Silicon rates and beneficial microorganism on blast suppression and productivity of upland rice

Rice is a crop grown throughout the world, occupying, in the year 2016 an area around 160 million hectares, with an annual production of around 741 million Mg of paddy grains. China is the world's largest rice producer, contributing about 28.3% of world production [1]. Globally, Brazil ranks ninth in the production of the rice grain, with an estimated production for 2017/2018 of 12 million Mg in an area of almost 2 million ha [2]. Rice is a crop that is easily adaptable to different soil and climate conditions, and because of this, it is the species with the most signi icant potential to increase production and to ight against global hunger [3-5].


Introduction
Rice is a crop grown throughout the world, occupying, in the year 2016 an area around 160 million hectares, with an annual production of around 741 million Mg of paddy grains. China is the world's largest rice producer, contributing about 28.3% of world production [1]. Globally, Brazil ranks ninth in the production of the rice grain, with an estimated production for 2017/2018 of 12 million Mg in an area of almost 2 million ha [2]. Rice is a crop that is easily adaptable to different soil and climate conditions, and because of this, it is the species with the most signi icant potential to increase production and to ight against global hunger [3][4][5].
In Brazil, rice can be grown in two types of ecosystems: the lowland and upland condition with supplemental irrigation or rainwater only [6]. In the upland ecosystem, rice is usually planted in rotation with soybeans, in the renewal of degraded pastures, or as the irst crop in opening up new areas for agriculture [7]. The production of upland rice is growing in global importance due to the reduction of irrigation water availability for irrigated lowland rice crops [8].
Among the factors that can provide a reduction in upland rice grain yield, there are diseases. Rice blast disease caused by Magnaporthe oryzae B. Couch [Pyricularia grisea (Cooke) Sacc.] is a signi icant constraint on worldwide rice production [9]. This disease is very aggressive and in the tropical region can cause yield losses of up to 100% [10]. In this sense, to avoid yield loss by blast farmers apply a high amount of fungicides [11][12][13]. The use of fungicides has ef iciency in blast control; however, its use as the only strategy to control this disease can cause environmental contamination, increase pathogen resistance to different active principles, besides increasing the costs of rice production [14]. Therefore, there is a need to search for alternatives to control rice blast in a sustainable way that avoids environmental polluting, reduce the production cost, increase grain yield and allow improving pro its for the farmers. The use of silicon and bene icial microorganisms, such as plant growth-promoting rhizobacteria (PGPR) and Trichoderma spp. fungus, has been highlighted in researches as sustainable alternatives, because they present high ef iciency in controlling diseases, improving plant development and increasing rice grain yield [8,[14][15][16][17][18][19][20][21].
Silicon is considered the second most abundant element in the Earth's crust and is absorbed by the roots of the plants as monossilicic acid (H 4 SiO 4 ), accumulating in different vegetal tissues, mainly of the wall of foliar cells [22]. The use of silicon in agriculture offers several bene its to the plants, such as resistance to diseases and pests, higher resistance to drought, salinity, and contributes to the improvement of nutritional status [23]. In resistance to diseases, silicon can act in a passive way, when the accumulation of silicon in the cell wall acts as a physical barrier to the penetration of pathogens and, actively, when the element is able to activate genes involved in the production of antimicrobial compounds [16][17][18][19][20][21][22][23][24]25].
Plant growth-promoting rhizobacteria and Trichoderma fungus also bring several bene its to plants, such as disease resistance, growth promotion, atmospheric nitrogen ixation, solubilization, and nutrient availability, siderophores production and plant growth regulating hormones [20,26,27] ( Figure 1). As a biocontrol, these bioagents act by different mechanisms of action, such as antibiosis, parasitism, production of antimicrobial compounds and also induction of resistance [13,21,28,29].
However, there is still a lack of information on the use of microorganisms associated with silicon in agriculture and speci ically in rice crops. There are a large number of species that inhabit the rhizosphere of plants, and therefore a large number of interactions could occur between microorganisms, silicon, and plants. For example, it is estimated that there are approximately 30,000 different species of bacteria and only 8% of them have been identi ied [30]. Thus, there are several interactions among silicon, plant, and microorganisms that are not yet known and, if investigated, can bring bene its to agricultural systems [26]. In this sense, the objective of this work was to study the effect of rates of silicon with bioagents in blast suppression and grain yield of upland rice.

Site description
The experiments were located in the Capivara and Palmital Farms. The Capivara Farm is located in Santo Antônio de Goiás, GO, Brazil, at 16º28'00" S and 49 º17'00" W, and 823 m of altitude. The Palmital Farm is located in the municipality of Goianira, GO, Brazil, at 16°26'14" S, 49°23'50" W, and 720 m of altitude. The climate is tropical savanna, considered Aw according to the Köppen classi ication [31]). There are two distinct seasons, usually dry from May to September (fall/ winter) and rainy from October to April (spring/summer). During the experiment, the temperature and amount of rainfall data were recorded ( Figure 2).
In both places soil was managed under no-tillage system for ive years with millet as last cultivated crop. The soil in both locations was classi ied as a sandy clay loam (kaolinitic, thermic Typic Haplorthox) and was corrected by liming. Before the application of treatments, the chemical characteristics of the soil were determined according to the methods described by Donagema, et al.

Experimental design and treatments
Trials were conducted in rainfed conditions using  the genotype of rice BRS Primavera, which is considered susceptible to blast [32]. Trials were arranged in a randomized block, arranged in a split-plot scheme, with four replications. Treatments consisted (main plots) of four silicon rates [ Rhizobacteria P. luorescens (R-46) and B. pyrrocinia (R-55) belong to the Microorganism Culture Collection at the Embrapa Rice and Bean Research Center [13]. The pool of T. asperellum isolates belongs to the Fungi Collection of the Plant Protection Laboratory of the Federal Rural University of Amazônia [33].
Calcium and magnesium silicate (SiCaMg) was used as the silicon source, containing 10.5% silicon, 27% calcium, 6% magnesium and 95% of useful calcium carbonate equivalence. Levels of calcium and magnesium in the soil were balanced, within each silicon rate tested, with the use of dolomitic limestone. Application of silicon and limming was made 30 days before rice sowing.

Application of bioagents
Three types of applications of bioagents in rice plants were carried out: seed (microbiolization), soil and foliar spraying. For the applications via seed and soil, the B. pyrrocinia bioagents and the T. asperellum pool were used, and for the foliar application, P. luorescens was used, according to Fillipi, et al., França, et al. [13,27].
B. pyrrocinia suspension was prepared with water, from cultures that had been growing for 24 hours on solid medium 523 [34], at 28 °C, and the concentration was set in a spectrophotometer to A540 = 0.5 (10 8 UFC). Rice seeds were immersed in the bacterial suspension, during 24 hours under constant agitation at 25 °C, and control seeds (bacteria-free) were immersed only in water, during 24 hours, in the same conditions according to Sperandio, et al. [21].
Each isolate of T. asperellum pool was grown in a Petri dish containing PDA (potato dextrose and agar) for 5 days and formulated as described by Silva, et al. [33]. The seed treatment was performed at concentrations of 10 g of powdered [33] T. asperellum per 1 kg of seed [13,33]. The concentration of the biological suspension was 10 8 conidia ml -1 . Soil application of B. pyrrocinia (10 8 UFC) or T. asperellum pool (10 8 conidia ml -1 ) was made at 15 days after sowing, with suspension prepared as described above, in the same concentration, at 200 l ha -1 , with a costal atomizer, according to Filippi, et al. [13].
P. luorescens solution for plant spray was done as the same as B. pyrrocinia. Plant spray pulverization of P. luorescens suspension was sprayed on the plants, using a manual backpack sprayer, at a constant pressure provided by a CO 2 pressure source, and a conical nozzle type (TX-VS2), at 15 days after sowing with a volume of 200 L ha -1 , according to Filippi, et al. [13].

Rice crop management
Rice was cultivated in the no-tillage system. So, millet was desiccated with a glyphosate application (1.8 kg ha -1 acid equivalent) 40 days before sowing of the upland rice. Fertilization at sowing furrows was calculated according to the soil's chemical characteristics [35]. Therefore, sowing fertilization was 20 kg ha -1 of N as urea, 120 kg ha -1 of P 2 O 5 as triple superphosphate and 60 kg ha -1 of K 2 O as potassium chloride, in both places. Nitrogen topdressing fertilization (as urea) was carried out 45 days after the rice emergence, using 60 kg ha -1 of N. Rice sowing was performed mechanically using 200 seeds m -2 , on December 15 th , 2015 (Capivara farm) and on January 14 th , 2016 (Palmital farm). Rice plant emergence occurred on average ive days after sowing for both places. Cultural practices were performed according to standard rice crop recommendations, keeping crop free from weeds, diseases, and insects.

Evaluations
The severity of leaf blast and panicles: For the evaluation of leaf blast, it was sampled ive times in each plot in periods of 48 hours. It was sampled ive rice tillers per subplots in each time, to estimate the area under the disease progress (AUDPC). For the evaluation of the panicle blast, 50 panicles were collected per subplot, at 80 days after sowing, in the maturing of the grains stage. In both evaluations, the samples were collected and taken to the Laboratory of Technical Support of Embrapa Rice and Beans for the quanti ication of the blast disease. The quanti ication of leaf blast severity was performed using a scanner, where rice plants were scanned in high-resolution image (200 dpi) and QUANT ® disease quanti ication program, obtaining the percentage of diseased and healthy area. For the severity of the panicle blast, a ivenote scale (1, 3, 5, 7 and 9) was used, taking into account the percentage of empty spikelets in the panicle, and the incidence of neck blast (panicle base) [36].
Silicon content in rice leaves: Rice leaf samples were collected from the lag leaf at the full blooming stage. Leaves from 50 plants per plot were collected, washed and then dried under forced-air circulation at 65 °C for 72 h before grinding and analyzing the samples for chemical composition. Besides, in each plot, one sample of 100 g of grains was taken for nutrient analyses. The concentration of Si was determined using methods described by Korndorfer, et al. [37]. The shoots were washed in water and then dried in an oven with forced air circulation at 65 ºC, then weighed to determine the dry shoot mass.
Grain yield: Rice harvest was carried out by hand after physiological maturity (March 3 rd , 2016 Capivara Farm and April 29 th , 2016, Palmital Farm) of the grains in the usable area of each subplot. Grain yield was determined by weighing the harvested grain of each subplot, corrected to 13% of the water content and converted to kg ha -1 .

Statistical analysis
For statistical analysis, the Statistical Package for the Social Sciences (SPSS), version 21.0. In the qualitative factor (with or without bioagents), data were subjected to an analysis of variance, and when the F test proved signi icant, the data were compared by a Tukey test at p < 0.10. In the quantitative factor (Si rates), results were submitted to regression analysis when p < 0.10. Plots, subplots, and all subplots interactions were considered as random effects. The Si rates, use of bioagents and places were considered as ixed effects.

Leaf blast severity
There was a triple interaction among silicon rates, bioagents and places for leaf blast severity (Table 1). In general, the severity of blast disease was higher in Palmital Farm than in Capivara Farm (Figure 3). At Capivara Farm, in all treatments, datas were explained by quadratic regressions, with tendencies for less rice blast severity when increasing silicon rates. However, a mixture of bioagents provided the lowest leaf blast severity. At Palmital Farm, increasing silicon rates did not provide a reduction in the rice blast severity at the control treatment (no bioagents). On the other hand, for all bioagents treatments, datas ited in quadratic regressions, with reducing values of rice blast severity when increasing silicon rates.

Area under disease progres curve (AUDPC)
Concerning AUDPC, there was an interaction between silicon rates and places, and a single effect of bioagents (Table 1). It was observed that, increasing silicon rates, reduced AUDPC, in both farms. AUPDC values were higher in Palmital Farm than in Capivara Farm (Figure 4). Regarding bioagents, all treatments with them allowed lower values of AUDPC and differed from the control (no bioagents) ( Table 2).

Severity of neck blast
There was an interaction between silicon rates and places for the severity of neck blast ( Table 1). The severity of neck blast was higher in Capivara Farm than in Palmital Farm ( Figure 5). In the Capivara and Palmital Farm, the treatment composed of plants fertilized with 2 ton ha -1 of silicon presented a lower neck blast severity compared to the control.

Silicon content in rice leaves
There was a triple interaction among silicon rates, bioagents and places for silicon content in rice leaves (Table 1). In both farms, there were tendencies to silicon content in rice leaves by increasing silicon rates (Figure 6). At Capivara Farm, the silicon contents in rice leaves in treatment composed of 8 ton ha -1 of silicon and the bioagent P. luore scence was highlighted concerning the control treatment (without bioagents). At Palmital Farm, the silicon contents in rice leaves in the treatment composed of 8 ton ha -1 of silicon with the bioagents mixture was signi icantly higher than the control treatment.

Dry biomass
There was an interaction between silicon rates and places for the dry mass of upland rice plants (Table 1). At Capivara Farm, dates itted in a quadratic regression, dry mass increased Table 1: F values of isolated and interaction factors for severity of foliar blast (SFB), area below disease progression curve (AUDPC), severity of neck blast (SNB), biomass (Biomass), Foliar silicon content (Si Foliar) and grain yield (GY).  up until 6 ton ha -1 and then, reduced the values (Figure 7). At Palmital Farm, dates also itted in a quadratic regression, however increasing silicon rates, increased dry biomass.

Grain yield
There was an interaction between silicon rates and places, and a single effect of bioagents for grain yield (Table 1). It was observed that grain yield improved by increasing silicon rates up to 6 ton ha -1 in Capivara Farm (Figure 8). On the other hand, in Palmital Farm, grain yield increased as increased silicon rates. Treatment with a mixture of bioagents allowed the highest upland rice grain yield and differed from all other treatments (Table 3). Control treatment had the lowest grain yield and differed from all other treatments.

Discussion
The two places in which the experiments were conducted differed, concerning the severity of the blast disease in the rice plants. Capivara Farm (576 mm of rain and 23.8 ºC average temperature during the trial, Figure 2) was sowed on December 15 th , 2015, and presented minor blast disease severity in relation to Palmital Farm (576.6 mm of rain and 23.3 ºC average temperature during the trial, Figure 2), which was sowed on January 14, 2016. According to Prabhu, et al. [22] and Robert and Walker, [38] in areas with similar environmental and agricultural conditions, blast pressure may vary when cultivated at different times, with a higher incidence of blast diseases in later sowing.
The combination of 2 tons ha -1 of silicon with a mix of bioagents provided less than 4% of the severity of leaf blast in the trial conducted at Capivara Farm. In the Palmital Farm, the combination of 2 ton ha -1 of silicon with B. pyrrocinia, P. luorecens, the pool of T. asperellum and the mix of all bioagents provided less than 5% of leaf blast severity. Therefore, we could see that the combination of silicon and bioagents was suf icient to control leaf blast. The use of bioagents and the use of silicon in rice plants had demonstrated ef iciency in the control of blast and other diseases [39][40][41][42][43]. Corroborating with this information, Cortes, et al. (2015) observed rice plants treated with the combination of 2 ton ha -1 of silicon with B. pyrrocinia, or with and T. asperellum, suppressed 92% and 93% of the blast disease, when compared the control treatment, under greenhouse conditions. Bueno, et al. [44], also showed, under greenhouse conditions, that 2 ton ha -1 of silicon combined with B. pyrrocinia suppressed 55% of leaf scald in rice plants.
Many studies have shown that silicon prevents and/ or delays the penetration of the pathogen through the formation of a mechanical barrier, composed of amorphous silica, depoused between the cuticle layer and the epidermis layer, in foliar cells, and also acts by systematically inducing the production of defense enzymes and proteins [45][46][47][48].
Other studies showed that bioagents act effectively on plant resistance to diseases, being able to control pathogens by antibiosis, parasitism, by competition of nutrients, producing enzymes that degrade pathogen structures and together activating latent defense mechanisms, inducing resistance in plants [28,29,49].
In our trial, another factor could potentiate the absorption and the effect of silicon, on the plants, by the bioagents to control leaf blast. Some sources of silicon, such as silicates, have in their composition much of the silicon in the polymerized form. According to Liu, et al. [50], PGPR's applied to soil act in the rhizosphere by degrading particles of the polymerized mineral through the excretion of polysaccharides, making the silicon available to the plants and potentializing their absorption. According to Curry and Perry [51], this absorption and silicon accumulation in the leaves can be regulated by an active process, which is stimulated and activated in the plant by protection against different stress conditions, among them the attack of pathogens. For example, Bueno, et al. [44] showed that rice plants fertilized with silicon, treated with B. pyrrocinia-treated, and inoculated with Monographella albescens, increased foliar silicon concentration and disease resistance, when compared to the control treatment (without silicon and bioagent) by 3 times. These af irmations corroborate to our indings on the content of foliar silicon, as applying silicon in the soil increases the foliar silicon content. Besides, the area below the disease progression curve was reduced by the use of silicon and with all the bioagents tested.
Concerning the effect of silicon rates at Capivara Farm, the dose of 4 ton. ha-¹ showed an increase of 27% of grain yield, when compared to the control (without silicon), representing a difference of 27.4 bags (50 Kg) per hectare. In the same area, the 17.4% increase in plant biomass of the same treatment was also observed concerning the control. At Palmital Farm, the treatment that stood out was the 8 ton ha -1 of silicon, which provided an increase in 17.5% in grain yield when compared to the control, representing a difference of 14 bags per hectare. In this area, silicon fertilization also in luenced the increase in biomass. According to Polleto, et al. (2011), the increased productivity provided by the use of silicon can be explained by improved availability of nutrients to plants, increased biomass, growth promotion and resistance to biotic stresses, mainly to diseases. Corroboration with this information, Ramos, et al. [53] used carbonized rice hull as a source of silicon in upland rice and observed an increase of 16% in grain yield, concerning the silicon-free treatment. Santos, et al. [51], similarly to the present study, recorded a increase of 50% in rice yield using 4 ton ha -1 of silicon. Sousa, et al. [55], in a greenhouse, reported a 37% of gain in the total growth of rice plants fertilized with 2 ton ha -1 of silicon, when compared to the silicon-free plants. According to Carvalho-Pupatto, et al. [56], plants fertilized with silicon increase the root system from 30 to 50 cm, allowing the plants to exploit higher volume of soil, increasing the absorption of nutrients, both with higher mobility and nitrogen, responsible for the photosynthetic gain.
Regarding the effect of bioagents on rice grain yield, the treatment that stood out was the mixture of the three bioagents, presenting grain yield 19.6% higher than the control (without bioagents). The various bene its can explain the ef iciency of these bioagents in increasing grain yield, such as the promotion of plant growth through increasing hormone's levels, the ixation of atmospheric nitrogen and the solubilization of minerals [57]. Trichoderma and PGPR's directly affect hormone concentrations in the plant's rhizosphere, with an increasing demand for gibberellin, cytokinin, abscisic acid and ethylene, promoting the increase and improving root mass, the assimilation of nutrients and water and, consequently, increasing nutrients absorption of [57][58][59]. França, et al. [27], using a mixture of the same species of T. asperellum of the present work in irrigated rice, reported a 70% increase in grain yield concerning the control (without bioagent). Nascente, et al. [20] applying B. pyrrocinia and P. luorescens in irrigated rice found higher concentrations of nitrogen, phosphorus, magnesium, copper, iron, and zinc concerning the control treatment. Rêgo, et al. [60], working with the mixture of the three bioagents of the present study, reported an increase of 14 % of roots and 27% of the diameter of the vascular cylinder of rice plants, when compared to the control treatment. Sousa, et al. [55] proved that the mixture of the three bioagents used in the present work in combination with 2 ton ha -1 of silicon increased by up to 68% the shoots, 37.04% the root system and 58, 38% in the total of rice plants biomass. According to the authors, this biomass improvement occurred due to the increase of sugar's level, chlorophyll, gibberellin and salicylic acid contents in rice plants treated with bioagents and silicon, improving the germination, tillering and vigor of the plants, essential characteristics for plants of rice cultivated in the ield, mainly in a no-tillage system.

Conclusion
Our data showed that the use of 2 ton ha -1 of silicon with a mixture of bioagents was the best treatment for controlling leaf blast. Besides, from rates, 2 to 6 ton ha -1 of silicon, in Capivara Farm, and up to 8 ton ha -1 of silicon in Palmital Farm, provided the highest grain yield. The mixture of bioagents provided the highest grain yield. In this sense, the best recommendation to connect blast control, grain yield and reduction of the amount of silicon, in the BRS Primavera rice variety, was the use of 2 ton ha -1 of silicon with the mixture of bioagents.