Components of Developing Soybean Seeds: Oil, Protein, Sugars, Starch, Organic Acids, and Amino Acids1

Abstract

A thorough chemical analysis on the composition of developing soybean [Glycine max (L.) Merr.] seeds does not appear in the literature. Several investigators have determined protein and oil, while others have determined the carbohydrates. The purpose of this study was to make a thorough analysis of the developing soybean seed, including protein, oil, sugars, starch, organic acids, and amino acids. Soybean seeds from ‘Harosoy 63’ grown in the field and ‘Steele’ grown in the greenhouse were analyzed for these components from 10 days after flowering, at 10‐day intervals, to maturity. These components accounted for 75% of the dry weight of the seed. The remaining dry weight of the seed is in cellulose, crude fiber, and ash. Oil and protein (mg/seed) increased until maturity in Harosoy 63 but leveled off before maturity in Steele. When oil and protein are expressed as percentage dry weight, similar patterns were observed for both cultivars. The sugars glucose, fructose, galactose, sucrose, raffinose, and stachyose were all detected in the soybean seed. Raffinose and stachyose were not detected until 50 and 40 days after flowering in Harosoy 63 and Steele, respectively. Starch reached a maximum value of 14.6 mg/seed in Harosoy 63 and 7.9 mg/seed in Steele at 40 and 30 days after flowering, respectively. Starch then declined sharply in both varieties. The water soluble polysaccharide (mg/seed) fraction showed a steady rise for both varieties throughout the study, but declined as percent of dry weight. The concentration of organic acids was low, with glycolic, succinic, malic, and citric being detected. Citric acid was detected in the greatest quantity. Seventeen amino acids and one unknown component were detected in the amino acid analysis. Of the amino acids detected, seven showed increases while three amino acids as well as the unknown showed decreases. Analysis of variances and standard error of the mean were performed on the data for each trait in each variety using a completely randomized design with specific tests for linear regression over the 70 or 60 days after flowering. In addition, quadratic regression was run for starch. Simple correlation coefficients were also found between all the traits studied and can be obtained upon request.