Phytochemistry 74 (2012) 90–99

Contents lists available at SciVerse ScienceDirect

Phytochemistry

journal homepage: www.elsevier .com/locate /phytochem

Molecular cloning and biochemical characterization of the UDP-glucose:Flavonoid 3-O-glucosyltransferase from Concord grape (Vitis labrusca)

Dawn Hall 1, Xiao Xin Yuan 2, Jun Murata 3, Vincenzo De Luca ⇑Department of Biological Sciences, Brock University, 500 Glenridge Ave., St. Catharines, ON, Canada L2S 3A1

a r t i c l e i n f o

Article history:Received 9 June 2011Received in revised form 31 August 2011Available online 16 November 2011

Keywords:Vitis labruscaVitaceaeAnthocyaninFlavonol3-O-GlucosyltransferaseGene expressionVeraisonEnzyme kineticsSubstrate specificityPhylogeny

0031-9422/$ - see front matter � 2011 Elsevier Ltd. Adoi:10.1016/j.phytochem.2011.10.007

⇑ Corresponding author. Tel.: +1 905 688 5550x455E-mail address: vdeluca@brocku.ca (V. De Luca).

1 Current address: Michael Smith Laboratories, The U#301 – 2185 East Mall, Vancouver, BC, Canada V6T 1Z

2 Current address: 175 Bamburgh Circle #1711, Sc3X8.

3 Current address: Suntory Foundation for Life SInstitute, Osaka 618-8503, Japan.

a b s t r a c t

Glucosylation of anthocyanidin substrates at the 3-O-position is crucial for the red pigmentation of grapeberries and wine. The gene that encodes the enzyme involved in this reaction has been cloned from Vitislabrusca cv. Concord, heterologously expressed, and the recombinant enzyme (rVL3GT) was character-ized. VL3GT has 96% amino acid sequence identity with Vitis vinifera VV3GT and groups phylogeneticallywith several other flavonoid 3-O-glycosyltransferases. In vitro substrate specificity studies and kineticanalyses of rVL3GT indicate that this enzyme preferentially glucosylates cyanidin as compared with quer-cetin. Crude protein extracts from several Concord grape tissues were assayed for glucosyltransferaseactivity with cyanidin and quercetin as acceptor substrates. A comparison of the VL3GT activities towardwith these substrates showed that the 3GT enzyme activity is consistent with the expression of VL3GT inthese tissues and is coincident with the biosynthesis of anthocyanins in both location and developmentalstages. Enzyme activities in grape mesocarp, pre-veraison exocarp, leaf, flower bud, and flower tissuesglucosylated quercetin but not cyanidin at high rates, suggesting the presence of additional enzymeswhich are able to glucosylate the 3-O-position of flavonols with higher specificity than anthocyanidins.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

The cultivation of grapes (viticulture) and the production ofwine are of cultural and economic importance worldwide. Thewine industry relies on several traditional European (Vitis vinifera)cultivars of grape which are recognized as the ‘‘elite’’ viticulturally-important varieties (De Luca, 2011). These traditional cultivars canbe classified based on their organoleptic attributes that include thedegree of coloration of various red wines to non-colored whitewines. Red wine is produced from red grapes that accumulateanthocyanin pigments mostly in the berry exocarp. In grapes, theanthocyanin profile is highly dependent on the species, cultivar,environmental growing conditions and developmental stage ofthe berry.

Anthocyanins are members of a large and diverse group ofphenylalanine-derived plant secondary metabolites, the flavonoids(Winkel-Shirley, 2001). Within the plant, secondary metabolites

ll rights reserved.

4; fax: +1 905 688 1855.

niversity of British Columbia,4.arborough, ON, Canada M1W

ciences, Bioorganic Research

mediate the plant’s interaction with its environment and haveimportant roles in attracting pollinators and seed dispersers, asvolatile aroma compounds, as structural components in cell wallbiosynthesis, and in the plant’s stress response (reviewed in Koeset al., 1994; Wink, 2003; Winkel-Shirley, 2001). The diversity ofplant natural products is increased by enzymatic modificationincluding hydroxylation, methylation, glucosylation and acylation,changing the biochemical and biological characteristics of thesecompounds.

Glucosylation is often the terminal reaction in flavonoid biosyn-thesis, and is catalyzed by a large, multi-gene family of cytosolicenzymes, the glucosyltransferases (Li et al., 2001; Bowles, 2002;Lim et al., 2003). These enzymes generally have broad but regio-selective substrate specificity and transfer glucose from uridine50-diphosphoglucose (UDPG) to low molecular weight moleculesincluding plant-derived or exogenous metabolites encounteredby the plant (Vogt and Jones, 2000; Jones and Vogt, 2001). Glucosy-lation changes the toxicity, stability, complexity, spectral charac-teristics, and solubility of these compounds (Vogt and Jones,2000). Within the plant, glucosylation is often required for metab-olite transport and storage, and glucosylated metabolites are fre-quently recognized by signal receptors within the same plant,bacteria and other plants (Jones and Vogt, 2001).

Grapes contain more than 200 glucosylated metabolites includ-ing flavonoids, hormones, stilbenes, terpenes and hydroxycin-namic and hydroxybenzoic acids (Sefton et al., 1993, 1994;

D. Hall et al. / Phytochemistry 74 (2012) 90–99 91

Monagas et al., 2005). Extensive studies with V. vinifera grapes re-port the presence of cyanidin (3), delphinidin (4), peonidin (2),petunidin and malvidin (1)-derived anthocyanins (see Fig. 1 forstructures) in certain varieties which have unique patterns of B-ring hydroxylation and methylation (reviewed in Mazza andMiniati, 1993; Boss et al., 1996a,b). Similarly mature grapes of Vitislabrusca cv. Concord contain a larger diversity of cyanidin (3),delphinidin (4), peonidin (2), petunidin and malvidin (1)-derivedsubstitution patterns than V. vinifera grapes (Wu and Prior, 2005;Wu et al., 2006; Hall et al., 2011).

The anthocyanidin aglycone is highly reactive, and is greatly af-fected by pH, temperature, solvents, light, enzymes and the pres-ence of other molecules which can function as co-pigments (Priorand Wu, 2006). Glucosylation of the anthocyanidin at the 3-O-position, increases its stability, decreases its reactivity, andchanges its spectral characteristics. The enzyme which catalyzesthis reaction, UDP-glucose: flavonoid 3-O-glucosyltransferase(3GT), has been identified and functionally characterized in severalplant species (Tanaka et al., 1996; Gong et al., 1997; Ford et al.,1998; Yamazaki et al., 2002; Yoshihara et al., 2005; Kim et al.,2006; Offen et al., 2006).

In V. vinifera cv. Shiraz, the 3GT has been cloned and charac-terized (VV3GT; Ford et al., 1998; Offen et al., 2006), andin vitro, glucosylates the 3-O-position of flavonols and anthocy-anidins (Ford et al., 1998). In the plant, VV3GT is expressed coin-cident with anthocyanin but not flavonol biosynthesis and kinetic

0100200300400500600700800900

1000

Isorham

netin (5

)

Querceti

n (6)

Dihydroquerc

etin (7

)

Kaempfer

ol (8)

0

5000

10000

15000

20000

25000

30000

35000

Malvidin (1) Peonidin (2) Cyanidin (3) Delphinidin (4)

Substrate

Substrate

ytivitcacifi cepS

)g

msl o

mp(1-

1-yti vit ca

cifi cepS)

gm

sl omp(

1-1-

A

B

O+OH

OH

OCH3OH

OCH3

OH

O+OH

OH

OCH3OH

OH

O+OH

OH

OHOH

OH

O+OH

OH

OHOH

OH

OH

OOH

OH

OCH3OH

OHO

OOH

OH

OHOH

OHO

OOH

OH

OH

OHO

OHO

OH

OHOH

OOH

Fig. 1. Substrate specificity of Vitis labrusca flavonoid 3-O-glucosyltransferase withanthocyanidin (A) and flavonol (B) substrates. Each bar respresents the mean ofthree independent trials ± sandard deviation.

analysis established that VV3GT glucosylates anthocyanidins at48-fold higher rates than their corresponding flavonols; suggest-ing that the in planta VV3GT is involved in the glucosylation ofanthocyanidins, but not flavonols (Ford et al., 1998). Consistentwith a role in anthocyanin biosynthesis, VV3GT gene expressionis crucial for anthocyanin accumulation in the berry exocarp ofred grapes; whereas this gene is not expressed in tissues whichdo not accumulate anthocyanins (berry mesocarp, white grapes,vegetative tissues) (Boss et al., 1996a,b; Kobayashi et al., 2001).Unlike several other plant species which exert transcriptionalcontrol of anthocyanin biosynthesis at earlier steps in the enzy-matic pathway (Mol et al., 1998; Koes et al., 2005), it appears thatanthocyanin biosynthesis in grapes is controlled by expression ofthe 3GT (Boss et al., 1996a,b; Kobayashi et al., 2002). Recently,the crystal structure of VV3GT has been resolved and severalamino acid residues which are essential for catalytic activity forboth donor and acceptor substrate binding have been identified(Offen et al., 2006).

This report describes the molecular cloning, biochemical char-acterization and gene expression analysis of a 3GT from the NorthAmerican grape V. labrusca (VL) cv. Concord that is frequently usedin the production of grape juices and jams, but has limited value inviticultural applications. VL3GT has 96% protein sequence identityto the previously characterized VV3GT, and glucosylates flavonolsand anthocyanidins in vitro (Fig. 2). Substrate specificity and ki-netic analysis indicate that this enzyme preferentially glucosylatesanthocyanidins in vitro, which corresponds with the expression ofthis gene in tissues that produce and accumulate anthocyanins. En-zyme assays with crude VL protein extracts reveal high levels ofanthocyanidin 3GT activity in post-veraison grape berry exocarp,whereas enzyme activities in grape flesh and other tissues showedhigher activity towards quercetin (6).

2. Results

2.1. Molecular cloning and functional characterization of a UDP-glucose: flavonoid 3-O-glucosyltransferase from Concord grape

Oligonucleotide primers were selected based on the previouslycharacterized V. vinifera (cv. Shiraz) UDP-glucose: flavonoid 3-O-glucosyltransferase (VV3GT) (Ford et al., 1998) and were used toamplify and clone the corresponding gene from cDNA preparedfrom RNA isolated from exocarps of V. labrusca (VL) week 12 AFgrape berries. The full-length VL3GT clone (1371 bp) encoded aputative protein of 456 amino acids, with an estimated pI of 6.29(compared to 6.52 for VV3GT) and a calculated molecular weightof 50.1 kDa. The nucleotide and amino acid sequence of VL3GTshowed 98% and 96% identity, respectively, to the previously char-acterized VV3GT from V. vinifera (Ford et al., 1998). A search of theNCBI database identified fourteen 3GT-like sequences from Vitisspp., of which 8 were unique, but highly similar to one another(96–99% nucleotide sequence identity) and to the VL3GT. Severalamino acid residues which are essential for catalytic activity andimportant for donor and acceptor substrate binding in VV3GT(Offen et al., 2006) are also conserved in VL3GT (Fig. 2). The VL3GTbelongs to the family 1 glucosyltransferase group based on thepresence of the Plant Secondary Product Glucosyltransferase(PSPG) consensus sequence at the C-terminus of the protein(Fig. 2; Vogt and Jones, 2000).

After cloning VL3GT into an N-terminal GST-fusion vector, it wasco-expressed with chaperone protein in Escherichia coli cells grownfor 20–24 h at 16–18 �C. The rVL3GT protein extracted from thesecells was purified by affinity chromatography and assayed for glu-cosyltransferase activity with UDPG and either cyanidin (3) orquercetin (6) as substrates. The substrate concentrations, enzyme

Fig. 2. Protein sequence alignment of the functionally characterized flavonoid 3-O-glucosyltransferases (3GT) from Vv3GT – Vitis vinifera cv. Shiraz and Vl3GT – Vitis labruscacv. Concord. The 14 unique amino acids between Concord and Shiraz are shaded in grey. Arrow indicates the catalytic base His20. Residues shaded in black are essential forenzyme activity and symbols indicate residues important for donor (�) and acceptor (�) substrate binding (Offen et al., 2006). The plant secondary product glucosyltransferase(PSPG) consensus sequence is underlined. Abbreviations: VV_3GT – V. vinifera flavonoid 3-O-glucosyltransferase (AAB81683; Ford et al., 1998). VVL_3GT – V. labrusca flavonoid3-O-glucosyltransferase (EF630356; this study).

92 D. Hall et al. / Phytochemistry 74 (2012) 90–99

amounts and reaction times were determined empirically to pro-duce suitable amounts of products for reproducible detection andidentification by HPLC. Enzyme assays with cyanidin (3) and UDPGsubstrates remained linear for 3 min and produced a monogluco-side as shown by mass spectrometry (449.1 m/z), while assays withpurified protein extracts from bacteria expressing the empty vector(pGEX lacking VL3GT), with boiled rVL3GT protein or withoutUDPG produced no glucosides in the presence of cyanidin. Thereaction product was confirmed to be cyanidin 3-O-monoglucoside(kuromanin (9)) based on its identical UV spectrum and HPLCretention time (RT = 14.6 min) as the kuromanin standard. WhenrVL3GT was incubated with quercetin as the flavonoid substratethe reaction remained linear for 30 min to produce a reaction prod-uct identified as quercetin 3-O-monoglucoside (10) (RT = 19.3 min)as determined by HPLC, UV spectral analysis (hypsochromic shiftto 356 nm) (Harborne, 1967; Vogt et al., 1997; Kramer et al.,2003) and mass spectrometry (463 m/z [M�H]�). When the assaywas carried out for 15 min or longer, two additional products(RT = 18.9 min and RT = 14.6 min) were also detected that elutedearlier than quercetin 3-O-monoglucoside (10). The product withRT = 18.9 min with an absorption maxima of 370 nm was quercetin7-O-monoglucoside (11) (463 m/z [M�H]�) and product withRT = 14.6 min was quercetin 3,7-O-diglucoside (12) (625 m/z[M�H]�) as also determined by their UV spectral properties (Har-borne, 1967; Vogt et al., 1997; Kramer et al., 2003). In order to min-imize the second glucosylation reaction, subsequent kineticanalyses were carried out for only 5 min.

2.2. Recombinant VL3GT glucosylates several anthocyanins andflavonols in vitro

Substrate specificity studies for rVL3GT were performed by test-ing several anthocyanidin- and flavonol-type substrates. TherVL3GT glucosylated several anthocyanidins (malvidin (1), peoni-din (2), cyanidin (3) and delphinidin (4)) (Fig. 1A; SupplementaryTable 1A) and flavonols (isorhamnetin (5), quercetin (6),dihydroquercetin (7) and kaempferol (8)) (Fig. 1B; SupplementaryTable 1A) when assayed in the presence of UDP-glucose. While theenzyme was able to glucosylate quercetin (6) to the 3-, 7- and the3,7 positions in a time-dependent manner (SupplementaryTable 1B), it was not able to further glucosylate cyanidin 3-O-glu-coside (9) and it could not use UDP-galactose as a donor substrate.While rVL3GT preferentially glucosylated the 3 position of querce-tin (6), it also was able to glucosylate position 7 to a small extent inthe early stages of the time course (Supplementary Table 1B). As 3-and 7-O-monoglucosides began to accumulate, increasing levels of3,7-di-O-glucosides also appeared later in the time course of theassay (Supplementary Table 1B). The rVL3GT accepted the antho-cyanidin substrates peonidin (2) and cyanidin (3) at 93.5- and6.4-fold higher rates than their corresponding flavonol substrates,isorhamnetin (5) and quercetin (6) (Fig. 1). In addition the O-meth-ylated substrates isorhamnetin (5), peonidin (2) and malvidin (1)were O-glucosylated at 3.6-, 14.6- and 115-fold higher ratesrespectively than their non-O-methylated derivatives (quercetin(6), delphinidin (4) and cyanidin (3)) (Fig. 1).

D. Hall et al. / Phytochemistry 74 (2012) 90–99 93

2.3. Kinetic analysis of the recombinant VL3GT

The kinetic parameters of rVL3GT were determined with theacceptor substrates cyanidin and quercetin and the donor sub-strate UDPG (Table 1). The recombinant enzyme shows high affin-ity (KM) for both quercetin (6) (2.2 lM) and cyanidin (3) (4.8 lM),but has a 6.2-fold higher turnover number (kcat) and a 2.8-foldhigher calculated catalytic efficiency (kcat KM

�1) for cyanidin (3)(kcat = 7.02 � 10�4 s�1; kcat KM

�1 = 146 M�1 s�1) than for quercetin(6) (kcat = 1.13 � 10�4 s�1; kcat KM

�1 = 52 M�1 s�1); which suggeststhat rVL3GT preferentially glucosylates cyanidin (3) compared toquercetin (6). Saturation kinetics of rVL3GT with 6 lM cyanidin(3) and varying concentrations of UDPG gave an apparent KM valueof 914 lM, a kcat value of 2.47 � 10�4 s�1, and a calculated catalyticefficiency of 0.27 M�1 s�1 (Table 1). The corresponding kineticparameters of VV3GT were previously calculated and establishedapparent KM values of 30 lM and 15 lM, kcat values of47.6 � 10�4 s�1 and 0.99 � 10�4 s�1, and kcat KM

�1 values of159 M�1 s�1 and 6.6 M�1 s�1 for cyanidin (3) and quercetin (6)respectively as acceptor substrates (Table 1; Ford et al., 1998).

2.4. VL3GT gene expression and enzyme activity throughout grapeberry development

Protein and total RNA was extracted from several developmen-tal stages of V. labrusca berry exocarp and mesocarp tissue, as wellas from leaves, flower buds and open flowers. Similar extracts werealso produced for comparative purposes from V. vinifera cv. PinotNoir (PN) and cv. Pinot Noir Droit (PND) post-veraison (week 14AF) berry exocarp and mesocarp tissue. The quality of RNA waschecked for contamination with DNA by performing PCR with theRNA template which had not been subjected to reverse transcrip-tion (data not shown) and primers designed to amplify V. viniferaactin were used with the cDNA templates to determine baselineexpression between samples. Glucosyltransferase gene expressionwas quantified and divided by the mean Actin gene expression val-ues to obtain the relative % expression of VL3GT compared to thoseof Actin throughout grape berry development in exocarp and meso-carp tissues (Fig. 3). While there was little expression of VL3GT be-fore veraison, these transcripts began to accumulate at week 8 and10 AF to reach maximal levels at week 12 AF and decreased there-after (Fig. 3A). Similarly, VV3GT expression was detected in exocarpextracts of week 14 AF V. vinifera cv. Pinot Noir (PN) and cv. PinotNoir Droit (PND) (Fig. 3B). VL3GT gene expression was low butdetectable in V. labrusca leaves, flower buds and open flowers(Fig. 3C) while it was not detected in the mesocarp tissues of eitherV. labrusca or V. vinifera berries (Fig. 3A and B).

2.5. Detection of VL3GT activity during grape berry ripening

No glucosyltransferase activity with cyanidin (3) as a substratewas detected in week 1 to week 6 AF berries in either V. labruscamesocarp or exocarp tissue (Fig. 4A and B). However, cyanidin(3) 3GT activity was first detected in V. labrusca grape berryexocarp, at 8 week AF, and rapidly increased to maximal levels

Table 1Kinetic analysis of VL3GT and VV3GT with cyanidin, quercetin and UDPG All values repre

Substrate KM (lM)

VL3GTa VV3GTb

Cyanidin (3) 4.79 ± 0.69 30Quercetin (6) 2.16 ± 0.28 15UDPG (cyanidin)a (quercetin)b 914 ± 0.38 1880

a Present study.b As determined in Ford et al. (1998).

between week 12 and 16 AF (Fig. 4B). Similarly, high levels ofcyanidin (3) 3GT activity were also detected in week 14 AF V. vinif-era Pinot Noir and Pinot Noir Droit exocarp tissue, respectively.Cyanidin (3) 3GT activity was first detected at week 10 AF in V.labrusca berry mesocarp tissue and it increased to maximal levelsat week 14 AF (Fig. 4A); whereas low levels of cyanidin (3) 3GTactivity were detected in V. vinifera Pinot Noir Droit mesocarp tis-sue and no activity was detected in V. vinifera cv. Pinot Noir meso-carp tissue (Fig. 4B and C). Cyanidin 3GT activity was also detectedat low levels in grape flowers and flower buds but not in leaf tissue(Fig. 4C), and in general, this activity was 10-fold lower in meso-carp than in exocarp tissue.

In contrast to cyanidin (3) 3GT activity, quercetin (6) 3GT activ-ity was detected before veraison in V. labrusca exocarp and meso-carp tissues (Fig. 4A and B). While quercetin (6) 3GT activity wasdetected in week 1 and 3 AF grape mesocarp, it decreased andwas absent in week 6 and 8 AF before rapidly increasing to maximallevels at week 10 AF and subsequently decreasing to a constant le-vel to week 12–16 AF (Fig. 4A). Quercetin (6) 3GT activity was de-tected at low levels in pre-veraison V. labrusca exocarp tissue andit increased throughout ripening to reach maximal levels at week16 AF (Fig. 4B). Similarly, quercetin (6) 3GT activity was detectedin post-veraison week 14 AF V. vinifera Pinot Noir and Pinot NoirDroit exocarp and mesocarp tissues (Fig. 4A and B); and is highly ac-tive in V. labrusca leaf, flower and flower bud tissues (Fig. 4C). Ingrape berry post-veraison exocarp tissue, cyanidin (3) 3GT activitywas 1.8- to 4.8-fold higher than quercetin (6) 3GT activity, in con-trast to pre-veraison exocarp tissue, pre- and post-veraison meso-carp tissue and leaf, flower bud and flower tissues wherequercetin glucosyltransferase activity was favored (Fig. 4).

2.6. Phylogenetic analyses of VL3GT with functionally characterizedgenes

The VL3GT gene was submitted to phylogenetic analysis with 50separate functionally characterized glycosyltransferases that glyco-sylate anthocyanins, flavonoids, phenols and other small molecules.The resulting neighbour-joining phylogenetic tree places VL3GT in aclade with several other GTs to form 2 subclades containing genesfrom Vitis vinifera, Malus domestica, Arabidopsis thaliana, Dianthuscaryophyllus, and from Solanum melongena, Petunia hybrida, For-sythia intermedia, Gentiana triflora, Vigna mungo, respectively, thatadd glucose, rhamnose or galactose to the 3-O-position of anthocy-anidins and flavonol substrates (Fig. 5). Remarkably the flavonoid 3-O-glucosyltransferases of Manihot esculenta, Hordeum vulgare andZea mays are placed in a separate clade together with other GTs thatglucosylate other small molecules. However the functions of thecassava, barley and corn genes are based on indirect genetic evi-dence and their actual biochemical roles remain to be established.

3. Discussion

The current report describes the molecular cloning, heterolo-gous expression and biochemical characterization of a UDP-glu-cose: flavonoid 3-O-glucosyltransferase (VL3GT) from V. labrusca.

sent the mean of at least three independent trials ± standard deviation.

kcat (10�4 s�1) kcat/KM (M�1 s�1)

VL3GTa VV3GTb VL3GTa VV3GTb

7.02 47.6 146 1591.13 0.99 52 6.62.47 0.99 0.27 0.014

0

20

40

60

80

100

1 3 6 8 10 12 14 16

MesocarpExocarp

05

101520253035

Vitis labruscaweeks after flowering

Vitis vinifera

enegevitale

R)

%(noisserpxe

Vitis labrusca

enege vitale

R)

%(noi sser px eR

elat

ive

gene

expr

essi

on (%

)

A

B C

01020304050607080

Leaf Flowerbud FlowerVV PN VV PND

MesocarpExocarp

Fig. 3. Relative gene expression of VL3GT in relation toVLActin. Developmental gene expression profile of Vl3GT in relation to VlActin gene expression in Vitis labrusca (VL) cv.Concord berry mesocarp and berry exocarp tissues (A) and in Vitis vinifera cvs. Pinot Noir (PN) and Pinot Noir Droit (PND) berry mesocarp and exocarp tissues (B). (C) Relativegene expression of VL3GT in relation to VLActin gene expression in V. labrusca leaf, flower bud and flower tissue. All values represent the mean of at least two (mesocarp, leaf,flower, flower bud) or four (exocarp) independent trials ± SD.

94 D. Hall et al. / Phytochemistry 74 (2012) 90–99

A search of the NCBI database identifies several 3GT-like sequencesfrom Vitis spp. including 8 unique but highly identical full-lengthgenes. The ORF of the V. labrusca sequence has 96% amino acid se-quence identity to the corresponding enzyme from V. vinifera cv.Shiraz (Fig. 1), that has been cloned, functionally characterized(Ford et al., 1998) and crystallized (Offen et al., 2006). Substratespecificity studies indicate that VL3GT glucosylates the 3-O-posi-tion of anthocyanidin and flavonol substrates (Fig. 2), while kineticanalyses suggests that this enzyme preferentially glucosylatesanthocyanidins (Table 1). VL3GT is expressed in post-veraison exo-carp tissue (Fig. 3), consistent with the high levels of cyanidin (3)glucosylation and anthocyanin accumulation in these tissues(Fig. 4).

3.1. Vitis labrusca grapes accumulate methylated anthocyanins thatreflect the in vitro substrate specificity of VL3GT

The high level of protein sequence identity which existsbetween VL3GT and VV3GT suggests a common function in theglucosylation of anthocyanidins and flavonols at the 3-O-position,which was confirmed in vitro. Despite having nearly identicalprotein sequences, in vitro biochemical characterization of VL3GTreveals several functional differences between these two proteins.Whereas rVV3GT can use several donor substrates including UDP-Gal (Offen et al., 2006), rVL3GT is specific for UDPG as a substrate.Interestingly, rVV3GT glucosylates the 7-O-position of flavonoidsat low levels in vitro (Offen et al., 2006), and incubation of rVL3GTwith quercetin (6) for 3–30 min identified low levels of a homolo-gous activity (producing quercetin 7-O-glucoside (11) and querce-tin 3, 7-O-diglucoside (12)) (Supplementary Table 1B).

Substrate specificity studies indicate that rVV3GT preferentiallyglucosylates hydroxylated acceptor substrates (cyanidin (3), del-phinidin (4), quercetin (6)) (Ford et al., 1998); whereas rVL3GT

preferentially glucosylates their O-methoxylated derivatives(peonidin (2), malvidin (1), isorhamnetin (5)) (Fig. 2). This differ-ence in substrate specificities is not reflected by published studiesof anthocyanin profiles of V. vinifera cv. Shiraz and V. labrusca cv.Concord grape berries; since both accumulate malvidin-3-O-gluco-sides (Mazza and Miniati, 1993; Boss et al., 1996a,b; Wu and Prior,2005; Wu et al., 2006; Hall et al., 2011) and raise interesting ques-tions about the order of reactions occurring in these 2 species thatlead to their accumulation.

Methylation generally precedes glucosylation in flavonoid bio-synthesis and it is likely that the anthocyanin profiles of grape re-flect the substrate specificity of the O-methyltransferase(s) thatO-methylate the B-ring hydroxyl groups of anthocyanidins andflavonols. However, recent studies (Hugueney et al., 2009) havedescribed a cation dependent S-adenosly-L-methionine-30, 50-flavo-nol/anthocyanin-O-methyltransferase (30,50OMT) in V. vinifera thatshowed very similar OMT activities with quercetin (6), quercetin 3-O-glucoside (10), cyanidin (3) and cyanidin 3-O-glucoside (9)substrates. This suggests that in V. vinifera grapes anthocyanin gly-cosylation may occur before methylation. The preference ofVL3OGT for O-methoxylated derivatives could be explained if theputative V. labrusca 30,50OMT showed a preference for aglycones.This hypothesis remains to be tested by isolation and characteriza-tion of such an OMT from Concord grapes.

3.2. Calculated kinetic values indicate that VL3GT preferentiallyglucosylates anthocyanidins

The kinetic parameters of rVL3GT were determined for the do-nor substrate UDP-glucose and the acceptor substrates quercetin(6) and cyanidin (3) in order to compare these to those obtainedfor rVV3GT (Ford et al., 1998). Both rVV3GT and rVL3GT have 2-foldlower apparent affinity for cyanidin (3) (KM VL = 4.8 lM; KM

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

1 3 6 8 10 12 14 16 VVPN

VVPND

QuercetinCyanidin

0

0.5

1

1.5

2

2.5

QuercetinCyanidin

QuercetinCyanidin

ytivitcacificepSmtakp(

)g-1

ytivitcacificepSmtakp(

)g-1

ytivitcacificepSmtakp(

)g-1

Vitis labruscaWeeks after flowering

1 3 6 8 10 12 14 16 VVPN

VVPNDVitis labrusca

Weeks after flowering

0

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Bud Leaf FlowerVitis labrusca

Tissue

C

B

A

Fig. 4. Developmental profile of flavonoid 3-O-glucosyltransferase activity withquercetin and cyanidin as substrates in Vitis labrusca and Vitis vinifera tissues.Flavonoid 3-O-glucosyltransferase activity in V. labrusca and V. vinifera mesocarp(A) and exocarp (B) tissues with quercetin (6) and cyanidin (3) as substrates. (C)Flavonoid 3-O-glucosyltransferase activity in V. labrusca flower bud, leaf and flowertissue. Abbreviations: VV PN – Week 14 after flowering V. vinifera cv. Pinot Noirtissue; VV PND – Week 14 after flowering Vitis vinifera cv. Pinot Noir droit tissue.Bars represent the mean of three independent trials ± SD.

D. Hall et al. / Phytochemistry 74 (2012) 90–99 95

VV = 30 lM) than for quercetin (6) (KM VL = 2.2 lM; KM VV = 15 lM)(Table 1); and both rVL3GT and rVV3GT have higher turnover num-bers with cyanidin (3) (kcat VL = 7.02 � 10�4 s�1; kcat VV = 47.6 �10�4 s�1) than with quercetin (6) (kcat VL = 1.13 � 10�4 s�1; kcat

VV = 0.99 � 10�4 s�1), although the rVV3GT is able to glucosylatecyanidin (3) at a faster rate than is the rVL3GT. The calculated cat-alytic efficiencies for the rVV3GT (kcat KM

�1 = 159 M�1 s�1) and therVL3GT (kcat KM

�1 = 146 M�1 s�1) with cyanidin (3) as a substrateare almost identical, whereas with quercetin, the kcat KM

�1 valuesindicated that quercetin (6) is a better substrate for rVL3GT(kcat KM

�1 = 52 M�1 s�1) than for rVV3GT (kcat KM�1 = 6.6 M�1 s�1)

(Table 1). In general, calculation of the kinetic parameters indicates

that both enzymes preferentially glucosylate cyanidin (3) com-pared to quercetin (6).

3.3. VL3GT gene expression profile is consistent with a role inanthocyanin modification

Vl3gt gene expression was monitored in Concord berry exocarpand mesocarp tissues throughout grape berry development (Fig. 3).VL3GT transcript was detected post-veraison in V. labrusca and V.vinifera exocarp tissues; at low levels in V. labrusca leaf, flowerbud and flower tissue; but was not detected in V. labrusca meso-carp tissue or in the exocarp of pre-veraison berries (Fig. 3). Theseresults are consistent with the well-documented expression ofVV3GT in post-veraison V. vinifera berry exocarp tissues from red(but not white) cultivars, concurrent with the biosynthesis andaccumulation of anthocyanins in these tissues (Boss et al.,1996a,b; Kobayashi et al., 2001).

3.4. Crude protein extracts from grape berry exocarp and mesocarpdifferentially glucosylate cyanidin and quercetin throughoutdevelopment

Protein was extracted from post-veraison V. vinifera and severalstages of pre- and post-veraison V. labrusca mesocarp and exocarptissue. The extracts were desalted and assayed for 3-O-glucosyla-tion of cyanidin (3) and quercetin (6). As was previously reportedfor V. vinifera, enzyme assays of crude desalted extracts were com-plicated by the presence of several endogenous enzymes capable ofglucosylating flavonols in vitro (Ford et al., 1998). In V. labrusca,glucosylation of quercetin (6) at the 3-O-position was detected inberry mesocarp, exocarp, leaf, flower bud and flower tissues, whileanthocyanidin 3GT activity was restricted to grape exocarp. Theseresults: (a) suggest that the quercetin (6) 3GT activity in the berrymesocarp, flower bud, leaf and flower is catalyzed by at least oneseparate enzyme that preferentially glucosylates flavonols; and(b) are consistent with the lack of anthocyanins and the accumula-tion of several 3-O-glucosylated flavonols in these tissues (Mooreand Giannasi, 1994; Park and Cha, 2003). Recent studies (Onoet al., 2010) in V. vinifera described a novel bifunctional UDP-glucose/UDP-galactose-flavonol 3-O-GT with 55% amino acid se-quence identity to VLRS3GT that may represent one of the otherenzymes responsible for glucosylation of Quercetin in grape tis-sues. Cyanidin (3) 3-O-glucosylation was detected at low levelsin post-veraison mesocarp tissue, and at 10-fold higher levels inpost-veraison exocarp tissue, consistent with the gene expressionprofile of VL3GT (Fig. 3) and the accumulation of anthocyanins inberry exocarp but not mesocarp tissue (Boss et al., 1996a,b; Kobay-ashi et al., 2001).

3.5. The VL3GT belongs to the same clade as other members of the 3GTfamily

Phylogenetic analysis places VL3GT in a clade with otherenzymes which transfer sugars from an activated donor to the 3-O-position of flavonoids (Fig. 5). The recent crystal structure ofrVV3GT has allowed the identification of several residues whichare important for catalytic activity and both acceptor and donorsubstrate binding (Fig. 1; Offen et al., 2006). Protein sequencealignment of rVL3GT with rVV3GT reveals the conservation ofthese essential residues, in both sequences (Fig. 1), which is consis-tent with their homologous function in glucosylating the 3-O-posi-tion of flavonoids.

In addition to glucosyltransferases, this clade also contains theArabidopsis thaliana flavonoid 3-O-rhamnosyltransferase (Joneset al., 2003) and the Vigna mungo and Petunia � hybrida flavonoid3-O-galactosyltransferases (Mato et al., 1998; Miller et al., 1999)

At76E12

Vm3GaT

Md3GT

Vv3GTVl3GT

Ph3GaT

Ph3GT

Sm3GT

Gt3GTFi3GT

Dc3GTAt3RT

At3GT

BpGlcATAt

72B1 At72E2

Me3GT

At71C1Db6GTNTGT1

Gt3’GTDb5GT

Sb7GT

At7GT

At73C6

Ac73G1VaABAGT Gm

Isoflv7GT

At5GTPf

5GTMp7GTNt

Gt2Ph5GT

VlOGT3

Ih5GT

VlOGT2VlOGT1TrVlOGT4

At74F1

VlRSGTCuLimGT

At84A1

Ip2’’GTPhRT

At80A2 PvHRA25

At89B1Rh5GT

OsRF5

Hv3GTZm3GT

OsRUGT5OsIsoflv7GT

3GT family

Fig. 5. Neighbour joining phylogenetic tree of several plant glycosyltransferases including flavonoid 3-O-glycosyltransferases. Arrows indicate genes which glucosylate the 3-O-position of flavonols and anthocyanidins which are not members of the 3GT gene family. Abbreviations: Vm 3GaT – Vigna mungo flavonoid 3-O-galactosyltransferase(AB009370); Ph 3GaT – Petunia hybrida flavonoid 3-O-galactosyltransferase (AFI65148); Gt 3GT – Gentiana triflora flavonoid 3-O-glucosyltransferase (D85186); Fi 3GT –Forsythia intermedia flavonoid 3-O-glucosyltransferase (AFI27218); Ph 3GT – Petunia hybrida flavonoid 3-O-glucosyltransferase (AB027454); Sm 3GT – Solanum melongenaflavonoid 3-O-glucosyltransferase (X77369); At 3GT – Arabidopsis thaliana UGT78D2 UDP-glucoseflavonoid 3-O-glucosyltransferase; Dc 3GT – Dianthus caryophyllus flavonoid3-O-glucosyltransferase (ABI91247); At 3RT – Arabidopsis thaliana UGT 78Dl UDP-rhamnose flavonol 3-O-glycosyltransferase; At 3GT – Arabidopsis thaliana UGT 78D2UDP-glucose flavonol 3-O-glucosyltransferase; Md 3GT – Malus domesticus flavonoid 3-O-glucosyltransferase (AY663784); Vv 3GT – V. vinifera flavonoid 3-O-glucosyltransferase(AF000372); VL3GT – V. labrusca flavonoid 3-O-glucosyltransferase (EF630356); BpGlcAT – Bellis perennis UDP-glucuronic acid:anthocyanin glucuronosyltransferase(ABI90262); At72Bl – Arabidopsis thaliana UGT72B1 hydroxybenzoic acid O-glucosyltransferase; Me 3GT – Manihot esculenta flavonoid 3-O-glucosyltransferase (CAA54612);At72E2 – Arabidopsis thaliana UGT72E2 TCP glucosyltransferase (NMI26067); At71Cl – Arabidopsis thaliana UGT71Cl phenolic alcohol glucosyltransferase (NMI28529);DbBet6gt – Dorotheanthus bellidiformis betanidin 6-O-glucosyltransferase (AF374004); NtGtl – Nicotiana tabacum glucosyltransferase 1 (AB052557); At89B1 – Arabidopsisthaliana UGT89B1 hydroxybenzoic acid glucosyltransferase (NMI06048); Rh5GT – Rosa hybrida anthocyanidin 5-O-glucosyltransferase (AB201050); Os Rf5 – Oryza sativaflavonoid O-glucosyltransferase Rf5 (NM_00I050705); Hv 3GT – Hordeum vulgare flavonoid 3-O-glucosyltransferase (XI5694); Zm 3GT – Zea mays flavonoid 3-O-glucosyltransferase (XI3501); Os RUGT – Oryza sativa flavonoid O-glucosyltransferase RUGT (XM_463383); OsIsoFlv7GT – Oryza sativa isoflavonoid 7-O-glucosyltransferase(BAC80066); Ac73G1 – Allium cepa 7-O-glucosyltransferase; UGT73G1(AY262062); At73C6 – Arabidopsis thaliana UGT73C6 flavonoid 7-O-glucosyltransferase; At7gt73Bl –Arabidopsis thaliana UGT73Bl flavonol 7-O-glucosyltransferase; Gt30GT – Gentiana triflora anthocyanin 30-O-glucosyltransferase (AB076697); DbBet5gt – Dorotheanthusbellidiformis betanidin 5-O-glucosyltransferase (AF374004), Sb7gt – Scutellaria baicalensis flavonoid 7-O-glucosyltransferase (AB031274); GmIsoflv7gt – Glycine maxisoflavonoid glucosyltransferase (DQ278439); VaABAgt – Vigna angularis abscisic acid glucosyltransferase (AB065190); At5GT – Arabidopsis thaliana UGT75Cl anthocyanin 5-O-glucosyltransferase; Pf5GT – Perilla frutescens anthocyanin 5-O-glucosyltransferase (ABOI3596); Mp7gt – Maclura pomifera flavonoid 7-O-glucosyltransferase (DQ985179);Vh5GT – Verbena � hybrida anthocyanin 5-O-glucosyltransferase (AbOI3598). NtGt2 – Nicotiana tabacum flavonol 7-O-glucosyltransferase (AB072919); Ph5gt – Petunia xhybrida anthocyanin 5-O-glucosyltransferase (AB027455); VLOGT3 – V. labrusca O-glucosyltransferase 3 (EF533706); Ih5gt – Iris hollandica anthocyanin 5-O-glucosyltrans-ferase (AB113664); trVLOGT4 – truncated V. labrusca O-glucosyltransferase 4 (EF533707); VLOGTI – V. labrusca O-glucosyltransferase 1 (EF533704); VLOGT2 – V. labrusca O-glucosyltransferase2 (EF533705); At84Al – Arabidopsis thaliana UGT74Fl salicylicacid glucosyltransferase (BTOI5796); At84Al – Arabidopsis thaliana UGT84Al hydroxycin-namic acid glucosyltransferase; CuLimgt – Citrus unshiu liminoid glucosyltransferase (AB033758); VLRSGT – V. labrusca resveratrol/hydroxycinnamic acid O-glucosyltrans-ferase (DQ832169); Ip200GT – Ipomoea purpurea anthocyanidin 3-O-glucoside-200-O-glucosyltransferase (Ab192315); PhRT – Petunia x hybrida anthocyanidin 3-O-glucoside-600-O-rhamnosyltransferase (CAA50377); At80A2 – Arabidopsis thaliana UGT80A2 sterol glucosyltransferase; PvHRA25 – Phaseolus vulgaris glucosyltransferase HRA25(AF303396); At76E12 – Arabidopsis thaliana UGT76E12 flavonol O-glucosyltransferase.

96 D. Hall et al. / Phytochemistry 74 (2012) 90–99

(Fig. 5). Recently, it was demonstrated that a single point mutation(G375H) at the C-terminus of the PSPG consensus sequence chan-ged the substrate specificity of a flavonoid 3-O-galactosyltransfer-ase, allowing it to also glucosylate the 3-O-position of flavonoids(Kubo et al., 2004) and site-directed mutagenesis of an isoflavonoid7-O-glucosyltransferase changed the activity and kinetic parame-ters of this protein towards its endogenous substrate (Noguchiet al., 2007). Similarly, domain swapping of the PSPG box fromthe curcumin O-glucosyltransferase from Catharanthus roseus withthe corresponding sequence from the Nicotiana tabacum NTGT1beliminated curcumin glucosyltransferase activity, which was

restored by site-directed mutagenesis of a single amino acid(R377C) (Masada et al., 2007); suggesting that non-conserved ami-no acid residues within the PSPG box are important for acceptorsubstrate specificity, in addition to the well-documented role ofconserved residues in this domain for donor substrate specificity.The necessity of single amino acid residues to establish the bio-chemical function of some GTs and the importance of these resi-dues for donor and acceptor substrate specificity are consistentwith the distinct functional properties of rVL3GT and rVV3GTdespite their 96% amino acid sequence identity. This should beinvestigated in future studies by modeling based on the crystal

D. Hall et al. / Phytochemistry 74 (2012) 90–99 97

structure of rVV3GT to guide site specific mutagenesis of one ormore of the 14 unique amino acid residues (Fig. 1) of either proteinand enzyme assay of the mutated proteins to determine changes insubstrate specificities.

4. Conclusion

Glucosylation of the 3-O-position of anthocyanidins is essentialfor their storage and the stability of their spectral characteristics(Prior and Wu, 2006). The molecular cloning and in vitro biochem-ical characterization of this enzyme from Concord grape berryshows its strict regiospecificity for the 3-O-position of anthocyani-dins and flavonols and its use of UDP-glucose, but not UDP-galact-ose as a donor substrate. Kinetic analysis reveals that rVL3GTpreferentially glucosylates anthocyanidins in vitro; and RT-PCRanalysis shows that VL3GT expression coincides with cyanidin3GT enzyme activity (but not quercetin 3GT activity) and the accu-mulation of anthocyanins in post-veraison Concord grape exocarptissues. A comparison of the in vitro biochemical parameters ob-tained with rVL3GT reveals distinct biochemical properties as com-pared the highly identical and previously characterized rVV3GT.

5. Experimental

5.1. Plant material

V. labrusca cv. Concord grape berries were harvested weeklyfrom June until October, 2003 to 2006 at the 3rd St. site of theG&H Wiley vineyard (St. Catharines, ON). V. vinifera cv. Pinot Noirand cv. Pinot Noir Droit berries were harvested from the Châteaudes Charmes Vineyard (Niagara on the Lake, ON) in September of2004. All stages of grape berry maturity are reported as weeks afterflowering (AF), and physiological parameters of grape berry devel-opment were determined as reported in Hall and De Luca, 2007.Concord pre-bloom flower buds were harvested June 12, 2006,grape flowers were harvested June 16, 2006 and grapevine leavesand stems were harvested June 21, 2006. Grapes were thawedbriefly, the exocarp and mesocarp were separated by hand, andthe seeds were removed. The exocarp was scraped to remove resid-ual mesocarp tissue and was quick frozen in liquid N2. All plantmaterial was stored at �80 �C for future use.

5.2. Chemicals

All chemicals were purchased from Sigma (http://www.sigmaaldrich.com) (Oakville, ON, Canada) or Indofine Chemical(http://www.indofinechemical.com) (Hillsborough, NJ, USA). Gen-erally, 2 mM stocks of all acceptor substrates were prepared inMeOH, while quercetin (6) and kaempferol (8) were first dilutedin DMSO then diluted to 2 mM final concentration with MeOH(1% final DMSO concentration). Stock solutions (180 mM) ofdiphosphoglucose (UDPG) and 50-diphosphogalactose (UDPGal)were prepared in H2O.

5.3. Extraction of crude protein and GT enzyme activity profiling ingrape tissue throughout development

Concord [pre-bloom flower bud (1 g), leaf (0.5 g), flower (1 g),week 1 AF whole berry (1 g), as well as week 3, 6, 8, 10, 12, 14,and 16 AF berries separated into exocarp (1 g) and mesocarp(2 g)] and V. vinifera cv. Pinot noir and cv. Pinot Noir Droit [week14 AF berries separated into exocarp (1 g) and mesocarp (2 g)] tis-sues were each ground to a fine powder in a mortar and pestle withliq. N2, and were extracted in 4 ml of grape extraction buffer [4 ml,500 mM Tris–HCl, pH 8.0, 0.1% b-mercaptoethanol, 1% polyvinyl-

pyrrolidone-10, 5 mM sodium metabisulfite (Na2S2O5), 10% glyc-erol (v/v), 1 mM phenylmethyl sulfonyl fluoride], and 10%insoluble polyvinylpyrrolidine. The extracts were centrifuged at21,000g for 10 min at 4 �C, and supernatants were desalted on aPD-10 Sephadex G-25 column (GE Healthcare; http://www.gehealthcare.com/caen), pre-equilibrated in desalting buffer(100 mM Tris–HCl; pH 8.0 + 0.1% b-mercaptoethanol). The amountof protein in each extract was determined using a protein assay kit(Bio-Rad Laboratories, http://www.bio-rad.com). Enzyme assayscontained 100 ll of protein with 16 lM quercetin (6) or 6 lMcyanidin (3) and 9 mM UDPG in a final volume of 125 ll. The as-says were incubated for 3 min (cyanidin (3)) or 5 min (quercetin(6)) at 30 �C, and were stopped with the addition of equal volume0.1% (0.12 N) HCl in MeOH or MeOH respectively.

5.4. HPLC analysis of enzyme assays

Prior to HPLC, reaction products were centrifuged at 21,000g for10 min at 4 �C, and supernatants were filtered through a 0.45 lMwhite nylon filter (Millipore, http://www.Millipore.com) for analy-sis by reversed phase HPLC on an Inertsil ODS-3 (GL Sciences,http://www.gls.co.jp/index-e.html) C18 column (4 � 250 mm)equipped with a 3 � 4 mm guard column (Phenomenex, http://www.phenomenex.com) using a method modified from thatdescribed by Ali and Strommer (2003). Between 1 and 100 lL ofeach reaction product was injected onto the column, and productswere eluted (1 ml min�1) with solvent A (5% HCO2H in H2O) andsolvent B (33:60:70 MeOH:CH3CN:H2O) using the following gradi-ent: 0–9 min, 10–47% B; 9–12 min, 47–85% B; 12–13 min, 85–100%B; 13–21 min, 100% B; 21–24 min, 100–50% B; 24–26 min, 50% B;26–27 min, 50–10% B; 27–37 min, 10% B. Assays with anthocyani-din substrates were monitored at 520 nm while assays withkaempferol (8), isorhamnetin (5), quercetin (6) and dihydroquerce-tin (7) were monitored at 365, 370, 370 and 290 nm respectively,using a 2996 photodiode array detector (Waters, Milford MA,USA; www.waters.com). Specific activities are reported aspicomoles of glucoside produced in 1 s by 1 mg of total protein.

5.5. RNA extraction

RNA was extracted from all grape tissues using the Plant RNAExtraction reagent (Invitrogen, http://www.invitrogen.com/) as de-scribed previously (Hall and De Luca, 2007) with the exceptionsthat 1 ll of linear acrylamide (5 lg/ll) was added with to theRNA as a co-precipitant and the extracted RNA was resuspendedon ice. RNA (0.5–1 lg) was reverse transcribed using either Super-script III reverse transcriptase (Invitrogen) or the Takara RNA PCRkit version 3.0 (Fisher Scientific, http://www.fishersci.ca/) as perthe manufacturer’s instructions, and was diluted with 5 volumesof water.

5.6. Molecular cloning of VL3GT

The previously characterized full-length V. vinifera (Vv) 3GT(AF000372; Ford et al., 1998) was used to design forward primer1 (F1-50ATGTCTCAAACCACCACCAAC30) and reverse primer 1(R1-50CTAGACATCCTTTGGTTTTGAC30). PCR with F1, F2 and 1 llof Concord week 12 AF exocarp-specific cDNA amplified a productof approximately 1400 bp, which was subcloned into a TA-vectorfor sequencing. The vector containing the desired sequence wasisolated, diluted 1000-fold and used as a template with forwardprimer 2 (F2-50CGAAGAATTCATGTCTCAAACCACCACC 30) and re-verse primer 2 (R2-50AACCAAAGGATGTCTAGCTCGAGAACG 30) tointroduce suitable restriction sites for cloning into the pGEX 4T-1GST fusion expression vector (GE Healthcare).

98 D. Hall et al. / Phytochemistry 74 (2012) 90–99

The vector containing the full-length gene-GST fusion (VL3GT)was transformed by electroporation into Escherichia coli (E. coli)DE3 pLys S cells (which express the groES–groEL chaperone pro-tein; CHAP2; Fisher Scientific), and the recombinant protein wasexpressed. Briefly, a 3 ml culture was inoculated in 2� yeast tryp-tone medium containing 50 lg ml�1 ampicillin and 25 lg ml�1

chloramphenicol (2YTAC) and grown to saturation at 37 �C. Fivehundred microliter of saturated culture was inoculated to 50 ml2YTAC culture and was induced to express the chaperone proteinswith L-arabinose. The culture was grown to an OD of 1.0 at 37 �C,was induced for recombinant GT protein expression with 0.1 mMof isopropyl-1-thio-b-D-galactoside (IPTG) and was grown for20–24 h at 16–18 �C. The cell pellets containing the recombinantVL3GT (rVL3GT) fusion proteins were centrifuged for 5 min at2800g and were stored at �20 �C.

Recombinant protein was extracted from bacterial cell pelletsand purified by Glutathione Sepharose 4B (GE Healthcare) affinitychromatography following the manufacturer’s instructions. Afterwashing cell pellets once with 2.5 ml of 1� PBS buffer containing0.1% b-mercaptoethanol, they were resuspended in the same buffercontaining lysozyme (1 mg ml�1 final) and were incubated for20 min at room temperature before sonication and centrifugationat 2800g at 4 �C for 10 min to remove cellular debris. The superna-tant was then incubated with glutathione Sepharose 4B (GEHealthcare) for 45 min at room temperature with gentle mixing.The protein was purified batchwise according to the manufac-turer’s instructions, except that bound protein was eluted with20 mM glutathione (50 mM Tris, pH 8.0, 0.1% b-mercaptoethanol).The eluted rVL3GT protein was desalted by MicroSpin G-25 col-umns (GE Healthcare) as per the manufacturer’s instructions andwas stored as a 50% glycerol stock at �20 �C.

5.7. Recombinant VL3GT enzyme activity assays

Enzyme activity assays were conducted using the purified GSTfusion protein since removal of the tag resulted in a rapid 100% lossin enzyme activity that prevented its use for kinetic analyses or forsubstrate specificity studies. The rVL3GT was assayed with 3.6 mMdonor substrate (UDPG or UDPGal) and 32 lM flavonol-related(quercetin (6), isorhamnetin (5), dihydroquercetin (7), kaempferol(8)) or 6 lM anthocyanidin (cyanidin (3), kuromanin, delphinidin(4), peonidin (2), malvidin (1)) acceptor substrates. Each assay con-tained 0.875 lg of rVL3GT for anthocyanidin substrates and1.75 lg rVL3GT for flavonol-related substrates in 100 mM Tris–HCl pH 8.0, 0.1% b-mercaptoethanol buffer to a final reaction vol-ume of 125 ll. All assays were repeated in at least triplicate andwere incubated and stopped as described above prior to processingand analysis by HPLC, with the exception that assays were incu-bated for 1–180 min to determine reaction linearity.

For the determination of kinetic parameters, all assays wereincubated for 3 min at 30 �C to a final reaction volume of 125 llwith 100 mM Tris–HCl, 0.1% b-mercaptoethanol. For acceptor sub-strates, each assay contained 9 mM UDPG with varying concentra-tions of quercetin (6) and cyanidin (3). UDPG kinetics weredetermined with 6 lM cyanidin (3) as the acceptor substrate.The reactions were stopped and analyzed as described above.

5.8. RT-PCR analysis of VL3GT gene expression

Vl3gt gene expression was monitored using the gene specificprimers: VL3GT – forward primer 3 (F3-50TGCAGGGCCTAACTCACTCT 30) and VL3GT reverse primer 3 (R3-50 GCAGTCGCCTTAGGTAG-CAC 30); and the gene specific primers for VLActin (Hall and De Luca,2007). VLActin and VL3GT gene expression was repeated in at leastduplicate, was quantified using Multigauge ver 3.0 (Fujifilm, Tokyo,Japan; www.fujifilm.ca), and the mean values were divided by the

mean actin gene expression value to obtain a relative value for geneexpression (% expression) in these tissues.

5.9. Sequence alignment and phylogenetic analysis

The protein sequence of VL3GT (EF630356) was aligned withthe full-length Vitis vinifera cv. Shiraz. flavonoid 3-O-glucosyltrans-ferase sequence (AAB81683 – Ford et al., 1998) using ClustalW 2(http://align.genome.jp/). Nucleotide sequence alignment andphylogenetic analysis of VL3GT with other glycosyltransferasesequences was analyzed using BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html) and the DNA distance neighbour-joiningphylogenetic analysis was visualized using Phylodraw (http://pearl.cs.pusan.ac.kr/phylodraw/).

Acknowledgements

This research was funded in part by a grant from the NaturalSciences and Engineering Research Council of Canada Discoveryand a Tier 1 Canada Research Chair to V.D.L., as well as an NSERCpost-graduate fellowship to D.H. We thank G&H Wiley Ltd. (St.Catharine’s, Ontario, Canada) for allowing us to collect Concordgrapes during the growing season as well as Dr. Helen Fisher ofthe University of Guelph Grape Research Station (Vineland, Ontar-io, Canada), and Dr. John Paroschy (Chateau des Charmes Winery,Niagara on the Lake, Ontario, Canada) for the supply of grape sam-ples. We also thank Dr. Jihong Wang, University Health Network,Toronto, Canada, for the picture in the graphical abstract of Con-cord grape that were harvested at different stages of development.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.phytochem.2011.10.007.

References

Ali, A., Strommer, J., 2003. A simple extraction and chromatographic system for thesimultaneous analysis of anthocyanins and stilbenes of Vitis species. J. Agric.Food Chem. 51, 7246–7251.

Bowles, D., 2002. A multigene family of glycosyltransferases in a model plant,Arabidopsis thaliana. Biochem. Soc. Trans. 30, 301–306.

Boss, P.K., Davies, C., Robinson, S.P., 1996a. Analysis of the expression ofanthocyanin pathway genes in developing Vitis vinifera L. Cv. Shiraz grapeberries and the implications for pathway regulation. Plant Physiol. 111, 1059–1066.

Boss, P.K., Davies, C., Robinson, S.P., 1996b. Expression of anthocyanin biosynthesispathway genes in red and white grapes. Plant Mol. Biol. 32, 565–569.

De Luca, V., 2011. Plant systems/wines. In: Murray Moo-Young (Ed.),Comprehensive Biotechnology, second ed. vol. 4. Elsevier, pp. 241–255.

Ford, C.M., Boss, P.K., Hoj, P.B., 1998. Cloning and characterization of Vitis viniferaUDP-Glucose:flavonoid 3-O-glucosyltransferase, a homologue of the enzymeencoded by the maize Bronze-1 locus that may primarily serve to glucosylateanthocyanidins in vivo. J. Biol. Chem. 273, 9224–9233.

Harborne, J.B., 1967. Comparative Biochemistry of the Flavonoids. Academic Press,London, UK.

Hugueney, P., Provenzano, S., Verriès, C., Ferrandino, A., Meudec, E., Batelli, G.,Merdinoglu, D., Cheynier, V., Schubert, A., Ageorges, A., 2009. A novel cation-dependent O-methyltransferase involved in anthocyanin methylation ingrapevine. Plant Physiol. 150, 2057–2070.

Gong, Z., Yamazaki, M., Sugiyama, M., Tanaka, Y., Saito, K., 1997. Cloning andmolecular analysis of structural genes involved in anthocyanin biosynthesis andexpressed in a forma-specific manner in Perilla frutescens. Plant Mol. Biol. 35,915–927.

Jones, P., Vogt, T., 2001. Glycosyltransferases in secondary plant metabolism:tranquilizers and stimulant controllers. Planta 213, 164–174.

Jones, P., Mebner, B., Nakajima, J.I., Schäffner, A.R., Saito, K., 2003. UGT73C6 andUGT78D1, glycosyltransferases involved in flavonol glycoside biosynthesis inArabidopsis thaliana. J. Biol. Chem. 278, 43910–43918.

Hall, D., De Luca, V., 2007. Mesocarp localization of a bi-functional resveratrol/hydroxycinnamic acid glucosyltransferase of Concord grape (Vitis labrusca).Plant J. 49, 579–591.

D. Hall et al. / Phytochemistry 74 (2012) 90–99 99

Hall, D. Kim, K.H., De Luca, V., 2011. Molecular cloning and biochemicalcharacterization of three Concord grape (Vitis labrusca) flavonol 7-O-glucosyltransferases. Planta. doi: 10.1007/s00425-011-1474-0.

Kim, J.H., Shin, K.H., Ko, J.H., Ahn, J.H., 2006. Glucosylation of flavonols by Escherichiacoli expressing glucosyltransferases from rice (Oryza sativa). J. Biosci. Bioeng.102, 135–137.

Kobayashi, S., Ishimaru, M., Ding, C.K., Yakushiju, H., Goto, N., 2001. Comparison ofUDP-glucose: flavonoid 3-O-glucosyltransferase (UFGT) gene sequencesbetween white grapes (Vitis vinifera) and their sports with red skin. Plant Sci.160, 543–550.

Kobayashi, S., Ishimaru, M., Hiraoka, K., Honda, C., 2002. Myb-related genes of theKyoho grape (Vitis labruscana) regulate anthocyanin biosynthesis. Planta 215,924–933.

Koes, R.E., Quattrocchio, F., Mol, J.N.M., et al., 1994. The flavonoid biosyntheticpathway in plants: function and evolution. BioEssays 16, 123–132.

Koes, R., Verweij, W., Quattrocchio, F., 2005. Flavonoids: a colorful model for theregulation and evolution of biochemical pathways. Trends Plant Sci. 10, 236–242.

Kramer, C.M., Prata, R.T.N., Willits, M.G., De Luca, V., Steffens, J.C., Graser, G., 2003.Cloning and regiospecificity studies of two flavonoid glucosyltransferases fromAllium cepa. Phytochemistry 64, 1069–1076.

Kubo, A., Arai, Y., Nagashima, S., Yahikawa, T., 2004. Alteration of sugar donorspecificities of plant glycosyltransferases by a single point mutation. Arch.Biochem. Biophys. 429, 198–203.

Li, Y., Baldauf, S., Lim, E.K., Bowles, D.J., 2001. Phylogenetic analysis of the UDP-glycosyltransferase multigene family of Arabidopsis thaliana. J. Biol. Chem. 276,4338–4343.

Lim, E.K., Baldauf, S., Li, Y., Elias, L., Worrall, D., Spencer, S.P., Jackson, R.G., Taguchi,G., Ross, J., Bowles, D.J., 2003. Evolution of substrate recognition across amultigene family of glycosyltransferases in Arabidopsis. Glycobiology 13, 139–145.

Masada, S., Terasaka, K., Mizukami, H., 2007. A single amino acid in the PSPG-boxplays and important role in the catalytic function of CaUGT2 (Curcuminglucosyltransferase), a Group D Family 1 glucosyltransferase from Catharanthusroseus. FEBS Lett. 581, 2605–2610.

Mato, M., Ozeki, Y., Itoh, Y., Higeta, D., Yoshitama, K., Teramoto, S., Aida, R., Ishikura,N., Shibata, M., 1998. Isolation and characterization of a cDNA clone of UDP-galactose: flavonoid 3-O-galactosyltransferase (UF3GaT) expressed in Vignamungo seedlings. Plant Cell Physiol. 39, 1145–1155.

Mazza, G., Miniati, E., 1993. Grapes. In: Anthocyanins in Fruits, Vegetables andGrains. CRC Press, Boca Raton, pp. 149–199 (Chapter 6).

Miller, K.D., Guyon, V., Evans, J.N.S., Shuttleworth, W.A., Taylor, L.P., 1999.Purification, cloning, and heterologous expression of a catalytically efficientflavonol 3-O-galactosyltransferase expressed in the male gametophyte ofPetunia hybrida. J. Biol. Chem. 274, 34011–34019.

Mol, J., Grotewold, E., Koes, R., 1998. How genes paint flowers and seeds. TrendsPlant Sci. 3, 212–216.

Monagas, M., Bartolome, B., Gomez-Cordoves, C., 2005. Updated knowledge aboutthe presence of phenolic compounds in wine. Crit. Rev. Food Sci. Nutr. 45, 85–118.

Moore, M.O., Giannasi, D.E., 1994. Foliar flavonoids of eastern North American Vitis(Vitaceae) north of Mexico. Plant Syst. Evol. 193, 21–36.

Noguchi, A., Saito, A., Homma, Y., Nakao, M., Sasaki, N., Nishino, T., Takahashi, S.,Nakayama, T., 2007. A UDP-glucose: isoflavone 7-O-glucosyltransferase fromthe roots of soybean (Glycine max) seedlings. J. Biol. Chem. 282, 23581–23590.

Offen, W., Marinez-Fleites, C., Yang, M., Lim, E., Davis, B.G., Tarling, C.A., Ford, C.M.,Bowles, D.J., Davies, G.J., 2006. Structure of a flavonoid glucosyltransferasereveals the basis for plant natural product modification. EMBO J. 25, 1396–1405.

Ono, E., Homma, Y., Horikawa, M., Kunikane-Doi, S., Imai, H., Takahashi, S., Kawai, Y.,Ishiguro, M., Fukui, Y., Nakayama, T., 2010. Functional differentiation of theglycosyltransferases that contribute to the chemical diversity of bioactiveflavonol glycosides in grapevines (Vitis vinifera). Plant Cell 22, 2556–2571.

Park, H.J., Cha, H.C., 2003. Flavonoids from leaves and exocarps of the grape Kyoho.Korean J. Biol. Sci. 7, 327–330.

Prior, R.L., Wu, X., 2006. Anthocyanins: structural characteristics that result inunique metabolic patterns and biological activities. Free Rad. Res. 40, 1014–1028.

Sefton, M.A., Francis, I.L., Williams, P.J., 1993. The volatile composition ofChardonnay juices: a study by flavor precursor analysis. Am. J. Enol. Vitic. 44,359–370.

Sefton, M.A., Francis, I.L., Williams, P.J., 1994. Free and bound volatile secondarymetabolites of Vitis vinifera grape cv. Sauvignon Blanc. J. Food Sci. 59, 142–147.

Tanaka, Y., Yonekura, K., Fukuchi-Mizutani, M., Fukui, Y., Fujiwara, H., Ashikari, T.,Kusumi, T., 1996. Molecular and biochemical characterization of threeanthocyanin synthetic enzymes from Gentiana triflora. Plant Cell Physiol. 37,711–716.

Vogt, T., Zimmermann, E., Grimm, R., Meyer, M., Strack, D., 1997. Are thecharacteristics of betanidin glucosyltransferases from cell-suspension culturesof Dorotheanthus bellidiformis indicative of their phylogenetic relationship withflavonoid glucosyltransferase? Planta 203, 349–361.

Vogt, T., Jones, P., 2000. Glycosyltransferases in plant natural product synthesis:characterization of a supergene family. Trends Plant Sci. 5, 380–388.

Wink, M., 2003. Evolution of secondary metabolites from an ecological andmolecular phylogenetic perspective. Phytochemistry 64, 3–19.

Winkel-Shirley, B., 2001. Flavonoid biosynthesis. A colorful model for genetics,biochemistry, cell biology, and biotechnology. Plant Physiol. 126, 485–493.

Wu, X., Prior, R.L., 2005. Systematic identification and characterization ofanthocyanins by HPLC–ESI-MS/MS in common foods in the United States:fruits and berries. J. Agric. Food Chem. 53, 2589–2599.

Wu, X., Beecher, G.G., Holden, J.M., Haytowitz, D.B., Gebhardt, S.E., Prior, R.L., 2006.Concentrations of anthocyanins in common foods in the United States andestimation of normal consumption. J. Agric. Food Chem. 54, 4069–4075.

Yamazaki, M., Yamagishi, E., Gong, Z., Fukuchi-Mizutani, M., Fukui, Y., Tanaka, Y.,Kusumi, T., Yamaguchi, M., Saito, K., 2002. Two flavonoid glucosyltransferasesfrom Petunia hybrida: molecular cloning, biochemical properties anddevelopmentally regulated expression. Plant Mol. Biol. 48, 401–411.

Yoshihara, N., Imayama, T., Fukuchi-Mizutani, M., Okuhara, H., Tanaka, Y., Ino, I.,Yabuya, T., 2005. CDNA cloning and characterization of UDP-glucose:anthocyanidin 3-O-glucosyltransferase in Iris hollandica. Plant Sci. 169, 496–501.