RNAi-induced silencing of gene expression in strawberry fruit (Fragaria × ananassa) by agroinfiltration: a rapid assay for gene function analysis


*(fax +49 (0)8161 548 595; e-mail schwab@wzw.tum.de).


Intron-containing constructs encoding self-complementary ‘hairpin’ RNA (ihpRNA) have the potential to efficiently silence genes in a range of plant species. In this study we demonstrate the silencing of a ripening-related chalcone synthase (CHS) gene in strawberry fruits (Fragaria × ananassa cv. Elsanta) by a construct (ihpRNA) containing the partial sense and corresponding antisense sequences of CHS separated by an intron obtained from a F. × ananassa quinone oxidoreductase gene. An Agrobacterium strain carrying a T-DNA expressing the ihpRNA transgene was injected with a syringe into the receptacles of growing fruits still attached to the plant about 14 days after pollination. As a consequence of the reduced levels of CHS mRNA and enzymatic CHS activity, the levels of anthocyanins were downregulated and precursors of the flavonoid pathway were shunted to the phenylpropanoid pathway leading to a large increases in levels of (hydroxy) cinnamoyl glucose esters. We anticipate that this technique in combination with metabolite profiling analysis will be useful for studying the function of unknown genes during the development and ripening of strawberry fruit.


Metabolite profiling and cDNA microarray analyses of strawberry fruit in different developmental stages have allowed the detection of ripening-related metabolites and genes (Aharoni and O'Connell, 2002; Aharoni et al., 2002). The functions of some of these genes have been identified directly by appropriate assays, or inferred based on their homology to genes of known functions in other organisms (Aharoni et al., 2000; Manning, 1998; Wein et al., 2002). Additionally, Agrobacterium-mediated genetic transformation resulting in the downregulation of an individual gene has provided some clues to its function in specific ripening pathways (Jiménez-Bermúdez et al., 2002; Lunkenbein et al., 2006a). However, these traditional ways of studying the functions of various genes, especially ripening-related genes in strawberry, are time-consuming, as it can take at least 15 months from the transformation experiment until the first ripe fruits are available for analyses.

Recently, Spolaore et al. (2001) have shown strawberry fruit to be amenable to transient gene expression mediated by Agrobacterium using a β-glucuronidase (GUS) reporter gene interrupted by an intron. This system combined with a biolistic transient transformation protocol has also been used for the functional analysis of homologous and heterologous promoters in strawberry fruits (Agius et al., 2005).

Ribonucleic acid interference (RNAi) of plant genes, initiated by the delivery of double-stranded RNA (dsRNA), is an attractive new reverse-genetics tool for the study of gene function (Waterhouse and Helliwell, 2003). The silence-inducing dsRNA can be delivered either by stably transforming plants with constructs that encode ‘hairpin’ RNAs (hpRNA) or viral RNA, or it can be transiently delivered by infiltrating plant cells with transgene-carrying Agrobacterium tumefaciens, or, alternatively, by infecting plants with a virus (Burch-Smith et al., 2004; Liu et al., 2002; Orzaez et al., 2006; Robertson, 2004; Wesley et al., 2001). Since most viruses used for virus-induced gene silencing (VIGS) have a limited number of hosts, and since the virus–host combination seems to be a crucial factor in determining the efficacy of silencing, we decided to develop an RNAi-based transient silencing method using intron-containing self-complementary hpRNA (ihpRNA) constructs, which appear to be the most effective constructs to date (Smith et al., 2000; Wesley et al., 2001). To our knowledge VIGS in strawberry has not been attempted until now and agroinfiltration of fruits for gene silencing has only been performed with tomato (Fu et al., 2005; Orzaez et al., 2006). The transient Agrobacterium-mediated ihpRNAi silencing technique has been used before for the downregulation of a co-delivered GFP reporter gene in leaves of Nicotiana benthamiana or in a N. benthamiana GFP line but not in fruits for the silencing of endogenous genes (Johansen and Carrington, 2001; Koscianska et al., 2005).

The efficacy of the ihp construct was tested in strawberry fruits by targeting the pigment biosynthesis gene chalcone synthase (CHS) that is involved in the biosynthesis of the major anthocyanins in strawberry fruits, pelargonidin 3-glucoside (purple) and pelargonidin 3-rutinoside (Hong and Wrolstad, 1990) (Figure 1a). CHS was chosen as the reporter gene because reduction of the CHS function using antisense technology leads immediately to the loss of pigmentation in flowers and fruit, and is thus easily detected (Deroles et al., 1998; Lunkenbein et al., 2006b; Que et al., 1998; Van der Krol et al., 1988).

Figure 1.

 Flavonoid biosynthetic pathway in strawberry fruit (a), and reaction catalysed by chalcone synthase (CHS) (b).
Arrows denote compounds that are up- (↑) or downregulated (↓) during intron-hpRNAi-mediated silencing of the FaCHS gene in strawberry fruit.

As a consequence of the reduced activity of CHS in strawberry plants transformed with an antisense CHS gene construct, the levels of anthocyanins, flavonols and proanthocyanidins were downregulated in the fruit and precursors of the flavonoid pathway were shunted to the phenylpropanoid pathway (Lunkenbein et al., 2006b). That the expression of the CHS gene in fruit tissue is developmentally regulated and associated with fruit colouring has been shown in Rubus (Kumar and Ellis, 2003), apple (Honda et al., 2002), bilberry (Jaakola et al., 2002), grapevine (Goto-Yamamoto et al., 2002) and strawberry (Aharoni and O'Connell, 2002; Manning, 1998). The phenomenon of co-suppression of homologous genes was first described in petunia, where the introduction of an additional CHS gene resulted unexpectedly in plants with totally white flowers (Metzlaff et al., 1997; Napoli et al., 1990; Van Blokland et al., 1994). Additionally, Arabidopsis plants transformed with a CHS-ihpRNA construct showed pronounced silencing resulting in the reduced production of flavonoid pigments (Wesley et al., 2001).

In the present study the capability of a construct encoding ihpRNA to produce gene silencing in octaploid Fragaria × ananassa was demonstrated using the strawberry CHS gene. The effect of the reduction in CHS function was confirmed on the mRNA, enzyme and metabolite level. This method may facilitate studies of gene function in strawberry fruits.

Results and discussion

Infection of strawberry fruit with pBI-Intron

To find an efficient method for transfection of strawberry fruit by Agrobacterium we tested different media and techniques for tissue infiltration and injected the Agrobacterium suspension after different intervals post-pollination. A GUS-intron gene (pBI-Intron) was used as a reporter to make sure that the observed GUS activity was not due to its expression inside the Agrobacterium cells (Figure 2a). The reporter was interrupted by an intron obtained from the strawberry quinone oxidoreductase gene (FaQR), which prevents expression of GUS activity in the prokaryote A. tumefaciens (Raab et al., 2006), while allowing its expression in plant cells due to their ability to splice out the intron and to produce a functional GUS mRNA. A modified injection method according to Spolaore et al. (2001) gave the best results. Strawberry fruits that were still attached to the plant were injected with Agrobacterium transformed with pBI-Intron for the expression of GUS (Jefferson, 1987). Fruits only remained viable when the injection of the Agrobacterium suspension occurred more than 10 days post-pollination. Deleterious side-effects of fruit agroinfiltration appeared in young fruit and consisted of growth arrest, premature ripening and abscission. Therefore, the developing fruit were routinely treated 14 days after pollination when the fruit had almost gained their maximum size (2–3 cm long). Time points of injection later than 14 days post-pollination were not tested because to suppress ripening-related genes it is mandatory to start silencing as soon as possible.

Figure 2.

Agrobacterium-mediated transient expression in F. × ananassa fruit.
Constructs used contained the 35S promoter and the terminator sequence (nos) (a). The pBI-Intron construct contained the GUS coding sequence interrupted by an intron obtained from the strawberry quinone oxidoreductase gene (FaQR) whereas the vector pBI-CHSi contained both sense and antisense FaCHS sequences separated by a FaQR gene intron.
β-Glucuronidase activity determined in fruit (b) and in protein extracts obtained from strawberry fruit injected with a suspension of Agrobacterium containing the pBI-Intron construct (c). Single fruit were analysed after defined intervals post-infiltration. Bars show the average activity of three fruits per time point and standard deviation.

Activity assays for GUS were performed 0–16 days after injection in order to assess the time period necessary to see optimum gene expression (Figure 2b,c). Though some variability could be observed in the values of GUS activity (2.2–10.9 fmol mg−1 min−1) it is apparent that even after 2 days reporter gene expression is detectable in every fruit (n = 24) with maximum activity at day 14. The results clearly demonstrate that after infiltration the fruit can be kept in physiological and viable conditions that permit gene silencing.

Silencing of the FaCHS gene

To test whether ihpRNA can be used to silence a gene expressed during strawberry fruit maturation, we targeted the F. × ananassaCHS (FaCHS) gene in the flavonoid and anthocyanin biosynthesis pathway (Figure 1). FaCHS catalyses the formation of naringenin chalcone, the precursor for major pigments in strawberry fruit like pelargonidin 3-glucoside and pelargonidin 3-glucoside-malonate, which are both derived from malonyl-CoA and p-coumaroyl-CoA.

The expression of the CHS gene in fruit tissue is developmentally regulated and associated with fruit colouring (Aharoni and O'Connell, 2002). Antisense suppression of the FaCHS gene in transgenic strawberry fruits results in the strong inhibition of anthocyanin production (Lunkenbein et al., 2006b). Therefore, the FaCHS gene is an ideal choice for a reporter gene for the successful application of ihpRNA in strawberry fruit.

To silence the FaCHS gene in strawberry fruit, the vector pBI-CHSi was generated by inserting a 303-bp fragment of the FaCHS gene in the sense and antisense orientation interrupted by a FaQR gene intron in the pBI-Intron to replace GUS (Figure 2a). An Agrobacterium suspension containing the pBI-CHSi construct was injected into fruit still attached to the plant 14 days after pollination. At this stage the fruit are green and FaCHS gene expression starts. About 10 to 14 days after treatment, the pBI-CHSi infiltrated fruit developed white regions, a clear sign of impaired anthocyanin accumulation where the injection of the Agrobacterium suspension occurred (Figure 3a). Fruit infiltrated with Agrobacterium containing the pBI-Intron control vector turned completely red like the untreated fruit of the wild type (Figure 3c). When fruit were injected three times with a suspension of Agrobacterium containing the pBI-CHSi construct, e.g. 10, 11 and 12 days post-pollination, we even obtained completely white fruits in the fully ripe stage, which was recognizable by their brown achenes (Figure 3b). Cross-sectional views confirm the lack of anthocyanin accumulation in the inner part of the fruit as well (Figure 3d,e). When the infiltration of Agrobacterium carrying the pBI-CHSi construct was conducted between 10 and 14 days post-pollination a transformation efficiency of 100% was achieved (n = 21). In all the following assays the entire fruits were used.

Figure 3.

 Silenced phenotypes of the FaCHS-ihpRNA transformed (pBI-CHSi) fruits and of a control fruit.
(a) A chimeric phenotype was obtained with white sectors where the solution of Agrobacterium containing the ihpRNA construct was injected.
(b) An almost white fruit was produced when the injection was repeated at three consecutive days.
(c) Bright red fruit of the control plants infiltrated with a suspension of Agrobacterium containing the control construct (pBI-Intron).
Cross-sectional views of two chimeric phenotypes (d, e) and a control fruit (f).
Photographs of the strawberry fruits were taken 14 days after injection.

Transcript levels, enzyme activity and metabolite concentrations

To confirm the suppression of FaCHS at the molecular level, we performed semiquantitative RT-PCR. Primers were used that anneal to the region outside the FaCHS gene targeted for silencing. The FaCHS message in the pBI-CHSi-infiltrated fruit was strongly reduced [by a factor of five according to t/c (the relative level of the target amplification product FaCHS, t, over the DBP control, c) corresponding to 20% of control] when compared with fruit infiltrated with Agrobacterium transformed with the control vector pBI-Intron and control fruit (untreated wild type) (Figure 4a). In contrast, the level of the loading control DBP (encoding for a DNA-binding protein) mRNA was similar in FaCHS-silenced fruit and control fruit (pBI-Intron and control) serving as an internal control for RNA quality and RT-PCR amplification.

Figure 4.

 Intron-hpRNAi-mediated silencing of the FaCHS gene in strawberry fruit attached to the plant.
(a) Reverse transcriptase-PCR analysis of the FaCHS and DBP (DNA binding protein) genes in the FaCHS-silenced strawberry fruit. The first-strand cDNA was generated from total RNA isolated from the FaCHS-silenced fruit (pBI-CHSi), fruit infiltrated with Agrobacterium containing the control vector (pBI-Intron), and untreated fruit (control) according to Asif et al. (2000) and used in a PCR reaction using FaCHS and DNB gene-specific primers. Lanes 1–5 correspond to products from PCR cycle numbers 20, 25, 30, 35 and 40. The t/c values indicate the relative level of the target amplification product FaCHS (t) over the DBP control (c).
(b) Detection of FaCHS activity by autoradiography in protein extracts from the FaCHS-silenced and control fruit. Lane 1 shows the level of the product formed by FaCHS activity in FaCHS-silenced fruit (pBI-CHSi), lane 2 the activity in fruit infiltrated with Agrobacterium containing the control vector (pBI121-Intron), lane 3 the activity in untreated control fruit and lane 4 the reference naringenin.

The reduction of the FaCHS transcript level led to strongly reduced FaCHS enzymatic activity as evidenced by comparison with the activities in control fruit (pBI-Intron and untreated wild type) (Figure 4b). FaCHS preparations isolated from control fruit formed greater amounts of the product naringenin chalcone which was converted in the assay to naringenin than extracts from pBI-CHSi-infiltrated fruit.

The first committed step in flavonoid biosynthesis is catalysed by chalcone synthase, which produces naringenin chalcone by the condensation of three molecules of malonyl-CoA and one molecule of 4-coumaroyl-CoA (Figure 1b). In this way the CHS catalysis represents a branching point in the phenylpropanoid metabolism leading to flavonoids (flavonols, anthocyanins and proanthocyanidins). The flavonoid and phenylpropanoid profiles of the different samples (untreated control fruit, pBI-Intron and pBI-CHSi) were determined by liquid chromatography–UV–electrospray ionization–tandem mass spectrometry (LC-UV-ESI-MSn) analysis. As long as the reaction of the corresponding enzyme of the downregulated gene is known, metabolic profiling of related compounds in the specific metabolic pathway provides valuable information about altered pool sizes.

As expected, FaCHS-silenced receptacles (pBI-CHSi fruit) produced low levels of pelargonidin 3-glucoside, pelargonidin 3-glucoside-malonate and (epi)-afzelechin-pelargonidin 3-glucoside, accumulating the phenylpropanoids caffeoyl glucose and feruloyl glucose instead (Figure 5a). The level of epicatechin was not affected by the treatment.

Figure 5.

 Effect of FaCHS gene silencing caused by ihpRNA on the metabolite levels.
(a) Fourteen days after pollination green strawberry fruit were infiltrated with Agrobacterium transformed with a construct encoding FaCHS-ihpRNA. Levels of metabolites were determined by LC-UV-ESI-MSn 14 days after infiltration in FaCHS-silenced fruit (pBI-CHSi; n = 17), untreated control fruit (control; n = 10) and fruit infiltrated with Agrobacterium containing the control vector (pBI-Intron; n = 12). Identities of the compounds were confirmed by references and literature data. Box plot graphs were designed for the peak areas of selected compounds with significantly reduced levels (left column), significantly enhanced levels (middle column) or unchanged levels (right column) in pBI-CHSi fruit. A horizontal line in the boxes indicates the medians and boxes the interquartile range. Whiskers extend to 10th and 90th percentiles. Outliers are displayed by black dots.
(b) Wilcoxon–Mann–Whitney U-test was used for non-parametric analysis of intergroup comparison of data from untreated strawberry, strawberry injected with Agrobacterium transformed with the pBI-Intron construct and Agrobacterium transformed with the pBI-CHSi construct. Statistical significance levels P < 5.0 × 10−4 are shaded.

Statistical evaluation

The determination of the metabolite levels (expressed as peak area) in 10 untreated control fruit, 12 pBI-Intron-infiltrated fruit and 17 fruit injected with Agrobacterium transformed with pBI-CHSi showed high variability between the individual compounds within each sample group. The distribution of the values indicated by the box plots can be explained mainly by biological variation but also by inhomogeneous silencing effects (Figures 5a and 3d,e). We applied statistical methods to identify those compounds with significantly altered metabolite levels in order to propose a procedure for the investigation of the downregulation of unknown genes.

Values of transcript and metabolite levels in stably transformed strawberry fruits are not normally distributed (Lunkenbein et al., 2006b). Therefore, the Wilcoxon–Mann–Whitney U-test was used for non-parametric analysis of intergroup comparison on the basis of metabolite levels (Hart, 2001) (Figure 5b). The intergroup comparison of individual metabolites between untreated and pBI-Intron-infiltrated fruit showed that the concentrations of all the metabolites listed in Figure 5(b) do not differ significantly (Figure 5b, second column, P geqslant R: gt-or-equal, slanted 1.7 × 10−2). However, significant differences were detected in levels of pelargonidin 3-glucoside (P = 6.4 × 10−5), pelargonidin 3-glucoside-malonate (P = 6.4 × 10−5), (epi)-afzelechin-pelargonidin 3-glucoside (P = 1.6 × 10−4) and phenylpropanoids [caffeoyl glucose (P = 2.4 × 10−7) and feruloyl glucose (P = 4.5 × 10−6)] in between untreated and pBI-CHSi fruit.

A similar result was obtained by the statistical evaluation of the metabolite levels in pBI-Intron- and pBI-CHSi-infiltrated fruit. However, the concentrations of caffeoyl glucose in pBI-Intron- and pBI-CHSi-injected fruit are not significantly different (P = 1.5 × 10−1). It is suspected that the level of this ester increased due to the injury caused by the injection (Figure 5a) so that a further increase as a consequence of CHS downregulation is less obvious. Feruloyl glucose shows a similar pattern (Figure 5a). Therefore, side-effects like the formation of phytoalexins and lipoxygenase-derived oxylipids that can be induced by the injection of the Agrobacterium cells have to be taken into account (Orzaez et al., 2006). Similarly, strawberry fruit transformed with an antisense construct of CHS produced significantly higher levels of caffeoyl glucose (816% of control) and feruloyl glucose (1092% of control) but significantly lower levels of pelargonidin-3-glucoside-malonate (6% of control), pelargonidin-3-glucoside (8% of control) and (epi)-afzelechin-pelargonidin-3-glucoside (2% of control) than the control fruit (Lunkenbein et al., 2006b).

The amounts of the other flavonoids (flavonols and proanthocyanidins) were not significantly affected by the silencing of the CHS gene, except for kaempferol glucoside, which decreased (Figure 5b, fourth column, P = 1.3 × 10−4). Since most flavonol glucosides and proanthocyanidins are primarily synthesized in unripe green fruit they are already present before the agroinjection occurs and the silencing of CHS takes place (Halbwirth et al., 2006). Therefore, we assume that only kaempferol glucoside is synthesized in appreciable amounts during the latter stages of fruit ripening. Strawberry fruits constitutively expressing an antisense construct against CHS contained lower levels of flavonoids than the control fruits, confirming their biosynthesis in early stages of fruit growth (Lunkenbein et al., 2006b).

We have developed a method that uses agroinfiltration to transiently downregulate a structural gene in strawberry fruit. Silencing effects were observed in transcript levels and enzymatic activities, while metabolite profiling confirmed the in vivo function of the gene. Thus, the method is suitable for analysing gene functions during the development of strawberry fruit even though appropriate additional experiments including agroinfiltration with control constructs should be performed. In the case when an unknown gene is silenced and the consequence is not phenotypically visible a comprehensive analysis of plant metabolites by gas chromatography–mass spectrometry (GC-MS), liquid chromatography–mass spectrometry (LC-MS) and other profiling techniques in combination with statistical methods can deliver functional clues to elucidate the catalysed reaction of the gene product. The ihpRNAi silencing technique is especially suitable for confirming in vivo functions of catalytically active proteins whose substrate preferences have been shown in enzyme assays. We assume that this Agrobacterium-mediated ihpRNAi silencing technique is also applicable to other fruits like tomato, apple, orange, grape, etc.

Experimental procedures

Plant material

The octaploid strawberry F. × ananassa cv. Elsanta was used for transfections. Standard growing conditions were maintained at 25°C and a 16-h photoperiod under 120 μmol m−2 sec−1 irradiance provided by Osram Fluora lamps (München, Germany). For genetic and molecular analysis fruit were injected 14 days after pollination and harvested 3–24 days after injection. Fruit harvested 28 days after pollination were used as controls.

Reference compounds

Phenylpropanoyl glucose esters were enzymatically synthesized with FaGT2 (Lunkenbein et al., 2006a). Pelargonidin 3-glucoside, quercetin glucuronide, quercetin glucoside, kaempferol glucuronide, kaempferol glucoside, catechin and epicatechin were obtained from Roth (Karlsruhe, Germany). Proanthocyanidins and pelargonidin 3-glucoside-malonate were isolated from strawberry and identified according to Fossen et al. (2004) and Gu et al. (2003).

Plasmid construction

pBI-Intron construction.  The first intron of the F. × ananassa quinone oxidoreductase gene (AY158836, nucleotides 4107–4561) was selected for the GUS control construct. The intron was PCR-amplified from strawberry cv. Elsanta genomic DNA using primers (forward: 5′-GTGAGTTCCTCCCTCTTTCT-3′ and reverse: 5′-CTGCAAACGAAAATGAAAATAAATGA-3′) and cloned into SnaBI-cut pBI121 (Jefferson, 1987) according to Vancanneyt et al. (1990).

pBI-CHSi construction.  The second intron of the F. × ananassa quinone oxidoreductase gene (AY158836, nucleotides 4886–4993) was PCR-amplified from strawberry cv. Elsanta genomic DNA using primers (forward: 5′-GAAGATCTGCTAGCAGGTACATTCTGATTTCATTATCC-3′ with a BglII and NheI restriction site and reverse: 5′-CTCACTAGTGCAAGCTGCATCACATAAAAGTACAC-3′ with a SpeI restriction site) and cloned into BamHI–Ecl136II cut pBI121 (Jefferson, 1987) to replace GUS. To clone sequences encoding the inverted-repeat RNA into the resulting vector, a 303-bp fragment corresponding to nucleotides 10 022–10 325 of pBINPlusCHSas (Lunkenbein et al., 2006b) was inserted into the 5′ and 3′ arms of the intron and the resulting plasmid was named pBI-CHSi.

Transfection of strawberry by agroinfiltration

The A. tumefaciens strain AGL0 (Lazo et al., 1991) containing the pBI-CHSi or pBI-Intron, respectively, was grown at 28°C in Luria–Bertani (LB) medium with appropriate antibiotics. When the culture reached an OD600 of about 0.8, Agrobacterium cells were harvested and resuspended in a modified MacConkey agar (MMA) medium (Murashige and Skoog salts, 10 mm morpholine ethanesulphonic acid pH = 5.6, 20 g l−1 sucrose, according to Spolaore et al., 2001). The Agrobacterium suspension was evenly injected throughout the entire fruit while it was still attached to the plant about 14 days after pollination by using a sterile 1 ml hypodermic syringe.

Isolation of nucleic acids and RT-PCR analysis

Genomic DNA was isolated from young leaves as described by Mercado et al. (1999). Fourteen days after injection total RNA was isolated from fruit tissue according to Asif et al. (2000) and treated with RNase-free DNase (Promega, Mannheim, Germany). The first strand cDNA was synthesized using 1 μg total RNA, oligo d(T) primer and superscript reverse transcriptase (Invitrogen, Karlsruhe, Germany). For detection of CHS in semi-quantitative PCR, forward 5′-GCTGTCAAGGCCATTAAGGA-3′ and reverse 5′-GAGCAAACAACGAGAACACG-3′CHS primers were designed (product size 245 bp). The forward primer anneals outside the region targeted for gene silencing to ensure amplification of endogenous gene transcripts. A strawberry cDNA sequence coding for a DNA-binding protein (DBP) was used as an internal control for loading (Schaart et al., 2002) and the amplification primers were 5′-GGCATCGGAGATGGTACTGT-3′ and 5′-CCAGCATTCCGAACTTCTTT-3′ (product size 247 bp).

Fluorometric assay of GUS activity

For the fluorometric GUS assay (Jefferson, 1987) strawberry fruits were ground in liquid nitrogen and protein was extracted with 1.5 ml g−1 fresh weight of modified GUS extraction buffer [500 mm Na-phosphate pH = 7.8, 10 mmβ-mercaptoethanol, 10 mm Na2EDTA, 0.1% sodium lauryl phosphate, 0.1% Triton X-100, 5% polyvinylpolypyrrolidone (PVPP)]. Aliquots of the extracts (50 μl) were added to 50 μl of assay buffer (extraction buffer supplemented with 1 mm 4-methylumbelliferyl-β-glucuronide). The released 4-methylumbelliferone (4-MU) was quantified with a Tecan Safire microplate reader (Crailsheim, Germany) according to the manufacturer's instructions. The GUS activity was expressed as fmol 4-MU released min−1 mg−1 fresh weight.

Chalcone synthase (CHS) enzyme assay

Protein was extracted according to Punyasiri et al. (2004). The resulting supernatant was used directly for the chalcone synthase assay. [2-14C]malonyl-CoA (3 nmol, specific activity 2222 Bq nmol−1), 4-coumaroyl-CoA (1 nmol) and enzyme extract (2 μg protein) were incubated for 30 min at 30°C in a final volume of 100 μl of 100 mm Tris-HCl (pH 7.5) buffer. The products were extracted by ethyl acetate (2 × 200 μl) and separated by cellulose thin layer chromatography (Merck GmbH, Darmstadt, Germany) with chloroform:acetic acid:water (10:9:1). Dried chromatograms were exposed to a screen and analysed using a Bio-Image Analyzer (Fuji BAS 1000, Raytest, Straubenhaurdt, Germany).

Metabolite analysis

Metabolites were analysed about 14 days after injection in the ripe stage. Strawberry fruit were ground in liquid nitrogen. Approximately 200 mg of frozen material was thawed, centrifuged (16 000 g, 10 min) and the supernatant was used directly for LC-UV-ESI-MSn analysis.

Liquid chromatography-UV-electrospray ionization-tandem mass spectrometry

The system used for LC-UV-ESI-MSn analysis was a Bruker Esquire 3000 plus mass spectrometer (Bruker, Bremen, Germany), equipped with an Agilent 1100 HPLC system composed of an Agilent 1100 quaternary pump and an Agilent 1100 variable wavelength detector (Agilent, Waldbronn, Germany). The column was a Eurospher C18 column, particle size 5 μm, 10 cm × 2 mm (Grom Analytik & HPLC GmbH, Rottenburg, Germany). The ionization parameters were as follows: the voltage of the capillary was 3074 V and the end plate was set to −500 V. The capillary exit was −109.8 V and the octopole RF amplitude 120 Vpp. The temperature of the dry gas (N2) was 300°C at a flow of 10 l min−1. The full scan mass spectra of the metabolites were measured from m/z50 to 800 until the ICC target reached 20 000 or 200 ms, whichever was reached first. Tandem mass spectrometry was performed using helium as the collision gas, and the collision energy was set at 1.00 V. Mass spectra were acquired in the negative and positive ionization mode. Auto-tandem mass spectrometry was used to break down the most abundant [M + H]+, [M − H] or the [M + HCOO] ion of the different compounds of the strawberry extracts. The LC parameters went from 0% acetonitrile and 100% water (acidified with 0.05% formic acid) to 50% acetonitrile and 50% acidic water in 50 min, then in 20 min to 100% acetonitrile, remained for 10 min at these conditions, returning to 100% water and 0% acetonitrile in 5 min at a flow rate of 0.2 ml min−1. The detection wavelength was 280 nm.

Relative quantification and statistical evaluation

Fruit juice obtained from untreated strawberries (control), strawberries injected with Agrobacterium transformed with pBI-Intron and strawberries infiltrated with Agrobacterium carrying the pBI-CHSi construct were analysed by LC-UV-ESI-MSn. All fruit were harvested at the same ripening stage. Metabolites were identified by their retention times, mass spectra and product ion spectra in comparison with the data determined for authentic reference material. Signals of the compounds were integrated in their [M + H]+, [M − H] or [M + HCOO] ion traces. Box plots of signal intensities were generated by Sigma Plot (SPSS, Chicago, IL, USA) and statistical significance levels were calculated using the Wilcoxon–Mann–Whitney U-test (Hart, 2001).


We would like to thank Thilo Fischer for helping us with the CHS enzyme assay, Robert Ludwig for offering the A. tumefaciens strain AGL0, Elma Salentijn for the pBINPlusCHSas vector and Erwin Grill for the GUS control construct and a number of other vectors. Heather Coiner is thanked for helpful comments on the manuscript during the proofreading. This work was supported by a grant (SCHW 634/10-1) from DFG.