Analysis of a Chemical Plant Defense Mechanism in Grasses

See allHide authors and affiliations

Science  01 Aug 1997:
Vol. 277, Issue 5326, pp. 696-699
DOI: 10.1126/science.277.5326.696


In the Gramineae, the cyclic hydroxamic acids 2,4-dihydroxy-1,4-benzoxazin-3-one (DIBOA) and 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one (DIMBOA) form part of the defense against insects and microbial pathogens. Five genes,Bx1 through Bx5, are required for DIBOA biosynthesis in maize. The functions of these five genes, clustered on chromosome 4, were demonstrated in vitro. Bx1 encodes a tryptophan synthase α homolog that catalyzes the formation of indole for the production of secondary metabolites rather than tryptophan, thereby defining the branch point from primary to secondary metabolism.Bx2 through Bx5 encode cytochrome P450–dependent monooxygenases that catalyze four consecutive hydroxylations and one ring expansion to form the highly oxidized DIBOA.

A substantial number of secondary metabolites in plants are dedicated to pathogen defense. These include the cyclic hydroxamic acids, which are found almost exclusively in Gramineae. For example, DIMBOA and its precursor DIBOA are present in maize. DIMBOA confers resistance to first-brood European corn borer (Ostrinia nubilalis), northern corn leaf blight (Helminthosporium turcicum), maize plant louse (Rhophalosiphum maydis), and stalk rot (Diplodia maydis), as well as to the herbicide atrazine (1). DIBOA is the main hydroxamic acid in rye, whereas DIMBOA is the predominant form in wheat and maize (1). The DIMBOA and tryptophan biosynthetic pathways share certain intermediates. Labeled tryptophan precursors such as anthranilic acid and indole are incorporated into DIMBOA, although labeled tryptophan is not incorporated (2). The maize mutation bx1 (benzoxazineless) abolishes DIMBOA synthesis (3). Plants homozygous for bx1 grow normally but are extremely susceptible to the pathogens mentioned above.

To clone the Bx1 gene, we used the Mutator(Mu) transposon tagging system (4). Approximately 150,000 seeds were produced from a cross of a Mu female line with the pollen from plants homozygous for the recessive bx1mutant allele (3). Seventeen putative mutants were identified and outcrossed to an inbred Bx1/Bx1 line. The segregation of the bx1 alleles in the progeny of the crosses was followed by a cleaved amplified polymorphic sequence (CAPS) marker (5) derived from the linked Bx4 gene (6). One of the putative mutants showed the expected 1:1 segregation for the bx1 allele and the newlyMu-induced recessive bx1 allele with respect to the linked marker. From this material, the Mu element cosegregating with the new bx1::Mu allele was identified and a flanking genomic DNA fragment was isolated by a polymerase chain reaction (PCR)–based method (7). This fragment was used to isolate the wild-type Bx1 and recessive bx1 (3) alleles from genomic λ libraries (8) as well as a full-length cDNA clone (Fig.1). DNA sequence analysis revealed that the bx1 allele harbors a deletion of 924 base pairs (bp) comprising 355 bp of the 5′ nontranscribed region, the first exon, the following intron, and 53 bp of the second exon. The position of theMu element in the bx1::Mu allele isolated in the transposon-tagging experiment was determined by PCR amplification of the flanking genomic sequences.

Figure 1

Structure and chromosomal location of theBx genes. (A) Schematic representation of theBx gene cluster on chromosome 4. Genetic distances are indicated (in centimorgans). (B) Exon-intron structure ofBx1 through Bx5. Exons are represented by boxes. Translation start and stop codons and polyadenylate addition sites are shown. Arrows represent insertion of a Mu element in the bx1::Mu and bx3::Mualleles. The deletion in the bx1 standard allele is indicated; it comprises nucleotides 1366 to 2289 of the published sequence (9). The distance from Bx1 toBx2 (2490 bp) is not drawn to scale. The complete sequences of the genes have been deposited in the European Molecular Biology Laboratory data bank [accession numbers X76713 (Bx1), Y11368 (Bx2),Y11404 (Bx3), X81828 (Bx4), and Y11403 (Bx5)]. (C) Insertion sites of Mu in bx1::Mu andbx3::Mu. The characteristic 9-bp host sequence duplication associated with Mu insertion is underlined with an arrow. The insertion occurred at position 2826 of the genomic DNA sequences in bx1::Mu and at position 1260 inbx3::Mu.

DNA sequence analysis revealed the exact Mu insertion site and the characteristic 9-bp host sequence duplication associated with integration of the transposon (Fig. 1). The exon sequences ofBx1 were found to be identical to a gene previously described (9) that is homologous to tryptophan synthase α (TSA) and, when expressed in Escherichia coli, complemented a tryptophan synthase α mutation. TSA catalyzes the conversion of indole-3-glycerol phosphate to indole, the penultimate reaction in tryptophan biosynthesis. However, indole was also implicated as an intermediate in DIMBOA biosynthesis (2). Consequently, the bx1 mutant should also be defective in the production of free indole.

The immersion of shoots of bx1 seedlings (4 days after imbibition) into a 1 mM solution of indole in the dark for 1 day restored the formation of DIMBOA (Fig.2A). Hence, the biosynthetic block in thebx1 mutant cannot be downstream of indole formation. Addition of tryptophan did not result in DIMBOA accumulation (Fig. 2B). When [3-13C]indole was administered to maize shoots, [2-13C]DIMBOA (75% incorporation) was recovered (Table1). This confirms that indole is an intermediate in DIMBOA biosynthesis (2).

Figure 2

Detection of metabolites of the DIMBOA pathway by HPLC. Metabolites are indicated at the position of chromatographic peaks. S represents the solvent peak. (A and B) Feeding of bx1 standard mutant seedling shoots with 1 mM indole (A) or 1 mM tryptophan (B). Seedling material (1 g) was extracted (21) and analyzed on a Merck LiChroCART RP-18 HPLC column (4 × 125 mm). Elution was for 5 min under isocratic conditions with solvent A (H2O/acetic acid, 9:1) followed by a linear gradient from 100% solvent A to 100% solvent B (methanol/H2O/acetic acid, 70:27:3) over 7 min. (C to F) Analysis of maize P-450 enzymes expressed in yeast microsomes. Reaction mixtures of 0.2 ml contained 50 mM potassium phosphate buffer (pH 7.5), 0.8 mM NADPH, 0.1 to 0.5 mM of the respective substrates, and 1 mg of microsomal protein. Incubation was for 30 min at 25°C. HPLC analysis was as described above. Shown are results for (C) BX2 microsomes incubated with indole, (D) BX3 microsomes incubated with indolin-2-one, (E) BX4 microsomes incubated with 3-hydroxyindolin-2-one, and (F) BX5 microsomes incubated with HBOA.

Table 1

1H-NMR analysis of indole-derived enzymatic products in maize (δ, chemical shift;JHH, coupling constant). The signal assignments are based on two-dimensional NMR analysis (28). We used1H-NMR spectroscopy to monitor 13C enrichment from [13C]indole and subsequent metabolites. The13C label from [3-13C]indole was incorporated into position 3 of indolin-2-one (100% incorporation) and position 2 of DIMBOA (75% incorporation). The label from [3-13C]indolin-2-one was incorporated into position 3 of 3-hydroxyindolin-2-one (100% incorporation). Some values ofJHH are dual because coupling acts on both protons over scalar bonds; others are missing because the signals were characterized by singlet multiplicity and JHHcould not be extracted.

View this table:

In bacteria, TSA activity is almost completely dependent on formation of an active α2β2 complex (10) with tryptophan synthase β (TSB), and indole is usually not released during tryptophan synthesis. An analogous heterosubunit complex exists in Arabidopsis (11). If the TSA homolog BX1 catalyzes the formation of free indole from indole-3-glycerol phosphate, BX1 should function independently of TSB. To test this assumption, we expressed BX1 in E. coli, then purified and assayed it for steady-state kinetic constants (12). We determined a Michaelis constantK m indole-3-glycerol phosphate of 0.013 mM and a catalytic rate constant k cat of 2.8 s−1. Comparison of these values with the constants for conversion of indole-3-glycerol phosphate to indole by E. coliα2β2 complex (K m indole-3-glycerol phosphate = 0.027 mM,k cat = 0.2 s−1) (13) demonstrates that BX1, independent of TSB, is ∼30 times as efficient as the bacterial complex in catalyzing the production of indole. Tryptophan is essential for the maize plant (14). Because the bx1 mutants are viable, Bx1 cannot be the only maize gene encoding TSA activity. We suggest that there are at least two TSA genes in maize: one that is involved in tryptophan biosynthesis, forming the α2β2 complex, and a second gene, Bx1, that is required for the production of free indole and secondary DIMBOA synthesis.

Four maize cytochrome P-450–dependent monooxygenase genes, one of which was isolated by subtractive cDNA cloning from high versus low DIMBOA-accumulating lines (6), are in the CYP71Csubfamily of plant cytochrome P-450 genes. These genes are strongly expressed in young maize seedlings, share an overall amino acid identity of 45 to 60%, and are clustered on the short arm of chromosome 4 (Fig. 1). The finding that all oxygen atoms of DIMBOA are incorporated from molecular oxygen (15) led to the speculation that these cytochrome P-450 enzymes might be involved in this pathway. The genes encoding these enzymes are here designatedBx2, Bx3, Bx4, and Bx5(6).

Direct evidence for the involvement of Bx3 in DIMBOA biosynthesis is provided by a mutant allele (Bx3::Mu) isolated by a reverse genetic approach to screen for Mu insertions in the P-450 genes (16). Sequencing of the PCR-amplified Mu-flanking genomic DNA fragments showed that Bx3::Mu has aMu transposon inserted in the second exon of the gene (Fig.1). In maize seedlings homozygous for the recessive mutant allele, no DIMBOA could be detected by high-performance liquid chromatography (HPLC) analysis. In contrast, DIMBOA was detected in seedlings that were either heterozygous or homozygous wild-type. Cosegregation of the recessive mutant phenotype was established by genomic blotting analysis of 27 F2 individuals (four homozygous recessives). These results demonstrate that an intact Bx3 gene is required for DIMBOA biosynthesis.

Genomic clones of Bx2 through Bx5 were isolated from a λ library and sequenced (8), and the exon-intron structure of the genes was determined (Fig. 1). The position of one intron is identical for Bx2 through Bx5, indicating a common evolutionary origin. The position of a second intron is conserved in Bx3 and Bx5. DNA sequence comparison showed that Bx1 and Bx2 are separated by only 2490 bp.

To investigate the function of the four P-450 enzymes in DIMBOA biosynthesis, we used a yeast expression system (17). The cDNAs of Bx2 through Bx5 were inserted into the pYeDP60 expression vector (18). These constructs were used to transform the WAT11 yeast strain. In WAT11, a galactose-inducibleArabidopsis thaliana microsomal NADPH (reduced form of nicotinamide adenine dinucleotide phosphate)–P-450 reductase (ATR1) replaces the yeast reductase (19).

Microsomes were isolated from the transgenic yeast strains and tested for enzymatic activity (17, 19). Indole was converted to DIBOA by the stepwise action of the four cytochrome P450 enzymes (Fig. 2, C to F). When [3-13C]indole was incubated with yeast microsomes containing BX2 protein, [3-13C]indolin-2-one was produced in the reaction assay. A sufficient amount of [3-13C]indolin-2-one was produced by this enzyme-catalyzed reaction to test for subsequent enzymatic conversions. Incubation of [3-13C]indolin-2-one with microsomes containing BX3 resulted in the production of [3-13C]hydroxyindolin-2-one. For further analysis, unlabeled 3-hydroxyindolin-2-one was obtained by reduction of isatin (20). The conversion of 3-hydroxyindolin-2-one to 2-hydroxy-1,4-benzoxazin-3-one (HBOA) was catalyzed by microsomes containing BX4. The reaction mechanism for this unusual ring expansion is as yet unknown. Finally, HBOA was converted to DIBOA by microsomes containing BX5. This reaction was previously described for maize microsomes (21). The identity of the reaction products was confirmed by cochromatography with the authentic substances and by their ultraviolet spectra. The reaction products indolin-2-one and 3-hydroxyindolin-2-one were further identified by their 1H nuclear magnetic resonance (NMR) spectra (Table 1). The identity of HBOA and DIBOA was corroborated by gas chromatography–mass spectrometry (GC-MS) analysis (22).

Although the four cytochrome P-450 enzymes are homologous proteins, they are substrate-specific. Only one substrate was converted by each respective P-450 enzyme to a specific product. No detectable conversions occurred in other enzyme-substrate combinations. Enzymatic reactions identical to the reactions with the different yeast microsomal preparations could be performed with maize microsomes, which indicates that these reactions occur natively in maize. These findings suggest an in vivo reaction sequence in maize from indole to HBOA (Fig.3). According to this scheme, benzoxazinone would not be a natural intermediate for DIMBOA synthesis, as proposed earlier on the basis of feeding experiments (23). Whether alternative routes for DIMBOA synthesis exist remains to be determined.

Figure 3

Biosynthetic pathways to DIMBOA and tryptophan. Names of gene products are indicated for each of the reactions. BX1 represents a tryptophan synthase α activity. BX2 through BX5 represent cytochrome P-450–dependent monooxygenases of the CYP71C subfamily. The sequence of N-hydroxylation and introduction of the methoxy group at C-7 of DIMBOA has not yet been elucidated.

Bx1 and the four cytochrome P-450 genes represent a sufficient set of genes for the conversion of indole-3-glycerol phosphate to DIBOA (Fig. 3). DIMBOA is the 7-methoxy derivative of DIBOA. Because the oxygen atom at C-7 is incorporated from molecular oxygen (15), hydroxylation by another cytochrome P-450 enzyme followed by a methyltransferase reaction would be expected for the conversion of DIBOA to DIMBOA.

We estimate that the DIMBOA concentration in maize seedlings is ∼0.1% of the fresh weight. This value exceeds the total tryptophan content of the seedling by a factor of about 10 to 20 (24). Hence, most of the metabolites from the early steps of the tryptophan pathway would end up in the DIMBOA pathway. Indole-3-glycerol phosphate represents the branch point from tryptophan biosynthesis, and BX1 enzyme would catalyze the committing step.

The synthesis of several other secondary metabolites in plants, such as the indole glucosinates, anthranilate-derived alkaloids, and tryptamine derivatives (24, 25), depends on the tryptophan pathway. Indole-3-glycerol phosphate was also proposed as a branch point from the tryptophan pathway for the synthesis of the indolic phytoalexin camalexin (3-thiazol-2′-yl-indole) in Arabidopsis thaliana(26, 27). It will be interesting to see whether a BX1 homologous enzyme also catalyzes the first specific step in camalexin synthesis. Such an observation would shed light on the role of indole-3-glycerol phosphate as an intermediate in a wide range of secondary metabolites in plants.


Stay Connected to Science

Navigate This Article