Properties, biosynthesis and processing of a sulfur-rich protein in Brazil nut (Bertholletia excelsa H.B.K.)

Transcription

1 Eur. J. Biochem. 162, (1987) 0 FEBS 1987 Properties, biosynthesis and processing of a sulfur-rich protein in Brazil nut (Bertholletia excelsa H.B.K.) Samuel S. M. SUN, Susan B. ALTENBACH and Filomena W. LEUNG ARCO Plant Cell Research Institute, Dublin, California (Received July 30,1986) - EJB An abundant seed protein, which is exceptionally rich in the sulfur-containing amino acids, methionine (18%) and cysteine (So/,), is synthesized in Brazil nut embryos about 9 months after flowering. This sulfur-rich protein consists of two low-molecular-mass polypeptide components, a 9-kDa polypeptide and a 3-kDa polypeptide. The two-subunit polypeptides associate through disulfide linkage(s) to form a 12-kDa protein molecule. We have demonstrated through in vitro translation studies, using RNA from 9-month-old embryos, that the sulfur-rich protein is synthesized as a larger precursor polypeptide of 18 kda. In addition, data from in vivo labelling studies of 9-month-old Brazil nuts suggest that there are two intermediate precursors of the sulfur-rich protein, one of 15 kda and another of 12 kda. One of these precursors, the 12-kDa polypeptide, accumulates for a 2-month period in the developing embryos. From these data we infer that at least three stepwise cleavages are involved in the maturation of the sulfur-rich protein from its 18-kDa precursor. The seeds of Brazil nut (Bertholletia excelsa H.B.K.) contain 15-17% protein by fresh weight [l] and some 50% protein by dry weight of the defatted flour [2]. Brazil nuts are probably one of the richest food sources of the sulfurcontaining amino acids; the total seed protein is reported to contain about 8.3% methionine and cysteine by weight [3]. The total protein of Brazil nuts can be fractionated into three size classes of proteins, the 11 S, 7S, and 2s proteins [4]. The 2 S fraction, which is water-soluble and thus classified as albumin, comprises about 30% of the total protein and is exceptionally rich in the sulfur amino acids; about 30% methionine and cysteine [4]. From this fraction we have purified a sulfur-rich protein, which contains 18% methionine and 8% cysteine. We have found that this protein is composed of two polypeptide components with molecular masses of about 9 kda and 3 kda (unpublished work). In this paper we report on some of the structural properties of the sulfur-rich protein. In addition we report on the biosynthesis of the sulfurrich protein in developing Brazil nut embryos. The maturation of Brazil nut fruits is a lengthy process, usually taking over a year (12-14 months) [l]. Mature Brazil nut embryos consist mainly of hypocotyl[5], in which the sulfur-rich protein, other seed proteins and the oil reserve (75% by fresh weight) are localized. We demonstrate that the sulfur-rich protein is synthesized and begins to accumulate in Brazil nut seeds about 8-9 months after flowering. In addition we have used both in vitro and in vivo labelling studies to show that the lowmolecular-mass sulfur-rich protein is synthesized initially as a larger precursor polypeptide, which undergoes stepwise cleavages to reach its mature form. ~~ Correspondence to S. S. M. Sun, ARCO Plant Cell Research Institute, 6560 Trinity Court, Dublin, California, USA Abbreviations. poly(a)-rich RNA, polyadenylated RNA; SDS, sodium dodecyl sulfate; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel electrophoresis. MATERIALS AND METHODS Plant materials The Brazil nut or castaiia tree (Bertholletia excelsa H.B.K.) is indigenous to the Amazon region; it is not grown anywhere in the United States. The materials used in the present study were from Brazil (Manaus) and Peru (Puerto Maldonado and Iquitos). Brazil nut trees are m tall with branches and fruits located only near the tops of the trees. Fruits were collected 6, 7, 8, 9 and 10months after flowering. Since it was not practical to tag and date the flowers, we estimated the developmental stages of the fruits using November, the month in which these trees usually flower, as the starting month. The fruits are generally round, cm in diameter, and have a tough outer shell. Each fruit contains about nuts (seeds) encased in brown triangular seed coats. After removal of the seed coat, the embryo was used as source material. Embryos from frozen fruits were used for preparation of mrna and protein. For pulse-labelling experiments, embryos from fresh fruits were used. Protein extraction and purification Brazil nut embryos were ground into a fine paste and defatted by extraction with hexane. The defatted flour was extracted with 1 M NaCl in M sodium phosphate buffer, ph 7.5 (1 g flour/lo ml buffer). After the slurry was filtered through four layers of cheesecloth, the homogenate was centrifuged at xg for 30 min at 4 C. The recovered supernatant fraction was designated as total extractable protein. The 2s protein fraction was purified from the total protein extract by sucrose gradient centrifugation as described by Youle and Huang [4]. To obtain pure sulfur-rich protein the 2 S fraction was further purified by extensive dialysis against

2 478 deionized water at 4 C followed by removal of the precipitated contaminating globulin proteins by centrifugation at x g for 30 min. The copurification of the mature sulfur-rich protein and its 12-kDa precursor polypeptide was as described for purification of the mature sulfur-rich protein, except that after two sucrose gradient centrifugations and fractionations the 95% pure mature protein and its precursor polypeptide were subjected to further gel filtration separation using a Sephadex G-150 column (1.6 x 56 cm) equilibrated with 1 M NaCl in M sodium phosphate buffer, ph 7.5. The same buffer was used for elution. Fractions containing the pure sulfur-rich protein and its precursor polypeptide were pooled and dialysed against deionized water before use. Labelling of the sulfur-rich protein Embryo slices (0.8 g/slice) from Brazil nuts at suitable developmental stages were incubated under sterile conditions with 8 p1 (925 kbq) [35S]methionine (3.8 x 1014 Bq mmol-', New England Nuclear, Boston, MA, USA) at 25 "C or 33 "C for the times indicated in the figure legends. When samples were labelled for longer than 1 h, 15 p1 sterile water was spotted onto the embryo slice hourly. At the end of the incubation the embryo slice was washed five times with sterile water, blotted dry and used for total protein extraction. For pulsechase labelling experiments, the embryo slice was labelled for 1 h at 33"C, then washed several times with sterile water, blotted dry and incubated with 25 p M Hepes/KOH buffer, ph 7.0, containing 25 mm methionine for the times indicated. At the end of the chase period the embryo slice was blotted dry, extracted for total protein, analyzed on SDS/20% polyacrylamide gels and visualized by autoradiography. The resulting autoradiogram was scanned using a Quick Scan R& D densitometer (Helena Laboratories, Beaumont, TX, USA). Antibody preparation and immunodetection The sulfur-rich protein purified from mature nuts proved not to be antigenic in rabbits in several trials. However, polyclonal antibodies (prepared by Berkeley Antibody Company, Emeryville, CA, USA) were produced by rabbits against a mixture of the mature protein and its 12-kDa precursor polypeptide purified from immature nuts. We were also able to obtain monoclonal antibodies (prepared by Lovelace Medical Foundation, Albuquerque, NM, USA) against the SDSdenatured 9-kDa and 3-kDa subunits of the sulfur-rich protein. By electrotransfer blotting/enzyme-linked immunoassay (Western blot analysis), both the monoclonal and the polyclonal antibodies were found to react with the 9-kDa polypeptide as well as with the 12-kDa precursor polypeptide but not with the 3-kDa polypeptide. Immunoprecipitation of the sulfur-rich protein and its precursor polypeptides in solution was performed as described previously [6]. Western blot analysis for the sulfur-rich protein was carried out as reported earlier [7]. Other methods The isolation of RNA from developing Brazil nut embryos and cell-free translation of RNA in a wheat germ system were performed as described previously for French beans [6]. SDS/ polyacrylamide slab gel electrophoresis (SDS-PAGE) was performed as described by Laemmli [8]. 20% acrylamide gels were used for the separation of the low-molecular-mass sulfur- rich protein, but still some of the 3-kDa subunit polypeptide might have been lost during staining and destaining. SDS- PAGE stained for glycoprotein was carried out by the periodic/schiffs base procedure [9]. The method of O'Farrell [lo] was used for isoelectric focusing gel electrophoresis. For studies of the possible proteolytic activity of the sulfur-rich protein, 10 pg bovine serum albumin was incubated with various concentrations (0-20 pg) of the sulfur-rich protein at 37 C for 2 h and then analyzed by SDS-PAGE. The sulfurrich protein was also examined for its possible trypsin inhibitor activity: 10 pg bovine serum albumin was incubated with trypsin (United States Biochemical Corp.) at 5 pg and 10 pg levels in the absence and presence of the sulfur-rich protein at various concentrations (2-20 pg) at 37 C for 1 h; after incubation, the protein mixtures were analyzed by SDS- PAGE. Molecular mass of the native sulfur-rich protein was determined by the sucrose gradient sedimentation procedure [I I] using cytochrome c (horse heart, 2S, 13 kda) and ribonuclease A (bovine pancreas, 2S, 13.6 kda) as markers, and by a gel filtration procedure using Sephadex G-50. The column (1.6 x 94 cm) was equilibrated and eluted with M sodium phosphate buffer, ph 7.2. Trypsin (bovine pancreas, 24 kda), myoglobin (whale muscle, 17.8 kda), cytochrome c (horse heart, 13 kda), and aprotinin (bovine lung, 6.5 kda) were used as molecular mass standards. RESULTS AND DISCUSSION Properties of the sulfur-rich protein The molecular mass of the native sulfur-rich protein was determined to be about 12 kda when analyzed by gel filtration methods using Sephadex G-50 (Fig. 1 A) and by sucrose gradient sedimentation (data not shown). The purified sulfurrich protein was then treated with SDS and 2-mercaptoethanol and analyzed by SDS-PAGE. Under these conditions the protein was resolved into two low-molecular-mass polypeptides of about 9 kda and 3 kda (Fig.lB, lane 1). If 2-mercaptoethanol is not included in the protein sample, the 3-kDa polypeptide disappears and a very diffuse protein band with a molecular mass of about 11 kda appears (Fig.lB, lane 2), suggesting that the 9-kDa and the 3-kDa polypeptides associate through disulfide linkage(s) to form a larger native protein molecule of about kda. Analysis of smaller quantities of the dissociated and reduced sulfur-rich protein by SDS-PAGE demonstrated that the broad 9-kDa component of the sulfur-rich protein could be separated into two bands; the slightly faster migrating polypeptide is about 4-5-fold more abundant than the slower migrating band (Fig. 2A). These two bands could represent two polypeptides, which differ by a few hundred Daltons or this difference in electrophoretic mobilities in SDS-PAGE could be caused by changes in just a few amino acid residues in the same polypeptide [12]. By isoelectric focusing, the sulfurrich protein is seen to be charge-heterogeneous and could be resolved into some 20 bands (Fig. 2B). The major polypeptide, amounting to about 80% of the total protein, has an isoelectric point of ph 6.4. This charge heterogeneity is a common property of many seed proteins and suggests that the sulfur-rich protein consists of a family of closely related polypeptides. It is possible that this heterogeneity may also account for the two 9-kDa polypeptide variants observed on SDS-PAGE and for the rather diffuse protein band of the non-reduced sulfur-rich protein as observed on SDS-PAGE

3 479 Ve/ Vo Fig.l. (A) Determination of the molecular mass of the sulfur-rich protein by gel filtration. (B) SDS-PAGE of the sulfur-rich protein before and afier reduction. (A) (0) The purified sulfur-rich protein. (0) Marker proteins; 1, bovine lung aprotinin (6.5 kda); 2, horse heart cytochrome c (13 kda); 3, whale muscle myoglobin (1 7.8 kda); 4, bovine pancreas trypsin (24 kda). (B) 1, 10 pg reduced protein; 2, 10 pg native protein. The reduction was done by treatment of the protein with 1% 2-mercaptoethanol and 1% SDS at 100 C for 2 min. The reduced protein sample was applied immediately to a SDS/20% polyacrylamide gel for electrophoresis (Fig. 1 B). Additional studies on the sulfur-rich protein have indicated that it is water soluble, contains no detectable carbohydrate, possesses no protease activity and is digestible by trypsin (data not shown, see Materials and Methods). Suljiur-rich protein synthesis during seed development Total proteins were extracted from Brazil nuts at various maturation stages and analyzed by SDS-PAGE (Fig. 3A). The mature nuts contain a large amount of the 9-kDa and 3-kDa polypeptide components of the sulfur-rich protein in addition to a number of other major polypeptides (lane d). A small amount of the 9-kDa subunit can be detected in the 8-monthold seeds (Fig. 3A, lane a), but we did not observe this 9-kDa polypeptide band in nuts at earlier maturation stages. Brazil nuts younger than 8 months contain various smaller sized and softer embryos floating in liquid endosperm. By 8 months, the nuts usually are fully occupied with solid embryos. As maturation continues, the quantity of the 9-kDa subunit increases in the seeds (Fig.3A, lanes b, c and d), until this polypeptide reaches some 30% of the total extractable seed protein (Fig. 3 A, lane d). As shown in Fig. 3 A, lanes a - d, the 3-kDa subunit of the sulfur-rich protein also increases in quantity during seed maturation. The biosynthesis of the sulfur-rich protein in the Brazil nuts clearly is developmentally regulated; the synthesis and accumulation of the 9-kDa component of the sulfur-rich protein coincides with the disappearance of the liquid endosperm and the filling up of the seed coats with embryos. For Brazil nuts, which require months for seed development, this time period represents a mid-maturation stage. The developmental stage in which the synthesis of the sulfur-rich protein in Brazil nut occurs is similar to that of the seed storage proteins of many other plants [13] even though the overall times required for seed maturation may be quite different. The sulfur-rich protein from Brazil nut differs from the abundant seed proteins of other plants in being synthesized and stored in the hypocotyl of the embryos rather than in the cotyledons or endosperm. When total proteins from 9-month-old Brazil nuts were reacted with monoclonal antibody prepared against the mature sulfur-rich protein, a polypeptide of 12 kda in addition to the 9-kDa polypeptide was found to react positively with the sulfur-rich protein antibody (Fig. 3 B, lane a). However, this 12-kDa polypeptide was not detected in the mature embryo (Fig.3B, lane b). In agreement with this finding, a predominant protein band of 12 kda is present on Coomassie-stained gels in total protein extracts from 9- month-old embryos but not from the mature ones (Fig. 3A, lanes b and d). During the course of this study we analyzed over 50 fruits at various developmental stages and found that the quantity of the 12-kDa polypeptide relative to the 9-kDa polypeptide varies in embryos around 9 months old. These results suggest that a 12-kDa polypeptide may be a precursor polypeptide of the sulfur-rich protein. This 12-kDa polypeptide accumulates in the developing seed and can be detected in the developing embryos over a 2-month period (Fig. 3A, lanes a and b). Poly(A)-rich RNAs isolated from embryos of various maturation stages were translated in vitro in a wheat germ system using [35S]methionine. When RNA from 9-monthold embryos was translated, little radioactive methionine was incorporated into polypeptides of 9 kda or 3 kda, the size of the sulfur-rich protein subunits which begin to accumulate in embryos of this age, or into a 12-kDa polypeptide, the size of a putative precursor of the sulfur-rich protein. Instead, a polypeptide about 18 kda is labelled heavily (Fig. 4A). This 38-kDa polypeptide is not translated from RNA from earlier stages nor from RNA from the mature nuts. To investigate whether this 18-kDa polypeptide represents a larger precursor polypeptide of the sulfur-rich protein, we immunoprecipitated cell-free protein products translated from 9-month-old embryo RNA with the polyclonal antiserum made against the sulfur-rich protein. As shown in Fig.4B, lanes b and e, the 18-kDa polypeptide is specifically precipitated by the antiserum. The specificity of our antiserum is demonstrated by the selective precipitation of the 18-kDa polypeptide from a mixture of the translation products directed by Brazil nut RNA and brome mosaic virus RNA (Fig.4B, lanes c and g). This result demonstrates that the sulfur-rich protein is initially synthesized as a larger 18-kDa precursor polypeptide. Processing of the sulfur-rich protein To further study the processing sequence of the sulfur-rich protein we labelled Brazil nut embryos in vivo for 2 h with [35S]methionine. Embyros from 9-month-old Brazil nut fruits incorporate much more of the tracer methionine into proteins than those from younger fruits (data not shown). As demonstrated earlier the 9-kDa subunit and a 12-kDa intermediate precursor are two of the major polypeptides synthesized and accumulated in the embryo at this stage (Fig.31, while the most prominent polypeptide synthesized by mrna isolated from embryos at this maturation stage is an 18-kDa precursor polypeptide (Fig.4). To our surprise over 90% of the label

4 480 Fig. 2. Charge and size heterogeneity of the sulfur-rich protein. (A) 2.0 pg purified sulfur-rich protein was analyzed by SDS-PAGE and visualized by Coomassie brilliant blue staining. (B) Isoelectric focusing of purified sulfur-rich protein. The sulfur-rich protein (50 pg) was separated by isoelectric focusing in a tube gel containing 2% Bio-Lyte, ph 3-10, 3% acrylamide, and 0.2% bisacrylamide according to the procedure as indicated under Materials and Methods. The ph profile of the gel was determined using a surface ph electrode (Bio-Rad) Fig. 3. (A) Electrophoretic separation of Brazil nut proteins from seeds of different developmental stages. (B) Immunodetection of the sulfur-rich protein and its precursor polypeptide in developing Brazil nut embryos. (A) Total proteins were isolated from seeds (a) 8 months, (b) 9 months, (c) 10 months or (d) 12 months after flowering and separated by electrophoresis on SDS/20% polyacrylamide gels. Lane e shows the 9-kDa and 3-kDa components of the purified sulfur-rich protein. (B) Total proteins from either 9-month-old Brazil nut embryos (lane a) or mature Brazil nut embryos (lane b) were separated by electrophoresis on SDS/20% polyacrylamide gels and electrotransferred onto nitrocellulose paper. The proteins were allowed to react with a monoclonal antibody directed against the denatured 9-kDa and 3-kDa components of the purified sulfur-rich protein. The polypeptides reacting with the antibody were visualized by autoradiography after reaction with goat anti- (mouse IgM) labelled with lz5i incorporated in vivo by the 9-month-old embryos was found in a polypeptide of 15 kda (Fig. 5, lane a) rather than in the 18-kDa, 12-kDa or 9-kDa polypeptides observed previously. The 15-kDa polypeptide was immunoprecipitated by the antiserum directed against the sulfur-rich protein (Fig. 5B, lane a), indicating that a 15-kDa polypeptide may also be involved as an intermediate precursor in the processing of this protein. However, we did not find this 15-kDa precursor polypeptide to accumulate in the nuts (by either protein staining or immunodetection), suggesting that the 15-kDa precursor rapidly undergoes additional processing. To investigate whether this 15-kDa polypeptide could be processed further into the 1ZkDa and 9-kDa forms of the sulfur-rich protein, we extended the duration of labelling from

5 48 1 Fig.4. (A) Polypeptidessynthesizedin vitro in the wheat germ system by RNA from developing Brazil nut embryos. (B) Selective immunoprecipitation of polypeptides synthesized by RNA from 9-month-old Brazil nut embryos with a polyclonal antibody prepared against the purified suljiurrich protein and its 12-kDa precursor. (A) RNA was isolated from Brazil nut embryos 9 months (lane a), 10 months (lane b) or 12 months (lane c) after flowering and 7.5 pg of each sample was translated in the wheat germ system. The products, labelled with [35S]methionine, were analyzed on SDS/ZOYO polyacrylamide gels and visualized by autoradiography. The positions of the 9-kDa and 3-kDa components of the sulfur-rich protein are indicated along the right side of the gel. (B) The [35S]methionine-labelled cell-free products, directed by 2 pg brome mosaic virus (BMV) RNA (lane a), 7.5 pg total Brazil nut RNA (lane b) or a mixture of 2 pg B W RNA and 7.5 pg Brazil nut RNA (lane c), were analyzed on SDS/20% polyacrylamide gels. In addition, these translation products were immunoprecipitated with a sulfur-rich protein antiserum. Lane d shows the immunoprecipitation of the BMV RNA-directed products, lane e shows the immunoprecipitation of the Brazil nut RNA-directed products and lane g shows the immunoprecipitation of a mixture of BMV and Brazil nut RNA-directed products. The products immunoprecipitated from the mixture of BMV and Brazil nut RNA-directed products with the preimmune serum are shown in lane f Fig. 5. (A) In vivo labelling of 9-month-old Brazil nut embryos with [3 'Sjmethionine. (B) The immunoprecipitation of the in vivo products labelled at 25 C with the polyclonal sulfur-rich protein antiserum. (A) Embryos were labelled in vivo for 2 h at 25 C (lane a), 17 h at 25 C (lane b), or 17 h at 33 C (lane c) and the labelled polypeptides were analyzed on SDS/20% polyacrylamide gels. The sizes estimated for the primary labelled products from comparison with known protein standards are indicated to the right of the gel. (B) Lane a shows the immunoprecipitation of the products labelled in vivo for 2 h at 25 C while lane b shows the immunoprecipitation of the products labelled in vivo for 17 h at 25 C 2 h to 17 h. At the end of this labelling some 90% of the incorporated label was found in a polypeptide of 12 kda (Fig. 5A, lane b), which can be immunoprecipitated with the sulfur-rich protein antiserum (Fig. 5B, lane b). Labelling for up to 24 h did not result in a marked change in this profile, except for a further decrease in the intensity of the 15-kDa polypeptide (data not shown). However, when the labelling experiment was carried out at an elevated temperature (33 C) for 17 h, the amount of label in the 12-kDa polypeptide decreased markedly while that in polypeptides of about 9 kda and 3 kda increased (Fig. 5A, lane c). The precursor/processed product relationship between the 15-kDa, 12-kDa, 9-kDa and 3-kDa polypeptides is demonstrated more convincingly in the pulse-chase labelling experiment shown in Fig.6A. In this experiment embryos from 9-month-old nuts were labelled with [35S]methionine for 1 h at 33 "C and then chased for increasing amounts of time with unlabelled methionine. Once again, over 90% of the radioactivity in the initial pulse-labelled products appears in a 15-kDa polypeptide (Fig. 6A, lane a). After a 6-h chase with unlabelled methionine, the label in the 15-kDa polypeptide decreases while that in the 12-kDa polypeptide increases correspondingly (Fig. 6A, lane b). This shifting of label is more obvious over the course of the next 6 h during which the label in the 15-kDa polypeptide decreases markedly while most of the radioactivity appears in the 12-kDa polypeptide (Fig. 6A, lane c). A longer chase results in another transition of the radioactivity among polypeptides: the label in the 12-kDa band decreases while that in the 9-kDa and a smaller 3-kDa polypeptide increases (Fig. 6A, lanes d, e). The biosynthesis and processing of the sulfur-rich protein in the developing Brazil nut seeds thus appears to occur in the following sequence: 18 kda.+ 15 kda + 12 kda + 9 kda + 3 kda (Fig. 6B). The 18-kDa precursor was not found among the polypeptides labelled in vivo in 9-month-old embyros. Since the cell-free translation system used in the in vitro experiment did not contain microsomal membranes for protein processing, it

6 482 A 3kDa t 12kDa m x x n I 3 kda 9 kda Small Subunit s-s(") Large Subunit Fig. 6. (A) In vivo pulse-chase labelling of 9-month-old Brazil nut embryos with [3SSJmethionine.(B) Proposedprocessing scheme of the 18-kDa precursor of the Brazil nut sulfur-rich protein. (A) Embryo slices, pulse labelled with [35S]methionine for 1 h at 33T, were incubated with unlabelled methionine for 0-26 h. The labelled polypeptides were then analyzed on a SDS/20% polyacrylamide gel and visualized by autoradiography. The autoradiogram was scanned with a densitometer. The positions of polypeptides of 15 kda, 32 kda, 9 kda and 3 kda are indicated. The duration of the chase with unlabelled methionine was (a) 0, (b) 6, (c) 12, (d) 18 and (e) 26 h. (B) Three step-wise cleavages are involved in the maturation of the sulfur-rich protein precursor protein. The first processed polypeptide (about 3 kda) represents a signal sequence. The resulting 15-kDa intermediate precursor is rapidly processed to a 12-kDa second intermediate precursor. The 12-kDa-precursor sulfur-rich protein accumulates in the developing seed for over 2 months before it is further processed to form the 9-kDa large-subunit and 3-kDa small-subunit polypeptides of the sulfur-rich protein appears likely that the 18-kDa primary translation product of the sulfur-rich protein may contain a signal sequence of about 3 kda, which the embryo is able to remove, leaving the 15-kDa sulfur-rich protein precursor polypeptide detected during in vivo embryo-labelling studies. Cytolocalization studies, using immuno-gold marking techniques, demonstrate that the Brazil nut sulfur-rich protein is located in membranebound structures of the hypocotyl (unpublished data), supporting the notion that a signal sequence might be present in the 18-kDa sulfur-rich protein precursor Many seed proteins undergo post-translational proteolytic processing, including some of the low-molecular-mass proteins found in seeds [13]. While the function of post-translational cleavage is unknown at present, the occurrence of two successive cleavage steps and the very different rate of the two processing events in the sulfur-rich protein (1 5 kda to 12 kda, rapid; and 12 kda to 9 kda and 3 kda, slow) is particularly intriguing. During in vivo labelling studies we found that an elevated temperature (33 "C) facilitates the processing of the 12-kDa precursor (Fig. 5A) whereas the processing of the 15-kDa precursor could take place at 25 "C. This higher temperature is about the same as that in the Amazon region where the sulfur-rich protein is synthesized in the Brazil nuts. The difference in temperature requirements between these two processing events suggests that two distinct enzyme systems may be involved. We thank Mr Bruce Nelson and Dr Marleni Flores for help in collecting the Brazil nut fruits, and Dr Phil Filner for helpful discussions and critical review of this manuscript. REFERENCES 1. Schreiber, W. R. (1950) The Amazon basin Brazil nut industry, Foreign Agriculture Report no.49, US Department of Agriculture, Washington DC. 2. Antunes, A. J. & Markakis, P. J. (1977) Agric. Food Chern. 25, Antunes, A. J. (1975) Protein supplement of navy bean with Brazil nuts, Ph.D. dissertation, p. 57. Department of Food Science and Human Nutrition, State University of Michigan, East Lansing, Michigan. 4. Youle, R. J. & Hudng, A. H. (1981) Am. J. Bot. 68, Prance, G. T. & Mori, S. A. (1979) in Flora neotropica, Monograph no. 21, Lecythidaceae, part 1, The actinomorphic-jlowered New World Lecythidaceae (Rogerson, C. T., ed.) p. 58, The New York Botanical Garden, Bronx, New York.

7 6. Hall, T. C., Ma, Y., Buchbinder, B. U., Pyne, J. W., Sun, S. M. & 10. O Farrell, P. H. (1975) J. Biol. Chem. 250, Bliss, F. A. (1978) Proc. Nut1 Acad Sci, USA 75, Martin, R. G. & Ames, B. N. (1981) J. Biol. Chem. 236, Adams, C. A,, Leung, F. W. &Sun, S. M. (1986) Pluntu (Bed.) 12. dejohn, W. W., Zweers, A. & Cohen, L. H. (1978) Biochern. 167, Biophys. Res. Commun. 82, Laemmli, U. K. (1970) Nature (Lond.) 227, Zacharius, R. M., Zell, T. E., Morrison, J. H. & Woodlock, J. J. (1978) Anal. Biochern. 30, Higgins, T. J. V. (1984) Annu. Rev. Plant. Physiol. 35,