HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Geater, C. W. Articles by Wilson, L. A. Search for Related Content PubMed Articles by Geater, C. W. Articles by Wilson, L. A. Agricola Articles by Geater, C. W. Articles by Wilson, L. A.

CROP BREEDING, GENETICS & CYTOLOGY

Association of Soybean Seed Traits with Physical Properties of Natto

^a Dep. of Agronomy, Ames, IA USA
^b Dep. of Food Science and Human Nutrition, Iowa State Univ., Ames, IA 50011 USA

wfehr{at}iastate.edu

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 
Breeding of soybean [Glycine max (L.) Merr.] cultivars suitable for natto involves selection for seed traits that influence the quality of the product. The objectives of this study were to evaluate the relationship of 11 seed traits to six natto quality traits and to assess the influence of genotypes and environment on the traits. Sixteen small-seeded genotypes were grown in replicated tests at three Iowa locations during 2 yr. The natto traits evaluated for each plot were water absorption, water loss, hardness of steamed seeds and natto, and darkness of steamed seeds and natto. Seed traits were total sugar, free sugar, sucrose, raffinose, stachyose, protein, oil, fiber, protein + oil, protein + oil + fiber, and seed size. There were significant differences among genotypes during one or both years for all the traits, except fiber content. Significant differences between years or among locations were observed for all traits, except water loss after steaming and stachyose content. None of the seed traits was significantly correlated with darkness of the steamed seeds or natto. All of the seed traits, except stachyose, oil, and seed size, were significantly correlated with one or more of the other natto quality traits. Protein + oil was significantly correlated with natto quality and other seed traits and can be a useful selection criterion in breeding cultivars for the natto industry.

Abbreviations: CORR, correlation procedure • F, fiber • GLM, general linear model • HPLC, high performance liquid chromatography • NIR, near-infrared reflectance • O, oil • P, protein • P + O, protein + oil • P + O + F, protein + oil + fiber

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 
NATTO IS A FERMENTED SOYBEAN PRODUCT consumed primarily in Japan (Taira, 1990). A seed size of 80 mg seed^-1 is commonly used for natto production and is a primary selection criterion in breeding cultivars for that market. Research on the composition of the seed that is most suitable for natto has been limited, which makes it difficult for the breeder to define the limits of acceptability for seed traits other than size.

The traits measured during the production of natto to assess the acceptability of a small-seeded cultivar are water absorption of the seeds, water loss of the seeds after steaming, hardness and darkness of the steamed seeds, hardness and darkness of the seeds in the natto after fermentation, and sensory evaluation of the natto for texture and flavor (H. Hasegawa, 1996, personal communication). High water absorption and low water loss are desired by natto manufacturers. Natto can be excessively hard, soft, dark, or light. The standards for hardness and color are determined by natto manufacturers on the basis of their perception of consumer preference. Taira (1990) reviewed studies she had conducted on the relationship of seed traits to the natto quality traits. She indicated that high water absorption of the seed was required to obtain soft steamed seeds. She considered relative amounts of the free sugars, sucrose, raffinose, and stachyose important to achieve the proper rate of fermentation. Taira (1990) concluded that the quality of natto was primarily associated with the seed traits of cultivars.

Significant differences among small-seeded genotypes for water absorption were reported by Cober et al. (1997). The year x location interaction component also was significant for water absorption because the relative differences among five locations in Canada were not consistent during 2 yr. No studies have been reported on the evaluation of genotype and environmental effects on the other natto quality traits.

Evaluation of natto quality traits is time consuming and would not be practical for selection among hundreds of genotypes in a cultivar development program. Indirect selection for natto quality based on seed traits that are measured more readily would be desirable. Protein, oil, and fiber can be evaluated readily by near-infrared reflectance (NIR) and would be desirable traits for indirect selection if they are correlated with the natto quality traits. Other seed traits that could be considered for indirect selection would be total sugar as determined by acid hydrolysis and the free or soluble sugars of sucrose, raffinose, and stachyose. The objectives of this study were to evaluate genotype and environmental effects on natto quality traits and to evaluate the relationship of seed traits to natto traits.

	Materials and methods

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 
Sixteen small-seeded cultivars and lines developed by Iowa State University were selected for the study. They were grown in a randomized complete-block design with two replications at Ames, Fairfield, and Stuart, IA, during 1995 and 1996. The soil types were a Nicollet loam (fine-loamy, mixed, superactive, mesic Aquic Hapludoll) at Ames, a Haig silty clay loam (fine, smectitic, mesic Vertic Argiaquoll) at Fairfield, and a Macksburg silty clay loam (fine, smectitic, mesic Aquic Argiudoll) at Stuart. In 1995, the entries were planted in two-row plots 2.8 m long with a row spacing of 69 cm between rows within the plot and 102 cm between rows of adjacent plots. The seeding rate was 33 seeds m^-1. In 1996, the entries were planted in four-row plots 3.7 m long with a row spacing of 69 cm and a seeding rate of 33 seeds m^-1. The center two rows of each plot were harvested. The harvested seed from both years was stored at 2°C at a relative humidity of 40% until seed analysis could be performed.

For analysis of the natto quality and seed traits, plots from each replication were tested consecutively from the lowest to the highest plot number as assigned for the randomized complete-block design in the field tests. A seed sample from each plot was processed into natto by a procedure obtained from the Ibaraki Industrial Technology Center, Mito City, Japan (H. Hasegawa, 1996, personal communication). A 50-g sample of soybean seed from each plot was soaked in water for 15 h at room temperature, during which the weight increased to greater than 220% of the original weight. The seeds were drained in a standard kitchen strainer for 1 min and weighed. The percentage of water absorption was calculated by dividing the weight of the seeds after soaking by the weight of the seeds before soaking and multiplying by 100.

After the soaked seeds were weighed, the samples were placed in a Dixie Canner Equipment Company (Athens, GA) Model RDTI-3 Retort and steamed at 131°C for 35 min. The samples were removed and cooled for 30 min to room temperature, then weighed to determine the amount of water loss after steaming. Water loss percentage was calculated by dividing the weight of each sample after steaming by the weight of the sample before steaming, multiplying by 100, and subtracting the product from 100.

The mean hardness of 10 individual seeds was determined after the samples had been weighed to determine water loss. Hardness was measured with an Instron Corporation (Canton, MA) Model 1122 Universal Testing Machine and expressed as the maximum force (g) required to break the seed. A seed was placed horizontally on a stationary bed below a probe with a cutting width of 0.5 mm attached to a 50-kg compression load cell. The probe was lowered through the seed at a constant speed of 10 cm min^-1 until the pressure was sufficient to break it. The mean hardness of the 10 seeds was used for data analysis.

The mean darkness was measured for five seeds of each sample after water loss had been determined with a Hunter Associates Laboratory, Inc. (Reston, VA), Labscan 6100 Spectrocolorimeter. The instrument was standardized with white and black tiles, Fcw illuminant, 10° standard observer, and 0.635-cm port size. Darkness was expressed on a Hunter L scale from 0 to 100, where 0 was black and 100 was white. Each sample was placed between two sheets of plastic cling wrap and pressed with a 600-mL glass beaker to flatten it. The mean L value of five seeds was used for data analysis.

After steaming, a separate 50-g sample of steamed seeds from each plot was inoculated with 0.5 mL of Bacillus natto starter culture provided by the Ibaraki Industrial Technology Institute, Mito City, Japan, packaged into a polystyrene natto container provided by Asaichiban Co., Ltd., Tsuchiura City, Japan, and covered with a perforated plastic sheet provided by Asaichiban Co., Ltd. The containers were closed and placed in a Controlled Environments LTD. (Winnipeg, Manitoba) Model E15 Growth Chamber for 18 h at a temperature of 39°C and a relative humidity of 90%. Following fermentation, the samples were refrigerated at 5°C for 24 h. Each sample was evaluated for hardness and darkness by the same procedure described for evaluation of the two traits for steamed seeds before fermentation.

Total sugar content was measured by the method described by Geater et al. (2001). For analyses of total sugar content, 10 g of seed from each plot was ground with a Cyclotec 1093 Sample Mill (Tecator Inc., Herndon, VA). A 150-mg sample of the powder was placed in a 16- by 125-mm screw cap tube. A 10-mL aliquot of distilled water and 1 mL of 25% (w/w) HCl were added to each tube. The tubes were capped with rubber-lined phenolic screw caps, vortexed at a speed setting of 4 for 5 s with a Fisher Scientific (Pittsburgh, PA) Vortex Genie 2 Mixer, and placed in an autoclave at 121°C for 20 min. The tubes were removed from the autoclave and cooled to room temperature in a 1°C water bath for 5 min. A 275-µL aliquot of 40% (w/v) NaOH was added to each tube. The tubes were capped, and the solution was mixed by inverting the tubes five times. The solution in each tube was emptied into a 500-mL volumetric flask. The flask was filled to 500 mL with distilled water, capped with a polyethylene snap cap, and mixed by inverting 25 times. Total sugar content was determined by the modified phenol-sulfuric acid method of Fox and Robyt (1990) and was expressed on a moisture-free basis.

Sucrose, raffinose, and stachyose contents were determined by high performance liquid chromatography (HPLC) by Pioneer Hi-Bred International, Inc. (Johnston, IA) with a method developed by the Dionex Corporation (Sunnyvale, CA). For each plot, 10 g of seed was ground for 1 min with a Stur-Dee Health Products (Oakdale, NY) Mitey-Mill. A 100-mg sample of the powder was placed into a 16- by 125-mm screw cap tube. To each tube, 5.0 mL of melezitose internal standard purchased from Sigma Aldrich (St. Louis, MO) was added. The melezitose internal standard was prepared in a ratio of 1.0 mg of D(+)melezitose to 1.0 mL of distilled water. The tubes were vortexed at a speed setting of 4 for 5 s with a Fisher Scientific Vortex Genie 2 Mixer. To each tube, 5.7 mL of HPLC-grade chloroform was added. The tubes were capped with rubber-lined phenolic screw caps, vortexed for 5 s, and incubated at room temperature for 4 h. A 100-µL aliquot of the upper layer of the tube was placed in a 16- by 125-mm screw cap tube, and 9.9 mL of distilled water was added. The tubes were capped with rubber-lined phenolic screw caps and mixed by shaking for 5 s by hand. A 0.5-mL aliquot of the test solution was placed in a Dionex 0.5-mL vial (P/N 038010) and capped with a Dionex 0.5 mL-filter cap (P/N 03011). The samples were stored at 5°C before analysis. HPLC analysis was performed with a Dionex Corporation DX500 High Performance Liquid Chromatographer with a Dionex Carbo Pac PA-1 4 by 250-mm column and a Dionex Carbo Pac PA-1 guard column with a 70:30 water to 600 mmol sodium hydroxide mobile phase.

Seed moisture, protein, oil, and fiber content were measured on a 100-g bulk sample of seed from each plot with a Tecator A/B Infratech 1221 whole-grain NIR analyzer. Protein, oil, and fiber contents were expressed on a moisture-free basis. Seed size was based on the weight of a random sample of 200 seeds.

The data were analyzed as a randomized complete-block design. Years, locations, and replications were considered random effects, and genotypes were considered fixed effects. Analyses of variance were conducted for each year combined across locations, for each location combined across years, and for the combined locations and years. The analyses of variance were performed with the general linear models procedure (GLM) of SAS (SAS Institute, 1992). Phenotypic correlation coefficients among the traits were calculated for genotype means averaged across years and locations with the correlation procedure (CORR) of SAS (SAS Institute, 1992).

	Results and discussion

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 
The significance of the main effects of years and locations for the natto and seed traits was dependent on the significance of the year x location interactions in the combined analysis of variance across years and locations (Table 1) . No significant differences were found between years or among locations when the year x location interaction was significant for a trait. To evaluate more fully year and location effects, analyses of variance were conducted independently for each location across the 2 yr and for each year across the three locations (Table 1). In those analyses, significant year or location effects were observed for all of the traits, except water loss after steaming and stachyose content. Cober et al. (1997) reported significant year x location interactions for water absorption, oil, and free sugar for two sets of small-seeded genotypes evaluated in Canada. Their year and location effects were not significant when the year x location interactions were significant, except for sugar content in one set of genotypes. They did not analyze the effect of years at individual locations or the effect of locations for individual years. Geater and Fehr (2000) observed significant differences among eight Iowa locations during 1 yr for the protein (P), oil (O), fiber (F), P + O, and P + O + F of 23 general-use and food-grade cultivars. Taira (1990) concluded from studies in Japan that the quality of soybeans for natto and other Japanese processed food was primarily influenced by cultivars instead of environmental conditions. The relative importance of genotype and environmental effects on natto quality and seed traits will depend on the genotypes and environments included in the study.

There were significant differences among locations and genotypes for hardness of the natto (Tables 1 and 2). The range among environments was 35 to 44 g. No significant interactions of genotypes with years and locations were found, the same as observed for steamed hardness. A significant positive correlation of 0.75 was observed between steamed and natto hardness of the 16 genotypes (Table 3). Steamed hardness would be useful for predicting genotypic differences for natto hardness, but the prediction has limitations. For example, IA2023 had the highest rank for steamed and natto hardness; on the other hand, IA3007 ranked 7 for steamed hardness and 16 for natto hardness (Table 2).

Darkness of the steamed seeds was significantly influenced by locations and genotypes (Tables 1 and 2). The range in L values among environments was 41.6 to 43.1. There were no significant interactions of genotypes with years or locations. Steamed darkness was not significantly associated with water absorption, water loss, steamed hardness, or natto hardness (Table 3).

There were significant differences among locations and genotypes for natto darkness (Tables 1 and 2). The range in L values among environments was 40.5 to 41.8. The interactions of genotypes with years and locations were not significant. Natto darkness was significantly correlated with steamed darkness (0.52), but not the other natto quality traits (Table 3). A95-686035 ranked 16 for darkness of both the steamed seeds and the natto (Table 2). On the other hand, IA2023 ranked 1 for darkness of the steamed seeds but 10 for darkness of the natto. Color of the steamed seeds can be a useful predictor of natto darkness, as long as the limitations of the relationship are considered during selection in a breeding program.

Significant differences among locations and genotypes were observed for total sugar (Table 1 and 4) . The range among environments was 209 to 248 g kg^-1. The differences among the genotypes were consistent among environments as indicated by the lack of significant genotype interactions with years and locations. The lack of genotype x environment interactions would facilitate selection for total sugar among lines in a breeding program or among cultivars for commercial soybean production. Total sugar was significantly correlated (–0.62) with natto hardness, but not with the other natto quality traits (Table 5) . A95-686001 ranked 1 for total sugar and 2 for natto darkness, but IA2024 ranked 13 for total sugar and 2 for natto darkness (Table 2 and 4).

There were significant negative associations of protein, P + O, and P + O + F contents with total sugar, free sugar, sucrose, and raffinose contents (Table 6). The negative correlation coefficients of protein with total sugar, free sugar, and sucrose were similar to those for oil. Hymowitz et al. (1972) observed a significant negative correlation of protein with sucrose (-0.38), a significant positive correlation with stachyose (0.41), and no significant correlations with free sugar (-0.19) or raffinose (-0.24). They reported significant positive correlations of oil with free sugar (0.26), sucrose (0.42), and raffinose (0.36), and a significant negative correlation with stachyose (-0.41). Openshaw and Hadley (1981) reported significant negative correlations of free sugar with protein (-0.59 and -0.50) and protein + oil (-0.39 and -0.32), and significant positive associations of free sugar with oil (0.40 and 0.21) in two F₃ populations. Hartwig et al. (1997) reported a significant negative correlation of protein with raffinose (-0.26) and sucrose (-0.78), and a significant positive correlation (0.67) between oil and sucrose for 40 general-use genotypes. These associations have important implications for determining the seed components that should be measured in a breeding program. All the sugar measurements are more time consuming and expensive than analysis of protein, oil, and fiber by NIR. It would be desirable if indirect selection for sugar could be conducted by selection for protein, oil, or some combination of the three traits. In our study, a major percentage of the variation in total sugar, free sugar, and sucrose was accounted for by the variation in protein, P + O, and P + O + F, which indicated that any of the three traits could be used effectively for indirect selection. In the studies of Hymowitz et al. (1972) and Geater and Fehr (2000), the correlations of protein with sugar traits were not sufficiently large to make indirect selection effective. Geater and Fehr (2000) found that the correlation of P + O + F with total sugar (-0.69) was influenced by seed size and would not be as effective for indirect selection as P + O, which was not influenced by seed size and had a correlation of –0.81 with total sugar. The consistently high correlation between P + O and total sugar suggested that it should be possible to use these NIR measurements to identify genotypes that are most likely to have the desired sugar contents, which would appreciably reduce the number of sugar analyses.

Significant differences were observed among years, locations, and genotypes for seed size (Table 1 and 4). The range among environments was 59 to 79 mg sd^-1. Seed size of the genotypes was not significantly correlated with any of the natto or seed traits (Tables 5 and 6). Selection for genotypes within the size range for this study would not be expected to influence the natto or seed traits.

The four natto quality characteristics influenced by seed traits were water absorption, water loss, and hardness of the steamed soybeans and natto. None of the seed traits had an influence on the darkness of the steamed soybeans or natto.

The data from the study were reviewed by two soybean manufacturers in Japan whose names are withheld at their request. The manufacturers indicated that the range for all the traits among genotypes, years, and locations was acceptable for natto production. They indicated that adjustments in their manufacturing processes could be made to accommodate the variation observed in this study.

The study did not address the relationship of seed traits to flavor and other sensory attributes of natto. That research would require collaboration with food scientists who are qualified in making and evaluating natto quality traits expected by consumers.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 
Journal Paper no. J-18746 of the Iowa Agric. and Home Econ. Exp. Stn. Project no. 3107, and supported by the Hatch Act, State of Iowa, and Iowa Soybean Promotion Board.

Received for publication February 3, 2000.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion REFERENCES

 

• Cober E.R., Frégeau-Reid J.A., Pietrzak L.N., McElroy A.R., Voldeng H.D. Genotype and environmental effects on natto soybean quality traits. Crop Sci. 1997;37:1151-1154.[Abstract/Free Full Text]
• Fox J.D., Robyt J.F. Miniaturization of three carbohydrate analyses using a microsample plate reader. Anal. Biochem. 1990;195:93-96.
• Geater C.W., Fehr W.R. Association of total sugar content with other seed traits of diverse soybean cultivars. Crop Sci. 2000;40:1552-1555 (this issue).[Abstract/Free Full Text]
• Geater C.W., Fehr W.R., Wilson L.A., Robyt J.F. An improved method of total sugar analysis for soybean seed. Crop Sci. 2001;in press.
• Hartwig E.E., Kuo T.M., Kenty M.M. Seed protein and its relationship to soluble sugars in soybean. Crop Sci. 1997;37:770-773.[Abstract/Free Full Text]
• Hymowitz T., Collins F.I., Panczner J., Walker W.M. Relationship between the content of oil, protein, and sugar in soybean seed. Agron. J. 1972;64:613-616.[Abstract/Free Full Text]
• Openshaw S.J., Hadley H.H. Selection to modify sugar content of soybean seeds. Crop Sci. 1981;21:805-808.[Abstract/Free Full Text]
• SAS Institute. The SAS system for Windows. Release 6.12. Cary, NC: SAS Inst, 1992.
• Taira H. Quality of soybeans for processed foods in Japan. Japan Agric. Res. Quart. 1990;24:224-230.

This article has been cited by other articles:

				B. Zhang, P. Chen, C. Y. Chen, D. Wang, A. Shi, A. Hou, and T. Ishibashi Quantitative Trait Loci Mapping of Seed Hardness in Soybean Crop Sci., July 1, 2008; 48(4): 1341 - 1349. [Abstract] [Full Text] [PDF]

				M. S. Cicek, P. Chen, M. A. Saghai Maroof, and G. R. Buss Interrelationships among Agronomic and Seed Quality Traits in an Interspecific Soybean Recombinant Inbred Population Crop Sci., April 25, 2006; 46(3): 1253 - 1259. [Abstract] [Full Text] [PDF]

				C. W. Geater, W. R. Fehr, L. A. Wilson, and J. F. Robyt A more rapid method of total sugar analysis for soybean seed Crop Sci., January 1, 2001; 41(1): 250 - 252. [Abstract] [Full Text] [PDF]

				C. W. Geater and W. R. Fehr Association of Total Sugar Content with Other Seed Traits of Diverse Soybean Cultivars Crop Sci., November 1, 2000; 40(6): 1552 - 1555. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (4) Citing Articles via Google Scholar Google Scholar Articles by Geater, C. W. Articles by Wilson, L. A. Search for Related Content PubMed Articles by Geater, C. W. Articles by Wilson, L. A. Agricola Articles by Geater, C. W. Articles by Wilson, L. A.