DOF transcription factor AtDof1.1 (OBP2) is part of a regulatory network controlling glucosinolate biosynthesis in Arabidopsis

Authors

  • Aleksandra Skirycz,

    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    2. Institute for Biochemistry and Biology, University of Potsdam, Karl-Liebknecht Str. 24-25, 14476 Golm, Germany
    Search for more papers by this author
  • Michael Reichelt,

    1. Max-Planck Institute for Chemical Ecology, Hans-Knöll-Straße 8, D-07745 Jena, Germany,
    Search for more papers by this author
  • Meike Burow,

    1. Max-Planck Institute for Chemical Ecology, Hans-Knöll-Straße 8, D-07745 Jena, Germany,
    Search for more papers by this author
  • Claudia Birkemeyer,

    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    Search for more papers by this author
  • Jakub Rolcik,

    1. Laboratory of Growth Regulators, Palacký University and Institute of Experimental Botany AS CR, Ślechtitelů 11, CZ-783 71 Olomouc, Czech Republic,
    Search for more papers by this author
  • Joachim Kopka,

    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    Search for more papers by this author
  • Maria Inès Zanor,

    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    Search for more papers by this author
  • Jonathan Gershenzon,

    1. Max-Planck Institute for Chemical Ecology, Hans-Knöll-Straße 8, D-07745 Jena, Germany,
    Search for more papers by this author
  • Miroslav Strnad,

    1. Laboratory of Growth Regulators, Palacký University and Institute of Experimental Botany AS CR, Ślechtitelů 11, CZ-783 71 Olomouc, Czech Republic,
    Search for more papers by this author
  • Jan Szopa,

    1. Institute of Biochemistry and Molecular Biology, University of Wroclaw, Przybyszewskiego 63-77, 51148 Wroclaw, Poland, and
    Search for more papers by this author
  • Bernd Mueller-Roeber,

    Corresponding author
    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    2. Institute for Biochemistry and Biology, University of Potsdam, Karl-Liebknecht Str. 24-25, 14476 Golm, Germany
    Search for more papers by this author
  • Isabell Witt

    1. Max-Planck Institute of Molecular Plant Physiology, Cooperative Research Group, Am Mühlenberg 1, D-14476 Golm, Germany,
    Search for more papers by this author

*(fax +49 331 977 2512; e-mail bmr@rz.uni-potsdam.de).

Summary

Glucosinolates are a group of secondary metabolites that function as defense substances against herbivores and micro-organisms in the plant order Capparales. Indole glucosinolates (IGS), derivatives of tryptophan, may also influence plant growth and development. In Arabidopsis thaliana, indole-3-acetaldoxime (IAOx) produced from tryptophan by the activity of two cytochrome P450 enzymes, CYP79B2 and CYP79B3, serves as a precursor for IGS biosynthesis but is also an intermediate in the biosynthetic pathway of indole-3-acetic acid (IAA). Another cytochrome P450 enzyme, CYP83B1, funnels IAOx into IGS. Although there is increasing information about the genes involved in this biochemical pathway, their regulation is not fully understood. OBP2 has recently been identified as a member of the DNA-binding-with-one-finger (DOF) transcription factors, but its function has not been studied in detail so far. Here we report that OBP2 is expressed in the vasculature of all Arabidopsis organs, including leaves, roots, flower stalks and petals. OBP2 expression is induced in response to a generalist herbivore, Spodoptera littoralis, and by treatment with the plant signalling molecule methyl jasmonate, both of which also trigger IGS accumulation. Constitutive and inducible over-expression of OBP2 activates expression of CYP83B1. In addition, auxin concentration is increased in leaves and seedlings of OBP2 over-expression lines relative to wild-type, and plant size is diminished due to a reduction in cell size. RNA interference-mediated OBP2 blockade leads to reduced expression of CYP83B1. Collectively, these data provide evidence that OBP2 is part of a regulatory network that regulates glucosinolate biosynthesis in Arabidopsis.

Introduction

Glucosinolates (GS) are a group of secondary metabolites that function as defense substances against herbivores and micro-organisms in the plant order Capparales, which includes the model plant Arabidopsis thaliana and economically important crop species such as oilseed rape (Brassica napus), Brassica fodder and vegetables (Brader et al., 2001; Fahey et al., 2001). Similar to other plant secondary compounds such as alkaloids or flavonoids, GS and its degradation products additionally have toxic as well as protective effects (e.g. as flavour compounds and cancer-preventive agents) in higher animals and humans. For that reason, there is a strong interest in the ability to regulate and optimize the levels of individual GS in specific tissues in order to improve the nutritional value and pest resistance of crops (Mithen, 2000; van Poppel, 1999). Extensive studies on GS metabolism in Arabidopsis thaliana have brought researchers closer to this goal, and several of the enzymes responsible for GS production have been identified and characterized (reviewed by Hansen and Halkier, 2005; Wittstock and Halkier, 2002). There is evidence that indole glucosinolates (IGS), derived from tryptophan, not only have an essential function in plant defense, but may also influence plant growth and development. Indole-3-acetaldoxime (IAOx), produced from tryptophan by the activity of the cytochrome P450 enzymes CYP79B2 and CYP79B3, is a precursor for IGS biosynthesis but is also an intermediate in the biosynthetic pathway of the plant hormone indole-3-acetic acid (IAA) (Hull et al., 2000; Mikkelsen et al., 2000; Zhao et al., 2002). CYP83B1, another cytochrome P450 enzyme, channels IAOx towards IGS (Barlier et al., 2000). Gain- and loss-of-function mutations of any of these enzymes changed both IGS and IAA levels, suggesting overlapping regulation of both pathways (Bak et al., 2001; Delarue et al., 1998; Zhao et al., 2002). As CYP83B1 acts at a metabolic branch point, it was proposed to be a key enzyme for IAA/IGS homeostasis in plants (Bak et al., 2001). However, IAA biosynthesis through the CYP79B2/B3–IAOx pathway may perhaps not be predominant in Arabidopsis because the free IAA level in CYP79B2/B3 double null mutants was not altered in comparison to the wild-type when plants were grown at 21°C, and was only partially affected at 26°C (Zhao et al., 2002). There is also evidence that IGS production is regulated both environmentally and developmentally by different signalling molecules [auxin, methyl jasmonate (MeJA), salicylic acid (SA)] and external stimuli (wounding, pathogen attack). Changes in the IGS level positively correlate with expression of genes crucial for IGS production, suggesting an involvement of transcriptional control in this regulatory network (Brader et al., 2001; Mikkelsen et al., 2003). However, except for the MYB transcription factor ATR1 that regulates some IGS pathway genes (Bender and Fink, 1998; Celenza et al., 2005), transcriptional regulators of this network are unknown.

The Arabidopsis genome encodes approximately 37 members of the DNA-binding-with-one-finger (DOF) transcription factors that play various roles in plant-specific processes. The amino acid sequence of DOF proteins is divergent except for the highly conserved N-terminal DOF region that acts as both a DNA-binding and protein–protein interaction domain. This domain binds the conserved DNA cis-element (A/T)AAAG or its complementary inverse sequence (summarized in Yanagisawa, 2002, 2004).

OBP2, also called AtDof1.1 (At1g07640), was identified as a member of the Arabidopsis DOF transcription factor family (Kang and Singh, 2000). OBP2 physically interacts with OBF4, a BZIP transcription factor, and in vitro stimulates its binding to ocs elements characteristic for pathogen-responsive plant promoters (Kang and Singh, 2000; Zhang et al., 1995). The biological function of OBP2 has not been investigated so far.

We are particularly interested in analyzing the role of DOF transcription factors expressed in leaves, and have previously provided evidence for their involvement in stomatal guard cell gene expression (Plesch et al., 2001). Using RT-PCR, we identified OBP2 among leaf-expressed DOF genes (data not shown), confirming previous observations (Kang and Singh, 2000). In this paper, we investigate the expression and function of OBP2. It is prominently expressed in the phloem of leaves and other organs. In wild-type plants, expression is stimulated by wounding, MeJA treatment and in response to insect feeding which precedes increased expression of GS biosynthetic genes and a corresponding accumulation of IGS. Accordingly over-expression of OBP2 in transgenic plants activates the IGS biosynthetic pathway and RNA interference of OBP2 leads to reduced expression of CYP83B1, a key gene in IGS biosynthesis. Collectively, these data provide evidence that OBP2 is part of a regulatory network that regulates GS biosynthesis in Arabidopsis.

Results

Identification of OBP2-regulated biological processes using transcript and metabolite profiling

As part of a wider screen to identify candidate genes regulated by DOF transcription factors, we performed Affymetrix expression profiling with the AtGenome array representing approximately 8200 genes using constitutive over-expression lines. Specifically, we compared two independent 35S:OBP2 lines to two independent empty-vector control lines and two biologically replicated wild-type samples. Genes were considered as changed if they met the suggested criteria in the Affymetrix Microarray suite 5 (MAS5) software (significant increase/decrease, appropriate ‘present’ call and greater than twofold change) in all eight comparisons between over-expressing lines and empty-vector control lines or wild-type. We identified 89 up-regulated and 43 down-regulated genes in 35S:OBP2 lines. Genes were classified according to their annotated function (see Tables S1 and S2). Of the up-regulated genes, 32 are related to stress responses (Table S3), some are regulated by abiotic stress such as the drought-induced protein D21, dehydrin Xero20, a putative galactinol synthase, heat shock protein 17.6A, heat shock protein 70kD and putative metal ion transporters. However, the majority of these genes have assigned functions in response to biotic stress and defense. These include genes involved in secondary metabolism (FAD-linked oxidoreductase), glycosyl hydrolase family members (1-α-glucosidase), cell wall modification enzymes (thaumatin and xyloglucan endotransglycosylase), a jasmonate biosynthetic enzyme (12-oxo-phytodienoic acid reductase), proteinases, PR proteins, and others such as vegetative storage proteins.

Among these biotic stress genes, the most prominent group of up-regulated genes were those encoding proteins of the glucosinolate/myrosinase system. Of the seven up-regulated genes, three encoded myrosinase-binding proteins (MBP) that possibly regulate myrosinase activity. Also up-regulated were the β-subunit of tryptophan synthase (TSB2), involved in tryptophan biosynthesis, and ATR1 AtMYB34) which regulates the expression of anthranilate synthase, both key enzymes of the tryptophan biosynthetic pathway (Bender and Fink, 1998). Biosynthesis of IGS initiates at the conversion of tryptophan to IAOx, catalyzed by CYP79B2 or CYP79B3 (Hull et al., 2000; Mikkelsen et al., 2000). Expression of both genes was up-regulated in 35S:OBP2 plants, suggesting changes in the IAOx pool.

To complement our transcript analysis, we performed metabolite profiling experiments which allowed the detection and quantification of a large number of low-molecular-weight compounds from leaf extracts, including hydroxyl- and amino acids, sugars, sugar alcohols, organic monophosphates, (poly)amines and aromatic acids (Roessner et al., 2000). We analyzed 35S:OBP2 plants using gas chromatography coupled to mass spectrometry (GC/MS) and applied principle component analysis (PCA) to the metabolite data set (Roessner et al., 2001). Metabolite data are given in Table S4. The metabolite data for 35S:OBP2 lines are clearly separated from wild-type (Figure S1). The five most important metabolites responsible for this separation were proline, galactinol, an unknown metabolite with high similarity to digalactosylglycerol, an unknown metabolite that may be a product of indole glucosinolate breakdown occurring during sample preparation (β-d-glucopyranose, 2,3,4,6-tetrakis-O-(trimethylsilyl)-, 1-(trimethylsilyl)-1H-indole-3-acetate) and glutamine. Moreover, transgenic samples clustered according to the level of OBP2 expression, with the stronger lines 1 and 10 forming a cluster separate from the weaker over-expression line 8 (Figure S1). Metabolites responsible for this separation were glycine, galactinol, fructose, raffinose and fumaric acid.

Many of these metabolites are known to accumulate in response to stress (reviewed by Loewus and Pushpalatha, 2000), and furthermore the identification of a putative IGS derivative supported a possible effect on GS metabolism.

Detailed analysis of the GS biosynthetic pathway

Our profiling analyses identified GS metabolism as the most likely biological target of OBP2 action. Targeted expression analysis on 35S:OBP2 plants using quantitative real-time PCR (Q-RTPCR) confirmed the increased transcript levels of CYP79B2,-B3 and ATR1 identified by transcript profiling (Figure 1a) and additionally showed increases for CYP83B1 and MAM-1, which were not represented on the array. CYP83B1 is the branching enzyme shifting the endogenous pool of IAOx towards IGS production, whereas MAM-1 catalyzes the condensing reactions of the first two methionine elongation cycles in short-chain aliphatic GS biosynthesis. Other enzymes from the aliphatic GS pathway, MAM-L, responsible for further elongation cycles of methionine (Field et al., 2004), CYP79F1/F2, which metabolize chain-elongated methionine derivatives into aliphatic oximes, and CYP83A1, which funnels these into aliphatic GS (Hansen et al., 2001; Naur et al., 2003), were not changed.

Figure 1.

 Expression levels of genes involved in GS biosynthesis or its regulation and GS content in transgenic plants with altered OBP2 expression.
Transcript level was measured using Q-RTPCR in: (a) 35S:OBP2 (OE) lines 1, 8 and 10, (d) DEX-inducible OBP2–GR lines 3 and 5, and empty-vector d143 control lines 3 and 5, and (e) RNAi–OBP2 lines 1, 2, 3, 5 and 11 (K, control treatment with 0.5% ethanol). (b, c) GS were determined using HPLC in 35S:OBP2 (OE) plants. Data are means ± SD for n = 3; no SD is indicated if n = 2. 35S:OBP2 and RNAi–OBP2 expression samples were from three independent experiments pooled from three plants each; all other samples were replicated with individual plants from one experiment. All samples were leaf samples taken from plants prior to bolting. 79B2, CYP79B2; 79B3, CYP79B3; 83B1, CYP83B1. Asterisks indicate values that are significantly different (P < 0.05) in comparison to respective controls.

To study the effect of these expression changes on GS biosynthesis, we measured GS in 35S:OBP2 plants in comparison to control plants. In agreement with the elevated expression of CYP79B2/B3 and CYP83B1, total IGS concentration was increased approximately twofold and the concentration of the major IGS indol-3-yl-methyl glucosinolate (I3M) was elevated approximately threefold on a leaf fresh weight basis (Figure 1b; Table S5). Concomitant with elevated MAM-1 expression, the aliphatic GS content was increased due to the accumulation of short-chain GS (Figure 1c, Table S5).

Inducible over-expression and RNA interference (RNAi) were used to modulate OBP2 expression. We used a dexamethasone (DEX)-inducible system consisting of CaMV 35S-driven expression of OBP2 fused to a glucocorticoid receptor (GR) domain, which upon DEX treatment becomes active due to nuclear targeting of the fusion protein (Lloyd et al., 1994), whereas the expression level of the chimeric gene itself remains unchanged (as seen in Figure 4d). Two OBP2–GR lines were compared to two empty-vector control lines either treated with a control solution or with 50 μm DEX. Plant material was harvested 10 and 24 h after treatment and used for Q-RTPCR experiments. In comparison to control lines, expression of CYP83B1 was consistently and significantly increased in all OBP2–GR lines 10 and 24 h after treatment (Figure 1d). Expression of CYP79B2, CYP79B3 and ATR1 were either not changed or increased in both OBP2–GR and the empty-vector control lines after DEX treatment (data not shown). MAM-1 was not changed in any of the lines tested. Indole GS were increased in a single line that had elevated expression of all three cytochrome P450 enzymes; however, in a separate experiment, only CYP83B1 expression was elevated and correspondingly the GS level was comparable to the control situation. In agreement with the absence of changes in MAM-1 expression, the level of aliphatic GS was not changed in any line.

Figure 4.

 Phenotype of plants with modified OBP2 expression
(a) Four-week-old wild-type plant (bar, 2 cm). (b) Four-week-old 35S:OBP2 transgenic line (bar, 2 cm). (c) Seven-week-old wild-type plant (bar, 3 cm). (d) Twelve-week-old 35S:OBP2 plant (bar, 3 cm). (e) Third mature leaf of 35S:OBP2 plant (bar, 0.5 cm). (f–i) Cells of the third mature leaf of 5-week-old wild-type and 35S:OBP2 plants were microscopically analyzed. (f) Adaxial epidermis of wild-type leaf (bar, 50 μm). (g) Adaxial epidermis of 35S:OBP2 leaf (bar, 50 μm). (h) Transverse section through the central part of a wild-type leaf (bar, 200 μm). (i) Transverse section through the central part of a 35S:OBP2 leaf (bar, 200 μm). (j) Five-week-old empty-vector control plants (d143) 1 week after DEX (left) or control (0.5% ethanol) treatment (right) (bar, 3 cm). (k) Five-week-old OBP2–GR plants 1 week after DEX (left) or control (0.5% ethanol) treatment (right) (bar, 3 cm). The magnification shows the newly forming, curly leaves that develop upon DEX treatment. (l) Seedlings of RNAi –OBP2 (left), wild-type (middle) and 35S:OBP2 (right) plants kept on agar plates. Note the reduced hypocotyl length in the case of the over-expression line (bar, 0.5 cm).

To reduce OBP2 transcript level, we transformed Arabidopsis plants with an RNAiOBP2 construct (see Experimental procedures). Several independent transformants with reduced OBP2 transcript level were identified using Q-RTPCR following BASTA selection (Figure 1e). CYP83B1 was significantly reduced in two lines, whilst CYP79B2 and B3 were significantly reduced in just one line and ATR1 and MAM-1 were not changed. In addition, the GS content of RNAi:OBP2 plants was not significantly different from wild-type (data not shown), possibly due to threshold levels in control of the GS pathway.

Biological function of OBP2

Changes of OBP2 expression in response to treatments known to stimulate IGS biosynthesis such as MeJA application, wounding and herbivore feeding were measured to investigate their biological significance (Brader et al., 2001; Mikkelsen et al., 2003; Reymond et al., 2004). Northern blot analysis indicated that the OBP2 transcript level increased within 4–6 h upon external MeJA application (Figure 2a). Similarly, mechanical wounding, which stimulates jasmonate biosynthesis, enhanced OBP2 transcript level (Figure 2a). These changes were confirmed by Q-RTPCR experiments indicating a threefold induction of OBP2 transcript level after MeJA treatment (Figure 2c), and a twofold induction following wounding (Figure 2d). We also investigated these changes using activity measurements for promOBP2:GUS lines (Figure 2b), which indicated that the enhanced OBP2 expression was at least partly due to transcriptional activation mediated by its promoter. The cellular expression pattern of OBP2 activity did not change (not shown). In addition, we characterized expression of OBP2 and genes involved in GS metabolism and regulation in leaves damaged by feeding by the generalist herbivore Spodoptera littoralis. Samples were taken 0.5, 1, 2, 4, 6, 8, 10 and 24 h after feeding started, and 48 h after insect removal from the plants. OBP2 expression increased after 6 h, followed by an increase in ATR1, CYP79B2/B3 and CYP83B1 expression. These increases were transient, with transcripts returning to control levels 24 h after the feeding started. There were no changes in the transcript level of MAM-1 (Figure 2e). In agreement with these transcript changes, we observed an increase in IGS content from 24 to 72 h (Figure 2f) but no changes in the concentration of aliphatic GS (data not shown).

Figure 2.

OBP2 expression is induced by feeding with Spodoptera littoralis, MeJA treatment and wounding.
(a) Northern blot analysis of OBP2 transcript level in Arabidopsis wild-type plants upon MeJA treatment and wounding. RNA was isolated from plants sprayed with 500 μm MeJA or 0.006% ethanol solution (control). After spraying, leaf samples were taken at the indicated time points. RNA was isolated from untreated leaves (control) or mechanically wounded leaves of soil-grown plants. The blots were also hybridized to radio-labelled plant defensin 1.2 (PDF1.2) cDNA as a positive control for wounding and MeJA treatment. Loading of gels with RNA is indicated by ethidium bromide staining (lower panels).
(b) GUS activity in three independent promOBP2:GUS lines 6 h after spraying with 500 μm MeJA or wounding. Proteins were extracted from four 6-week-old treated plants each in three biological replicates.
Transcript level was measured using Q-RTPCR for (c) MeJA treatment, (d) wounding, and (e) herbivore feeding.
(f) IGS were determined using HPLC after herbivore feeding.
For herbivore feeding, data are means ± SD (n = 3), and samples were replicated with individual plants from one experiment. For MeJA and wounding experiments, samples were taken from two independent (MeJA) or one (wounding) experiments pooled from four plants. All samples were leaf samples taken from plants prior to bolting. Asterisks indicate values that are significantly different (P < 0.05) in comparison to respective controls.

OBP2 expression pattern

Using semi-quantitative RT-PCR, OBP2 was previously found to be expressed in all organs of 6-week-old Arabidopsis plants, revealing the highest expression in roots and leaves, and a comparatively weak expression in stems and flowers (Kang and Singh, 2000). To investigate OBP2 expression at the cell and tissue level, approximately 1 kb of the 5′ upstream regulatory region of OBP2 was fused to the E. coliβ-glucuronidase (GUS) reporter gene and transferred to the nuclear genome of Arabidopsis. GUS activity was observed in the central cylinder (vascular tissue and pericycle) of both main and lateral roots, but was absent from root hairs and the root cap (Figure 3b–d). In leaves, GUS expression in the vasculature was strongest in phloem cells (Figure 3e,f). GUS staining was also found in the vasculature of stems (Figure 3g,h) and in stamen filaments of flowers, whereas weaker staining was detectable in the vasculature of petals and carpels (Figure 3i–l). GUS activity in the vascular tissue was observed from the seedling stage to the mature plant (Figure 3a–l).

Figure 3.

OBP2 is expressed in the vasculature throughout the Arabidopsis plant.
Plants were transformed with the promOBP2:GUS fusion construct.
(a) Seedling kept for 5 days on sterile agar medium (bar, 5 mm). (b) Root from 2-week-old plant (bar, 5 mm). (c) Longitudinal view of root. (d) Root, transverse section (bar, 200 μm). (e) Mature leaf (bar, 5 mm). (f) Vascular bundle of a leaf, cross-section (bar, 50 μm). (g) Flower stalk, longitudinal view (bar, 5 mm). Note the blue GUS staining along the vascular bundles. (h) Flower stalk, cross-section through a vascular bundle (bar, 200 μm). (i) Flower (bar, 0.5 mm). (j) Petal (bar, 0.5 mm). (k) Stamen (bar, 0.5 mm). (l) Style (bar, 0.4 mm). ph, phloem; xy, xylem; ep, epidermis; co, cortex; en, endodermis; cc, central cylinder.

Phenotypic effects of altered OBP2 expression

Mature 35S:OBP2 plants exhibited a strong apical dominance, reduced overall height, and an increased number of curly, small rosette leaves; however, the total leaf number was slightly decreased compared with wild type (Figure 4a–e). Flower morphology was not affected but flower number was reduced and fewer and smaller siliques were produced reducing seed yield. Under long-day conditions, 35S:OBP2 plants bolted 4 weeks later than wild-type plants, but once the inflorescence was initiated no further retardation was observed (Figure 4c,d). To investigate the basis of reduced plant size, we counted and measured leaf cells. In fully expanded third leaves of 35S:OBP2 plants, epidermal and parenchyma cell size was significantly reduced and the cell number was significantly increased in comparison to the wild-type (Table 1, Figure 4f–i). Interestingly, there was an increased ratio of spongy to palisade parenchyma cells compared to wild-type. This, combined with the reduction of intracellular space in 35S:OBP2 leaves, is likely to contribute to the observed upward leaf curvature. This leaf phenotype was also observed in OBP2–GR lines with DEX-inducible OBP2 activation. Leaves started to curl 3–4 days after DEX application (Figure 4j,k). In addition, we found opposite effects in mature RNAiOBP2 plants, which were slightly larger due to increased leaf size and had increased epidermal cell area (Table 1).

Table 1.   Analysis of epidermis, palisade and spongy cells of third rosette leaves of 5-week-old wild-type (C24), 35S:OBP2 (OE) and RNAi–OBP2 plants
 C24 (n = 6)OE#8 (n = 5)OE#10 (n = 5)RNAi#5 (n = 5)RNAi#11 (n = 5)
Epidermal cells
Adaxial epidermis
 Total cellsa470.4 ± 72.2599 ± 25.7*983 ± 124*426 ± 59464 ± 62
 Cell area (mm2)b4327 ± 6501581 ± 65*1439 ± 72*5041 ± 720*5127 ± 673
Abaxial epidermis
 Total cellsa521 ± 43.8622 ± 58.3*990.1 ± 157*504 ± 57486 ± 72
 Cell area (mm2)b3915 ± 348.71526 ± 74*1446 ± 70*4728 ± 634*4812 ± 578*
 C24 (n = 3)35S:OBP2#8 (n = 3)35S:OBP2#10 (n = 3)
  1. aData are mean values ± SE from n plants.

  2. bData are mean values ± SE for approximately 100 cells from n plants.

  3. cTotal number of cells in the section, excluding epidermal cells, xylem and phloem cells.

  4. Data for epidermis cells were obtained by analyzing dental resin imprints, whereas data on palisade and spongy cells were obtained from transverse sections of leaves embedded in Technovit. Asterisks indicate values that are significantly different from the wild-type plants (Student's t test; P < 0.05).

Palisade and spongy cells
Total cellsc367.8 ± 21.15474.8 ± 26.2*794.6 ± 24.3*
Palisade cellsa83.84 ± 7.3289.53 ± 8150.3 ± 8.65*
Spongy cellsa283.9 ± 13.83385.2 ± 18.2*644.3 ± 15.65*
Spongy cells/palisade cells 3.38  4.34.28
Cell lengthb53.25 ± 0.7332.8 ± 0.81*35.7 ± 1.87*
Cell widthb28.71 ± 0.8316.9 ± 0.52 *19.3 ± 0.66*

Seedlings of 35S:OBP2 lines had shorter hypocotyls and a reduced number of lateral roots, and this was also seen when OBP2–GR plants were germinated on DEX-containing media (Figure 5a,b). An opposite phenotype was seen in RNAiOBP2 lines, which had longer hypocotyls and increased lateral root formation (Figure 5a).

Figure 5.

 Analysis of lateral root formation in wild-type (C24), 35S:OBP2, OBP2-GR and RNAi–OBP2 lines.
Given is the number of lateral roots produced per mm of primary root. Lateral roots were counted 12 days (a) or 10 days (b, c) after germination (DAG) and data are mean values ± SD (n = 10–15). (a) Comparison of C24, 35S:OBP2 (OE#1, OE#8, OE#10) and RNAi–OBP2 (R#1, R#5, R#11) lines. (b) Comparison of two empty-vector control lines (K#3, K#4) and two OBP2–GR lines (IOE#3, IOE#5) grown with (+ ) and without (-) DEX. (c) Quantitative analysis of lateral root formation in C24 and two different 35S:OBP2 lines (OE#1 and#10) grown with and without tryptamine. Asterisks indicate values that are significantly different (P < 0.05) in comparison to respective controls.

Overlap between GS and auxin metabolism

IGS were shown to have a dual role in auxin biosynthesis. IAOx is a precursor of both IGS and IAA, and IGS themselves have been speculated to be a source of the IAA precursor indole-3-acetonitrile (IAN) (Normanly and Bartel, 1999). It was previously shown that 0.1 μm 1-naphthaleneacetic acid (1-NAA) is sufficient to rescue lateral root development in the Arabidopsis aux1 mutant (Marchant et al., 2002). This concentration of 1-NAA led to a highly increased number of lateral and adventitious roots in both wild-type and 35S:OBP2 plants such that there was no remaining root phenotypic difference (data not shown). Lateral root formation in 35S:OBP2 lines was also restored by treatment with 100 μm tryptamine, an inhibitor of CYP83B1, which led to two- to threefold more lateral roots per mm of primary root in comparison to plants grown in the absence of tryptamine (Figure 5c). However, in contrast to what these data might suggest, free auxin was actually increased in 35S:OBP2 seedlings, which is also confirmed in seedlings super-transformed with an auxin-indicator DR5GFP construct (data not shown). Free and total auxin concentration was also elevated in mature leaves of 35S:OBP2 plants, but we measured no significant changes in seedlings or mature leaves of RNAi plants (Table 2).

Table 2.   Auxin levels in fully developed wild-type (C24), 35S:OBP2 and RNAi–OBP2 plants (mature leaves), and free auxin levels in liquid culture-grown seedlings of wild-type (C24), empty-vector 35S control line (K#2), 35S:OBP2 and RNAi–OBP2 plants
Mature leavesSeedlings
RNAi – OBP2Fold change in IAAa35S: OBP2Fold change in IAAaTotal auxinb (fmol g−1 DW)RNAi – OBP2Free IAA (pmol g−1 DW)35S: OBP2Free IAA (pmol g−1 DW)
  1. DW, dry weight; ND, not determined. For mature leaves, data for total auxin content are means ± SD (n = 5). Data for free auxin are represented as fold differences. Each sample represents 35–40 pooled plants harvested in one biological experiment.

  2. For seedlings, data are means ± SD (n = 3 or 4); where no SD is indicated, n = 2. Each sample represents a single flask.

  3. aFree IAA determined by GC/MS measurements.

  4. bTotal auxin determined by ELISA of immunoaffinity-purified samples.

  5. Asterisks indicate values that are significantly different from the wild-type plants (Student's t-test; P < 0.05).

#51.62*#15.35*172* ± 41.7#5810 ± 65#1937* ± 146
#111.20#85.77*221.5* ± 16#11ND#81158
# 261.05#109.11*295* ± 88#26ND#10945* ± 173
C241C241105.6 ± 6.9C24761 ± 92C24761 ± 92
K#2NDK#2NDNDK#2734 ± 75K#2734 ± 75

Discussion

OBP2, originally identified through its homology to OBP1, is a member of the DOF transcription factor family in Arabidopsis (Kang and Singh, 2000). We identified OBP2 in a search for leaf-expressed DOF genes. OBP2 was previously reported to be most highly expressed in leaves and roots (Kang and Singh, 2000). Promoter–GUS studies showed that OBP2 exhibits phloem-associated expression in all organs examined, including cotyledons, leaves, flower stalks, anther filaments, siliques and roots. This expression pattern is intriguing, and suggests that a common upstream transcription factor triggers OBP2 expression in vascular strands at different developmental stages and in different tissues. The identification of cis element(s) and the transcriptional regulator(s) binding these will require further work.

To elucidate the functional role of OBP2 in Arabidopsis, we constitutively over-expressed it and performed array-based analysis of transcripts in fully developed leaves. This revealed an induction of genes important for plant defense responses. Comparison of these genes with a study investigating the response to herbivore feeding and related treatments (Reymond et al., 2004) identified significant overlap. Of the 2964 genes present on both arrays, 45 were induced in the herbivore study and 37 were induced in 35S:OBP2 plants. Six genes were induced in both studies (VSP2, CYP79B2, galactinol synthase, putative lectin, myrosinase-binding protein and OPR3), 10.7-fold more than would be expected by chance (P = 2.432 × 10−5, Fisher exact test). Remarkably, four of these genes belong to a single expression cluster containing 26 genes shown to increase in response to insect feeding and MeJA treatment but not punctual wounding and to require COI1 for induction by Pieris rapae. CYP83B1, not present on the Affymetrix array but confirmed to increase by Q-RTPCR in our experiments (and Northern blots, data not shown), also belongs to this expression cluster. These data argue for a role of OBP2 in the response to herbivore attack. Furthermore, the mutual genes include four involved in GS metabolism, possibly regulated by OBP2. Targeted expression studies on GS biosynthesis showed that CYP79B2/B3, CYP83B1 and MAM-1 were up-regulated in 35S:OBP2 plants. Consistent with these findings, there was an accumulation of indole and short-chain aliphatic glucosinolates.

Transcript profiling and GS measurements of 35S:OBP2 plants reflect long-term changes in the steady-state mRNA levels resulting from permanent higher OBP2 expression. To investigate whether these changes were directly related to changes in OBP2 expression, we used DEX-inducible over-expression and RNAi of OBP2. Expression of CYP83B1 was up-regulated in OBP2–GR transgenic lines 10 h after induction and increased further within 24 h. However, the expression of ATR1, CYP79B2 and CYP79B3 was not consistently changed. CYP83B1 transcript level was also consistently decreased by RNAi of OBP2 expression. Consistent with the report that 35S promoter-driven over-expression or null mutation of CYP83B1 alone is not sufficient to change IGS levels in leaves harvested from the plants before bolting (Naur et al., 2003), there were no consistent changes in IGS in response to inducible over-expression or RNAi of OBP2. However, IGS levels were increased in an inducible line when the expression of CYP79B2 and CYP79B3 was also increased. Enhanced CYP79B2 and CYP79B3 expression was variable, sometimes occurring in control lines after DEX treatment. This indicates that the primary effect of OBP2 is on CYP83B1, an enzyme that catalyzes the first dedicated step of IGS biosynthesis. However, changes in IGS levels are dependent on the preceding enzymes, CYP79B2 and CYP79B3, which are commonly controlled by the transcription factor ATR1, and in our experiments their expression changes were always co-regulated. Constitutive 35S promoter-driven expression of ATR1 also increases the expression of CYP79B2, CYP79B3 and CYP83B1 in seedlings, and their expression is reduced, but still detected, in leaves of atr1 null mutants (Celenza et al., 2005). However, there is no effect of the null mutation on the expression of these genes at the seedling stage. There is complex regulation of this pathway as cyp79B2/cyp79B3 double mutants have enhanced expression of ATR1, CYP79B2,CYP79B3 and CYP83B1, whilst cyp83B1 mutants have elevated expression of ATR1, CYP79B2, CYP79B3 and mutant CYP83B1, and ASA1 and TSB1 which encode enzymes of tryptophan biosynthesis (Celenza et al., 2005). Although ASA1 and TSB1 expression are returned to near wild-type levels in cyp83B1/atr1 double mutants, the steady-state transcript levels of CYP79B2, CYP79B3 and CYP83B1 remains substantially elevated (Celenza et al., 2005). This provides clear evidence that these enzymes are regulated by factors additional to, and independent of, ATR1. Differential regulation of ATR1 and CYP83B1 has already been demonstrated following IAA treatment, which up-regulates CYP83B1 (Barlier et al., 2000) but down-regulates ATR1 (Smolen and Bender, 2002). It is possible that increased CYP83B1 expression in 35S:ATR1 plants is related to the elevated auxin levels found in these plants. In our inducible over-expression experiments, CYP83B1 was induced independently of ATR1, although ATR1 was co-regulated with CYP79B2 and CYP79B3. Independent regulation of CYP79B2/B3 versus CYP83B1 would allow more precise and separate regulation of the IGS and IAA biosynthetic pathways. CYP83B1 is a regulator of auxin production in Arabidopsis by controlling the flux of indole-3-acetaldoxime into IAA and IGS biosynthesis, and as a branching enzyme CYP83B1 should be able to precisely respond to the plant status and as such requires tight regulation (Bak et al., 2001). The induction of ATR1 and CYP79B2/B3 in 35S:OBP2 but not OBP2–GR plants may be explained by the higher expression in the constitutive plants. Alternatively, massive and sustained over-expression of transcription factors may have secondary effects, causing stress that may affect gene expression. This would be consistent with the accumulation of stress-related transcripts and metabolites in 35S:OBP2 plants. This is a problem that cannot be excluded from any study using constitutive over-expression, and here we report the regulation of a GS biosynthetic enzyme in the short-term using an inducible expression system.

There is further evidence for a role of OBP2 in biotic stress responses, possibly as part of a network that regulates GS biosynthesis in Arabidopsis. The biosynthesis of GS appears to be phloem-localized. CYP79B2, CYP79B3, UGT74B1 and IQD genes (Douglas Grubb et al., 2004; Glawischnig et al., 2004; Levy et al., 2005; Mikkelsen et al., 2000) exhibit the same expression patterns in leaves of Arabidopsis plants transformed with promoter–reporter (GUS) gene constructs. All five genes are strongly expressed in the entire leaf vasculature, including major and minor veins as well as free-ending veinlets extending into the intercostal leaf regions. The detailed expression pattern of the CYP83B1 gene is not known. Recent analysis of the vasculature transcriptome using expression profiling also showed very high phloem expression of OBP2 and CYP79B2/B3, as well as MAM-1, CYP79F1/F2 and CYP83A1 (Zhao et al., 2005), with CYP83B1, but not ATR1, also being more highly expressed in phloem than xylem. In addition, OBP2 and CYP79B2, CYP79B3 and CYP83B1 are all induced by the same spectrum of treatments, including wounding, MeJA and herbivore feeding. Expression of OBP2 was induced after 6 h of herbivore feeding, while expression of CYP83B1, CYP79B2 and CYP79B3 increased after 8 h. Auxin 2,4-dichlorophenoxyacetic acid (2,4-D) has been demonstrated to enhance expression of OBP2 in Arabidopsis leaves (Kang and Singh, 2000). Similarly, CYP83B1 transcription was shown to be induced by IAA (Barlier et al., 2000). Finally, CYP83B1 contains 11 DOF-binding sites in its 1 kb upstream 5′ regulatory region. The promoter of CYP83A1, which is not up-regulated by OBP2, contains only five potential DOF-binding sites, and a random sequence of approximately 1 kb sequence would be predicted to have less than four such elements (Kang et al., 2003). A working model for OBP2 action is depicted in Figure 6.

Figure 6.

 Proposed model of OBP2 function in plants.
OBP2 is involved in biotic stress responses by regulating GS metabolism. Candidate downstream targets include the key enzyme of IGS synthesis, CYP83B1. Expression of OBP2 is induced by herbivore feeding, MeJA treatment, wounding and auxin.

In contrast to the clear role of OBP2 in GS metabolism and biotic stress responses, its role in the observed phenotypic changes is less clear. Certainly it is likely that OBP2 regulates more genes than just CYP83B1, and these may influence the phenotype of 35S:OBP2 plants. In addition, the relationship between these changes and possible changes in auxin metabolism is also uncertain, especially considering the redundancy and complexity of auxin biosynthesis and action. Increased IAA in mature leaves of 35S:OBP2 plants could result either from increased metabolic flow through IAOx due to enhanced CYP79B2/B3 expression, or the speculated role of IGS as an additional source of the IAA precursor indole-3-acetonitrile (IAN) (Normanly and Bartel, 1999). Similar shoot phenotypes have been reported for 35S:CYP79B2 and rnt1-1 plants, which have elevated auxin concentrations (Bak et al., 2001; Mikkelsen et al., 2000). In seedlings, the phenotypic data are not consistent with the measured increase in auxin, although the restoration of lateral root growth by 1-NAA or tryptamine inhibition of CYP83B1 indicates that 35S:OBP2 plants have the capacity to respond to further increased auxin levels. The proximity of auxin to GS metabolism makes it an interesting candidate for study in relation to OBP2 function, but the evident complexity and the interaction of auxin with many other signaling pathways put this beyond the scope of the present study.

It should be pointed out that, in contrast to gain-of-function mutants of CYP79B2/B3 and CYP83B1, plants over-expressing OBP2 are characterized by up-regulation of both CYP79B2/B3 and CYP83B1 and an increase in both indole and aliphatic GS. Although our inducible data indicate that not all of these changes are likely to be primary effects, these plants may still be of considerable interest for studying the contribution of elevated GS levels to plant protection against pathogens.

Experimental procedures

General methods

Standard molecular techniques were performed as described previously (Sambrook et al., 1989). Oligonucleotides were obtained from TibMolbiol (Berlin, Germany). DNA sequencing was performed by AGOWA (Berlin, Germany). Unless otherwise indicated, other chemicals were purchased from Roche, Merck (Darmstadt, Germany) or Sigma (Deisenhofen, Germany). For sequence analyses, the tools provided by the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/), MIPS (http://mips.gsf.de/) and The Arabidopsis Information Resource (TAIR; http://www.arabidopsis.org/) were used.

Plant material

Arabidopsis seeds were sown in soil (Einheitserde GS90; Gebrüder Patzer, Sinntal-Jossa, Germany) and grown in a growth chamber with a 16 h day length provided by fluorescent light at 80 or 120 μmol m−2 sec−1 and a day/night temperature of 20/16°C and relative humidity of 60/75%. In tissue culture, seedlings were grown in half-strength Murashige and Skoog (1962) medium (0.5 MS) supplemented with 1% sucrose and solidified with 0.7% agar under a 16 h day (140 μmol m−2 sec−1)/8 h night regime (22°C). For root experiments, plants were grown vertically under the same conditions except 1% agar was used. Agrobacterium tumefaciens strain GV3101 (pMP90) was used to transform Arabidopsis thaliana (L.) Heynh. cv. C24 (Clough, 1998). If not indicated otherwise, measurements were performed on fully developed rosette leaves harvested from randomized 5- to 6-week-old plants before bolting.

Constructs

35S:OBP2.  PCR was used to amplify the OBP2 coding region using Arabidopsis C24 leaf cDNA as template. Primer sequences were as follows: OBP2-fwd, 5′-GTTTAAACATGGCGGAGAGAGCAAGGCAGG-3′ (added PmeI restriction site underlined); OBP2-rev, 5′-TTAATTAATTACCGGAGCGTCTGATAAACC-3′ (added PacI restriction site underlined). The OBP2 cDNA was inserted into pUni/V5-His-TOPO (Invitrogen, Karlsruhe, Germany), and, after sequence confirmation, cloned via the PmeI/PacI sites into a modified pGreen0229 plant transformation vector (http://www.pgreen.ac.uk) containing a cauliflower mosaic virus (CaMV) 35S promoter and PmeI/PacI restriction sites.

OBP2–GR.  Primer sequences were as follows: OBP2-fwd, 5′-GGATCCATGGGTGGATCGATGGCGGAGAG-3′ (added XbaI restriction site underlined); OBP2-rev, 5′-GGATCCCACAAGAGATCATTAGAAGGACCC-3′ (added XbaI restriction site underlined). The OBP2 cDNA was inserted into TOPO-TA (Invitrogen), and, after sequence confirmation, cloned via the XbaI sites into a d143 (pBI-GR) vector (Lloyd et al., 1994).

RNAi–OBP2. A 139 bp 3′-UTR OBP2 fragment was PCR-amplified from a leaf cDNA library using the oligonucleotides RNAi-fwd 5′-CACCGATTGTGGTTTTACAACTTAAATTCG-3′ and RNAi-rev 5′-GGACTGTTACACACATGTATTGAGGC-3′. The PCR product was cloned into pENTR/D-TOPO of the Gateway System (Invitrogen, Karlsruhe, Germany). After sequencing, the OBP2 fragment was transferred to the destination vector pJawohl8-RNAi (kindly provided by Dr Imre Somssich, MPI for Plant Breeding, Cologne, Germany).

Promoter–GUS fusion

An approximately 1 kb 5′ genomic fragment upstream of the ATG start codon was amplified by PCR using primers OBP2GUS-fwd (5′-AAGCTTAATTTGCTCCAATAATAACACATC-3′) and OBP2GUS-rev (5′-CCCGGGTGTTCCTTCTACCCTTTTTTTTTA-3′) from Arabidopsis C24 genomic DNA. The promoter fragment was inserted into plasmid pCR-Blunt II-TOPO (Invitrogen) and verified by sequencing. Subsequently, the promoter was fused to the E. coliβ-glucuronidase (GUS) reporter gene in pGPTV-Kan (Becker et al., 1992), previously cut with HindIII and SmaI, resulting in plasmid promOBP2:GUS.

GUS assays

β-Glucuronidase activity was determined histochemically as described previously (Plesch et al., 2001) using 5-bromo-4-chloro-3-indolyl-β-d-glucuronic acid (X-Gluc) as substrate (Jefferson et al., 1987). Fluorometric determination of β-glucuronidase activity was performed using 4-MUG (4-methylumbelliferyl β-d-glucuronide) as substrate.

Induction experiments

For dexamethasone induction experiments, soil-grown plants were sprayed with 50 μm DEX or control solution (0.5% ethanol) and harvested after the indicated time. For plate experiments, 5 μm DEX or 0.05% ethanol was added to the medium.

For MeJA and wounding experiments, plants were sprayed with 500 μm MeJA or control solution (0.006% ethanol), and mechanically wounded with tweezers.

For herbivore treatment, eggs of Spodoptera littoralis were obtained from Syngenta Crop Protection AG (Stein, Switzerland). Larvae were kept on an artificial diet based on ground white beans at a temperature between 20 and 22°C. Three 1-week-old larvae of S. littoralis were placed on each rosette of A. thaliana and allowed to feed for 24 h. Plants were covered with ‘breathable’ plastic bags to prevent insects from escaping, and placed in a controlled environment chamber with a 16 h light/8 h dark photoperiod, 70% relative humidity, and a constant temperature of 22°C. Untreated control plants were kept under the same conditions. After 0.5, 1, 2, 4, 6, 8, 10, 24 and 72 h, six herbivore-damaged plants and six control plants were harvested for RNA extraction and GS analysis.

Anatomical analysis using light microscopy

To determine the number and area of epidermis cells, dental resin imprints were taken from the abaxial and adaxial leaf surfaces. Nail polish copies prepared from the imprints were analyzed by light microscopy. Data evaluation was based on imprints from five or six individual plants. For each plant, five separate fields of 0.31 mm2 were analyzed. For cross-sections, tissues were fixed in FAA solution (5% v/v acetic acid, 50% v/v ethanol, 3.7% v/v formaldehyde), and left overnight at room temperature. Dehydration was carried out through an ethanol series (50%, 70%, 90%, 95% and 100%) at room temperature with 2–4 h per step. After dehydration, the tissue was infiltrated for 2 h in a solution containing 50% v/v ethanol and 50% v/v Technovit 7100. The samples were then incubated overnight in a solution of 1% Hardener I in 100% Technovit 7100. Polymerization was carried out by adding Hardener II. Sections of 4 μm thickness were cut using histo-knives with a microtome and stained with 0.1% (w/v) toluidine blue. Dental resin imprints and leaf cross-sections were analyzed using the Olympus BX 41 System microscope (Olympus, Hamburg, Germany) and Meta Value software (Visitron Systems GmbH – Imaging Microscopy, Puchheim, Germany).

RNA gel blot analysis

Total RNA was prepared using TRIzol reagent (Gibco/BRL, Karlsruhe, Germany) according to the manufacturer's instructions. RNA gel blot analysis was performed as described by Gomez-Merino et al. (2004).

Expression profiling

Two 35S:OBP2 lines were analyzed using Affymetrix GeneChips representing approximately 8200 Arabidopsis genes. Total RNA was isolated from fully developed rosette leaves harvested from individual plants of two independent 35S:OBP2 lines, two different wild-type plants, and two independent empty-vector lines plants. Quality-checked RNA was sent to the German Resource Center for Genomic Research (RZPD, Berlin, Germany) for probe preparation and Arabidopsis GeneChip hybridization (Affymetrix, Santa Clara, CA, USA). Aliquots (20 μg) of total RNA were used for double-strand cDNA synthesis (SuperScript Choice system, Gibco/BRL). Biotin-labelled cRNA was synthesized using the BioArray High Yield RNA Transcript Labeling Kit (Enzo Life Sciences, Farmingdale, NY, USA). Affymetrix GeneChip experiments, including washing and scanning procedures, were performed as described in the Affymetrix technical manual. For each micro-array, overall intensity normalization for the entire probe set was performed using MicroArray Suite software 5.0. Chip files were generated using the program's default parameters. Pair-wise comparison between the files for the 35S:OBP2 lines, two different control lines, and two wild-type plants was carried out. Genes were considered as altered when the signal was more than twofold changed, had an appropriate ‘present’ call, and was defined as increased or decreased by the software. Only genes that met these criteria in all eight comparisons (35S:OBP2 line 8 × WT 18; 35S:OBP2 line 8 × WT 27; 35S:OBP2 line 8 × CTRL 11; 35S:OBP2 line 8 × CTRL 24; 35S:OBP2 line 10 × WT 18; 35S:OBP2 line 10 × WT 27; 35S:OBP2 line 10 × CTRL 11; 35S:OBP2 line 10 × CTRL 24; WT representing untransformed wild-type plants and CTRL representing vector-alone transformed transgenic lines) were considered further.

Quantitative real-time PCR

Total RNA was isolated using TRIzol reagent as described above. RNA (1 μg) was then reverse-transcribed with Superscript II reverse transcriptase (Invitrogen) in a reaction volume of 20 μl to generate first-strand cDNA. Every cDNA preparation was tested for contamination with genomic DNA by PCR with primers LEH–fw (5′-AACAGCAACAACAATGCAACTACTGATT-3′) and LEH–rev (5′-ACAAACAGAGACAAGAGACAAGACATGG-3′) that span an intron of the LATE ELONGATED HYPOCOTYL gene. Occasionally, genomic PCR products appeared after 35 amplification cycles; such samples were discarded. Real-time RT-PCR was performed with 1 μl of a 1:2 dilution of the first-strand cDNA reaction and SYBR Green reagent (Applied Biosystems, Foster City, CA, USA), in a 20 μl volume, on a Applied Biosystems Real Time PCR 7300 machine, with the following primer pairs: ACTIN2, actin-fw (5′-ATGGCTGAGGCTGATGATATTCAAC-3′) and actin-rev (5′-TACAAGGAGAGAAC AGCTTGGATG-3′); UBQ10, UBQ-fw (5′-ATGCAGATCTTTGTTAAGACTCTCAC-3′) and UBQ-rev (5′-ATAGTCTTTCCGGTGAGAGTCTTC-3′); OBP2-fw (5′-GCATCCGTTGGATCTTTGAGC-3′) and OBP2-rev (5′-AAAGCGTATAGCCCCGTCGTT-3′); CYP79B2-fw (5′-TTTGATGGATTGTCTGGCGC-3′) and CYP79B2-rev (5×-CAAAGACGAACAAGGCAACC-3′); CYP79B3-fw (5′-CGGTTTGTTTATCATCTCCGC-3′) and CYP79B3-rev (5′-TTGCTTACCGCTGATGAAATC-3′); CYP83B1-fw (5′ TCCGACCTTTTCCCTTATTTCG-3′) and CYP83B1-rev (5′TTGAGACGTGCACTGAGACCAG-3′); ATR1-fw (5′-CGGGTCTTAAGTAATTAGCC-3′); ATR1-rev (5′-AAGAAAGGAGCTTGGACTCC-3′); MAM1-fw (5-CATGTTGCTCTTCTGTGTCC-3′) and MAM1-rev (5′-ACATACCGAACAAGCTTCCC-3′).

Data were normalized to ACTIN2: nCt = Ctgene–CtACTIN2, where Ct refers to the number of cycles at which SYBR Green fluorescence in a PCR reaches an arbitrary value during the exponential phase of DNA amplification, set at 0.3 in all experiments, and then compared according to the formula: Cr (change in signal log ratio) = nCtcontrol– nCtsample. Use of the UBQ10 control gene did not affect the results; therefore, only the ACTIN2 data are presented.

Metabolite profiling

Leaves (60 mg fresh weight) of eight individual plants were frozen in liquid nitrogen and powdered in a Retsch mill. Metabolite extraction and measurement were performed as described previously (Roessner et al., 2000). For data analysis, a retention time and mass spectral library for automatic peak quantification of metabolite derivatives was implemented within the MASSLAB (ThermoQuest, Manchester, UK) method format. The t-tests were performed using the algorithm incorporated into Microsoft EXCEL (Microsoft Corporation, Seattle, WA, USA). The word ‘significant’ is used in the text when the change in question has been confirmed to be statistically significant (P < 0.01) with the t-test.

Glucosinolate measurements

Leaves were harvested, frozen in liquid nitrogen and lyophilized to dryness. GS analysis was performed as described previously (Brown et al., 2003), with the following modifications: metabolites were extracted with 80% (v/v) methanol, and 4-hydroxybenzyl-glucosinolate was used as an internal standard.

Hormone measurements

Leaves were harvested and frozen in liquid nitrogen. 2H5-indole-3-acetic acid was used as an internal standard. Samples were extracted with 5 ml of Bieleski solvent for 30 min at 70°C, and then were passed through C18-u columns (Phenomenex Ltd., Berlin, Germany), pre-conditioned with methanol and Bieleski solvent. Eluates were dried under vacuum, re-dissolved in diethylether and successively applied to a aminopropyl cartridge (pre-conditioned in diethylether), then washed with chloroform:2-propanol 2:1 (v/v) and eluted with 10% formic acid in diethylether (Müller et al., 2002). Resulting eluates were dried and re-dissolved in water at pH 2–3 (adjusted with HCl) and applied to ENV+ columns (pre-conditioned with water, pH 3). Elution was carried out with water (pH 7) followed by 0.35 m hydroxylamine in 60% methanol/water (v/v) (Dobrev and Kaminek, 2002). Combined eluates from both steps were evaporated, and then derivatized with N-methyl-N-(tert-butyldimethylsilyl)trifluoroacetamide (MTBSTFA) according to the method described by Birkemeyer et al. (2003). Derivatized samples were measured on an ion trap system (Saturn 2000; Varian Inc. PaloAlto, CA, USA) in EI-MRM-Modus using the following parameters: 30 m DB 5-MS fused-silica column with 0.25 mm inner diameter, 0.25 μm film thickness (Agilent Technologies, Palo Alto, CA, USA); injection temperature 230°C; injection volume 1 μl; splitless mode; carrier gas helium; flow 1 ml min−1; temperature program of 5 min at 70°C, then rising at 10°C min−1 to 200°C, at 15°C min−1 to 310°C, hold for 10 min; transfer line temperature 250°C; ion trap temperature 200°C; parent ions 232 234; excitation amplitude 0.60.

We also determined auxin content using an alternative method. Plant material was processed as described by Prinsen et al. (2000). After the methylation with diazomethane, the samples were dried under a nitrogen stream, dissolved in a mixture of 50 μl ethanol (70% v/v) and 450 μl of phosphate buffer (50 mm, pH 7.2) and subjected to an auxin-specific immunoaffinity extraction that was almost identical, in terms of its execution, to that for cytokinins (Novák et al., 2003). The final analysis was done by a competitive ELISA (Strnad, 1996) using anti-C1-IAA antibodies and IAA-alkaline phosphatase tracer. The assay was calibrated using [3H]IAA.

AGI code numbers

The AGI code numbers for the genes tested in this study are as follows: At1g07640 (OBP2); At4g39950 (CYP79B2); At2g22330 (CYP79B3); At4g31500 (CYP83B1/SUR2); At5g60890 (ATR1); At1g01060 (LATE ELONGATED HYPOCOTYL); At3g18780 (ACTIN2); At4g05320 (UBQ10); At5g44420 (PDF1.2); At5g23010 (MAM-1).

Acknowledgements

Funding of this research was provided by BMBF (GABI Program, FKZ 0312276M), BayerCrop Science, the Czech Academy of Sciences (grant number IBS5038351) and the Fonds der Chemischen Industrie (number 0164389). Further support was provided by the Interdisciplinary Research Centre ‘Advanced Protein Technologies’ (IZ-APT) of the University of Potsdam. A.S. thanks the Ernst Schering Foundation, Berlin, for providing a doctoral fellowship. She is also member of the International PhD Programme ‘Integrative Plant Science’ (IPP-IPS) funded by the DAAD (Deutscher Akademischer Austauschdienst) and the DFG (Deutsche Forschungsgemeinschaft) under grant number DAAD Az. D/04/01336. We thank Imre Somssich (MPI for Plant Breeding, Cologne, Germany) for providing the pJawohl8-RNAi vector, Gerd Jürgens (University of Tübingen) for providing the DR5–GFP construct, Bettina Sittlinger and Ursula Krause for their excellent technical support, Karin Koehl and her MPI Green Team for plant care, and Josef Bergstein for expert photography.

Ancillary