Current GM crops are thus far the almost exclusive domain of herbicide and insect resistance traits. The Bt toxins used for insect control have a narrow specificity against lepidopteran and coleopteran pests only. Yet, aphids and thrips are highly important pests worldwide, causing severe direct losses and transmitting devastating viruses such as Tomato Spotted Wilt Virus (TSWV). So far, few useful traits against aphids or thrips have been reported. The ideal of an insecticide-free culture of GM crops like cotton or potato is, therefore, currently compromised by the continued need in those crops to fight sucking pests using chemical means.

At Plant Research International in Wageningen, The Netherlands, we have identified two new types of genes to fight sucking pests. The first involves protease inhibitors and the other involves mono- and sesquiterpene synthase genes. Protease inhibitors interfere with protein digestion, causing stunted growth, increased mortality, and reduced fecundity. Mono- and sesquiterpenes act primarily as cues emitted by plants in response to insect attack. They determine food choices and call in the help of predators and parasites to fight the herbivore. We found that manipulation of these traits can be a successful way of controlling major sucking insect pests such as western flower thrips and aphids.

Figure 1. Peach aphid (Myzus persicae) on the left panel, and the western flower thrips (Frankliniella occidentalis, female adult) on the right panel.

Protease inhibitors
Protease inhibitors were successfully applied for the first time in transgenic tobacco in 1987 against Heliothis zea. This initiated a rush of research to employ these commonly found genes in plants as insect resistance traits. However, it became quite evident that overexpression of most plant protease inhibitors was quite ineffective and was resulting at the most in a minor slowdown of growth rate. In 1995 it was demonstrated that in response to dietary inhibitors the insects were able to induce protease genes that were insensitive to them. Recently, we published a detailed analysis of how these resistant enzymes evolved from their sensitive ancestral genes. It was clear that there was a need to find inhibitors still effective against such "resistant" enzymes. To find a source of such inhibitors, two approaches were proposed. In the first approach, synthetic libraries of inhibitor variants were selected using phage display in order to generate novel structures¹. In the other approach, inhibitors from mainly the animal kingdom were tested against the insect proteases^2,3.

The phage display method, although elegant in principle, suffered from a lack of sufficient quantities of resistant enzymes to be used in the selection experiments. Also, it became evident that the tertiary protein fold of the inhibitors was more crucial than the primary amino acid sequence in blocking inhibitors from entering the active site. Changing the folds of proteins was not a realistic option, and, thus, the successes using phage display remained few. Nevertheless, Ceci et al.¹ demonstrated that they could select a chymotrypsin inhibitor (Chy8), which was five times more effective against pea and peach aphid than the parent trypsin inhibitor molecule MTI-2. For pea aphid the IC50 and LC50 were both around 75 ug/ml, which translates into an expression level in plants of 0.5 – 1% of total protein (Table 1).

The use of inhibitors from the animal kingdom proved to be an easier way of finding novel molecules with potency against insect pests. A large range of known cysteine and aspartic protease inhibitors was tested against aphids and thrips. Several inhibitors appeared to be potentially useful against these insects and, in the case of western flower thrips, were investigated in detail. Particularly effective was a dual inhibitor from sea anemone, called equistatin. This inhibitor represented a new class of protease inhibitors with a novel fold that was very good at blocking both cysteine and aspartic gut proteases of many insects and had good results in in vitro bioassays. Upon overexpression in some plants like potato, this inhibitor, however, was quite susceptible to cleavage by asparagine-specific plant proteases called legumains. The combination of equistatin with a number of different cystatins (which also act as legumain inhibitors) in the form of fusion proteins of four to seven independent domains prevented degradation, and in addition, proved to be much more effective against thrips than any of the single domains. Greenhouse trials, which monitored the survival of adult insects and the number of offspring produced during the first 14 days, demonstrated that the multidomain transgenic potato and chrysanthemum plants had fewer adults and 80% less offspring. From the data it was predicted that the population would eventually die out³ (Figure 2). In vitro assays had only found effects on fecundity and not on adult mortality. Choice assays had, however, indicated that protease inhibitors not only reduce the growth of larvae and fecundity of adults, but are also strongly deterrent to adult insects in a dose dependent fashion². So the disappearance of the adults from their cages in the greenhouse was explained as a result of deterrence and not mortality. If insects even try to escape their only food source in a no-choice situation, deterrence or repellence may prove an effective, additional way of protecting plants against herbivores. Volatile organic compounds emitted by plants are an interesting second strategy in that respect.

Figure 2. Effect of overexpression of an engineered 7-domain protease inhibitor in chrysanthemum on the average number of larvae found on the plants two weeks after inoculation with females. Tests were performed in the greenhouse with individual caged plants. Data are based on a replicate of six plants. P-values represent significance by t-test.

Terpene synthases
Volatile organic compounds emitted by plants are known to provide strong cues to predators and parasites of herbivores to locate their prey. We recently published⁴ that the herbivore itself is affected by these compounds. In the article, we demonstrate that in choice assays aphids are deterred from Arabidopsis plants that constitutively produce high levels of linalool. Recent unpublished data further corroborate these results on other plant species such as potato and chrysanthemum (Figure 3). On those plants, the deterrence was higher, with 75% of aphids and 90% of adult thrips preferring the control over the transgenic plants. In the next year, greenhouse and field trials will be carried out to measure the effects of overexpression of linalool in a realistic situation and on other insects. In the future, the challenge will be to find the right balance between cost for the plant and effect in terms of resistance. These aspects will also hinge on our ability to find the most active volatiles and to manage their expression in the right tissue and with the proper timing, just like plants have done over millions of years.

Figure 3. Effect of overexpression of the monoterpene linalool in chrysanthemum based on the linalool synthase gene from strawberry. The graph shows that 90% of thrips given a choice between the transgenic line, which emits linalool very strongly, and the control line in which linalool is not detectable, prefer the control line. After the first 15 minutes of orientation by the insects, effects are significant.

Conclusion
Recently, the ecological roles of both monoterpenes such as linalool and protease inhibitors on oviposition success of moths and seed set success of plants were demonstrated in field experiments using wild tobacco species by the group of Ian Baldwin. This has improved our understanding of the decisive role of these genes for the survival of plants in a natural setting. We have demonstrated that engineering these traits to make them more effective by the selection of more active inhibitors or by promoting the emission of higher levels of specific volatile organic compounds can make an even stronger difference to the success of sucking insect pests on transgenic plants.

References

1. Ceci LR, Volpicella M, Conti S, Gallerani R, Beekwilder MJ, Jongsma M.A. (2003) Selection by phage display of a mustard chymotrypsin inhibitor toxic to pea aphid. Plant Journal 33: 557-566.

2. Outchkourov NS, de Kogel WJ, Schuurman-de Bruin A, Abrahamson M, Jongsma MA (2004a) Specific cysteine protease inhibitors act as deterrents of Western flower thrips Frankliniella occidentalis (Pergande) in transgenic potato. Plant Biotechnology Journal 2: 439-448.

3. Outchkourov NS, de Kogel WJ, Wiegers GL, Abrahamson M, and Jongsma MA. (2004b) Engineered multidomain cysteine protease inhibitors yield resistance against western flower thrips (Frankliniella occidentalis) in greenhouse trials. Plant Biotechnology Journal 2: 449-458.

4. Aharoni A et al. (2003) Terpenoid metabolism in wild-type and transgenic Arabidopsis thaliana plants. Plant Cell 15: 2866-2884.

Maarten A. Jongsma
Plant Research International, Wageningen, The Netherlands
Maarten.jongsma@wur.nl
http://www.plant.wur.nl; www.impactvector.com