Storage proteins of the albumin solubility fraction from seeds of Bertholletia excelsa H.B.K. were separated by reversed-phase high-performance liquid chromatography and their primary structures were determined by gasphase sequencing on intact polypeptides and on the overlapping tryptic and thermolysin peptides. The 2 s storage proteins consist of two subunits linked by disulphide bridges. The large subunit (8.5 kDa) is expressed in at least six different isoforms while the small subunit (3.6 kDa) consists of only one form. These proteins are extremely rich in glutamine, glutamic acid, arginine and the sulphur-containing amino acids cysteine and methionine. One of the variants even contains a sequence of six methionine residues in a row. Comparison with known sequences of 2s proteins of other dicotyledonous plants shows limited but distinct sequence homology. In particular, the positions of the cysteine residues relative to each other appear to be completely conserved, suggesting that tertiary structure constraints imposed by disulphide bridges dominate sequence conservation. It has been proposed that the two subunits of a related protein (the Brassica n a p s storage protein) is cleaved from a precursor polypeptideThe amino acid sequence homology of the Brazil nut protein with the former suggests that a similar protein processing event could occur.Storage proteins are major components of seeds. They are expressed during seed development as long precursors, which undergo co-and post-translational processing before they are deposited, often as protein bodies, in the endosperm, the cotyledons or other tissues of the embryo [l].Several classes of storage proteins are distinguished according to their solubility [2]. The major seed proteins of dicotyledonous plants are the water-soluble albumins, with a sedimentation coefficient of 2s (the 2s proteins), and the globulins which are soluble in high-salt buffers and have a sedimentation coefficient ranging from 7 s to 12s. The plant globulin fraction has been extensively characterized, particularly in nutritionally important legumes and oilseeds. They are generally of high molecular mass and extremely rich in arginine, glutamine and asparagine, amino acids thought to act as nitrogen sources for the developing seedlings [l]. The 2s proteins differ from the globulins in forming a class of lowmolecular-mass proteins which are rich in cysteine. Although some of these proteins have only been recognized recently as storage proteins [3], others, such as the proteinase inhibitors of soybean, have been studied for a long time To date, the complete amino acid sequence of the castor bean (Ricinus communis) and pea (Pisum sativum L.) proteins have been published [6, 71 and the amino acid sequence has been predicted from a cDNA for the rapeseed protein, napin [8]. The castor bean and rapeseed proteins show the same structural organization characterized by two subunits of different size, interlinked by sulphur bridges, while for the pea protein a homodimeric structure has been propo...