistance (R) genes in rice breeding has been proved to be the most economic, effective and environment-friendly strategy for blast management. But, after the release of blast resistant varieties, the emergence of virulent races of the pathogen often cause the rapid loss of effectiveness of resistance conferred by monogenic resistance [5]. Few exceptions of monogenic durable resistance exist [6]. Monogenic resistances may also contribute to durable resistance if appropriate management strategies are used: agronomic conditions, rotation and/or mixtures of varieties, etc. [7][8][9][10]. In addition, the pyramiding of multiple R genes with different resistance spectra to races of M. oryzae into a single variety through marker-assisted selection strategy is one of the most effective methods to breed durable varieties for durable control [11]. Pyramiding requires the characterization and identification of markers closely linked to the R genes of interest.
In the past decades, genetic studies on blast resistance in rice have been extensively conducted, and over 100 major blast R genes from O. sativa and its wild relatives have been identified and mapped on the 12 chromosomes of rice [12][13][14][15][16]. Clusters of functional genes were identified on chromosomes 6, 11 and 12. Most of the R genes are dominant, except 3 recessive genes, pi21 [17], pi55(t) [18] and pi66(t) [16]. The availability of rice genome sequences of two subspecies of cultivated rice, O. sativa ssp. japonica cultivar Nipponbare [19] and indica cultivar 9311 [20], greatly facilitate the development of molecular markers for fine mapping of R genes, and comparison of R gene positions between mapping populations.
Blast R gene Pi57(t) is carried by a introgression line IL-E1454, and was introgressed from O. longistaminata into indica cultivar RD23. Previously, this gene was preliminary mapped on chromosome 12 of rice using a BC 4 F 2 population derived from the cross between IL-E1454 and the recurrent parent RD23. Pi57(t) was mapped to a 6.07 Mb region between molecular marker RM27892 and RM28093 [21]. Although Pi57(t) can be differentiated from known R genes Pita, Pita2, Pi12, Pi19 and Pi20 also located on chromosome 12, through pathogen-testing with different M. oryzae isolates [21], its exact genomic position on chromosome 12 remains unclear. In this study, two mapping populations from ILE1454/RD23 and IL-E1454/ Lijiangxintuanheigu (LTH) were used for further mapping of Pi57(t).

Mapping population construction, planting and resistance evaluation
Resistant donor parent IL-E1454 was crossed with susceptible cultivars RD23 (indica) and LTH (japonica), respectively. The IL-E1454/RD23 and IL-E1454/LTH F 1 plants were grown in the greenhouse to generate F 2 populations for gene mapping. The germinated F 2 seeds of the IL-E1454/RD23 and IL-E1454/LTH cross combinations were sown in trays of 20×12×5 cm filled with compost, and each tray sowed with 95 seeds. Seedlings were inoculated with M. oryzae strain HN09-1C-7 by spraying at 4 leaf stages with 20 ml conidial suspension per tray. The inoculated rice plants were stored for one night in a controlled dark chamber at 25˚C with 95% relative humidity, and then transferred back to the greenhouse. Lesion types on rice leaves were observed 6-7 days after inoculation and scored according to a standard reference scale [3]. Plants scored from 1 to 3 were considered to be resistant and scored from 4 to 6 were considered to be susceptible. Four hundred and seventy-five seedlings and 570 of the IL-E1454/ RD23 and IL-E1454/LTH populations respectively were inoculated and evaluated for the expected 3:1 resistant: susceptible segregation ratio in F 2 populations [21]. To determine the resistance spectrum of Pi57(t) locus, IL-E1454 and 10 monogenic lines were inoculated with 322 isolates from 6 countries (S1 Table).

M. oryzae isolate cultivation
M. oryzae isolate HN09-1C-7, virulent to RD23 and LTH but avirulent to IL-E1454, and previously used to map Pi57(t) [21] was cultured on oatmeal medium (20 g of oatmeal, 15 g of agar, 10 g of sucrose and 1 L of distilled water) for 7 days in dark incubator at 25˚C, and then aerial mycelia were washed off by gentle rubbing with distilled water and paintbrush. The colonies were then successively exposed to fluorescent light for 3 days to induce sporulation at 25˚C. Conidia were harvested by softly scraping and flooding the medium surface with distilled water containing 0.01% Tween 20 detergent. The concentration of conidial suspension was adjusted to 50000 conidia/ml for inoculation.

Marker development and genetic map construction
Total DNA was extracted from fresh leaves of each plant following the method of Edwards et al. [22]. The SSR markers located in the genomic region carrying Pi57(t) and producing a polymorphic band between parents were used to genotype the mapping population. Sequencetagged site (STS) markers were developed based on the alignment (using BLAST) within the critical region of the genomic sequences of 93-11 and Nipponbare.
PCR amplification conditions consisted of a denaturing step of 94˚C/3 min, followed by 35 cycles of 94˚C/30 s, annealing temperature (see Table 1)/30 s, and 72˚C/1 min, ending with an extension step of 72˚C/7 min. Amplicons were separated by 8% polyacrylamide gel electrophoresis and visualised by silver staining. Primer sequences and other relevant properties of the marker assays are summarized in Table 1. The polymorphism determined by all STS markers developed in this study among resistant donor IL-E1454, and susceptible parents RD23 and LTH were showed in S1 Fig. The genetic and linkage map of polymorphic markers was constructed using MAPMAKER/EXP 3.0 [23]. The Kosambi mapping function was used to transform recombination frequency to genetic distance (cM).

Physical map construction in silico and candidate gene prediction
To construct a physical map of Pi57(t) locus, all molecular markers used for gene mapping were landed on the IRGSP1.0 pseudomolecule of reference cv. Nipponbare released by IRGSP through BLASTN search (https://www.ncbi.nlm.nih.gov/Blast.cgi). Subsequently, the physical map spanning Pi57(t) locus was constructed based on the reference genomic sequence of Nipponbare. The candidate R genes in the target region were predicted based on the annotation information by Rice Genome Annotation Project (http://rice.plantbiology.msu.edu/), GEN-SCAN (http://genes.mit.edu) and FGENSH (http://www.softberry.com/) software.

Amplification of candidate R genes from IL-E1454 by PCR
To amplify the candidate R genes from IL-E1454, PCR primers were designed according to the reference genomic sequence of Nipponbare. The long-range enzyme (PrimeSTAR GXL DNA polymerase, TAKARA BIO INC.) was used to amplify the target DNA fragments. The PCR amplification conditions and primers information were summarized in Table 1. After amplification, the PCR products were then sequenced and analyzed.

Fine mapping of Pi57(t) locus
In a previous study, Pi57(t) gene were preliminary mapped in a region spanning the centromere of chromosome 12, and delimited between two SSR markers (RM27892 and RM28093).

Construction of physical map of Pi57(t) locus, candidate gene prediction and amplification
All the molecular markers closely linked with Pi57(t) gene were landed to the genome sequence of chromosome 12 of reference cultivar Nipponbare by BLAST analysis (Fig 1c), and the phenotypes of recombinants between STS57-320 and STS57-372 were showed in Fig 1b. Subsequently, Pi57(t) locus defined by two flanking markers (STS57-320 and STS57-372), and co-segregated with two markers (STS57-336 and STS57-4). The resulting physical map is shown in Fig 1c with a physical distance of ca. 51.7 kb in the target region.
Based on the annotated Nipponbare genome sequence, 6 genes (LOC_os12g18690, LOC_os12g18700, LOC_os12g18710, LOC_os12g18729, LOC_os12g18750, and LOC_os12g18760) were predicted in the target region flanked by STS57-320 and STS57-372 (Chr12:10799294 to10850958). Among these candidate genes, all genes encode uncharacterized protein with the exception of LOC_os12g18760 that encode a putative peptidase family C78 domain containing protein. When amplifying these candidate genes in IL-E1454 using primers designed based on the genome sequence of Nipponbare, the candidate genes LOC_os12g18690, LOC_os12g18710, LOC_os12g18729, and LOC_os12g18760 were successfully obtained, and showed high homology to the corresponding gene loci in Nipponbare at a level of 98.1%, 84.1%, 94.5% and 98.4%, respectively (Fig 1d). The candidate gene LOC_os12g18700 and LOC_os12g18750 could not be amplified. When new PCR primers were designed to amplify the remaining target region with IL-E1454 DNA as the template, two gaps located between LOC_os12g18690/LOC_os12g1871, and LOC_os12g18729/LOC_os12g18760could not be successfully filled out (data not shown), these suggested that LOC_os12g18700 and LOC_os12g18750 genes would be absent or inserted with large DNA fragments in the gaps region, respectively.

Resistance spectrum determination of Pi57(t) gene
To determine the resistance spectrum, identity and potential usefulness of Pi57(t) in rice breeding for disease resistance, IL-E1454 and 10 monogenic lines were tested with 322 M. oryzae isolates from Cambodia, Laos, Myanmar, Thailand, Vietnam and China ( Table 2). The inoculation results showed that IL-E1454 was resistant to 300 isolates (93.17%) of the total tested isolates. Compared with monogenic lines carrying R genes located on chromosome 12, IL-E1454 was resistant to all isolates from Laos and Myanmar, indicated that Pi57(t) could be differentiated from Pi12, Pi19, Pi20, Pita and Pita-2, due to their susceptibility to part of the isolates from Laos and/or Myanmar. Meanwhile, Pi57(t) showed a high resistant frequency to the isolates used in this study with the comparison of the known broad-spectrum R genes (Pi5, Piz, Piz-5, Piz-t and Pi9), suggesting that Pi57(t) gene conferred a broad spectrum resistance against M. oryzae.  Fine mapping of the blast resistance gene Pi57(t) in Oryza longistaminata

Discussion
In a previous study, Pi57(t) was identified and preliminary mapped in 6.07 Mb region on chromosome 12 of rice [21]. In this study, this O. longistaminata-derived gene was finely mapped to a region of 51.7 kb on the short arm proximal to centromeric position of chromosome 12 of rice, by using two mapping population from IL-E1454/RD23 and IL-E1454/ LTH. It has been well documented that the recombination frequency along a chromosome is quite different in plant, and that the chromosomal recombination was significantly suppressed in the region with more repetitive DNA sequences and/or close to the centromeric regions than other regions [26][27][28]. Pi57(t) was located in the region close to centromere of chromosome 12 and we observed low recombination frequency: after mapping with 13395 F 2 individuals, Pi57(t) locus was still mapped in a large chromosomal region flanked by molecular markers STS57-1 and STS57-2. Although the population size used in this study are relatively larger than those used in other genes mapping [13,25], increasing mapping population consisting of 39445 F 2 individuals could finally delimit this locus to an estimated 51.7 kb, based on the physical distance determined by in silico mapping on O. sativa reference genome. Classical genetics and molecular data have demonstrated that many resistance genes in plant are often clustered in a certain chromosomal region as a complex locus [29]. To date, 19 R genes Pita, Pita2, Pitq6, PiGD-3, Pi6(t), Pi12(t), Pi19(t), Pi20(t), Pi21(t), Pi24(t), Pi31(t), Pi32 (t), Pi39(t), Pi41, Pi42(t), Pi57(t), Pi58(t), Pi61(t) and Pi157(t) have been mapped on chromosome 12, and most of them are concentrated around the centromere region [12,13,21,30,[31][32][33][34]. Most of them were mapped to a relative large chromosomal region spanning over several Mb on the short arm of chromosome 12 [12,31,[33][34]. By using a large number of M. oryzae strains, Pi57(t) could be differentiated from genes located in the same genomic regions and introgressed in monogenic lines (Pita, Pita-2, Pi12, Pi19 and/or Pi20). However, the positional or allelism relationship among these genes could not be compared with each other in detail, due to the limited information about their rough mapping position and different M. oryzae strains used in gene mapping research [13,21,31,[33][34]. Fine mapping of R genes provide direct information about the relationship among the genes in a cluster. For example, by comparing with the chromosomal position of cloned Pita gene, Pi61(t) gene was mapped at ca. 200 kb region on the telomere side, and Pi39(t) was localized at 37 kb region on the centromeric side in the short arm of chromosome 12 [13,[24][25]. Pi57(t) was mapped to the proximal side to the centromere compared with Pi39(t) location, indicating that this is a new locus conferring resistance to rice blast. With the characterization of broad spectrum resistance against M. oryzae, this gene would be a very useful gene resource for improvement of resistance to rice blast in rice breeding program.
Most of the R genes cloned from plants so far encode protein with nucleotide-binding site and leucine rice repeat (NBS-LRR), LRR-kinase or kinase structure [35]. To date, all cloned rice blast R genes encode NBS-LRR proteins, except for Pid-2 and the recessive pi21, which encode a receptor-like kinase protein and a proline-rich protein, respectively [36,37]. In this study, Pi57(t) gene was mapped in a region containing 6 predicted genes without any similarity to known R genes, based on the gene annotation results of reference genomic sequence of O. sativa ssp. japonica cultivar Nipponbare. In the present study, because two gaps exit in target region in IL-E1454 with the comparison of Nipponbare genomic sequence, whether Pi57(t) encodes protein with either novel structure similar to the annotated candidate gene, or a known R gene structure but located in gap region in IL-E1454 remains to be clarified. Currently, the gap-filling with genome walking strategy and genetic transformation for candidate genes are undergoing.

S1 Fig. Polymorphic analysis of resistance donor IL-E1454, and susceptible parents RD23 and LTH with STS markers developed in this study.
(PDF) S1 Table. Resistance reaction of IL-E1454 and 10 monogenic lines to 322 Magnaporthe oryzae strains. (DOC)