Protein Quality and Antinutritional Factors of Wild Legume Seeds from the Sonoran Desert Magdalena Ortega-Nieblas* Center of Scientific and Technologic Research, University of Sonora (CICTUS), A.P. 1819, Hermosillo, Sonora, Mexico 83000 Luz Va ´ zquez-Moreno and Marı ´a R. Robles-Burguen ˜o Research Center for Food and Development (CIAD), A.P. 1735, Hermosillo, Sonora, Mexico 83000 Seeds from seven species of legume trees of the Sonoran Desert were studied to determine their potential value as a nonconventional food and/or as a feed source. Seed proximal composition and protein quality were evaluated by conventional methodology. Defatted flours from Acacia farnesiana, Cercidium microphyllium, Cercidium sonorae, Mimosa grahamii, Olneya tesota, Parkinsonia aculeata, and Prosopis juliflora have protein contents that ranged from 20 to 30%. These proteins were deficient in sulfur amino acids but high in lysine and phenylalanine content. Except for M. grahamii seeds, trypsin inhibitors were relatively high in all species and varied from 50 to 70 TIU/ g. Other antinutritional factors as lectins, alkaloids, saponins, and phenols were detected in all but Mimosa seeds. Protein in vitro digestibility was found to be 77-84%, and these values increased (5-10%) after a thermic treatment. The calculated C-PER values were found to be 1.8-2.1%. From the seeds of the seven species evaluated, M. grahamii showed the highest potential as a protein source. Keywords: Desert legume seeds; protein quality; antinutritional factors INTRODUCTION Intense efforts to find alternative sources of proteins from plants adapted to adverse conditions are been conducted around the world (Siddhuraju et al., 1995; Bravo et al., 1994; Bhattacharya et al., 1994). Seeds from Cassia laevigata and Tamaridus indica, both legumes from India, were reported as a low-cost protein source to alleviate the protein malnutrition among people living in developing countries (Siddhuraju et al., 1995). Seeds and pods from the genera of Acacia, Olneya (palo fierro, iron wood), and Prosopis (mesquite) have been used in several countries for human con- sumption, livestock food, and fodder production (Figueire- do, 1990; Degen et al., 1995; Felker, 1981). Studies of mesquites from South America and the Mediterranean confirmed that these plants produce pods and seeds that are palatable to human and animals. In a recent comparison of mesquite pods with carob pods (the latter is widely used in the food industry), it was demonstrated that mesquite pods had better nutritional properties than carob pods, including lower polyphenolic content and higher protein digestibility (Bravo et al., 1994). In Mexico, the Sonoran Desert has a large group of perennial trees from the Leguminosae family that produce large amounts of indehiscent pods; however, the chemical composition and protein quality of their seeds remain unknown. These plants have important agro- nomic characteristics applicable under unfavorable conditions such as reduced availability of water and fertilizers (Felker, 1981; Meyer et al., 1986; Bresanni and Elias, 1974). In the United States, some of these species have been used for arid land reforestation and erosion control (Estrada and Marroquin, 1992). Fur- thermore, Prosopis pod yield has been estimated up to 10 000 kg/ha (Felker, 1981; Felger and Nabham, 1978). In this work, seeds of seven legume species of the Sonora Desert were selected based on their abundance in the natural habitat and productivity, and their seed com- position and protein quality were determined by con- ventional methods. MATERIALS AND METHODS Seeds from selected species were collected from various regions of the Sonoran Desert (Rodney et al., 1972; Ortega- Nieblas, 1993). The species were Acacia farnesiana (huizache), Cercidium microphyllium (palo verde 1), Cercidium sonorae (brea), Mimosa grahamii (gatun ˜ a), Olneya tesota (palo fierro, iron wood) Parkinsonia aculeata (palo verde 2), and Prosopis juliflora (mesquite). Seeds were manually separated from husks and ground on a Wiley grinder using a 150 μm mesh. All reagents used were of analytical grade. Chemical Analysis. Fat was extracted from 100 g of meal in a Soxhlet apparatus using hexane (AOAC, 1990). Defatted flours were stored at 4 °C until used. The nitrogen content of flours was determined using the micro-Kjeldahl method (AOAC, 1990). A conversion factor of 6.25 was used to calculate protein content. Moisture, crude fiber, and ash were determined using the official methods (AOAC, 1990). Carbo- hydrate content was calculated by difference. Antinutritional Factors. Trypsin inhibitors were deter- mined by using the procedure described by Hamerstrand and Black (1981). Phenols, alkaloids, and saponins were deter- mined using the methods described by Goldstein and Swain (1975), Waldi and Schnackerz (1973), and Harbone (1984). Hemagglutinin Activity. Crude extracts were obtained by extraction of flours in saline solution, 1:10 (w/v), for 2 h at room temperature and overnight at 4 °C. Crude extracts were recovered after centrifugation, 20 min at 5000g (Lis and Sharon, 1986). Hemagglutinin activity was determined by the serial double-dilution method using human B + erythrocytes treated with trypsin (Turner and Liener, 1975). * Author to whom correspondence should be ad- dressed [fax (62) 80-0058]. 3130 J. Agric. Food Chem. 1996, 44, 3130-3132 S0021-8561(95)00613-3 CCC: $12.00 © 1996 American Chemical Society