Cinnamate Metabolism in Ripening Fruit. Characterization of a UDP-Glucose:Cinnamate Glucosyltransferase from Strawberry

Abstract

Strawberry (Fragaria × ananassa) fruit accumulate (hydroxy)cinnamoyl glucose (Glc) esters, which may serve as the biogenetic precursors of diverse secondary metabolites, such as the flavor constituents methyl cinnamate and ethyl cinnamate. Here, we report on the isolation of a cDNA encoding a UDP-Glc:cinnamate glucosyltransferase (Fragaria × ananassa glucosyltransferase 2 [FaGT2]) from ripe strawberry cv Elsanta that catalyzes the formation of 1-O-acyl-Glc esters of cinnamic acid, benzoic acid, and their derivatives in vitro. Quantitative real-time PCR analysis indicated that FaGT2 transcripts accumulate to high levels during strawberry fruit ripening and to lower levels in flowers. The levels in fruits positively correlated with the in planta concentration of cinnamoyl, p-coumaroyl, and caffeoyl Glc. In the leaf, high amounts of Glc esters were detected, but FaGT2 mRNA was not observed. The expression of FaGT2 is negatively regulated by auxin, induced by oxidative stress, and by hydroxycinnamic acids. Although FaGT2 glucosylates a number of aromatic acids in vitro, quantitative analysis in transgenic lines containing an antisense construct of FaGT2 under the control of the constitutive 35S cauliflower mosaic virus promoter demonstrated that the enzyme is only involved in the formation of cinnamoyl Glc and p-coumaroyl Glc during ripening.

Plants synthesize large amounts of phenylpropanoid acids, mainly hydroxycinnamic acids, which are often found in conjugated forms, such as glycosides or Glc esters. These conjugates have been identified in numerous plants (Molgaard and Ravn, 1988; Herrmann, 1989). Glucosides may be bioactive by themselves as defense compounds or they may be storage forms (Dixon, 2001). On the other hand, 1-O-acyl Glc esters may serve as activated intermediates analogous to CoA thioesters in plant secondary metabolism (Villegas and Kojima, 1986; Lehfeldt et al., 2000).

Glycosylation of hydroxycinnamic acids to form both glycosides and Glc esters is catalyzed by a group of enzymes called glucosyltransferases (GTs), which transfer the Glc residue from mostly UDP-activated Glc (Mock and Strack, 1993). Related GTs are known that glycosylate other compounds, such as flavonoids (Cheng et al., 1994), alkaloids (Moehs et al., 1997), terpenoids (Kita et al., 2000), cyanohydrins (Jones et al., 1999), thiohydroxymates (Reed et al., 1993), and plant hormones (Jackson et al., 2001). Many glycosyltransferases are able to glycosylate more than one aglycon, and they appear to recognize only the part of the molecule where glycosylation takes place (Hoesel, 1981). Glycosylation normally takes place in the cytosol, but Glc conjugates are found in the vacuole (Vogt and Jones, 2000).

Glycosides and Glc esters of hydroxycinnamic acids and the enzymes that catalyze their formation have been described from many species (Vogt and Jones, 2000; Lim et al., 2001, 2003). However, the Glc ester of cinnamic acid, which contains no substitutions in the benzene ring, has been described only in a few species (Latza et al., 1996), and no information is available regarding its formation or role in plants. In strawberries (Fragaria × ananassa), methyl and ethyl esters of cinnamate constitute important flavor compounds in the fruit (Schreier, 1980; Hirvi and Honkanen, 1982; Da Silva and Das Neves, 1999). Cinnamoyl Glc ester has also been found in strawberry fruit (Latza et al., 1996). It is possible that the methyl and ethyl esters of cinnamate are formed through a cinnamoyl Glc ester intermediate because these high-energy compounds have been observed to serve as precursors in similar transfer reactions (Latza and Berger, 1997). However, at present, no biochemical data are available concerning the synthesis of any of these compounds in strawberry fruit. Here we describe the identification and characterization of a strawberry glycosyltransferase that catalyzes the formation of cinnamoyl Glc ester and its possible role in the fruit.

RESULTS

Identification of Putative Glycosyltransferases from Strawberry Fruit

We identified four cDNAs among a set of 1,100 strawberry expressed sequence tags generated previously (Aharoni and O'Connell, 2002) as encoding putative glycosyltransferases by homology BLAST searches with published sequence information. Full-length cDNAs were obtained and named FaGT1 to 3 (Fragaria × ananassa glucosyltransferase 1 to 3) and FaRT1 (Fragaria × ananassa rhamnosyltransferase 1) based on the results of the similarity searches. FaGT1 showed the highest identity to several UDP-Glc:flavonoid 3-O-GTs from white grapes (Vitis vinifera; Ford et al., 1998; Kobayashi et al., 2001). FaGT3 showed the highest similarity to three tobacco (Nicotiana tabacum) GTs and one sweet honey leaf (Stevia rebaudiana) GT (Taguchi et al., 2001, 2003; Richman et al., 2005), and FaRT1 displayed homology to a petunia (Petunia hybrida) rhamnosyltransferase (Brugliera et al., 1994; Fig. 1

Figure 1.

Phylogenetic relationship of the four glycosyltransferases cloned from Fragaria (boxed) with 42 other sequences of glycosyltransferases, all of which have already been biochemically characterized. The dendrogram was created using the Clustal sequence alignment program of the Lasergene software package (DNASTAR). Lengths of lines indicate the relative distance between nodes. Indicated are the family and the GenBank accession number and, in the case of several Arabidopsis GTs with the same accession number, also the name of the UGT according to the superfamily nomenclature system (Mackenzie et al., 1997).

).

FaGT2 displayed the highest similarity to a functionally characterized limonoid GT from Satsuma mandarin (Citrus unshiu; GenBank accession no. AB033758; 68.3% identity; Kita et al., 2000), to several ester-forming Arabidopsis (Arabidopsis thaliana) GTs (GenBank accession nos. AB019232, Z97339; 49.6%–55.0% identity; Milkowski et al., 2000a; Sato et al., 2000), and to a rapeseed (Brassica napus) ester-forming GT (GenBank accession no. AF287143; 50.3% identity; Milkowski et al., 2000b; Figs. 1 and 2

Figure 2.

Sequence alignment of FaGT2 (Fragaria_AY663785_FaGT2), Satsuma mandarin GT (Citrus_AB033758), rapeseed GT (Brassica_AF287143), and three Arabidopsis (Arabidopsis_AB019232_UGT84A2, Arabidopsis_Z97339_UGT84A1, and Arabidopsis_Z97339_UGT84A3) deduced protein sequences. The ORF of FaGT2 consists of 1,668 bp encoding for a protein with 555 amino acids. The plant secondary product glycosyltransferase box, characteristic for GTs involved in natural product conjugation, is underlined.

). On the other hand, FaGT1, FaGT3, and FaRT1 showed the highest similarity to glucoside-forming GTs. The four strawberry glycosyltransferases showed sequence identity to each other ranging from 19.0% to 21.9% at the amino acid level. Because FaGT2 forms a separate clade with five GTs, four of which are known to form d-Glc esters from their various substrates, whereas the other strawberry GTs clustered with glucoside-forming GTs, we investigated FaGT2 further.

Enzymatic Activity of the Recombinant FaGT2 Enzyme

Full-length FaGT2 cDNA encodes a protein of 555 amino acids with an estimated molecular mass of 61.4 kD. The entire FaGT2 open reading frame (ORF) was used for generating the recombinant protein in Escherichia coli cells. SDS-PAGE confirmed the predicted 65-kD molecular mass of the FaGT2 fusion protein, consisting of a 61.4-kD FaGT2 protein and 3-kD N-terminal tag. Enzyme assays were performed with the purified recombinant protein, and substrates previously identified as glycosylated derivatives in strawberry fruit as well as others not reported (Wintoch et al., 1991). Of the 35 substrates tested, 15 were accepted as substrates by the recombinant enzyme, including benzoic acid, cinnamic acid, and their respective derivatives, whereas about 20 were not accepted, including several flavonoids (quercetin, kaempferol, pelargonidin), alcohols (benzyl alcohol, 2-phenylethanol), nonaromatic acids (2-methylbutyric acid, hexanoic acid, abscisic acid), aromatic acids (indoleacetic acid, naphthylacetic acid [NAA], o-coumaric acid, salicylic acid, 2-amino benzoic acid, gallic acid, gentisic acid, phenylacetic acid), phenol, and 4-hydroxy-2,5-dimethyl-3(2H)-furanone. Limonin showed only trace activity with the recombinant FaGT2 enzyme. Only compounds containing a carboxylic group and an aromatic ring structure were accepted as substrates (Fig. 3

Figure 3.

Chemical structures of some substrates and products of FaGT2. FaGT2 converts benzoic acid, cinnamic acid, and their respective derivatives to the corresponding Glc esters, using UDP-activated Glc. A hydroxyl group at position 2 in the benzoic acids has an inhibitory effect on the activity.

; Table I

Table I.

Michaelis-Menten kinetics (K_m and V_max values) of recombinant FaGT2

Assays were performed as described in “Materials and Methods.”

Substrate	Km	Vmax	KcatKm−1
	μm	nkat mg−1	s−1 μm−1
Cinnamic acid	356.9 ± 82.1	2.34 ± 0.53	0.42
p-Coumaric acid	603.5 ± 120.7	2.69 ± 0.54	0.29
Caffeic acid	707.7 ± 127.3	2.59 ± 0.46	0.24
Ferulic acid	358.5 ± 71.0	1.65 ± 0.31	0.30
5-Hydroxyferulic acid	315.7 ± 37.8	2.24 ± 0.33	0.05
Sinapic acid	300.2 ± 42.1	1.21 ± 0.17	0.26
Benzoic acid	502.6 ± 55.2	1.08 ± 0.12	0.14
3-Hydroxybenzoic acid	464.4 ± 55.7	1.05 ± 0.13	0.15
p-Hydroxybenzoic acid	642.4 ± 89.5	0.77 ± 0.11	0.07
Vanillic acid	515.3 ± 97.4	1.69 ± 0.30	0.21
3,4-Dimethoxycinnamic acid	108.0 ± 14.0	0.68 ± 0.09	0.41
Phenylpropionic acid	411.2 ± 45.2	0.17 ± 0.02	0.03
Phenylbutyric acid	431.0 ± 56.0	0.22 ± 0.03	0.03
3-Aminobenzoic acid	488.2 ± 61.1	0.75 ± 0.09	0.09
4-Aminobenzoic acid	437.1 ± 78.7	1.73 ± 0.32	0.25
UDP-Glc (with cinnamic acid)	80.8 ± 9.7	4.88 ± 0.57	3.89

). Substituents at the aromatic ring were tolerated as long as they were not located in the ortho position. Thus, 3- and 4-hydroxy as well as 3- and 4-aminobenzoic acid and p-coumaric acid were glucosylated, but 2-hydroxy and 2-amino benzoic acid and o-coumaric acid were not. FaGT2 shows a strict regioselectivity as already observed for other GTs (Lim et al., 2003). Protein extracts obtained from E. coli cells carrying the empty pRSETB vector showed no activity with the tested substrates. No product could be detected if UDP-Glc or the acceptor was omitted. The enzyme could be inhibited by heating the assay to 95°C for 5 min immediately after addition of the enzyme solution.

Kinetic Parameters of the Recombinant FaGT2 Enzyme

The kinetic parameters of the recombinant GT enzyme were obtained from hyperbolic Michaelis-Menten saturation curves for both the donor and acceptor substrates (Table I) under optimal conditions, which were determined to be pH 8 to 8.5 and 21°C. Product formation was linear between 10 to 30 min and saturated after 120 min. The apparent K_m and V_max for cinnamic acid were 357 μm and 2.34 nkat mg⁻¹, respectively, and for UDP-Glc, 80.8 μm and 4.88 nkat mg⁻¹, respectively. The kinetic data indicate that the enzyme shows the highest affinity for 3,4-dimethoxy cinnamic acid (K_m = 108 μm; V_max = 0.69 nkat mg⁻¹), and converts p-coumaric acid (K_m = 604 μm; V_max = 2.69 nkat mg⁻¹) the fastest (Table I). The addition of salts (chlorides) to buffered standard assays at concentrations ranging from 0.01 to 0.1 mm had only a slight effect on the glucosylation of cinnamic acid. The addition of Cu²⁺, Fe²⁺, or Zn²⁺ to 1 mm led to a loss of 30% to 50% of the activity, whereas the addition to 10 mm caused a loss of more than 90% in the case of Cu²⁺ and Zn²⁺, 60% in the case of Fe²⁺, and 30% in the case of Mn²⁺. No effect could be observed by adding 10 mm Mg²⁺ to the assay.

Elucidating the Structure of the Products Generated by the Recombinant FaGT2 Enzyme

Glycosyltransferases form either glycosides (O-acetales) or Glc esters (Vogt and Jones, 2000). Because FaGT2 acts on cinnamic acid carrying only a carboxyl function, it seemed likely that the enzyme transfers the Glc molecule to the carboxyl group of cinnamic and benzoic acids rather than to other functional groups. We confirmed that FaGT2 catalyzed the formation of 1-O-cinnamoyl-β-d-Glc by liquid chromatography (LC)-UV-electrospray ionization (ESI)-tandem mass spectrometry (MSⁿ) and comparison with an authentic standard. Similar fragmentation patterns were observed for the remaining products forming either the formic acid adduct or the anion of the Glc ester (Table II

Table II.

Retention times and fragmentation patterns of Glc ester formed by FaGT2 and their corresponding aromatic acids

Compound names, retention times under the LC conditions described in “Materials and Methods,” and the fragments with intensities in percent are indicated. Identity of the glucosylated compounds of plant extracts was confirmed by comparing the chromatographic properties and mass fragmentation patterns with the reference compounds synthesized by FaGT2.

Compound	Retention Time	MS	MS2	MS3
	min
Cinnamoyl-d-Glc	15.2	355, 100% [M+HCOO]-	309, 100% [M-H]-, 207, 70% 147, 60% [aglycon]-	147, 100%, [aglycon]-
p-Coumaroyl-d-Glc	12.4	325, 100% [M-H]-	163, 100%, [aglycon]-, 145, 90% 187, 85%	119, 100% [aglycon-COO]-
Caffeoyl-d-Glc	11.4	341, 100% [M-H]-	179, 100% [aglycon]-, 161, 30% 203, 15%	135, 100% [aglycon-COO]-
Feruloyl-d-Glc	12.8	355, 100% [M-H]-	193, 100% [aglycon]-, 217, 75% 175, 25%	149, 100% [aglycon-COO]-
Sinapoyl-d-Glc	12.9	385, 100% [M-H]-	223, 100% [aglycon]-, 247, 90% 205, 50%	205, 100% 179, 98%, [aglycon-COO]-, 126, 20%
Benzoyl-d-Glc	10.0	329, 100% [M+HCOO]-	121, 100% [aglycon]-
p-Hydroxy-benzoyl-d-Glc	10.2	299, 100% [M-H]-	137, 100% [aglycon]-, 179, 45% 239, 40%	93, 100% [aglycon-COO]-
Cinnamic acid	19.1	147, 100% [M-H]-	103, 100% [M-COO]-	–
p-Coumaric acid	15.2	163, 100% [M-H]-	119, 100% [M-COO]-	–
Caffeic acid	13.6	179, 100% [M-H]-	135, 100% [M-COO]-	–
Ferulic acid	15.7	193, 100% [M-H]-	149, 100% [M-COO]-, 134, 75% 178, 60%	–
Sinapic acid	15.6	223, 100% [M-H]-	163, 100% 208, 90% 179, 50% [M-COO]-	–
Benzoic acid	16.2	121, 100% [M-H]-	–	–
p-Hydroxy benzoic acid	12.2	137, 100% [M-H]-	93, 100% [M-COO]-	–

). The identities of the enzymatically formed products as esters were additionally confirmed by alkaline hydrolysis. Due to the high free energy of the hydrolysis 1-O-Glc esters of cinnamic and benzoic acids, derivatives are unstable in the presence of 1 n NaOH, unlike the O-glucosides of the same substrates. All products were easily hydrolyzed by 1 n NaOH. We also detected isomers of the Glc esters, already observed by Jackson et al. (2001) during their experiments using indoleacetic acid as a substrate for a Glc ester-forming enzyme.

Developmental Expression of FaGT2 and Its Correlation with the Levels of Free and Glucosidically Bound Cinnamic and Benzoic Acids and Derivatives

The spatial and temporal expression pattern of the FaGT2 gene was studied by the quantitative real-time PCR (qRT-PCR) approach. The amount of FaGT2 mRNA increased strongly along the fruit-ripening stages, with maximal expression value in fully ripe red fruit (Fig. 4A

Figure 4.

Developmental and spatial expression pattern of FaGT2 in different plant parts of strawberry cv Elsanta, analyzed by qRT-PCR and parallel analyses of metabolites. A, Relative expression in receptacles, vegetative tissues, and achenes. The increase in mRNA values is shown relative to the root Ct value, which is referred to as 1. Ct represents the cycle at which sample crosses threshold value. Error bars show the sd of three independent experiments. B, Quantification of free and glucosylated aromatic acids by LC-ESI-UV-MSⁿ (−, not detectable).

). Northern-blot analysis confirmed the results (data not shown). The expression pattern correlated well with the concentration of Glc esters in strawberry fruit (Fig. 4B). Levels of cinnamoyl-, p-coumaroyl-, caffeoyl-, and p-hydroxybenzoyl Glc increased during fruit ripening in the receptacle, indicating that they are ripening related. FaGT2 expression in achenes was very low compared to expression in the receptacles, and it did not increase during development. Expression of the FaGT2 gene was also observed in flowers and petioles, but FaGT2 mRNA was barely detectable in leaves and roots (Fig. 4A). Although leaves and flowers contain high levels of p-coumaroyl- and caffeoyl Glc esters, FaGT2 expression levels in these organs were much lower than those in ripe fruit. The amount of aromatic acids bound as Glc esters exceeded that of free acids by a factor of 100 in most tissues.

Induction of Expression

To determine whether the expression of the FaGT2 gene is regulated by auxin, we performed qRT-PCR analyses of FaGT2 mRNA of de-achened green fruit kept on the plant during the following 5 d and also of de-achened fruit treated with the auxin NAA. De-achened green fruit start ripening and become pink within 5 d due to anthocyanin formation (Aharoni et al., 2002). A clear increase in FaGT2 expression was detected in fruit after 5 d of removing the auxin-producing achenes (Fig 5A

Figure 5.

A, Effect of removing achenes and auxin treatment of green fruit on FaGT2 gene expression analyzed by qRT-PCR. The increase in FaGT2 mRNA in de-achened fruit and de-achened fruit treated with 1-NAA is shown relative to control green receptacle fruit Ct value (sampled at day 0, achenes still attached), which is referred to as 1. Ct represents the cycle at which sample crosses threshold value. Mean values ± sd of three independent experiments. B, Effect of oxidative stress caused by menadione on FaGT2 gene expression in fruits, analyzed by qRT-PCR. Relative expression in white fruit injected with menadione (black bars) in comparison with white fruit injected with water (control; white bars) is shown. The increase in mRNA values is shown relative to control white fruit Ct value, which is referred to as 1. Mean values ± sd of three independent experiments. C, Relative expression in strawberry cell cultures treated with menadione (black bars) in comparison with untreated controls (white bars). The increase in mRNA values is shown relative to control 0-h Ct value, which is referred to as 1. Mean values ± sd of three independent experiments. D, FaGT2 gene expression in strawberry cell culture treated with free acids, analyzed by qRT-PCR. The increase in mRNA values is shown relative to control at the 4- and 8-h Ct values, which are referred to as 1. Mean values ± sd of three independent experiments.

). However, these increases in FaGT2 transcript level were partially reversed in green de-achened fruit treated with NAA. Therefore, expression of the FaGT2 gene is negatively regulated by auxins.

Previous studies have demonstrated a potential relationship between oxidative stress and changes in gene expression during strawberry fruit ripening (Aharoni et al., 2002). To determine whether FaGT2 gene expression is under the influence of oxidative stress conditions, we injected free radical (reactive oxygen species)-generating compounds such as H₂O₂ and menadione into strawberry fruit or added the compounds to strawberry cell cultures and quantified the levels of FaGT2 mRNA. Expression of the FaGT2 gene was clearly induced by these oxidative stress treatments in both strawberry fruit and strawberry cell cultures (Fig. 5, B and C). These results are in agreement with those obtained for other strawberry ripening-related genes (Aharoni et al., 2002) and support the possibility that oxidative stress can play a role in the strawberry fruit-ripening process (Aharoni et al., 2002).

To examine whether free hydroxycinnamic acids whose Glc esters accumulate during ripening (Fig. 4) influence the levels of expression of FaGT2, we added free acids to strawberry cell cultures and quantified the levels of FaGT2 mRNA. All the free acids tested were able to induce FaGT2 expression (Fig. 5D). Benzoic acid, whose ester does not accumulate in the fruit, did not induce expression of FaGT2 (Fig. 5D).

Function of FaGT2 in Strawberry Fruit

To examine the function of FaGT in planta, we generated transgenic strawberry plants expressing the FaGT2 gene in antisense (AS) orientation under the control of the cauliflower mosaic virus 35S promoter. Eleven kanamycin-resistant lines were transferred to the greenhouse. Four lines producing sufficient amounts of fruit for analyses were quantified for their relative FaGT2 mRNA levels in red fruits by TaqMan reverse transcription-PCR. Two plants transformed with the AS construct; AS lines 6 and 9 showed lower levels of the FaGT2 transcript in the fruits of different harvests corresponding to only 34% and 53% of the levels in control plants, respectively (Fig. 6A

Figure 6.

A, Relative quantification of FaGT2 transcripts in AS transgenic strawberry fruits using the comparative Ct method. Expression levels of FaGT2 mRNA in control plants are compared with those from transgenic plants (FaGT2 AS6 and 9). Ct represents the cycle at which sample crosses threshold value. The dbp mRNA was used for standardization. B, Quantitative determination of Glc esters in ripe fruit of control plants cv Calypso (white bars) and strawberry plants cv Calypso transformed with AS FaGT2 constructs (FaGT2 AS6, gray bars; FaGT2 AS9, black bars). The amount of at least three different purifications from different harvests is indicated with its sd. ANOVA performed on these results revealed a significant [F(8,54) = 5.81, P < 0.05] difference in the levels of cinnamoyl- and p-coumaroyl-d-Glc in fruits among the plants transformed with the AS construct and control plants (*). Details are described in “Materials and Methods.”

). These lines with a constant transgenic effect on FaGT2 expression were selected for further analysis of cinnamoyl-, p-coumaroyl-, caffeoyl-, feruloyl-, and p-hydroxybenzoyl-d-Glc levels in ripe fruit (Fig. 6B). AS lines 6 and 9 contained reduced levels of cinnamoyl- and p-coumaroyl-d-Glc in comparison with the control group. The sds of the Glc esters in the controls were larger than those in the transgenic plants because fruits from different control plants and locations were used for analyses. Transgenic lines did not significantly differ in the concentration of caffeoyl-, feruloyl-, and p-hydroxybenzoyl-d-Glc when compared with controls. The absolute amounts of metabolites in Figure 6B differ slightly from those presented in Figure 4B, a difference that might be explained by the different cultivars used in each experiment.

DISCUSSION

Similarity of FaGT2 to Other GTs

We show here that the strawberry gene FaGT2 encodes a protein that catalyzes the transfer of a UDP-activated Glc residue to a variety of aromatic acids to form the corresponding Glc esters. FaGT2 contains a PSPG box found in many other GTs, and is most closely related in sequence to other ester-forming GTs from Arabidopsis (Jackson et al., 2001; Lim et al., 2001) and from rapeseed (Milkowski et al., 2000b) as well as to a Satsuma mandarin GT shown to produce glucosides (Kita et al., 2000; Fig. 1). FaGT2 can be assigned to group L in the phylogenetic tree of the Arabidopsis GT multigene family (Li et al., 2001), a branch that consists of 107 Arabidopsis sequences. Other functionally characterized members of the L group come from tobacco, rapeseed, and maize (Zea mays) capable of forming Glc esters, whereas others from Satsuma mandarin and Asian beefsteak (Perilla frutescens) produce O-glucosides and one from rapeseed forms an S-glucoside (Ross et al., 2001). The enzymatic activity of FaGT2 is comparable to UGT84A1-A3 isolated from Arabidopsis with respect to substrate specificity to aromatic acids and enzyme kinetics (K_m and V_max), although it seems that FaGT2 displays a slight preference for cinnamic acid (Lee and Raskin, 1999; Milkowski et al., 2000a; Lim et al., 2001). All of the positively tested substrates for FaGT2 contained both a carboxyl group and an aromatic ring system, which would be consistent with a function in phenylpropanoid metabolism.

Ripening-Related Expression of FaGT2

The observation that the FaGT2 gene is predominantly expressed in fully ripe fruit receptacles with a negligible expression in achenes suggests that FaGT2 is involved in metabolic pathways strongly activated in ripe receptacle tissue. This hypothesis is also supported by the weak expression found for the FaGT2 gene in vegetative tissues. The expression of some strawberry fruit ripening-related genes has been shown to be negatively regulated by auxins synthesized by achenes. It is assumed that the phytohormone auxin is the main signal molecule coordinating the growth and initiation of ripening in strawberry fruit (Given et al., 1988). During the early stages of fruit development, auxin promotes fruit growth. On the other hand, strawberry ripening is triggered by a decline in the levels of auxin in the receptacle, probably due to the cessation of auxin synthesis and transport from the maturing achenes. The expression results demonstrate that FaGT2 is also negatively controlled by auxin. The expression characteristics of the FaGT2 gene during strawberry fruit growth and ripening and its suppression by auxins further support its role in the fruit-ripening process. (Medina-Escobar et al., 1997; Moyano et al., 1998; Trainotti et al., 1999; Aharoni et al., 2002; Blanco-Portales et al., 2002; Benítez-Burraco et al., 2003).

Recently, a study showed that oxidative stress conditions during strawberry fruit development induce the expression of ripening-related genes (Aharoni et al., 2002). A potential source of oxidative stress might be the lignification of vascular tissue (Aharoni et al., 2002). Expression of a strawberry lignification-related FaCAD gene has been shown to be greatly induced in ripe receptacle tissue (Blanco-Portales et al., 2002). Similarly, expression of the FaGT2 gene was also increased by oxidative stress. Thus, the FaGT2 gene is under antagonistic transcriptional regulation for both auxins and oxidative stress.

Reaction Catalyzed in Vivo

We observed a positive correlation during fruit development between accumulation of FaGT2 transcripts and increasing concentrations of cinnamoyl-, p-coumaroyl-, and caffeoyl-d-Glc. The remaining Glc esters detected in the fruit did not show a clear ripening-related pattern (i.e. they were either present in constant amounts during fruit development or showed, in the case of p-hydroxybenzoyl Glc, a biphasic accumulation pattern). p-Coumaroyl-d-Glc and caffeoyl-d-Glc were also detected in leaves and flowers where FaGT2 was not expressed or barely expressed. Expression of additional GTs in leaf and flower tissue or the transport of Glc esters to the foliage or flowers might explain the detection of these metabolites. However, the absence of Glc esters in petioles suggests that transport is unlikely.

The concentrations of free aromatic acids in strawberry fruit and in other parts of the plant were only small, about two orders of magnitude less than that of the Glc esters (Schuster and Herrmann, 1985; Hakkinen and Torronen, 2000; Kosar et al., 2004; Olsson et al., 2004). No correlation of the amounts of free aromatic acids and Glc esters could be deduced. These data clearly demonstrate that most of the aromatic acids in the strawberry plant are immediately conjugated, thus preparing them for further biosynthetic steps. However, it is not known how much of the free acids are channeled into the general phenylpropanoid pathway and branching ways in strawberry fruit.

The large reduction in the concentration of cinnamoyl-d-Glc in transgenic AS lines (up to 75%) suggests that synthesis of this compound in the fruit is controlled mostly by FaGT2. Because transgenic fruits transformed with the AS construct also contained reduced levels of p-coumaroyl-d-Glc, it appears that FaGT2 also contributes to the formation of this ester. However, the extent of the down-regulation was not as potent as for the cinnamoyl derivative (about 50%). This observation suggests that this ester is probably also formed by additional GTs that provide a basic level of p-coumaroyl-d-Glc in the ripening fruit. The identification of p-coumaroyl-d-Glc in immature fruits, leaves, and flowers, where FaGT2 is not or barely expressed, strongly supports this hypothesis. Concentrations of the other metabolites, such as caffeoyl-, feruloyl-, and p-hydroxybenzoyl-d-Glc, were not significantly affected. Preliminary data for FaGT1, FaGT3, and FaRT1 show that they do not form Glc esters and therefore are not involved in the biosynthesis of the phenylpropanoyl- and benzoyl-d-Glc derivatives.

Role of FaGT2 and Possible Function of Cinnamoyl-d-Glc in Strawberry Fruit

An 1-O-trans-cinnamoyl-β-d-glucopyranose:alcohol cinnamoyltransferase has been found in the fruit of cape gooseberry (Physalis peruviana; Latza and Berger, 1997). This enzyme catalyzes the transfer of cinnamic acid, which is activated as a Glc ester, to various short-chain alcohols (e.g. methanol and ethanol), thus forming the volatile esters methyl- and ethyl cinnamate. These compounds have also been found in strawberry fruit and they contribute to the strawberry flavor (Schreier, 1980; Hirvi and Honkanen, 1982; Da Silva and Das Neves, 1999). The recently identified strawberry alcohol acyl-CoA transferase cannot account for these esters due its different substrate specificity (Aharoni et al., 2000). But it is possible that 1-O-trans-cinnamoyl-β-d-Glc, another activated acid, might be processed in a later biosynthetic step to these flavor compounds in strawberry fruit. Unfortunately, because fruits of the strawberry variety Calypso that were used in our transformation experiments produce only small amounts of methyl and ethyl cinnamate (≤0.01 mg kg⁻¹), we were not able to determine their accurate concentrations and compare their levels in the fruits of the transgenic plants.

In conclusion, we demonstrate that the accumulation of cinnamoyl-d-Glc during strawberry fruit ripening is caused by a UDP-Glc:cinnamate GT whose expression is negatively regulated by auxin and induced by hydroxycinnamic acids. Future experiments will examine whether this enzyme provides the precursor for the biosynthesis of methyl and ethyl cinnamate, two constituents of strawberry flavor.

MATERIALS AND METHODS

Chemicals

Except when noted, all chemicals, salts, solvents, and phenolic compounds were purchased from Sigma, Aldrich, Fluka, and Roth. HPLC gradient-grade acetonitrile was from Fisher and Roth. Radiolabeled [6-³H]UDP-Glc (1 mCi/mL; 60 Ci/mmol) was purchased from Biotrend Chemikalien GmbH.

Plant Material

Strawberry (Fragaria × ananassa) cv Calypso or Elsanta plants were provided by Plant Research International, Wageningen, The Netherlands. They were grown under standard greenhouse conditions with a 16-h photoperiod. Tissues harvested for RNA extraction or chemical analysis were immediately frozen in liquid nitrogen and stored at −80°C. Homogenization was carried out with liquid nitrogen.

Construction of the FaGT Expression Plasmid

Construction of the cDNA library and sequence analysis of the expressed sequence tags has been described elsewhere (Aharoni and O'Connell, 2002). GenBank accession numbers are FaGT1, AY663784; FaGT2, AY663785; FaGT3, AY663786; and FaRT1, AY663787. The entire ORF of the FaGT2 gene was cloned using RACE-PCR (CLONTECH), subcloned to the pGEM vector (Promega), and sequenced from both directions. The coding region of FaGT2 was amplified using the forward primer 5′-AAAAAGGATCCGGGTTCCGAATCATTGGTTCA-3′ and the reverse primer 5′-AAAAAAAGCTTCCTTACGACTCGACTAGTTCA-3′. These primers added BamHI and HindIII restriction sites at the 5′ and the 3′ end of the coding region, respectively, so that the PCR fragments could be ligated to the same sites in the pRSETB expression vector (Invitrogen). The forward primer eliminated the native Met ATG codon of FaGT2 and formed a fusion protein at the N terminus with a peptide that included an ATG translation initiation codon, a series of six His residues (His tag), and an anti-Xpress (Invitrogen) epitope.

Strawberry FaGT2 Gene Cloning

Isolation of the strawberry FaGT2 genomic clone was performed by differential screening of a strawberry genomic library according to Blanco-Portales et al. (2002). A ³²P-labeled FaGT2-cDNA was used as a probe. One positive clone ranging between 12 and 17 kb in size was isolated and analyzed by restriction mapping. Then suitable fragments for DNA sequencing were obtained by subcloning into the pBluescript vector (Stratagene), and their DNA inserts were completely sequenced on both strands.

Heterologous Expression of the Protein

Escherichia coli BL21 (DE3) pLysS (Novagen, Merck Biosciences) cells containing the GT sequence in the pRSETB vector were incubated overnight in Luria-Bertani medium containing the appropriate antibiotics (ampicillin 100 μg mL⁻¹ and chloramphenicol 34 μg mL⁻¹) at 37°C. The next day the culture was diluted 1:40 with Luria-Bertani medium supplemented with the antibiotics and incubated at 37°C on a centrifuge until the OD₆₀₀ reached 0.4 to 0.6. Expression was initiated with the addition of IPTG to 1 mm, and the culture was centrifuged at 16°C to 18°C overnight. For purification, the talon resin from CLONTECH (BD Biosciences) was used. The manufacturer's instructions were followed with slight modifications. Briefly, the cells were harvested by centrifugation and disrupted using a chilled mortar and pestle with glass beads. The supernatant was added to the pre-equilibrated resin after centrifugation, and the resin was washed several times. The His-tagged protein was eluted with imidazol buffer and used for SDS-PAGE, western-blot analysis, or enzyme activity assays. Protein concentration was determined by the Bradford method (Bradford, 1976) using bovine serum albumin as a standard. As a negative control, BL 21 cells were transformed using an empty pRSETB plasmid.

Enzyme Assay

Enzyme activity was assayed in 100 mm Tris-HCl buffer, pH 8.0, containing 5 mm 2-mercaptoethanol, 10% glycerol, and 5 mm sodium disulfite. The aglycon concentration was varied from 2 μm to 2 mm, at constant (5 mm) UDP-Glc concentration. A mixture of unlabeled and [6-³H]UDP-Glc (1.0 × 10⁵ dpm) was used. The reaction was started by adding 2 to 20 μg protein to a final volume of 250 μL. Assays were performed for 30 min at 21°C. To elucidate the influence of salts, assays were conducted in the presence of 0.01, 0.1, 1, and 10 mm Cu²⁺, Mg²⁺, Fe²⁺, Zn²⁺, and Mn²⁺. The reaction mixtures were extracted twice with water-saturated n-butanol (500 μL), and radioactivity of the products was determined by liquid scintillation counting (LKB Rackbeta 1214) after addition of 10 mL of an emulsifier-safe scintillation cocktail (Perkin-Elmer). As a control, BL 21 cells transformed with an empty pRSETB were purified and assayed for activity under the same conditions. As a second control, purified enzyme solution was added to the reaction mixture and immediately heated to 95°C for 5 min.

Identification of the FaGT2 Products

1-O-trans-cinnamoyl-β-d-glucopyranose was synthesized as described in detail elsewhere (Plusquellec et al., 1986). Identity of the product was confirmed with ¹H-NMR and ¹³C-NMR (data not shown). HPLC data acquisition and analysis of both the synthetic and enzymatically formed products were performed using Eurochrom 2000 software and a HPLC system equipped with a Eurospher C-18 column (25 cm × 4.0 mm i.d., particle size 5 μm) connected to a Maxistar pump and variable wavelength monitor (Knauer). Analyses were performed using a linear gradient proceeding from 95% water acidified with 0.05% formic acid and 5% acetonitrile to 100% acetonitrile in 30 min at a flow rate of 1 mL min⁻¹. Alkaline hydrolysis of formed products was performed by adding 1 m NaOH to a final concentration of 0.1 m NaOH. Products were analyzed using reversed-phase HPLC UV and online radioactivity detection (Radiomatic A100; Canberra Packard) with a yttrium-silicate solid cell.

Auxin Treatments

Achenes of green-stage strawberry fruits still attached to the growing plant were carefully removed using the tip of a scalpel blade. One de-achened fruit was treated with a lanoline paste containing 1 mm 1-NAA in DMSO 1% (v/v). The other de-achened fruits (reference group) were treated with the same paste, but without NAA. Fruit samples were harvested at 1, 2, 3, 4, and 5 d after treatment, immediately frozen in liquid nitrogen, and stored at −80°C until use.

Strawberry Cell Cultures

To initiate callus cultures, young leaves from micropropagated strawberry plants were removed and cut into strips (3–4 mm wide) with sterile scalpel and forceps. These explants were transferred to petri dishes on the surface of solid medium (Murashige and Skoog, 1962), supplemented with 2.5 mg L⁻¹ 2,4-D, with lower epidermis downward. The dishes were sealed with parafilm and incubated at 25°C in a growth chamber under diffuse light. When sufficient callus was developed from the explants (4–6 weeks), small pieces (0.3–0.5 cm diameter) were excised and transferred to fresh medium and cultured as before. In this way, callus stocks were maintained indefinitely by subculturing to fresh medium at monthly intervals. To initiate cell suspension cultures, 1 g of callus was excised, disaggregated, and transferred to 30 mL of liquid medium in a 100-mL Erlenmeyer flask. These cultures were incubated at 25°C on a rotary centrifuge at 100 rpm under diffuse light. The cell cultures were maintained by subculturing 1 g of filtered cells in fresh liquid medium every 15 d.

Oxidative Stress Treatments

One milliliter of a sterile solution of 1 mm menadione, 0.5 mm hydrogen peroxide, or sterile water (control) was carefully injected into white-stage fruit (still attached to the plants) using a hypodermic syringe with a needle gauge. Fruit was harvested at 2, 4, and 8 h, immediately frozen in liquid nitrogen, and stored at −80°C until RNA isolation. Additionally, the compounds were added to 3-d-old subcultured strawberry cells. Cell cultures treated with sterile water were used as controls. Cells were filtered and harvested at 2 and 4 h, immediately frozen in liquid nitrogen, and stored at −80°C until RNA isolation.

Treatment with Phenolics Compounds

Three-day-old strawberry cell cultures were treated with the following phenolics compounds: cinnamic acid, p-coumaric acid, caffeic acid, ferulic acid, and benzoic acid. All compounds were added to 1 mm final concentration. Cell cultures treated with water were used as controls. Cells were filtered and harvested at 4 and 8 h, immediately frozen in liquid nitrogen, and stored at −80°C until RNA isolation.

RNA Isolation and Gene Expression Analysis

Total RNA from a pool of six to seven strawberry fruit at different ripening stages, roots, leaves, flowers, and petioles were isolated according to Manning (1991). To investigate the differential expression of the strawberry FaGT2 gene during the strawberry ripening process, qRT-PCR analyses were carried out.

qRT-PCR Analysis

qRT-PCR studies of gene expression were carried out as previously described for other strawberry fruit ripening-related genes (Blanco-Portales et al., 2002; Benítez-Burraco et al., 2003). Two specific primers (5′-GGCTTTGAACTTCTGGTTCTGCCGGAAGG-3′ and 5′-GAGTTCCACCCGCAATGAGTCACAAAGC-3′) corresponding to the FaGT2 cDNA sequence were employed. As an endogenous control, we used a strawberry gene showing strong sequence homology with that of a gene coding a putative submergence-induced protein (sip) from banana (Musa acuminate [AF414129]; for expression analysis in the different plant parts) or a cDNA from strawberry that showed strong homology to a putative DNA-binding protein (dbp) gene from Arabidopsis (Arabidopsis thaliana; for expression analysis in the transgenic lines). These genes displayed constitutive expression in all conditions analyzed in this study. The specific primers (5′-ATTTCAGGCTGCAGCATCAACTGCGGTG-3′) and (5′-GCAATGCGGCTTTTCATTGGTGATCCAG-3′) for the sip gene were employed for normalization purposes. For monitoring dbp cDNA levels, a forward dbp primer (5′-TTGGCAGCGGGACTTTACC-3′), a reverse dbp primer (5′-CGGTTGTGTGACGCTGTCAT-3′), and a fluorogenic TaqMan dbp probe (FAM-5′-TCCTTCTTCAGCTGCATCAATATGTTTTCCA-3′-TAMRA) were used in the TaqMan quantitative PCR reaction. In qRT-PCR analysis, quantification is based on cycle threshold (Ct) values. Ct is a measurement taken during the exponential phase of amplification when limiting reagents and small differences in starting amount have not yet influenced PCR efficiency. Ct is defined as the cycle at which fluorescence is first detectable above background and it is inversely proportional to the logarithm of the initial copy number. In our system, each 10-fold difference in initial copy number produced a 3.2 cycle difference in Ct. Each reaction was done in triplicate, and the corresponding Ct values were determined. The Ct values of each qRT-PCR reaction were normalized using the Ct value corresponding to one of the endogenous controls. Moreover, the efficiency of each qRT-PCR was also calculated. All these values were utilized to determine the increases in gene expression according to the following expression:

\[\mathrm{Fold}{\,}\mathrm{change}{=}2^{{-}{\Delta}({\Delta}\mathrm{Ct})}\]

\[{\Delta}\mathrm{Ct}{=}\mathrm{Ct}(\mathrm{target}{\,}\mathrm{gene}){-}\mathrm{Ct}(\mathrm{endogenous}{\,}\mathrm{control})\]

\[{\Delta}({\Delta}\mathrm{Ct}){=}{\Delta}\mathrm{Ct}(\mathrm{condition}{\,}1){-}{\Delta}\mathrm{Ct}(\mathrm{condition}{\,}2).\]

The expression levels corresponding to the FaGT2 gene in transgenic strawberry plants were analyzed according to TaqMan quantitative PCR as described previously (Schaart et al., 2002). Because FaGT2 and the corresponding transcripts are only formed in ripe fruit, we decided to determine the down-regulation of FaGT2 in fruits of the AS lines. Briefly, TaqMan quantitative PCR was set up in three replicates of 20-μL volumes containing first-strand cDNA from 10 ng total RNA, 800 nm forward FaGT2 primer (5′-GAGGGTCGTCGGACAGGAA-3′), and 100 nm reverse FaGT2 primer (5′-TGATCCTCCTAACCTCGTCCA-3); 10× TaqMan buffer A, including a passive internal reference dye; 5.5 mm MgCl₂; 200 μm dATP; 200 μm dCTP; 200 μm dGTP; 400 μm dUTP; 0.01 unit/μL AmpErase uracil-n-glycosylase; and 0.025 unit μL⁻¹ AmpliTaq Gold DNA polymerase. All reagents were supplied with the TaqMan PCR core reagent kit (Perkin-Elmer, Applied Biosystems). Cycling parameters for the PCR reaction were 50°C for 2 min, 95°C for 10 min, followed by 40 cycles of 95°C for 15 s and 60°C for 1 min. All PCR reactions were incubated in the ABI Prism 7700 sequence detection system (Perkin-Elmer, Applied Biosystems).

Identification of Aromatic Acids and Their Glucosylated Derivatives

A glass column (50 × 2.5 cm) filled with Amberlite XAD-2 polymeric adsorbent (20–60 mesh; Aldrich) was successively washed with methanol and water. Approximately 2 to 5 g of frozen strawberry fruit, root, flower, leaf, and petiole was homogenized with 20 mL water using Ultra Turrax (T18 basic; IKA Works), centrifuged (3,500g, 10 min), and the supernatant applied to the XAD-2-column. The residue was resuspended in 20 mL water and the extraction was repeated twice. After rinsing the column with 100 mL distilled water, the semipolar compounds were eluted with 50 mL diethyl ether and glycosides by 80 mL methanol. The methanolic extract was concentrated in vacuo to 1 mL, whereas the diethyl ether extract was distilled using a Vigreux column to 1 mL. For LC-UV-ESI-MSⁿ analysis, the entire extract was pipetted onto an equivalent volume of distilled water, and the volatile organic phase was removed with a stream of nitrogen. The methanolic extract was used directly for LC-UV-ESI-MSⁿ analysis.

LC-UV-ESI-MSⁿ

The system used for LC-UV-ESI-MSⁿ analysis was a Bruker esquire 3000 plus mass spectrometer, equipped with an Agilent 1100 HPLC system composed of an Agilent 1100 quaternary pump and an Agilent 1100 variable wavelength detector. The Eurospher C18 column (10 cm × 2 mm) had a particle size of 5 μm (Grom Analytik and HPLC GmbH). The ionization voltage of the capillary was 3,074 V, the end plate was set to −500 V, the capillary exit was −109.8 V, and the Octopole radiofrequency amplitude was 120 Vpp. The temperature of the dry gas (N₂) was set to 300°C at a flow of 10 L min⁻¹. The full-scan mass spectra of the glycosides ranging from m/z 50 to 500 were measured until the ICC target reached 20,000 or 200 ms, whichever was reached first. Fragmentation of ions was performed using helium as the collision gas and the collision energy was set at 1.0 V. All mass spectra were acquired in the negative ionization mode. Auto mass spectrometry was used to break down the most abundant pseudomolecular [M−H]⁻ or [M+HCOO]⁻ ions of the different compounds of the strawberry extracts. Identity of the compounds was confirmed by comparing the retention times and fragmentation patterns with the enzymatically synthesized reference compounds (Table II). The LC gradient proceeded from 0% acetonitrile and 100% water acidified with 0.05% formic acid to 50% acetonitrile and 50% acidic water in 35 min, in 2.5 min to 100% acetonitrile, 2.5 min at these conditions, then back to 100% water and 0% acetonitrile in 5 min at a flow rate of 0.2 mL. The detection wavelength was 280 nm.

Quantification and Statistical Analysis

Fruit harvested from each transgenic plant was pooled regardless of harvest date and at least three samples from each plant were independently analyzed by LC-UV-ESI-MSⁿ. Control fruit from several strawberry cv Calypso plants were pooled regardless of harvest date and purified in parallel throughout the course of the experiment. Due to the lack of reference compounds, standard curves were generated with synthesized cinnamoyl-d-Glc for UV and MS signals and used to calculate all compounds as cinnamoyl-d-Glc equivalents. Where possible, we integrated the UV signal using the MSⁿ signal to verify the purity. Where coelution of several compounds took place, or where the UV signal was too small to quantify, we integrated the peak area in the isolated MS ion trace. Statistical differences between the variable groups and the control group were calculated using one-way ANOVA (P < 0.05) with StatCrunch free online software (www.statcrunch.com) together with the Dunnett's post hoc comparison (computed by hand).

Plant Transformation

Plant transformation was carried out as described in detail elsewhere (Schaart et al., 2002). In short, young folded leaves of strawberry (cv Calypso) were collected and surface sterilized using a 2% (w/v) sodium hypochlorite solution. Segmented (4–6 mm) leaflets were transfected using supervirulent Agrobacterium tumefaciens strain AGL0 containing the FaGT2 gene in the AS direction in a derivative of pBINPLUS plasmid under the control of a cauliflower mosaic virus 35S promoter. After cocultivation for 4 d in the dark at 25°C, the leaf explants were transferred to selection medium. Regenerated shoots were multiplied on proliferation medium and subsequently transferred to the greenhouse. During the first growing season, plants were phenotyped. Several ripe fruits (n = 3) were harvested from each transgenic line and analyzed for FaGT2 mRNA level by quantitative PCR. The following year different batches of ripe fruit were harvested from selected transgenic lines at different dates for further analysis of metabolites.

In most plant species, a generative progeny is produced for the selection of transgenic plants that are used for further study. In the case of octoploid strawberry, the high genetic background variation hinders the analysis of a generative progeny. If a primary transgenic octoploid strawberry is selfed, the offspring is genetically diverse with respect to the genetic background and the cultivar characteristics are lost (Mathews et al., 1998). However, an advantage of a vegetatively propagated crop like strawberry is that the selected transgenic lines with their specific expression levels of the target gene can be maintained forever in the same genetic background. But, as is always the case with callus-derived plants, one should be aware of transgenic instability and somaclonal variation (Mathews et al., 1998; Bath and Srinivasan, 2002; Jiménez-Bermúdez et al., 2002; Houde et al., 2004; Labra et al., 2004). In our transformation system, we used the ever-bearing cultivar Calypso that can produce up to eight harvests per year under greenhouse conditions. This allowed the detection of chimerism and transgene instability of primary transgenics over the year by monitoring harvest-to-harvest variation for FaGT2 transcripts or metabolite levels. In this way, chimeric plants and plants with unstable AS expression were eliminated.

Sequence data from this article can be found in the GenBank/EMBL data libraries under accession numbers FaGT1, AY663784; FaGT2, AY663785; FaGT3, AY663786; and FaRT1, AY663787.

ACKNOWLEDGMENTS

We thank Eran Pichersky for helpful discussions and valuable advice during the revision of the manuscript.

Author notes

1

This work was supported by Degussa AG.

2

These authors contributed equally to the paper.

3

Present address: Department of Plant Sciences, Weizmann Institute of Science, 76100 Rehovot, Israel.

^*

Corresponding author; e-mail ; fax [phone].

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: Wilfried Schwab ().

Article, publication date, and citation information can be found at www.plantphysiol.org/cgi/doi/10.1104/pp.105.074955.