Deficiency of mitochondrial fumarase activity in tomato plants impairs photosynthesis via an effect on stomatal function


(fax + 49 331 567 8408; e-mail


Transgenic tomato (Solanum lycopersicum) plants expressing a fragment of a fumarate hydratase (fumarase) gene in the antisense orientation and exhibiting considerable reductions in the mitochondrial activity of this enzyme show impaired photosynthesis. The rate of the tricarboxylic acid cycle was reduced in the transformants relative to the other major pathways of carbohydrate oxidation and the plants were characterized by a restricted rate of dark respiration. However, biochemical analyses revealed relatively little alteration in leaf metabolism as a consequence of reducing the fumarase activity. That said, in comparison to wild-type plants, CO2 assimilation was reduced by up to 50% under atmospheric conditions and plants were characterized by a reduced biomass on a whole plant basis. Analysis of further photosynthetic parameters revealed that there was little difference in pigment content in the transformants but that the rate of transpiration and stomatal conductance was markedly reduced. Analysis of the response of the rate of photosynthesis to variation in the concentration of CO2 confirmed that this restriction was due to a deficiency in stomatal function.


Fumarase (fumarate hydratase; E.C. catalyses the reversible hydration of fumarate to malate, one of the constituent reactions of the tricarboxylic acid (TCA) cycle. Carbon flux through the TCA cycle can follow two different routes. Fumarase is, however, required for both the normal state, characterized by the decarboxylating energy-producing reactions of the cycle, and the carbon-conserving state, during which carbon is shunted through the glyoxylate cycle (Pracharoenwattana et al., 2005). In the early stages of oil-seed germination the majority of carbon flux has been reported to be via the glyoxylate cycle and the respective expression and activity of fumarase and NAD-dependent isocitrate dehydrogenase have been postulated to be diagnostic for this switch (Falk et al., 1998). Moreover, fumarase activities have been documented to be very high in guard cells in comparison to mesophyll cells in both Vicia faba and Pisum sativum (see Outlaw, 2003, for a review). Allosteric properties of purified pea fumarase, which revealed inhibition by physiological concentrations of pyruvate, 2-oxoglutarate and the adenine nucleotides ATP, ADP and AMP, are consistent with this step being an important control point in the cycle (Behal and Oliver, 1997). Furthermore, the importance of the fumarase reaction has recently been demonstrated in correlative studies demonstrating that this enzyme activity is a reliable diagnostic for the degree of seed dormancy in seeds of Picea abies, Pinus contorta, Betula pendula and Fagus sylvatica (Shen and Oden, 2002). Moreover, studies in cultured carrot cells suggest that fumarase can be exuded from cells and at least under these conditions is important in the utilization of extracellular malate (Kim and Lee, 2002). Despite the presence of these correlative reports, there has been very little analysis of this enzyme at the molecular genetic level. Although genes encoding fumarase have been cloned from many plant species including Arabidopsis, potato, apple, rice and soybean, only a single transgenic analysis, restricted to analysis of the protein content, has been carried out to date (Nast and Müller-Röber, 1996).

Unlike the situation in microbial systems, such as yeast, for which there is clear documentation of a dual targeted fumarase present in the cytosol and mitochondrion (Sass et al., 2003), the situation in plants is less well understood. The prevailing opinion is that plant fumarase is exclusively localized in the mitochondria (Gout et al., 1993). That said, in certain species such as Arabidopsis thaliana and Glycine max, fumarate has been suggested as a significant intercellular transport molecule (Chia et al., 2000). The recent cloning of a transporter of the mitochondrial membrane which is able to mediate transfer of succinate and fumarate (Catoni et al., 2003), and the fact that analysis of the Arabidopsis genome reveals two genes encoding fumarases (with only one of the encoded proteins bearing features indicative of mitochondrial targeting; Millar et al., 2004), give further credence to theories of intercellular transport of fumarate.

As part of our continuing investigation into the role of TCA cycle enzymes in the illuminated leaf, in this paper we describe the generation and characterization of transgenic tomato plants exhibiting antisense inhibition of a fumarase gene. We have previously comprehensively phenotyped the wild tomato species (Solanum pennellii) mutant Aco1, which exhibited a deficiency in expression of one of the two isoforms of aconitase present in the tomato (Carrari et al., 2003). This study revealed that despite exhibiting a decreased flux through the TCA cycle, this mutant was characterized by an increased rate of CO2 assimilation (Carrari et al., 2003) as well as an increase in fruit size. Given that S. pennellii is a green-fruited species bearing very small fruits, we also evaluated the effect of downregulating the TCA cycle in the elite cultivated species Solanum lycopersicum by generating and characterizing plants in which the mitochondrial malate dehydrogenase was antisense inhibited (Nunes-Nesi et al., 2005). The results of these two studies suggested an important role for respiratory pathways in photosynthetic metabolism and as such are in close agreement with results from other researchers (reviewed in Fernie et al., 2004a). Here we demonstrate that the reduction in mitochondrial fumarase activity has a negative impact on photosynthetic performance, reflecting the potential importance of malate metabolism in C3 photosynthesis (Martinoia and Rentsch, 1994). However, they suggest that its role is fairly specific, with detailed physiological, transcriptional and biochemical characterization revealing that the major impairment is in stomatal function.


Cloning of a cDNA encoding fumarase from tomato

Searching tomato expressed sequence tag (EST) collections (Van der Hoeven et al., 2003) revealed the presence of eight ESTs encoding fumarase, all of which belonged to a single tentative consensus (TC) sequence. Assembly and sequence analysis of this gene revealed an open reading frame of 493 amino acids. Comparison with the functionally characterized fumarases revealed high identity (97%) with the potato protein (StFUM1), whilst it also showed high homology to two of the Arabidopsis gene products (87% and 86% for AtFUM1 and AtFUM2, respectively) and weak homology to human (71%), yeast (67%) and Escherichia coli fumarase C (60%); no significant similarity was found between the deduced amino acid sequence of tomato fumarase and those of E. coli fumarases A and B. The presence of a single TC for fumarase in EST databases strongly suggests that, in contrast to A. thaliana which contains two genes (Millar et al., 2004), S. lycopersicum, like Solanum tuberosum (Nast and Müller-Röber, 1996), contains a single fumarase gene SlFUM1. However, results of a Southern blot in which we hybridized genomic DNA, digested with four different endonucleases, to the entire coding region of SlFUM1, yielded restriction fragments consistent with there being more than a single fumarase gene (Figure 1a). The most likely explanation for this discrepancy is that only a single isoform of fumarase is expressed in tissues from which the EST library was constructed. In support of this hypothesis Northern and Western blots (Figure 1b,c) indicated the presence of unique hybridizing signals which, given the high conservation amongst fumarase sequences, suggests that only one isoform is expressed in the leaves. Sequence analysis indicated that SlFUM1 has a putative mitochondrial transit peptide, whilst analysis of mRNA by Northern blots exhibits constitutive expression of the gene, with the transcript present at approximately equivalent levels in young and mature leaves, stems, roots and fruits (data not shown). In addition, the transcript is ubiquitous during fruit development (data not shown). Analysis of fumarase activity revealed that it was relatively high in roots and red fruits but generally lower in photosynthetic tissues and epidermal fragments harvested from leaves, irrespective of whether this was assessed on a per milligram protein or a per gram fresh weight basis. In addition, the activity in floral tissues was high but only when assessed on a per gram fresh weight basis (data not shown).

Figure 1.

 Characterization and expression of tomato fumarase.(a) Southern blot of genomic DNA from tomato (S. lycopersicum, cv. Moneymaker) digested with the restriction enzymes indicated and probed with the 1860-bp fragment of the S. lycopersicum fumarase gene. Numbers on the right indicate molecular masses in kilobase pairs.(b) Northern blot containing total RNA extracted from fully expanded leaves of 5-week-old plants.(c) Western blot analysis of leaves of 5-week-old transgenic plants with altered mitochondrial fumarase activity as compared with wild type (WT).(d) Fumarase activity determined in 5-week-old leaves taken from fully expanded source leaves of transgenic plants with altered expression of fumarase as compared with wild type. Values are presented as means ± standard error (SE) of five individual plants per line; an asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type.(e) Mitochondrial fumarase activity of WT and transgenic tomato leaves. Fumarase activity was determined across Percoll gradients. Subcellular fractionation was carried out in leaves taken from two 6-week-old plants per genotype; activities were calculated with respect to the level of contamination as assessed by marker enzyme assays. Values presented are the means ± SE of these two preparations. Black and grey bars represent total and mitochondrial fumarase activity respectively.

Transgenic plants show minor retardation in root growth and a reduced fruit yield

A 1860-bp fragment of the cDNA encoding fumarase was cloned in the antisense orientation into the transformation vector pBinAR between the cauliflower mosaic virus (CaMV) promoter and the ocs terminator. We then transferred 70 transgenic tomato plants obtained by Agrobacterium-mediated transformation to the greenhouse.

Screening of the lines by activity gel assays for a reduction of fumarase activity yielded eleven lines that displayed considerable reduction in activity, and Northern blot analysis confirmed that they additionally displayed reduced expression of the SlFUM1 gene. These lines were amplified in tissue culture and then transferred to the greenhouse. Following 5 weeks growth, the fumarase activity was determined by direct assay, using a recently established cycling assay (Gibon et al., 2004). Statistical analysis revealed that three of these lines, FL41, FL11 and FL56, exhibited reductions in enzyme activity that rendered them suitable for further analysis (Figure 1d), whilst line FL63 was also included for comparative purposes. Subcellular fractionation studies on lines FL11 and FL41 demonstrated that this reduction was a consequence of a specific reduction in mitochondrial activity (Figure 1e).

Analysis of the maximal catalytic activities of other important enzymes of photosynthetic carbohydrate metabolism revealed no consistent changes in important enzymes of photosynthesis (rubisco and transketolase), or the TCA cycle (citrate synthase, NAD-dependent malate dehydrogenase and succinyl-CoA ligase), but a minor reduction in enzymes of glycolysis (pyruvate kinase and ATP- and pyrophosphate (PPi)-dependent and phosphofructokinases), starch and sucrose synthesis [ADP-glucose pyrophosphorylase (AGPase) and sucrose phosphate synthase (SPS)] (Table 1). In addition, there were no major changes in the initial and total activities of the NADP-dependent malate dehydrogenase of the chloroplast – a commonly used diagnostic marker for alterations in plastidial redox status (Scheibe et al., 2005). An interesting feature of the data is the gross disparity in the activities of the tricarboxylic acid cycle enzymes, with the activities of fumarase, citrate synthase and NAD-dependent malate dehydrogenase being greatly in excess of succinyl-CoA ligase and NAD-dependent isocitrate dehydrogenase (Table 1).

Table 1.   Enzyme activities in antisense fumarase transgenic lines. Activities were determined in 6-week-old fully expanded source leaves harvested 6 h into the photoperiod. Data presented are means ± SE of measurements from six independent plants per genotype; values set in bold type were determined by the t-test to be significantly different (P < 0.05) from the wild type (WT)
  1. FW, fresh weight; MDH, malate dehydrogenase.

nmol min−1 g FW−1
AGPase1685.9 ± 213.01655.1 ± 212.81286.7 ± 237.8985.6 ± 80.0
Citrate synthase317.9 ± 27.5371.3 ± 53.2203.0 ± 24.7294.6 ± 12.2
NAD-isocitrate dehydrogenase21.3 ± 3.526.4 ± 7.245.0 ± 14.535.3 ± 4.8
NAD-MDH57402.7 ± 6108.659012.1 ± 5589.546372.6 ± 6921.543484.6 ± 2731.9
NADP-MDH initial activity106.0 ± 6.198.0 ± 2.380.8 ± 5.284.6 ± 3.1
NADP-MDH total activity164.7 ± 12.3209.8 ± 53.7146.5 ± 21.5135.7 ± 3.7
NADP-MDH % of activation64.9 ± 2.260.3 ± 12.258.9 ± 5.562.4 ± 2.0
Phosphofructokinase (ATP)360.9 ± 46.5151.7 ± 34.8204.9 ± 70.4219.6 ± 32.7
Phosphofructokinase (PPi)275.1 ± 32.4232.9 ± 26.9225.2 ± 44.6166.6 ± 9.4
PEP carboxylase590.6 ± 45.1571.6 ± 61.7475.7 ± 70.8454.5 ± 18.2
Pyruvate kinase275.3 ± 19.8251.2 ± 16.3213.2 ± 30.0224.2 ± 14.8
SPS1353.3 ± 218.01338.1 ± 134.41198.2 ± 190.2782.2 ± 73.3
Succinyl-CoA ligase40.0 ± 8.724.9 ± 5.229.0 ± 6.717.2 ± 5.4
Transketolase5351.8 ± 549.46279.2 ± 649.64877.5 ± 732.44591.3 ± 418.4
Rubisco2151.4 ± 258.51463.6 ± 315.22131.0 ± 321.71866.8 ± 249.9

When we grew the transgenic plants in the greenhouse side by side with wild-type controls, the transformants were taller at early developmental stages. Close examination of the transgenic plants revealed that the most severely inhibited lines were significantly taller (Figure 2a), due to a larger internodal interval (Figure 2b). Assessment of biomass revealed that the transformants displayed a trend of reduced total fruit mass (significantly so in lines FL41 and FL11; Figure 2d) but no consistent change in stem (Figure 2e) or leaf (Figure 2f) mass. These changes sum to a decreased total dry weight in the transformants (Figure 2g). When fruit weight was assessed on an individual fruit basis it was apparent that the fruits of the transformants were significantly smaller (Figure 2h). There was, however, no marked difference in leaf formation, onset of senescence or flowering time (data not shown).

Figure 2.

 Growth phenotype of 10-week-old antisense fumarase tomato plants.
(a) Height of plant. (b) Internode length after 5 weeks’ growth. (c) Total root dry weight. (d) Total fruit dry weight. (e) Total stem dry weight. (f) Total leaf dry weight. (g) Total plant dry weight. (h) Mean fruit weight.
Values are presented as means ± SE of six individual plants per line; an asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type.

Inhibition of fumarase results in a reduced flux through the TCA cycle

In our previous studies (Carrari et al., 2003; Nunes-Nesi et al., 2005) the analysis of the incorporation and subsequent metabolism of 14CO2 in genotypes deficient in the expression of TCA cycle enzymes suggested a reduction in flux through the cycle. However, analysis of the fumarase antisense plants revealed no such change in incorporation of radiolabel in TCA cycle intermediates, or downstream metabolites thereof, under CO2-saturating conditions (Table 2). In order to assess the rate of respiration more directly under normal growth conditions we took three complementary approaches. First, we measured the rate of dark respiration via infra-red gas exchange analyses in the wild type and lines FL11 and FL41 (Figure 3a). These measurements revealed a reduction in the rate of CO2 production with that of the transformants being less than half that observed in the wild type. Secondly, we incubated leaf discs taken from plants in the light and supplied these with [1-14C]-glucose, [2-14C]-glucose, [3,4-14C]-glucose or [6-14C]-glucose over a period of 6 h. During this time, we collected the 14CO2 evolved at hourly intervals. Carbon dioxide can be released from the C1 position by the action of enzymes that are not associated with mitochondrial respiration but CO2 released from the C3,4 positions of glucose cannot (Nunes-Nesi et al., 2005). Thus, the ratio of CO2 evolution from the C1 position of glucose to that from the C3,4 positions of glucose provides an indication of the relative rate of the TCA cycle with respect to other processes of carbohydrate oxidation. When the relative 14CO2 release of the transgenic and wild-type lines is compared for the various fed substrates, an interesting pattern emerges (Figure 3b). The rate of 14CO2 evolution is always highest in leaves incubated in [1-14C]-glucose; however, with the exception of line FL41, the total release of 14CO2 is reduced in this incubation. Similarly, the release of 14CO2 from the other positional labelled glucoses was significantly lower in lines FL11 and FL41. In addition to these changes in absolute release, there was a shift in the evolution of 14CO2 from the variously labelled glucose, with the relative release from the C3,4 positions much lower in the transgenic lines, FL11 and FL41, than in the wild type (after 6 h the C3,4/C1 ratios were: wild type, 0.92 ± 0.05; FL11, 0.73 ± 0.11; FL41, 0.61 ± 0.04). Thus, these data reveal that a lower proportion of carbohydrate oxidation is carried out by the TCA cycle in the transgenic lines in keeping with the observation of reduced dark respiration in these plants. The differences in release from C2 and C6 positions were far less marked, suggesting that there were no major alterations in metabolic fluxes involved in cycling through the pentose phosphate pathway or in pentan synthesis (Keeling et al., 1988). Thirdly, NMR experiments on isolated mitochondria extracted from young fruits, analogous to those previously carried out on mitochondrial malate dehydrogenase antisense plants (Nunes-Nesi et al., 2005), revealed an increase in the labelling of fumarate, consistent with a restriction in the flux through fumarase in vivo (data not shown).

Table 2.   Effect of decreased mitochondrial fumarase activity on photosynthetic carbon partitioning at the onset of illumination of 6-week-old fully expanded source leaves. Leaf discs were cut from six separate plants of each genotype at the end of the night and illuminated at 700 μmol photons m−2 sec−1 of photosynthetically active radiation in an oxygen electrode chamber containing air saturated with 14CO2. After 30 min, the leaf discs were extracted and fractionated. Values presented are the mean ± SE of measurements from six individual plants per genotype. Values in bold were determined by the t-test to be significantly different (P < 0.05) from the wild type (WT)
Label incorporated (Bq)
Total uptake200.6 ± 20.4189.3 ± 18.1206.7 ± 26.5187.8 ± 17.3
Starch106.1 ± 9.1109.0 ± 8.298.1 ± 9.3107.8 ± 8.3
Organic acids21.5 ± 2.615.1 ± 1.622.4 ± 3.115.9 ± 1.3
Amino acids21.9 ± 2.918.9 ± 2.025.7 ± 3.822.6 ± 3.2
Soluble sugars51.2 ± 5.946.3 ± 6.360.5 ± 10.241.4 ± 4.5
Redistribution of radiolabel (as percentage of total assimilated)
Starch53.0 ± 3.857.8 ± 2.448.2 ± 4.557.6 ± 2.7
Organic acids10.7 ± 1.17.9 ± 0.510.8 ± 1.28.5 ± 0.2
Amino acids10.9 ± 1.310.1 ± 1.012.4 ± 1.611.9 ± 1.1
Soluble sugars25.4 ± 2.224.1 ± 2.028.6 ± 2.622.1 ± 1.7
Figure 3.

 Respiration parameters in leaves of the antisense fumarase tomato plants.
(a) Dark respiration measurements performed in 5–6-week-old plants. Values are presented as mean ± SE of four determinations per line; an asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type.
(b) Evolution of 14CO2 from isolated leaf discs in the light. The leaf discs were taken from 10-week-old plants and were incubated in 10 mm MES–KOH solution, pH 6.5, 0.3 mm Glc supplemented with 2.32 kBq ml−1 of [1-14C]-, [2-14C]-,[3,4-14C]-, or [6-14C]-Glc at an irradiance of 200 μmol m−2 sec−1. The 14CO2 liberated was captured (at hourly intervals) in a KOH trap and the amount of radiolabel released was subsequently quantified by liquid scintillation counting. Values are presented as means ± SE of determinations on six individual plants per line.

Reduction in mitochondrial fumarase activity results in a reduction of photosynthesis that is independent of the chloroplast electron transport rate and correlates strongly with reductions in stomatal conductance

Given the reduction in plant growth, we next analysed whether the transformants exhibited altered photosynthetic rates. First, we studied the metabolism of 14CO2 by leaf discs excised from the wild-type and transformant plants. We were surprised to see that discs from the transformants displayed neither difference in assimilation rate nor in radiolabel distribution with respect to the wild type (Table 2). The lack of difference in assimilation in leaf discs incubated in saturating concentrations of 14CO2 suggests that the photosynthetic machinery is not compromised in the transformants. To investigate this hypothesis, chlorophyll fluorescence was measured in vivo, in a subset of the lines, using a pulse amplitude modulation (PAM) fluorimeter to calculate relative electron transport rates (ETRs) across a range of irradiances. However, whatever the photon flux density (PFD), the fumarase transformants did not display altered chloroplast electron transport (Figure 4a). Finally, gas exchange was measured directly, in the same lines as used for the electron transport measurements, under PFDs that ranged from 100 to 1000 μmol m−2 sec−1 (Figure 4b–d). The transformants exhibited assimilation rates that were lower than the wild type under all conditions, with the exception of the lowest light intensity in the case of line FL41 (Figure 4b). Analysis of other parameters of gas exchange revealed that the transformants were also characterized by reduced transpiration rates, regardless of the level of irradiance (Figure 4c), and a reduced stomatal conductance (Figure 4d). When taken together, these data indicate that the photosynthesis is reduced in these lines by a mechanism that restricts CO2 uptake via the stomata, but which is independent of the photosynthetic capacity of the chloroplast. Consistent with this finding is the high correlation between stomatal conductance and assimilation across the genotypes (Figure 4e).

Figure 4.

 Effect of decreased mitochondrial fumarase activity on photosynthetic parameters.
(a) In vivo chlorophyll fluorescence was measured as an indicator of the ETR by use of a PAM fluorometer at PFDs ranging from 100 to 800 μmol m−2 sec−1. (b) Assimilation rate as a function of light intensity. (c) Transpiration rate as a function of light intensity. (d) Stomatal conductance as a function of light intensity. (e) Assimilation rate as a function of stomatal conductance under conditions of varying light intensity. Each point represents a value from an individual measurement. (f) Assimilation rate as a function of CO2 concentration. (g) Stomatal conductance as a function of CO2 concentration. Each point is a mean ± SE of values from six replicates. (h) Time taken for stomatal opening following a dark to light transition. (i) Time taken for stomatal closure following a light to dark transition.
Values are presented as means ± SE of six individual determinations per line. All measurements were performed in 5–6-week-old plants. The lines used were: WT, black circles; FL41, dark grey triangles; FL11, grey circles. An asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type.

To further characterize photosynthesis in these lines we evaluated gas exchange over CO2 concentrations ranging from 100 to 2000 μmol mol−1. These experiments revealed that the transformants were only significantly compromised in assimilation rate (Figure 4f) and stomatal conductance (Figure 4g) at or below 400 μmol mol−1, thus supporting the conclusions from the experiments described above. As a final gas exchange experiment, we investigated the duration of stomatal opening and closing in the transgenics following dark to light and light to dark transitions. Interestingly, in both instances the transformants exhibited slower rates (Figure 4h,i), significantly so in the case of stomatal closing.

Photosynthetic carbon metabolism in the fumarase transformants

Analysis of the carbohydrate content of leaves from 5-week-old plants during a diurnal cycle revealed that the transformants were characterized by little change in the levels of sucrose (Figure 5a) but a significant reduction in the level of starch (Figure 5b), glucose and fructose (data not shown). The line FL11 was also characterized as exhibiting increased levels of malate (Figure 5c) and fumarate (Figure 5d). Evaluation of the content of leaf pigments revealed that they did not change in a manner consistent with the altered activity of fumarase (Table 3). In all cases metabolite contents were in a similar range to those previously reported for tomato (e.g. Nunes-Nesi et al., 2005).

Figure 5.

 Diurnal changes in sucrose (a), starch (b), malate (c) and fumarate (d) content in leaves of 5-week-old transgenic and wild-type tomato plants.At each time point, samples were taken from mature source leaves. The lines used were: WT, black circles; FL41, dark grey triangles; FL11, grey circles. The data presented are means ± SE of measurements from six individual plants per line; an asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type. Grey bars indicate the dark period; white bars indicate the light period.

Table 3.   Pigment contents in antisense fumarase tomato plants. Pigments were determined in 6-week-old fully expanded source leaves. Samples used were harvested at exactly the same time as those for enzyme determinations presented in Table 1. Values presented are means ± SE of six individual plants per line. Values in bold were determined by the t-test to be significantly different (P < 0.05) from the wild type (WT).
  1. FW, fresh weight.

nmol g FW−1
Neoxanthin106.0 ± 15.596.6 ± 10.0105.3 ± 11.893.9 ± 7.2
Violaxanthin101.0 ± 6.897.4 ± 8.2105.9 ± 7.075.2 ± 5.9
Antheraxanthin5.2 ± 0.45.6 ± 1.17.7 ± 0.74.3 ± 0.3
Lutein445.3 ± 21.8408.5 ± 24.7412.7 ± 25.0343.8 ± 25.4
Zeaxanthin19.8 ± 1.518.9 ± 1.320.7 ± 1.617.1 ± 1.0
Chlorophyll b1304.3 ± 77.61238.9 ± 43.71197.6 ± 80.31060.4 ± 84.7
Chlorophyll a4104.5 ± 235.74007.0 ± 129.83845.6 ± 222.83500.6 ± 211.2
β-Carotene260.3 ± 53.3267.0 ± 52.4255.4 ± 55.2233.8 ± 46.0
Chlorophyll a/b3.1 ± 0.03.2 ± 0.03.2 ± 0.03.3 ± 0.1

We next decided to extend this study to major primary pathways of plant photosynthetic metabolism by utilizing an established gas-chromatography mass-spectrometry (GC-MS) protocol for metabolic profiling (Fernie et al., 2004a). These studies revealed only minor changes in the levels of a wide range of organic acids, amino acids and sugars (Figure 6, Table S1). Notably alanine (line FL11), aspartate (lines FL11 and FL41), cysteine (line FL41), glycine (line FL41), methionine (line FL63) and valine (lines FL63 and FL11) were significantly increased whilst proline (lines FL11 and FL41) and tyrosine (line FL41) were significantly decreased. As would perhaps be expected, there was also considerable variation in the relative pool sizes of the organic acids measured, with citrate, isocitrate and threonate (all line FL41) being significantly increased and galacturonate (line FL41), maleate (all lines) and shikimate (line FL63) being significantly decreased.

Figure 6.

 Relative metabolite content of the fully expanded leaves from 6-week-old-plants of the antisense fumarase plants.
Metabolites were determined as described in Experimental procedures. The full data sets from these metabolic profiling studies are available as Table S1. Data are normalized with respect to the mean response calculated for the wild type (to allow statistical assessment, individual plants from this set were normalized in the same way). The lines used were: WT, black bars; FL63, dark grey bars; FL11, grey bars; FL41, white bars. Values are presented as means ± SE of six individual plants per line; an asterisk indicates values that were determined by the t-test to be significantly different (P < 0.05) from the wild type.

Since GC-MS does not readily allow the determination of phosphorylated intermediates, which are important diagnostic markers for alterations in photosynthesis (Stitt, 1997), we analysed these by means of cycling assays, regular linear spectrophotometric assays and HPLC methods (Supplementary Table S1). However, as was the case in the antisense mitochondrial malate dehydrogenase genotypes (Nunes-Nesi et al., 2005), none of these metabolites were significantly altered in the fumarase transformants.


The importance of mitochondrial metabolism in photosynthesis in the illuminated leaf has received much attention in the form of reverse genetic and inhibitor studies (Fernie et al., 2004b; Plaxton and Podesta, 2006). Respiration has been proposed to optimize photosynthesis in a number of ways, including facilitation of the export of excess reducing equivalents from the chloroplast, acceleration of photosynthetic induction and supplying ATP for anabolic processes of the cytoplasm (Plaxton and Podesta, 2006). Whilst reductions in the activity of the mitochondrial malate dehydrogenase or aconitase have been shown to result in an enhanced photosynthetic performance and increased harvest index of cultivated (S. lycopersicum) and wild (S. pennellii) tomato species, respectively (Carrari et al., 2003; Nunes-Nesi et al., 2005). Here we report on the consequences of the antisense inhibition of fumarase, a further enzyme of the TCA cycle.

Clones covering the full-length 1860-bp open reading frame of fumarase were isolated and utilized in the creation of transgenic lines containing significant reductions in the activity of this enzyme (Figure 1d). When the expression of this gene was altered using the antisense technique, changes in metabolism were relatively limited (Tables 1–3 and Figures 3 and 6). To summarize, the inhibition of fumarase activity was coupled to a reduction of dark respiration in addition to a relatively mild decrease in the flux through the TCA cycle in the light (Figure 3). The data presented here suggest that fumarase catalyses the conversion of fumarate to malate in vivo. However, the relatively small changes in other aspects of photosynthetic metabolism suggest that it is relatively unimportant in terms of cellular homeostasis. Interestingly, as was previously observed in the other genotypes deficient in enzymes of the TCA cycle, the activities of other key enzymes of carbon metabolism were largely unaltered, as were the levels of phosphorylated intermediates, suggesting that the restriction of photosynthesis was not mediated by a classical metabolic mechanism. Indeed, despite the fact that the previously characterized lines displayed large alterations in the expression of genes associated with photosynthesis (Urbanczyk-Wochniak et al., 2006), the fumarase lines described here were characterized by very few transcriptional changes (Nunes-Nesi and Fernie, unpublished). Furthermore, a broad GC-MS based metabolite profile revealed relatively few changes in metabolites, and those that were observed were relatively mild (Figure 6). The lack of major changes in the levels of organic acids of the TCA cycle is at first sight perplexing, the most likely explanation being the upregulation of other pathways of malate formation in the transformants. Other changes of note in the metabolite profiles were the decreases in maleate and proline and putrescine. The decrease in maleate concentration is surprising as it is most probably formed directly from fumarate by maleate isomerase. However, maleate could potentially be utilized by a maleate hydratase to help meet the cellular requirement for malate. The reasons behind the decrease in proline and putrescine are also unclear, although they, like maleate, are closely associated with reactions of the TCA cycle.

Despite lacking an obvious metabolic phenotype, the transformants were observed to be taller, yet to harbour less total biomass than the wild type, with this reduction being most noticeable in the fruits and roots (Figure 2). A reduction in root growth has previously been reported for plants exhibiting reduced expression of aconitase (Carrari et al., 2003) and the mitochondrial malate dehydrogenase (Nunes-Nesi et al., 2005). However, this characteristic was confined to a single line in the current study. In contrast, the reduction in fruit yield in the fumarase lines was directly opposite to the situations observed previously. Both the Aco1 mutant and the mitochondrial malate dehydrogenase antisense line were characterized by an elevated fruit yield that was essentially proportional to the increased rate of photosynthesis observed in these lines (Carrari et al., 2003; Nunes-Nesi et al., 2005). Analysis of the rate of assimilation in the lines described here revealed that it was reduced, as well as fruit yield (at least under the conditions described here). This observation is interesting, given that several further studies have indicated that fruit yield in tomato is highly dependent on photosynthetic efficiency (Baxter et al., 2005).

Whilst these phenotypes are interesting, we focused our efforts here on understanding the restriction in photosynthesis. As described above, this was clearly not due to a classical metabolic restriction of photosynthesis given that we observed an altered rate of assimilation under normal atmospheric conditions (Figure 4b), but not under saturating CO2 concentrations (Table 2), suggesting that the inhibition of photosynthesis was due to an impaired stomatal functioning. In keeping with this, plotting the assimilation rate against stomatal conductance across all light intensities and genotypes used revealed a very strong correlation between the parameters (Figure 4e). Although this evidence is largely circumstantial, strong support for it was obtained by demonstrating that the rate of assimilation (and stomatal conductance) was only affected at relatively low concentrations of CO2. Interestingly, our studies also showed that the dynamics of stomatal opening were altered, with the transformants displaying significantly reduced rates of closure. Whether this is the result of alteration in the guard cells alone, which have been reported to have exceptionally high mitochondrial density (Outlaw, 2003), or is a mesophyll or even non-leaf derived signal, is, however, not apparent from the current study.

In summary, the work presented here reveals that the phenotype obtained following reduction of mitochondrial fumarase activity is distinct from that observed in the Aco1 mutant and the mitochondrial malate dehydrogenase antisense lines (Carrari et al., 2003; Nunes-Nesi et al., 2005). The maximal catalytic activity of fumarase in the wild type is relatively high and it seems that the residual activity of fumarase in the transgenics is more than enough to ensure relatively normal metabolic function. Despite the relatively minor metabolic consequences, decreased expression of fumarase leads to a restriction in photosynthesis and a consequent reduction in total plant biomass and harvest index. Detailed biochemical and physiological characterization of the lines reveals that the mechanism by which the inhibition of photosynthesis is achieved, is a defect in stomatal efficiency. However, although some of the metabolic changes observed, such as the decrease in putrescine, can be postulated to influence ion channel-mediated stomatal opening (Liu et al., 2000), the precise factors mediating this phenotype remain unknown. Whilst these data provide a clear link between altered mitochondrial metabolism and stomatal functioning, future experimentation including a deeper focus at the level of the guard cell will be required in order to elucidate the precise factors underlying this phenomenon.

Experimental procedures


Tomato (S. lycopersicum L.) cv. Moneymaker was obtained from Meyer Beck (Berlin). Plants were handled as described in the literature (Carrari et al., 2003; Nunes-Nesi et al., 2005). All chemicals and enzymes were obtained from Roche Diagnostics (, with the exception of radiolabelled sodium bicarbonate and d-[1-14C]-, d-[3,4-14C]- and d-[6-14C]-glucose (Glc), which were from Amersham International (; d-[2-14C]-Glc was from American Radiolabeled Chemicals (; and the [3-13C]-pyruvate was from Aldrich Chemical (

Complementary DNA cloning and expression

The 1860-bp fragment of the SlFUM1 was cloned in antisense orientation into the vector pBinAR (Liu et al., 1999) between the CaMV 35S promoter and the ocs terminator. This construct was introduced into plants by an Agrobacterium-mediated transformation protocol, and plants were selected and maintained as described in the literature (Tauberger et al., 2000). Initial screening of 70 lines was carried out using activity gel assays as described by Shaw and Prasad (1970) and Northern blot analyses. These screens allowed the selection of 11 lines, which were taken to the next generation. Total fumarase activity was confirmed in the second harvest of these lines after which four lines were chosen for detailed physiological and biochemical analyses.

Immunodetection of fumarase protein

Western analysis of fumarase protein was carried out on 25 μg of crude protein extract, exactly as previously described (Nunes-Nesi et al., 2005) using antibodies raised against Arabidopsis fumarase (Behal and Oliver, 1997).

Northern blot analysis

Total RNA was isolated using the commercially available TRIzol kit (Invitrogen, according to the manufacturer’s suggestions for the extraction from plant material. Hybridization using standard conditions was carried out using the ESTs for fumarase obtained from the Clemson University (Clemson, SC, USA) collection.

Southern blot analysis

Genomic DNA (15 μg) from S. lycopersicum was isolated and Southern blotted using standard conditions as described by Sambrook et al. (1989). The blot was hybridized with a PCR product of 1860 bp amplified using the oligonucleotide sequences 5′-CTCAGACTCAGCATCAAC-3′ and 5′-ATTCATGGCCCAACCTCATA-3′ (forward and reverse primer, respectively), corresponding to the full-length sequence of the SlFUM-1 gene. The hybridization was performed at 65°C and washes were done at low stringency (2 × SSC and 1 ×  SSC, 0.1% w/v SDS) for 2 h and high stringency (0.5 × SSC and 0.2 ×  SSC, 0.1% w/v SDS) for 1 h at 60°C.

Analysis of enzyme activities

Enzyme extracts were prepared as described previously (Gibon et al., 2004), except that Triton X100 was used at a concentration of 1% and glycerol at 20%. Fumarase, AGPase, phosphofructokinase (ATP- and PPi-dependent), phosphoenolpyruvate (PEP) carboxylase, pyruvate kinase, SPS and transketolase activities were determined as described by Gibon et al. (2004). Rubisco activity was determined as described by Sharkey et al. (1991). Citrate synthase was assayed by incubating for 40 min crude extract or NADH standards in a freshly prepared medium containing malate 0.3 mm, malate dehydrogenase 5 U ml−1, 0.25 mm NAD+, 0.25 mm acetyl-CoA in 50 mm Tricine buffer pH 8.5. The reaction was stopped by addition of an equal volume of 0.5 m NaOH. Isocitrate dehydrogenase (NAD) was assayed by incubating crude extract or NADH standards in a freshly prepared medium containing 5 mm MgSO4, 1 mm NAD+, 1 mm isocitrate in 50 mm 3-(N-morpholino)-propanesulphonic acid (MOPS) buffer pH 7.5 for a period of 40 min. The reaction was stopped by addition of an equal volume of 0.5 m NaOH. The NADH produced by citrate synthase and isocitrate dehydrogenase (NAD) was then determined using the NAD+-based cycling protocol described in Gibon et al. (2004). Succinyl-CoA ligase was assayed as described in Studart-Guimarães et al. (2005). Malate dehydrogenase (NADP) was assayed as described in Scheibe and Stitt (1988) and malate dehydrogenase (NAD) as described by Jenner et al. (2001).

The mitochondrial activity was subsequently determined applying the same method to mitochondrial fractions following the protocol for mitochondrial isolation described by Sweetlove et al. (2002). Activities are given per gram fresh weight based on the amount of tissue from which the mitochondria were prepared. The purity of the mitochondrial preparations was confirmed by measurement of cytochrome c oxidase (CCO) and UDP-glucose pyrophosphorylase (UGPase) (Carrari et al., 2003), which serve as marker enzymes for the mitochondria and cytoplasm, respectively. In all instances, the contamination of the mitochondria was less than 5%, and recoveries of both marker enzymes were 88% for UGPase and 90% for CCO.

Determination of metabolite levels

Leaf samples were taken at the time point indicated, immediately frozen in liquid nitrogen and stored at –80°C until further analysis. Extraction was performed by rapid grinding of tissue in liquid nitrogen and immediate addition of the appropriate extraction buffer. The levels of starch, sucrose, fructose and glucose in the leaf tissue were determined exactly as described previously (Fernie et al., 2001). Levels of glycolytic intermediates were measured as described in Gibon et al. (2002). Nucleotides and nucleosides were determined by HPLC as detailed by Fernie et al. (2001), while phosphate was determined using the protocol described by Sharkey and Vanderveer (1989). Malate and fumarate were determined using aliquots of 10-μl extract or standards (ranging from 0 to 20 nmol) which were pipetted into a microplate containing 100 mm Tricine/KOH pH 9, 3 mm NAD+, 1 mm methylthiazolyldiphenyl-tetrazolium bromide, 0.4 mm phenazine ethosulphate and 0.5% v/v Triton X100. After reading the absorbance at 570 nm for 5 min, 1 U of malate dehydrogenase was added and the absorbance was read until stability. Subsequently, 0.1 U of fumarase was added and the absorbance read until stability. The levels of all other metabolites were quantified by GC-MS as described by Roessner et al. (2001), with the exception that the peak identification was optimized for tomato tissues (Roessner-Tunali et al., 2003) and the metabolites studied included recent additions to our mass spectral libraries (Schauer et al., 2005). Photosynthetic pigments were determined exactly as described in Bender-Machado et al. (2004).

Measurements of photosynthetic parameters

The 14C-labelling pattern of sucrose, starch and other cellular constituents was performed by illuminating leaf discs (10-mm diameter) in an oxygen electrode chamber (Hansatech, containing saturated level of 14CO2 at a PFD of 700 μmol photons m−2 sec−1 of photosynthetically active radiation at 22°C for 30 min, and subsequent fractionation was performed exactly as detailed by Lytovchenko et al. (2002). Fluorescence emission was measured in vivo using a PAM fluorometer (Walz; on 5-week-old plants maintained at fixed irradiance (250 and 700 μmol photons m−2 sec−1) for 30 min prior to measurement of chlorophyll fluorescence yield and relative ETR, which were calculated using the WinControl software package (Walz). Gas-exchange measurements were performed in a special custom-designed open system (Lytovchenko et al., 2002). The CO2 response curves were measured at saturating irradiance with an open-flow gas exchange system (LI-COR, model LI-6400;

Measurement of respiratory parameters

Dark respiration was measured using the same gas exchange system as defined above. Estimations of the TCA cycle flux on the basis of 14CO2 evolution were carried out following incubation of isolated leaf discs in10 mm 2-(N-morpholine)-ethanesulphonic acid (MES)–KOH, pH 6.5, containing 0.3 mm Glc supplemented with 2.32 kBq ml−1 of [1-14C]-, [2-14C]-,[3,4-14C]- or [6-14C]-Glc at an irradiance of 200 μmol m−2 sec−1. Evolved 14CO2 was trapped in KOH and quantified by liquid scintillation counting. Estimations of the TCA cycle flux were also performed in isolated mitochondria. For this purpose, mitochondria were isolated from fruit following the method of Holtzapffel et al. (2002). After isolation and purifications steps, the mitochondria were pelleted and resuspended in the wash buffer with 10% DMSO and rapidly frozen and stored at –80°C prior to NMR analysis described below.

Metabolism of [3-13C]-pyruvate

The metabolism of [3-13C]-pyruvate by isolated mitochondria was analysed by 13C NMR spectroscopy using the protocol employed by Nunes-Nesi et al. (2005), except for an increase in the hexokinase concentration from 0.1 to 0.15 U ml−1.

Statistical analysis

The t-tests were performed using the algorithm embedded into Microsoft Excel (Microsoft, The term significant is used in the text only when the change in question has been confirmed to be significant (P < 0.05) with the t-test.


We are grateful to Lothar Willmitzer (Max-Planck-Institut für Molekulare Pflanzenphysiologie) for support and discussions, as well as Fábio Murilo Da Matta and Marcelo Ehlers Loureiro (Departamento de Biologia Vegetal, Universidade Federal de Viçosa, Minas Gerais, Brazil). We thank Drs Robert Behal and David Oliver (Iowa State University, Ames, IA, USA) for provision of the fumarase antibody. We are also thankful to Dirk Buessis (Max-Planck-Institut für Molekulare Pflanzenphysiologie) for help in the organization of gas-exchange measurements. We are indebted to Antje Lohmann and Peter Doermann (Max-Planck-Institut für Molekulare Pflanzenphysiologie) for pigment measurements and Janneke Hendriks (Max-Planck-Institut für Molekulare Pflanzenphysiologie) for the optimization of the fumarate and malate assays. In addition we acknowledge the excellent care of the plants by Helga Kulka (Max-Planck-Institut für Molekulare Pflanzenphysiologie).

This work was supported by grants from the Max Planck Gesellschaft (to ANN, FC, YG and RS), the Conselho Nacional de Desenvolvimento Científico e Technológico (CNPq, Brazil; to ANN) and the BBSRC (to LJS and RGR).