Identification of the Acidic and Basic Subunit Complexes of Glycinin*

Five complexes consisting of one acidic and one basic subunit that were linked via disulfide bonds were purified from unreduced S-alkylated glycinin. The acidic and basic subunits were identified unambiguously using NHz-terminal sequence analysis, sodium dodecyl sulfate (SDS)-electrophoresis, and analytical isoelectric focusing. The subunit pairings are A1,B2, AlbBlb, A2B1,, A3B4, and F2(2)B3. Polypeptide & was not linked to a corresponding basic subunit via a disulfide bond. The study shows that pairing between subunits is nonran- dom, which is consistent with evidence that glycinin is synthesized as a M, = 60,000 precursor that undergoes post-translational modification to form the individual linked subunits. attempts been made to determine the struc-tural relationships between polypeptides contained within the protein complex called glycinin, one of the primary storage proteins in soybean seeds. These studies demonstrated that glycinin consists of two major groups of polypeptides which are designated acidic and basic subunits on the basis of their isoelectric points (1). Resolution of the polypeptides making up each of the two groups has been difficult, however, due to similarities in their size and charge. Recently, the major polypeptides present in each of the two major subunit groups were purified and identified unambiguously on the basis of their NH2-terminal and partial internal amino acid sequences (2, 3). Six acidic (Ala,

Numerous attempts have been made to determine the structural relationships between polypeptides contained within the protein complex called glycinin, one of the primary storage proteins in soybean seeds. These studies demonstrated that glycinin consists of two major groups of polypeptides which are designated acidic and basic subunits on the basis of their isoelectric points (1). Resolution of the polypeptides making up each of the two groups has been difficult, however, due to similarities in their size and charge. Recently, the major polypeptides present in each of the two major subunit groups were purified and identified unambiguously on the basis of their NH2-terminal and partial internal amino acid sequences (2,3). Six acidic (Ala, Alb, Az, Aa, A4, and F2(2)) and five basic (Bla, Blh, Bz, Ba, and B4) polypeptides having homologous but distinct sequences were found. These polypeptides could each be the products of more than one gene even though they are homologous in the regions sequenced, since each of the 11 purified acidic and basic subunits exhibited charge heterogeneity upon isoelectric focusing (3).
Evidence has been reported that at least some of the acidic and basic subunits are linked with one another via disulfide bonds (4,5) to form acidic-basic polypeptide complexes (ABcomplexes). Reconstitution studies of reductively denatured glycinin as well as the electrophoretic properties of partially purified complexes suggested that pairing between subunits could be nonrandom (6), although precise identification of the * This work was supported by the United States Department of Agriculture Competitive Grants Program (Grant 78-59-2181-0-020-01) and the American Soybean Association Research Foundation. Cooperative research of the United States Department of Agriculture and the Purdue Agricultural Experiment Station. This is Journal Paper No. 8332 from the Purdue Agricultural Experiment Station. Mention of a trademark or proprietary product does not constitute a guarantee or warranty of the product by either Purdue University or of the United States Department of Agriculture. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. individual polypeptides involved in forming them was not possible with the methods used for those studies.
It is important to determine whether pairing between acidic and basic subunits is specific and, if so, which subunits are involved, since this would not only bear on the assembly process for the glycinin complex, but also on the genetic relationships of the subunits. We report here the use of NH2terminal sequence analysis to distinguish between the various polypeptides and identify precisely which subunits are covalently linked.

RESULTS AND DISCUSSION
Preparation of Nonreduced Glycinin-Nonreduced glycinin was purified using the procedure of Moreira et al. (2), except that sulfhydryl-reducing agents were removed from the buffers. Sodium dodecyl sulfate-polyacrylamide electrophoresis of purified unreduced glycinin yielded a major band at M , = 57,000 and a minor one a t about 30,000 (Fig. 1, lane I ) . Exposure of the unreduced glycinin to 2-mercaptoethanol prior to electrophoresis resulted in breakdown of the complexes into the component acidic (Mr = 37,000-42,000) and basic (Mr = 20,000) subunits ( Fig. 1, Eane IO). A minor band was present in lane 10 a t about 10,000 daltons but was generally lost during destaining of the gels. It was due to acidic polypeptide F2(2). These results indicated that the acidic and basic subunits were linked to each other in vivo via one or more disulfide bonds. We refer to the linked subunits as ABcomplexes.
Five AB-complexes were purified using anion exchange chromatography and then characterized using NHs-terminal amino acid sequence analysis, sodium dodecyl sulfate electrophoresis, and isoelectric focusing. Unambiguous identifications of the acidic subunits associated with each complex were made on the basis of the NH2-terminal sequence data published previously (2, 3), while identification of the basic subunits relied on differences in the isoelectric focusing patterns of the purified proteins (3). The results of these experiments are summarized in Table I, and the procedural details supporting our conclusions are contained in the miniprint supplement. ' The results establish that the acidic and basic subunits of glycinin are nonrandomly associated with each other via disulfide bonding. Each of the acidic and basic polypeptides of ' Portions of this paper (including "Materials and Methods" and some of the "Results" including Table I1  Those in lanes [7][8][9][10] corresponded to the ones in lanes 1-5 following treatment with 2% 2-mercaptoethanol. F4 and Fs had the same mobility as F3 and are not shown. Fraction 7 was also unaffected by treatment with 2-mercaptoethanol.

TABLE I Comparison of the methionine content of the acidic and basic subunits and the AB-comwlexes ofdvcinin
glycinin which have been identified thus far on the basis of differences in their primary structures was accounted for in these experiments. With the exception of A,, which is apparently not covalently bound to a basic subunit, each acidic subunit is linked to only one of the basic polypeptides. Given the high degree of sequence homology among the acidic and basic subunits, it is difficult to envision a simple mechanism whereby previously unlinked molecules could associate with one another with such high specificity. We propose that an alternative view accounting for this specificity is that the pairing arises because certain acidic and basic polypeptides are synthesized together from a single gene as a high molecular weight precursor. Evidence supporting this hypothesis is presented in the accompanying paper where it is shown that translating mRNA purified from developing seeds results in the synthesis of a M , = 60, OOO polypeptide which can be selectively immunoprecipitated by antiglycinin-IgG (7). An unanswered question is whether or not & has a unique basic polypeptide associated with it as all other acidic polypeptides do. If so, then the two are either not disulfide linked rtit Complexes 8753 like the other complexes or else its disulfide bonds are particularly sensitive to cleavage. In this regard, it should be noted that several soybean cultivars including Raiden lack F2(2), A,,, and B3.2 Since F2(2) makes up only a small proportion of the total 11 S protein and does not appear to account for all of the B3 which can be purified from glycinin preparations,' we propose that A, is associated with B3 in vivo.
Both F2(2) and A, exhibited size heterogeneity (Fig. 1). While such heterogeneity had not previously been observed for these polypeptides, A2 had been reported to be heterogeneous in size even though it had a single NH2-terminal sequence. While steps were taken to limit proteolysis during the experiment by addition of diisopropylphosphorofluoridate and phenanthroline, an artifactual origin of the size heterogeneity cannot be ruled out. Other plausible explanations are either that several different gene products have the same NH2terminal sequence or there is some variability in the posttranslational modification of precursors to the AB-complexes.
Since the soybean storage proteins are nutritionally limiting in methionine and cysteine when fed to monogastric animals, it is important to consider variations in the content of this amino acid among the various complexes which have been identified ( Table I). There is an 8-fold difference between the complexes having the highest and lowest methionine content. More importantly, the high methionine acidic and basic subunits pair together, while those low in methionine are also paired. This observation may ultimately be quite important in improving the methionine content through plant breeding, particularly if the ratios of the various glycinin complexes present in the seed can be altered with respect to one another. Since glycinin is the most prevalent protein in soybeans, eliminating the methionine-deficient AB-complexes can be expected to exert marked effect on the nutritional quality of the seed. In contralt to free baric subunits which d i d dmunts th?OUghOUt the salt gradlent. They would COnsequently not hlve been rerelved from the backgmund during sequence I n l l y l i s .