Field performance of new cowpea cultivars inoculated with efficient nitrogen ‐ fixing rhizobial strains in the Brazilian Semiarid

The objective of  this work was to evaluate the contribution of efficient nitrogen‐fixing rhizobial strains to grain yield of new cowpea cultivars, indicated for cultivation in the Brazilian Semiarid region, in the sub‐medium of the São Francisco River Valley. Two experiments were set up at the irrigated perimeters of Mandacaru  (Juazeiro,  state  of  Bahia)  and  Bebedouro  (Petrolina,  state  of  Pernambuco).  The  treatments consisted  of  single  inoculation of five  rhizobial  strains  – BR 3267, BR 3262,  INPA 03‐11B, UFLA 03‐84 (Bradyrhizobium sp.), and BR 3299T (Microvirga vignae) –, besides a treatment with nitrogen and a control without  inoculation  or  N  application.  The  following  cowpea  cultivars  were  evaluated:  BRS  Pujante, BRS Tapaihum, BRS Carijó, and BRS Acauã. A randomized complete block design, with four replicates, was used. Inoculated plants showed similar grain yield to the one observed with plants fertilized with 80 kg ha‐1 N. The cultivars BRS Tapaihum and BRS Pujante stood out in grain yield and protein contents when inoculated, showing their potential for cultivation in the sub‐medium of the São Francisco River Valley.


Introduction
Cowpea [Vigna unguiculata (L.) Walp] has great importance in the Brazilian Semiarid region, especially for small farmers.This crop is mainly cultivated for the production of dry and green beans for human consumption, and it is considered an essential source of proteins, carbohydrates, fibers, vitamins, and minerals for the populations in the Semiarid region of northeastern Brazil (Santos et al., 2008).According to Freire Filho (2011), the area cultivated with cowpea in the Northeast region achieved almost 1.3 million hectares and produced more than 382 thousand tons, with productivity around 330 kg ha -1 in 2009.
Water and thermal stresses in the Semiarid region, as well as the few technological resources available for use in cowpea cropping systems, are the main responsible for the low productivity in the region (Fernandes Júnior & Reis, 2008;Fernandes Júnior et al., 2012).Some low-cost technologies can contribute to increase cowpea production in the Brazilian Northeast.Among these, the use of inoculants containing selected efficient rhizobia strains can be highlighted.The use of rhizobial inoculants has beneficial economic and environmental impacts for leguminous crops (Moreira & Siqueira, 2006;Araújo et al., 2012).Currently, four Bradyrhizobium strains (BR 3267,BR 3262, are authorized for use in the commercial production of cowpea inoculants in Brazil (Martins et al., 2003;Lacerda et al., 2004;Zilli et al., 2009).However, the rather insignificant productive response of cowpea genotypes to inoculation resulted in low diffusion of the technology among farmers in the Semiarid region.Nevertheless, over the past few years, researches have shown that some rhizobial strains, with high agronomic efficiency, can increase the productivity of cowpea in the region (Almeida et al., 2010;Costa et al., 2011;Fernandes Júnior et al., 2012;Freitas et al., 2012;Ferreira et al., 2013;Alcantara et al., 2014).
Despite these studies, there is still a lack of knowledge on the efficiency of these strains at the sub-medium of the São Francisco River Valley.Martins et al. (2003) studied the agronomic efficiency and the competitive ability of the strain BR 3267 and those results supported its inclusion in the list of bacteria used as cowpea inoculants.Up to now, field evaluations of the other three strains available for cowpea inoculation were not carried out in the region.Additionally, the field performance of the promising new strain BR 3299 T , isolated from cowpea grown in the state of Sergipe and recently described as the new specie Microvirga vignae (Radl et al., 2014), in the São Francisco River Valley, still needs to be assessed.
Moreover, the interaction between macro-and microsymbionts is an important factor that affects productive response to inoculation (Alcantara et al., 2014).The evaluation of this interaction might make it possible to more properly indicate specific inoculants, enhancing the technical recommendations for cowpea genotypes.
The inoculation response of new cowpea genotypes developed for the Brazilian Semiarid region has not yet been assessed.Among the newly developed cultivars, only the "Mulato" type, BRS Pujante, had its responses to inoculation studied and has been proven to be highly responsive to the practice (Chagas Junior et al., 2010).The "Canapu" type, 'BRS Acauã'; the "Fradinho" type, 'BRS Carijó'; and the black-coated 'BRS Tapaihum' have short cycle and high levels of minerals and micronutrients in grains (Santos et al., 2008;Santos, 2011); but, in spite of their favorable technological features, the efficiency of their association with rhizobia is still unknown.
The objective of this work was to evaluate the contribution of efficient nitrogen-fixing rhizobial strains to the development and grain yield of new cowpea cultivars, indicated for cultivation in the Brazilian Semiarid region, in the sub-medium of the São Francisco River Valley.

Materials and Methods
Two experiments were carried out under field conditions.In both assays, the inoculation treatments consisted of four strains of Bradyrhizobium sp.: BR 3267 (SEMIA 6462), BR 3262 (SEMIA 6464), INPA 03-11B (SEMIA 6463), and UFLA 03-84 (SEMIA 6461).These strains are officially indicated for cowpea inoculant production in Brazil (Brasil, 2011).Besides them, the strain BR 3299 T was also evaluated.Besides the inoculation treatments, a treatment with N fertilization and a control without inoculation and without N were assessed.The N fertilization treatment was equivalent to 80 kg ha -1 N, applied as urea, split in two times.The cowpea cultivars BRS Pujante, BRS Tapaihum, BRS Acauã, and BRS Carijó were evaluated.The rhizobial strains were supplied by the Johanna Döbereiner Culture Collection of Diazotrophic Bacteria, from Embrapa Agrobiologia, Seropédica, state of Rio de Janeiro.The cowpea seeds were acquired in Petrolina, state of Pernambuco, from Embrapa Produtos e Mercados.
The experiments used a factorial arrangement of five inoculation plus control treatments per four cultivars (7x4), in a total of 28 treatments.A randomized complete block design was used, with four replicates.
The experimental plots measured 12 m 2 and consisted of eight 3-m lines.A 0.50x0.25-mspacing between plants was adopted, in accordance to the agronomical recommendations for the crop.
The first experiment was carried out from June to August 2012, in the irrigated perimeter of Mandacaru, at the Mandacaru Experimental Field (MEF) (09º24'S, 40º26'W) of Embrapa Semiárido, in Juazeiro, state of Bahia.The soil of the area is classified as Vertissolo Háplico (Vertisol) (Santos et al., 2013), with a clayey texture.The second experiment was carried out from December 2012 to February 2013, at the Bebedouro Experimental Field (BEF) (09º09'S, 40º22'W), located at the Bebedouro irrigated perimeter of Embrapa Semiárido, in Petrolina, state of Pernambuco.The soil is classified as Argissolo Amarelo distrófico (Ultisol), with a sandy loam texture.The soils from both experimental areas were prepared with a plowing and a harrowing.
Soil samples were collected during planting in order to determine rhizobial populations.This analysis was carried out with the most probable number (MPN) method, using infected cowpea plants ('BRS Pujante'), as described in Hungria & Araújo (1994).
To prepare the inoculant, the bacteria were cultivated in YM liquid medium (Vincent, 1970), under constant stirring, for five days.Then, 10 mL of the culture broth of each bacteria were inoculated individually in plastic bags containing 30 g of sterilized peat, in order to reach the cell concentration of 10 9 viable cells per gram of inoculant.The inoculation was performed with 40 g of inoculant per 1 kg of seeds, in plastic bags containing the seeds, the inoculant, and a sugar solution (sucrose 10% w/v), and then hand mixed.The inoculated seeds were shade-dried, and sowing was performed manually soon after inoculation.
Weed control was done manually, according to necessity.The control of plagues and diseases was done with products recommended for the crop in the region.For the experiment implemented at the MEF, a furrow irrigation system was used, applying a daily irrigation depth of approximately 5.5 mm.For the experiment implemented at the BEF, a dripper irrigation system was used, with drippers spaced at 0.5 m and with a 1.6 L h -1 water flow.During the execution of the experiments, average monthly precipitation was between 1.4 and 27.6 mm at the MEF and BEF, respectively, with a daily average temperature of 24 to 28°C.
A first evaluation of plant nodulation was performed during flowering (45 days after the emergence of the plants).Ten plants were collected in 1.0 m of the second row of each plot.These plants were cut at soil level, and the radicular system was separated and packed in plastic bags (Fernandes Júnior et al., 2012).The roots were washed, and the nodules were separated from the radicular system and counted.The aerial part and the nodules were packed in paper bags, dried at 65°C in a forced air chamber until constant weight, and then weighed.The following characteristics were determined: number of nodules (NN), shoot dry matter mass (DMS), nodule dry matter mass (DMN), and nitrogen accumulated in the shoots (NAS) using the semimicro Kjeldahl method (Liao, 1981).
A second evaluation was done during harvest.Beans were gathered from the useful plot (4 m 2 , central to each plot).The grains were threshed for weighing, and grain yield (GY) was calculated.Grain protein content (GPC) was determined according to Williams (1984).
Data were subjected to analysis of variance, using the Sisvar 4.2 statistical analysis system (Ufla, Lavras, MG, Brazil).Means were compared by Student's test, at 10% probability.Previously to the analysis, data were transformed by (x+1.0) 0.5 .

Results and Discussion
Rhizobial populations of the experimental areas were estimated at 2.8x10 3 and 4.3x10 3 cells per gram of soil, for the MEF and BEF, respectively.These populations can be considered high.In soils with established native rhizobia populations, the efficiency of the inoculation may be compromised because the introduced rhizobial strains need to compete with the indigenous isolates for nodulation sites (Zilli et al., 2013).
At the MEF, the NN in the cultivar BRS Pujante was influenced by the inoculated strain (Table 1).Inoculation treatments had higher NN than the fertilization treatment with 80 kg ha -1 N. BR 3262, BR 3299 T , INPA 03-11B, and UFLA 03-84 provided greater NN than the control treatment (without inoculation and N fertilization).Nodule formation inhibited by N fertilization was also observed by Silva et al. (2012), when evaluating rhizobia inoculation in the cultivar BRS Pujante.
Cultivars BRS Tapaihum, BRS Carijó, and BRS Acauã did not respond to the inoculation treatments, regarding NN. Chagas Junior et al. (2010) found that the cultivar BRS Pujante nodulated abundantly with all evaluated strains, with lower nodulation rates observed in the nitrogen fertilization and in the absolute control treatments.
For DMN, BRS Pujante stood out, whereas the other cultivars showed equal means.Inoculation of the strain BR 3262 increased DMN of the cultivar BRS Pujante, when compared to the control treatment, but it did not differ from inoculation with the strains BR 3299 T , UFLA 03-84, and INPA 03-11B.
Regarding DMS, inoculated treatments did not differ from the control treatment.However, a significant increase in NAS was observed for the cultivar BRS Pujante inoculated with BR 3267, and for BRS Carijó, inoculated with BR 3262, when compared to the N fertilization and to the control treatment, respectively.Apart from nodulation, nodule efficiency may be influenced by plant genotype and by the efficiency of the strain in the nodules, which also depends on the macrosymbiont genotype (Alcantara et al., 2014).Therefore, the difference in nitrogen accumulation in shoots indicates that the genotypes BRS Carijó and BRS Pujante can benefit from inoculation.
At the BEF, NN differed according to the treatments (Table 2).Cultivar BRS Pujante, inoculated with BR 3267 and BR 3262, showed higher NN.Once again, N fertilization inhibited the formation of nodules, and the strains BR 3267 and BR 3262 provided an increase in nodule mass compared to the other treatments.
For the cultivar BRS Carijó, the strains BR 3267 and BR 3262 stood out regarding NN and DMN, respectively.For 'BRS Acauã', however, only the strain BR 3262 differed from the treatment with nitrogen fertilization, but it did not differ from the other treatments.'BRS Tapaihum' was not significantly affected by the treatments, regarding NN and DMN.The inoculation with the strain BR 3299 T increased DMS of the four studied cultivars.
but did not differ from the N fertilization and the control treatments.For 'BRS Tapaihum', the strain BR 3262 provided the highest NAS.For 'BRS Carijó', inoculation treatments did not differ from the control.For 'BRS Acauã', BR 3262 provided higher NAS than the nitrogen fertilization and the other inoculated treatments.
The similarity of nodulation parameters and vegetative development between some inoculated and non-inoculated treatments indicates that the native rhizobial population was efficient to establish symbiosis with cowpea, which is able to nodulate with a wide range of tropical soil rhizobia (Leite et al., 2009;Jaramillo et al., 2013).The capacity of the native rhizobial population to establish an efficient symbiosis with cowpea has already been observed in the Brazilian Semiarid (Martins et al., 2003;Freitas et al., 2012;Alcantara et al., 2014).
Grain yield results suffered the interaction between inoculation treatments and cowpea genotypes, in both experiments.Grain yield ranged from 1,091 to 1,629 kg ha -1 in the MEF, and from 817 to 1,824 kg ha -1 in the BEF (Table 3).Grain yield levels were considerably higher than 330 kg ha -1 , which is the average for the northeastern region, but within the expected levels for inoculation experiments under irrigated conditions in the Brazilian Semiarid (Santos et al., 2008;Santos, 2011).
In the MEF experiment, when cultivar BRS Pujante was inoculated with the strains BR 3267 and UFLA 03-84, productivity ranged from 1,496 to 1,629 kg ha -1 , higher than the ones observed with the strains BR 3262 and INPA 03-11B.
In the BEF, grain yield of non-inoculated treatments did not differ from the ones obtained with the strains BR 3267, BR 3299 T , and UFLA 03-84.Inoculation treatments had a similar yield to the one of the control and the N fertilization treatments.However, the strains UFLA 03-84 and BR 3262 had an increase of 30 and 31% in grain yield, respectively, compared to the control treatment.Ferreira et al. (2013) reported that the cultivar BR 17 Gurguéia, inoculated with the strain BR 3262, had a significant increase in grain yield, of 50.17%, in comparison to the untreated mineral N treatment without inoculation.
Inoculation treatments with the cultivar BRS Tapaihum had similar grain yield to that of the control treatment.At the BEF, however, the association of this cultivar with the strains BR 3262, UFLA 03-84, and INPA 03-11-B resulted in grain yield equal to the one obtained with the N-supplied control, but greater than the one of the absolute control.Inoculation with the strain INPA 03-11B stood out, increasing yield in 89%, when compared to the absolute control.
For 'BRS Carijó', the inoculation of the strains BR 3262, BR 3299 T , and UFLA 03-84 provided higher means at the MEF than the strains BR 3267 and INPA 03-11B, but similar yield to that of the absolute control.At the BEF, cultivars did not differ also.For 'BRS Acauã', grain yield at the MEF did not differ.
Table 2. Response of cowpea (Vigna unguiculta) cultivars to the inoculation of rhizobial strains, at the experimental field of Bebedouro, Petrolina, state of Pernambuco, Brazil (1) .-----------------(mg per plant)------------------( However, at the BEF, the association with the strains BR 3267, BR 3299 T , and INPA 03-11B increased productivity in 88, 83, and 83%, respectively, achieving higher yields than the one of the control treatment, but similar ones to that of the N fertilization treatment. The obtained results are similar to the ones observed in regions of the Brazilian Semiarid.Costa et al. (2011) found that inoculation with INPA 03-11B promoted similar yields to that of N fertilization, in Bom Jesus, state of Piauí, with productivity of up to 1,604 kg ha -1 .Almeida et al. (2010) in an experiment carried out at Teresina, state of Piaui, reported a grain yield of 1,637 kg ha -1 for the cultivar BR 17 Gurguéia inoculated with the strain BR 3267, and of 1,823 kg ha -1 when inoculated with the strain BR 3262, a 24.6 and 38.9% increase, respectively, compared to that of the control treatment.
Grain protein contents in 'BRS Pujante' were influenced by the inoculation treatments and by N fertilization (Table 3).In both trials, these treatments showed higher GPC than those observed in the absolute control, especially with the strain BR 3299 T , which increased GPC in 58.8 and 37.1%, at the MEF and BEF, respectively.For 'BRS Tapaihum', higher GPC at the MEF was obtained with the strains INPA 03-11B, BR 3299T, and UFLA 03-84, with 26.7, 24.9, and 24.8%, respectively.These GPC values represented a respective increase of 38.3, 29, and 28.5% in comparison to the control treatment, which showed GPC of 19.3%.At the BEF, this genotype performed better with the strains UFLA 03-84, BR 3262, and INPA 03-11B, which increased GPC in 39.8, 37.7, and 29.8%, respectively, compared to the control.Therefore, 'BRS Tapaihum', which has a peculiar black-coated seed, despite not responding to inoculation regarding grain yield, had more nutritious grains with these treatments.For genotypes BRS Carijó end BRS Acauã, inoculation treatments had no effect on GPC, except when BRS Carijó was inoculated with INPA 03-11B, which provided lower values.In the Semiarid region, GPC is an important trait, since the production is mostly destined for human consumption and cowpea is considered one of the main protein sources for rural populations (Santos et al., 2008;Freire Filho, 2011).
The observed dependence of the inoculation treatments on genotypes, rhizobia strain, and environment is corroborated by other studies (Melo & Zilli, 2009;Alcantara et al., 2014).Therefore, plant breeding may have an important role in improving N supply for cowpea by biological fixation (Xavier et al., 2006;Araújo et al., 2012).The present study also emphasizes the importance of testing the response to inoculation of different varieties, under distinct edaphoclimatic conditions.
For cowpea, a lot of attention has been given to the selection of the microsymbionts, since the selection of the strain is the main resource for optimization of biological nitrogen fixation (BNF) (Fernandes Júnior Table 3. Grain yield and grain protein content of the evaluated cowpea (Vigna unguiculta) cultivars, according to the inoculation treatments in the experimental fields of Mandacaru (MEF) and Bebedouro (BEF) (1) ., 2012).However, an adequate interaction between symbiotic partners is essential to BNF efficiency and, for this reason, it should be prioritized in breeding programs for this species, taking the well-succeeded example of soybean (Glycine max L.) in Brazil.

Conclusions
1. Cowpea cultivars respond differently to the inoculation of rhizobial strains, in the Brazilian Semiarid.
2. The inoculation of rhizobial strains can be beneficial to grain yield and protein contents in cowpea, mainly for the cultivars BRS Pujante and BRS Tapaihum, in the sub-medium of the São Francisco River Valley.