Proc. Natl. Acad. Sci. USA
Vol. 92, pp. 2-8, January 1995
Colloquium Paper
This paper was presented at a coUoquium entitled "Chemical Ecology: The Chemistry of Biotic Interaction, " organized
by a committee chaired by Jerrold Meinwald and Thomas Eisner, held March 25 and 26, 1994, at the National
Academy of Sciences, Washington, DC.
The chemistry of defense: Theory and practice
MAY R. BERENBAUM
Department of Entomology, 320 Morrill Hall, University of Illinois, 505 South Goodwin, Urbana, IL 61801-3795
ABSTRACT Defensive chemicals used by organisms for
protection against potential consumers are generally products
ofsecondary metabolism. Such chemicals are characteristic of
free-living organisms with a limited range of movement or
limited control over their movements. Despite the fact that
chemical defense is widespread among animals as well as
plants, the vast majority of theories advanced to account for
patterns of allocation of energy and materials to defensive
chemistry derive exclusively from studies of plant-herbivore
interactions. Many such theories place an undue emphasis on
primary physiological processes that are unique to plants
(e.g., photosynthesis), rendering such theories limited in their
utility or predictive power. The general failure of any single
all-encompassing theory to gain acceptance to date may
indicate that such a theory might not be a biologically realistic
expectation. In lieu of refining theory, focusing attention on
the genetic and biochemical mechanisms that underlie chem-
ical defense allocation is likely to provide greater insights into
understanding patterns across taxa. In particular, generali-
zations derived from understanding such mechanisms in
natural systems have immediate applications in altering pat-
terns ofhuman use ofnatural and synthetic chemicals for pest
control.
Irrespective of taxon, the chemicals that play a prominent role
in interspecific interactions are rarely the same substances
used by an organism to meet the daily challenges of living, such
as respiration, digestion, excretion, or, in the case of plants,
photosynthesis. They are, in both plants and animals, of "a
more secondary character" [to borrow a phrase from Czapek
(1)]. These secondary compounds are generally derived from
metabolites that do participate in primary physiological pro-
cesses. In plants, for example, secondary compounds such as
alkaloids, coumarins, cyanogenic glycosides, and glucosino-
lates derive from amino acid; tricarboxylic acid cycle constit-
uents are involved in the formation of polyacetylenes and
polyphenols; glucose, aliphatic acids and other "primordial
molecules" (2) play a role not only in primary metabolism but
in secondary metabolism as well. In insects, many defensive
secretions are derived from the same amino acids used to
construct proteins [among them, quinones in many beetles and
cockroaches derive from tyrosine, formic acid in ants from
serine, isobutyric acid in swallowtail caterpillars from isoleu-
cine and valine, and alkyl sulfides in ants from methionine (3,
4)]. Presumably, secondary compounds are physiologically
active in nonconspecific organisms precisely because of their
secondary nature; it is to be expected that most organisms
possess effective means for metabolizing, shunting around, or
otherwise processing primary metabolites and it is the unusual
chemical that circumvents these mechanisms to cause toxicity.
The publication costs of this article were defrayed in part by page charge
payment. This article must therefore be hereby marked "advertisement" in
accordance with 18 U.S.C. §1734 solely to indicate this fact.
Unlike primary metabolites, which are practically universal
constituents of cells, tissues, and organs, secondary com-
pounds are generally idiosyncratic in distribution, both taxo-
nomically and ontogenetically. Chlorophyll, for example, the
principal photosynthetic pigment, is found in virtually all
species of angiosperms, in virtually all life stages of virtually all
individuals. In contrast, the furanocoumarins are secondary
compounds known from only a handful of angiosperm families
(5). Within a species (e.g., Pastinaca sativa), there is variability
in furanocoumarin content and composition among popula-
tions (6, 7); within an individual, there is variation among body
parts during any particular life stage (8) and temporal variation
in the appearance of these compounds over the course of
development (9); there are even differences in the content of
individual seeds, depending upon their location in an umbel
(10), fertilization history (11), and their position within the
schizocarp (12).
Secondary chemicals are by definition taxonomically re-
stricted in distribution, yet despite this fact there are patterns
in production and allocation that transcend taxa (13). Their
presence in an organism is generally characterized by special-
ized synthesis, transport, or storage. Levels of abundance are
subject to environmental or developmental regulation and,
unlike primary constituents, which may be present in virtually
all cells of an organism, chemical defenses are typically com-
partmentalized, even in those cases in which the chemicals are
acquired exogenously, as when sequestered from a food
source. There often exists a system for external discharge,
delivery, or activation, not only as a means of ensuring contact
with a potential consumer but also as a means of avoiding
autotoxicity until a confrontation arises; and of course these
compounds are almost invariably, by virtue of structure,
chemically reactive (e.g., able to be taken up by a living system,
to interact with a receptor or molecular target, and to effect a
change in the structure of the molecular target). The remark-
able convergence of structural types in plant and insect
secondary metabolites is at least suggestive that the processes
leading to biological activity in both groups share certain
fundamental similarities (14).
Secondary chemicals can be said to be defensive in function
only if they protect their producers from the life-threatening
activities of another organism. Distinguishing between offen-
sive and defensive use of chemicals is difficult, and present
terminology does little to assist in making that distinction. The
term "allomone" is frequently used synonymously with "chem-
ical defense," yet allomones are not necessarily defensive in
function. An allomone has been defined as a chemical sub-
stance beneficial to its producer and detrimental to its recip-
ient (15), so chemicals used by a predator to lure prey (16) are
rightly regarded as allomonal but are not obviously defensive.
By the same token, chemicals that reduce competition for
limited resources, clearly beneficial to the producer, may be
defensive of those resources but are not necessarily defensive
in the life of the organism producing them. Allelopathic
2

Proc. NatL Acad Sci USA 92 (1995) 3
compounds produced by a plant species may increase fitness
of a plant by preempting a resource, such as water or soil
nitrogen, that might otherwise be exploited by other plants
(17), but in the sense that such compounds can kill potential
competitors (such as nonconspecific seedlings) they are used
in an offensive fashion, as for range expansion at the expense
of another organism.
A defensive chemical, then, is a substance produced in order
to reduce the risk of bodily harm. As such, most are poisons-
defined as "any agent which, introduced (especially in small
amount) into an organism, may chemically produce an inju-
rious or deadly effect" (18). This rather restrictive definition
may not be universally embraced by chemical ecologists. On
one hand, the definition implies an interaction with another
organism and, particularly with respect to plants, secondary
compounds may fulfill many functions in the life of the
producer organism other than producing injurious or deadly
effects on other organisms (19, 20). Many plant secondary
compounds, for example, are inducible by UV light and
presumably serve to protect (or "defend") plants from dam-
aging effects of UV exposure (21); by no stretch of the
imagination can such compounds be considered poisons, since
they exert no injurious effects on the damaging agent, the sun.
In this context, they can no more be considered "defenses"
than cell wall constituents can be considered "defenses"
against gravity. On the other hand, some investigators, while
acknowledging the fact that secondary chemicals have dele-
terious effects on other organisms, are reluctant to ascribe
their presence, particularly in plants, to selection pressure
exerted by those organisms (22-24). Calling certain secondary
chemicals "defenses" would be giving credence to the assertion
that they exist only by virtue of the selection pressures exerted
by consumers. Nonetheless, an examination of the distribution,
pattern of allocation, chemical structure, and modes of action
of secondary compounds in a broad cross section of organisms
reveals so many striking convergences and similarities that the
notion that variation in the distribution and abundance of
chemicals that act as poisons results at least in part from
selection by consumer organisms certainly seems tenable, if
not inescapable.
Distribution of Defenses
One line of evidence, admittedly circumstantial, that consum-
ers have influenced the evolution of chemical defenses is their
taxonomic distribution. There are entire phyla in which chem-
ical defenses have never been identified (Table 1). Undoubt-
edly, in many cases this absence of chemical defenses may
result simply from an absence of studies explicitly designed to
discover them-for many small, obscure organisms, life his-
tories, let alone chemistries, are poorly known. This problem
may not be as severe a problem as it might appear, because
chemically defended organisms often call attention to them-
selves byway of aposematic coloration (Table 1) (in fact, it may
well be that effective defenses, particularly chemical ones, may
be a prerequisite for a conspicuous life-style among smaller
organisms). Nonetheless, any reported absence of chemical
defense may be artifactual due to incomplete information.
With that caveat in mind, it is interesting to note that con-
spicuously abundant on the list of the chemically defenseless
are phyla comprised exclusively of parasitic animals. As well,
chemical defenses are absent in entirely parasitic orders within
classes (Phthiraptera and Siphonaptera in the class Insecta, for
example). These organisms are subject to mortality almost
exclusively by their hosts, and poisoning or otherwise severely
impairing a host is unlikely to enhance lifetime fitness of a
parasite (particularly those parasites that cannot survive more
than a few hours without one).
Chemical defenses are also rare in organisms at the top of
the food chain-organisms that are themselves at low risk of
being consumed. Large vertebrates, by virtue of size, speed,
and strength, often occupy that position in both terrestrial and
aquatic ecosystems (carnivores and odontocetes, for example).
Chemically defended mammals include skunks and the duck-
billed platypus, both opportunistic scavengers (32). Among
birds, chemical defense has been demonstrated to date only in
the pitohui (25), which feeds on leaf litter invertebrates (J.
Daly; ref. 79), but likely exists in the conspicuously colored
female hoopoe, which "has a strongly repulsive musty smell
that emanates from her preen gland, and is believed to have a
protective function like attar of skunk" (33). Hoopoes are also
opportunistic feeders that consume debris along with insects
and other invertebrates. It is somewhat surprising that chem-
ical defenses are not more frequently encountered among
small birds, but the absence of reports may be due to the
tendency of investigators to assume conspicuous plumage
results from sexual selection, rather than aposematism and
distastefulness (25).
In contrast with fast, strong predators, organisms with a
limited range of movement, or limited control over their
movements-those that cannot run away from potential pred-
ators-are well represented among the chemically defended
(Table 1). Sessile marine invertebrates are particularly accom-
plished chemists; these include in their ranks sponges, antho-
zoan corals, crinoid echinoderms, polychaetes, bryozoans,
brachiopods, and tunicates (26, 34). Completely consistent
with the pattern is the virtually universal presence of toxins in
plants, ranging from mosses to angiosperms (4), most of which
remain firmly rooted to the ground for most of their lives and
occupy the bottom rung of most food chains. It is interesting to
note that chemically defended taxa tend to be more speciose than
those lacking chemical defenses, but whether this relationship
reflects sampling vagaries or causation is anybody’s guess.
Patterns of Allocation
Secondary chemistry differs from primary chemistry princi-
pally in its distributional variability and it is this variability that
has intrigued ecologists for the past 30 years. Theories [or
provisional hypotheses (35)] to account for the structural
differentiation and function of secondary metabolites, as well
as the differential allocation of energy and materials to
defensive chemistry, abound, but they are almost exclusively
derived from studies of plant-herbivore interactions (Table 2).
This emphasis may be because the function of secondary
chemicals in plants is less immediately apparent to humans,
who have historically consumed a broad array of plants without
ill effects, so alternative explanations of their presence readily
come to mind. The fact that animals upon disturbance often
squirt, dribble, spray, or otherwise release noxious substances
at humans and cause pain leads to readier acceptance of a
defensive function [although there are skeptics who are un-
convinced of a defensive function of certain animal secondary
compounds-Portier (48), for example, reports that "Certains
auteurs voient dans les glandes nucales (of swallowtail cater-
pillars) un appareil d’elimination de substances toxiques ou
tout au moins inutilisables contenus dans la nourriture"]. Why
plants, by virtue of their ability to photosynthesize, should
occupy a unique place in theories of chemical defense alloca-
tion is unclear. Plants produce secondary compounds as
derivatives of primary metabolism; animals do the same. In
fact, plants may be rather unrepresentative of chemical de-
fense strategies as a whole in that they rarely coopt defense
compounds from other organisms via sequestration, although
there are exceptions to the general rule [e.g., parasitic plants
(49-51)].
The relative importance of consumer selection pressure in
determining patterns of production of secondary compounds
varies with the theory. Coley et at (44) suggest that resource
availability and the concomitant growth rate of a plant, more
Colloquium Paper: Berenbaum