The transcription factor HIG1/MYB51 regulates indolic glucosinolate biosynthesis in Arabidopsis thaliana


For correspondence (fax + 49 221 4705039; e-mail


Glucosinolates are a class of plant secondary metabolites that serve as antiherbivore compounds in plant defence. A previously identified Arabidopsis thaliana activation-tagged line, displaying altered levels of secondary metabolites, was shown here to be affected in the content of indolic and aliphatic glucosinolates. The observed chemotype was caused by activation of the R2R3-MYB transcription factor gene HIG1 (HIGH INDOLIC GLUCOSINOLATE 1, also referred to as MYB51). HIG1/MYB51 was shown to activate promoters of indolic glucosinolate biosynthetic genes leading to increased accumulation of indolic glucosinolates. The corresponding loss-of-function mutant hig1-1 contained low levels of glucosinolates. Overexpression of the related transcription factor ATR1/MYB34, which had previously been described as a regulator of indolic glucosinolate and indole-3-acetic acid homeostasis, in the hig1-1 mutant background led to a partial rescue of the mutant chemotype along with a severe high-auxin growth phenotype. Overexpression of MYB122, another close homologue of HIG1/MYB51, did not rescue the hig1-1 chemotype, but caused a high-auxin phenotype and increased levels of indolic glucosinolates in the wild-type. By contrast, overexpression of HIG1/MYB51 resulted in the specific accumulation of indolic glucosinolates without affecting auxin metabolism and plant morphology. Mechanical stimuli such as touch or wounding transiently induced the expression of HIG1/MYB51 but not of ATR1/MYB34, and HIG1/MYB51 overexpression reduced insect herbivory as revealed by dual-choice assays with the generalist lepidopteran herbivore, Spodoptera exigua. We hypothesize that HIG1/MYB51 is a regulator of indolic glucosinolate biosynthesis that also controls responses to biotic challenges.


Glucosinolates are a class of nitrogen- and sulphur-containing plant secondary metabolites, which are mainly found in members of the Brassicaceae family including the model plant Arabidopsis thaliana (Rask et al., 2000; Kliebenstein et al., 2001; Mithen, 2001). Their role in plant defence against microorganisms (Mari et al., 1996; Manici et al., 2000) and herbivores (Kliebenstein et al., 2002a; Levy et al., 2005; Mewis et al., 2005), as well as their importance in the human diet as inducers of anticarcinogenic enzymes (Shapiro et al., 1998; Gross et al., 2000; Mithen et al., 2003; Grubb and Abel, 2006), attracted increased attention to this particular group of plant secondary metabolites.

Depending on the nature of the amino acid side chain, glucosinolates can be grouped into aliphatic, aromatic and indolic glucosinolates. The biosynthesis of glucosinolates comprises (i) chain elongation of the amino acids, (ii) formation of the glucosinolates by cytochrome P450 monooxygenases (CYPs), C-S lyases, S-glucosyltransferases and sulfotransferases, and (iii) secondary modifications (reviewed by Grubb and Abel, 2006).

The composition of glucosinolates can drastically vary in different ecotypes of A. thaliana (Kliebenstein et al., 2001; Reichelt et al., 2002), during ontogenesis (Petersen et al., 2002; Brown et al., 2003) or in response to environmental stimuli (Brader et al., 2001; Kliebenstein et al., 2002b; Mikkelsen et al., 2003; Bednarek et al., 2005; Mewis et al., 2005). Indolic glucosinolates are mainly found in vegetative parts of plants, rosette leaves and roots, whereas highest concentrations of aliphatic glucosinolates are found in seeds, flowers and siliques (Brown et al., 2003). Recent studies have shown that genes encoding enzymes of the indolic glucosinolate biosynthetic pathway form stable co-expression clusters, and group together with tryptophan biosynthetic genes in response to stress conditions (Gachon et al., 2005).

Synthesis of indolic glucosinolates in A. thaliana starts with the formation of indole-3-acetaldoxime (IAOx) from tryptophan, a step catalysed by CYP79B2/CYP79B3. IAOx can subsequently serve as a substrate for CYP83B1 resulting in the formation of 1-aci-nitro-2-indolyl-ethane, but is simultaneously a precursor for indole-3-acetic acid (IAA, auxin) biosynthesis. An interplay between the IAA and indolic glucosinolate biosynthetic pathways could be demonstrated by the analysis of several loss-of-function and gain-of-function mutants. For example, A. thaliana plants overexpressing CYP79B2 showed an increased accumulation of both indolic glucosinolates and IAA; conversely, the cyp79B2/cyp79B3 double mutant is impaired in the synthesis of indolic glucosinolates, but only partially in IAA biosynthesis, suggesting the existence of other IAA precursors besides IAOx (Zhao et al., 2002). Mutants defective in CYP83B1 (SUR2) show a high-IAA phenotype accompanied by reduced levels of indolic glucosinolates (Barlier et al., 2000; Bak and Feyereisen, 2001; Smolen and Bender, 2002). A total lack of indolic glucosinolates was not observed, presumably because CYP83A1, involved in the synthesis of aliphatic glucosinolates, can partially compensate for the defect in CYP83B1 (Naur et al., 2003). Interestingly, the cyp83B1 phenotype correlates with that of the ugt74B1 mutant, which is defective in the next and penultimate step of indolic glucosinolate biosynthesis: the formation of desulfo-glucosinolate (Grubb et al., 2004).

Although the functional role of several indolic glucosinolate biosynthetic pathway genes has been studied for quite some time, the first regulatory components for this pathway were identified only recently. A calmodulin-binding nuclear protein, IQD1, was shown to be a positive regulator of aliphatic and indolic glucosinolate formation (Levy et al., 2005). Overexpression of IQD1 resulted in increased levels of both indolic and aliphatic glucosinolates, thereby enhancing plant defence against generalist herbivores, whereas the levels of both types of glucosinolates were decreased in iqd loss-of-function mutants. Likewise, the R2R3-MYB transcription factor ATR1/MYB34, along with its known role in tryptophan biosynthesis (Bender and Fink, 1998), appears to control homeostasis between indolic glucosinolate and IAA biosyntheses (Celenza et al., 2005). Furthermore, a DNA-binding-with-one-finger (AtDof1.1) transcription factor has recently been shown to be part of the regulatory network controlling glucosinolate biosynthesis (Skirycz et al., 2006). Here, we present functional evidence that HIG1/MYB51 represents a transcriptional regulator of indolic glucosinolate biosynthesis genes in A. thaliana. Expression of HIG1/MYB51 responds to mechanical stimuli like touch or wounding, and overexpression of HIG1/MYB51 results in increased plant resistance against generalist herbivores.


Identification of the HIG1/MYB51 gene and properties of the HIG1 protein

The HIGH PHENOLIC COMPOUND1-1 DOMINANT (HPC1-1D) mutant was isolated in a previously described screen of the A. thaliana activation-tagged TAMARA population for mutant plants containing increased levels of phenolic compounds (Schneider et al., 2005). Although the HPC1-1D activation-tagged mutant did not show any informative growth phenotype, HPLC profiling of HPC1-1D revealed an increased accumulation of an unknown metabolite in the mutant by between four- and sixfold when compared with wild-type plants (Schneider et al., 2005). Using preparative and analytical HPLC analysis in combination with mass spectrometry, this compound was subsequently identified as desulfo-indolic glucosinolate or desulfo-glucobrassicin (indol-3-ylmethyl glucosinolate, see Figure S1). The HPC1-1D mutant was therefore renamed as HIG1-1D (HIGH INDOLIC GLUCOSINOLATE1-1 DOMINANT).

The insertion position of the dSpm-Act element in the HIG1-1D mutant has been localized to chromosome 1, approximately 1.4 kb upstream of At1g18570. Several genes encoding proteins of unknown or hypothetical functions are located adjacent to the insertion site of the non-autonomous recombinant transposon carrying the 4x35S enhancer. Analysis of mRNA levels by RT-PCR demonstrated an obvious overexpression of At1g18570 in the activation-tagged HIG1-1D mutant relative to the wild-type, whereas expression of the other adjacent gene At1g18560 encoding a transposase-related protein was neither detected in wild-type plants nor in the mutant line (Schneider et al., 2005). Steady-state mRNA expression levels of other neighbouring genes – At1g18590 (the desulfoglucosinolate sulfotransferase AtST5c; Piotrowski et al., 2004), At1g18580 (a glucosyltransferase family protein) and At1g18550 (a kinesin motor related protein) – were not affected in HIG1-1D mutant plants (not shown), corroborating our previous finding that the observed metabolic phenotype of the mutant is caused by ectopic overexpression of At1g18570.

The results summarized above demonstrate that At1g18570 is the gene responsible for the HIG1-1D phenotype. The gene product of At1g18570 is MYB51, a member of subgroup 12 of the large family of R2R3-MYB transcription factors (Figure 1). We refer to this gene as HIGH INDOLIC GLUCOSINOLATE 1 (HIG1/MYB51). The HIG1/MYB51 gene structure exhibiting three exons and two introns was confirmed by isolation and sequencing of the full-length cDNA. The predicted protein is 352 amino acid residues in length with a molecular mass of about 39 kDa and a calculated isoelectric point of 5.6. The sequence displays similarities in PFAM and PROSITE motifs (PS00037, PS50090, PS00334 and PF00249.18) to known nuclear localized proteins (as found at HIG1/MYB51 does not contain a typical nuclear localization signal (NLS) as revealed by PredictNLS (; Nair and Rost, 2005); however, an amino acid residue stretch KKRLIKK was detected (amino acid residues 114–120) that might act as a SV40-type NLS. Similar motifs are also present in sequences of other members of the HIG1/MYB51 subfamily (KKRLKQK and KKCLVKK in the case of ATR1/MYB34 and MYB122, respectively), suggesting that these motifs could actually direct the proteins to the nucleus.

Figure 1.

HIG1/MYB51 belongs to the subgroup 12 of the Arabidopsis thaliana R2R3-MYB gene family.
Amino acid sequences from the six members of R2R3-MYB factors assigned to subgroup 12 that share the common amino acid motif [L/F]LN[K/R]VA (Stracke et al., 2001) and the amino acid sequence of GL1 (Glabrous1) were downloaded from Genbank. The distance tree was constructed by aligning the sequences with CLUSTAW and DIALIGN-T, and by using PHYLIP with the BLOSUM62 matrix. The bootstrap support for all nodes is better than 87, except for the separation of MYB76 and MYB28.

To determine the subcellular localization of the HIG1/MYB51 transcription factor in plant cells, a translational fusion of the full-length HIG1/MYB51 coding sequence with the open reading frame encoding the GFP was generated. The localization of the fusion protein was analyzed in transfected BY2 tobacco protoplasts, as well as in cultured A. thaliana (Col-0) cells and in transiently transformed A. thaliana leaves. In all systems, a nuclear localization of the HIG1:GFP fusion protein could be demonstrated (Figure S2).

The glucosinolate contents of HIG1-1D plants, HIG1/MYB51 overexpression lines and the hig1-1 loss-of-function mutant

We constructed lines that overexpress HIG1/MYB51 under control of the CaMV 35S promoter in the wild-type background, and isolated a hig1 null mutant (GABI-Kat line 228B12, Col-0 background, harbouring a T-DNA insertion in the second exon of HIG1/MYB51; the mutant and the allele is referred to as hig1-1).

To analyze whether the glucosinolate levels are affected by changes in HIG1/MYB51 expression, the glucosinolate accumulation pattern of HIG1-1D was analyzed along with that of HIG1/MYB51 overexpression lines and that of the homozygous hig1-1 mutant, which does not express HIG1/MYB51 (see Figure 2a). Three out of ten representative overexpressing lines [35S:HIG1-3(wt), -7(wt) and -8(wt)] are shown in more detail together with HIG1-1D and hig1-1 (Figure 2b and c). As expected, indol-3-ylmethyl glucosinolate (I3M) was accumulated to sixfold higher levels in the HIG1-1D mutant (Figure 2b) than in the wild-type. Likewise, the overexpression lines contained increased levels of I3M by between three- and eightfold compared with the wild-type. There was also an increase in the other main indolic glucosinolate 4MOI3M (and 1MOI3M) level in both HIG1-1D and the overexpression lines compared with the wild-type.

Figure 2.

 Glucosinolate (GS) contents and steady state levels of HIG1/MYB51 mRNA transcripts in rosette leaves of 5-week-old wild-type plants, loss-of-function mutant hig1-1, 35S:HIG1 overexpressing lines and the activation-tagged line HIG1-1D.
(a) Semi-quantitative analysis of steady-state mRNA level of HIG1/MYB51 in wild-type (wt) (Col-0), hig1-1, three independent overexpression lines [35S:HIG1-3(wt), 35S:HIG1-7(wt) and 35S:HIG1-8(wt)], and HIG1-1D. Total RNA was prepared from rosette leaves of 5-week-old plants, and gene-specific primers for HIG1/MYB51 and Actin2 genes were used. Each PCR assay (22–26 cycles) was repeated three times with two independent sets of plants.
(b) Glucosinolate contents in the different lines (means ± SD, n = 3). GS was extracted from freeze-dried rosette leaves of 5-week-old plants. 3MSOP, 3-methylsulfinylpropyl-GS; 4MSOB, 4-methylsulfinylbutyl-GS; 5MSOP, 5-methylsulfinylpentyl-GS; 4MTB, 4-methylthiobutyl-GS; I3M, indol-3-yl-methyl-GS; 8MSOO, 8-methylsulfinyloctyl-GS; 4MOI3M, 4-methoxyindol-3-ylmethyl-GS; 1MOI3M, 1-methoxyindol-3-ylmethyl-GS.
(c) Content of bulk of glucosinolates in wild-type, hig1-1, overexpressing lines 35S:HIG1-3(wt), 35S:HIG1-7(wt) and 35S:HIG1-8(wt), and HIG1-1D.
*Significantly different (Student’s t test; P < 0.05) in comparison with wild-type.

Notably, the activation of indolic glucosinolate biosynthesis in both HIG1-1D and the overexpression lines was accompanied by the downregulation of the main aliphatic glucosinolate 4MSOB (Figure 2b). However, the total quantity of glucosinolates (aliphatic plus indolic) was increased in HIG1-1D and the overexpression lines up to threefold (Figure 2c), with HIG1/MYB51 transcript levels being correlated with the indolic glucosinolate levels in the respective lines (Figure 2a). Hence, the chemotype of 35S:HIG1 overexpression lines reproduced that of HIG1-1D. All overexpressing lines possessed an unchanged growth phenotype, as is the case for HIG1-1D and the hig1-1 mutant (see below).

Figure 5.

 Growth phenotypes of wild-type plants and hig1-1 mutants expressing ATR1/MYB34 or MYB122 constructs, and determination of corresponding auxin levels.
(a) Different independent transgenic lines are shown: the abbreviation given in brackets indicates the genetic background, i.e either wild-type (wt) or the hig1-1 background (m). Upper row, from the left to the right: HIG1-1D, 35S:HIG1, 35S:ATR1-17(wt), 35S:ATR1-1(wt) [lines harbouring the construct in the Col-0 background exhibiting a high indole-3-acetic acid (IAA) phenotype], 35S:ATR1(wt) (line harbouring the 35S:ATR1 construct in the Col-0 background exhibiting a strong growth phenotype), 35S:ATR1-22(m) and 35S:ATR1-24(m) (lines harbouring the 35S:ATR1 construct in the hig1-1 background exhibiting a high-IAA phenotype). Lower row from the left to the right: wild-type (Col-0), hig1-1, 35S:MYB122-6(wt) and 35S:MYB122-16(wt) (lines harbouring the 35S:MYB122 construct in the Col-0 background exhibiting a high-IAA phenotype and a strong growth phenotype), 35S:MYB122-8(m) and 35S:MYB122-11(m) (lines harbouring the 35S:MYB122 construct in the hig1-1 background [no IAA phenotype]). Inset: defective flowers and siliques peculiar for 35S:ATR1 overexpressing plants, independent of the genetic background, and for 35S:MYB122 in the wild-type background.
(b) Free indole-3-acetic acid (IAA) concentrations in leaves of wild-type, HIG1-1D, hig1-1, ATR1/MYB34 and MYB122 overexpression plants. 35S:ATR1-17(wt) and 35S:ATR1-1(wt), ATR1/MYB34 overexpressed in the wild-type background; 35S:ATR1-22(m) and 35S:ATR1-24(m), ATR1/MYB34 overexpressed in the hig1-1 background; 35S:MYB122-16(wt) and 35S:MYB122-6(wt), MYB122 overexpressed in the wild-type background; 35S:MYB122-8(m) and 35S:MYB122-11(m), MYB122 overexpressed in the hig1-1 background. Data are mean values ± SD of three individual experiments (n = 3). *Significantly different (Student’s t test; P < 0.05) in comparison with the wild-type. For details, see Experimental procedures.

As also shown in Figure 2, the accumulation of tryptophan- and methionine-derived glucosinolates was decreased in hig1-1 (see 4MSOB and I3 M in Figure 2b), and the total glucosinolate content was reduced to almost one third compared with the wild-type. Plants that overexpress HIG1/MYB51 in the background of hig1-1 are indistinguishable from HIG1/MYB51 overexpression lines in the Col-0 background with respect to all analysed chemotypic and growth phenotype parameters (Figure S3).

HIG1/MYB51 trans-activatesthe expression of indolic glucosinolate biosynthetic pathway genes

We analyzed the ability of HIG1/MYB51 to activate promoters of genes involved in the indolic glucosinolate biosynthetic pathway (Figure 3) using co-transformation assays. A. thaliana leaves were infiltrated with (i) a supervirulent Agrobacteria strain LBA4404.pBBR1MCS.virGN54D (kindly provided by Dr Memelink, University of Leiden) carrying the 35S:HIG1 construct for effector expression and/or (ii) one of several different reporter constructs containing the uidA (GUS) gene driven by the promoter of putative target genes (DHS1, ASA1, TSB1, CYP79B2, CYP79B3, CYP83B1 and AtST5a). Leaves transiently expressing only the GUS reporter constructs fused to promoters of target genes showed only weak GUS activity, whereas co-transformation with 35S:HIG1 led to a significant increase in GUS expression demonstrating the potential of HIG1/MYB51 to trans-activate the respective promoters (Figure 4). All tested promoters except for ASA1 were activated by HIG1/MYB51 indicating that HIG1/MYB51 positively regulates indolic glucosinolate biosynthetic pathway genes.

Figure 3.

 Proposed pathways for indolic glucosinolate and indole-3-acetic acid (IAA) biosyntheses.
Targets genes of HIG1/MYB51 are shown in bold. Proposed IAA synthesis from the indolic glucosinolate indol-3-ylmethyl glucosinolate is indicated by a dotted arrow (Pollmann et al., 2002). IAN, indole-3-acetonitrile.

Figure 4.

HIG1/MYB51 activates glucosinolate biosynthetic pathway genes. Co-transformation assays to determine the target-gene specificity for HIG1/MYB51 (effector) towards target promoters of indolic biosynthetic pathway genes (means of GUS activity in nmol methylumbelliferone (MU) per min and mg protein ± SD, n = 5). The promoters of DHS1, ASA1, TSB1, CYP79B2, CYP79B3, CYP83B1 and AtST5a genes were fused to the uidA (GUS) reporter gene (TargetPromoter:GUS vectors). Fully expanded leaves of Arabidopsis thaliana plants were infiltrated with the supervirulent Agrobacterium strain LBA4404.pBBR1MCS.virGN54D containing either the reporter construct (TargetPromoter:GUS:pGWB3i) and a ‘null’ effector (empty vector without HIG1/MYB51, 35S:pGWB2) or the target gene fused to the reporter (TargetPromoter:GUS:pGWB3i) and, in addition, the HIG1/MYB51 effector (35S:HIG1:pGWB2). Black bars represent expression of only the TargetPromoter:GUS constructs with ‘null’ effector, grey bars represent the expression of TargetPromoter:GUS constructs co-transformed with 35S:HIG1.
*Significantly different (Student’s t test; P < 0.05) in comparison with the ‘null vector’ control.

Overexpression of ATR1/MYB34 and MYB122 in wild-type and the hig1-1 mutant: analysis of growth phenotypes and auxin contents

As evident from Figure 1, HIG1/MYB51 is a member of a group of related R2R3-MYB genes (subgroup 12), with MYB122 as the closest and ATR1/MYB34 as the next homologue. MYB122 has to our knowledge not yet been functionally analyzed, whereas ATR1/MYB34 has been implicated in the regulation of indolic glucosinolate biosynthesis. Therefore, we studied the effect of ectopic overexpression of ATR1/MYB34 and MYB122, respectively, in both hig1-1 mutant (m) and wild-type (wt) genetic backgrounds. Both coding sequences were driven by the CaMV 35S promoter.

Interestingly, all lines overexpressing ATR1/MYB34 in the hig1-1 mutant background [from 70 primary transformants, nine lines were analyzed in detail, two of which, 35S:ATR1-22(m) and 35S:ATR1-24(m), are presented in Figure 5a] showed highly retarded shoot and root growth, curly leaves, a bushy stature and were unable to produce seeds. This phenotype is clearly indicative of high IAA levels in the plants.

Ectopic overexpression of ATR1/MYB34 in wild-type background [from 70 primary transformants, eight lines were analyzed in detail, three of which, 35S:ATR1-17(wt), 35S:ATR1-1(wt) and 35S:ATR1(wt) (not further analyzed), are presented in Figure 5a] led to a similar high-IAA phenotype. Generally, the expression level of the transgene in both genetic backgrounds correlated with the strength of the high-IAA phenotype. Anyway, most of the flowers of 35S:ATR1 overexpressing plants stayed closed until senescence or developed tiny mal-developed siliques (Figure 5a, inset).

In contrast to plants overexpressing ATR1/MYB34 in hig1-1 or wild-type background, overexpression of MYB122 in the hig1-1 mutant background [from 70 primary transformants, eight lines were analyzed in detail, two of which, 35S:MYB122-8(m) and 35S:MYB122-11(m), are presented in Figure 5a] did not result in an aberrant growth phenotype. However, in the wild-type background (Col-0), several MYB122 overexpressing lines {from 70 primary transformants, seven lines were analyzed in detail, two representative lines with either a moderate [35S:MYB122-6(wt)] or a strong [35S:MYB122-16(wt)] phenotype are shown in Figure 5a} displayed a high-IAA phenotype, even though this was less pronounced than in the case for ATR1/MYB34 overexpressing lines. 35S:MYB122-16(wt) plants were retarded in growth: most of the flowers stayed undeveloped and were almost unable to produce healthy siliques, albeit some seeds could be collected for further propagation.

We also measured the levels of free IAA in these different lines (Figure 5b). Although the IAA levels in the activation-tagged HIG1-1D mutant and in hig1-1 did not significantly differ from that in the wild-type, the ATR1/MYB34 overexpression lines possessed threefold [35S:ATR1-17(wt), moderate phenotype in the wild-type background] and up to sevenfold higher IAA levels in plants with a more severe growth phenotype [35S:ATR1-1(wt), 35S:ATR1-22(m) and 35S:ATR1-24(m)]. Interestingly, auxin levels of MYB122 overexpression lines in the hig1-1 background remained unchanged, but increased by between two- and fivefold when MYB122 was overexpressed in the wild-type background. Thus, the levels of free IAA nicely corresponded to the observed growth phenotypes: none of the HIG1/MYB51 overexpressing lines showed an aberrant growth phenotype neither in the wild-type nor in the hig1-1 background (Figure 5a; Figure S3). By contrast, all ATR1/MYB34 overexpression lines and strong MYB122 overexpression lines in the wild-type background showed a high-IAA phenotype.

It has been previously suggested that nitrilases could be involved in the production of additional IAA under particular physiological conditions (Bartel and Fink, 1994; Grsic-Rausch et al., 2000; Pollmann et al., 2006). Nitrilases convert indole-3-acetonitrile (IAN), produced from the indolic glucosinolate I3M, into IAA (Figure 3). The observation that the expression of the NITRILASE 2 gene is dramatically increased in ATR1/MYB34 overexpression lines, but not in HIG1/MYB51 and MYB122 overexpression lines, might shed new light on the high-IAA chemotype of ATR1/MYB34 overexpressing lines (Figure S4).

Overexpression of ATR1/MYB34 and MYB122 in wild-type and the hig1-1 mutant: analysis of chemotypes and indolic glucosinolate pathway genes

We also analyzed if the hig1-1 mutant chemotype characterized by lower contents of glucosinolates could be rescued by overexpression of ATR1/MYB34 or MYB122. Results are shown in Figure 6a for the two ATR1/MYB34 overexpressing lines in the hig1-1 background [35S:ATR1-22(m) and 35S:ATR1-24(m)] and, for comparison, in the wild-type background [35S:ATR1-1(wt) and 35S:ATR1-17(wt)], respectively. All lines except for 35S:ATR1-17(wt) showed higher contents of both the main indolic and aliphatic glucosinolates I3M and 4MSOB compared with the hig1-1 mutant. The levels of I3M are also up to 1.7-fold higher compared with the wild-type, but clearly lower compared with the overexpressing line HIG1-1D (which shows a sixfold increase in the I3M level). The level of 4MOI3M, which is enhanced in HIG1-1D, was not affected in ATR1/MYB34 overexpression lines. It should be noted that the level of aliphatic glucosinolates (e.g. 4MSOB) are lower in HIG1-1D compared with the wild-type (see Figure 2).

Figure 6.

 Glucosinolate contents and transcript levels of indolic glucosinolate pathway genes in rosette leaves of 5-week-old ATR1/MYB34 and MYB122 overexpression plants.
(a), (b) Glucosinolate contents (GS) in hig1-1, ATR1/MYB34 and MYB122 overexpressing lines. For comparison, the glucosinolate contents in the wild-type and HIG1-1D are also shown. 35S:ATR1-1(wt) and 35S:ATR1-17(wt), wild-type background; 35S:ATR1-22(m) and 35S:ATR1-24(m), hig1-1 background. 35S:MYB122-16(wt) and 35S:MYB122-6(wt), wild-type background; 35S:MYB122-8(m) and 35S:MYB122-11(m), hig1-1 background. 3MSOP, 3-methylsulfinylpropyl-GS; 4MSOB, 4-methylsulfinylbutyl-GS; 5MSOP, 5-methylsulfinylpentyl-GS; 4MTB, 4-methylthiobutyl–GS; I3M, indol-3-yl-methyl-GS; 8MSOO, 8-methylsulfinyloctyl-GS; 4MOI3M, 4-methoxyindol-3-ylmethyl-GS; 1MOI3M, 1-methoxyindol-3-ylmethyl-GS. Means ± SD, n = 3.
(c), (d) Real-Time RT-PCR analysis of indolic glucosinolate pathway genes measured in the overexpression lines 35S:ATR1-1(wt), 35S:ATR1-22/24(m), 35S:MYB122-16(wt) and 35S:MYB122-8/11(m). For comparison, the gene expression levels in hig1-1 and HIG1-1D are also presented. Relative gene expression values in comparison with the wild-type are shown (WT = 1). Means ± SD, n = 3). For details, see Experimental procedures.
*Significantly different (Student’s t test; P < 0.05) in comparison with the wild-type.

Figure 6b shows that the low indolic glucosinolate chemotype of the hig1-1 mutant could not be converted towards the wild-type phenotype by overexpression of MYB122 [lines 35S:MYB122-8(m) and 35S:MYB122-11(m)]. However, strong overexpression of MYB122 in the wild-type background [line 35S:MYB122-16(wt)] led to a significant increase in I3M levels.

We also addressed the question of whether or not the recorded metabolite and IAA profiles are supported by expression profiling data of indolic glucosinolate and IAA biosynthetic genes. As shown in Figure 6c and d, transcript levels of ASA1, TSB1, CYP79B2, CYP79B3 and CYP83B1 were upregulated in all ATR1/MYB34 overexpression lines, independent of the genetic background, and in MYB122 overexpression lines in the wild-type background [35S:MYB122-16(wt)]. In the hig1-1 background [lines 35S:MYB122-8(m) and 35S:MYB122-11(m)], MYB122, although expressed to high levels, was obviously not able to upregulate any of the target genes that could be activated in the wild-type background. Notably, expression levels of indolic glucosinolate biosynthetic pathway genes further downstream of CYP83B1, i.e. UGT74B1 and AtST5a, were not significantly affected, neither by ATR1/MYB34 nor by MYB122 overexpression.

Taken together, all genes of indolic glucosinolate biosynthesis (from TSB1 to the last enzyme AtST5a) are activated by HIG1/MYB51 (Figures 4 and 6c,d). ATR1/MYB34 and MYB122 can also positively regulate glucosinolate biosynthesis genes, but overexpression of ATR1/MYB34, independently of the genetic background, led to a high-IAA phenotype, as is the case for overexpression of MYB122 in the wild-type background. This might be caused by the accumulation of IAOx, which could be shuttled into alternative metabolic pathways such as the biosynthesis of IAA.

Tissue-specific expression of HIG1/MYB51

To study the tissue-specific expression of HIG1/MYB51, the promoter region (from –1676 to + 342 bp), including the first intron and the first exon of the HIG1/MYB51 coding sequence, was used to generate a translational fusion with the uidA (GUS) reporter gene. This construct was stably transferred to A. thaliana plants by Agrobacterium tumefaciens-mediated transformation. Of the 15 transgenic ProHIG1:uidA lines that were analyzed in detail, 12 lines showed a similar tissue-specific pattern of HIG1/MYB51 expression. Histochemical analysis of GUS activity revealed a strong reporter gene expression already in the hypocotyl of 3-day-old seedlings (Figure 7). Although the cotyledons showed no GUS activity at this stage, they were clearly stained at a later stage of ontogenesis. Furthermore, GUS activity was detected in the roots of young seedlings. In newly emerging rosette leaves, GUS staining was only faint and primarily visible around the midvein and at the base of trichomes. Reporter gene activity increased gradually in expanding leaves, reaching a maximum in fully expanded leaves, where the mesophyll tissue also showed GUS staining. During senescence, GUS expression disappeared almost completely, but was still detectable in the vasculature. Furthermore, expression was present in primary and lateral roots of plants at the rosette stage. In the floral tissue, GUS activity was only detected in young flowers but not at later stages of development. Siliques showed very faint staining in an early stage, and GUS activity was restricted to the abscission zone in fully developed siliques.

Figure 7.

 Histochemical GUS staining in tissues of ProHIG1:GUS plants:
(a) 3-day-old seedling;
(b) 14-day-old seedling;
(c) 3-week-old plant;
(d) 5-week-old plant;
(e) adult leaves (GUS expression in midvein and trichomes);
(f) roots of adult plants;
(g) fully expanded leaf;
(h) immature flower;
(i) mature flower;
(j) developing and mature siliques.

Overall,HIG1/MYB51 promoter activity was strongest in the vegetative parts of the plants, mainly in mature rosette leaves. HIG1/MYB51 expression can also be observed in roots, but not in mature flowers or siliques. By contrast, analysis of ProATR1:uidA plants revealed that ATR1/MYB34 is mainly expressed in meristematic tissues and young flowers, but not in leaves, which are the main site of indolic glucosinolate biosynthesis and accumulation (Figure S5). Notably, the tissue-specific expression data for both HIG1/MYB51 and ATR1/MYB34 are consistent with AtGenExpress data from the Genevestigator microarray database (Zimmermann et al., 2004;

HIG1/MYB51 expression is induced by mechanical stimuli like touch and wounding

Glucosinolates play a major role in plant defence against herbivore attack, normally accompanied by wounding. Because HIG1/MYB51 is potentially involved in the control of indolic glucosinolate biosynthesis, HIG1/MYB51 expression was analyzed in response to mechanical stimuli like wounding or touch. As shown in Figure 8a, mechanical puncturing already induced an increase in HIG1/MYB51 gene expression in the wounded rosette leaves within 10 min, reaching a maximum after 30 min and fading away after 1 h, suggesting a transient response of HIG1/MYB51 to mechanical stimuli. Remarkably, expression of ATR1/MYB34 did not respond at all to wounding, and the induction of MYB122 expression was observed only after a delay of about 1 h.

Figure 8.

 Induction of HIG1/MYB51 expression by mechanical stimuli such as touch and wounding in fully expanded leaves, senescent leaves and siliques.
(a) Real-time RT-PCR analysis of HIG1/MYB51, ATR1/MYB34 and MYB122 gene expression upon wounding. Rosette leaves of 4–5-week-old plants were punctured and harvested at the indicated time points. Total RNA was reverse transcribed into cDNA and used as a template for quantitative RT-PCR as described in Experimental procedures; Actin2 primers were used as control. Relative gene expression values are shown compared with non-wounded leaves (0 min = 1). Data are mean values ± SD of three individual experiments (n = 3).
(b to d) Induction of GUS expression in ProHIG1:GUS plants at the cutting sites.
(e to g) Induction of GUS expression in ProHIG1:GUS plants at the touch site. For details, see Experimental procedures.
(e) Untouched fully expanded leaf.
(f), (g) Different touch sites in fully expanded leaf.

The activity of the HIG1/MYB51 promoter in response to touch was analyzed in more detail using the HIG1/MYB51-promoter GUS reporter lines. As shown in Figure 8 (b–g), adult rosette leaves showed strong GUS staining at the position of the touch stimulus. Furthermore, we could observe intense GUS staining at the cutting sites of samples, including leaves and stems. These data suggest that HIG1/MYB51 expression is induced in response to mechanical stimuli and might respond in a similar way to herbivore attack.

HIG1/MYB51 overexpression reduces leaf consumption by a generalist herbivore

The breakdown of glucosinolates by myrosinase upon tissue damage is known as a defence mechanism against herbivores (Rask et al., 2000; Mithen, 2001; Martin and Müller, 2007). As HIG1/MYB51 expression is induced by mechanical stimuli and HIG1/MYB51 overexpressing lines accumulate high levels of indolic glucosinolates, one can speculate about the role of HIG1/MYB51 in the plant defence system. To test this hypothesis, we assayed the consumption preference of the generalist lepidopteran herbivore Spodoptera exigua (Lepidoptera: Noctuidae) in dual-choice assays, measuring the leaf area consumed by fourth-instar larvae (n = 20 for each assay) maintained on an artificial diet before testing. When offering wild-type leaves and leaves of the hig1-1 mutant, there was no feeding preference towards one of the lines, i.e. similar quantities of leaves from both lines were consumed (Figure 9). However, when leaves of the HIG1-1D overexpressor were offered together with wild-type or hig1-1 mutant leaves, the larvae consumed significantly lower quantities of high indolic glucosinolate leaves (HIG1-1D) than of the other leaves offered. This clearly shows that HIG1/MYB51 overexpression linked to a high indolic glucosinolate chemotype can increase plant resistance against generalist herbivores, and suggests a role of HIG1/MYB51 in plant defence against biotic challenges.

Figure 9.

 Dual-choice assay with the generalist herbivore Spodoptera exigua. The graph shows the preference (mean leaf consumption in FW ± SD) of larvae for leaves from either wild-type (Col-0) and hig1-1 loss-of-function mutant (n = 20; P = 0.757) (upper bar), HIG1-1D gain-of-function and hig1-1 loss-of function mutants (n = 20; P = 0.0113) (middle bar) and HIG1-1D gain-of-function and wild-type (n = 20; P = 0.0026) (lower bar). P values calculated using Student’s t tests show significant avoidance of the high-indolic glucosinolate line HIG1-1D.


HIG1/MYB51 is a regulator of glucosinolate biosynthetic pathways

Substantial progress has been achieved in the understanding of glucosinolate biosynthesis using forward and reverse genetics in the model plant A. thaliana. However, little is known about underlying regulatory mechanisms and genes that coordinate and control glucosinolate compound accumulation during development and in response to environmental challenges. Recently, the first regulatory factors involved in the control of glucosinolate biosynthesis, including IQD1 were described (Levy et al., 2005; see Introduction). In addition, the R2R-MYB transcription factor ATR1/MYB34, which is known to activate tryptophan biosynthetic genes (Bender and Fink, 1998), has been reported to act as a regulator of indolic glucosinolate homeostasis (Celenza et al., 2005).

Here, we present HIG1/MYB51 as a new regulator of glucosinolate biosynthesis identified as an activation-tagged mutant (HPC1-1D, HIG1-1D), that was initially isolated in a screen for mutants with altered accumulation of secondary metabolites (Schneider et al., 2005). The chemotype caused by the dominant HIG1-1D allele, namely the accumulation of I3M, was shown to be caused by 4x35S enhancer-mediated overexpression of the R2R3-MYB transcription factor HIG1/MYB51. 4x35S enhancer elements have been demonstrated to increase quantitatively the original expression pattern of a gene, and not to lead to constitutive overexpression (Neff et al., 1999; van der Graaff et al., 2003). Remarkably, ectopic overexpression of HIG1/MYB51 in the wild-type background could phenocopy the chemotype of the original activation-tagged line HIG1-1D (Figure 2). On the other hand, a T-DNA insertion allele of HIG1/MYB51 (hig1-1) caused a significant reduction in I3M levels, and this chemotype could be counteracted by overexpression of HIG1/MYB51 (Figure S3). Thus, HIG1/MYB51 represents an important, but certainly not the only, regulatory component controlling indolic glucosinolate biosynthesis in A. thaliana. Notably, neither the hig1-1 mutant nor the HIG1-1D or HIG1/MYB51 overexpressing plants showed significant changes in IAA levels or aberrant growth phenotypes indicative for distorted auxin levels (Figure 5).

We also demonstrate that genes involved in the biosynthesis of tryptophan and indolic glucosinolate biosynthesis, i.e. DHS1, TSB1, CYP79B2, CYP79B3, CYP83B1 and AtST5a, are activated by HIG1/MYB51 in trans (Figure 4). Thus, HIG1/MYB51 appears to serve as a general activator of both tryptophan synthesis genes and tryptophan secondary metabolism genes. A co-regulation of indolic glucosinolate pathway enzymes and tryptophan biosynthetic enzymes has been observed by analyzing microarray data derived from different stress experiments (Gachon et al., 2005). Hence, the activation of the tryptophan biosynthetic pathway seems to be required for providing sufficient levels of the precursor tryptophan for the increased formation of indolic glucosinolates in response to stress.

The level of the main short-chain aliphatic glucosinolate 4MSOB is lower in HIG1-1D and in HIG1/MYB51 overexpression lines compared with the wild-type (Figures 2 and 6). It thus appears that HIG1/MYB51 possesses opposite effects on the biosynthetic pathways of indolic and aliphatic glucosinolates. One may speculate that the decreased accumulation of methionine-derived glucosinolates may result from a metabolic crosstalk between both branches of glucosinolate biosynthesis, by which a distinct ratio of the different glucosinolates is maintained. The ‘limiting electron hypothesis’, which has been recently proposed, could serve as an explanation for these observations (Grubb and Abel, 2006): the competition of cytochrome P450 monooxygenases involved in aliphatic (CYP79F1, CYP79F2 and CYP83A1) and indolic glucosinolate (CYP79B2, CYP79B3 and CYP83B1) biosynthetic pathways for electrons to reduce dioxygen to water could be the reason for a reciprocal negative feedback regulation between both branches of glucosinolate biosynthesis.

Strikingly, the hig1-1 null allele caused not only reduced levels of indolic glucosinolates, but also of short-chain aliphatic glucosinolates, resulting in a reduction of the total glucosinolate content by approximately 65% (Figure 2). On the other hand, HIG1/MYB51 overexpression in both the wild-type and the hig1-1 mutant background led to the expected increase in the level of indolic glucosinolates, with aliphatic glucosinolates remaining at low levels (Figure 2; Figure S3). The reason for this is not known yet but might point to a regulatory interplay between HIG1/MYB51 and other factors involved in the control of both indolic and aliphatic glucosinolate biosynthetic genes. It is feasible, for example, that the activity of putative regulators of aliphatic glucosinolate biosynthesis and corresponding biosynthetic genes is altered in the hig1-1 mutant.

Roles of HIG1/MYB51, ATR1/MYB34 and MYB122 in the regulation of glucosinolate biosynthesis are divergent

A. thaliana mutants defective in the early steps of the glucosinolate biosynthetic pathway, or overexpressing these enzymes, all show an altered morphology reminiscent of IAA overexpression or IAA repression phenotypes, respectively (Zhao et al., 2002; Barlier et al., 2000; Bak et al., 2001; Reintanz et al., 2001). These observations can be attributed to the fact that IAOx as a product of CYP79B2 and CYP79B3 actions is an important intermediate of both the indolic glucosinolate and IAA pathways (Figure 3). Overexpression of the transcription factor AtDof1.1 causes changes in glucosinolate profiles, but also affects auxin accumulation (Skirycz et al., 2006). Overexpression of ATR1/MYB34 was shown earlier to result in a hyperactivation of indolic glucosinolate pathway genes, accompanied by only a modest elevation of IAA levels with no impact on plant growth (Celenza et al., 2005).

The question of how glucosinolate biosynthesis is regulated (for example, in response to biotic stresses) without a strong impact on auxin biosynthesis is still open. It does not appear implicitly favourable to plants to alter IAA biosynthesis when an increased biosynthesis of indolic glucosinolates is required. Our data demonstrate that even though both IAA and indolic glucosinolate biosynthetic pathways share common enzymes, both pathways can be specifically regulated to allow appropriate plant responses to environmental challenges. HIG1/MYB51 can activate genes both upstream and downstream of IAOx, leading to increased indolic glucosinolate levels without a significant effect on IAA contents (Figures 2, 5b and 6). Even though ATR1/MYB34 can partially rescue the low-indolic glucosinolate hig1-1 mutant chemotype, overexpression of ATR1/MYB34 in the hig1-1 or wild-type background in our hands led to up to sevenfold higher IAA levels, reflected by pronounced high-IAA growth phenotypes (Figure 5a). In addition, all ATR1/MYB34 overexpressing plants were defective in the development of generative organs, the root system, as well as the vegetative biomass. In contrast, HIG1/MYB51 overexpressors and hig1-1 plants display unaltered morphology compared with the wild-type, and also the IAA levels of these plants are, if at all, only slightly altered (Figure 5).

Remarkably, ASA1 is not trans-activated by HIG1/MYB51 (Figure 4), and the upregulation of ASA1 therefore does not seem to be required for an increased biosynthesis of indolic glucosinolates. However, ASA1 appears to play a pivotal role in the regulation of auxin biosynthesis. For example, the ethylene-triggered activation of auxin biosynthesis was recently shown to be strictly dependent on ASA1. Likewise, the high-IAA phenotype of sur1 and sur2 mutants with defects in C-S lyase and CYP83B1 functions, respectively, is suppressed in mutants defective in ASA1/WEI2 function (Stepanova et al., 2005). Also, these results support the hypothesis that ATR1/MYB34 overexpression, which does cause increased steady-state levels of ASA1 transcripts (Celenza et al., 2005; Figure 6c), might primarily be linked to the regulation of IAA homeostasis.

In addition, the observation that ATR1/MYB34 does not show any expression in vegetative parts of the plant, e.g. rosette leaves, argues against a major role of ATR1/MYB34 in the control of indolic glucosinolate biosynthesis in leaves of adult plants (Figure S5; AtGenExpress data analyzed using Genevestigator,

In contrast to ATR1/MYB34, MYB122 could not rescue the low-indolic glucosinolate chemotype of the hig1-1 mutant (Figure 6b). Likewise, MYB122 overexpression did not cause a high-IAA phenotype in hig1-1 mutants, which raises the question about the role of MYB122 in the tryptophan pathway. Overexpression of MYB122 in the wild-type background, however, led to the enhanced transcription of several tryptophan pathway genes, including ASA1, TSB1, CYP79B2, CYP79B3 and CYP83B1, along with elevated auxin levels (Figures 5b and 6). A number of scenarios may explain this result. Increased IAA production could cause a positive feedback on the expression of these genes. Alternatively, MYB122 could directly enhance the expression of these genes, although only in concert with HIG1/MYB51. In addition, or alternatively, HIG1/MYB51 and/or MYB122 could act as activators or repressors, depending on the context, in a similar way as it has been shown for MYB4 (Jin et al., 2000).

It can be concluded that ATR1/MYB34, HIG1/MYB51 and MYB122 apparently have different roles in the regulation of indolic glucosinolate and IAA biosynthetic pathways. All three factors have the potential to upregulate glucosinolate biosynthetic pathway genes, e.g. TSB1, CYP79B2 and CYP79B3, and can positively regulate indolic glucosinolate accumulation. However, along with increased levels of indolic glucosinolates, overexpression of ATR1/MYB34 and MYB122 led to high-IAA phenotypes. Only the overexpression of HIG1/MYB51 led additionally to the activation of genes further downstream of CYP83B1, i.e. of UGT74B1 and AtST5a. These plants did not exhibit an aberrant growth phenotype. Obviously, the next step is a detailed analysis of the interplay of subgroup 12 R2R3-MYB factors, and potentially other factors, in controlling distinct but partially overlapping sets of target genes from IAA and glucosinolate biosynthetic pathways.

HIG1/MYB51 is expressed at sites of indolic glucosinolate accumulation and plays a role in biotic stress responses

The expression of glucosinolate biosynthetic genes in A. thaliana, e.g. that of CYP79B2, UGT74B1, CYP79F1 and CYP79F2, and also of recently discovered regulators of glucosinolate biosynthesis, IQD1 and AtDof1.1, often appeared to be restricted to vascular tissues (Mikkelsen et al., 2000; Grubb et al., 2004; Reintanz et al., 2001; Levy et al., 2005; Skirycz et al., 2006). HIG1/MYB51 expression overlaps with the expression of these genes but, in addition, is also present in the mesophyll of mature rosette leaves, the pavement cells of young rosette leaves and trichomes (Figure 7). Thus, HIG1/MYB51 expression correlates at least to a great extent with the sites of indolic glucosinolates biosynthesis and accumulation, an observation that is in accordance with the role of HIG1/MYB51 as a positive regulator of this pathway.

Environmental stimuli such as herbivore attack or wounding are known to have a great impact on the regulation of glucosinolate biosynthesis. Several glucosinolate biosynthetic pathway genes were shown to be induced upon mechanical stimuli or hormone treatment (Brader et al., 2001; Kliebenstein et al., 2002b; Mikkelsen et al., 2003; Mewis et al., 2005). Here we show that expression of HIG1/MYB51, but not of ATR1/MYB34, in leaves is rapidly induced by wounding or touch (Figure 8). It is tempting to speculate that HIG1/MYB51 is a key player in the signal transduction chain leading from the touch perception output to an increased biosynthesis of indolic glucosinolates, thereby rendering the plant more resistant to herbivores. It could indeed be shown that generalist herbivores avoided the HIG1-1D line with higher contents of indolic glucosinolates in dual-choice assays (Figure 9). Furthermore, QTL (quantitative trait locus) mapping analyses in A. thaliana (Kliebenstein et al., 2002a) provided data concerning a QTL controlling herbivore resistance in a Ler × Col population, which could be mapped on chromosome I between 14 and 28 cM. This QTL nicely fits to the position of the locus At1g18570 encoding HIG1/MYB51. This finding adds further evidence to the assignment of HIG1/MYB51 as an important regulatory component in controlling glucosinolate biosynthesis upon biotic challenges.

Until now it was assumed that the common pathway for the biosynthesis of both IAA and indolic glucosinolates in plants from the Brassicaceae family is simultaneously regulated by ATR1/MYB34. As shown in this study, HIG1/MYB51 obviously represents a key component in controlling the mechanical-induced regulation of the indolic glucosinolate biosynthetic pathway in A. thaliana without affecting IAA biosynthesis and plant morphology.

Experimental procedures

Plant materials and growth conditions

Plants (A. thaliana ecotype Columbia) were grown in a temperature-controlled greenhouse under a 16-h light/8-h dark regime or in a growth chamber under a 12-h light/12-h dark regime at day/night temperatures of 21°C/18°C and at 40% humidity. The HIG1-1D activation-tagged mutant was previously identified as HPC1-1D in a population of activation-tagged lines (Schneider et al., 2005). The hig1-1 line GK-228B12 is homozygous for a T-DNA insertion in the second exon of HIG1/MYB51, and was supplied by GABI-Kat (Rosso et al., 2003); the HIG1/MYB51 transcript is not detectable in hig1-1 plants (see Figure 2a).

Preparation of methanolic extracts and HPLC analysis of desulfoglucosinolates

Leaves (50–80 mg) were placed in a 1.5-mL reaction tube and frozen in liquid nitrogen. Frozen leaf samples were lyophilized and homogenized in a mill (MM301; Retsch, Glucosinolates were extracted in 80% methanol after the addition of 20 μl of a 5 mm solution of benzyl glucosinolate ( as an internal standard. Extracts were loaded on a DEAE Sephadex A25-column (0.1 g powder equilibrated in 0.5 m acetic acid/NaOH, pH 5). Glucosinolate analysis was performed by conversion to desulfoglucosinolates through overnight incubation with purified sulfatase (EC designated ‘type H-1, from Helix pomatia, 16 400 U g–1 solid’ (Sigma, For analysis of desulfo glucosinolates, samples were subjected to HPLC analysis on a 1100 Series chromatograph (Hewlett-Packard, with a quaternary pump and a 1040M diode-array detector. Elution was accomplished on a Supelco C-18 column (Supelcosil LC-18, 250 × 4.6 mm, 5 μm; Supelco, with a gradient (solvent A, water; solvent B, methanol) of 0–5% B (10 min), 5–38% B (24 min), followed by a cleaning cycle (38–100% B in 4 min, with a 6-min hold, then 100–0% B in 5 min, with a 7-min hold). Peaks were quantified by the peak area at 229 nm (bandwidth 4 nm) relative to the area of the internal standard peak, applying the response factors as described by Brown et al. (2003).

Construction of transgenic recapitulation A. thaliana plants and constructs for ATR1/MYB34 and MYB122 overexpression

To generate HIG1/MYB51 overexpression (recapitulation) plants, the HIG1/MYB51 full-length coding sequence (without the stop codon) was amplified by RT-PCR and cloned into the Entry TOPO vector (Invitrogen, For primer sequences, see Table S1. To drive expression of the HIG1/MYB51 coding sequence under control of the CaMV 35S promoter, the binary Gateway compatible plant transformation vector pGWB2 was used. To recombine the insert from the entry clone into the destination vector, an LR reaction (Invitrogen) between both clones was performed. The final 35S:HIG1:pGWB2 construct was transformed into Agrobacterium tumefaciens (strain GV3101) by electroporation and into A. thaliana plants by vacuum infiltration.

To complement the hig1-1 loss-of-function mutant with the ATR1/MYB34 and MYB122 genes, the ATR1/MYB34 and MYB122 full-length coding sequences were amplified by RT-PCR and cloned into the Entry TOPO vector (for primer sequences, see Table S1). To recombine the insert from the entry clone into the destination vector, an LR reaction between the Entry vectors containing ATR1/MYB34 and MYB122, and the pGWB2 destination vector were performed. The final 35S:ATR1:pGWB2 and 35S:MYB122:pGWB2 constructs were transformed into A. tumefaciens (strain GV3101) by electroporation, and finally into A. thaliana wild-type and the hig1-1 mutant by vacuum infiltration. All transformants were selected with kanamycin and verified by PCR analysis. At least 70 independent primary transformants were analyzed by RT-PCR.

Histochemical analysis of transgenic plants expressing the ProHIG1:uidA fusion construct: wounding and touch treatments

The HIG1/MYB51-promoter region (from –1676 to + 342 bp) was amplified from genomic DNA of A. thaliana plants and cloned into the Entry TOPO vector (Invitrogen) (for primer sequences, see Table S1). To drive expression of uidA under control of the HIG1/MYB51 promoter, the binary Gateway compatible plant transformation vector pGWB3 was recombined with the Topo Entry vector using an LR reaction. The ProHIG1:uidA clone in pGWB3 was used to transform A. tumefaciens and A. thaliana.

Transformants were selected with kanamycin and verified by PCR analysis. Histochemical localization of GUS in transgenic plants harbouring the ProHIG1:uidA construct was performed as described by Jefferson et al. (1987) with some modifications. Sample tissues were infiltrated with the reaction buffer [50 mm Na2HPO4-NaH2PO4, pH 7.0, 0.5 mm K3Fe(CN)6, 0.5 mm K4Fe(CN)6, containing 2 mm 5-bromo-4-chloro-3-indolyl-β-D-glucuronic acid (X-Gluc) as substrate] under vacuum and incubated at 37°C overnight. Plant pigments were destained with 80% ethanol, and the GUS staining patterns were recorded under a binocular microscope (SMZ-U; Nikon,

Wounding of plants involved the mechanical treatment of leaves by cutting with a scalpel followed by GUS staining. For the leaf touch experiments, pressure was exerted by gently pushing part of the leaf between thumb and forefinger. Approximately 50% of the leaf area was left untouched. After 3–10 min, leaves were collected and infiltrated for GUS staining. Control plants were carefully collected avoiding any kind of mechanical stimuli and were subjected to the same infiltration procedure.

Real-time PCR analysis

The expression of glucosinolate biosynthesis genes was analyzed by real-time quantitative RT-PCR analysis using the fluorescent intercalating dye SYBR-Green in a GeneAmp® 5700 Sequence Detection System (Applied Biosystems, The Arabidopsis ACTIN2 gene was used as a standard. A two-step RT-PCR analysis was performed. First, total RNAs (10 μg per reaction) were reversely transcribed into cDNAs, using the First-Strand cDNA Synthesis SSII Kit (Invitrogen) according to the manufacturer’s instructions. Subsequently, the cDNAs were used as templates in real-time PCR reactions with gene-specific primers (Table S2). The real-time PCR reaction was performed using SYBR-Green master Kit System (Applied Biosystems) according to the manufacturer´s instructions. The C t, defined as the PCR cycle at which a statistically significant increase of reporter fluorescence is detected, is used as a measure for the starting copy number of the target gene. Relative quantification of expression levels was performed using the comparative C t method (see manufacturer’s instructions, Bulletin #2, Applied Biosystems). The relative value for the expression level of each gene was calculated by the equation Y = 2–ΔΔCt, where ΔC t is the difference between control and target products (ΔC t = C tGENE – C tACT,) and ΔΔC t = ΔC tmutant – ΔC twt. Thus, the calculated relative expression values are normalized to the wild-type expression level, WT = 1. The efficiency of each primer pair was tested using wild-type (Col-0) cDNA as a standard template, and the RT-PCR data were normalized dependent on the relative efficiency of each primer pair.

Auxin measurements

For IAA analysis, about 200 mg of leaf material was collected, immediately frozen in liquid nitrogen and homogenized. Subsequently, 1 ml of methanol (60°C) containing 10 pmol ml–1 deuterated [2H]2-IAA was added and the sample was further homogenized. The samples were shaken at 20°C for 60 min and then centrifuged for 15 min at 20 000 × g. The supernatant was concentrated to dryness using a vacuum evaporator, and the further processing of samples and the subsequent quantitative analysis of IAA as its methyl ester by chemical ionization (methanol) gas-chromatography tandem-mass-spectrometry (GC-MS/MS) were performed as described by Müller and Weiler (2000). All spectra were recorded on a Varian Saturn 2000 ion-trap mass spectrometer connected to a Varian CP-3800 gas chromatograph fitted with a CombiPal autoinjector (Varian, The following mother ions were used for quantification: IAA, m/z = 190 [M + H]+, [2H]2-IAA m/z = 192 [M + H]+. The quantities of endogenous compounds were calculated from the signal ratios of the unlabelled daughter ion (m/z = 130) over the stable isotope-containing daughter ion (m/z = 132) for each sample (Müller and Weiler, 2000).

Effector and reporter construction for transient co-transformation experiments

Promoter regions of DHS1 (from –1152 to + 351 bp), ASA1 (from –1210 to + 96 bp), TSB1 (from –2320 to + 132 bp), CYP79B2 (from –1383 to + 81 bp), CYP79B3 (from –1353 to + 84 bp), CYP83B1 (from –995 to + 27 bp) and AtST5a (from –1134 to + 141 bp) genes were amplified from genomic DNA of A. thaliana plants. The corresponding primer sequences are listed in Table S1. Promoters of Dsh1, ASA1, CYP79B2, CYP79B3, CYP83B1 and AtST5a genes were cloned into pEntry TOPO vector (Invitrogen). To drive Agrobacteria-mediated expression of uidA under control of those promoters, the binary plant transformation vector pGWB3i containing an intron within the GUS gene was generated. An 189-bp intron fragment was amplified from pPCV 6NFHyg GUS Int Vector (Vancanneyt et al., 1990) using a proof-reading polymerase. The blunt-ended PCR product was ligated to pGWB3 using the SnaBI restriction site. Using LR reactions (Invitrogen), pGWB3i was recombined with the pEntry Topo vectors containing the different promoters. The final ProDSH1:uidA, ProASA1:uidA, ProTSB1:uidA, ProCYP79B2:uidA, ProCYP79B3:uidA, ProCYP83B1:uidA and ProAtST5a:uidA clones in pGWB3i, as well as 35S:HIG1 in pGWB2, were used to transform the supervirulent A. tumefaciens strain LBA4404.pBBR1MCS.virGN54D.

Plant transfection via infiltration and fluorimetric GUS activity assay

To estimate the transactivation potential of HIG1/MYB51 towards promoters of target genes, the supervirulent Agrobacteria containing 35S:HIG1 in pGWB2 (effector), the same strain containing the empty pGWB2, and the antisilencing Agrobacteria strain 19K (Voinnet et al., 1999), and each of the reporter promoters in pGWB3i vector within the supervirulent Agrobacteria were taken from fresh YEB plates, grown overnight, sedimented, re-suspended in 10 mm MgCl2, 10 mm 2-(N-morpholine)-ethanesulphonic acid (MES), pH 5.6, and adjusted to an OD of 0.7–0.8. Two working solutions were prepared for each promoter and at least six plants were infiltrated. Working solution 1 contained a suspension with effector and reporter constructs together with the Agrobacteria strain 19K in a 1:1:1 ratio. Working solution 2 contained a suspension with an empty vector (without effector), reporter and 19K Agrobacteria strain in a 1:1:1 ratio. Acetosyringon was added (0.15 mm, final concentration) and the suspension was incubated for 2–4 h at room temperature.

Leaves of three plants were infiltrated with each working solution using a syringe. Following infiltration, plants were exposed for 12–24 h in darkness. Either five or six leaf probes were taken randomly from infiltrated plants for protein isolation and GUS activity measurements after 4 days of transient expression.

One part of the transfected leaf was subjected to histochemical staining and the other part to fluorimetric determination of GUS activity. Quantitative fluorimetric GUS assays were performed using 4-methylumbelliferyl-β-glucuronide (MUG) as a substrate dissolved in 50 mm sodium phosphate buffer (pH 7), 1 mm EDTA and 0.1% Triton X-100, as described by Jefferson et al. (1987). Protein content was determined using the BCA kit (Pierce Biotechnology, and bovine serum albumin as a standard. The activity of the reporter gene was expressed in nmol 4-methylumbelliferone (4-MU) per min and mg of extracted protein.

Dual-choice assays

Dual-choice assays were performed to study the consumptional preference of the generalist lepidopteran herbivore, S. exigua (Lepidoptera: Noctuidae). Eggs of S. exigua were received from Bayer CropScience ( and larvae were maintained on an artificial diet. Fourth-instar larvae were used in dual-choice assays. Larvae were tested individually in Petri dishes (5.5 cm in diameter), offering them two leaves of equivalent age of different Arabidopsis lines on moistened filter paper for 8 h at 25°C. Leaves were weighed and scanned before and after feeding. Leaf area was analyzed using Winfolia (Regent Instruments Inc., Three different pair combinations of leaves were provided to each of 20 larvae: HIG1-1D and wild-type (Col-0), HIG1-1D and hig1-1, and Col-0 and hig1-1. The consumed FW was calculated using the following: [(weight begin*area end)/area begin].


We thank Professor E. Weiler (University Bochum, Germany) for IAA determinations, B. Kleinhenz for excellent technical assistance and R. Yatusevich for cloning promoter GUS constructs used in the co-transformation assays. The Gateway destination vectors used in this work (pGWB2, pGWB3 and pGWB5) were kindly provided by T. Nakagawa (Shimane University, Japan). Eggs of Spodoptera exigua were received from Bayer CropScience (Monheim, Germany). This work was supported by grants from the European Union (QLK1-CT-2001-01 080), the Deutsche Forschungsgemeinschaft and the Fonds der Chemischen Industrie.