SEARCH

SEARCH BY CITATION

Keywords:

  • vacuolar sorting determinant;
  • beta-conglycinin;
  • protein storage vacuole;
  • seed storage protein;
  • soybean;
  • crystalloid-like structure

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In maturing seed cells, many newly synthesized proteins are transported to the protein storage vacuoles (PSVs) via vesicles unique to seed cells. Vacuolar sorting determinants (VSDs) in most of these proteins have been determined using leaf, root or suspension-cultured cells apart from seed cells. In this study, we examined the VSD of the α′ subunit of β-conglycinin (7S globulin), one of the major seed storage proteins of soybean, using Arabidopsis and soybean seeds. The wild-type α′ was transported to the matrix of the PSVs in seed cells of transgenic Arabidopsis, and it formed crystalloid-like structures. Some of the wild-type α′ was also transported to the translucent compartments (TLCs) in the PSV presumed to be the globoid compartments. However, a derivative lacking the C-terminal 10 amino acids was not transported to the PSV matrix, and was secreted out of the cells, although a portion was also transported to the TLCs. The C-terminal region of α′ was sufficient to transport a green fluorescent protein (GFP) to the PSV matrix. These indicate that α′ contains two VSDs: one is present in the C-terminal 10 amino acids and is for the PSV matrix; and the other is for the TLC (the globoid compartment). We further verified that the C-terminal 10 amino acids were sufficient to transport GFP to the PSV matrix in soybean seed cells by using a transient expression system.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Vacuoles of plant cells are multifunctional organelles. It has been shown that there are at least two types of vacuoles: lytic vacuoles (LV) and protein storage vacuoles (PSV) (Vitale and Raikhel, 1999). Co-existence of the LVs and the PSVs is observed in barley and pea root tip cells and maturing pea cotyledon cells (Hoh et al., 1995; Paris et al., 1996). In pea cotyledon cells, the LVs disappear during development and become replaced by PSVs (Hoh et al., 1995). The PSVs contain lectins, proteases and protease inhibitors, as well as major storage proteins (7S and 11S globulins; Herman and Larkins, 1999).

The transport mechanisms of soluble proteins to the LVs or the PSVs are thought to be different from each other. Proteins are transported to the LVs by clathrin-coated vesicles (CCVs) formed at the exit of the Golgi complex. When proteins are sorted to the CCVs, specific amino acid sequences (vacuolar sorting determinants, VSDs) in the proteins are recognized by receptor proteins (Ahmed et al., 2000; Vitale and Raikhel, 1999). On the other hand, transport to the PSVs is generally mediated by dense vesicles (DVs) formed at the cis-cisternae of the Golgi complex (Chrispeels, 1983; Hillmer et al., 2001; Hinz et al., 1999; Hohl et al., 1996). In the case of maturing pumpkin cotyledon cells, it is suggested that PSV proteins are transported bypassing the Golgi complex by precursor accumulating (PAC) vesicles formed at the endoplasmic reticulum (ER) (Hara-Nishimura et al., 1998). The mechanism of transport to the PSVs has not been well studied compared to that of the LVs.

VSDs identified so far are classified into three types: sequence-specific VSDs (ssVSDs), C-terminal VSDs (ct-VSDs), and physical-structure VSDs (ps-VSDs). The ssVSDs contain conserved amino acid sequences, such as the NPIR-like motif, which are considered to be necessary for recognition by a sorting receptor. The ct-VSDs are present in C-terminal regions of polypeptides, and have variable lengths and sequences. They are often enriched in hydrophobic amino acids and are present in C-terminal propeptides. The sorting mechanism related to the ct-VSDs has hardly been elucidated. Although little information has been obtained about the ps-VSDs, they are postulated to be hydrophobic surface patches on the molecular surface resulting from the folding of polypeptides into the three-dimensional protein structures (Neuhaus and Rogers, 1998).

VSDs of various PSV proteins, such as tobacco chitinase (Neuhaus et al., 1991), tobacco glucanase (Melchers et al., 1993), barley lectin (Bednarek et al., 1990), common bean phaseolin (Frigerio et al., 1998a), Brazil nut 2S albumin (Saalbach et al., 1996), castor bean ricin (Frigerio et al., 1998b), and barley phytepsin (Törmäkangas et al., 2001), have been reported. As for ricin, the VSD exists in the internal propeptide region, which contains NPIR-like motif (LLIRP; Frigerio et al., 2001b). It is thought that the VSD of this protein is an ssVSD because Ile in the motif, which is conserved in the internal propeptide regions of proricin and proricin-related proteins, was shown to be essential for vacuolar sorting (Frigerio et al., 2001b). On the other hand, VSDs of chitinase, glucanase, lectin, phaseolin, and Brazil nut 2S albumin are located in their C-terminal propeptide regions and have no conserved amino acid sequence. Therefore, they are considered to be ct-VSDs.

However, these determinants were identified using leaf mesophyll cells, root cells and suspension-cultured cells, and not maturing seed cells. It is difficult to say whether the mechanism of transport of protein in these cells is the same as in seed cells, where proteins are transported by unique DVs or PAC vesicles, as the mechanism of DV or PAC vesicle formation has not been elucidated yet. Therefore, VSDs of the proteins that are transported to the PSVs in seed cells should be determined using seed cells.

Soybean β-conglycinin (7S globulin), one of the major storage proteins of soybean, is a trimeric protein composed of α (approximately 67 kDa), α′ (approximately 71 kDa) and β (approximately 50 kDa) subunits and exhibits molecular heterogeneity (Thanh and Shibasaki, 1976, 1978): 10 molecular species having different subunit compositions are present in the seeds. The α and α′ subunits are synthesized as pre-proproteins and the β subunit as a pre-protein. The propeptides are present in the N-terminal regions. The α and α′ subunits contain extension regions (α, 125 residues; α′, 141 residues) at the N-termini of the mature subunits in addition to the core regions, which exhibit high absolute homologies among all subunits (more than 70%; Maruyama et al., 1998). The extension regions are rich in acidic amino acids and are considered to protrude from molecular surfaces (Maruyama et al., 1998). The pro- and extension regions are not common among legume 7S globulins. The α and α′ subunits contain two high-mannose glycans in the core regions, whereas the β subunit contains one (Utsumi et al., 1997). Previously, using an Escherichia coli expression system, we have shown that neither the pro-region, the extension region, nor the glycans is essential for folding and self-assembly of β-conglycinin (Maruyama et al., 1998, 2001). The physiological roles of these regions remain to be elucidated. Although it is not known whether these subunits have C-terminal propeptides, their C-terminal regions exhibit similarity in sequences with the ctVSD (AFVY) of common bean phaseolin (Figure 1), which is composed of only the β-type subunit (Doyle et al., 1986; Frigerio et al., 1998a, 2001a; Maruyama et al., 1998; Sebastiani et al., 1990).

image

Figure 1. Amino acid sequence alignment of C-terminal regions of β-conglycinin α′, α and β subunits and phaseolin.

Italics indicates a VSD of phaseolin examined using tobacco leaves, tobacco protoplast (Frigerio et al., 1998a), and Arabidopsis suspension-cultured cells (Frigerio et al., 2001a).

Download figure to PowerPoint

To elucidate the contribution of each region of the α′ subunit to vacuolar sorting, we examined the vacuolar sorting of its deletion derivatives, which lack one, two or all three of the following – the propeptide, the extension region and the glycans – or lack the C-terminal 6 or 10 amino acids in transgenic Arabidopsis seed cells by immunogold electron microscopy. We show here that α′ contains two VSDs, and one of them resides in the C-terminal region comprised of 10 amino acids. The C-terminal 10 amino acid sequence is also sufficient for the transport of a reporter protein GFP to the PSVs, examined by a newly developed transient expression assay using maturing cotyledons of soybean. This transient expression method will become a useful tool for elucidating the complex mechanism of vesicular traffic in seed cells.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Accumulation and assembly of wild type α′ and its derivatives in transgenic Arabidopsis seeds

A precursor of the wild-type α′ subunit (Wild) of β-conglycinin consists of a signal peptide, a propeptide, an extension region, and a core region in which two high-mannose-type glycans are linked at Asn 277 and Asn 533 (Figure 2, Wild). Therefore, to examine their contribution to vacuolar sorting, we constructed α′ deletion derivatives, which lack the following: one region, either the propeptide (ΔPropep), the extension region (ΔExt) or the glycans (ΔGlyc); two regions, without the propeptide and extension regions (Core); three regions (CoreΔGlyc); C-terminal 6 (ILRAFY) (ΔCT6) and 10 (PLSSILRAFY) (ΔCT10) residues (Figure 2). All coding sequences were placed downstream of the α′ promoter, which regulates temporal and tissue-specific expression of the α′ gene in soybean and some other transgenic plants (Beachy et al., 1985). We decided to use Arabidopsis as a host plant because sucrose density gradient centrifugation of seed extract and subsequent SDS–PAGE showed an undetectable amount of 7S globulin (data not shown), although it has been reported that there are several genes for putative 7S globulin-like proteins in Arabidopsis genome, which might form a heterotrimer with a vacuolar sorting-impotent α′ derivative and transport it to the PSV. The constructs were transferred into Arabidopsis by an Agrobacterium tumefaciens-mediated gene transfer procedure.

image

Figure 2. Schematic diagram of the wild-type α′ subunit and its derivatives used in this work.

We constructed genes for the wild-type α′, its deletion derivatives, and fusion proteins with GFP.

Download figure to PowerPoint

As shown in Figure 3, the expression and the accumulation of the wild type and the derivatives in the T2 seeds were examined by immunoblotting of non-transformed (lane 4) and transgenic (lanes 5–12) seed extracts. The wild type (lane 5) and all derivatives (lanes 6–12) were detected at their expected sizes (compare the sizes with those of the controls, lanes 1–3), although, probably, the extension regions were partially degraded, resulting in fragments whose sizes were larger than or the same as that of the core region (lanes 5–9). This degradation seems to occur in the seeds, because the degradation pattern turned out to be the same as that shown in Figure 3, when we attempted several extraction methods. Similar degradation was also observed when α′ was expressed in transgenic petunia seeds (Beachy et al., 1985). However, any bands smaller than that of the core did not arise, indicating that the expressed proteins were stably accumulated in the seeds.

image

Figure 3. Accumulation of wild-type α′ and the derivatives in the transgenic Arabidopsis seeds.

The protein extracts (20 µg) from dry seeds of non-transformed (lane 4) and transgenic Arabidopsis for the wild-type α′ (lane 5), ΔPropep (lane 6), ΔGlyc (lane 7), ΔCT6 (lane8), ΔCT10 (lane 9), ΔExt (lane 10), Core (lane 11), and CoreΔGlyc (lane 12) were subjected to SDS–PAGE followed by immunoblotting with an antiserum against α′. As a size marker, we used native α′ purified from soybean seeds (lane 1) and recombinant α′ (lane 2) or α′ core (lane 3) expressed in Escherichia coli. Recombinant proteins from E. coli have no glycans.

Download figure to PowerPoint

To eliminate the possibility that the α′ derivatives might be defective in folding in transgenic Arabidopsis seeds because of the deletion mutation, the self-assemblies of α′ derivatives were examined by sucrose density gradient centrifugation and subsequent immunoblotting (Figure 4), as it could be thought that correct assembly is a consequence of correct folding of α′ derivatives. ΔPropep, ΔGlyc, ΔCT6, and ΔCT10, which have the extension region, sedimented dominantly at 7–8S similar to the wild type, indicating that they were able to self-assemble into trimers. ΔExt, Core, and CoreΔGlyc, which lack the extension region, sedimented at 9–10S, indicating that they assembled into hexamers. We previously showed that the core region of α′ produced in E. coli assembled into a hexamer (a dimer of a trimer) under the conditions examined here (Maruyama et al., 1998). Therefore, all the α′ derivatives folded correctly and assembled in a way similar to that for the wild type.

image

Figure 4. Assembly of the wild-type α′ and the derivatives in the transgenic Arabidopsis seeds.

The seed extracts were subjected to centrifugation on a 12 ml 10–30% (w/v) linear sucrose density gradient and analyzed by SDS–PAGE followed by immunoblotting. Sedimentation is from left to right. Sedimentation co-efficiency is indicated on the top of the figure.

Download figure to PowerPoint

C-terminal region of α′ contributes to vacuolar sorting in transgenic Arabidopsis seeds

The final destinations of the wild-type α′ and its derivatives in the transgenic mature seeds were examined by immunogold electron microscopy with an antiserum against α′ (Figure 5). The PSVs of T2 dry seeds expressing the wild-type α′ were heavily labeled (Figure 5a,b) by the gold particles coupled with a secondary antibody, but the other parts of the cells and outside of the cells were not labeled. In contrast, no specific labeling was detected in the cells of non-transformed Arabidopsis seeds (Figure 5c). These results demonstrate that the wild-type α′ was sorted to the PSVs. Interestingly, the gold particles were not equally distributed in the PSVs of the seeds expressing the wild-type α′. A large part of the gold particles were detected in the circular (maybe spherical) structures (we call this structure a ‘crystalloid-like structure’ in this report; Figure 5a,b, arrow heads), and the rest were detected in translucent compartments (TLCs; Figure 5a,b, arrows and an inset) in the PSVs. Much less gold particles were detected in the crystalloid-like structures when the sections were treated with an antiserum against soybean proglycinin (Figure 5d, arrow head), which would cross-react with Arabidopsis 11S globulin. The crystalloid-like structures were never detected in the PSVs of non-transformed Arabidopsis seeds (Figure 5c). Therefore, this structure was formed mainly by accumulation of α′. On the other hand, the TLCs exist in the PSVs of non-transformed Arabidopsis (Figure 5c). As shown in the inset of Figure 5(b), the inside of the TLCs is actually slightly electron dense, and therefore it means that there is something inside the TLCs. To exclude the possibility that the existence of gold particles in the TLCs of the transgenic seeds was because of non-specific binding of the α′ antibody or the secondary antibody, the number of gold particles within the TLCs of randomly selected PSVs in both the non-transformed and the transgenic seeds immunolabeled with α′ antiserum was quantified. A total of 263 gold particles were found in 825 TLCs of 61 randomly selected PSVs of the non-transformed Arabidopsis seeds; 97.6% of the TLCs had two or less gold particles. On the other hand, 1751 gold particles were found in 675 TLCs of 39 randomly selected PSVs of the transgenic seeds; 77.5% of the TLCs had two or less gold particles. In the transgenic seeds, the number of the TLCs labeled by the antiserum in each PSV varied, with some TLCs containing several dozens of gold particles (Figure 5a,b). In a control experiment, the sections of the transgenic seeds were immunolabeled using the buffer for dilution of antiserum instead of the primary antibody. Almost no gold particles were detected (data not shown). These results show that binding was specific, and that the presence of the gold particles in the TLCs of the transgenic seeds is because of the presence of α′ in the TLCs. The gold particles were hardly detected in the TLCs (Figure 5e) when the sections were treated with antiserum against soybean proglycinin, indicating that the Arabidopsis 11S globulin does not reside in the TLCs. ΔPropep, ΔExt, ΔGlyc, and Core were also detected in crystalloid-like structures (Figure 5f–i) and in the TLCs of the PSVs, but were not detected outside of the cells. These indicate that neither the propeptide, the extension region, nor the glycan was essential for vacuolar sorting of α′. This was further confirmed by the fact that CoreΔGlyc was also detected in the crystalloid-like structures (Figure 5j) and in the TLCs of the PSVs, but was not detected outside of the cells. Therefore, the core region without the glycans is sufficient for sorting of α′ to the PSVs.

imageimageimage

Figure 5. Location of the wild-type α′ and the derivatives in the transgenic Arabidopsis seeds.

Ultrathin sections were treated with antisera against α′ (a–c, f–n) or soybean proglycinin (d, e, o) followed by secondary antibody conjugated to 15 nm gold particles.

(a) A PSV accumulating wild-type α′. This PSV contains only one TLC having more than two gold particles.

(b) A PSV accumulating wild-type α′. This PSV contains four TLCs having more than two gold particles. Inset represents the blow-up image of the localization of α′ to the TLC.

(c) Non-transformed Arabidopsis.

(d) A crystalloid-like structure treated with an antiserum against proglycinin.

(e) Non-transformed Arabidopsis treated with an antiserum against proglycinin.

(f) ΔPropep.

(g) ΔExt.

(h) ΔGlyc.

(i) Core.

(j) CoreΔGlyc.

(k) ΔCT6.

(l) ΔCT6.

(m) ΔCT10.

(n) The intercellular space (ICS) in (m) presented more clearly.

(o) A section of ΔCT10 treated with an antiserum against proglycinin. Arrowheads and arrows indicate the crystalloid-like structures and the translucent compartments (TLC) in the PSVs.

ICS, intercellular space. Bars: (a–e, n, o) 1 µm; (b, inset) 250 nm; (f–k) 500 nm; (l, m) 2 µm.

ΔCT6 was also detected in the crystalloid-like structures and in the TLCs of the PSVs, but not outside the cells (Figure 5k,l), indicating that ΔCT6 was sorted to the PSVs. ΔCT10 was also detected in the TLCs of the PSVs (Figure 5m, arrows) similar to the wild-type α′ and the other derivatives, but it was hardly detected in the other parts of the PSVs. Also, the crystalloid-like structure was not formed (Figure 5m). Moreover, a large amount of ΔCT10 was detected in the intercellular spaces (ICSs) which were highly electron dense (Figure 5m,n), while it was not electron dense at all in the ICSs of seed cells of non-transformed Arabidopsis and of the other transgenic plants (Figure 5l). When the sections were treated with an antiserum against soybean proglycinin, gold particles were localized exclusively in the PSVs, but not in the ICSs (Figure 5o), indicating that the electron-dense aggregates in the ICSs are composed of ΔCT10. We observed the same results from three independent transgenic lines. These indicate that the secretion of ΔCT10 was not caused by an abnormal secretion system of the transgenic plant, but as a result of the lack of the C-terminal 10 amino acids. Thus, we conclude that the C-terminal 10 amino acids are necessary for complete sorting of α′ to the PSVs.

C-terminal 24 amino acids of α′ direct GFP to the PSVs in Arabidopsis seed cells

Recently, our group determined the three-dimensional structure of the core region of α′, and it is considered that its C-terminal 24 amino acid residues are in a state of disorder outside the molecule (unpublished data). It is possible that VSDs reside on molecular surfaces or outside the molecule in order to be recognized by the sorting equipment. We therefore examined whether the C-terminal 24 amino acids are sufficient for sorting of a reporter protein GFP to the PSVs. We constructed a gene for a fusion protein containing the signal peptide and the C-terminal 24 amino acids of α′ at the N- and C-termini of GFP, respectively (Figure 2, GFP-CT24). As a control, we constructed a gene for a fusion protein that consisted of GFP having only the α′ signal peptide at its N-terminus (Figure 2, spGFP). These were introduced into Arabidopsis by A. tumefaciens-mediated gene transfer procedure.

The expression and accumulation of these fusion proteins in the T2 seeds were confirmed by immunoblotting of dry seed extracts with anti-GFP antiserum (data not shown). The location of the fusion proteins was examined by confocal microscopy of T2 dry seeds. Fluorescence of spGFP was detected in the ER network and the ICSs, but hardly in the PSVs (Figure 6a), although in a few cells, weak fluorescence was also detected in the PSVs probably caused by the tendency for GFP to be transported to the vacuole. On the other hand, fluorescence of GFP-CT24 was detected in the PSVs, but not in the ICSs (Figure 6b). These indicate that the C-terminal 24 amino acids of α′ can direct GFP to the PSVs. Interestingly, fluorescence was detected almost all over the PSV, except for some parts which appeared as dark spots (Figure 6c, arrows), in contrast to the fact that the wild-type α′ and its derivatives, except for ΔCT10, were localized in the crystalloid-like structures and in the TLCs (Figure 5a,b,f–l). The dark spots are round in shape (maybe spherical), and the diameter is 0.5–1 µm. They are uniformly distributed in every PSV. From their appearance, these dark spots might represent the TLCs in the PSVs observed by the electron microscopic analysis.

image

Figure 6. Location of GFP fusion proteins in the transgenic Arabidopsis seed cells.

(a) Cells expressing spGFP.

(b) Cells expressing GFP-CT24.

(c) A blow-up image of a cell expressing GFP-CT24. Arrows indicate dark spots in the PSVs.

V, PSV; ICS, intercellular space. Bars: (a, b) 10 µm; (c) 5 µm.

Download figure to PowerPoint

C-terminal 10 amino acids of α′ direct GFP to the PSVs in soybean cotyledon cells

We examined whether the C-terminal amino acids of α′ direct GFP also in soybean cotyledon cells by utilizing a newly developed soybean transient expression assay system involving modified GFP and particle bombardment.

In the transgenic Arabidopsis seeds, a small portion of spGFP was transported to the PSVs. It is possible that this was caused by the same mechanism as sorting of ctVSDs because the C-terminal sequence of GFP (MDELYK, underline indicates hydrophobic residues) is somewhat hydrophobic. In this experiment, we used GFP modified by adding four contiguous Gly (GGGG) at its C-terminus (mGFP), as the contiguous Gly has been reported to inhibit the function of barley lectin ctVSD (Dombrowski et al., 1993). We constructed a gene for a fusion protein containing the signal peptide and the C-terminal 24 amino acids of α′ at the N- and C-termini of mGFP, respectively (Figure 2, mGFP-CT24). As a control, we constructed a gene for a fusion protein that consisted of mGFP having only the α′ signal peptide at its N-terminus (Figure 2, spmGFP). These plasmids were introduced into maturing soybean cotyledons by particle bombardment. The tissue was observed by confocal microscopy 24–40 h after bombardment. In the cells where spmGFP was expressed, the fluorescence was detected mainly in the ER network and in the ICSs, indicating that spmGFP was well secreted out of the cell and, as expected, almost no fluorescence was detected in the PSVs (Figure 7a,b). Fluorescence was not detected in the cells adjacent to the fluorescent cell, indicating that only the cells where the plasmids were introduced gave fluorescence. Fluorescence of mGFP-CT24 was detected in the PSVs (Figure 7c), indicating that the C-terminal 24 amino acids of α′ is sufficient for transporting mGFP to the PSVs. Therefore, the procedure developed here is useful for analysis of vacuolar sorting.

image

Figure 7. Location of modified GFP fusion proteins in soybean cotyledon cells.

(a) spmGFP.

(b) The cell in (a) was optically sectioned at the deeper position of z-axis.

(c) mGFP-CT24.

(d) mGFP-CT24Δ10.

(e) mGFP-CT10.

(f) A visible light image collected simultaneously with (e).

V, PSV; ICS, intercellular space. Bars = 10 µm.

Download figure to PowerPoint

To examine whether the C-terminal 10 amino acids are essential for sorting of mGFP-CT24 to the PSVs, we constructed a gene for a fusion protein that lacks the C-terminal 10 amino acids of mGFP-CT24 (mGFP-CT24Δ10), and introduced this gene into maturing soybean cotyledons. As shown in Figure 7(d), the fluorescent pattern of mGFP-CT24Δ10 was similar to that of spmGFP (Figure 7a,b). The fluorescence was exclusively detected in the ER and in the ICSs, but not in the PSVs, indicating that mGFP-CT24Δ10 was secreted out of the cells. Thus, the C-terminal 10 amino acids of α′ were again shown to be necessary for sorting to the PSVs.

Using the same transient expression system, the C-terminal 10 amino acids alone (mGFP-CT10) were shown to direct mGFP to the PSVs in maturing soybean cotyledons. The fluorescence was exclusively detected in the PSVs (Figure 7e). Figure 7(f) shows a visible light image of the same section as Figure 7(e). There were some intercellular spaces between the adjacent cells, but no fluorescence was detected there as shown in Figure 7(e). This clearly demonstrates that the C-terminal 10 amino acids of α′ are sufficient for directing mGFP to the PSVs in maturing cotyledons of soybean.

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Accumulation of α′ in seeds of transgenic Arabidopsis

The α′ subunit has a propeptide, an extension region, and two glycans (Utsumi et al., 1997). By using an E. coli expression system, it was shown that these parts are not essential for correct folding or assembly (Maruyama et al., 1998, 2001). In this study, the wild-type α′ and all the derivatives lacking one, two, or all of them correctly folded, assembled, and accumulated in seeds of transgenic Arabidopsis, confirming the earlier observation. In addition, these parts were demonstrated not to be essential for sorting of α′ to the PSVs. In other words, increase of surface hydrophobicity and decrease of solubility caused by deletion of the extension region and the glycans (Maruyama et al., 1999) had no effect on sorting of α′ to the PSVs.

The extension region of α′ was partially degraded in Arabidopsis seeds, resulting in fragments whose sizes were larger than or similar to that of the core region detected by immunoblotting. Such degradation does not occur in soybean seeds; thus, it seems to be degraded by an Arabidopsis endogenous enzyme. Hence, the degradation may be a property of the host plants, and not of α′. Similar degradation was also observed when α′ was expressed in transgenic petunia seeds (Beachy et al., 1985). ΔCT10 and ΔPropep largely accumulated in their intact form, while the wild-type α′ was largely degraded to the core size. The ratio of the intact size of ΔCT6 was between those of ΔCT10 and the wild type. ΔPropep and ΔCT6 were sorted to the PSVs as was the wild type, suggesting that the extent of the degradation does not reflect the final destination of the proteins. The size of the degradation products of ΔCT10 secreted out of the cells were similar to those of the others sorted to the PSVs, implying that they were degraded before arrival at the bifurcation on the secretory pathway (i.e. the ER or the Golgi apparatus) by the same enzyme or, after arrival at their final destination, by homologous enzymes.

Sorting determinants of α′ to the multiple compartments in the PSV

The wild-type α′ and all the derivatives other than ΔCT10 were sorted to the PSVs, localized in the crystalloid-like structures and in the TLCs, and were never secreted out of the cells, while a large part of ΔCT10 was secreted out of the cells and some resided in the TLCs of the PSVs. Further, endogenous 11S globulin was not detected in the TLCs. Hence, it is considered that multiple compartments co-exist in the PSV of Arabidopsis: the TLC and the compartment other than the TLC (composed of the matrix where the endogenous 11S globulin accumulates and the crystalloid-like structure formed by α′). Protein transport to the TLC is not by bulk flow because spGFP was secreted out of the cells, and not sorted to the TLCs. Thus, there must be a selective pathway to each compartment, and α′ has two independent VSDs that function in each pathway.

The core region without the C-terminal 10 amino acids that is common between ΔCT10 and CoreΔGlyc must contain the sorting determinant to the TLC, as the wild-type α′ and all the derivatives were sorted to the TLCs. The TLCs are likely to be globoid compartments. The globoid compartments storing highly electron dense phytic acid crystals are found in PSVs of various plants, such as Arabidopsis (Jiang and Rogers, 2001; Otegui et al., 2002). In the globoid compartment, proteins different from those existing in the matrix and the crystalloid are also stored (Jiang et al., 2001). Similarly, we also observed highly electron dense objects in several TLCs (data not shown). However, in our experiment, the inside of most of the TLCs were slightly electron dense (Figure 5b, inset). It is known that the preservation of the phytic acid crystals during preparation for electron microscopy is difficult (Jiang and Rogers, 2001). Therefore, a decline in the electron density of the TLC might be a result of the dissolution and the partial loss of phytic acid crystals during preparation for electron microscopy. Hence, it is conceivable that the core region without the C-terminal 10 amino acids of α′ contains the sorting determinant to the globoid compartment.

The fact that ΔCT10 was secreted out of the cells and was not detected in the compartments other than the TLCs in the PSVs suggests that the C-terminal region of α′, composed of 10 amino acids, works as the sorting determinant for the compartments other than the TLCs (i.e. the matrix and the crystalloid-like structure). This is also supported by the fact that GFP-CT24 was localized mainly in the matrix and hardly in the TLCs (represented by dark spots in the PSVs in Figure 6c) in seed cells of transgenic Arabidopsis. mGFP-CT24 and mGFP-CT10 were also transported to the PSVs in soybean seed cells. However, fluorescence was detected in the whole PSVs, and dark spots, which existed in the PSVs of Arabidopsis seeds (Figure 6c) and represent the TLCs, were not detected. The PSVs of soybean seed cells also contain the globoid compartments (Boatright and Kim, 2000; Lott and Buttrose, 1977). However, they are much smaller (i.d. < 400 nm) (Boatright and Kim, 2000; Lott and Buttrose, 1977) than those (i.d. = 0.5–1 µm) in Arabidopsis seed cells. The dark spots (the globoid compartments) might not be detected with a confocal laser scanning microscope, although they existed in the PSVs.

Formation of the crystalloid-like structure in the PSV

The wild-type α′ and all the derivatives, except for ΔCT10, formed crystalloid-like structures in the PSVs, which were never detected in the seed cells of non-transformed Arabidopsis. The PSVs in seed cells of some plants contain crystalloids (Jiang and Rogers, 2001). A crystalloid exhibits higher electron density than that of the matrix, and contains 11S globulin (although 11S globulin also exists in the matrix; Hara-Nishimura et al., 1987; Jiang et al., 2000; Lord, 1985). Phaseolin (7S globulin of common bean) was mainly detected in the matrix in the PSV of transgenic tobacco seeds (Greenwood and Chrispeels, 1985). In our study, the electron density of the crystalloid-like structure was only slightly different from that of the matrix, and the crystalloid-like structure primarily contained α′, and not the endogenous 11S globulin. Thus, the crystalloid-like structure in this study is a structure with characteristics different from those of the crystalloid. Crystalloid-like structures formed by exogenous protein were also found in the glutelin inclusion bodies when pea legumin (11S globulin) was expressed in wheat endosperm (Stöger et al., 2001), although, unlike the crystalloid-like structure of α′, they contained 11S globulin, were highly electron dense, and grew into crystals in vivo at a high accumulation level.

GFP-CT24 was localized in the matrix, and no crystalloid-like structures were formed in the PSVs of Arabidopsis seed cells. On the other hand, the wild-type α′ and all the derivatives, except for ΔCT10, existed in the crystalloid-like structures and hardly existed in the matrix. These suggest that the formation of the crystalloid-like structures is not essential for the vacuolar sorting performed by recognition of the VSD, and that the crystalloid-like structures might be formed in the matrix after α′ and the derivatives are transported into the matrix. An important factor for the formation of the crystalloid-like structures could be the physicochemical properties of α′, such as surface net charge and hydrophobicity under the environment in the matrix.

Characteristics of the VSD that resides in the C-terminus of α′

The wild-type α′ and all the deletion derivatives, except for ΔCT10, were transported to the PSV matrix and formed the crystalloid-like structures there, whereas ΔCT10 was not detected in the matrix and was secreted out of the cells. The C-terminal 10 amino acid residues (PLSSILRAFY: amino acids with hydrophobic nature are underlined) were sufficient to transport mGFP to the PSV matrix. Thus, it is concluded that the sequence composed of the C-terminal 10 amino acid residues contains the sorting determinant for the PSV matrix where a large amount of storage proteins accumulate. Although it is not known whether α′ has a C-terminal propeptide, it is possible that the VSD is a kind of ctVSD because of the following points, which are consistent with the characteristics of the ctVSDs identified so far (Neuhaus and Rogers, 1998). The VSD of α′: (i) resides in the C-terminus, (ii) has a highly hydrophobic nature, and (iii) has no similarity in sequence and length to those of the other VSDs, except for the VSD of phaseolin. The C-terminal three amino acids (AFY) of α′, similar to the VSD of phaseolin (AFVY) (Frigerio et al., 1998a, 2001a), are not essential for the sorting of α′ to the PSV. This corresponds to the fact that the VSD of barley lectin, known to be one of the ctVSDs, works even when the C-terminal 12 amino acids are removed (Dombrowski et al., 1993); that is, not all amino acids comprising the ctVSD are essential for the function of the ctVSD. However, a possibility that the C-terminal sequence (PLSS) of ΔCT6 that is not present in phaseolin might work as the ssVSD cannot be excluded. In addition, another possibility that the difference in sequence and length might reflect the difference of sorting mechanisms between storage tissues and vegetative tissues cannot be excluded.

Based on the X-ray analysis, the C-terminal region of α′ is disordered (unpublished data) similar to that of phaseolin (Lawrence et al., 1994). However, the disorder of the C-terminus is not a sole requirement for ctVSD-dependent vacuolar sorting, as a contiguous two- or three-Gly peptide, which is strongly expected to be disordered, inhibited the function of barley lectin ctVSD when it was linked to the C-terminus of the ctVSD (Dombrowski et al., 1993). Some other property like hydrophobicity and/or an unknown property of the C-terminal region may be required for correct function of the ctVSDs.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plasmid construction

To construct plasmids for the wild-type α′, DNA consisting of the α′ promoter (−963 to −1) and the α′-coding region was inserted between BamHI and KpnI sites of pBI101 (Clontech, Palo Alto, CA, USA), whose SacI site had been replaced by a KpnI site by means of inserting a KpnI linker (TaKaRa, Tokyo, Japan). The promoter region and the signal peptide and propeptide coding region of α′ were amplified from genomic DNA of soybean (Glycine max L. cv. Wasesuzunari), which had been prepared from young leaves by the cetyltrimethylammonium bromide (CTAB) nucleic acid extraction procedure (Rogers and Bendich, 1988), by PCR using the following primers: primer 1, 5′-GGGGGATCCGTTTTCAAATTTGAATTTTAATGTGTGTTGTAA-3′ (underline indicates a BamHI site introduced upstream of −963 of α′ promoter); primer 2, 5′- CTTAAGGAGGTTGCAACGAGCGTGG-3′. The extension and core coding region of α′ was amplified using α′ cDNA (Maruyama et al., 1998) as a template and the following primers: primer 3, 5′-GTGGAGGAAGAAGAAGAATGCGAG-3′; and primer 4, 5′-CGCGGTACCCGATACTTTCCTCGCTCACT-3′ (underline indicates a KpnI site introduced at 64 bases downstream of stop codon). Both amplified fragments were inserted together between BamHI and KpnI sites of pBluescript SK (Stratagene, La Lolla, CA, USA), creating pBSα′, and then the BamHI–KpnI fragment of pBSα′ was inserted into pBI101 with BamHI/KpnI sites to construct pBIα′.

To construct plasmids for α′ derivatives, ΔPropep, ΔExt and Core, we used the following primers: primers 1 and 2 for the promoter region and the signal and propeptide coding region using genomic DNA as a template for ΔExt; primers 1 and 5 (5′-AATGCCAAATGAGACAGAAACTGATGC-3′) for the promoter region and the signal peptide coding region using genomic DNA as a template for core and ΔPropep; primers 3 and 4 for the extension and core coding region using α′ cDNA as a template for ΔPropep; and primers 4 and 6 (5′-CGAAGACATAAGAATAAGAACCCTTTTC-3′) for the core coding region using α′ cDNA as a template for Core and ΔExt. Two of the PCR products were inserted together between BamHI and KpnI sites of pBluescript to construct pBSΔPropep, pBSΔExt, and pBSCore, and then the BamHI–KpnI fragments were inserted into pBI101 with BamHI/KpnI sites to construct pBIΔPropep, pBIΔExt, and pBICore.

According to the following procedures, plasmids for ΔCT6 and ΔCT10 were constructed. By PCR, a DNA fragment from a SacI site (+1396 from the initiation codon) in the core coding region to 18 (ΔCT6) or 30 (ΔCT10) base pairs upstream of the stop codon were amplified using the following primers: primer 7 (5′-GGACTTGGATGTCTTCCTCAGTG-3′) and primer 8 (5′-TGAAGACAAAGGACCCTTTCTTC-3′) for ΔCT6; and primers 7 and 9 (5′-ACCCTTTCTTCCCTTGTTCCC-3′) for ΔCT10 using pBSα′ as a template. Primer 10 (5′-TGAATAAGTATGTAGTACTAAAATTATG-3′; underline indicates a stop codon) and primer 11 (5′-CGCTTTCTTCCCTTCCTTTCTCGC-3′) were used for amplifying from the stop codon to a DraIII site which is on the f1(–) origin of pBSα′. These PCR fragments were inserted between SacI and DraIII sites of pBluescript SK, and then the SacI–KpnI fragments were inserted into pBIα′ with SacI/KpnI sites to construct pBIΔCT6 and pBIΔCT10.

pBSΔGlyc was constructed by means of mutation of the sites corresponding to two N-linked glycosylation signals of pBSα′. Codons for Asn 277 and Asn 533 were replaced with codons for Ser and Thr, respectively. Before mutation, we confirmed that these replacements do not disturb the folding of this derivative based on the three-dimensional structure of β homotrimers of β-conglycinin (Maruyama et al., 2001). We used pBSα′ as a template and the following primers: primer 12 (5′-GAACCAAGCATGCCACGCTCGTTGC-3′) and primer 13 (5′-AGTCCCGCTAAGGATAACGATGAGG-3′; underline indicates the codon for Ser substituted for Asn) were used for mutation of first glycosylation site, and primer 14 (5′-GCCATTCTTACCTTGGTGAACAACG-3′) and primer 15 (5′-GAAATTCAGATCTGAGGTAGCGGTG-3′; underline indicates the codon for Thr substituted for Asn) were used for mutation of second glycosylation site. Both amplified fragments were inserted together between AflII in the propeptide coding region and BglII sites in the core coding region of pBSα′ to construct pBSΔGlyc, and then the BamHI–KpnI fragment was inserted into pBI101 with BamHI/KpnI sites to construct pBIΔGlyc.

To construct the plasmid for CoreΔGlyc, we used pBSΔGlyc as a template and the following primers: primers 5 and 16 (5′-CAGTTACTTATCCTTCCTCCA-3′) for amplifying a fragment containing the region from an EcoT22I site in the promoter region to the signal peptide coding region, and primers 4 and 6 for amplifying the core coding region. These PCR fragments were inserted together between EcoT22I and KpnI sites of pBSα′ to construct pBSCoreΔGlyc, and then the BamHI–-KpnI fragment was inserted into pBI101 with BamHI/KpnI sites to construct pBICoreΔGlyc.

According to the following procedures, we constructed a plasmid for spGFP. By PCR, we prepared fragments of the α′ promoter and the signal peptide coding region using primers 1 and 5, and pBSα′ as a template. We also prepared a fragment of a GFP coding region using primer 17: 5′-GTGAGCAAGGGCGAGGAGCTGTTCA-3′ and T3 promoter primer, and GFP (S65T)/pBluescript SK (Chiu et al., 1996) as a template, kindly gifted by Dr Y. Niwa, University of Shizuoka. These fragments were inserted together between BamHI and KpnI sites of pBluescript to construct pBSspGFP, and the BamHI–KpnI fragment was inserted into pBI101 with BamHI/KpnI sites to construct pBIspGFP.

According to the following procedures, we constructed a plasmid for GFP-CT 24. We used primer 18 (5′-GCTCTAGAGCGTGAGCAAGGGCGAGGAGCTGTTCA-3′; underline indicates an XbaI site introduced upstream of the codon for the second residue Val of GFP), primer 19 (5′-CTTGTACAGCTCGTCCATGCCGT-3′) and GFP (S65T)/pBluescript SK as a template for amplifying a fragment coding GFP, and primers 11 and 20 (5′-CAGCCTCAGCAGAAAGAGGAGGGGAA-3′), and pBSα′ as a template for amplifying C terminal 24 amino acids coding region of α′ cDNA. Both fragments were inserted together between XbaI and KpnI sites of pBluescript to construct pBSgfp-ct24. Then, using primer 17 and T3 promoter primer, a fragment of GFP having the C-terminal region of α′ was amplified from pBSgfp-ct24. The fragment was inserted with the fragment of the α′ promoter and the signal peptide coding region between BamHI and KpnI sites of pBluescript to construct pBSGFP-CT24. Then, the BamHI–KpnI fragment was inserted into pBI101 with BamHI/KpnI sites to construct pBIGFP-CT24.

According to the following procedures, we constructed a plasmid for spmGFP containing four contiguous Glys at its C-terminus. We used primer 21 (5′-TAAAGCGGCCGCCCGGCTGCAG-3′; italics indicates the stop codon), primer 22 (5′-TCCTCCTCCTCCCTTGTACAGCTCGTCCAT-3′, underline indicates codons for glycines introduced before the stop codon) and pBSspGFP as a template. The amplified fragment was self-ligated to construct pBSspmGFP.

According to the following procedures, we constructed plasmids for mGFP-CT24 and mGFP-CT10 having the C-terminal 24 and 10 amino acids of α′, respectively. We used primer 22 and each of primer 20 or primer 23 (5′-CCTTTGTCTTCAATTTTGAGGGCTTTTTACTGA-3′), and pBSGFP-CT24 as a template. The amplified fragments were self-ligated. To exchange the 3′ untranslated region of these plasmids, which were derived from α′, with the nopaline synthase (nos) terminator, a DNA fragment from a BstYI site (+572 from the initiation codon) in the GFP region of these plasmids to the stop codon was amplified using the following primers: primer 24 (5′-GGTGAACTTCAAGATCCGCCAC-3′, underline indicates a BstYI site), primer 25 (NotI) (5′-ATAGTTTAGCGGCCGCTCAGTAAAAAGCCCT-3′; underline indicates introduced NotI site, and italics indicates the stop codon). The amplified fragments were inserted into pBSspGFP with BstYI/NotI sites to construct pBSmGFP-CT24 and pBSmGFP-CT10.

To construct a plasmid for mGFP-CT24Δ10, we used primer 24 and primer 26 (5′-ATAGTTTAGCGGCCGCTTAACCCTTTCTTCCCTTGTT-3′; underline indicates NotI site, and italics indicates the stop codon introduced after the codon for the eleventh amino acids from the last), and pBSmGFP-CT24 as a template. The amplified fragment was inserted into pBSspGFP with BstYI/NotI sites to construct pBSmGFP-C24Δ10.

Sequences of the DNA regions inserted into pBluescript were confirmed by ABI Prism 310 DNA Analyzer (Applied Biosystems, Foster City, CA, USA) before transfer to pBI101. All pBI plasmids were transformed into A. tumefaciens LBA4404 strain by electroporation using Gene Pulser Transfection Apparatus (Bio-Rad, Hercules, CA, USA).

Transformation of Arabidopsis

Transformation of Arabidopsis thaliana (ecotype Col-0) was performed by an in planta Agrobacterium vacuum infiltration method (Bechtold et al., 1993). Transformed seedlings (T1 plants) were selected on Murashige–Skoog (MS) agar plates containing 200 µg ml−1 carbenicillin and 30 µg ml−1 kanamycin, and were grown to bear T2 seeds. T2 dry seeds were collected and subjected to the following analysis.

Preparation of specific antisera

To prepare specific antiserum against the α′ subunit, recombinant α′ expressed in E. coli was purified as described by Maruyama et al. (1998), and used for the immunization of a rabbit as described by Katsube et al. (1999).

Specific antiserum against soybean proglycinin A1aB1b (Utsumi et al., 1994) was also used in this study.

Protein extraction

Seeds were homogenized at 4°C in 35 mm sodium phosphate buffer (pH 7.6) containing 0.4 m NaCl, 1 mm EDTA, 0.02% (w/v) NaN3, 0.1 µm pepstatin A, 10 mm 2-mercaptoethanol, and 0.1 mm (p-amidinophenyl)methanesulfonyl fluoride (p-APMSF). The homogenates were centrifuged at 13 000 g for 10 min at 4°C to remove cellular debris, and the supernatants were used as crude cell extracts. The amount of total protein in the crude extract was determined using the Wako protein assay kit (Wako, Osaka, Japan).

Immunoblotting

Total protein (20 µg) of each crude cell extract was dissolved in SDS sample buffer composed of 50 mm Tris–HCl (pH 6.8), 2% (w/v) SDS, 10% (v/v) glycerol, 1% (v/v) 2-mercaptoethanol, and 0.004% (w/v) Bromophenol Blue. After boiling, the samples were subjected to SDS–PAGE on 11% polyacrylamide gel. The separated proteins on gels were transferred electrophoretically to a nitrocellulose membrane (0.45 µm; Schleicher & Schuell Inc., Dassel, Germany) and detected with anti-α′ subunit serum or anti-GFP serum (Invitrogen, Carlsbad, CA, USA) followed by a goat antirabbit IgG-alkaline phosphatase conjugate (Promega, Madison, WI, USA).

Analysis of self-assembly

To analyze self-assembly of α′ expressed in Arabidopsis seeds, crude seed extracts were subjected to 10–30% sucrose-density gradient centrifugation as described by Utsumi et al. (1988). Crude cell extract from soybean (cv. Shirotsurunoko) was run in parallel as size markers. Separated fractions were subjected to immunoblotting.

Electron microscopy

Arabidopsis T2 dry seeds were imbibed in water for 20 min before processing. Seeds were vacuum-infiltrated for 1 h with a fixative containing 50 mm sodium phosphate buffer (pH 7.2) and 1.5% (w/v) glutaraldehyde. The seeds were cut into halves and fixed for another 2 h in the fixative, followed by dehydration in an ethanol series, and embedded in epoxy resin (Quetol-812; Nisshin EM, Tokyo, Japan). Ultrathin sections were mounted on copper grids previously coated with carbon-coated formvar films, etched with 3% (v/v) aqueous hydrogen peroxide for 10 min, and immersed in blocking solution composed of PBS containing 1% (w/v) BSA for detection of α′ or 5% (w/v) BSA for detection of 11S globulin for 30 min at room temperature. Then, they were incubated with antiserum against α′ (1 : 2000) diluted in blocking solution or antiserum against glycinin (1 : 50) diluted with diluting solution composed of PBS containing 1% (w/v) BSA and 0.25% (v/v) Tween 20 for 1 h at room temperature. After washing with blocking solution or diluting solution, sections were incubated for 30 min at room temperature with antirabbit IgG secondary antibody conjugated to 15 nm colloidal gold (Amersham Biosciences, Piscataway, NJ, USA) diluted 1 : 25 in blocking solution or diluting solution. The sections were rinsed with PBS and distilled water, and then stained with 4% (w/v) uranyl acetate and lead citrate. After staining, the sections were examined with a transmission electron microscope (model H-700H; Hitachi, Tokyo, Japan) at 100 kV. Control experiment was performed by omitting the primary antibody.

Transient expression assays

Developing seeds of soybean were taken out of their pods and immersed in 70% (v/v) ethanol to sterilize their surfaces. After rinsing with sterile water, they were vertically cut into halves and placed on MS agar plates. Particle bombardment was performed with Biolistic PDS-1000/He (Bio-Rad, Hercules, CA, USA) following the instructions of the manufacturer. Five hundred micrograms of gold microcarriers (particle size of 1.0 µm), coated with 0.5 µg of plasmids of pBSspmGFP, pBSmGFP-CT24, pBSmGFP-CT24Δ10 or pBSmGFP-CT10, was used for each bombardment. Each sample was bombarded twice under the following conditions: vacuum, 26–28 in. Hg; target distance, 6 cm; and Helium pressure, 1100 psi. After bombardment, cotyledons were incubated on the MS agar plates at 25°C in dark for 24 h.

Confocal laser scanning microscopy

Before microscopy, thin sections (less than 1 mm) cut with razor blades from the surfaces of the soybean cotyledons that had been bombarded and incubated were placed on slide glasses and covered with cover glasses. Dry seeds of the transgenic Arabidopsis were pealed and chopped in drops of water on slide glasses, and then covered with cover glasses.

Fluorescent and visible light images were obtained using MRC-1024 confocal laser scanning microscope (Bio-Rad, Hercules, CA, USA). GFP was excited with laser wavelength of 488 nm and detected through a filter for laser wavelength of 506–538 nm.

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Dr T. Aoyama (Kyoto University) for guidance on in planta Agrobacterium vacuum infiltration method, Dr Y. Niwa (University of Shizuoka) for providing the GFP (S65T) gene, and Dr T. Fujiwara (University of Tokyo) for providing Arabidopsis seeds. This work was supported, in part, by grants from the Ministry of Education, Culture, Sports, Science and Technology (to N.M.), and the Program for Promotion of Basic Research Activities for Innovative Biosciences (to S.U.).

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • Ahmed, S.U., Rojo, E., Kovaleva, V., Venkataraman, S., Dombrowski, J.E., Matsuoka, K. and Raikhel, N.V. (2000) The vacuolar sorting receptor AtELP is involved in transport of NH2-terminal propeptide-containing vacuolar proteins in Arabidopsis thaliana. J. Cell Biol. 149, 13351344.
  • Beachy, R.N., Chen, Z.-L., Horsch, R.B., Rogers, S.G., Hoffmann, N.J. and Fraley, R.T. (1985) Accumulation and assembly of soybean β-conglycinin in seeds of transformed petunia plants. EMBO J. 4, 30473053.
  • Bechtold, N., Ellis, J. and Pelletier, G. (1993) In planta Agrobacterium mediated gene transfer by infiltration of adult Arabidopsis thaliana plants. C.R. Acad. Sci. Paris Life Sci. 316, 11941199.
  • Bednarek, S.Y., Wilkins, T.A., Dombrowski, J.E. and Raikhel, N.V. (1990) A carboxyl-terminal propeptide is necessary for proper sorting of barley lectin to vacuoles of tobacco. Plant Cell, 2, 11451155.
  • Boatright, W.L. and Kim, K.S. (2000) Effect of electron microscopy fixation pH on the ultrastructure of soybean protein bodies. J. Agric. Food Chem. 48, 302304.
  • Chiu, W.-I., Niwa, Y., Zeng, W., Hirano, T., Kobayashi, H. and Sheen, J. (1996) Engineered GFP as a vital reporter in plants. Curr. Biol. 6, 325330.
  • Chrispeels, M.J. (1983) The Golgi apparatus mediates the transport of phytohemagglutinin to the protein bodies in bean cotyledons. Planta, 158, 140151.
  • Dombrowski, J.E., Schroeder, M.R., Bendnarek, S.Y. and Raikhel, N.V. (1993) Determination of the functional elements within the vacuolar targeting signal of barley lectin. Plant Cell, 5, 587596.
  • Doyle, J.J., Schuler, M.A., Godette, W.D., Zenger, V., Beachy, R.N. and Slightom, J.L. (1986) The glycosylated seed storage proteins of Glycine max and Phaseolous vulgaris. J. Biol. Chem. 261, 92259238.
  • Frigerio, L., Virgilio, M., Prada, A., Faoro, F. and Vitale, A. (1998a) Sorting of phaseolin to the vacuole is saturable and requires a short C-terminal peptide. Plant Cell, 10, 10311042.
  • Frigerio, L., Vitale, A., Lord, J.M., Ceriotti, A. and Roberts, L.M. (1998b) Free ricin A chain, proricin and native toxin have different cellular fates when expressed in tobacco protoplasts. J. Biol. Chem. 273, 1419414199.
  • Frigerio, L., Foresti, O., Felipe, D.H., Neuhaus, J.-M. and Vitale, A. (2001a) The C-terminal tetrapeptide of phaseolin is sufficient to target green fluorescent protein to the vacuole. J. Plant Physiol. 158, 499503.
  • Frigerio, L., Jolliffe, N.A., Cola, A.D., Felipe, D.G., Paris, N., Neuhaus, J.-M., Lord, J.M., Ceriotti, A. and Roberts, L.M. (2001b) The internal propeptide of the ricin precursor carries a sequence-specific determinant for vacuolar sorting. Plant Physiol. 126, 167175.
  • Greenwood, J.S. and Chrispeels, M.J. (1985) Correct targeting of the bean storage protein phaseolin in the seeds of transformed tobacco. Plant Physiol. 79, 6571.
  • Hara-Nishimura, I., Hayashi, M., Nishimura, M. and Akazawa, T. (1987) Biogenesis of protein bodies from vacuoles in developing pumpkin cotyledons. Protoplasma, 136, 4955.
  • Hara-Nishimura, I., Shimada, T., Hatano, K., Takeuchi, Y. and Nishimura, M. (1998) Transport of storage proteins to protein storage vacuoles is mediated by large precursor-accumulating vesicles. Plant Cell, 10, 825836.
  • Herman, E.M. and Larkins, B.A. (1999) Protein storage bodies and vacuoles. Plant Cell, 11, 601613.
  • Hillmer, S., Movafeghi, A., Robinson, D.G. and Hinz, G. (2001) Vacuolar storage proteins are sorted in the cis-cisternae of the pea cotyledon Golgi apparatus. J. Cell Biol. 152, 4150.
  • Hinz, G., Hillmer, S., Bäumer, M. and Hohl, I. (1999) Vacuolar storage proteins and the putative vacuolar sorting receptor BP-80 exit the Golgi apparatus of developing pea cotyledons in different transport vesicles. Plant Cell, 11, 15091524.
  • Hoh, B., Hinz, G., Jeong, B.-K. and Robinson, D.J. (1995) Protein storage vacuoles form de novo during pea cotyledon development. J. Cell Sci. 108, 299310.
  • Hohl, I., Robinson, D.G., Chrispeels, M.J. and Hinz, G. (1996) Transport of storage proteins to the vacuole is mediated by vesicles without a clathrin coat. J. Cell Sci. 109, 25392550.
  • Jiang, L. and Rogers, J.C. (2001) Compartmentation of proteins in the protein storage vacuole, a compound organelle in plant cells. Adv. Bot. Res. 35, 139173.
  • Jiang, L., Phillips, T.E., Rogers, S.W. and Rogers, J.C. (2000) Biogenesis of the protein storage vacuole crystalloid. J. Cell Biol. 150, 755769.
  • Jiang, L., Phillips, T.E., Hamm, C.A., Drozdowicz, Y.D., Rea, P.A., Maeshima, M., Rogers, S.W. and Rogers, J.C. (2001) The protein storage vacuole: a unique compound organelle. J. Cell Biol. 155, 9911002.
  • Katsube, T., Kurisaka, N., Ogawa, M., Maruyama, N., Ohtsuka, R., Utsumi, S. and Takaiwa, F. (1999) Accumulation of soybean glycinin and its assembly with the glutelins in rice. Plant Physiol. 120, 10631073.
  • Lawrence, M.C., Izard, T., Beuchat, M., Blafrove, R.J. and Coleman, P.M. (1994) Structure of phaseolin at 2.2 Å resolution. Implications for a common vicilin/legumin structure and the genetic engineering of seed storage proteins. J. Mol. Biol. 238, 748776.
  • Lord, J.M. (1985) Synthesis and intracellular transport of lectin and storage protein precursors in endosperm from castor bean. Eur. J. Biochem. 146, 403409.
  • Lott, J.N.A. and Buttrose, M.S. (1977) Globoids in protein bodies of legume seed cotyledons. Aust. J. Plant Physiol. 5, 89111.
  • Maruyama, N., Katsube, T., Wada, Y., Oh, M.H., Barba de la Rosa, A.P., Okuda, E., Nakagawa, S. and Utsumi, S. (1998) The roles of the N-linked glycans and extension regions of soybean β-conglycinin in folding, assembly and structural features. Eur. J. Biochem. 258, 854862.
  • Maruyama, N., Sato, R., Wada, Y., Matsumura, Y., Goto, H., Okuda, E., Nakagawa, S. and Utsumi, S. (1999) Structure-physicochemical function relationships of soybean β-conglycinin constituent subunits. J. Agric. Food Chem. 47, 52785284.
  • Maruyama, N., Adachi, M., Takahashi, K., Yagasaki, K., Kohno, M., Takenaka, Y., Okuda, E., Nakagawa, S., Mikami, B. and Utsumi, S. (2001) Crystal structures of recombinant and native soybean β-conglycinin β homotrimers. Eur. J. Biochem. 268, 35953604.
  • Melchers, L.S., Sela-Buurlage, M.B., Vloemans, S.A., Woloshuk, C.P., Van Roekel, J.S.C., Pen, J., Van Den Elzen, P.J.M. and Cornelissen, B.J.C. (1993) Extracellular targeting of the vacuolar tobacco proteins AP24, chitinase and β-1,3-glucanase in transgenic plants. Plant Mol. Biol. 21, 583593.
  • Neuhaus, J.-M. and Rogers, J.C. (1998) Sorting of proteins to vacuoles in plant cells. Plant Mol. Biol. 38, 127144.
  • Neuhaus, J.-M., Sticher, L., Meins, F., Jr, and Boller, T. (1991) A short C-terminal sequence is necessary and sufficient for the targeting of chitinases to the plant vacuole. Proc. Natl. Acad. Sci. USA, 88, 1036210366.
  • Otegui, M.S., Capp, R. and Staehelin, L.A. (2002) Developing seeds of Arabidopsis store different minerals in two types of vacuoles and in the endoplasmic reticulum. Plant Cell, 14, 13111327.
  • Paris, N., Stanley, C.M., Jones, R.L. and Rogers, J.C. (1996) Plant cells contain two functionally distinct vacuolar compartments. Cell, 85, 563572.
  • Rogers, S.O. and Bendich, A.J. (1988) Extraction of DNA from plant tissues. In Plant Molecular Biology Manual, (Gevin, S.B., Schilperoort, R.A. and Verma, D.P.S., eds). Dordercht: Kluwer Academic Publishers, pp. A6/1A6/10.
  • Saalbach, G., Rosso, M. and Schumann, U. (1996) The vacuolar targeting signal of the 2S albumin from Brazil nut resides at the C terminus and involves the C-terminal propeptide as an essential element. Plant Physiol. 112, 975985.
  • Sebastiani, F.J., Farrell, L.B., Schuler, M.A. and Beachy, R.N. (1990) Complete sequence of cDNA of α subunit of soybean β-conglycinin. Plant Mol. Biol. 15, 197201.
  • Stöger, E., Parker, M., Christou, P. and Casey, R. (2001) Pea legumin overexpressed in wheat endosperm assembles into an ordered paracrystalline matrix. Plant Physiol. 125, 17321742.
  • Thanh, V.H. and Shibasaki, K. (1976) Heterogeneity of beta-conglycinin. Biochim. Biophys. Acta, 439, 326338.
  • Thanh, V.H. and Shibasaki, K. (1978) Major proteins of soybean seeds. Subunit structure of β-conglycinin. J. Agric. Food Chem. 26, 692695.
  • Törmäkangas, K., Hadlington, J.L., Pimpl, P., Hillmer, S., Brandizzi, F., Teeri, T.H. and Denecke, J. (2001) A vacuolar sorting domain may also influence the way in which proteins leave the endoplasmic reticulum. Plant Cell, 13, 20212032.
  • Utsumi, S., Kim, C.S., Sato, T. and Kito, M. (1988) Signal sequence of preproglycinin affects production of the expressed protein in Escherichia coli. Gene, 71, 349358.
  • Utsumi, S., Kitagawa, S., Katsube, T., Higasa, T., Kito, M., Takaiwa, F. and Ishige, T. (1994) Expression and accumulation of normal and modified soybean glycinins in potato tubers. Plant Sci. 102, 181188.
  • Utsumi, S., Matsumura, Y. and Mori, T. (1997) Structure-function relationships of soy proteins. In Food Proteins and Their Applications (Damodaran, S. and Paraf, A., eds). New York: Marcel Dekker, pp. 257291.
  • Vitale, A. and Raikhel, N.V. (1999) What do proteins need to reach different vacuoles? Trends Plant Sci. 4, 149155.