INTRODUCTION

Most insects depend on plants for their food, and so insects are among the most aggressive and destructive adversaries of plants. Terrestrial plants characteristically lack mobility and cannot elude these destructive pests by flight. Yet plants are far from passive participants in the struggle that has gone on between insects and modern plants for more than 65 million years. The primary weapons of plants against the feeding ravages of insects are the plants' own chemical constituents. The thousands of distinct compounds higher plants synthesize in their metabolism are classified either as primary metabolites or as secondary metabolites. Primary metabolites are substances such as nucleic acids, adenosine triphosphate (ATP) and glucose that are common to all life. Secondary metabolites vary among plants, but plants would be unable to wage chemical warfare without them. Many biologists believe the feeding activity of insects and other herbivores and the ravages of a host of pathogenic organisms have provided the selection pressure for the elaboration and maintenance of these sophisticated chemical barriers to predation and disease. In such a situation natural selection would have capitalized on the inherent variability in secondary plant chemistry to favor the metabolites that can enhance the organism's Darwinian fitness by affording an effective level of protection. Indeed, in considering the relation betwocn certain insects and plants, Paul P. Feeny of Cornell University has said: "We are thus witnessing an evolutionary arms race in which the plants, for survival, must deploy a fraction of their metabolic budgets on defense (physical as well as chemical) and the insects must devote a portion of their assimilated energy and nutrients to various devices for host location and attack."The point is aptly illustrated by the plant Dioclea megacarpa a vinelike legume. It has only one insect predator: Caryedes brasiliensis. a small beetle of the family Bruchidae, seed-eating beetles of worldwide distribution. One reason the plant is so remarkably successful in repelling other insects is that it stores L-canavanine, an insecticidal amino acid that among other things disrupts the production of normal insect proteins. What is even more remarkable is the adaptation of C. brasiliensis to the chemical defense of the plant.

CANAVANINE AS AN ALLELOCHEMICAL

Our work at the University of Kentucky has yielded some insight into the interactions between insects and the secondary metabolites of higher plants. Canavanine is one of about 250 amino acids that are synthesized by higher plants but are not utilized as building blocks of proteins. This nonprotein amino acid is synthesized by members of the Lotoideae (Fabaceae), a group of leguminous plants, and is found in such agronomically important crops as alfalfa and numerous trefoils, including.clover. It is also synthesized by annual and perennial ornamental and arborescent species such as Wisteria and Robinia. There is considerable evidence that canavanine functions as an important nitrogen-storing metabolite, particularly in the seed, where it supports the growth of the newly developing plant. The molecule L-canavanine is a structural analogue of L-arginine, one of the 20 protein-forming amino acids. In canavanine the terminal -CH₂ group of arginine is replaced by oxygen.

This difference is of little consequence in the metabolism of canavanine, which therefore serves in virtually every enzyme controlled reaction for which arginine is the preferred substrate. For example, canavanine is activated by arginyl transfer RNA synthetase, an enzyme that activates arginine and then links it to its transfer RNA. Once canavanine has been linked to the transfer RNA, which normally transports arginine to the protein-assembly sites on the ribosomes of the cell, canavanine is inevitably inserted into the growing protein chain in place of arginine.

Canavanine is chemically much less basic than arginine, and under physiological conditions it would be less positively charged than arginine. This dillerence in charge can affect the interactions that cause the protein chain to fold up into the uniquely correct conformation of a given protein molecule. There is increasing evidence that proteins structurally altered by the incorporation of canavanine instead of arginine do not function properly. Herein lies an important toxic effect of canavanine on most insects. My colleague Douglas L. Dahlman and I have established experimentally the potent insecticidal properties of canavanine. In our work we employed the tobacco hornworm, Manduca sexta, a lepidopteran consumer of plants such as tobacco and the tomato, none of which store canavanine. Hence M. sexta has lacked the opportunity to adapt to this toxic amino acid and nullify it.

EFFECTS OF CANAVANINE are seen in Manduca sexta, the tobacco hornworm. The abnormal pupal and adult forms resulted from the inclusionof canavanine in the diet fed to the larvae in the laboratory many days earlier. Comparably severe developmental abnormalities are elicited by injuecting canavanine directly into the hemolymph, the circulatory fluid of the insect.

When we injected canavanine labeled with radioactive carbon (C¹⁴) into the hemolymph, or circulatory fluid, of the tobacco hornworm, we found that at least 3.5 percent of the labeled canavanine was incorporated into newly formed insect proteins. Providing the fat body, an internal organ, of the migratory locust Locusta migratoria migratorioides with canavanine results in its incorporation into vitellogenin, a major protein of the fat body that is critically important in egg development. The canavanine-containing vitellogenin of the locust exhibits a higher mobility in an electric field than the natural protein does. This finding reveals a significant change in the physicochemical properties of the protein resulting from the incorporation of canavanine. The canavanine is assimilated exclusively at the expense of the protein's content of arginine.

CARYEDES BRASILIENSIS AND DIOCLEA MEGACARPA

In 1973 I learned of the work of Daniel H. Janzen, a tropical ecologist who is now at the University of Pennsylvania. Part of Janzen's work in Costa Rica involved studies of the interaction of the bruchid beetle Caryedes brasiliensis and the vinelike legume Dioclea megacarpa. My interest in this particular insect-plant interaction developed from Janzen's report that the seeds of the plantcontain a considerable amount of canavanine. (Later I determined that canavanine can account for as much as 13
percent of the dry weight of the seed,exclusive of the seed coat, and for as much as 55 percent of all the nitrogen in the seed.) The accumulation of so much canavanine provides a highly effective chemical barrier to predation: the only known insect consumers of the seeds are the larvae of C. brasiliensis.

Natural History. The female beetle lays her eggs on the wall of the newly ripe fruit of the plant, often near the fissure (i.e. the suture along the pericarp) that is characteristic of legumes. The eggs are protected in structures termed oothecae.The time of laying is late fall, at the end of the rainy season and about a month before the pods mature. D. megacarpa pods are suitable for the deposition of eggs for only a few months. Bruchid beetles do not lay their eggs on matureor rotting pods, and if the females miss the right period, the beetle population must await the end of the next rainy season before the life cycle is completed with the deposition of eggs on the next crop of pods.

The newly hatched larva boresthrough the wall of the fruit and the seed coat and makes a small chamber in the storage tissues of the seed. The chamber is the larval home for the several months required for the larva to grow, pupate and emerge as an adult. As many as 50 of the beetles can develop in a single seed.

SEED OF DIOCLEA MEGACARPA infested with larvae of the brucid beetle, Caryedes brasiliensis. A distinct puparium can be seen in the low end of the infested seed.

The young beetle carves an exit portal in the seed coat and escapes ultimately through the suture along the pod or through an opening created by a vertebrate that tears apart the pod to gain access to the seeds

The adult beetle, released from April through June, feeds solely on pollen and does not ingest canavanine. This seed predator is an ideal system for investigating questions of how insects adapt to the toxicants produced by plants. The ability of the beetle to pursue its specialized feeding habit so successfully results from biochemical adaptations that enable it to cope with the canavanine, which would otherwise be highly poisonous.

As part of a research program, in which I was sponsored by the National Science Foundation, I undertook a long-term, systematic investigation of these biochemical adaptations. As a beginning my associates and I asked whether or not this bruchid beetle incorporates canavanine into its proteins. If it avoids producing aberrant proteins, how does it manage to do so? To answer these questions we injected larvae, obtained from infected D. megacarpa seeds, with radioactively labeled canavanine. Only the terminal carbon atom of the injected canavanine was radioactive. This unique labeling pattern was important because the treatment of radioactive canavanine with the enzymes arginase and urease releases the radioactive carbon atom as carbon dioxide, a gas that can be trapped chemically and then quantified by liquid-scintillation spectroscopy. 

The newly synthesized and slightly radioactive proteins obtained from the larvae were isolated and digested with a strong acid to release their constituent amino acid building blocks. The protein digest was then purified by ion-exchange chromatography, which isolated basic amino acids, such as canavanine. Treatment of the isolated amino acids with arginase and urease failed to cause the release of appreciable amounts of radioactive carbon dioxide. This finding establishcd that the larvae were not incorporating canavanine into protein in significant amounts.

The biochemical basis for the finding became apparent when the arginyl transfer RNA synthetase of this insect was compared with that of Manduca sexta the tobacco hornworm. The enzyme from M. sexta attached both arginine and canavanine to arginyl transfer RNA, but the arginine-activating enzyme from C. brasiliensis attached only arginine. It thus appeared likely that among the biochemical adaptations achieved by the beetle was the development of an arginyl transfer RNA synthetase capable of discriminating between canavanine and arginine. In this way canavanine is not linked to the transfer RNA for arginine and the beetle avoids the production of aberrant, canavanine-containing proteins.

In a recent companion study we gained additional insight into the discriminatory capability of the protein-synthesizing system of the bruchid beetle. The study required the synthesis of a group of radioactively labeled amino acids related structurally to arginine and their testing, along with canavanine and arginine, to determine if any of them were incorporated into the protein of M. sexta and C. brasiliensis. The tobacco hornworm larvae fixed each of the compounds into newly synthesized protein; the bruchid beetle larvae did not.

Many important questions remained, but the one we directed our attention to was how the bruchid beetle larva handies canavanine. It may excrete the substance or in some other way avoid its adverse biochemical effects. An exciting alternative possibility was that the beetle had developed some means of utilizing the toxic natural chemical allomone) as a food resource (kairomone). It was reasonable to think the insect might not sacrifice such an abundant source of nitrogen. (Recall that canavanine contains 4 nitrogen atoms and thus about 30% of its mass is made of nitrogen).

Our first effort was to analyze the canavanine content of D. megacarpa seeds infested with bruchid beetle larvae. We applied a statistical technique ( i.e. regression analysis) to estimate the weight of the seed prior to insect attack from the weight of the intact seed coat. The amount of canavanine in the uneaten part of the seed and in the frass (the fecal matter of the larvae) enabled us to calculate how much canavanine the seed had contained originally. These calculations indicated that more than half of the original seed canavanine was consumed in larval feeding.

This finding still left the question of how the ingested canavanine was metabolized. Arginase, an enzyme that cleaves L-arginine into L-ornithine and urea, also acts on L-canavanine, cleaving it into L-canaline and urea. The enzyme is distributed widely among insects and has several important functions, one of them being to provide ornithine from arginine for the formation of glutamic acid. The bruchid beetle produces arginase and would be able to convert canavanine into canaline and urea:

(put structures here)

Urease, the enzyme that acts on urea, is seldom found in insects. Our analysis of the bruchid-beetle larvae disclosed an extraordinarily high urease activity. The larvae cleave canavanine into canaline and urea and then draw on urease to generate ammonia and urea. In this way half of the nitrogen stored in canavanine becomes available as ammonia for metabolic reactions.

Does the larva utilize the ammonia as a source of nitrogen for its metabolic reactions? Is it not only detoxifying canavanine but also incorporating part of the nitrogen from it into newly made amino acids? These are important questions because with the possible exception of the incorporation of cyanide into L-asparagine, an amino acid of proteins, no other instance is known of an insect utilizing a toxic plant compound for amino acid production.

The commercial availability of urea labeled with the heavy isotope of nitrogen (N¹⁵) and the facile ability of the bruchid beetle larva to convert canavanine into urea made it possible to test the hypothesis experimentally. Larvae were injected with urea containing the heavy nitrogen isotope, and the newly synthesized free amino acids were isolated and converted chemically into a more volatile form. This conversion made it possible to separate the amino acids by gas chromatography. As each compound emerged from the chromatographic apparatus it entered-a mass spectrometer. (These experiments were done at the excellent mass-spectroscopy facilities of the Tobacco and Health Research Institute of the University of Kentucky.) In the mass spectrometer the compound is bombarded by electrons that break the amino acid into a unique pattern of fragments. Computer analyses yield an accurate characterization and compilation of the fragments obtained for each compound. Knowing the relative abundance of the fragments containing the two isotopes of nitrogenmakes it possible to accurately reconstruct the proportions of N¹⁵ and N¹⁴ in the original amino acid.

These determinations revealed a significant incorporation of heavy nitrogen into alanine, glycine, serine, proline, methionine, aspartic acid (and/or asparagine) and glutamic acid (and/or glutamine). They are the protein-forming amino acids that insects can synthesize from suitable precursors. On the other hand, no appreciable heavy nitrogen was found in threonine, leucine, isoleucine, histidine, Iysine or hydroxyproline.They are the protein-forming amino acids that insects cannot synthesize and must get from their food.

L-CANALINE

A further problem is that L-canaline, a nonprotein amino acid produced by the cleavage of L-canavanine into urea, is highly toxic in its own right. It is distinctive in being the only nonprotein amino acid with a free terminal aminooxygroup: -ONH2.

Experiments with larvae of the tobacco hornworm revealed that L-canaline retards growth, causes severe developmental aberrations, increases mortality and interferes with nerve function. Canaline reacts with the aldehyde grop of pyridoxal phosphate, a cofactor essential for the function of certain enzymes.
The canaline-pyridoxal-phosphate complex is stable, and its formation essentially shuts down the catalytic action of certain enzymes that contain pyridoxal phosphate. On first consideration it appears that the bruchid beetle is merely exchanging one poison for another.

By examining the metabolic fate of canaline in the bruchid-beetle larvae we found that the insect has an enzyme capable of cleaving canaline to yield homoserine and ammonia:

A fascinating aspect of these biochemical processes is that the bruchid beetle has adapted to a detrimental substance in its food (canavanine) by functioning in much the same way as a plant synthesizing the same substance. This convergent evolution is revealed by four lines of evidence.

Another conclusion to be drawn from our investigations is that this seed-eating bruchid beetle has achieved many distinct but interrelated biochemical adaptations in its utilization of D. megacarpa as a food source. It is reasonable to propose that among the first of the adaptations was the development of the discriminatory ability enabling the insect to avoid the synthesis of dysfunctional proteins. The bruchid beetles that originally invaded the legume may have simply excreted canavanine and canaline or somehow avoided them. Other nitrogen-rich metabolites stored in the seed (such as proteins) could have satisfied the need of the growing larva for nitrogen. Over a period of time, through natural selection, the insects may have become progressively better equipped to cope with canavanine and canaline and ultimately could utilize the nitrogen they contain.

MOVEMENT TOWARD OBLIGATORY MUTUALISM

Such adaptive successes would have fostered the dependence of the beetle on D. megacarpa. In time the benefits of its association with the plant may have led to the failure to lay its eggs on other seeds in its habitat. In considering the narrow feeding range of C. brasiliensis and the danger inherent in depending on a single food and egg-laying resource, it is important to think of the potential advantages. The insect is relatively free from competition for the seeds and does not have to invest in a capacity to process a wide range of plant toxins. The developing larvae are assured a reasonably safe haven, since the toxicity of canavanine should protect the seed from the larvae of other species.

A final point on which these studies bear is the remarkable ability of insects to adapt to a host of detrimental substances, including the insecticides devised by human beings. Is it possible that microbes living symbiotically in the insect gut have a major role in processing toxic compounds such as canavanine? The seed of D. megacarpa has an array of microbial symbionts and pathogens that C. brasiliensis could have acquired, and they may account for some or even all of the biochemical adaptations required for the utilization of canavanine and canaline. We have isolated microbial symbionts from the beetle larvae and found several that can subsist on canavanine and canaline (or one of them) as the sole source of both carbon and nitrogen.

Our continuing studies of the relation between the bruchid beetle and its leguminous host have provided a basic understanding of how an insect has adapted at the biochemical level to toxins presented by its host plant. It is regrettable that similar biochemical investigations have not been done, so that the present information is limited to this single association. Nevertheless, what we have found is of practical value in helping to understand the response of insects not only to natural toxins but also to chemical control measures. In addition the studies reveal the kinds of insight that can be gained by the union of ecological and biochemical approaches.