A total of 968 wheat genotypes (CIMMYT and national wheat improvement program; Supplementary Table 1) were compiled (referred to here as CRP genotypes as they are the part of CGIAR Research Program on Wheat funded project) and screened in 2 years in sets of 484 genotypes each in the years 2012-13 and 2013-2014, at five independent geographical locations in South-Asia as detailed below (Figures 1A,B). All the plant material collected and used in this study is summarized in Table 1A. The first set of germplasm of 484 genotypes (screened in 2012–13; Supplementary Table 1A) consisted of the WAM, 3rd CSISA SB, CIMCOG and 44th IBWSN populations (detailed description in Table 1A). The second set of 484 genotypes (Supplementary Table 1B; screened in 2013-14), were from 4th CSISA-SB (Singh et al., 2015), 4th CSISA HT-EM, 45th IBWSN, 33rd ESWYT, 2nd SATYN, 20th SAWYT, and 1st WYCYT populations, as detailed in Table 1A. All the genotypes were screened at 3 sites in India: Banaras Hindu University (BHU; 25.2 °N, 83.0 °E), Rajendra Agriculture University (RAU; 25.2 °N, 83.0 °E), and Uttar Banga Krishi Viswavidyalaya (UBKV; 26.2 °N, 89.2 °E), as well as at the Wheat Research Centre (WRC) of Bangladesh Agriculture Research Institute (BARI, Dinajpur, Bangladesh; 27.5 °N, 83.44 °E) and National Wheat Research Program (NWPR) of Nepal Agriculture Research Council (NARC, Bhairahawa, Nepal; 27.5 °N, 83.4 °E) in alpha-lattice design, where each genotype was planted in two rows of 2 m each. Alpha-lattice design was generated using PROC PLAN method in SAS software v9.2. α(0, 1) lattice method was used as blocking criterion. In total, 22 blocks, having 22 genotypes in each block, thus comprising 484 genotypes in each replication, were planted. Row-to-row and plant-to-plant distance was 25 and 5 cm, respectively. The whole setup was independently duplicated at all the five sites. Fertilizers (nitrogen (N), phosphorus (P) and potassium (K) in a ratio of 120:60:40 were applied according to standard agronomic practice.

We used a stringent, two-pronged criterion for selection of resistant genotypes by analyzing the area under disease progress curve (AUDPC, detailed below; Supplementary Table 2). First, a cutoff [equal to lowest AUDPC + LSD (least significant difference)] was used to review the genotype performance in terms of resistance to spot blotch pathogen at a given location. All the genotypes having AUDPC of less than the cutoff were selected as a primarily resistant group of genotypes at a given location. Next, such groups of genotypes across all the five locations were compared. Only the genotypes that were “resistant” in at least 3 of 5 locations were selected for further investigations. Same exercise was conducted in both the years, thus a total of 31 genotypes were selected in year 1, and 15 were selected in year 2 (Supplementary Table 2). Additionally, a principal component analysis (PCA) of performance of 484 genotypes in each year was also conducted (Supplementary Figure 1).

“Resistant” nature of these selected genotypes (a total of 46 genotypes) was reconfirmed by independent screenings for three consecutive years (2014–2017) in field conditions at all the five locations (Figure 1). Yangmai #6, which is proven as a resistant genotype (Sahu et al., 2016b), was also included in the confirmatory trial, making the total number of resistant genotypes as 47. Also, eight susceptible genotypes were included as controls or reference set for comparison of traits in the confirmatory trials. Thus, for confirmatory analysis, a total of 55 lines were evaluated in randomized block design with three replicates where 2 m single row was assigned for each line. Row-to-row and plant-to-plant distance was 25 and 5 cm, respectively. These 55 genotypes were also used in biochemical characterizations.

A pure culture of B. sorokiniana (strain HD-3069) was maintained on PDA (potato dextrose agar) media (Chand et al., 2003; Pandey et al., 2008). Sorghum grains were used for large-scale multiplication of spores of B. sorokiniana (Chand et al., 2013). For creating artificial epiphytotic in wheat lines, spores suspension was inoculated by spraying at a concentration of 10⁴ spore ml⁻¹ of water at the growth stage (GS) 55 (Zadoks et al., 1974) during evening time. After inoculation, a light irrigation was given to maintain a conducive environment for disease development.

Disease was scored at three different growth stages of wheat (Zadoks et al., 1974), GS 63 (beginning of anthesis to half complete), GS 69 (anthesis complete) and GS 77 (late milking), using double digit (DD, 00–99) scale (Saari and Prescott, 1975). The first digit (D1) indicates vertical disease progress on the plant, and the second digit (D2) indicates severity, measured in diseased leaf area. The DS percentage for each score was based on the following formula:

Disease progression or AUDPC was calculated using the percent severity estimations corresponding to the disease rating (Shaner and Finney, 1977; Madden et al., 2007) as:

Where Y_i = disease level at time t_i

(t_i+1 − t_i) = days between two disease scores

n = number of readings.

In addition to disease severity, days to heading (DH), thousand kernel weight (TKW) and plot yield (amount of harvested grain per crop area) were also recorded to evaluate plant performance.

Time course experiments were conducted to explore the elicitation of H₂O₂, SOD, MDA, SA, and phenolic acids (which may also act as phytoalexins) during the spot blotch progression. Flag-leaves were collected after inoculation of B. sorokiniana. Three-five biological replicates at 0, 12, 24, 48, and 72 h post inoculation (hpi) were harvested in liquid nitrogen and stored in −80°C until further use. Samples at 0 hpi (just at the time of inoculation) served as the control/references.

H₂O₂ was estimated by modified protocol of Patterson et al. (1984). Leaf samples (200 mg) were homogenized in 2 ml of sodium-phosphate buffer (pH 6.5). The extract was centrifuged at 10,000 rpm for 15 min. Reaction was initiated by adding 1 ml of TiSO₄ (HiMedia, GRM2484) (1% in 20% H₂SO₄) to supernatant. Reaction mixture was centrifuged at 6,000 rpm for 15 min. Absorbance was recorded at 410 nm on double beam UV-VIS spectrophotometer (UV-1800, Shimadzu). The quantity of H₂O₂ was expressed as unit per gram fresh weight using extinction coefficient 53.5714.

SA was evaluated as previously described (Sahu et al., 2016a,b). Briefly, 150–200 mg of flag leaves were pulverized with liquid nitrogen and extracted with 1 ml of 90% methanol followed by centrifuged at 13,200 rpm for 15 min at 4°C. Supernatant was recovered and pellet was re-extracted with 1 ml of absolute methanol and supernatants were pooled and dried. 5% trichloroacetic acid was added to the dried samples, followed by addition of 800 μL extraction buffer (ethyl acetate:cyclohexane, 1:1). Upper organic layer was separated and dried for SA estimation by adding UPLC eluent and transferred to a clean vial for analysis.

For total phenolic content and UPLC-based analysis, phenolic acids were extracted as described earlier (Sahu et al., 2016b) in both free and bound form. Flag leaf samples (100–150 mg) were pulverized with liquid N₂ and extracted with 1 ml of 70% methanol. After centrifugation, the process was repeated and supernatants were combined and used for free phenolic acid contents where pellets were analyzed for bound phenolic acid contents after hydrolysis. Total phenolic contents were determined with the Folin-Ciocalteu's assay (Kofalvi and Nassuth, 1995; Pandey and Baldwin, 2008). Total phenolics content was expressed as μg gallic acid equivalents (GAE). Estimation of individual phenolic acids (4-hydroxybenzoic acid, syringic acid, vanillic acid, p-coumaric acid, chlorogenic acid, caffeic acid, sinapic acid and ferulic acid) was investigated using UPLC-DAD method using C18 column (Sahu et al., 2016b). Water and acetonitrile (80:20 v/v) with 0.1% formic acid (pH 3.0) was used as mobile phase at the flow rate of 0.3 ml min⁻¹ where column was maintained at 40°C. DAD detector was set with four wavelengths (275, 285, 310, and 320 nm) for UPLC analysis.

SOD was estimated as described earlier (Dhindsa et al., 1981; Xu et al., 2013). 100 mg leaf samples were crushed in 5 ml of extraction buffer (0.1 M phosphate buffer), centrifuged at 9,000 rpm for 15 min and supernatant (source of enzyme) was collected. Three ml of the reaction mixture containing 0.1 ml of 1.5 M sodium carbonate, 0.2 ml of 200 mM methionine, 0.1 ml of 2.25 mM NBT, 0.1 ml of 3 mM EDTA, 1.5 ml of 100 mM potassium phosphate buffer, 1 ml of distilled water and 0.1 ml of enzyme extract. Two test tubes without enzyme extracts were taken as control. The reaction was started by adding 0.1 ml of riboflavin (60 μM) and placing the test tubes below a light source of two tungsten incandescent light bulb having illuminance of 8072.93 lux for 15 min. A non-irradiated complete mixture that did not develop color served as a blank. Absorbance was recorded at 560 nm in spectrophotometer. The enzyme unit (EU) was expressed on per gram fresh weight basis.

Lipid peroxidation was estimated as the MDA content, determined by the method of Heath and Packer (1968). Plant samples (100 mg) were homogenized in 5 ml of 0.1% (w/v) of trichloro-acetic acid solution. The homogenate was centrifuged at 9,000 rpm for 20 min and 0.5 ml supernatant was added to 1 ml of 0.5% (w/v) thiobarbituric acid in 20% TCA. The mixture was heated 95°C for 20 min in a water bath and immediately cooled on ice. The samples were again centrifuged at 9,000 rpm for 5 min. The absorbance was recorded at 532 and 600 nm. The amount of MDA-TBA complex (red pigment) was calculated from the extinction coefficient as 155 mM⁻¹ cm⁻¹. Results are presented as μmoles MDA g⁻¹ fresh weight.

PDA plates containing phenolic acids in a dilution series in the concentrations of 0–1,000 μg/ml of syringic acid were prepared and B. sorokiniana was allowed to grow on these plates. Diameter (cm) of growth was recorded upto 10 day to determine inhibition of growth. Cultures growing on PDA only (0 μg/ml of phenolic acids) were used as controls for comparison.

Role of syringic acid in plant defense was validated by exogenously complementing the spot blotch susceptible Sonalika genotype with syringic acid and determining the gain in resistance (or reduction in susceptibility). Sonalika plants were treated with a foliar spray of syringic acid (TCI Chemicals, India; product number, G0014) at concentrations of 0.1, 0.2, 0.5 mM until run-off. Syringic acid was prepared in 100% methanol and diluted in double distilled water. For comparisons, Sonalika plants (control, 0 mM) were sprayed with only double distilled water containing equal amount of methanol as required to dissolve syringic acid. Twenty four hours after spraying syringic acid, the plants were inoculated with B. sorokiniana spore suspension (foliar spray as described above, at concentrations of 10⁶ spores/ml until run-off). The disease severity (DS) (%) was scored on 5, 7, 9 days post inoculation (dpi). Six-seven replicates were used for each condition. The plants were scored on the basis of the Saari and Prescott double digit scale (DD, 00–99), disease severity (DS; %) was scored, and AUDPC was calculated as described above. Plants were grown under controlled conditions at a photoperiod of 14(25°C)/8(22°C) at a relative humidity of 73-75%. Fertilizers (NPK, at a proportion of 19:19:19) were supplied at 15 and 45 days of germination.

Analysis of lattice design was done using PROC MIXED model in SAS v9.2. Analysis of variance (ANOVA) was performed in general linear model (GLM). Dunnett's test; p ≤ 0.05 and Tukey-Kramer test; p ≤ 0.05 was used to detect significant differences between each genotype under different treatments. Data was expressed as the mean ± standard deviation. R statistical package was also used for PCA analysis. Analysis of correlation was performed by calculating Pearson correlation coefficient (r) between the highest levels of accumulated metabolite and the AUDPC, similar to one described earlier (Sahu et al., 2016b). SA, H₂O₂, vanilic acid, chlorogenic acid, p-caumaric acid, and free ferrulic acid were evaluated for correlation analysis at 12 hpi, total phenolic content, 4-hydroxybenzoic acid, syringic acid, caffic acid and bound ferrulic acid were evaluated at 24 hpi, whereas SOD and MDA contents were evaluated at 72 hpi (further detailed in results).

A total of 968 wheat genotypes (Supplementary Table 1) were screened at five geographical locations in 2 years to evaluate the variability for spot blotch resistance (Figure 1). Disease progression (in form of AUDPC) was calculated and genotypes were categorized into resistant and susceptible (detailed in Methods sections). From the field studies of year 1 (2012–13), 31 genotypes were found resistant (Supplementary Figure 1 and Supplementary Table 2): Three genotypes (192, 335, and 337) were resistant at all the five locations, 8 genotypes (89, 228, 236, 294, 297, 346, 390, and 483) were resistant at four locations, whereas 20 were screened as resistant at the three locations. All the resistant genotypes showed large variation for plot yield (149.1–283.1 g/plant), days to heading (75.7–87.7) and TKW (24.5–41.7 g) (Supplementary Table 1). Similarly, 15 wheat genotypes were screened as “resistant” from the second year (2013–14) of field studies (Supplementary Figure 1 and Supplementary Table 2). PCA clusters also largely coincided with the phenotypic selection (Supplementary Figure 1).

A total of 46 wheat genotypes were finally identified as spot blotch resistant from the field screening (Supplementary Table 1). All these 46 genotypes were confirmed as “resistant” in subsequently independent field studies at all the five locations over 3 years (2014–17; Figure 1 and Table 1C and Supplementary Table 3). Pedigree (along with genotype identifiers of field trails the entry and CRP numbers) of these 46 genotypes, a previously characterized resistant genotype (check), Yangmai #6, and eight susceptible controls are given in Table 1B. Hereafter, these genotypes are referred as CRP1-55. The mean AUDPC varied from 134.6 to 369.5 for resistant genotypes, including the 47th resistant genotype (Yangmai #6). As expected, on the other hand, AUDPC values of most of the susceptible genotypes were much higher and varied in the range of 524.1–861.0 (with an exception of genotype CRP39 at BHU Varanasi location with an AUDPC of 363) (Table 1C). The mean variation in plot yield for resistant (150.6-411 g/plant) and susceptible (143.8–374.3 g/plant) genotypes were also measured (Supplementary Table 4). A comparative analysis of 55 genotypes on the basis of AUDPC separated the susceptible and resistant genotypes and ranked them differently for disease progression response (Supplementary Figure 2, Supplementary Table 3). These observations suggest that the 46 genotypes identified in field trails could be used as sources of resistance against spot blotch across South-Asia.

To determine the basis of differential response of wheat genotypes to spot blotch infection, we measured several signaling components involved in disease resistance. Since ROS play an important role in plant defense against pathogens (Liu et al., 2010), a time course study was performed to analyze H₂O₂ accumulation during spot blotch infection in wheat genotypes (to evaluate the ROS activity). For the resistant genotypes, enhanced H₂O₂ accumulation was observed as early as 12 hpi (Figure 2A and Supplementary Table 5). Amongst the resistant genotypes, CRP21 showed highest H₂O₂ accumulation after 12 hpi (2.4-fold), whereas CRP41 had the lowest H₂O₂ accumulation (Figure 2A and Supplementary Table 5). All susceptible genotypes failed to accumulate H₂O₂ over the time course of infection. Except CIANO T79 and CRP55, all the susceptible lines clustered together in a single group (Figure 2A). Further, a negative correlation between H₂O₂ accumulation and AUDPC of 55 CRP lines was recorded (r = −0.60; P ≤ 0.05; Figure 2B), indicating that H₂O₂ elicitation may be linked to elicitation of defenses against spot blotch infection.

SOD is an antioxidant enzyme that maintains the steady state level of ROS in the plant cells under various stresses. The level of SOD was almost similar in resistant as well as susceptible genotypes before infection (0 hpi), whereas, the activity was elevated in resistant genotypes at 24 hpi, which further increased at 48 and 72 hpi (Figure 2C and Supplementary Table 5). In contrast, susceptible genotypes failed to elicit significant changes in the SOD level within 24 hpi. Although, marginal increase in SOD levels was noted at 72 hpi in susceptible genotypes, these levels were significantly lower than in the resistant lines (Figure 2C and Supplementary Table 5). Overall, SOD level showed a negative correlation to AUDPC (r = −0.45; P ≤ 0.05; Figure 2D).

Malondialdehyde (MDA), a secondary product of lipid peroxidation, is a commonly used marker of cell membrane damage, and its levels usually increase in response to biotrophic infection (Torres et al., 2006). When we examined MDA content during the time course of spot blotch infection, no significant differences were observed across the genotypes at 0 hpi (Figure 2E and Supplementary Table 5). The MDA levels were increased at 24 hpi; moreover, the levels were high throughout the time course of infection in susceptible genotypes as compared to resistant lines (Figure 2E and Supplementary Table 5). CRP5, a susceptible genotype, showed highest MDA content (0.945 μmoles/g FW) at 72 hpi, followed by CIANO T79 (0.823 μmoles/g FW). All the susceptible genotypes, except Sonalika and CRP39, formed a single cluster (Figure 2E). A strong positive correlation of MDA content and AUDPC was observed (r = 0.74; P ≤ 0.05; Figure 2F).

We determined the natural variation in pathogen-induced SA accumulation in wheat and its association to resistance. Our results indicate that the elicitation of spot blotch-induced SA is directly associated with resistance. At 0 hpi, no significant differences were observed for SA levels across 55 wheat genotypes (Figure 3A and Supplementary Table 6). However, at 12 and 24 hpi, the resistant lines displayed an increase in the SA levels (49.62–93.79 μg/g fresh weight; Figure 3A and Supplementary Table 6). CRP45 showed maximum elicitation of 4.5-fold after spot blotch infection at 12 hpi, which remained high at 24 hpi (as compared to time 0). All the susceptible genotypes, including CIANO T79 and Sonalika, did not show significant increase in accumulation of SA after spot blotch infection and formed a single cluster (Figure 3A). The resistant genotypes were grouped in two clusters, with their maximum SA contents at 12 and 24 hpi, respectively (Figure 3A). SA displayed a strong negative correlation of r = −0.73 (P ≤ 0.05) with AUDPC (i.e., susceptibility; Figure 3B). It may be hypothesized that the failure of eliciting SA-dependent defense signaling may result in increased disease progression (susceptibility of genotypes).

Since phenolic acids and their derivatives are known to play multiple roles in plant pathogen interaction, and that their elicitation may be regulated by SA, we measured spot blotch-induced phenolic acid accumulation in 55 genotypes. A clear increase in accumulation of phenolic acids after B. sorokiniana infection was recorded in resistant genotypes than the susceptible genotypes at 12 and 24 hpi (Figure 4 and Supplementary Table 7). Resistant and susceptible genotypes formed two distinct clusters, each, for 4-hydroxybenzoic acid and syringic acid.

Further statistical analysis showed strong inverse correlation between pathogen-induced content of phenolic compounds and the AUDPC (Figure 5). Most strong negative correlation was observed for syringic acid (r = −0.84; P ≤ 0.05), followed by 4-hydroxybenzoic acid (r = −0.67; P ≤ 0.05; Figure 5).

Due to a strong negative correlation of syringic acid and AUDPC (Figure 5), we hypothesized that it may act as a defensive phytochemical with antimicrobial property. To test direct inhibition of B. sorokiniana growth by syringic acid, we performed in vitro pharmacological plate assays. The inhibitory effect of syringic acid on fungal growth was evident. A significant inhibition of >40% was observed with as little as 125 μg/ml concentration (Figures 6A,B; one-way ANOVA; Tukey's test; P ≤ 0.05), whereas higher concentrations of syringic acid further reduced the growth of B. sorokiniana (>80%, compared to control plates, 0 μg/ml; Figures 6A,B; one-way ANOVA; Tukey's test; P ≤ 0.05).

To further validate the role of syringic acid in plant defense against spot blotch disease, we performed complementation assays. Sonalika plants, which are highly susceptible to spot blotch and fail to elicit any significant changes in syringic acid after B. sorokiniana attack (Figure 4), were subjected to foliar application of 0.1, 0.2, and 0.5 mM syringic acid concentrations. Sonalika plants, exogenously complemented with syringic acid, had lower disease severity as well as AUDPC than the Sonalika control (0 mM) plants treated in identical manner (Figure 6C). Most significant difference was observed at 9 dpi where 0.5 mM syringic acid treated wheat had 31.4% reduced spot blotch severity than untreated plants (Figure 6D). Thus, the strong negative correlation of syringic acid to disease progression, the in-vitro growth inhibition assays and exogenous complementation assays, together, show that syringic acid contributes to spot blotch resistance in wheat by, for example, possibly acting as an antifungal compound, whose accumulation is enhanced after spot blotch infection.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.