Quantitative trait loci (QTL) mapping is a step towards the identification of factors regulating traits such as fruit ascorbic acid content. A previously identified QTL controlling variations in tomato fruit ascorbic acid has been fine mapped and reveals that the QTL has a polygenic and epistatic architecture. A monodehydroascorbate reductase (MDHAR) allele is a candidate for a proportion of the increase in fruit ascorbic acid content. The MDHAR enzyme is active in different stages of fruit ripening, shows increased activity in the introgression lines containing the wild-type (Solanum pennellii) allele, and responds to chilling injury in tomato along with the reduced/oxidized ascorbate ratio. Low temperature storage of different tomato introgression lines with all or part of the QTL for ascorbic acid and with or without the wild MDHAR allele shows that enzyme activity explains 84% of the variation in the reduced ascorbic acid levels of tomato fruit following storage at 4 °C, compared with 38% at harvest under non-stress conditions. A role is indicated for MDHAR in the maintenance of ascorbate levels in fruit under stress conditions. Furthermore, an increased fruit MDHAR activity and a lower oxidation level of the fruit ascorbate pool are correlated with decreased loss of firmness because of chilling injury.
Chilling injury is a type of oxidative stress that occurs during storage below 10 °C to which fleshy fruit, such as tomato, are particularly sensitive. Although low temperatures can be used to extend fruit shelf life, they may trigger physiological disorders and loss of quality, ultimately resulting in chilling injury (Lyons 1973). As a result of chilling injury, fruits develop symptoms such as a rubbery texture, watery flesh and irregular ripening. Improving the antioxidant content of sensitive fruits might improve fruit quality during post-harvest storage (Hodges et al. 2004), and correlations in apple indicate this is the case (Hodges et al. 2004; Davey et al. 2007). Scavenging reactive oxygen species (ROS) during chilling in tomato solicits a complex network of molecules and enzymes related to the antioxidative response systems, whose efficiency might be related to fruit shelf life (Malacrida et al. 2006). Chilling injury has been shown to lead to cellular stress and oxidation of cellular components such as the ascorbate pool (Hodges et al. 2004).
Fruit ascorbic acid concentration is highly influenced by the environment in both fruit and leaves (Dumas et al. 2003; Bartoli et al. 2006; Davey et al. 2007). The regulation of ascorbate levels in cells is also tightly controlled by the level of synthesis (Smirnoff, Conklin & Loewus 2001), recycling, degradation (Pallanca & Smirnoff 2000; Green & Fry 2005) and transport of this molecule within the cell or between organs (Horemans, Foyer & Asard 2000). The recycling pathway is especially important during stress responses and adaptation; because of its role as an antioxidant, the reduced ascorbate is oxidized into an unstable radical (monodehydroascorbate) during oxidative stress, which dissociates into ascorbate and dehydroascorbate. Dehydroascorbate is also unstable and rapidly degraded so the ascorbate pool can be depleted if the oxidized forms of monodehydroascorbate and dehydroascorbate are not recovered by two reductases: monodehydroascorbate reductase (MDHAR) and dehydroascorbate reductase (DHAR) (Smirnoff & Wheeler 2000). Recently, the role of recycling genes, such as MDHAR and DHAR, has been shown to be of interest: the overexpression of a DHAR enzyme in tobacco increased the ascorbic acid levels from two- to fourfold (Chen et al. 2003) and significantly increased the ascorbate redox state (Chen et al. 2003; Kwon et al. 2003). Furthermore, the overexpression of MDHAR in tobacco increases the reduced ascorbate levels and confers stress tolerance to the transformed plants (Eltayeb et al. 2007).
Quantitative trait loci analysis is valuable for identifying the regions of the genome that control a measurable characteristic, and identification of QTLs may help to highlight key regulators of fruit quality traits. In particular, in tomato, many QTL studies have been performed for fruit traits and reveal more than 30 QTL regions involved in the variation of fruit size (Grandillo, Ku & Tanksley 1999) or soluble solid content (Fulton et al. 2002). Ascorbic acid content lends itself to QTL analysis as it is a character exhibiting quantitative variation, which is controlled by several genes, more or less influenced by the environment (Stevens et al. 2007a). In order to determine the precise gene or regulatory region responsible for the quantitative character, approaches, such as fine mapping or a candidate gene approach, are essential to identify the genes segregating around a locus (Pflieger, Lefebvre & Causse 2001; Stevens et al. 2007b). Validation of the gene or regulatory region then involves correlating the expression or activity of the regulator with the character it controls and/or identifying the polymorphism responsible for the trait variation.
Previous work has identified an ascorbic acid QTL on tomato chromosome 9; the wild species alleles increase the fruit ascorbic acid content by up to 100% depending on the year when compared with the Solanum lycopersicum alleles from the processing cultivar M82 (Stevens et al. 2007b). A MDHAR gene colocates with this QTL, and in this paper, we confirmed the role of this gene in a QTL for fruit ascorbic acid content through fine mapping. MDHAR is an enzyme involved in the response to oxidative stress, and similar genes involved in regenerating reduced ascorbate from oxidized ascorbate are known for their roles in protection against oxidative stress (Smirnoff & Wheeler 2000). Therefore, MDHAR activity may only become a regulator of ascorbic acid levels once plants are under stress. We therefore stored tomatoes at low temperatures sufficient to induce chilling injury, and show that MDHAR activity is linked to ascorbic acid content and shelf life in tomato fruit following chilling.
MATERIALS AND METHODS
Culture of introgression lines for fine mapping and characterization during ripening
The population of introgression lines (IL) consisted of 75 lines each containing a single introgression fragment from the wild tomato accession Solanum pennellii LA716 in the genetic background of M82, a processing tomato variety with a determinate growth as previously described (Eshed & Zamir 1995). Eleven ‘sub’ IL (subIL) derived from the IL9.2.5 (Fridman et al. 2002) were cultured over 2 years (a greenhouse trial in autumn 2002 then a field trial in spring 2003). Thirty fully ripe fruits were harvested from 12 plants per line, and 5 bulks of 6 fruits were ground to a fine powder in liquid nitrogen and were analysed for ascorbic acid content as described further. For the characterization of ascorbic acid content and MDHAR activity during ripening, M82 and IL9.2.5 were cultured in the greenhouse in spring and autumn 2004, and a minimum of six fruits were harvested at mature green, breaker or red ripe stages, pooled, ground to a fine powder in liquid nitrogen, and analysed as described further.
Culture and conditions for low temperature storage
Severe chilling stress: culture of M82 plants and 40-d chilling stress
M82 plants were grown in the greenhouse in spring 2005. Fruits were harvested when they turned orange. A minimum of 12 fruits per ripening stage and per temperature were harvested with the experiment being repeated twice. Firmness measurements were carried out on all fruits. The harvested fruits were then divided into three groups: one for grinding in liquid nitrogen for biochemical measurements at harvest, and the others for storage at 4 or 20 °C. The fruits were stored for 40 d at 4 °C or 8 d at 20 °C. The 20 °C fruits were used as a ripening control.
Moderate chilling stress: culture of M82 and IL, and 28-d chilling stress
M82, IL9.2.5, subIL18.104.22.168 and subIL22.214.171.124 plants were grown in the greenhouse in spring 2007. The fruits were harvested on three dates: 21 May 2007 (harvest A), 29 May 2007 (harvest B) and 4 June 2007 (harvest C); and each harvest was analysed separately. At each harvest for each IL, a minimum of 20 fruits were harvested between breaker and orange stage from 10 plants. Firmness measurements were carried out on all fruits. The harvested fruits were then divided into two groups: one for biochemical measurements at harvest, and the other for storage at 4 °C. The fruits were stored for 28 d at 4 °C. Tomatoes were subjected to a less severe stress in 2007 than in 2005 (28 d at 4 °C instead of 40 d) because of the occurrence of pathogen attack on the tomatoes left longer than this storage period.
Transformation of the MDHAR allele into a polymerase chain reaction marker
The S. pennellii MDHAR allele corresponds to the previously mapped MDHAR3 gene (SGN-U145504) (Stevens et al. 2007a). The following primers were designed based on the available sequence and were used to amplify a fragment of 443 bp of the gene from genomic DNA of M82 and S. pennellii: 415F 5′-GGGCAAGTTGAAGAGGAGAA GGG and 415R 5′-GCAGGGGGTTGAACCTTTGC. A HinfI site was identified in the S. pennellii sequence that was absent from the M82 sequence, and this generated a direct cleavage amplified polymorphism marker. Digestion of the polymerase chain reaction (PCR) product produced two fragments if the HinfI site in the S. pennellii sequence was present and allowed the fine mapping of the allele using genomic DNA purified from the subIL.
Sequencing of MDHAR cDNAs
Three overlapping partial cDNA fragments, which covered the entire MDHAR cDNA sequence from the start ‘ATG’ to the stop ‘TAA’ codons (1302 bp) based on the sequence already available in Genbank for tomato (L41345) (Grantz, Brummell & Bennett 1995), were amplified by reverse transcriptase-PCR from M82 and S. pennellii using RNA extracted from the fruit of either accession by standard techniques. The pairs of primers used are listed with their start position relative to the ATG codon and the fragment size:
1252F: 5′-AAAAAGCTCTAACTATTTTCACTG, and 1252R: 5′-TCCTCCACTTCCAACACACA (−37 bp from ATG, 259 bp product); 416F: 5′-GGAGCTGCTAGACT CCCAGG, and 415R (see previous description) (+174 bp from ATG, 1076 bp); 415F (see previous description), and 996R: 5′-GGCAGAAACTTCAAAGAGCTG (+807 bp from ATG, 543 bp). The partial cDNAs were sequenced, and alignment of the M82 and S. pennellii sequences was carried out.
Transversal sections from the middle of the fruit, including pericarp, columella and the jellylike locular tissues, were frozen and ground to a fine powder in liquid nitrogen, and were stored at −80 °C for the following assays:
Ascorbic acid content
Measurements of ascorbic acid content were carried out as described (Stevens et al. 2006) on the material conserved at −80 °C. Briefly, tomato tissue was ground in liquid nitrogen, and 1 g of powder was homogenized with 600 µL of ice-cold 6% trichloroacetic acid (TCA). Samples were centrifuged for 15 min at 25 000 g at 4 °C. Twenty microlitres of the supernatant was used in each assay. The ascorbate standards were prepared fresh: a solution of 1 mg/mL sodium ascorbate was diluted in 6% TCA to give a concentration in 20 µL of 0, 5, 10, 15, 20 and 30 nm, allowing a standard curve of absorption values of between 0 and 1 to be generated after the addition of the appropriate reagents (see further). Two assays were carried out on each sample, one to measure the total ascorbate [including the addition of 5 mm dithiothreitol (DTT)] and one to quantify the reduced ascorbate content (omission of DTT from the assay). Twenty microlitres of each sample or standard was distributed into at least two wells (for two repetitions) of a 96-well microplate and was mixed with 20 µL of 5 mm DTT (total ascorbate assay) or 0.4 m phosphate buffer, pH 7.4 (reduced ascorbate assay). The plate was incubated at 37 °C for 20 min. Ten microlitres of N-ethyl malemide (total ascorbate assay) or 0.4 m phosphate buffer, pH 7.4 (reduced ascorbate assay) was added and mixed, followed by the addition of 80 µL colour reagent (see further). After incubation at 37 °C for 40 min, the absorbance was read at 550 nm using a microplate reader (Thermo Electron, Courtaboeuf, France). The colour reagent was made up as follows: solution A: 31% orthophosphoric acid, 4.6% w/v TCA and 0.6% w/v iron chloride; solution B: 4% 2,2-dipyridyl (w/v made up in 70% ethanol). Solutions A and B were mixed 2.75 parts (A) to 1 part (B). The standard curve obtained from the standard solution values allowed calculation of the ascorbate concentration of the samples after correction for the quantity of water introduced by the tomato fruit sample.
Assay of MDHAR activity was carried out according to a previously described method (Arrigoni, Dipierro & Borraccino 1981) based on the oxidation of NADH at 30 °C measured at 340 nm. Approximately 150 mg of frozen powder was homogenized with 600 µL of extraction buffer (50 mm Tris-HCl pH 7.8). After centrifuging at 25 000 g, 4 °C for 15 min, 50 µL of the clear supernatant was assayed in a final volume of 1 mL containing 1 mm ascorbate, 0.2 mm NADH and enough ascorbate oxidase enzyme to give a linear rate of production of the monodehydroascorbate radical as measured at 260 nm (approximately 0.1 units). The initial rate without the ascorbate oxidase enzyme was subtracted from the final rate once this enzyme was added to the reaction. Results were expressed as per gram fresh weight and not per milligram protein because of changes in the protein concentration of a fruit during ripening (Jimenez et al. 2002). Assays were carried out in triplicate. Measures of MDHAR activity represent the sum of the activities of different MDHAR isoforms present in the fruit.
Measurements of tomato firmness
Firmness is an overall estimation of fruit resistance to compression, and is a combination of skin resistance and flesh firmness (Grotte et al. 2001). Fruit firmness was assessed with a texturometer (Texture analyser TAplus; Ametek, Lloyd Instruments Ltd., Fareham, UK). This apparatus registered force/deformation curves by measuring the reaction force in response to an increasing mechanical constraint applied to the fruit by a 5 cm flat disc supported by a motorized arm. Fruit firmness was equivalent to the force necessary to obtain a deformation corresponding to 3% of the fruit diameter. The probe speed was 20 mm min-1. The measurement was first carried out on each fruit at harvest and then after the low temperature storage. The total firmness lost corresponded to the difference between the two measurements and was expressed as a percentage of the fruit pressure measured at harvest. The result therefore takes account of the differences in fruit firmness of the lines at harvest. To avoid artefacts because of the cold temperature of fruits following chilling, the fruits were allowed to warm up at 20 °C in the light for 24 h before measurement of their firmness.
Statistical analysis was carried out using the R package (http://www.R-project.org) (R-Development-Core-Team 2007). Pairwise comparisons with M82 for physical measurements on firmness were carried out using the non-parametric method of Wilcoxon, and an adjustment of type Benjamini and Yekutieli. Pearson correlation coefficients between means of the physical and biochemical variables were estimated. For the physical measurements, data were shown from each harvest as the harvest date introduced a large source of variability.
Fine mapping of the ascorbic acid QTL on chromosome 9
Previous work identified a colocation between a QTL for ascorbic acid content in fruits on chromosome 9 (bin 9D) and an MDHAR gene in two different tomato populations (Stevens et al. 2007a). In order to confirm the role of MDHAR and ascorbate recycling in the ascorbate content of fruit, a high resolution mapping of this gene was carried out using 11 independently derived recombinants in an M82 genetic background (Fridman et al. 2002). These subIL were derived from the IL9.2.5, which harbours the ascorbic acid QTL (between 10 and 100% increase in fruit ascorbic acid content at the red ripe stage). A PCR marker was generated for the S. pennellii MDHAR allele and was used to map this allele to subIL 126.96.36.199 to 188.8.131.52 (Fig. 1) in an interval of less than 0.1 cM in the same region as markers TG225 and CT283. In parallel, the ascorbic acid quantification was carried out on fruits of these subIL in 2002 and 2003, and revealed a significantly increased ascorbic acid content in subIL groups B, D and E, but not in group C (lines 184.108.40.206 and 220.127.116.11; Fig. 1). This result indicates that at least two QTL exist, which contribute to the increased ascorbic acid content of IL9.2.5 (AA9.1+ and AA9.3+, Fig. 1) separated by a third QTL in a region of 0.2 cM between markers CT283 and GP263 (AA9.2−) exerting a negative influence and explaining the low ascorbic acid content of group C subIL.
A difference in enzyme activity of the S. pennellii and M82 MDHAR alleles
The protein coding regions of the M82 and S. pennellii cDNAs have been sequenced, and three base pairs differed. Two of these changes had no effect on the protein sequence, while the third led to a non-conservative amino acid change (proline in M82 replaced with alanine in S. pennellii) at amino acid 416 (position relative to start methionine). To investigate more fully the involvement of MDHAR in the ascorbic acid content of IL9.2.5 and M82, MDHAR activity measurements were carried out on three late ripening stages (mature green, breaker and red) of M82 and IL9.2.5 tomatoes grown in spring and autumn of 2004, and were compared with the fruit ascorbic acid concentrations (Fig. 2). IL9.2.5 has increased fruit ascorbic acid content compared with M82 in both seasons, the greatest difference with M82 being seen at the red ripe stage. MDHAR activity decreases during ripening and is higher in IL9.2.5 than in M82 in autumn but not in springtime. Therefore, the difference identified in MDHAR enzyme activity may be under environmental control.
Monodehydroascorbate reductase activity and reduced ascorbic acid levels under severe chilling stress
As a role for MDHAR in response to the environment is indicated, and similar genes involved in regenerating reduced ascorbate from oxidized ascorbate are known for their roles in protection against oxidative stress (Smirnoff & Wheeler 2000), MDHAR activity may only become a regulator of ascorbic acid levels once plants are under stress. To test this hypothesis, chilling stress was chosen as an easily applied stress to which tomato fruits are particularly sensitive. To determine the role of the ascorbate pool, its redox state and the MDHAR gene in protection against chilling injury, M82 tomatoes were harvested between breaker and orange stage, and were subjected to a severe chilling stress by storage at 4 °C for 40 d. Fruit firmness after the chilling stress was compared with the fruit stored for 8 d at 20 °C (a normal ripening control). Figure 3 shows that firmness decreased after the long-term cold storage, beyond the firmness lost during a normal ripening at 20 °C (Fig. 3a), and that the oxidized ascorbate increased at 4 °C (Fig. 3b) in parallel with a decrease in MDHAR activity (Fig. 3c). The total ascorbate content of the fruits remained unchanged (Fig. 3d), and so the increase in oxidized ascorbate was mirrored by a decrease in the reduced ascorbate content of fruits. Photos of tomatoes after storage at 20 or 4 °C show that at 4 °C, symptoms of chilling injury appear. A tomato following a normal ripening has a uniform colour and texture (Fig. 3e), whereas after a cold storage, non-uniform ripening (Fig. 3f) and loss of firmness (Fig. 3g) occur.
Induction of chilling stress in the tomato IL
In order to investigate the role of the ascorbic acid QTL in the maintenance of post-harvest quality, a moderate chilling stress was applied to M82 and three IL: IL9.2.5 containing the entire ascorbic acid QTL and lin5, the sugar QTL, and two subIL chosen for their elevated fruit ascorbic acid content (subIL18.104.22.168 and subIL22.214.171.124). The two subIL were contrasted for the MDHAR allele present (wild-type allele in 126.96.36.199, M82 allele in 188.8.131.52) and the lin5 allele responsible for the sugar QTL and colocating with part of the ascorbic acid QTL (wild-type allele in 184.108.40.206, M82 allele in 220.127.116.11, Fig. 1) (Fridman et al. 2002). SubIL 18.104.22.168 and 22.214.171.124 were not chosen for these experiments because initially, the emphasis was on using lines with the highest ascorbate content and also on the putative role of the sugar QTL (Brix9-2-5, Fig. 1) in the increased ascorbate content of the QTL found in IL9.2.5. No significant differences in fruit weight were found between the lines at harvest. All IL contained significantly more fruit ascorbic acid than M82 (Fig. 4a, open shapes) as they contain all or part of the QTL. IL9.2.5 and subIL126.96.36.199 had greater MDHAR activity than M82 as expected, as IL9.2.5 and subIL188.8.131.52 carried the more active S. pennellii MDHAR allele. SubIL184.108.40.206 had an intermediate activity (Fig. 4b). Firmness measurements, ascorbic acid content and MDHAR activity assays were carried out at each harvest and after 28 d at 4 °C (Table 1 and Fig. 4). The MDHAR activity varied considerably with the harvest date as shown in Fig. 4c and as already noted in Fig. 2, further evidence that the activity of this enzyme is under a strong environmental control. Tomatoes from the IL had more MDHAR activity at later harvests, which was not the case for M82.
Table 1. Fruit firmness at harvest and after the low temperature storage
Firmness at harvest (kPa)
Total firmness lost (%)
Significantly different at P < 0.05.
The average values and standard deviations (in brackets; SD) for M82 and each introgression line at the three harvests are shown. Pairwise comparisons between M82 and the different introgression lines for each harvest were carried out.
Chilling injury resulted in a loss of firmness beyond the softening that occurs during normal ripening as shown for the severe stress experiment previously carried out (Fig. 3). Under the moderate stress regime, when the tomato lines were compared at harvest, no significant differences were found in their firmness (Table 1). After conservation at 4 °C, if the lines IL9.2.5, subIL220.127.116.11 and subIL18.104.22.168 are compared with M82, significant differences (P < 0.05) in firmness appeared (Table 1). The firmness lost in the three IL was lower than the firmness lost by M82, showing that the IL may tolerate cold storage better than M82.
Effect of moderate chilling on ascorbic acid content and MDHAR activity
The evolution of fruit ascorbic acid content following storage at 4 °C for M82 and the IL is shown in Fig. 4a (filled shapes show 4 °C values). During the chilling period, the total ascorbic acid levels increased in all the lines probably because of the normal ripening process. Although the total ascorbate increased in all lines, the percentage of the reduced form decreased for M82, and by deduction, the percentage of the oxidized form of ascorbate increased, similar to the results in 2005 following a severe chilling (Fig. 3). The percentage of the reduced form after chilling compared with at harvest is maintained for the IL, particularly for IL9.2.5 and subIL22.214.171.124. The MDHAR activity of the lines after chilling averaged over the different harvests is shown in Fig. 4b. The total MDHAR activity did not increase significantly after chilling. M82, however, had less MDHAR activity than the other lines, with subIL126.96.36.199 having an intermediate activity, and IL9.2.5 and subIL188.8.131.52 having the highest activity.
Correlations between fruit ascorbic acid content, MDHAR activity and loss of firmness: MDHAR activity is positively correlated with ascorbic acid levels after chilling stress
The two hypotheses tested are that firmness after a low temperature storage at 4 °C is affected (1) by ascorbic acid or MDHAR activity at harvest or (2) by the response of ascorbic acid or MDHAR activity levels at chilling temperatures. The correlations between firmness, ascorbic acid levels and MDHAR activity generated by the different lines, harvests and by moderate chilling are shown in Table 2. Neither ascorbic acid content (total or reduced) nor MDHAR activity was correlated with firmness at harvest, whereas after chilling, all the correlations became significant. The strongest correlations were found with the reduced ascorbic acid content, which was inversely correlated with loss of firmness (correlation coefficient of −0.73, P < 0.01, Table 2) and therefore potentially resistance to chilling injury. The correlation was also significant between firmness and ascorbic acid content, and firmness and MDHAR activity although at a lower level (P < 0.05).
Table 2. The correlations between fruit firmness and ascorbic acid content or MDHAR activity, at harvest and after low temperature storage
Firmness at harvest
Total firmness lost
P < 0.05;
P < 0.01.
Each data point represents the correlation between the 12 mean values obtained from the four lines at the three harvests.
MDHAR, monodehydroascorbate reductase; NS, not significant.
In order to evaluate the role of the candidate gene MDHAR in tomato fruit ascorbic acid content, at harvest (non-stressed conditions) or after a moderate chilling stress (28 d at 4 °C), correlations between ascorbic acid levels and MDHAR activity were analysed. At harvest, the MDHAR activity explained 38% of the variation in the reduced ascorbic acid levels. However, after chilling, the MDHAR activity explained 84% of the variation in the reduced ascorbic acid levels (Fig. 5). Total ascorbic levels were also similarly associated with MDHAR activity levels as only 36% of their variation was explained by the MDHAR activity at harvest, while the percentage reached 77% after chilling (data not shown).
The 9.2.5 locus is composed of at least three QTLs affecting fruit ascorbic acid content
The previously identified QTL for tomato fruit ascorbic acid content (Stevens et al. 2007a) has been fine mapped within a region of approximately 0.3 cM and has been shown to be made up of several QTL, with non-additive and epistatic effects, contributing to the overall QTL found in IL9.2.5. Our observation on the complexity of the QTL in the 9.2.5 region is similar to the work by Kroymann & Mitchell-Olds (2005), who found that a 1 cM chromosome interval in Arabidopsis had a highly polygenic and epistatic architecture for a quantitative variation of growth rate. In this paper, we show that an ascorbic acid QTL is made up of several QTLs with smaller effects, and a minimum of three introgressed regions from S. pennellii may be involved in fruit ascorbic acid content (AA9.1+, 9.2− and 9.3+, Fig. 1). The first region labelled ‘AA9.1+’ covered by the IL of groups B, C and D had an overall positive effect on tomato fruit ascorbic acid content as evidenced by the ascorbic acid content of the lines subIL184.108.40.206 and subIL220.127.116.11, which carry only this region. The MDHAR candidate gene colocates with the QTL in this region, and therefore has been fine mapped to a region of less than 0.1 cM. The second region (AA9.2−, Fig. 1) corresponds to the IL of groups C and D: the S. pennellii alleles of group C may exert a negative influence on the ascorbic acid content at the AA9.1+ locus explaining the difference in the ascorbic acid content of the fruits from subIL of group C. A QTL for plant weight is also found in this region, colocating with the self-pruning orthologue (SP-9D) (Fridman et al. 2002). The S. pennellii allele of this gene is associated with a semi-determinate, rather than determinate, growth leading to increased leaf numbers and plant weight. The SP gene is a candidate gene for a QTL for plant weight and may modulate the strength of the photosynthetic vegetative source. Therefore, the wild allele may also contribute positively to the fruit ascorbic acid content explaining the increased ascorbic acid content of the fruits from the IL of group D. A final region (AA9.3+) between the plant weight and sugar QTLs, shown in Fig. 1, present in the IL of groups D and E may also increase the fruit ascorbic acid content. Lin5, the gene controlling the total soluble solids (mainly sugar content) of the 9.2.5 region (Fridman et al. 2002), could also conceivably have a positive effect on the ascorbic acid content QTL of this region. An increase in fruit sugar could increase fruit ascorbic acid as the ascorbic acid synthesis pathway starts from glucose.
Monodehydroascorbate reductase activity in IL9.2.5 is increased compared with M82
Ascorbic acid has a major role as an antioxidant, donating electrons to a wide range of substrates. In particular, ascorbate is involved in the enzymatic and non-enzymatic scavenging of damaging ROS reviewed in Davey et al. (2000). The production of ROS is increased under conditions of environmental stress, and the ascorbate pool needs to be maintained in a reduced state to protect against cellular damage arising from the overproduction of ROS. Reductases, such as MDHAR, have therefore important roles in stress tolerance. The MDHAR activity of the fruits from IL9.2.5 increased compared with the parent line M82 under certain seasons and at certain harvest dates. Therefore, the polymorphism responsible for the increase in MDHAR activity of the S. pennellii allele is likely to be found in a regulatory region of the gene, such as the promoter, or may result from an increased stability of the protein resulting from the single amino acid change in the S. pennellii enzyme. A hypothesis is that common mechanisms are involved in the response to oxidative stresses, such as the chilling stress imposed in this study, that require an increased activity of enzymes involved in oxidative stress tolerance such as MDHAR. The S. pennellii MDHAR allele present in IL9.2.5 may show a different response under conditions leading to oxidative stress, for example, by transcript induction or increased protein activity.
We have shown in 2005 that under a severe chilling stress (40 d at 4 °C), tomato fruits lose firmness above the levels found during a normal ripening at 20 °C, and the reduced ascorbate levels fall in parallel with MDHAR activity (Fig. 3). A role for MDHAR in the maintenance of ascorbate levels in fruit under stress conditions is therefore indicated. Results from 2007, when the chilling stress was moderate, also indicate that MDHAR may be involved in fruit reduced ascorbic acid content during chilling as shown in Fig. 4a. The IL containing the more active S. pennellii allele maintain a higher percentage of reduced ascorbic acid following chilling stress. This MDHAR cDNA has already been isolated, the gene has shown to be expressed in fruit, and mRNA levels to be positively correlated with MDHAR enzyme activity, but negatively correlated with ascorbate levels in different tissues (Grantz et al. 1995). Like us, the authors suggest that this enzyme may contribute to maintaining the levels of ascorbic acid for protection against oxidative stress arising from wounding. It is likely that turnover genes have a role in the regulation of ascorbate levels, for example, under stress conditions, particularly as the elevated expression of the other enzyme involved in ascorbate turnover, DHAR, has been shown in tobacco to enhance the ascorbate content of plants (Chen et al. 2003). The overexpression of MDHAR has also been shown to increase the reduced ascorbate content of tobacco leaves (Eltayeb et al. 2007). In this paper, we showed positive correlations between fruit MDHAR activity and fruit ascorbic acid content under stress conditions (after chilling compared with at harvest), but the ultimate validation of this gene and its role in the trait would involve the production of transgenic plants under- or overexpressing the gene in fruit.
It is also of note that at least two other MDHAR genes exist in tomato (Stevens et al. 2007a), but only the gene studied in this paper colocates with an ascorbic acid QTL. The gene is likely to encode a cytosolic MDHAR isoform (as suggested by Grantz et al. 1995) or indeed a peroxisomal MDHAR isoform because of the presence of an ‘serine-lysine-isoleucine’ putative peroxisomal targeting sequence at the C-terminus of the protein. The protein sequence from tomato also shares the most homology with a peroxisomal isoform of MDHAR from Arabidopsis thaliana where six MDHAR polypeptides have been identified (Lisenbee, Lingard & Trelease 2005). Of course, measures of MDHAR activity represent the sum of the activities of different MDHAR isoforms present in the fruit.
Tomato post-harvest quality and fruit softening during chilling
Low temperatures are widely used to extend fruit shelf life before consumption. Low temperatures slow down fruit metabolism and the development of microflora responsible for fruit spoilage. However, below a certain threshold depending on the fruit, low temperatures may trigger physiological disorders and loss of quality during storage, the phenomenon known as chilling injury (Lyons 1973). Improving the chilling tolerance of sensitive fruits may avoid loss of fruit quality because of inappropriate storage. The precise mechanisms that trigger chilling injury remain unclear. Membrane composition was an initial emphasis, especially the saturation state of constitutive lipids, which was thought to regulate membrane fluidity and permeability at low temperature (Lyons 1973). Substantial alterations in cell wall metabolism have also been measured in response to cold storage, and numerous cell wall-modifying enzymes are affected, with important consequences for pectin metabolism (Brummell et al. 2004). Ascorbate has been suggested to have a role in cell wall loosening during cell expansion or fruit ripening (Fry 1998), and it is generally considered that ripening and post-harvest storage are oxidative processes (Jimenez et al. 2002; Hodges et al. 2004). Therefore, connections between antioxidant potential and post-harvest quality have already been made in several species, for example, apple (Davey et al. 2007). Our results indicate first that tomato fruits lose firmness following severe chilling injury above the amount because of normal ripening (Fig. 3). Second, when the loss of firmness following moderate chilling from the fruits of four lines, which vary for ascorbic acid content and MDHAR activity, is observed over different harvests, a negative correlation exists between the reduced ascorbic acid levels, and to a lesser extent total ascorbic acid levels, and loss of firmness. The results in this paper indicate that the response or maintenance of the antioxidative system during chilling is more important than the antioxidant quality (reduced ascorbate levels and MDHAR activity in this case) at harvest. In other studies on other fruits, it has been shown that, generally, cultivars that could maintain their reduced ascorbate and glutathione pools also had better storage properties (Davey & Keulemans 2004).
Acclimation and defence against chilling injury in tomato
In 2005, we were able to apply a severe chilling stress (tomatoes were left for 40 d at 4 °C), whereas in 2007, tomatoes were only left for 28 d, as longer periods at 4 °C led to a pathogen attack. As a consequence, the chilling injury received in 2005 was more severe, and MDHAR activity levels had already started to fall and oxidized ascorbate levels to increase. It is therefore interesting in 2007 that even though the stress was less severe, we were still able to correlate the reduced ascorbic acid levels and MDHAR activity with the maintenance of firmness. It would also be interesting to look at other components of the antioxidative system under conditions of severe and moderate chilling stress. Data from the glutathione system (Tausz, Sircelj & Grill 2004) suggested that a period of stress acclimation precedes degradation of the antioxidative defences.
We have shown that the ascorbic acid QTL on chromosome 9 is composed of several QTL with sometimes opposing effects, but that the MDHAR gene remains a candidate for part of this QTL. Ultimate validation by production of transgenic plants may confirm or disprove this hypothesis. We have also shown a good correlation between MDHAR activity and reduced fruit ascorbic acid levels under stress conditions: MDHAR activity explains 84% of the variation in the reduced ascorbic acid levels in fruits compared with 38% at harvest under non-stress conditions, showing that part of the ascorbic acid QTL on chromosome 9 may be under environmental control. The maintenance of MDHAR activity and the reduced ascorbate pool in tomato fruits is correlated with the reduced loss of firmness during chilling, and therefore has implications for the post-harvest quality of tomatoes.
We would like to thank Patrice Reling and Maggy Grotte for their help with the physical measurements, Michel Buret for his help with the ascorbic acid assays, Philippe Duffé for technical assistance, and Yolande Carretero and the experimental team at INRA, St Maurice, for the plant management.