Intramolecular disulfide bonds between conserved cysteines in wheat gliadins control their deposition into protein bodies.

Following synthesis, wheat gliadin storage proteins are deposited into protein bodies inside the endomembrane system in a way that enables not only their efficient accumulation and dehydration during seed maturation, but also their rapid rehydration and degradation during germination. In the present report, we studied the mechanism of gliadin deposition and whether it was controlled by the conformation of these proteins. Although gliadins are generally known to be insoluble in aqueous solutions, sucrose gradient analysis showed that a considerable amount of these proteins appeared as relatively soluble monomers in developing grains. In vitro reduction of the intramolecular disulfide bonds that are present in natural monomeric gliadins caused their precipitation into insoluble aggregates. In addition, pulse-chase experiments in the absence or presence of reducing agents showed that formation of intramolecular disulfide bonds also played a major role in folding and deposition of the gliadins in vivo. Our results imply that following sequestration into the endoplasmic reticulum, the gliadins fold into relatively soluble monomers, which are incompetent for rapid aggregation and gradually assemble into protein bodies. This pattern of deposition apparently depends on the conformation of the gliadins, which is stabilized by intramolecular disulfide bonds formed between the conserved cysteines. The contribution of this study to the understanding of the evolution and function of gliadins is discussed.

four subgroups of ␣, ␤, ␥, and aggregated gliadins. All three protein classes contain a domain that is composed of small amino acid sequence repeats, rich in glutamine and proline (2)(3)(4). In the S-poor gliadins, this domain apparently accounts for the majority of the polypeptide. In the HMW-GS, the repetitive domain appears in the center and is flanked by small Nand C-terminal nonrepetitive domains. These nonrepetitive domains contain several cysteine residues that form intermolecular disulfide bonds between the different glutenin subunits resulting in their polymerization (2)(3)(4). In the S-rich gliadins, the repetitive domain is present in the N-terminal part of the polypeptide. This domain varies between the different S-rich gliadin subclasses in the number and consensus sequences of the repeats, but all of the repeats are thought to be arranged in a ␤-turn configuration (3,4). The C-terminal domain of the S-rich gliadins is apparently arranged predominantly in ␣ helix and ␤-sheet configurations and also contains three to four intramolecular disulfide bonds formed between six to eight evolutionary conserved cysteine residues (3,4).
Despite extensive studies, the fine structure of wheat gliadins and their mechanism of deposition into PB are still not clearly understood. As gliadins and glutenins extracted from mature grains are largely insoluble in aqueous solutions, it was suggested that these proteins spontaneously precipitate and aggregate into insoluble deposits immediately after insertion into the ER (see Ref. 5 for a review). However, in recent years, various lines of evidence suggested that the maturation of wheat storage proteins may not be spontaneous, but rather assisted by molecular chaperones that are present inside the ER (1,6). Of these molecular chaperones, the binding protein (BiP) was suggested to assist in the general maturation of storage proteins from wheat and other plant species (7)(8)(9)(10)(11)(12). A second molecular chaperone, protein-disulfide isomerase (PDI), was suggested to assist in the formation of correct intramolecular disulfide bonds in wheat S-rich gliadins (13). Nevertheless, it is still unclear whether the molecular chaperones assist only in the folding of the storage proteins or also in various steps of their deposition into PB.
Studying the mechanism of wheat storage protein deposition, we have previously expressed wild-type and modified Srich gliadins in Xenopus oocytes (14). This study demonstrated that, following insertion into the ER, the gliadins could diffuse rather freely within the organelle for a few hours, suggesting that deposition into PB does not occur by their mere precipitation and aggregation. In addition, density gradient analysis showed that not all of the gliadins were present in dense aggregates, suggesting that the deposition occurs by a slow regulated process. In contrast to wild type proteins, mutant gliadins, lacking conserved cysteines in the C-terminal regions, rapidly aggregated into nondiffusible complexes within the ER, suggesting that intramolecular disulfide bonds played a major role in the deposition of the gliadins. Based on these observations (14), we have presently studied the mechanism of depo-* This research was supported by grant US-2334-93 from the United States-Israel Binational Agricultural Research and Development (BARD) Fund as well as by a grant from the Ministry of Science and Technology (Israel) and the Commission of the European Communities. 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.
‡ Incumbent of the Bronfman Chair of Plant Sciences. To whom correspondence should be addressed. Fax: 972-8-9344181; E-mail: lpgad@wiccmail.weizmann.ac.il. 1 The abbreviations used are: PB, protein bodies; BiP, binding protein; ER, endoplasmic reticulum; HMW-GS, high molecular weight glutenin subunit(s); PDI, protein-disulfide isomerase; DTT, dithiothreitol; PAGE, polyacrylamide gel electrophoresis. sition of gliadins in the homologous system of developing wheat grains. We found that not all of the gliadins were present in dense complexes, but a significant amount appeared as relatively soluble monomers. Moreover, we found that the intramolecular disulfide bonds play a major role in the conformation of the gliadins and control the process of their deposition. The relevance of these findings to the function and evolution of wheat storage proteins is discussed.

EXPERIMENTAL PROCEDURES
Plant Material-Developing grains were obtained from hexaploid bread wheat cultivars Deganit and Atir.
Protein Extraction-Single grains or thin grain slices, at about 17 days after anthesis, were gently ground by a mortar and pestle in 1 ml of 10 mM Tris-HCl, pH 7.5, 150 mM NaCl (aqueous buffer). In some experiments, as indicated in the text, this buffer was supplemented with 0.5% Triton X-100. The grain extract was then passed through cheesecloth in order to clear it from debris. In experiments depicted in Figs. 2, 3, and 9, the resulting homogenate was further centrifuged for 10 min at 13,000 ϫ g in a bench-top centrifuge. The resulting supernatant was defined as "soluble" proteins.
Purified ␣ and ␥ gliadins were kindly provided by Dr. J. Mosse (15). Freeze-dried purified gliadin was resuspended in "aqueous buffer" at a concentration of 1 mg/ml, and soluble purified gliadin was obtained as described above.
Sucrose Gradient Fractionation-Eight hundred l from the wheat grain homogenates or soluble proteins were layered onto a 5.2-ml linear density gradient of 5-20% sucrose in aqueous buffer. In some experiments, DTT was added to the loading material at concentrations of 150 mM and 10 mM, for soluble proteins from developing grains and biochemically purified soluble gliadins, respectively. In these experiments, DTT in the same concentrations was included also within the sucrose gradient itself. In the initial experiments (Figs. 1, 3, and 5), the gradients were centrifuged for 18 h at 45,000 rpm and 4°C in a Beckman sw55Ti rotor, and divided into 13 fractions. In subsequent experiments (Figs. 6 -8), gradients were centrifuged for 2 h and divided into 10 fractions. Both ways yielded good separation between monomeric and complexed gliadins. Individual fractions were trichloroacetic acid-precipitated and washed with ice-cold acetone and 0.1 N NaOH. For enrichment of gliadins, the trichloroacetic acid-precipitated pellets were resuspended in 70% ethanol, incubated for 30 min at 60°C, and centrifuged at 13,000 ϫ g for 15 min. The gliadin-enriched supernatant was then acetone-precipitated.
SDS-PAGE, Western Blot Analysis, and Radiolabel Detection-Proteins were dissolved in sample buffer and fractionated by SDS-PAGE on 10% polyacrylamide gels as described previously (16). In "nonreducing" SDS-PAGE, ␤-mercaptoethanol was omitted from the sample buffer. After electrophoresis, gels were either stained overnight with Coomassie Blue or transferred to nitrocellulose membranes and immunoblotted in Western blots with anti-␥-gliadin (17), anti-HMW-GS (17) and antiyeast BiP (18) antiserum, all at dilutions of 1:1000, or anti-alfalfa PDI (19) at a dilution of 1:20,000. Immunoreacting bands were detected using the ECL kit (Amersham Corp.) according to the instructions of the manufacturer. When needed, blots were stripped of bound antibodies in 100 mM ␤-mercaptoethanol, 2% SDS, 62.5 mM Tris-HCl, pH 6.7, and incubated at 60°C for 30 min with agitation.
Gels loaded with radioactive labeled samples were dried following the SDS-PAGE. The dried gels were then exposed and quantitated using a PhosphorImager™ (Fujix BAS 1000) coupled to an image analysis software system (MacBAS). Following exposure to the imaging plate, gels were usually exposed to an x-ray film.
Metabolic Labeling-Developing wheat grains at about 17 days after anthesis were cross-sectioned with a razor blade. The thin slices were labeled at room temperature with 50 -100 Ci of a [ 35 S]methionine ϩ [ 35 S] cysteine mixture (EXPRE 35 S 35 S, DuPont NEN) in the aqueous buffer for 10 min. In one set of experiments (Fig. 8), the slices were preincubated in 100 mM DTT in aqueous buffer for 30 min prior to labeling. Excess DTT and label were removed by washing in 1 ml of aqueous buffer. In pulse-chase experiments, the grain slice was left in the aqueous buffer for the indicated period of time following the pulse labeling and washing. Incorporation of radiolabel was terminated by rapidly freezing the grain slice in liquid nitrogen. Labeling of proteins at steady state was performed by placing a flap that was cut through the stem of a developing spike, into a tube containing the [ 35 S]methionine and cysteine mixture, for 2 h. The spike was then supplied with water for 24 h, and grains were collected and frozen.
Detection of Reduced Gliadins-Gliadins containing surface thiols were detected by covalent chromatography using activated thiopropyl-Sepharose 6B (Pharmacia Biotech Inc.). Developing wheat grain slices were pulse labeled with the [ 35 S]methionine and cysteine mixture as described above, and then gently extracted in 500 l of deaerated aqueous buffer, to avoid oxidation of free thiol groups. Extracts were centrifuged and pellets (now deprived of water soluble proteins) were resuspended in 350 l of de-aerated aqueous buffer containing 1% SDS, and recentrifuged. The supernatant (enriched in gliadins) was added to 10 mg of preswelled thiopropyl-Sepharose 6B beads. Coupling was then performed by incubating at room temperature for 30 min with gentle shaking. The beads were pelleted at 1000 ϫ g for 1 min, and supernatant containing unbound protein was kept. Beads were washed in 1% SDS-containing aqueous buffer, and the bound protein was then eluted in the SDS-containing buffer supplemented with 5% ␤-mercaptoethanol. Supernatants containing bound and unbound gliadins were trichloroacetic acid-precipitated and washed with acetone and NaOH as described, and then resuspended in protein sample buffer.

Analysis of Storage Protein Complexes in Developing Wheat
Grains-To study the mechanism of wheat storage protein deposition into PB, we first fractionated homogenates from developing grains on 5-20% sucrose gradients. Individual fractions were then reacted in Western blots with anti-HMW-GS and anti-␥-gliadin sera. These sera, which were used for the general identification of HMW-GS and S-rich gliadins, are not entirely type-specific and also detect other proteins in the size range of ϳ50 -60 kDa (apparently S-poor gliadins). In addition, the anti-␥-gliadin serum apparently also detects aggregated gliadins that are similar in size and amino acid sequence to the ␥-gliadins. Thus, hereafter, ␥-gliadins detected by these antibodies will refer to all ␥-type S-rich gliadins. As shown in Fig.  1A, the HMW-GS were entirely present in complexes that sedimented to the bottom of the gradient. A considerable amount of the ␥-gliadins also sedimented to the bottom of the sucrose gradients (Fig. 1B), but, in contrast to the HMW-GS, a significant proportion of the gliadins also sedimented in lighter fractions. Although it is difficult to measure the accurate molecular weight of proteins in sucrose gradients, the gliadins sedimenting in the light fractions were apparently monomers, based on the co-sedimentation of protein markers. Thus, for simplicity, these lightly migrating gliadins were defined as "monomeric gliadins" throughout the rest of the study. Wheat gliadins are generally known to be insoluble in aqueous solutions (3), but whether this is related to the structure of the proteins or to their aggregation state is not clear. We therefore wished to test the solubility of the monomeric gliadins from developing grains in an aqueous solution that is apparently similar to the conditions present inside the ER lumen, where the initial maturation steps of the storage proteins take place. Developing grains were homogenized in aqueous buffer, and, upon centrifugation, soluble (supernatant) and insoluble (pellet) proteins were separated in SDS-PAGE. The gels were then reacted in Western blots with anti-HMW-GS and anti-␥-gliadin sera. As shown in Fig. 2, while the HMW-GS subunits were insoluble, a significant proportion of the ␥-gliadins were soluble. Sucrose gradient fractionation also showed that the soluble ␥-gliadins were present entirely as monomers (Fig. 3A).

Reduction of Soluble Gliadins Causes Their Aggregation into Insoluble Deposits in Vitro-
We have previously demonstrated that elimination of conserved cysteines present in the C-terminal region of a wheat ␥-gliadin triggered its aggregation into nondiffusible complexes in Xenopus oocytes (14). Therefore, we tested whether the presence of a significant amount of gliadins as soluble monomers was related to their conformation as determined by intramolecular disulfide bond formation. Soluble gliadins, extracted from developing wheat grains, were divided into two equal portions, and to one portion we added DTT to dissociate their disulfide bonds. The homogenates were then fractionated on 5-20% sucrose gradients, and individual fractions were reacted in Western blots with anti-␥-gliadin serum. As shown in Fig. 3, the natural soluble gliadins appeared as monomers (panel A), while reduction of the disulfide bonds in vitro caused their aggregation into complexes that sedimented to the bottom of the gradient (panel B). Notwithstanding, two cross-reacting bands with molecular mass larger than 43 kDa still appeared as monomers upon reduction (Fig. 3A, bands marked by an asterisk). These proteins are bigger than the expected size of the ␥-gliadins and may belong to the related fraction of the S-poor gliadins, which do not possess cysteine residues and therefore are not affected by reducing agents (3,4).
As the soluble extract from developing wheat grains apparently contained a mixture of proteins, including molecular chaperones, it was impossible to deduce from Fig. 3 whether aggregation resulted directly from the conformational change of the gliadins upon reduction, or was assisted by additional factors. To address this, we tested the change in conformation and aggregation of purified ␣ and ␥-gliadins, upon reduction of their intramolecular disulfide bonds. First, purified gliadins were dissolved in SDS sample buffer lacking or containing DTT and the proteins were separated by SDS-PAGE. Reduced glia-dins migrate slightly slower in the gel, apparently due to their extended conformation (13). Indeed, as shown in Fig. 4, the DTT-treated purified gliadins migrated slower than the nontreated ones, indicating that the purified gliadins were oxidized and that the DTT reduced their disulfide bonds. Next, the purified gliadins were dissolved in aqueous buffer and soluble gliadins were obtained by centrifugation. Each fraction of soluble gliadins was divided into two equal portions, and one of them was supplemented with DTT to reduce the intramolecular disulfide bonds. The reduced or oxidized gliadins were then fractionated on 5-20% sucrose gradients, and individual fractions were separated by SDS-PAGE and stained with Coomassie Blue. As shown in Fig. 5, both soluble purified ␣and ␥-gliadins migrated as monomers (panels A, C, and E), while reduction with DTT caused their aggregation into complexes that sedimented to the bottom of the gradient (panels B, D, and F).
Maturation of Newly Synthesized Gliadins-The mechanism of deposition into protein bodies was studied further by analyzing the maturation of newly synthesized gliadins in vivo. Developing grains were pulse-labeled with [ 35 S]methionine ϩ [ 35 S]cysteine either for 10 min or for 24 h as a control, and their homogenate was separated on a 5-20% sucrose gradient. Individual fractions were then ethanol-extracted to enrich for the gliadins, and the ethanol-soluble proteins were separated by SDS-PAGE and detected by autoradiography. Notably, a considerable amount of the radioactive gliadins, labeled either for 24 h or for 10 min, sedimented to the bottom of the gradient (Fig. 6, A and B). Western blot analysis of total protein from the same fractions with anti-BiP and anti-PDI sera showed that these two ER-resident molecular chaperones were also enriched at the bottom of the gradient (Fig. 6C). This suggested that sedimentation of the newly synthesized gliadins into the dense fractions may not have been necessarily due to their interactions with each other, but rather due to their association with the ER or with ER-resident molecular chaperones. To eliminate such associations, developing grains were homogenized with aqueous buffer containing 0.5% Triton X-100 and fractionated on 5-20% sucrose gradients. This detergent caused the dissociation of the ER membrane as deduced from the release of most of the PDI and BiP to the top fractions of the gradients (Fig. 6, cf. panels C and F). The Triton X-100 also released a significant amount of monomeric newly synthesized  gliadins (pulse-labeled for 10 min) that sedimented at the top of the gradient (Fig. 6, cf. panels B and E). Nevertheless, also in the Triton X-100-containing extracts, a small but significant amount of the newly synthesized gliadins was still present in complexes that sedimented to the bottom of the gradient. Analysis of the gliadins labeled for 24 h showed that 0.5% Triton X-100 caused also some dissociation of complexes into monomers (Fig. 6, cf. panels A and D), but the degree of dissociation was much smaller than that of the newly synthesized gliadins.
The results of this experiment implied that a significant amount of the newly synthesized gliadins were present as monomers associated with the ER and that their assembly took place later after synthesis. We also analyzed in more detail the time course of gliadin assembly by 5-20% sucrose gradient analysis of Triton X-100-containing homogenates from developing grains that were pulse-labeled with [ 35 S]methionine plus [ 35 S]cysteine for 10 min and chased for 0, 10, and 50 min. As shown in Fig. 7, with the maturation of the gliadins, progressively increasing amount of radioactivity was incorporated both into the top and bottom fractions, while the ratio between monomers and complexes was not altered significantly. This suggested that assembly of the gliadins occurs by a gradual process that continues for hours after their synthesis.

Inhibition of Disulfide Bond Formation in Newly Synthesized Gliadins Causes Their Aggregation in Vivo-
The results presented in Figs. 3 and 5 showed that in vitro, dissociation of the intramolecular disulfide bonds caused the aggregation of monomeric gliadins. We therefore wished to test whether prevention of disulfide bond formation in newly synthesized gliadins would cause their aggregation also in vivo. To address this, developing grains were preincubated with aqueous buffer containing 100 mM DTT for 30 min followed by pulse labeling for 10 min with [ 35 S]methionine plus [ 35 S]cysteine. Control grains were preincubated with aqueous buffer alone. Following the labeling, grains were washed extensively in aqueous buffer to remove traces of DTT. In vivo reduction of the gliadins was first assessed by their binding to thiopropyl-Sepharose beads, which bind to proteins containing free thiol groups. Indeed, while under normal conditions most of the newly synthesized gliadins did not bind to the beads, upon preincubation with DTT, about half of the newly synthesized gliadins were bound to the beads, confirming their in vivo reduction (data not shown). Homogenates from these grains were then fractionated on 5-20% sucrose gradients, fractions were ethanol-extracted, and radioactive proteins in each fraction were analyzed by SDS-PAGE and exposed to a PhosphorImager plate. As shown in Fig. 8, pretreatment with DTT reduced the efficiency of the gliadin labeling. Still, a significantly higher proportion of the gliadins sedimented to the bottom of the gradient, indicating that prevention of disulfide bond formation caused the aggregation of the newly synthesized gliadins in vivo. Although some traces of DTT were probably still present inside the cells after   A, B, D, and E). The other two gels were subjected to Western blot analysis, using simultaneously, both anti-BiP and anti-PDI antibodies (C and F). P, insoluble pellets that sediment to the bottom of the gradient. The location of BiP and PDI is indicated on the right side of panels C and F. the extensive washing, they clearly did not have any in vitro effect on the aggregation of the gliadins because upon homogenization of each small grain slice in 0.8 ml of aqueous buffer the maximal expected concentration of DTT was in the micromolar range, which is much below the minimal concentration (ϳ1 mM) that can effectively cause gliadin aggregation in vitro under these conditions (data not shown).
Insoluble Aggregates of Reduced Gliadins Possess Different Properties than Complexes of Natural Ones-The interactions between gliadins in the insoluble aggregates of the malfolded reduced polypeptides and in the insoluble complexes formed by natural gliadins were studied by the ability of the gliadins to be solubilized from the insoluble pellets upon re-extraction with aqueous buffer. Developing grains were homogenized in aqueous buffer and divided into soluble supernatant and insoluble pellet upon centrifugation (Fig. 9, lanes 1 and 2). The soluble fraction was divided into two fractions, DTT was added to one of them, and the supernatants were then recentrifuged to obtain soluble supernatant and insoluble pellet (Fig. 9, lanes 3  and 4 and lanes 5 and 6). Insoluble pellets of the natural gliadins (Fig. 9, lane 2) and of the reduced ones (Fig. 9, lane 6) were re-extracted in aqueous buffer and divided again into soluble supernatants and insoluble pellets upon centrifugation (Fig. 9, lanes 7 and 8 and lanes 9 and 10). While a considerable amount of gliadins from the insoluble pellets of the natural gliadins were resolubilized, no resolubilization of gliadins was obtained from the insoluble pellets of the reduced polypeptides (Fig. 9, cf. lanes 7 and 8 and lanes 9 and 10).

Maturation of Wheat Storage
Proteins--In the present report, we have shown that the S-rich gliadins and HMW-GS follow different maturation steps during grain development. The HMW-GS efficiently assemble into relatively large insoluble complexes that sediment to the bottom of the 5-20% sucrose gradients. This observation is in close agreement with previous findings that the HMW-GS are deposited in highly dense PB (17,20). Assembly of the HMW-GS apparently occurs both by noncovalent interactions and by formation of intermolecular disulfide bonds between cysteines present on their N and C termini (3,4).
Unlike the HMW-GS, our results show that the newly synthesized S-rich gliadins mature mostly as soluble monomers and a significant amount of them remain as monomers at steady state, in the immature grains. These results support our previous observation (14) that S-rich gliadins can diffuse rather efficiently for several hours in the ER of Xenopus oocytes, and also argue against their spontaneous precipitation and aggregation due to polypeptide structure. In addition, our results support a previous work in which nearly 10% of the gliadins were estimated to be soluble upon extraction of wheat flour in 0.5 M NaCl (21). The presence of monomeric gliadins at steady state in developing grains, and the fact that natural gliadins are largely incompetent for assembly in vitro, also suggest that these proteins possess a specific conformation, which does not favor strong protein-protein interactions. This observation is of particular interest taking into account that the S-rich gliadins are highly enriched in glutamine residues (3,4), which may enhance aggregation by formation of intermolecular hydrogen bonds (22). It is thus possible that many of these residues favor the formation of either intramolecular hydrogen bonds or hydrogen bonds with water. The involvement of intermolecular hydrogen bonds and other noncovalent interactions in the aggregation of the malfolded reduced gliadins was also supported by our finding that this process was prevented when caotropic agents or detergents (i.e. potassium thiocyanate, Triton X-100, and SDS) were present in the aqueous buffer (data not shown).
Not all of the S-rich gliadins remained as monomers. During maturation, some of the proteins progressively assembled into insoluble complexes that sedimented into the bottom of a 5-20% sucrose gradient (complexes containing at least eight polypeptides based on an average gliadin size of ϳ30 kDa and on the sedimentation of molecular markers in the gradient, see Fig. 1). The assembly and precipitation of the gliadins may result from increased gliadin concentration within the PB upon grain maturation. It is also possible that gliadin assembly may be furthermore enhanced by a change in the pH, which occurs following their transport from the ER to the vacuoles (1), similar to the case of legume storage proteins (23). Nevertheless, the interactions between the gliadins are presumably relatively weak as these polypeptides can evidently be rapidly rehydrated and degraded upon seed germination.
The present observation that S-rich gliadins are apparently in an equilibrium between monomers and complexes is in agreement with our previous finding that PB containing them possess variable densities, as deduced from sedimentation in metrizamide gradients (17). However, it is impossible to draw a simpler correlation between the density of the PB and the oligomerization state of the gliadins because the dense PB in metrizamide gradients apparently contain both gliadins and HMW-GS (17), while the HMW-GS are sufficient by themselves to form the dense PB (20).

The Intramolecular Disulfide Bonds in S-rich Gliadins Stabilize Their Conformation and Prevent Their Rapid Aggregation-
The high evolutionary conservation of cysteine residues suggests that they are important for the maturation of the S-rich gliadins. However, as the gliadins possess no known function besides storage, the functional role of these cysteines was not clear. In the present report, we showed that the intramolecular disulfide bonds, formed between these cysteines, play a major role in the conformation of the gliadins and in fact also control their deposition by ensuring slow assembly rather than rapid aggregation and precipitation. This was concluded by two lines of evidence: (i) reduction of soluble monomeric gliadins caused their aggregation in vitro, and (ii) prevention of disulfide bond formation in newly synthesized gliadins enhanced their aggregation in vivo. In this regard, we wish to stress that despite of the fact that both the natural insoluble complexes of the gliadins, as well as the ones formed by the aggregation of malfolded gliadins, sedimented to the bottom of the 5-20% sucrose gradients, the properties of each type of these complexes were entirely different due to the following: (i) the natural complexes contained oxidized gliadins, while the aggregates of the malfolded proteins contained reduced polypeptides; and (ii) when the insoluble pellets of the natural and malfolded reduced gliadins were resuspended in the aqueous buffer, only the natural gliadins could be released as monomers (Fig. 9), suggesting that the interactions responsible for the oligomerization of the natural gliadins were much weaker than those causing the aggregation of the reduced malfolded polypeptides. Our conclusion regarding the role of intramolecular disulfide bonds in the deposition of S-rich gliadins is also supported by biophysical studies (24), suggesting that these disulfide bonds play a role in the conformational structure of these proteins.
Functional Evolution of Wheat Storage Proteins-The major function of storage proteins is to fix amino acids into proteinbound forms that can accumulate to high levels in dense protein bodies within the limiting space of the seed storage cells. However, these storage proteins, which are dessicated during seed maturation, are also efficiently rehydrated, degraded, and mobilized into the germinating embryo upon germination. In-deed, the composition of amino acids in the coding capacity of the storage proteins is noncommon and generally reflects their composition in the free amino acid pool in the seeds. The noncommon amino acid composition of storage proteins, particularly those of cereal grains, that are highly enriched in glutamine residues, may result in tight aggregation and hence interfere with subsequent rehydration and degradation during germination. In the present report, we have shown that the S-rich gliadins have apparently evolved to obtain a specific conformation that enables their folding into forms that are rather incompetent for aggregation despite of their high glutamine content. This appears to be controlled by the formation of intramolecular disulfide bonds between cysteine residues that were highly conserved during gliadin evolution.
Wheat storage proteins also contain a class of HMW-GS that contrarily to the gliadins, efficiently assemble into large insoluble polymers, linked by noncovalent and intermolecular disulfide bonds (Fig. 1) (3, 25). Notwithstanding, in all wheat species, the HMW-GS represent only a minor proportion of about 5% of the total storage proteins and are controlled by loci containing only two genes, in contrast to the gliadins, which appear in multigene families (26). Thus, it is tempting to speculate that the HMW-GS evolved to represent a small fraction of an insoluble core to which the major fraction of the S-rich gliadins can join for the initiation of PB formation. This may have enabled the efficient accumulation of the storage proteins in dense PB within the limiting space of maturing endosperm cells, and their subsequent rapid rehydration and degradation during germination.