Eur. J. Biochem. 267, 2649±2657 (2000) q FEBS 2000

Purification, molecular cloning and ethylene-inducible expression of asoluble-type epoxide hydrolase from soybean (Glycine max [L.] Merr.)

Masaomi Arahira, Van Hai Nong, Kyoko Udaka and Chikafusa Fukazawa

Genetic Engineering Laboratory, National Food Research Institute, Ministry of Agriculture, Forestry and Fisheries, Tsukuba Science City,

Ibaraki, Japan

A soybean protein was purified from mature dry seeds. Amino-acid sequencing of the nine internal peptides

derived from this N-terminally blocked protein showed that it has a significant similarity to the soluble epoxide

hydrolases known to date. A degenerate series of 23-mer oligonucleotides with sequences corresponding to an

internal region of eight amino-acid residues was synthesized as a probe mixture for detection of a putative

epoxide hydrolase cDNA in a developing cotyledon cDNA library. The 1332-bp cDNA obtained was found to

have an open-reading frame encoding the seed epoxide hydrolase-like precursor consisting of 341 amino-acid

residues, suggesting that 25 amino-acid residues upstream from the second methionine correspond to a transit

peptide. Employing an Escherichia coli expression system, the putative mature epoxide hydrolase-like protein

was overexpressed and purified to homogeneity. This recombinant protein was confirmed to exhibit its epoxide-

diol converting activity using styrene oxide as substrate. The Vmax and Km values for styrene oxide are

1.36 mmol´min21´mg21 and 1500 mm, respectively. Sedimentation equilibrium experiments showed that the

active form of this epoxide hydrolase is monomeric in solution. Using the above cDNA as a probe, a 12-kb

genomic clone was selected and the sequence of a 1933-bp fragment from this clone was found to cover the entire

coding region together with 5 0- and 3 0-flanking regions of the soybean epoxide hydrolase gene. The coding

region of the gene, interrupted by two short introns, was identical to the corresponding regions of the cDNA.

Northern blot analyses showed that this epoxide hydrolase gene was expressed strongly at a very early stage

(13 days after flowering) and then the level of expression gradually decreased and almost ceased at a very late

stage (58 days after flowering) of seed development, whereas its expression was markedly up-regulated by

ethylene treatment. In stems (hypocotyl portion), the epoxide hydrolase transcript was detected at significant

levels and was also up-regulated in response to ethylene. On the other hand, it is hardly expressed in leaves, even

though they were treated with the phytohormone. Overall, the results obtained may indicate that soluble-type

epoxide hydrolase mRNA is expressed at the maximum level in an early stage of seed development. Later, oil

bodies are formed and subsequently epoxy fatty acids, naturally occurring metabolites, accumulate within those

bodies. The temporal induction of this epoxide hydrolase transcript in some tissues in response to ethylene also

indicates that this epoxide hydrolase may play a crucial role in self-defense systems of plant.

Keywords: enzyme purification; epoxide hydrolase; expression analysis; molecular cloning; soybean seeds.

During the development of plant seeds to maturation, the seedsare exposed to various environmental stresses, both abiotic andbiotic, including attack by fungal, bacterial and viral pathogens.As in the case of defense of vegetative tissues, for seed defense,plants synthesize and utilize a number of defense proteins [1,2].Lectins and protease inhibitors, occurring in seeds in relativelylarge amounts compared to storage proteins, are among the

diverse type of seed defense-related proteins that have beenrelatively well-characterized both biochemically and in termsof molecular biology [2±4]. However, the complexity of theplant defense gene expression system and the coordinatedregulation of different defense-related genes involved in theresponse to different environmental stimuli during seeddevelopment remain mostly to be elucidated.

We are particularly interested in the structure of defense-related genes and their expression during soybean seeddevelopment. Recently, we cloned and characterized a soybeanseed-specific chitinase belonging to the PR-8 chitinase family,which was proposed to function only in seed defense [5]. Wealso identified several transcripts known to be expressed inresponse to oxidative stress. These include genes encoding acopper/zinc superoxide dismutase (Cu/Zn-SOD), an ironsuperoxide dismutase (Fe-SOD) [6] and a lipoxygenase(M. Arahira, V. H. Nong, K. Udaka and C. Fukazawa,unpublished data) from developing soybean seeds.

Plant seeds are known to contain various epoxy fatty acidsand their derivatives that probably function as naturalfungicides for seed defense against fungi [7±9]. One of the

Correspondence to C. Fukazawa, Genetic Engineering Laboratory,

National Food Research Institute, Ministry of Agriculture, Forestry

and Fisheries, Tsukuba Science City, Ibaraki 305-8642, Japan.

Fax: 1 81 298 38 8137, Tel.: 1 81 298 38 8137,

E-mail: fukazawa@ss.nfri.affrc.go.jp

Abbreviation: DAF, day(s) after flowering.

Enzymes: epoxide hydrolase (EC 3.3.2.3); arginyl endopeptidase,

lysyl endoprotease (EC 3.4.21.50); pyroglutamyl-peptidase, 5 0-oxoprolyl-

peptidase (EC 3.4.21.50).

Note: the nucleotide sequences reported in this paper have been submitted

to the EMBL/GenBankTM/DDBJ and are available under the accession

numbers X78547and X78548.

(Received 10 January 2000, accepted 6 March 2000)

2650 M. Arahira et al. (Eur. J. Biochem. 267) q FEBS 2000

enzyme families involved in the pathway for biosynthesis anddegradation of epoxides is epoxide hydrolase, which catalyzesthe conversion of various epoxides to their corresponding diolsby the addition of a water molecule [10]. Such epoxideactivating or detoxifying enzymes are well-characterized inmammals, where they appear in four principal forms, includingsoluble and microsomal forms, in terms of substrate specifity,and are classified as a/b-hydrolase-fold family enzymes [10].In the case of plants, epoxide hydrolase activity has beendetected in spinach [11], apple [12] and broad bean [9]. Theactivity in broad bean is the only case in which an epoxidehydrolase activity has been found in an extract of seeds.However, this enzyme has not yet been purified andcharacterized. Although a soluble epoxide hydrolase has beenpurified from soybean seedlings and characterized enzymati-cally [13±15], the molecular structure of the epoxide hydrolaseand its expression in soybean has not been reported. Untilrecently, cDNAs encoding soluble epoxide hydrolases fromseedlings of Arabidopsis subjected to water-stress, potatotubers, and from tobacco mosaic virus-infected tobacco leaveshave been cloned [16±18].

We here report the purification of a soluble epoxidehydrolase from soybean seeds, cloning of the gene, and itsspatial and temporal expression during seed development andupon the ethylene treatment.

M A T E R I A L S A N D M E T H O D S

Materials

Fresh defatted flour prepared from mature soybean seeds(Glycine max), kindly provided by Fuji Seiyu Co. (Japan), wasused for the enzyme purification in this study. Immaturesoybean seeds, var. Bonminori, were harvested at specifiedintervals ranging from a very early stage of development (< 13days after flowering, DAF) to a very late stage (< 58 DAF). Asample of the seeds harvested at 48 DAF was exposed toethylene gas for 6 h at room temperature. The seeds wereaseptically removed from the pods, frozen in liquid nitrogenand stored at 280 8C until use. Stems (hypocotyl portion) werecollected from 5-day-old seedlings grown in the dark (withoutand with ethylene gas treatment for 6 h in double plastic bags)and leaves from 20-day-old plants (without and with ethylenegas treatment for 8 h), were used freshly. [a-32P]dCTP(3000 Ci´mmol21) and [g-32P] ATP (5000 Ci´mmol21) werepurchased from Amersham. Reverse transcriptase (RousAssociated Virus 2) and Taq DNA polymerases for molecularcloning were obtained from Takara Shuzo. Restriction endo-nucleases were obtained from Nippon Gene. Arginyl endo-peptidase and lysyl endoprotease were from Takara Shuzo andWako Chemicals, respectively. Oligonucleotides were syn-thesized and purified by Sawady Technology as customsyntheses. Tris (2-amino-2-hydroxymethyl-1,3-propandiol),and styrene oxide were purchased from Wako Chemicals.Styrene glycol was from Aldrich.

Purification of soybean seed epoxide hydrolase-like proteinsoybean seed epoxide hydrolase-like protein, which was tightlycopurified with the seed-specific chitinase Chib2, was achievedby the procedure employed for the purification of the latterenzyme [5]. Briefly, after extraction of soybean flour with20 mm sodium acetate buffer, pH 5.0 at 4 8C (5 kg flour in15 L buffer), precipitation at 40% (w/v) ammonium sulfate anddialysis against the same buffer supplemented with 1 mm2-mercaptoethanol, the dialyzed protein preparation was loadedonto a butyl Toyopearl 650 m column (Tosoh, Japan). Upon

elution of the adsorbed proteins from the column, the 28-kDachitinase [5] was accompanied by a 33-kDa protein. The latterprotein was separated completely from chitinase Chib2 bychromatofocusing using a Mono P column (Pharmacia LKBBiotechnology Ltd, Sweden). Chromatofocusing separationwas achieved in 2 h by a linear pH gradient in therange 4.0±6.2 using the histidine buffer, pH 6.2, and Polybuffer74-HCl, pH 4.0. Because the N-terminal amino acid of this33-kDa protein was blocked, several internal peptides,generated by arginyl and/or lysyl endopeptidases, wereseparated and sequenced. Peptide isolation and sequencingwere performed as described previously [19]. The C-terminalsequencing runs were also performed on a PerkinElmer/Applied Biosystems Procise 494C sequencer by the methodof Boyd et al. [20,21]. Protein concentrations were determinedby the method of Bradford [22] using a protein assay kit(Bio-Rad Laboratories) with crystalline bovine serum albumin(Seikagaku Kogyo Co.) as a standard protein. SDS/PAGEpatterns were analyzed by Laemmli's procedure [23].

Mass spectrometric analysis

To measure the molecular masses of native and recombinantepoxide hydrolases, respectively, each sample (about 40 mg)was put onto a reversed-phase HPLC column (Silica ODS120T, f 4.6 � 150 mm, Tosoh Co.) for desalting the samples,equilibrated with 0.1% trifluoroacetic acid and eluted with a60-min linear gradient of acetonitrile (0±60%) at a flow rate of1 mL´min21. The soybean epoxide hydrolase protein wasmonitored by measuring the absorbance at 280 nm. Themolecular mass of the desalted protein was measured byusing a PerkinElmer Sciex triple stage quadrupole ion spraymass spectrometer (model API 365) calibrated over the range of200±3000 mass units using polypropyleneglycols according tothe manufacture's instructions.

Sedimentation equilibrium

To examine whether the purified epoxide hydrolases derivedfrom mature seeds and from an Escherichia coli overexpressionsystem are essentially monomeric in solution, the low-speedsedimentation equilibrium experiments were performed in aBeckman Optima XL-A analytical ultracentrifuge at 7100 g at20 8C. Highly purified epoxide hydrolase protein in 50 mmTris/HCl buffer (pH 8.8) containing 1 mm EDTA wasdiluted to concentrations ranging from 0.4 to 1.2 mg´mL21

with the Tris/HCl buffer. The epoxide hydrolase samples wereloaded into six-channel charcoal-Epon Yphantis cells. Theconcentration gradient in the cell was determined by UVabsorption at 278 nm. The length of time to reach equilibriumwas determined empirically by making successive overlayscans 10 min apart.

Data analysis was carried out employing a nonlinear least-squares program [24] to fit all of the data globally. A density of1.005 g´mL21 for the buffer, measured with a Mettler-PaarDMA-02 density meter, and polypeptide partial specific volumeof the epoxide hydrolase, calculated as 0.736 mL´g21 from theamino-acid composition [25], were used for a computationalcalculation of the molecular mass.

Preparation of polyclonal antibodies

Polyclonal antibodies against the purified soybean seed epoxidehydrolase were raised in random-bred female albino rabbits byimmunizing about 200 mg of purified epoxide hydrolase in1 mL of Freund's complete adjuvant (Yatron), intradermally,

q FEBS 2000 Soybean soluble epoxide hydrolase (Eur. J. Biochem. 267) 2651

according to the immunization schedule described previously[26]. The antiserum prepared was used at a dilution in the range1 : 1000 to 1 : 1500 for immunoblot analysis (Bio-Rad HRPImmunoblotting kit).

Assay of the epoxide hydrolase activity

Epoxide hydrolase activity was assayed using styrene oxide as asubstrate. Styrene oxide is volatile, unstable on storage,unstable in buffer, and reacts with sulfhydril groups. Therefore,it was thought that significant levels of styrene oxide might belost due to evaporation and that some might be lost due toalkylation of proteins during the assay, leading to a change inthe concentration of styrene oxide used. We monitored theepoxide hydrolase-converted styrene glycol and calculated theamount from a calibration curve of authentic styrene glycol bya HPLC system. Under the assay conditions described below, ithas been confirmed that there were no significant peaks derivedfrom styrene oxide degradation products, including theautooxidized product, in the presence and in the absence of aheat-denatured epoxide hydrolase on a chart of the HPLCrecorder (C-R6A, Shimazu). The minimum threshold fordetecting styrene glycol in the HPLC system was around2 nmol. The standard reaction mixture (50 mL) was composedof 100 mm Tris/HCl buffer (pH 8.8), 1 mm EDTA, 2 mmstyrene oxide in 1% acetonitrile and an aliquot of the enzymesolution in a microtube (200 mL). After incubation at 37 8C for5 min, the reaction mixture was centrifuged briefly, frozen inliquid nitrogen and then kept on ice until use. After thawingout, the resulting sample was immediately applied onto areversed-phase HPLC column (Silica ODS 120T, f4.6 � 150 mm, Tosoh Co.) for separating styrene glycol fromits oxide. The column was pre-equilibrated with 4.8%acetonitrile and run with a 30-min linear gradient of acetonitrile(4.8±60%) at a flow rate of 1 mL´min21. We could not detectthe loss of significant levels of styrene oxide throughout theassay system when styrene oxide was incubated in an epoxidehydrolase-omitted reaction mixture for at least 17 min.Substrate and product were monitored by measuring theabsorbance at 258 nm and were eluted at 20.5 min as aretention time and 8.6 min, respectively, under the conditionsdescribed above. Employing a Hans±Wilkinson plot, kineticconstants were calculated from the amount of styrene glycol atsubstrate concentrations varying between 500 mm and 2500 mm[21]. One unit was defined as the amount of enzyme catalyzingthe production of 1mmol of styrene glycol per min in the aboveassay system.

mRNA isolation, construction of cDNA libraries, andscreening of soybean epoxide hydrolase-like protein cDNA

Total RNA was isolated from immature seeds, leaves and stemsof soybean and poly(A)1 RNA was isolated from individualtotal RNA preparations [26]. The cDNA libraries wereconstructed according to the method described previously[19]. To obtain reliable plaque-hybridization signals fromepoxide hydrolase-like protein cDNA clones, three independentcDNA libraries were constructed using mRNAs correspondingto early (18 DAF), middle (38 DAF), and late (58 DAF) stagesof seed development and a 23-mer degenerate oligonucleotide(antisense), corresponding to all possible codons for theselected part of the determined internal 23 amino-acidsequence, NH2-Phe-Gln-Phe-Asn-Pro-Glu-Met-Pro-COOH,conformed to the following: 5 0-GGCAT(T/C)TCIGG(A/G)TT(A/G)AA(T/C)TG(A/G)AA-3 0 was synthesized as a probe.After labeling the mixed oligonucleotide probe at the 5 0-OH

end with T4 polynucleotide kinase (Nippon Gene Inc.) and[g-32P]ATP to a specific activity of greater than4 � 108 c.p.m.´mg21, the purified probe was hybridized tonitrocellulose filters (Schleicher & Schuell) containing replicasof recombinant phage plaques [19,26]. A positive recombinantphage was amplified and purified on CsCl gradients. The insertDNA derived from the recombinant lgt10 phage DNA byEcoRI cleavage was subcloned into the same restrictionenzyme site of pUC18 and pUC19 and then sequencedaccording to the method described previously [19]. DNAsequencing was performed by an automated DNA sequen-cer DSQ 1000 (Shimadzu, Japan) using a ThermoSequenase Cycle Sequencing kit (Amersham). Computeranalyses of DNA sequences obtained were performed usingthe genetyx programs (Software Development Co.) asdecribed previously [27].

Overexpression and purification of the recombinant epoxidehydrolase protein

We assumed that soybean seed epoxide hydrolase-like proteinmight lose its enzymic activity during a time-consuming multi

Fig. 1. Purity analyses of soluble epoxide hydrolase (EH) preparation

isolated from soybean seeds. (A) SDS/PAGE pattern. Protein separation

was carried out on a 10±20% gradient gel (Daiichi Kagaku Yakuhin Co.).

After electrophoresis, protein was stained with Coomassie brilliant blue R-

250. The left lane shows molecular mass markers (Pharmacia Biotech AB).

(B) Isoelectric focusing pattern showing an apparent pI of 5.2 on an

Immobiline thin-layer gel. (C) Determination of the precise molecular mass

of soybean seed EH by means of an ion-spray mass-spectrometer.

2652 M. Arahira et al. (Eur. J. Biochem. 267) q FEBS 2000

step process for the purification, so we attempted to over-produce the protein in E. coli. Based upon the putative matureN-terminus (the second methionine shown in Fig. 2), twoprimers [primer I (sense): 5 0-ATATACATATGGAGCAAA-TAAAGCACAGAACA-3 0, which was constructed to corre-spond to codons for amino-acid residues 26±33 of the epoxidehydrolase-like precursor with 11 nucleotides containing anNdeI site (bases underlined) at the upstream; primer II(antisense): 5 0-GGTTGAATTCGTTTTTGGACAAGATCAG-AACTTC-3 0, which corresponds to codons for the C-terminalregion with the following 3 0-untranslated 27 nucleotidescontaining the stop codon, TGA, and an EcoRI site (basesunderlined)] were used to carry out a PCR amplificationemploying the KpnI-cut cDNA insert isolated from a positiverecombinant plasmid DNA (pSEH 221) as a template. Afterpurification and double-digestion with EcoRI and NdeI, theresultant PCR product was subcloned into the correspondingsites of a expression vector pRSET (Invitrogen). This expres-sion plasmid, designated as pRSEH310, was transformed intoE. coli strain BL21 (DE3), and cells harboring the aboveplasmid were grown at 37 8C in Luria±Bertani mediumsupplemented with ampicillin (50 mg´mL21). When the D650

of the culture reached 0.8, isopropyl thio-b-d-galactoside wasadded to the culture medium at a final concentration of 1 mmand the incubation continued for a further 12 h. The washedcells were suspended in an aliquot of 200 mm acetate (pH 5.0)

containing 0.1 mg´mL21 lysozyme, 1 mm EDTA and 0.1 mNaCl, followed by disrupting by an ultra-sonicator (NihonseikiSeisakushyo Co.) as mentioned previously [28]. After centri-fugation, the supernatant was heated to 55 8C for 10 min,whereby most bacterial proteins were denatured and sub-sequently removed by centrifugation at 8000 g for 20 min at4 8C. The overexpressed 33-kDa soluble recombinant epoxidehydrolase-like protein was purified by gel-filtration on aSephacryl S-200 HR column, followed by chromatofocusingon Mono P (Pharmacia LKB Biotechnology Ltd) according tothe procedure described for seed materials.

Southern analysis of soybean epoxide hydrolase gene andisolation of its genomic clone

Genomic DNA was extracted from young soybean leaves andwas purified by ultracentrifugation on cesium chloridegradients according to the methods described previously [27].The DNA (< 10 mg) was digested with appropriate restrictionenzymes, fractionated on a 0.8% (w/v) agarose gel, transferredonto a Zeta-Probe membrane using a VACUGENE 2016vacuum blotting unit (Pharmacia LKB Biotechnology Ltd),and baked under vacuum. Hybridization using the epoxidehydrolase cDNA fragment labeled with [a-32P] dCTP as aprobe (DNA labeling kit, Nippon Gene), was essentially asdescribed previously [5]. A soybean genomic library in phage

Fig. 2. Comparison of cDNA-deduced primary sequence of soybean soluble epoxide hydrolase with those of three plant homologues extracted from

the databases. SOYBEAN: Glycine max epoxide hydrolase, EMBL X75847; ARABIDOPSIS: Arabidopsis thaliana homologue, EMBL D16628; POTATO:

Solanum tuberosum homologue (clone 3.1), EMBL U02494; TOBACCO: Nicotiana tabacum homologue, EMBL U57350. The initial and second methionines

are highlighted. The amino-acid segments that could be identified by protein sequencing are underlined. The amino-acid residues that may form the so-called

catalytic triad composed of a nucleophilic acid residue (N), an acidic residue (A) and a histidine (H) are indicated by arrowheads. The arrows indicate the

positions where intervening sequences are inserted. Identical amino acids are shown by asterisks.

q FEBS 2000 Soybean soluble epoxide hydrolase (Eur. J. Biochem. 267) 2653

vector EMBL3 (Stratagene) was used for screening the epoxidehydrolase gene. A cDNA probe of about 1.3 kb derived fromthe pSEH226 was separated electrophoretically, excised andthen electro-eluted. The purified insert was then labeledwith [a-32P]dCTP according to the random priming methodemploying Nippon Gene labeling kit. The labeled probe washybridized to nitrocellulose-filters (Schleicher & Schuell),containing replicas of recombinant phage plaques at 42 8Covernight, washed and then exposed to an X-ray film (Konica).After purification, positive recombinant phages were amplifiedand purified on CsCl gradients. Approximately 10 mg of thephage DNA was digested by several restriction enzymes,indicating that SalI cut the entire insert having a 12-kb as a onefragment out from the positive recombinant phage. ThisSalI-generated recombinant DNA fragment was subcloned intothe same restriction enzyme site of pUC19 vector, followed bysequencing [5].

Northern hybridization samples of poly(A)1 RNA (0.5 mg)were electrophoresed on a 1% (w/v) agarose gel afterdenaturation with glyoxal and dimethyl sulfoxide and then,blotted onto a nylon membrane Zeta-Probe (Bio-Rad). Themembrane was baked at 80 8C under vacuum for 2 h.Hybridization was performed using a 32P-labeled epoxidehydrolase cDNA probe and then the membrane was washedaccording to the procedure described previously [26]. To assaythe amount of mRNA contained in each mRNA preparationused, we measured Cerenkov's counts of 32P-incorporatedcDNA synthesized by a procedure described previously [26].As a standard, we used 4.7 � 105 c.p.m. of the synthesizedcDNAs as the equivalent to 1 mg of mRNA.

R E S U LT S

Purification and characterization of a soluble-type epoxidehydrolase from soybean mature seeds

A protein, copurified with a novel type PR-8 chitinase termedChib2 [5,27], was separated into homogeneous state as judgedby several separation techniques based upon different principles(Fig. 1). This single 33-kDa SDS/PAGE-purified protein had amolecular mass of 36 171 ^ 11 Da with a pI of 5.2 (Fig. 1).On the other hand, sedimentation equilibrium experimentsshowed that molecular mass of this protein is around 36 000 Da(data not shown). The data suggest that the protein existslargely as a monomer in solution. The molecular disparitybetween an SDS/PAGE analysis and electrospray ionizationmass spectrum may suggest that this protein has a rigidlystacked molecular shape. To further characterize this proteinwe performed protein sequencing experiments, followed bycloning the corresponding cDNA. The N-terminus of thisprotein was blocked, so that several deblocking methods suchas formic acid treatment [29], pyroglutamyl-peptidase [30], andprotein N-terminal deblocking [31] were applied to thisprotein using a Takara deblocking kit (Takara Shuzo).However, we could not release a modified group from theprotein N-terminus. Therefore, nine internal peptides generatedby lysyl- and/or arginyl-endopeptidases were separated fromindividual digest and then, eight to 25 amino-acid residues ofeach purified peptide were determined (shown by underlines inFig. 2). The C-terminal analysis showed that its terminaltripeptidyl sequence is NH2-Asn-Lys-Phe-COOH. A compu-tational analysis based upon the determined amino-acidsequences suggested that this protein has significant similarityto soluble-type epoxide hydrolase family. To detect whetherthis epoxide hydrolase-like protein derived from soybean

mature seeds catalyzes the conversion of an epoxide into itsdiol, styrene oxide, a compound commonly used in earlyresearch on epoxide hydrolase, was used as a substrate andallows comparison of our data with that of other studies [32].However, no epoxide hydrolase activity was observed at theprotein amounts varying between 2.0 mg and 10.5 mg under theassay conditions used here.

Isolation and characterization of a cDNA encoding theepoxide hydrolase-like protein

To clarify whether this purified protein corresponds to a soluble(cytosolic)-type epoxide hydrolase family of plants in terms ofthe primary sequence, we have attempted to isolate thecorresponding cDNA. Because we could not speculate on theoptimal seed-developmental stage at which a high levelexpression of mRNA species encoding the epoxide hydrolase-like protein occurred, three separated recombinant phagelibraries were constructed by using mRNAs derived from 18DAF, 38 DAF and 58 DAF seeds. A plaque that showedmoderate but clear hybridization signal was selected from9.8 � 105 independent recombinant plaques of the 18 DAFcDNA library, but from the other two libraries no positiveplaques were obtained after screening up to about 2 � 106

independent recombinant plaques. The resulting 1313-bpcDNA insert (accession number, X78547), except for thepoly(A) segment, of a recombinant plasmid, pSEH 221, wasfound to contain an open reading frame composed of 341 aminoacids (calculated molecular mass, 39 012 Da). It is important tonote that the cDNA-deduced protein contained exactly theamino-acid segments of the authentic protein determined bydirect protein sequencing (see Fig. 2, underlined regions). Atthe 5 0-terminus, 72 untranslated nucleotides were foundpreceding the initial AUG codon and at the 3 0-terminus, a

Fig. 3. Purification of recombinant soybean seed epoxide hydrolase

produced in Escherichia coli. CBB, SDS/PAGE pattern of a recombinant

epoxide hydrolase preparation stained by Coomassie brilliant blue R-250;

WB, Western blot of the recombinant epoxide hydrolase using polyclonal

antibodies against the purified authentic protein (see Fig. 1).

2654 M. Arahira et al. (Eur. J. Biochem. 267) q FEBS 2000

218 nucleotide untranslated region adjacent to the 19 nucleotidepoly(A) tail was found. The comparison of the amino-acidsequence deduced from the cDNA with those of three plantepoxide hydrolases known to date proved that the overallidentity at the amino-acid sequence permitting introduction ofproper hypothetical deletions was 57% (Arabidopsis), 54%(potato), and 36% (tobacco). Importantly, this primarysequence also contains the catalytic triad consisting of anucleophilic amino acid (Asp126), an acidic catalytic residue(Asp 285) and a catalytic histidine (His320) in conservedpositions as compared to those determined for mammalian andbacterial enzymes [33±35]. These confirm that this proteinfrom soybean mature seeds is a plant soluble-type epoxidehydrolase. The first methionine codon is located 25-amino-acidresidues upstream of the second methionine which is conservedtightly together with the following residue, Glu, among planthomologs (Fig. 2). To investigate the hydropathic character ofthe N-terminal 40-amino-acid residues of the soybean epoxidehydrolase, we applied the method of Kyte and Doolittle [36].The pattern, expressed as a function of hydropathic index,showed that the amino-acid segments spanning residues 1±15and 20±23 of the epoxide hydrolase were hydrophobic andthe segments spanning residues 16±19 and 24±33 werehydrophilic, indicating the typical alternating pattern ofhydrophobicity and hydrophilicity (data not shown). Because of

the N-terminal blockage, we have no information on theN-terminus of authentic epoxide hydrolase, but Ile13 waspredicted as the mature N-terminus of the cDNA-deducedprotein according to the method of von Heijne [37]. Thus, theabove analyses strongly suggest that soybean seed epoxidehydrolase contains an N-terminal transit-peptide. Based upon amolecular mass of 36 171 ^ 11, Da determined by massspectrometric analysis, and Phe341 as the intact C-terminuswe estimated the putative N-terminal position of the matureepoxide hydrolase from the inferred entire sequence. The cal-culated molecular mass of the amino-acid sequence spanningresidues 26±341 (from the second Met to the C-terminal Phe) ismore close to the experimental value than any other cases, andis 36 116 Da. This molecular disparity is beyond the error ofmass spectrometric analysis (substantial deviation is less than0.03%) and is 55 ^ 11 Da, suggesting that this deviation maybe possibly due to the acetyl group (43 Da).

Confirmation of catalyzing the conversion of an epoxide intoits diol by E. coli-produced epoxide hydrolase protein

Authentic epoxide hydrolase derived from mature soybeanseeds did not convert styrene oxide into its diol under the assayconditions used. In attempting to define the causes that affectthe oxide-converting activity, the first point to consider is the

Fig. 4. Southern blot analysis of soybean

EH, epoxide hydrolase (EH) genes (A) and

structure and restriction map of a 1993-bp

genomic DNA fragment containing the EH

gene as compared to EH cDNA (B).

(A) Samples of 10 mg of DNA were digested with

appropriate restriction enzymes in single or

double digestions. The size markers are indicated

on the left. (B) Two introns are shown as bold

lines. This fragment, isolated from a 12-kb insert

of a soybean genomic clone, was initially

analyzed by sequencing. Further upstream and

downstream regions are not shown and are

marked as double flashes.

q FEBS 2000 Soybean soluble epoxide hydrolase (Eur. J. Biochem. 267) 2655

substrate used, styrene oxide. As styrene oxide is both anirreversible inhibitor (alkylation of the enzyme) and a poorsubstrate for the murine soluble enzyme [38], relatively largeamount of the enzyme may be necessary to catalyze theconversion of styrene oxide into the diol. The second is due todenaturation through the time-consuming multi step processfor the purification. Employing a epoxide hydrolase-expression plasmid, pRSEH310, covering the amino-acidsegment spanning residues 26±341 of the deduced entireepoxide hydrolase protein (Fig. 2), we tried to produce therecombinant enzyme, followed by confirming the oxide±diolconverting activity. The recombinant epoxide hydrolase wasoverproduced and subsequently purified to homogeneity asjudged by SDS/PAGE (Fig. 3), mass spectrometric analysis(not shown), sedimentation equilibrium analysis (not shown),and immunoblotting (Fig. 3). The yield of the purifiedrecombinant epoxide hydrolase was about 80 mg per 1 L ofthe culture medium. Employing sufficient amount of therecombinant epoxide hydrolase (protein amounts varyingbetween 5.2 mg and 64 mg per reaction mixture), we couldmeasure its catalyic activity using styrene oxide as substrate(data not shown). Kinetic analysis showed that the Vmax andKm values of the recombinant epoxide hydrolase were1.36 mmol´min21´mg21 and 1500 mm, respectively. Thesevalues resemble the corresponding values from murine enzyme[38]. This recombinant epoxide hydrolase also exhibited themolecular mass of 36 000 Da by sedimentation equilibriumanalysis, indicating that the active form of this enzyme ismonomeric.

Organization of the soybean epoxide hydrolase gene

Southern blot analysis revealed the occurrence of several majorDNA fragments capable of hybridizing with the cDNA probe ineach restriction digest (Fig. 4A), suggesting the existence of asmall family of epoxide hydrolase genes in soybean. From apositive recombinant phage containing a 12-kb insert, a1933-bp-DNA region covering the entire soybean epoxidehydrolase gene was sequenced (accession number, X78548). Asshown in Fig. 4B, soybean epoxide hydrolase gene spans1630 bp from the putative cap site (G) to the polyadenylationsite, and is interrupted by two short intervening sequences(IV-I, 168 bp and IV-II, 148 bp) occurring after the codon forGln180 and in the codon for Leu260 (between the second andthird bases of the codon). The TATA motif is located 63 bpupstream from the putative cap site. The nucleotide sequence ofthe cDNA and that of the corresponding region including thetranslated and the untranslated parts of this genomic structureare identical, indicating that soybean epoxide hydrolase mRNA

species cloned here is the transcript of this gene. To ourknowledge, this is the first report on the structure of a geneencoding plant epoxide hydrolase.

Spatial gene expression of the epoxide hydrolase duringsoybean development and temporal response to ethylenetreatment Northern analysis (Fig. 5) showed that the highestlevel of epoxide hydrolase gene expression occurs at a veryearly stage of seed development (13±18 DAF). Thereafter, itgradually decreased and almost disappeared at the late stage (58DAF). However, the expression was reactivated in response toethylene-treatment (C2H4/48 DAF). In case of stems, thismRNA expressed ubiquitously was also up-regulated by thephytohormone, whilst the mRNA was hardly detectable inleaves without and with ethylene treatment. As it is thought thatethylene is synthesized as a signal substance of wounding, theseresults imply that soybean epoxide hydrolase genes areexpressed in response to environmental stimuli in a mannerregulated spatially and temporally as suggested by previouswork with Arabidopsis and tobacco [16,18].

D I S C U S S I O N

It has been thought that plant epoxide hydrolase is involved inthe biosynthetic pathway of cutin monomer, which consists of alipoxygenase/peroxygenase/epoxide hydrolase cascade [39].Cutin covers all aerial parts of a plant body as one of thestructural defense systems; plant seeds are usually not exposeddirectly to the air so that their surface parts are not composed ofcutin-related materials. Thus, an involvement of the epoxidehydrolase in cutin biosynthesis in seeds, correlated with thehigh level of gene expression in early stages of soybean seeddevelopment (Fig. 5), is unlikely. Styrene oxide is used here asa convenient substrate for assaying the oxide-convertingactivity of soybean epoxide hydrolase. It is, however, labile,poor substrate, similar to its properties as a substrate of themurine epoxide hydrolase case [38]. It is commonly thoughtthat all epoxides are reactive alkylating agents. In contrast tothis, most epoxy fatty acids are relatively stable in the absenceof degrading enzymes. Various epoxy fatty acids and their diolderivatives occur naturally in seeds. Although the physiologicalrole of epoxy linoleic acids such as 9,10-epoxy-12(Z )-octadecenoic acid and 12,13-epoxy-9(Z )-octadecenoic acid isnot clear, they have been found to have antifungal effects asinhibitors of the spore germination and germ tube growth ofrice blast disease (Pyricularia oryzae) [40±42]. Also, theircorresponding diol derivatives exhibit similar antifungal effects[40]. Although the natural substrate for the epoxide hydrolasesin plants has not been identified, it is possible that epoxy fattyacids such as oleic and linoleic acid oxides formed by

Fig. 5. Northern hybridization. The mRNA

preparations examined were from soybean

developing seeds harvested at 13, 18, 28, 38, 48

and 58 DAF, as well as from leaves (L) and

stems (S) without ethylene treatment.

C2H4/48 DAF, C2H4/L and C2H4/S mean that

mRNA isolated from individual tissue after

ethylene-treatment was loaded onto the gel.

The quantity of mRNA loaded on each slot is

the same (0.5 mg each).

2656 M. Arahira et al. (Eur. J. Biochem. 267) q FEBS 2000

cytochrome P450-dependent mixed-function oxidases similarto those identified in mammals [43] are endogeneous substratesfor plant epoxide hydrolases [15,44]. Blee and Schuberdemonstrated that peroxygenase, a strictly hydroperoxide-dependent membrane-associated oxidase occurring in higherplants, catalyzes the epoxidation of mono- and polyunsaturatedfatty acids [44]. In soybean, it has been confirmed thatperoxygenase forms preferentially (9R,10S )-epoxy-12(Z )-octadecenoic acid and (12R,13S )-epoxy-9(Z )-octadecenoicacid, which exist in vivo as major metabolites from linoleicacid [45]. Interestingly, 9,10-epoxy-12(Z )-octadecenoic acid isalso produced in mammals [46], and it is commonly known asleukotoxin (a burn toxin) that is synthesized by leukocytes incorrelation with multiple organ failure and adult respiratorydistress syndrome that appears in severe burn patients [47,48].On the contrary, it has been reported, very recently, that theabove linoleic acid oxide is not directly cytotoxic and the toxiccompound is its diol metabolite which was converted byepoxide hydrolase in in vitro systems and then injected intoanimals intravenously [49].

In soybean, the seed-developmental period is one of the mostproductive and active in the life cycle. Fatty acids are one of thestorage substances and they accumulate in oil bodies of the seedat around middle stage of the development. Preceding themaximum stage for producing the vegetable oil, epoxidehydrolase mRNA is expressed at the maximum level (Fig. 5),supporting the hypothesis that this epoxide activating ordetoxifying enzyme is key to the regulation of the dynamiclevel of epoxy fatty acids and may be transferred into oil bodiesin which epoxide hydrolase may work after the removal of itstransit peptide. Although we have not found either epoxy fattyacids or their diol derivatives to be more toxic to soybean cellsand/or pathogens, the finding that the expression level ofsoybean epoxide hydrolase is up-regulated by ethylene indeveloping seeds and young stems, leads us to speculate thatepoxide hydrolase-converted diols located in oil bodies aretoxic to pathogens such as fungi and invertebrate pests, butare not so toxic to the host cells. Clearly, further experimentsare required in order to establish the full significance of ourobservations.

A C K N O W L E D G E M E N T S

This work was supported by a Basic Research Core System grant from the

Science and Technology Agency of Japan (to C. F.). We are indebted to Mrs

Michiyo Yokota for preparation of this manuscript. V. H. N. is from

Institute of Biotechnology, National Centre for Natural Science and

Technology, Hanoi, Vietnam, and K. U. from Tokyo Kasei University,

Itabashi-ku, Tokyo, Japan.

R E F E R E N C E S

1. Ryan, C.A. & Jagendorf, A. (1995) Self defense by plants. Proc. Natl

Acad. Sci. USA 92, 4075±4075.

2. Bowles, D.J. (1990) Defense-related proteins in higher plants. Annu.

Rev. Biochem. 59, 873±907.

3. Peumans, W.J. & Van Damme, E.J. (1995) Lectins as plant defense

proteins. Plant Physiol. 109, 347±352.

4. Birk, Y. (1996) Protease inhibitors in legume seeds ± overview. Arch.

Latinoam. Nutr. 44, 26S±30S.

5. Yeboah, N.A., Arahira, M., Nong, V.H., Zhang, D.-Y., Kadokura, K.,

Watanabe, A. & Fukazawa, C. (1998) A class III acidic endochitinase

is specifically expressed in the developing seeds of soybean (Glycine

max [L.] Merr.). Plant. Mol. Biol. 36, 407±415.

6. Arahira, M., Nong, V.H., Kadokura, K., Kimura, K., Udaka, K. &

Fukazawa, C. (1998) Molecular cloning and expression patterns of

Cu /Zn-superoxide dismutases in developing soybean seeds. Biosci.

Biotech. Biochem. (Tokyo) 62, 1018±1021.

7. Badami, R.C. & Patil, K.B. (1981) Structure and occurrence of unusual

fatty acids in minor seed oils. Prog. Lipid Res. 19, 119±153.

8. Ohta, H., Shida, K., Peng, Y.-L., Furusawa, I., Shishiyama, J., Aibara,

S. & Morita, Y. (1990) The occurrence of lipid hydroperoxide-

decomposing activities in rice and the relationship of such activities

to the formation of antifungal substances. Plant Cell Physiol. 31,

1117±1122.

9. Hamberg, M. & Fahlstadius, P. (1992) On the specificity of a fatty acid

epoxygenase in broad bean. Plant Physiol. 99, 987±995.

10. Beetham, J.K., Grant, D., Arand, M., Garbarino, J.E., Kiyosue, T.,

Pinot, F., Oesch, F., Belknap, W.R., Shinozaki, K. & Hammock, B.D.

(1995) Gene evolution of epoxide hydrolyzes and recommended

nomenclature DNA. Cell. Biol. 14, 61±71.

11. Croteau, R. & Kolattukudy, P.E. (1975) Biosynthesis of hydroxyfatty

acids polymers. Enzymatic epoxidation of 18-hydroxyoleic acid to

18-hydroxy-cis-9,10-epoxystearic acid by a particulate preparation

from spinach (Spinacia oleracea). Arch. Biochem. Biophys. 170,

61±72.

12. Croteau, R. & Kolattukudy, P.E. (1975) Biosynthesis of hydroxyfatty

acids polymers. Enzymatic hydration of 18-hydroxy-cis-9,10-epoxy-

stearic acid to threo-9,10,18-trihydroxystearic acid by a particulate

preparation from apple (Malus pumila). Arch. Biochem. Biophys. 170,

73±81.

13. Blee, E. & Schuber, F. (1992) Occurrence of fatty acid epoxide

hydrolases in soybean (Glycine max). Biochem. J. 282, 711±714.

14. Blee, E. & Schuber, F. (1992) Regio- and enantioselectivity of soybean

fatty acid epoxide hydrolase. J. Biol. Chem. 267, 11881±11887.

15. Blee, E. & Schuber, F. (1995) Stereocontrolled hydrolysis of the

linolenic acid monoepoxide regioisomers catalyzed by soybean

epoxide hydrolase. Eur. J. Biochem. 230, 229±234.

16. Kiyosue, T., Beetham, J.K., Pinot, F., Hammock, B.D., Yamaguchi-

Shinozaki, K. & Shinozaki, K. (1994) Characterization of an

Arabidopsis cDNA for a soluble epoxide hydrolase gene that is

inducible by auxin and water stress. Plant J. 6, 259±269.

17. Stapleton, A., Beetham, J.K., Pinot, F., Garbarino, J.E., Rockhold,

D.R., Friedman, M., Hammock, B.D. & Belknap, W.R. (1994)

Cloning and expression of soluble epoxide hydrolase from potato.

Plant J. 6, 251±258.

18. Guo, A., Durner, J. & Klessig, D.F. (1998) Characterization of a

tobacco epoxide hydrolase gene induced during the resistance

response to TMV. Plant J. 15, 647±656.

19. Arahira, M. & Fukazawa, C. (1994) Ginkgo 11S seed storage proteins

family mRNA: unusual Asn-Asn linkage as post-translational

cleavage site. Plant. Mol. Biol. 25, 597±605.

20. Boyd, V.L., Bozziri, M., Zon, G., Noble, R.L. & Mattaliano, R.J.

(1992) Sequencing of peptides and proteins from the carboxy

terminus. Anal. Biochem. 206, 344±352.

21. Wang, Y., Arahira, M. & Fukazawa, C. (1999) An isoelectric separation

of soybean b-amylase isoforms and their enzymic characteristics.

Biosci. Biotech. Biochem. (Tokyo) 63, 726±730.

22. Bradford, M.M. (1976) A rapid and sensitive method for the

quantitation of microgram quantities of protein utilizing the principle

of protein-dye binding. Anal. Biochem. 72, 248±254.

23. Laemmli, U.K. (1970) Cleavage of structural proteins during the

assembly of the head of bacteriophage T4. Nature. 227, 680±685.

24. Johnson, M.L., Correia, J., Yphantis, D.A. & Halvorson, H.R. (1981)

Analysis of data from the analytical ultracentrifuge by nonlinear

least-squares techniques. Biophys. J. 36, 575±588.

25. Laue, T.M., Shah, B.D., Ridgeway, T.M. & Pelletier, S.L. (1992)

In Analytical Ultracentrifugation in Biochemistry and Polymer

Science (Harding, S.E., Rowe, A.J. & Horton, J.C., eds), pp.

90±125. Royal Society of Chemistry, Cambridge, UK.

26. Fukazawa, C., Momma, T., Hirano, H., Harada, K. & Udaka, K. (1985)

Cloning and sequencing of double-stranded cDNA complementary to

a soybean storage protein Glycinin A3B4 mRNA. J. Biol. Chem. 260,

6334±6339.

27. Watanabe, A., Nong, V.H., Zhang, D.-Y., Arahira, M., Yeboah, N.A.,

q FEBS 2000 Soybean soluble epoxide hydrolase (Eur. J. Biochem. 267) 2657

Udaka, K. & Fukazawa, C. (1999) Molecular cloning and ethylene-

inducible expression of Chib1 chitinase from soybean (Glycine max

(L.) Merr.). Biosci. Biotech. Biochem. (Tokyo) 63, 251±256.

28. Totsuka, A. & Fukazawa, C. (1993) Expression and mutation of

soybean b-amylase in Escherichia coli. Eur. J. Biochem. 214,

787±794.

29. Ikeuchi, M. & Inoue, Y. (1988) A new photosystem II reaction center

component (4.8 kDa protein) encoded by chloroplast genome. FEBS

Lett. 241, 99±104.

30. Moyer, M., Harper, A., Payne, G., Ryals, J. & Fowler, E. (1990) In situ

digestion with pyroglutamate aminopeptidase for N-terminal

sequencing electroblotted proteins. J. Protein Chem. 9, 282±283.

31. Tsunazawa, S., Takakura, H. & Sakiyama, F. (1990) Microsequence

analysis of N-acetylated proteins. J. Protein Chem. 9, 265±266.

32. Gill, S., Ota, K. & Hammock, B.D. (1983) Radiometric assays for

mammalian epoxide hydrolases and glutathione S-transferase. Anal.

Biochem. 131, 273±282.

33. Pinot, F., Grant, D.F., Beetham, J.K., Parker, A.G., Borhan, B., Landt,

S., Jones, A.D. & Hammock, B.D. (1995) Molecular and biochemical

evidence for the involvement of the Asp333-His523 pair in catalytic

mechanism of soluble epoxide hydrolase. J. Biol. Chem. 270,

7968±7974.

34. Arand, M., Wagner, H. & Oesch, F. (1996) Asp333, Asp495, and

His523 form the catalytic triad of rat soluble epoxide hydrolase.

J. Biol. Chem. 271, 4223±4229.

35. Rink, R., Fennema, M., Smids, M., Dehmel, U. & Janssen, D.B. (1997)

Primary structure and catalytic mechanism of the epoxide

hydrolase from Agrobacterium radiobacter AD1. J. Biol. Chem.

272, 14650±14657.

36. Kyte, J. & Doolittle, R.F. (1982) A simple method for displaying the

hydropathic character of a protein. J. Mol. Biol. 157, 105±132.

37. Von Heijne, G. (1986) A new method for predicting signal sequence

cleavage site. Nucleic Acids Res. 14, 4683±4690.

38. Dietze, E.C., Kuwano, E. & Hammock, B.D. (1994) Spectrophotomeric

substrates for cytosolic epoxide hydrolase. Anal. Biochem. 216,

176±187.

39. Blee, E. & Schuber, F. (1993) Biosynthesis of cutin monomers:

involvement of a lipoxygenase/peroxygenase pathway. Plant J. 4,

113±123.

40. Kato, T., Yamaguchi, Y., Uehara, T., Yokoyama, T., Namai, T. &

Yamanaka, S. (1983) Defense mechanism of the rice plant against

rice blast disease. Naturwissenschaften 70, 200±201.

41. Kato, T., Yamaguchi, Y., Abe, N., Uehara, T., Namai, T., Kodama, M.

& Shiobara, Y. (1985) Structure and synthesis of unsaturated

trihydroxy C18 fatty acids in rice plants suffering from rice blast

disease. Tetrahedron Lett. 26, 2357±2360.

42. Masui, H., Kondo, T. & Kojima, M. (1989) An antifungal compound

9,12,13-trihydroxy-(E )-10-octadecenoic acid from Colocasia anti-

quorum inoculated with Ceratocystis fimbriata. Phytochemistry. 28,

2613±2615.

43. McCarthy, M.B. & White, R.E. (1983) Functional differences between

peroxidase compound I and the cytochrome P-450 reactive oxygen

intermediate. J. Biol. Chem. 258, 9153±9158.

44. Blee, E. & Schuber, F. (1990) Efficient epoxidation of unsaturated fatty

acids by a hydroperoxide-dependent oxygenase. J. Biol. Chem. 265,

12887±12894.

45. Blee, E. & Schuber, F. (1990) Stereochemistry of the epoxidation of

fatty acids catalyzed by soybean peroxygenase. Biochem. Biophys.

Res. Commun. 173, 1354±1360.

46. Hayakawa, M., Sugiyama, S., Takamura, T., Yokoo, K., Iwata, M.,

Suzuki, K., Taki, F., Takahashi, S. & Ozawa, T. (1986) Neutrophils

biosynthesize leukotoxin, 9,10-epoxy-12-octadecenoate. Biochem.

Biophys. Res. Commun. 137, 424±430.

47. Ozawa, T., Hayakawa, M., Kosaka, K., Sugiyama, S., Ogawa, T.,

Yokoo, K., Aoyama, H. & Izawa, Y. (1991) Leukotoxin, 9,10-epoxy-

12-octadecenoate, as a burn toxin causing adult respiratory distress

syndrome. Adv. Prostaglandin Thromboxane Leukot Res. 21B,

569±572.

48. Kosaka, K., Suzuki, K., Hayakawa, M., Sugiyama, S. & Ozawa, T.

(1994) Leukotoxin, a linoleate epoxide: its implication in the late

death of patients with extensive burns. Mol. Cell. Biochem. 139,

141±148.

49. Moghaddam, M.F., Grant, D.F., Cheek, J.M., Greene, J.F., Williamson,

K.C. & Hammock, B.D. (1997) Bioactivation of leukotoxins to their

toxic diols by epoxide hydrolase. Nat. Med. 3, 562±566.