Antisense inhibition of enolase strongly limits the metabolism of aromatic amino acids, but has only minor effects on respiration in leaves of transgenic tobacco plants


Author for correspondence:
Frederik Börnke
Tel: +49 9131 85 25239


  • • Enolase catalyses the reversible conversion of 2-phosphoglycerate and phosphoenolpyruvate in glycolysis. Phosphoenolpyruvate constitutes an important branch point in plant metabolism. It is converted to pyruvate by pyruvate kinase and organic acids by phosphoenolpyruvate carboxylase. Phosphoenolpyruvate also acts as a precursor for the synthesis of aromatic amino acids in plastids.
  • • Tobacco (Nicotiana tabacum) enolase antisense plants were analysed for changes in metabolite composition, respiration and photosynthetic parameters.
  • • Antisense repression resulted in up to a 95% reduction in total enolase activity. It also resulted in fundamental changes in foliar metabolism. Although 2-phosphoglycerate remained largely unaltered, there was a substantial decrease in phosphoenolpyruvate. The levels of aromatic amino acids and secondary phenylpropanoid metabolites that are derived from these compounds decreased strongly, as did branched chain amino acids. The level of pyruvate was unaltered, as was the rate of respiration. There were substantial increases in tricarboxylic acid cycle intermediates, including a 16-fold increase in isocitrate, an increase in the total free amino acid content, including a 14-fold increase in asparagine and glutamine, and a 50% decrease in free sugars.
  • • We conclude that a decrease in enolase activity affects secondary pathways, such as the shikimate branch of amino acid biosynthesis, but does not inhibit the rate of respiration.


The glycolytic pathway, comprising the 10 consecutive enzymatic steps oxidizing glucose to pyruvate in the plant cytosol, fulfils two fundamental roles. First, it generates adenosine triphosphate (ATP) and reducing equivalents and, second, it produces building blocks for anabolism. Glycolysis is assumed to be particularly important in actively growing autotrophic tissue, where its intermediates are utilized in the biosynthesis of numerous compounds, such as phenylpropanoids, isoprenoids, amino acids and fatty acids (ap Rees, 1990). In the past few years, a number of reverse genetic approaches have been used to analyse the importance and control of the glycolytic pathway in plants (reviewed in Plaxton & Podestá, 2006). Such studies have often focused on those enzymes that catalyse irreversible reactions and that are subject to allosteric regulation, such as ATP-dependent phosphofructokinase (ATP-PFK), pyrophosphate-dependent phosphofructokinase (PFP) and cytosolic pyruvate kinase (PKc). Experiments with transgenic plants, examining these candidate pacemaker enzymes in the glycolytic pathway, have not revealed major changes in metabolite levels or fluxes in response to a decrease in PFP or PKc or an increase in ATP-PFK enzyme activity (Gottlob-McHugh et al., 1992; Hajirezaei et al., 1993; Burell et al., 1994). Instead, it has been found that the antisense repression of enzymes catalysing reversible reactions, such as cytosolic phosphoglucomutase and cytosolic phosphoglycerate mutase by 60% and 75%, respectively, reduces the glycolytic flux in potato tubers and leaves (Fernie et al., 2002; Westram et al., 2002).

Enolase (2-phospho-d-glycerate hydratase; E.C. catalyses the reversible dehydration of 2-phosphoglycerate (2PGA) to phosphoenolpyruvate (PEP) in glycolysis. Genes encoding plant enolases have been cloned from several species, such as Arabidopsis, tomato (Van der Straeten et al., 1991), castor bean (Blakeley et al., 1994), Mesembryanthemum cristallinum (Forsthoefel et al., 1995), maize (Lal et al., 1991, 1998) and Echinochloa phyllopogon (Fox et al., 1995). Although plant enolase is not subject to allosteric regulation, it has been reported to be regulated by a cytosolic thioredoxin in M. crystallinum and Arabidopsis thaliana (Anderson et al., 1998). Among other glycolytic enzymes, enolase has been implicated to play an important role during adaptation to anaerobiosis (Lal et al., 1998). Two cDNAs encoding cytosolic enolases have been cloned from maize: ZmEno1 and ZmEno2 (Lal et al., 1991, 1998). Although both maize isoforms share nearly 90% identity at the amino acid level, they have remarkably different expression patterns. The expression of ZmEno2 is constitutive under aerobic conditions, whereas ZmEno1 levels were induced 10-fold in maize roots after 24 h of anaerobic treatment.

PEP generated through the enolase reaction in the cytosol represents a central metabolite in plant primary and secondary metabolism, and has several possible fates. It can either be converted by PK to pyruvate, which is subsequently further respired in the tricarboxylic acid (TCA) cycle, or converted to oxaloacetate by phosphoenolpyruvate carboxylase (PEPC). The latter is an essential anaplerotic reaction and is needed to replenish the TCA cycle when organic acids are being used as building blocks for the synthesis of amino acids and other biomolecules. PEP can also be transported via the phosphoenolpyruvate/phosphate translocator (PPT) of the inner envelope membrane (Fischer et al., 1997) into the plastid stroma, where it serves as a precursor for the biosynthesis of aromatic amino acids in the shikimate pathway. Aromatic amino acids are themselves precursors for the synthesis of phenolics, phenylpropanoids and prenylquinones (tocochromanols, plastoquinone) (Herrmann & Weaver, 1999). Further, after conversion to pyruvate by plastid PK, PEP can also act as a precursor for the biosynthesis of fatty acids (Qui et al., 1994), branched chain amino acids (Schulze-Siebert et al., 1984) and isoprenoids (Lichtenthaler et al., 1997).

Although the glycolytic sequence is complete in the plant cytosol, the set of glycolytic enzymes has been found to be incomplete in some types of plastid. Although all enzymes of the glycolytic pathway are present in plastids of developing wheat and castor bean seeds (Plaxton, 1996), pea, spinach and Arabidopsis chloroplasts and cauliflower bud amyloplasts lack phosphoglycerate mutase and/or enolase (Stitt & ap Rees, 1978; Journet & Douce, 1985; Bagge & Larsson, 1986; Van der Straeten et al., 1991; Prabhakar et al., 2009). Based on the experimental evidence gathered so far, it appears likely that mature chloroplasts are unable to convert the primary Calvin cycle product 3PGA to PEP, and may therefore rely on the import of PEP from the cytosol. The characterization of the Arabidopsis cue1 mutant demonstrated that the provision of cytosolic PEP to the stroma via PPT is essential for the shikimate pathway (Streatfield et al., 1999), and more recent studies have indicated that the resulting shortage of PEP in the chloroplast stroma can be bypassed by overexpressing pyruvate, orthophosphate dikinase (PPDK) in the cue1 background (Voll et al., 2003).

In this study, we sought to investigate how a restriction of PEP supply within the cytosol through antisense repression of enolase affects the PEP-consuming pathways, such as mitochondrial respiration and the synthesis of aromatic amino acids, within the plastid. To this end, we carried out a detailed physiological analysis of transgenic tobacco plants with reduced expression of cytosolic enolase (Eno antisense plants). The transgenics display a visual phenocopy of the reticulate Arabidopsis cue1 mutant phenotype, and many changes in metabolism (Streatfield et al., 1999). It was found that, although pyruvate levels and respiration were not decreased, there was a strong decrease in PEP and in metabolites synthesized from PEP via the shikimate pathway. Further, there were major changes in the levels of sugars, organic acids and amino acids. These results indicate that the glycolytic PEP supply is a pacemaker for plastidic amino acid biosynthesis.

Materials and Methods

Transgenic plants, growth and maintenance

Tobacco plants (Nicotiana tabacum L. cv. Samsun NN) were grown in tissue culture under a 16-h light : 8-h dark period (50 μmol m−2 s−1 light, 21°C) at 50% relative humidity on Murashige and Skoog medium (Sigma) containing 2% (w/v) sucrose, and were transferred to soil after 14 d of cultivation. In the glasshouse, plants were grown in soil (Fruhstorfer Erde Typ T; Hawita Gruppe GmbH, Vechta, Germany) with daily watering, and subjected to a 16-h light : 8-h dark cycle (25°C : 21°C) at 300 μmol m−2 s−1 light and 75% relative humidity for an additional 6–7 wk until sampling was performed.

Transformation of tobacco plants by Agrobacterium-mediated gene transfer using Agrobacterium tumefaciens strain C58C1:pGV2260 was carried out as described previously (Rosahl et al., 1987).

Construction of the tobacco enolase antisense vector and phenotyping of transgenic plants were performed as described in Lein et al. (2008).

RNA isolation and reverse transcriptase-polymerase chain reaction (RT-PCR) analysis

Total RNA was extracted from tobacco leaf material as described by Logemann et al. (1987). For RT-PCR experiments, 2.5 μg of DNAse-treated total RNA was reverse transcribed into cDNA with oligo(dT) (30-mer) using M-MLV[H] reverse transcriptase (Fermentas, St Leon-Roth, Germany). A fraction (c. 1/10) of the first-strand cDNAs was used as a template for PCR with gene-specific primers in a volume of 100 μl with 1 U of Taq-polymerase (Takara, Shouzo, Japan), 20 μm of each deoxynucleoside triphosphate (dNTP) and 0.25 μm of each primer. An initial denaturation step for 5 min at 95°C was followed by 25–35 cycles of 5 s at 95°C, 45 s at 55°C and 1 min at 72°C. PCR products were separated on 1% (w/v) agarose gels containing ethidium bromide and visualized by UV light using a transilluminator (PeqLab GmbH, Erlangen, Germany). Amplification of ubiquitin using primers 5′-ATGCAGAT(C/T)TTTGTGAAGAC-3′ and 5′-ACCACCACG(A/G)AGACGGAG-3′ served as an internal control.

Cloning of a cDNA fragment encoding cytosolic enolase from tobacco

The entire coding region of tobacco enolase was amplified from cDNA as described above using primers FB464 (5′-CACCATGGCAACTATCAAATCCATTAAG-3′) and FB466 (5′-GTAGGGCTCGACAGGCTTGCGGAA-3′). The resulting fragment was cloned into pENTR-D/TOPO (Invitrogen) and entirely sequenced. The tobacco enolase sequence was submitted to Genbank/DDBJ/EMBL under accession number FJ979826.

Determination of protein content

Protein content was determined according to Bradford (1976) with bovine γ-globulin as the standard.

Measurement of enzymatic activities

Leaf punches (0.63 cm2) were extracted in 5 vol of N2-purged extraction buffer [50 mmN-2-hydroxyethylpiperazine-N‘-2-ethanesulphonic acid (Hepes)/KOH, pH 7.5; 5 mm MgCl2; 2 mm EDTA; 5 mm dithiothreitol (DTT); 10 μm phenylmethylsulphonylfluoride (PMSF); 0.1% (v/v) Triton-X-100; 50% (v/v) glycerol], immediately stored at −20°C and assayed for the indicated enzyme activities in a microtitre plate reader within 2 h after clarification of the extract for 2 min at 20 000 g. The activities of PK, PEPC, PPDK and NADP-GAPDH (Voll et al., 2003), fructose-1,6-bisphosphatase (FBPase) (Zrenner et al., 1996), phosphoenolpyruvate carboxykinase (PEPCK) (Walker et al., 1999), enolase (Mujer et al., 1995) and hexokinase (Wiese et al., 1999) were determined from crude extracts, as described previously.

Determination of metabolite levels

For metabolite determination, leaf material was harvested either 5 h into the photoperiod or 4 h after the beginning of the dark period and immediately frozen in liquid nitrogen. Phosphorylated intermediates were determined from neutralized perchloric acid extracts in a fluorescence microtitre plate reader, as established by Häusler et al. (2000). Carboxylic acid contents were determined from the same extracts according to Schneidereit et al. (2006).

Measurement of photosynthesis and respiration

Photosynthetic performance was determined with a combined infrared gas exchange analyser/chlorophyll imaging system (GFS-3000 and MINI-Imaging-PAM Chlorophyll Fluorometer; Walz, Effeltrich, Germany), as described in Abbasi et al. (2009). The assimilation rate (A) was calculated according to von Caemmerer & Farquhar (1981). The chlorophyll fluorescence parameters of electron transport rate (ETR) and nonphotochemical thermal energy dissipation (NPQ) were calculated according to Genty et al. (1989), and Y(NO), the quantum yield of nonregulated energy dissipation, was calculated according to Kramer et al. (2004). For the measurement of dark respiration, plants were allowed to dark adapt for 2 h before steady-state respiration was assessed over a period of 30 min with the GFS-3000.

Determination of leaf carbohydrates, amino acids and UV-absorbing substances

Carbohydrates and amino acid contents were measured as in Abbasi et al. (2009); UV-absorbing compounds were measured according to Pinto et al. (1999).


Identification of essential genes for photosynthetic development

Previously, we have described a functional genomics effort that combines large-scale expressed sequence tag (EST) sequencing, high-throughput gene silencing and visual phenotyping to identify genes whose expression is necessary for normal photosynthetic development of tobacco leaves (Lein et al., 2008). To this end, c. 64 000 ESTs were sequenced from three normalized cDNA libraries constructed from various tissues of N. tabacum, and c. 20 000 cDNA clones were randomly selected and used in pairs to generate antisense or co-suppression transgenic tobacco plants. After transfer to the glasshouse, the transgenic plants were scored visually after 10–14 d for changes in growth, leaf form and the occurrence of chloroses or necroses.

One of the transformations led to a visual phenotype in several of the T0 lines, which included growth retardation and a reticulate leaf phenotype, similar to that seen in the Arabidopsis cue1 mutant (data not shown). As the initial set of plants was generated by transforming two constructs at the same time, we next had to establish which was responsible for the observed phenotype. To this end, both constructs were used individually to generate two independent new sets of transgenic tobacco plants. The phenotype could be reproduced after independent transformation of antisense construct no. 002212070. Three plants out of approximately 70 primary transformants harbouring the antisense construct clearly displayed the reticulate leaf phenotype (Fig. 1a) and were chosen for further analysis. The T1 generation of these three independent transgenic lines was used for further studies, as described below.

Figure 1.

 Basic characterization of enolase antisense tobacco (Nicotiana tabacum) plants. (a) Phenotype of a 6-wk-old transgenic plant from line Eno1. (b) Reverse transcriptase-polymerase chain reaction (RT-PCR) analysis of enolase expression in tobacco enolase antisense plants. Total RNA was isolated from leaves and used for RT-PCR analysis as described in Materials and Methods applying 35 PCR cycles. Ubiquitin served as an internal standard to ensure that equal amounts of cDNA were used for the reactions. Samples from two independent plants per line were analysed. (c) Enolase activity in leaves. Enzyme activity was determined in leaves from six individual plants per line (± SD). ME, transgenic control.

Identification of cytosolic enolase as the affected locus

BlastX analysis (Altschul et al., 1990) of the tobacco cDNA clone 002212070 revealed the presence of a partial open reading frame (ORF) ranging from position 246 to 473 of the nucleotide sequence. This ORF displayed 97% identity to the C-terminal part (amino acid 369 to the terminal amino acid 444) of tomato enolase (GenBank acc. no. CAA41115.1; Van der Straeten et al., 1991). The GenBank tobacco EST dataset was searched with the tobacco cDNA sequence to obtain overlapping EST clones. One EST clone contained the putative translational start site and was used to deduce oligonucleotides flanking the potential tobacco enolase coding region, together with the clone 002212070. These oligonucleotides were used to amplify a full-length enolase sequence by RT-PCR from tobacco source leaf cDNA. Subsequent cloning and sequence analysis of the PCR product revealed a contiguous ORF of 1335 bp, encoding a protein with a predicted molecular mass of 48 kDa which represents the cytosolic enolase isoform. The overall identity of the polypeptide to cytosolic enolase from tomato was 95%. By contrast, the deduced coding sequence of cytosolic tobacco enolase showed only c. 60% identity to plastidic enolases from other plant species, including a partial tobacco EST coding for a putative plastidic isoform.

Thus, we conclude that the cDNA clone 002212070 corresponds to a fragment of cytosolic tobacco enolase, which is very unlikely to target the putative plastidic isoform based on the low degree of identity. Accordingly, the three transgenic lines used in this study were designated Eno-1, Eno-2 and Eno-3, respectively.

The reticulate leaf phenotype correlates with residual enolase activity in Eno transgenics

Seeds of lines Eno-1, Eno-2 and Eno-3 were germinated on kanamycin-containing medium alongside ME-1, a transgenic control line containing the β-glucuronidase (GUS) gene driven by a cytosolic FBPase promoter from potato (Ebneth, 1996). Kanamycin-resistant plants were analysed 6–7 wk after transfer to the glasshouse. All Eno plants displayed the reticulate leaf phenotype, albeit to a different degree. Eno-1 plants were retarded in growth compared with the controls and uniformly showed dark-green paraveinal and light-green interveinal regions on all leaves (Fig. 1a). Eno-2 and Eno-3 plants grew at the same rate as control plants and displayed the reticulate phenotype only on young leaves up to c. 10 cm in length.

To monitor enolase transcript levels in leaves of Eno plants, RT-PCR with gene-specific primers was conducted using ubiquitin as an internal control. The result showed that the enolase mRNA steady-state level was most strongly reduced in line Eno-1, followed by lines Eno-3 and Eno-2, respectively (Fig. 1b). Foliar enolase activity was reduced to 5% (Eno-1), 13% (Eno-3) and 38% (Eno-2) of control plants, respectively (Fig. 1c), indicating that cytosolic enolase is the predominant, if not the only, enolase activity in tobacco leaves. For all analyses reported hereafter, we harvested samples from phenotypic leaves to investigate the impact of enolase deficiency on tobacco leaf metabolism.

Inhibition of enolase activity leads to reduced PEP levels and limits the accumulation of shikimate pathway products

Enolase catalyses the reversible interconversion of 2PGA and PEP. Thus, the reduction in enolase activity in the Eno antisense plants should be accompanied by an increase in the 2PGA : PEP ratio. Before performing any metabolite measurements, we determined whether the phenotypic alterations observed in Eno antisense plants would lead to changes in leaf mass per unit leaf area when compared with control plants. These measurements revealed no differences in specific leaf area fresh weight or specific leaf area dry weight between transgenic and control plants (data not shown).

In samples harvested during the light period, the 2PGA level increased by c. 50% in all three transgenic lines and the PEP content decreased by more than 70% in the enolase transgenics, compared with controls (Table 1). The 2PGA : PEP ratio in the two strongest Eno lines was increased by more than 6.5-fold (2PGA : PEP), indicating that the enolase reaction is displaced from equilibrium compared with the control. In samples taken during the dark period, the 2PGA level remained unaltered, but PEP decreased again, resulting in a 2.4–3.7-fold increase in the 2PGA : PEP ratio in the enolase antisense plants relative to controls (Table 2). PEP was further converted to pyruvate by PK. Pyruvate levels did not show any consistent change in the transgenic lines.

Table 1.   Metabolite levels in leaves of 6-wk-old Eno antisense transformants compared with control tobacco (Nicotiana tabacum) plants in the light
  1. Leaves were harvested 5 h into the photoperiod. Contents are given in μmol m−2, except for the soluble sugar and malate values, which are given in mmol m−2, and starch contents, which are given in mmol m−2 hexose equivalents. Values represent the mean (± SD) from five different plants per line. Significant differences between Eno transgenics and the wild-type were calculated using a Welch–Satterthwaite t-test (Junker et al., 2006; Klukas et al., 2006) and are indicated by *, < 0.05 and **, < 0.01.

  2. aKG, a-ketoglutarate; Fru6P, fructose-6-phosphate; Glc1P, glucose-1-phosphate; Glc6P, glucose-6-phosphate; PEP, phosphoenolpyruvate; 6PG, 6-phosphogluconate; PGA, phosphoglycerate; TCA, tricarboxylic acid cycle.

 Sucrose2.88 ± 0.391.44 ± 0.18**2.18 ± 0.221.75 ± 0.13*
 Glucose8.09 ± 0.941.88 ± 0.38**6.61 ± 1.114.46 ± 0.69*
 Fructose3.12 ± 0.141.02 ± 0.16*1.97 ± 0.21*1.81 ± 0.26*
 Starch56.8 ± 3.719.7 ± 2.0**30.2 ± 5.9**15.3 ± 1.1**
Phosphorylated intermediates
 3PGA283.9 ± 2.1210.0 ± 6.1**267.0 ± 5.3242.9 ± 9.6*
 2PGA8.4 ± 0.813.0 ± 1.2*12.9 ± 1.3*13.3 ± 2.3*
 PEP21.6 ± 1.75.1 ± 1.5**7.9 ± 1.3**4.7 ± 0.5**
 Pyruvate21.1 ± 4.926.0 ± 3.718.2 ± 1.922.2 ± 2.6
 Glc6P46.0 ± 5.2171.7 ± 19.9**114.8 ± 25.6*152.6 ± 17.9**
 Fru6P11.4 ± 3.737.1 ± 8.4*12.1 ± 2.351.6 ± 16.5
 Glc1P2.0 ± 1.218.2 ± 6.1*12.6 ± 4.74.0 ± 0.6
 6PG0.5 ± 0.011.5 ± 0.2*0.6 ± 0.041.5 ± 0.2*
TCC intermediates
 aKG25.0 ± 2.638.5 ± 3.5*36.9 ± 2.7*38.0 ± 2.5*
 Isocitrate16.9 ± 1.4189.2 ± 10.4**132.7 ± 39.6**270.6 ± 50.8**
 Citrate317.9 ± 22.1513.3 ± 47.3284.7 ± 22.8328.5 ± 49.4
 Malate0.98 ± 0.031.02 ± 0.03 1.09 ± 0.041.05 ± 0.05
 Succinate24.7 ± 0.85.6 ± 0.5**14.9 ± 1.1**5.5 ± 0.4**
 Fumarate17.2 ± 0.6107.8 ± 13.3**101.0 ± 14.9**148.8 ± 7.9**
Metabolite ratios
 2PGA : PEP0.42 ± 0.032.72 ± 0.58*1.79 ± 0.613.09 ± 0.58**
 PEP : pyruvate1.19 ± 0.210.16 ± 0.07*0.47 ± 0.11*0.23 ± 0.04*
 Glc6P : Fru6P2.93 ± 0.215.45 ± 1.218.52 ± 2.304.83 ± 1.45
 Citrate : isocitrate18.3 ± 5.51.06 ± 0.25**1.98 ± 0.79*0.78 ± 0.11**
 Isocitrate : aKG0.4 ± 0.13.5 ± 1.41.7 ± 0.87.6 ± 1.4*
 aKG : succinate1.03 ± 0.266.90 ± 0.35**2.46 ± 0.21**7.31 ± 1.04**
Table 2.   Metabolite levels in leaves of 6-wk-old Eno antisense transformants compared with control tobacco (Nicotiana tabacum) plants in the dark
  1. Leaves were harvested 5 h into the dark period. Contents are given in μmol m−2, except for malate, which is given in mmol m−2, and the values represent the mean (± SD) from five different plants per line. Significant differences between Eno transgenics and the wild-type were calculated using a Welch–Satterthwaite t-test (Junker et al., 2006; Klukas et al., 2006) and are indicated by *, < 0.05 and **, < 0.01.

  2. aKG, a-ketoglutarate; Fru6P, fructose-6-phosphate; Glc1P, glucose-1-phosphate; Glc6P, glucose-6-phosphate; PEP, phosphoenolpyruvate; 6PG, 6-phosphogluconate; PGA, phosphoglycerate; TCC, tricarboxylic acid cycle.

Phosphorylated intermediates
 3PGA49.9 ± 2.045.5 ± 4.454.2 ± 4.939.5 ± 1.5
 2PGA7.3 ± 2.26.9 ± 0.27.5 ± 0.88.3 ± 1.3
 PEP10.8 ± 1.76.4 ± 1.76.6 ± 1.36.3 ± 0.5
 Pyruvate11.7 ± 2.412.1 ± 0.615.0 ± 2.715.5 ± 1.4
 Glc6P60.7 ± 3.1149.6 ± 6.5**119.4 ± 11.5*135.2 ± 13.5**
 Fru6P12.4 ± 2.431.7 ± 2.8**23.6 ± 4.727.8 ± 4.3**
 Glc1P5.8 ± 0.89.6 ± 2.59.9 ± 2.611.7 ± 2.7
 6PG0.3 ± 0.030.56 ± 0.05*0.28 ± 0.030.33 ± 0.02
TCC intermediates
 aKG36.8 ± 3.433.9 ± 6.031.7 ± 2.738.7 ± 3.4
 Isocitrate16.9 ± 1.4275.1 ± 35.6**72.7 ± 20.7193.2 ± 48.3*
 Citrate290.7 ± 19.2375.8 ± 93.8229.2 ± 50.2189.5 ± 41.9
 Malate0.59 ± 0.020.56 ± 0.060.59 ± 0.190.53 ± 0.03
 Succinate8.7 ± 0.85.1 ± 1.14.8 ± 0.75.4 ± 0.7
 Fumarate7.9 ± 0.518.5 ± 3.7**24.6 ± 4.3**33.4 ± 2.1**
Metabolite ratios
 2PGA : PEP0.55 ± 0.112.01 ± 0.701.55 ± 0.31*1.30 ± 0.17**
 PEP : pyruvate1.15 ± 0.120.43 ± 0.120.46 ± 0.090.42 ± 0.05
 Glc6P : Fru6P3.97 ± 0.354.56 ± 0.534.04 ± 0.375.07 ± 0.50
 Citrate : isocitrate11.3 ± 2.61.13 ± 0.16**8.49 ± 2.751.59 ± 0.67**
 Isocitrate : aKG0.5 ± 0.19.5 ± 2.9*2.3 ± 0.75.2 ± 1.4*
 aKG : succinate1.50 ± 0.200.38 ± 0.08**1.16 ± 0.122.65 ± 0.29**

As the initial steps of the shikimate pathway are exclusively localized in the plastid stroma (Herrmann & Weaver, 1999), we next investigated whether the reduced PEP content in the transgenic plants affected the synthesis of the aromatic amino acids phenylalanine and tyrosine, that is, the primary products of the shikimate pathway. The phenylalanine content was considerably reduced by more than 80% in the two strongest Eno lines, Eno-1 and Eno-3, whereas the phenylalanine contents in Eno-2 plants were reduced by c. 30% (Table 3). Tyrosine levels were generally relatively low, and there was no significant difference between transgenic and control plants.

Table 3.   Amino acid contents in leaves of 6-wk-old Eno antisense transformants compared with control tobacco (Nicotiana tabacum) plants
  1. Leaves were harvested 5 h into the photoperiod. Concentrations are given in μmol m−2. Values represent the mean (± SD) from four different plants per line. Significant differences between Eno transgenics and the wild-type were calculated using a Welch–Satterthwaite t-test (Junker et al., 2006; Klukas et al., 2006) and are indicated by *, < 0.05 and **, < 0.01.

  2. aa, amino acid.

Asp37.9 ± 4.8108.8 ± 13.9**69.6 ± 7.2**72.6 ± 7.9**
Glu201.6 ± 22.6139.1 ± 14.8244.3 ± 15.2206.0 ± 17.1
Asn8.2 ± 1.6106.2 ± 18.2**14.7 ± 1.6*23.3 ± 7.0
Gln20.9 ± 2.2455.4 ± 94.2*37.3 ± 6.279.9 ± 23.9
Ala27.8 ± 4.244.0 ± 6.8*48.7 ± 3.9**52.9 ± 5.7*
Ser37.0 ± 6.7102.4 ± 9.5**65.2 ± 7.8*120.4 ± 17.5**
Gly12.5 ± 1.882.9 ± 9.5**30.6 ± 5.5*38.4 ± 7.5*
Arg50.9 ± 6.037.3 ± 4.234.2 ± 2.042.9 ± 9.8
Pro23.6 ± 2.623.1 ± 3.230.3 ± 3.213.0 ± 1.2**
His4.9 ± 0.76.1 ± 1.15.1 ± 0.63.8 ± 0.4
Thr8.8 ± 2.828.3 ± 3.4*17.6 ± 2.324.5 ± 2.9*
Lys5.3 ± 0.64.7 ± 0.89.7 ± 0.8**4.2 ± 0.4
Val9.6 ± 0.67.4 ± 0.6*10.3 ± 1.08.8 ± 0.9
Ile8.6 ± 1.63.6 ± 0.6*4.5 ± 0.7*4.4 ± 0.9**
Leu8.0 ± 0.88.4 ± 0.77.6 ± 0.63.8 ± 0.3**
Phe12.2 ± 1.01.6 ± 0.2**8.7 ± 1.0*2.1 ± 0.4**
Tyr1.8 ± 0.71.4 ± 0.12.9 ± 0.30.9 ± 0.1
Total free aa443 ± 61886 ± 195*711 ± 63**756 ± 100**
Protein (mg m−2)1.47 ± 0.461.24 ± 0.271.31 ± 0.361.56 ± 0.43

Phenylalanine can be further metabolized into a large number of different phenolic compounds. To determine the effect of reduced phenylalanine content on secondary metabolism, the levels of total UV-absorbing substances, for example, representing phenylpropanoids and flavonoids, were used as a proxy. As shown in Fig. 2, the amount of UV-absorbing substances was decreased by c. 50% in Eno-1 and Eno-3 plants, whereas Eno-2 plants showed only a reduction of c. 20% compared with controls.

Figure 2.

 Amount of total UV-absorbing substances in enolase antisense tobacco (Nicotiana tabacum) plants. Leaf samples of the defined area were taken from 6-wk-old plants and analysed as described in Materials and Methods. ME, transgenic control.

Taken together, these data indicate that a restriction of PEP generation within the cytosol limits the flux through the shikimate pathway in the chloroplast, eventually decreasing the amount of phenylalanine-derived secondary compounds in the transgenic tobacco plants.

It has been described that PEP is capable of fuelling the biosynthesis of branched chain amino acids in isolated chloroplasts (Schulze-Siebert et al., 1984). Pyruvate acts as a building block for valine and isoleucine. We assessed the contents of this subclass of amino acids, and found that the contents of isoleucine and valine were significantly reduced in at least one of the Eno transgenics (Table 3). This may indicate a shortage of PEP-derived pyruvate inside the stroma of the transgenics.

The balance of central carbohydrate and amino acid metabolism and the TCA cycle is altered in Eno transgenics

We next examined whether the restricted conversion of 2PGA into PEP in the leaves of enolase antisense plants led to alterations in central metabolic routes. To this end, leaf samples harvested during the light period, as well as samples harvested in the dark, were analysed. Profound changes were observed within the pools of carboxylates and hexose-phosphates (hexose-P).

In the light, we observed a strong increase in glucose-6-phosphate (Glc6P), fructose-6-phosphate (Frc6P) and glucose-1-phosphate (Glc1P) in all transgenics, which amounted to four- to nine-fold for the strongest enolase antisense line Eno-1. By contrast, starch, sucrose, glucose and fructose contents were decreased in the transgenics, up to three-fold in Eno-1, compared with controls (Table 1). This indicates that the net synthesis of carbohydrates during photosynthesis is inhibited, even though hexose-P precursors accumulate. In the dark, hexose phosphates were also elevated in the antisense plants with respect to the control, suggesting that carbohydrate mobilization is stimulated in the transgenics.

Catabolic flux via glycolysis feeds carbon skeletons into the TCA cycle for subsequent respiration. In order to investigate whether the repression of enolase activity also affected the TCA cycle, the contents of TCA cycle intermediates were determined. During the light period, the a-ketoglutarate (aKG), citrate and malate content were largely unaffected in transformants when compared with the control specimen (Table 1). By contrast, isocitrate and fumarate levels were increased between six- and 16-fold in the transgenics, and the succinate content was decreased up to four-fold compared with controls. This picture essentially remained the same in the dark, although the changes in fumarate and succinate observed during the light were much less pronounced and absent, respectively (Table 2).

Both the abundance of sugars and the availability of carbon skeletons derived from the TCA cycle control nitrogen assimilation into amino acids. Surprisingly, total free amino acid content was increased by two-fold in enolase antisense plants compared with control plants, and leaf protein content was unaltered in the transgenics (Table 3). Asparagine and glutamine were increased by up to 14- and 22-fold in Eno-1 transgenics, indicating an imbalance between the foliar storage of free sugars and organic nitrogen. Likewise, serine and glycine were substantially increased compared with controls, which may indicate increased flux through the photorespiratory C2 cycle.

Impact on the activities of other enzymes

PEP and pyruvate represent central metabolites and serve as important building blocks in various biosynthetic routes in plants. To assess whether the imbalance between 2PGA and PEP in the Eno transgenics led to changes in other enzyme activities that metabolize PEP and pyruvate, we determined the activities of PK, NADP-malic enzyme (ME), NADP-isocitrate dehydrogenase, PEPC, PEPCK and PPDK in crude extracts of Eno transgenics and control leaves (Table 4). As controls, the activities of the Calvin cycle enzymes NADP-GAPDH and FBPase were analysed. Although the two Calvin cycle activities and most of the examined enzyme activities remained unaltered between Eno and control plants, the activities of plastidic PK, which could be differentiated from the cytosolic activity by its different pH optimum (pH 6.9 for the cytosolic enzyme and pH 8.1 for the plastidic enzyme) and PEPCK were significantly increased in Eno-1 leaves.

Table 4.   Activities of selected enzymes of carbohydrate metabolism measured in leaf samples from transgenic enolase antisense tobacco (Nicotiana tabacum) plants
  1. Leaf samples were taken 5 h into the light period, and enzyme activities were determined from crude extracts. Activities are given as U m−2. Values represent the mean (± SD) from five measurements. Significant differences between Eno transgenics and the wild-type were calculated using a Welch–Satterthwaite t-test (Junker et al., 2006; Klukas et al., 2006) and are indicated by *, < 0.05 and **, < 0.01.

  2. cyt, cytosolic; FBPase, fructose-1,6-bisphosphatase; Hxk, hexokinase; NADP-GAPDH, NADP glyceraldehyde-3-phosphate dehydrogenase; NADP-IcDH, NADP isocitrate dehydrogenase; NADP-ME, NADP malic enzyme; PEPC, phosphoenolpyruvate carboxylase; PEPCK, phosphoenolpyruvate carboxykinase; PK, pyruvate kinase; pl, plastidic; PPDK, pyruvate,orthophosphate dikinase.

NADP-GAPDH228 ± 22274 ± 26218 ± 18231 ± 17
FBPase pl230 ± 13227 ± 21249 ± 18230 ± 13
PK pl43.5 ± 15.281.3 ± 13.3*77.7 ± 20.674.0 ± 15.3
PK cyt58.2 ± 21.951.3 ± 10.656.3 ± 8.954.8 ± 18.7
NADP-ME13.8 ± 2.610.3 ± 3.38.5 ± 1.68.1 ± 1.7
NADP-IcDH49.6 ± 5.459.6 ± 5.748.9 ± 3.055.0 ± 2.7
PEPC15.3 ± 2.617.0 ± 2.114.5 ± 0.814.2 ± 1.4
PEPCK35.9 ± 3.259.6 ± 4.2*36.6 ± 1.652.8 ± 7.0
PPDK22.8 ± 1.227.6 ± 2.224.8 ± 0.727.7 ± 1.8
Hxk3.9 ± 0.79.5 ± 1.8*8.1 ± 0.4**9.4 ± 0.4**

Photosynthetic performance of gas exchange of Eno transgenic lines

To reveal whether reduced PEP availability in enolase antisense plants limited respiration, and to investigate whether the observed carbohydrate shortage for nitrogen assimilation could be explained by diminished photosynthetic capacity, we determined foliar gas exchange rates at an ambient photon flux density of 200 μmol quanta m−2 s−1 in combination with chlorophyll fluorescence imaging and recorded dark respiration rates.

Only in Eno-1 were the CO2 assimilation rate, electron transport rate and reduction state of the plastoquinone pool significantly reduced compared with controls, which was paralleled by increased nonregulated energy dissipation (Table 5). Photochemistry was even more impaired in the chlorotic intercostal fields of Eno-1 when compared with the overall average in Eno-1 leaves. Measurements at a saturated photon flux density of 1200 μmol quanta m−2 s−1 resulted in similar results. Surprisingly, respiration rates in the dark were comparable between controls and Eno transgenics, irrespective of whether they were recorded at the beginning (Table 5) or in the middle of the dark period (not shown), indicating that gross CO2 production was not altered.

Table 5.   Photosynthetic performance of young mature source leaves of enolase antisense tobacco (Nicotiana tabacum) plants
LineA (μmol m−2 s−1)Rd (μmol m−2 s−1)ETR (μmol m−2 s−1)%QA redNPQY (NO)
  1. Measurements were conducted at a photon flux density of 200 μmol quanta m−2 s−1, with gas exchange rates and chlorophyll fluorescence imaging data recorded simultaneously. For chlorotic Eno-1 intercostal fields, chlorophyll fluorescence was calculated separately (see line Eno-1 meso). Please note that gas exchange data cannot be spatially differentiated. Values represent the mean (± SD) of four measurements. Significant differences between Eno transgenics and the wild-type were calculated using a Welch–Satterthwaite t-test (Junker et al., 2006; Klukas et al., 2006) and are indicated by *, < 0.05 and **, < 0.01.

  2. A, CO2 assimilation rate; Eno-1 meso, chlorotic mesophyll area in Eno-1 leaves; ETR, electron transport rate; nd, not determined; NPQ, nonphotochemical quenching; %QA red, % reduced free plastoquinone; Rd, apparent dark respiration rate; Y (NO), nonregulated energy dissipation.

Control3.80 ± 0.130.28 ± 0.0451.1 ± 0.287.4 ± 0.60.404 ± 0.0270.230 ± 0.004
Eno-12.92 ± 0.210.39 ± 0.1045.9 ± 2.085.6 ± 0.90.443 ± 0.0590.248 ± 0.004*
Eno-1 mesondnd42.1 ± 1.1*84.3 ± 0.40.465 ± 0.0670.262 ± 0.003**
Eno-23.62 ± 0.110.24 ± 0.0450.7 ± 0.987.4 ± 1.00.400 ± 0.0040.230 ± 0.005
Eno-33.71 ± 0.300.25 ± 0.0750.3 ± 0.688.3 ± 1.20.400 ± 0.0230.232 ± 0.004


Foliar PEP production is decreased on transgenic suppression of cytosolic enolase

In this study, we have presented a detailed physiological analysis of transgenic tobacco plants with reduced enolase activity, which were generated during the course of a large-scale effort to identify essential leaf function in tobacco (Lein et al., 2008). Plants contain plastidic and cytosolic enolases, which share less than 60% identity. Although there is experimental evidence that some nongreen plastids contain enolase activity (Plaxton, 1996), there is no experimental evidence for substantial enolase activity in pea and spinach chloroplasts (Stitt & ap Rees, 1978, 1980; Bagge & Larsson, 1986). Furthermore, the Arabidopsis genome contains a single gene encoding a plastid-targeted enolase (Prabhakar et al., 2009). Expression of the Arabidopsis plastidial enolase was detected in heterotrophic tissues, including trichomes and nonroot hair cells, but not in the mesophyll of leaves. This led to the conclusion that plastids in the former tissues contain a complete set of glycolytic enzymes, whereas, in chloroplasts, the conversion of 3PGA to PEP is blocked as a result of the absence of plastidial enolase activity (Prabhakar et al., 2009). Furthermore, Arabidopsis T-DNA insertion mutants lacking plastidial enolase expression showed no reduction in enolase activity in leaves compared with the wild-type (Prabhakar et al., 2009). Thus, it appears that plastidial and cytosolic enolase, respectively, have specific functions in metabolism and development in Arabidopsis.

We used an antisense construct of a cloned tobacco enolase gene to obtain a progressive decrease in enolase activity. The tobacco enolase sequence showed strong similarity to the identified cDNA of cytosolic enolases from other species, and lacked a transit peptide. We therefore conclude that the cloned tobacco enolase is also localized in the cytosol. The fact that up to 95% of the total enolase activity could be repressed whilst a residual transcript amount of the cytosolic target gene was still detectable indicates that enolase activity is most probably absent from tobacco chloroplasts as well. This presumption is also reinforced by our observation that the 2PGA : PEP ratio is massively displaced from equilibrium in the enolase transgenics, whereas it is close to the equilibrium constant in the controls (Tables 1 and 2). Although we currently cannot exclude a subtle contribution of the plastidial enzyme to the overall enolase activity, it appears that PEP generation by cytosolic enolase in tobacco leaves cannot be compensated by plastidial enolase (similar to Arabidopsis, the tobacco gene is most probably not expressed in green tissues) or by any other reaction.

PEP produced by the enolase-catalysed reaction occupies a central position in plant metabolism. As outlined in the Introduction, PEP can drive mitochondrial respiration after conversion to pyruvate by PKc, be diverted into the carboxylate pool to support anaplerotic provision of carbon skeletons for nitrogen assimilation and other metabolic routes by the TCA cycle, or be transported into the plastid stroma by PPT to serve as a precursor for the biosynthesis of aromatic amino acid and phenolics via the shikimate pathway, as well as for the biosynthesis of branched chain amino acids, fatty acids or isoprenoids after its conversion to pyruvate by plastidic PK. The 2PGA : PEP ratio is close to equilibrium in wild-type plants (2.5; Table 1), and markedly displaced in Eno transgenic lines (0.35–0.6; Table 1). Enolase catalyses a readily reversible reaction in vivo, in which the net flux (v) is equal to the difference between the forward (v+1) and reverse (v−1) fluxes. From the metabolite data alone, it is not possible to decide whether the flux from 2PGA to PEP has decreased, or whether the increase in the substrate relative to the product (which will increase v+1 relative to v−1) allows net flux to be maintained. Two scenarios are therefore possible to explain the changes in metabolism in Eno-1 lines: they are a result of decreased flux from 2PGA to PEP, or they are caused by the decrease in PEP, which kinetically limits other enzymes or acts as a regulatory ligand. The fact that enolase cannot be replaced by any other reaction makes it an especially interesting and powerful way to investigate both of these possibilities.

Repression of cytosolic enolase activity limits the plastidic shikimate pathway

The decreased level of PEP in the Eno transgenics leads to a strong reduction in the content of the primary shikimate pathway products phenylalanine and phenylpropanoids, as well as downstream derivatives (Table 4, Fig. 2), similar to that observed in the Arabidopsis cue1 mutant (Streatfield et al., 1999; Voll et al., 2003). In both the cue1 mutants and the Eno transgenics, tyrosine contents remained low and unaltered, indicating that the availability of the substrate phenylalanine becomes limiting for the flux into phenylpropanoids via phenylalanine ammonia lyase in both species, as generally suggested (Da Cunha, 1987). Interestingly, the inhibition of transketolase which supplies 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase, the committed step of the shikimate pathway, with the complementary precursor erythrose-4-phosphate also leads to severely decreased levels of phenylalanine and phenylpropanoids in tobacco (Henkes et al., 2001), strongly indicating that the entire shikimate pathway can be substrate limited when the availability of the precursors PEP or erythrose-4-phosphate decreases below a critical threshold.

Although the abundance of most other free amino acids increases compared with controls, the contents of the plastid-synthesized branched chain amino acids, such as isoleucine and valine, are decreased in Eno antisense plants (Table 4). This may also be related to limited PEP availability. Branched chain amino acids are synthesized from pyruvate in the plastid, and pyruvate is probably synthesized from PEP by plastidic PK.

Taken together, our data clearly demonstrate that PEP availability strongly impacts on amino acid biosynthesis in the chloroplasts. This is presumably a result of the restriction of the rate of PEP uptake by PPT. Transport of metabolites between subcellular compartments is often one of the limiting steps in metabolic pathways (Riesmeier et al., 1993; Tjaden et al., 1998). Furthermore, the Km(PEP) value of PPT is about six times lower than the Km(PEP) value of PKc (300 μmvs 50 μm; Fischer et al., 1997; Plaxton & Podestá, 2006), thus favouring PEP conversion to pyruvate.

Pyruvate provision for respiration is independent of enolase in the transgenics

The level of pyruvate and the rate of respiration are similar in Eno transgenics and control plants. Further, the Eno transgenics contained higher levels of many organic acids, especially isocitrate and fumarate. The supply of substrates to the mitochondria via the reactions catalysed by PK and PEPC is obviously not impeded by a decrease in enolase activity. It implies that plants are able to bypass enolase, or are able in some other way to compensate for reduced enolase. This is reminiscent of the results obtained by Gottlob-McHugh et al. (1992), who showed that antisense reduction of PK in transgenic tobacco plants had no major effect on pyruvate levels and respiratory metabolism in leaves, indicating that plants are also able to compensate for a shortfall in pyruvate formation.

In plants, glycolysis is feedback inhibited through PEP inhibition of phosphofructokinase (PFK) and 3PGA-mediated inhibition of PFP (Plaxton & Podestá, 2006 and references therein). The lower PEP levels in the Eno transgenic lines would be expected to lead to allosteric activation of PFK and the stimulation of glycolysis. This should lead to an increase in 2PGA and at least partially compensate for the decreased activity of enolase. There was an increase in 2PGA in the Eno transgenic lines in the light, when it is derived from the Calvin cycle, but not in the dark. This implies that the decrease in PEP does not strongly stimulate glycolysis. Unexpectedly, the Eno transgenic lines showed increased levels of hexose-P pools in both the light and dark (Tables 1 and 2), and significantly increased hexokinase activities (Table 3). This increase in hexose-P could well be explained by a combination of increased hexokinase activity and reduced downstream metabolism. It is unclear whether PEP itself or some other change in the Eno transgenic lines is responsible for the increase in HK activity and hexose-P levels.

Even if glycolysis is activated in the Eno transgenics, this does not prevent a decrease in PEP, and therefore cannot on its own explain how pyruvate levels are maintained (Tables 1 and 2). More generally, our data imply that the flux to pyruvate, organic acids and respiration is not controlled by PEP itself but, rather, by events downstream of PEP. This implies that plant cells are able to regulate carbon utilization via respiration, independent from the control of carbon entry into glycolysis. This also involves preferential allocation of PEP to respiration rather than (see above) biosynthesis in the plastid.

Several possible explanations can be offered as to why PK and PEPC are able to effectively compete for PEP. One is that these are highly regulated enzymes, and the relaxation of feedback regulation may allow their Km(PEP) values to be decreased to enable flux to be maintained at the lower PEP levels found in the Eno transgenic lines. Intriguingly, it has also been shown recently that enzymes of the cytosolic glycolytic pathway (including enolase and PK) are physically associated with the outer mitochondrial membrane (Giegéet al., 2003), and their degree of association is dynamically responsive to the respiratory demand of the mitochondria (Graham et al., 2007). This microcompartmentation of glycolysis might allow substrate channelling between sequential enzymes within the pathway, which then restricts the use of intermediates by competing metabolic pathways, for example, the demand of PEP for the synthesis of aromatic amino acids in the chloroplast. This could potentially explain why the pyruvate level and respiration rate in enolase antisense tobacco plants remain unaffected, but the supply of PEP to biosynthetic reactions in the plastid is restricted.

As the steady-state level of a given metabolite reflects the balance between its synthesis and degradation, the regulation of enzyme activities lying further downstream might play a role in determining the pyruvate level in enolase antisense plants. The mitochondrial pyruvate decarboxylase complex (mtPDC) links glycolytic carbon metabolism with the TCA cycle by catalysing the oxidative decarboxylation of pyruvate to acetyl-CoA. Evidence suggests that there is significant potential for the regulation of mtPDC through product feedback by NADH and acetyl-CoA (Randall et al., 1989). It has been shown for several plant species that, in the light, mtPDC activity rapidly decreases in leaves and this down-regulation has been linked to photorespiratory glycine oxidation, which also generates NADH (Gemel & Randall, 1992). Given the increase in serine and glycine in the Eno plants, it appears likely that there is an increased flux through the photorespiratory C2 cycle. In turn, this could lead to increased NADH levels within the mitochondrion, further decreasing mtPDC activity and thus buffering pyruvate levels during the photoperiod. However, based on the currently available data, it appears most likely that the residual enolase activity in transgenic Eno tobacco plants is sufficient to generate the pyruvate that is needed for respiration.


Taken together, the reported work elucidates two fundamental aspects of glycolysis. First, we have observed that the restriction of glycolysis hampers PEP supply for the plastidic shikimate pathway, but not pyruvate provision for mitochondrial respiration. As glycolysis is crucial for the production of both PEP and pyruvate, it was surprising that PEP contents substantially decreased on enolase limitation, whereas pyruvate contents remained unaffected. This implies that cellular pyruvate production underlies certain homeostasis, whereas PEP does not. Second, changes in the level of PEP may play a major role in regulating and coordinating glycolysis and respiratory metabolism.


The authors gratefully acknowledge excellent technical assistance in the enzyme and metabolite assays by Alexandra Saur and Hildegard Voll (Chair of Biochemistry, FAU Erlangen-Nuremberg), as well as the gardener team at the IPK Gatersleben and Christine Hösl at the FAU Erlangen-Nuremberg for assistance in the glasshouse.