The Phosphorylation State and Expression of Soybean BiP Isoforms Are Differentially Regulated following Abiotic Stresses*

The mammalian BiP is regulated by phosphorylation, and it is generally accepted that its unmodified form constitutes the biologically active species. In fact, the glycosylation inhibitor tunicamycin induces dephosphorylation of mammalian BiP. The stress-induced phosphorylation state of plant BiP has not been examined. Here, we demonstrated that soybean BiP exists in interconvertible phosphorylated and nonphosphorylated forms, and the equilibrium can be shift to either direction in response to different stimuli. In contrast to tunicamycin treatment, water stress condition stimulated phosphorylation of BiP species in soybean cultured cells and stressed leaves. Despite their phosphorylation state, we demonstrated that BiP isoforms from water-stressed leaves exhibit protein binding activity, suggesting that plant BiP functional regulation may differ from other eukaryotic BiPs. We also compared the induction of the soybean BiP gene family, which con-sists of at least four members designated soyBiP A, soyBiP B, soyBiP C, and soyBiP D, by tunicamycin and osmotic stress. Although all soybean BiP genes were induced by tunicamycin, just the soyBiP A RNA was up-regulated by osmotic stress. In addition,

The endoplasmic reticulum (ER) 1 provides the folding environment that facilitates the acquisition of proper folding and assembly of secretory proteins, a prerequisite for them to move further through the secretory apparatus (for reviews, see Refs. [1][2][3][4]. A set of ER-resident proteins, including molecular chaperones and folding enzymes, associates with newly synthesized polypeptides to assist proper folding and assembly of oligomeric secretory proteins (reviewed in Refs. 1, 3, and 5). The binding protein (BiP) represents one of the best-characterized molecular chaperones from the ER.
The mechanism of BiP binding to and release from nascent polypeptide is believed to be analogous to that described for the cytosolic HSP70 protein (6,7). HSP70-related proteins exist in equilibrium between self-assembled forms and monomers in which the binding site is either free or associated with other proteins (8). This equilibrium is regulated by cycles of ADP/ATP binding and hydrolysis as well as cycles of protein substrate binding and release (9). BiP also exists in interconvertible oligomeric and monomeric forms and is post-translationally modified and regulated by ADP-ribosylation and phosphorylation (10,11). These modifications occur upon the ATP-dependent release of BiP from associated proteins and can be reversed under stress conditions that increase the level of unfolded polypeptides in the ER. Because phosphorylation and ADP-ribosylation appear to be restricted to oligomeric forms of mammalian BiP that are not bound to nascent polypeptides, the monomeric, unmodified form of BiP is thought to be the biologically active species (12). In spinach, three forms of BiP, the 85-kDa monomer, a 280-kDa multimeric form, and a 650-kDa oligomeric form, have been described (13). However, only the oligomeric form of BiP is phosphorylated in vitro, suggesting that, like mammals, the level of functional plant BiP is regulated by post-translational modification.
The ER molecular chaperone proteins are expressed constitutively at low levels in all cells but are induced upon accumulation of unfolded protein in the lumen of the ER (5,7). The expression of folding-defective mutant secretory proteins or treatment of cells with agents that impair protein folding activates a signaling cascade to allow communication between the ER and nucleus (14). This signal transduction pathway, designated the unfolded protein response (UPR) pathway, is characterized by the coordinated transcriptional up-regulation of BiP and other ER proteins that are involved in folding and assembly of nascent proteins. The inter-organelle signaling cascade, which has been elucidated in yeast, involves an ER transmembrane kinase and a basic leucine zipper transcription factor, Hac1p, whose level is modulated by a regulated spliceosome-independent mRNA splicing event (reviewed in Ref. 15).
In plants, like mammals and yeast, the expression of BiP is regulated according to cellular requirements for chaperone activity. Thus, both the increase of secretory activity and the accumulation of unfolded proteins within the ER result in the induction of BiP synthesis in plants (reviewed in Refs. 1 and 16). In the floury-2 mutant of maize, the synthesis of a zein-like storage protein variant, which contains an uncleavable signal sequence, is associated with increased accumulation of BiP (17)(18)(19)(20). Expression of an assembly-defective mutant of the * This research was supported by the Brazilian Government Agency, Financiadora de Estudos e Projetos/Fundo Nacional para Desenvolvimento Científico e Tecnológico, Grant 64.94.0113.00 (to E. P. B. F.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  bean storage protein phaseolin also induces BiP synthesis in tobacco leaf protoplasts (21). Furthermore, tunicamycin, a potent activator of the UPR pathway, efficiently induces BiP expression at both mRNA and protein level in several plant systems (18,22). These results have led to the conclusion that, like mammals and yeast BiP, plant BiP is most likely regulated through an unfolded protein response pathway. This idea is supported by the observation that, as in mammalian cells, overexpression of BiP in tobacco leaf protoplasts attenuates ER stress caused by tunicamycin and prevents activation of the unfolded protein response pathway (23,24). However, in some plant species, specific stress conditions and developmental events alter BiP mRNA and protein levels to different extents, suggesting that post-transcriptional mechanisms are also involved in the regulation of BiP synthesis in plants (13,25). Alternatively or additionally, these discrepancies between the level of BiP mRNA and protein may reflect differential expression and regulation of plant BiP gene families, since the genome of several plant species is represented by multiple BiP genes (25)(26)(27). Both alternatives support the argument that multiple, complex regulatory mechanisms control BiP gene expression in plants.
In soybean, three distinct BiP cDNAs have been isolated from a leaf library (25), and one has been identified in a seed cDNA expression library (28). In this paper, we used twodimensional gel electrophoresis and reverse transcription (RT)-PCR assays to examine the potential BiP regulatory mechanisms in plants. We compared the mechanisms controlling BiP up-regulation by tunicamycin treatment and water stress as well as the phosphorylation state of the induced BiP forms.

EXPERIMENTAL PROCEDURES
Plant Material and Greenhouse Experiments-Soybean plants (Glycine max cv. Cristalina) were germinated in 5-liter pots containing a mixture of soil, sand, and dung (3:1:1) and grown in standardized greenhouse conditions. Plant tissues were harvested, immediately frozen in liquid nitrogen, and stored at Ϫ80°C.
Water stress condition was induced in 40-day-old plants by withholding watering for 8 days before harvesting of leaves. Half of the leaves were used to measure the relative water content (29), and the other half was frozen in liquid nitrogen.
Cell Culture and Induction of Soybean BiP-Cotyledons cells from the soybean variety IAC-12 were cultured as described previously (30). Tunicamycin was added to cultures at 4 days after passage by dilution of a 5 mg/ml stock in Me 2 SO into normal growth medium to 10 g/ml and incubated for the intervals indicated in the figure legends. For water stress, the cells were washed and then cultured with normal growth medium containing 10% (w/v) PEG-8000 (polyethylene glycol), which corresponds to a water potential of Ϫ1.4 megapascals for the indicated intervals.
For two-dimensional gel electrophoresis, SDS-PAGE, and RT-PCR assays, the cells were filtered under vacuum and washed with an isotonic solution (0.25 M NaCl) to remove any adhered medium and PEG. The cells were then frozen in liquid nitrogen before protein and RNA extractions.
Two-dimensional gel electrophoresis was performed as described (33). For the first dimension, 30 -50 g of protein were loaded on isoelectric focusing tube gels (Bio-Rad) in which the pH gradient was established with 80% pH 5-7 and 20% pH 3.5-10 Ampholines (Amersham Pharmacia Biotech). After electrophoresis at 750 V for 3.5 h, the gels were equilibrated with buffer A (5 mM Tris-HCl (pH 6.8), 6 M urea, 30% (v/v) glycerol, and 2% (w/v) SDS) and stored at Ϫ80°C. Prior to the second dimension, the gels were re-equilibrated for 15 min with buffer A containing 65 mM dithiothreitol and then for 15 min with buffer A containing 65 mM dithiothreitol and 260 mM iodoacetamide. SDS-PAGE was carried out as described previously (34), and the proteins were transferred from 10% SDS-polyacrylamide gels to nitrocellulose membranes by electroblotting. Immunoblot analyses were performed using polyclonal anti-BiP-carboxy antibodies (28) at a 1:1000 dilution and a goat anti-rabbit IgG alkaline phosphatase conjugate (Sigma) at a 1:5000 dilution. Alkaline phosphatase activity was assayed using 5-bromo-4-chloro-3-indolyl phosphate (Life Technologies, Inc.) and p-nitro blue tetrazolium (Life Technologies, Inc.).
The pH gradient of the isoelectric focusing gels was determined by incubating 0.2-cm mock-loaded gel slices in 2 ml of H 2 O at room temperature for 12 h with shaking. The pH of each gel slice was measured and plotted as a function of the distance from the anode. Pre-stained molecular markers were electrophoresed in the second dimension and served as a reference point for comparison of the different gels.
For the dephosphorylation assays, total protein extracts were dialyzed against 10 mM Tris-HCl (pH 7.5), 10 mM MgCl 2 , 50 mM potassium acetate, and 4 mM PMSF and treated with 4 units of alkaline phosphatase (Amersham Pharmacia Biotech) for 2 h at 37°C. After phosphatase treatment, the proteins were dialyzed against 40 mM Tris-HCl (pH 7.5) and 1 mM PMSF, concentrated by freeze-drying, and resuspended in lysis buffer. The integrity of the proteins was monitored by SDS-PAGE.
RT-PCR-Total RNA from cells was extracted as described (35) or using an RNAeasy kit (Qiagen). The RNA was treated with 2 units of RNase-free DNase (Promega) in 20 mM Tris-HCl (pH 8.4), 50 mM KCl, and 2 mM MgCl 2 at 37°C for 1 h and recovered by ethanol precipitation. First-strand cDNA was synthesized from 2-5 g of total RNA using the SuperScript II Kit (Life Technologies, Inc.) according to the manufacturer's instructions. PCR assays were performed with soyBiPA-, soy-BiPB-, soyBiPC-(25), or soyBiPD-(28) specific primers (Table I) PCR assays were also carried out with soybean Actin 3 gene-specific primers (GenBank™ accession number J01297) to assess the quantity and quality of the cDNA. The upstream primer 5Ј-cccctcaacccaaaggtcaacag-3Ј (coordinates 614 to 636) and the downstream primer 5Јggaatctctctgccccaattgtg-3Ј (positions 2011 to 2024) amplify a 440-bp fragment from the actin 3 cDNA and a 520-bp fragment, including an 81-bp intron, from genomic DNA.
RNA Gel Blotting-RNA for gel blot analysis was isolated from control, tunicamycin, and water-stressed cells as described (35). Equal amounts of total RNA were denatured by formamide/formaldehyde and resolved on 1% agarose gels containing formaldehyde (36). The RNA was transferred to a nylon membrane by capillary transfer and immobilized by UV cross-linking (Stratalinker, Stratagene). The membrane was hybridized at high stringency conditions (18) using the soyBiPD or soyBiPA cDNA as probes. The hybridization probe was radiolabeled with [␣-32 P]dCTP by random primed labeling (Amersham Pharmacia Biotech). Autoradiography was performed at Ϫ80°C using a Lightning-Plus intensifying screen (Sigma). The amount of RNA and the integrity of ribosomal RNA were monitored by rehybridizing the membranes with a pea 18 S rDNA probe.
Co-immunoprecipitation Assays-All of the procedures were conducted at 4°C. Protein extracts were prepared by homogenization of water-stressed leaves with 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 0.5% (v/v) Triton X-100, 1 mM PMSF, and 2 units/ml apyrase or 2 mM ATP at a ratio of 1 g of the tissue/2 ml of extraction buffer. After incubation of the protein extracts for 10 min, cell debris was removed by centrifugation at 13,000 ϫ g for 15 min.
Immunoprecipitations were performed as described (37) with some modifications. Protein extracts (1 ml) were precleared by incubating with 100 l of 50% (v/v) protein A-Sepharose (Amersham Pharmacia Biotech) and goat anti-rabbit IgG-agarose (Sigma) suspensions in 20 mM Tris-HCl (pH 7.5), 140 mM NaCl for 1 h. After incubation, proteins that bound nonspecifically to protein A or anti-IgG were removed by centrifugation. The precleared supernatant was then incubated with 100 l of anti-BiP-carboxy antibodies (28) for 2 h under agitation at 4°C, followed by incubation with 30 l of 50% (v/v) protein A-Sepharose and anti-IgG-Sepharose suspensions for 4 h. The immunocomplexes were recovered by centrifugation at 8,000 ϫ g for 5 min. Pelleted Sepharose beads were washed extensively with 1 ml of 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 1% (v/v) Triton X-100, and 1 mM PMSF and resuspended in 40 l of SDS-PAGE sample buffer (34). Immunoprecipitated proteins were analyzed by immunoblotting using a chicken anti-WSBiP antibody and a rabbit anti-chicken IgG alkaline phosphatase conjugate (Sigma). The chicken anti-WSBiP antibody was raised against a BiP fraction that was purified from water-stressed leaves essentially as described (18), except that the ATP-agarose affinity chromatography was replaced by an immunoaffinity batch step using an anti-BiP antibody (28) cross-linked to protein A-Sepharose beads (Amersham Pharmacia Biotech). About 50 g of BiP preparations from water-stressed leaves were emulsified with complete Freund's adjuvant and injected subcutaneously under the wing of a chicken. For subsequent injections at 14-day intervals, incomplete Freund's adjuvant was used. After the third injection, eggs were collected daily, and the anti-WSBiP antibody was purified from egg yolk as described (38,39).

The Phosphorylation State of Water Stress-induced BiP Forms Is Distinct from That of Tunicamycin-induced BiP
Forms-Previously we showed that soybean BiP is induced by water stress (28). In water-stressed leaves, BiP synthesis is up-regulated to the same extent as in nutritionally stressed leaves and in pathogen-infected leaves. However, the synthesis of spinach BiP, which is a product of a single gene, is unaffected by water stress (13). In fact, in spinach, water stress resulted in disappearance of the BiP mRNA, although the level of protein remained unaltered. To characterize the water stress regulation of the soybean BiP family, we examined the BiP isoforms in control (relative water content 85%) and in water-stressed (relative water content 50%) soybean leaves by immunoblotting assays of two-dimensional gels (Fig. 1). The pI range of the BiP species detected in water-stressed leaves was more acidic than that of the BiP isoforms from control leaves (compare Fig. 1, B and A). To identify whether water stress resulted in the appearance of phosphorylated BiP isoforms or caused the induction of more acidic, distinct BiP species, the leaf protein extract was treated with phosphatase before electrophoresis (Fig. 1C). Dephosphorylation of BiP caused a shift in the isoelectric focusing migrations of the water-stressed-induced forms (pI 5.4 -5.8) toward a less acidic pI cluster (pI 6.1-6.25) and a single acidic species, pI 6.5 (Fig. 1, B and C). The effectiveness of the alkaline phosphatase treatment was confirmed by complete removal of radiolabeled phosphate incorporated into the protein (data not shown). These results indicated that water stress stimulated phosphorylation of the induced BiP species.
Although our result may suggest that phosphorylation of BiP is part of the ER stress response in plants, it would be in marked contrast with the stress-mediated dephosphorylation of mammalian BiP by tunicamycin treatment of cultured cells (11). To directly compare the induction and phosphorylation state of the water stress-and tunicamycin-induced BiP forms of soybean, we used cultured soybean cells treated with either 10% (w/v) PEG or tunicamycin. After the treatments, the cells were immediately frozen in liquid nitrogen, and whole cell protein extracts were displayed in two-dimensional gels and immunoblotted with an anti-BiP serum (Fig. 2). Apparently all the BiP forms were induced upon treatment of suspension cells with tunicamycin, because the cross-reactive induced proteins were resolved in several isoelectric states ranging from pI 5.8 to 6.5 (compare Fig. 2, A and C, Bs and In). In contrast, treatment of the cells with PEG promoted the induction of a subset of BiP forms, which were resolved in a more acidic isoelectric focusing position cluster, pI 5.4 -5.7 (Fig. 2E, Ac).
The relative differences in the isoelectric focusing migrations between the water stress-and tunicamycin-induced forms prompted us to examine whether their phosphorylation state changed following stress conditions. In untreated control cells ( Fig. 2A), BiP occurred in several isoelectric states that resolved as an acidic cluster (3-4 forms, pI 5.8 -6.1, In) and a more basic species, pI 6.5 (Bs). The acidic cluster represents different phosphorylated forms of the same protein because dephosphorylation assays caused their co-migration as a less acidic single species (compare BiP forms under In in Fig. 2, A  and B). In contrast, phosphatase treatment of the samples did not cause a shift in the isoelectric point of the more basic species (Fig. 2B, Bs). Together, these results indicated that soybean BiP exists as a pool of phosphorylated and nonphosphorylated forms in untreated soybean cells. Because treatment of the cell protein extracts with alkaline phosphatase before electrophoresis did not alter their relative migration and amounts in two-dimensional gels, the tunicamycin-induced BiP forms are nonphosphorylated (Fig. 2, C and D). In contrast, osmotic stress resulted in the appearance of phosphorylated BiP forms, as judged by the complete conversion of the acid induced forms (Fig. 2E, Ac) to less acidic forms upon alkaline phosphatase treatment (Fig. 2F, In). The acid-induced cluster (pI 5.4 -5.7, Ac) migrated at a position consistent with phosphorylated forms of the less acidic cluster (pI 6.0 -6.3, In). These results demonstrated that, unlike tunicamycin, water stress induces BiP phosphorylation in cultured cells.
BiP Isoforms from Water-stressed Leaves Exhibit Protein Binding Activity-The differential post translational modification of water stress-and tunicamycin-induced soybean BiP forms prompted us to investigate whether the phosphorylated water stress-induced species were functional. An inherent property of molecular chaperones is their capacity to associate with protein substrates in an ATP-dependent manner. To analyze the protein binding activity of water stress-induced BiP isoforms, we took advantage of water stress-induced proteins as targets for co-immunoprecipitation assays. In waterstressed leaves, we detected the induction of a 28-kDa polypeptide (Fig. 3, A and B, lanes 3), which was not detected in normal leaves (lanes 1) and insect-attacked leaves (lanes 2). Antibodies prepared against a purified BiP fraction from water-stressed leaves, here referred as anti-WSBiP, cross-reacted with the 28-kDa polypeptide (Fig. 3B, lane 3), whereas the anti-soybean BiP serum prepared against the BiP carboxyl terminus (28), here referred as anti-BiP, was specific to BiP and failed to recognize the water stress-induced 28-kDa polypeptide (lane 5). The 28-kDa protein was not a BiP degradation product, because antibodies prepared against BiP purified from soybean (44) or maize (18) seeds cross-reacted with all soybean BiP isoforms from water-stressed leaves but did not recognize the 28-kDa water stress-induced polypeptide (data not shown). This result indicated that the 28-kDa polypeptide did not share conserved epitopes with BiP. More likely, it was a contaminant of the BiP-purified fraction used to raise the anti-WSBiP antibody. In view of this observation, the cross-reactivity of the anti-WSBiP antibody raised the possibility that the 28-kDacontaminant polypeptide had apparently co-purified with BiP from water-stressed leaves as a result of a previous association between these proteins.
Specific interaction between BiP and the water stress-induced 28-kDa polypeptide was confirmed by immunoprecipita-tion of water-stressed leaf protein extracts with anti-BiP followed by Western blotting with anti-WSBiP (Fig. 3C). Although anti-BiP serum did not cross-react with the 28-kDa polypeptide, this water stress-induced polypeptide was co-immunoprecipitated by the BiP-specific antibody in ATP-depleted conditions (lane 3). The precipitation of the 28-kDa polypeptide by anti-BiP serum was not due to nonspecific interactions, because antibodies to an unrelated protein failed to precipitate the water stress-induced protein (data not shown). Inclusion of ATP in the immunoprecipitation assays prevented BiP⅐28-kDa polypeptide association (lanes 1 and 2). This result is in agreement with previous data showing that the addition of ATP to protein extracts causes the disruption of BiP-substrates complexes (1,2,16,37). In addition, the 28-kDa water stressinduced protein was found to be strongly enriched in microsomal vesicles derived from the endomembrane system (Fig. 3D,  lane 2). The subcellular localization of the in vitro BiP substrate further indicates that association of BiP from waterstressed leaves and the 28-kDa-induced polypeptide may be biologically relevant.

The Kinetics of BiP Induction by Water Stress in Soybeancultured Cells Differs from the Delayed Kinetics of BiP Induction by
Tunicamycin-In the previous experiment, we demonstrated that tunicamycin treatment induced all BiP species, whereas osmotic stress only promoted the accumulation of a subset of soybean BiP isoforms in cell cultures. This observation is consistent with different mechanisms controlling BiP expression under abiotic stresses. As a first step in characterizing the mechanisms by which water stress regulates BiP gene expression, we performed a time course experiment to compare the kinetics of BiP induction by water stress and tunicamycin. Levels of BiP protein and RNA were examined at various times after supplementation of the normal growth medium with PEG or tunicamycin. Although both treatments promoted BiP induction, the kinetics were different (Fig. 4). PEG treatment resulted in increased BiP protein and RNA levels as early as 30 min after the treatment, with maximal accumulation occurring after 2 h (Fig. 4, A and B). In contrast to the rapid induction of BiP by PEG, the induction of BiP by tunicamycin occurred with delayed response kinetics. A slight increase in BiP protein was first detected 6 h post-treatment and increased gradually until saturation after 24 -36 h (Fig. 4D). BiP mRNA induction was initially detected by 2 h post-treatment and reached full induction by 6 -12 h but declined with the accumulation of the protein as the glycosylation block stress persisted (Fig. 4E). As RNA loading controls, the membranes were reprobed with a 32 P-labeled ribosomal cDNA, and the induction level was normalized to the 18 S rRNA signal (Fig. 4, C and F).
Differential Regulation of the BiP Gene Family in Response to Water Stress-The level of BiP induction by PEG was significantly lower than its induction by tunicamycin (Fig. 4). As a possible explanation for this difference in BiP induction, we analyzed the individual contribution of the members of the soybean BiP gene family to the general pattern of BiP mRNA up-regulation by PEG and tunicamycin treatments through RT-PCR with gene-specific primers. The gene-specific primers were designed to take advantage of the most divergent sequences of the known BiP cDNAs (GenBank™ accessions U08382, U08383, U08384, and AF031241), which differ most in their 5Ј-and the 3Ј-untranslated sequences. Consequently, primer sets were designed to amplify small fragments from either 3Ј or 5Ј ends of the genes ( Table I). The set of forward and reverse primers were 100% complementary to the annealing sequences of their cognate cDNA, whereas they were a minimum of four mismatches to the other cDNAs.
The specificity of the primers were further confirmed in control PCR reactions performed with soybean genomic DNA, soyBiPA-and soyBiPD-specific DNA sequences as templates (data not shown). Our results indicate that under the conditions used for the PCR analyses, the A and D primers are gene-specific and do not cross-amplify a non-cognate BiPA or BiPD gene sequence. The sets of B and C primers are capable of discriminating between sequences present in the soybean genome and BiP A-or BiP D-specific sequences. Because soy-BiPB and soyBiPC DNA are no longer available, the specificity of B and C primers were confirmed further by sequencing the amplified fragments from leaf cDNA. For the RT-PCR assays, the integrity and amount of the cDNA prepared from soybeancultured cells were routinely assessed with actin-specific primers. The actin primers were designed to amplify a 580-bp fragment from genomic DNA and a 440-bp fragment from actin cDNA such that the presence of contaminating genomic DNA could be easily assessed in our cDNA preparations.
All RT-PCR reactions were repeated with different numbers of cycles to ensure a quantitative linear amplification (data not shown). In addition, the integrity and amount of the cDNA from different treatments were routinely assessed with actinspecific primers (Fig. 5, Actin). Because the four sets of genespecific primers were effective to detect increased amounts of the amplified fragment from cDNA prepared from RNA of tunicamycin-treated cells (BiPA, BiPC, BiPD, BiPB, lanes T), all soybean BiP mRNAs were up-regulated by tunicamycin. In contrast, osmotic stress did not induce the soyBiPB, soyBiPC, and soyBiPD expression (BiPC, BiPD, BiPB, lanes P) but caused the up-regulation of soyBiPA transcripts (BiPA, lane P). These results suggested that the soybean BiP gene family is differentially regulated by abiotic stresses through distinct signaling pathways. In addition, they may explain the lower level of BiP induction by PEG compared with its induction by tunicamycin (Fig. 4).
To test further the hypothesis that water stress induces the synthesis of BiP by a pathway distinct from the unfolded response pathway, we analyzed whether the simultaneous treatment of the cell with both PEG and tunicamycin promoted a synergistic effect on BiP induction. Analysis of the induction of soyBiPA-specific transcripts by RT-PCR demonstrated that the combination of PEG and tunicamycin treatments had an additive effect on BiP mRNA levels (Fig. 5, BiPA, lane PT). Fig. 6 shows the induction of BiP protein accumulation by tunicamycin (lanes 3 and 4), PEG (lanes 5 and 6), and a combination of both treatments (lanes 7 and 8) in soybean suspension cell cultures. When cells were treated with both PEG and tunicamycin, the two stimuli appeared to act synergistically to increase BiP accumulation. Collectively, these results may indicate that induction of soybean BiP in response to water stress occurs by a nonlinear or distinct pathway from that regulating the BiP response to the accumulation of unfolded proteins in the ER. and protein have been observed in spinach and soybean plants under specific stress conditions or developmental stages (13,25). These observations suggest that post-transcriptional mechanisms and/or differential expression of plant BiP gene family are involved in regulation of BiP synthesis in plants. Both alternatives are consistent with the existence of multiple, complex regulatory mechanisms controlling plant BiP gene expression. In support of this hypothesis, we showed that the soybean BiP gene family exhibits differential regulation in response to abiotic stresses. Three lines of evidence indicated that stimulation of BiP expression by water stress occurs through a pathway distinct from the UPR signaling cascade. First, although all BiP forms were up-regulated by tunicamycin, only a subset of the BiP forms were induced by osmotic stress in cell cultures. Similarly, the mRNA levels of all four soyBiP genes were controlled by tunicamycin, but only the soyBiPA RNA was up-regulated by osmotic stress. Water stress represents the first example of a stimulus that differentially up-regulates BiP genes of the same organism. The absence of soyBiPC, soyBiPB, and soyBiPD induction in PEG-treated cells suggests that the UPR and water-stress regulated pathways are independently controlled. Second, the rapid induction of BiP by PEG was distinct from the delayed tunicamycin-induction kinetics. The difference in the kinetics of BiP induction suggests that different components from the UPR pathway are involved in the signaling pathway that regulates BiP expression under osmotic stress. Finally, treatment of soybean suspension cell culture with both PEG and tunicamycin promoted a synergistic effect on the level of BiP induction. These data indicate that the BiP inductions by osmotic stress and a glycosylation block are additive and support the notion that the regulation of BiP expression by these abiotic stresses functions in distinct signal transudation pathways.
We have also demonstrated that the soybean BiP forms induced by water stress are post-translationally phosphorylated. This result was unexpected because tunicamycin-induced BiP or BiP bound to nascent proteins is unmodified in animal cells, and it is generally accepted that the nonphosphorylated form is the biologically active BiP species in the folding pathway (10 -12). In fact, in tunicamycin-treated soybean cell cultures the induced BiP forms were also found to lack phosphorylation. Thus, the modification of plant BiP protein in response to water stress differs from the usual pattern of posttranslational modifications of eukaryotic BiPs.
Our data do not allow us to reconcile the apparent contradiction between dephosphorylation in response to some ER stresses (like tunicamycin) and phosphorylation in response to water stress. The results are further confounded by the differential induction of BiP gene expression and the apparently uniform phosphorylation of all of the BiP proteins under water deficit. One possibility is that, in addition to its chaperone function, soybean BiP exhibits a regulatory, distinct biological function under osmotic stress. This hypothesis was raised because of the selective up-regulation of soyBiPA mRNA and its rapid induction by osmotic stress, which is in marked contrast with the delayed kinetics of induction along with the coordinated induction of all of the BiP RNAs by tunicamycin. These differences are inconsistent with the water stress response being due to accumulation of denatured proteins in the ER. More likely, induction of soyBiPA mRNA by osmotic stress may represent a primary response to water stress that is activated TABLE I Gene-specific primers The annealing position corresponds to the nucleotide position in the cognate cDNA in which the 5Ј nucleotide of the primer sequence anneals. The numbering scheme was taken considering the first nucleotide of the BiP cDNA sequence in the GenBank™ as the nucleotide ϩ1. F and R following the name of the primers refer to forward and reverse, respectively. ND, not determined.  as soon as the stress is sensed and may accommodate a regulatory function.
Alternatively, the soyBiPA gene may have evolved independent regulatory mechanisms to coordinate efficiently the demand of BiP chaperone function with the induction of a selected group of secretory proteins by osmotic stress, which are probably involved in the osmotic response mechanism (45). Although the relationship between phosphorylation of soybean BiP and central aspects of its molecular function, such as oligomerization and protein substrate binding, has not yet been fully addressed, we showed that BiP isoforms from waterstressed leaves exhibit protein binding activity and associates with a water stress-induced 28-kDa polypeptide. Although the identity of the water stress-induced polypeptide is unknown at present, we provided two lines of evidence suggesting that the association between BiP and the 28-kDa water stress-induced polypepide is not an in vitro artifact and may be physiologically relevant. First, the 28-kDa polypeptide was localized in microsomal membranes composed primarily of endomembrane vesicles derived from the ER, Golgi, and tonoplast. As a secretory protein, the 28-kDa polypeptide is expected to be transiently co-localized with BiP as it enters the ER. Second, the complex BiP⅐28-kDa polypeptide was sensitive to ATP, a property of chaperone-mediated interactions. Therefore, phosphorylation of BiP by osmotic stress cannot be attributed simply to inactivation of induced BiP isoforms. The identification of water stress-induced secretory proteins encoded by early response genes and their association with BiP would provide further support to this possibility.
We demonstrated that soybean BiP exists in interconvertible phosphorylated and nonphosphorylated forms, and the equilibrium can be shift to either direction in response to different stimuli. Although water stress-induced phosphorylation of BiP could be a mechanism to confer different biological function or substrate specificity to induced BiP isoforms, the dehydration process may eventually cause general denaturation of proteins, mimicking the tunicamycin-induced stress condition. These observations raise another possibility that the phosphorylated BiP forms might also serve as a storage pool, which could be rapidly dephosphorylated by unfolded protein as they are accumulated in the ER. This would provide an efficient means to control the effective BiP concentration in the ER before the unfolded protein response was activated. In support of this hypothesis, synthetic peptides have been shown to bind to oligomeric, modified mammalian BiP in vitro and induce their conversion to active, unmodified species (46). Comparison of protein binding activities, oligomerization state, and phosphorylation sites between normal, PEG-induced, and tunicamycininduced BiP forms in cultured cells will allow us to address these possibilities, providing insight into the role of the differential post-translational modification on BiP function during water stress.
Although the overall ER stress response in plants is thought to be similar to that of yeast and mammals, our results suggest that the BiP stress response may differ significantly in plants.
We have identified several key differences in the expression and regulation of the ER molecular chaperone soybean BiP under abiotic stress. These differences may be related to the existence of multiple BiP genes in plants and to the unique challenge that stress conditions represent to plants compared with other eukaryotes. Because plants cannot avoid environmental changes, they are constantly subjected to a variety of stress conditions. Acclimation to environmental changes requires responses against cell damages, such as preservation of membrane and protein structures, which enable the plant to tolerate and minimize the deleterious effect of abiotic stress.
Possibly the members of the plant BiP gene family have evolved independent regulatory mechanisms to ensure a high level of expression under a broad range of biotic and abiotic stress conditions to protect the plant against cell damage. In this context, it would be interesting to know whether other soybean ER molecular chaperones follow similar response to osmotic stress.