Metabolically engineered oilseed crops with enhanced seed tocopherol

ARTICLE IN PRESS

1096-7176/$ - se

doi:10.1016/j.ym

�CorrespondE-mail addr

1These autho2Current add

Maryland Heig3Current add

bolism, 14 Cam4Current add

Ferry Road, Po5Current add

Fifth Street, Da

Metabolic Engineering 7 (2005) 384–400

www.elsevier.com/locate/ymben

Metabolically engineered oilseed crops with enhanced seed tocopherol

Balasulojini Karunanandaaa,1, Qungang Qia,1, Ming Haoa, Susan R. Baszisa,Pamela K. Jensena, Yun-Hua H. Wonga,2, Jian Jianga,3, Mylavarapu Venkatrameshb,4,

Kenneth J. Gruysa,5, Farhad Moshiria, Dusty Post-Beittenmillera,James D. Weissa, Henry E. Valentina,�,5

aMonsanto Company, 800 N. Lindbergh Boulevard, St. Louis, MO 63167, USAbRenessen LLC, 800 N Lindbergh Boulevard, St. Louis, MO 63167, USA

Received 30 January 2005; received in revised form 8 May 2005; accepted 18 May 2005

Available online 25 August 2005

Abstract

Tocochromanols (tocopherols and tocotrienols) are important lipid soluble antioxidants and are an essential part of the

mammalian diet. Oilseeds are particularly rich in tocochromanols with an average concentration 10-fold higher than other plant

tissues. Here we describe a systematic approach to identify rate-limiting reactions in the tocochromanol biosynthetic pathway, and

the application of this knowledge to engineer tocochromanol biosynthesis in oilseed crops. Seed-specific expression of genes

encoding limiting tocochromanol pathway enzymes in soybean increased total tocochromanols up to 15-fold from 320 ng/mg in WT

seed to 4800 ng/mg in seed from the best performing event. Although WT soybean seed contain only traces of tocotrienols, these

transgenic soybean accumulated up to 94% of their tocochromanols as tocotrienols. Upon crossing transgenic high tocochromanol

soybean with transgenic high a-tocopherol soybean, the vitamin E activity in the best performing F2-seed was calculated to be 11-

fold higher than the average WT soybean seed vitamin E activity.

r 2005 Elsevier Inc. All rights reserved.

Keywords: Vitamin E; Tocochromanol; Tocopherol; Tocotrienol; Oilseed crop; Nutrition

1. Introduction

Tocochromanols are lipophilic antioxidants that aresynthesized in the plastids of plants, and by some greenphotosynthetic bacteria. They are composed of a groupof structurally related compounds with eight majorforms that can be distinguished by the number and

e front matter r 2005 Elsevier Inc. All rights reserved.

ben.2005.05.005

ing author.

ess: [email protected] (H.E. Valentin).

rs contributed equally to the manuscript.

ress: Chemir Analytical Services, 2672 Metro Blvd.

hts, MO 63043, USA

ress: BiogenIDEC, Pharmacokinetics and Drug Meta-

bridge Center, Cambridge, MA 02142, USA

ress: Exelixis Plant Sciences, 16160 SW Upper Boones

rtland, OR 97224, USA

ress: Monsanto Company, Calgene Campus, 1920

vis, CA 95616, USA

position of methyl groups on the aromatic head group(a-, b-, g-, d-tocopherol, and a-, b-, g-, d-tocotrienol) andthe presence of a saturated (tocopherols) or unsaturated(tocotrienols) isoprenoid side chain. Their biopotencyfor animals and humans is expressed as vitamin Eactivity (Nomenclature rules for vitamin E (972.31),1990; IUPAC-IUP Joint Commission on BiochemicalNomenclature, 1982) with a-tocopherol having thehighest vitamin E activity (Traber and Sies, 1996;Sheppard et al., 1993; Bramley et al., 2000). The vitaminE activity of 1mg chemically synthesized all racemic a-tocopherol is defined as 1 IU. Natural isomeric pureRRR-a-tocopherol is 1.5-fold more active than syntheticvitamin E (Chow, 2001). The National Institute ofHealth (NIH) currently suggests a recommended dailyallowance (RDA) of 22–28 IU for humans (http://www.cc.nih.gov/ccc/supplements/vite.html#rda).

http://www.cc.nih.gov/ccc/supplements/vite.html#rda

http://www.cc.nih.gov/ccc/supplements/vite.html#rda

www.elsevier.com/locate/ymben

ARTICLE IN PRESS

Fig. 1. Schematic drawing of the tocochromanol biosynthetic pathway. Abbreviations: DMAPP, dimithylallyldiphosphate; GGH, geranylger-

anyldiphosphate hydratase; HPPD, p-hydroxyphenylpyruvate dioxygenase; IPP, isopentenyldiphosphate; MEP, 2-C-methyl-D-erythritol 4-

phosphate; TYRA, bifunctional chorismate mutase-prephenate dehydrogenase; VTE1, tocopherol cyclase; VTE2, homogentisate phytyltransferase;

VTE3, 2-methyl-6-phytylbenzoquinol methyltransferase; VTE4, g-tocopherol methyltransferase.

B. Karunanandaa et al. / Metabolic Engineering 7 (2005) 384–400 385

Higher vitamin E doses of 100–1000 IU have beenassociated with cancer reduction, improved immuneresponse, and cardiovascular benefits (Buring andHennekens, 1997; Tangney, 1997; Bramley et al.,2000). Such high doses are currently realized onlythrough supplemental tocopherol intake. Tocopherolsare also used in the pharmaceutical and cosmeticsindustry (Edwards, 2001), and as animal feed additivesto improve the quality and shelf life of meat (Sanderset al., 1997).

The combined annual consumption of tocopherols forhuman and animal applications was estimated at about40,000 t in 2002, with only 10% coming from naturalsources. Natural vitamin E originates almost exclusivelyfrom soybean oil, and is used predominantly for humanapplications due to its premium price and limitedavailability.

The first committed reaction in tocochromanolbiosynthesis is the prenylation of homogentisic acid(HGA) catalyzed by the homogentisate phytyltransfer-ase (VTE2) (Fig. 1) (Schledz et al., 2001; Collakova andDellaPenna, 2001; Savidge et al., 2002), or the homo-gentisate geranylgeranyl transferase (HGGT) (Cahoonet al., 2003). It is generally accepted that prenylation ofHGA with phytyldiphosphate (PDP) results in theformation of a�, b�, g�, and d–tocopherol, whileHGA-prenylation with geranylgeranyldiphosphate(GGDP) results in the formation of the correspondingtocotrienols (Fig. 1). While most of the biochemicalreactions leading to tocochromanol formation have

been known for more than two decades (Soll andSchultz, 1980; Soll et al., 1980, 1983, 1985), the genesinvolved in tocochromanol biosynthesis have beendiscovered only very recently (Rohmer, 2003; Shintaniand DellaPenna, 1998; Collakova and DellaPenna,2001; Savidge et al., 2002; Porfirova et al., 2002; Chenget al., 2003; Van Eenennaam et al., 2003), and analysisof tocochromanol pathway regulation and identificationof rate limiting reactions of the pathway has just beeninitiated (Savidge et al., 2002; Collakova and DellaPen-na, 2001, 2003a,b).

This manuscript provides experimental evidenceindicating limiting tocochromanol intermediates andrate-limiting biochemical reactions for tocochromanolbiosynthesis and it presents strategies and results onhow to apply this information to engineer commercialoilseed to produce oils with substantially enhancedtocochromanol and vitamin E content.

2. Materials and methods

2.1. Plant growth conditions, and transformation

procedures

Arabidopsis thaliana var. Columbia was grown under16 h photoperiod at 130–180 mE PPFD, 20 1C, and 70%relative humidity in Conviron MTPC432 chambers(Conviron, Winnipeg, Canada). The plants were grownin Metro Mix 200 soil (Hummert International, Earth

ARTICLE IN PRESSB. Karunanandaa et al. / Metabolic Engineering 7 (2005) 384–400386

City, MO) in 2.500 pots at a density of 32 plants/flat,three flats to a group of 96. The three flats of a groupwere grown side by side at all times. For Agrobacterium

mediated germline transformation, Arabidopsis plantswere grown in a 20 1C chamber under 130–180 mE PPFDwith a 16 h photoperiod. Transgenic plants were selectedby plating on agar plates with kanamycin selection.Transformed hemizygous plants were grown to maturityand the R2 segregating seed were harvested fortocochromanol analysis.

Agrobacterium mediated germline transformation ofsoybean (Glycine max) cultivar A3244 was carried out aspreviously described (Martinell et al., 2002) using freshlygerminated soybean meristems that were induced toform shoots directly. Transgenic plants were grown in agreenhouse with a 10 h photoperiod, a daytime tem-perature of 29 1C and a nighttime temperature of 24 1C.Plants were grown to maturity and seeds were harvestedfor analysis.

Canola, cultivar Ebony, was transformed as describedby Radke et al. (1992), using glyphosate resistance asselectable marker.

2.2. Bacterial strains, growth conditions and cell sample

preparation

Wild type and recombinant cells of Synechocystis sp.PCC 6803 (ATCC 27184) were cultivated photoauto-trophically in BG-11 medium (Sigma, St Louis, MO)buffered with 10mM TES-NaOH (pH 8.0) at 30 1Cunder a light intensity of 1500 LUX and shaken at225 rpm on a rotary shaker unless otherwise indicated.For growth on solid medium, 1.5% (wt/vol) agar(Difco) was supplemented to the medium. Recombinantplasmids were transformed into Synechocystis viaconjugative method as described by Poster (1988) andstabilized by supplementation of the growth mediumwith 25 mg/ml of kanamycin, 10 mg/ml of gentamycin, or30 mg/ml of spectinomycin, depending on the plasmidused in the experiments. The transformed cells werespread on a 0.45 mM cellulose membrane filter (What-mans) and placed on non-selective solid medium,incubated for 24 h as described above, and transferredto selective medium plates containing the appropriateantibiotics. Positive colonies were inoculated and grownin 2ml of BG-11 liquid medium supplemented with theappropriate antibiotic for 2 days and then transferred to50ml liquid culture. These cell cultures served as pre-cultures for the final 300ml liquid cultures. Cell densitywas monitored spectrophotometrically (Spectra MAX,Molecular Devices, CA) at a wavelength of 730 nm.When the A730 value of the cell culture reached 0.6–0.8,cells were sub cultured in 300ml of fresh BG-11 mediumsupplemented with the appropriate antibiotics, and A730

was adjusted to 0.2. Cell samples were harvested atdifferent time points for measurements of tocopherols,

tocotrienols, intermediates, gene expression and enzymeactivities. Gene expression of the bifunctional Erwinia

herbicola prephenate dehydrogenase (Eh-TYRA), ma-ture Arabidopsis thaliana homogentisate dioxygenase(At-HPPD), and mature Arabidopsis thaliana GGDPdehydrogenase (At-GGH) was confirmed by immunoblot analysis (Harlow and Lane, 1988).

2.3. Synechocystis expression vectors and gene

expression

Gene expression in Synechocystis was driven by theLac promoter (Dickson et al., 1975), which expressedconstitutively in this organism. The coding regions oftocochromanol related genes from plant sources wereexpressed as mature genes. For the expression ofmultiple genes, trans-genes were arranged as operonstructures downstream of the Lac promoter. Geneexpression for Eh-TYRA, At-HPPD, and At-GGH wasconfirmed by immuno blot analysis using peptidedirected antibodies specific for each gene. The growthcharacteristics of all transgenic Synechocystis culturesused for these experiments remained unchanged com-pared to Synechocystis WT-cultures.

The base vector for expression of pathway genes waspSL1211 (Ng et al., 2000). Trans-genes tested in thepresent study were Eh-TYRA, At-HPPD, a mature At-

GGH, and Synechocystis GGH (Syn-GGH), a matureArabidopsis thaliana VTE2 (At-VTE2), and Synechocys-

tis sp. PCC6803 VTE2 (Syn-VTE2). The pSL1211-derived expression vectors harboring single or multiplegene(s) were constructed as operon structures usingstandard molecular cloning techniques and designatedas pMON36531, pMON36550, pMON36538,pMON36567, and pMON36549 (Table 1). The structur-al genes of Eh-TYRA, mature At-GGH, Syn-VTE2, andmature At-VTE2 were designed to be preceded by thenucleotide sequence AGGAAACAGCC ATG, withAGGA serving as Shine-Dalgarno sequence. At-HPPD

was preceded by AGGAGGACAGCC ATG, withAGGAGG as Shine-Dalgarno sequence. The trans-geneorganization of tocopherol pathway genes in theseexpression vectors was as depicted in Table 1. Restric-tion digestions and DNA sequencing were performed toverify correct orientations and sequences of geneticelements and gene(s) in these plasmids.

2.4. Tocochromanol analysis

Tocochromanol analysis was performed as describedby Van Eenennaam et al. (2003). The tocochromanolcomposition is provided on a wt/wt base. Statistical dataanalysis was performed using JMP statistical discoverysoftware (www.jmp.com).

http://www.jmp.com

ARTICLE IN PRESS

Table 1

Synechocystis expression vectors used in this study

Plasmid Relevant genetic elements Reference

pSL1211 Kanr Ng et al. (2000)

pMON36550 Specr, PLac-At-HPPD This study

pMON36531 Specr, PLac-Eh-TYRA This study

pMON36538 Specr, PLac-Eh-TYRA-At-HPPD This study

pMON36567 Specr, PLac-Eh-TYRA-At-HPPD-At-GGHm This study

pMON36549 Kanr, PLac-Eh-TYRA-At-HPPD-Syn-GGH-Syn-VTE2 This study

Synechocystis multi-gene vectors were assembled as operon structures.

Abbreviations: At-GGHm, mature A. thaliana geranylgeranyldiphosphate hydratase (accession # NM_106107); At-HPPD, A. thaliana p-

hydroxyphenylpyruvate dioxygenase (accession # AF060481); Kanr, kanamycin resistance gene; PLac, E. coli Lac promoter (Dickson et al., 1975);

Specr, spectinomycin resistance gene; Syn-GGH, Synechocystis sp. PCC 6803 GGH (accession # NP_441659); Syn-VTE2, Synechocystis sp. PCC 6803

homogentisate phytyltransferase (accession # NP_441094).


2.5. Enzyme assays

To confirm gene-functionality of At-VTE2, Syn-

VTE2, At-HPPD, Eh-TYRA, At-GGH, and Syn-GGH

prior to transformation of these genes into a plantsystem, enzyme assays were performed on E. coli crudeextracts harboring expression constructs for each ofthese enzymes, and assays were performed according tothe procedures described in Savidge et al. (2002), Secor(1994), Cotton and Gibson (1970), and Keller et al.(1998); respectively. Using this approach functionalitywas confirmed for of At-VTE2, Syn-VTE2, At-HPPD,and Eh-TYRA. Enzyme activity of E. coli expressed At-

GGH and Syn-GGH was measured using radiolabeledGGDP (Amersham Biosciences UK Ltd, Buckingham-shire, UK), or geranylgeranyl bacteriochlorophyll iso-lated from Rhodospirillum rubrum (Katz et al., 1972).However, neither Syn-GGH, nor At-GGH-functionalitycould be confirmed using the assay procedures asdescribed by Keller et al. (1998).

2.6. PCR screen of F1 soybean crosses for presence of

trans-genes

F1 plants obtained by crossing homozygous soybeanlines transformed with pMON67227 (At-VTE4 and At-

VTE3) and pMON69943 (Eh-TYRA, At-HPPD, andSyn-VTE2) were screened by genomic PCR analysis.Leaf tissue (30–50mg) harvested from the F1 plants wasground in liquid nitrogen and genomic DNA wasisolated using the Qiagens genomic DNA isolationKit (Qiagen Inc., Valencia, CA). Eh-TYRA and At-

VTE4 served as representative genes for the trans-genesets in pMON69943, and pMON67227, respectively.The presence of these two genes was confirmed by PCRamplification using approximately 30 ng genomic DNAand the Invitrogen PCR SuperMix system (InvitrogenCorporation, Carlsbad, CA). Primer pairs Eh-TYRA-primers (forward: 50- GATCTAGAACAATGGCTTC-CTCTAT-30 and reverse: 50- CTTATTATGGGCG-

GCTGTCATTGG-30) and At-VTE4-primers (forward:50- AGAACCGGGTACCGAGCTCGAGAT -30 andreverse: 50- TATTAGAGTGGCTTCTGGCAA -30)were used for the screen. The PCR program consistedof 5min incubation at 94 1C, followed by 25 cycles of94 1C for 15 s, 55 1C for 15 s, and 72 1C for 2.15min. Thefinal PCR-product was incubated for 10min at 72 1C.The PCR products were separated on 0.8% agarose gelsto identify double positive lines.

2.7. Northern blot analysis

Gene expression was usually confirmed on selectedevents by western blot analysis. However, antibodies forSyn-VTE2 and At-VTE2 were not sufficiently specific toconfirm expression of these genes in plant tissues.Expression of these genes was therefore confirmed onselected Arabidopsis and canola events by NorthernBlot analysis.

RNA isolation was performed on 100mg fullymatured but not dried, Arabidopsis silique tissue and50mg immature canola seed collected 40 days afterpollination. Samples were ground with 0.25 g polyvi-nylpolypyrrolidone in liquid nitrogen. While theground tissues were frozen, 10ml REC.8+ buffercontaining 50mM Tris-HCl (pH 9.0), 800mM NaCl,10mM EDTA, 0.5% CTAB and 0.5% b-mercaptoetha-nol was added to each sample, mixed and centrifugedfor 5min at 10,000 rpm in a Sorvall centrifuge, modelSuper T21. The supernatant was filtered through mira-cloth (Calbiochem, http://www.emdbiosciences.com/html/CBC/home.html). After extraction with 3mlchloroform, the supernatant was separated and ex-tracted twice with equal volume of phenol:chloroformmix (1:1) and then ethanol precipitated. The ethanolpellet was re-suspended in 50mM EDTA. Equalamounts of total RNA (20 mg) were fractionatedon 1.2% agarose gels containing 1M formaldehyde.Gels were blotted onto nylon membrane (Ambion,Inc., Austin, TX) according to the manufacturer’s

http://www.emdbiosciences.com/html/CBC/home.html

http://www.emdbiosciences.com/html/CBC/home.html


instructions and hybridized to a random primer-labeledDNA probe in Sigma PerfectHybTM buffer (Sigma,St. Louis, MO, catalog # H-7033), containing 10mg/mlof salmon sperm DNA at 65 1C. A cDNA fragment(PvuI/PvuI) of At-VTE2 was labeled and used as agene-specific probe. Probe labeling with 32P-dCTP(50 mCi/reaction) was done using Amersham Redipri-meTM labeling kit (Amersham Biosciences Corp., Piscat-away, NJ, catalog # RPN1633). Blots were washed in2X SSC, 0.1% SDS at 65 1C and exposed to X-ray film(BioMax MS film, Sigma, St. Louis, MO, catalog# Z36,307-3).

2.8. Immuno blot analysis

To extract the total protein from seed, single soybeanseed or 20–22mg of Arabidopsis or a comparableamount of canola seed were mixed with 0.5mmzirconia/silica micro beads (Biospec Products, Bartles-ville, OK) and 500 ml protein extraction buffer [50mMTris (pH 7.4), 1mM EDTA, 0.1% Triton X-100,Complete protease inhibitor (Roche Molecular Bio-chemicals, Indianapolis, IN; http://www.biochem.ro-che.com) (1 tablet/50ml buffer)] and shaken in a beadbeater for 45 s twice and subsequently cooled on ice. Theextracted sample was spun in an Eppendorf table topcentrifuge at 14,000 rpm for 10min. The supernatantwas removed. Protein measurements were performedaccording to Bradford (1976), using bovine serumalbumin as standard, and spectrophotometric measure-ments were done in a computer-directed microtiter platereader (Spectra Max Plus 384, Molecular Devices Corp.,Sunnyvale, CA). Sample buffer containing 100mM Tris-HCl, pH 6.8, 4% SDS, 0.1% bromophenol blue, 20%glycerol, and 10% b-mercaptoethanol, was mixed with80mg total protein from each sample, heated for 2minat 100 1C, and centrifuged at 12,000 rpm for 5min at4 1C (Sorvall centrifuge, model Super T21). Totalproteins from each sample were separated in 10% SDSPAGE gel and blotted onto PVDF membrane (Bio-RadLaboratories, Hercules, CA) at 300mA for 2 h using aBio-Rad Transblot SD wet electroblotting apparatus(Bio-Rad Laboratories, Hercules, CA). Eh-TYRA, At-HPPD, and At-GGH specific antibodies were obtainedfrom Sigma-Genosys (http://www.sigma-genosys.co.uk/)by injecting rabbits with synthetic peptidesCSMLASRRKEAEALG (Eh-TYRA), CMMKDEEG-KAYQSGG (At-HPPD), CLPPEIIDRRVRKMK(Syn-GGH), and CDAYLRERAEKSGAT (At-GGH).Blots were treated with rabbit Eh-TYRA-, At-HPPD-,or GGH-specific antiserum (1:1000), and immunecomplexes were detected using alkaline phosphatase-conjugated goat anti-rabbit secondary antibodies (Sig-ma-Alrich, www.sigma-aldrich.com) and nitroblue tet-razolium plus 5-bromo-4-chloro-3-indolyl phosphate(Harlow and Lane, 1988).

2.9. Extraction of intermediates for liquid

chromatography/mass spectrometry (LC/MS) analysis

Extraction from Arabidopsis thaliana seeds was accom-plished by weighing 1274mg of seeds into a 2mlmicrotube. Approximately 1 g of 0.5mm zirconia/silicamicrobeads (Biospec Products, Bartlesville, OK) and500ml of the appropriate extraction solvent was added tothe tube and shaken in a Fast Prep (FP120; SavantInstruments Inc., Holbrook, NY) for two sessions of 45 seach. The extract was allowed to settle and the super-natant was filtered, using the appropriate membrane, intoan LC vial for analysis. For tyrosine, HPPA and HGAanalysis, the extraction solvent was MeOH:water (50:50)with 0.1% formic acid, and the filter material used wasNylon. In the case of 2M6PBQ and 2M6GGBQ, theextraction solvent was pure ethanol and the filter wasmade of polytetrafluoroethylene (PTFE).

For soybean and canola, the seeds were ground priorto extraction in order to obtain homogenous flour usinga machine built in-house. For grinding bulk canola, 1 gseed were placed in a 30ml polypropylene tube with a34in steel ball. The tubes were shaken for 30 s at an

amplitude of 3.81 cm at 1000 strokes per minute. Forgrinding bulk soybean, 5–10 seed were treated similarlywith the shaking frequency set at 1200 strokes perminute. For grinding single soybean, one seed was placedin a 14ml polypropylene tube with a 3

8 in steel ball. Thetubes were shaken for 90 s at an amplitude of 3.81 cm at1000 strokes per minute. After grinding, approximately30710mg soybean or canola meal was extracted andfiltered as described above for Arabidopsis.

2.10. Determination of tocopherol intermediates using

liquid chromatography/mass spectrometry

The liquid chromatography/mass spectrometry (LC/MS) system consisted of an HP1100 series HPLC(Agilent Technologies, Palo Alto, CA) connected to anAPI QSTAR Pulsar-i (Applied Biosystems, Foster City,CA) mass spectrometer using their TurboIonspray (i.e.electrospray) interface. The system was operated usingthe Analyst QS software (Applied Biosystems, FosterCity, CA). All methods described here utilized anAlltima C18 column (4.6� 50mm, 3 mm; Alltech As-sociates, Inc., Deerfield, IL) operated at a flow rate of1ml/min connected to the TurboIonspray source with asplit of approximately 1:5 (MS:waste). The massspectrometer was operated in negative ion mode for allmethods. The injection volumes for each method wereset to 20 ml for all standards and samples. Analytes werequantified against external standard curves generatedwith the sample sets. Standards for HGA, p-hydro-xyphenylpyruvate (HPPA), and tyrosine were purchasedfrom Sigma Chemicals (Sigma St. Louis, MO). Stan-dards for 2M6PBQ and 2M6GGBQ were synthesized as

http://www.biochem.roche.com

http://www.biochem.roche.com

http://www.sigma-genosys.co.uk/

http://www.sigma-aldrich.com

ARTICLE IN PRESSB. Karunanandaa et al. / Metabolic Engineering 7 (2005) 384–400 389

described previously (Soll and Schultz, 1980; Henry etal., 1987). The LC and MS parameters for each methodwere as follows.

Tyrosine and HPPA were analyzed by LC/MS usinggradient conditions. The HPLC mobile phases werewater (A), acetonitrile (B) and 10% formic acid (C). Thegradient method began with 0.1% C (throughout run)and 4% B isocratic for 0.2min, increasing to 95% B at0.6min, which was held until 0.8min, then returned toinitial conditions (4% B) at 1.0min and equilibrateduntil 3.0min. The mass spectrometer used ‘‘TOF MS’’mode to scan from 160 to 190 amu. The quantitationwas based on extracted ion chromatograms centeredaround the appropriate quasi-molecular ion of eachanalyte (tyrosine �180.15; HPPA �179.10). The reten-tion times were 1.22 and 1.14min for tyrosine andHPPA, respectively. This method could also monitor forHGA, but interferences from the seed matrix would notallow for reliable quantitation.

To obtain specificity for HGA, an MS/MS methodwas used. The mobile phases for this LC/MS/MSmethod were water (A), methanol (B) and 0.5% formicacid (C). The 5min method used 1% C throughout thegradient and started at 5% B. The gradient increased to98% B over 1.80min, was held at 98% B until 2.0min,and then returned to 5% B at 2.2min, where itconditioned the column for the remaining time. HGAeluted around 2.2min. The mass spectrometer used‘‘Product Ion’’ mode, where 167.1 amu was selected asthe parent ion. The mass spectrometer was set to useLOW resolution and a collision energy (CE) of �15 kV.The TOF mass range scanned from 123.05 to123.15 amu to detect the daughter ion mass. Thesesettings allowed for the specificity needed to quantitateHGA in matrix. The HGA detection limit under theseconditions was 0.25 ppm.

For LC/MS analysis of the 2M6PBQ and 2M6GGBQ,the HPLC separation was performed isocratically at96% B for 4.5min, where mobile phase (A) was waterand (B) was 0.1% acetic acid in methanol. The retentiontimes for 2M6GGBQ and 2M6PBQ were approximately1.5 and 1.9min, respectively. The mass spectrometer wasset to ‘‘TOFMS’’ mode and used two scan functions (i.e.‘‘Experiments’’): Experiment 1 scanned from 401.25 to401.45 amu to monitor for 2M6PBQ, and Experiment 2scanned from 395.20 to 395.45 amu to monitor for2M6GGBQ. Detection limits for 2M6PBQ and2M6GGBQ were 0.2 and 2 ppm, respectively.

3. Results

3.1. Feeding experiments/identification of rate limiting

tocopherol intermediates

To identify tocochromanol pathway intermediatesand biochemical reactions that limit tocochromanol

biosynthesis, cotyledon-derived soybean suspensioncultures were subjected to feeding experiments usingthe tocopherol precursors chorismic acid, tyrosine,HPPA, HGA, phytol, geranylgeraniol, and 1-deoxy-D-xylulose. All metabolites were supplied in concentra-tions ranging from 0.25 to 5mM. The strongest effectswere obtained with phytol and HGA when provided atthe 2mM level resulting each in a two-fold increase intotal tocopherol levels. Geranylgeraniol was not aseffective as phytol, and increased total tocopherols byup to 1.5-fold only. When phytol and HGA wereprovided in combination, each at 2mM, total tocochro-manol levels increased by five-fold, suggesting that theavailability of these two intermediates is critical fortocochromanol biosynthesis.

Data obtained in feeding experiments are often hardto interpret, as the rates for metabolite uptake are oftennot available. In addition some intermediates such astyrosine and chorismate exhibited inhibitory effects atthe concentration range tested above, while otherintermediates such as HPPA are unstable in aqueoussolutions. Therefore, subsequent experiments to confirmthe importance of HGA and PDP availability fortocochromanol biosynthesis were performed in trans-genic models.

3.2. Synechocystis sp. PCC 6803 as transgenic model for

tocochromanol pathway engineering

Potential key genes of the tocochromanol biosyntheticpathway, identified based on previous feeding experi-ments and literature data, were cloned from bacterialand plant sources, and expressed as single genes and incombination in the Synechocystis model. Eh-TYRA, andAt-HPPD were chosen as key genes to increase HGA,and At-GGH was chosen to increase PDP availability.Expression of these genes was combined with expressionof Synechocystis VTE2 (Syn-VTE2), which had beenidentified as rate limiting for tocochromanol biosynth-esis previously (Collakova and DellaPenna 2001,2003a,b; Savidge et al., 2002).

When expressing single genes in Synechocystis, At-

HPPD had the strongest impact on tocochromanollevels, resulting in up to seven-fold WT-tocochromanollevels (Fig. 2A). Combined expression of the key genesidentified above increased tocochromanol levels with theaddition of each new gene, and tocochromanol levels intransgenic cultures increased over the entire cultivationperiod of 24 days (Fig. 2A and B). Expression of all fourgenes (Eh-TYRA, At-HPPD, At-VTE2, and At-GGH)resulted in up to 16.5-fold WT tocopherol levels. Thesecultures contained up to 2200 ng tocopherol per mgdry cell mass (Fig. 2B). A small fraction (5–10%)of these tocopherols was accumulated in the form ofg-tocopherol, while the remaining fraction consisted

ARTICLE IN PRESS

Fig. 2. Tocopherol and tocotrienol formation in transgenic Synecho-

cystis cultures: (A) tocochromanol timecourse over a period of 24 days;

(B) tocopherol and tocotrienol content of lyophilized Synechocystis

cultures after a cultivation period of 24 days; (C) Synechocystis culture

supernatants after a cultivation period of 24 days.

B. Karunanandaa et al. / Metabolic Engineering 7 (2005) 384–400390

of a-tocopherol. g-Tocopherol is not detected in WTSynechocystis cells.

Transgenic expression of At-HPPD, or At-HPPD andEh-TYRA, coincided with the occurrence of a brownculture supernatant, suggesting the presence of melaninlike pigment, a reaction product of HGA (Fig. 2C)Coon et al., 1994). Tocopherols and tocotrienols do notexhibit substantial absorption properties in the visiblelight spectrum and are not water-soluble. They aretherefore likely not directly associated with the observedcolor changes. Interestingly, these cultures contained10–20% of their tocochromanols in the form oftocotrienols, while tocotrienols in WT-cultures werebelow 2% of total tocochromanols (Fig. 2B). When At-

GGH was co-expressed with At-HPPD and Eh-TYRA,the culture supernatant was not brown, tocotrienolswere not detected, and total tocochromanol accumula-tion increased compared to the double gene construct(Fig. 2B and C).

3.3. Engineering oil seed tocochromanol content

Genes identified to be critical for tocochromanolpathway engineering in Synechocystis were expressed assingle genes and in combination in Arabidopsis, canolaand soybean. In Arabidopsis, each gene was placed intwo constructs using either the constitutive enhancedcauliflower mosaic virus promoter (Pe35S, McPherson

and Kay, 1994) or the seed-specific napin promoter(PNapin, Kridl et al., 1991). Single gene expression of At-

GGH did not result in significant changes of seedtocochromanol levels (Table 2). In contrast, seed-specific expression of At-HPPD, Eh-TYRA, At-VTE2,or Syn-VTE2 as single genes resulted in significant seedtocochromanol increases of 1.09-fold, 1.20-fold, 1.41-fold, and 1.36-fold in R2 seed populations, respectively.In our hands, expression of the same genes under thecontrol of a constitutive promoter did not significantlyincrease seed tocochromanol levels in the population ofall events (Table 2). For this reason, all subsequentexpression cassettes for the enhancement of seedtocochromanol production were designed using seed-specific promoters (Tables 3, 4, and 5).

When single genes were expressed in crops, only seed-specific expression of Eh-TYRA in canola resulted in asignificant increase of the average seed tocochromanolcontent of all events, with up to two-fold WT-tocochromanol levels for the best performing event(Table 4). However, when combinations of two or moretocochromanol genes were expressed, seed tocochroma-nol levels increased significantly in Arabidopsis, andcanola for all gene combinations (Tables 3 and 4). Thebest performing events harboring expression constructsfor Eh-TYRA, At-HPPD, and At-VTE2, or Syn-VTE2

in Arabidopsis and canola contained 5- and 3.7-foldWT-tocochromanol levels, respectively (Tables 3 and 4).In soybean seed populations tocochromanol increasedsignificantly when Eh-TYRA, At-HPPD, and Syn-VTE2

expression was combined (Table 4). The best performingconstruct harbored At-GGH in addition to Eh-TYRA,AT-HPPD, and At-VTE2. Four events obtained withthis construct (pMON77637) had on average more than10-fold WT tocochromanol levels, with up to 15-foldWT-levels for seed from the best performing event(Table 5). Interestingly, the increase in total tocochro-manols was obtained exclusively through tocotrienolaccumulation, with an average of 94% tocotrienols inseed from this event. In these seed the pool of a, b, g, andd-tocopherol was slightly reduced compared to WT-levels (Table 5). Overall d-tocotrienol was the dominat-ing tocochromanol species in these seed. Among thetocopherol pool, d-tocopherol was the dominatingtocochromanol species (Table 5). Both, Arabidopsisand soybean seed harboring At-HPPD and Eh-TYRA

expression constructs combined exhibited a dark brownseed phenotypre (Fig. 3).

Single soybean seed from transformation events withpMON66682 had total tocochromanol levels up to 4.30-fold higher than WT-seed (Table 4). Up to 82% of totaltocochromanols in these seed were accumulated astocotrienols. The transgenic expression of At-GGH wasexpected to reduce tocotrienol formation in these seed.Western analysis on pMON66682 seed extracts con-firmed transgenic expression of the At-GGH for four out

ARTICLE IN PRESS

Table 2

Impact of seed-specific and constitutive expression of single genes on tocochromanol levels in Arabidopsis R2 seed populations.

Vector designation Relevant genetic elements Reference Mean tocochromanol

level [ng/mg seed]

P-Value Average fold

tocochromanol

increase

pMON36525 PNapin-CTP2-At-HPPD-napin 30 Valentin and Mitsky (2002) 599748.9 0.0448 1.0970.10

control 552735.9

pMON26586 Pe35S-CTP2-At-HPPD-E9 30 This study 556735.4 0.3959 1.0370.07

control 541724.7

pMON36520 PNapin-CTP1-Eh-TYRA-napin 30 Valentin and Mitsky (2002) 524778.5 0.0084 1.2070.19

control 437728.6

pMON36511 Pe35S-CTP1-Eh-TYRA-E9 30 Valentin and Mitsky (2002) 406792.7 0.4112 0.9370.21

control 437728.6

pMON43852 PNapin-At-GGH-napin 30 This study 535748.6 0.4168 1.0370.09

control 519715.4

pMON26583 Pe35S- At-GGH-E9 30 This study 514742.3 0.6330 0.9870.08

control 523733.3

pCGN10822 PNapin-At-VTE2-napin 30 Savidge et al. (2002) 809767.0 0.0001 1.4170.12

control 574738.2

pCGN10800 Pe35S- At-VTE2-TML 30 Lassner et al. (2003) 493753.2 0.3469 0.9570.10

control 52078.36

pMON16602 PNapin-Syn-VTE2-napin 30 Valentin and Mitsky (2002) 679779.0 0.0002 1.3670.16

control 500739.9

pMON21698 Pe35S-Syn-VTE2-E9 30 This study 552718.4 0.5522 1.0270.03

control 544735.4

The original Arabidopsis seed tocochromanol data set that is summarized in Tables 2 and 3 is provided in Supplemental Table 1. To establish the

significance of changes in seed tocochromanol levels upon expression of various trans-genes in Arabidopsis, seed tocochromanol data from

transgenic populations were compared with seed tocochromanol data from populations of WT- seed or seed transformed with a vector control which

had been grown together with the transgenic seed. Means, standard error, and significance were calculated using Tukey–Kramer HSD. P-values

o0.05 are significant.

Abbreviations: At-GGH, A. thaliana geranylgeranyldiphosphate hydratase (accession # NM_106107); At-HPPD, A. thaliana p-hydroxyphenylpyr-

uvate dioxygenase (accession # AF060481); At-VTE2, A. thaliana homogentisate phytyltransferase (accession # AAL35412); CTP1, modified

chloroplast target peptide from the small subunit of the Arabidopsis ribulose bisphosphate carboxylase (Barry and Kishore, 1998); CTP2, chloroplast

target peptide from the Arabidopsis 5-enolpyruvylshikimate-3-phosphate synthase (Klee et al., 1987); E9 30, 408 nt of Pea rubisco 30 sequence

(Coruzzi et al., 1984); Eh-TYRA, Erwinia herbicola bifunctional chorismate mutase, prephenate dehydrogenase (accession # X60420); napin 30, napin

30 untranslated region (Kridl et al., 1991, accession # M64632); Pe35S, enhanced cauliflower mosaic virus promoter (Mc Pherson and Kay, 1994);

PNapin, Napin seed-specific promoter (Kridl et al., 1991, accession # M64632); Syn-VTE2, Synechocystis sp. PCC 6803 homogentisate

phytyltransferase (accession # NP_441094); TML 30, Agrobacterium tumefaciens (accession # AF242881, nucleotides 10303 to 9203).


of seven tested events. The GGH western signal obtainedfrom transgenic soybean seed comigrated with an E. coli

expressed mature At-GGH protein (Fig. 4), suggestingthat the soybean-expressed protein was properly pro-cessed, and therefore, presumably properly targeted tothe plastid. Nevertheless, these events did not show atocotrienol reduction in favor of increased tocopherolswhen compared to transformation events from the samevector, not expressing GGH (Fig. 4). The phenotypicalchanges in seed color, shape, and seed germinationobserved previously on seed harboring the combined At-

HPPD and Eh-TYRA expression constructs were main-tained in seed from pMON77637 and pMON66682.

3.4. Accumulation of tocopherol precursors and

intermediates

LC-MS analysis of Arabidopsis seed for tocopherolprecursors and intermediates revealed that HGA incontrol seed was below the limit of detection

(o0.25 ppm), and HGA levels in seed expressing At-

HPPD under napin promoter control were approxi-mately 0.25 ppm. Expression of Eh-TYRA correspondedwith a two-fold increase in seed HPPA (Po0.02), andthree-fold increase in free tyrosine levels (Po0.0001).

Interestingly all seed harboring Eh-TYRA and At-

HPPD expression constructs combined, accumulatedtocotrienols, and all Arabidopsis and soybean seedharboring these two genes exhibited a brown to darkbrown seed phenotype, were unevenly shaped, andgerminated at lower rates than WT-seed. LC-MSanalysis of WT and brown seed revealed HGA increasesof 60 and 800-fold over the detection limit for matureArabidopsis and soybean seed, respectively (Fig. 3).

Transgenic high tocochromanol seed from Arabidop-sis, canola, and soybean exhibited a strong shift intocochromanol composition towards d-tocopherol andd-tocotrienol, suggesting a rate limitation by VTE3(Table 5, and above). In order to test this hypothesis, 2-methyl-6-phytylbenzoquinol (2M6PBQ) and 2-methyl-

ARTICLE IN PRESS

Table

3

Impact

oftrans-geneexpressiononArabidopsisseed

tocochromanollevels

Vector

designation

Relevantgenetic

elem

ents

Reference

Max.tocochromanol

increase

a[fold

wild-

typelevel]

Meantocochromanol

levelb[fold

wild-type

level]

Significance

c

pMON36525

PNapin-C

TP2-A

t-H

PP

D-napin

30

ValentinandMitsky

(2002)

1.26

1.0970.10

A

pMON36520

PNapin-C

TP1-E

h-T

YR

A-napin

30

ValentinandMitsky

(2002)

1.53

1.2070.19

AB

pCGN10822

PNapin-A

t-V

TE

2-napin

30

Savidgeet

al.(2002)

1.62

1.4170.11

BC

pMON36528

PNapin-C

TP2-A

t-H

PP

D-napin

30 -

PNapin-A

t-V

TE

2-napin

30

ValentinandMitsky

(2002)

1.75

1.4670.15

C

pMON36596

PNapin-C

TP2-A

t-H

PP

D-napin

30 -

PNapin-C

TP1-E

h-T

YR

A-napin

30

ValentinandMitsky

(2002)

2.24

1.7870.22

D

pMON69907

PNapin-C

TP1-E

h-T

YR

A-napin

30 -

PNapin-A

t-V

TE

2-napin

30

ValentinandMitsky

(2002)

2.41

1.7970.38

D

pMON69909

PNapin-C

TP2-A

t-H

PP

D-napin

30 -

PNapin-C

TP1-E

h-T

YR

A-napin

30 -

PNapin-A

t-V

TE

2-napin

30

ValentinandMitsky

(2002)

5.00

2.8870.66

E

Abbreviationsusedhereare

asdefined

under

Table

1.Abbreviations:

At-

VT

E2,

A.

tha

lia

nahomogentisate

phytyltransferase

(accession#AAL35412);CTP1,modified

chloroplast

target

peptide

from

thesm

allsubunitoftheArabidopsisribulose

bisphosphate

carboxylase

(BarryandKishore,1998);CTP2,chloroplasttarget

peptidefrom

theArabidopsis5-enolpyruvylshikim

ate-3-phosphate

synthase

(Kleeet

al.,1987);

Eh

-TY

RA,

Erw

inia

her

bic

ola

bifunctionalchorism

ate

mutase,prephenate

dehydrogenase

(accession#X60420);napin

30 ,napin

30untranslatedregion(K

ridlet

al.,1991,

accession#M64632);

PNapin,Napin

seed-specificpromoter(K

ridlet

al.,1991,accession#M64632).

aMaxim

um

seed

tocochromanolincrease

ofbestperform

ingeventanalyzedon10–20mgbulk

seed

oftheR2-generation

bAverageseed

tocochromanolincrease

ofalleventsoftheR2seed

population(approxim

ately

20–30events).Means,standard

error,andsignificance

werecalculatedusingTukey–Kramer

HSD.

cMeansfollowed

bythesamelettersare

notsignificantlydifferentfrom

each

other

(a¼

0.05).


ARTICLE IN PRESS

Table

4

R1seed

tocochromanollevelsin

transgenic

canola

andsoybeanseed

populations

Vectordesignation

Relevantgenotype

Reference

Tocochromanollevel

(ng/m

g)

Bestperform

ingeventa

Meanofallevents

Significance

b

Ca

no

la

WT-control

273

227722.4

A

pMON58175

PNapin-C

TP1-E

h-T

YR

A-napin

30

Thisstudy

540

370756.6

B

pMON58173

PNapin-C

TP1-S

yn

-VT

E2-napin

30

Thisstudy

299

247719.5

A

pMON58187

PNapin-C

TP2-A

t-H

PP

D-napin

30

Thisstudy

251

224723.4

A

pMON58178

PNapin-C

TP2-A

t-H

PP

D-napin

30 -

PNapin-

CTP1-E

h-T

YR

A-napin

30

ValentinandMitsky(2002)

763

5557177

D

pMON58186

PNapin-C

TP2-A

t-HPPD-napin

30 -

PNapin-

CTP1-E

h-T

YR

A-napin

30 -

PNapin-C

TP1-

Sy

n-V

TE

2-napin

30


829

4677197

C

So

yb

ean

WT-control

355

320733.4

A

pMON36575

P7Sa’-C

TP1-E

h-T

YR

A-E930


435

355763.2

A

pMON36576

P7Sa’-C

TP2-A

t-H

PP

D-E930


482

336756.7

A

pMON69933

PArc-5-C

TP1-S

yn

-VT

E2-A

rc30


357

321721.7

A

pMON36581

PArc-5-A

t-V

TE

2-A

rc30


456

374748.3

A

pMON69924

P7Sa’-C

TP2-A

t-H

PP

D-E930 ,

P7Sa’-C

TP1-

Eh

-TY

RA-E930


1090

8257336

AB

pMON69943

P7Sa’-C

TP2-A

t-H

PP

D-E930 ,

P7Sa’-C

TP1-

Eh

-TY

RA-E930 ,

PArc-5-C

TP1-S

yn

-VT

E2

-

Arc

30


1318

10547228

B

pMON77637

P7sa’-CTP2-A

t-H

PP

D-E930 ,

P7sa’-CTP1-

Eh

-TY

RA-E930 ,

PArc-5-

At-

VT

E2-A

rc30 ,

P7Sa–

At-

GG

H-N

os30

Thisstudy

3765

96171154

B

pMON66682

POleosin-A

t-G

GH-O

leosin30 ,

PFAE-A

t-

VT

E2-A

rc30 ,

P7Sa’-C

TP2-A

t-H

PP

D-E930 ,

P7Sa’-C

TP1-E

h-T

YR

A-E930

Thisstudy

1398

5747399

AB

Fortocochromanolanalysis,10–20mgofcanola

seed

or5-pooledsoybeanseed

per

eventwereused.Allexpressioncassettesin

multigenevectors

wereorganized

intheheadto

tailorientationin

the

order

provided

above.

Abbreviations:

Arc

30 ,

Pha

seo

lus

vulg

ari

sarcelin-5

30untranslated

region

(Wanget

al.,2002,accession

#Z50202);

At-

GG

H,

A.

tha

lia

nageranylgeranyldiphosphate

hydratase

(accession

#

NM_106107);

At-

GG

Hm,mature

A.

tha

lia

nageranylgeranyldiphosphate

hydratase;

At-

HP

PD,

A.

tha

lia

na

p-hydroxyphenylpyruvate

dioxygenase

(accession#AF060481);

At-

VT

E2,

A.

tha

lia

na

homogentisate

phytyltransferase

(accession#AAL35412);CTP1,modified

chloroplast

target

peptidefrom

thesm

allsubunitoftheArabidopsisribulose

bisphosphate

carboxylase

(Barryand

Kishore,1998);CTP2,chloroplasttarget

peptidefrom

theArabidopsis5-enolpyruvylshikim

ate-3-phosphatesynthase

(Kleeet

al.,1987);E930 ,408ntofPea

rubisco

30sequence

(Coruzziet

al.,1984);

Eh-T

YR

A,

Erw

inia

her

bic

ola

bifunctionalchorism

ate

mutase,prephenate

dehydrogenase

(accession#X60420);Kanr ,kanamycinresistance

gene;

NOS,nopalinesynthase

30UTR

(accession#

AF465641)Oleosin30 ,Oleosin30untranslatedregion(accession#AL161562,nucleotides

175567–176040);napin

30 ,napin

30untranslatedregion(K

ridlet

al.,1991,accession#M64632);P7sa’,

Gly

cine

ma

xseed

storageprotein

promoter(C

hen

etal.,1986);P7sa,

Gly

cine

ma

xseed

storageprotein

promoter(W

angandDubois,2003);PArc-5,

Ph

ase

olu

svu

lga

risarcelin-5

promoter(W

anget

al.,

2002);

PFAE,Arabidopsisfattyacidelongase

promoter(accession#AF355982);

PNapin,Napin

seed-specificpromoter(K

ridlet

al.,1991,accession#M64632);

POleosin,

A.

tha

lia

naoleosinpromoter

(accession#X62353);

Sy

n-V

TE

2,

Sy

nec

ho

cyst

issp.PCC

6803homogentisate

phytyltransferase

(accession#NP_441094).

aBestperform

ingeventforcanola,andmeantocochromanolcontentofbestperform

ingeventforsoybean;Means,standard



HSD.

bMeansfollowed

bythesamelettersare

notsignificantlydifferentfrom

each

other

(a¼

0.05).


ARTICLE IN PRESS

Table

5

Seedtocochromanolcontentandcompositionofsingle

seed

ofselected

soybeanlines

transform

edwithpMON77637

Event

Averagetocopherolcontent(ng/m

g)

a–,g–,and

Averagetocotrienolcontent

a–,g–

and

Totaltocochromanolcontent

d-tocopherol

[ng/m

g]

d-tocotrienol

(ng/m

g)

a–g–

d–(ng/m

g)

a–g–

d–(ng/m

g)

Maxim

um

aMeanb

P-value

WT-control

33.073.03

22974.5

57.371.86

319717.8

5.8372.99

0.3370.82

0.6771.63

6.874.5

362

326720.6

–

GM_A40061:@

.22.676.50

124717.2

13971.18

285722.1

27.473.97

714742.2

1,1127265

18547293

2429

21397282

6.791�10�8

GM_A40056:@

.23.478.79

114719.9

17378.87

311725.6

35.479.96

845744.3

1,759798.9

26407118

3095

29517107

2.793�10111

GM_A40044:@

.22.073.39

110726.0

162712.6

294736.9

34.672.61

1,024798.3

2,0447387

31027473

3795

33967441

6.739�10�13

GM_A40049:@

.14.874.44

69.4721.8

165727.2

249750.1

12.4717.0

947798.9

2,4997232

34587196

3901

37077208

6.209�10�14

GM_A40059:@

.18.475.59

72.0713.2

165715.7

255729.5

45.8713.9

1,000792.2

2,4647145

35107163

3980

37657144

4.036�10�14

GM_A40048:@

.20.0713.3

52.6726.2

142729.4

214760.7

56.2726.0

7527197

2,7257856

35337930

4806

37487956

4.586�10�14

Fivedark

seed

per

eventwerechosenandanalyzedseparately

fortocochromanolcontentandcomposition.b-tocopherolandb-tocotrienolwerebelow

2%

oftotaltocochromanols.

aMaxim

um

tocochromanolcontentin

bestperform

ingsingle

soybeanseed.

bMeantocochromanolcontentoffivesingleseed.Means,standard



HSD.

P-values

are

provided

formeantocochromanolcontentof

transgenic

seed

comparedto

WT-seed,andvalues

o0.05are

significant.Plasm

idpMON77637harbors

seed-specificexpressioncassettesfor

At-

HP

PD,

Eh-T

YR

A,

At-

VT

E2,and

At-

GG

H.For

elem

entdetailseeTable

2.

Fig. 3. Seed phenotype and HGA accumulation in transgenic

Arabidopsis and soybean seed. Panels A and B show segregating

Arabidopsis and soybean seed transformed with pMON69909, and

pMON69943, respectively. These plasmids contained seed-specific

expression constructs for Eh-TYRA, At-HPPD, and Syn-VTE2

(Tables 3 and 4). Panels C and D show LC-MS chromatograms for

HGA from Arabidopsis and soybean seed extracts harboring

pMON69909, and pMON69943, respectively.


6-geranylgeranylbenzoquinol (2M6GGBQ) levels wereanalyzed in selected seed samples from all three plantspecies. Detection limits for 2M6PBQ and 2M6GGBQwere 0.2 and 2 ppm, respectively. In Arabidopsis andcanola seed harboring At-HPPD, and Eh-TYRA or At-

HPPD, Eh-TYRA, and Syn-VTE2 expression con-structs, 2M6PBQ levels were indeed four and six-foldincreased above the detection limit. In WT-seed2M6PBQ was below the limit of detection. In Arabi-dopsis seed 2M6GGBQ were also increased from 10.2 to15.8 ppm. In WT and transgenic soybean seed 2M6PBQand 2M6GGBQ were not detectable.

3.5. Soybean optimized for vitamin E production

Previous studies in Arabidopsis and soybeanhave demonstrated that WT-seed tocopherols can beconverted to nearly all a-tocopherol by expression of

ARTICLE IN PRESS

Fig. 4. Western blot analysis of soybean seed transformed with

pMON66682. Vector pMON66682 harbors seed-specific expression

cassettes for At-HPPD, Eh-TYRA, At-VTE2, and At-GGH. Seed

extracts were immuno detected with antisera against At-HPPD, Eh-

TYRA, and At-GGH. At-VTE2-expression was not immuno blot

confirmed due to the lack of specific antibodies. Abbreviations: MW,

molecular mass; nd, below the limit of detection.

Fig. 5. Correlation between a-tocopherol and total tocochromanol

levels in F2 segregating soybean seed obtained through crossing

homozygous soybean lines transformed with pMON69943, and

pMON67227. Plasmids pMON69943 and pMON67227 harbored the

P7Sa‘-CTP2-At-HPPD-E9 30, P7Sa‘-CTP1-Eh-TYRA-E9 30, PArc�5-

CTP1-Syn-VTE2-Arc 30, and P7sa’-At-VTE4-E9 30, P7sa’-At-VTE3-E9

30 expression cassettes, respectively.


one (Shintani and DellaPenna, 1998) or two methyl-transferases (Van Eenennaam et al., 2003). To test if thisconversion can still be accomplished with seed havingsubstantially increased total seed tocochromanol levels,homozygous soybean plants transformed withpMON69943, harboring seed specific Eh-TYRA, At-

HPPD, and Syn-VTE2 expression constructs werecrossed with homozygous soybean lines transformedwith pMON67227, harboring seed-specific expressionconstructs for At-VTE3 and the Arabidopsis g-methyl-transferase (At-VTE4) (Van Eenennaam et al., 2003).

Single seed from the phenotypically segregating F2-generation were analyzed for seed tocochromanol

content and composition. The tocochromanol data fromthese seed indicated the presence of four phenotypicallydistinct seed populations: a population of WT-seed, asecond population of seed containing nearly 100%a-tocopherol, a third seed population exhibiting theelevated total tocochromanol seed phenotype, and afourth seed population combining both transgenic seedtocochromanol phenotypes (Fig. 5). Four seed out of atotal population of 300 seed did not fit this pattern andtheir tocochromanol values were not included in furthercalculations. The seed population combining the in-creased a-tocopherol phenotype with the increased totaltocochromanol phenotype exhibited a significant aver-age tocochromanol increase by 1.2-fold compared to thehigh total tocochromanol population. In addition, theaverage a-tocopherol and a-tocotrienol content in theseseed increased to 83.8%.

Using vitamin E activities for the different tocochroma-nol isoforms provided by Chow (Chow, 2001), the totalvitamin E content in F2 seed combining the high totaltocochromanol phenotype with the high a-tocopherol andhigh a-tocotrienol phenotype, was calculated to be up to11-fold higher than in WT-soybean seed. The phenotypesin seed color and shape observed previously in seedharboring both, At-HPPD and Eh-TYRA expressionconstructs (see above) were maintained. However, thegermination vigor of F1 seed from this crossing experimentwas substantially improved, and was under green houseconditions comparable to non-transgenic soybean seed.

4. Discussion

The present paper describes a systematic approach toidentify limiting factors of tocochromanol biosynthesis,and a new strategy to use this knowledge for engineering


increased seed tocochromanol and vitamin E levels inoilseed crops. Feeding experiments performed withsoybean suspension cultures indicated that the avail-ability of HGA, phytol, and presumably PDP limittocopherol biosynthesis. In our hands these twometabolites appeared to have comparable effects ontocochromanol biosynthesis. Similar experiments per-formed with safflower suspension cultures also identifiedthese two metabolites as key intermediates for increasedtocochromanol biosynthesis, but found phytol toprovide a stronger effect than HGA (Furuya et al.,1987). If this difference reflects variations in the twodifferent cell suspension cultures is not known at thistime.

We chose the unicellular cyanobacterium Synechocys-

tis sp. PCC 6803 as a model to verify and extend theseexperimental data to a transgenic system with a fastturnaround time. In Synechocystis, overexpression ofAt-HPPD as a single trans gene had by far the strongestimpact on total tocochromanol levels, resulting in up toseven-fold WT-tocochromanol levels. This was in starkcontrast to HPPD-expression in oilseed. In Arabidopsisseed tocochromanol increased only by 1.09 and 1.25-fold compared to WT-levels for the population of allevents, and for the best performing event, respectively.This correlates well with data published by Tsegaye etal. (2002) who found up to 1.28-fold increased seedtocochromanol levels in Arabidopsis expressing At-

HPPD under the control of the seed-specific DC3promoter. The differences on tocochromanol accumula-tion upon HPPD expression in Synechocystis versusplants suggest substantial differences in either theregulation of tocochromanol biosynthesis or the toco-chromanol biosynthetic pathways. One known differ-ence in biochemical pathways for HGA biosynthesis inbacteria and higher plants is the presence of abifunctional prephenate dehydrogenase in bacteria,which allows HGA biosynthesis from chorismate viaprephenate and HPPA (Fig. 1). In fact, the Synecho-

cystis genome includes an open reading frame (slr2081)with a deduced amino acid sequence that is 28%identical to the deduced Bacillus subtilis TYRA aminoacid sequence (Henner et al., 1984). In plants, HGA issynthesized from chorismate, via prephenate, arogenate,tyrosine, and HPPA (Fig. 1) (Siehl, 1999).

Expression of Eh-TYRA in Synechocystis cultures andseed-specific expression of Eh-TYRA in Arabidopsis,canola, and soybean resulted in a moderate tocochro-manol increase of 1.60-, 1.20-, 1.63-, and 1.11-fold,respectively. While the average tocochromanol changesin Arabidopsis and canola were significant, the toco-chromanol increase in Eh-TYRA expressing soybeanwas not (Tables 2 and 4). The best performing eventsfrom Eh-TYRA expressing Arabidopsis and canola hadtotal seed tocochromanol levels increased to 1.53-foldand 2.37-fold WT-levels, respectively. The reasons for

the differences in the impact on total tocochromanols inArabidopsis, canola and soybean are not known at thistime. Traces of tocotrienols (o5%) were detected insome Eh-TYRA expressing seed from canola andsoybean.

Combined seed-specific expression of At-HPPD andEh-TYRA in Arabidopsis, canola, and soybean estab-lished the direct pathway for HGA biosynthesis as it ispresumed to exist in Synechocystis when expressingHPPD. Interestingly, this gene combination increasedthe average total seed tocochromanol levels to 1.78-fold,2.44-fold, and 2.58-fold versus WT-levels for Arabidop-sis, canola, and soybean seed, respectively. Thisrepresented a significant increase in total seed tocochro-manol levels in all populations when compared to WT-seed. When compared to seed populations harboringsingle gene expression constructs, the level of totaltocochromanol increase remained significant in Arabi-dopsis and canola. In addition, Arabidopsis andsoybean seed showed a brown seed phenotype thatcorrelated with a substantial increase of HGA levels(Fig. 3). In summary, these data provided strongevidence for the potency of the HPPD and TYRA genecombination, and that HGA limits tocochromanolbiosynthesis in these systems. A role of HGA as limitingprecursor is consistent with previous feeding experi-ments, and with a conclusion recently communicated byCollakova and DellaPenna, which is based on mRNAlevels of tocopherol pathway enzymes during stressconditions (Collakova and DellaPenna, 2003a,b). It isalso consistent with data published by Rippert et al.(2004), that describe substantially increased tocochro-manol levels in tobacco leaf harboring expressionconstructs for yeast TYRA, and At-HPPD.

The brown seed color in Arabidopsis and soybeanwas reminiscent of the brown coloration of culturesupernatant of Synechocystis cultures expressing At-

HPPD, or At-HPPD and Eh-TYRA. These parallels incoloration, and the impact on tocochromanol levels inSynechocystis expressing At-HPPD, and in dicotyledo-nous plants expressing Eh-TYRA and At-HPPD sup-port the previous working theory that TYRA andHPPD enzymes can create a potent shunt for HGA-biosynthesis, which can increase HGA-levels by 60 and800-fold in mature Arabidopsis and soybean seed,respectively. The high efficiency of the Eh-TYRA-At-

HPPD combination for HGA-biosynthesis presumablyresults from bypassing the tyrosine regulatory mechan-ism. Tyrosine biosynthesis is feed-back inhibited bytyrosine on the level of the chorismate mutase reaction(Siehl, 1999). Co-expression of Eh-TYRA and At-HPPD

appears to direct intermediates from chorismate toHGA without substantially altering tyrosine levels.Because tyrosine levels are affected only marginally,the increased flux towards HGA is not subjected to feed-back inhibition. In Synechocystis, the endogenous


TYRA activity was apparently not limiting, and there-fore, the high HGA- and high tocochromanol pheno-type was achieved by expression of At-HPPD alone.Plants are not known to express a TYRA like gene andtherefore, TYRA and HPPD are both required toestablish this phenotype. Interestingly, the elevatedHGA-phenotype was accompanied by the accumulationof tocotrienols, which accounted for up to 20%, 42%,47%, and 72% of total tocochromanols in Synechocys-

tis, Arabidopsis, canola, and soybean, respectively. Thisobservation is consistent with observations from Rip-pert et al. (2004), who also reported the accumulation oftocotrienols in tobacco leaves harboring yeast TYRA

and At-HPPD expression constructs.Collakova and DellaPenna (2001) reported that At-

VTE2 does not utilize GGDP as substrate. In our handshowever, the Arabidopsis homogentisate phytyltransfer-ase utilizes both PDP and GGDP as substrates(unpublished results), but displays a marked preferencefor PDP over GGDP when assayed in vitro (S. Hunterand E. Cahoon, personal communication). With theobserved preference of At-VTE2 for PDP it is surprisinghow such large tocotrienol levels were produced inArabidopsis. An alternative might be the existence of aVTE2-paralog, as it has been detected by recentbioinformatics analysis of the Arabidopsis genome(Valentin et al., 2003). The large amounts of HGAwhich always coincided tocotrienol biosynthesis mightact as a stress signal, that either activates the VTE2

paralog, or alters At-VTE2 substrate specificity. How-ever, only altered substrate specificity would be con-sistent with the tocotrienol increase that was observedupon co-expression of At-VTE2 in addition of Eh-

TYRA and At-HPPD in Arabidopsis, and soybean(Tables 3 and 4).

Another interesting variation in tocopherol pathwayengineering between Synechocystis and oilseed wasfound in the conversion of tocotrienols to tocopherols.In Synechocystis cultures expressing At-HPPD and Eh-

TYRA, the additional expression of At-GGH appearedto be sufficient to convert tocotrienols to tocopherols.This was concomitant with the disappearance of browncoloration in the culture supernatant, and with theutilization of excess HGA for increased tocopherolformation. HGA utilization was presumably improveddue to increased availability of PDP. Increased totaltocopherols in these cultures are also consistent withPDP being the preferred substrate for Syn-VTE2 (datanot shown). A role for GGH in supplying precursors fortocopherol biosynthesis was suggested by Tanaka et al.(1999) where antisense suppression of GGH in trans-genic tobacco resulted in reduced tocopherol levels, andenzyme assay data, which suggest that E. coli expressedAt-GGH can reduce GGDP and geranylated chloro-phyll (Keller et al., 1998). In our studies however,conversion of tocotrienols to tocopherols by transgenic

expression of the At-GGH was clearly not obtained intransgenic seed. Co-expression of At-GGH with At-

HPPD, Eh-TYRA, and At-VTE2 did not reducetocotrienol levels in soybean seed transformed withpMON66682, although immuno blot analysis of ex-tracts from these soybean seed was consistent with At-

GGH being expressed, and plastid targeted, andsequence analysis of the trans-gene confirmed thesequence integrity. Comparison of the Synechocystis

results to plant seed tocochromanol data suggests thatthere may be additional unidentified factors thatfacilitate the conversion of tocotrienols to tocopherolsin presence of GGH in the bacterial system, which arepossibly lacking in Arabidopsis, canola and soybeanseed. Additional experiments are needed to explain thesedifferences.

The highest total seed tocochromanol levels wereobtained in transgenic seed systems through combinedseed-specific expression of one or two genes in additionto the TYRA and HPPD combination. In Arabidopsis,expression of At-VTE2 in addition to Eh-TYRA, and At-

HPPD increased the average seed tocochromanol levelof the R2 seed populations 2.88 fold compared to WT-seed. A comparable gene combination, using Syn-VTE2

in canola and soybean provided events with higher totaltocochromanol levels than events that harbored only Eh-

TYRA, and At-HPPD expression constructs. However,the impact on all R1 seed populations was notstatistically significant (Table 4), or was even negative,as in canola. In our data set, this may be the result oflarge individual variations in total tocochromanol levelsfrom event to event, resulting in a high standarddeviation. The reasons for high variability from eventto event are not well understood yet, but the correlationof the increase in variability from event to event withincreasing number of transgenes, and therefore, vectorssize may indicate a connection to T-DNA stability.

In soybean the highest total seed tocochromanollevels were obtained with seed transformed withpMON77637. This binary vector harbored seed specificexpression constructs for At-HPPD, Eh-TYRA, At-

VTE2, and At-GGH. Four transformation events fromthis construct produced on average 10.4–11.5-fold WT-seed tocochromanol levels (Table 3). Whether At-GGH

expression had a positive impact on total tocochroma-nol accumulation in these seed is currently unknown.However, it appears unlikely, because the tocochroma-nol increase observed in these seed appears to resultfrom tocotrienol accumulation only (Table 5). In facttocotrienols appear to increase in these seed whiletocopherol levels appear to decrease resulting in seedwith an average of 494% tocotrienols for eventGM_A40048:@. In contrast, WT soybean seed containonly traces of tocotrienols.

Compared to previous studies of tocopherol pathwayengineering the total fold tocochromanol increase


obtained here does not appear unusual, as Cahoon et al.(2003) achieved a 15-fold total tocochromanol increasesin Arabidopsis leaf, and 4–6-fold total tocochromanolincrease in corn seed by transgenic expression of abarley homogentisate geranylgeranyl transferase. Evenhigher tocochromanol increases have been reported as aresult of stress conditions. Collakova and DellaPenna(2003b) obtained up to 18-fold tocochromanol increasein Arabidopsis leaf as a result of abiotic stress. However,the final tocochromanol levels of 900 ng/mg dryArabidopsis leaf material, and 334 ng/mg in corn seedreported by Cahoon et al. appear low compared to4800 ng/mg obtained in this study in soybean seed. Themaximum tocochromanol levels obtained by Cahoon etal. in Arabidopsis leaf appear to be of the samemagnitude as the 926 ng/mg seed tocopherol reportedby Savidge et al. (2002) in Arabidopsis seed expressingAt-VTE2. Whether these numbers suggest that expres-sion of the enzyme catalyzing the first committedreaction in tocochromanol biosynthesis (VTE2, orHGGT) reaches its maximum at 900–1000 ng/mg tissuematerial under non stress conditions remains to betested in further experiments. However, the combinedexpression of enzymes, which increase tocochromanolprecursor availability with VTE2 expression, as pre-sented here, appears to provide a potent strategy toincrease total tocochromanol levels in seed and cyano-bacterial systems.

The crossing experiment performed between soybeanlines that accumulate almost exclusively a-tocopherol asa result of transgenic At-VTE3 and At-VTE4 expression,and soybean lines with about 3-fold increased total seedtocochromanol levels as a result of At-HPPD, Eh-

TYRA, and Syn-VTE2 expression demonstrated afurther average seed tocochromanol increase by 1.2-fold in seed combining both sets of genes as compared toseed exhibiting the high tocochromanol phenotype only.This result is consistent with a flux limitation by VTE3,which had been predicted based on a shift in tocochro-manol composition towards d-tocopherol and d-toco-trienol in seed with increased total tocochromanol levels(Table 5), and it is consistent with conclusions reportedby Collakova and DellaPenna (2003b) that VTE3 maylimit tocopherol biosynthesis under high stress condi-tions. High stress conditions were correlated withincreased tocopherol biosynthesis.

A compositional tocochromanol shift towards theaccumulation of d-tocopherol and d-tocotrienol wasobserved in all plant seed when tocochromanol levelsexceeded 2-fold WT-levels (dataset for canola notshown, for Arabidopsis and soybean seed see Supple-mental Table 1, and Table 5, respectively). Thissuggested that the substrate competition betweenVTE1 and VTE3 favored the cyclase reaction, indicatingthat VTE3 was the dominating factor in flux limitationin seed systems. This appears to contrast observations

from Kumar et al. (2005) who saw increased leaftocopherol levels upon transgenic expression of VTE1

but may be the result of using different metabolicbackgrounds (leaf versus oil seed). Oil seed optimizedfor vitamin E production require that all tocopherols areconverted to the tocopherol isoforms with the highestvitamin E activity, a-tocopherol. Overexpression ofVTE1 is expected to divert flux towards d-tocopherolas observed by Kumar et al. (2005), and as documentedin this manuscript. Transgenic expression of VTE1 forengineering enhanced seed vitamin E content wouldtherefore require careful finetuning with transgenicVTE3 expression in order to keep the tocopherolcomposition at its optimum.

Interestingly, F1-seed from the crossing experimentsgerminated much more vigorously than high tocochro-manol seed harboring only the combined At-HPPD, Eh-

TYRA, and Syn-VTE2 expression constructs. LC-MSanalysis of such seed had provided some evidence forincreased 2M6PBQ- and 2M6GGBQ-levels in Arabi-dopsis and canola. In WT and transgenic soybean seedthese compounds remained below or at the limit ofdetection, making it impossible to decide if theseintermediates increased or decreased as a result of transgene expression. The recent characterization of At-

VTE3 (Motohashi et al., 2003; Cheng et al., 2003)revealed that this enzyme is required for both themethylation of 2M6PBQ, and 2-methyl-6-solanylbenzo-quinol (2M6SBQ), the precursor of plastoquinone.Plastoquinones are essential for plant metabolism. It istherefore possible that the shift in metabolite poolsdisrupts the balance between 2M6PBQ, 2M6GGBQ,and 2M6SBQ. Vastly increased 2M6PBQ- and2M6GGBQ-pools may inhibit plastoquinol formation,and thus result in poor seed vigor.

The work presented here demonstrates that oilseedtocochromanol levels can be increased substantially.The 4800 ppm seed tocochromanol obtained in soybeantranslates to a concentration of 2.4% tocochromanol inthe oil. Under these conditions the vast majority oftocochromanols were accumulated as d-tocotrienol,suggesting limitations by the availability of PDP andby VTE3 enzyme activity. The limitation by VTE3 issupported by our observation of increased totaltocochromanols obtained in soybean seed from thecrossing experiment of high tocochromanol soybeanlines with all a-tocopherol soybean lines. To which levelseed tocochromanols can be increased when theselimitations are removed remains to be tested in futureexperiments.

Acknowledgments

We wish to thank our colleagues in plant transforma-tion, and the plant growth facilities for their expert


assistance and support. We also wish to thank ourcolleagues at Renessen for their support and advice.

Appendix A. Supplementary Materials

The online version of this article contains additionalsupplementary data. Please visit doi:10.1016/j.ym-ben.2005.05.005.

References

Nomenclature rules for vitamin E (972.31), 1990. Official methods of

analysis. Assoc. Off Anal Chem, 15th ed., 109, 8–15.

IUPAC-IUP Joint Commission on Biochemical Nomenclature, 1982.

Nomenclature of tocopherols and related compounds Recommen-

dation. Eur. J. Biochem. 123, 473–475.

Barry, G.F., Kishore, G.M., 1998. Glyphosate tolerant plants. US

patent US5776760.

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.

Bramley, P.M., Elmadfa, I., Kafatos, A., Kelly, F.J., Manios, Y.,

Roxborough, H.E., Schuch, W., Sheehy, P.J.A., Wagner, K.-H.,

2000. Vitamin E. J. Sci. Food Agric. 80, 913–938.

Buring, J.E., Hennekens, C.H., 1997. Antioxidant vitamins and

cardiovascular disease. Nutr. Rev. 55, 53–60.

Cahoon, E.B., Hall, S.H., Ripp, K.G., Ganzke, T.S., Hitz, W.D.,

Coughlan, S.J., 2003. Metabolic redesign of vitamin E biosynthesis

in plants for tocotrienol production and increased antioxidant

content. Nature Biotech 21, 1082–1087.

Chen, Z.L., Schuler, M.A., Beachy, R.N., 1986. Functional analysis of

regulatory elements in a plant embryo-specific gene. Proc. Natl.

Acad. Sci. USA 83, 8560–8564.

Cheng, Z., Sattler, S., Maeda, H., Sakuragi, Y., Bryant, D.A.,

DellaPenna, D., 2003. Highly divergent methyltransferases catalyze

a conserved reaction in tocopherol and plastoquinone synthesis in

cyanobacteria and photosynthetic eukaryotes. Plant Cell 15,

2343–2356.

Chow, C.K., 2001. Vitamin E. In: Rucker, R.B., Suttie, J.W.,

McCormick, D.B., Machlin, L.J. (Eds.), Handbook of Vitamins

third ed. Marcel Dekker Inc, New York, pp. 165–197.

Collakova, E., DellaPenna, D., 2001. Isolation and functional analysis

of homogentisate phytyltransferase from Synechocystis sp. PCC

6803 and Arabidopsis. Plant Physiol 127, 1113–1124.

Collakova, E., DellaPenna, D., 2003a. Homogentisate phytyltransfer-

ase activity is limiting for tocopherol biosynthesis in Arabidopsis.

Plant Physiol 131, 632–642.

Collakova, E., DellaPenna, D., 2003b. The role of homogentisate

phytyltransferase and other tocopherol pathway enzymes in

regulation of tocopherol biosynthesis during abiotic stress. Plant

Physiol 133, 930–940.

Coon, S.L., Kotob, S., Jarvis, B.B., Wang, S., Fuqua, W.C., Weiner,

R.M., 1994. Homogentisic acid is the product of MelA, which

mediates melanogenesis in the marine bacterium Shewanella

colwelliana D. Appl. Environ. Microbiol. 60, 3006–3010.

Coruzzi, G., Broglie, R., Edwards, C., Chua, N.H., 1984. Tissue-

specific and light-regulated expression of a pea nuclear gene

encoding the small subunit of ribulose-1,5-bisphosphate carbox-

ylase. EMBO J 3, 1671–1679.

Cotton, R.G.H., Gibson, F., 1970. Tyrosine and phenylalanine

biosynthesis. T and P proteins (Aerobacter aerogenes); chorismate

mutase (Pisum sativum). Methods Enzymol 17 (Part A), 564–574.

Dickson, R.C., Abelson, J., Barnes, W.M., Reznikoff, W.S., 1975.

Genetic regulation: the Lac control region. Science 187, 27–35.

Edwards, H., 2001. Vitamin E: an important anti-oxidant in the skin?

Retin lipid-soluble Vitamin. Clin. Pract. 17 (2), 43–47.

Furuya, T., Yoshikawa, T., Kimura, T., Kaneko, H., 1987. Production

of tocopherols by cell culture of safflower. Phytochemistry 26

2741–2727.

Harlow, E., Lane, D., 1988. Antibodies: A Laboratory Manual. Cold

Spring Harbor Laboratory Press, Cold Spring Harbor NY, p. 349.

Henner, D.J., Band, L., Flaggs, G., Chen, E., 1984. The organization

and nucleotide sequence of the Bacillus subtilis hisH, tyrA and aroE

genes. Gene 49 (1), 147–152.

Henry, A., Powls, R., Pennock, J.F., 1987. Intermediates of tocopherol

biosynthesis in the unicellular alga Scenedesmus obliquus. The

presence of three isomeric methylphytylbenzoquinones. Biochem.

J. 242, 367–373.

Katz, J.J., Strain, H.H., Harkness, A.L., Studier, M.H., Svec, W.A.,

Janson, T.R., Cope, B.T., 1972. Esterifying alcohols in the

chlorophylls of purple photosynthetic bacteria. A new chlorophyll,

bacteriochlorophyll(gg), all-trans-geranylgeranyl bacteriochloro-

phyllide a. J. Am. Chem. Soc. 94, 7938–7939.

Keller, Y., Bouvier, F., d’Harlingue, A., Camara, B., 1998. Metabolic

compartmentation of plastid prenyllipid biosynthesis. Evidence for

the involvement of a multifunctional geranylgeranyl reductase.

Eur. J. Biochem. 251, 413–417.

Klee, H.J., Muskopf, Y.M., Gasser, C.S., 1987. Cloning of an

Arabidopsis thaliana gene encoding 5-enolpyruvylshikimate-3-

phosphate synthase: sequence analysis and manipulation to obtain

glyphosate-tolerant plants. MGG 210, 437–442.

Kridl, J.C., McCarter, D.W., Rose, R.E., Scherer, D.E., Knutzon,

D.S., Radke, S.E., Knauf, V.C., 1991. Isolation and characteriza-

tion of an expressed napin gene from Brassica rapa. Seed Sci. Res.

1, 209–219.

Kumar, R., Raclaru, M., Schuaeler, T., Gruber, J., Sadre, R., Luhs,

W., Zarhloul, K.M., Friedt, W., Enders, D., Frentzen, M., Weier,

D., 2005. Characterization of plant tocopherol cyclases and their

overexpression in transgenic Brassica napus seeds. FEBS Lett 579,

1357–1364.

Lassner, M.W., Savidge, B., Mitsky, T., Weiss, J., Post-Beittenmiller,

M.A., 2003. Nucleic acid sequences to proteins involved in

isoprenoid biosynthesis. US Patent No. 6,541,259 B1.

Martinell, B.J., Julson, L.S., Emler, C.A., Huang, Y., McCabe, D.E.,

Williams, E.J., 2002. Soybean Agrobacterium transformation

method. In US Patent 6,384,301.

McPherson, J.C., and Kay, R., 1994. DNA sequence for enhancing the

efficiency of transcription. US Patent 5,322,938.

Motohashi, R., Ito, T., Kobayashi, M., Taji, T., Nagata, N., Asami,

T., Yoshida, S., Yamaguchi-Shinozaki, K., Shinozaki, K., 2003.

Functional analysis of the 37 kDa inner envelope membrane

polypeptide in chloroplast biogenesis using a Ds-tagged Arabidop-

sis pale-green mutant. Plant J 34, 719–731.

Ng, W.-O., Zentella, R., Wang, Y., Taylor, J.S.A., Pakrasi, H.B., 2000.

phrA, the major photoreactivating factor in the cyanobacterium

Synechocystis sp. strain PCC 6803 codes for a cyclobutane-

pyrimidine-dimer-specific DNA photolyase. Arch. Microbiol. 173,

412–417.

Porfirova, S., Bergmuller, E., Tropf, S., Lemke, R., Dormann, P.,

2002. Isolation of an Arabidopsis mutant lacking vitamin E and

identification of a cyclase essential for all tocopherol biosynthesis.

PNAS USA 99, 12495–12500.

Poster, R.D., 1988. DNA transformation. In: Packer, L., Glazer, A.N.

(Eds.), Methods in Enzymology, vol 167. Academic Press Inc, San

Diego, pp. 703–712.

Radke, S.E., Turner, J.T., Facciotti, D., 1992. Transformation and

regeneration of Brassica rapa using Agrobacterium tumefaciens.

Plant Cell Rep 11, 499–505.

dx.doi.org/10.1016/j.ymben.2005.05.005

dx.doi.org/10.1016/j.ymben.2005.05.005


Rippert, P., Scimemi, C., Dubald, M., Matringe, M., 2004. Engineer-

ing plant shikimate pathway for production of tocotrienol and

improving herbicide resistance. Plant Physiol. 134, 92–100.

Rohmer, M., 2003. Mevalonate-independent methylerythritol phos-

phate pathway for isoprenoid biosynthesis. Elucidation and

distribution. Pure Appl. Chem. 75 (2–3), 375–388.

Sanders, S.K., Morgan, J.B., Wulf, D.M., Tatum, J.D., Williams,

S.N., Smith, G.C., 1997. Vitamin E supplementation of cattle and

shelf-life of beef for the Japanese market. J. Anim. Sci. 75,

2634–2640.

Savidge, B., Weiss, J.D., Wong, Y.-H.H., Lassner, M.W., Mitsky,

T.A., Shewmaker, C.K., Post-Beittenmiller, D., Valentin, H.E.,

2002. Isolation and characterization of tocopherol phytyltransfer-

ase genes from Synechocystis sp. PCC 6803 and Arabidopsis. Plant

Physiol 129, 321–333.

Schledz, M., Seidler, A., Beyer, P., Neuhaus, G., 2001. A novel

phytyltransferase from Synechocystis sp. PCC 6803 involved in

tocopherol biosynthesis. FEBS Let 499, 15–20.

Secor, J., 1994. Inhibition of barnyardgrass 4-hydroxyphenylpyruvate

dioxygenase by sulcotrione. Plant Physiol 106, 1429–1433.

Sheppard, A.J., Pennington, J.A., Weihrauch, J.L., 1993. Analysis and

distribution of Vitamin E in vegetable oils and foods. In: Packer,

L., Fuchs, J. (Eds.), Vitamin E in health and disease. Marcel

Dekker, New York, pp. 9–31.

Shintani, D., DellaPenna, D., 1998. Elevating the vitamin E content of

plants through metabolic engineering. Science 282, 2098–2100.

Siehl, D.L., 1999. The biosynthesis of tryptophan, tyrosine, and

phenylalanine from chorismate. In: Singh, B.K. (Ed.), Plant amino

acids: biochemistry and biotechnology. Marcel Dekker Inc, New

York, pp. 171–204.

Soll, J., Schultz, G., 1980. 2-Methyl-6-phytylquinol and 2,3-dimethyl-

5-phytylquinol as precursors of tocopherol synthesis in spinach

chloroplasts. Phytochemistry 19, 215–218.

Soll, J., Kemmerling, M., Schultz, G., 1980. Tocopherol and

plastoquinone synthesis in spinach chloroplasts subfractions. Arch.

Biochem. Biophys. 204, 544–550.

Soll, J., Schultz, G., Rudiger, W., Benz, J., 1983. Hydrogenation of

geranylgeraniol. Plant Physiol 71, 849–854.

Soll, J., Schultz, G., Joyard, J., Douce, R., Block, M.A., 1985.

Localization and synthesis of prenylquinones in isolated outer and

inner envelope membranes from spinach chloroplasts. Arch.

Biochem. Biophys. 238, 290–299.

Tanaka, R., Oster, U., Kruse, E., Rudiger, W., Grimm, B., 1999.

Reduced activity of geranylgeranyl reductase leads to loss of

chlorophyll and tocopherol and to partially geranylgeranylated

chlorophyll in transgenic tobacco plants expressing antisense RNA

for geranylgeranyl reductase. Plant Physiol 120, 695–704.

Tangney, C.C., 1997. Vitamin E and cardiovascular disease. Nutr.

Today 32, 13–22.

Traber, M.G., Sies, H., 1996. Vitamin E in humans: demand and

delivery. Ann. Rev. Nutr. 16, 321–347.

Tsegaye, Y., Shintani, D.K., DellaPenna, D., 2002. Overexpression of

the enzyme p-hydroxyphenylpyruvate dioxygenase in Arabidopsis

and its relation to tocopherol biosynthesis. Plant Physiol. Biochem.

40, 913–920.

Valentin, H.E., Venkatesh, T.V., Karunanandaa, B., 2003. cDNAs

encoding homogentisate prenyltransferase sequence homologs and

their use in improving tocopherol synthesis in transgenic plants.

International patent application WO 2003080647.

Valentin, H.E., Mitsky, T.A., 2002. TYRA genes and uses thereof.

International patent application WO 02/089561 A1.

Van Eenennaam, A.L., Lincoln, K., Durrett, T.P., Valentin, H.E.,

Shewmaker, C.K., Thorne, G.M., Jiang, J., Baszis, S.R., Levering,

C.K., Aasen, E.D., Hao, M., Stein, J.C., Norris, S.R., Last, R.L.,

2003. Engineering vitamin E content: from Arabidopsis mutant to

soy oil. Plant Cell 15, 3007–3019.

Wang, Q., Dubois, P., 2003. Seed specific 7Sa promoter for expressing

genes in plants. International patent application WO 03/020016.

Wang, Q., Dubois, P., Liang, J., Oulmassov, T., 2002. Sequence of

Phaseolus vulgaris arcelin gene promoter T-Arc5 and the uses of

the promoter for enhancement of gene expression in transgenic

plants. International patent application WO 0250295 A2.

Documents

Metabolically engineered oilseed crops with enhanced seed tocopherol