J. Nutr. Sci. Vitaminol., 38, 215-220, 1992 Note A Pitfall in Determining the Globulin/Albumin Ratio in Amaranth Grains Yotaro KONISHI,Junko AZUMAYA,Kazuko HORIKAWA, and Nobuji NAKATANI Departmentof Food and Nutrition, Osaka City University, 3-3-138 Sugimoto,Sumiyoshi-ku, Osaka 558, Japan (Received December9, 1991) Summary In amaranth grains (Amaranthus hypochondriacus L. and Amaranthus cruentus L.), there were two albumin (Alb) fractions, Alb-1 and Alb-2; Alb-1 was extracted with water or 0.5M NaCl, and Alb-2 was extracted with water after extracting Alb-1 and globulin. The amount of Alb-2 comprised about 30% of the total albumin and globulin content. Little or no measurable protein corresponding to the amaranth Alb-2 was extracted from seeds of buckwheat (Fagopyrum esculentum Moench), quinoa (Chenopodium quinoa L.), or barley (Hordeum vulgare L. cv Sekitorisai-1). That the globulin/albumin ratio (G/A) in amaranth grains has been conflicting among investigatiors probably accounts for the consideration of Alb-2 during extraction. We here report the revised G/ A ratio of 0.3-0.4. Key Words globulin/albumin ratio, Amaranthus hypochondriacus L., Amaranthus cruentus L., buckwheat (Fagopyrum esculentum Moench), quinoa (Chenopodium quinoa L.), barley (Hordeum vulgare L.) Amaranth (genus Amaranthus) that was a staple food of the Aztecs now attracts world attention because of its high agronomic and food potentials. Am aranth is a fast-growing C4 plant and produces good yield of seeds which contain about 16% protein with high biological value (1-4) In amaranth grain, albumin and globulin are major fractions and prolamins are trace in amount (5-7). How ever, the globulin/albumin ratio (G/A) in amaranth grains has been conflicting among investigators; it has been reported as 0.28 (8), 2.1(5, 9), 0.59 (10), or 0.9 0.95 (6). The reason why there is large variation might be ascribed to differences in species, cultivating conditions, sample preparation, or extraction procedures, although it is not adequately explained. We have recently observed two albumin fractions (Alb-1 and Alb-2) in amaranth grains (11). Alb-2 was extracted with water after extracting Alb-1 and Abbreviations: Alb-1, albumin-1 fraction; Alb-2, albumin-2 fraction. 215 216 Y. KONISHI et al. globulin from the defatted amaranth flour; it was estimated as about 30% of total albumin and globulin contents. Comparison of some physicochemical properties between Alb-1 and Alb-2 fractions has been previously done (11). The Alb-2 was found to be insoluble in more than 0.1M NaCl or 1mM various metal ions; Ab-1 , however, was not insoluble by these. The Alb-2 was shown to comprise 4 major proteins of 56, 36, 32, and 22 kDa by SDS-polyacrylamide gel electrophoresis (11). In this paper, we propose that the variation of the G/A ratio previously reported should be primarily ascribed to the Alb-2 content rather than the above mentioned factors. We here report the revised G/A ratio in amaranth grains . We also report that there are few or no measurable proteins corresponding to the Alb 2 in barley, buckwheat, or quinoa (the last two are taxonomically near to Ama ranthaceae). Materials and methods. 1) Samples: A. hypochondriacus (mercado grain type) and A. cruentus seeds were a generous gift from Dr. R. Becker, USDA Agricultural Research Service, CA. Buckwheat (Fagopyrum esculentum Moench) seeds were purchased locally. Quinoa (Chenopodium quinoa) was obtained at a local market in Peru. Barley (Hordeum vulgare cv. Sekitorisai-1) was a generous gift from Dr. K. Komae (National Agriculture Research Center, Tsukuba, Japan). The crude protein contents of these samples are shown in Table 1; they were determined by the micro-Kjeldahl method. Whole grains of amaranth and quinoa and dehulled buckwheat and barley kernels were ground using a mill to pass 32 mesh sieves, defatted with n-hexane at room temperature, followed by grinding to pass 60 mesh sieves, as described previously (11). 2) Fractionation and measurement of proteins: The defatted samples (0 .5g) were mixed with 5ml solvents under magnetic stirring and the protein was extracted at 4•Ž by two different procedures. In the first method (method A) , 0.5 M NaCl was used as the first solvent for 4 times extraction (60, 30, 30, 30min), water as the second (30, 30, 30min), and 0.5M NaCl as the third (30, 30, 30min) . In the second method (method B), starting from water, 0.5M NaCl followed by water. Duration and the number of extractions were similar to those in method A . All solvents contained 0.1mM phenylmethylsulfonyl flouride, a protease inhibitor , to protect proteolysis. Protein was measured by the method of Bradford (12) using bovine serum albumin as a standard. This method gives little variation in color Table 1. Crude protein content of samples. 1N•~6.25, 2N•~6.31, 3N•~5.83. J. Nutr. Sci. Vitaminol. GLOBULIN/ALBUMIN RATIO IN AMARANTH GRAIN 217 Fig. 1. Fraction of albumin and globulin from defatted flours of A.•@hypochondriacus ( A, D), A, cruentus (B, E), buckwheat (C, F), quinoa (G), and barley (H), by a sequential extraction method, method A (upper column, A-C) or method B (lower column, D-H). Open and closed columns show water and 0.5M NaCl-soluble proteins, respectively. ation among various sources of protein and show high correlation with values obtained by the Kj eldahl method (13). Results. Figure 1 shows typical patterns of protein fractionation of A. hyp ochondriacus, A. cruentus, and buckwheat, by method A and method B. By method A, a large amount of protein was extracted with the first 0.5 M NaCI (Fr. 1-4), which contained albumin (defined as Alb-1) and globulin upon dialysis against water. After removing the salt-soluble proteins, a water-soluble protein fraction (Alb-2) was obtained subsequently (Fr. 5-7). Thereafter, few or no measurable proteins were extracted with 0.5M NaCI (Fr. 8-10). On the other hand, by method B, three protein fractions were obtained by each solvent from both amaranth grains, showing that two albumin fractions were extracted separately. The last extracted albumin fraction (Fr. 8-10) showed similar electrophoretic pattern to that of the Alb-2 fraction obtained by method A, as reported previously (11). The amount of Alb-2 obtained by method A or method B was about 30% of the total albumin and globulin contents. The pattern of the protein extraction from buckwheat was quite different from that from amaranth grains (Fig. 1). Regardless of extraction method, the amount of protein decreased with the number of extractions, except that small amount of salt-soluble proteins (Fr. 5-7) appeared by method B. No proteins corresponding to the amaranth Alb-2 were detected from buckwheat, or barley (Fig. 1). For quinoa grain, however, a small amount (13% of total albumin and globulin content) of protein corresponding to the Alb-2 of amaranth grains was extracted (Fig. 1). Table 2 summarizes the albumin and globulin contents and the G/A ratio in Vol. 38, No. 2, 1992 218 Y. KONISHI et al. Table 2. Albumin and globulin content in amaranth grains. Protein was fractionated by method B (see the text) and measured by the method of Bradford. The values are shown as mean•}SD of 3 different experiments . 1 Globulin/Albumin -1+Albumin-2. amaranth grains. The total albumin and globulin contents were different between A. hypochondriacus and A. cruentus, which was mostly due to the Alb-1 content (see also Fig. 1D and E). The contents of globulin and Alb-2, however, did not differ between the two samples. From these values, the G/A (G/Alb-1+Alb-2) ratio was calculated as 0.41 for A. hypochondriacus and 0.31 for A, cruentus. Discussion. The Alb-2 specifically found in amaranth grains contributes to determine the G/A ratio. Alb-2 has not so far been extracted from amaranth grains , since it shows unique behavior upon extraction. Indeed, it was recovered after extracting Alb-1 and globulin from the grains, irrespective of species .Furthermore, the Alb-2 was not easily extracted upon substituting 0.5M NaCl by water, since it was found to be salt-insoluble albumin (11). These results and our previous findings that Alb-2, but not Alb-1, was protected from pronase attack when amaranth flour was treated with pronase before protein fractionation (11) , sugge sted the possibility that Alb-2, like polymeric globulin (5,14), is associated with protein bodies (2,15). Bressani and Garcia (6) reported the G/A ratio of 0.9-0.95 of several ama ranth grains, by twice extraction with water, followed by twice extraction with 0 .5 M NaCl, without the third extraction with water as we have done in this study . Under these conditions, however, Alb-2 fraction would be little extracted, while it would be recovered in the alkali-soluble fractions (glutelins) subsequently . In fact, these authors obtained high glutelin content (6). Another pitfall on determining the G/A ratio may be ascribed to an apparent pI5-6 of Alb-2 (11), which was similar to the p1 value of amaranth globulin (5). In our previous report (5), albumin and globulin were extracted with 0.5M NaCl for 4 times and then with water twice according to the method of Landry and Moureaux (16), combined these proteins, and dialyzed against distilled water to separate albumin from globulin. As a result, we obtained the G/A ratio of 2.1, which was quite in agreement with the value obtained by Gorinstein et al . (9). However, these procedures would precipitate Alb-2 together with globulin , since J.