QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online May 26, 2006
Journal of Experimental Biology 209, 2344-2361 (2006)
Published by The Company of Biologists 2006
doi: 10.1242/jeb.02244

This Article Summary Full Text Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in JEB Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Garland, T. Articles by Kelly, S. A. Search for Related Content PubMed PubMed Citation Articles by Garland, T., Jr Articles by Kelly, S. A.

Phenotypic plasticity and experimental evolution

Theodore Garland, Jr^* and Scott A. Kelly

Department of Biology, University of California, Riverside, Riverside, CA 92521, USA

View larger version (31K): [in a new window]   Fig. 1. Complex traits, such as behavior, are composed of numerous lower-level (subordinate) traits, themselves interrelated in a strongly hierarchical fashion. In general, natural and sexual selection will tend to act more strongly at higher levels of biological organization, as indicated by the relative thickness of the black arrows. As typically viewed by organismal and evolutionary biologists, selection acts on phenotypic variation (which reflects variation in gene expression), but does not generally act directly on genetic variation (e.g. at the level of DNA sequences). Exceptions to this point can occur via such phenomena as genomic conflict (e.g. Stearns and Hoekstra, 2005).

View larger version (23K): [in a new window]   Fig. 2. Hypothetical example of the effects of positive directional selection favoring individuals with higher values for a particular trait on the mean value of that trait (A) and on the plasticity of that trait or of a subordinate trait (B). (A) The standard expectation for the effects of positive directional selection on the distribution of a trait (for example, heat tolerance) across several generations. During generation one, a selective event – high temperature lasting for several days – kills a majority of the individuals in the population (G1) before they can breed. The survivors (S1) of this selective event then breed and the mean heat tolerance in their offspring (G2) is somewhat higher than for their parents (G1). The difference in population mean phenotype between generations one (G1) and two (G2) indicates that evolution has occurred (assuming that the environment in which the organisms are living has not changed in a way that causes the altered phenotypes via direct environmental effects). This process continues for several generations such that the mean value of the trait in generation five (G5) is substantially higher than in generation one. (B) A hypothesis regarding the correlated evolution of the plasticity of heat tolerance or of a subordinate trait that supports heat tolerance (e.g. expression of heat shock proteins). In the original population, exposure to high temperatures for a few hours or days causes some individuals to increase in heat tolerance (which would probably be adaptive if the high temperatures continued) while an equal number of other individuals actually exhibit a decrease in heat tolerance, which would be maladaptive (inappropriate) if high temperatures persisted. For the population as a whole, the average plastic response is zero. Following a selective event and subsequent breeding of the survivors (S1), which produces the next generation (G2), the average plastic response in this new generation tends to be an increase in heat tolerance. Thus, natural selection has caused an evolutionary increase in both the average `innate' (or `constitutive' or `intrinsic') heat tolerance (A) and a shift in the average plasticity of individuals (B) such that, on average, they become more heat tolerant when exposed (acutely) to high temperatures. This constitutes the evolution of adaptive plasticity. See text for discussion of possible genetic mechanisms of such a correlated response to selection.

View larger version (39K): [in a new window]   Fig. 3. Example of a selection experiment (20 generations) with Drosophila melanogaster (Harshman et al., 1991) in which plasticity evolved to be higher in the Selected lines (S; N=3) as compared with the Control lines (C; N=3). C flies (left) were exposed to either standard medium or lemon for 24 h prior to sacrifice for measurement of a detoxification enzyme, and S flies (right) were similarly treated. Values are means ± s.d. For C lines, the magnitude of the induction caused by lemon exposure, as indicated by glutathionase S-transferase activity on lemon divided by the value on standard medium, was 1.16, whereas for S lines the value was 2.58. The greater induction caused by lemon exposure in the S lines relative to the C lines is an example of a genotype-by-environment interaction. TSO, trans-stilbene oxide. See text for further details.

View larger version (12K): [in a new window]   Fig. 4. Hypothetical trajectories for the amount of voluntary wheel running across a 6-day test period, as implemented in the selective breeding experiment with house mice. (A,B) Examples of greater plasticity in the Selected lines (S; black circles) than in the Control lines (C; gray squares). (C) Similar relative increases in wheel running on a day-to-day basis, but greater absolute increases in the S lines. See text for discussion.

View larger version (13K): [in a new window]   Fig. 5. Wheel running of 48 female mice from generation 23 (Belter et al., 2004), with Selected lines (S) depicted by black circles and Control lines (C) as gray squares. Values are least-squares (adjusted) means from ANOVA, as shown in Table 1. Compare with Fig. 4C.

View larger version (26K): [in a new window]   Fig. 6. Hypothetical relations between a phenotypic trait and the amount of running exhibited during the final week of a multi-week exposure to wheels. In both A and B, it is assumed that mice housed without wheels (not shown) would have values of the phenotype lower than or equal to (about 24) those exhibited by Control mice housed with wheels. (A) The greater phenotypic values for the Selected lines (S; black circles) as compared with Control lines (C; gray squares) are explainable statistically by their greater amount of running (`more pain, more gain'). A real example of this pattern involves the level of brain-derived neurotrophic factor (BDNF) in the hippocampus of S and C mice after 1 week of access to running wheels [see fig. 2 (Johnson et al., 2003)]. (B) There is no relation between the amount of running and phenotype within either group and for a given amount of running the increase in phenotype (relative to the values when mice do not have wheel access) is greater for selected lines than for C lines. Hence, S lines exhibit a greater plastic response. See text for further explanation.

View larger version (13K): [in a new window]   Fig. 7. Hematocrit of female mice from a study reported elsewhere (Swallow et al., 2005). Mice were given wheel access for 8 weeks beginning at weaning. Additional mice (not shown) were housed without access to running wheels, and they showed no Selected (S) versus Control (C) difference in hematocrit. Thus, S mice are more responsive (more plastic) when granted wheel access, but not by virtue of a simple linear relation with amount of running that applies to all animals. Compare with Fig. 6B.

View larger version (15K): [in a new window]   Fig. 8. Complicated hypothetical relations between a phenotypic trait and the amount of running exhibited during the final week of a multi-week exposure to wheels. Mice from Control lines (gray squares) exhibit a positive and linear quantitative relation with the amount of wheel running, but this relation is lost in the Selected lines (black circles). Some traits, such as the amount of neurogenesis in the hippocampus [see fig. 2D (Rhodes et al., 2003)], actually show this sort of complicated pattern.

View larger version (16K): [in a new window]   Fig. 9. Comparison of the effects of 21 generations of selective breeding for high voluntary wheel running (`Evolution') with the effects of 8 weeks of wheel access (`Phenotypic plasticity'), based on least-squares (adjusted) means for hindlimb bones [as presented in table 3 (Kelly et al., 2006)]. Bone lengths were unaffected by either selective breeding or chronic wheel access, but both factors increased bone diameters (one exception), with the former generally having somewhat larger effects. The magnitude of training effects is similar to what has been reported previously for mammalian bone in response to either forced or voluntary exercise (see Kelly et al., 2006).