Gut Reactions: How Caterpillars and People Disarm Alarming Substances with Cytochrome P540

May R. Berenbaum

A glance through the tabloids at the local grocery store reveals dramatic and surprising new developments in food science and nutrition. According to these publications, ordinary vegetables have the power to cure cancer, promote longevity, improve your memory, and even make you more attractive to the opposite sex. Surprisingly, many of the tabloid reports are based, at least loosely, on scientific studies. Scientists in a broad range of disciplines have come to recognize the remarkable pharmacological properties of phytochemicals--active components of plants that play no known role in the daily physiological life of the plant (that is, they don't offer nutrition or structural support to the plant). Nutritionists and physicians, recognizing that herbs and other plants have been used for millennia for curative purposes, have begun to explore the health-promoting effects of these constituents in detail. However, scholars know that phytochemicals have a "dark side": physiological activity that can be bad for, or downright incompatible with, life. Thus, accompanying the enthusiastic reports of miraculous therapeutic effects, there has been occasional media concern about undesirable complications linked to the incautious use of herbal remedies.

Phytochemists--biologists who studied the phytochemicals in plants long before they were discovered by the popular press--have long known of the detrimental effects of many plant constituents. Over a century ago, German biologist Ernst Stahl suggested that these phytochemicals serve plants in their interactions with the animal world in the same manner that spines and thorns do. In other words, the ecological function of phytochemicals is to protect plants against predation by herbivores. Because they are mostly firmly rooted to the ground, plants cannot escape predators by running away; because they lack a nervous system and muscle tissue, most cannot deter potential consumers by physically subduing them. Thus, the principal defense that plants have against being eaten is the production of toxic substances: phytochemicals. The magnitude of the selection pressure that plants face from consumers is reflected in the diversity of phytochemicals produced. Over 50,000 different kinds of toxins have been isolated from angiosperm (flowering) plants alone.

Although the diversity of phytochemicals is mind- boggling, these diverse plant constituents share several characteristics. They all have taxonomically restricted distribution among plant families (some alkaloids, for example, are manufactured only by a handful of genera in a single plant family); idiosyncratic production within and between individual plants, varying with age, locality, and genotype; production by specialized pathways, with transformations catalyzed by dedicated enzymes; and localization in specialized organs or vacuoles (where physiologically active chemicals can be separated from vulnerable plant tissues).

To eat plants is to eat, along with essential nutrients, a variety of potential toxins. To exploit a particular species of plant as food, then, a herbivore must deal with the particular toxic substances produced by that species. Coping mechanisms vary with the type and size of the herbivore. Large, mobile herbivores (such as mammalian grazers, including humans) minimize the ingestion of any particular toxin by consuming a variety of plants. Due to the idiosyncratic taxonomic distribution of most phytochemicals, a diverse diet of plants virtually guarantees ingestion of a diverse array of phytochemicals, with minimal intake of any single class. Many mammalian herbivores are long-lived, learn, and can modify their behavior; thus, they further reduce their intake of dangerous substances by remembering unpleasant tastes or postingestive experiences and then avoiding undesirable plants. Some mammals even learn by observing the experiences of other members of the species.

To some degree, the fabulous omnivory of humans provides some protection against the bad effects of phytochemical ingestion. Humans are arguably the most omnivorous of all plant-feeding animals. A single slice of pizza, for example, may, depending on topping selection, contain representatives of over a dozen plant families (and, with cheese, pepperoni, and mushrooms, representatives of three kingdoms) in a single bite, exposing the consumer to hundreds of phytochemicals. The enormous variety of the average human diet (with the possible exception of the diet of a fussy four-year-old who refuses to eat anything other than peanut butter and jelly sandwiches) protects against ingestion of potentially dangerous quantities of any group of toxins. Humans can further reduce their exposure to toxins behaviorally by their unparalleled capacity for learning; even reading tabloid newspapers can lead them to change dietary habits.

In contrast with large, long-lived, mobile mammalian herbivores, insect herbivores are small, short-lived, and--particularly in immature, larval, feeding stages--relatively immobile. These characteristics have contributed to the evolution of oligophagy, or dietary specialization, in herbivorous insects, particularly relative to other kinds of herbivorous animals. Many caterpillars, for example, complete their entire larval development inside a single seed or leaf of one plant. It has been estimated that almost 90 percent of all insect herbivores are oligophagous--that is, they feed on three or fewer plant families. Moreover, to support their extremely rapid growth, caterpillars can eat over 4,000 times their body weight in plant food over the course of their development. Thus, most insect herbivores typically encounter, over the course of their lives, a highly predictable and relatively narrow range of plant chemicals, but they eat proportionately huge amounts of these. While many insects can learn, such behavior is less likely to be helpful to short-lived organisms than to long-lived mammalian herbivores.

Although behavior can modify the range and amount of phytochemicals ingested, physiological and biochemical mechanisms help almost all herbivores to disarm potentially toxic materials. Biochemical resistance mechanisms are those that change the chemical structure of the toxin to render it unable to cause damage. Metabolic detoxification is generally a two-step process. In phase one, the structure of the toxin is chemically altered with the aid of enzymes. In phase two, the altered toxin is joined to a carrier molecule for export from the body. Among the most important phase-one systems in herbivores ranging from microbes to mammals are the enzymes known as cytochrome P450 monooxygenases, or P450s. These will be the subject of the balance of my presentation.

Cytochrome P450s make up a large family of hemoproteins. More than 400 genes that code for various P450 proteins have been characterized. Despite the apparent diversity of P450s, several conserved (i.e., constant throughout evolutionary change) structural elements are common to virtually all. The P450 enzymes catalyze a tremendous diversity of reactions, including biosynthetic and detoxification reactions.

Because they are involved both in the biosynthesis of phytochemicals and in their detoxification, P450s occupy a unique role in the interaction between plants and plant eaters. This important position may reflect the remarkable versatility of these enzymes, involving both recognition of the target compounds they modify and regulation of the reading of the genetic code. It is thought that over 50 gene duplication events during 400 million years of evolutionary history have led to the present diversity of P450s. This variety offers organisms important biochemical flexibility.

In herbivorous insects, the variation in the ability to metabolize phytochemicals reflects the evolutionary association between insects and the plants producing the phytochemicals. Oligophagous herbivores, which feed on a narrow range of plants, tend to have higher levels of P450 activity against their host phytochemicals than do polyphagous (generalized) herbivores, which rarely encounter those chemicals. This increased level of activity is probably the result of reciprocal adaptive evolution, or coevolution. In 1964 Paul Ehrlich and Peter Raven examined patterns of diversity to show that herbivorous insects and angiosperm plants interact in a stepwise, coevolutionary way. Plants that have a novel form of phytochemical may gain protection from herbivores that are unable to detoxify the novel chemical. This chemically mediated plant resistance then selects for insect behavioral, physiological, or biochemical resistance mechanisms; insects that acquire metabolic resistance to the erstwhile toxin can exploit a plant resource that is unavailable to others. Thus, the evolution of chemical resistance in plants selects for toxicological resistance in insects, and the evolution of insect resistance then selects for novel forms of phytochemical resistance in plants.

This reciprocal process has been called an evolutionary "arms race." It has contributed not only to the diversification of phytochemicals in plants but also to the diversification of resistance strategies, including cytochrome P450-based mechanisms, in herbivores.

Comparisons of P450-mediated metabolic resistance to phytochemicals show the different challenges faced by large, mobile, polyphagous plant consumers, such as humans, and by small, relatively sedentary, primarily oligophagous plant consumers, such as insects. Although many plant species and plant parts are eaten by insects but not regarded as food by humans (bark comes to mind), some foods are shared by both groups. Indeed, that insects eat plants that humans grow for their own food has been a source of conflict for more than eight thousand years. One group of plants regarded as fodder enthusiastically by both humans and insects is the family Apiaceae, the family to which carrots belong. People and insects routinely consume a variety of tissues from these plants: roots (parsnips, carrots), seeds (caraway, dill), and stems and leaves (fennel, celery). In terms of phytochemistry, the Apiaceae are distinct among plants in that many species in the family produce phytochemicals called furanocoumarins.

Furanocoumarins are typical phytochemicals in that they are restricted in distribution, found in less than a dozen plant families, and structurally diverse in only two families, the Apiaceae and Rutaceae (the rue family, including citrus species). Also typical of phytochemicals is their distribution within plants; furanocoumarin content and composition vary with genotype, plant age, soil conditions, light exposure, and tissue type. There are more than 200 different plant furanocoumarins. Generally, they occur as mixtures, with individual plants producing up to a dozen or more different types, although a few plants may produce only one furanocoumarin.

Plants that produce furanocoumarins present a distinct toxicological challenge to plant eaters. Furanocoumarins can absorb ultraviolet light to form an energy-containing molecule that can react with DNA to form crosslinks and interfere with transcription (the reading of the genetic code), with amino acids to denature proteins, with unsaturated fatty acids to form specialized cyclic organic compounds that disrupt membrane integrity, and with ground-state oxygen to generate toxic oxyradicals that damage essential macromolecules. Thus, furanocoumarins, following exposure to light, are bad; they are broadly biocidal and cause mortality in bacteria, nematodes (threadlike worms), protozoans, insects, snails, fish, birds, and mammals. The degree and mode of toxicity vary with the type of furanocoumarin. Those known as linear furanocoumarins can damage the structure of DNA. In contrast, those known as angular furanocoumarins cannot form crosslinks between DNA base pairs; however, they are mutagenic and toxic to many organisms.

Yet despite their impressive toxicological properties, furanocoumarins are eaten routinely by herbivores. Indeed, many species attacking furanocoumarin-containing plants are oligophagous, restricted to feeding on plants containing these compounds. How is this possible? These organisms have remarkable biochemical means for disarming the toxins.

For example, species in at least three families of Lepidoptera, which include moths and butterflies and their juvenile, caterpillar larval forms, consume plants with furanocoumarins. These species rely on cytochrome P450-mediated detoxification. Substantial differences exist among these species, not only in ongoing, constitutive levels of furanocoumarin metabolism but also in the inducibility of detoxification metabolism triggered in response to furanocoumarin ingestion. In general, the levels of constitutive and inducible metabolism correspond to the frequency of furanocoumarin consumption. For example, within the family Papilionidae--the swallowtail butterflies--ongoing P450-mediated activity against xanthotoxin, a linear furanocoumarin present in both apiaceous and rutaceous hostplants, is high in Papilio cresphontes, a butterfly associated with furanocoumarin-containing Rutaceae, and in Papilio polyxenes and Papilio brevicauda, two butterfly species associated with furanocoumarin-containing Apiaceae. In contrast, P450-mediated metabolism of xanthotoxin is not detectable in Papilio troilus, a species associated with host plants in Lauraceae, laurel trees and their relatives that lack furanocoumarins. Papilionid species outside the genus Papilio that do not feed on furanocoumarin-containing plants (Battus philenor, Eurytides marcellus) also do not metabolize xanthotoxin.

P. polyxenes, the black swallowtail, feeds exclusively on plants containing furanocoumarins and, relative to insects that rarely come across these compounds, can metabolize furanocoumarins at very high rates. This butterfly metabolizes both linear and angular furanocoumarins, although it metabolizes the angular variety, which it encounters less frequently, with activities only one-half to one-third those recorded for linear furanocoumarins. The black swallowtail is one of about 200 species in the genus Papilio, and more than 75 percent of the species in that genus are associated with furanocoumarin-containing plants. Evidence suggests that certain P450s are important in all of these furanocoumarin-associated species.

Although the black swallowtail copes with furanocoumarins by using many behavioral and biochemical means (including inducible antioxidant enzymes and rapid excretion), the chief form of metabolic resistance to furanocoumarins in that species is cytochrome P450-mediated metabolism. Indeed, growth and development in species that feed on furanocoumarin-containing plants are proportional to levels of furanocoumarin-metabolic activity. Two complementary DNAs (i.e., DNA fragments whose structure complements the natural gene) from two alleles, CYP6B1v1 and CYP6B1v2, both induced by xanthotoxin, have been described. Expression of the protein CYP6B1 has demonstrated that these P450s metabolize linear furanocoumarins frequently found in swallowtail host plants but metabolize the less frequently encountered angular furanocoumarins poorly, if at all.

What is known about the molecular regulation of the CYP6B1 gene in the black swallowtail is consistent with a reciprocal relationship with furanocoumarin-containing plants. This gene possesses a regulatory element that is responsive to furanocoumarins and does not respond to other natural or synthetic potential inducers.

The tiger swallowtail, Papilio glaucus, is a species that is an exception to the pattern in which furanocoumarin-metabolizing ability is associated with frequent feeding on furanocoumarin-containing plants. This polyphagous species feeds on tree species in about a dozen families but encounters furanocoumarins in only one of its many hosts, the hoptree (Ptelea trifoliata). Although tiger swallowtails rarely encounter furanocoumarins, individuals in some populations can metabolize xanthotoxin with 13-fold inducibility--a level of inducibility comparable to that of the furanocoumarin specialist P. polyxenes, the black swallowtail. Not surprisingly, the responsible protein is very good at metabolizing the specific furanocoumarins found in the hoptree--the one furanocoumarin-containing plant in the diet of the tiger swallowtail.

At the molecular level, there are two P450 genes associated with furanocoumarin metabolism in the tiger swallowtail. These are 99.3 percent identical with respect to their coding nucleotides. These two genes are close, mapping to within 10 kilobases of each other, and probably arose via a recent duplication event.

Comparisons of amino acid sequences and phylogenetic analyses of known CYP6B proteins suggests that the genes CYP6B1 and CYP6B3 in the black swallowtail share with CYP6B4 in the tiger swallowtail a common ancestor not shared by P450s from species of Lepidoptera in other families.

The presence of conserved sequences in the CYP6B1 and CYP6B4 coding regions also suggest that P450-mediated metabolism of furanocoumarins may be under strong stabilizing selection in an evolutionary lineage that is diversifying together with chemically related plants. Indeed, conservation of host plant utilization patterns, with host plant shifts associated with shared phytochemical traits, is typical of lepidopteran evolution.

Caterpillars are not, of course, the only animals to make meals out of furanocoumarin-containing plants. A range of vertebrate species use cytochrome P450-mediated metabolism for detoxifying furanocoumarins in their food. Goats, chickens, dogs, rats, mice, humans, and bovine rumens all rely on cytochrome P450s for the metabolism of xanthotoxin.

The extraordinarily broad diet of humans ensures that intake of any given furanocoumarin is likely to be very low on a daily basis. Also, virtually no apiaceous plant serves as a dietary staple for humans, at least not in contemporary times. Most plants in this family are consumed as herbs or spices in modest quantity (e.g., parsley, dill, caraway), and even those plants eaten in greater quantity (e.g., fennel, celery, parsnip, carrot) are rarely consumed on a daily basis. Furanocoumarin toxicity as a consequence of ingestion remains a relatively rare phenomenon. When it does occur, it is generally the result of unusual circumstances and is newsworthy. For example, the Archives of Internal Medicine reported the severe photosensitization of a woman who, for reasons not indicated in the article, consumed almost half a kilogram of celery root, drank the water it was prepared in, and then an hour later went to a tanning salon, where she received massive exposure to photoactivating ultraviolet light rays. For the vast majority of people consuming a typical Western-style diet, ingestion of the photoactive furanocoumarins is generally low.

Exactly which P450 is responsible for furanocoumarin metabolism in humans is not clear. Although the study of cytochrome P450-mediated metabolism in humans presents fewer operational challenges than that in insects (which, as very small organisms, offer very little tissue to work with), the evolutionary interpretation of studies of human P450s presents unique difficulties. In striking contrast with insect feeding habits, which are often extremely specific and very conservative over wide geographical ranges (e.g., black swallowtails in New York and in Illinois consume basically the same plants), human feeding habits defy easy generalization and evolutionary interpretation. While there may be differences in the diets of human residents of New York and Illinois, it is just as likely that the diets among New Yorkers vary just as dramatically. Because the human diet is so broad and so profoundly altered from its original form, it is impossible to absolutely identify elements of the ancestral diet that may be associated with the evolution of a particular P450.

There is no question that, in contemporary humans, P450s play an important role in the metabolism of phytochemicals. However, their roles in these metabolic conversions have been studied primarily in the context of drug disposition. Many phytochemicals are used as drugs therapeutically (e.g., taxol for breast cancer, codeine for pain) or recreationally (e.g., tetrahydrocannabinol, nicotine, caffeine). The cytochrome P450 CYP2D6 is enormously important in drug disposition and contributes to the metabolism of over twenty substrates, including phytochemicals (e.g., nicotine, codeine, sparteine) and synthetic drugs (e.g., debrisoquine, nortryptiline, propanolol, captopril, dextromethorphan, 4-methoxyamphetamine). Both interracial and interpopulational polymorphisms in structure and function are found, and there are at least eleven alleles (alternative forms of a gene) in which mutations affect enzyme activity. Some of these mutations decrease or altogether eliminate enzyme activity. Some people who are "slow metabolizers" carry alleles characterized by several mutations; in one variant, the entire gene is missing. Other mutations enhance metabolic rate, producing "superfast metabolizers" who can process substrates at five times the normal rate.

In another example, the cytochrome P450 CYP2C9 contributes to metabolism of tetrahydrocannabinol, the mind-altering compound in marijuana (Cannabis sativa). This enzyme is also involved in the metabolism of several synthetic drugs, including the anticoagulant and rat poison warfarin. This enzyme also displays dramatic interracial differences in rates of metabolism. And another CYP2 enzyme, CYP2E1, which is responsible in part for the metabolism of caffeine--an alkaloid found in beverages produced from the plants Cola nitida and Coffea arabica--displays interpopulation variation and interracial variation even within a single locality. This pattern may account for differences in how people from various ethnic groups respond to coffee and other caffeine-containing beverages.

To date, there has been no systematic investigation of P450-mediated metabolism of phytochemicals with respect to consumption of plant material by humans. The widespread occurrence of variations in enzyme activity levels corresponding to racial or ethnic groups raises the tantalizing possibility that, insofar as ethnicity is associated with dietary food patterns, historical associations with particular diets may have been important in selecting for and in maintaining such polymorphisms. These polymorphisms, which occur at frequencies too high to attribute to genetic drift or to spontaneous mutation, appear to confer no selective advantage in the absence of particular drugs. Differentiation among ethnic groups is consistent with historical differentiation in dietary exposure to particular plant toxins. It is an intriguing possibility that the current distribution of P450 genotypes reflects variation in selection pressures generated by exposure to dietary phytochemicals and their effects on fitness.

The ability to metabolize phytochemicals via cytochrome P450 pathways has clearly evolved in different ways in different organisms. Although many differences exist in the ways in which vertebrate and insect herbivores consume phytochemicals, there are several aspects of cytochrome P450-mediated metabolism of these compounds that are shared. In both mammals and insects, P450s are the chief metabolic enzymes involved in the detoxification of dietary phytochemicals. Also, structural and functional polymorphisms exist in both vertebrate and invertebrate consumers of phytochemicals. Allelic variation is perhaps better documented in humans than in insects, but differences in frequency of polymorphism probably reflect differences in intensity of study rather than fundamental biological difference. Even the small amount of work done on insects reveals polymorphism in phytochemical-metabolizing loci (positions in chromosomes); for example, within the genus Papilio alone, three allelic variants of the enzyme CYP6B1 have been found, and at least two allelic variants of CYP6B4 exist.

As phytochemicals occur unpredictably in the diet because of their idiosyncratic distribution among plants, rates of encounter with any particular compound may vary. Plant populations vary in their phytochemical content, and so also do the frequency and intensity with which particular phytochemicals are encountered. Thus, variations in P450 enzymes may well aid herbivores in coping with the inevitable variability associated with a plant diet. In both humans and insects, cytochrome P450 enzymes are substrate-inducible; that is, synthesis of the enzymes is triggered by the presence of the target substrate. The enzyme CYP1A2, for example, is induced by caffeine and contributes to its metabolism. All known CYP6B genes in Papilio species are induced by least one furanocoumarin, xanthotoxin.

Regulation of enzyme activity via dietary exposure is likely to be adaptive for both polyphagous and oligophagous plant consumers. Maintaining continuously high levels of P450s could be problematical, both physiologically and ecologically. For example, while many P450-mediated conversions are detoxifications, some reactions are bioactivations that enhance toxicity. Also, high ecological costs of ongoing production of P450 enzymes may result from diversion of energy, materials, or resources away from growth and reproduction and into detoxification enzyme systems.

Although similarities are found, there are unquestionable differences between insect and vertebrate P450s. The human P450s involved in metabolism of phytochemicals are, as a rule, broadly substrate-specific. CYP3A4, for example, is known to metabolize over 50 different substrates. On the other hand, the few insect P450s with defined substrate specificities have a narrower pattern of substrate specificity. CYP6B1 and CYP6B4, for example, metabolize only a subset of linear furanocoumarins that are found in plants. This extreme substrate specificity may reflect differences between humans and insects in breadth of diet, intensity of selection, and rates of evolution.

Such differences are to be expected; most herbivorous insects have evolved in the context of a narrow diet of host plants, whereas most vertebrate herbivores (with the notable exception of koalas and a handful of others) have evolved with few restrictions on diet. Differences in the specificity and activity of P450-mediated metabolism in these two groups may be the result of differences in the evolutionary history of association between plant consumers and the plants consumed.

Elucidating and appreciating these differences may reap considerable rewards. Humans and insects have competed for a common food supply since agriculture was invented over eight thousand years ago. The process by which herbivorous insects coevolve with their host plants--the result of a highly intimate association--differs substantially from the process by which humans interact with plants. Humans can manipulate plant evolution via artificial selection in a way that is not reciprocal. Altering plant chemistry for the purpose of suppressing insect populations (as, for example, by breeding plants for resistance or by genetically engineering plants to produce a bacterial toxin) can affect human health through the inevitable encounters between phytochemicals and human detoxification systems.

Understanding and appreciating the differences in P450 structure, function, and regulation in humans and insects may help us develop new approaches to protecting plants against insects and provide insights into ways to use phytochemicals safely in the diet. The tabloids may have to look elsewhere for sensational headlines.

Author's Acknowledgments:

This report is based largely on the talk I presented at the fall Stated Meeting of the Midwest Center. That talk, in turn, was a condensed and popularized version of my paper titled "Animal-Plant Warfare: Molecular Basis for Cytochrome P450-Mediated Natural Adaptation," in A. Puga and K. B. Wallace, eds., Molecular Biology of the Toxic Response (Philadelphia: Taylor & Francis, 1999: 553-71). I thank Roger Myerson and Anne Simon Moffat of the Academy for their support and encouragement in the preparation of this paper.

©1999 by May R. Berenbaum.

Back to the March/April 2000 Bulletin