Broad and novel genetic resources for M. graminicola resistance in rice
The wild relatives of rice harbor rich and novel genetic resources which can be used to improve pest and disease resistance in cultivated rice [30, 31]. For example, resistance to grassy stunt virus, bacterial leaf blight, neck blast and brown plant hopper was successfully introgressed into the cultivated rice from their wild relatives [32–34]. A number of beneficial traits including tolerance to biotic and abiotic stresses have been lost in cultivated rice which possesses a narrow genetic base because of domestication and breeding bottlenecks [23, 24]. According to an estimate, modern rice varieties have retained only 20% of the genetic diversity present in their wild relatives [35]. India has unprecedented diversity of wild rice germplasm and landraces that are spread over fifteen diverse agro-climatic zones [23, 24]. Therefore, these untapped genetic resources were taken as the prime candidates for the association panel in our study in order to unravel the conserved loci that govern M. graminicola resistance in rice. Indeed, a large proportion (14.8%; 40 of 270 accessions) of the wild accessions was found to be highly resistant to M. graminicola infection in the present study.
Herein, we screened 272 diverse rice accessions collected from ten agro-climatic zones for M. graminicola resistance by measuring the relative numbers of galls, endoparasites, egg mass, eggs per egg mass and MF ratio in PF-127 medium. A significant variation in the susceptibility level of rice accessions to RRKN infection was documented. A repeat experiment with fifty randomly selected accessions from this diversity panel showed strong correlation with the initial screening results. This suggests that PF-127-based screening is a robust and reproducible method for dissecting the rice-RRKN interaction, in agreement with the previous reports from our laboratory [36, 37]. In addition, 40 highly resistant accessions were further evaluated for M. graminicola resistance in large plug trays containing soil as the medium. Taken together the data of both soil- and PF-127-based screening, O. nivara accessions NKSWR 30 and NKSWR 259 showed almost immune response to M. graminicola with zero galls and zero MF in majority of the replicates at 16 dpi. Other O. nivara accessions such as NKSWR 43, NKSWR 123, NKSWR 19, NKSWR 108, NKSWR 25, NKSWR 18 and a O. rufipogon accession IC 336687 also supported extremely low population level of M. graminicola. The negligible susceptibility of these accessions cannot be attributed to poor root growth because root weight of these accessions was comparable with that of susceptible accessions (data not shown). The differential susceptibility of 332 O. sativa accessions to M. graminicola was earlier reported via soil-based screening in which two accessions LD 24 (indica) and Khao Pahk Maw (aus) were found to be almost immune to RRKN infection [22, 38].
Among 40 highly resistant wild accessions, 34 belonged to O. nivara type (33 of them collected from middle Gangetic plains agro-climatic zone) whereas 3 each belonged to O. rufipogon and O. sativa f. spontanea types. According to PF-127-based screening, a clear difference in resistance level in 272 accessions was also evident when genotypes were categorized into different taxonomic groups, i.e. nivara-, rufipogon- and spontanea-type. These differences can be explained by the possibility of different selection pressures among the geographic regions. When compared within different agro-climatic zones, accessions from MGP showed least nematode infection. The majority of resistant accessions originally from middle Gangetic plains (occupies the eastern part of Uttar Pradesh and Bihar) presumably have experienced higher selection pressure from RRKN infection. A similar stratification of cyst nematode (Heterodera glycines) resistance level across the different geographic location of soybean crop was reported when 235 wild accessions were included for association mapping analysis [21].
The genetic variation among Indian wild rice accessions were studied via model-based population structure analysis using genome-wide unlinked SNP markers (unlinked markers provide high reproducibility and success rate of population structure analysis; [39]) which grouped the accessions into three distinct subpopulations, namely Pro-Aus, Pro-Indica and Mid-Gangetic populations. Additionally, genetic diversity and cluster analysis based on SNP markers revealed the current wild accessions were highly diverse indicating each populations of an agro-climatic zone constituted mixture of genetically diverse individuals. According to the Fst values, Pro-Indica and Mid-Gangetic populations contained the greatest and least proportion of admixture types, respectively. Thus, accessions collected from middle Gangetic plains agro-climatic zone represented a conserved wild rice subpopulation whereas relatively greater level of gene flow and outcrossing was speculated among the geographically adjacent than the distant populations. Notably, overlapping geographic distribution pattern of wild rice subpopulations was reported by several studies [29, 40, 41]. Due to their open inflorescence the sympatric species, O. nivara and O. rufipogon can outcross with each other and also with cultivated O. sativa [42, 43]. Nevertheless, the greater genetic diversity of present wild rice accessions is in line with the hypothesis that wild rice are the source of useful new genes for future varietal improvement program [30, 44].
Phenotyping of the association panel has offered important information about the degree and distribution of RRKN susceptibility in Indian wild rice species including O. nivara, O. rufipogon and O. sativa f. spontanea, and in two cultivated species, i.e. O. sativa indica and japonica. This provided new insights into the evolution of resistance/susceptibility of wild and cultivated rice species to M. graminicola parasitism. The higher susceptibility of indica and especially japonica rice is not surprising maybe because no selection is in function to introduce resistance in japonica as RRKN is not a pest of temperate rice cultivation [22]. Intriguingly, a majority of O. nivara accessions from middle Gangetic plains (a suspected genetic diversity hotspot; [23]) has shown the least susceptibility to RRKN infection. This may be because of their eco-geographical isolation in accordance with the rationale that gene flow is inversely related to geographic distance of natural populations [45].
Genetic loci determining the parasitic success of M. graminicola in rice
By analyzing the marker-trait associations in 272 rice accessions via GWAS, we identified 17 significant SNPs that governed the RRKN resistance-related traits such as numbers of galls, egg mass, eggs per egg mass and MF ratio. By applying the –log10(P) score of 4 as a threshold for significance, we identified two QTLs at chromosome 1 and 4 of rice were associated with all the traits. A similar stringency level or a less stringent threshold (–log10(P) > 3) was adopted while reporting significant SNPs associated with nematode resistance/susceptibility in rice [22], wheat [20], soybean [21] and Arabidopsis [46]. In particular, four SNPs in chromosome 4, three SNPs each in chromosome 1, 2, 11, two SNPs in chromosome 6, and one SNP each in chromosome 3 and 10 were found to be associated with different traits. The trait such as total endoparasite counts in rice accessions was not taken into consideration as only one SNP was detected for this trait at –log10(P) > 2.75. Lowering the threshold for significance in GWAS study may increase the false discovery rate which in turn reveals more common alleles with smaller effect size in test populations [19].
A number of transcription factors (bZIP, SCARECROW, MYB, MADS-box, WRKY, ARF, GRAS etc.) that regulate plant defense responses to nematode and various biotic stress [47–54], were located on different chromosomes within 200-kb LD of the suspected QTLs in this study. The 200 kb window was adopted because of the fact that LD decay occurs in rice at 50–500 kb [25–28]. Although in wild rice the decay is much higher at 1-200 kb, O. nivara maintains LD over larger distance due to their high level of self-pollination [29]. Notably, O. nivara constituted 66.67% of total wild rice accessions in our study. Other important candidate genes located in different QTLs include resistance genes such as NBS-LRR, Cf2/Cf5, RGA3, Lr34 analog ABC transporter, zinc finger, WD repeat, leucine-rich repeat, HEAT repeat etc. All these genes putatively function to elicit plant innate immunity during pathogen invasion [55–58]. A number of candidate genes putatively involved in plant defense response to biotic stress were located within the SNPs in QTL 1.3 (ubiquitin-conjugating receptor kinase; [59, 60]), 2.1 (ADP-ribosylation factor; [61]), 2.3 (DEAD/DEAH-box; [62]), 4.2 (zinc finger; [57]) and 6.2 (GTP-binding Rac protein; [63, 64]) in our study.
We assume that chromosomes 1 (QTL 1.3 harbored a number of WRKY TF), 2, 4 (QTL 4.2 contained several zinc finger motifs), 6 and 11 (QTL 11.2 displayed several NBS-LRR resistance genes) might play a decisive role in contributing resistance to M. graminicola in rice because a number of significant SNP loci were identified in each of these chromosomes. A previous association mapping analysis have identified chromosomes 4 and 11 to harbor candidate genes such as lectin and homolog of stripe rust resistance protein [22]. Using RIL-based bi-parental mapping study, QTLs for RRKN resistance were identified on almost all the 12 chromosomes of rice to date [7, 8, 12, 13]. It is evident that RRKN resistance in rice is a complex trait that involves a number of genes which regulate a cascade of plant defense responses. Therefore, unravelling the conserved and novel RRKN-resistance loci from the non-domesticated wild rice population could improve our understanding of molecular mechanisms underlying the rice-RRKN interaction.
Candidate genes involved in M. graminicola resistance in rice
Unlike a report of homologous stripe rust and powdery mildew resistance genes in chromosome 11 of O. sativa [22], no information is available about the canonical nucleotide-binding leucine-rich repeat (NB-LRR)-type genes that impart resistance against M. graminicola in rice. Herein, we report a number of NBS-LRR proteins (LOC_Os11g10550, 10570, 10610, 10620 and 10760) which was mapped on chromosome 11 of rice within 200-kb LD of QTL 11.2. Incidentally, Hsa-1Og gene that conferred resistance against the cyst nematode, H. sacchari, was mapped on chromosome 11 of O. glaberrima [65]. The nucleotide-binding state of NBS-LRR proteins regulates the activity of plant resistance (R) proteins that are involved in activation of plant innate immunity upon pathogen recognition [66]. The R gene in tomato contains Cf2/Cf5 locus (encode LRR) that confer resistance in tomato to M. incognita [55]. Upon pathogen invasion in host plants, the activated membrane-bound NBS-LRR immune receptors translocate to the cell nucleus and interact with specific transcription factors (although not exclusive, comprise members of ERF, bHLH, bZIP, MYB, NAC and WRKY families) which modulate the plant immunity [50]. In our qRT-PCR analysis, compared to uninfected plants the expression of NBS-LRR and Cf2/Cf5 protein was positively regulated in M. graminicola-infected Pusa 1121 at 7 dpi. In coherence, other defense response regulatory genes such as MYB TF, SCARECROW, zinc fingers, bZIP TF and 14-3-3 were upregulated in nematode-infected plants than the uninfected ones.
Overexpression of a rice ADP-ribosylation factor induced pathogen resistance in tobacco by regulating the transcript accumulation of pathogenesis-related (PR) genes and salicylic acid (SA) [61]. OsRac1 (GTP binding Rac protein) was shown to be a component of disease resistance pathway acting downstream of R gene when OsRac1 transformed japonica rice was infected with the blast fungus, Magnaporthe oryzae [63]. Increased resistance to tobacco mosaic virus was documented when tobacco was transformed with OsRac1 [64, 67]. In our qRT-PCR study, both ADP-ribosylating and Rac protein was substantially upregulated in M. graminicola infected plants than the uninfected ones at 7 dpi. Together, our data suggest changes in expression level of these candidate genes can be a contributory factor to confer resistance in rice to M. graminicola. However, expression of ARF18 TF (auxin response factor) was downregulated in nematode-infected plants compared to uninfected plants at 7 dpi in our qRT-PCR study. Auxin-regulated proteins are known to coordinate the balance between plant root growth and disease resistance by promoting the auxin biosynthesis and suppressing the benzoxazinoid-based defense compound formation [68]. The direct role of auxin influx (AUX1, LAX3) and efflux (PIN3) proteins in giant cell formation were unraveled during Arabidopsis-M. incognita compatible interaction [69].