Polypeptide Components of Oligomeric Legumin-Like ... - Springer Link

Journal of Protein Chemistry, Vol. 22, No. 2, February 2003 (© 2003)

Polypeptide Components of Oligomeric Legumin-Like Thiamin-Binding Protein from Buckwheat Seeds Characterized by Partial Amino Acid Sequencing and Photoaffinity Labeling Maria Rapala-Kozik,1,3 Katarzyna Ostrowska,2 Katarzyna Bednarczyk,1 Robert Dulinski,1 and Andrzej Kozik1 Received December 6, 2002

Among thiamin-binding proteins that ubiquitously occur in plant seeds, that of common buckwheat became a model of extensive studies of the chemical mechanism of ligand-protein interaction. In this work, the polypeptide components of buckwheat seed thiamin-binding protein (BSTBP) are identified and characterized. We suggest that BSTBP is probably a fraction of major storage 13 S globulin (legumin), has an average molecular mass of 235 kDa and comprises hexamers of 57-kDa and 38-kDa subunits in variable combinations. Each subunit is a pair of disulfide-linked polypeptide chains, 36 kDa plus 24 kDa and two-times 22 kDa, respectively. The N-terminal sequences of 22-kDa and 24-kDa components show strict homology with those reported for “basic subunits” of buckwheat legumin. By photoaffinity labeling of BSTBP with 4-azido-2-nitrobenzoylthiamine, it is shown that the 36-kDa chain plays the major role in thiamin binding, but the other chains may also be variably involved. Putative thiamin-binding fragments are identified and sequenced. KEY WORDS: Thiamin-binding proteins; storage globulins; legumins; photoaffinity labeling; buckwheat seeds.

Plant seeds contain specific proteins that form noncovalent complexes with thiamin (Mitsunaga et al., 1986a; Kozik and Rapala-Kozik, 1995, 1996). These thiamin-binding proteins (TBPs)4 probably play a physiological role of the store of thiamin, which is apparently essential for seed germination and seedling growth (Mitsunaga et al., 1987; Shimizu et al., 1990). Homogeneous TBP preparations were obtained from numerous species (Adachi et al., 2000; Adamek-Swierczynska et al., 2000; AdamekSwierczynska and Kozik, 2002; Mitsunaga et al., 1986b;

Nishimura et al., 1984; Rapala-Kozik and Kozik, 1999; Shimizu et al., 1995; Watanabe et al., 1998a). Their basic molecular properties such as the native and subunit molecular masses, isoelectric points, amino acid compositions, etc. were reported. Except for the unique thiaminbinding albumins of sesame seeds (Shimizu et al., 1995) and of pea seeds (Adamek-Swierczynska and Kozik, 2002), all other seed TBPs belong to the globulin class but their actual status among the major seed storage globulins remains to be established. In each case, the basic parameters of thiamin-protein interaction (the binding capacities and the dissociation constants) were also determined. The

1

4

1. INTRODUCTION

Department of Analytical Biochemistry, Faculty of Biotechnology, Jagiellonian University, Gronostajowa 7, 30-387 Kraków, Poland. 2 Department of Organic Chemistry, Faculty of Chemistry, Jagiellonian University, Ingardena 3, 31-045 Kraków, Poland. 3 To whom correspondence should be addressed at: Jagiellonian University, Faculty of Biotechnology, Gronostajowa 7, 30-387 Kraków, Poland. E-mail: [email protected]

Abbreviations: ANBT, 4-azido-2-nitrobenzoylthiamin; BSTBP, buckwheat seed thiamin-binding protein; DTT, dithiothreitol; HPLC, high-performance liquid chromatography; PAGE, polyacrylamide gel electrophoresis; PVDF, polyvinylidene difluoride; SDS, sodium dodecyl sulfate; TBP, thiamin-binding protein; TEMED, N,N,N⬘,N⬘tetramethylethylenediamine; TFA, trifluoroacetic acid; TPCK, N-tosylL-phenylalanine chloromethyl ketone.

167 0277-8033/03/0200-0167/0 © 2003 Plenum Publishing Corporation

168

Rapala-Kozik, Ostrowska, Bednarczyk, Dulinski, and Kozik

detailed chemical characterization of ligand-protein interaction was performed on TBP from seeds of buckwheat (Fagopyrum esculentum Moench) and, for that protein only, a conceptual model of thiamin-binding center was proposed (Kozik, 1995; Rapala-Kozik and Kozik, 1992, 1996; Rapala-Kozik et al., 1999). However, this protein itself is poorly characterized, with controversies in the literature even about its very basic molecular properties such as the subunit composition or the native molecular mass (Kozik, 1995; Mitsunaga et al., 1986b; RapalaKozik and Kozik, 1992; Watanabe et al., 1998b). The present work aimed at the characterization of polypeptide chains that built the oligomeric molecule of buckwheat seed thiamin-binding protein (BSTBP). The major polypeptide components of BSTBP were identified and their partial N-terminal amino acid sequences were determined, clearly showing that BSTBP seemed to be a fraction of the major buckwheat seed storage legumin. Additionally, the protein was affinity labeled with a photoactivatable thiamin derivative to determine the partial sequences that may contribute to the structure of thiamin-binding center. 2. MATERIALS AND METHODS 2.1. Chemicals 4-Azido-2-nitrobenzoylthiamine (ANBT) was synthesized according to the published method (Sempuku, 1988). The preparation obtained seemed to be at least 95% pure as analyzed by high-performance liquid chromatography (HPLC) (Kozik and Rapala-Kozik, 1993). Acrylamide, bisacrylamide, dithiothreitol (DTT), glycine, molecular mass standard mixture for gel filtration, sodium dodecyl sulfate (SDS), Tris, and tricine were purchased from Bio-Rad (Richmond, CA). Ammonium persulfate, Coomassie Brilliant R-250, molecular mass standards (14,000–70,000 Da) for electrophoresis, thiamin, N-tosyl-L-phenylalanine chloromethyl ketone (TPCK)treated trypsin, N,N,N⬘,N⬘-tetramethylethylenediamine (TEMED), and trifluoroacetic acid (TFA) were from Sigma (St. Louis, MO). Polyvinylidene difluoride (PVDF) membrane (ProBlott Membrane) was from Applied Biosystems, and HPLC-grade solvents were from Merck (Darmstadt, Germany). Standard chemicals, of at least ACS grade, were purchased from Merck or Sigma. 2.2. Buckwheat Seed Thiamin-Binding Protein BSTBP was purified by the method of Mitsunaga et al. (1986b). The preparations obtained appeared to be

homogeneous in polyacrylamide gel electrophoresis (PAGE) in nondenaturing system of Davis (1964) and had a thiamin-binding capacity of 9.5–11 nmol thiamin bound/mg protein. Protein was determined by the method of Lowry et al. (1951), and thiamin-binding capacity was determined by the saturation ligand-binding assay involving the ultrafiltration and flow-injection fluorometry (see below).

2.3. Electrophoretic Techniques “Native” PAGE was performed according to Davis (1964) in 7% separating gel. Denaturing PAGE in the presence of SDS (SDS-PAGE) was carried out according to the system of Schägger and von Jagow (1987), with 10% separating gel and without “spacer” gel. For both electrophoretic techniques, a Bio-Rad MiniProtean II apparatus was applied and the manufacturer’s instructions were followed. The gels were routinely stained with Coomassie Brilliant Blue R-250. For molecular mass estimations, a mixture of standard proteins (Sigma) containing ␣-lactalbumin (14,200 Da), soybean trypsin inhibitor (20,100 Da), trypsinogen (24,000 Da), carbonic anhydrase (29,000 Da), glyceraldehyde-3-phosphate dehydrogenase (36,000 Da), ovalbumin (45,000 Da), and bovine albumin (66,000 Da) was used. Individual polypeptide components of BSTBP were obtained by electroelution from unstained gels after SDSPAGE. A Bio-Rad Model 422 electroeluter was used. The elution was carried out in 25 mM Tris, 192 mM glycine, 0.1% SDS buffer, for 3 hr at a current of 10 mA per tube. Blots of unstained gels after SDS-PAGE of BSTBP on PVDF membrane (Matsudaira, 1987) were prepared using a Bio-Rad Mini Trans-Blot apparatus and were stained with Coomassie Brilliant R-250.

2.4. N-Terminal Sequencing N-terminal amino acid sequences of protein bands dissected from PVDF blots or of tryptic peptides purified by HPLC (see below) were determined using a gas-phase Model 491 sequencer (Perkin Elmer–Applied Biosystems, Foster City, CA) in the BioCenter Instrumental Facility (Jagiellonian University, Kraków, Poland). The phenylthiohydantoin derivatives of amino acids were analyzed with an HPLC system containing a Microgradient Delivery System Model 140C and a Programmable Absorbance Detector Model 785A (Perkin Elmer–Applied Biosystems).

Polypeptide Components of Oligomeric Legumin-Like Thiamin-Binding Protein

169

2.5. Gel Filtration

2.8. Flow-Injection Fluorometry

High-resolution gel filtration of BSTBP was performed on a Superdex-200 HR 10/30 column (Pharmacia, Uppsala, Sweden) eluted with 0.05 M sodium phosphate buffer, pH 8.0, containing 0.15 M NaCl at a flow rate of 0.5 ml/min at ambient temperature. The chromatographic system used consisted of (i) a Well Chrom K-1001 HPLC pump (Knauer, Germany), (ii) a Rheodyne 9725 injector equipped with a 200 ␮l sample loop, (iii) a UV-1 photometric monitor (Pharmacia), and (iv) Chromed PC software (Poznan, Poland) for data acquisition and analysis. The column was calibrated with a Bio-Rad molecular mass standard mixture containing thyroglobulin (670 kDa), ␥-globulin (158 kDa), ovalbumin (44 kDa), and myoglobin (17 kDa).

Thiamin and ANBT were routinely determined by a semiautomated version of standard thiochrome method using the following equipment. A carrier liquid (water) and an oxidizing reagent [0.0025% potassium hexacyjanoferrate(III) in 2.17% sodium hydroxide] were pumped at a flow rate of 0.8 ml/min per channel with a Gilson Minipuls 3 two-channel peristaltic pump. Samples were injected in 2-min intervals into the water stream through a Rheodyne 7125 injector equipped with a 50 ␮l sample loop. Both liquids met in a tie and then flew through a knitted 3 m ⫻ 0.5 mm Teflon delay tube (Supelco) to the flow cell of a Shimadzu RF-535 fluorescence monitor set at 365-nm excitation wavelength and 430-nm emission wavelength. The detector signal was acquired and analyzed with a Chroma software (PolLab, Warsaw, Poland). The quantification was based on the peak areas.

2.6. Dynamic Light Scattering The size of native BSTBP molecule was determined using a DynaPro MS-800 dynamic light scattering instrument (Protein Solutions, Charlottesville, VA) equipped with a temperature control module and a Dynamics v. 5.25.44 software for data acquisition and analysis. The fluctuations of scattered light were measured at 20°C, from a 12-␮l cell with BSTBP samples in 0.05 M sodium phosphate buffer, pH 8.0, containing 0.15 M NaCl. 2.7. Ligand-Binding Assays The parameters of BSTBP interaction with thiamin or with ANBT were usually determined by isotope-free saturation binding experiments, involving the ultrafiltration for separation of free ligand from ligand-protein complex (Mickowska et al., 2000). In a single experiment, a series of BSTBP samples (0.5 ml, approximately 0.1 mg protein) in 0.05 M sodium phosphate buffer, pH 8.0, were incubated with increasing ligand concentrations (10⫺6–10⫺4 M) at room temperature for 2 hr, and then the samples were briefly filtered through the Nanosep-10 (Pall Filtron) protein centrifugal microconcentrators. In the filtrates collected (approximately 0.1 ml), the free ligand concentration was determined by flow injection fluorometry (see below). For the determination of ANBT binding to BSTBP by an alternative thiamin-displacement method (Mickowska et al., 2000), BSTBP samples were incubated with constant amount of thiamin (4 ⫻ 10⫺7 M) and increasing concentrations of ANBT (10⫺6–10⫺3 M). After ultrafiltration, free thiamin was determined in the filtrates by HPLC as described previously.

2.9. Photolabeling of BSTBP with ANBT BSTBP (approximately 1 mg/ml) was incubated with ANBT (10⫺6 M, 10⫺5 M, or 10⫺4 M) for 2 hr in the dark. The samples were then placed on ice and illuminated with a Vilber Lourmat (Cedex, France) Model VL-6-LC UV lamp (365 nm, 6 W) from a distance of approximately 30 cm for 10 min. Finally, the samples were dialyzed against 0.05 M sodium phosphate buffer, pH 8.0, for 24 hr at 4°C. Qualitative visualization of the label distribution between protein bands after SDS-PAGE was possible by immersing the unstained gel in 0.125% potassium hexacyjanoferrate(III) in 2.17% sodium hydroxide and the observation of the bluish band fluorescence under the UV lamp. For quantitative estimation of the label content, the bands were dissected from the unstained SDSPAGE gels, the protein was electroeluted, and the samples were analyzed for thiochrome-like fluorescence by flow-injection fluorometry.

2.10. Isolation of Thiamin-Labeled Peptides from Tryptic Digest of ANBT-Labeled BSTBP ANBT-labeled BSTBP (approximately 0.5 mg/ml) in 0.05 M sodium phosphate buffer, pH 8.0, was incubated with TPCK-treated trypsin (20 ␮g/ml) for 24 hr at 37°C. The digest was then analyzed by HPLC in a system commonly used for peptide separations. A Supelcosil LC-318 (ODS) column (4 mm ⫻ 250 mm) equipped with a Supelguard LC-318 cartridge precolumn (4 mm ⫻ 20 mm)

170


was applied. The solvent A was 0.1% trifluoroacetic acid (TFA) in water, and the solvent B was 0.08% TFA in 80% acetonitrile. The separation was performed in a gradient of B from 10% to 70% in 40 min, at a flow rate of 1 ml/min at ambient temperature. The HPLC instrument used consisted of (i) a Shimadzu DGU-14A solvent degasser, (ii) a Shimadzu low-pressure quaternary gradient solvent delivery unit (LC-9A HPLC pump and FCV-9AL proportioning valve), (iii) a Knauer Model A0263 injector equipped with a 100 ␮l sample loop, (iv) a Merck-Hitachi L-4000A UV detector, (v) a Shimadzu RF-535 fluorescence detector, and (vi) Schimadzu Class VP (v. 4.0) hardware/software package for the pump control and data acquisition and analysis. The absorbance of the eluate was monitored at 215 nm. In pilot, analytical separations, downstream of UV detector the eluate was mixed in a stainless steel tie with the oxidizing solution of 0.005% potassium hexacyjanoferrate in 2.17% sodium hydroxide, pumped at a flow rate of 0.8 ml/min by the Gilson Minipuls 3 peristaltic pump. The mixture flew through a knitted 3 m ⫻ 0.5 mm Teflon delay tube before reaching the flow cell of fluorescence detector. The excitation wavelength was set at 365 nm, and the emission wavelength at 430 nm. After identification of fluorogenic (i.e., thiamin-labeled) peptides, the postcolumn derivatization system and the fluorescence detector were removed. Separations were then performed with the absorbance detection only; the thiamin-labeled peptides were collected, and the samples were lyophilized and subjected to amino acid sequence analysis.

3. RESULTS 3.1. Characterization of BSTBP Components by SDS-PAGE Two major bands of apparent molecular masses of 57 kDa and 38 kDa were detected by SDS-PAGE in all BSTBP preparations provided that the samples were not reduced (Fig. 1). As might be judged from band intensities, the smaller species was in a nearly twofold excess over the larger one in most of our BSTBP preparations. While a number of minor bands of smaller molecular mass were also seen, they were estimated not to exceed 15% of all protein. In the presence of a reducing agent (DTT), a band of apparent molecular mass of 36 kDa, and another, stronger and apparently double band of approximately 23 kDa were detected. When the 57-kDa and 38-kDa components, separated in the absence of the reducing agent, were electroeluted from the gel and then individually reanalyzed after reduction, the 57-kDa

Fig. 1. SDS-PAGE characteristics of buckwheat seed thiamin-binding protein (BSTBP). The discontinuous system of Schägger and von Jagow (1987) with 10% separating gel was applied. Lane 1, BSTBP sample, without reduction; lane 2, BSTBP sample reduced with DTT; lane 3, BSTBP larger subunit, electroeluted from unstained gel after SDS-PAGE (without reduction) and subjected (without reduction) to the next SDS-PAGE; lane 4, the same sample but reduced for the last SDS-PAGE; lane 5, BSTBP smaller subunit, electroeluted from unstained gel after SDS-PAGE (without reduction) and subjected without reduction to the next SDS-PAGE; lane 6, the same sample, reduced for the last SDS-PAGE; lane 7, molecular mass standards.

component presented two 36-kDa and 24-kDa bands and the 38-kDa component showed a single though diffused 22-kDa band. 3.2. Partial N-Terminal Amino Acid Sequencing of BSTBP Polypeptide Chains All individual polypeptide chains that seemed to constitute the BSTBP molecule were subjected to automated N-terminal sequencing based on Edman degradation. The determined N-terminal sequences of the 22-kDa chain (first 33 residues) and the 24-kDa chain (first 15 residues) were GLEQAFXNLKFKQNVNRPSRADVFNPRAGRINT and GLEESFXNLR(F/Q)RQNL, respectively. These sequences were very similar to each other; at least 8 of the first 15 residues were identical. Attempts to sequence the 36-kDa chain were rather unsuccessful as only low levels of some short sequences, e.g., STEGQQQGN, could be detected. Hence, the N terminus of this component seemed to be blocked. 3.3. Determination of the Native Molecular Mass of BSTBP Estimations of the molecular mass of native BSTBP molecule by high-performance gel filtration on Superdex200 column (results not shown) gave results slightly variable between preparations, in a range of 220–254 kDa (7 BSTBP preparations). This variability seemed to be

Polypeptide Components of Oligomeric Legumin-Like Thiamin-Binding Protein

171

correlated with the ratio of 57-kDa and 38-kDa bands in the SDS-PAGE pattern. An independent estimation of the size of native BSTBP molecule by a physical method of dynamic light scattering (results not shown) gave the values of hydrodynamic radius in the range of 5.85–6.46 nm (five preparations), which corresponded to a molecular mass range of 210–265 kDa, based on the calibration curve provided by the instrument manufacturer. The mean ⫾ SD value from 12 independent estimations (all preparations and both methods used) was 235 ⫾ 22 kDa. 3.4. Characteristics of ANBT Binding to BSTBP ANBT, a photoactivatable thiamin derivative, had previously been applied for covalent labeling of a thiamintransport protein of yeast cell membrane (Sempuku, 1988). The chemical structure of ANBT is shown in Fig. 2; this compound contains the arylazide group, which becomes highly reactive under near UV illumination. In the dark, ANBT strongly bound to BSTBP (Fig. 3, plot 1), with an apparent dissociation constant of 4.9 ␮M. Comparable values of the dissociation constant were repeatedly reported for the complex of BSTBP with thiamin (Mitsunaga et al., 1986b; Rapala-Kozik and Kozik, 1992; Rapala-Kozik et al., 1999). Moreover, ANBT effectively displaced thiamin from the complex with BSTBP (Fig. 3, plot 2). When the mixtures of BSTBP and ANBT were illuminated, the subsequent tests for thiamin binding by dialyzed protein showed irreversible inactivation of BSTBP (Fig. 3, plots 3-6). However, when the illuminated samples additionally contained excess of thiamin (2 ⫻ 10⫺3 M), the protein, after removing all small reactants by dialysis, was able to bind thiamin with the affinity and to the capacity comparable to those of the native BSTBP (results not shown). 3.5. Photoaffinity Labeling of BSTBP with ANBT BSTBP was photolabeled with ANBT and subjected to SDS-PAGE without sample reduction to separate the 57-kDa and 38-kDa subunits or, after reduction

Fig. 3. Scatchard plots for the binding of thiamin and 4-azido-2nitrobenzoylthiamine (ANBT) to native or ANBT-labeled buckwheat seed thiamin-binding protein (BSTBP). Protein samples were prepared in 0.05 M sodium phosphate buffer, pH 8.0. The determinations of ANBT binding to native BSTBP (in the dark) were based on the saturation of protein with ANBT (1) or on the displacement of thiamin by ANBT from the protein (2). The saturation type experiments on thiamin binding were performed for native BSTBP (3) and for samples of BSTBP that was photolabeled with 1 ␮M ANBT (4), 10 ␮M ANBT (5), or 100 ␮M ANBT (6). The following binding parameters, the dissociation constants (␮M) and the thiamin-binding capacities (nmol thiamin bound/mg protein), were determined: (1) 4.9 and 7.8; (2) 5.9 and 9.4; (3), 4.4 and 10.5; (4) 12.4 and 6.4; (5) 41.7 and 4.2; (6) 153 and 2.1. B and F denote the bound and free ligand concentrations, respectively.

with DTT, to separate the 36-kDa and 22/24-kDa chains. Individual BSTBP components were then electroeluted from the gel. The content of thiamin label in each sample was estimated by flow-injection fluorometry. In terms of total fluorescence intensity, the thiamin label was nearly equally distributed between the 57-kDa and 38-kDa subunits or between the 36-kDa and 22/24-kDa chains. However, the ratio of fluorescence intensity to apparent protein content in the band (Fig. 4) was highest for the 36-kDa chain and, second, for the 58-kDa subunit. When ANBT labeling was performed in the presence of the excess of thiamin (2 ⫻ 10⫺3 M), the label content was lowered by at least 60% in the 57-kDa subunit and by 70% in the 36-kDa chain but only by 35% and 31% in the 38-kDa subunit and the 22/24-kDa chains, respectively.

3.6. Detection of Putative Thiamin-Binding Peptides of BSTBP

Fig. 2. Structural formula of 4-azido-2-nitrobenzoylthiamine (ANBT).

The ANBT labeled protein was digested with trypsin and the derived peptide fragments were separated by HPLC. The fragments bearing the thiamin label were

172


Fig. 4. The distribution of thiamin label between polypeptide components of 4-azido-2-nitrobenzoylthiamine (ANBT)-labeled buckwheat seed thiamin-binding protein (BSTBP). Protein was photolabeled with 100 ␮M ANBT (filled bars) or with 100 ␮M ANBT in the presence of 2 ⫻ 10⫺3 M thiamin (open bars) and subjected to SDSPAGE in two variants, with or without sample reduction. Polypeptide components were electroeluted from unstained gels and analyzed for thiochrome-like fluorescence by flow injection fluorometry. The ratios of fluorescence intensity to protein content in the band are expressed relatively to that obtained for the 36-kDa chain.

detected by post-column oxidation to thiochrome-like derivative and the fluorometric detection (Fig. 5). Of three major peaks on the fluorescence trace, the last one (3) seemed to be of free ANBT. The peaks 1 and 2 corresponded to peptide fragments, which were collected and subjected to the automated N-terminal sequence analysis. The sequences GDEEPQQGF and MDDEXVLEWMKDI were determined for peptides 1 and 2, respectively.

4. DISCUSSION 4.1. Relation of BSTBP to the Major Buckwheat-Seed Storage Legumin All isolated seed TBPs had basic molecular properties similar to those of the major seed storage proteins of the given species. The straightforward verification of those similarities should be the determination of at least partial amino acid sequences of TBP polypeptide chains and the homology searching within the available protein databases. Studies of that type were performed on unique sesame seed TBPs, which were clearly shown to be 2 S albumins (Watanabe et al., 1999, 2001). Recently, a weak thiaminbinding activity was assigned to the major pea seed albumin PA2 (Adamek-Swierczynska and Kozik, 2002), but a physiological significance of that finding remains obscure. Nevertheless, most of TBPs are globulins. Albeit recently partial N-terminal sequences of all subunits of a sunflower

Fig. 5. HPLC analysis of a tryptic digest of 4-azido-2-nitrobenzoylthiamine (ANBT)-labeled buckwheat seed thiamin-binding protein (BSTBP). Protein was labeled with 100 ␮M ANBT and treated with TPCK-treated trypsin (1:25 enzyme/substrate mass ratio) in 0.05 M sodium phosphate buffer, pH 8.0, for 24 hr at 37°C. The digest was then analyzed on an analytical (4 mm ⫻ 250 mm) ODS column in a standard TFA/water/acetonitrile peptide-separation system. The eluate was analyzed by monitoring the absorbance at 215 nm (dashed trace) or by postcolumn oxidation and monitoring the thiochrome-like fluorescence (solid trace).

seed thiamin-binding globulin were determined, they did not appear to show significant homologies to any known seed protein (Watanabe et al., 2002). In this work we could unequivocally determine the partial N-terminal sequences of two from at least three polypeptide chains that seemed to constitute the oligomeric molecule of BSTBP. The BLAST homology searching revealed a practical identity of the N-terminal sequence (33 residues) of 22-kDa BSTBP chain to an internal fragment of a large (565 amino acid residues) “legumin-like” buckwheat protein (NCBI database accession Nos. T10696, BAA21758, and AAD32713), also known as the “major allergen” of this species. The gene, encoding the latter protein, was shown to be expressed during buckwheat seed development (Fujino et al., 2001). According to the possibly universal mechanism of legumin-precursor processing (Casey, 1999; Müntz, 1996), this protein might be predicted to generate a 41-kDa Nterminal chain (“␣” or “acidic”) and a 21-kDa C-terminal chain (“␤” or “basic”) by a cleavage after an asparagine residue just upstream of the fragment homologous to the N-terminal sequence of 22-kDa BSTBP chain. The N-terminal sequence (15 residues) of 24-kDa BSTBP chain was very similar to that of 22-kDa chain. Moreover, both sequences showed a clear homology to another reported N-terminal sequence (GIDENVCTMKLRENI) of the basic 26-kDa subunit of major buckwheat legumin (Rout et al., 1997). The seventh residue in our BSTBP sequences probably was cysteine,

Polypeptide Components of Oligomeric Legumin-Like Thiamin-Binding Protein which could not be determined by the sequencing procedure applied. This residue is evolutionary conserved in legumins (13 S storage globulins) and forms the only disulfide bridge that links the acidic and basic chains (Casey, 1999; Müntz, 1996). The fourteenth asparagine residue and the G-(L/I)-(E/N)-E motif are also highly conserved among the flowering plants (Rout et al., 1997). The above data strongly suggest that the 22-kDa and 24-kDa BSTBP components have properties of “typical” basic (␤) legumin chain and that they may originate from precursors common to those of the major buckwheat seed legumin. The major buckwheat seed storage 13 S globulin (legumin) is actually a spectrum of proteins that have at least eight types of subunits of molecular masses ranging from 43 kDa to 68 kDa (Radovic et al., 1996). Each subunit is a pair of an acidic chain (32–43 kDa) and a basic chain (20–23 kDa) linked via a single disulfide bridge (Radovic et al., 1996). The 57-kDa BSTBP subunit, comprising the 36-kDa acidic chain and 24-kDa basic chain, seems to fit that scheme. In principle, these chains should originate from the same, single precursor molecule (Müntz, 1996). Unfortunately, that assumption could not be verified in our study as the N terminus of the 36-kDa chain seemed to be blocked. A significance of the “residual” sequence STEGQQQGN detected in this band was uncertain; it could result from some proteolytic truncations near the N terminus of the original 36-kDa chain that might have occurred during the BSTBP purification procedure. The motif (E/Q)GQQ(Q/E)G could be found in many seed storage proteins, including acidic chains of several legumins. On the other hand, no significant similarity of the STEGQQQGN sequence to any fragment of “acidic” part of the buckwheat seed legumin precursor (Fujino et al., 2001) could be detected. In contrast to the 57-kDa BSTBP component, the 38-kDa species seems to be atypical as composed of two chains of similar size. Possibly, the 38-kDa subunit is actually a heterodimer, containing both the basic and acidic chains, the latter having the N terminus blocked and being truncated at the C terminus so that its molecular mass is coincidentally similar to that of the basic chain. A partial confirmation of that hypothesis could be obtained by SDS-PAGE analysis of tryptic digestion of BSTBP (results not shown). At early digestion stages, the 57-kDa band rapidly disappeared, leaving apparently untouched the 38-kDa band and relatively insignificant amounts of small digestion products, suggesting a conversion of the large subunit into a species of approximately 38 kDa. SDS-PAGE characteristics of BSTBP did not depend on the presence of proteinase inhibitor cocktail over the purification procedure. Thus, it could be

173

hypothesized that the 38-kDa BSTBP subunit might be produced from a large, “typical” legumin subunit of approximately 60 kDa due to the action of some endogenous seed proteinases. On the other hand, in one report (Maksimovic et al., 1996) it was suggested that buckwheat might represent an exceptional case of the separate synthesis of the acidic and basic polypeptide chains of a legumin-type protein. If that holds true, the assembly of the acidic and basic chain would be random and a possibility of some homodimer fraction could not be excluded.

4.2. Molecular Mass and Subunit Composition of Native BSTBP Molecule The hypothesis that BSTBP is a fraction of the major storage 13 S globulin (legumin) of buckwheat seeds helps to define the association state of native BSTBP molecule. To be compatible with the “typical” legumin quaternary structure (Casey, 1999; Plietz et al., 1984) and with the available characteristics of the buckwheat legumin (Belozersky, 1975; Radovic et al., 1996), the BSTBP preparation should be a rather heterogeneous population of essentially hexameric molecules composed of various combinations of the 57-kDa and 38-kDa subunits. Because of the presence of the atypically small, possibly proteolytically truncated, 38-kDa subunits, BSTBP is smaller than the “average” buckwheat-seed 13 S globulin of 280 kDa (Belozersky, 1975). The average BSTBP molecular mass (approximately 235 kDa), estimated by the gel filtration and dynamic light scattering methods, seems to be roughly consistent with at least twofold excess of the 38-kDa subunit over the large one, apparent from the SDS-PAGE pattern of most of our BSTBP preparations. Nevertheless, a possibility that BSTBP also contains some fraction of aggregates lower than hexamers cannot be excluded. A recent report from the other laboratory (Watanabe et al., 1998b) suggested that BSTBP had a significantly higher molecular mass of 320 kDa and was composed of 56-kDa and 50-kDa subunits. That picture was consistent with a hexameric quaternary structure of BSTBP. Unfortunately, only a single 25-kDa band was detected in SDSPAGE patterns of reduced BSTBP samples. On the other hand, a much lower native molecular mass (140 kDa) of BSTBP was also reported (Mitsunaga et al., 1986b; Rapala-Kozik and Kozik, 1992). According to the interpretation presented in this work, a variable amount of 38-kDa subunit, a putative product of a proteolytic truncation of a larger approximately 60 kDa species, in different BSTBP preparations, may account for those inconsistencies.

174


4.3. Contribution of BSTBP Polypeptide Components to Thiamin Binding As calculated from the thiamin-binding capacity and from the estimated molecular mass of BSTBP preparations, the stoichiometry of thiamin binding to TBP was between two and three thiamin molecules bound per protein molecule. Hence, the individual (12) polypeptide chains of BSTBP molecule must variably contribute to the structure of thiamin-binding center. This contribution was studied by labeling BSTBP with a photoaffinity reagent, ANBT. The labeling of BSTBP with ANBT was preceded by several tests intended to prove that ANBT would specifically label the thiamin-binding site. As shown in Fig. 3, ANBT bound to BSTBP (in the dark) competitively with thiamin and with the affinity similar to that of thiamin. Moreover, ANBT-photolabeled BSTBP lost the thiamin-binding affinity unless the binding site was protected by the presence of thiamin. The distribution of fluorogenic label between SDS-PAGE detectable BSTBP components, presented in Fig. 4, suggested that it was the 57-kDa subunit and its 36-kDa acidic chain that accommodated the highest proportion of the thiamin label. Thus, the 36-kDa chain might be the major contributor to the BSTBP thiaminbinding activity, although the involvement of the other chains could not be definitely excluded. As expected for the true “affinity labeling,” it was significantly prevented by the presence of thiamin excess (2 ⫻ 10⫺3 M). Consistently with the highest ANBT label content, the highest thiamin protection against ANBT labeling was assigned to the 57-kDa subunit and the 36-kDa chain.

4.4. Polypeptide Fragments Putatively Involved in Thiamin Binding The thiamin-labeled peptides isolated from tryptic digests of ANBT-labeled BSTBP preparations (Fig. 5) are likely to contribute to the formation of thiaminbinding center. In principle, their amino acid sequences should distinguish BSTBP from the other molecular versions of buckwheat legumin. Indeed, neither of two putative thiamin-binding peptides had a sequence similar to any fragment of the buckwheat legumin precursor (Fujino et al., 2001). The peptide 1 (Fig. 5) had an EEPQQ motif that could be found in acidic parts of several legumins, including soybean glycin or fava bean legumin. A significance of these similarities remains uncertain. In contrast, peptide 2 (Fig. 5) did not contain any motif that would be popular in seed storage globulins. It should, finally, be noted that the presence of acidic amino acid residues in both putative thiamin-binding

peptides is compatible with the previous reports that carboxyl groups play an essential role in the binding mechanism (Rapala-Kozik and Kozik, 1996). 4.5. Conclusions SDS-PAGE analysis of BSTBP shows that this protein is composed of two major subunits of apparent molecular mass of 57 kDa and 38 kDa. Each subunit is a dimer of disulfide-linked polypeptide chains. The 57-kDa subunit is composed of 36-kDa and 24-kDa chains; the 38-kDa subunit contains two chains of molecular mass of approximately 22 kDa. The 22-kDa and 24-kDa components have N-terminal amino acid sequences very similar or identical to those of some versions of basic chains of the buckwheat seed major storage 13 globulin (legumin). As a fraction of major buckwheat legumin, BSTBP seems to be a heterogeneous population of essentially hexameric molecules, each containing some combination of 57-kDa and 38-kDa subunits. The average molecular mass of the BSTBP is approximately 235 kDa. The individual polypeptide chains variably contribute to the thiamin-binding activity of BSTBP. Results of the photoaffinity labeling of BSTBP with an azidobenzoyl derivative of thiamin (ANBT) indicate that the 36-kDa acidic chain may play the major role in the formation of thiamin-binding center. The sequences of thiamin-labeled peptides isolated from tryptic digests of ANBT-labeled BSTBP are not clearly similar to any fragment of the buckwheat legumin or of any seed globulin from other species. Hence, these putative thiaminbinding sequences may determine the unique status of BSTBP among the other forms of the major storage legumin of buckwheat seeds. ACKNOWLEDGMENTS This work was supported in part by grant 6 P04C 0039 19 from Polish State Committee for Scientific Research (K.B.N.). REFERENCES Adachi, T., Watanabe, K., and Mitsunaga, T. (2000). Cereal Chem. 77: 578–581. Adamek-Swierczynska, S., and Kozik, A. (2002). Plant Physiol. Biochem. 40: 735–741. Adamek-Swierczynska, S., Rapala-Kozik, M., and Kozik, A. (2000). J. Plant. Physiol 156: 635–639. Belozersky, M. A. (1975). Biosynthesis of Storage Proteins, Nauka, Moscow, pp. 152–156. Casey, R. (1999). In: Shewry, P. R., and Casey, R. (eds), Seed Proteins, Kluwer Academic Publishers, Dordrecht/Boston, pp. 159–169.

Polypeptide Components of Oligomeric Legumin-Like Thiamin-Binding Protein Davis, B. J. (1964). Ann. N. Y. Acad. Sci. 121: 404–427. Fujino, K., Funatsuki, H., Inada, M., Shimono, Y., and Kikuta, Y. (2001). J. Agric. Food Chem. 49: 1825–1829. Kozik, A. (1995). In: Matano, T., and Ujihara, A. (eds.), Current Advances in Buckwheat Research, Shinshu University Press, Asahi Matsumoto City, pp. 823–831. Kozik, A., and Rapala-Kozik, M. (1993). J. Chromatogr. 648: 349–356. Kozik, A., and Rapala-Kozik, M. (1995). J. Plant. Physiol 146: 760–762. Kozik, A., and Rapala-Kozik, M. (1996). Plant Physiol. Biochem. 34: 779–786. Lowry, O. M., Rosebrough, N. J., Farr, A. L., and Randall, R. J. (1951). J. Biol. Chem. 193: 265–275. Maksimovic, V. R., Varkonji-Gaˇsic, E. I., Radovic, S. R., and Savic, A. P. (1996). J. Plant Physiol. 147: 759–761. Matsudaira, P. (1987). J. Biol. Chem. 262: 10035–10053. Mickowska, B., Dulinski, R., and Kozik, A. (2000). J. Biochem. Biophys. Methods 44: 95–107. Mitsunaga, T., Shimizu, M., and Iwashima, A. (1986a). J. Plant. Physiol. 124: 177–180. Mitsunaga, T., Shimizu, M., and Iwashima, A. (1987). J Plant Physiol. 130: 279–284. Mitsunaga, T., Shimizu, M., Matsuda, M., and Iwashima, A. (1986b). Cereal Chem. 63: 332–335. Müntz, K. (1996). J. Exp. Bot. 298: 605–622. Nishimura, H., Uehara, Y., Sempuku, K., and Iwashima, A. (1984). J. Nutr. Sci. Vitaminol. 30: 1–10. Plietz, P., Damashun, G., Zirwer, D., Gast, K., and Schleiser, B. (1984). Biochim. Biophys. Acta 784: 140–146.

175

Radovic, S. R., Maksimovic, V. R., and Varconji-Gaˇsic, E. I. (1996). J. Agric. Food Chem. 44: 972–974. Rapala-Kozik, M., and Kozik, A. (1992). Biochim. Biophys. Acta 1159: 209–214. Rapala-Kozik, M., and Kozik, A. (1996). Biochemie 78: 77–84. Rapala-Kozik, M., and Kozik, A. (1999). Plant Physiol. Biochem. 37: 539–544. Rapala-Kozik, M., Chernikevich, I. P., and Kozik, A. (1999). J. Protein Chem. 18: 721–728. Rout, M. K., Chrungoo, N. K., and Roa, K. S. (1997). Phytochemistry 45: 865–867. Schägger, H., and von Jagow, G. (1987). Anal. Biochem. 166: 368–379. Sempuku, K. (1988). Biochim. Biophys. Acta 944: 177–184. Shimizu, M., Inaba, K., Yoshida, T., Toda, T., Iwashima, A., and Mitsunaga, T. (1995). Physiol. Plant. 93: 93–98. Shimizu, M., Mitsunaga, T., Inaba, K., Yoshida, T., and Iwashima, A. (1990). J. Plant Physiol. 137: 123–124. Watanabe, K., Chikushi, K., Adachi, T., Shimizu, M., Yoshida, T., and Mitsunaga, T. (1998a). J. Nutr. Sci. Vitaminol. 44: 665– 672. Watanabe, K., Chikushi, K., Adachi, T., Shimizu, M., Yoshida, T., and Mitsunaga, T. (1999). Physiol. Plant. 107: 8–13. Watanabe, K., Konishi, A., and Mistunaga, T. (2002). Plant Physiol. Biochem. 40: 417–421. Watanabe, K., Shimizu, M., Adachi, T., Yoshida, T., and Mitsunaga, T. (1998b). J. Nutr. Sci. Vitaminol 44: 323–328. Watanabe, K., Takahashi, H., and Mitsunaga, T. (2001). Physiol. Plant. 112: 546–551.