Indole Glucosinolate Biosynthesis Limits Phenylpropanoid Accumulation in Arabidopsis thaliana

Abstract

Plants produce an array of metabolites (including lignin monomers and soluble UV-protective metabolites) from phenylalanine through the phenylpropanoid biosynthetic pathway. A subset of plants, including many related to Arabidopsis thaliana, synthesizes glucosinolates, nitrogen- and sulfur-containing secondary metabolites that serve as components of a plant defense system that deters herbivores and pathogens. Here, we report that the Arabidopsis thaliana reduced epidermal fluorescence5 (ref5-1) mutant, identified in a screen for plants with defects in soluble phenylpropanoid accumulation, has a missense mutation in CYP83B1 and displays defects in glucosinolate biosynthesis and in phenylpropanoid accumulation. CYP79B2 and CYP79B3 are responsible for the production of the CYP83B1 substrate indole-3-acetaldoxime (IAOx), and we found that the phenylpropanoid content of cyp79b2 cyp79b3 and ref5-1 cyp79b2 cyp79b3 plants is increased compared with the wild type. These data suggest that levels of IAOx or a subsequent metabolite negatively influence phenylpropanoid accumulation in ref5 and more importantly that this crosstalk is relevant in the wild type. Additional biochemical and genetic evidence indicates that this inhibition impacts the early steps of the phenylpropanoid biosynthetic pathway and restoration of phenylpropanoid accumulation in a ref5-1 med5a/b triple mutant suggests that the function of the Mediator complex is required for the crosstalk.

Figure 1.

The Arabidopsis ref5-1 Mutant Is Defective in CYP83B1. (A) Representative photographs of wild-type (Col-0) and ref5-1 plants under white light or UV light. Plants were compared 3 weeks after planting. (B) Map-based cloning of the REF5 gene. The REF5 locus was mapped to a 6-BAC interval at the bottom of chromosome 4 defined by markers Cer449753 and Cer452102 (vertical bars). The annotation of three candidate genes within this interval, a cinnamoyl-CoA reductase-like protein (CCR like; At4g30470), a serine carboxypeptidase II-like protein (SCPL; At4g30610), and a cytochrome P450 monooxygenase 83B1 (CYP83B1; At4g31500), suggested that they might be involved in phenylpropanoid synthesis. (C) Gene organization of REF5 and schematic representation of ref5-1 and ref5-2 mutants. White box indicates exons, gray box in second exon indicates the heme binding region, and the intervening line denotes an intron. The location of the T-DNA insertion in ref5-2 (SALK_028573) is shown above the genomic structure. The arrow indicates mutation position in ref5-1.

Figure 2.

The ref5-1 and ref5-2 Mutants Are Defective in Indole Glucosinolate Biosynthesis and Phenylpropanoid Biosynthesis. (A) Sinapoylmalate content in wild-type, ref5 mutant, and ref5-1 × ref5-2 F1 plants. Data represent mean ± sd (n = 4). (B) Hypocotyl length of 7-d-old wild-type, ref5 mutant, and ref5-1 × ref5-2 F1 plants. Seedlings were grown on MS plates for 7 d. Data represent mean ± sd (n = 10). (C) Quantification of desulfoglucosinolate content in rosette leaves of ref5-1 and ref5-2 compared with the wild type. Glucosinolates are identified as follows: I3M, indol-3-methyl-glucosinolate; 4MOI3M, 4-methoxyindol-3-ylmethyl-glucosinolate. Data represent mean ± sd (n = 4). * and ** indicate P < 0.05 and P < 0.01 by a two-tailed Student’s t test compared with the wild type, respectively. (D) Three-week-old ref5 mutants grown on MS plates or directly planted on soil compared with the wild type. When seeds were directly planted on soil, ref5-1 mutants display a less severe growth defect than ref5-2 and sur2-1.

Figure 3.

Phenylpropanoid Accumulation Is Limited by IAOx or an Aldoxime Derivative. (A) Scheme of the indole glucosinolate biosynthesis pathway. The oxidation of Trp to IAOx is catalyzed by CYP79B2 and CYP79B3. CYP83B1/REF5 subsequently oxidizes IAOx to a product that spontaneously reacts to produce an S-alkylthiohydroximate. Further modifications, such as glucosylation and sulfation, give rise to the final indole glucosinolates. (B) Desulfoglucosinolate content measured from rosette leaves of 3-week-old plants. Data represent mean ± sd (n = 4). (C) Quantification by HPLC of sinapoylmalate in leaves of the wild type, ref5-1, cyp79b2 cyp79b3 double mutant, ref5-1 cyp79b2 cyp79b3 triple mutant, and yuc6-1D, an auxin accumulation mutant. Data represent mean ± sd (n = 5). * and ** indicate P < 0.05 and P < 0.01 compared with the wild type, respectively. * and ** indicate P < 0.05 and P < 0.01 by a two-tailed Student’s t test compared with the wild type, respectively. (D) Impact of overexpression of CYP79B2 on levels of sinapoylmalate. Sinapoylmalate content in wild-type (open circle) and transgenic plants overexpressing CYP79B2 (closed circle) is shown.

Figure 4.

The ref2-1 ref5-1 Double Mutants Display Severe Growth Defects and a Reduction of Sinapoylmalate Contents. (A) A representative ref2-1 ref5-1 double mutant compared with Col-0, ref2-1, and ref5-1. Four-week-old wild-type, ref2-1, ref5-1, and ref2-1 ref5-1 plants were grown under long-day conditions (16 h light/8 h dark). Bar = 1 cm. (B) Quantification of sinapoylmalate content in leaves of genotyped members of a ref2-1 × ref5-1 F2 segregating population. Four-week-old F2 individual lines were analyzed for sinapoylmalate content and genotyped for the ref2-1 and ref5-1 mutations. Data represent mean ± sd (n = 4). a, b, and c indicate P < 0.05 compared with the wild type, ref2-1, and ref5-1, respectively, by a Student’s t test.

Figure 5.

Analysis of Soluble Metabolites in ref5-1 comt1 and ref2-1 comt1 Double Mutants. (A) and (B) The level of 5-OH-feruloylmalate (A) and sinapoylmalate (B) in the wild type, ref5-1, comt1, ref5-1 comt1, ref2-1, and ref2-1 comt1. Data represent mean ± sd (n = 5). a and b indicate P < 0.01 compared with the wild type and comt1, respectively, from a Student’s t test. (C) Morphology of 4-week-old ref5-1 comt1 plants compared with the respective single mutants.

Figure 6.

The Phenylpropanoid Content of ref5 Mutant Roots Is Reduced. Coniferin (A) and syringin (B) content in seedling roots as determined by HPLC. Data represent mean ± sd (n = 3). Syringin content in comt1, fah1-2, ref5-1 comt1, and ref5-1 fah1-2 mutants was below the limit of detection. nd, not detected. Asterisks indicate P < 0.01 by a two-tailed Student’s t test compared with the wild type.

Figure 7.

Flavonoid Content in ref5 Mutants and cyp79b2 cyp79b3 Double Mutants Compared with the Wild Type. HPLC profiles of UV absorbing metabolites (A) and quantification of flavonoids (B) from leaves of 4-week-old wild type, ref5-1, and cyp79b2 cyp79b3 double mutants. Two flavonoid peaks are shown as K1 and K3. K1, kaempferol 3-O-[6’’-O-(rhamnosyl)glucoside] 7-O-rhamnoside; K3, kaempferol 3-O-rhamnoside 7-O-rhamnoside. Data represent mean ± sd (n = 4). * and ** indicate P < 0.05 and P < 0.01 by a two-tailed Student’s t test compared with the wild type, respectively.

Figure 8.

Phenylalanine Content in ref5 Mutants and cyp79b2 cyp79b3 Double Mutants. Data represent mean ± sd (n = 3). Asterisks indicate P < 0.01 by a two-tailed Student’s t test compared with the wild type.

Figure 9.

PAL Activity in ref5 Mutants Compared with the Wild Type and Growth Phenotypes of the ref5-1 pal1 pal2 Triple Mutant and the ref5-1 pal1 and ref5-1 pal2 Double Mutants. (A) PAL activity in crude extracts of 2-week-old wild-type and ref5 seedlings. PAL activity is expressed as average PAL activity in pkat µg⁻¹ protein ± sd. Asterisks indicate P < 0.01 by a two-tailed Student’s t test compared with the wild type. (B) Four-week-old soil-grown ref5-1 pal1 pal2 triple mutant (ref5-1 pal1/2), ref5-1 pal1, and ref5 pal2 double mutants compared with the wild type, ref5-1, and the pal1 pal2 double mutant.

Figure 10.

Growth Phenotypes of ref5-1 ref3-2 Double Mutants.

Figure 11.

Relative Expression of MYB4 and Phenylpropanoid Biosynthetic Genes in ref5 Mutants. At1g13320 was used for internal control. The relative expression was calculated with 2^−ΔΔCT. Data represent mean ± sd (n = 3). * and ** indicate P < 0.05 and P < 0.01, respectively, compared with the wild type.

Figure 12.

Three MED5b Mutants Were Isolated from a ref5 Suppressor Screen. (A) Morphological phenotype of the three ref5 suppressors. Three-week-old suppressors are compared with the wild type and the ref5-1 mutant. (B) The location of mutations in ref5 suppressors is shown above the schematic of the MED5b (At3g23590) genomic locus. The two mutations in the three suppressors result in early stops. (C) Sinapoylmalate content measured in whole rosette leaves of 3-week-old plants. Data represent mean ± sd (n = 4). Asterisks indicate P < 0.01 by a two-tailed Student’s t test compared with ref5-1.

Figure 13.

The Disruption of MED5a/b Restores Phenylpropanoid Content in ref5-1. (A) Morphology and ref phenotype in 3-week-old ref5-1 med5a/b triple mutant compared with its single mutants. ref5-1 med5a/b plants exhibit the high auxin phenotypes similar to ref5-1 but the ref phenotype is suppressed. (B) to (E) Soluble metabolite analysis in mature leaves and seedling roots. Sinapoylmalate and sinapoylglucose content was measured in 3-week-old whole rosette leaves ([B] and [C]). Syringin and coniferin content was measured in 2-week-old seedling roots ([D] and [E]). Asterisks indicate P < 0.01 by a two-tailed Student’s t test compared with ref5-1.