Endophytic Streptomyces hygroscopicus OsiSh-2-Mediated Balancing between Growth and Disease Resistance in Host Rice

ABSTRACT Plants fine-tune the growth-defense trade-off to survive when facing pathogens. Meanwhile, plant-associated microbes, such as the endophytes inside plant tissues, can benefit plant growth and stress resilience. However, the mechanisms for the beneficial microbes to increase stress resistance with little yield penalty in host plants remain poorly understood. In the present study, we report that endophytic Streptomyces hygroscopicus OsiSh-2 can form a sophisticated interaction with host rice, maintaining cellular homeostasis under pathogen-infection stress, and optimize plant growth and disease resistance in rice. Four-year field trials consistently showed that OsiSh-2 could boost host resistance to rice blast pathogen Magnaporthe oryzae while still maintaining a high yield. The integration of the proteomic, physiological, and transcriptional profiling analysis revealed that OsiSh-2 induced rice defense priming and controlled the expression of energy-consuming defense-related proteins, thus increasing the defense capability with the minimized costs of plant immunity. Meanwhile, OsiSh-2 improved the chloroplast development and optimally maintained the expression of proteins related to plant growth under pathogen stress, thus promoting the crop yield. Our results provided a representative example of an endophyte-mediated modulation of disease resistance and fitness in the host plant. The multilayer effects of OsiSh-2 implicate a promising future of using endophytic actinobacteria for disease control and crop yield promotion.

surrounded by infected seedlings to allow the airborne spores to serve as inoculum. As shown in Fig. 1, OsiSh-2 foliar spraying treatment with M. oryzae infection (E1M1) significantly reduced the disease indexes of the seedling blast ( Fig. 1A and B) and panicle blast ( Fig. 1D and E) for up to 45.3 and 61.0%, respectively, compared to the OsiSh-2 untreated (E-M1) rice. Correspondingly, the seedling height increased up to 26.5% after the seedling blast (Fig. 1C), and the yield loss rate reduced up to 57.0% after the panicle blast (Fig. 1F).
In the year 2018 and 2019, we further complemented a positive-control treatment, i.e., pesticide tricyclazole (Trz1M1), one of the commonly used fungicides for rice blast disease. Interestingly, although Trz1M1 treatment showed a better disease control efficiency than E1M1 treatment with an average 9.7% lower disease index between 2018 and 2019 (Fig. 1E), the yield loss rate of the OsiSh-2 treatment was much lower (average 15.8%) than that of the tricyclazole (Fig. 1F). These results suggest a direct stimulation of rice growth by OsiSh-2, not just an indirect effect of the disease resistance, thereby alleviating the reduction of growth.
To test whether OsiSh-2 conferred a desirable growth-promoting and yieldincreases for its host rice, the growth parameters of E1M-and E-M-rice without M. oryzae infection were determined in greenhouse and field conditions. As shown in Fig. 2, the growth parameters of E1M-rice were significantly higher than those of E-M-rice at 28 days posttreatment (dpt) ( Fig. 2A and B). The production of E1M-rice was also remarkably improved at the yellow ripening stage with the thousand-grain weight (TGW) significantly increased by 2.1% ( Fig. 2C and D). In the field experiment, the agronomic traits of E1M-rice were also significantly better than that of E-M-rice, with the TGW markedly increased by 2.2% ( Fig. 2E and F).
The inoculation of OsiSh-2 was also confirmed. In the greenhouse, the reisolation frequency of OsiSh-2 from inner tissues of E1M-rice seedling was 50.0%, while no OsiSh-2 was isolated from the E-M-rice (see Fig. S1A in the supplemental material). The relative OsiSh-2 biomass in rice, indicated by the ratio of ShRpoA gene (a DNAdirected RNA polymerase subunit alpha of OsiSh-2) versus OsUbq gene, was 74.9-fold higher than that in E-M-rice (see Fig. S1B). In the field samples, the reisolation frequencies of OsiSh-2 from E1M-and E-M-rice seedling were 53.8% and zero, respectively (see Fig. S1A).
For plants, the defense and growth pathways are intertwined and conflicted. To cope with the trade-off between growth and defense, plants developed symbiotic relationships with microbes. Some rhizobacteria were reported with the capacity of plant growth stimulation and rice blast resistance induction in greenhouse conditions. However, the experiments in the laboratory do not necessarily reflect the potential of microbes under variable field conditions. Among the very limited reports about the field performances, the recent examples including the rhizobacteria Bacillus spp. (34) The numbers of effective tillers (ET) and the TGW (g) were determined to represent agronomic traits. (E) Representative photograph of all panicles in one plant (three plants per treatment) and all grains in a panicle (left, filled grains; right, flat grains) for E-M-, E1M-, and dead (DE1M-) OsiSh-2 and cell-free culture filtrate (CF1M-) rice. (F) The grain yield of rice in different treatments is presented as the number of effective tillers, the panicle length (cm), and the thousand-grain weight (g). All of the rice panicles were sampled. Values are means 6 the SD (from 20 independent rice plants). Bars in panels B and C with asterisks are significantly different, as determined by the Tukey-Kramer test (*, P , 0.05; **, P , 0.01; ***, P , 0.001). Bars in panel F with different letters are significantly different (ANOVA, P , 0.05) according to Duncan's multiple-range test. The percentage changes followed by "1" or "-" above the bars were calculated by using the following formula: percent change = [(value of treated rice) -(value of untreated rice)]/(value of untreated rice) Â 100%. and Cladosporium cladosporioides C24G (35) showed the capability of significantly reducing rice blast disease severity with increased grain yield during 2-year field experiments. In this study, our consecutive greenhouse and field trials revealed that rice endophytic OsiSh-2 also has positive effects on rice growth as well as disease resistance. Endophytic actinobacteria might represent a novel element in coping with the trade-off between growth and defense in the host plants. We thus used OsiSh-2rice symbiont to study the involved mechanisms of this balance modification of the growth and defense.
Endophytic OsiSh-2 establishes a flexible interaction with host rice and the beneficial effects on rice is an active process. Although OsiSh-2 is a beneficial microbe of rice, it also might be recognized as an alien organism by rice and need confront the rice defense system to colonize (24). On the other hand, the plant immune system is tuned, and the activation of immune response is costly, which might lessen plant fitness (36). Thus, the level of immune activation should be carefully regulated between OsiSh-2 and rice.
In the greenhouse test without pathogen stress, the growth of E1M-rice was impaired compared to the E-M-rice at 7 dpt by OsiSh-2. However, E1M-and E-M-rice showed a comparable phenomenon at 14 dpt, and E1M-rice possessed a more flourishing growth than E-M-rice at 21 dpt (see Fig. S2). This result indicated that the initial defects in plant growth in OsiSh-2-infected plants are overcome and reversed in later growth stages.
It is also reasonable to assume that the benefits of OsiSh-2 in host rice are related to a dynamic regulated process and rely on its active form. To test this hypothesis, the effects of OsiSh-2 spores (E1M1), dead OsiSh-2 spores (DE1M1), and cell-free culture filtrate (CF1M1) on rice after M. oryzae infection were examined. E-M1 rice was used as a blank control. The detached rice leaves were collected and inoculated with blast by using a punch method. As shown in Fig. 3A to C, the lesions of E1M1, DE1M1, and CF1M1 rice were all significantly smaller than those of the E-M1 rice, and the relative fungal biomass, as indicated by the ratio of MoPot2 gene (an inverted repeat transposon of M. oryzae) to the OsUbq gene (a rice genomic ubiquitin gene) in treated rice, was correspondingly decreased. E1M1 rice, in particular, developed a 56.9% smaller lesion area at the wound inoculation sites, with a 48.5% reduction in fungal biomass ( Fig. 3B and C). When we replanted the rice plants in the field with blast disease occurrence, the growing status and decreases in yield loss rates for three OsiSh-2-treated rice were still better than for E-M1 rice ( Fig. 3D to F). Similarly, the yield of E1M1 rice was much better than other compared groups, with the lowest yield loss rate 49.2% (Fig. 3F) and the highest TGW of 22.7 g (Fig. 3G).
We also investigated the growth of the rice treated as described above without pathogen stress. Similarly, all of the growth parameters of E1M-rice were best in the greenhouse test (see Fig. S3). In the field condition, the best agronomic traits were found in E1 rice, while those of DE1M-and CF1M-rice did not surpass those of E-M-rice ( Fig. 2E and F). For example, the TGW of DE1M-rice was 1.8% lower than that of E1M-rice and only 0.4% higher than that of E-M-rice. The TGW of CF1M-rice was even 0.5% lower than that of E-M-rice.
In short, the active OsiSh-2 exhibited a greater advantage in the enhancement of both growth promotion and defense resistance than did its dead cells and metabolites. This might benefit from the symbiosis relationship between OsiSh-2 and host rice, by which the dynamic accommodation and adaption would happen in facing the various environmental changes. That is, OsiSh-2 might participate in the balance modification of the growth and defense in host rice.
Cytological, physiological, and genetic transcription characterization confirm the promotion of rice growth and defense in the presence of OsiSh-2. The growthpromoting characters of OsiSh-2 were then confirmed at the subcellular level by transmission electron microscope (TEM) observation of the rice leaves. As shown in Fig. 4A, E1 rice showed a better development of chloroplasts compared to E-rice in both M1 and M-situations, with the relative chloroplast sizes increased by 47.0 and 63.2% (Fig. 4B) and the numbers of starch grains per cell increased by 748.8 and 89.4%, respectively (Fig. 4C). The chloroplast is the organelle where photosynthesis happens and sunlight energy is converted and stored for plant growth and development. As expected, this morphological change was correlated with functional improvement: the growth parameters involved with chloroplasts, including the chlorophyll content and net photosynthetic rate of E1 rice, were all significantly enhanced compared to those in E-rice ( Fig. 4D to G).
It should be noted that M. oryzae infection led to obvious damage in the chloroplast structure of E-M1 rice leaves, evident as grana revealing an obscure boundary in the laminar system. However, the grana in E1M1 rice still maintained sharply defined areas (Fig. 4A), indicating a protective effect on chloroplast integrity by OsiSh-2. As we knew, in addition to the function of photoautotrophy, chloroplasts also play crucial roles in the activation of defense responses (37). The biosynthesis of defense-related molecules such as salicylic acid and jasmonic acid, as well as secondary messengers, including calcium and reactive oxygen species (ROS), all take place in the chloroplast (38). A recent study showed that a light-harvesting complex II protein LHCB5 in the The chlorophyll contents, including total chlorophyll and carotenoid in E-and E1 rice at 168 h after M-or M1 treatment, were measured by using spectrophotometry. (F and G) The photosynthesis efficiency analysis, including the net photosynthetic rate and intercellular carbon dioxide concentrate in rice samples collected as described for panel C, was determined by using portable photosynthesis system. Values are means 6 the SD (n = 20). Bars with different letters are significantly different (ANOVA, P , 0.05) according to Duncan's multiple-range test. The percent changes followed by "1" or "-" above the bars were calculated by using the following formula: percentage change = [(value of treated rice) -(value of untreated rice)]/(value of untreated rice) Â 100%.
Endophyte Modulates Host Rice Defense and Fitness ® chloroplast is subject to specially enhance the resistance to M. oryzae by the accumulation of ROS. Conversely, the successful colonization and invasion of M. oryzae in rice tissues are always accompanied by the destruction of chloroplasts in rice plants (39). Consistent with these reports, our data suggested that OsiSh-2 could not only promote host rice growth by effectively improving the rice chloroplast development but also help the host to enhance disease resistance by alleviating the damaging effects on chloroplast-caused pathogen infection.
In addition to cytological management, plants can develop induced resistance in response to various types of stress stimulation, including pathogen infection or specific beneficial microbes' colonization (40,41). Thus, the ROS signaling, which is involved in the immune response, was determined. Spraying OsiSh-2 on the rice leaves led to a quick accumulation of H 2 O 2 at 0.5 h posttreatment (hpt) in the veins of E1M1 rice leaves, whereas no H 2 O 2 accumulation was detected in E-M1 rice (see Fig. S4A and B). As a result, the downstream immune responses of H 2 O 2 messenger, such as callose deposition, which works as an effective barrier in the plant cell wall for blocking the pathogen invasion (42), was induced. By 12 hpt, the number of callose deposition around the stomata was noticeably increased in E1M1 rice leaves, whereas no callose deposition was observed in E-M1 rice (see Fig. S4C and D).
We next investigated the rice responsive processes during pathogen infection. M. oryzae infection led to significant ROS production in rice. Interestingly, this response in E1M1 rice was faster and stronger than that in E-M1 rice. A large amount of H 2 O 2 accumulated in the E1M1 rice leaves at 1 to 2 h postinfection (hpi) by M. oryzae, while the H 2 O 2 in E-M1 rice just began to accumulate at 24 to 36 hpi ( Fig. 5A; see also Fig. S5A). Consistently, the expression of OsRbohB, an NADPH oxidase gene involved in deliberating generation of ROS in plants during defense response (43), reached its first peak at 2 hpi in E1M1 rice, which was 10 h earlier than that in E-M1 rice (Fig. 5B). The callose deposition also happened at least 4 h before that in E-M1 rice, and the amount was also dramatically improved in E1M1 rice during 1 to 24 hpi ( Fig. 5C; see also Fig. S5B). These results, especially the callose data, presented a typical defense priming response pattern in OsiSh-2-rice symbiont, indicating that OsiSh-2 could raise an immune response in rice by priming (26).
Except inducing priming, we thought OsiSh-2 might also have contributed the first line of defense against the pathogen on host rice. In our previous study, both OsiSh-2 and its culture filtrate can strongly hinder the appressorial formation and inhibit the conidium germination of M. oryzae in vitro, two main steps for the foliar pathogen M. oryzae to infect the rice leaf (33). Correspondingly, the scanning electron microscope observation revealed that M. oryzae infection seriously destroyed the surface structures of E-M1 rice leaves. However, the inoculation of OsiSh-2 prior to the blast treatment led to a decreased distribution of M. oryzae conidia, and the damaging effects by pathogen infection were alleviated. We even could not observe the mycelia of M. oryzae by the first 12 hpi (Fig. 6). The genome annotations revealed that among the 74 secondary metabolites-synthesize gene clusters in actinobacteria OsiSh-2, 67 clusters were involved with the biosynthesis of antibiotic substance, including polyketide synthase and nonribosomal peptide synthetase. We have confirmed the capacity of OsiSh-2 on producing antimicrobial enzymes, plant hormones, antibiotics and siderophores (26)(27)(28). These compounds might directly or indirectly inhibit the pathogen, thus decreasing additional transcription and translation costs of the defense-related gene of plants. As a consequence, the saved energy could drive the growth of the host. To elucidate this direct protection of OsiSh-2 on rice, further study of the isolation and purification of the antagonistic products of OsiSh-2 is under way.
OsiSh-2 dynamically balanced the expression of host proteins involved in defense and growth activity. To gain a deeper understanding of the active OsiSh-2 in the modification of the growth and defense in host rice from the molecular level, the proteomic profiles of E-M1 and E1M1 rice at 48 hpi after M. oryzae infection (about 9 dpt by OsiSh-2 or water control) were identified by using mass spectrometry-based TMT labeling quantitative proteomics analysis. The mass spectrometry proteomics data are available via the ProteomeXchange Consortium with the data set identifier PXD026348. The workflow is shown in Fig. 7A. Altogether, 6,392 proteins were identified and 5,163 proteins were quantified (see Fig. S6A). The numbers of differently expressed proteins (DEPs) in each compared group are shown in Fig. 7B and C. The identities of DEPs and the magnitude of their changes are listed in Table S1. Volcano plots show the quality of the identified peptides and the change levels of the DEPs that had significantly altered in each group, respectively (see Fig. S6B and C).
Compared to the M. oryzae uninfected control, 1,022 DEPs were found in the M. oryzae-infected E-rice (E-M1/E-M-). However, when pretreated with OsiSh-2 (E1M1/ E1M-), the number of DEPs remarkably decreased to 512. These data indicated that compared to the OsiSh-2-pretreated rice, nontreated rice might induce a more extensive process in response to M. oryzae infection. Recently, it was considered that desensitizing the stress responses could be a more efficient strategy to achieve both stress resistance and growth (44). Such an idea matches our results. From the perspective of energy metabolism, this result indicated that E1M1 rice was in a more homeostatic state compared to E-M1 rice when facing the pathogen stress. Taking the functionannotated DEPs into consideration, some proteins likely regulating the defense response, energy metabolism, and nutrient metabolism were analyzed emphatically.
There are 566 upregulated DEPs in E-rice after M. oryzae infection (E-M1/E-M-) (Fig. 7B). KEGG analysis showed that these proteins are mainly involved with energy-supply processes, such as oxidoreductase (3.8-fold) in fatty acid degradation and pyruvate kinase Endophyte Modulates Host Rice Defense and Fitness ® (2.1-fold) in glycolysis/gluconeogenesis, and energy-consuming processes, such as 30S ribosomal protein S4 (1.8-fold) in ribosome, asparagine synthetase (7.1-fold) in amino acid metabolism, putative dynamin homolog (1.8-fold) in endocytosis, and stemod-13(17)-ene synthase (4.5-fold) in diterpenoid biosynthesis (Fig. 7D). This result indicated that the pathogen stress induced the defense responses, which might be correlated with an energy consumption. This was in line with the previous reports that the disease resistance is an energy-consuming process, leading to the growth-defense trade-off in plants (4,9,10,13). Correspondingly, we found that the 456 downregulated DEPs in E-M1 rice were mainly related to the energy fixation and nutrient absorption pathways, such as ferredoxin (0.4fold) and ATP synthase epsilon chain (0.3-fold) in photosynthesis, phosphoenolpyruvate carboxylase (0.5-fold) and ribulose bisphosphate carboxylase large chain (0.5-fold) in carbon metabolism, and glutamine synthetase (0.6-fold) in nitrogen metabolism ( Fig. 7B and D).
However, the colonization of OsiSh-2 in rice mitigated the overconsumption of energy caused by M. oryzae. Among induced proteins in M. oryzae-infected rice described above, most DEPs (n = 387) were not activated when OsiSh-2 was inoculated in E1M1 rice (Fig. 7C). These findings indicated again the role of OsiSh-2 in maintaining a relative homeostasis in the stress response. For instance, ribosome assembly is the most energy-demanding process linked to cell growth and defense response due to its requirement for coordinated production of processed ribosomal RNAs, proteins, and biogenesis factors (45). A total of 25 ribosomal proteins significantly accumulated in E-M1 rice (upregulated 1.3-to 1.8-fold), while they were not activated expression in E1M1 rice (see Table S2A). Endocytosis is also an energy-consuming process by which cells internalize substances from their external environment. There were also 17 endocytosis proteins (1.3-to 1.8-fold) only activated in E-M1 rice (see Table S2A). These results indicated that during the battle with M. oryzae, the presence of OsiSh-2 led to less activation of energy-consuming pathways than in E-M1 rice. The saved energy might be shifted to maintain normal rice growth. Correspondingly, we found that more than half of downregulated DEPs (n = 347) in E-M1 rice did not exhibit a reduced expression level in E1M1 rice (Fig. 7C). There were 15 proteins in the photosynthesis process, 31 proteins in the carbon metabolism pathway, and 6 proteins involved in nitrogen metabolism, which all decreased the expression by 0.5-to 0.7-fold in E-M1 rice (see Table S2B).
To be clearer, we then compared the expression levels of DEPs between E1M1 and E-M1 rice. The upregulated pathways in E1M1/E-M1 group were enriched in photosynthesis and nitrogen metabolism, e.g., the ATP synthase epsilon chain (2.6fold) and ferredoxin (2.5-fold) in photosynthesis and glutamine synthetase (1.4-fold) in nitrogen metabolism (Fig. 7E). On the other hand, the downregulated pathways in the E1M1/E-M1 group were mainly related to energy supply or energy-consuming processes, such as alcohol dehydrogenase 1 (0.7-fold) in glycolysis/gluconeogenesis, 40S ribosomal protein S30 (0.5-fold) in ribosome, putative dynamin homolog (0.6-fold) in endocytosis, and protein disulfide isomerase-like 1-5 (0.6-fold) in protein processing in the endoplasmic reticulum (Fig. 7E). Combined with the results presented above from field and greenhouse trials, these results indicated that OsiSh-2 pretreatment might positively modulate the expression profile of host rice to optimize the allocation of resources between growth and defense.
Interestingly, we found that the expression levels of proteins involved in the glutathione and ascorbate/aldarate metabolism were significantly higher in E1M1 rice than in E-M1 rice. These metabolic processes play crucial roles in maintaining cellular redox homeostasis and detoxifying in living organisms (46). In addition, three peroxidases involved in oxidative stress response showed 1.3-to 2.0-fold overproduction only in E1M1 rice. Peroxidases participate in the release or consumption of ROS, which can repair cell damage caused by ROS and catalyze the formation of lignin in plants (46). In addition, a recent study showed that the modulation of apoplastic ROS homeostasis by a growth-related transcription factor (Homolog of BEE2 Interacting with IBH 1 [HBI1]) can regulate the trade-off between growth and immunity in Arabidopsis thaliana (9). Thus, these results indicated that OsiSh-2 might also be conducive to maintaining the ROS homeostasis in rice during pathogen infection, consequently contributing to the balance between growth and defense in the host.
To verify the proteomic data, the expression levels of the genes encoding 12 proteins involved in the energy-supply or energy-consuming pathways and 6 proteins related to the photosynthesis or nitrogen/carbon metabolism were detected at 48 hpi. In agreement with the proteomic analysis, the evoked levels of eight genes, such as ADH2 and PGAM-i in glycolysis/gluconeogenesis metabolism, CNX in the protein processing in endoplasmic reticulum process, FDH in the biosynthesis of secondary metabolite, and RABA2a in the endocytosis, were much lower in E1M1 rice than in E-M1 rice, when respectively compared to untreated E-M-rice. The transcript abundances of atpE and LFNR2 related to the photosynthesis, GLN2 and CAs in the nitrogen metabolism, and ALDOA in the carbon metabolism showed significantly higher levels in E1M1 rice than in E-M1 rice (see Fig. S7). Some genes, such as Rps4-30 and RPS17 in the ribosome, were not detected to have the same change tendency as their encoded proteins, likely reflecting their fast turnover inside the cell.
The protein expression profiles confirmed that pathogen infection led to a balance between growth and defense, but OsiSh-2 mitigates this conflict by attenuating energy-consuming defense-related responses and shifting toward growth activities. As shown in Fig. 8, a schematic regulation model of OsiSh-2 was constructed based on the proteomic data.
Conclusions. We demonstrated the contribution of endophytic actinobacteria OsiSh-2 to promoting both yield and disease resistance in host rice by sustaining a balance between growth and immune defense. We confirmed that the enhanced protection effect of the OsiSh-2 is based on a priming mechanism instead of consuming energy to express many host defense-related genes and proteins. The saved energy is beneficial for plant growth, such as improving photosynthesis activity. A schematic model of OsiSh-2-rice symbiont in responding to blast pathogen M. oryzae was constructed (Fig. 9). The ability of OsiSh-2 to achieve both disease resistance and high productivity is likely to increase with a better understanding of the relationship between disease-response and fitness pathways. This sophisticated regulation through a mutual reaction between OsiSh-2 and rice supports the application of endophytic actinobacterial strains as plant inoculants or plant-strengthening agents in agriculture and forestry.

MATERIALS AND METHODS
Plant materials and growth conditions. Rice (Oryza sativa) was cultured in the field at the Yahua Seeds Science Research Institute, Longping High-Tech in Changsha, Hunan, China (YSSRI; 28°1194999N, 112°5894299E), to assess disease resistance and for yield evaluation. Plants in growth chambers were grown for M. oryzae inoculations and immune and comparative proteomic assays. The growth chambers were set as follows: 16/8-h light/dark cycles at 26°C and 80% humidity. The rice cultivar used in the study is Xiang'aizao 7, a cultivar susceptible to M. oryzae.
Endophyte and pathogenic fungus materials and growth conditions. Endophytic Streptomyces hygroscopicus subsp. hygroscopicus OsiSh-2 (GenBank accession number GCA_001705785.1, China General Microbiological Collection Center accession number CGMCC-8716) was isolated from rice sheath and routinely cultured in International Streptomyces Program 2 solid medium at 30°C, as described previously (33). The rice blast pathogenic fungus M. oryzae 70-15, which is compatible with Nip, was routinely cultured on solid complete medium at 28°C, as described previously (47). Spore suspensions of OsiSh-2 and 70-15 were prepared as described previously (48).
Greenhouse tests. Surface-sterilized rice seeds were sown into pots containing autoclaved soil and grown in a growth chamber. A spore suspension of OsiSh-2 (10 8 spores ml 21 ) was sprayed on rice leaves at the two-leaf stage. Water containing 0.2% (vol/vol) Tween 20 was used as a control. A conidial suspension of M. oryzae (10 5 conidia ml 21 ) was then sprayed a few days later when the rice plants were at the three-leaf stage. Infected rice plants were placed in darkness overnight at 26°C and then transferred to the growth chamber. The disease index and plant height were determined at 7 days postinfection by M. oryzae, as described previously (48). Independent experiments were repeated twice with three replicates and 50 plants per replicate.
Field tests. Surface-sterilized seeds were sown in the field in Liuyang, Hunan, China (YSSRI, 28°1 090299N, 113°3891699E). The OsiSh-2 treatment was the same as that done in the greenhouse assays, and spraying commercial fungicide tricyclazole (1.5 mg/ml) containing 0.2% (vol/vol) Tween 20 was added as a positive control (49). The rice seedlings were surrounded by infected seedlings to mimic a natural infection. The disease index and plant height for 1-month-old rice plants after the seedling blasts were determined, as described previously (48). In detail, the disease severity of each seedling blast was evaluated according to the standard evaluation system for rice (International Rice Research Institute, 2013). The disease level was scored on a scale from 0 to 9 as follows: 0, no lesions observed; 1, small brown specks of pin-point size without a sporulating center; 2, larger brown specks but ,1 mm in diameter; 3, small (roundish to slightly elongated) necrotic gray spots 1 to 2 mm; 4, typical susceptible blast lesions (spindle-shaped) 3 mm or longer, infecting ,4.0% of the leaf area; 5, typical blast lesions infect-  Endophyte Modulates Host Rice Defense and Fitness ® experiment was repeated twice with three replicates per treatment and 100 plants/replicate. The disease index, yield loss rate, and agronomic traits, including the effective tiller number, panicle length, and thousand-grain weight (TGW) for 4-month-old rice after the panicle blast were detected. The panicle with at least five filled grains is defined as effective panicle. The disease index of the panicle blast = R(number of panicles at all levels Â assessed disease level value)/(total number of sampled panicles Â highest disease level value) Â 100%. The yield loss rate = R [(number of empty grains per panicle/total grains per panicle)]/total number of collected panicles Â 100%. Each treatment was carried out in two different fields, and three replicates per treatment in each field using a randomized complete block design. Each replicate contained 100 rice plants for seedling blasts and 20 for panicle blasts.
The field experiment without pathogen stress was conducted in paddy fields located at YSSRI (28°3  1670N, 112°67820E). Approximately 20-day-old rice seedlings were transplanted to the paddy fields, with three repeats for each treatment. The planting density was set as 35 hills/m 2 (one seedling per hill, 16.5-cm hill spacing, 19.8-cm inter-row spacing), as described previously (50). Data were collected from the 20 plants from the middle two rows. Several agronomic traits, including the number of effective tillers, the panicle length, and the TGW, were measured.
Punch inoculation analysis. The basal resistance of rice plants against M. oryzae was evaluated by the punch inoculation method, as described previously, with some modifications (51). In brief, detached leaves of rice at the three-leaf stage under the greenhouse conditions were lightly wounded with a mouse ear punch. Then, 10 ml of spore suspension of M. oryzae was added to the wound. The inoculated rice leaves were placed in plates with water for restoring humidity, kept in darkness overnight at 26°C, and then transferred to the growth chamber for 5 days. The lengths of the lesions were then measured.
Fungal and endophytic biomass analysis. The fungal and endophytic biomass in rice plants was quantified as described previously with a slight modification (51). In brief, 0.05-g rice tissues were collected for DNA extraction using a Dzup (plant) genomic DNA isolation reagent kit (Sangon Biotech, Shanghai, China) according to the manufacturer's specifications. DNA-based quantitative real-time PCR (qRT-PCR) was performed using a CFX96 real-time system instrument (Bio-Rad, USA). The reactions were conducted using SYBR Premix Ex Taq TM II (Tli RnaseH Plus; TaKaRa, Dalian, China) according to the manufacturer's specifications. The relative fungal growth was calculated using the threshold cycle value (C T ) of MoPot2 DNA (an inverted repeat transposon of M. oryzae) against the C T of OsUbq DNA (a rice genomic ubiquitin gene), while the relative OsiSh-2 growth was calculated using the C T of ShRpoA DNA of OsiSh-2 against the C T of OsUbq DNA of rice as a ratio (ShRpoA/OsUbq), represented by the equation 2 CT(OsUbq) -CT(ShRpoA) , as described in a previous report (51). The primers for DNA-based qRT-PCR are listed in Table S3.
Endophyte reisolation analysis. For OsiSh-2 reisolation from inner host plants, the rice tissues were excised and subjected to four-step surface sterilization as described previously (52). In brief, the rice tissues were (i) washed in 4% NaClO for 6 min, (ii) washed in absolute ethanol for 30 s, (iii) rinsed in deionized water for 30 s, (iv) washed in 5% Na 2 S 2 O 3 Á5H 2 O for 5 min, and (v) finally rinsed in deionized water. After drying under laminar airflow for1 to 3 h, the rice tissues were distributed onto the Mannitol soya flour (MS) medium, a commonly used medium for actinomycetes isolation, and incubated at 27°C for up to 14 days. Benomyl and nalidixic acid were added to the isolation media at 50 mg liter 21 to control endophytic fungal and bacterial growth, respectively. The efficiency of the surface sterilization procedure was assessed by adding 200 ml of the final rinse water to the MS medium. Surface sterilization method was considered successful when no microbes grew on the MS plates.
Microscopy imaging analysis. For the scanning electron microscopy observation, the leaf samples were fixed and observed as previously described (48). For the transmission electron microscope observation, leaf pieces of approximately 1 to 2 mm 2 were cut from the middle of leaves and then immediately soaked in a fixation buffer (2.5% glutaraldehyde in 100 mM phosphate buffer [pH 7.4]). Polymerization and staining of the leaf samples were performed as described previously (38). The machine (HT-7700; Hitachi, Japan) was operated at an accelerating voltage of 100 kV. The percentage of relative chloroplast size was determined by counting the total chloroplast intensity in a single cell versus the total intensity of the corresponding cell from 10 to 15 cells of rice using Photoshop CS6 (Adobe Systems) and ImageJ v1.49 (National Institutes of Health), as described previously (48).
Analysis of the effects of OsiSh-2 on rice growth and yield. Under the growth chamber condition, surface-sterilized rice seeds were germinated, and one part was transferred to plates with International Rice Research Institute nutrient solution for morphologic observation and for evaluation of growth parameters, including the height of shoots, the dry weights of whole plants, and the net photosynthetic rate, as described previously (50). Another part was sown into pots containing autoclaved soil for yield assessment. The panicles of 4-month-old rice were collected to determine the panicle length and TGW. The morphologies of rice seedling roots were determined by computerized scanning (STD 1600; Regent Instruments, Quebec, Canada) and analyzed using WinRHIZO software (Regent Instruments). The independent experiment was repeated twice with three replicates (50 plants per replicate for seedlings and 3 plants per replicate for panicle collection).
Immune response analysis. The rice leaves were stained by diaminobenzidine (DAB; Biotopped) and aniline blue (Macklin), respectively, for the H 2 O 2 accumulation and callose deposition assay. Average H 2 O 2 and callose measurements were performed as described previously (48). Three replicates were performed for each treatment.
RNA extraction and expression analysis. Rice leaves (0.05 g) were collected for total RNA extraction by using a Plant Total RNA isolation kit (Sangon Biotech, China). The resulting RNA was reverse transcribed using a PrimeScript RT reagent kit with gDNA Eraser (TaKaRa, Japan). The genes expressions were analyzed by qRT-PCR, as described previously (48). The primers used for qRT-PCR are listed in Table S3. The transcript data were normalized using b-actin mRNA expression levels as the internal reference. Independent experiments were repeated twice, and all reactions were performed in triplicate.
Protein extraction and identification. The leaves of rice seedlings in each treatment (0.1 g) were harvested. Total protein extraction and trypsin digestion were performed according to a previous report (53). After trypsin digestion, the peptides were desalted using a Strata X C 18 SPE column (Phenomenex) and vacuum dried. To identify endogenous rice proteins, peptides were reconstituted in 0.5 M TEAB (triethylammonium bicarbonate buffer) and processed according to the manufacturer's protocol for the tandem mass tag kit. High-pressure liquid chromatography fractionation and liquid chromatographytandem mass spectrometry (LC-MS/MS) analysis of the tryptic peptides were performed as described previously (54). The resulting MS/MS data were processed using the MaxQuant search engine (v.1.5.2.8). MS/MS spectra were searched against the Uniprot_OSjaponica 39947 database (63,195 sequences) concatenated with the Universal Protein Resource (UniProt) database for rice were performed as described previously (55). Independent experiments were repeated twice.
Data analysis and bioinformatic annotation. In the MS quantitative analysis, the intensities of multiple unique peptides of each protein were detected and are presented as the quantitative value of the corresponding proteins. The intensities of multiple peptides of each protein in the two compared samples were taken to log 2 (so that the data conform to the normal distribution), and then we used a twosample two-tailed t test method to calculates the P value. When the P value was ,0.05, a change in differential expression of .1.3 is regarded as the change threshold of significant upregulation, and ,1/1.3 is regarded as the change threshold of significant downregulation. The functions of the proteins were annotated based on gene ontology terms from the UniProt-GOA database. The pathway enrichment of the proteins was analyzed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) database.
Statistical analysis. Statistical parameters are reported in the figures and figure legends. Normally, statistical analysis of the data was performed by one-way repeated-measures analysis of variance (ANOVA) using SPSS software (Chicago, IL), followed by Tukey's post hoc test or Duncan's multiple-range tests, which were used to compare the means for the multiple groups. A P value of ,0.05 was considered statistically significant.
Data availability. The data that support the findings of this study are available from the corresponding author upon reasonable request.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.