2000/01/01 by ARY A. HOFFMANN, Ary A. Hoffmann, MIRIAM J. HERCUS +1 · 33 citations
Biochemistry, Genetics and Molecular Biology · Agricultural and Biological Sciences · Environmental Science · #Evolution and Genetic Dynamics #Animal Behavior and Reproduction #Animal Ecology and Behavior Studies
paper · pdf · doi:10.1641/0006-3568(2000)050[0217:esaaef]2.3.co;2
Stressful environmental conditions can be defined as those that lead to a sharp reduction in fitness in populations. That is, when changed environmental conditions cause a drastic reduction in reproductive output, and when persistence of the conditions leads to permanent damage, these conditions constitute an environmental stress. Physical stresses that are encountered rarely in populations—such as periods of drought or extreme cold—or that are encountered by a minority of a species—such as in populations that are located at distribution borders or are exposed to local chemical stresses arising from human activities—can, through their direct or indirect effects, lead to marked reductions in the size of populations (Glynn 1988, Hoffmann and Parsons 1991) and to repeated cycles of colonization and extinction (Andrewartha and Birch 1954). Because of their effects on fitness, stressful conditions can be extremely effective in shifting the mean of a trait by imposing directional selection. There are many examples of such shifts in natural populations (Hoffmann and Parsons 1997), including responses to selection arising from human activities, such as the evolution of pesticide resistance in insects and the evolution of resistance to heavy metals in plants and invertebrates (Macnair 1993, McKenzie and Batterham 1994). In addition, there is evidence for rapid shifts in morphological traits due to periodic exposure to climatic stresses, particularly in bird populations. For instance, selection led to increased body size in Darwin's finches after approximately 80% of the population died during a drought (Grant and Grant 1989); selection also increased body size in cliff swallows during an extreme cold spell that resulted in more than 50% mortality (Brown and Brown 1998). However, although stressful conditions can in some instances lead to rapid evolution, stress is not necessarily required to explain the observed rates of evolutionary change in the fossil record and in historic times. Selection experiments performed in the laboratory have shown that most traits can respond fairly rapidly to directional selection, even when selection intensities are only moderate. The experiments of Weber on a range of traits in Drosophila (e.g., Weber 1990, Weber and Diggins 1990) illustrate the large phenotypic changes in morphological and physiological traits that can be achieved through stress selection. Selection responses in such laboratory experiments involve evolutionary rates that far exceed those seen in the fossil record (Gingerich 1983). Thus, moderate or even weak selection (i.e., selection imposed by factors other than stress) may be sufficient to account for observed rates of evolution in the fossil record and in historic times. Periodically stressful conditions may influence evolutionary rates by generating and maintaining variability and by overcoming adaptation limits caused by gene flow, helping to explain diversification patterns in the fossil record In this article, we explore ways in which stressful conditions may stimulate evolutionary change and explain evolutionary patterns other than by simply increasing the selection intensity on some traits. Although stresses can be physical or biotic, we focus on physical stresses. We first argue that evolutionary stasis may be common in populations in the absence of periodically stressful conditions. Populations may be prevented from adapting to environmental changes because of the effects of gene flow, deleterious mutations, tradeoffs, and lack of variability. We then show how stressful conditions can provide a way of overcoming this stasis and promoting adaptive changes. They do so by promoting the expression of variability in traits, influencing adaptive changes by restricting gene flow, and allowing the persistence of genetic variation. Nevertheless, we conclude that the extent to which stressful conditions have contributed to actual evolutionary shifts remains an open question. Natural selection is a powerful force that results in organisms being well adapted to their environments. Evolutionary biology textbooks are full of examples of the power of natural selection in explaining patterns of variation among organisms. Nevertheless, it is also clear that adaptation is often unsuccessful. When faced with a change in environmental conditions, populations may migrate, become extinct, or evolve. Many fossil studies suggest that movement of populations away from the stressful environment is most common, whereas adaptation is least common, at least when evolutionary patterns are discerned from morphology. For instance, Coope's (1979) study of Coleoptera in the Quaternary showed that few beetles exhibited morphological changes over this time period, whereas there were marked changes in the distribution of beetle species, suggesting that species shifted distributions instead of evolving when climatic changes occurred. For mammals in the Eocene and lower Oligocene (37–30 million years ago), when marked climatic change took place, only 3 out of 177 species showed continual morphological changes; for most species there was no morphological change throughout this period, despite shifts in distribution (Prothero and Heaton 1996). There are many additional examples of fossil assemblages that show little morphological change and virtually the same species composition over extended periods (Schopf 1996). Of course, the absence of morphological change in a lineage does not necessarily imply the absence of evolutionary change, particularly because physiological adaptation can occur independently of morphology and therefore not be reflected in the fossil record. One way of discerning the extent to which physiological evolution has occurred is to consider whether changes in distributions involve entire species assemblages (Coope 1979). If so, then large-scale physiological evolution seems unlikely because not all species possess the same ability to adapt physiologically. Schopf (1996) and Coope (1979) emphasized that groups of organisms in the fossil record tend to occur in the same assemblages despite environmental changes and that even the dominance of certain species appears constant over an extended time. However, other authors have emphasized that species tend to respond individually rather than as an assemblage; such responses may reflect physiological adaptation without morphological change (Graham 1992, Nowak et al. 1994). Limits to morphological and physiological adaptation are evident in extant populations as well as in fossil studies. Although many pest species have evolved resistance to agricultural chemicals, others have failed to evolve resistance, even when chemicals have been applied for many years. Limits are also evident at species borders, which have tended to be relatively constant over many decades (Bull 1991). In the absence of evolutionary limits to adaptation, borders would be expected to continually shift in response to evolutionary changes in geographically marginal populations. Instead, ecological range expansion appears to be rare. Given the evidence for adaptive limits, how can the apparent stasis of fossil assemblages and the inability of populations to overcome marginal conditions be explained? One possibility is that there is insufficient genetic variability for adaptive changes to occur. This hypothesis appears to run counter to both the efficacy of laboratory selection in changing population means and the high level of genetic variability that is common in natural populations (Nevo 1988). Nevertheless, a limitation of laboratory selection experiments is that they have tended to focus on traits that can respond to selection. Many traits exhibit an extremely low level of variability and do not respond readily to selection. Some examples include floral morphology in Linanthus, vibrissa number in mice, and embryonic developmental rates in Drosophila (see references in Scharloo 1991, Hoffmann and Parsons 1997). Moreover, even when genetic variability is present in populations, it may not be used. For example, it may be of the wrong type, as illustrated by the evolution of insecticide resistance (Roush and McKenzie 1987). In field populations of insects, resistance is usually determined by a single gene. However, there are situations in which resistance has not evolved or has not persisted, despite the presence of genetic variability. In such cases, it appears that resistance is determined by several genes, each of which has relatively small et al. high intensity of selection an for the that resistance does not evolve when it is determined Selection for resistance is in the mortality to to effective of When the selection intensity is extremely only may have high resistance to pesticide whereas weak selection all can has been for the evolution of resistance to other such as heavy metals (Macnair 1991). That is, resistance is more to evolve when it is determined by than when it is determined by several are also more to resistance to chemical stresses than are because of gene (Roush and McKenzie 1987). When a population that been exposed to chemicals, the of the of resistance This is to be when resistance from a of many (i.e., than when it from a single gene. When with resistance with the gene for resistance be in the and the of with resistance is therefore This has a lower when resistance is by or a few For instance, when a single resistance, of the of a and be for the resistance However, when are only in be Thus, for traits variation in populations is it may be to a response to selection, even when genetic variation may also adaptive among populations. The of gene in adaptation in a population has been for a number of traits and organisms including variation in and in size in and in Moreover, even when selection among populations, these may not has that local change over in with gene flow, populations. may only reproductive populations. The of gene in adaptive changes is to be particularly at the of a and and and have emphasized that natural selection as a increasing fitness in the encountered by organisms not in more marginal environments. conditions are by the minority of of species, particularly because the of species the a as for populations. populations therefore to adapt to marginal conditions, well adapted only to common conditions away from the Moreover, the effects of gene in adaptation can be by In increased fitness in is with fitness in environments. For instance, resistance to and stresses in Drosophila is often with an in and or with a in (Hoffmann and Parsons 1991, et al. et al. 1996). physiological changes lead to a in an in or the fitness of organisms conditions. There is evidence for such in insects and other (Hoffmann and Parsons 1991, 1997), although few field have been In the ability to in is often with low which ability field conditions and The that marginal conditions are often with low fitness conditions, and the of gene on the inability of marginal populations to evidence that the effects of deleterious may be in only some and as a that are not deleterious in the environment may be so in a environment and to environment Although selection deleterious mutations, this is more in that are encountered by organisms than in those that are rarely Thus, that are in a environment to a in the of populations to marginal conditions et al. 1997). The of all of these is a population that is well adapted to common conditions adapted to conditions or to conditions that are by a minority of of a That is, although natural selection can be extremely effective in genetic effects are often by gene and fitness among environments. that evolutionary change in a population are in The of with on evolutionary changes in a include gene from populations, an in the of deleterious mutations, and the presence of traits and environments. can also be by a lack of genetic variation. the of that stressful conditions can also have a on evolutionary change, as we of that can have or on rates of adaptation in a and the of stress on these for The that evolution and large phenotypic shifts because of with large effects was by and by some developmental However, this has been by as because with large effects are often rapidly in populations due to their low Nevertheless, it does that large phenotypic shifts can be in populations by repeated exposure to stressful conditions. has been that stresses can lead to the increased expression of phenotypic variability particularly in morphological traits that are examples include the of the and of patterns in Drosophila by exposure to chemical and stresses can be for and in in a population as as they have a genetic and they can become even in the absence of the a to as In the of phenotypic in which a is by the changes this have been to involve in the gene and 1996). of stress on variation. variation can be increased rates of and can also the expression of variation at the phenotypic level by for the expression of traits, by influencing or or by other by which genetic occur a a phenotypic is only a certain level of which then the to be these in they in gene in which results in the expression of the phenotypic in the absence of the stress. Thus, stressful conditions can to phenotypic that can then be in populations, to evolutionary to account for genetic selection, a phenotypic is stressful conditions when a genetic is not conditions, which are with a The genetic is determined by the number of present that lead to the expression of the selection for the this leads to an in the of the for the expression of the The distribution of the population is and the is even when stress is not One with on these experiments is that stresses chemicals or applied to were required to phenotypic shifts by genetic However, and have a more for morphological stress. They the effects of in of the Drosophila genes, on morphological This gene for the which to that in The authors showed that these well as a of morphological variation that can then be exposed to morphological are their expression no on the presence of The by which this occur is as stress may cause a in because these are by The reduction in would then lead to the increased expression of morphological the would be even in the absence of changing that the expression of This is to the for genetic in that environmental lead to increased variability. and that this rapid morphological However, as in the of genetic this would the morphological to have a fitness In addition, that evolutionary change, such as gene flow, deleterious mutations, and traits and would to be Stressful conditions can also influence evolution by increasing and In the organisms exposed to stressful conditions can more than organisms not exposed to such conditions because of an increased of and of 1988, Hoffmann and Parsons 1991). There is also evidence that rates natural populations exposed to of stress. For instance, et al. rates in a from of a The is and more than the which and relatively constant conditions. rates were in from the than in from the some of this variation was a was through of Thus, both and rates were in from the more stressful that in natural populations may also be with stress from the in the increasing stress is with increasing of as by (Nevo et al. 1996). in the rates of both and rates of evolutionary When populations are directional selection, there is evidence that a in the response to selection Moreover, increased rates are also with rapid evolutionary responses in the laboratory and Thus, in and rates stress be an adaptive response by organisms. However, changes in and rates stress are not necessarily Some changes in have been with the of that the of a and gene There is evidence that the of some of environmental stress et al. In addition, the in and rates may be an indirect of organisms for when they are stress. Moreover, as well as are with an in Because most are an in a in a One way that organisms this is a that that fitness in response to the stressful conditions an is that such has from that have and are therefore to without a a the the of a that these been and 1993, 1998). The is that are not in the of arising to with environmental conditions. Instead, it appears that stressful conditions rates to a extent in of that not include those selection. number of explain this For instance, has that stress the of of some genes, particularly those these are in more to because that is is particularly to as a of being single for of the time. Although these results and others (see Hoffmann and Parsons that stress phenotypic and genetic there remains a large these effects and that they have a in adaptive all of the in laboratory experiments are unlikely to in In addition, a few by stresses are to be by gene in populations, the have large effects on the fitness of organisms. One way to explore whether stress has more effects on phenotypic variation and on adaptive change is to consider stress effects on variation in traits to be selection. this a number of experiments have the of morphological and traits both stressful and conditions. the of variation in a trait that to phenotypic and their it is therefore an for the effects of natural selection is defined as the genetic is and the phenotypic is is the of a which is defined as is the genetic which as well as genetic The of a trait an of the of phenotypic that is this is than for evolutionary change because not all of genetic variation to the and If the of a trait directional selection is increased by stressful conditions, then that trait is to evolve more rapidly than a trait is not increased by stress. the results of experiments this have been (Hoffmann and studies in Drosophila suggest that the of traits may be increased high or there are also to this particularly in and suggest that both the and of morphological traits may conditions. For traits, the are also In a Drosophila study that the effects of a stress arising from a of cold and and Hoffmann variation for increased stressful conditions. However, other studies have shown a in the of and time stressful conditions 1991, et al. 1998). One limitation of studies on stress and is that they have tended to focus on traits with moderate and fairly high of phenotypic variation. on with low and low phenotypic more changes in with stress. would also be to whether traits that have a selection (i.e., traits that to respond to directional can to evolve after stress such experiments to be traits that are to have adaptive shifts in natural populations rather than morphological traits of adaptive the effects of stress on variation suggest a more the environment and genetic than has been This environmental is also apparent in and 1998). are with and that are by environmental changes to be of include the many by The environment can influence the patterns of these genes, and their in their patterns can be and many The environment therefore not only on variation natural selection also to the expression of genetic and phenotypic that can then be exposed to selection. In some can in stress a on evolutionary rates When a stressful environment is it can an in the of and all of in can lead to increased phenotypic variation exposed to selection. Moreover, stress can the expression of variation in traits that can then be exposed to selection. In to influencing of stress may also influence rates of evolution by the effects of gene on Stressful conditions are to gene among populations because these conditions both in species to and in (i.e., because of a reproductive with a high fitness in marginal conditions may then in because they are no being to a on and have a in which a trait is selection in a species the of a They showed that populations become selection can a few to shift the population mean several away from that of the restricting gene can have a on there have been few to the of this many examples of gene restricting In addition, there is direct evidence that restricting gene can lead to rapid evolutionary not in the of stress. For instance, that from a to in that they showed low of even were and high of were therefore The authors that the evolution of was by gene with a population in the an environment were and low of were therefore When gene was the and populations for a in the population increased That is, the of gene an adaptive The effects of stress on gene to account for the high rates of evolution and diversification that are often seen in environments. include the in and the in 1994). are with from periodically stressful conditions, which in are to have led to marked changes in the of and size of populations. However, it is also that other direct and indirect effects of stress have contributed to diversification in such The effects of stress on gene and population size can have as well as If populations become then genetic variation can become and is expected to The only is which can in a of genetic variability a population 1988). small population size also lead to which when there is an in the of for deleterious in a There is evidence from agricultural studies that is increased stressful conditions. This is also apparent in more experiments with other including and plants et al. et al. 1998). In stressful conditions may rates of adaptation when gene is by the of populations adaptive responses there is a reduction in population size is to whether adaptive changes are by a reduction in gene For instance, it would be to whether populations at the of species distributions results in their adapting to marginal conditions. Periodically stressful conditions can evolutionary rates by the expression of fitness among and particularly in have shown that variation among for to have no fitness conditions, whereas fitness among these become evident more stressful conditions et al. In addition, it appears that the the of at several and fitness, which is often weak can become moderate of stress. include an in the of on the fitness of the stress and and on the fitness of the and stress and The fitness effects by stressful conditions can influence the persistence of genetic variation in populations and evolutionary effects on persistence can occur in more environment When at a are in a Because the and fitness to be stressful conditions than stress may the persistence of genetic variation. However, can at a only fairly conditions and Stressful conditions can also persistence when environment in relatively fitness in and The evidence for fitness in and as that often change that include stressful conditions. In addition, there is evidence that stresses to some and 1998). There are several ways in which stressful periods may have a in evolution by extinction from Hoffmann and Parsons in the of in species to evolutionary of ecological for adapted expression of increased genetic and phenotypic increasing rates of evolutionary in gene of to If stressful conditions to genetic there be an of genetic variation and the of populations such conditions. and have evidence of this in a number of organisms (Nevo 1998). For instance, in is with increasing from a an and (Nevo et al. 1994). However, it be emphasized that there are to be as well as effects of stressful conditions on genetic variation. If these conditions and population are of genetic variation may be genetic the adaptive of a Thus, some of the in as well as effects of stressful conditions on genetic variation. populations are rapid to account for rates of evolutionary change in the fossil record (Gingerich without the to a for stressful conditions. Nevertheless, although patterns in the fossil record are to there is some evidence that periodic stresses evolutionary changes and to counter evolutionary There is particularly from that rates of evolutionary change environments. tend to in that are both and In that tend to be relatively constant or in which stressful conditions are present and is low organisms that have only morphological changes. from from the and far more in to the was not than in the more more The then to the other well as being with the of evolutionary stress has also been to patterns of evolution in the fossil record. is whether evolution patterns which periods of rapid change are by periods of or patterns of evolutionary has that these patterns on the of environmental changes. When environmental conditions change and the environment does not (i.e., periodically stressful conditions are evolutionary changes be whereas stasis rapid evolutionary change occur when extremely stressful conditions have been in promoting evolution because of the evolutionary diversification that often extinction Although these often out a large of all species, they tend to be by the of and periods of rapid number of (see this have been to explain how evolutionary diversification occur Hoffmann and Parsons 1997). One possibility is that stressful conditions among particularly and because population are and many populations become This may evolutionary to stressful conditions. possibility is that that are after a extinction can become by evolutionary that are by the presence of other organisms. In addition, the direct effects of stress in generating phenotypic by the may be during periods of these the stresses to extinction have a in that Although fossil patterns suggest that evolutionary change is or by stressful conditions, it is to fossil patterns with stress effects on patterns in extant populations because of the time That is, because rapid changes in the fossil record would as relatively changes in extant populations, it is to these However, it does that evolutionary shifts are in some of that are relatively and to be effects of stress on phenotypic variability and indirect effects on gene may well a in promoting those this the of how stressful conditions evolutionary change is conditions can influence the expression of variation at the and phenotypic and the expression of fitness among However, it remains to be seen such can be to the of stresses that occur in as well as to the of adaptive among populations. The focus on the genetic of among species to at least the of the in adaptive among species and their to such as genetic The of the indirect effects of stress on evolutionary change be to because of the that stress can cause both and effects on evolutionary If population small due to there is a that of genetic variation can be and effects a reduction in gene stressful conditions may rates of One way of this is to experiments to whether evolutionary limits in natural populations are with gene Populations at species provide with which to such is to of the being imposed on natural populations because of human to this adaptive changes in populations. on the responses of organisms to stress in laboratory is by the We and for the to this