Protein disulfide isomerase family proteins involved in soybean protein biogenesis Hiroyuki Wadahama 1, *, Shinya Kamauchi 1, *, Masao Ishimoto 2 , Teruo Kawada 1 and Reiko Urade 1 1 Graduate School of Agriculture, Kyoto University, Uji, Japan 2 National Agricultural Research Center for Hokkaido Region, Sapporo, Japan Many proteins that are synthesized in the endoplasmic reticulum (ER) are folded with an accompanying forma- tion of intramolecular disulfide bonds, aided by protein disulfide isomerase (PDI) and related proteins, which are characterized by thioredoxin motifs within their pri- mary structure [1,2]. Both yeast and mammalian PDIs are known to be multifunctional folding catalysts and molecular chaperones, which catalyze the formation and rearrangement of disulfide bonds between correct pairs of cysteine residues in nascent polypeptide chains within the ER [3]. Mammalian PDI functions not only as a catalytic enzyme, but also as a subunit of both microsomal triacylglycerol transfer protein [4] and prolylhydroxylase [5]. The mammalian PDI family, ER-60 ⁄ ERp57, which also has a protein oxidoreductase activity, interacts and cooperates with calnexin and cal- reticulin for oxidative folding of N-glycosylated proteins [6–8]. The genes of these PDI families are unfolded pro- tein response (UPR) genes, which are induced by the accumulation of unfolded proteins in the ER [9]. Keywords endoplasmic reticulum; protein disulfide isomerase; soybean; storage protein; unfolded protein response Correspondence R. Urade, Graduate School of Agriculture, Kyoto University, Uji, Kyoto 611-0011, Japan Fax: +81 774 38 3758 Tel: +81 774 38 3757 E-mail: urade@kais.kyoto-u.ac.jp Database The nucleotide sequence data for GmPDIS-1, GmPDIS-2, BiP, calreticulin and b-conglycinin a¢ are available in the DDBJ ⁄ EMBL ⁄ GenBank databases under accession numbers AB182630, AB182631, AB210900, AB196794 and AB113351 *These authors contributed equally to this work (Received 23 July 2006, revised 5 October 2006, accepted 22 November 2006) doi:10.1111/j.1742-4658.2006.05613.x Protein disulfide isomerase family proteins are known to play important roles in the folding of nascent polypeptides and the formation of disulfide bonds in the endoplasmic reticulum. In this study, we cloned two similar protein disulfide isomerase family genes from soybean leaf (Glycine max L. Merrill cv. Jack) mRNA by RT-PCR using forward and reverse primers designed from the expressed sequence tag clone sequences. The cDNA encodes a protein of either 364 or 362 amino acids, named GmPDIS-1 or GmPDIS-2, respectively. The nucleotide and amino acid sequence identities of GmPDIS-1 and GmPDIS-2 were 68% and 74%, respectively. Both pro- teins lack the C-terminal, endoplasmic reticulum-retrieval signal, KDEL. Recombinant proteins of both GmPDIS-1 and GmPDIS-2 were expressed in Escherichia coli as soluble folded proteins that showed both an oxidative refolding activity of denatured ribonuclease A and a chaperone activity. Their domain structures were identified as containing two thioredoxin-like domains, a and a¢, and an ERp29c domain by peptide mapping with either trypsin or V8 protease. In cotyledon cells, both proteins were shown to dis- tribute to the endoplasmic reticulum and protein storage vacuoles by con- focal microscopy. Data from coimmunoprecipitation and crosslinking experiments suggested that GmPDIS-1 associates with proglycinin, a pre- cursor of the seed storage protein glycinin, in the cotyledon. Levels of GmPDIS-1, but not of GmPDIS-2, were increased in cotyledons, where glycinin accumulates during seed development. GmPDIS-1, but not GmPDIS-2, was induced under endoplasmic reticulum-stress conditions. Abbreviations Ab, amyloid b-peptide; AZC, L-azetidine-2-carboxylic acid; DSP, dithiobis(succinimidylpropionate); ER, endoplasmic reticulum; PDI, protein disulfide isomerase; PSV, protein storage vacuole; UPR, unfolded protein response. FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 687 In plants, a genome-wide search of Arabidopsis thali- ana identified a set of 22 orthologs of known PDI-like proteins that was separated into 10 phylogenetic groups [10]. Among these groups, five groups (I–V) have two thioredoxin domains and show structural similarities to PDI-like proteins in other higher eukary- otes. The PDI family proteins that are categorized in group IV are the smallest molecules (approximately 360 amino acids) in groups I–V of these proteins. The amino acid sequences of the group IV proteins, other than the sequence of the thioredoxin domain, were dif- ferent from those of groups I–III and group V. Group IV proteins lack a KDEL-like ER retrieval signal. The genes of group IV have been identified only in plant and Dictyostelium genomes. Previously, it was shown that mRNA of the Arabidopsis group IV ortholog was induced by ER stress [11,12]. This finding implies that group IV proteins play an important role in quality control of proteins in the ER; however, their cellular localization, function and physiologic roles remain unclear. In this study, we report the isolation of cDNA clones that encode two soybean group IV PDI ortho- logs, GmPDIS-1 and GmPDIS-2. The identification of their domain structures, tissue distribution, cellular localization and changes in expression during soybean seed embryogenesis are described. In addition, we pro- vide evidence to suggest an association between GmPDIS-1 and proglycinin, a seed storage protein, in the course of the folding process. Results cDNA cloning and expression of GmPDIS-1 and GmPDIS-2 To clone the soybean ortholog of Arabidopsis PDI- like2-1 categorized in group IV [10], a blast search was performed using the nucleotide sequence of PDI- like2-1 cDNA from The Institute for Genomic Research Soybean Index. As a result, two tentative consensus sequences, TC176086 and TC176115, were found. Using two primer sets designed from their nuc- leotide sequences, we cloned two cDNAs from the RNA extracted from young soybean leaves by RT-PCR. These cDNAs encoded proteins, named GmPDIS-1 and GmPDIS-2, which consisted of 364 and 362 amino acids, respectively (Fig. 1). The nucleotide and amino acid sequence identities of GmPDIS-1 and GmPDIS-2 were 68% and 74%, respectively. Both proteins possess a putative N-terminal secretory signal sequence and two tandem thioredoxin-like motifs, with a CGHC active site. Arginine residues (R122 and R241 of GmPDIS-1, and R121 and R240 of GmPDIS-2), which have been demonstrated to be involved in the regula- tion of the active site redox potential in human PDI [13,14], were conserved. In addition, glutamic acid resi- dues (E51 and E170 of GmPDIS-1, and E50 and E169 of GmPDIS-2), which have been suggested to facilitate the escape of the active site from a mixed disulfide with the substrate [15], were also conserved. Most PDI family proteins found in eukaryotic cells have C-ter- minal, KDEL-related sequences that act as a signal for retention in the ER [16,17]. However, GmPDIS-1 and GmPDIS-2 lack this type of C-terminal signal. An amino acid sequence similar to the C-terminal domain of ERp29, an animal PDI-related protein [18,19], was present in the C-terminal region of both GmPDIS-1 and GmPDIS-2. The recombinant GmPDIS-1 and GmPDIS-2 pro- teins were expressed in Escherichia coli and purified (Fig. 2A,B). Both recombinant proteins were soluble and eluted in a monomeric form from a gel-filtration column (data not shown). To determine whether both recombinant proteins were folded, far-UV CD was performed. Both GmPDIS-1 and GmPDIS-2 yielded CD spectra typical of a folded protein (data not shown). The activity of recombinant GmPDIS-1 and GmPDIS-2 (i.e. the catalysis of oxidative refolding of the reduced, denatured RNaseA) was measured. The specific activities of GmPDIS-1 and GmPDIS-2 were 66 and 43 mmol RNaseAÆmin )1 Æmol )1 , respectively (Fig. 2C). The specific activity of bovine PDI was 431 mmol RNaseAÆmin )1 Æmol )1 . The domain structures of GmPDIS-1 and Gm- PDIS-2 were predicted to be a linear sequence of three domains in an a–a¢–ERp29 C-terminal-like domain (ERp29c) from the region of sequence homology to the conserved domains. Hence, we subjected the recombinant GmPDIS-1 and GmPDIS-2 proteins to limited proteolysis with either trypsin or V8 protease to determine their domain boundaries. After proteoly- sis for various time periods, the native recombinant proteins were gradually degraded, resulting in the gen- eration of smaller peptide fragments (data not shown). The sites of proteolytic cleavage were deter- mined to be Lys140 and Ile141 of GmPDIS-1 and Lys139 and Ile140 of GmPDIS-2 by N-terminal sequencing of the trypsin peptide fragments. The N-terminal amino acid sequences of other peptide fragments were AHHHHH, corresponding to the N-terminal histidine tag of the recombinant proteins. We then determined the C-terminal amino acid resi- dues of the peptide fragments by measuring their masses by MALDI-TOF MS. Most cleavage sites resi- ded in two narrow regions, overlapping the putative Soybean protein disulfide isomerase family H. Wadahama et al. 688 FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS boundary regions in GmPDIS-1 and GmPDIS-2 between a and a¢ and a¢ and ERp29c, respectively (Fig. 3). From these results, we concluded that both GmPDIS-1 and GmPDIS-2 have a linear sequence of three domains in an a–a¢–ERp29c pattern. Several mammalian and yeast PDI family proteins are known to function as molecular chaperones. Therefore, we measured the molecular chaperone activity, which prevents the aggregation of amyloid b-peptide (Ab) (1–40) monomers. Such aggregation can be initiated by the addition of ‘seed’, which was obtained by sonication of Ab(1–40) aggregates into 50 lm Ab(1–40) monomers. This aggregation was monitored as an increase in thioflavin T fluorescence. The intensity of the fluorescence increased almost lin- early over 60 min. The seed-dependent aggregation of Ab(1–40) monomers was inhibited by b oth GmPDIS-1 and GmPDIS-2 in a concentration-dependent manner (Fig. 4). In the presence of 2 lm GmPDIS-1 or 0.5 lm GmPDIS-2 (molar ratio of 1 : 25 or 1 : 100 to Ab), almost all Ab aggregation was inhibited for at least 40 min. Tissue distribution and cellular localization of GmPDIS-1 and GmPDIS-2 We prepared antibodies against recombinant GmPDIS-1 and GmPDIS-2. Anti-GmPDIS-1 serum specifically Fig. 1. Multiple amino acid sequence alignment of GmPDIS-1, GmPDIS-2, Arabidopsis PDI-like2-1 (AtPDIL2-1), and alfalfa G1 (MsG1) [64]. A multiple alignment of the polypeptides was generated using CLUSTAL W. Numbers refer to the amino acid number, asterisks indicate amino acid matches, and dashes represent gaps between the sequences. The putative signal sequence (underlined), active site CGHC motifs (sha- ded in black), conserved arginine (shaded in gray) and conserved glutamic acid (boxes) are indicated. H. Wadahama et al. Soybean protein disulfide isomerase family FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 689 immunoreacted to recombinant GmPDIS-1, but not to recombinant GmPDIS-2, whereas anti-GmPDIS-2 serum immunoreacted strongly to recombinant GmPDIS-2 and weakly to recombinant GmPDIS-1 (Fig. 5A, lanes 1–4). Immunoglobulin molecules that immunoreact to GmPDIS-1 in anti-GmPDIS-2 serum were eliminated by pretreatment of the serum with purified recombinant GmPDIS-1 (Fig. 5A, lanes 5 and 6). Hence, we used pretreated anti-GmPDIS-2 serum for the experiments described below. Anti-GmPDIS-1 serum or anti-Gm- PDIS-2 serum was immunoreacted with a single 40 kDa or 38 kDa band in roots, stems, trifoliolate leaves, flow- ers and cotyledons by western blotting (Fig. 5B). The amounts of these proteins in leaves decreased during leaf expansion. GmPDIS-1 and GmPDIS-2 have an N-terminal signal sequence for targeting to the ER, but lack a typical ER-retention signal sequence, like the C-terminal KDEL. We immunostained soybean cotyledons with either rabbit anti-GmPDIS-1 serum or rabbit anti-GmP- DIS-2 serum, and then clarified the subcellular localiza- tion of GmPDIS-1 and GmPDIS-2 by confocal microscopy. The specimens were double-stained with guinea pig anti-BiP serum, as BiP is a well-known ER resident protein [20–22]. To confirm the specificity of the anti-BiP or anti-calreticulin serum, we performed west- ern blotting analysis using soybean protein extracts. Anti-BiP or anti-calreticulin serum immunoreacted with a single 70 kDa or 54 kDa band corresponding to BiP or calreticulin, respectively, in cotyledon extracts (Fig. 6). In the immature cotyledon from an 80 mg bean that was initiating the accumulation of seed storage pro- teins, such as glycinin [23,24] and b-conglycinin [25,26], in its protein storage vacuole (PSV), GmPDIS-1, Gm- PDIS-2 and BiP were localized mainly to the ER (Fig. 7A–D). Interestingly, the PSVs were also slightly stained with anti-BiP serum. To confirm residence of GmPDIS-1 and GmPDIS-2 in the lumen of the ER, microsomes prepared from cotyledon cells were treated with proteinase K in the absence or presence of Triton AB C Fig. 2. Activity of the recombinant GmPDIS-1 and GmPDIS-2. The recombinant GmPDIS-1 (A) and GmPDIS-2 (B) in E. coli (lane 1) were purified by His-tag column chromatography (lane 2), followed by gel filtration chromatography (lane 3). Proteins in each effluent were separated by 10% SDS ⁄ PAGE and stained with Coomassie Blue. (C) PDI activity of the recombinant GmPDIS-1 (left bar) and GmPDIS-2 (right bar). The activity was assayed by the measure- ment of RNase activity produced through the regeneration of the active form from reduced RNaseA. Each value represents the mean of six (GmPDIS-1) or eight experiments (GmPDIS-2). A B Fig. 3. Schematic representation of cleavage sites in GmPDIS-1 (A) and GmPDIS-2 (B) by limited proteolysis. The upper line represents recombinant protein. The boxes below indicate the domain boundaries predicted by an NCBI conserved domain search. The arrows indicate the determined cleavage sites. Black boxes in domain a and a¢ represent the CGHC motif. Soybean protein disulfide isomerase family H. Wadahama et al. 690 FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS X-100. Both GmPDIS-1 and GmPDIS-2 were resistant to protease treatment in the absence of detergent. Because of disruption of microsome membranes in the presence of Triton X-100, GmPDIS-1 and GmPDIS-2 were degraded by the protease treatment (Fig. 8). Sim- ilar phenomena were observed in the case of BiP and calreticulin, which are well known as luminal proteins of the ER. In the cotyledon from the 220 mg bean that heavily accumulated seed storage proteins in its PSVs, BiP was visualized in both the PSV and the ER (Fig. 7E,F). Images of GmPDIS-1 and GmPDIS-2 over- lapped with those of BiP (Fig. 7G,H). GmPDIS-1 associates with proglycinin in the cotyledon cells GmPDIS-1 and GmPDIS-2 were shown to have oxida- tive folding activity in vitro and to be localized to the ER of the cotyledon, suggesting that they may func- tion in protein folding that is accompanied by the for- mation of intramolecular disulfide bonds like those of proglycinin [27]. We then attempted to detect an association between GmPDIS-1 or GmPDIS-2 and glycinin in the cotyledon cells by immunoprecipitation with antibodies against GmPDIS-1, GmPDIS-2 and glycinin after treatment with the protein crosslinker dithiobis[succinimidylpropionate] (DSP). First, we confirmed the immunoprecipitation of GmPDIS-1, GmPDIS-2 and glycinin from the microsomal extract of cotyledons from 150 mg beans by western blotting analysis. The efficiencies of immunoprecipitation of GmPDIS-1 and GmPDIS-2 were not influenced by crosslinking treatment of the microsomes with DSP prior to immunoprecipitation (Fig. 9A,B). Immunopre- cipitation of glycinin acidic subunits was also con- firmed (Fig. 9C). Second, the processing of proglycinin to mature glycinin was monitored by pulse-chase experiments in order to determine the labeling time of glycinin with [ 35 S]methionine and [ 35 S]cysteine. Glyci- nin molecules are synthesized as a single polypeptide chain and associate as trimers in the ER [28,29]. These trimers move to the PSV, where a processing enzyme cleaves them into acidic and basic polypeptide chains [26,30]. Nascent proteins in the isolated cotyledons Fig. 4. Inhibition of Ab aggregation by GmPDIS-1 and GmPDIS-2. Seed-dependent aggregation of 50 l M Ab(1–40) in the pres- ence of GmPDIS-1 (A) or GmPDIS-2 (B) (closed circle, 0 l M; closed square, 0.2 lM; open circle, 0.5 l M; closed triangle, 1.0 lM; open triangle, 2.0 l M). Each value repre- sents the mean of two experiments. AB Fig. 5. Expression of GmPDIS-1 and GmPDIS-2 in soybean tissues. (A) Crossreactivity of the antibody prepared against recombinant GmPDIS-1 or GmPDIS-2 with recombinant GmPDIS-1 (2.4 lg) (lanes 1, 3 and 5) and GmPDIS-2 (2.4 lg) (lanes 2, 4 and 6). Anti-GmPDIS-2* represents anti-GmPDIS-2 serum (1 lL) treated with purified recombinant GmPDIS-1 (5 lg) to eliminate the antibodies that crossreact with GmPDIS-1. (B) Detection of GmPDIS-1 and GmPDIS-2 in soybean tissues. Thirty-microgram samples of protein extracted from the cotyledon (80 mg bean), root, stem, 3 cm length leaf (Leaf-3), 6 cm leaf (Leaf-6), 10 cm leaf (Leaf-10) and flower were separated by 10% SDS ⁄ PAGE and immunostained with anti-GmPDIS-1 serum (lanes 1–7) or anti-GmPDIS-2 (lanes 8–14) serum pretreated with recombinant GmPDIS-1 as described under (A). H. Wadahama et al. Soybean protein disulfide isomerase family FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 691 were metabolically pulse-labeled with [ 35 S]methionine and [ 35 S]cysteine for 15 min and then chased in the presence of cold methionine and cysteine. The labeled glycinin was immunoprecipitated with anti-(glycinin acidic subunit) serum. Immediately after pulse labeling for 15 min, most of the label was in proglycinin (Fig. 9D, lane 1). After a 6 h chase, the labeled pro- glycinin decreased and the processed products, i.e. the acidic and basic subunits of mature glycinin, appeared (Fig. 9D, lane 4). On the basis of these results, we labeled the cotyledons with [ 35 S]methionine and [ 35 S]cysteine for 6 h to detect simultaneously proglyci- nin and the acidic and basic subunits of mature glyci- nin in the immunoprecipitation experiments. After labeling, the microsomes from the cotyledons were treated with the crosslinker DSP, solubilized, and immunoprecipitated with nonimmune, rabbit anti- GmPDIS-1 serum or rabbit anti-GmPDIS-2 serum. The immunoprecipitants were treated with dithiothrei- tol to reduce the disulfide bonds formed by crosslink- ing with DSP, and then were subjected to a second immunoprecipitation with anti-(glycinin acidic subunit) serum. No band corresponding to glycinin was obser- ved in immunoprecipitation experiment with the non- immune serum (Fig. 9E, lane 2), whereas a 50–53 kDa band corresponding to proglycinin was detected in the immunoprecipitant with anti-GmPDIS-1 serum (Fig. 9E, lane 4). These results suggest that GmPDIS-1 molecules associate with proglycinin in the lumen of the ER. As coimmunoprecipitation of proglycinin with GmPDIS-1 was dependent on DSP treatment, GmPDIS-1 may noncovalently associate with proglyci- nin in the ER. On the other hand, slight coimmuno- precipitation of proglycinin with GmPDIS-2 was detected by DSP treatment (Fig. 9E, lane 6). Changes in the levels of GmPDIS-1 and GmPDIS-2 during seed development Very large amounts of seed storage proteins are syn- thesized and translocated to the ER during the matur- ation stage of embryogenesis. Under such conditions, the folding machinery, composed of molecular chaper- ones and foldases, may be strengthened for folding of de novo synthesized seed storage proteins. Therefore, we determined the relationships between changes in expression levels of both GmPDIS-1 and GmPDIS-2 and the synthesis of storage proteins during develop- ment of soybeans by western blotting. In addition, the expression levels of both BiP, a universal ER chaper- one [31], and calreticulin, known as a chaperone for glycoprotein folding [32,33], were determined. Pro- b-conglycinin, possessing an N-terminal prosequence, and proglycinin are transient protein forms that are present in the ER prior to processing in the PSV [28,29]. Hence, the amounts of pro-b-conglycinin and proglycinin are considered to be nearly equivalent to the synthesis levels of both b-conglycinin and glycinin. The synthesis of proglycinin was initiated when the seeds achieved a mass of 50 mg, and increased gradu- ally until they grew to 300 mg (Fig. 10G). On the other hand, the synthesis of pro-b-conglycinin was ini- tiated when the seeds achieved a mass of 40 mg. The synthesis of pro-b-conglycinin increased until the seeds grew to 70 mg and then decreased (Fig. 10E). GmPDIS-1, GmPDIS-2, BiP and calreticulin were expressed in the early stages of embryogenesis (Fig. 10A–D). In the cotyledons of seeds with a mass greater than 100 mg, the levels of GmPDIS-1 and BiP increased proportionally to the synthesis of proglyci- nin. Thus, GmPDIS-1 and BiP may be expressed to enhance the machinery for the folding of seed storage proteins such as proglycinin. However, this event appeared to be independent of transcriptional regula- tion, as the amounts of GmPDIS-1 and BiP mRNA did not correlate with the levels of GmPDIS-1 and BiP expression (Fig. 11A,C). The level of GmPDIS-2 did not correlate with the synthesis of proglycinin (Fig. 10B,G). The amount of GmPDIS-2 mRNA cor- related with the amount of protein (Fig. 11B). The Fig. 6. Analysis of the specificities of anti-BiP serum and anti- calreticulin serum. The purified recombinant BiP (lane 1), calreticu- lin (lane 3) and soybean cotyledon extracts (20 lg of protein) (lanes 2 and 4) were subjected to 10% SDS ⁄ PAGE. Recombinant pro- teins were stained with Coomassie Blue. Cotyledon proteins were immunostained with anti-BiP serum (lane 2) or anti-calreticulin serum (lane 4). Soybean protein disulfide isomerase family H. Wadahama et al. 692 FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS levels of calreticulin (Fig. 10D) and its mRNA (Fig. 11D) were slightly higher in the cotyledons of 70 mg seeds, in which the synthesis of the pro-b-con- glycinin glycoprotein had reached the most advanced stage (Fig. 10E). GmPDIS-1, but not GmPDIS-2, is induced by ER stress Many ER-resident proteins are upregulated by the accumulation of unfolded protein in the ER (i.e. ER A B C D E F G H Fig. 7. Localization of GmPDIS-1 and GmP- DIS-2 in soybean cotyledons. Cotyledons at early (80 mg bean, A–D) or late (220 mg bean, E–H) stages of seed development were immunostained with a combination of anti-b-conglycinin a¢ serum (green) and anti- BiP serum (red) (A, E), anti-(glycinin acidic subunit) serum (green) and anti-BiP serum (red) (B, F), anti-GmPDIS-1 serum (green) and anti-BiP serum (red) (C, G), or anti- GmPDIS-2 serum (green) and anti-BiP serum (red) (D, H). Visible light images collected simultaneously are shown on the right. Asterisks and arrows indicate PSVs and ER networks, respectively. Bars: 10 lm. H. Wadahama et al. Soybean protein disulfide isomerase family FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 693 stress) [34]. Thus, they are the products of UPR genes. Both Arabidopsis PDI-like2-1 and alfalfa G1 expression have been shown to be upregulated by tunicamycin treatment [11,12]. In order to determine whether GmP- DIS-1 and GmPDIS-2 respond to ER stress, we treated soybean cotyledons with tunicamycin, dithiothreitol, and l-azetidine 2-carboxylic acid (AZC), and measured GmPDIS-1, GmPDIS-2, BiP and calreticulin mRNA levels by real time RT-PCR. Both BiP and calreticulin are encoded by well-known UPR genes [35,36]. GmP- DIS-1, BiP and calreticulin expression were upregulat- ed by all treatments with tunicamycin, dithiothreitol, and AZC, suggesting that they are encoded by UPR genes (Fig. 12). There were differences in the extent of GmPDIS-1, BiP and calreticulin induction between the stages of seed development. Under tunicamycin- induced ER stress, GmPDIS-1, BiP and calreticulin transcriptional induction were highest in the most immature cotyledons, whereas under ER stress induced with the proline analog AZC, transcriptional responses were highest in the most mature cotyledons. On the other hand, the expression of GmPDIS-2 was hardly affected by treatment with tunicamycin, dithiothreitol, or AZC. Discussion In this study, we cloned and characterized the cDNAs of GmPDIS-1 and GmPDIS-2 as members of the PDI family. The amino acid sequences of GmPDIS-1 and GmPDIS-2 were similar to each other. Both had two tandem thioredoxin-like domains, a and a¢. Their thio- redoxin-like domain organization occurred in tandem at the N-terminus and was the same as that of mam- malian P5 [37]. However, the C-terminal regions of both GmPDIS-1 and GmPDIS-2 had no sequence similarity to P5. Their C-terminal domains were similar to the C-terminal domain of mammalian ERp29 [16,17]. Recombinant GmPDIS-1 and GmPDIS-2 showed molecular chaperone-like activity that inhibited the aggregation of Ab(1–40). We detected oxidative refolding of the unfolded RNaseA by both recombin- ant GmPDIS-1 and GmPDIS-2. This refolding corres- ponded to 15% and 10% of bovine PDI activity, respectively. The amino acid sequences of the two thio- redoxin domains of GmPDIS-1 and GmPDIS-2 were similar to those of the other PDI family proteins that exhibited an oxidative refolding activity for RNase [38]. In addition, a group of amino acids that are essential for oxidative refolding activity was con- served between GmPDIS-1 and GmPDIS-2. Therefore, it seems possible that the low activities of both GmPDIS-1 and GmPDIS-2 may be due to their low affinity for unfolded RNaseA. Most of the identified PDI family proteins are resi- dents of the ER. GmPDIS-1 and GmPDIS-2 lack the C-terminal ER-retention signal. However, they are colocalized with BiP to the ER. It is unclear whether the existence of ER luminal proteins such as GmPDIS-1 and GmPDIS-2, which lack the KDEL sequence, results from retention or retardation. The importance of the ERp29c domain of Dictyostelium Dd-PDI for ER retention was demonstrated by deletion mutation experiments. In addition, it was demonstrated that the ERp29c domain was sufficient to localize a green fluor- escent protein chimera to the ER [39]. The C-terminal ERp29c domains of both GmPDIS-1 and GmPDIS-2 may possibly play a similar role. Alternatively, these proteins may be retained in the ER by association with other ER-resident proteins, such as BiP. In addition to the localization in the ER, localization of BiP, Gm- PDIS-1 and GmPDIS-2 in the PSVs of the cotyledon from the 220 mg bean was observed. Pimpl et al. indi- cated that BiP was constitutively transported from the ER to vacuoles via the Golgi [40]. The saturation of the HDEL receptor with HDEL or KDEL proteins was assumed to cause the BiP transport to vacuoles via Golgi bodies. Tamura et al. reported that BiP and the 62 kDa PDI that has a KDEL ER-retention signal were constitutively transported to vacuoles in Arabid- opsis cultured cells [41]. In this case, these proteins were presumably transported to the vacuoles independ- ently of the medial ⁄ trans-Golgi complex, as PDI that Fig. 8. Localization of GmPDIS-1 and GmPDIS-2 in the lumen of microsomes. Cytosol (lane 1) and microsomes (lanes 2–4) were iso- lated from cotyledons (100 mg beans). Microsomes were treated with proreinase K (lanes 3 and 4) in the absence (lane 3) or pres- ence (lane 4) of Triton X-100. Proteins (10 lg) were separated by 10% SDS ⁄ PAGE, blotted on a poly(vinylidene difluoride) membrane and immunostained with specific antibodies against GmPDIS-1, GmPDIS-2, BiP and calreticulin, respectively. The asterisk indicates a band of degraded calreticulin. Soybean protein disulfide isomerase family H. Wadahama et al. 694 FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS was located in the vacuoles had high-mannose glycans but not Golgi-processed complex glycans. In soybean cotyledon, two routes were identified for transporting proglycinin from the ER directly to the PSV and from the ER to the PSV via the Golgi [42,43]. It remains unclear, however, how GmPDIS-1, GmPDIS-2 and BiP are transported from the ER to the PSV and what role these proteins play in the PSV. In this study, the association of GmPDIS-1 with pro- glycinin molecules was demonstrated. The proglycinin protomer has two disulfide bonds, an intrachain bond and an interchain bond between the acidic and basic subunits [27]. The importance of the disulfide bonds has been demonstrated for the assembly of glycinins [44]. The structural stability of the mature glycinin molecule required both disulfide bridges to be intact [45]. In general, PDI family proteins catalyze the for- mation of disulfide bonds on nascent polypeptide chains in the ER. Hence, GmPDIS-1 may support proglycinin folding that accompanies the formation of AB C ED Fig. 9. Coimmunoprecipitation of GmPDIS-1 and proglycinin. Confirmation of immunoprecipitation of GmPDIS-1 (A), GmPDIS-2 (B) and glyci- nin (C) with each specific antibody. Microsomes were isolated from cotyledons (150 mg beans) and treated with (+) or without (–) DSP. Pro- teins were extracted and immunoprecipitated with anti-GmPDIS-1 serum (A), anti-GmPDIS-2 serum (B), or anti-(glycinin acidic subunit) serum (C). The proteins extracted from the ER (lane 1) and the immunoprecipitants (lanes 2 and 3) were separated by SDS ⁄ PAGE and immunoblotted with anti-GmPDIS-1 serum (A), anti-GmPDIS-2 serum (B), or anti-(glycinin acidic subunit) serum (C). Asterisks indicate rabbit serum immunoglobulins recovered by the first immunoprecipitation in the immunoprecipitant. (D) Time-dependent processing of proglycinin in the cotyledon. Cotyledons were labeled with Pro-mix L-[ 35 S] in vitro labeling mix for 15 min (lane 1) and chased for 1 h (lane 2), 2 h (lane 3) or 6 h (lane 4) at 25 °C. The extracts from the microsomes were subjected to immunoprecipitation with anti-(glycinin acidic subunit) serum. The proteins in the precipitants were separated by SDS ⁄ PAGE and detected by fluorography. Pro 11S, proglycinin; 11S-A, glycinin acidic subunits; 11S-B, glycinin basic subunits. (E) Coimmunoprecipitation experiments. Cotyledons were labeled with Pro-mix L-[ 35 S] in vitro labeling mix for 6 h. After labeling, microsomes were isolated and treated with (+) or without (–) DSP. The extracts from the microsomes were subjected to immunoprecipitation with nonimmune serum (lanes 1 and 2), anti-GmPDIS-1 serum (lanes 3 and 4), or anti-GmPDIS-2 serum (lanes 5 and 6). The precipitants were treated with dithiothreitol and then subjected to a second immunoprecipitation with anti-(glyci- nin acidic subunit) serum. The final precipitants were subjected to SDS ⁄ PAGE and analyzed by fluorography. H. Wadahama et al. Soybean protein disulfide isomerase family FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 695 disulfide bonds in the ER of the cotyledon cells. In addition, there is a possibility that GmPDIS-1 may function as a molecular chaperone, as GmPDIS-1 had molecular chaperone-like activity. In mammalian cells, PDI family proteins were shown to be present in the folding complex. For example, human PDI is a member of the BiP system, composed of BiP, GRP94, ERdj3, GRP170, etc. [46,47]. Another PDI family, ER- 60 ⁄ ERp57, forms a complex with either calnexin or cal- reticulin to fold N-glycosylated proteins [6,48]. The for- mation of such complexes is thought to facilitate the attachment of the PDI family to substrates, resulting in an increase in the folding rate of the substrate [7,49]. GmPDIS-1 may form a complex with other ER chaper- one proteins in the ER. It was suggested that the expression of both Gm- PDIS-1 and GmPDIS-2 was differentially regulated in the cotyledons during seed development. The level of GmPDIS-1 was dramatically increased in the late sta- ges of seed maturation, in contrast to GmPDIS-2, which was present at a low level during the same stage. These results may suggest the importance of GmPDIS-1 in the folding of proglycinin. The regula- tion of GmPDIS-1 levels during this stage was a post- transcriptional event rather than a transcriptional one. It is unknown how the levels of GmPDIS-1 are con- trolled. The transcriptional responses of the GmPDIS-1 and GmPDIS-2 genes to ER stress were also different. Thus, GmPDIS-1 mRNA, but not GmPDIS-2 mRNA, was induced by treatment with tunicamycin, dithio- threitol, and AZC. Only one ortholog to GmPDISs [PDI-like2-1 (At2g47470)], a UPR gene [11,12], is present on nonduplicated region of chromosome 1 in Arabidopsis [10], and the amino acid sequence identi- ties between PDI-like2-1 and GmPDIS-1 or GmPDIS-2 were 75% and 70%, respectively. Hence, it is pre- sumed that GmPDIS-1 is the soybean ortholog of PDI-like2-1 and that GmPDIS-2 may be a paralogous gene generated from GmPDIS-1 by a gene duplication event. In both rice and maize, the presence of two paralogs has been reported [10]. As GmPDIS-1 and well-known UPR genes, such as BiP and calreticulin, were induced in the cotyledons during seed maturation after all treatments with tunicamycin, dithiothreitol, or AZC, it is assumed that a mechanism that counters ER stress exists in soybean cotyledons. The details of such mechanisms in plants remain unknown. Arabidop- sis and rice orthologs of Ire1, a sensor protein in sign- aling pathways for transcriptional responses against ER stress in yeast and mammals [50], have been identi- fied and shown to be capable of acting as sensors of ER stress in yeast cells [51,52]. In addition, an Arabid- opsis transcription factor, AtbZIP60, has been found to activate promoters through UPR cis -elements under ER stress [53]. We found expressed sequence tag clones that were predicted to be soybean orthologs of Ire1 (AW459105) and bZIP60 (TC226837). It is likely that the pathways composed of these orthologs play main roles in the induction of UPR genes in soybean cotyle- dons, although such cis-elements have not yet been identified. The extent of induction of GmPDIS-1, BiP and calreticulin varied, and was dependent upon the stage of bean development. It is likely that such differ- ences in induction depend on the amount of misfolded protein that accumulates after reagent treatment rather A B C D E F G H Fig. 10. Expression of GmPDIS-1 and GmP- DIS-2 in soybean cotyledons during matur- ation. Thirty micrograms (A–E) or 5 lg (F–H) of proteins extracted from cotyledons were separated by 10% SDS ⁄ PAGE and immuno- stained with specific antibodies against GmPDIS-1 (A), GmPDIS-2 (B), BiP (C), cal- reticulin (D), the prosequence of b-conglyci- nin a¢ (E), b-conglycinin a¢ (F), and glycinin acidic subunits, respectively. Proglycinin (G) and mature glycinin acidic subunits (H) were the 53 and 35 kDa bands on the same blot. Soybean protein disulfide isomerase family H. Wadahama et al. 696 FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS [...]... classes of the protein disulfide isomerase family in plants, including single-domain protein disulfide isomerase- related proteins Plant Physiol 137, 762–778 Kamauchi S, Nakatani H, Nakano C & Urade R (2005) Gene expression in response to endoplasmic reticulum stress in Arabidopsis thaliana FEBS J 272, 3461–3476 Martı` nez IM & Chrispeels MJ (2003) Genomic analysis of the unfolded protein response in Arabidopsis... 2006 FEBS 699 Soybean protein disulfide isomerase family H Wadahama et al onto a polyvinylidene difluoride membrane The GmPDIS-1, GmPDIS1-2, BiP, calreticulin, glycinin acidic subunit, b-conglycinin a¢ and pro-b-conglycinin a¢ proteins were then immunostained with specific antibodies and horseradish peroxidase-conjugated IgG antibody (Promega Corporation), using the Western Lightning Chemiluminescence Reagent... under liquid nitrogen were ground into fine powders with a micropestle SK-100 Proteins were extracted from 100 mg of tissue by boiling for 5 min in 200 lL of SDS ⁄ PAGE buffer [58] containing a 1% cocktail of protease inhibitors (Sigma-Aldrich Inc.) The concentrations of proteins were measured with a protein assay kit (RC DC protein assay; Bio-Rad Laboratories) Proteins were subjected to SDS ⁄ PAGE and... of Escherichia coli thioredoxin in the activation of the mixed disulfide intermediate J Biol Chem 272, 29998–30001 Munro S & Pelham HR (1986) An Hsp70-like protein in the ER: identity with the 78 kd glucose-regulated protein and immunoglobulin heavy chain binding protein Cell 46, 291–300 Munro S & Pelham HR (1987) A C-terminal signal prevents secretion of luminal ER proteins Cell 48, 899– 907 Demmer... sonication in 8 mL of 20 mm Tris ⁄ HCl buffer (pH 8), containing 5 mm imidazole and 0.5 m NaCl (binding buffer), and then centrifuged at 10 000 g for 30 min at 4 °C on an RA-200J rotor by a Kubota 1710 (Kubota, Tokyo, Japan) The supernatant was applied to a column packed with His-Bind resin (EMD Biosciences, Inc.) After washing of the column with binding buffer containing 60 mm imidazole, recombinant proteins. .. Creighton TE, Zapun A & Darby NJ (1995) Mechanisms and catalysts of disulfide bond formation in proteins Trends Biotechnol 13, 18–23 3 Gilbert HF (1998) Protein disulfide isomerase Methods Enzymol 290, 26–50 4 Wetterau JR, Combs KA, Spinner SN & Joiner BJ (1990) Protein disulfide isomerase is a component of the microsomal triglyceride transfer protein complex J Biol Chem 265, 9801–9807 5 Tasanen K, Parkkonen... be important in early embryogenesis, whereas GmPDIS-1 may function in the folding of seed storage proteins and the alleviation of ER stress Because both GmPDIS-1 and GmPDIS-2 are distributed ubiquitouly in other tissues, they might assist folding of various proteins FEBS Journal 274 (2007) 687–703 ª 2006 The Authors Journal compilation ª 2006 FEBS 697 Soybean protein disulfide isomerase family H Wadahama... The obtained microsomal pellet was resuspended in buffer A The suspension was treated with 0.5 mgÆmL)1 proteinase K in the presence or absence of 1% Triton X-100 for 5 min at 4 °C Proteins in the samples treated without Triton X-100 were precipitated with 10% trichloroacetic acid for 30 min at 4 °C Proteins in the cytosol and microsomes were analyzed by western blotting as described above Labeling of... absorption coefficient of a protein Protein Sci 4, 2411–2423 Creighton TE (1977) Kinetics of refolding of reduced ribonuclease J Mol Biol 113, 329–341 Lyles MM & Gilbert HF (1991) Catalysis of the oxidative folding of ribonuclease A by protein disulfide isomerase: dependence of the rate on the composition of the redox buffer Biochemistry 30, 613–619 Soybean protein disulfide isomerase family 58 Laemmli UK (1970)... calreticulin The amplified DNA fragment was subcloned into the ligation-independent cloning site of the pET46Ek ⁄ LIC vector (EMD Biosciences, Inc., San Diego, CA) The recombinant proteins have the Histag linked to the N-terminus An expression plasmid encoding His-tagged mature b-conglycinin a¢ without the signal peptide (Met1–Lys30) or prosequence (Gln31–Lys62) was constructed as follows b-Conglycinin a¢ . Protein disulfide isomerase family proteins involved in soybean protein biogenesis Hiroyuki Wadahama 1, *, Shinya Kamauchi 1, *, Masao. identify 10 classes of the protein disulfide isomerase family in plants, including single-domain protein disulfide isomer- ase-related proteins. Plant Physiol 137,