Introduction

Rice (Oryza sativa L.) is an important cereal crop, representing the food for the sustenance of more than half of the world’s population. The normal growth and development of the plant are often deterred when the nutrient supplies are limited; healthy plants are often synonymous with an optimum supply of nitrogen (N), which is fundamental to the growth and productivity of cereal crops. The nitrogen nutrient forms an important constituent of key molecules such as nucleic acids, amino acids, ATP, chlorophyll and several plant hormones. The deficiency of N is often visualized by the pale green- or yellow-colored leaf tissues, indicating a decrease in the photosynthetic pigments and vital enzymes related to the precursors (Good et al. 2004).

Farmers in many parts of the world prefer to apply N fertilizer in excess in an attempt to increase the yields. They fail to realize that crops use only a small quantity of N fertilizer that is applied during a season and the rest leaches into the underground water reservoirs, leading to environmental pollution. Also, plants get predisposed to disease caused by obligatory or parasitic pathogens, when excessive N is applied (Gothoskar et al. 1995). The blast disease caused by Magnaporthe oryzae in rice is devastating, especially when the rice plants are irrigated or they receive higher amounts of rainfall or higher levels of nitrogen fertilizer. The epidemics of rice blast disease in different parts of the world and their pervasiveness reduce the productivity by 50–90%, as the pathogen can infect all aerial tissues, including leaves, stem, and panicles (neck blast disease). The distinct characteristics of the disease include large ellipsoid lesions on the surface of rice leaves. In older rice plants, neck blast symptoms can result in the complete loss of the rice crop, when the fungus spreads into the panicles (Talbot 2003). M. oryzae is an ascomycete reproducing asexually by producing conidia in nature, with its sexual stage being amenable to genetic analysis of its pathogenicity in the laboratory. Under humid conditions, the three-celled spores produced from lesions get transported to new hosts by wind or dispersed through splash; such spores after landing on leaves start germination, beginning the life cycle of M. oryzae (Ribot et al. 2008). Efforts are being continually made to develop resistant/tolerant cultivars by incorporating the broad-spectrum resistance genes through breeding approaches (Ou 1980; Onaga and Asea 2016). Newer strategies include marker-assisted backcross breeding and gene pyramiding, which lead to more durable resistant cultivars (Koide et al. 2010; Ellur et al. 2016). Genetic study-based programs led to the development of near-isogenic lines, with single blast resistance genes by backcrossing donor cultivars such as CO-39 (Mackill and Bonman 1992), while conventional breeding combined with marker-assisted selection has helped in pyramiding genes governing resistance to blast (Lv et al. 2013).

Phyllosphere refers to the exposed surface of the plant, which supports a wide assortment of microorganisms ranging from eubacteria, archaea, cyanobacteria, filamentous fungi, protozoa and even nematodes. The microbial populations in this niche show interesting dynamics in different environments (Yadav et al. 2004), with specialized functions which make them more adapted to this milieu, such as nitrogen fixation or enzyme activities related to utilization of carbon sources (Sebastian et al. 1987; Giri and Pati 2004). Nitrogen fixation in the phyllosphere has been investigated using both culture-based and molecular tools and several diazotrophs have been isolated (Ruinen 1971; Patti and Chandra 1981; Murty 1983; Knief et al. 2012). Wu et al. (2009) observed in their studies related to the influence of N fertilizer application in rice cultivars, that the spatial or niche related aspects, such as rice roots versus rhizosphere, surface vs. the deeper layer soils, exert a greater influence on the soil archaeal community composition. However, the abundance of microbial groups or nitrogen fixers in pathogen-challenged rice phyllosphere of different cultivars, as influenced by different N doses, using molecular tools is less investigated.

Sharma (2007) observed that the excessive use of N fertilizers results in greater leaf growth and over-succulence; this makes it more susceptible to certain diseases. The nitrogen disease hypothesis states that an increased susceptibility of plants to pathogens is observed in plants grown under high nitrogen (N) availability, due to increased foliar nitrogen concentrations (Mitchell et al. 2003). Although fertilization is also known to decrease the disease resistance of plants occasionally, it may also increase disease tolerance, as a result of growth stimulating effect on plants in an environment with increased nutrient availability (Panique et al. 1997). On the contrary, the deficiency of N induces the modification of several morphological and physiological parameters, growth limitation, and even reduction in the numbers and area of leaves (Ruckstuhl 1998). In the present study, the interactive effect of fertilization (doses of N) and pathogen inoculation on plant growth, elicitation of defense responses and the abundance of different phyllosphere microbial communities of rice cultivars (blast resistant and susceptible) and nifH and amoA genes was investigated. We examined the hypothesis that the rice cultivars may respond differentially to increased doses of nitrogen fertilizer and pathogen inoculation, both in terms of functional traits and the ability to support microbial members with different functions, particularly related to N metabolism, on the leaf surface.

Materials and methods

Environmental chamber and conditions

The experiments were undertaken at the National Phytotron Facility, Indian Council of Agricultural Research (ICAR)-Indian Agricultural Research Institute (IARI), New Delhi (28°40′N, 77°12′E), which is located at an altitude of 228.6 m above mean sea level. The growth conditions of the environmental chamber were 28 °C, 90% RH, 14/10 h day/night. Mesocosm experiments were set up, using Pro-Trays with 18 cells, each supporting 300 g soil belonging to Typic Haplusterts, having pH 7.4. Macronutrient content of the soil was: available N − 213.25 kg ha−1, available P − 16 kg ha−1, organic C − 1.01%. Micronutrient content was Zn 24.5, Fe 11.7, Cu 21.42, Mn 35.67 µg g−1 soil. The seeds of four rice cultivars, namely, CO-39, CO-39I, PB-1, and P-1609 were obtained from the Division of Genetics, ICAR-IARI, New Delhi. The cultivars of PB-1 and P-1609 are of basmati (aromatic) type, while CO-39 and CO-39I are non-basmati type; within these two categories, one each was blast susceptible (CO-39 and PB-1), and the others blast resistant (CO-39I and P-1609). Basal doses of fertilizers, urea as a source of N, single super phosphate and muriate of potash for P and K, respectively, were applied over and above the native nutrient concentrations of the soil. Three doses of nitrogen along with P and K were applied: (1) 0 N (N 0:P 60:K 60 kg ha−1), (2) recommended dose 120 N (N 120:P 60:K 60 kg ha−1) and (3) 180 N (N 180:P 60:K 60 kg ha−1).

Planting

The seeds of rice were surface-sterilized using sodium hypochlorite for 3 min, followed by rinsing with distilled water, washing with 100% ethanol for 30 s, and then repeated rinsing with distilled water, to remove any traces of ethanol. Seeds were then soaked in sterile water for 24 h and were allowed to germinate in water agar (0.75% agar) for about 3 days. The germinated seeds were then transferred into the Pro-Trays having a thin film of standing water over the soil.

Pathogen inoculation

Magnaporthe oryzae RMg-D1 isolate (Kumar et al. 2017) was procured from the Division of Plant Pathology, ICAR-IARI, New Delhi. The fungus was maintained using the rice straw extract oat meal agar (rice extract 1 L; sucrose 40 g; oat meal agar 33 g and agar 13.3 g), supplemented with streptomycin (100 µg mL−1). The fungus was allowed to grow at 25 °C for 7–10 days and the number of spores was checked regularly; when the density of spores was about 106 mL−1, the spores were harvested in a mixture of Tween-20 (2–3 drops) and sterile distilled water.

The rice plants were challenged by the inoculation of the fungal pathogen, using the foliar spray of spores, at 3-leaf stage (15 days after transplanting). Spraying was done until the leaf surface was wet uniformly, after which the plants were left in the dark for 24 h. After the dark period, the light was normally provided in cycles of 14/10 h light/dark. Sterile distilled water was sprayed at intervals on the uninoculated plants. The humidity of more than 90% was maintained uniformly inside the environmental growth chamber.

Pigment and nutrient estimation

The leaf samples were collected 15–20 days after pathogen inoculation. Chlorophyll a, b and carotenoids were determined spectrophotometrically at 480, 510, 645, 663 nm after extraction of pigments from leaves with dimethyl sulphoxide (DMSO) as described by Jeffrey and Humphrey (1975). The leaf samples were dried and finely ground before being used for estimating total nitrogen content by the Kjeldahl’s method, which was recorded using N-autoanalyzer and phosphorus content spectrophotometrically (Jackson 1967). Micronutrients were also estimated in the leaves, after grinding to fine powder and digestion with di-acid mixture (nitric acid and perchloric acid). These samples were measured for micronutrient concentrations (Fe, Zn, Cu, and Mn) using an Atomic Absorption Spectrophotometer, at the respective wavelengths for Fe (248.7 nm), Zn (213.7 nm), Cu (324.6 nm), and Mn (279.5 nm) (Lindsay and Norvell 1978).

Biochemical analyses

The leaf samples collected 15–20 days after pathogen inoculation were used for the preparation of leaf extracts using 0.1 M sodium phosphate buffer (pH 7.2), for the analyses of hydrolytic enzyme activities as described by Prasanna et al. (2013, 2016). All samples were taken in triplicate and values expressed as IU g−1 fresh weight of tissues. After washing the leaf samples (1 g) in running tap water, they were homogenized using 5 mL of sodium phosphate buffer (0.1 M, pH 5.2) in a mortar and pestle. After centrifugation at 14,000g for 20 min at 4 °C, the supernatant was transferred to vials (5 mL) and stored at − 20 °C. Glycol chitosan and carboxymethyl cellulose served as the substrates for determining the activities of chitosanase (EC 3.2.1.99) and β-1,4 glucanase (EGases, EC 3.2.1.4), respectively. Chitosanase activity was determined by adding 100 µL of 1% glycol chitosan to aliquots of leaf extracts (500 µL) as described by Ohtakara (1988), and absorbance was measured at 530 nm and was expressed as µmol glucose released min−1 g−1 fresh weight of tissues. The activity of β-1,4 glucanase was determined using the leaf extracts (500 µL) and carboxymethyl cellulose (500 µL, 1%) as given by Ghosh et al. (1983). One IU (International unit) of hydrolytic enzyme activity was defined as 1 µmol of glucose (for β 1,3 endoglucanases/CMCase) or glucosamine (for chitosanase) released min−1.

The activity of polyphenol oxidase (PPO) was measured using catechol as the substrate (Ghosh et al. 1983). The enzyme activity was determined spectrophotometrically at 546 nm by monitoring the changes in absorbance at 30 s intervals for 3 min. The peroxidase (PO) activity was measured using guaiacol (molar extinction coefficient 26.6 mM−1 cm−1) as the hydrogen donor (Ghosh et al. 1983) and the changes in absorbance at 470 nm were recorded at 30 s intervals for 3 min. The phenylalanine ammonia lyase (PAL) activity was assayed in leaf extracts by measuring the amount of trans-cinnamic acid formed from l-phenylalanine spectrophotometrically at a wavelength of 290 nm against the blank, as given by Prasanna et al. (2013, 2016). In the blank, 0.1 mL of distilled water was used in place of the test sample. The enzyme activity was expressed as nmole of cinnamic acid h−1 g−1 freshweight.

Antioxidant enzyme activities in the leaf samples were analyzed after homogenization using phosphate buffer in liquid nitrogen. Inhibition of photochemical reduction of nitrotetrazolium blue chloride (NBT), measured at 560 nm was taken as an index of superoxide dismutase activity (Beauchamp and Fridovich 1971). Catalase activity, based on the initial rate of disappearance of H2O2 at 240 nm (Bergmeyer 1970), was measured using a reaction mixture containing 0.05 M Na-phosphate buffer (pH 7.0), mixed with 0.1 mM EDTA and 3% H2O2. One unit of catalase activity represented the rate at which 1 mol H2O2 was destroyed per min.

Disease scoring indices

The disease scoring was based on the details given by Mackill and Bonman (1992). The disease reactions were scored 15 days after pathogen inoculation on a scale ranging from 0 to 5, in which 0 represented no evidence of infection, 5 was representative of appearance of typical spindle-shaped blast lesions, 3 mm or longer with little or no coalescence of lesions, along with more than half the number of leaves killed by the coalescence of lesions.

DNA extraction and qPCR analyses

Phyllosphere suspensions were prepared using the leaf samples of 1–2 g, cut into 1 × 1 cm pieces and placed in sterile Tris–EDTA buffer (10 mmol L−1 Tris–HCl, 1 mmol L−1 EDTA, pH 8.0), as described earlier (Thapa et al. 2017, 2018). After centrifugation at high speed to concentrate the cells as the pellets, aliquots (500 µL) which included the pellets along with minimal amounts of leaf washings, were used for extracting DNA using the Bacterial/Fungal isolation kit (Nucleopore, Genetix Biotech Asia), following the manufacturer’s instructions. The 16S rRNA gene copies of bacteria, archaea, and cyanobacteria, and the functional gene copies of nifH and bacterial amoA genes were quantified using the qPCR method. The qRT-PCR experiments were conducted with the Roche Light Cycler® 96 instrument (Roche Diagnostics Corp., Indianapolis, IN, USA). Fast SYBR® Green dye was employed to determine the gene copy number, as it is an intercalating dye which binds to the double-stranded DNA, and fluorescence can be measured at the end of each amplification cycle, indicating quantitatively, the amplification of the desired gene. Primers used in this study and thermal cycling profile are listed in Table S1. The specificity of amplification was confirmed by melt-curve analysis after 40 cycles, i.e., 95 °C for 15 s and 65 °C for 60 s, 97 °C for 1 s. A standard curve was generated by ten-fold serial dilutions of purified genomic DNA. CT values were plotted against log (copy number) and the values of CT were plotted against log (copy number). The concentration of genomic DNA and the cloned plasmid was determined using Nanodrop 3300 spectrofluorometer (Waltham, MA, USA) and the gene copy numbers were calculated and expressed in terms of per cm2 leaf area.

Statistical analyses

The experimental data were analyzed using the statistical program WASP version 2.0 (Web Agri Stat Package, Indian Council of Agricultural Research) in a three-factorial design. Three-factor analysis of variance (ANOVA) using Factor A (pathogen challenge), Factor B (rice cultivars) and Factor C (nitrogen doses) was performed and the details are given in the legends of figures and table. Paired t test was performed using Microsoft Excel, to investigate the correlation between the different nitrogen doses and the disease scores.

Results

Photosynthetic pigments

The concentrations of chlorophyll a varied significantly among the different rice cultivars tested, especially when challenged with the fungal pathogen. In general, the PB-1 and P-1609 showed much lower values, which were enhanced significantly by pathogen inoculation. The susceptible cultivar CO-39 showed significant decreases (by 0.4–1.3 mg g−1 fresh weight) after pathogen inoculation at higher N doses (Fig. 1a). In contrast, the concentrations of chlorophyll a in the resistant cultivars, CO-39I and P-1609, and the susceptible cultivar PB-1 showed increased concentrations of chlorophyll a after the pathogen inoculation. The cultivar CO-39 showed a minor reduction with the increasing nitrogen doses, and pathogen inoculation also led to a reduction, but N doses were not significantly influential. The cultivar CO-39I showed a stimulatory response to increasing N doses, particularly when there was no inoculation of the fungal pathogen. The highest concentration of chlorophyll a was recorded in the cultivar PB-1 with 0 N dose, while the cultivar P-1609 exhibited the lowest concentrations of chlorophyll a in samples without the pathogen inoculation, but almost two-fold higher values after pathogen inoculation at all N doses.

Fig. 1
figure 1

Effect of disease incidence, varietal differences and nitrogen rate (kg ha−1) on a chlorophyll a content and b PAL activity. Error bars indicate the standard deviation of the mean from three biological replicates, each having five technical replicates. Factor A (disease challenge); Factor B (rice cultivars); Factor C (nitrogen doses); CD critical difference, **significant at 1%, NS not significant

Full size image

The concentrations of chlorophyll b differed among the cultivars tested, with higher concentrations observed in CO-39I, followed by CO-39 (Table 1). In pathogen-inoculated plants, lower values were recorded only for CO-39, whereas 0 and 120 N (recommended dose of fertilizers) led to higher values in CO-39I. Lower values were recorded in the basmati cultivars—PB-1 and P-1609 as compared to the other two cultivars—but the values increased with pathogen inoculation in both the cultivars. N doses showed a differential response; highest values were observed in absolute control with the cultivars CO-39 and CO-39I, while higher dose showed more chlorophyll b in case of PB-1 and P-1609. The response of the different cultivars, in terms of carotenoid content was strikingly different, with highest values observed in CO-39I, followed by CO-39 (Table 1). Pathogen inoculation led to lowering of values, in the case of CO-39, but increased in CO-39I, with 0 N treatment showing the highest values in both the cases. P-1609 and PB-1 also showed higher values with 180 N and pathogen inoculation increased these values.

Table 1 Plant pigments in different rice cultivars, as influenced by N doses and pathogen challenge
Full size table

Pathogen-challenged plants showed non-significant effects of fertilizer doses in terms of both chlorophyll b and carotenoids. However, cultivars showed significant effects and interacted with both fertilizer doses and pathogen challenge significantly, based on ANOVA.

Lytic and defense enzymes

CMCase values differed across various cultivars, with CO-39 recording the highest value followed by CO-39I, P-1609 and PB-1, in control plants (Table 2), while cultivars CO-39 and P-1609 recorded highest values in pathogen-challenged plants. Pathogen inoculation led to an enhancement in the values in most treatments, with a sharp increase in P-1609 for the 0 N samples. ANOVA illustrated that the cultivar type had a more prominent effect on these values, followed by fertilization and pathogen inoculation and their interactions were significant at 1% level of probability.

Table 2 Activity of hydrolytic enzymes in different rice cultivars, as influenced by N doses and pathogen challenge
Full size table

Chitosanase values were also different between the various cultivars, and higher values were observed in case of CO-39 followed by CO-39I, P-1609 and PB-1 (Table 2). Fertilization also had an important role, and higher dose was found to be stimulatory for CO-39 and PB-1. Moreover, pathogen challenge increased the chitosanase values in case of CO-39I, PB-1, and P-1609, whereas uninoculated plants showed higher values in CO-39. Cultivar-based differences did not show any significant interactions with fertilization or pathogen challenge. In case of phenylalanine lyase activity, the plants belonging to cultivar CO-39I showed significantly higher values, followed by CO-39, P-1609 and PB-1 (Fig. 1b). Pathogen inoculation led to significantly greater elicitation in CO-39 and PB-1, while N doses, 120 N in CO-39I and 180 N in P-1609 were more influential in elicitation. Only cultivars and pathogen inoculation individually had a significant role; the effect of fertilizers and the tripartite interactions were not significant.

In case of the activity of polyphenol oxidase, cultivar-specific differences were significant, with the highest activity being observed in CO-39I (Table S3). Pathogen inoculation also increased the enzyme activity in all the cultivars, except PB-1. Although, with an increase in the dose of N, higher values were recorded in pathogen-challenged plants, no significant effect of fertilizers was recorded. Peroxidase activity was significantly enhanced by pathogen challenge, as observed in CO-39, followed by CO-39I, P-1609 and PB-1 (Table S3). Catalase values did not vary significantly across the cultivars (Table S2).

Superoxide dismutase activity was highest in PB-1, followed by P-1609, CO-39, and CO-39I (Table 3), and was enhanced by pathogen inoculation in all the cultivars, except P-1609 and CO-39. Although the effect of N dose alone was not significant, its interactions with cultivars were only significant at 5% level of probability. Interaction effect of all three cultivars, pathogen inoculation and fertilization, the interaction of cultivars and pathogen inoculation, as well as the interaction of pathogen inoculation and fertilization, were all significant at 1% level of probability.

Table 3 Antioxidant enzyme activity, as elicited in different rice cultivars, and influenced by N doses and pathogen challenge
Full size table

Visibly, the pathogen-susceptible cultivars exhibited a higher incidence of disease, as compared to the pathogen-resistant cultivars and the disease severity scores showed a significant increase with higher doses of N, except for P-1609 (Supplementary Figure). The cultivar CO-39I exhibited a three-fold enhancement, with an increase in N level from 120 to 180. The susceptible cultivar CO-39 recorded the highest values, in terms of incidence of disease/disease scores, followed by PB-1 (a susceptible cultivar) at 180 N (Table 4). Significant differences were recorded which were cultivar-specific, besides the effect of N doses; the interaction of fertilizer doses with the cultivars was significant at 1% probability.

Table 4 Disease scoring indices, as influenced by fertilizer doses and rice cultivars
Full size table

Leaf macro and micronutrients

The cultivars differed widely in terms of leaf N values, with highest values observed in CO-39I, followed by CO-39, P-1609 and PB-1. Leaf N values increased with the higher doses of N in control treatments, except for P-1609 where the values decreased slightly for 180 N. as compared to RDF. When the leaves were challenged with pathogen, leaf N values were much higher at the 120 and 180 N doses, except in CO-39. Although the tripartite interaction was observed to be significant, the interactions between cultivars or fertilizer doses and pathogen challenge, were not significant (Fig. 2a).

Fig. 2
figure 2

Effect of pathogen challenge, varietal differences and nitrogen rate (kg ha−1) on leaf macronutrients; a leaf total nitrogen and b leaf total phosphorus. Error bars indicate the standard deviation of the mean from three biological replicates, each having five technical replicates. Factor A (pathogen challenge); Factor B (rice cultivars); Factor C (nitrogen doses); CD critical difference, *significant at 5%, **significant at 1, NS not significant

Full size image

P content of the leaf varied across different cultivars with highest values being observed in CO-39I, followed by CO-39, PB-1 and P-1609 (Fig. 2b). After pathogen challenge, the values of P increased in different cultivars, except P-1609. Both the cultivars—CO-39 and CO-39I, exhibited highest values in both sets of treatments. In general, despite a minimal increase at 120 N, over 0 N; plants from 180 N showed lower values and the interactions among the cultivars, N doses and pathogen challenge were significant.

In the case of leaf zinc content, CO-39 samples had highest values followed by CO-39I, PB-1, and P-1609 (Table 5). Cultivars, fertilizer N dose, and pathogen challenge significantly influenced the zinc content of leaves, but no correlation/trend between the factors could be observed. In terms of leaf copper, PB-1 had the highest values followed by CO-39I, CO-39, and P-1609. Leaf iron content also differed significantly across cultivars with highest values observed for CO-39I, followed by CO-39, P-1609, and PB-1 (Table S4). PB-1 showed a sharp increase at 180 N and showed much lower values in pathogen-challenged plants, while both CO-39 and PB-1 exhibited higher values in the pathogen-challenged plants. PB-1 showed a similar trend for leaf manganese, as observed for iron. Although the interaction of N doses with pathogen challenge alone was not significant, the interactions among the cultivars, N doses and pathogen challenges were significant in terms of micronutrient content of leaves.

Table 5 Leaf micronutrient content as influenced by fertilizer doses, rice cultivars and pathogen challenge
Full size table

Abundance of microbial members

The abundance of eubacteria (log copy number of 16S rRNA gene copies cm−2 leaf) was highest in var. PB-1 and P-1609, under the control treatments (no pathogen challenge) with 120 N, and pathogen inoculation also enhanced the abundance in var. PB-1. The dose of 120 N application also led to enhancement in the basmati cultivars: PB-1 and P-1609 (Fig. 3a). At a higher N dose of 180 N, a decrease was recorded in the cultivars CO-39, PB-1, and P-1609 in the control treatments (no pathogen challenge), and pathogen inoculation brought about a distinct increase in the abundances of eubacteria in var. CO-39 and PB-1, at all nitrogen doses. The application of higher N doses in the control treatment led to increased abundances of archaea in var. CO-39 and PB-1, while there were decreases in var. CO-39I and P-1609. Pathogen inoculation led to increased abundances of archaea in all the cultivars tested at 180 N dose (Fig. 3b).

Fig. 3
figure 3

Real-time PCR analyses of 16S rRNA genes of various groups of microorganisms, as elicited by pathogen challenge, nitrogen doses and varietal differences; a eubacteria, (b) archaea and c phylum Cyanobacteria. Error bars indicate the standard deviation of the mean from three biological replicates, each having five technical replicates. Factor A (pathogen challenge); Factor B (rice cultivars); Factor C (nitrogen doses); CD critical difference, **significant at 1%, NS not significant

Full size image

The abundance of cyanobacterial gene copies was higher in var. CO-39I and PB-1, relative to those of CO-39 and P-1609 in the control treatments. Increased N application led to overall decreased abundances of cyanobacterial gene copies in all the cultivars tested for the 0 N treatment, but increases, at least by one-fold in fungal pathogen-treated samples for CO-39I and PB-1 was observed. However, a combination of higher doses of N and pathogen-challenged treatments/cultivars led to decreased abundances of cyanobacteria (Fig. 3c). The copies of functional gene nifH also showed a decrease in samples from higher N treatments, with or without the fungal inoculation, relative to the respective control treatments for the non-basmati cultivars, but an overall increase was recorded with pathogen inoculation for the basmati cultivars (Fig. 4). However, the copies of functional gene bacterial amoA increased due to the application of higher N, with or without the fungal inoculation. An increase, at least by one-fold was recorded in the cultivar P-1609, with higher N application under both the pathogen inoculated or control treatment.

Fig. 4
figure 4

Real-time PCR analyses of selected genes; a nifH and b amoA, as elicited by pathogen challenge, nitrogen doses and varietal differences. Error bars indicate the standard deviation of the mean from three biological replicates, each having five technical replicates. Factor A (pathogen challenge); Factor B (rice cultivars); Factor C (nitrogen doses); CD critical difference, *significant at 5%, **significant at 1%, NS not significant

Full size image

Discussion

Rice crop faces several biotic and abiotic stresses which reduce its productivity; among which blast and bacterial blight diseases caused by M. oryzae and Xanthomonas oryzae, respectively, are a major scourge (Talbot 2003; Variar et al. 2009). The present study was conducted to study the influence of nitrogen fertilization on the blast pathogen-susceptible and -resistant cultivars and their interactive effects on the plant growth, elicitation of plant defense/antioxidant machinery and the phyllosphere microbial communities. Efforts in India have focused on developing durable, resistant cultivars through the incorporation of resistance genes, particularly in basmati cultivars, mainly through molecular breeding, including gene pyramiding and marker-assisted selection (Variar et al. 2009; Ellur et al. 2016). Pusa Basmati 1 (PB-1) is a popular basmati cultivar, susceptible to blast disease, while Pusa 1609 (P-1609) is the first MAS-derived neck blast-resistant basmati rice cultivar; both these are grown in several hectares in India (Ellur et al. 2016). Another approach has been the development of near-isogenic lines (NILs), which have proved valuable in breeding programs, as they can be used for mapping resistance genes with molecular markers (Yu et al. 1987). CO-39I is a near-isogenic line, developed by backcrossing using the indica cultivar CO-39 (highly blast susceptible).

Nitrogen (N) is one of the most important nutrients, which is a major component of chlorophyll and correlated with leaf color, crop growth status and yield. The use of N fertilizers has been an integral component of the package of practices recommended for increasing the crop productivity over the past few decades (Kant et al. 2011). Inputs of N are known to elicit and stimulate plant defense reactions, but the excess N also becomes available for exploitation by the pathogenic microorganisms (Tavernier et al. 2007). Host–pathogen interaction, particularly in conditions with greater availability of N is known to elicit and stimulate plant defense reactions, but is also exploited by the pathogenic microorganisms (Mur et al. 2016; Tavernier et al. 2007). Higher supply of N to the plant was found to lead to higher spore production by the powdery mildew fungus (Jensen and Munk 1997) and enhance the leaf colonization by Pseudomonas syringae (Gupta et al. 2013). In the present investigation, increased N fertilization led to differential responses by the cultivars, in terms of photosynthetic pigment content of leaves; however, cultivars interacted significantly with both fertilizer doses and pathogen challenge. In our earlier work (Thapa et al. 2017), a significant correlation was recorded between the nutritional content of leaf and soil, with the abundance of bacteria and their functional attributes; in the present investigation, the effects of tripartite interactions of cultivars with pathogen challenge or N doses on micronutrient content were mostly significant.

Loper and Lindow (1994) observed that the phyllosphere is a unique niche, with a major portion of the leaves harboring scanty, with non-uniform amounts distribution of nutrients, while few “oases” of relatively abundant nutrients hosted the majority of inhabitants. Often, the generous use of nitrogenous fertilizers has been responsible for enhancing disease progression and development (Solomon et al. 2003). Contradictory reports regarding the role of N in pathogenesis are available; some researchers suggest that starvation of N controls pathogenicity genes and can act as cue for disease development (Lopez-Berges et al. 2010), although a majority suggest that excessive supply of N may stimulate pathogen growth (Pageau et al. 2006; Walters and Bingham 2007). This can be attributed to cultivars, pathogen types or timing of N application, etc. (Long et al. 2000).

Higher doses of N application is often responsible for the increase in the leaf N content, leading to the greater disease susceptibility to yellow rust of wheat caused by Puccinia striiformis (Neumann et al. 2004). Higher supply of N to the plant may also lead to the higher spore production by the powdery mildew fungus (Jensen and Munk 1997) and increased leaf colonization by Pseudomonas syringae (Gupta et al. 2013). Leaf N, which is an inherent trait, showed distinct differences among the cultivars; however, PB-1 and CO-39 exhibited a more linear response to doses of N application, with/without pathogen inoculation. Despite highest values of leaf N in CO-39I, higher doses of N did not lead to concomitant increases in the leaf N, in with/without pathogen inoculation treatments; rather a marginal lowering was recorded. This was supported by our observations that the leaf photosynthetic pigments showed stimulation at 120 N over 0 N, but in general, showed a reduction at 180 N.

Increased N can often contribute to the level of plant defense; constitutive or induced. The toxicity of NH4+ is likely caused by oxidative stress from the excessive accumulation of reactive oxygen species (ROS). ROS formation and induction of oxidative stress-related genes represent an important plant defense response, aimed to control such oxidative stress and plants have developed numerous strategies for the detoxification of ROS. Among anti-oxidative enzymes, superoxide dismutases (SODs) and peroxidases (PODs) play key roles in ROS detoxification in cells. Kong et al. (2017) found that, urea, which is the most commonly used source of organic N, decreased the levels of two antioxidant enzymes (SODs and PODs) in the leaves of wheat under the excessive N condition, compared with those in wheat treated with an appropriate level of N. In the present study also, urea was used as a nitrogen source, but its doses did not significantly influence the values of SOD and POD activities; however, the interactive effect of N doses with cultivars or pathogen challenge or their tripartite interactions were significant.

Plants are known to manipulate their primary and secondary metabolism in response to pathogen infection which is influenced by the level of N (Ward et al. 2010). Biological mineralization of organic materials can also influence dynamics of N availability which may become extreme with changes in pH (Fu et al. 1987), soil humidity and oxygen levels (Smith et al. 1998). Production of plant volatiles or ROS-scavenging biochemical chain, represented by the H2O2-producing and -degrading enzymes such as SOD and catalase enzymes, are modulated in pathogen-challenged plants or those inoculated with biocontrol agents (Abanda-Nkpwatt et al. 2006; Trotel-Aziz et al. 2008). An increase in SOD activity together with no significant changes in catalase activity leads to moderate hydrogen peroxide accumulation, which serves as both the signal and effector molecule in plant immunity (Gechev et al. 2006). A similar trend was recorded with the pathogen inoculation in our study. Endophyic bacteria are known to prime potato (Solanum tuberosum) plants by elicitation of antioxidant enzymes (Pavlo et al. 2011). PAL activity is also elicited in the presence of fungus which enables the plant to withstand various biotic and abiotic stresses (Forlani 2010). However, the activity of elicited defense/hydrolytic/antioxidant enzymes and disease scores were not significantly influenced by fertilizer doses alone, although varietal differences and pathogen inoculation brought about significant differences as a result of their interactions with one another. Disease scoring results revealed that with the increasing levels of N, the severity also gradually progressed; disease intensity of greater severity being diagnosed in susceptible cultivars, as compared to the resistant cultivars. Paired t test illustrated the significant role of N levels, with the addition of 120 and 180 N showing sharper effects over 0 N samples, in terms of disease scores, followed by those in 180 N, over 120 N. In the present study, a sharp increase in nitrogen dose from 0 to 120 exhibited a significant correlation with disease scores [Paired test values of 8.14; P (T ≤ t) 0 and 120 N − 3.39E−09], illustrative of this aspect. The tripartite interaction was more significant, viz., cultivars–inoculation–fertilization, followed by interaction of cultivars and fertilization, or cultivars and pathogen inoculation, or pathogen inoculation and fertilization; subsequently by individual factors such as cultivars, fertilization and pathogen inoculation. This highlights that the selection of cultivar with its characteristic microbiome maybe a major driver for reducing crop losses due to pathogen challenge.

Ikeda et al. (2014) suggested that low N fertilization management could change the rice root microbiomes relevant to methane cycles, methanogenesis and methanotrophs; however, phyllosphere populations were not investigated. In our study, qPCR analyses of the abundance of eubacteria (log copy number of 16S rRNA gene copies cm−2) in the phyllosphere samples revealed a gradual change in cv. CO-39I with N doses, with/without pathogen challenge. Although higher dose of 120 N application led to rapid increases in cv. PB-I and P-1609, any further increase in N (180 N) led to decreases in the abundance of eubacteria in cv. CO-39, PB-I, and P-1609. Although a slight increase was recorded as a result of pathogen inoculation in cv. PB-1 and P-1609, the combined treatments of higher N and pathogen inoculation brought about a decrease in the abundance of eubacteria in all the cultivars tested, despite minor differences in cv. PB-1 and P-1609. Sasaki et al. (2013) analyzed the profiles of microbial communities from the rhizosphere and phyllosphere, across 9 cultivars of rice, using automated ribosomal intergenic spacer analysis (ARISA) and different levels of N fertilization. Their results revealed that shoot bacterial communities were significantly affected by plant genotype, whereas root bacterial communities were largely affected by fertilization level. In the present investigation, the responses in terms of pigments or enzyme activity elicited were mainly influenced by the cultivar and nitrogen doses, and influenced by pathogen challenge; however, disease-related indices were found to be modulated by nitrogen doses, but its interactions with cultivars was also significant.

In terms of archaeal communities, application of higher N doses led to increased abundances only in cv. CO-39 and PB-1, while cv. CO-39I and P-1609 exhibited a drop in the values. Pathogen challenge decreased the abundances of archaea in all the cultivars under the 0 N dose; however, an increase was recorded in all the cultivars tested under the higher N dose treatments. Wu et al. (2009) observed that spatial or niche factor had the greatest influence on the archaeal community composition, but Breidenbach and Conrad (2014) observed only minor changes in the bacterial and archaeal communities, during various growth stages of rice crop or change from flooded to aerobic conditions, and emphasized the existence of a resilient native community. Several researchers (Bates et al. 2011; Ikeda et al. 2015; Kanchan et al. 2018) pointed that low N (in soil) or lower C:N ratios strongly influence the structure of microbial communities, particularly through selective abundance of Bacilli, besides being consistently correlated with archaeal relative abundances. In our investigation, a similar correlation with leaf N concentrations could be observed, and their presence in the leaves, despite in lower numbers in the pathogen challenged treatments, highlights their important role in ammonia oxidization, which in turn not only influences the local ammonia availability and pH, but has been speculated to play a role in defense against pathogens (Müller et al. 2015). Recent reports suggest their significance to the plant holobiome, particularly in terms of promoting plant growth though auxin biosynthesis, facilitating nutrient transformations and protecting against abiotic/biotic stress, besides aiding in the maintenance of a healthy microbiome (Taffner et al. 2018).

Cyanobacteria are a less encountered group in the phyllosphere, and are more abundant in the humid environs of rain forests. In our study, the abundance of cyanobacterial 16S rRNA gene copies was cultivar-specific, with higher values in cv. CO-39I and PB-I, relative to those of the blast susceptible cv. CO-39 and P-1609. Metaproteomic analyses of the rice rhizosphere and phyllosphere microbiome revealed that although dinitrogenase reductase proteins were only detected in the rhizosphere samples, nifH was identified in several phyllosphere bacteria (Knief et al. 2012). Contrasting reports on the drivers for cyanobacterial communities in terms of composition and richness in various forest trees are documented. Rigonato et al. (2012) investigated the cyanobacterial communities in the Brazilian mangrove forests and through clone library analysis, identified 19 genera of cyanobacteria, besides several uncultivated taxa; most of the sequences were affiliated with the orders Nostocales and Oscillatoriales. They concluded that spatial location, rather than plant species was the stronger regulator of the composition of cyanobacterial communities. However, Rigonato et al. (2016) observed the significance of plant species, when comparing across various species in an Atlantic forest. Furnkranz et al. (2008) observed that the leaf colonizing epiphytes contributing towards significant nitrogen inputs into the rainforest, mainly belonging to Cyanobacteria and Alphaproteobacteria, based on phylogenetic analyses, but expression studies were not undertaken. In the present investigation, only one host species, i.e., rice, but challenged by increased N doses led to a decrease in the abundance of cyanobacterial 16S rRNA gene copies in all the cultivars tested, and when challenged with the fungal pathogen, there was an increase by at least by one-fold in the control treatment. However, the application of higher doses of N led to decreased abundances of cyanobacteria in most of the pathogen-challenged treatments, irrespective of cultivars. As there is no published information on the effect on the cyanobacterial populations to nitrogen doses and/ pathogen challenge, in relation to phyllosphere communities, more in-depth studies are warranted on this aspect in future.

Most of the studies on nifH genes or quantification of their expression are restricted to mostly rhizosphere, or in the phyllosphere of forest species or marine vegetation. In our investigation, the analyses of number of copies of functional gene nifH showed decreases due to the application of higher N, with or without the fungal inoculation, relative to the respective control treatments. Nitrogen fixation has been reported as a major activity in the phyllosphere by several researchers (Ruinen 1971; Patti and Chandra 1981; Murty 1983; Abril et al. 2015). Quantification of nifH in the phyllosphere, particularly stems of rice, supports these observations (Elbeltagy et al. 2001). Studies on root and rhizosphere microbiomes have revealed that a complex and nested array of factors at varying spatial scales, including plant community, plant host, soil edaphics and microbial taxon and community characteristics modulate the structure and function of microbial communities (Schlatter et al. 2015); however, there are no similar studies in rice phyllosphere. Ferrando et al. (2012) compared the microbiome of the leaves of physiologically different rice cultivars, and observed that they were highly similar, being composed of a reduced group of strongly associated and persistent bacteria that were partially recovered by cultivation; however archaeal/cyanobacterial members were not identified.

Interestingly, the copies of functional gene bacterial amoA increased due to the application of higher N, with or without the pathogen challenge. An increase of almost one-fold was recorded in the cv. CO-39I and cv. P-1609 with higher N application under the pathogen inoculation treatment. Wang et al. (2009) observed that although the community composition of archaeal ammonia oxidizers in rice paddies was not influenced by soil depth or N fertilizers, the bacterial ammonia oxidizers were influenced distinctly. Ammonia oxidizing microorganisms (AOM) are known for their significant role in the nitrification of rice paddy soil, as the ammonia monooxygenase alpha subunit (amoA) gene is responsible for catalyzing the first step of ammonia oxidation, and their dynamics and populations have been investigated by several researchers, particularly in the rice rhizosphere. Ke et al. (2013) observed that more than season or nitrogen amendment, the compartmentalization or niche investigated played a crucial role in rice crop. This is also consistent and positively correlated with the archaeal abundance values, and illustrates the significance of archaeal ammonia oxidisers in defense against pathogens and maintaining a healthy phyllosphere microbiome.

Most of the efforts towards controlling the serious losses due to rice blast disease caused by M. oryzae focus on developing resistant lines which are durable, through the introduction of blast-resistant gene(s) through backcrossing or pyramiding genes. However, there is a need to include a dose of N as a factor for screening, and understand plant–microbiome interactions for obtaining better responses to management practices in the newly developed cultivars; thereby, a better understanding of the basis of varietal resistance to foliar pathogens. Our further investigations are directed towards developing effective biological control measures in a selected cultivar, as phyllosphere is often the zone where both the pathogen and biocontrol agents may coexist and interact with each other.