Report

Resistance to an Herbivore Through Engineered Cyanogenic Glucoside Synthesis

See allHide authors and affiliations

Science  07 Sep 2001:
Vol. 293, Issue 5536, pp. 1826-1828
DOI: 10.1126/science.1062249

Abstract

The entire pathway for synthesis of the tyrosine-derived cyanogenic glucoside dhurrin has been transferred from Sorghum bicolorto Arabidopsis thaliana. Here, we document that genetically engineered plants are able to synthesize and store large amounts of new natural products. The presence of dhurrin in the transgenicA. thaliana plants confers resistance to the flea beetlePhyllotreta nemorum, which is a natural pest of other members of the crucifer group, demonstrating the potential utility of cyanogenic glucosides in plant defense.

Cyanogenic glucosides are a group of amino acid–derived secondary metabolites that are widely distributed in the plant kingdom (1, 2). When the plant tissue is disrupted by herbivore attack, the cyanogenic glucosides are degraded into a sugar, a keto compound, and hydrogen cyanide (HCN). This cyanogenesis confers protection against some, but not all, herbivore attacks (1, 3). Insects feeding on cyanogenic plants may have evolved mechanisms to detoxify or to sequester cyanogenic glucosides (4–6), which in turn protect the insect against predators (7). To assess the effect of cyanogenesis, it is necessary to study insects that have not coevolved with cyanogenic glucosides. Such insects are found among those that specifically feed on cruciferous plants, which do not produce cyanogenic glucosides. To render investigations in such an experimental system possible, we transferred the pathway for cyanogenic glucoside biosynthesis into the cruciferous plant speciesArabidopsis thaliana by genetic engineering and studied the effect of cyanogenic glucosides in plant defense against the crucifer-specialist flea beetle Phyllotreta nemorum.

Biosynthesis of the cyanogenic glucoside dhurrin in Sorghum bicolor is highly channeled (8) and catalyzed by two multifunctional microsomal cytochromes P450 (CYP79A1 and CYP71E1) (9–13) and a soluble UDPG-glucosyltransferase (sbHMNGT) (14) [Web fig. 1 (15)]. We transformedA. thaliana plants expressing the genes encoding CYP79A1 and CYP71E1 (16) with a recombinant plasmid conferring gentamycin resistance and encoding sbHMNGT (17). Newly generated plants containing the dhurrin pathway were kanamycin- and gentamycin-resistant, which facilitated their selection.

The strategy resulted in several independent lines that were examined for their ability to synthesize dhurrin. Upon administration of radiolabeled tyrosine to detached leaves and analysis of methanol (MeOH) extracts by thin-layer chromatography, one predominant radioactive product comigrating with authentic dhurrin was observed. High-performance liquid chromatography–mass spectrometry (HPLC-MS) analyses (18) showed that the product had the same retention time, ultraviolet (UV) spectral properties, and molecular mass as the dhurrin standard (Fig. 1).

Figure 1

HPLC analysis of A. thaliana plants containing either the entire dhurrin biosynthetic pathway or the cytochrome P450–catalyzed part. a, CYP79A1 and CYP71E1; b, CYP79A1, CYP71E1, and sbHMNGT; c, wt. Each UV trace represents equal amounts of leaf material.

Four-week-old transgenic A. thaliana plants contained as much as 4 ± 0.5 mg of dhurrin per gram of fresh weight (gfw) (19). This level of dhurrin is similar to that found in seedlings of S. bicolor (20), demonstrating the ability to efficiently integrate the dhurrin biosynthetic pathway intoA. thaliana. Transgenic A. thaliana lines containing ∼1 mg of dhurrin/gfw displayed no apparent phenotypic differences as compared with the wild-type (wt) plants, and lines containing higher levels of dhurrin had only small reductions in growth [Web fig. 2 (15)]. Therefore, the diversion of tyrosine toward dhurrin biosynthesis and the storage of dhurrin did not cause inherent metabolic problems. As with other plants containing cyanogenic glucosides, the transgenic A. thaliana plants released high levels of HCN, up to ∼2 μmol/gfw, upon tissue damage (21). Likewise, dialyzed protein extracts from A. thaliana hydrolyzed dhurrin (22). Therefore, an endogenous β-glucosidase with dhurrin hydrolyzing activity is present in A. thaliana.

The expression of the CYP79A1 gene in A. thalianaresults in the production of p-hydroxybenzylglucosinolate as a result of metabolic cross talk between the pathways for cyanogenic glucoside and glucosinolate synthesis (23). A. thaliana plants expressing all three dhurrin biosynthetic pathway genes also accumulated p-hydroxybenzylglucosinolate, although at much lower levels (22), whereas dhurrin was the only product seen to accumulate in high abundance (Fig. 1). As withS. bicolor (8), pathway intermediates in theseA. thaliana plants were hardly detectable. Thus, although all three sorghum sequences were driven by the cauliflower mosaic virus (CaMV) 35S promoter in the transgenic A. thalianaplants, the individual enzyme activities of the pathway for dhurrin biosynthesis seemed to be adequately balanced. Transgenic plants lacking sbHMNGT produce p-glucosyloxy-benzoylglucose,p-glucosyloxy-benzoic acid, andp-hydroxybenzoylglucose (Fig. 1). Radiolabeling experiments showed that these glucosides are derived fromp-hydroxymandelonitrile that had decomposed into HCN andp-hydroxybenzaldehyde, of which the latter had been finally oxidized into p-hydroxybenzoic acid (Fig. 2). None of the hundred or more predicted A. thaliana(24) secondary plant product glucosyltransferases are capable of converting the aglyconep-hydroxymandelonitrile into the corresponding cyanogenic glucoside in planta. Thus, the glucosyltransferase activity provided by sbHMNGT is necessary to obtain dhurrin production in transgenicA. thaliana.

Figure 2

A. thaliana plants expressing both CYP79A1 and CYP71E1, but not sbHMNGT, accumulate glucosides of benzoic acid, a product of p-hydroxymandelonitrile decomposition.

The flea beetle P. nemorum (Coleoptera: Chrysomelidae: Alticinae) accepts wt A. thaliana as a food source (25) and as a specialist crucifer-feeder is not expected to have encountered cyanogenic glucosides during its recent evolutionary history. In choice tests with adult beetles (26), the consumption of leaf-disc material from transgenic A. thaliana plants containing dhurrin was compared to consumption of wt plants (Fig. 3A). The beetles consumed up to 80% less of the transgenic leaf-disc material (D ± 95%, confidence limit = 0.70 ± 0.10). Consumption of leaf-disc material from the transgenic lines expressing the two cytochrome P450 genes (CYP79A1 and CYP71E1) (D = 0.06 ± 0.16), or the UDPG-glucosyltransferase gene (sbHMNGT) (D = 0.03 ± 0.06), or containing the two empty expression vectors (D = –0.18 ± 0.23) was not significantly different from the consumption of leaf-disc material from wt plants. Therefore, the deterrent effect was directly attributable to the presence of dhurrin.

Figure 3

A. thaliana leaves containing dhurrin inhibit flea beetle and larvae feeding. (A) Adult beetles fed extensively only on leaves containing no dhurrin. (B) Larvae (indicated by arrows) frequently initiated no mines on leaves containing dhurrin, although attempts were made to feed (indicated by circles). Scale bar, 2.5 mm. (C) Nearly all larvae (98%) presented to leaves containing about 4 mg of dhurrin/gfw died.

Normally, after hatching from eggs laid in the soil, the larvae of P. nemorum climb a plant in search of a suitable site for initiation of a leaf mine. In nonchoice bioassays with newly emerged flea beetle larvae (27), the presence of dhurrin in the transgenic plants reduced the number of leaf mines initiated (G = 190.0, df = 5, P < 0.0001) (Fig. 3, B and C). All larvae that did not initiate mines died. Of the larvae that did initiate mines, a higher mortality was observed on leaves containing dhurrin (G = 149.9, df = 5,P < 0.0001) (Fig. 3C). As for the adult insects, mine initiation and larval survival on transgenic plants expressing the cytochrome P450 genes only, the glucosyltransferase gene only, or containing the two empty expression vectors were not significantly different from those on wt plants (mine initiation rates:G = 4.39, df = 3, P > 0.05; survival rates: G = 1.93, df = 3,P > 0.05).

Thus, the pathway for biosynthesis of the cyanogenic glucoside dhurrin can be transferred from sorghum into the acyanogenic model plant A. thaliana by the use of genetic engineering. The accumulation of substantial amounts of dhurrin does not appear to pose any inherent physiological problems for the transgenic A. thaliana and confers resistance to the flea beetle P. nemorum, demonstrating that cyanogenic glucosides can promote plant defense. Such engineering of cyanogenic glucosides into acyanogenic crop plants may prove useful for pest control purposes.

  • * These authors contributed equally to this work.

  • To whom correspondence should be addressed. E-mail: blm{at}kvl.dk

REFERENCES AND NOTES

View Abstract

Navigate This Article