Agronomic characteristics, isoflavone content and Kunitz trypsin inhibitor of vegetable soybean genotypes

Five vegetable soybean genotypes were evaluated for agronomic characteristics, functional and antinutritional properties. The experiments were conducted in the field, in the School of Agricultural and Veterinary Sciences, in Jaboticabal, São Paulo State, Brazil. The experimental design was a randomized block design with five treatments (genotypes) and four replications. The genotypes JLM003, JLM004, JLM010, JLM024 and JLM030 were evaluated in the following characteristics: number of pods per plant, number of seeds per pod, fresh weight of pods per plant (g), fresh weight of 100 seeds (g), estimated total yield of immature seeds (kg ha-1), moisture content, proteins, lipids, ash, carbohydrates, isoflavones and Kunitz trypsin inhibitor. In the assessment of the major agronomic traits, in the conditions of this study, the genotype JLM010 was the most recommended for the region of Jaboticabal. This genotype showed better agronomic characteristics of production as a vegetable, producing 136.04 g of fresh weight of pods per plant, 92.52 g of fresh weight/100 seeds and 11.12 t/ha of total immature seeds, and better functional properties.


C urrently, soybean (Glycine max)
is one of the most important agricultural products in the Brazilian economy, highlighting the country's export tariff. Only about 2% of the national production is for direct consumption, while 70% of the production is exported as in natura grains (Brasil, 2007).

Palavras
For this type of products, some characteristics could be considered as large seeds, high protein content; yellow color hilum; absence of lipoxygenases (better flavor), and reduced level of Kunitz trypsin inhibitor, which allows reduced heat treatment (Embrapa, 2003).
Production system for vegetable soybeans differs from common soybeans only in the stage of harvest and postharvest procedures. The ideal harvest period is between 35 and 40 days after flowering, when the levels of saccharose are elevated, conferring sweeter taste to the seeds (Chiba, 1991).
The vegetable soybean or edamame can have good acceptability and constitute a healthy habit, as it is rich in protein (13%), contains low oil content (5.7%), non cholesterol and hydrogenated fat. It also presents reasonable levels of phosphorus, calcium and vitamins B 1 and B 2 (Shanmugasundaram & Yan, 2004).
Considering all these advantages, studies are needed to characterize the vegetable soybean cultivars. Therefore, evaluations of the agronomic performance, chemical composition, functional properties, and presence of antinutritional factors should be conducted. The present work evaluated performance of five vegetable soybean genotypes.

MATERIAL AND METHODS
The experiments were conducted in the field, from November 1, 2006to March 19, 2007, in the experimental area of the Universidade Estadual de São Paulo, in Jaboticabal, São Paulo State (21° 14' 05''S, 48° 17' 09''W, altitude of 614 m). The climate, according to the Köppen classification, is of the Aw type with transition to Cwa (Volpe 1 ), and the soil of the experimental area was classified as being eutroferric latossolo (Embrapa, 1999).
The experimental design was randomized blocks with five treatments (genotypes) and four replicates. The utilized genotypes were previously tested in earlier studies and, that one presenting the best characteristics for consumption as vegetable, was used in the current experiment: JLM003 originated from the germplasm collection of Embrapa Hortaliças (medium size, dry weight of 100 seeds = 30 g, purple flower, yellow hilum, yellow grain); JLM004 from Embrapa Hortaliças (medium size, dry weight of 100 seeds = 37 g, purple flower; light brown hilum); JLM010 originated from Asian Vegetable Research and Development Center, Taiwan (medium size, dry weight of 100 seeds = 41 g, white-flower; yellow hilum and seed), and JLM024 from germplasm collection of Embrapa Cenargen (tall size, dry weight of 100 seed = 19 g,) and JLM030 from Embrapa Soja (medium size, dry weight of 100 seeds = 27.8 g, yellow hilum).
Each experimental plot consisted of four rows 6 meters length, 0.60 m between rows and 0.15 m between plants. In each plot, 20 plants from two central rows were evaluated for all characteristics.
Due to intensive attacks of pigeons in previous direct seeding plantings, we decided to grow seedlings in Polystyrene trays with 128 cells, utilizing the commercial substrate Plantmax Hortaliças ® HT, where one seed was planted per cell. Seeds were previously inoculated with Bradyrhizobium spp, in the proportion of 500 g of innoculum per 50 kg of vegetable soybean.
The seedlings were transplanted 20 days after planting date, on November 21, 2006, when plants presented 2 or 3 definitive leaves and 12 cm height. At that time, they were removed from the trays and planted in the definitive location that was prepared according to conventional system of plowing and harrowing. Soil was fertilized five days before transplanting, based on soil analysis and recommendations of Raij et al. (1997). Fertilization consisted of 50 kg ha -1 of N, 40 kg ha -1 of P 2 O 5 , 60 kg ha -1 of K 2 O and 1 kg ha -1 of B, which were provided by the sources of urea, simple superphosphate, potassium chloride and borax, respectively.
Sprinkler irrigation was used daily to keep the soil continuously moist and adequate for better crop development.
The harvests were carried out when pods were at the R 6 reproductive stage, using the reference scale of Fehr & Caviness (1977), adapted by Costa & Marchezan (1982). The plants were cut just above the soil surface and taken to the laboratory where the pods were removed.
We evaluated the number of pods/ plant; number of seeds/pod; fresh weight of pods/plant; fresh weight of 100 seeds; estimated total productivity of immature seeds (obtained from the number of plants/m 2 , number of pods/plant, mean number of seeds/pod and mean mass of 100 seeds). Moisture content was determined by the gravimetric method, protein was quantified by the micro Kjeldahl method, according to AOCS (1988), and lipids were determined by Soxhlet extraction according to the method of AOCS (1988). Ash content was determined according to the method proposed by the AOCS (1988), through complete incineration of the sample in a muffle oven, with an initial temperature of 150°C, an increasing of 50 to 50°C every 30 minutes, until reaching the final temperature of 550°C. Carbohydrates were determined by the difference as proposed by the AOCS (1988). Isoflavones were quantified by high performance liquid chromatography (HPLC) following the method of Berhow (2002). Kunitz trypsin inhibitor (KSTI) was determined in extracts obtained from defatted flour, by using a standard trypsin with BPNA as substrate and reading absorbance at λ= 410 nm, according to Kakade et al. (1974) changed per Hamerstrand et al. (1981. The data were submitted to analysis of variance by the F test, according to the proposed design, and the means were compared by Tukey's test at the 5% probability level, using the Estat program (System for Statistical Analysis), version 2.0.

RESULTS AND DISCUSSION
Significant differences were found at the 5% probability level among treatments for all agronomic characteristics examined, except for number of seeds per pod (Table 1).
Regarding the number of pods per plant, the genotypes were not different: JLM004 presented 48.56 pods, while JLM003 presented 48.94 and JLM010 51 pods. The highest value for this characteristic was observed in genotypes JLM024 (70,62) and JLM030 (75,10), not differing from both. In general, vegetable soybean genotypes with smaller numbers of pods have seeds that are relatively large. All the evaluated genotypes can be utilized commercially, because according to Nguyen (2001) Similar trend was noted in other study of Castoldi et al. (2008), in wich despite significant difference among genotypes, the number of seeds per pod Means followed by the same letter in the column do not differ according to Tukey's test at the 5% probability level (médias seguidas da mesma letra na coluna não diferem entre si pelo teste de Tukey ao nível de 5% de probabilidade); **, NS significant at the 1% probability level and not significant at the 5% probability level by the F test, respectively (**, NS significativo ao nível de 1% de probabilidade e não significativo ao nível de 5% de probabilidade pelo teste F, respectivamente); DMS: Least Significant Difference (diferença minima significativa). 1 Calculated based on fresh weight ( 1 cálculo baseado na matéria fresca); 2 Means followed by the same letter in the column do not differ according to Tukey's test at the 5% probability level ( 2 médias seguidas da mesma letra na coluna não diferem entre si pelo teste de Tukey ao nível de 5% de probabilidade); *,**, NS significant at the 5% probability level, significant at the 1% probability level, and not significant at 5% probability level by the F test, respectively (*,**, NS significativo ao nível de 5% de probabilidade, significativo ao nível de 1% de probabilidade e não significativo ao nível de 5% de probabilidade pelo teste F, respectivamente); DMS: Least Significant Difference (diferença minima significativa). Such explanation is in line with one of the results reported by Castoldi et al. (2006) that when the genotypes are planted with a spacing of 0.15 m, the weight of the pods per plant increases significantly compared to that with spacings of 0.05 m and 0.10 m.
One reason for the increase in the weight of pods of plants cultivated in larger spaces may be due to a greater light interception, in plants more spaced, between stages R 1 (beginning the flowering) and R 5 (beginning the grain filling), which permits greater movement of nutrients into the pods and consequently greater share of these.
For fresh weight (g) of 100 seeds we observed 92.52 g for the genotype JLM010, indicating that this genotype has large seeds. According to Carrão-Panizzi (2006), large seeds facilitate cooking and eating soybeans as a vegetable because the larger the volume the easier it is to shuck the seeds. Therefore, it is important to select genotypes with greater fresh weight of 100 seeds. Charlo et al. (2008) also found that the genotype JLM010 has large seeds (85.25 g) in relation to other genotypes, which is very desirable in vegetable soybean cultivars. These findings are similar to those reported by Castoldi et al. (2006) who showed that for the December planting, the largest mean fresh weights of 100 seeds were observed in the genotypes JLM010 (68.19 g) and JLM003 (62.20 g). For the September planting, the genotype JLM010 was the one that showed the best performance (86.97 g) of all the genotypes.
The values above are similar to those found by Castoldi et al. (2009), who showed that the largest mean fresh weights of 100 seeds were observed in the genotype JLM010 (68.2 g) sown in September, and that corroborates with the study of Smiderle et al. (2007) in assessments of productivity and quality of seeds of eight vegetable soybean genotypes in the Roraima's cerrado, where the greatest fresh weight (g) of 100 seeds occurred for the genotype JLM 10 (93 g).
The genotype JLM010 showed a higher value of estimated total productivity of immature seeds (11.12 t/ha), not differing statistically from the genotypes JLM030 (10.01 t/ha) and JLM024 (7.48 t/ha).
As in the present study, Costa et al. (2005), evaluating 10 varieties of vegetable soybean in Montes Claros, found that the genotype JLM003, showed the lowest productivity (Table  1).
In the experiment of Smiderle et al. (2007), the yields were significantly lower (3.45 t/ha to 5.33 t/ha, respectively for genotypes JLM 18 and JLM 08), because following the productivity calculation, these authors separated pods with 0 seed, 1 seed, 2 seeds and 3 seeds. This demonstrates the importance of separating the pods, so that the calculation of productivity is not overestimated, since for vegetable soybean are recommended only pods with two or more seeds.
Means for the chemical proximate composition, levels of isoflavones and Kunitz trypsin inhibitor evaluated for five genotypes of vegetable soybean are presented in Table 2.
C r a n c i a n i n o v e t a l . ( 2 0 0 5 ) determined the chemical composition of the vegetable soybean line BRM94-52273, in R 6 reproductive stage, and found that the moisture content was 67%; this R 6 moisture level is similar to that reported here. It was also reported that the consumption of green soybean at stage R 6 is healthy, because it constitutes a vegetable with high nutritive value, besides having an elevated moisture content which confers good texture to the seeds.
Similar results were found in the study conducted by Bates et al. (1977) who evaluated the protein quality of mature green soybean, mature dry soybean and sprouting soybean at the University of Florida, where they found that the moisture level for the mature green soybean was 69.4%.
All these values are in agreement with those found by USDA (2008) which reports that the quantity of water present in green soybean, boiled with salt and drained is 68.60 g/100 g (67.50%).
In relation to protein content, the genotype JLM010 (16.54%) showed the highest value, differing statistically from the others except for JLM024 (16.33%) ( Table 2).
Protein contents in the current experiment are lower than that one found by Castro et al. (1973) (29.2 to 57.9%), Pereira & Oliveira (2004) (30 to 45%) and Kamizake et al. (2006) (42.2 to 45.4%). Differences detected by these authors are due to analyzes of the chemical composition of matured grain soybean cultivars, which have a maximum of 12% moisture, in contrast to vegetable soybean, which has more than 60% moisture. This affirmation is in accordance with Masuda (1991), who reported that the chemical compounds vary with the stages of growth of the plant, therefore the observed differences between grain and vegetable soybean. Crancianinov et al. (2005) reported that the protein level in vegetable soybean was 12%, value close to that obtained in the current experiment, which means that it would require the ingestion of 210 g of vegetable soybean, according to FDA (1991) for a healthy diet (25 g of soybean protein).
The protein composition of vegetable soybean cultivars studied here is also in accordance to values found by Masuda (1991) (11.40%) and by the USDA (2008) (12.95%), where this product stands out as an excellent source of proteins.
According to Moraes et al. (2006), as the level of protein increases, the level of oils decreases, which was noted also in the present study. Therefore, we can state that for vegetable soybean, the ideal would be seeds with high concentration of protein and consequently low concentration of oils, thereby making this a vegetable not only having substantial flavor and being rich in protein but also being low in calories. Crancianinov et al. (2005) reported that lipid concentration increases from stage R 6 (5%) to stage R 8 (17%). The grain type soybean thereby has a higher oil content compared to the vegetable type soybean, which makes the former more caloric, when considering the dry mass.
The USDA (2008) reports that the amount of lipids present in green soybean is 6.80 g/100 g (6.80%), demonstrating the low caloric character of the genotypes in question. Average lipid level in this experiment was 3.92%, a value lower than the obtained by the USDA (2008). This may have occurred due to different temperatures during the filling stage of pods, since according to Sediyama et al. (1985), the oil and protein contents, of soybean grain are influenced by temperature.
Data presented in Table 2 indicate that there was no significant difference among the genotypes evaluated for ash content, which showed values of 1.53% (JLM004 and JLM030) to 1.90% (JLM010). A similar data of 1.70% was reported by the USDA (2008).
Comparing the percentage of ash present in grain and vegetable soybean, studies of Ciabotti et al. (2006) and Kamizake et al. (2006) reported that the grain soybean presents higher quantities of ash (between 4.51% to 5.24% and 3.64% to 3.84, respectively). This is due to the fact that since the beginning of the reproductive stage of grain soybean to seed maturity, the plant will have available a longer period of time for the accumulation of chemical components.
Carbohydrate level was not statistically different among the genotypes. The observed values were 18.60% for JLM030, 18.20% for JLM024 and 17.06% for JLM010. These values are higher than those found by USDA (2008) (11.05 g/100g) where the quantity of dietary fiber (4.20 g/100g) was reported separate from the quantity of carbohydrates, which was not done in the present results.
With respect to isoflavones, the genotype JLM024 had a high content (92.62 mg 100g -1 of sample), not differing statistically from the g e n o t y p e J L M 0 1 0 ( 6 9 . 5 4 m g 100g -1 ) . Crancianinov et al. (2005) found higher isoflavone concentrations in the stages R 6 (42.63 mg/100 g) and R 7 (49.20 mg/100 g). Despite that the cultivar studied by Crancianinov et al. (2005) showed a high quantity of total isoflavones in stage R 7 , but this value was lower than the highest value observed in the present study (92.62 mg/100g of sample) and higher than the lowest observed value (22.67 mg 100 g -1 of sample). This can be explained by the environmental and genetic differences that affect concentrations of isoflavones in soybean seeds (Tsukamoto et al., 1995;Carrão-Panizzi et al., 1999). Simone et al. (2000) studied the retention and changes of soybean isoflavones and carotenoids in immature soybean seeds (Edamame) during processing; they observed that the level of total isoflavones varied from 70.2 mg/100 g to 230.8 mg/100 g, which were higher than those found in the present experiments. It was because they were calculated based on dry weight and not on fresh weight as in the study in question.
In relation to the amount of Kunitz trypsin inhibitor (KSTI), the genotype JLM024 showed the highest mean (17.80 mg/100 g of sample), differing from the other genotypes which showed mean values of 13.9 mg/100 g to 15.71 mg/100 g sample. Kumar et al. (2006) evaluated the physicochemical properties of immature pods of Japanese soybean cultivars and found that the amount of Kunitz trypsin inhibitor varied from 5.6 mg/g (Enrei) to 25.5 mg/g (Hatsataka), demonstrating different values for each genotype.
According to Morais & Silva (1996) this anti-nutritional factor is actually not a problem, because soybeans, for human consumption, are always cooked. The absence of this factor in the plant, however, results in less cost in processing and greater safety in the direct utilization of soybean seeds.
According to the results of the agronomic, functional and antinutritional characteristics evaluated in this study, we suggest the genotype JLM010 as the most recommended for the region of Jaboticabal. This genotype also showed better agronomic characteristics for production of vegetable soybean and also displayed better functional properties.