Composition and Protein Quality of Sweet Lupin Seed

J. Sci. Fd Agric. 1974, 25, 851859 Composition and Protein Quality of Sweet Lupin Seed Edwin L. Hove Applied Biochemistry Division, DSIR, Palmerston North, New Zealand (Manuscrigt received 28 January I974 and accepted 7 March 1974) The proportion of seed coats (hulls) of four varieties of sweet lupin seeds ranged from 19 to 25 %. The amount of hull varied inversely with the weight of the seed, within each variety. Lupinus luteus cv Weiko I11 had 40.1 % protein (dry basis); of 3 cultivars of L. angustifolius cv Uniharvest had the highest protein (36.2 %), while cv Uniwhite had the lowest (28.6%). The protein content of dehulled lupins was about 20% higher than that of the whole seeds. L. angustifolius cv Uniwhite and L. luteus cv Weiko I11 were analysed for moisture, protein, lipid, ash, fibre, amino acids, carbohydrates, calcium, phosphorus, zinc, iron, copper and manganese. Whole seed, hulls and kernels were examined. Sulphur amino acids were limiting. The protein efficiency ratios of the two varieties, supplemented with methionine, were 2.40 and 2.56, respectively; these values increased by about 10% when the protein of the dehulled lupin seeds were tested. The requirement for added methionine for maximum p.e.r. was 0.65 g/kg diet. When dehulled lupin seed was steeped in water for a day and boiled and included in the diet of rats, slightly superior growth and p.e.r. resulted. The two varieties of dehulled lupin seeds effectively supplemented the protein of barley meal in the diet of rats, especially when methionine supplements were added, but were less efficient in supplemental value than was casein. 1. Introduction Lupin is a legume whose use has been limited largely to green manure and green fodder for ruminants because of the presence of toxic and bitter alkaloids. However, genetic selections by plant breeders have resulted in a number of sweet varieties with seeds of improved nutritional value. A descriptive history and present status of lupins has been published by Gladstones.' Lupinus luteus, cultivar (cv) Weiko I11 was approved for cropping in Germany in 1951. L. angustifolius, cv Uniwhite was released in Western Australia in 1967, followed in 1971 by cv Uniharvest and last year by cv Unicrop. The protein content of different varieties of lupin seeds range from 25 to 44% of dry matter.' The use of lupin seed as a protein supplement in mixed feed for pigs and poultry is beginning to develop in Australia and New Zealand, and may be able to replace imported soya bean meal for most uses. The protein efficiency ratio (p.e.r.) for lupin seed protein as listed in an F.A.O. publication3 is very low (0.57), even when supplemented with methionine (1.06), as compared with a standard casein p.e.r. of 3.05. The quoted data are those of Tannous and Cowan4 who tested samples of bitter lupin from the Mediterannean region, after debittering by a process of prolonged steeping and boiling. The p.e.r. values were 851

852 E. L. HoFe much lower than predictable from the amino acid pattern3 Perhaps residual toxins remained in the samples, or it may be that the process of debittering damaged the protein. An objective of the present paper is to report on p.e.r. values of some of the newer sweet lupin strains, and to determine whether the procedure for debittering does lessen the nutritional value of the protein. The lupin seed coat, or hull, makes up 20 to 25% of the weight of the seed and is very low in protein and high in fibre. Its chemical composition and digestibility in the diet of monogastric animals such as the rat are described el~ewhere.~ A correlation between the size of the seed and the proportion of seed coat for different varieties is reported in the present paper. 2.1. Lupin seed samples 2. Experimental The Agronomy Department of Massey University provided samples of Lupinus luteus, cv Weiko 111, and L. angustifolius, cv Uniwhite, cv Uniharvest and cv Unicrop, from their experimental plots designed to establish comparative crop yield data. To establish the relation between the size of the seed and the relative amount of hull, SO seeds of the four cultivars were randomly picked from 500 g samples, weighed individually and hand dissected to get the weight of the parts. The correlation coefficients (r) between the dry matter of seed and hull and their linear regression equations, were calculated by the procedures listed in Snedecor.6 Six other samples of L. angustifolius cv Uniwhite whose planting times were at fortnightly intervals were analysed. A larger quantity (SO kg) of Uniwhite was obtained from a local feed company who also dehulled half of the sample to remove the seed coatings cleanly and yield the kernels in hemispheres. To simulate the debittering process4 3 kg of the dehulled Uniwhite kernels were mixed with 12 1 water, and soaked with mixing at 15 "C for 24 h. A third of the sample was freezedried, another third was ovendried at 107 "C for 12 h, and the remaining third was boiled for 2 h and then freezedried. The dried samples were friable and easily ground to 20 mesh. 2.2. Analytical procedures Moisture, ash, crude protein (microkjeldah1 N times 6.25), crude lipid (ethyl ether extract) and crude fibre were determined by AOAC method^.^ Individual carbohydrates were determined as outlined by Bailey.5 Amino acids were determined by column chromatography in a Beckman 120C amino acid analyser on samples hydrolysed overnight in 6 MHCI. Mineral elements were determined by atomic absorption spectroscopy as described in the handbook supplied with the PerkinElmer instrument. 2.3. Protein quality by rat assay Protein efficiency ratios (p.e.r.) were estimated by the AOAC rneth~d.~ Male weanling albino rats, individually caged in groups of 5 or 10 and weighing 45 to 55 g, were given the test diets for 4 weeks. Food and water were continuously available. The gain in body weight divided by the weight of the protein in the food consumed gave values for p.e.r. Room temperature was 21 1 "C.

Protein quality of lupin seed 853 The test diets contained ground lupin to supply 100 to 115 g of crude protein (total N x 6.25)/kg diet. The remainder of the diet was 50 g of mineral mix, 50 g of vitamin mix, SO g of maize oil and wheat starch to make up the kg. Composition of the mineral and vitamin mixtures were described previ~usly.~ A casein control group was included with each experiment. In one experiment, increasing amounts of DLmethionine were added to a basal diet containing 100 g of lupin protein (N x S.5)/kg; the purpose was to determine the minimum amount of added methionine needed to give maximum p.e.r. Dehulled lupin cv Uniwhite was used, with five ratslgroup. 2.4. Supplemental value of lupin protein A basal diet that contained 650 g barley meal/kg (equivalent to 60 g crude protein/kg) was supplemented with ground dehulled lupin, cv Uniwhite and cv Weiko 111, or with casein, to supply an additional 40, 60 or 80 g protein/kg. Total crude protein levels were, therefore, 100, 120 and 140 g/kg diet. The series was repeated with the addition of 3 g of DLmethionine/kg diet. Four weanling male rats were used for each treatment, and comparisons between groups were evaluated by the Student t test for significance. 3. Results 3.1. Relation of seed coat to whole seed weights Within varieties, larger lupin seeds had lesser proportions of seed coats (Table 1). The negative correlation was highly significant statistically, with t values well in excess of 6 for 48 degrees of freedom. The seed of L. luteus cv Weiko I11 was significantly smaller than the others, although the proportion of seed coat was only slightly larger than that of Uniwhite. The Weiko seed was relatively flat, whereas the others were round. The dry matter of kernels and hulls were uniform and all fell within the range of 87.6 to 89.0 %. The crude protein of the samples are given in Table 1. TABLE 1. Comparison of four cultivars of lupin seed as to crude protein content, correlation coefficients between seed weight and % of seed coat and the linear regression equation between seed weight and the % seed coat (dry weight basis) Variety and cultivar Protein N x 6.25 ( %) (XI Mean g weight of seed (Y) Mean % seed coat (X, Y) Correlation coefficient Linear regression (r) equation L. angustifolius Uniwhite Uniharvest Unicrop L. luteus Weiko I11 28.6 36.2 32.4 40.0 0.162 0.169 0.206 0.122 23.2 21.7 19.7 23.7 0.72 0.88 0.62 0.73 Y = 0.028X + 25.2 Y = 0.035X + 22.5 Y = 0.08OX + 20.1 Y = 0.021X + 24.7 a 48 degrees of freedom, P < 0.001 for r = 0.6.

~ 854 E. L. Hove 3.2. Chemical composition of lupin seed Proximate analyses were carried out on samples of Weiko I11 and Uniwhite, and Table 2 lists the values. Weiko I11 had substantially higher protein and somewhat more ash and crude fibre than Uniwhite. Lupins contained almost no starch and little lipid. The complex carbohydrates varied in concentration with the variety but no lignins were present. Kernels had less calcium and more phosphorous than the whole seed. The seed coats were unique in having no detectable phosphorous. Concentrations of trace elements were sufficiently high to make lupins a valuable contributor of these essential nutrients. TABLE 2. Composition of two varieties of lupin seed, seed coats and kernels, as to proximate analyses, mineral and trace element contents, and individual carbohydrates of kernels L. angustifolius cv Uniwhite L. luteus cv Weiko 111 Component Seed Seed coat Kernel Seed Seed coat KerneI Crude protein 29.6 & 0.17 2.1 38.1 40.1 3.O 50.5 Ether extract 6.5 k 0.50 1.o 7.2 5.2 1.o 6.1 Ash 3.5 * 0.12 3.1 3.7 4.0 2.2 4.4 Crude fibre 16.8 k 0.20 54.9 3.5 17.7 56.5 4.4 Minerals Calcium 0.25 0.77 0.13 0.27 0.56 0.20 Phosphorus 0.46 0 0.53 0.48 0 0.54 Magnesium 0.15 0.09 0.16 0.21 0.15 0.24 Iron (parts/million) 55 65 71 73 77 87 Zinc (parts/million) 38 37 38 54 20 68 Manganese (parts/million) 40 17 37 115 68 152 Copper (parts/million) 6.6 8.9 7.2 Carbohydrates Starch 1.5 1.5 Cellulose 1.3 1.8 Hemicellulose 3.1 6.0 Pentosans 12.5 3.6 Soluble sugars 7.2 5.6 Pectins 4.9 1.3 Values are means of six separately grown samples of Uniwhite seed with S.E.M. and duplicates of all other samples, expressed as % dry weight, except where noted. (Moisture of all samples were within the range of 10.9 to 12.6%.) The amino acid contents listed in Table 3 indicate low levels of the sulphur amino acids, but otherwise the amount and balance of the essential amino acids seem adeq~ate.~ The protein of hulls had relatively higher proportions of lysine, threonine and valine, as compared with the values for kernels. The fatty acid composition of lipids in seed, seed coat and kernel have been previously reported.l0 A remarkable similarity to the fatty acid composition of soy bean was noted. 3.3. P.e.r. of lupin protein With added methionine the p.e.r. of lupin protein was about 2.70. In Table 4 the data for lupin protein was significantly less than the p.e.r. of casein (P < 0.05), whereas in Table 5 there was no difference between the p.e.r. of lupin and casein. The data in

Protein quality of lupin seed 855 TABLE 3. Amino acid composition (mg/g N) of lupin seed, seed coats and kernels, expressed as means f S.E.M. of six separately grown Uniwhite seeds samples, and duplicates of other samples L. angustiyolius cv Uniwhite L. luteus f cv Weiko I11 Amino acid Seed Seed coat Kernels Kernels Lysine Histidine Threonine Valine Isoleucine Leucine 3Cystine Methionine Tyrosine Phenylalanine Arginine Aspartate Glutamate Serine Proline Glycine Alanine Ammonia 309 f 12.9 163 f 7.8 214+ 6.8 246 k 9.0 268+ 7.7 453 k 12.1 54f 3.8 38 f 3.4 188 f 11.2 257 & 8.9 629 f 36.9 648 & 15.9 1440 f 50.1 368 k 8.0 305 k 7.3 264 k 7.0 226f 6.7 87 + 2.5 590 153 277 329 275 449 n.d. 45 153 274 321 649 846 405 317 309 300 81 286 167 21 6 245 279 471 79 38 181 241 642 63 1 1470 326 297 239 208 92 395 205 237 263 300 661 125 50 163 292 775 732 1850 359 315 277 233 120 TABLE 4. Protein efficiency ratios (p.e.r.) of Lupinus luteus cv Weiko 111 and L. ungustifolius cv Uniwhite at 110 g protein (N x 6.25)/kg diet, and casein at 97 g protein/kg diet DLmethionine p.e.r. added p.e.r. of dehulled Protein source (g/kg diet) of seed seed kernel Uniwhite lupin 0 1.29 f 0.08 1.43 k 0.16 Weiko I11 lupin 0 1.39 f 0.06 1.83 f 0.04 Uniwhite lupin 3 2.40 & 0.18 2.65 k 0.05 Weiko I11 lupin 3 2.56 k 0.09 2.78 f 0.05 Casein control 0 2.90 f 0.04 Means with S.E.M. of groups of five male weanling rats, 4 weeks. Table 4 indicate a consistently higher p.e.r. for cv Weiko I11 as compared with cv Uniwhite, but the difference was not statistically significant. The data in this Table also reveal a slight but consistent superiority of dehulled lupin seed over the whole seed, for both varieties. The effect of processing dehulled lupin seed cv Uniwhite is shown in Table 5. When steeped in water for a day the lupin gave significantly better growth but did not improve the p.e.r. However, when in addition the lupin was boiled for 2 h, not only was the growth rate improved but a significantly (P < 0.05) better p.e.r. resulted. Oven drying at 107 "C for 12 h damaged the protein severely.

856 E. L. Hove TABLE 5. Effect of steeping, boiling and drying on the p.e.r. of dehulled lupin (L. angustifolius cv Uniwhite) Protein Gain in Protein p.e.r. (N x 6.25) weight eaten (g gain/g protein eaten) Treatment (g/kg) (g) (g) N x 6.25 N x 5.5" Raw ground lupin kernels 113 I19 f 6.2 43.8 2.70 f 0.03 3.09 Steeped 24 h, freezedried 115 138 & 5.3 51.0 2.68 k 0.04 3.06 Steeped 24 h, boiled 2 h 110 132+ 1.9 47.0 2.82 f 0.06 3.23 Steeped 24 h, ovendry 12 h 115 73 f 2.9 36.8 1.92 k 0.03 2.20 Casein control 97 94 f 4.6 34.4 2.78 f 0.08 2.71 Except for casein where the conventional factor of 6.38 was used. Values are means with S.E.M. for 10 male weanling rats/group given the diets with 2 to 4 g DLmethionine added/kg for 4 weeks. TABLE 6. Supplemental value of protein of dehulled L. angustifolius cv Uniwhite, dehulled L. lureus cv Weiko 111 and casein, added at three levels to a basal diet containing 650 g of barley meal/kg g Protein/kg diet Supplemental protein From From, supplemental added DL Uniwhite Weiko 111 Standard Total protein methionine lupin lupin casein 100 120 140 103 123 143 100 120 140 103 123 143 40 60 80 40 60 80 40 60 80 40 60 80 Mean growth rate of rats (g/day) 4.1 (0.39) 3.9 (0.22) 5.2 (0.63) 4.1 (0.26) 3.7 (0.33) 6.5 (0.31) 4.7 (0.31) 5.5 (0.48) 7.3 (0.60) 5.3 (0.43) 5.0 (0.36) 6.1 (0.49) 6.4 (0.46) 6.4 (0.50) 7.3 (0.58) 7.8 (0.53) 7.5 (0.50) 7.3 (0.58) Mean p.e.r. for rats with S.E.M. in parens 2.07 (0.10) 2.26 (0.07) 2.62 (0.12) 2.08 (0.06) 2.33 (0.13) 2.68 (0.09) I.90 (0.11) 2.34 (0.08) 2.80 (0.03) 2.47 (0.07) 2.30 (0.03) 2.67 (0.09) 2.41 (0.12) 2.44 (0.08) 2.49 (0.10) 2.39 (0.06) 2.29 (0.07) 2.09 (0.05) Values are means with S.E.M. for groups of four male weanling rats fed the diets for 2 weeks. The data on p.e.r. for lupin protein were obtained using the factor of 6.25 to convert N to protein values. If the more nearly correct factor of 5.5 was used for calculating lupin protein, as suggested9 for all legumes, and the factor of 6.38 used for the casein standard, as is commonly done in the dairy industry, then all diets of Table 4 and 5 would have contained just on 10% protein and the p.e.r. for lupin would have been considerably higher, as shown in the last column of Table 5.

Protein quality of lupin seed 857 3.4. Methionine requirement for maximum p.e.r. To obtain maximum p.e.r. of lupin cv Uniwhite the addition of 0.65 g of DLmethionine/kg diet was necessary (Figure l), under the conditions used. The protein source contributed 0.21 g of methionine and 0.43 g of cystinelkg diet. 28 T 0 I I I II I I I I I I I I I 05 10 20 30 Figure rats with (N x 5.5) the diets of young x 6.25) or 100 g/kg 3.5. Supplemental value of lupin protein in barley diets Lupin protein was not quite as efficient as casein in supplementing the protein of barley (Table 6), although when methionine was added the p.e.r. values were not greatly different. With the highest level of lupin in the presence of methionine growth rates of over 7 g/day were attained. With casein, this maximum rate of growth was attained with the middle level of supplementation. Obviously the essential amino acids of lupin are diluted out sooner than are those of casein. This is also seen in the wastage of the supplemental p.e.r. value of casein as the casein level (with added methi onine) increased. 4. Discussion If lupin seed is to become a major protein supplement to cereal grains in mixed feed for poultry and pigs there are at least three questions that need to be resolved. First: is there a need or advantage to use dehulled seeds? The seed coat has negligible

858 E. L. Hove protein and more than 50 % fibre. At 20 to 25 % of the seed, it substantially dilutes the protein nutritional value. The seed coat added to a normal diet depressed protein and organic matter apparent dige~tibility,~ and as shown by data in the present paper, the p.e.r. of whole seed was slightly but consistently less than that of the dehulled seed. On the other hand, the hull contained no growth depressants5 even at high levels. and substantial amounts of the fibre were digested by rats and pigs. This was attributed to the absence of lignan. Whether or not to dehull will probably rest on economic grounds and whether there is a market for the separated seed coats. A peculiarity of the seed coat was the apparent absence of phosphorus, although it had most of the calcium of the seed. The observed higher levels of iron and copper in kernels plus seed coats than in the original seed must be attributed to contamination during dissecting. Second: which variety of lupin seed is the best? This will depend mainly on agronomic factors outside the scope of this paper. Of the two varieties studied in detail, the L. luteus cv Weiko I11 had marked advantage of higher protein content of somewhat better quality. The comparison was made with L. angustifolius cv Uniwhite which was the only cultivar of this species cropped locally. The newer cultivars, especially Uniharvest, have substantially higher protein contents, due in part to significantly lower amounts of seed coat. Uniharvest is said by its developer' to have nonshattering pods resulting in higher yields; this, together with its higher protein ought to make it the variety of choice. Third: how does lupin compare with other protein supplements as to protein quality and supplemental potential to the proteins of cereal grains in mixed diets? As is the case with all legumes and most oil seeds, methionine is the limiting amino acid in lupin protein and needs to be added at about 1 % of the protein for maximum p.e.r. values in rat studies. The p.e.r. based on crude protein (N x 6.25) is deceptive. A good case has been made9 for the use of a more realistic factor to express protein content as N x 5.5 for all seeds. This tends to be resisted since it results in a lower figure for the crude protein content of the commodities in trade; for a similar reason the Dairy Industry always uses the factor of N x 6.38 to ensure higher calculated crude protein in milk products, including casein. When the more nearly accurate conversion factors were used, it was observed that lupin protein (with methionine added) was superior to casein as sole dietary protein and was equal to casein as supplements to the protein of barley meal in rat diets. When lupin kernels were steeped in water and freezedried they promoted superior growth in rats due to increased food intake, but with no change in p.e.r.; this was probably a palatability factor. However, when the lupin kernels were boiled as well as steeped, the increased growth was accompanied by a modest increase in p.e.r. This indicated that boiling had destroyed some toxic factor present in slight amounts. Acknowledgements Mr W. Whitehead of the Manawatu Mills Ltd, Palmerston North supplied and dehulled a large sample of L. angustifolius, cv Uniwhite. Mr Neville Withers of the Agronomy Department of Massey University supplied a number of other samples, including the four varieties reported in Table 1. Sue Mills and Sue King did the carbohydrate and proximate analyses and Mrs Pru Cooper did the amino acid and mineral assays.

Protein quality of lupin seed 859 References 1. Gladstones, J. S. Lupins in Western Australia, Dept. Agric., Western Australia, Perth. 1972, Bull. 3834. 2. Gardiner, M. R.; Gorman, R. C. Aust. J. exp. Agric. Anim. Husbandry 1965,5, 410. 3. Food and Agricultural Organisation of the United Nations. Aminoacid Contents of Foods and Biological Data on Proteins Nutritional Studies No. 24, Rome 1970, pp. 282 and 284. 4. Tannous, R.; Cowan, J. Proc. 7th Int. Congr. Nutr., Hamburg 1967, p. 272. 5. Bailey, R. W.; Mills, S.; Hove, E. L. J. Sci. Fd Aguic. 1974 (in press). 6. Snedecor, G. W. Statistical Methods Iowa State College Press, Ames. 1955, 5th edition. 7. Ofticid Methods of Analyses Association of Official Agricultural Chemists, Washington D.C. 1965, 10th edition. 8. Hove, E. L.; Lohrey, E.; Urs, K.; Allison, R. M. Br. J. Nutr. 1974, 31, 147. 9. Sosulski, F. W.; Sarwar, G. Can. Inst. Fd Sci. Technol. J. 1973, 6, 1. 10. Hansen, R.; Czochanska, Z. J. Sci. Fd Agric. 1974,25, 409.