A diminution in ascorbate oxidase activity affects carbon allocation and improves yield in tomato under water deficit

Authors

  • CÉCILE GARCHERY,

    1. INRA, UR1052, Génétique et amélioration des fruits et légumes, Domaine St Maurice BP94, 84143 Montfavet, France
    Search for more papers by this author
    • Joint authorship.

  • NOÉ GEST,

    1. INRA, UR1052, Génétique et amélioration des fruits et légumes, Domaine St Maurice BP94, 84143 Montfavet, France
    Search for more papers by this author
    • Joint authorship.

  • PHUC T. DO,

    1. Max-Planck-Institute of Molecular Plant Physiology, Am Muehlenberg 1, 14476 Potsdam-Golm, Germany
    Search for more papers by this author
  • MOFTAH ALHAGDOW,

    1. INRA, UMR 1332 Biologie du Fruit et Pathologie, 33140 Villenave d'Ornon, France, and University of Bordeaux, UMR 1332, 33140 Villenave d'Ornon, France
    Search for more papers by this author
    • Present address: Moftah Alhagdow, Faculty of Agriculture, Al-Fateh University, PO Box 13040, Tripoli, Libya.

  • PIERRE BALDET,

    1. INRA, UMR 1332 Biologie du Fruit et Pathologie, 33140 Villenave d'Ornon, France, and University of Bordeaux, UMR 1332, 33140 Villenave d'Ornon, France
    Search for more papers by this author
  • GUILLAUME MENARD,

    1. INRA, UMR 1332 Biologie du Fruit et Pathologie, 33140 Villenave d'Ornon, France, and University of Bordeaux, UMR 1332, 33140 Villenave d'Ornon, France
    Search for more papers by this author
  • CHRISTOPHE ROTHAN,

    1. INRA, UMR 1332 Biologie du Fruit et Pathologie, 33140 Villenave d'Ornon, France, and University of Bordeaux, UMR 1332, 33140 Villenave d'Ornon, France
    Search for more papers by this author
  • CAPUCINE MASSOT,

    1. INRA, UR1115, Plantes et systèmes de culture horticoles, Domaine St Paul, Site Agroparc, 84914 Avignon, France
    Search for more papers by this author
  • HÉLÈNE GAUTIER,

    1. INRA, UR1115, Plantes et systèmes de culture horticoles, Domaine St Paul, Site Agroparc, 84914 Avignon, France
    Search for more papers by this author
  • JAWAD AARROUF,

    1. UAPV Avignon, EA4279 Laboratoire de Physiologie des Fruits et des Légumes, Pôle Agrosciences, BP 21239, 84916 Avignon Cedex 9, France
    Search for more papers by this author
  • ALISDAIR R. FERNIE,

    1. Max-Planck-Institute of Molecular Plant Physiology, Am Muehlenberg 1, 14476 Potsdam-Golm, Germany
    Search for more papers by this author
  • REBECCA STEVENS

    Corresponding author
    1. INRA, UR1052, Génétique et amélioration des fruits et légumes, Domaine St Maurice BP94, 84143 Montfavet, France
      R. Stevens. Fax: +33 4 32 72 27 02; e-mail: rebecca.stevens@avignon.inra.fr
    Search for more papers by this author

R. Stevens. Fax: +33 4 32 72 27 02; e-mail: rebecca.stevens@avignon.inra.fr

ABSTRACT

The regulation of carbon allocation between photosynthetic source leaves and sink tissues in response to stress is an important factor controlling plant yield. Ascorbate oxidase is an apoplastic enzyme, which controls the redox state of the apoplastic ascorbate pool. RNA interference was used to decrease ascorbate oxidase activity in tomato (Solanum lycopersicum L.). Fruit yield was increased in these lines under three conditions where assimilate became limiting for wild-type plants: when fruit trusses were left unpruned, when leaves were removed or when water supply was limited. Several alterations in the transgenic lines could contribute to the improved yield and favour transport of assimilate from leaves to fruits in the ascorbate oxidase lines. Ascorbate oxidase plants showed increases in stomatal conductance and leaf and fruit sugar content, as well as an altered apoplastic hexose : sucrose ratio. Modifications in gene expression, enzyme activity and the fruit metabolome were coherent with the notion of the ascorbate oxidase RNAi lines showing altered sink strength. Ascorbate oxidase may therefore be a target for strategies aimed at improving water productivity in crop species.

INTRODUCTION

Drought is a major factor limiting plant yield as water is a crucial determinant of carbon allocation (Reynolds & Tuberosa 2008; Rosa et al. 2009; Witt et al. 2012). Genes and genotypes that have an impact on productivity under conditions of limiting water are therefore of importance for world food production but a limited number of examples of single genes that improve drought tolerance are present in the scientific literature, an example validated in the field is a transcription factor of the NF-YB family (Nelson et al. 2007). Plant yield per volume of water used is known as water productivity and is considered to be a polygenic trait (Morison et al. 2008) as many processes can contribute to an improvement of yield under stressful conditions (Lopes et al. 2011). Examples of these processes include photoprotection, control of transpiration or water uptake efficiency and control of carbon partitioning between source and sink tissue (Reynolds & Tuberosa 2008).

Carbon partitioning is directly linked to the sugar content of the source and sink tissues, as sucrose may up-regulate genes involved in sucrose hydrolysis and growth in sink organs, but is also regulated by abiotic and biotic stress stimuli, which affect plant growth and yield (Koch 2004; Muller et al. 2011). Transfer of assimilates from source organs, such as mature photosynthetic leaves, to sink organs such as roots, fruits or tubers occurs either via symplastic or apoplastic pathways. The symplastic pathway relies on a concentration gradient of solute between source and sink organs and transport of sucrose via plasmodesmata, whereas the apoplastic pathway involves transport of sucrose or hexoses across the plasma membrane by H+/sucrose transporters or monosaccharide transporters (Ayre 2011). Changes in carbon partitioning or switches between the apoplastic and symplastic pathways occur throughout development (Viola et al. 2001; Godt & Roitsch 2006; Zhang et al. 2006) or as a response to the environment (Roitsch 1999; Geigenberger, Kolbe & Tiessen 2005; Couee et al. 2006; Hermans et al. 2006).

The apoplast also forms the link between the external environment and the cell and thus has roles in cell growth, defence and signal transduction (Pignocchi & Foyer 2003). Many of the signals are a result of redox changes within this compartment, as a consequence of environmental changes or production of reactive oxygen intermediates, and therefore redox buffering of this compartment is essential (Zhang & Guo 2012). Ascorbic acid is considered to be the most abundant antioxidant in the apoplast: the apoplast contains up to 5% of the leaf's ascorbate pool (Veljovic-Jovanovic et al. 2001), a pool which is generally highly oxidized in contrast to the symplastic pool. Regulation of the apoplastic ascorbate pool occurs via transport of ascorbate and dehydroascorbate between the cytosol and the apoplast (Horemans, Foyer & Asard 2000; Pignocchi et al. 2003; Pignocchi & Foyer 2003). Dehydroascorbate is reduced in the symplast by dehydroascorbate reductase (Smirnoff & Wheeler 2000). The activity of ascorbate oxidase, an apoplastic enzyme that catalyses the oxidation of ascorbic acid to monodehydroascorbate, a radical which rapidly degrades to dehydroascorbate, plays a major role in controlling the redox state of the apoplast. Previous studies have shown that this enzyme has a role in the perception of the environment or stress responses (Pignocchi & Foyer 2003; Pignocchi et al. 2006) and may be linked to the cellular ascorbate redox state (defined as the ratio of reduced to total ascorbate). The ascorbate redox state has been shown to control processes including stomatal function (Chen & Gallie 2004; Fotopoulos et al. 2008), cell division (Kato & Esaka 2000; Potters et al. 2002) and changes as a result of stress (Stevens et al. 2008; Gest et al. 2010).

Ascorbate oxidase transcripts are regulated by oxidative stress and also hormones (Esaka et al. 1992; Pignocchi et al. 2003; Sanmartin et al. 2007). Ascorbate oxidase activity is proportional to light intensity in Cucurbita pepo (De Tullio et al. 2007) and tobacco (Pignocchi et al. 2003) and the transcript is highly expressed in roots and young fruits (Sanmartin et al. 2007; Ioannidi et al. 2009). Transgenic plants have shed light on the role of ascorbate oxidase in stress responses and regulation of the apoplastic ascorbate pool; for example, in tobacco underexpressing ascorbate oxidase, the apoplastic redox state was more reduced than in wild-type plants, also there was more total ascorbic acid in the apoplast of antisense lines compared to wild type (Pignocchi et al. 2003). Overexpression of cucumber ascorbate oxidase in tobacco did not change total leaf ascorbate content but reduced the redox state of the apoplast to 3% suggesting that the large increase in apoplastic dehydroascorbate content far exceeded the capacity for its transport to the cytosol (Pignocchi & Foyer 2003). Underexpression of this enzyme often leads to increased stress tolerance such as to salt in Arabidopsis and tobacco (Yamamoto et al. 2005). In this study, the authors also found that seed yield was increased in the antisense plants under stress conditions compared to wild type. Cucumber ascorbate oxidase overexpression in tobacco decreased stress tolerance to ozone (Sanmartin et al. 2003) and increased sensitivity to reactive oxygen species and fungal attack by Botrytis cinerea (Fotopoulos, Sanmartin & Kanellis 2006). Ascorbate oxidase is also a candidate for an ozone tolerance quantitative trait loci (QTL) in rice QTL (Frei et al. 2010). The hypothesis to explain these observations was that the increased redox state of the apoplast of the antisense plants protected against the rise in hydrogen peroxide levels following stress, the reverse being true in the overexpressing plants. In addition, the effect of overexpression of ascorbate oxidase and the increased sensitivity to stress was correlated with the suppressed expression of genes involved in ascorbate acid recycling (Fotopoulos et al. 2006). We show in this study that a reduction in ascorbate oxidase activity in tomato is correlated with increased final fruit yield under unfavourable growing conditions and that changes to both plant physiology and sugar metabolism occur that might explain this phenotype.

MATERIALS AND METHODS

Standard plant growth conditions and sampling

Solanum lycopersicum L. variety West Virginia 106 (WVa106, a cherry tomato) plants were grown in a multispan Venlo-type greenhouse, orientated N-S in 5l pots (potting compost P3 Tref, Tref EGO substrates BV) in either spring (May harvest) or autumn (November harvest) in Southern France. Plant nutrition and chemical pest and disease control were in accordance with commercial practices. Water and nutrients were supplied to the plants using a drip irrigation system to maintain 20–30% drainage (percentage of liquid provided that drains through the pot). Light intensities of 300–700 photosynthetically active radiation (PAR) were obtained over the culture period; a maximum of 700 PAR was obtained on sunny days. Flowers were mechanically pollinated three times a week and side shoots removed as they appeared. Plant material was harvested at solar noon on a sunny day unless otherwise stated. Harvested material was immediately frozen in liquid nitrogen and stored at −80 °C. Prior to the molecular and biochemical analysis, plant material was ground in liquid nitrogen. For all physiology experiments, a minimum of five plants, making five biological replicates, per genotype and per condition were used. The leaf apoplast experiments shown in Fig. 6 were carried out on the same plants, same leaf age on two dates 1 week apart (date 1: 6 October 2009; date 2: 30 September 2009).

Adaptation of growing conditions to limit assimilate production

To test the impact of increased fruit load, trusses were either left unpruned (a maximum of 10 fruits per truss) or pruned to five fruits per truss once fruit had set. For the leaf removal experiment, alternate leaves were removed once fruit were set on trusses one to six. For plants grown under water limitation, drip irrigation (water and nutrients) was reduced by one third (to 0% drainage) from the appearance of the first truss until the end of the growing season. For each experiment, a minimum of five plants per condition per genotype were used.

RNA extraction and semi-quantitative RT-PCR

RNA was extracted from frozen tomato powder (leaves or fruit) conserved at −80 °C using TRI Reagent™ (Euromedex, Souffelweyersheim, France) according to the manufacturer's instructions. Contaminating DNA was removed by treatment with RQ1 RNAse free DNAse (Promega, Madison, WI, USA). Reverse transcription was performed with 5 µg of total RNA with Superscript III Reverse Transcriptase (Invitrogen, Carlsbad, CA, USA), according to the manufacturer's instructions. For semi-quantitative RT-PCR, 2 µL of undiluted cDNA was used for PCR in a final volume of 20 µL ensuring that for the control, the amount of amplified product was proportional to the initial concentration of template present in the reaction, in this case after 20 cycles. After electrophoresis on a 1% agarose gel and blotting onto Hybond N+ membrane, hybridization was performed with 32P labelled probes, labelled using the random primer method. Hybridizations were performed at 65 °C with actin (SGN-U580422) and ascorbate oxidase probes corresponding to the respective coding sequences.

Ascorbate oxidase assays

Total ascorbate oxidase activity was measured based on a previously described protocol (Pignocchi et al. 2003). Leaf material (approximately 50 mg) was homogenized with 0.1 m sodium phosphate pH 5.6, 0.5 mm ethylenediaminetetraacetic acid (EDTA), 1 m NaCl and 100 µL of spun extract was assayed rapidly at 265 nm in a final volume of 1 mL containing 0.1 mm ascorbate in the same phosphate buffer. The absorbance decrease at 265 nm was followed on addition of the extract. Protein assays were carried out in triplicate using the method of Bradford (1976). Bovine serum albumin was used as a standard.

Construction of the RNAi ascorbate oxidase plasmid

The ascorbate oxidase gene identified corresponds to the accession AY971876 (SGN-U581990), which has previously been mapped in tomato (Zou et al. 2006; Stevens et al. 2007). BLAST analysis revealed the presence of one further ascorbate oxidase genes in tomato: SGN-U581988. A fragment of 503 bp, starting 2 bp upstream of the ATG start codon, was amplified from tomato first strand cDNA by PCR using the following primers:

  • AOF2: 5′-AAAAAGCAGGCTACATGGTTGAGCATGATTTTCATC

  • AOR3: 5′-AGAAAGCTGGGTATGATCATATGAAAATGGCTCTAA

The PCR product obtained was cloned into vector pDONR™201 (Invitrogen, LifeTechnologies, Saint Aubin, France) (BP reaction, Gateway according to the manufacturer's instructions) and afterwards into the destination vector pK7WIWG2(1),0 (Karimi, Inze & Depicker 2002; LR reaction, Gateway according to the manufacturer's instructions). This construct was used to transform Agrobacterium strain GV3101.

Tomato transformation and selection of RNAi lines

WVa106 tomato cotyledons were transformed based on a previously described method (Hamza & Chupeau 1993). Leaf tissue was tested for ploidy by flow cytometry (by use of a Ploidy Analyser, Partec, Münster, Germany, according to the manufacturer's instructions) and for the presence of the transgene by PCR. Plants not containing the transgene and non-diploid plants were eliminated. Segregation analysis showed that lines AO15 and AO42 had only one copy of the transgene whereas AO16 had a least two copies.

Photosynthesis, transpiration and stomatal conductance

Net CO2 assimilation rates (photosynthesis), transpiration and stomatal conductance were measured using an LCA4 (Analytical Development Corporation, ADC BioScientific Ltd., Hoddesdon, UK) at ambient light levels. Measurements were carried out on three leaves per plant of a minimum of five plants per line.

Measurement of pericarp cell size and cell number

Fruit harvested 3 days post-anthesis (dpa) were measured with an electronic sliding scale before being immersed in a cold fixative (formalin-acetic acid-alcohol at 1:1:8 v/v/v) for 48 h at 4 °C. Fruits were then rinsed in distilled water, dehydrated in alcohol (25-50-70-85 and 100%) and embedded in resin (Kit Technovit 7100, Wehrheim, Germany). Sections (5 µm thickness) were serially cut with a Leica microtome (Reichert-Jung Leica instruments, Nussloch, Germany), collected on microscope slides and allowed to dry (Aarrouf et al. 2008). Pericarp sections were stained with 0.5% toluidine blue and observed with a light microscope (Optiphot-2, Nikon). The photographic images were captured using a Leica automatic camera. Images were analysed with ImageJ software (US National Institutes of Health, Bethesda, MD, USA) and the number of cells in a given area counted. A minimum of five fruits per genotype were analysed. Five sections were analysed from each fruit.

Measurement of fruit growth rate and final diameter

Fruit expansion was measured twice weekly using electronic callipers from 10 dpa throughout the ripening period on six fruits per plant with a minimum of three plants per line. Fruit growth versus time was described with a logistic curve (Gautier, Varlet-Grancher & Membre 2001). This empirical function D(t) provides the quantitative estimation of the final fruit diameter D and the maximal growth rate µ.

image

The parameters of the regression were estimated by the maximum likelihood method using Excel.

Extraction of apoplastic fluid

Extracellular fluid was recovered from three fully expanded and well exposed folioles (midrib removed) or fruit pericarp halves (two to three fruits per plant from which locular gel and seeds were removed) in 30 mL ice-cold 20 mm glycine pH 2.5, 100 mm KCl by vacuum infiltrating for 5 min at −0.1 MPa according to a previously published method (Turcsanyi et al. 2000). The plant material was blotted dry, inserted into a 15 mL tube with a small hole punctured in the base and contained in a larger tube, and the extracellular fluid was collected by centrifuging at 1000g for 10 min at 4 °C. Cytoplasmic contamination of apoplast fractions was measured by assay of the cytosolic marker glucose-6-phosphate in 50 mm HEPES (4-[2-hydroxyethyl]-1-piperazineethanesulfonic acid) pH 7.2, 5 mm MgCl2, 1 mm NADP containing 1 unit of glucose-6-phosphate dehydrogenase. The appearance of NADPH at 340 nm in samples was recorded and compared to a standard curve of glucose-6-phosphate. The contamination was calculated as a percentage of the glucose-6-phosphate in the fruit from which the apoplastic fraction was obtained. Cytosolic contamination of below 5% was recorded independently of the line chosen. Concentrations of other metabolites (ascorbate and sugars) were calculated directly for the extracellular fluid recovered.

Ascorbic acid content

Measurements of ascorbic acid content were carried out as described (Stevens et al. 2006) on material conserved at −80 °C. Extractions and assays were carried out in ice-cold 6% trichloroacetic acid (TCA) in triplicate. The microplate assay used was a spectrophotometric assay based on the detection of dipyridyl-Fe2+ complexes following the reduction of Fe3+ to Fe2+ by the reduced form of ascorbate present in the samples and comparison with standards of known concentrations. Total ascorbate (reduced + oxidized forms) was measured by mixing the sample with 5 mm dithiothreitol (DTT), to reduce dehydroascorbate, prior to the assay. Non-specific background fluorescence was eliminated by measuring the absorbance obtained by replacing the DTT with 0.3 units of ascorbate oxidase. The efficiency of the oxidation was checked on ascorbate standards treated in the same way. This step was important to measure the iron-reducing activity that was not due to ascorbate: in leaf apoplast samples, this was found to be about 30% of total iron-reducing activity but was homogeneous between transgenic and non-transgenic lines.

Gas chromatography–mass spectrometry (GC-MS) analysis

Metabolite extraction, derivatization, GC-MS analysis and data processing were performed as described previously (Roessner et al. 2001; Lisec et al. 2006). Metabolites were identified in comparison to database entries of authentic standards (Kopka et al. 2005; Schauer et al. 2005). Leaf samples corresponded to fully expanded but non-senescent leaves. Red fruits were harvested 40 dpa and green fruits 14 dpa.

Sucrose, fructose and glucose assays

Sugars were quantified using enzymatic-based assays in 96-well microplates or by high-performance liquid chromatography (HPLC) according to Gomez et al. (2007) except that extractions were carried out on non-lyophilized material.

Real-time quantitative PCR

A dilution series of pooled cDNAs, obtained as described above, was used to provide standards to calculate the efficiency of the reaction for each set of primers tested. qPCRs were carried out in a final volume of 25 µL containing Master Mix GoTaq (Promega SYBER Green), 10 µm primers and 2 µL of 1:5 diluted cDNA using standard conditions in a Stratagene Mx3005P® Thermocycler (Stratagene, Cedar Creek, TX, USA). A thermal denaturation curve of the amplified DNA was used to measure the melting temperature of the PCR product. Relative gene expression was calculated by the 2−ΔΔCt method (Livak & Schmittgen 2001), with the genes encoding translation initiation factor eIF-4A and actin as internal constitutive controls. Primer sequences for qPCR reactions are found in Supporting Information Table S1.

Assay of enzymes involved in sugar metabolism

Measurement of adenosine diphosphate (ADP) glucose pyrophosphorylase, acid and neutral invertase and sucrose phosphate synthase activities were carried out on leaves and fruit as previously described (Gibon et al. 2004).

RESULTS

Ascorbate oxidase gene expression and enzyme activity in tomato

An ascorbate oxidase gene was identified in tomato (SGN-U581990) based on sequence homology with the protein sequence available in GenBank: (AY971876; Zou et al. 2006). We examined the expression of this gene in different organs of tomato (Fig. 1a). Expression was high in roots, 7-day-old fruits and fruit pericarp but relatively low in other fruit samples and leaves, stems and flowers. Analysis of enzyme activity revealed similar results; activity was high in young tomato fruit and lower in leaves and ripe fruit (Fig. 1b).

Figure 1.

Ascorbate oxidase transcript levels and enzyme activity in tomato tissues. (a) Semi-quantitative RT-PCR on total RNA from mature leaves (ML), young leaves (YL), roots (R), shoots (S), flowers (Fl), young fruits 7 days post-anthesis (dpa; 7 d), 14 dpa (14 d), mature green fruits (MG), orange fruits (O), red fruits (R) or pericarp tissue of mature green, orange and red fruits. Transcript levels of the ascorbate oxidase gene (AO) compared to an actin constitutive control are shown. (b) Specific total ascorbate oxidase activity in extracts of young leaves, whole young fruit 7 dpa and whole ripe fruit. Bars represent standard errors, a minimum of five independent samples per tissue type were used.

RNAi tomato lines have reduced ascorbate oxidase transcript and activity in leaves

To characterize the role of ascorbate oxidase in tomato, we generated lines with decreased ascorbate oxidase activity by using an RNA interference (RNAi) strategy. Ten independent RNAi lines were generated, none of which were immediately clearly distinguishable from wild type. Of these 10 T0 lines, three (AO15, AO16 and AO42) were chosen for more detailed analysis in the T1 and T2 generations. Initial evaluation of lines involved verification that transcript levels and enzyme activity were reduced. Results are shown in Fig. 2. Transcript levels are reduced in leaves of independent plants of all three lines (Fig. 2a). The ascorbate oxidase activity of leaves was also measured (Fig. 2b). All lines showed a significant reduction in ascorbate oxidase activity, the strongest reduction being observed for line AO42 (60% of wild-type activity) and the lowest reduction for line AO16 (80% of wild-type activity).

Figure 2.

Tomato ascorbate oxidase RNAi lines have reduced transcript levels and reduced ascorbate oxidase activity levels. (a) Transcript levels of the ascorbate oxidase gene (AO) compared to an actin constitutive control. Semi-quantitative RT-PCR on total RNA from young leaves of four independent plants from AO15 (AO15.1–4), AO16 (AO16.1–4) and AO42 (AO42.1–4) with two wild-type (WT) controls. (b) Specific total ascorbate oxidase activity in wild type, AO15, AO16 and AO42 young leaves. Bars represent standard errors, a minimum of five independent samples per tissue type were used. Statistical differences (Kruskal and Wallis non-parametric test) between wild-type and transgenic lines are indicated by a star (P-value <0.05).

Ascorbate oxidase RNAi lines show similar plant growth to wild type

As ascorbate oxidase activity is potentially correlated with plant growth, we monitored growth of transgenic lines: final plant dry and fresh weight were not significantly different between the RNAi ascorbate oxidase lines and wild type (Fig. 3) under both normal conditions or stress conditions. Under the stress condition chosen, where water input to the plants was limited by one third (see Materials and Methods), the decreased irrigation schedule led to an overall reduction in dry plant biomass of 40%.

Figure 3.

Final plant dry weight is not changed under normal greenhouse conditions (control) or when drip irrigation is reduced by one third (stress). Final plant dry weight (aerial parts without fruits) is shown for a minimum of 10 plants per genotype grown under normal greenhouse conditions or grown under a stress condition where irrigation was reduced by one third (see Materials and Methods). Plants were harvested at the end of the growing season (18 weeks after sowing). Bars represent standard errors. Results with the same superscript were not significantly different (P < 0.05) according to the classification obtained by the Tukey test.

Early fruit size is affected in ascorbate oxidase RNAi lines

Fruit size was monitored in the transgenic lines to examine the effects of reduced ascorbate oxidase activity on fruit development. Early fruit growth was measured 3 dpa and was significantly increased for two of the three transgenic lines (AO15 and AO42) compared to wild type and empty vector transformed controls (Fig. 4a). Given the increase in size of young fruit, we checked whether this was correlated with an increase in cell number or cell size. Fruit pericarp cell number and size for lines AO15 and AO42 were compared to wild-type values. The increase in fruit size seen in ascorbate oxidase RNAi lines is correlated with an increase in pericarp cell size (Fig. 4b), which is significant for line AO15.

Figure 4.

Early fruit growth is increased in ascorbate oxidase RNAi lines. (a) Flowers were tagged at anthesis and fruits harvested 3 d later. The diameter of at least 50 fruits per genotype was measured for wild-type (WT) and empty vector transformed lines (EV) as well as the three ascorbate oxidase RNAi lines. (b) Pericarp cell area and number 3 days post-anthesis (dpa). The left panel shows an example of a 3 dpa fruit (false colours, bar represents 1 mm), the right panel the pericarp cells (bar represents 100 µm). EV is a transformed control. Cell area and cell number were counted using ImageJ software, data for wild-type fruits were normalized to 100 and the other genotypes compared to the normalized data. At least five fruit were analysed from each genotype and between four and six sections per fruit. Stars indicate statistical differences between wild-type and transgenic lines (*P < 0.05; **P < 0.01, ***P < 0.001) using a standard analysis of variance parametric test.

Final fruit growth or yield can be altered depending on environmental conditions

The increase in fruit size at 3 dpa led us to examine the effect of ascorbate oxidase RNAi on final fruit size under both normal conditions and three conditions where assimilates may become limiting (trusses left unpruned, plants grown under reduced water or mature leaves removed; see Materials and Methods).

In the first case, fruit trusses were allowed to grow normally (a maximum of 10 fruits per truss) or pruned to five fruits per truss (to fix the number of fruit per plant and thus limit competition between fruits for assimilates) and the final maximum fruit diameter and maximum growth rate were calculated. When pruned and unpruned trusses were compared, wild-type fruits grown under pruned conditions were significantly larger than fruits from plants whose trusses were not pruned (Fig. 5a). However, in the case of the ascorbate oxidase RNAi lines, no significant differences were found in terms of final fruit diameter under pruned or unpruned conditions (Fig. 5a). No significant differences were found in terms of fruit growth rate under pruned or unpruned conditions for any of the lines.

Figure 5.

Fruit yield in ascorbate oxidase RNAi lines and wild type under conditions of limiting assimilate concentration (competition between fruits, reduced watering or leaf removal). (a) Wild type (WT), but not RNAi ascorbate oxidase, fruits benefit from limitation of fruit number. The unpruned condition corresponds to 10 fruits per truss and for the pruned condition fruit number was limited to five per truss. Fruits from three unpruned trusses per plant were compared with fruits from three pruned trusses per plant. A minimum of three plants per genotype and per condition were used. Stars show statistical differences intra-line between pruned and unpruned conditions (*P < 0.05). (b) Average total fruit yield per plant from a minimum of 10 plants per genotype grown under a normal watering regime (control) and a reduced watering regime – reduction of one third compared to a normal watering regime (stress; see Materials and Methods). Results with the same superscript were not significantly different (P < 0.05) according to the classification obtained by the Tukey test. (c) Effects of leaf pruning on total fruit yield per plant for the ascorbate oxidase and wild-type lines. For the leaf removal experiment, alternate leaves were removed once fruit were set. A minimum of five plants per condition per genotype were used. Results with the same superscript were not significantly different (P < 0.05) according to the classification obtained by the Tukey test. Only wild type shows significant difference between normal and conditions where leaves are removed.

The advantage presented by ascorbate oxidase RNAi lines under unpruned conditions and during early fruit growth could be due to an increased availability of assimilates for fruit growth in these lines. To test this hypothesis, ascorbate oxidase RNAi lines were grown under both normal conditions and conditions where assimilate availability was limited either by reducing water input (a reduction of one third in the drip irrigation bringing both water and fertilizer to the plants, from the appearance of the first flowers until the end of the growing season) or by reducing leaf number (removal of alternate mature photosynthetic leaves once fruit set). Under normal conditions, total fruit yield was not significantly different between wild-type and the three ascorbate oxidase lines (Fig. 5b). However, when plants were grown with decreased irrigation, the total fruit yield of wild-type plants decreased by 30% whereas the fruit yield of ascorbate oxidase lines decreased by 7% (AO15), 9% (AO16) and 0% (AO42). Ascorbate oxidase RNAi lines showed similar improved yield compared to wild type under other conditions that limited assimilate production, such as when 50% of photosynthetic leaves were removed from plants (Fig. 5c): only wild-type plants show a significant decrease in fruit yield following leaf removal.

The increased early fruit size in RNAi ascorbate oxidase lines, the decreased benefit of pruning on fruit size and the improved yield following reduction of irrigation led us to think that the source–sink relationship may be altered in ascorbate oxidase RNAi plants and so we investigated parameters involved in apoplastic metabolism, carbon fixation and sugar levels and in both leaves and fruit.

Total and apoplastic ascorbic acid levels in leaves and fruit and effects of ascorbate oxidase on the apoplastic redox state

Reducing the activity of the ascorbate oxidase enzyme would be expected to increase the concentration of the reduced form of ascorbate and thus the redox state (defined as the ratio of reduced to total ascorbate) of the apoplastic pool. We therefore assayed ascorbate content in leaves and fruit, both in whole tissues and apoplastic fluid. Total and reduced ascorbate levels for whole tissues (leaves and fruit) of the lines are shown in Table 1. One of the three lines, AO15, showed significantly increased total ascorbate levels in leaves. The redox state of the two tissues was not significantly changed in whole leaves or fruit (Table 1).

Table 1.  Total and reduced ascorbic acid levels and redox state in leaves and fruit of wild-type and ascorbate oxidase lines
 LineTotal AA mg/100 gfwtSD P Reduced AA mg/100 gfwtSD P RedoxSD P
  1. Leaf samples: a minimum of three folioles from fully expanded leaves from five plants per line (five replicates; 12-week-old plants). Fruit samples: a minimum of three orange fruit per plant, five plants per line (five replicates). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05; Kruskal and Wallis non-parametric test).

  2. gfwt, gram fresh weight; ns, comparison with wild type non-significant.

 WT1056 9210 0.880.07 
Whole leafAO1512421 * 10821ns0.870.08ns
AO161126ns9617ns0.860.12ns
AO4210610ns9513ns0.890.05ns
 WT335 306 0.910.07 
Whole fruitAO15314ns294ns0.930.04ns
AO16324ns304ns0.920.02ns
AO42283ns264ns0.920.03ns

The apoplastic redox state in leaves was found to fluctuate in leaves of tomato plants. Under certain conditions (samples harvested at solar noon on a sunny day; date 1; 6 October 2009), leaf apoplastic ascorbate redox state did not differ significantly between wild-type and transgenic lines (Fig. 6a, pale bars). A second experiment carried out on the same plants (samples also harvested at solar noon on a sunny day; date 2; 30 September 2009) revealed differences in the leaf apoplastic ascorbate redox state between transgenic and wild-type plants (Fig. 6a, dark bars) for lines AO16 and AO42, the line AO42 showing the greatest difference compared to wild type. When the environmental data recordings from the greenhouse were checked, the only measurable difference between the two sample dates was a difference of ambient hygrometry (55% for date 1 and 35% for date 2), indicating that plants may have been under water stress on the second date. The total and reduced forms of ascorbate are presented in Fig. 6b for the second date and only show significant differences in total ascorbate (line AO15, an increase) or an increase in reduced ascorbate (line AO16). The redox state of the fruit apoplastic ascorbate pool tended to be more reduced in transgenic lines, the difference being significant for fruit of the lines AO15 and AO42 (Fig. 6c; results from four independent experiments) although total and reduced ascorbate levels were not significantly different to wild-type fruit (Fig. 6d).

Figure 6.

Apoplastic ascorbate levels and redox state in leaves and fruit of wild-type and ascorbate oxidase RNAi lines. (a) Leaf apoplastic ascorbate redox state is shown on two dates for the same plants, leaf stage and harvest hour. Values with standard errors are for a minimum of five independent plants per genotype, three mature folioles from 12-week-old plants were harvested per plant at solar noon on a sunny day. Date 1: 6 October 2009 (pale bars); Date 2: 30 September 2009 (dark bars). (b) Total and reduced apoplastic ascorbate concentrations for the samples shown in a (date 2). (c) Fruit apoplastic ascorbate redox state (mature green fruit). Values with standard errors are for a minimum of five independent plants per genotype. 1–2 fruit of the correct developmental stage were harvested per plant at solar noon on a sunny day. Results are from four independent experiments. (d) Total and reduced apoplastic ascorbate concentrations for the samples shown in (c). Stars indicate statistical differences between wild-type and each transgenic line according to a Kruskal and Wallis non-parametric test (*P < 0.05; **P < 0.01).

Stomatal conductance is increased in transgenic ascorbate oxidase lines

Stomata affect carbon fixation as they are the control point for the entrance of carbon dioxide into the plant. We therefore carried out measurements of stomatal conductance, transpiration and photosynthesis on wild-type and transgenic ascorbate oxidase RNAi lines grown under normal conditions. Ascorbate oxidase transgenic lines show differences in the measured parameters compared to wild-type plants (Table 2). Stomatal conductance was significantly increased in both AO15 and AO42 lines and transpiration was significantly increased in the AO42 lines.

Table 2.  Stomatal conductance, transpiration and photosynthetic rates of leaves of wild-type and ascorbate oxidase lines at ambient light levels under normal watering conditions in the greenhouse
LinePAR measured (µmol m−2 s−1)SD P Stomatal conductance (mol m−2 s−1)SD P Transpiration (mmol m−2 s−1)SD P Photosynthetic rate (µmol CO2 m−2 s−1)SD P
  1. Net CO2 assimilation rates (photosynthesis), transpiration and stomatal conductance were measured at ambient light levels. Measurements were carried out on three leaves per plant on 12-week old plants of a minimum of five plants per line, for the indicated light levels (PAR, photosynthetically active radiation). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05) or ** (P < 0.01; Kruskal and Wallis non-parametric test).

  2. ns, comparison with wild type non-significant.

WT34245 0.50.3 4.11.2 8.51.3 
AO1533741ns0.90.5 * 4.60.8ns9.01.2ns
AO4237849ns2.41.3 ** 6.00.4 ** 9.61.5ns

Ascorbate oxidase RNAi plants have increased hexose and sucrose levels in leaves and fruit and an increased apoplastic hexose : sucrose ratio

The increase in stomatal conductance as well as the improved yield and increased early fruit size of the ascorbate oxidase RNAi plants led us to examine sugar levels in the plants. In leaves of transgenic ascorbate oxidase lines, both hexose and sucrose levels were significantly increased (Table 3) and a similar increase in hexose concentration was found in fruit (Table 3). Transport of sugars from leaves to source organs can either occur via symplastic or apoplastic pathways. In the case of an apoplastic pathway, it is the apoplastic hexose : sucrose ratio that will affect the sucrose concentration gradient. As ascorbate oxidase is an apoplastic enzyme, we measured apoplastic sugar levels in wild-type and ascorbate oxidase RNAi lines. As stated previously, the redox state of the apoplastic pool was less oxidized in the three ascorbate oxidase RNAi lines under certain conditions (Fig. 6). The differences in the apoplastic hexose : sucrose ratio for the three ascorbate oxidase lines are shown in Fig. 7a (leaves; data from date 2) and 7b (fruits) and Table 4 (leaves) and Table 5 (fruit). In both fruit and leaves, the apoplastic hexose : sucrose ratio increases in ascorbate oxidase transgenic lines compared to wild type. The variability found in the apoplastic ascorbate redox state has been correlated with hexose : sucrose, hexose and sucrose levels for leaves of individual plants in four independent experiments. The apoplastic ascorbate redox state is best correlated with the hexose : sucrose ratio (R2 = 0.43, P = 1 × 10−5; Fig. 7c) and the hexose concentration in the leaf apoplast but not the apoplastic sucrose concentration (Table 4). Whereas the increased leaf apoplast hexose : sucrose ratio resulted from an increase in apoplastic leaf hexoses, the increased fruit apoplast hexose : sucrose ratio resulted from a decreased sucrose concentration (Table 5).

Table 3.  Hexose and sucrose concentrations in leaves and fruit of wild-type and ascorbate oxidase RNAi lines
  Glucose (g/100 g FW)SD P Fructose (g/100 g FW)SD P Sucrose (g/100 g FW)SD P
  1. Each sample represents a minimum of three folioles from fully expanded leaves from five plants per line (five replicates; 12-week-old plants) or for fruit: three orange fruit per plant, five plants per line (five replicates). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05) or ** (P < 0.01; Kruskal and Wallis non-parametric test).

  2. nd = not detectable (sucrose concentrations in ripe tomato fruit are extremely low in most varieties); ns: comparison with wild type non-significant.

 WT0.530.11 0.570.16 0.660.17 
LeafAO150.660.14 ** 0.710.15 * 0.820.22 *
AO420.610.14ns0.700.13 * 0.860.26 *
 WT1.240.12 1.440.18 nd  
Orange fruitAO151.350.12ns1.350.11nsnd  
AO161.370.08 * 1.600.12nsnd  
AO421.430.14 * 1.630.21nsnd  
Figure 7.

Hexose : sucrose ratios in the leaf and fruit apoplast of WT and RNAi ascorbate oxidase lines. (a) Leaf apoplastic hexose : sucrose ratio under conditions of changing redox (date 2; Fig. 6a), samples harvested as described in Fig. 6a legend. A minimum of five independent samples per genotype were used. (b) Fruit apoplastic hexose : sucrose ratio. A minimum of five independent samples per genotype were used, samples were harvested as described in the legend for Fig. 6c. (c) Correlation between leaf apoplastic ascorbate redox state and leaf apoplastic hexose : sucrose ratio. Data is obtained from WT, AO15, AO16, AO42 and empty vector transformed (EV) lines: five plants per line, two independent experiments. An R2 of 0.4266 is obtained with a P-value of 1 × 10−5. An equally good correlation is obtained between leaf apoplastic hexose and leaf apoplastic ascorbate redox but no correlation is obtained between leaf apoplastic sucrose levels and ascorbate redox. Stars indicate statistical differences between wild-type and transgenic lines according to a Kruskal and Wallis non-parametric test (*P < 0.05; **P < 0.01; ***P < 0.001).

Table 4.  Leaf apoplastic hexose and sucrose concentrations for wild-type and transgenic ascorbate oxidase RNAi lines
 Glucose (g L−1)SD P Fructose (g L−1)SD P Sucrose (g L−1)SD P
  1. Each sample represents a minimum of three folioles per plant from five plants per line (five replicates). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05) or ** (P < 0.01; Kruskal and Wallis non-parametric test). ‘Redox stable’ corresponds to the samples in Fig. 6a, date 1. ‘Redox increase’ to the samples in Fig. 6a, date 2.

  2. Ns, comparison with wild type non-significant.

Redox stable         
 WT0.350.06 0.290.08 0.830.10 
 AO150.350.05ns0.280.04ns0.940.12ns
 AO160.440.06ns0.390.08ns0.860.05ns
 AO420.430.03ns0.380.08ns1.140.11ns
Redox increase         
 WT0.320.11 0.160.11 1.380.40 
 AO150.580.18 * 0.480.21 * 1.640.54ns
 AO160.410.13ns0.300.13ns1.380.29ns
 AO420.370.13ns0.320.17ns1.110.36ns
Table 5.  Fruit apoplastic hexose and sucrose concentrations for wild-type and transgenic ascorbate oxidase RNAi lines
 Glucose (g L−1)SD P Fructose (g L−1)SD P Sucrose (g L−1)SD P Hexose: sucrose ratioSD P
  1. Each sample represents a minimum of two mature green fruit per plant from five plants per line (five replicates). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05; Kruskal and Wallis non-parametric test). ns, comparison with wild type non-significant.

WT9.451.45 9.161.52 0.690.09 27.204.33 
AO158.580.86ns8.410.67ns0.540.13ns33.048.76ns
AO168.291.73ns8.101.68ns0.550.10 * 30.438.33ns
AO428.180.86ns8.240.78ns0.420.03 * 39.755.71 *

Ascorbate oxidase RNAi plants show broader changes in the fruit metabolome

In order to learn more about the changes in sink metabolism of the ascorbate oxidase RNAi lines, we used a GC-MS protocol (Fernie et al. 2004) to measure the levels of 52 metabolites including amino acids, TCA cycle intermediates, organic acids and primary metabolites in leaves, green and red fruit from both transgenic and wild-type plants (Supporting Information Tables S2–S4). The levels of metabolites in the leaves of the transgenics were similar to those observed in the wild type (Supporting Information Table S2). In contrast, the levels of metabolites in the green fruits of the transformants showed significant increases in several amino acids (valine, asparagine, glutamine and pyroglutamic acid), whereas only alanine and beta-alanine decreased in the lines (Supporting Information Table S3). Changes in organic acid levels in green fruit also occurred, both lines displayed significant reductions in the level of succinic and benzoic acid and levels of nicotinic acid, glucuronic acid, phosphoric, maleic, nonanoic and threonic acids decreased in one or both lines. Both lines also showed a decrease in myo-inositol levels. The metabolite profiles of the transformants from red fruit also varied from wild type, although, as seen previously (Carrari et al. 2006; Do et al. 2010), were unpredictable from the changes in green fruit. In red fruit, both transgenic lines displayed decreased levels of beta-alanine and γ-aminobutyrate (GABA; Supporting Information Table S4) but far fewer changes were observed in levels of the organic acids. As seen in green fruits, myo-inositol was decreased in both transgenic lines.

Expression of genes involved in sugar metabolism and the source–sink relationship are altered in ascorbate oxidase lines

Both redox state and sugars have important signalling functions and may affect gene expression. We therefore looked at gene expression in leaves and fruit of wild type, AO15 and AO42 plants. The genes investigated have either been previously identified as responding to source–sink changes (Baldet et al. 2006; Prudent et al. 2010), or had been found to be under- or overexpressed in fruit in a transcriptome study (Garcia et al. 2009). Results are shown in Table 6. Similarly to the metabolite data, more changes were seen in fruit gene expression than in leaf gene expression. In leaves, decreases in uridine diphosphate (UDP)-galactose-epimerase expression [an enzyme involved in channelling intermediates into cell wall biosynthesis; (Dormann & Benning 1998; Seifert et al. 2002)] and decreases in β-amylase expression were only seen in the most severely affected line, AO42, compared to wild type. In fruit, a similar decrease in UDP-galactose-epimerase expression is seen in line AO42. However, in contrast to the leaves, an overall increase is seen in β-amylase gene expression in fruits from both lines. Starch phosphorylase expression is significantly increased for AO42 compared to wild type. Notably in fruit, the expression of two auxin-responsive genes is increased in transgenic lines and the expression of cell division regulator fw2.2 (Dahan et al. 2010; Guo et al. 2010) is significantly increased for AO42 compared to wild type. Finally in fruits, the expression of several sucrose phosphate synthase isoforms is increased in transgenic lines and expression of lin6– a cell wall invertase – is significantly increased for AO42 (Proels & Roitsch 2009). Finally, although ascorbate oxidase transcripts are decreased in leaves (significantly for the line AO15), significant ascorbate oxidase transcript reduction is not detected in orange fruit implying that gene regulation or RNAi in fruit tissue may be complex.

Table 6.  Expression of genes involved in sugar metabolism, fruit growth or the source–sink relationship in leaves and fruit of ascorbate oxidase lines compared to wild type
  AO15 leafAO42 leafAO15 fruitAO42 fruit
SGN unigeneDescriptionRatio WT P-valueRatio WT P-valueRatio WT P-valueRatio WT P-value
  1. The ratios shown represent ascorbate oxidase transgenic/wild type.

  2. Leaf samples: a minimum of three folioles from fully expanded leaves from five plants per line (five replicates; 12-week-old plants). Fruit samples: a minimum of three orange fruit per plant, five plants per line (five replicates). P-values are given according to a Kruskal and Wallis non-parametric test.

U578305Acid invertase0.50.080.90.641.60.490.80.82
U580029Acid invertase1.00.341.20.810.50.311.10.74
U578195Acid invertase0.90.311.30.681.10.931.60.65
U569800ADP glucose pyrophosphorylase LSU11.20.891.00.872.00.310.50.58
U581813ADP glucose pyrophosphorylase LSU21.20.580.90.781.00.703.20.15
U581990Ascorbate oxidase0.10.010.40.522.10.303.10.16
U579749Auxin regulated protein IAA10.40.340.60.281.70.091.30.58
U575393Auxin responsive protein0.80.550.40.154.30.0023.00.01
U577967Auxin responsive protein0.90.760.40.211.60.333.00.004
U579298Beta-amylase0.90.590.30.0021.70.010.020.001
U576412Beta-amylase0.70.300.40.082.10.253.20.12
U579102Beta-fructofuranosidase (invertase)0.90.270.60.350.90.242.60.14
U583476cyclinD31.90.011.10.691.50.234.70.09
AF261774fw2.20.50.090.90.842.20.248.30.02
U584136lin6, cell wall invertase1.10.410.50.683.50.114.60.04
U568728Starch phosphorylase2.20.850.60.633.40.135.50.002
U567040Sucrose phosphate synthase1.70.920.50.331.10.234.30.15
U568014Sucrose phosphate synthase1.20.350.50.325.00.075.70.05
U582735Sucrose phosphate synthase1.00.350.80.632.70.195.30.17
U601027Sucrose phosphate synthase4.80.370.70.480.80.0012.80.22
U577970Sucrose synthase0.80.711.10.640.70.820.40.84
U581233Sucrose transporter sut10.80.641.10.940.90.752.00.11
U564745UDP galactose-epimerase0.40.100.30.0041.30.120.10.001
U579867UDP glucose-pyrophosphorylase2.10.150.60.561.00.542.80.23

Activities of enzymes involved in the regulation of sugar metabolism are altered in ascorbate oxidase lines

Several enzymes are known to have regulatory roles in the control of sugar metabolism, for example, ADP glucose pyrophosphorylase (entry into starch synthesis), acid and neutral invertases (degradation of sucrose into hexoses) and sucrose phosphate synthase (entry into sucrose synthesis). The activity of these four enzymes was measured in leaves and fruit of wild-type and ascorbate oxidase RNAi lines, results are shown in Table 7. Neither ADP glucose pyrophosphorylase activity nor acid invertase activities show differences between the different lines. Acid invertases include cell wall and vacuolar activities. However, neutral invertase activity (representing cytosolic invertases) increased significantly for two of the three transgenic lines in leaves compared to wild type but not in fruits. Furthermore, sucrose phosphate synthase activity showed highly significant differences in the transgenic lines compared to wild type, its activity decreased in transgenic fruits of all three lines and increased in transgenic leaves for two of the three lines.

Table 7.  Activity of enzymes involved in the regulation of sugar metabolism in leaves and fruit of ascorbate oxidase lines and wild type
  ADP glucose pyro-phosphorylase (nmol g−1 min−1)SD P Acid invertase (nmol g−1 min−1)SD P Neutral invertase (nmol g−1 min−1)SD P Sucrose phosphate synthase (nmol g−1 min−1)SD P
  1. Sucrose synthase activity was not detected in leaves or fruits.

  2. Leaf samples: a minimum of three folioles from fully expanded leaves from five plants per line (five replicates; 12-week-old plants). Fruit samples: a minimum of three orange fruit per plant, five plants per line (five replicates). Values with standard deviations are shown. Significant differences as compared to wild type are indicated by * (P < 0.05) or ** (P < 0.01; Kruskal and Wallis non-parametric test). ns, comparison with wild type non-significant.

 WT2408.5 51938 147081 1109 
LeafAO152446.2ns51025ns150180ns794 *
AO1625013.5ns54327ns1699111ns24737 **
AO422258.1ns50730ns193937 ** 18615 **
 WT5.80.4 125447 19051364 1809 
FruitAO156.20.3ns120933ns18392255ns1365 **
AO166.60.6ns118739ns18153277ns13612 *
AO425.40.3ns126022ns17862501ns13117 *

DISCUSSION

Ascorbate oxidase RNAi plants have increased stomatal conductance, leaf and fruit sugar content in the symplastic space and an increased hexose : sucrose ratio in the apoplastic space and increased fruit yield under unfavourable growing conditions. The changes seen in gene expression, enzyme activity and the fruit metabolome confirmed that alterations in carbon allocation were occurring in the transgenic plants. The pleiotropic nature of these effects means that deciding which effect is at the origin of the effect on yield is relatively complex. However, a number of important conclusions can be drawn, including the fact that a change in the activity of a single apoplastic enzyme affects allocation of carbon between source and sink organs.

Control of carbon allocation between source and sink tissues

The increased stomatal conductance seen in ascorbate oxidase RNAi lines should improve entry of carbon dioxide into leaves and may improve carbon fixation, a consequence of which could be increased sugar levels in photosynthesizing tissues under conditions such as high carbon dioxide or drought, which would favour stomatal closure in wild type. Links between the ascorbate redox state and stomatal opening have already been shown (Chen & Gallie 2004; Fotopoulos et al. 2008). The changes in symplastic or apoplastic sucrose and hexose concentrations seen in this study will also affect carbon allocation within the plant by altering solute gradients. For example, sucrose levels have been shown to be negatively correlated with the rate of carbon import into fruit (Klann, Hall & Bennett 1996) and so the low sucrose levels in the fruit apoplast seen in this study are likely to be why the ascorbate oxidase RNAi fruits download sucrose from the phloem more efficiently. The sucrose is then converted rapidly to hexoses in the ripening fruit, explaining the increased hexose content of ascorbate oxidase RNAi fruit. This hypothesis implies a major role for apoplastic transfer in tomato fruit development and in explaining fruit yield, indeed tomato is an apoplastic loading species like Arabidopsis (Adams et al. 2007).

Carbohydrate partitioning is a dynamic process influenced by both environmental and developmental cues (Hermans et al. 2006). Generally, a plant will reduce growth and competitiveness under stress conditions and growth can therefore be regulated by signals such as redox changes (Pedreira et al. 2004). In Arabidopsis, underexpression of ascorbate oxidase led to increased seed yield under stress conditions compared to wild type (Yamamoto et al. 2005) showing that ascorbate oxidase affects resource allocation in other species. The correlation found in our study between ascorbate oxidase activity and apoplastic sugar levels in fruit implied that the changes in apoplastic sucrose concentration (a relative decrease in apoplastic sucrose concentration correlated with decreased ascorbate oxidase activity in fruit) might also be one of the mechanisms improving yield under conditions where assimilate is limiting.

As expected, ascorbate oxidase activity had a direct effect on the apoplastic redox state, which was less oxidized when ascorbate oxidase activity decreased, and this enzyme therefore plays a role in responses linked to the concentration of the oxidized form of ascorbate in the apoplast. The variability in the leaf apoplastic redox state in the transgenic lines may be because of the presence of a second ascorbate oxidase gene in tomato or because the apoplastic redox state is not only controlled by ascorbate oxidase activity. Indeed, transport of ascorbate or dehydroascorbate across the plasma membrane (Horemans et al. 2000) will normally function to maintain a status quo in the apoplast. The transport of dehydroascorbate/ascorbate across the plasma membrane means that changes occurring in apoplastic ascorbate levels can have direct effects on cellular (symplastic) metabolism. Dehydroascorbate and monodehydroascorbate can enhance vacuolation and cell-wall loosening or even exert redox control over the cell cycle (Hidalgo, Gonzalez-Reyes & Navas 1989; Burhans & Heintz 2009; Foyer et al. 2009). Ascorbate oxidase activity is also correlated with cell expansion (Kato & Esaka 2000) but does not explain the increased cell size of early fruits, which must result from another phenomenon such as resource allocation.

The altered resource allocation seen in ascorbate oxidase lines is supported by changes in the fruit metabolome, gene expression and enzyme activity

As previously observed in transgenic tomato lines deficient in the expression of L-galactone-1,4-lactone dehydrogenase, alterations in the fruit metabolome did not necessarily match those in the leaf (Alhagdow et al. 2007). The metabolites that change in the ascorbate oxidase lines included organic acids, which are known to act as alternative respiratory substrates when carbon supply is limited [when sugar levels decline, these metabolites accumulate; (Ishizaki et al. 2005)], which tended to decrease in the transformants. The contrast between the leaf and fruit metabolome is also mirrored in the gene expression profiles of the two tissues suggesting that leaves and fruit respond differently to redox changes or sugar signalling and also highlighting the differences between source and sink tissues. In fruit, the increases in β-amylase and starch phosphorylase expression imply that starch breakdown may also be under redox control in these plants. The variations in expression of genes such as fw2.2 (Frary et al. 2000; Nesbitt & Tanksley 2001) and UDP-galactose-epimerase (Dormann & Benning 1998; Seifert et al. 2002) support hypotheses concerning altered resource allocation in ascorbate oxidase fruits. Notably, auxin-responsive gene expression was also increased in transgenic lines: independent links between auxin signalling, sugars and redox have already been shown in the literature (Esaka et al. 1992; den Boer & Murray 2000; De Tullio, Jiang & Feldman 2010). The increased expression of several sucrose phosphate synthase isoforms in transgenic lines was not reflected in the corresponding enzyme activities. Enzyme activity trends do not necessarily follow the trends shown by the corresponding genes because the activity is the result of a collection of isoforms, each responding differently, and also because of post-translational modifications. These differences between transcript and enzyme have already been seen for metabolic enzymes (Keurentjes et al. 2008; Piques et al. 2009). In our study, strong regulatory effects were seen on the activity of sucrose phosphate synthase, a key enzyme controlling sucrose synthesis (Lunn & MacRae 2003) and carbon partitioning (Worrell et al. 1991; Galtier et al. 1993). The differences are likely to be due to post-translational regulation, in common with the regulation of many enzymes involved in sugar metabolism.

Diverse effects of a diminution in ascorbate oxidase have been shown on gene transcription, the activity of enzymes involved in sugar metabolism and the fruit metabolome. As the changes in ascorbate redox occur in the apoplast, the majority of these effects must be indirect and probably result from changes within the cell, resulting from signals from the apoplast or plasma membrane, or are a consequence of changes in the concentrations of sugars themselves, as sugars are well-recognized signalling molecules (Ehness et al. 1997; Sinha et al. 2002). In conclusion, many interesting questions remain concerning the mechanisms controlling carbon allocation: in this study, we have linked the activity of an apoplastic enzyme, ascorbate oxidase, to sugar metabolism and carbon partitioning in tomato.

ACKNOWLEDGMENTS

We acknowledge the help of the Metabolomic Facility of the Bordeaux Functional Genomics Center, IBVM, INRA, Bordeaux. We thank Doriane Bancel and Sylvie Serino for help with sugar assays; David Page and Naima Dlalah for helping us to compare our ascorbate apoplast assays to an HPLC procedure; Jean-Jacques Longuenesse for help with the measurement of physiological parameters, Manu Botton for management of tomato plants; Alain Goujon for providing environmental greenhouse data; Gisèle Riqueau for technical assistance and the experimental team at INRA, St Maurice for plant management. We thank Mathilde Causse and Jean-Luc Gallois for helpful discussions. The work was financed by an INRA incentive project ‘VTCFruit’ obtained by C Rothan. The PhDs of N Gest and C Massot were financed by INRA and the region ‘Provence-Alpes-Côte d'Azur’.

Ancillary