Expression of a bacterial feedback-insensitive 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase of the shikimate pathway in Arabidopsis elucidates potential metabolic bottlenecks between primary and secondary metabolism


Author for correspondence:
Gad Galili
Tel: +972 8 9343511


  • The shikimate pathway of plants mediates the conversion of primary carbon metabolites via chorismate into the three aromatic amino acids and to numerous secondary metabolites derived from them. However, the regulation of the shikimate pathway is still far from being understood. We hypothesized that 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase (DAHPS) is a key enzyme regulating flux through the shikimate pathway.
  • To test this hypothesis, we expressed a mutant bacterial AroG gene encoding a feedback-insensitive DAHPS in transgenic Arabidopsis plants. The plants were subjected to detailed analysis of primary metabolism, using GC-MS, as well as secondary metabolism, using LC-MS.
  • Our results exposed a major effect of bacterial AroG expression on the levels of shikimate intermediate metabolites, phenylalanine, tryptophan and broad classes of secondary metabolite, such as phenylpropanoids, glucosinolates, auxin and other hormone conjugates.
  • We propose that DAHPS is a key regulatory enzyme of the shikimate pathway. Moreover, our results shed light on additional potential metabolic bottlenecks bridging plant primary and secondary metabolism.


The shikimate pathway is ubiquitous in multiple organisms, including plants, converting primary carbon metabolites via shikimate into chorismate (Fig. 1; for reviews, see Herrmann & Weaver, 1999; Tzin & Galili, 2010b). Chorismate serves mainly as a common precursor for the synthesis of the three aromatic amino acids (AAAs) phenylalanine (Phe), tyrosine (Tyr) and tryptophan (Trp) (Bentley, 1990). In plants, the shikimate and AAA biosynthesis pathways provide the substrates for multiple secondary metabolites, such as alkaloids, flavonoids, lignin, coumarins, indole derivatives and other phenolic compounds (Gilchrist & Kosuge, 1980), rendering the shikimate pathway a bridge between primary and secondary metabolism. Although the regulation of the synthesis of AAAs from chorismate, the end product of the shikimate pathway, has been studied extensively in plants (Cho et al., 2007; Yamada et al., 2008; Tzin et al., 2009; Maeda et al., 2010), the regulation of flux through the shikimate pathway itself in plants is much less well understood. The most studied enzyme of the shikimate pathway is enolpyruvylshikimate 3-phosphate synthase (EPSPS), catalyzing the formation of enolpyruvylshikimate 3-phosphate (Fig. 1), because of its association with resistance to the herbicide glyphosate, which is the basis for the Roundup-Ready transgenic crops (Singer & McDaniel, 1985; Smart et al., 1985; Klee et al., 1987; Duke & Powles, 2008; Vivancos et al., 2011).

Figure 1.

Schematic diagram of the shikimate and aromatic amino acid (AAA) metabolic networks in plants. Arrows represent a single enzymatic step and dashed arrows represent several enzymatic steps. AS, anthranilate synthase; CM, chorismate mutase; CS, chorismate synthase; DAHPS, 3-deoxy-d-arabino-2-heptulosonate 7-phosphate synthase; DHQS, 3-dehydroquinate synthase; DHQ/SDH, 3-dehydroquinate dehydratase/shikimate 5-dehydrogenase; EPSPS, 5-enolpyruvylshikimate 3-phosphate synthase; SK, shikimate kinase.

The first committed enzyme of the shikimate pathway is 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase (DAHPS), which converts phosphoenolpyruvate (PEP) and erythrose 4-phosphate (E-4P) into 3-deoxy-d-arabino-heptulosonate 7-phosphate (DAHP; Fig. 1; Herrmann, 1995). Arabidopsis thaliana plants possess two DAHPS genes, DHS1 (AT4G39980) and DHS2 (AT4G33510), as well as an additional putative gene (AT1G22410) with high similarity to DHS1. The presence of amino-terminal extensions characteristic of chloroplast transit peptides in the Arabidopsis proteins encoded by DHS1 and DHS2 supports the notion that both proteins are localized in the chloroplast (Keith et al., 1991). Previous studies have shown that these genes are differentially expressed in tomato (Solanum esculentum) (Gorlach et al., 1993), and DHS1 is highly induced by physical wounding or by infiltration with pathogen in both Arabidopsis (Keith et al., 1991) and tomato (Gorlach et al., 1993). Nevertheless, despite the available information on DAHPS expression and activity in plants, it is still unknown whether DAHPS serves as a major regulator of flux through the shikimate pathway, and hence whether it is a key enzyme bridging between plant primary and secondary metabolism. As opposed to plants, extensive studies have been performed on microbial DAHPS enzymes, and have described DAHPS as a key enzyme controlling the shikimate pathway (Wu & Woodard, 2006).

Escherichia coli possesses three different DAHPS isoforms, encoded by the AroF, AroG and AroH genes, which are feedback inhibited by the individual AAAs Tyr, Phe and Trp, respectively (Brown, 1968). The major isoform is AroG, which makes up 80% of the total DAHPS activity, and the less abundant isoenzymes, AroH and AroF, make up 1% and 19% of the total bacterial DAHPS activity, respectively (Wallace & Pittard, 1967; Jensen & Nasser, 1968; McCandliss et al., 1978; Zurawski et al., 1981; Hu et al., 2003). In contrast with E. coli and many other bacterial species, the allosteric regulation of plant DAHPS is still questionable (Gilchrist & Kosuge, 1980; Herrmann & Weaver, 1999). Yet, in vitro activities of DAHPS from different plant species are weakly inhibited by Trp (Graziana & Boudet, 1980; Rubin & Jensen, 1985) and Tyr (Reinink & Borstap, 1982), or weakly enhanced by either Trp or Tyr (Suzich et al., 1984; Pinto et al., 1986). In addition, the Vignaradiata (bean) DAHPS activity is weakly inhibited by prephenate and arogenate, the precursors of Phe and Tyr biosynthesis (Rubin & Jensen, 1985), but whether this is caused by the inhibition level or activity of the enzyme is still unknown (Herrmann, 1995; Tzin & Galili, 2010b). In this research, we overexpressed in Arabidopsis plants, under the CaMV 35S constitutive promoter, two forms of chimeric genes: one encoding a plastid-targeted bacterial DAHPS (35S-AroGWT) and the other encoding a plastid-targeted mutant bacterial DAHPS possessing a single point mutation converting leucine (Leu) at position 175 into glutamine (Gln) (35S-AroG175). This point mutation renders the bacterial enzyme feedback insensitive to Phe regulation (Hu et al., 2003). Our results show that, in Arabidopsis plants, DAHPS represents a rate-limiting factor in the flux of the shikimate pathway, and additional potential metabolic bottlenecks bridging plant primary and secondary metabolism are also identified.

Materials and Methods

Plasmid construction and plant transformation

The coding DNA sequence of the E. coli AroG gene, encoding DAHPS, was amplified by PCR with the following oligonucleotides: forward 5’-CATGCATGCTGATGAATTATCAGAACGACGA-3’, which introduces an SphI restriction site (italic), and reverse 5’-GGAATTCCCCGCGACGCGCTTTTACTG-3’, which introduces an EcoRI restriction site (italic). Two varieties of recombinant gene were constructed: AroGWT (original sequence) and the feedback-insensitive point mutation at 524 bp, AroG175 (Leu175Gln, italic), using PCR with the following oligonucleotides: forward 5’-GTGCACCGCGAACAGGCATCAGGGCTT-3 and reverse 5’-AAGCCCTGATGCCTGTTCGCGGTGCAC-3’. The RuBisCO small subunit-3A plastid transit peptide (Shaul & Galili, 1993) was fused in frame to the 5’ end of the AroG open reading frame. The AroG 3’ was fused to three copies of a hemagglutinin (HA) epitope tag fused to an octopine synthase (OCS) terminator. The entire fragment was subcloned into the Ti plasmid pBART, a derivative of pART27 (Gleave, 1992). For sequence alignment, the National Center for Biotechnology Information (NCBI) database was used ( Agrobacterium tumefaciens strain EHA-105 was transformed with the plasmids by electrophoresis using a gene pulser apparatus (Bio-Rad). Wild-type (WT) Arabidopsis thaliana (L.) Heynh., ecotype Colombia, plants were transformed as described previously (Clough & Bent, 1998).

Plant material and growth conditions

Seeds were imbibed for 48 h at 4°C and transferred to a climate-controlled growth room with a regime of 16 h of light (c. 100 μmol m-2s-1) and 8 h of dark (long-day conditions) at 22°C. Plant selection was performed by spraying the leaves with 100 μg ml−1 Basta (Glufosinate ammonium; Bayer CropScience, T2 generation plants were measured for progeny test, and lines with a single gene insertion were selected on the basis of 3 : 1 genetic segregation. Resistance to the 5MT Trp analog (Sigma/Aldrich) was performed as described previously (Li & Last, 1996; Tzin et al., 2009).

Immunoblot analysis

Immunoblots were performed as described previously (Stepansky & Galili, 2003) using monoclonal anti-HA antibodies (Sigma-Aldrich). Equal amounts of protein were loaded on each lane of the gel.

Ultra-performance liquid chromatography/quadrupole time of flight-mass spectrometry (UPLC/qTOF-MS) sample extraction, data profiling and statistical analysis

Metabolic analysis was performed on aerial tissues of 10-d-old Arabidopsis seedlings (100 mg frozen powder) expressing the AroGWT, AroG175 and control lines (= 5–6). UPLC/qTOF-MS samples were extracted in 450 μl of methanol 80%.

The mixture was sonicated for 20 min at room temperature, centrifuged (3000 g, 10 min) and filtered through a 0.22-μm polytetrafluoroethylene membrane filter (Acrodisc CR 13 mm; PALL, before injection into the UPLC/qTOF-MS instrument. Mass spectral analyses were carried out by the UPLC/qTOF instrument (Waters Premier QTOF,, with the UPLC column connected online to a UV detector (measuring at 240 nm; Waters Acquity) and then to the MS detector. The separation of metabolites was performed on a 100 mm × 2.1 mm i.d., 1.7-μm UPLC BEH C18 column (Waters Acquity; Mintz-Oron et al., 2008). Injections of samples in the negative and positive ionization modes were performed in separate injection sets, and preprocessing was carried out for each ionization mode independently. The analysis of the raw UPLC/qTOF-MS data was carried out using the XCMS software, which performs chromatogram alignment, mass signal detection and peak integration (Smith et al., 2006), from the Bioconductor package (v. 2.1) for the R statistical language (v. 2.6.1). XCMS was used with the following parameters: fwhm = 10.8, step = 0.05, steps = 4, mzdiff = 0.07, snthresh = 8, max = 1000. Metabolites were identified using standard compounds by comparison of their retention times, UV spectra, MS/MS fragments and dual-energy fragments. The identification of metabolites for which standards were not available was carried out as for our previously reported compounds (Malitsky et al., 2008; Tzin et al., 2009). To assess whether the different lines in the analysis vary in their composition of metabolites, Student’s t-test (JMP software, was performed. The resulting P values were controlled for multiple hypotheses testing using a 5% false discovery rate (FDR) cut-off (Hochberg & Benjamini, 1990; see Supporting Information Table S2 for details regarding putative identification). The principal component analysis (PCA) plots were generated using TMEV4 software (Saeed et al., 2003; Scholz et al., 2004).

GC-MS profiling of derivatized extracts and statistical analysis

Metabolic analysis was performed with aerial tissues of 10- or 20-d-old Arabidopsis seedlings (100 mg frozen powder). Plants, expressing the AroGWT lines, AroG175 lines and control (n = 5–6), were harvested, frozen, ground, extracted and analyzed as described previously (Malitsky et al., 2008). To assess whether the different lines in the analysis vary in their composition of metabolites, Student’s t-test (JMP software) was performed. The resulting P values were controlled for multiple hypotheses testing using a 5% FDR cut-off (Hochberg & Benjamini, 1990; see Table S1 for details regarding putative identification). The PCA plots were generated using TMEV4 software (Saeed et al., 2003; Scholz et al., 2004).

Determination of lignin composition

Thioacidolysis was performed as described previously (Lapierre et al., 1985; Chen et al., 2006), using 15 mg of extract-free samples reacted with 3 ml of 0.2 M BF3 etherate in an 8.75 : 1 dioxane–ethanethiol mixture. Lignin-derived monomers were identified by GC-MS and quantified by GC as their trimethylsilyl derivatives (Hewlett Packard 5890 series II GC with a 5971 series mass-selective detector; column: HP-1, 60 m × 0.25 mm × 0.25 μm film thickness). Mass spectra were recorded in electron impact mode (70 eV) with a scanning range of 60–650 m/z (Lapierre et al., 1985; Chen et al., 2006).


Generation and characterization of transgenic Arabidopsis plants expressing a bacterial DAHPS enzyme

To study the importance of DAHPS in regulating the fluxes bridging primary and secondary metabolism in plants, we generated two chimeric genes under the control of the 35S promoter and the OCS terminator (Fig. S1a). The first chimeric gene (35S-AroGWT) contained the coding DNA sequence of a WT E. coli AroG gene, encoding the WT bacterial DAHPS (accession number AAA23492), fused in frame at its 5’ end to a DNA encoding a plastid transit peptide derived from a RuBisCO small subunit gene-3A plastid transit peptide (Shaul & Galili, 1993). The second chimeric gene (35S-AroG175) was principally the same construct, except that it contained the coding DNA sequence of a mutant, bacterial DAHPS encoding a Phe-feedback-insensitive DAHPS enzyme, resulting from a substitution of the amino acid Leu to Gln at position 175 of the bacterial enzyme. The coding DNA sequences of the WT and mutant bacterial enzymes were fused at their 3’ end to a DNA encoding an HA epitope tag to allow the detection of the recombinant proteins in the transgenic plants. To examine the polypeptides produced by the chimeric AroG transgenes, proteins from independently transformed plants were subjected to immunoblot analysis with anti-HA antibodies. As shown in Fig. S1b, the immunoblot analysis revealed that the transgenic plants expressing the plastidic AroG constructs produced two AroG-derived polypeptide bands: the lower band corresponded in size to the mature AroG polypeptide (42.5 kDa) and the upper band, which migrated slightly more slowly, corresponded in size to the unprocessed AroG containing the plastid transit peptide (48.1 kDa). These results indicated that a major portion of the AroG polypeptide, produced by this transgene, was processed and translocated into the plastids, as documented previously for genes encoding other bacterial enzymes fused to a plastid transit peptide (Tzin et al., 2009). Homozygous T2 plants were generated that contained a single insertion, based on their 3 : 1 segregation for resistance to Basta selection in the heterozygous state, and were used for further analysis. The transgenic AroGWT- and AroG175-expressing plants had comparable phenotypes to the control plants and were fully fertile (data not shown).

To obtain a global view of the effect of expression of the AroGWT and AroG175 transgenes on plant metabolism, we performed a GC-MS analysis of polar compounds of derivatized extracts using four independently transformed lines from each of these transgenes. As shown in the PCA plot (Fig. 2a), the metabolic profiles of the control and AroGWT lines were grouped together, and were completely separated from the AroG175 Phe-feedback-insensitive lines. To test further whether AroGWT and AroG175 expression altered the levels of metabolites associated with the shikimate and AAA biosynthesis pathways, we focused on related metabolites that could be detected by the GC-MS platform. As shown in Fig. 2b, the levels of shikimate, prephenate and Phe were higher in the different AroG175 lines than in the AroGWT lines, expressing the natural feedback-sensitive bacterial enzyme, and the control plants. These results implied that the bacterial AroG enzyme was active in the transgenic plants and, similar to its operation in bacteria, the feedback insensitivity trait of the AroG175 enzyme was fundamental for enhancement of the flow of primary carbon metabolites via the shikimate pathway into the production of AAAs also in the plant. We therefore continued the study with plants expressing AroG175 polypeptide.

Figure 2.

Initial metabolic characterization of Arabidopsis thaliana plants expressing the AroGWT and AroG175 genes. Samples of 20-d-old Arabidopsis seedlings were extracted from aerial tissue and subjected to GC-MS analysis. (a) Principal component analysis (PCA) plot of datasets obtained from 90 known (targeted) metabolites derived from the AroGWT (four independent lines), AroG175 (four independent lines) and control plants. The combined percentages of the variance of the first two dimensions (X, Y) are indicated. (b) Fold change of the metabolites shikimate (white bars), prephenate (gray bars) and phenylalanine (Phe) (black bars) in AroGWT (four lines) and AroG175 (four lines) compared with control. Error bars represent the standard error (n = 3).

Metabolic analysis elucidates a significant effect of AroG175 expression on both primary and secondary metabolism

To study the effect of AroG175 expression on a wide range of primary and secondary metabolites, we selected two independently transformed homozygous lines (AroG175-2 and AroG175-21) which showed a similar trend in the metabolite fold changes presented in Fig. 2. However, line AroG175-2 displayed milder changes and line AroG175-21 displayed strong metabolic changes. Aerial tissues of 10-d-old Arabidopsis seedlings were collected and subjected to GC-MS as well as targeted and nontargeted LC-MS analyses. A PCA plot of the GC-MS data displayed significant separation between the two AroG175 lines and the control (Fig. 3a). Furthermore, PCA of the LC-MS data demonstrated a relatively small separation between AroG175-2 and the control line, but a considerably greater separation between AroG175-21 and both the control and AroG175-2 lines (Fig. 3b). Analysis of the levels of specific metabolites by GC-MS and LC-MS (Fig. 4) also showed that, as expected, the changes were principally stronger in the AroG175-21 line than in the AroG175-2 line. GC-MS analysis revealed that the levels of seven metabolites were increased significantly in the AroG lines, including the primary metabolites shikimate, prephenate, Phe and Ala (Fig. 4a–c,f), as well as the secondary metabolites phenylacetonitrile (isothiocyanate) and homogentisate (Tyr derivative which is a precursor of tocopherols and tocotrienols) (Fig. 4d,e) and 4-hydroxybenzoate (Fig. 4g). In the LC-MS analysis, the AroG175-2 line generally displayed milder changes in the levels of the different metabolites relative to the AroG175-21 line (Table S2). Thus, in the nontargeted LC-MS analysis, we focused only on the metabolites whose levels were significantly different (fold change of more than two) between the AroG175-21 line and the control, but also showed similar trends in the AroG175-2 line. The nontargeted LC-MS analysis of AroG175-21 revealed a total of 4473 mass signals, the levels of 2414 of which were altered significantly (up- or down-regulated) when compared with the control. Assuming an average of five mass signals per single metabolite (Malitsky et al., 2008), AroG175-21 expression resulted in altered levels of c. 500 metabolites. We identified 45 metabolites according to standards and cross-reference with chemical databases; 17 presented a significant induction (fold change of more than two), and none presented a reduction in fold change.

Figure 3.

Principal component analysis (PCA) plots of datasets obtained from metabolic analyses of the control (squares) and AroG175 lines 2 (triangles) and 21 (circles). The samples from the control and two independently transformed AroG175 lines (n = 5–6) were extracted from aerial tissues of 10-d-old Arabidopsis thaliana seedlings and metabolites were detected by GC-MS analysis (137 known and unknown integrated metabolites) (a) and ultra-performance liquid chromatography/quadrupole time of flight-mass spectrometry (UPLC/qTOF-MS) (4474 mass signals) (b). The combined percentages of the variance of the first two dimensions (X, Y) are indicated at the bottom of each panel.

Figure 4.

Fold change of metabolites detected by GC-MS and LC-MS in the control and transgenic plants expressing the AroG175 gene. The metabolite levels represent the fold change in two AroG175 lines (AroG175-2, A2 and AroG175-21, A21) compared with control (Con) plants (= 5–6). The samples were extracted from aerial tissues of 10-d-old Arabidopsis thaliana seedlings and metabolites were detected by GC-MS (a–g) and LC-MS (h–y). Asterisks indicate statistically significant differences between the two AroG175 lines and the control, using Student’s t-test; P = 0.05 with a false discovery rate (FDR) cut-off (P < 0.05). Bars on top of the histograms indicate standard errors.

With regard to the other two AAAs, this analysis showed that the level of Trp was 2.6-fold higher in the AroG175-21 line than in the control, but its level in the AroG175-2 line was not significantly different from that in the control (Fig. 4h). The level of Tyr was not altered in the two transgenic lines when compared with the control (Table S2), although homogentisate was increased in both lines.

The LC-MS analysis also showed that the levels of a number of Phe-derived phenylpropanoid secondary metabolites were significantly higher in AroG175-21 than in control plants (Fig. 4i–v). These included coumarate hexose derivatives (I–II), ferulate hexose derivatives (I–II), ferulic acid derivatives (I–II), a sinapoyl hexose derivative, sinapate, sinapyl alcohol, coniferin, kaempferol deoxyhexose derivatives and 2-phenylethyl glucosinolate (see also Table S2). The levels of some secondary metabolites derived from Trp were also significantly higher in the AroG175-21 line, including the Trp-derived glucosinolate 4-methoxyindole glucosinolate (Fig. 4x) and the indole-3-acetate (IAA) conjugate 4-O-(indole-3-acetyl)-dihexose (Fig. 4w). Additional metabolites whose levels were higher in the AroG175-21 line included derivatives of the hormone conjugates jasmonate (12-hydroxy jasmonate-hexose; Fig. 4y) and salicylate (hydroxybenzoate hexose; Fig. 4u).

In addition to the LC-MS target analysis for semi-polar compounds, we measured the lignin monomer composition in stems and flowers of Arabidopsis plants (stems and flowers were selected as mature plant tissues with high lignin). As several precursors of lignin monomers were strongly increased in both AroG175 lines (sinapate, coniferin and sinapoyl hexose; Fig. 4), we analyzed only one line, AroG175-2. The total lignin monomer content in the stems was similar between the control and AroG175-2 lines. However, in flowers, the total lignin content was significantly higher in the AroG175-2 line than in the control, particularly because of the induction of higher levels of monomers H and G (Fig. 5).

Figure 5.

Effect of expression of the AroG175 gene on lignin monomer composition in Arabidopsis thaliana stems and flowers. The following lignin monomers are shown: guaiacyl (white bars) units derived from coniferyl alcohol; syringyl (gray bars) units derived from sinapyl alcohol; and p-hydroxyphenyl (black bars) units derived from p-coumaryl alcohol in μmol g−1 of cell wall-extracted residues (CWR). Asterisks indicate statistically significant differences between the AroG175-2 line and the control, using Student’s t-test; P = 0.05. Error bars indicate standard errors.

We also studied the effect of AroG expression on the Trp flux by growing young plants of the four AroG175 lines (2, 8, 21 and 28), showing high levels of shikimate, prephenate and Phe (Fig. 2), for 1 month on medium containing the toxic Trp analog, 5-methyl-Trp (5MT). 5MT inhibits the first enzyme from Trp biosynthesis, anthranilate synthetase (Widholm, 1972; Kisaka et al., 1996). and has been used previously to show that the expression of the E. coli PheA* gene (encoding a bifunctional feedback-insensitive chorismate mutase/prephenate dehydratase which converts chorismate via prephenate to phenylpyruvate) in Arabidopsis renders the plants more sensitive to growth on medium containing 5MT (Tzin et al., 2009). As shown in Fig. 6, all Arabidopsis genotypes (the four AroG175 lines and the control) grew on medium containing 75 and 100 μM 5MT, whereas only the AroG175-expressing lines grew on medium containing 150 μM 5MT. Among the four AroG175-expressing lines grown on medium containing 150 μM 5MT, the AroG175-21 and AroG175-2 lines exhibited better growth than the AroG175-8 and AroG175-28 lines, and the AroG175-21 line seemed to grow slightly better than the AroG175-2 line. These results imply that AroG175 expression stimulates the flux from chorismate to Trp biosynthesis.

Figure 6.

Effect of 5-methyl-tryptophan (5MT) on the growth of Arabidopsis thaliana plants expressing the AroG175 gene. Seeds were germinated on media containing three concentrations of 5MT (75, 100, 150 μM) and medium without treatment. The different lines tested are indicated.


Expression of the bacterial feedback-insensitive DAHPS differentially influences the production of Phe and Trp

Despite the major importance of AAA metabolism in plant primary and secondary metabolism, very little is known about its regulation. The observation that AroG175 expression enhances the levels of Phe and Trp in Arabidopsis plants implies that DAHPS is a limiting enzyme regulating the conversion of primary carbon metabolites into the biosynthesis of chorismate en route to the production of these amino acids. Previous studies dealing with carbon consumption into Phe vs Tyr in Arabidopsis plants have suggested that c. 30% of the carbon fixed in photosynthesis courses down the Phe branch towards lignin biosynthesis, with the flux down the Tyr branch being far smaller (Rippert & Matringe, 2002; Pribat et al., 2010). Of the two AAAs, Phe accumulation was stimulated to a much higher degree than Trp in the AroG175-expressing plants (Fig. 4). This implies that, under favorable (nonstress) growth conditions, the Phe biosynthesis pathway efficiently competes with the Trp biosynthesis pathway over their common precursor metabolite chorismate (such competition may also be related to the extensive synthesis of lignin), or, alternatively, Trp is more rapidly catabolized than Phe (Fig. 7). This can be explained by the existence of either elevated expression and/or higher enzymatic properties of chorismate mutase of Phe/Tyr biosynthesis over anthranilate synthase of Trp biosynthesis (Fig. 1). Interestingly, despite the fact that chorismate is channeled more efficiently towards lignin and Phe than Trp and Tyr under favorable growth conditions, some stress conditions may also require extensive synthesis of Trp and its catabolic products. A variety of biotic and abiotic stresses that stimulate the production of Trp-derived secondary metabolites are associated with the induced expression of genes controlling both Trp biosynthesis and catabolism enzymes (Zhao & Last, 1996; Less & Galili, 2008). By contrast, stresses that stimulate the production of Phe-derived secondary metabolites are associated only with stimulated expression of the gene encoding the Phe-catabolic enzyme Phe ammonia lyase (PAL), which converts Phe to multiple secondary aromatic metabolites, but not of genes of Phe biosynthesis (Shadle et al., 2003; Less & Galili, 2008). Increased expression of the Trp biosynthetic enzymes under stress conditions may enable a more efficient competition of this pathway with Phe/Tyr biosynthesis in stresses that require extensive production of Trp-derived secondary metabolites, such as auxin (Tzin & Galili, 2010a).

Figure 7.

A metabolic map describing changes in the levels of specific metabolites in plants expressing the AroG175 gene, compared with the control. Metabolites whose levels were increased are marked in small colored squares, pink representing a two- to four-fold increase and red representing a > four-fold increase [correction added after online publication 15 March 2012: the figure key and legend have been corrected to note that the increase in metabolites indicated by red squares represents a greater than four-fold increase]. None of the identified metabolites were significantly decreased. The dashed black arrows represent several consecutive enzymatic steps. The gray arrows represent unknown enzymatic steps. The classes of metabolites are shown in italics. Background colors: yellow, primary metabolite pathways; gray, secondary metabolite pathways derived from aromatic amino acids (AAAs); green, secondary metabolite pathways derived from other substrates (not AAAs). DAHPS, 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase.

AroG175 expression reveals potential metabolic bottlenecks within the shikimate pathway and between primary and secondary metabolism

We found that AroG175 expression lines demonstrate a significant increase in the accumulation of shikimate, as well as prephenate, the precursor for Phe biosynthesis (Figs 1, 2b, 4a–c). A possible explanation of this phenomenon is that the enzymatic steps involving shikimate kinase and prephenate aminotransferase, converting the metabolites shikimate and prephenate to their respective downstream metabolites shikimate-3-phosphate and arogenate/phenylpyruvate (Fig. 1), represent potential metabolic bottlenecks of the shikimate and AAA biosynthesis pathways, at least when DAHPS activity is not a limiting factor. Yet, the accumulation of shikimate appears to be more complex, as a recent publication has shown that RNA interference-mediated suppression of the bifunctional shikimate pathway enzyme 3-dehydroquinate dehydratase/shikimate 5-dehydrogenase (DHQ/SDH), which converts dehydroquinate into shikimate (Fig. 1), causes an accumulation rather than reduction in the level of shikimate in transgenic tobacco plants (Ding et al., 2007). Moreover, shikimate has also been shown to unexpectedly accumulate in herbicide-resistant plants possessing a mutant EPSPS enzyme, whose product is located downstream of shikimate in the shikimate pathway, although this may have been a result of altered biochemical properties of the mutant enzyme (Pline et al., 2002; Mueller et al., 2003; Buehring et al., 2007) (Fig. 1). Metabolic profiles of Lolium perenne have shown the induction of shikimate, AAAs and several phenylpropanoids under fungal endophyte infection (Rasmussen et al., 2008). Thus, we suggest that shikimate is an essential metabolite that may indicate the metabolic status of the plant, as well as the regulatory complexity of the whole shikimate pathway. As the levels of some shikimate pathway intermediate metabolites were below detection levels, it is impossible to elucidate whether other enzymatic steps of the shikimate pathway also participate in the regulation of primary to secondary metabolism.

AroG expression influences the production of various secondary metabolites derived from AAAs

In this work, we also observed that AroG175 expression increases the levels of Phe and Trp, as well as a number of Phe-derived secondary metabolites, including lignin precursors and their derivatives, flavonoids, Phe glucosinolates, indole glucosinolate, and isothiocyanate and salicylate derivatives (Figs 4, 7). Interestingly, shikimate also serves as a substrate for hydroxycinnamoyl-CoA shikimate/quinate hydroxycinnamoyltransferase (HCT), an enzyme which plays a significant role in lignification (Hoffmann et al., 2004). HCT down-regulation in plants caused a large reduction in lignin content and altered lignin monomer content (Hoffmann et al., 2004; Chen et al., 2006). In floral organs of AroG175-2-expressing plants, the total lignin content, as well as the H and G monomer content, increased significantly. Shikimate is the substrate of HCT, which may explain the induction of monomer G (greater abundance than monomer S in both stem and flowers). However, the increase in the total content of monomer H, but not monomer S, and only in flowers, may indicate other regulatory factors and bottlenecks in the pathways. This suggests that shikimate is a regulator of the metabolic status of the shikimate pathway and/or lignin biosynthesis.

Salicylate can be synthesized from chorismate, cinnamate or benzoate (Chen et al., 2009) and its conjugates include several glucosylated forms, such as salicyloyl glucose ester and salicyloyl glucoside (Erb & Glauser, 2010). Altered production of Phe-derived secondary metabolites has also been observed previously on expression of a bacterial bifunctional PheA gene (Tzin et al., 2009). These data hypothesize the presence of a metabolic cross-interaction between the fluxes of the shikimate and AAA biosynthesis pathways, and their further metabolism into various secondary metabolites. They also indicate that DAHPS functions as an important regulatory step in the conversion of primary to secondary metabolism in plants. However, although plants are apparently able to convert a significant amount of primary metabolites into secondary metabolites (Haslam, 1993), the metabolic profiling analysis showed that both PheA* expression (Tzin et al., 2009) and AroG175 expression (this article) had essentially no major effect on the levels of primary metabolites (Figs 4, 7 and Table S1). This may indicate that enzymatic steps other than those involving DAHPS regulate the conversion of primary metabolism.


We thank Hanna Levanony and Clarita BenDayan for excellent technical assistance, Ester Feldmesser for assistance with the bioinformatics analysis and Arye Tishbee for operating the LC-MS instrument. This study was supported by: A Magnet Program of the Israeli Ministry of Industry, Trade and Labor and the Israeli Bio-TOV Consortium including HAZERA GENETICS Ltd., Evogene Ltd., FRUTAROM Ltd., RAHAN MERISTEM (1998) Ltd. and ZERAIM GEDERA Ltd; and AERI Alternative Sustainable Energy Research Initiative of the Weizmann Institute of Science. The research in the laboratory of A.A. was also supported by research grants from the European Research Council (ERC) project SAMIT (FP7 program), Sir Harry Djanogly, CBE, Mrs Louise Gartner, Dallas, TX, USA, and Mr and Mrs Mordechai Segal, Israel. G.G. is the incumbent of the Bronfman Chair in Plant Sciences. A.A. is the incumbent of the Adolpho and Evelyn Blum Career Development Chair.