Alteration of the interconversion of pyruvate and malate in the plastid or cytosol of ripening tomato fruit invokes diverse consequences on sugar but similar effects on cellular organic acid, metabolism, and transitory starch accumulation.

Summary: Normal tomato ripening is influenced by alterations on both cytosolic phosphoenolpyruvate carboxykinase and plastidic NADP-malic enzyme. This study provides compelling evidence of their roles in starch biosynthesis, respiration rates, and tricarboxylic acid cycle flux. The aim of this work was to investigate the effect of decreased cytosolic phosphoenolpyruvate carboxykinase (PEPCK) and plastidic NADP-dependent malic enzyme (ME) on tomato (Solanum lycopersicum) ripening. Transgenic tomato plants with strongly reduced levels of PEPCK and plastidic NADP-ME were generated by RNA interference gene silencing under the control of a ripening-specific E8 promoter. While these genetic modifications had relatively little effect on the total fruit yield and size, they had strong effects on fruit metabolism. Both transformants were characterized by lower levels of starch at breaker stage. Analysis of the activation state of ADP-glucose pyrophosphorylase correlated with the decrease of starch in both transformants, which suggests that it is due to an altered cellular redox status. Moreover, metabolic profiling and feeding experiments involving positionally labeled glucoses of fruits lacking in plastidic NADP-ME and cytosolic PEPCK activities revealed differential changes in overall respiration rates and tricarboxylic acid (TCA) cycle flux. Inactivation of cytosolic PEPCK affected the respiration rate, which suggests that an excess of oxaloacetate is converted to aspartate and reintroduced in the TCA cycle via 2-oxoglutarate/glutamate. On the other hand, the plastidic NADP-ME antisense lines were characterized by no changes in respiration rates and TCA cycle flux, which together with increases of pyruvate kinase and phosphoenolpyruvate carboxylase activities indicate that pyruvate is supplied through these enzymes to the TCA cycle. These results are discussed in the context of current models of the importance of malate during tomato fruit ripening.

Fruit ripening and development is well established to be under tight genetic control (Matas et al., 2009;Karlova et al., 2011;Klee and Giovannoni, 2011). A growing body of evidence suggests that, in parallel, large metabolic shifts additionally occur during this process (Carrari and Fernie, 2006;Osorio et al., 2011). Insights into the genetic mechanisms that mediate fruit ripening-related processes, such as cell wall metabolism, pigment synthesis, and sugar metabolism, have resulted from studies that collectively span a wide range of plant species and indicate that they are broadly conserved (Seymour et al., 2008;Moing et al., 2011;Zhang et al., 2011). However, certain speciesspecific differences exist in the dynamics of other metabolite pools across ripening, with, for example, grape (Vitis vinifera), strawberry (Fragaria 3 ananassa), prune (Prunus domestica), and pepper (Capsicum annuum) displaying slightly different metabolic programs from tomato (Solanum lycopersicum; Zamboni et al., 2010;Lombardo et al., 2011;Osorio et al., 2011Osorio et al., , 2012Zhang et al., 2011). This fact notwithstanding, tomato has become the primary experimental model in which to study the development and ripening of fleshy fruits (Giovannoni, 2004;Fernandez et al., 2009). Furthermore, it is arguably also the best characterized fruit at the biochemical level (Rose and Bennett, 1999;Fraser and Bickmore, 2007;Moco et al., 2007).
One interesting observation is that the levels of several organic acids correlate strongly with transcriptional programs of tomato fruit development . This finding led to follow-up work demonstrating that the level of malate played an important role in tomato fruit ripening influencing starch accumulation, total soluble solid levels at ripening, and postharvest properties (Centeno et al., 2011), when taken alongside the results of many recent studies on the control of stomatal aperture by malate (Nunes-Nesi et al., 2007;Lee et al., 2008;Araújo et al., 2011;Penfield et al., 2012), thus expanding the documented biological roles for this acid beyond those previously documented (Fernie and Martinoia, 2009;Meyer et al., 2010). This finding thus suggests that malate metabolism may exert a key influence on the normal ripening and metabolism of tomato fruit. Many of our previous studies were targeted at mitochondrial reactions of the tricarboxylic acid (TCA) cycle (for review, see Sweetlove et al., 2010;Nunes-Nesi et al., 2011); here, however, we took RNA interference (RNAi) strategies to silence either the cytosolic phosphoenolpyruvate carboxykinase (PEPCK) or the plastidic NADP-dependent malic enzyme (ME) under the control of the ripening-specific E8 promoter. In plants, PEPCK is a cytosolic enzyme that catalyzes the ATP-dependent decarboxylation of oxaloacetate (OAA) to phosphoenolpyruvate (PEP; Chollet et al., 1996). In tomato, PEPCK abundance increases at the start of ripening, during which there is a decrease in malate content (Bahrami et al., 2001). This has led to the suggestion that PEPCK is involved in this dissimilation and that part of this dissimilated malate may be used in gluconeogenesis (Leegood and Walker, 2003;Famiani et al., 2005Famiani et al., , 2009. In contrast, NADP-ME isoforms function in chloroplasts and the cytosol (Drincovich et al., 2001;Gerrard Wheeler et al., 2005) and catalyze the reversible conversion of malate and pyruvate (Farineau, 1977;Drincovich et al., 2001). In tomato, NADP-ME has been implicated in respiration during ripening, providing pyruvate and/or NADPH as a substrate for respiration (Farineau, 1977;Edwards and Andreo, 1992).
In this study, the two generated tomato lines (cv MicroTom) with reduced cytosolic PEPCK and plastidic NADP-ME gene expression were characterized at morphological and biochemical levels, including comprehensive metabolite profiling and an evaluation of the effect of these genetic interventions on cellular fluxes. The results are discussed with respect to current models of energy metabolism within the fruit and with respect to the roles of carboxylic acids during fruit ripening and development.

Generation of Tomato Plants with Fruit-Specific Inhibition of PEPCK and NADP-ME
An 560-bp fragment of the complementary DNA (cDNA) encoding cytosolic PEPCK (accession no. AY007226) and a 601-bp fragment of the cDNA encoding a tomato plastidic NADP-ME (accession no. AF001269) were independently cloned using RNAi design into the pBINplus under the control of the ripening promoter E8. Plants (cv Micro-Tom) obtained by an Agrobacterium tumefaciens-mediated transformation were grown in the greenhouse, and seeds were collected. Eight lines for RNAi-PEPCK and nine lines for RNAi-ME were selected on the basis of PEPCK and NADP-ME1 gene expression by quantitative real-time (qRT)-PCR (Fig. 1, A and B). Three lines per construct were chosen: PEPCK7, PEPCK10, and PEPCK11 (PEPCK lines) and ME2, ME7, and ME13 (ME lines). Eight T2 plants per line were subsequently grown in the greenhouse, and fruits at breaker, breaker + 5 d after pollination (DAP), and breaker + 10 DAP were harvested. Assays of PEPCK and NADP-ME activities revealed that the lines displayed considerable reduction in activity during ripening (Tables I and II), and qRT-PCR analysis confirmed that they additionally displayed reduced expression of the PEPCK and NADP-ME1 genes, respectively (Fig. 2). We additionally observed an increase in PEPCK level during normal tomato ripening, while the NADP-ME1 level declined gradually at the onset of ripening as described previously (Bahrami et al., 2001;Tomato Genome Consortium, 2012;Fig. 2).

Specific Expression of SlNADP-ME Genes
The Arabidopsis (Arabidopsis thaliana) genome is known to contain four genes encoding functional NADP-ME isoforms (Gerrard Wheeler et al., 2005). NADP-ME4 is localized to plastids, while NADP-ME1, NADP-ME2, and NADP-ME3 are cytosolic isoforms (Gerrard Wheeler et al., 2005). The genome of tomato possesses three more genes as part of NADP-ME1 encoding putative NADP-MEs: cytosolic NADP-ME2 (AF001270; Knee and Finger, 1992), plastidic NADP-ME3 (Solyc03g120990; Tomato Genome Consortium, 2012), and a putative NAD(P)-ME4 (Tomato Genome Consortium, 2012). We compared the abundance of all three NADP-MEs and NADP-ME1 transcripts using published tomato (cv Heinz) RNAseq data from three ripening stages (green, breaker, and breaker + 10 DAP; Tomato Genome Consortium, 2012). The expression levels of NADP-ME2 and NADP-ME1 were similar, while the expression of NADP-ME3 and NADP-ME4 accounted for less than 100 times that of NADP-ME1 and NADP-ME2 (Tomato Genome Consortium, 2012). Interestingly, NADP-ME2 and NADP-ME3 expression was decreased during ripening, while NADP-ME4 expression was unaltered during fruit ripening (Supplemental Fig. S1; Tomato Genome Consortium, 2012). Importantly, qRT-PCR evaluation of the expression levels of cytosolic NADP-ME2, plastidic NADP-ME3, and NADP-ME4 during ripening in the transgenic ME lines revealed that the plastidic NADP-ME3 was decreased in all transgenic lines and in all three ripening stages in comparison with the wild type (Supplemental Fig. S1). However, the levels of the transcripts of NADP-ME2 and NADP-ME4 were unaltered in the transformant lines (Supplemental Fig. S1).

Fruit Size and Yield
Fruit size and total fruit yield were determined in fully mature fruit (breaker + 10 DAP). The ME transformants exhibited a significantly lower fruit yield consistent across all three transgenic lines (Fig. 1C). Given that fruit size was essentially unaltered (Fig. 1D), it is apparent that this was a result of a minor decrease in total fruit number in the ME transformants (data not shown). By contrast, PEPCK transformants exhibited unaltered fruit size and yield (Fig. 1,C and D).

Metabolic Profiling of Breaker Fruits of PEPCK and NADP-ME Lines
In order to gain a deeper comprehension of the effects of reducing the expression of these two enzymes, we next determined metabolite levels in the pericarp tissue of breaker-stage fruit using an established gas chromatography-mass spectrometry (GC-MS)-based metabolite profiling method . In the PEPCK lines, we observed a reduction in the levels of the amino acids Asp (all three lines), Ile (lines PEPCK7 and PEPCK11), and Thr (lines PEPCK7 and PEPCK10; Fig. 3). Similarly, a decrease in the levels of citric acid (all three lines), Fru and Glc (all three lines), as well as putrescine (lines PEPCK7 and PEPCK11) was observed (Fig. 3). Additionally, PEPCK lines showed an increase in the levels of Glu (all three lines) and galactinol (PEPCK10 and PEPCK11; Fig. 3). In the ME lines, a smaller number of changes were observed. These lines were characterized by increases of some amino acids, such as Asp (lines ME7 and ME13), Ser (lines ME7 and ME13), and Val (lines ME7 and ME13; Fig. 4), sugars and sugar derivatives such as Fuc (lines ME2 and ME7) and myoinositol (lines ME2 and ME13; Fig. 4), as well as putrescine (lines ME2 and ME13).

Metabolic Profiling during Ripening of PEPCK and NADP-ME Fruits
Given that changes in the content of primary metabolites during the ripening process are highly dynamic Osorio et al., 2011), we next evaluated metabolite contents during ripening, breaker + 5 DAP, and breaker + 10 DAP stages. In the PEPCK lines, many metabolites at breaker + 5 DAP and breaker + 10 DAP stages displayed similar patterns of change to those observed at breaker stage. This behavior was observed for amino acids such as Asp (all three lines) and Glu (all three Figure 1. Characterization and expression of tomato cytosolic PEPCK and NADP-ME. A and B, Screening by qRT-PCR in primary PEPCK (A) and ME (B) transformants in extracts of tissue sampled at breaker + 10 DAP. Black arrows indicate lines selected for further analysis. C and D, Total fruit yield (C) and fruit size (D) of the PEPCK and ME lines. For all parameters, values are presented as means 6 SE of eight biological replicates (one biological replicate is represented by one individual plant). Asterisks indicate values that were determined by Student's t test to be significantly different (P , 0.05) from the wild type (WT).  lines) as well as for sugars such as Fru (all three lines), galactinol (all three lines), Glc (all three lines), Fuc (all three lines for breaker + 5 DAP and lines PEPCK7 and PEPCK10 for breaker + 10 DAP), and putrescine (all three lines; Fig. 3). Other changes of note in the metabolite profiles at breaker + 5 DAP stage were the significant increases in dehydroascorbic acid (all three lines), malic acid (all three lines), and Rha (all three lines; Fig. 3) as well as decreases in Trp (all three lines) and GalA (all three lines; Fig. 3). Notably, at breaker + 10 DAP stage, Asn (all three lines), Met (lines PEPCK7 and PEPCK11), Pro (lines PEPCK7 and PEPCK10), g-aminobutyrate (GABA; lines PEPCK7 and PEPCK10), and myoinositol (lines PEPCK7 and PEPCK11) were significantly decreased, while Tyr was significantly increased (all three lines; Fig. 3). By contrast, evaluation of the primary metabolite composition of ME lines revealed that the pattern of changes in metabolites was more conserved in the comparison between breaker + 5 DAP and breaker + 10 DAP stages than with the breaker stage (Fig. 4). Only a reduction in Asp was conserved between the three studied fruit stages. In addition, other conserved changes between breaker + 5 DAP and breaker + 10 DAP stages were noted, such as decreases in the levels of some amino acids, such as GABA, Met, and Tyr, as well as in the organic acid GalA (Fig. 4). Increases of dehydroascorbic acid and Rha were observed in the transformants at these developmental stages (Fig. 4).

Deficiency in PEPCK and NADP-ME Leads to Alteration in Nucleotide and Starch Metabolism
To further characterize these lines, we evaluated one of the major products of the TCA cycle, namely pyrimidinereducing equivalents. This analysis revealed a decrease in NADPH levels in breaker stage of the PEPCK lines (lines PEPCK7 and PEPCK11) but an increase in NAD + (lines PEPCK10 and PEPCK11). Thus, the NADPH/ NADP + and NADH/NAD + ratios slightly decreased and increased in PEPCK11, respectively (Table III).
By contrast, the pattern of change of pyrimidine levels in the ME lines was markedly different. At breaker stage, the levels of NADP + increased in one line (ME13). As a consequence, the NADPH/NADP + ratio decreased (Table IV).
Since pyruvate and PEP metabolites are transformed via the gluconeogenic pathway to sugar phosphates, which can subsequently be converted to starch (Lara et al., 2004), and the fact that previous studies demonstrated that phosphoenolpyruvate carboxylase (PEPC) activity plays an important role within potato (Solanum tuberosum) starch metabolism (Rademacher et al., 2002), we next evaluated the starch content of both the wild type and transformants at breaker stage (Fig. 5). A significant decrease in starch levels was observed in both transgenic sets, PEPCK and ME lines, suggesting that the rate of transient starch degradation in the fruit is, at least partially, dependent on the total reserves of starch available (Fig. 5).

Enzymes Involved in Starch and Organic Acid Metabolism
To better understand the above-described metabolic alterations, we next surveyed the activities of a broad range of enzymes from central metabolism (Tables V and VI). In the PEPCK lines, these activities were generally invariant from the wild type, with the exception of the decrease in ADPglucose pyrophosphorylase (AGPase) and the total NADP-malate dehydrogenase (MDH) activities in two of the lines (PEPCK7 and PEPCK10) at breaker. In addition, a significant reduction was also observed in the level (or activity) of pyruvate kinase (PK) in line PEPCK11 and in NAD-ME and NADP-ME in line PEPCK7 in the same ripening stage (Table V).  Similar to the situation in the PEPCK lines, analysis of the ME lines at breaker stage revealed that the activities of both AGPase and NADP-MDH decreased in the lines ME2 and ME7 (Table VI). Moreover, a significant increase in PEPC (all three lines) and PK (lines ME2 and ME7) activities was also observed (Table VI).

Respiration in the Transgenic Fruits
In order to evaluate the effect of deficiencies in the expression of these enzymes on the respiration rate, we next evaluated the relative rate of flux through the major pathways of carbohydrate oxidation at breaker + 5 DAP. For this purpose, we incubated pericarp discs taken from fruits in the light and supplied them with [1-14 C]Glc, [3:4-14 C]Glc, or [6-14 C]Glc for a period of 6 h. During this period, we collected the 14 CO 2 evolved at hourly intervals. Carbon dioxide can be released from position C1 by the action of enzymes that are not associated with mitochondrial respiration, but carbon dioxide released from the C3,4 positions of Glc cannot . Thus, the ratio of carbon dioxide evolution from C1 to C3,4 of Glc provides an indication of the relative rate of the TCA cycle in comparison with the other routes of carbohydrate oxidation.
When the relative 14 CO 2 release of the transgenic and wild-type lines was compared, an interesting pattern emerges, with a relatively unaltered carbon dioxide evolution in the ME lines but the PEPCK lines displaying considerably altered carbohydrate oxidation. Both lines PEPCK7 and PEPCK10 displayed a greater relative release of 14 CO 2 from C3,4 positions than the wild type, indicating a greater flux through the reactions catalyzed by pyruvate dehydrogenase and/or ME (Fig. 6). The difference in release from the C6 position was less marked, however, suggesting that there was no alteration in fluxes through pentan synthesis (Stitt and ap Rees, 1979). Despite there being little change in the absolute release of 14 CO 2 from C3,4 in the MADP-ME lines, the relative release from the C3,4 positions was considerably higher in line ME2.

Effect of Decreased PEPCK and NADP-ME Activities on Metabolic Fluxes
The effect of reduced PEPCK and NADP-ME activities on carbon fluxes was determined by incubating pericarp discs at breaker + 10 DAP stage from the wild type and two lines of the PEPCK (PEPCK7 and PEPCK10) and ME (ME2 and ME7) transformants with [U-14 C]Glc for 2 h. Following this incubation, the discs were washed and frozen prior by fractionation of the labeled material to determine the label redistribution (Table VII). Neither line from the PEPCK or the ME transformants displayed differences from the wild type in the total label uptake or in the metabolized label incorporated into cell wall, amino acids, or Suc (Table VII). However, there was decreased label incorporation into starch, protein, and organic acids in lines of both transformants (Table VII). Moreover, PEPCK7 and PEPCK10 lines showed significant decreases of label into the pool of hexose phosphates (Table VII). Determination of the specific activity of the hexose phosphate pools reveals, as expected, that the PEPCK transformants had proportionally lower label in this metabolic pool. On the calculation of the metabolic fluxes, it emerged that both transformants, PEPCK and ME, displayed dramatic decreases in the rate of starch synthesis (Table VII). A similar pattern was found for the fluxes to organic acids and protein, although these changes are considerably less marked than for starch. To further investigate changes in primary metabolism, we fed isotopically labeled [ 13 C]pyruvic acid (Pyr) to breaker + 10 DAP fruits from PEPCK and ME lines. The absolute redistribution of 13 C was measured by assessing the incorporation of the isotope into each intermediate within the relevant pathways (Fig. 7) in metabolites closely associated with Pyr using an adapted GC-MS protocol that facilitates isotope tracing (Roessner-Tunali et al., 2004). Interestingly, the changes in redistribution of isotope were conserved across the transformants.
The PEPCK lines were characterized by elevated label incorporation in fumaric acid, malic acid, and Glu in addition to dramatic elevation in label incorporation in Asp. Moreover, they displayed reduced incorporation in Ala, GABA, Fru, Glc, and succinic acid (Fig. 7). However, the pattern of changes in label accumulation in the ME lines was different. This suggested that there were significant increases in label incorporation to citric acid and succinic acid. In contrast, decreased label incorporation in Ala, Glu, fumaric acid, and malic acid was observed (Fig. 7).

DISCUSSION
During the ripening of tomato, there is a large decrease in the organic acid content of the pericarp (Knee and Finger, 1992;. This decrease in organic acids could be brought about either by a restriction of their synthesis or an enhanced degradation potentially initially catalyzed by NAD-ME, NADP-ME, or PEPCK. Malic acid has long been implicated in the physiology of fruit ripening (Carrari and Fernie, 2006;Sweetman et al., 2009). It has recently been demonstrated that the level of malic acid plays an important role in the regulation of starch biosynthesis and thereby the accumulation of total soluble solid during tomato ripening. Detailed biochemical characterization revealed that this was due to regulation of the activation state of the first committed step of starch synthesis that is catalyzed by AGPase (Centeno et al., 2011). Intriguingly, however, this is not the first study to implicate reactions beyond the direct pathway of starch biosynthesis, since the mitochondrial NAD-ME (Jenner et al., 2001), the plastidial adenylate kinase (Regierer et al., 2002), and the cytosolic UMP synthase  have been demonstrated to   In our previous studies, we focused on mitochondrial reactions of the TCA cycle, and here, we evaluated the effect of modifying malate-pyruvate balances at different subcellular locations in a fruit ripening-specific manner by silencing either cytosolic PEPCK or the plastidic NADP-ME.
In tomato, as previously mentioned, a differential regulation of starch biosynthesis was observed in transgenic lines exhibiting different fruit malic acid levels in which the activation stage of AGPase correlated with the accumulation of transitory starch (Centeno et al., 2011). Similar, in this context, was the fact that the down-regulation of cytosolic PEPCK during ripening resulted in a significantly higher malic acid level at the breaker + 5 DAP stage. However, similar down-regulation of the plastidic NADP-ME during ripening displayed no significant changes in the malic acid level at steady state.
The observed decrease in starch content in the cytosolic PEPCK antisense lines at the breaker stage seems to result from an alteration in the malic acid level, as observed previously in the TCA cycle transformants (Centeno et al., 2011). However, this correlation was not apparent in plastidic NADP-ME antisense lines, which also revealed decreased starch content. Interestingly, however, we observed a significant decrease of NADPH in the cytosolic PEPCK lines, which leads to a slight decrease of the NADPH/NADP + ratio. On the other hand, only one line showed changes in the plastidial redox state in the plastidic NADP-ME antisense lines. Consistent with these changes in the plastidial redox status were the alterations in the activities and activation state of AGPase (in PEPCK and ME lines) and plastidial malate dehydrogenase (in ME lines), which are well documented diagnostics of alteration in redox status (Scheible and Stitt, 1988;Tiessen et al., 2002;Nashilevitz et al., 2010). The lack of absolute starch accumulation in both lines was also corroborated with a decrease in the starch biosynthesis flux. Thus, the observations described here demonstrate that, as we postulated previously, starch metabolism is strongly affected by changes in plastidial redox state but seems not to alter malate levels per se.
When other areas of metabolism are considered, some interesting observations are apparent. It is suggested that in fruit ripening, NADP-ME may produce the respiratory substrates pyruvate and NADPH from malate and NADP + (Ruffner, 1982;Drincovich et al., 2001) in cytosol and maybe thus partake in malate degradation during ripening (Ruffner et al., 1984;Franke and Adams, 1995). Here, feeding experiments involving the incubation of positionally labeled Glcs  showed that the strong decrease in plastidic NADP-ME levels did not lead to substantial changes in overall respiration rates and TCA cycle flux in tomato fruits. However, it is important to note that one line (ME2) was characterized by an increase in the C3,4-C1 ratio, also suggestive of an increased flux through pyruvate dehydrogenase (albeit one not mediated by an elevated enzymatic capacity of the NAD-ME). It is conceivable that the remaining activity of plastidic NADP-ME might still be high enough to supply the TCA cycle with sufficient pyruvate to maintain the overall respiration rate or that there are other alternative sources that may be able to continue to supply the mitochondria with pyruvate or some other carbon source, such as cytosolic NADP-ME, PK, and PEPC. In light of these observations, it is interesting that cytosolic NADP-ME expression was not modified in the ME lines. Interestingly, the activities of PK and PEPC enzymes displayed a clear increase in the lines with reduced plastidic NADP-ME activity. These observations indicate that the lack of plastidic NADP-ME activity increased the flow through these enzymes, which suggests that plastidic NADP-ME activity is likely an important source of pyruvate synthesis in ripening tomato fruits. Intriguingly, the large decrease in the organic acid content frequently documented as occurring in the pericarp on ripening (Knee and Finger, 1992;Carrari Figure 6. Respiratory parameters in breaker + 10 DAP fruits of PEPCK and NADP-ME lines. A and B, Evolution of 14 CO 2 from breaker + 10 DAP pericarp discs of PEPCK (A) and ME (B) lines. C and D, Ratio of dioxide evolution from C3,4/C1 positions of Glc in breaker + 10 DAP pericarp discs of PEPCK (C) and ME (D) lines. WT, Wild type. et al., 2006) correlates with a large increase in PEPCK (Bahrami et al., 2001). This prompts the question, what is the contribution of this enzyme to flux through the TCA cycle? If the TCA cycle is drained via OAA, PEPCK activity requires support from neither PEPC nor PK nor from any part of the TCA cycle. However, we can exclude an increase of PK and PEPC activities in compensation for the decrease in cytosolic PEPCK. Moreover, in the PEPCK lines, uniformly labeled Glc feeding experiments also revealed a reduced flux to protein. Furthermore, feeding experiments involving the incubation of positionally labeled Glcs suggested that the PEPCK lines displayed higher flux through the reactions catalyzed by pyruvate dehydrogenase, ME, or both. This can be concluded from the increased C3,4-C1 ratios displayed by PEPCK lines, since the carbon release from the C3,4 positions occurs via decarboxylation catalyzed by these enzymes. Both the mitochondrial NAD-ME and plastidial NADP-ME activities showed no differences in the PEPCK lines, suggesting that these fruits showed high flux through the TCA cycle via pyruvate dehydrogenase.
PEP required for gluconeogenesis in fruit may originate from malate through the activities of MDH and PEPCK or potentially MEs (Ruffner, 1982). Evidence from radiolabeling work in tomato fruit (Farineau, 1977;Halinska and Frenkel, 1991) suggested that gluconeogenesis does occur during ripening stages, when sugars are accumulating rapidly. Here, we observed that when cytosolic PEPCK was reduced (PEPCK lines), this led to reduced levels of the major sugars, Glc and Fru, during fruit ripening, accompanied by an accumulation of malate, which provides further evidence for gluconeogenesis from organic acids (Halinska and Frenkel, 1991) and the important role of PEPCK enzymes in this process.
In order to address the extent to which PEPCK activity contributes to the regulation of the TCA cycle, it is first necessary to consider how individual carbon fluxes around the complete TCA cycle differ depending on the particular biosynthetic output. When biosynthesis is absent, the net flux of carbon through each TCA cycle step will be equal, likewise if the TCA cycle's biosynthetic output is exclusively from OAA to Asp. By contrast, if the biosynthetic output of the TCA cycle is exclusively from 2-oxoglutarate to Glu, then the flux from malate to 2-oxoglutarate will be proportionately higher than the flux from 2-oxoglutarate to malate. Bearing these considerations in mind, it is interesting that we observed that during ripening, the level of Asp derived from OAA was highly reduced in fruits deficient in PEPCK activity compared with the wild type. By contrast, the steady-state level of Glu was higher than in the wild type during ripening. However, 13 C feeding experiments revealed a higher redistribution of the label in both metabolites (Asp and Glu). Interestingly, a higher redistribution of 13 C label in malic and fumaric acids as well as lower label in succinic acid were also displayed in the PEPCK lines. Although with these data, the flux of carbon through the TCA cycle from malate to 2-oxoglutarate and from 2-oxoglutarate to malate cannot be calculated, these data indicate that when carbon in the form of OAA cannot be converted into PEP through the activity of PEPCK and thus cannot completely flow through the TCA cycle, the flux from TCA intermediates toward Table VII. Metabolism of [U-14 C]Glc in pericarp discs from PEPCK and ME lines in breaker + 10 DAP fruits Pericarp discs were cut from six separate plants per genotype. After 2 h of incubation, the discs were extracted and fractioned. The specific activity of the hexose phosphate pool was estimated by dividing the label retained in the phosphate ester pool by the summed carbon of the hexose phosphates and used to calculate absolute fluxes. Values represent means 6 SE of six biological determinations. Values that are significantly different by Student's t test from the wild type are set in boldface (P , 0.05).

Parameter
Wild Type PEPCK7 PEPCK10 ME2 ME7 Label incorporated (Bq) Total uptake 287.2 6 16.5 235.2 6 12.2 208.3 6 14.8 260.6 6 19.2 269.5 6 5.4 Insoluble Starch 10.2 6 0.2 5.5 6 0.6 6.3 6 1.0 4.8 6 0.6 5.1 6 0.2 Cell wall 5.1 6 0.7 5.2 6 0.5 6.4 6 1.0 4.1 6 0.5 6.0 6 0.7 Protein 3.0 6 0.3 1.7 6 0.1 1.4 6 0.1 0.9 6 0.1 1.7 6 0.2 Soluble Organic acids 130.7 6 8.6 96.9 6 11.3 65.0 6 3.8 76.7 6 6.4 86.7 6 6.6 Total hexose phosphates 11.3 6 2. the synthesis of Asp is increased and reintroduced in the TCA cycle via 2-oxoglutarate/Glu. In conclusion, we present here compelling evidence of the importance of both cytosolic PEPCK and plastidic NADP-ME in normal tomato ripening metabolism. Both enzymes exhibited considerable influence on starch biosynthesis metabolism. Their functions in other areas of metabolism, however, are somewhat divergent with our results, suggesting that the lack of accumulation of transitory starch was reflected by a decrease of major sugars in the PEPCK lines. This observation was not seen in plastidic NADP-ME. Moreover, the inactivation of PEPCK affected the respiration rate, suggesting that excess OAA was converted to Asp and reintroduced in the TCA cycle via 2-oxoglutarate/Glu. These results indicate that the cytosolic PEPCK plays a role in carbon and nitrogen remobilization pathways during fruit ripening. However, further studies will be required to investigate these processes, and detailed analysis of starch metabolism revealed that this was impacted by the effect of alterations in the plastid redox status on AGPase. As such, these data provide direct support for our earlier model (Centeno et al., 2011), suggesting that malatemediated effects on starch metabolism are caused by cellular redox changes, AGPase, and pyrimidine nucleotide levels (in PEPCK lines). Rather unexpected, in this context, was the fact that the down-regulation of plastidic NADP-ME had an effect on starch and AGPase activation state but no effect was observed on malate levels. This observation demonstrates that changes in plastidial redox state may be totally independent of changes in the malate levels. Another explanation is that reduction in plastidic NADP-ME activity provokes a small reduction in plastidial malate that was not observed in the global malate pool, but it was enough to display a change in redox state to affect the activation state of AGPase. Increasing plastidial malate dehydrogenase corroborated the change in redox state in these plants (Scheible and Stitt, 1988;Tiessen et al., 2002;Nashilevitz et al., 2010).
Additionally, we also found that the negligible impact of NADPH-ME activity on respiratory metabolism can be explained by increasing PK and PEPC activities, which can compensate for the reduction in pyruvate derived from malate via plastidic NADP-ME.
Intriguingly, the changes in redox state observed here are not as severe as those described previously for TCA cycle manipulations, despite the fact that the inhibition of the enzymes modified here is at least as efficient as that in our previous study. For example, the lack of apparent changes in pigmentation or postharvest characteristics indicates that the plastidic NADP-ME and cytosolic PEPCK reactions are quantitatively less important for malate-mediated ripening control than those catalyzed by the mitochondrial malate dehydrogenase and fumarase enzymes. There have been recent advances in our understanding of the regulation of starch synthesis in response to metabolic signals.
However, our knowledge of the signal transduction cascades has not been elucidated, although this work highlights the complex interactions that exist between different metabolic pathways and suggests novel approaches for the manipulation of starch synthesis.

Generation of Transgenic Plants
A 560-bp fragment of SlPEPCK (accession no. AY007226) and a 601-bp fragment of SlNADP-ME (accession no. AF001269) were independently cloned using the RNAi approach into the vector PBINplus, between the E8 promoter and the octopine synthase gene terminator. These constructs were subsequently introduced into tomato (Solanum lycopersicum) plants by the Agrobacterium tumefaciens-mediated transformation protocol. Plants were selected and maintained as described in the literature (Tauberger et al., 2000). For initial screening, eight RNAi-PEPCK lines and nine RNAi-ME lines were selected on the basis of PEPCK and NADP-ME1 gene expression by qRT-PCR. This screening allowed the identification of three lines per construct that were taken to the next generation.

Analysis of PEPCK and NADP-ME mRNA Expression by qRT-PCR
Total RNA was extracted according to Bugos et al. (1995) with minor modifications. The integrity of the extracted RNA was checked by electrophoresis under denaturing conditions after treating the RNA with RNase-free DNaseI (Roche). First-strand cDNA synthesis of 1 mg of RNA in a final volume of 20 mL was performed with Moloney murine leukemia virus reverse transcriptase, Point Mutant RNase H Minus (Promega), according to the supplier's protocol using oligo(dT) T19 primer.

Metabolite Profiling
Metabolite extraction, derivatization, standard addition, and sample injection for GC-MS were performed according to Schauer et al. (2006). Both chromatograms and mass spectra were evaluated using TAGFINDER (Luedemann et al., 2008). The level of starch in the tissue was determined exactly as described previously (Fernie et al., 2001). NAD + , NADP + , NADH, and NADPH were determined as described previously (Schippers et al., 2008).

NAD-MDH and PK
These proteins were extracted and assayed as described by Centeno et al. (2011).

AGPase
AGPase activity was measured in the pyrophosphorolysis direction with a spectrophotometric assay as described (Tiessen et al., 2002(Tiessen et al., , 2003. Frozen tissues were homogenized in liquid nitrogen, and approximately 100 mg was extracted rapidly (1 min) with 1 mL of extraction buffer (50 mM HEPES-KOH, pH 7.8, and 5 mM MgCl 2 ) at 4°C. After centrifugation (30 s at 13,000g at 4°C), 10 mL of the supernatant was used for the AGPase assay. The reaction was performed in a total volume of 200 mL containing 50 mM HEPES-KOH, pH 7.8, 5 mM MgCl 2 , 10 mM Glc-1,6-bisP, 0.6 mM NADP + , 2.5 mM sodium-inorganic pyrophosphate, 1 unit mL 21 phosphoglucomutase, 2.5 units mL 21 Glc-6-P dehydrogenase, and a range of concentrations of ADP-Glc (0.4-1 mM) in the absence of inorganic phosphate and 3PGA, with or without dithiothreitol (10 mM) for activation assay. Reactions were followed online at 340 nm and were linear up to 30 min. AGPase activity was measured in the ADP-Glccleaving direction in the absence (Vsel) or presence of dithiothreitol (Vred). The ratio between the activities in these two assays (Vsel/Vred) is termed "activation state".

PEPCK
Protein extraction was performed as described by Bahrami et al. (2001). The activity of PEPCK was measured in the carboxylation direction by following oxidation of NADH at 340 nm at room temperature (Walker et al., 1999).

NAD-ME and PEPC
Protein extraction and enzyme activity were analyzed as described (Borsani et al., 2009).

Analysis of [ 13 C]Pyr-Labeled Samples
Discs (10-mm-diameter latitudinal core) from tomato fruit pericarp at breaker + 10 DAP stage from both transgenic lines (PEPCK and ME lines) and the wild type were fed with 34 mM [ 13 C]Pyr (Cambridge Isotope Laboratories) by incubation in buffered solution (10 mM MES-KOH, pH 6.5) for a period of 6 h. At the end of the incubation, discs were snap frozen in liquid nitrogen. They were subsequently extracted in 100% methanol and processed exactly as described (Timm et al., 2008). The metabolic fate of these substrates was subsequently assessed exactly as described previously (Tieman et al., 2006). A total of four biological replicates were used.

Incubation of Plant Material with [U-14 C]Glc
Discs (10 mm-diameter latitudinal core) from tomato fruit pericarp at breaker + 10 DAP stage from both transgenic lines (PEPCK and ME lines) and the wild type were taken. Discs were washed three times in fresh incubation medium (10 mM MES-KOH, pH 6.5), and then eight discs were incubated in 5 mL of incubation medium containing 1.00 mCi of [U-14 C]Glc (specific activity of 8.11 MBq mmol -1 ). Samples were then incubated for 2 h before washing again three times in unlabeled incubation medium and freezing in liquid nitrogen until further analysis. All incubations were performed in a sealed 100-mL flask at 25°C and shaken at 150 rpm. The evolved 14 CO 2 was collected (in hourly intervals) in 0.5 mL of 10% (w/v) KOH, and the amount of radiolabel was subsequently quantified by liquid scintillation counting. After this step, the discs were harvested, washed three times in buffer (eight discs per 100 mL), and frozen in liquid nitrogen to enable further analysis. A total of four biological replicates were used.

Fractionation of 14 C-Labeled Material
Tissue was fractionated exactly as described by Fernie et al. (2001), with the exception that hexoses were fractionated enzymatically rather than utilizing thin-layer chromatography as described by . Labeled Suc levels were determined after 4 h of incubation of 200 mL of total neutral fraction with 4 units mL 21 hexokinase in 50 mM Tris-HCl, pH 8.0, containing 13.3 mM MgCl 2 and 3.0 mM ATP at 25°C. For labeled Glc and Fru levels, 200 mL of neutral fraction was incubated with 1 unit mL 21 Glc oxidase and 32 units mL 21 peroxidase in 0.1 M potassium phosphate buffer, pH 6, for a period of 6 h at 25°C. After the incubation time, all reactions were stopped by heating at 95°C for 5 min. The label was separated by ion-exchange chromatography as described by Fernie et al. (2001). The reliability of these fractionation techniques has been thoroughly documented previously (Runquist and Kruger, 1999;Fernie et al., 2001;. Fluxes were calculated as described by Fernie et al. (2001), following the assumptions detailed by Geigenberger et al. (1997Geigenberger et al. ( , 2000.

Supplemental Data
The following materials are available in the online version of this article.
Supplemental Figure S1. Expression of SlNADP-ME genes during ripening.
Supplemental Figure S2. NADP-ME activity of tomato fruit at different pH ranges.