Environmental and genetic variation of isoflavone content of soybean seeds grown in Brazil

The objective of this work was to evaluate isoflavone concentrations in seeds of different Brazilian soybean cultivars grown in a range of locations and environmental conditions in Brazil. Seeds of 233 cultivars grown in Ponta Grossa, PR, Brazil, during the 2001/2002 soybean season, and of 22 cultivars sown in different locations of Brazilian Northeast, Southeast on South regions were analyzed for  total  isoflavones,  including daidzin, glycitin, genistin and acetylgenistin. The total isoflavones ranged from 12 mg 100 g‐1 (cv. Embrapa 48) to 461 mg 100 g‐1 (cv. CS 305) among the 233 cultivars grown in Ponta Grossa, and the differences among them are due to genetic effects since all cultivars were grown and collected at the same locatation and year. This is an indication of the possibility of breeding for isoflavone content. Differences in isoflavone content observed in the cultivars grown in different locations permit the selection of locations for optimum isoflavone content (low or high), depending on the uses of soybean. In the Northeast region (5–8°S), higher concentrations of  total  isoflavones were observed at São Raimundo das Mangabeiras (232 mg 100 g‐1) and Tasso Fragoso (284 mg 100 g‐1) municipalities, and in the South  (23–30°S), isoflavones were higher in Guarapuava, Canoinhas, Vacaria and Campos Novos municipalities, ranging from 130 to 409 mg 100 g‐1.


Introduction
Three structurally distinct isoflavones (daidzein, genistein, and glycitein) are accumulated in soybean seeds as their acylated 7-O-glycoside forms (daidzin, genistin, and glycitin) (MacLeod & Ames, 1988).The 6-O-malonyl derivatives of these glucosides have been reported by Kudou et al. (1991) as the predominant form present in intact soybean seeds.The malonyl isoflavones are heat labile, and easily converted to the corresponding glycosides and aglycones when submitted to heat (Kudou et al., 1991).
Several studies indicate that genetic variability affects seed isoflavone content of soybean cultivars, and isoflavone content variations can also be affected by annual changes in environmental conditions Pesq.agropec.bras., Brasília, v.44, n.11, p.1444Brasília, v.44, n.11, p. -1451Brasília, v.44, n.11, p. , nov. 2009 from different locations.Air temperature during the seed-filling period is the most critical environmental factor affecting the final isoflavone concentration.Other factors that affect isoflavones are variations in temperature -both levels and length of time at high or low temperatures -, humidity, light intensity, rainfall -duration and number of rainy days -, and physical damage caused by insects and diseases (Wang & Murphy, 1994;Tsukamoto et al., 1995Tsukamoto et al., , 2001;;Carrão-Panizzi et al., 1998;Daydé et al., 2004).
Soybeans grown at low latitudes, generally with higher average temperatures, tend to have relatively lower isoflavone content.The average isoflavone content of Brazilian soybean cultivars is of about 150 mg 100 g -1 , whereas for American soybeans it is 200 mg 100 g -1 , while soybean grown in Japan and Canada have higher concentrations (450 and 550 mg 100 g -1 respectively) (Tsukamoto et al., 2001).Although average isoflavone concentrations of Brazilian soybeans are relatively low (Carrão-Panizzi et al., 1996), they vary considerably among cultivars (Silva et al., 2009) and in cultivars grown at different locations over the years.A range of 227 to 105 mg 100 g -1 was observed for the same cultivar (IAS 5) grown, respectively, in a location with low (Vacaria, RS, 19°C), and high (Palotina, PR, 24°C) average air temperatures during the seed-filling months (Carrão-Panizzi et al., 1999).Daydé et al. (2004) found that low temperatures coupled with late precipitation (rain or irrigation) increased isoflavone concentrations.
Isoflavones in the diet may have effects on the health of humans, through their antioxidant and estrogenic activities, and a wide spectrum of pharmacological properties (Messina, 1999).Isoflavones are also reported as influencing the development and remediation of some of the chronic diseases associated with aging, such as osteoporosis, cognitive disfunction, coronary heart disease and cancer (Messina, 1999).They also have an effective antixenotic, antioxidant and antibiotic action in mammalian systems (Graham, 1991).
Much research is underway with soybean and its isoflavones regarding human health issues.In order to certify isoflavone content and composition in dietary supplements or soyfoods, the qualitative determination of these compounds in various soybean cultivars grown under different environmental conditions is needed.
The objective of this work was to evaluate isoflavone concentrations in seeds of different Brazilian soybean cultivars, verifying genetic differences and effects of different sowing locations in Brazil.
Seed lots from each location and cultivar were analyzed for isoflavone concentration.Seed samples were dried in an oven with forced air circulation at 50°C for three days, before grinding.Weighed samples of ground soybean were defatted overnight with hexane, then washed with three volumes of hexane and allowed to dry.The defatted samples were then extracted overnight with 4x DMSO:MeOH (1:1, volume:mass) at 50°C.Sample extracts were centrifuged at 20,800 g for 15 min at 5ºC.After centrifugation, an aliquot of each sample was filtered through a 0.45-µm Millipore filter into autoinjector vials for HPLC analysis.
The isoflavone concentrations were determined according to Berhow (2002) and Barnes et al. (1994) on an Agilent 1100 Series HPLC system equipped with a photodiode array detector system running under HP Chemstation software (A.06.01)(Agilent Technologies, Wilmington, DE, USA) and a 5-µm particle size Inertsil ODS-3 C-18 reverse-phase column (4.6x250 mm) (Metachem, Torrance, CA) with a Metaguard column (Metachem, Torrance, CA).The column was equilibrated with 20% methanol (0.025% Pesq.agropec.bras., Brasília, v.44, n.11, p.1444Brasília, v.44, n.11, p. -1451Brasília, v.44, n.11, p. , nov. 2009 trifluroacetic acid) and 80% water (0.025% trifluroacetic acid) at a flow rate of 1 mL min -1 .After injection, the column was run for 2 min at the initial conditions, then developed with a linear gradient to 100% methanol (0.025% trifluroacetic acid) over a period of 50 min.The wavelength monitored and used for integration was of 285 nm.Isoflavone peaks were identified by their spectra and retention time and confirmed by LC-MS.Standard curves were prepared with pure standards of daidzin, glycitin and genistin isolated at the National Center for Agricultural Utilization Research -NCAUR (USDA/ARS, USA).Molar extinction coefficients obtained from these curves were used to determine the levels of all forms of the isoflavones measured in the samples on a molar basis.
The multivariate statistics method was applied to the 233 soybean cultivars using a k-means partition cluster analysis (Souza, 2004) by means of the Statistica software, version 7.1 (Statsoft, 1995).For cluster analysis, each isoflavone content datum was standardized by subtracting the average value and dividing the result by the standard deviation.Differences in isoflavone content -daidzin, glycitin, genistin and acetylgenistin, the only forms found above detection limits in these samples -were evaluated in 233 cultivars by cluster analysis through the multivariate exploratory method of k-means (Andeberg, 1973).The method determines the greatest distance between any two cultivars at different clusters.
A completely randomized design was used to evaluate the experimental data from different planting locations.To determine significant differences among the mean values, ANOVA and Tukey tests at 5% probability were used for comparing data from cultivars and locations by using SAS (SAS Institute, 1995).

Results and Discussion
The 233 soybean cultivars were divided in nine groups according to their isoflavone contents (Table 1).The analysis of variance was significant (p≤0,000) for all the studied variables of the nine groups.The evaluation of the classified cultivar groups showed that the multivariate method was efficient in 94.4% of the cases.For each of the nine groups, cultivars were arranged as presented in Table 2.The descriptive statistical analysis showed large variability for daidzin and genistin.
Isoflavones are highly affected by environmental conditions, as has already been observed in other soybean cultivars and in alternate growing locations (Tsukamoto et al., 1995;Carrão-Panizzi et al., 1998;Daydé et al., 2004).In Brazil, soybean is grown in locations within latitudes 32°S to 5°N.This wide environmental range supports the need for further research aiming at determining location effects on isoflavone content of soybean seeds.
In the locations in the Northeast of Brazil (5-8°S), the weather is very similar, with average temperatures of 27°C for the seed-filling period (February to April).Some variation in the local conditions (precipitation or temperature) could cause the differences in isoflavone concentrations among locations.No significant correlation was found between cultivars and locations for this region.However, in the municipalities of São Raimundo Table 2. Soybean cultivars that compose each of the nine groups defined by cluster analysis.
das Mangabeiras, MA, and Tasso Fragoso, MA, higher concentrations of total isoflavones were found for all cultivars (Table 3).In general, differences among cultivars were not remarkable in the Northeast, but cultivars BRS Sambaíba and BRS Tracajá showed the highest and lowest total average values for isoflavone content.
Growing locations in Minas Gerais, MG, in the Southeast region, are much closer to each other (only 18°-19°S), and all had an average temperature of 24°C, and precipitation of about 400 mm.Nevertheless, higher concentrations of isoflavones were observed in Sacramento, MG, for all cultivars (Table 4).Cultivar Preciosa had higher concentrations of isoflavones in all locations except for Uberlândia, MG.
In the South, with more weather variation among locations (Figure 1), all cultivars grown in Londrina and Palotina, PR, with higher average temperatures during the filling period (26°C), showed reduced isoflavone concentrations.Higher concentrations of isoflavones were observed in Guarapuava, PR, and Campos Novos, SC, which had the lowest average temperatures (21°C).Similar observations have encouraged nutriceutical-grade soybean production at locations in the South of Paraná state, where reduced average temperatures have resulted in increased isoflavone content (Carrão-Panizzi et al., 2003).Cultivars Codetec 202 and IAS-5 had the highest isoflavone concentrations (Figure 1).Cultivar IAS-5 showed higher concentrations of isoflavones in other studies, which suggests this may be a genetic trait (Carrão-Panizzi et al., 1999).The highest concentrations of isoflavones were observed for cultivars IAS-5 and BRS 66, grown in most of Rio Grande do Sul (Figure 2), where average temperatures are generally lower.Cultivars grown in these locations had higher isoflavone concentrations, except for soybeans grown in Seberi and São Luiz, RS, where warmer average temperatures were observed (Table 5).
Genetic variability in isoflavone concentration is observed when different soybean cultivars are grown in the same location.Of 12 different cultivars grown in Conquista, MG, cultivars BRS Piraíba and BRS Preciosa had the highest concentrations of isoflavones, while BRS Aurora, BRS Seleta, and BRS Pirarará had the lowest (Table 6).The concentrations of isoflavone aglycones (daidzein, glycitein and genistein) were either small or undetected, as expected for soybean seeds before processing.Malonyl and some acethyl forms of the isoflavones were not found in this study, because they were transformed to glycoside forms (1) Mean values by equal letters, small in the lines and capital in the columns, do not differ the Tukey's test at 5% probability. (1)Mean values followed by equal letters, small in the lines and capital in the columns, do not differ by the Tukey's test at 5% probability.
during heat treatment, when seed samples were dried for three days at 50°C before grinding.Reduction of malonyl forms was also observed when soybean grains were soaked for 12 hours at 50°C (Carrão-Panizzi et al., 2003).
The evaluation of the soybean crops grown in Londrina, PR (23°18'S), during the soybean season of 2004/2005 showed differences in isoflavone content for crops initiated at three different planting dates (October, November and December).The filling periods for the three crops were January, February and March respectively.The average temperatures and precipitations for these three crops were 23.4°C and 376 mm, 24.2°C and 52 mm, and 24.5°C and 48 mm respectively (Brasil, 2007).All cultivars except BRS 262 yielded higher concentrations of isoflavones when sown in October (Table 7), with lower average temperatures and higher precipitation during the January seed-filling period.Adequate water supply and lower temperatures enhance isoflavone concentrations in the  Table 5.Environmental data of some of the locations in the South region of Brazil, in 2002.
(1) Average of daily minimal and maximal temperatures and (2) total precipitation in the three months (February, March and April) in 2002 (Brasil, 2007).