Background

Rice is a staple food of more than half of all people worldwide (Gnanamanickam 2009). However, rice production is affected by many diseases that threaten the food security of the increasing world population. Rice blast, caused by the fungal pathogen Magnaporthe oryzae, is a destructive disease of rice (Valent and Chumley 1991). It typically causes an annual yield loss of 10–30% and leads to large economic losses in many countries (Skamnioti and Gurr 2009). Rice varieties with blast resistance can help control this pathogen (Hulbert et al. 2001).

Two types of resistance genes are responsible for rice blast resistance: major resistance (R) genes that confer race-specific resistance and quantitative trait loci (QTLs) that control partial, nonrace-specific resistance (Skamnioti and Gurr 2009). More than 100 blast resistance loci or genes have been mapped to rice chromosomes (Fang et al. 2016). Among these, only 28 R genes and 2 QTLs have been cloned and characterized (** and cloning of blast resistance genes in rice (Oryza sativa L.). Sci Agric Sin 42:1601–1615" href="/article/10.1186/s12284-019-0309-7#ref-CR36" id="ref-link-section-d53690707e676">2009). All of the cloned R genes except for Pi-d2, pi21 and Ptr contain nucleotide-binding domain leucine-rich repeat (NLR) proteins (Liu et al. 2010; Zhao et al. 2018).

Among the cloned R genes, the Pik locus is especially important because it harbors a number of blast R genes used in rice breeding (Zhai et al. 2011). The Pik locus has at least six alleles (Pik, Pikm, Pikp, Piks, Pikh, and Pi1) that cluster on the end of the long arm of chromosome 11 (Zhai et al. 2011; Liu et al. 2014). Three of these alleles (Pikm, Pikp, and Pik) have been isolated and characterized (Zhai et al. 2011). Because R genes are highly specific to M. oryzae races, resistance of a single R gene is often rapidly overcome by the selection of compatible pathogen races (Hittalmani et al. 2000; Oliveira-Garcia and Valent 2015). In response to the rapid evolution of M. oryzae, the rice genome has evolved R gene polymorphism, which confers multiple forms of race-specific resistance (Hayashi et al. 2004). For example, the physical interaction of Pik alleles with specific Avr-Pik alleles can be explained by the coevolution of pathogen and host (Yoshida et al. 2009; Kanzaki et al. 2012; Wu et al. 2014). The Pik gene comprises two NBS-LRR genes, Pik-1 and Pik-2, and the former acts as the senor for the interaction with the corresponding AvrPik protein and the latter is responsible for defense activation and signaling. A recent study showed that polymorphic residues in Pik-1 determine the resistance specificity (Carlos et al. 2018)

Genome-wide association study (GWAS) has recently been used for assessing associations between genetic markers and blast resistance in rice. GWAS was first used to identify genes underlying complex diseases in humans (Altshuler et al. 2008). With its wider use and the development of GWAS modeling (Price et al. 2006; Yu et al. 2006; Liu et al. 2016), GWAS has become a powerful approach for map** a number of traits of rice, including agronomic traits (Huang et al. 2010; Zhao et al. 2011) and tolerance to abiotic stress (Zhu et al. 2015; Lv et al. 2016a, 2016b; Kaler et al. 2017). In rice, GWAS in combination with high-throughput sequencing and gene knockout techniques has been used to rapidly identify new functional genes that influence yield, heading, awn length, and other agronomic traits (Si et al. 2016; Yano et al. 2016).

In their study of resistance to 16 representative blast strains collected from all over China, researchers recently identified 30 loci associated with blast resistance using an indica rice population (Wang et al. 2014). The rice diversity panel 1 (RDP1), which comprises over 400 rice cultivars from 82 countries, was developed for GWAS of agronomical traits, and is publicly available (Zhao et al. 2011). Using RDP1, our team previously performed an association study of rice resistance to blast isolates from Asia (China, India, the Philippines, and South Korea) and the Americas (Colombia), which led to the identification of 97 loci associated with blast resistance along with two new Pi5 alleles (Kang et al. 2016). Similarly, our team analyzed QTLs associated with blast resistance in the field and identified 16 loci associated with field blast resistance using RDP1 (Zhu et al. 2016).

In the current study, we performed GWAS on rice blast using the 234 RDP1 cultivars and three isolates of the pathogen from the major rice production regions in Hunan Province, China. A total of 56 QTLs associated with blast resistance were identified in the rice genome. Only one QTL associated with resistance to all three isolates was found, and it was localized with the known R gene Pik locus. Genotype analysis of the significantly associated SNP-11.27701887 on the Pik locus in 234 cultivars confirmed the association between genotype and resistance to the three isolates. In addition, sequence analysis and protein-protein interaction analysis indicated that this gene is a previously unreported allele on the Pik locus. The new allele was named Pikx.

Methods

Plant and fungal materials

A total of 234 rice accessions from RDP1 were used, and these comprised 59 tropic japonica (TRJ), 35 temperate japonica (TEJ), 53 indica (IND), 45 aus (AUS), 10 aromatic (ARO), and 32 admixture (ADMIX) accessions (Zhao et al. 2011). Fifteen seeds of each accession were germinated and sowed in small pots with soil. The rice seedlings were grown in a growth chamber under controlled conditions (26 °C, 75% relative humidity, and a 10 h light/14 h dark photoperiod). Three M. oryzae isolates, 110–2, 193–1-1, and 87–4, were collected in Dong’an County, Taojiang County, and Hanshou County, respectively, in Hunan Province (Fig. 1e). The M. oryzae isolates were cultured on an oat medium to obtain conidia for inoculation of 3-week-old rice seedlings.

Fig. 1
figure 1

The information of three M. oryzae isolates and blast disease scores of the 234 RDP1 cultivars. a-c The distribution of disease scores of the 234 RDP1 cultivars against isolates 110-2, 193-1-1 and 87-4. d PCA of resistance phenotypes of the 234 RDP1 cultivars to the three isolates. e Geographic distribution of three M. oryzae isolates in Hunan Province Map

Full size image

Blast resistance phenoty** of rice cultivars

M. oryzae conidia at a spore concentration of 1 × 105 were sprayed onto rice leaves as described by Park et al. (2012). Each cultivar was inoculated with the three isolates in three replications. Approximately 7 days after inoculation with M. oryzae, disease was scored on a scale from 0 to 9 based on the size and area ratio of lesions as described previously (Kang et al. 2016a, 2016b). The signal peptide-truncated cDNA fragments of AvrPik-A, −C, −D, and -E were synthesized by Genereate Ltd. and were then inserted into the pGBKT7 (for Y2H) vector and into the pYBA1152 (for Co-IP) vector. Fragments of Pikx-1-cc that were amplified from the cDNA of rice seedlings of cultivar NSFTV_131 using PrimeSTAR GXL DNA polymerase (Takara) were cloned into pGADT7 (for Y2H) and PYBA1144 (for Co-IP) using specific primers (Additional file 1: Table S1).

Results

Resistance of the 234 RDP1 cultivars to the three M. oryzae isolates

The distribution of blast disease scores of the 234 cultivars inoculated with the three M. oryzae isolates is shown in Fig. 1a, b, and c. Among the inoculated rice cultivars, 11 were highly resistant (disease scores ≤1), and 14 were highly susceptible (disease scores ≥8) to all the three isolates (highlighted with yellow and blue, respectively, in Additional file 2: Table S2). Principal component analysis (PCA) showed that the levels of resistance of the cultivars to the three isolates were diverse (Fig. 1d). However, the disease scores of the cultivars to isolates 193–1-1 and 87–4 were clustered perhaps because the two isolates were collected in fields that were near each other and that might therefore contain similar avirulence genes.

Association map** of rice QTLs linked to resistance to M. oryzae

To identify genomic regions that are associated with blast resistance to the three isolates, we performed a GWAS using the disease scores and the 700 K SNP genotypes of the inoculated cultivars. A total of 56 QTLs associated with blast resistance to the three isolates were detected in the rice genome (−Log10 P ≥ 4.0) (Fig. 2). Among the QTLs, 24 loci were associated with resistance to isolate 110–2, 32 were associated with resistance to isolate 87–4, and 22 were associated with resistance to isolate 193–1-1 (Additional file 3: Table S3). Some of the loci were associated with resistance to two isolates, i.e., three loci were associated with resistance to both 110–2 and 87–4, two loci to both 110–2 and 193–1-1, and one locus to 87–4 and 193–1-1. Among all of the mapped QTLs, 16 were located at sites of five previously cloned resistance genes, and 40 were loci identified for the first time in this study. Interestingly, seven QTLs (46–52) were located in the Pik region on chromosome 11 (Fjellstrom et al. 2004) (Additional file 3: Table S3). Among them, QTL48, QTL50, and QTL51 were associated with resistance to all three isolates. In addition, seven of the QTLs (QTL21, QTL48, QTL49, QTL50, QTL51, QTL52 and QTL53) were also identified in a previous study (Kang et al. 2006; Hayashi et al. 2004; Hittalmani et al. 2000; Kang et al. 2016). For example, five loci were detected in the same rice population when inoculated with different blast isolates (Additional file 4: Table S4) (Kang et al. 2016). One of the five common loci is Pik on chromosome 11, which has been frequently studied in the last several years. Some of the alleles at the Pik locus such as Pik, Pik-1, Pik-p, and Pik-m have been isolated and characterized (Ashikawa et al. 2008; Yuan et al. 2011; Zhai et al. 2011; Hua et al. 2012). In this study, we identified a new allele, Pikx, that is significantly associated with rice blast resistance at the Pik locus. Because it confers resistance to the three isolates collected in Hunan, we speculate that this new gene may be useful for rice breeding against the M. oryzae population in Hunan and other provinces in South China. The function of this gene in blast resistance will be confirmed by rice transformation in the future.

Our Y2H and CoIP assays also indicated that Pikx-1 cc, which is similar to Pik-m-1-cc reported in Kanzaki et al. (2012), can interact with AvrPik-A, −D, and -E. These interactions may be possible because the key sites of the Avr-R-recognition domain HMA in the Pikx-1 cc are identical to those of the Pik-m-1 cc domain (Maqbool et al. 2015). Avr-Pik is a 113 amino acid protein with a 21 amino acid signal peptide at its N terminus (Yoshida et al. 2009). The latter study identified five alleles of Avr-Pik (Avr-Pik-A, B, C, D, and E) in 21 isolates of M. oryzae from Japan and found that the five Avr-Pik alleles differ from one other by a total of five DNA substitutions, all of which cause amino acid changes. According to Yoshida et al. (2009), the Avr-Pik-D allele is likely the ancestral allele from which the Avr-Pik-E, −C, −A, and -B alleles are derived. As noted earlier, Yoshida et al. (2009) found Avr-Pik-B in an isolate (isolate 9505–3) from Japan. In the present study, the Y2H assay showed that AvrPik-A, −D, and -E strongly interact with Pikx-1-cc (Fig. 5a). In the Co-IP assay, however, Pikx-1-cc strongly interacted only with AvrPik-E and weakly interacted with AvrPik-A and -D (Fig. 5b). The reason for the difference warrants further investigation.