Foliar application of Azospirillum brasilense in soybean and seed physiological quality

Bacteria of the Azospirillum genus have considerable potential for application in agricultural systems, either in co-inoculation and foliar application to increase crop yields, due to its role in the production of phytohormones. The objective of this research was to evaluate the yield and quality of soybean seeds produced under the effect of different doses of Azospirillum brasilense bacteria applied to the leaves. Seeds of four soybean cultivars (Anta 82 RR ® , BRS Favorita RR ® , BRS 780 RR ® and BRS 820 RR ® ) were produced in the 2013/2014 crop year in Lavras, Minas Gerais, Brazil, with application of six doses of inoculants based on A. brasilense (0, 300, 400, 500, 600 and 700 mL ha -1 ) of the strains, AbV5 and AbV6. Besides the yield, the mass of a thousand seeds, moisture content, germination, emergence at 7 and 15 days after sowing, emergence speed index, accelerated aging, strength, viability and mechanical damage were evaluated by tetrazolium and the hypochlorite test. Regardless of the soybean cultivar, application of up to 700 mLha -1 of A. brasilense innoculant at v3 stage (second open trefoil) of the plants, the yield, physiological potential and seeds were not affected.


INTRODUCTION
In the last decades, the soybean crops [Glycine max (L.) Merrill] have expanded to several regions of the world.In this scenario, Brazil stands out in the production of this oilseed, being the second largest producer of soybeans.In 2014/2015 cropping year, the soybean crops occupied an area of 31.33 million hectares, which reached the total production of 93.25 million tons.In the Southeast, the state of Minas Gerais is the largest producer of soybeans, with an area of about 1.31 million hectares (CONAB, 2015).
In commercial crops and seed production fields, inoculation with Bradyrhizobium japonicum bacteria is common eliminating the use of nitrogen fertilization due to the efficiency of biological nitrogen fixation (BNF).Recently, it has been speculated that the use of Azospirillum brasilense bacteria can increase the soybean crops performance (Hungria et al., 2007).Bacteria of the genus Azospirillum, produce phytohormones, *Corresponding author.E-mail: alan_zuffo@hotmail.com.
Author(s) agree that this article remains permanently open access under the terms of the Creative Commons Attribution License 4.0 International License including auxins, gibberellins, cytokinins, that under "in vitro" conditions (Araujo, 2008;Masciarelli et al., 2013), can reduce the need of chemical inputs application, provide the fixing of atmospheric nitrogen, reduce biotic and abiotic stresses, and increase crop yield (Hungria, 2011).These benefits can improve the physiological potential of seeds.Cassán et al. (2009) found that in corn (Zea mays L.) and soybean crops, the treatment of seeds with A. brasilense strains led to significant increase in germination and vigor.However, in wheat crop(Triticum aestivum L.), Brzezinski et al. (2014) found that the sanitary quality of wheat seed is not influenced by A. brasilense inoculation.The inoculation of wheat seeds with A. brasilense favors the vigor (accelerated aging) and seedlings shoot mass.
According to Azevedo et al. (2007), little has been studied on the effects of cultural practices on the physiological quality of soybean seed.It is known that high quality seeds are desirable for the success of agriculture.Thus, the objective of the research was to evaluate the yield and quality of soybean seed produced under the effect of different doses of A. brasilense bacteria inoculation applied to the leaves.

MATERIALS AND METHODS
The seeds were produced in 2013/2014 crop year, in Lavras -MG, Brazil, at the Scientific and Technological Development Center of Agriculture -at UFLA, located at 21°12'S latitude, 44°58'W longitude and altitude of 918 m in soil classified as distroferric Red Latosol -LVdf.
The climate is Cwa, according to the Köppen classification, with an average annual temperature of 19.3°C and normal annual rainfall of 1,530 mm (Dantas et al., 2007).During the seed production process, the climatic data was collected at the meteorological station of the National Institute of Meteorology (INMET) located at the Federal University of Lavras-UFLA and are presented in Figure 1.
The experiment was performed in a randomized block design with three replications.The treatments were arranged in a 4 x 6 factorial, four soybean cultivars seeds were produced (Anta 82 RR ® , BRS Favorita RR ® , BRS 780 RR ® , BRS 820 RR ® ) with the application of six doses of the inoculants based on A. brasilense (0, 300, 400, 500, 600 and 700 mL ha -1 ) applied to the leaves in the V3 stage (second open trefoil).Each plot consisted of four sowing rows of 5 m in length, spaced at 0.50 m, and the area of each plot was of 10 m 2 (5 m x 2 m).The two central rows were considered as useful area.
The sowing was performed in November 2013.The fertilization consisted of 350 kg ha -1 of N-P2O5-K2O (02-30-20), applied via groove.The Bradyrhizobium japonicum bacteria were inoculated via groove after soybean seeding.The dose of B. japonicum was 18 mL p. c. kg -1 of seedstrains SEMIA 5079, containing 10.8 x 10 6 CFU/seed of Nitragin Cell Tech HC ® (3x10 9 CFU/mL).the inoculant Azo ® (1 x 10 8 CFU/mL) strains estirpes AbV5 and AbV6 were used.The application of microorganisms was performed using a motor-driven backpack sprayer, coupled to a bar with four spray nozzles XR 110.02, applying spray volume equivalent to 150 L ha -1 .The legumes were harvested manually and threshed by stationary threshing in order to simulate the mechanical harvesting held by the harvester.The grain yield (kg ha -1 ), being standardized for grainmoisture of 13% was determined.Subsequently, the seeds were separated from impurities with the assistance of sieves, and packaged in 'Kraft' model paper bags.Seeds with moisture content above 13% were placed in the shade to slow drying.After determination of the adequate moisture content, mixing of samples and separation in sieves were done/performed.For the laboratory tests, seeds classified in the 6.00 mm diameter sieve were used.Laboratory evaluations were performed using a completely randomized design, all the tests were performed with two subsamples of 50 seeds per replication, totaling 300 seeds per treatment.The physiological and physical quality of the seeds were determined using the following determinations: Thousand seeds mass and moisture content Recommendation of Brasil (2009) was followed.

Germination test
This was done in the paper roll form with a water volume in the amount of 2.5 times the dry mass of the substrate at 2°C.The evaluations were performed on the 8th day after sowing, according to the criteria established by the Rules for Seed (BRASIL, 2009).

Emergence test under controlled conditions
The substrate used was composed of soil and sand (2:1).After sowing, the trains were kept in a greenhouse at 25°and relative air humidity of ±70%.After the emergence of the first seedling (visible cotyledon), daily evaluations were made for the number of emerged plants, with the final counting 15 days after sowing.For the speed emergence index (SEI) calculation, the Maguire (1962) formula was adopted.

Accelerating aging test
This was realized according to Marcos Filho (1999) recommendations, with 42 g seeds per treatment on an adapted stainless steel grille in gerbox boxes, with 40 mL of water in the bottom, maintained at 41°C for 72 h in a BOD and were evaluated at germination.

Tetrazolium test
The seeds were placed in a germitest paper moistened for 16 h at 25°C.After this period, they were submerged in 0.075% tetrazolium salt solution for three hours at 40°C in an incubation chamber.The viability, the damage percentage and the vigor were determined according to Neto et al. (1998).

Sodium hypochlorite test
For the sodium hypochlorite test, the seeds were kept for 10 min in a sodium hypochlorite solution (0.2%) and evaluated in accordance with Krzyzanowski et al. (2004).

Statistical analysis
The variance analysis was realized adopting the statistical model and the analysis procedure similar to Ramalho et al. (2012).The means were grouped by the Scott-Knott test (1974).The statistical analysis was realized with SISVAR ® statistical package (Ferreira, 2011).

RESULTS AND DISCUSSION
The inoculation doses with isolated A. brasilense bacteria and the interaction of cultivars (C) and bacteria (A) did not lead to a significant difference on the studied tests (Table 1).According to this fact, it can be inferred that the physiological quality of the soybean seeds, on average, does not depend on the A. brasilense inoculation dose and on the studied cultivar.Similar results were also obtained by Zuffo et al. (2015a) and Zuffo et al. (2015b).The authors did not observe the A. brasilense bacteria on the soybean agronomical characters.
A possible explanation for this fact is that the used doses did not promote any effect on the physiological quality because of the absence of effect on agronomical characters, as verified by Gitti et al. ( 2012), Zuffo et al. (2015a, b).With no influence on agronomical characteristics on the field, the plant produced seeds with a similar physiological potential.On the other hand, when different seeds from different cultivars were evaluated, differences on the physiological quality were observed (Table 1).The cultivar effect on the soybean seed physiological quality was verified by Zambuzzi et al. (2014).The cultivars have different characteristics related to the genetic background, growth habit, maturation group and other attributes, promoting the existence of some variations.
For cultivars, a high amplitude between the means for thousand seed mass was observed, from 187.22 g to 127.77 g for BRS Favorita RR ® and Anta 82 RR ® , respectively (Table 1), however, all the cultivars presented satisfactory productive performance, with results above the average mean for the crop in Minas Gerais state -2658 kg ha -1achieved during the 2014/2015 crop year (CONAB, 2015).
The moisture contents of seeds from different cultivars were similar (Table 1).According to Leoffler et al. (1998), the uniformity on the moisture content is very important to standardize the evaluations and provide consistent results.
In a general manner, the studied cultivars presented a lower germination percentage as compared to the required standard to commercialize seeds in Brazil, which is 80%, established by the normative instruction no.45 (BRASIL, 2013).Seeds with low or medium germination can generate less competitive seedlings in the field (Neto et al., 2010).The main reason for the low germination can be related to the higher percentage of dead and infected seeds verified during the germination test (Figure 2).According to Binotti et al. (2008), the seed pathogens can increase the deterioration and reduce the vigor and germination of the seeds.During the emergence tests, 7 and 15 days after emergence, higher means were observed as compared to germination.The cultivars Anta 82 RR ® and BRS 780 RR ® demonstrated more emergence percentage.The higher means from emergence test, when compared with germination test have already been mentioned in literature by Henning and Neto (1980), Bizzetto and Homechin (1997) and Zambiazzi et al. (2014).The authors related that the seedlings, on emergence, releases the infected tegument on the soil, but during the germination test, on the paper, the tegument remains associated with the cotyledons and the fungus promote the seed deterioration.

Source of variation
Yield (kg ha was expected because the seedlings that emerge first can grow more and have more biomass because of photosynthesis in the first growth stages.Besides this, in field conditions, the seedlings from seeds with a higher speed emergence index promote a faster closure between the field lines, leading to better weed control (França et al., 2010).
For the accelerating aging test, the cultivar BRS 820 RR ® followed by BRS Favorita RR ® was more sensitive to high temperature and relative humidity conditions.For the vigor percentage and seed viability, measured by tetrazolium, no significant difference between cultivars was observed.For the mechanical damages, moisture deterioration and stink bug damage were found to be the main problems for the soybean; different damages were observed according to cultivar (Figure 3).Thus, cultivars have distinct genetic characteristics that can lead to higher or lower susceptibility to damages from bugs, harvest or adverse conditions.For the damage percentage by hypochlorite, Anta 82 RR ® and BRS 780 RR ® presented lower means.This data support the results obtained for mechanical damages by tetrazolium, demonstrating consistency between the tests, but without observing effect of different doses of inoculant with A. brasilense on the physical and chemical seed quality.

Conclusion
Despite the soybean cultivar, the inoculant application with A. brasilense bacteria up to 700 mL/ha applied on plants at V 3 stage did not affect the yield, physiological potential and seed  damage incidence.

Figure 1 .
Figure 1.Monthly means for rainfall and air temperature in Lavras, MG, Brazil, in 2013/2014 crop year during soybean production.Source: National Meteorology Institute (INMET).

Figure 2 .
Figure 2. Total percentage of dead and infected seeds during germination test for each cultivar during the 2013/2014 crop year.Lavras.MG.Brazil.