Changes in the heritability of five morphological traits under combined environmental stresses in Drosophila melanogaster.
Second, environmental conditions may influence the way genes affect a trait irrespective of [V.sub.E]. For instance, under stressful conditions, genetic variability might be increased if new phenotypes appear that are genetically determined. The view that stress increases genetic variability was championed by Waddington (1961) and has been extended by Parsons (1987). In addition, [V.sub.A] may change when organisms are under novel conditions (which are often stressful). Because novel conditions are not normally encountered, there is no history of stabilizing selection to decrease levels of genetic variability (Hoffmann and Parsons 1991). This "selection history" hypothesis has been used to explain an increase in genetic variance in a novel environment (Holloway et al. 1990; Kawecki 1995; Guntrip et al. 1997). The hypothesis has been around for a long time (Mather 1943) and has also been invoked to explain the relative importance of dominance versus additive effects under extreme conditions (Jinks et al. 1973).
Finally, changes in heritability due to the genetic variance may occur because of the way genes control the response to environmental conditions. Gavrilets and Scheiner (1993) propose a general model to account for heritability changes with environmental conditions by separating effects on genetic variation in traits from genetic variation in the slope of a reaction norm. They show that heritabilities will depend both on the environment and the genetic correlation between the trait mean and its plasticity. However, it is unclear if heritabilities increase or decrease with stress under this model. Under some conditions, heritabilities may increase as environments become more or less favorable.
Empirical data on heritability changes with environmental conditions are inconsistent, which is perhaps not surprising given the variety of genetic approaches, traits, and environments that have been considered (Hoffmann and Parsons 1991). Changes in [V.sub.E] have been shown to decrease heritabilities under stressful conditions in plants (Blum 1988), although the data are not entirely consistent (Ceccarelli 1987). [V.sub.E] changes have also been used to explain differences in heritabilities for morphological traits under field and laboratory conditions in Drosophila (e.g., Prout and Barker 1989). Increases in [V.sub.A] under stress have been implicated in several studies. For instance, Sgro and Hoffmann (1998) found that [V.sub.A] for development time in Drosophila increased under some culture temperature extremes. Changes in the heritability of morphology when birds are placed under nutritional stress may also be due to genetic components rather than changes in [V.sub.E] (Merila 1997).
Here we test patterns exhibited by morphological traits in Drosophila melanogaster reared under two sets of conditions. We consider a control environment which has constituted the normal rearing environment of the population for several years, and a stress environment which consisted of a combination of three stresses that substantially reduced fitness as evidenced by changes in viability and development time. We use a combination of stresses rather than a single stress because combinations are often likely to be encountered by Drosophila under field conditions. We show that changes in the environmental variance are sufficient to account for the marked heritability changes (but relatively constant evolvabilities) across environments for wing traits.
The laboratory D. melanogaster stock was founded from 100 inseminated females collected near Melbourne. The population had been maintained as discrete generations by turning over around 800 flies each generation. Cultures were maintained at 25 [degrees] C and continuous light, and had been held in the laboratory on a sugar-dead yeast-agar medium for around 70 generations by the time experiments started.
Heritabilities of five morphological traits were estimates under two environmental regimes. In the control regime, groups of 30 eggs from paired males and females were placed on medium with a normal amount of yeast (3.2%), no ethanol, and without exposure to cold stress. The medium also contained sugar (4.8%), agar (1.6%), and both methylhydroxybenzoate and propionic acid as preservatives. In the stress regime, vials contained 12% (v/v) ethanol and a reduced (1.6% w/v) amount of yeast. Vials were stressed at IC for 90 min immediately after eggs had been transferred, and again three days later when larvae should have reached the second instar stage. In preliminary experiments, this combined treatment reduced viability from [greater than] 70% to [less than] 10% and increased mean development time by more than three days. Eggs for both regimes were obtained by exposing flies to medium with live yeast for two days and then allowing females to oviposit overnight on a small amount of medium (2 ml) contained in a spoon placed in a vial. All vials were placed at 25 [degrees] C for emergence.
When adults eclosed, sexes were separated to ensure virginity, and flies were aged for three to five days. We then paired 155 females with males from a different family (but from the same treatment) and collected eggs from them. We obtained 30 eggs from each pair and transferred these to stress or control medium depending on environments experienced by the parents. Parents were stored in Eppendorf tubes at -20 [degrees] C. Upon eclosion, F1s were aged up to five days, and two males as well as two females from each family were randomly selected and frozen. For the stress treatment, not all vials produced sufficient offspring but all families that produced at least one offspring were kept.
Three wing traits were measured on parents and their offspring. Wing width was taken from the intersection of the wing margin and vein II to the intersection of the margin and vein V. Wing length was taken from the intersection of the wing margin and vein III to the intersection between this vein and the anterior crossvein. Finally, the length of the posterior crossvein was measured. All measurements were made with an image analysis system (Video Trace version 1.6) attached to a Panasonic digital camera (WV-CP610). Correlations between repeat measures for all traits based on 20 females indicated a high degree of repeatability (r [greater than] 0.97, P [less than] 0.001 for each trait). Both wings were measured for all individuals, and measurements were averaged over the two wings. Two bristle traits were measured on both sides of the fly under a dissecting microscope, the number of sternopleural bristles and the number of outer orbital bristles. These bristle counts are normally measured without any error.
Means and Phenotypic Variances
The combined stress treatment substantially reduced egg-to-adult viability, for the controls an average of 76.9% (parentals) and 75.3% ([F.sub.1]s) of the eggs hatched per vial. In contrast, the equivalent values for the stress treatment were 8.4% and 7.2%. Stress also markedly increased development time, extending it by several days.
Means for wing traits in the parental generation were lower in the stress environment compared to the control environment (Table 1). Differences between environments were significant (P [less than] 0.001) for crossvein, wing width and female wing length. In contrast, differences in sternopleural bristle number and male wing length were not significant. For the orbital bristle numbers, the opposite trend was apparent; means were significantly lower for the control treatment than the stress treatment. The same patterns were evident in the F1s (data not shown).
Turning to variances, these were relatively larger in the stress environment for the wing traits (see SDs in Table 2). Differences were significant for crossvein length (females, [F.sub.154,147] = 2.33, P [less than] 0.001; males, [F.sub.152,147] = 2.04, P [less than] 0.001), wing width (females, [F.sub.152,147] = 3.12, P [less than] 0.001; males, [F.sub.151,147] = 1.67, P = 0.001), and wing length (females, [F.sub.153,147] = 2.44, P [less than] 0.001; males, [F.sub.152,147] = 1.94, P [less than] 0.001). Orbital bristle variances were also larger in the stress environment compared to the control environment (females, [F.sub.154,147] = 1.51, P = 0.005; males, [F.sub.152,147] = 1.73, P [less than] 0.001). [TABULAR DATA FOR TABLE 1 OMITTED] In contrast, variances for sternopleural bristle number did not differ significantly between environments (females, [F.sub.154,147] = 1.01; males, [F.sub.151,146] = 1.18).
Of the 155 families set up under each set of conditions, 137 produced offspring for the heritability analysis in the control environment, and 122 were available for the stress environment analysis. We initially undertook regressions for male and female parents separately, but because there were no maternal effects we only present the mid-parental regression coefficients.
Phenotypic correlations among the wing traits were positive and significant in most cases. For instance, phenotypic correlations for females in the parental generation of the controls were all significant (P [less than] 0.001) and were 0.33 (crossvein and wing width), 0.64 (crossvein and wing length), and 0.53 (wing width and length). Genetic correlations estimated from the parent-offspring covariances also tended to be positive. [TABULAR DATA FOR TABLE 2 OMITTED] For instance, for the control environment they ranged from 0.21 (sons, crossvein and wing width) to 0.84 (daughters, crossvein and wing length). This suggests that variation in the wing traits are partly influenced by the same genes affecting wing size. For this reason, we undertook a principal component analysis of the wing data separately. The first component accounted for 72.5% and 75.7% of the variance for females and males respectively. In all cases, wing traits correlated strongly (r [greater than] 0.80) with the components. Subsequent analyses were undertaken on component scores as well as on the individual traits.
In the control environment, regression coefficients for all traits and the principal components differed significantly from 0 and heritabilities (given by the regression coefficients) tended to be in the medium to high range (Table 2). In the stress environment, regression coefficients for the bristle traits were similar to those in the control environment and were always significantly greater than 0. These coefficients did not differ significantly between environments when compared following Sokal and Rohlf (1995), after correcting for [TABULAR DATA FOR TABLE 3 OMITTED] multiple comparisons due to the number of traits measured using the Dunn-Sidak method (Table 2). In contrast, coefficients for the wing traits in the stress environment often did not differ significantly from 0 and were always less than those in the control environments. Differences between regression coefficients were significant for all wing traits (after controlling for multiple comparisons), except for the wing length comparison involving sons.
To directly compare genetic variances in the traits across environments, we followed the approach in Shaw and Billington (1991), and undertook a Restricted Maximum Likelihood analysis (Shaw 1987) with program pcr2 from the Quercus package. In this analysis, female and male parental data were treated separately by including a sex factor in the model. By comparing constrained and unconstrained models allowing the same or separate [V.sub.A] estimates in the two environments, it was possible using the log likelihood statistic to test if [V.sub.A] was constant across environments. These analyses (Table 3) indicated that estimates of [V.sub.A] across environments only differed significantly for length of the crossvein and principal component for sons; in these cases, estimates of [V.sub.A] are significantly larger in the control environment than in the stress environment. [V.sub.A] estimates are similar across environments for the other wing comparisons. However, there are large differences in the [V.sub.e] estimates for the wing traits. In all cases, estimates are much larger in the stress environment than in the control environment. The increase in phenotypic variance in the stress environment is therefore largely due to changes in [V.sub.e] rather than in [V.sub.A], with the exception of crossvein length. In contrast, the two bristle traits show similar estimates of [V.sub.A] across environments as anticipated from the absence of any changes in the heritabilities of the traits in the two environments (Table 2).
We also computed evolvabilities in the two environments. The evolvability of a trait ([I.sub.A]) is computed as [Mathematical Expression Omitted], following Houle (1992), where [Mathematical Expression Omitted] is the mean of the trait before selection. Houle (1992) has argued that this measure provides a better indicator than heritability of the extent to which traits are likely to respond to selection and be influenced by forces maintaining variability. Evolvabilities computed using the maximum likelihood estimates of [V.sub.A] were similar across environments, with the exception of evolvabilities for crossvein length which were lower under stress (Table 2). The similar estimates reflect the fact that [V.sub.A]s were relatively constant across environments.
This study has shown marked changes in the heritabilities of wing traits between environments. Our estimates in the control environment are high relative to those for other morphological traits in Drosophila (Roff and Mousseau 1987), indicating high levels of genetic variability in our base population. The heritabilities decrease to low levels in the combined stress treatment, and this change can be largely explained by changes in environmental variance. In addition, there was a decrease in the additive genetic variance for crossvein length in the stressful environment. These findings suggest that additive genetic variances for morphological traits are not increased under stressful conditions in D. melanogaster. Instead heritability changes for these traits are due mainly to the environmental variance increasing under stress. This largely accounts for the increase in the phenotypic variance seen here and may account for such increases previously noted in other studies on the effects of environmental extremes on morphological traits in Drosophila (Tantawy and Mallah 1961; David et al. 1994; Imasheva et al. 1997).
The fact that stressful conditions do not increase levels of genetic variability and the rate of evolution (as measured by evolvability) of morphological traits is consistent with other data. For instance, in D. melanogaster Sgro and Hoffmann (1998) found that culture temperature extremes did not influence the heritability of wing length, and in mussels Toro and Paredes (1996) found decreasing heritabilities for larval shell length with decreasing food quality. Several bird studies have found that size-related traits tend to have a lower heritability under low growth conditions compared to favorable conditions (Larsson et al. 1997; Merila 1997). Nevertheless, Moed et al. (1997) recently found an increase in variation in body size among D. melanogaster isofemale lines reared under low temperatures and poor food conditions. Perhaps changes in genetic variability depend on the environmental stresses being considered, or else epistatic effects that contribute to isofemale line variation respond differently to stressful conditions than additive effects (c.f., Blows and Hoffmann 1996).
The absence of environmental effects on the additive genetic variance for morphological traits may be relevant to comparisons of heritabilities in field and laboratory environments. Heritabilities are similar in many cases (Weigensberg and Roff 1996), although there are exceptions in Drosophila where heritabilities for wing traits tend to be relatively lower in the field. As in the present experiments, the lower field values appear to be mainly associated with an increase in [V.sub.E] under field conditions, as argued by Prout and Barker (1989) for the heritability of thorax length in D. buzzatii. Interestingly, unlike for wing traits, heritabilities for bristle traits do not necessarily decline under field conditions (Coyne and Beecham 1987; Woods et al. 1998). These field-lab comparisons and our lab results therefore suggest that the effect of environmental conditions on heritabilities varies with different classes of morphological traits.
Other types of traits may show different patterns to those evident for morphological traits. Sgro and Hoffmann (1997) found that some culture temperatures in D. melanogaster can increase heritable variation in fecundity and development time. Westerman and Parsons (1973) compared inbred strains for longevity under radiation stress and found that the additive genetic variance was relatively larger at the extremes. Neyfakh and Hartl (1993) showed that developmental patterns could only be selected at elevated temperatures. However, in cowpea weevils, changes in heritable variation with environmental quality vary for different life history traits (Kawecki 1995).
In summary, our data suggest that stressful conditions do not increase the expression of genetic variance in wing traits and that they have little effect on [V.sub.A] and [V.sub.e] for bristle traits. Stressful conditions do increase the expression of environmental variance in wing traits. Perhaps increases in [V.sub.A] under stress only apply to morphological traits that are normally highly canalized, such as wing venation patterns in Drosophila and vibrissae number in mice (see Hoffmann and Parsons 1997).
We are grateful to F. Shaw for help with Quercus and two anonymous referees for comments on the manuscript. This research was supported by a grant from the Australian Research Council.
BLOWS, M. W., AND A. A. HOFFMANN. 1996. Evidence for an association between nonadditive genetic variation and extreme expression of a trait. Am. Nat. 148:576-587.
BLUM, A. 1988. Plant breeding for stress environments. CRC Press, Boca Raton, FL.
CECCARELLI, S. 1987. Yield potential and drought tolerance of segregating populations of barley in contrasting environments. Euphytica 36:265-273.
COYNE, J. A., AND E. BEECHAM. 1987. Heritability of two morphological characters within and among natural populations of Drosophila melanogaster. Genetics 117:727-737.
DAVID, J. R., B. MORETEAU, J. P. GAUTIER, G. PETAVY, A. STOCKEL, AND A. G. IMASHEVA. 1994. Reaction norms of size characters in relation to growth temperature in Drosophila melanogaster: an isofemale line analysis. Genet. Sel. Evol. 26:229-251.
GAVRILETS, S., AND S. M. SCHEINER. 1993. The genetics of phenotypic plasticity. VI. Theoretical predictions for directional selection. J. Evol. Biol. 6:49-68.
GUNTRIP, J., R. M. SIBLY, AND G. J. HOLLOWAY. 1997. The effect of novel environment and sex on the additive genetic variation and covariation in and between emergence body weight and development period in the cowpea weevil, Collosobruchus maculatus (Coleoptera, Bruchidae). Heredity 78:158-165.
HOFFMANN, A. A., AND P. A. PARSONS. 1991. Evolutionary genetics and environmental stress. Oxford Univ. Press, Oxford, U.K.
-----. 1997. Extreme environmental change and evolution. Cambridge Univ. Press, Cambridge, U.K.
HOLLOWAY, G. J., S. R. POVEY, AND R. M. SIBLY. 1990. The effect of new environment on adapted genetic architecture. Heredity 64:323-330.
HOULE, D. 1992. Comparing evolvability and variability of quantitative traits. Genetics 130:195-204.
IMASHEVA, A. G., V. LOESCHCKE, L. A. ZHIVOTOVSKY, AND O. E. LAZENBY. 1997. Effects of extreme temperatures on phenotypic variation and developmental stability in Drosophila melanogaster and D. buzzatii. Biol. J. Linn. Soc. 61:117-126.
JINKS, J. L., J. M. PERKINS, AND H. S. POONI. 1973. The incidence of epistasis in normal and extreme environments. Heredity 31:263-269.
KAWECKI, T. J. 1995. Expression of genetic and environmental variation for life history characters on the usual and novel hosts in Callosobruchus maculatus (Coleoptera, Bruchidae). Heredity 75:70-76.
LARSSON, K., K. RATTISTE, AND V. LILLELEHT. 1997. Heritability of head size in the common gull Larus canus in relation to environmental conditions during offspring growth. Heredity 79: 201-207.
MATHER, K. 1943. Polygenic inheritance and natural selection. Biol. Rev. 18:32-64.
MERILA, J. 1997. Expression of genetic variation in body size of the collared flycatcher under different environmental conditions. Evolution 51:526-536.
MOED, G. H., G. DE JONG, AND W. SCHARLOO. 1997. Environmental effects on body size variation in Drosophila melanogaster and its cellular basis. Genet. Res. 70:35-43.
NEYFAKH, A. A., AND D. L. HARTL. 1993. Genetic control of the rate of embryonic development: selection for faster development at elevated temperatures. Evolution 47:1625-1631.
PARSONS, P. A. 1987. Evolutionary rates under environmental stress. Evol. Biol. 21:311-347.
PROUT, T., AND J. S. F. BARKER. 1989. Ecological aspects of the heritability of body size in Drosophila buzzatii. Genetics 123: 803-813.
ROFF, D. A., AND T. A. Mousseau. 1987. Quantitative genetics and fitness: lessons from Drosophila. Heredity 58:103-118.
SGRO, C. M., AND A. A. HOFFMANN. 1998. Effects of temperature extremes on genetic variances for life history traits in Drosophila melanogaster as determined from parent-offspring regression. J. Evol. Biol. 11:1-20.
SHAW, R. G. 1987. Maximum-likelihood approaches applied to quantitative genetics of natural populations. Evolution 41:812-826.
SHAW, R. G., AND H. L. BILLINGTON. 1991. Comparison of variance components between two populations of Holcus lanatus: a reanalysis. Evolution 45:1287-1289.
SOKAL, R. R., AND F. J. ROHLF. 1995. Biometry, 3d ed. Freeman, New York.
TANTAWY, A. O., AND G. S. MALLAH. 1961. Studies on natural populations of Drosophila. I. Heat resistance and geographical variation in Drosophila melanogaster and Drosophila simulans. Evolution 15:1-14.
TORO, J. E., AND L. I. PAREDES. 1996. Heritability estimates of larval shell length in the Chilean blue mussel Mytilus chilensis, under different food densities. Aquat. Living Resour. 9:347-350.
WADDINGTON, C. H. 1961. Genetic assimilation. Adv. Genet. 10: 257-293.
WEIGENSBERG, I., AND D. A. ROFF. 1996. Natural heritabilities: can they be reliably estimated in the laboratory? Evolution 50:2149-2157.
WESTERMAN, J. M., AND P. A. PARSONS. 1973. Variation in genetic architecture at different doses of g-radiation as measured by longevity in Drosophila melanogaster. Can. J. Genet. Cytol. 15: 289-298.
WOODS, R. E., M. J. HERCUS, AND A. A. HOFFMANN. 1998. Estimating the heritability of fluctuating asymmetry in field Drosophila. Evolution 52:816-824.
|Printer friendly Cite/link Email Feedback|
|Author:||Hoffmann, Ary A.; Schiffer, Michele|
|Date:||Aug 1, 1998|
|Previous Article:||Does the trade-off between growth and reproduction select for female-biased sexual allocation in cosexual plants?|
|Next Article:||Cryptic female choice: criteria for establishing female sperm choice.|