Coevolution

Introductory article Article Contents

John N Thompson, Washington State University, Pullman, Washington, USA

. The Importance of Coevolution

Coevolution is the process of reciprocal evolutionary change between interacting species, driven by natural selection. It is one of the major processes organizing the Earth’s biodiversity into interrelated communities of species.

The Importance of Coevolution Coevolution and the organization of biological communities The evolution of biological communities is a history of the development of coevolved relationships. Research in recent decades has indicated that coevolved interactions form the basis for much of the organization of communities. Primary succession from rock in terrestrial environments often begins through colonization by lichens, which are coevolved mutualisms between fungi and algae or Cyanobacteria. These mutualisms appear to have arisen on many occasions during the history of life, but they are all based upon a relationship in which the fungi provide a habitat for growth of the symbiont and the algae or Cyanobacteria provide nutrition through photosynthesis. More than 13 000 lichen species occur worldwide, and they are capable of colonizing environments of extreme temperature and desiccation. Similar coevolved symbiotic relationships form the basis of community structure in aquatic environments. Coral reefs, which form the substrate for much of the diversity of life found in tropical ocean environments, develop through coevolved symbioses between corals and dinoflagellate algae. Coral reefs are found in warm, shallow, nutrient-poor regions of oceans, and the photosynthetic algae provide the nutrition used by the corals for growth. The colonization of terrestrial communities by vascular plants and some bryophytes also relies upon complex coevolved relationships, including mycorrhizae and rhizobia. Mycorrhizae are structures formed by associations between fungi and the roots of plants. The exact role of these relationships in plant and fungal growth appears to vary among taxa, but many of them supplement plant nutrition. Unlike mycorrhizae, which are associated with the majority of vascular plants, rhizobia are genetically complex coevolved associations between bacteria and legumes or a few taxa in other plant families. These associations are crucial in plant succession in many communities, because they allow plants to fix nitrogen into a form they can use for growth. The food webs of communities rely upon coevolved symbioses at every trophic level. Plants are eaten by animals, which rely for digestion upon coevolved symbioses with microorganisms in their gut. Similar symbioses

. Coevolution and the Diversification of Life . The Coevolutionary Process . Coevolving with Multiple Species

are required by predators and parasites that feed on other animals. Blood-sucking insects, for example, rely upon intracellular bacterial symbionts to provide B vitamins in their diet. Even the breakdown of decaying animals and plants relies upon coevolved symbioses. When plants decay, they are eaten by detritivores such as termites that rely upon coevolved gut symbionts to break down plant cellulose. In addition to shaping intimate symbioses, coevolution also shapes the organization of interactions among freeliving taxa. Competition for food or other resources has been shown to lead to coevolutionary specialization by competitors for different resources, thereby decreasing competitive interactions. Predators and prey, flowering plants and pollinators, and birds and fleshy fruits are all involved in complex, coevolving relationships that often link groups rather than simply pairs of species. The processes by which these complex, multispecific coevolved relationships develop are still poorly understood, but there is now little question that reciprocal evolutionary change shapes many of these relationships.

Rapid coevolutionary change Coevolution is an ongoing process. It is important not only in the long-term organization of biological communities, but also in the rapid evolutionary dynamics that occur on the time scales of decades. Genes in plants and animals that provide resistance against pathogen attack are now known sometimes to change rapidly in frequency within natural populations over less than a decade as patterns of attack by pathogens change. The frequencies of different genes conferring virulence in pathogens also can change rapidly. When myxoma virus was introduced into Australia in the 1950s as a way of controlling the abundance of introduced rabbits, it was highly virulent. Within a decade, however, evolution of the virus and evolution of increased defence in the rabbit populations led to a coevolved change in the outcome of the interaction, resulting in this case in a drop in the level of virulence. Analyses of coevolution are increasing in importance as ecologists, epidemiologists and conservation biologists have begun to realize how rapidly the outcomes of interactions between species can change. Work on the dynamics of biological communities, studies of emerging diseases, and protocols for design of nature reserves are

ENCYCLOPEDIA OF LIFE SCIENCES © 2001, John Wiley & Sons, Ltd. www.els.net

1

Coevolution

beginning to show that there is no real separation between ecological time and evolutionary time. Although many of the coevolutionary changes occurring over short time scales may not lead to major long-term changes in the interactions between species, they are none the less crucial components of the ongoing dynamics of these relationships. As a result, studies of coevolution have become part of the developing field of applied evolutionary ecology.

Coevolution and the Diversification of Life Some of the most crucial events in the diversification of life have been a direct result of coevolution between species. The evolutionary origin of the eukaryotic cell is partially the result of a coevolutionary symbiosis in which one of the species became obligatory organelles now called mitochondria. The mitochondria allowed aerobic respiration in eukaryotes and therefore subsequent colonization of a wide range of environments. The mitochondrial genes and host nuclear genes eventually became so integrated that they came to function as a single organism, incapable of survival and reproduction without the other. Plants originated as some eukaryotes coevolved with yet another intracellular symbiont, in this case one capable of photosynthesizing. Combining the photosynthesizing capability of intracellular Cyanobacteria with the multicellular complexity of eukaryotes created the opportunity for diversification of plants into a vast array of forms. Subsequent diversification of these coevolved symbioses through changes in the biochemistry of photosynthetic pathways allowed plants to thrive in habitats ranging from hot to cold and sunny to shady. The evolution and spread of sexual reproduction among taxa, which was one of the most important events in the history of life, may also be partially a result of the importance of coevolution in biological communities. There are currently two major views (and a number of subsidiary views) on why sexual reproduction is so common among species. One is that it has been favoured by natural selection as a mechanism of DNA repair. Sexually reproducing females are more likely to have viable offspring than asexual females, whose offspring may all harbour some deleterious mutation inherited from their mother. The other view is that sexual reproduction is favoured through coevolution with enemies, especially parasites. Parasitism may be the most common way of life on Earth, and most eukaryotic organisms are attacked by one or more parasitic species. Many parasites have much faster generation times than their hosts and are therefore potentially capable of evolving much faster. By producing genetically diverse offspring through sexual reproduction, a sexual female stands a better chance than an asexual 2

female of producing at least some offspring with genotypes resistant to attack by well-adapted parasites. If coevolution is partially responsible for the evolution and maintenance of sexual reproduction in species, then it has also augmented the conditions for diversification of life. Sexual reproduction rapidly produces novel genetic combinations, which may subsequently evolve in different directions in different populations. Enemies may drive that divergence, because hosts that differ from nearby individuals and nearby populations may be more likely to survive and reproduce than hosts that are similar to neighbours. Consequently, ongoing sexual reproduction driven by coevolution may be one of the most important causes of the diversification of life.

The Coevolutionary Process Coevolution is a hierarchical process. It occurs among local interacting populations of species, it is reshaped by connections among populations over broad geographic scales, and it creates patterns in the diversification of lineages of species. The local, geographic and species-level scales in the coevolutionary process each generate processes and patterns not evident at other levels in the hierarchy of life. Appreciation of the hierarchical structure of the process of evolution, including coevolution, has been one of the fundamental changes in evolutionary theory in recent decades.

Coevolution within local communities Most species are collections of genetically distinct populations connected by the occasional movement of individuals between them. Extrapolating from published studies of genetic differences among a wide range of taxa, a recent study estimated that species are commonly composed of more than 200 genetically distinct populations. The researchers further estimated that the number of genetically different populations of all species worldwide may be between 1.1 and 6.6 billion. The coevolving interactions between local populations of different species provide the raw material for coevolutionary processes that occur at higher levels in the hierarchical structure of life. The basic unit of coevolutionary change is the local population and its interactions with other species. It is at this level in the hierarchical organization of life that local adaptation between species takes place. Nevertheless, current research suggests that reciprocal selection on local populations of interacting species is often erratic, occurring in some years and generations but not in others. During some years, populations of one of the species numbers may be low or environmental conditions may not favour reciprocal selection on the interacting populations.

Coevolution

The geographic mosaic of coevolution The outcomes of coevolution at the level of local populations form the basis of processes occurring over broad geographic scales. As individuals move among populations, novel defences or counterdefences evolved in one population may become incorporated into other populations. As genes spread among populations, they create a geographic mosaic in coevolving traits. Different populations of each species may harbour different combinations of traits. The geographic structure of coevolving interactions may be further augmented by differences in expression of genes. Genetic traits evolved in one population may be expressed differently when they occur within the different genetic backgrounds of other populations or in the different physical environments in which that population occurs. The geographic mosaic formed by populations harbouring different combinations of genes and expressing them in different ways has the potential to make the coevolutionary process much more dynamic than is evident at the level of any single local population. Only a few studies have so far analysed the ongoing dynamics of coevolution over broad geographic scales for pairs or groups of interacting species, but the current available evidence suggests that much of the dynamics of coevolution may occur at this level in the hierarchical organization of life. The geographic mosaic theory of coevolution suggests that the coevolutionary process has three crucial components that drive ongoing reciprocal change. 1. Natural selection on interactions differs among populations. These ‘selection mosaics’ form the fundamental geographic structure of coevolving interactions. 2. Reciprocal change in an interaction occurs in some communities but not in others. That is, across the geographic landscape there are coevolutionary hotspots, where reciprocal selection shapes an interaction, and coevolutionary coldspots, where natural selection is acting on only one or none of the interacting species. 3. As individuals move among populations and some populations become extinct, the geographic distribution of coevolved traits is continually changing. This ‘trait remixing’ among populations constantly changes the overall coevolved structure of interacting species and the genetic landscape within which future evolution proceeds. Some genes may become restricted to a few populations, while other genes become common, at least for some time, across many populations. The geographic mosaic theory makes three ecological predictions about the coevolutionary process. First, populations of interacting species will differ in their coevolved traits. Second, traits of interacting species will

be well matched in some communities and mismatched in other communities. Mismatches should occur because interacting species differ in how their populations are distributed geographically and because new genes will continually be added from neighbouring populations. Some local maladaptation of populations may therefore result from the geographic mosaic of coevolution. Third, few coevolved traits should spread throughout all populations of interacting species, because few traits will be favoured in all environments. Examples are now accumulating of interactions showing strong geographic mosaics. Wild Drosophila melanogaster flies differ geographically across Europe in their ability to resist attack by parasitic wasps that lay their eggs in larval Drosophila. The wasps, in turn, differ geographically in their ability to overcome the defences of their Drosophila hosts. Wild flax in Australia differs among populations in the combinations of genes that the plants harbour for resistance to flax rust, and the rust differs among populations in its ability to overcome various combinations of resistance genes. Similarly, wild parsnip, an introduced weed in North America, differs among populations in the profile of chemical defences it musters against its major herbivore, the moth Depresaria pastinacella. In turn, some populations of the moth have chemical counterdefences that match the chemical profile of their local host population. Another moth, Greya politella, differs geographically in the plants on which it feeds, and can even differentiate among plants taken from different populations. The plants differ geographically in a variety of traits and adjacent populations with different genotypes differ in the levels of attack by the moths. These kinds of examples suggest that the ongoing dynamics of coevolution that occur across geographic landscapes may be crucial for maintaining the genetic diversity of species.

The coevolutionary process and speciation As coevolution proceeds, it may sometimes lead to the formation of new species as some of the genetically distinct populations become reproductively isolated from the other populations. Local coevolution between species may produce specialized populations with traits so different from other populations that hybrids between them would fare poorly. Through this kind of diversifying coevolution, the hierarchical coevolutionary process may scale up directly from coadaptation in local populations through geographic mosaics of genetically differentiated populations to the diversification of life through speciation. For example, recent work on coevolution between lodgepole pines and red crossbills (Loxia curvirostra) in North America suggests that speciation in red crossbills may have resulted from diversifying coevolution. Lodgepole pine cones are attacked heavily both by squirrels and red crossbills throughout the Rocky Mountains. Over 3

Coevolution

much of the geographic range of these pines, squirrels seem to have been the most important selective agents, selecting for cones that have wide bases that are difficult for squirrels to open. In some outlying mountain ranges, however, squirrels have been absent for at least 6000–10 000 years, and red crossbills have been the major selective agents on cone shape. In these isolated mountain ranges the pines have evolved cone shapes and large sizes that make it difficult for the crossbills to extract seeds. In response, natural selection has favoured in these populations largebilled crossbills capable of extracting seeds from these wide cones. Because the efficiency of seed extraction in these birds depends heavily on their bill size and shape, hybrids between these birds and those from other populations may be at a selective disadvantage. Work on these and other crossbill populations throughout western North America over the past decade has suggested that ‘red crossbills’ are actually a collection of very closely related species, each specialized to different conifer species and, perhaps in the case of lodgepole pine, different populations of a single conifer species. Exactly how coevolution shapes even broader patterns of diversification in species is only now starting to be studied in detail. In 1964 Paul Ehrlich and Peter Raven suggested that one of the ways is through a process now called escape-and-radiate coevolution. Using plants and butterflies as examples, they noted that plant families differ greatly in the chemical compounds that they use as defences against enemies, and butterfly taxa that have broken through those defences have often diversified in species by colonizing the plants harbouring those defences. From those observations, they suggested that coevolution may sometimes proceed through a multistep process involving both adaptation and diversification. It begins with a plant population that evolves a novel defence that blocks attack by its enemies. Now free of its enemies, the plant population enters a new ‘adaptive zone’ where it spreads geographically and diversifies into a number of different species. The plant population has therefore escaped from its enemies and radiated into a new lineage of species. This, in turn, creates a new opportunity for any former enemy population that is able to break through the new defences. A butterfly population with a new counterdefence can colonize the new plant lineage and diversify in species as different populations of these butterflies specialize on different species in the new plant lineage. Once the plant lineage has been colonized, the process begins anew as natural selection favours plants with novel defences that block attack by the new butterfly lineages These repeated cycles of escape, radiation in host species, colonization by enemies, and radiation in enemy species should be visible in the phylogeny of interacting species as starbursts of speciation rather than as a one-toone matching of each speciation event in the host lineage with a speciation event in the enemy lineage. Such one-toone matching, called cospeciation, can be important in the 4

diversification of species linked in obligate symbiotic relationships, but it may also occur in the absence of true coevolution (i.e. actual reciprocal evolutionary change driven by natural selection). Some lice, for example, may speciate along with their mammalian hosts even if the lice and mammals are not imposing reciprocal selection on one another. Lice speciation may simply be a consequence of speciation of their hosts. Therefore, escape-and-radiate coevolution is likely to produce much more complexity in the diversification of life than divergence through cospeciation. The current evidence for escape-and-radiate coevolution comes mostly from studies of interactions between some plants such as milkweeds and associated insects. In these cases, novel innovations in plant defences appear to have led to a radiation in species harbouring those defences and subsequent colonization and radiation in the insects. Nevertheless, no full analysis of the hypotheses has yet been completed, taking into account the phylogenies of the plants and insects, the geographic structuring of the interactions, and ecological consequences of the defences and counterdefences.

Coevolving with Multiple Species Besides the problem of understanding the hierarchical structure of coevolution, the other major current challenge is to understand how species may coevolve simultaneously with multiple species. Almost all species must interact with more than one other species during their lifetimes. The overall evolutionary direction of the traits of a species is the result of the combined selection pressures exerted by these interactions and the additional pressures exerted by the physical environment. Some evolutionary biologists have argued that there are so many selection pressures acting on most species, that coevolution between pairs or small groups of species may be unusual. Three specific views have been proposed as alternatives, or additional processes, that shape the ongoing importance of reciprocal selection in determining interspecific interactions. The hypothesis of sequential evolution suggests that parasites such as insects feeding on plants are often too rare to exert significant selection pressures on their hosts. Hence, much of the diversification of insects, which may include the majority of species on Earth, has come about through insects simply tracking the speciation of their plant hosts. Nevertheless, research in evolutionary ecology since the 1980s has demonstrated that insects can be powerful selection agents on plant evolution, and that parasites in general may be very important selection agents in host evolution. The general hypothesis of diffuse coevolution suggests that species are faced with too many conflicting pressures in their interactions with other species for natural selection

Coevolution

to favour coevolutionary responses to individual enemy or mutualist species. Instead, natural selection favours genes that provide the best evolutionary solution to the sum total of these selection pressures exerted by other species. By this view, natural selection favours general defences that best defend against all enemies rather than specific defences against particular enemies (e.g. the human immune system). Just as with diffuse coevolution, the hypothesis of escalation suggests that species are faced with such a range of conflicting selection pressures that responses to individual species are not of evolutionary importance. But, in addition, the escalation hypothesis suggests that there is a long-term pattern in the evolution of interspecific interactions. As life has diversified, species have been faced with enemies that are increasingly dangerous. Consequently, much of the evolution of interspecific interactions is toward ever greater levels of generalized defence in an increasingly hostile world. Escalation is an important hypotheses about macroevolutionary patterns in the history of life. The hypotheses of sequential evolution, diffuse coevolution and escalation are not mutually exclusive alternatives to strict coevolution involving reciprocal evolutionary change among species. Instead, all the forms of multispecific evolution may have occurred throughout the diversification of life. The problem is understanding which of these scenarios for evolutionary change explain most of the evolution of interspecific interactions. Recent research is beginning to dissect general views such as diffuse coevolution into more testable, specific hypotheses on multispecific evolution involving reciprocal change. For example, the hypothesis of coevolutionary alternation suggests that parasites may coevolve with multiple hosts by alternating among them over hundreds or thousands of years. As the most common host evolves higher levels of defences, natural selection will favour parasites that attack a different host with lower defences. The process may continue until the original host, now unattacked for many generations, evolves lowered defences and again becomes vulnerable. In this way, a parasite may cycle among a group of hosts over millennia. Yet other parasites may coevolve with multiple hosts by cycling among them throughout their life cycle. By specializing during each stage of a life cycle on a different host species, parasites may turn on and off specific genes that allow them to attack and resist defences of each host. Such cycling is common in many parasitic taxa. Other organisms without complex life cycles may also coevolve

with multiple species by specializing on different species at different ages or sizes. In contrast to antagonistic interactions among species, mutualistic interactions may directly collect new species into them as rewards offered by one species (e.g. nectar for pollinators) are exploited by yet other species. In this way, some mutualisms may spread to include new species over evolutionary time. The differences in how antagonistic and mutualistic interactions coevolve in multispecific interactions are only now beginning to be studied in detail within natural communities. The overall question of how organisms coevolve with multiple species remains one of the major problems in developing the theory of the evolution of community organization. Resolving the problem requires approaches to coevolution that take into account the partitioning of interactions throughout organisms’ lives, the potential for coevolutionary alternation over millennia, and the geographic structure of interactions creating the opportunity for different populations of a species to specialize on a different range of other species.

Further Reading Burdon JJ (1997) The evolution of gene-for-gene interactions in natural pathosystems. In: Crute IR, Holub EB and Burdon JJ (eds) The Genefor-Gene Relationship in Plant–Parasite Interactions, pp. 245–262. Walingford, UK: CAB International. Davies NB and Brooke M de L (1989) An experimental study of coevolution between the cuckoo, Cuculus canornus, and its hosts. II. Host egg markings, chick discrimination and general discussion. Journal of Animal Ecology 58: 225–236. Douglas AE (1994) Symbiotic Interactions. Oxford: Oxford University Press. Ebert D and Hamilton WD (1996) Sex against virulence: the coevolution of parasitic diseases. Trends in Ecology and Evolution 11: 79–82. Farrell BD and Mitter C (1998) The timing of insect/plant diversification: might Tetraopes (Coleoptera: Cerambycidae) and Asclepias (Asclepiadaceae) have co-evolved? Biological Journal of the Linnean Society 63: 553–577. Lively CM (1996) Host–parasite coevolution and sex. BioScience 46: 107–114. Margulis L (1993) Symbiosis and Cell Evolution, 2nd edn. New York: WH Freeman & Sons. May RM and Anderson RM (1990) Parasite–host coevolution. Parasitology 100: 89–101. Pellmyr O, Thompson JN, Brown JM and Harrison RG (1996) Evolution of pollination mutualism in the yucca moth lineage. American Naturalist 148: 827–847. Schluter D (1994) Experimental evidence that competition promotes divergence in adaptive radiation. Science 266: 798–801. Thompson JN (1994) The Coevolutionary Process. Chicago, IL: University of Chicago Press.

5