 |  |
| ||||
Copyright © 2000, American Society of Plant Physiologists A Novel Alliinase from Onion Roots. Biochemical Characterization and cDNA Cloning1 *Corresponding author; e-mail jlancaster/at/xtra.co.nz; fax 64–3–3296888. Received July 17, 1999; Accepted November 30, 1999.  This article has been cited by other articles in PMC. | |||||||||||||
Abstract We have purified a novel alliinase (EC 4.4.1.4) from roots of onion (Allium cepa L.). Two isoforms with alliinase activity (I and II) were separated by concanavalin A-Sepharose and had molecular masses of 52.7 (I) and 50.5 (II) kD on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and 51 (I) and 57.5 (II) kD by gel filtration fast-protein liquid chromatography. Isoform I had an isoelectric point of 9.3, while isoform II had isoelectric points of 7.6, 7.9, 8.1, and 8.3. The isoforms differed in their glycosylation. Both contained xylose/fucose containing complex-type N-linked glycans, and isoform II also contained terminal mannose structures. Both isoforms had activity with S-alk(en)yl-l-cysteine sulfoxides. Unlike other allium alliinases, A. cepa root isoforms had cystine lyase activity. We cloned a gene from A. cepa root cDNA and show that it codes for A. cepa root alliinase protein. Homology to other reported allium alliinase genes is 50%. The gene coded for a protein of mass 51.2 kD, with two regions of deduced amino acid sequence identical to a 25- and a 40-amino acid region, as determined experimentally. The A. cepa root alliinase cDNA was expressed mainly in A. cepa roots. The structure and function of the alliinase gene family is discussed. | |||||||||||||
The characteristic flavor of onion (Allium cepa L.), garlic (Allium sativum L.), and related alliums occurs when the enzyme alliinase (EC 4.4.1.4) hydrolyzes the S-alk(en)yl-l-Cys sulfoxides (ACSOs) to form pyruvate, ammonia, and sulfur-containing volatiles. The ACSO substrates are located in the cytoplasm, while alliinase is compartmentalized in the vacuole until cell rupture (Lancaster and Collin, 1981). The production of volatile sulfur products by this reaction is central to the characteristic and commercially important flavor of these plants. Alliinase (synonyms: alliin alkyl-sulfonate-lyase, alliin lyase, alkyl Cys sulfoxide lyase, Cys sulfoxide lyase, and C-S lyase) is present in A. cepa bulbs at high levels, representing up to 6% by weight of the soluble protein in the bulb (Nock and Mazelis, 1987). Alliinase enzymes from garlic cloves (Mazelis and Crews, 1968), Allium tuberosum (chinese chives) (Manabe et al., 1998), Allium ursinum (ramson or wild garlic) (Landshuter et al., 1994), and Allium porrum (leek) (Lohmüller et al., 1994) have also been purified to homogeneity and characterized. Related enzymes have also been identified in Acacia farnesiana seedlings (Mazelis and Creveling, 1975), Brassica oleracea buds (Hamamoto and Mazelis, 1968), bacteria (Nomura et al., 1963; Kamitani et al., 1990, 1991), and fungi (Iwami and Yasumoto, 1980). Alliinase is a glycoprotein with an estimated carbohydrate content of 4.6% in onion and 5.5% in garlic (Nock and Mazelis, 1987). Pyridoxal 5′-P (PyrP) has been demonstrated as an essential co-factor with empirical measurements predicting one very tightly bound PyrP per subunit (Tobkin and Mazelis, 1979) at Lys-285 in the amino acid sequence deduced from A. cepa cDNA (Kitamura et al., 1997). Genes encoding alliinase have been isolated from A. cepa (Van Damme et al., 1992; Clark, 1993; Gilpin et al., 1995; King et al., 1998), A. tuberosum (Chinese chives) (Manabe et al., 1998), Allium ascalonicum (shallot) (Van Damme et al., 1992), and A. sativum (Van Damme et al., 1992; Rabinkov et al., 1994). Northern analysis in A. sativum tissues showed that alliinase cloned from clove cDNA was expressed in cloves and leaves but not in roots (Rabinkov et al., 1994), although they were shown to contain high alliinase activity. This indicated the existence in A. sativum roots of a non-homologous alliinase gene. In this paper we report the purification and characterization of A. cepa root alliinase protein and the cloning of a gene from A. cepa root cDNA encoding this protein. The gene is substantially different in sequence, and the protein has novel structure and function compared with other allium alliinases. The role of this novel alliinase in roots is discussed. | |||||||||||||
MATERIALS AND METHODS Protein Purification and Sequencing Alliinase was purified from onion (Allium cepa cv Pukekohe Longkeeper) roots grown in hydroponic solution culture (Randle et al., 1995). Roots (typically 300 g) were harvested, washed in tap water, and blotted dry before being frozen in liquid N2 and homogenized in an equivalent volume (w/v) of buffer A (0.02 m sodium phosphate, 30% [v/v] ethylene glycol, and 2.5 μg mL−1 PyrP, pH 6.5) containing 0.05% (w/v) 2-mercaptoethanol, 1 mm phenylmethylsulfonyl fluoride (PMSF), 5 mm 6-amino-n-hexanoic acid, and 5 mm EDTA. All purification and column steps were performed at 4°C.The homogenate was squeezed through two layers of cheesecloth and the cell debris removed by centrifugation. The supernatant was applied to a CM-Sepharose CL6B (11.0- × 3.0-cm) column equilibrated with buffer A. After washing off unbound material, active fractions were eluted with buffer A containing 0.5 m NaCl and pooled. Pooled fractions were adjusted to 1 mm MnCl2 and 1 mm CaCl2 and applied to a concanavalin A (ConA)-Sepharose 4B column (9 × 1.5 cm) previously equilibrated in buffer B (0.05 m Tris, 0.5 m NaCl, 30% [v/v] ethylene glycol, 2.5 μg mL−1 PyrP, 1 mm MnCl2, and 1 mm CaCl2, pH 7.5). After washing off unbound material with buffer B, the column was then eluted with buffer B containing 5 mm methyl α-d-glucopyranoside and fractions pooled and concentrated. Further elution was with buffer B containing 0.2 m methyl α-d-mannopyranoside. Both concentrated fractions (about 5 mL) were dialyzed against buffer A. Each was applied to its own CM-Sepharose CL6B (4.5- × 1.5-cm) column equilibrated in buffer A, and, after washing off unbound material, active enzyme was eluted with buffer A containing 0.5 m NaCl. Sample purity was established by SDS-PAGE (Laemmli, 1970) followed by visualization with silver staining (Giulian et al., 1983). Subunit sizes were determined using a 10-kD molecular mass ladder (Gibco-BRL, Cleveland). Synthesis of Potential Alliinase Substrates The substrates S-methyl-l-Cys sulfoxide (MCSO), S-ethyl-l-Cys sulfoxide (ECSO), S-propyl-l-Cys sulfoxide (PCSO), and S-butyl-l-Cys sulfoxide (BCSO) were synthesized by the method of Lancaster and Kelly (1983). S-Propenyl- l-Cys sulfoxide (PrenCSO) was prepared by hydrolyzing γ-glutamyl-propenyl-L-Cys sulfoxide with γ-glutamyl transpeptidase (Kuttan et al., 1974). S-Allyl-l-Cys sulfoxide (AllylCSO) was obtained from Extrasynthese (Genay, France), and l-Cys, L-cystine, methyl Cys, and cystathionine were purchased from Sigma (St. Louis).Measurement of Alliinase Activity Alliinase activity was measured by a coupled NADH/lactate dehydrogenase (LDH) assay in 0.2 m N-[2-hydroxy-1,1-Bis(hydroxymethyl)ethyl]glycine (Tricine)-KOH, pH 8.0, 0.1 mm NADH, 12.5 units of LDH, 0.1% (v/v) alliinase preparation, depending on the stage of preparation and activity of sample, and, 0.01 m ECSO at 20°C. Activity was calculated from the conversion of NADH to NAD+ as measured by a decrease in A340 (Schwimmer and Mazelis, 1963) and is reported as nanokatals (1 nkat = 1 nmol s−1). The protein concentration of eluants was determined using a refinement of the Bradford dye binding assay (Spector, 1978).Molecular Mass Determination by Gel Filtration Fast-Protein Liquid Chromatography (FPLC) Protein samples (1–2 mg) eluted with 5 mm methyl α-d-glucopyranoside (alliinase isoform I) and 0.20 m methyl α-d-mannopyranoside (alliinase isoform II) from the Con A-Sepharose 4B column were dissolved in 0.05 m Tris-HCl buffer (pH 7.5) containing 0.15 m NaCl and 15% (v/v) ethylene glycol, (1 mL), and centrifuged at 15,000g for 10 min. The samples were analyzed by FPLC on a 1- × 30-cm Sepharose 12 column (Pharmacia LKB Biotechnology, Uppsala) at 20°C. Samples (100 μL) were injected onto the column and eluted with the above buffer at 0.4 mL min−1 and 224 psi. Eluting proteins were detected at 280 nm, and fractions (0.2 mL) were collected at 4°C. Alliinase activity was measured in each fraction to determine the elution times, and the molecular mass of alliinase was estimated from a calibration graph of markers (from Sigma): apoferritin (443 kD), β-amylase (200 kD), bovine serum albumin (66 kD), ovalbumin (43 kD), and carbonic anhydrase (29 kD).Isoelectric Focusing (IEF) IEF of alliinase isoforms I and II was performed on 0.2-mm-thick 5% (w/v) T, 0.75% (w/v) C (Bis) acrylamide gels, containing 5% (w/v) glycerol, 2% (w/v) Bio-Lyte 3/10 Ampholyte cast onto Gel Bond support film and run on a Multiphor flat bed gel apparatus (Pharmacia, Piscataway, NJ). Typically, a 124- × 110-mm gel was run at 8 W of constant power for 3,300 V h−1 using a cathode buffer of 0.02 m Lys, 0.02 m Arg, and 2 m ethylene diamine, and an anode buffer of 0.02 m Asp and 0.02 m Glu. PIs were estimated from IEF standard markers (Bio-Rad, Hercules, CA): cytochrome c, pI 9.60; lentil lectin, pI 7.8, 8.0, and 8.2; human hemoglobin C, pI, 7.50; and human hemoglobin A, pI 7.1. Gels were visualized with silver staining (Giulian et al., 1983) and also for alliinase activity on the substrate ECSO by staining for thiol reaction products (see below) (Ukai and Sekiya, 1997).Determination of Substrate Specificity Alliinase isoforms I and II were separated by IEF as above, and subsequently reacted with 1, 5, and 10 mm of each of the substrates PrenCSO, MCSO, and ECSO in situ. Alliinase activity in situ in IEF gels was demonstrated by staining for thiol reaction products (Ukai and Sekiya, 1997). The gel was soaked in 0.05 m Tris-HCl, pH 8.0, containing 2.5 μg mL−1 PyrP with shaking for 10 min, and put into a 15-mL Falcon tube. Into each tube was added the following solution: 0.5 mL of 1 m Tris-HCl, pH 8.0, 2.5 mg of 3-(4′,5′-dimethyl-2-thiazol)-2,5-diphenyl-2H-tetrazolium bromide, 2.5 mg of phenazine methosulfate, and 10 μL of 10 mm PyrP made up to 9 mL with water. The reaction was started and timed by the introduction of each of the above substrates in a total volume of 1 mL. After incubation, the solution was poured off and the strips were washed in water until no color was left. The strips were then removed and dried.Kinetic analysis of alliinase isoforms I and II was carried out in vitro in solution with the substrates MCSO, ECSO, PrenCSO, PCSO, AllylCSO, cystine, methyl Cys, Cys, and cystathione. Approximately 0.05 to 0.33 μg of protein, purified as in Table I, was used for each incubation. Substrate concentrations ranged between 1 and 25 mm. Alliinase and substrates were incubated at 30°C in 0.2 m Tricine-KOH (pH 8.0) buffer containing 0.01 mm PyrP in a total volume of 200 μL. Incubation was for 1 min for the ACSO substrates and for 20 min (with 0.2 m Tris-HCl, pH 8.8) for cystine. Methyl Cys, Cys, and cystathionine were incubated for 60 min. The reaction was stopped by adding 200 μL of 2,4 dinitrophenylhydrazine (0.0125% [w/v] in 2 n HCl). The reaction was incubated at 37°C for 10 min, 1.0 mL of 0.6 n NaOH was added, and A420 determined after 5 min. A pyruvic acid (0.1–0.8 μmol mL−1) standard curve was generated and used for calibration of the enzymically developed pyruvate (Randle and Bussard, 1993). Activity was calculated as nanokatals per milligram of protein (1 nkat = 1 nmol pyruvate s−1). Calculation of Km and Vmax was carried out using a computerized least-squares-fit method.
Western Blotting of IEF Gels Alliinase isoform I (0.128 μg) and alliinase isoform II (1.313 μg) were loaded into lanes and focused by IEF as above. The isoforms were blotted onto a polyvinylidene difluoride (PVDF) membrane. The gel was washed in 0.01 m 3-(cyclohexylamino)propanesulfonic acid (CAPS), 10% (w/v) MeOH, pH 11.0, transfer buffer for 15 min with a change every 5 min. The gel was removed from the backing film by placing a dry piece of blotting paper on the surface and carefully peeling it off. Blotting was done at 0.8 mA cm−2 for 30 min in a semi-dry apparatus. Bands were visualized by Coomassie staining and excised. N-Terminal sequences were determined using an on-line gas phase sequencer (model 470A, Applied Biosystems, Foster City, CA) using the standard protocol of the manufacturer.Determination of N-Linked Glycosylation of Alliinase Isoforms To probe each alliinase with glycan-specific antibodies, isoforms I and II were separated by IEF and blotted onto PVDF membrane as described above. Antibody staining with monoclonal antibody 2.23 (McManus et al., 1988) was performed at a dilution of 1:1,000, and immune recognition was visualized using alkaline phosphatase secondary antibodies. Blots with digoxigenin-labeled Galanthus nivalis agglutinin (GNA) lectin (Boehringer Mannheim, Mannheim, Germany) and alkaline-phosphatase conjugated anti-digoxigenin (Boehringer Mannheim) were performed according to the manufacturer's instructions. For digestion with endoglycosidase H (endo H) (Boehringer Mannheim), 10 μg of alliinase isoform II or the soybean agglutinin lectin (L1395, Sigma) was denatured in 0.1 m 2-mercaptoethanol and 0.05% (w/v) SDS by heating at 100°C for 5 min. After cooling, the denatured protein mixture was buffered with 50 mm sodium citrate, pH 5.6, and PMSF was added to a final concentration of 0.4% (w/v). Endo H (0.4 unit) was then added and the reaction incubated at 25°C for 24 h. The reduction in molecular mass at the conclusion of the endo H treatment was determined by SDS-PAGE (Laemmli, 1970).Cyanogen Bromide Cleavage of Alliinase Alliinase isoform II, eluted with 0.2 m methyl α-d-mannopyranoside from Con A-Sepharose, was further purified on CM-Sepharose CL6B. Cyanogen bromide peptides for amino acid sequencing were produced by dissolving 40 to 50 μg of alliinase isoform II in 25 μL of 70% (w/v) formic acid containing 1.5 mg of cyanogen bromide and incubating overnight at 4°C in the dark. The mixture was diluted 10-fold with water and vacuum-dried under centrifugation. The peptides were separated on 16.5% (w/v) Tris-Tricine polyacrylamide gels (Bio-Rad) and blotted onto PVDF membrane. The bands were excised and N-terminal sequences were determined using an on-line gas phase sequencer (model 470A, Applied Biosystems) using standard protocols.The peptides were also separated by SDS-PAGE and transferred to PVDF membrane (Immobilon-P, Millipore, Bedford, MA) by blotting at 0.8 mA cm−2 for 30 min in a semi-dry apparatus. The membrane was probed for oligosaccharide side chains with Con A-horseradish peroxide conjugate and visualized with 4-chloro-l-napthol and H2O2 (Clegg, 1982; DeBlas and Cherwinski, 1983). Analysis of ACSOs ACSOs from A. cepa roots were extracted in methanol:chloroform:water (12:5:3; v/v) and ethanol (80%; v/v) according to the method of Lancaster and Kelly (1983). BCSO, which is not normally found in allium species, was added as internal standard at 1 mg g−1 fresh weight to determine the percentage recovery through the extraction, derivatization, and analysis procedures. The onion aqueous extract (1 g fresh weight equivalent) was lyophilized at −36°C in vacuo and then redissolved in 1 mL of deionized water. Sample fractionation was performed by ion-exchange chromatography with Dowex 1×8 (200–400 mesh, 10- × 40-mm column) resin in the acetate form. A 0.5-mL sample of extract was loaded onto the column and eluted using acetic acid (HOAc) at 0.1, 1.0, and 2.0 n. The 0.1 n HOAc fraction contained the ACSOs. The 1.0 and 2.0 n fractions contained the γ-glutamyl peptide biosynthetic intermediates. Samples were lyophilized and derivatized with phenyl isothiocyanate and analyzed by reverse-phase HPLC (Randle et al., 1995).Cloning of a Partial Alliinase cDNA from A. cepa Root for Library Screening Oligonucleotide primers were designed from a 438-bp A. cepa leaf cDNA AOB249 (accession no. AA451570; King et al., 1998). Primers were synthesized by Gibco-BRL: AOB249-1 (5′-GGCTGGTAGCGGCAGTCTACT-3′) situated at the 5′ end, and AOB249-R (5′-TGTCGTAGTTGTACCCAGACG-3′) situated at the 3′ end. Total RNA was extracted from root tips of hydroponically grown A. cepa using TRIZOL (Gibco-BRL). First-strand cDNA was generated with a NotI-d(T)18 primer using Ready-To-Go T-Primed First Strand Beads (Amersham-Pharmacia Biotech, Uppsala). The first-strand cDNA was used as template to amplify a 570-bp product by PCR with the 5′ primer AOB249-1. Amplification conditions were 30 s at 95°C, 30 s at 60°C, and 30 s at 72°C for 40 cycles. PCR products were purified with the Prep-A-Gene kit (Bio-Rad) and then cloned into the pGEM-T vector using the pGEM-T vector system (Promega, Madison, WI). Plasmid DNA was isolated using a plasmid isolation kit (High Pure, Roche, Basel) from colonies that had the correct insert size, as determined by PCR with M13 forward and reverse primers. Sequencing reactions were performed using a terminator cycle sequencing kit (ABI PRISM Big Dye kit, Perkin-Elmer Applied Biosystems, Foster City, CA) and analyzed on an automated sequencer (model 377, Perkin-Elmer-Applied Biosystems). A 300-bp fragment (rootAOB) was PCR amplified from the vector for probe with the primers AOB249-1 and AOB249-R (conditions as above) and purified with a PCR product purification kit (High Pure, Roche).Construction and Screening of a cDNA Library Total RNA was extracted from root tips of hydroponically grown A. cepa as described above. The mRNA fraction was isolated from total RNA by double selection using a mRNA isolation kit (Messagemaker, Gibco-BRL). A cDNA library was constructed from 5 μg of mRNA using the ZAP cDNA synthesis kit (Stratagene, La Jolla, CA), and packaged in vitro with Gigapack II Gold packaging extract (Stratagene). The resulting cDNA library (1.8 × 106 primary recombinant phages) was excised in vivo.A total of 5 × 104 excised phagemids were plated with host strain SOLR at a density of 5 × 103 colonies per 10 cm2 plate and transferred to nitrocellulose membranes (Hybond-C Extra, Amersham-Pharmacia Biotech). Membranes were placed colony side up onto 3MM filter paper (Whatman, Clifton, NJ) prewetted with a 5% (w/v) SDS, 2× SSC solution and heated for 3 min on medium-high in a 850 W microwave, followed by vacuum drying at 80°C for 30 min. Membranes were hybridized with 32P-labeled root AOB fragment randomly primed with the Megaprime labeling kit (Amersham-Pharmacia Biotech). Hybridization was carried out in 6× SSC, 10% (w/v) dextran sulfate, 0.1% (w/v) Na2H2P2O7, 1% (w/v) SDS, and 0.3 mg mL−1 denatured herring sperm DNA at 65°C for 16 h. Post-hybridization washes were carried out at 65°C for 30 min each: 2× SSC, 0.1% (w/v) SDS, 1× SSC, 0.1% (w/v) SDS, and 0.1× SSC, 0.1% (w/v) SDS. Autoradiography signals were obtained after 16 h of exposure at −80°C using Kodak X-Omat AR film (Eastman Kodak, Rochester, NY) with special intensifying screens (Agfa Curix CX, Agfa Gevaert, Mortsel, Belgium). Colonies that gave a positive signal from hybridization were checked for insert size by PCR using M13 forward and reverse primers. Sequence Analysis of the Full-Length Root Alliinase cDNA Plasmid DNA was isolated from colonies that had an insert size of approximately 1,600 bp as determined by PCR. Restriction digests were also carried out on the plasmid DNA with EcoRI and XhoI to check insert size. Sequencing reactions were performed as previously described. Sequencing primers used were: SK (Stratagene), M13 reverse, and AOB249-1. Further primers were designed to sequence the complete gene: Alliroot1 (5′-TACGAATGGAAGGGAAATGC-3′) at position 686, and Alliroot2 (5′-GCTGATGCCACTACTGGTGA-3′) at position 329.Northern Hybridization Total RNA was extracted from root, leaf, and bulb tissues of A. cepa as described above. Approximately 10 μg of RNA was separated by electrophoresis through a 1% (w/v) agarose gel containing formaldehyde and transferred to Zeta-Probe GT nylon membrane (Bio-Rad) (Sambrook et al., 1989). Three duplicate membranes were hybridized with separate probes as previously described: the full-length alliinase A. cepa root cDNA, a full-length alliinase A. cepa leaf cDNA (Clark, 1993), and a pea ubiquitin cDNA (Watts and Moore, 1989). | |||||||||||||
RESULTS Purification and Characterization of Alliinase Protein Molecular Mass of Purified Alliinase Alliinase was purified from onion roots using ion-exchange and affinity chromatography. The results of a typical purification of alliinase are summarized in Table I. The separation of alliinase into two peaks of activity on the Con A-Sepharose column indicated the presence of isoforms (Fig. 1a). Alliinase activity eluted at 5 mm methyl α-d-glucopyranoside (isoform I) and 0.2 m methyl α-d-mannopyranoside (isoform II). Further purification on CM-Sepharose resulted in isoform I being 109-fold purified, with a specific activity of 3,380 nkat mg−1, and isoform II being 64-fold purified, with a specific activity of 1,995 nkat mg−1. The specific activity of isoform II appeared to decrease after this final purification step. However, visualization after SDS-PAGE showed that contaminating proteins had been removed by the final CM-Sepharose purification. It is possible that factors important to activity may also be removed by this step. On gel filtration FPLC, alliinase isoforms I and II eluted at a molecular mass of 51 and 57.5 kD, respectively (Fig. 1b). Isoforms I and II were separated by SDS-PAGE (Fig. 1c). Each isoform was homogeneous on SDS-PAGE. Isoforms I and II gave submit masses of 52.7 and 50.5 kD, respectively.
IEF The alliinase isoforms I and II were separated by IEF and silver stained (Fig. 1d). Isoform I gave one band at pI 9.3. Isoform II gave four bands at pIs 7.6, 7.9, 8.1, and 8.3. The isoform II bands with alliinase activity were N-terminal sequenced, and all gave the sequence LTWTTK (Fig. 1d). The 9.3-pI band from isoform I could not be sequenced, possibly due to blockage at the N terminus.Alliinase Glycosylation Western blotting using glycan-specific antibodies was used to determine the nature of N-linked glycosylation of each alliinase isoform. Isoform I was recognized by monoclonal antibody 2.23 but not by the GNA lectin, while isoform II was recognized by both probes (Fig. 2a). Isoform II was subject to endo H digestion, but no reduction in molecular mass of the protein was observed (Fig. 2b). In contrast, a decrease in molecular mass of the endo H-digested soybean agglutinin lectin was observed, a protein known to be glycosylated with an oligo-Man N-linked glycan structure (Dorland et al., 1981).
Internal Amino Acid Sequence Alliinase isoform II was digested with cyanogen bromide and the amino acid sequence of the resultant peptides was determined. There were two main peptides of 25 and 12 kD. The 12-kD peptide gave the amino acid sequence LTWTTKAATEAEKVAAIYCSDHGRA. This band gave a negative response when western blotted and probed with Con A, indicating an absence of oligosaccharide side chains. The 25-kD peptide gave the amino acid sequence SDELDRHIRLLHNAVGNAKVDD KFLVFGNG VT QLL NGVII. This peptide gave a positive response when western blotted and probed with Con A, indicating the presence of oligosaccharide side chains.Alliinase Activity and Substrate Specificity Alliinase isoform II was separated by IEF gel into four bands and reacted in situ with each of MCSO, ECSO, and PrenCSO. All four bands, which were of equal loading, gave comparable activity staining with each of the substrates. Alliinase isoform I, separated by IEF gel, was inactive in situ with each of the same substrates. Purified alliinase isoforms were then reacted in vitro in solution, with each of the substrates MCSO, ECSO, PCSO, AllylCSO, PrenCSO, cystine, methyl Cys, Cys, and cystathionine (Table II). Km, Vmax, and Km/Vmax were calculated from the enzyme reaction data. Both alliinase isoforms showed a similar order of activity with the substrates. They were most active with PrenCSO and showed decreasing activity through the sequence Allyl CSO, PCSO, ECSO, and MCSO. Cystine was the least active. Both isoforms were inactive with methyl Cys, Cys, and cystathionine.
Root Flavor Precursors Roots contained the flavor precursors MCS and PrenCSO at concentrations of 0.04 and 0.11 mg g−1 fresh weight, respectively. PCSO was below detection. There was an absence of γ-glutamyl sulfur peptides.Cloning, Sequencing, and Expression of Alliinase cDNA Alliinase cDNA Cloning A partial A. cepa leaf cDNA clone (AOB249) encoding a putative alliinase gene with approximately 60% homology to A. cepa bulb alliinases was previously identified during sequencing of seedling leaf cDNA clones used to develop the onion linkage map (King et al., 1998). This sequence was used to design a primer for reverse transcripase PCR that would not anneal with A. cepa bulb alliinases. When used in reverse transcripase PCR reactions with root cDNA, this gave a product that was cloned and found to be homologous with AOB249. A Uni-ZAP XR cDNA library was constructed containing 1.8 × 107 plaque forming units. Approximately 1.8 × 106 plaque forming units were mass excised, and 5 × 104 phagemids were screened using the 300-bp root-derived AOB probe. Thirty-one positive signals were detected from colony hybridization. Of the 31 clones, three had an expected insert size of approximately 1,600 bp, as determined by PCR with M 13 forward and reverse primers and restriction digest with EcoRI and XhoI.Alliinase Nucleotide Sequence Two of the cDNA clones were sequenced and found to be identical. The complete nucleotide sequence of the root alliinase cDNA is shown in Figure 3 (accession no. gbAF126049). The clone is 1,700 bp with an open reading frame of 1,493 bp. There is a possible translation termination codon at 1,517 bp and a 3′ untranslated region of 180 bp. Deduced amino acid sequence gave a mature protein of 453 amino acids and 51.2 kD with a predicted pI of 7.23. Two regions of deduced amino acid sequence are identical to the 25 and 40 amino acids determined experimentally from the purified protein, indicating that the cDNA cloned codes for A. cepa root alliinase isoform II (Fig. 3). Upstream of the N terminus (LTWTTK) is a leader sequence of 132 bp and 44 amino acids. The A. cepa root-alliinase-deduced amino acid sequence contains four potential glycosylation sites of the Asn-X-Ser/Thr (NXS/T) conformation at amino acid positions 55, 143, 182, and 188. A Lys residue at position 248 from the N terminus indicates a possible PyrP-binding residue within the PyrP consensus sequence [GD]-X-D-[LIV]-X(3)-[STAGCM](2)-[TS]-K-[FYW]-[LI]-X-G-[HQ]-[SGN] (Manabe et al., 1998).
Alignment with Other Alliinases The deduced alliinase isoform II amino acid sequence was aligned with other alliinase amino acid sequences from alliums (Fig. 4). Homology with A. cepa alliinases are 52.8% for leaf shoots (Clark, 1993) and 53.4% for bulb (Van Damme et al., 1992). Homology with garlic (Allium sativum) clove alliinase is 51.4% (Rabinkov et al., 1994) and 52.5% (Van Damme et al., 1992). Homology with shallot (Allium ascalonicum) is 51.8% (Van Damme et al., 1992), and with chinese chives (Allium tuberosum) homology is 49.4% (Manabe et al., 1998). In contrast, the sequences (amino acids 238–253) around the predicted PyrP Lys-248 binding residue of the A. cepa root alliinase show strong agreement with the predicted PyrP binding motif of other alliinases (Fig. 4). A. cepa root alliinase cDNA has four predicted glycosylation NXS/T sites (amino acids 55, 143, 182, and 188), of which two are in common with glycosylation sites of the other allium alliinase cDNA sequences (143 and 188) (Fig. 4). The site at amino acid 327 in A. cepa, A. sativum, and A. ascalonicum, is absent from A. cepa root alliinase cDNA. The A. cepa root alliinase leader sequence is 44 amino acids long. There is little amino acid sequence homology between the A. cepa root alliinase leader and the leader sequences of A. tuberosum, A. cepa, A. sativum, and A. ascalonicum alliinases, although the latter three have high sequence homology with each other. However, allium alliinase leader sequences reported so far have the hydrophobic regions indicative of glycoproteins, and the A. cepa root alliinase is similar in this feature. A. cepa root alliinase cDNA also has a C-terminal extension of 11 amino acids, YGGDEGSYEST, which is absent in the other allium alliinase sequences.
Expression of Root Alliinase A. cepa root alliinase cDNA was highly expressed in A. cepa roots, but expressed to a very low extent in A. cepa leaf and bulb tissue (Fig. 5A). Alliinase cDNA from A. cepa leaf tissue was expressed in leaf and bulb tissue, but not in roots (Fig. 5B). Loadings of RNA were similar for both root and leaf tissue, but lesser for bulb tissue (Fig. 5C).
| |||||||||||||
DISCUSSION We report here the purification and characterization of an allium protein and show, for the first time to our knowledge, that the enzyme has both alliinase and cystine lyase activity. The protein from A. cepa roots is coded for by a cDNA with low sequence homology to other allium alliinases. The existence of a protein with alliinase activity, but with low sequence homology to A. sativum alliinase cDNA was postulated by Rabinkov et al. (1994) for roots of A. sativum. A. cepa root alliinase was purified in three steps into two isoforms of alliinase specific activities 3,380 (isoform I) and 1,995 (isoform II) nkat mg−1 protein. The isoform specific activities are considerably greater than that reported for bulb alliinase of approximately 750 nkat mg−1 protein (Nock and Mazelis, 1987; Hanum et al., 1995; Clark et al., 1998). On SDS-PAGE the A. cepa root alliinase isoforms I and II separated at 52.7 and 50.5 kD, respectively. The A. cepa bulb alliinase is a doublet on SDS-PAGE with masses of 53.3 and 51.6 kD (Clark et al., 1998). A. cepa root alliinase isoforms I and II gave activity at 51 and 57.4 kD on gel filtration FPLC, indicating that root alliinase is active as a monomer. Alliinase in A. cepa has previously been reported as a trimer or tetramer (Nock and Mazelis, 1987; Lancaster and Boland, 1990; Hanum et al., 1995), while Clark et al. (1998) showed A. cepa bulb alliinase was also active as a monomer as well as a trimer and tetramer. In A. sativum and wild garlic (Allium ursinum) alliinase has been shown to aggregate with low-molecular-mass lectins into stable, active complexes (Smeets et al., 1997). Aggregation of alliinase into multimeric forms does not seem to occur in A. cepa roots. The separation of isoforms of alliinase with IEF was reported for A. cepa bulbs by Nock and Mazelis (1987), although the pI of the bands was not specified. Our own experiments (J.E. Lancaster and M.L. Shaw, unpublished results) showed that it was difficult to focus A. cepa bulb alliinase into bands, although they were present. The tendency of alliinase to aggregate with lectins may have contributed to the technical difficulty of focusing A. cepa bulb alliinase protein. In contrast, the A. cepa root alliinase protein focused well into one band for alliinase isoform I and four different bands for alliinase isoform II. Leek (Allium porrum) alliinase protein gave bands of pI 7.5 and 7.6 (Landshuter et al., 1994), similar to the pIs of 7.6, 7.9, 8.1, and 8.3 for A. cepa root isoform II. In contrast, A. ursinum alliinase protein had a low pI of 4.7 (Lohmüller et al., 1994). This low pI for A. ursinum alliinase could be because of the complexing of the protein with lectins. The presence of lectins is a feature of allium cloves/bulbs. Lectins focus at acidic pIs, and lectin alliinase complexes tend to migrate toward the anode. A. cepa root alliinase separated into two isoforms on Con A-Sepharose chromatography with 5 mm methyl-α-d-glucopyranoside, and 0.2 m methyl-α-d-mannopyranoside, indicating the existence of forms of alliinase with significantly different glycosylation. Alliinase isoform I was recognized by monoclonal antibody 2.23, a monoclonal antibody which specifically recognizes Xyl/Fuc-containing complex-type N-linked glycans (McManus et al., 1988) with the α-1,3 Fuc sugar a critical part of the monoclonal antibody epitope (Wilson et al., 1998). The occurrence of complex type glycans is confirmed by elution of isoform I from the Con A-Sepharose column with 5 mm methyl-α-d-glucopyranoside. Alliinase isoform II reacts both with monoclonal antibody 2.23 and the GNA lectin probe, indicating the occurrence of Xyl/Fuc containing complex-type glycans (monoclonal antibody 2.23-recognized) and oligo-Man structures in at least one of the potential glycosylation sites and/or hybrid chains with terminal Man structures. Treatment of isoform II with endo H did not cause a reduction in molecular mass, suggesting that oligo-Man structures are not present. Recognition by both monoclonal antibody 2.23 and the GNA lectin is most likely due to the presence of hybrid chains with terminal Man structures that have been characterized on other plant glycoproteins (Lerouge et al., 1998). The occurrence of accessible (terminal) Man structures is confirmed by the requirement for 0.2 m methyl-α-d-mannopyranoside to elute isoform II from ConA-Sepharose. A. cepa bulb alliinase eluted from Con A-Sepharose with 5 mm methyl-α-d-glucopyranoside (Clark et al., 1998). Both A. cepa root alliinases (I and II) showed similar enzymatic activity with the range of substrates tested. For reasons that were not apparent, A. cepa root alliinase isoform I was inactive when tested for alliinase activity in situ in the IEF gel. A. cepa roots contain PrenCSO and MCSO, but PCSO was not detected. A. cepa bulb alliinase gave Km values for PrenCSO, PCSO, and MCSO of 6, 11, and 34 mm, respectively (Schwimmer, 1969). These are of the same magnitude and order as A. cepa root alliinase I and II, although the Km for AllylCSO was not reported for bulb alliinase. A. cepa bulb alliinase was reported to be inactive with cystine, Cys, and methyl Cys (Schwimmer and Mazelis, 1963). A major difference between the A. cepa root and bulb alliinases is the activity toward cystine. A. cepa root alliinase isoforms I and II were both active with cystine, and thus have both cystine lyase and Cys sulfoxide lyase activities. In contrast, only C-S lyases with Cys sulfoxide lyase activity have been reported for alliums (Ramirez and Whitaker, 1998). In Brassica species, C-S lyases have both Cys sulfoxide and cystine lyase activity (Ramirez and Whitaker, 1998). To our knowledge, the A. cepa root alliinase isoforms I and II are the first reported C-S lyases from alliums to have both functions. The A. cepa root alliinase did not have cystathionase activity. A. cepa bulb alliinase has also been shown to have activity with Cys-S conjugates such as S-chloronitrophenyl-L-Cys (Kitamura et al., 1997). A. cepa root alliinase was not reacted with these substrates. A. cepa root alliinase cDNA contained an open reading frame of 1,493 bp, which translated to a mature protein of 453 amino acids, a predicted mass of 51.2 kD, and a leader sequence of 44 amino acids. The predicted mass was similar to that determined experimentally for A. cepa root alliinase protein, and the predicted pI was similar to those determined experimentally. The homology between the deduced amino acid sequence and the 25 and 40 amino acid sequences experimentally derived indicated that the cDNA was indeed alliinase. The A. cepa root cDNA also contains four potential Asn glycosylation sites at amino acid positions 55, 143, 182, and 188. A. cepa root alliinase protein has an actual glycosylation site at Asn-143 but not Asn-55, and possible sites at Asn-182 and Asn-188. Rabinkov et al. (1995) generated peptides from A. sativum alliinase with cyanogen bromide. The peptides were separated on SDS-PAGE, stained to detect carbohydrate moieties, and amino acid sequenced. Of the four possible glycosylation sites, only one, Asn-146, was actually glycosylated in the A. sativum mature protein. Similarly, we have found that in A. cepa bulb alliinase, Asn-328 is glycosylated and Asn-146 and/or Asn-191. Asn-19 was not glycosylated, (U. Wäfler, M.L. Shaw, and J.G. Lancaster, unpublished results). The alignment of allium alliinase deduced amino acid sequences showed a consensus Asn glycosylation site only at Asn-146 (or Asn-143 for A. cepa root alliinase). It is likely that the glycosylation of site Asn-146 is necessary for alliinase activity. Glycosylation of A. cepa root alliinase at Asn-182 and Asn-188 is possible, but unlikely. Site-directed mutagenesis experiments in A. tuberosum (Manabe et al., 1998) and PyrP-labeling studies in A. cepa (Kitamura et al., 1997) have shown that a Lys in the region of 280 to 285 amino acids from the N terminus is essential for alliinase activity. The allium consensus sequence around this site occurs also around Lys-248 for A. cepa root alliinase cDNA. Lys-248 is likely to be the PyrP-binding site for A. cepa root alliinase. The region around Lys-280 to -285 in allium alliinase cDNAs is also conserved in C-S lyases for the metabolism of Cys, homo-Cys, and Met in Arabidopsis cystathionine β-lyase, and cystathionine γ-synthase; Saccharomyces cerevisiae cystathionine γ-lyase, cystathionine β-lyase, and cystathionine γ-synthase; Pseudomonas aeruginosa O-succinyl homo-Ser sulfhydrylase and Met γ-lyase (Manabe et al., 1998). The reactions catalyzed by these enzymes comprise the cleavage or formation of Cβ-S or Cγ-S bonds of the amino acid derivatives. Despite the conservation of this sequence for C-S lyase activity, the A. cepa root alliinase is the only allium protein shown to have cystine activity and Cys sulfoxide activity. The other allium alliinases are more restricted in their substrate specificity. Homology between alliinase cDNA-deduced amino acid sequences of A. cepa (bulb and leaf), A. sativum, and A. ascalonicum was very high at > 90% (Van Damme et al., 1992, Clark, 1993). The A. tuberosum alliinase-deduced amino acid sequence was only 66% to 69% homologous to these other alliinases. A. cepa root cDNA deduced amino acid sequence was the most divergent at about 50%. The A. cepa root alliinase sequence had high homology to the AOB249 partial clone. This clone mapped to a different locus from A. cepa bulb alliinase (King et al., 1998). If the remainder of the AOB249 sequence is homologous to A. cepa root alliinase, then this clones' different locus is further evidence for the dissimilarity of A. cepa root alliinase from other alliinases. Expression studies using RNA and northern analysis showed that A. cepa root alliinase cDNA was expressed to a much greater extent in roots than in leaves and bulb, and confirmed that alliinase leaf cDNA was not expressed in roots. As A. cepa alliinase bulb cDNA has a very high homology to leaf cDNA (approximately 90%), such a probe, had it been used, would not hybridize to A. cepa root RNA. Thus, we have confirmed in A. cepa what Rabinkov et al. (1994) postulated for garlic: the existence of a cDNA coding for a protein with alliinase activity that did not hybridize with alliinase bulb cDNA. As well as being divergent in sequence from the other allium alliinases, the A. cepa root alliinase has wider C-S lyase activity. A. cepa root alliinase may have a function in sulfur assimilation and remobilization in roots. It would be of interest to know if alliinase cDNAs with wide sequence divergence and wider C-S lyase activity are found in the roots of other alliums. | |||||||||||||
ACKNOWLEDGMENTS We thank Dr. Michael Havey (University of Wisconsin, Madison) for the gift of AOB 249 probe, Lyn Watson (Institute of Molecular Biosciences, Massey University, Palmerston North, New Zealand) for technical assistance, and Catriona Knight (Department of Biochemistry, Auckland University, New Zealand) for the amino acid sequencing. | |||||||||||||
Footnotes 1This work was funded by New Zealand Foundation for Research, Science and Technology (contract no. CO2603). | |||||||||||||
LITERATURE CITED
| |||||||||||||
