Adaptive patterns of phenotypic plasticity in laboratory and field environments in Drosophila melanogaster

Abstract

Identifying mechanisms of adaptation to variable environments is essential in developing a comprehensive understanding of evolutionary dynamics in natural populations. Phenotypic plasticity allows for phenotypic change in response to changes in the environment, and as such may play a major role in adaptation to environmental heterogeneity. Here, the plasticity of stress response in D. melanogaster originating from two distinct geographic regions and ecological habitats was examined. Adults were given a short-term, 5-day exposure to combinations of temperature and photoperiod to elicit a plastic response for three fundamental aspects of stress tolerance that vary adaptively with geography. This was replicated in both the laboratory and in outdoor enclosures in the field. In the laboratory, geographic origin was the primary determinant of the stress response. Temperature and the interaction between temperature and photoperiod also significantly affected stress resistance. In the outdoor enclosures, plasticity was distinct among traits and between geographic regions. These results demonstrate that short-term exposure of adults to ecologically relevant environmental cues results in predictable effects on multiple aspects of fitness. These patterns of plasticity vary among traits and are highly distinct between the two examined geographic regions, consistent with patterns of local adaptation to climate and associated environmental parameters.

INTRODUCTION

Adaptation to environmental heterogeneity remains a fundamental issue in evolutionary biology. Many ecologically relevant environmental parameters vary predictably over spatial gradients such as latitude or altitude; similarly, such parameters also change predictably over seasonal time. Such environmental variance can result in spatially and temporally varying selection regimes that can maintain variation in natural populations (e.g., Levene 1953; Haldane and Jayakar 1963). For organisms that rely on environmental cues for their development and life-history strategies, anticipation of future environmental conditions can be essential to fitness (e.g. Kostal 2006; Stinchcombe et al. 2004). Alternatively, the phenotype expressed by a given genotype can be directly modulated by the specific environmental conditions experienced. Phenotypic plasticity, the change in the expressed phenotype of a genotype as a function of the environment (Scheiner 1993), is one such mechanism; plasticity can manifest upon exposure to an environmental stressor and increase performance and fitness after exposure (Hoffman and Parsons 1991). Evolutionarily, heterogeneous environments are predicted to favor individuals that have an ability to modify their phenotype in response to relevant environmental cues (Via and Lande 1985; Scheiner 1993). Thus, adaptive plasticity plays a major role in adaptation to variable or novel environments (Ghalambor et al. 2007; Charmantier et al. 2008; Gomez-Mestre and Jovani 2013).

Plasticity varies widely within and among taxa, and can also operate at different life-stages of an organism. Developmental plasticity is commonly associated with non-reversible phenotypic changes in response to the developmental environment. In contrast, short term adult plasticity is commonly manifested as a reversible response to short term exposure to specific changes in the environment (Wilson and Franklin 2002; Fischer et al. 2003); this short term response represents a form of acclimation, a physiological response resulting from sensory detection of environmental change and subsequent gene-regulated change in phenotypic expression (Wilson and Franklin 2002). Both basal tolerance and acclimation to environmental heterogeneity are important for local adaptation (Gerken et al. 2015). Thus, the adaptation to novel or fluctuating environmental conditions may reflect a combination of genetically determined basal tolerance levels as well as the ability to respond via plasticity over short time scales (Nyamukondiwa et al. 2011); however, the role of short term plasticity in the adaptation to variable climatic environments remains unresolved.

Drosophila melanogaster is an excellent model to examine evolutionary dynamics associated with environmental heterogeneity, as it is a genetic model with an extensive geographic range that spans various habitat types. Latitudinal clines are pervasive, both for fitness associated phenotypes (e.g., James et al. 1997; Hoffmann et al. 2002; Schmidt et al. 2005; Pool and Aquadro 2007; Schmidt and Paaby 2008) as well as allele frequencies across the genome (e.g., Kolaczkowski et al. 2011; Fabian et al. 2012; Bergland et al. 2014). The presumed absence of population structure in D. melanogaster suggests that these clines are driven by natural selection and local adaptation; however, latitudinal clines may also be generated by demography and secondary contact (Kao et al. 2015; Bergland et al. 2016).

While mechanisms generating these observed clines are largely unknown, temperature is a primary determinant of performance and fitness; many clines for thermal-related traits have been identified (e.g., Hoffmann and Watson 1993; Guerra et al. 1997; Hoffmann et al. 2002). In contrast, in various Drosophila species most of the variance in short term, physiological acclimation appears to exist among taxa rather than among populations within taxa or among genotypes within populations (Levins 1969; Hoffmann and Watson 1993). D. melanogaster exhibits rapid, pronounced, and reversible physiological acclimation to temperature exposure in the adult stage (Levins 1969), but geographically distinct populations demonstrate similar acclimation responses in response to both low and high temperature stress (Hoffmann and Watson, 1993). These results suggest that thermal, adaptive plasticity in D. melanogaster may reflect more generalized patterns of physiological acclimation rather than genetic differentiation among populations for the plastic response. However, many previous studies examining the effects of temperature on plasticity have ignored photoperiod as a key covariate. As in many other arthropods, photoperiod affects stress tolerance and life-history traits in drosophilids (e.g., Hori and Kimura 1998; Vesala et al. 2012; Bauerfeind et al. 2014). In temperate habitats, temperature and photoperiod covary across latitude and season and can influence physiology as well as life history traits (e.g., Lanciani et al. 1990, 1992; Bradshaw and Holzapfel 2001; Pegoraro et al. 2014); these parameters also elicit pronounced developmental plasticity (e.g., Giesel 1989; Giesel et al. 1989). Thus, the general role of short term plasticity in local adaptation to spatially varying temperature and photoperiod has not been comprehensively addressed.

If short term plasticity plays a role in local adaptation to the thermal/photoperiodic environment we can make the following predictions: 1) short term exposure to variable environmental cues will result in physiologically mediated plasticity; 2) this physiological plasticity will have predictable effects on performance and fitness; 3) populations that are locally adapted to distinct habitats will exhibit predictable variation in their short term plastic response; 4) this variation in plasticity among populations will be consistent with predictions based on local adaptation (e.g., temperate populations of D.melanogaster that are more cold tolerant will exhibit a more robust plastic response following cold exposure).

To explore the potential role of short term, adult plasticity in the adaptation to distinct climatic environments, we manipulated both temperature and photoperiod and exposed D. melanogaster adult females from two distinct climatic environments to various treatment combinations for a short term duration of five days. We utilized a combination of laboratory and field-based assays to examine variation in the plastic response across traits and among populations from two distinct geographic origins. Our results demonstrate: 1) short term exposure of adults to various temperatures and photoperiods has a pronounced and predictable effect on multiple aspects of performance and fitness; 2) geographically disparate populations exhibit distinct patterns of plasticity; and 3) these patterns of differential plasticity are consistent with local adaptation to thermal/photoperiodic regimes. Together, our results suggest that the short term, physiological acclimation to temperature and photoperiod in this species is modulated by natural selection and may represent an important component in the suite of traits underlying adaptation to environmental heterogeneity.

MATERIALS AND METHODS

Collection and maintenance

Natural populations of D. melanogaster were collected from replicate high and low latitude regions of the east coast of the U.S. by direct aspiration and subsequently established as isofemale lines in the field. High latitude, temperate populations were collected as follows: 1) in October 2009 from Bowdoin, Maine (44.05 °N, 69.97 °W); 2) in October 2009 from Shoreham, Vermont (43.89 °N, 73.31 °W); 3) in October 2010 from Harvard, Massachusetts (42.5 °N, 71.58 °W); and 4) in October 2010 from Bowdoin, Maine (44.05 °N, 69.97 °W). These four independent collections were used to construct two outbred, biological replicates that were created by combining 10 age-controlled cohort females from each of 50 isofemale lines from two populations. The first biological replicate was comprised of 100 isofemale lines from the 2009 Maine and Vermont collections (50 lines per population), and the second biological replicate was comprised of 100 isofemale lines from the 2010 Massachusetts and Maine collections (50 lines per population). After five generations of outcrossing, the two independent biological replicates were then split into experimental replicates and cultured for an additional five generations prior to use in experimental assays. The experimental design is illustrated in Figure S1.

The same experimental design was used to construct both biological and experimental replicate, outbred populations for the low latitude, subtropical geographic region. Populations were collected: 1) in July 2010 from Jacksonville, Florida (30.33 °N, 81.65 °W); 2) in July 2010 from Homestead, Florida (25.46 °N, 80.47 °W); 3) in October 2010 from Macon, Georgia (32.84 °N, 83.63 °W); and 4) in October 2010 from Homestead, Florida (25.46 °N, 80.47 °W). The first biological replicate was composed of 50 isofemale lines from the July 2010 Jacksonville, FL collection and 50 isofemale lines from the July 2010 Homestead, FL collection. The second biological replicate included 50 isofemale lines from the October 2010 Macon, GA population and 50 lines from the October 2010 Homestead, FL collection. All experimental populations (N = 12) were cultured at 25°C, 12L:12D on standard cornmeal-molasses medium in population cages (30cm × 30cm × 30cm; Bioquip Products, Gardena, CA) with a generation time of 21d.

Laboratory manipulations

The experimental conditions in the laboratory environment consisted of four treatment combinations in which the temperature and photoperiod were manipulated using an orthogonal design. Flies were exposed to either 29°C (high temperature) or 14°C (low temperature). An upper limit of 29°C was chosen as it is the highest temperature that does not cause temporary male sterility in laboratory culture (Chakir et al. 2002; Vollmer et al. 2003). The lower temperature was set at 14°C for the cold treatment to avoid expression of reproductive dormancy, which is elicited when flies are exposed to temperatures less than 13°C (Emerson et al. 2009). In addition, flies were exposed to long day (LD, 15L: 9D) or short day (SD, 9L: 15D) photoperiod regimes, representing the extreme seasonal photoperiods for the study site (Philadelphia, PA, USA).

To generate experimental flies, embryos were collected over a 24h period from each of the twelve experimental cages and cultured at 25°C and 12L: 12D. Upon eclosion, virgin females were collected, sorted into groups of five, and held under control conditions for 24h before randomly assigning to one of the four experimental treatment combinations: 29°C LD, 29°C SD, 14°C LD, 14°C SD. Flies were then exposed to the appropriate temperature and photoperiod for a period of 5d, and subsequently directly transferred without anesthesia into glass vials to perform stress tolerance assays.

Field experiment

In addition to evaluating the effects of short-term exposure to temperature and photoperiod in the laboratory, we examined the effects of exposure to variable field conditions on the stress response. Field experiments were conducted over a period of 25 months (June 2012 – July 2014). Flies were cultured as before: embryos were collected over a 24h period from each of the twelve experimental cages and cultured at 25°C and 12L: 12D under control conditions. Upon eclosion, virgin females were collected and held at 25°C for a 24-hour period before releasing them into outdoor cages (Bioquip Products, Gardena, CA). The floor of each cage was layered with grass and leaf litter and experimental flies were provided with 25 ml of water in a beaker and 50 ml of standard medium. Each cage was randomly placed in an experimental garden on the University of Pennsylvania campus in areas without direct exposure to solar radiation. The outdoor treatments lasted for a period of 5d, consistent with the treatment duration in the lab experiments. Temperature was recorded using iButton® device (Maxim Integrated) and the day length data for the five-day treatment was recorded from the National Weather Service website (http://www.weather.gov/) for Philadelphia, PA. At the end of the 5d treatment, each fly was collected by direct aspiration and randomly assigned and split for examination of either heat shock tolerance (N = 1607) or chill coma recovery time (N = 1347).

Field assays were conducted on all experimental populations at 11 timepoints over the duration of the experiment: June, July, October 2012; April, June, July, August, September, October 2013; April, July 2014. If individuals exhibited short-term plasticity and physiological responses to temperature or photoperiod, we hypothesized a positive association between exposure to these experimental variables and subsequent performance in stress assays. Differential, predictable responses between temperate (northern) and subtropical (southern) populations indicate that patterns of short-term, adult physiological plasticity may be shaped by natural selection and contribute to local adaptation to climatic conditions.

Phenotypic Assays

To evaluate the plastic response associated with temperature and photoperiod, we conducted three stress tolerance assays that have been widely used in the examination of climatic adaptation in Drosophila: heat shock tolerance, chill coma recovery time, and starvation tolerance. To evaluate chill coma recovery time, experimental flies from both the laboratory manipulations and field assays were exposed in groups of 5 in empty glass vials to 0°C for a period of 2h (David et al. 1998). Subsequently, vials were transferred to room temperature and allowed to recover; recovery events, defined as the active return to an upright posture, were observed and analyzed using digital video recording (Total N = 1678).

In the heat tolerance assays, experimental females from both the laboratory manipulations and field experiments were transferred in groups of 5 to glass vials. These vials were then immersed in a 37°C water bath for a period of 2h (Schmidt and Paaby 2008). After the exposure to high temperature, flies were transferred to new vials containing standard medium and allowed to recover at room temperature for 1h. Subsequently, survivorship in each vial was recorded by counting the number of flies that were still alive (total N = 2059).

To examine the starvation tolerance response, virgin female flies in groups of 5 were transferred to glass vials containing a water soaked (1 mL) cotton ball (Chippendale et al. 1996). The vials were then placed under control conditions of 25°C, 12L: 12D and mortality was recorded every 3h until all flies had died (Total N = 1658). The starvation tolerance assay was only performed for the experiments conducted for the laboratory-based manipulations of temperature and photoperiod. All data associated with this manuscript are available through the Dryad Digital Repository at http://dx.doi.org/10.5061/dryad.hj5m4 (Mathur and Schmidt 2016).

Statistical analysis

Analyses were performed using JMP v.10 (SAS Raleigh, NC, USA). To examine the among population variation in chill coma recovery and starvation tolerance, a three-way nested ANOVA was performed, with region, temperature and photoperiod as fixed factors; experimental replicate was nested within biological replicate and geographic region, and biological replicate was nested within geographic region. For the heat shock assay we performed a nominal logistical regression, modeling the effects of population, temperature and photoperiod on the log odds of survivorship.

Temperature data were recorded every 10 minutes for all cages in all field assays; these data were subsequently used to calculate mean temperature, absolute maximum temperature, absolute minimum temperature, temperature range, mean maximum temperature, mean minimum temperature and the mean temperature difference across the 5d field exposure for each assay. A Principal Components Analysis (PCA) was performed to determine which environmental variables (the various temperature variables in addition to photoperiod) exhibited the strongest association with variation observed in the stress tolerance assays. Subsequently, we used ANOVA, with geographic region and minimum temperature (Tmin) as main effects, to examine whether the flies from the two sampled geographic regions exhibited equivalent patterns of cold and heat tolerance as a function of temperature exposure in the outdoor field enclosures.

RESULTS

Laboratory Assays

We examined the role of physiological acclimation at the adult life stage of D. melanogaster on variation in the stress response between temperate and subtropical populations of D. melanogaster exposed to distinct temperature and photoperiod regimes. We had predicted that if the perception of environmental conditions and subsequent physiological modification represents a component in the adaptive response to local environmental conditions then we would observe differential patterns of plasticity between the geographic regions consistent with previously documented patterns of baseline, phenotypic differentiation (e.g., Schmidt et al. 2005; Schmidt and Paaby 2008).

Chill coma recovery

As expected, temperature and geographic region were significant predictors of patterns of chill coma recovery in both populations (Table 1): flies from high latitude populations recovered faster than flies from low latitude populations, and exposure to low temperature also resulted in faster recovery from cold at both long and short day photoperiods (Figure 1 A,B). However, the effects of temperature and region exhibited a significant interaction, and also varied significantly with photoperiod (Table 1). Populations from the two geographic regions were phenotypically identical when exposed to short day photoperiods at both high and low temperatures, but were significantly distinct when exposed to long day photoperiods (Figure 1 A,B; Table 2). Similarly, photoperiod had no effect on chill coma recovery in the low temperature treatment, but when exposed to high temperatures, exposure to short days resulted in a decrease in recovery time relative to exposure to long days (Table 2); the combination of high temperatures and long day photoperiods was also the treatment combination that resulted in the largest observed difference between geographic regions. Together, these results demonstrate that both short term exposure of adults to distinct photoperiods and temperatures resulted in physiological plasticity that had a significant effect on the response to cold. Furthermore, the responses between temperate and subtropical populations were qualitatively distinct in some treatment combinations.

TABLE 1.

Effects of geographic region, temperature, photoperiod and the interaction terms on the three stress tolerance phenotypes assayed under laboratory conditions.

Effect tests			Chill coma			Heat shock		Starvation tolerance
Source	N	DF	SS	F	P	L-R ChiS	P	DF Den	F	P
Region	1	1	180198	4.4865	0.0343*	21.698	<.0001*	141.5	6.9767	0.0092*
Temperature	1	1	34610790	861.7248	<.0001*	11.375	0.0007*	1907	224.8572	<.0001*
Photoperiod	1	1	513	0.0128	0.91	5.780	0.0162*	1838	15.9171	<.0001*
Replicate[Region]	2	2	742261	9.2403	0.0001*	5.023	0.0811	136.5	1.2943	0.2774
Cage[Region, Replicate]	8	8	2641925	8.2222	<.0001*	52.528	<.0001*	136.7	0.5549	0.813
Region × Temperature	1	1	2493455	62.081	<.0001*	20.629	<.0001*	1907	23.6726	<.0001*
Region × Photoperiod	1	1	6092861	151.6975	<.0001*	0.521	0.4703	1838	1.161	0.2814
Temperature × Photoperiod	1	1	290304	7.2279	0.0072*	10.318	0.0013*	1851	54.7555	<.0001*
Region × Temperature X Photoperiod	1	1	125219	3.1177	0.0776	15.829	<.0001*	1851	0.0101	0.9197

Figure 1.

Patterns of phenotypic plasticity in the laboratory assays for populations originating from high and low latitude geographic regions. Graphs depict response across the two temperature treatments (14 and 29° C) for long day photoperiods (15L:9D; left panels) and short day photoperiods (9L:15D; right panels). Chill coma recovery time is depicted in A, B; survival following heat shock in C, D; starvation tolerance in E, F. Plotted are means ± s.e.

TABLE 2.

Analysis of variance for the effects of geographic region, Tmin and their interaction on the stress tolerance phenotypes in the outdoor enclosure field assays.

Effect tests			Chill coma			Heat shock
Source	N	DF	SS	F	P	SS	F	P
Region	1	1	0.6	0.00001	0.9954	0.06989091	1.0935	0.3095
Tmin	1	1	247502.17	13.6334	0.0002*	0.08142389	1.274	0.2738
Region*Tmin	1	1	283860.17	15.6362	<.0001*	0.00038219	0.006	0.9392

Heat shock

As was observed with response to cold, both temperature exposure and geographic region had a significant effect on survivorship following exposure to high temperature; photoperiod also affected tolerance to heat. However, the main effects also exhibited significant interactions, including a three-way interaction between temperature, photoperiod, and geographic region (Table 1). Under exposure to long day photoperiods, flies from the high and low latitude regions exhibited parallel response to temperature, with short term exposure to high temperature resulting in elevated heat tolerance, and the temperate populations again being more stress tolerant (Figure 1C). A very different pattern was observed, however, under short day photoperiods: tolerance to heat was equivalent between geographic regions when exposed to low temperature, but survivorship was significantly distinct between regions when flies were exposed to elevated temperature (Figure 1D). The non-parallel responses between regions demonstrate that geographic origin, temperature and photoperiod all had significant effects when flies were exposed to heat shock.

Starvation tolerance

Similar to chill coma recovery and heat tolerance, starvation tolerance was distinct between geographic regions; it was also affected by both temperature and photoperiod (Table 1). Also similar to the patterns observed for the temperature tolerance phenotypes, resistance to starvation demonstrated significant interactions of temperature with region and photoperiod. Following exposure to low temperatures at both long and short day photoperiods, flies from the high and low latitude geographic regions exhibited qualitatively identical starvation tolerance. However, the short term exposure to high temperatures resulted in a significant difference between regions, with the temperate populations exhibiting significantly higher tolerance (Figure 1 E,F; Table 2). This demonstrates that higher temperature exposure negatively affected starvation tolerance in the low latitude populations.

Field Experiment

PC1 explained 67.3% of the variation for chill coma recovery times and 66.1% of the variation for heat shock tolerance (Table S1), which directed further analysis. Of the various temperature variables measured (Tavg (average temperature), Tmax (maximum temperature), Tmin (minimum temperature), Tvar (range of temperature), Tmax_avg (average maximum temperature), Tmin_avg (average minimum temperature) and Delta_T (difference between Tmax_avg and Tmin_avg), Tavg and Tmin were the two variables most strongly associated with observed variance for both cold and heat tolerance in the outdoor assays. We selected Tmin as the predictor variable for subsequent analyses as it is a more ecologically relevant parameter than the average temperature (Kelty and Lee 2001). Each field experimental data point had a unique Tmin value as well as other associated environmental variables.

In the 11 independent assays conducted in the outdoor field enclosures, the minimum temperature experienced by flies over the 5day field exposure had a significant effect on the time to recovery from chill coma; as expected, exposure to lower temperatures in the field resulted in a faster time to recovery from subsequent cold exposure in the laboratory based assay (Table 2.1; Fig 2.1). Additionally, while the main effect of geographic region was not significant, a significant interaction between minimum temperature and region was evident. Flies from high latitude populations exhibited a stronger response to temperature exposure, and recovered faster as a function of decreasing temperature in the field enclosures. In contrast, flies from the low latitude geographic region exhibited a shallower response to minimum temperature experienced in the outdoor field enclosures and did not recover as quickly when exposed to low temperatures in the field. This again demonstrates a non-parallel response associated with geographic origin, in which patterns of physiological acclimation vary predictably between temperate and subtropical geographic regions. In the heat shock assay, there was no effect of environmental exposure in the field on the subsequent survivorship of the flies after treatment.

These differential patterns from the two populations in the chill coma recovery assay are indicative of adaptive plasticity in flies from high latitude population, but not for populations originating from low latitudes. This plasticity in the cold response is an indication of potential cold adaptation in the temperate populations. Similar to the results seen in the laboratory manipulations, field-based experiments suggested that temperature was a major driver of the subsequent plastic response in populations from both geographic regions.

DISCUSSION

Our results demonstrated the following: 1) D. melanogaster populations from two distinct habitats and geographic regions exhibited baseline differences in three aspects of stress response; 2) short term acclimation of adults to both temperature and photoperiod, hallmarks of seasonal change, had a significant effect on the measured traits; 3) these effects of adult acclimation and resulting plasticity were distinct between populations originating from high and low latitude habitats, and patterns of cold tolerance in outdoor field enclosures were consistent with the hypothesis that short term, physiological plasticity plays a role in local adaptation to the climatic environment. In general, the high latitude, northern populations were more stress tolerant: in the treatment combinations in which there was a significant difference between the geographic regions, these populations were more starvation, cold, and heat tolerant. High latitude populations are exposed to extremities of temperature changes over seasonal as well as diurnal time scales; in these habitats, flies also experience larger fluctuations in photoperiod throughout the year in comparison to the more thermally stable environments at subtropical latitudes. The patterns of elevated stress resistance for high latitude populations are consistent with previous work in North American populations (e.g., Schmidt and Paaby 2008). Research on thermal clines in D. melanogaster populations in eastern Australia show an increased high temperature tolerance in tropical populations (e.g., Hoffmann et al. 2002). The difference in patterns exhibited on the two continents may be associated with differences in the severity of the winter and/or high temperature stress, colonization history (Bock and Parson 1981; Kao et al. 2015; Bergland et al. 2016), or a host of additional factors that may affect patterns of local adaptation.

Differences in thermal tolerance between temperate and tropical D. melanogaster populations have been well documented (e.g., Chown and Terblanche 2006; Sgrò et al. 2010). This could be explained in part by genetically based, differential expression of genes that respond to temperature acclimation (e.g., Goto 2001; Sinclair et al. 2007). However, it remains unclear whether thermal plasticity also varies predictably with latitude, and what role different forms of plasticity play in population differentiation and local adaptation (Hoffmann and Watson 1993). We have shown that there are different plastic responses in the way D. melanogaster from geographically distinct populations respond to variation in both temperature and photoperiod, as well as a significant interaction between the two parameters. Additionally, these patterns of short-term plasticity were also observed in response to environmental conditions experienced in outdoor enclosures in the field; our field experiment illuminated responses to natural conditions and examined ecological relevance. Studies that focus on modulating temperature and other environmental parameters in laboratory conditions aim to interpret the results in the context of the environmental conditions that an organism is expected to experience in the field. However, laboratory and field results may not always be equivalent (e.g., Kristensen et al. 2008; Vanin et al. 2012). Our outdoor results demonstrate similar non – parallel responses and reactions norms for populations from the two geographic regions that had been observed in laboratory assays. The patterns observed validate that there is a differential response in these populations of different geographic origin that is trait specific and suggests that this may result from local adaptation. Similar to patterns observed in the lab, the high latitude populations were more responsive to the cold exposure treatment and exhibited faster chill coma recovery time in response to cold temperatures in the field. This response was not observed in populations from the low latitude, southern region.

The observed effects of physiological acclimation to both high and low temperature conditions on stress tolerance are consistent with temperature being a major factor in modulating subsequent performance and fitness in D. melanogaster (e.g., Hoffmann et al. 2005a; Kristensen et al. 2003; Kristensen et al. 2008; Parkash and Ranga 2013; Cossins et al. 2002; Sisodia and Singh 2010, Ballard et al. 2008). Similarly, exposure to distinct, seasonal photoperiods had predictable effects on multiple aspects of stress tolerance. Exposure to short day photoperiods can increase cold tolerance (e.g., Lanciani et al. 1992), whereas acclimation under long day photoperiods can be associated with increased heat tolerance (e.g., Fischer et al. 2012). These responses may be mediated by changes in fundamental aspects of physiology such as basal metabolic rates (Lanciani et al. 1990).

One of the more intriguing aspects of our data is the pronounced interaction between temperature and photoperiod observed for all three traits investigated. Previous studies on short term acclimation on the stress response has largely focused on either temperature or photoperiod independently. This may, in part, explain the discrepancy between our data, suggesting that patterns of acclimation vary predictably among geographically distinct populations, and studies that have not observed such variation among populations for acclimation to such factors as temperature (e.g., Levins 1968; Hoffman and Watson 1993) and desiccation (Hoffmann 1990). The non-parallel responses for the two geographic regions we assayed also suggest that natural populations may exhibit distinct responses when one factor changes but the other remains constant. The combinations of short day photoperiod and low temperature (associated with winter) as well as long day photoperiod and high temperature (associated with summer) are commonly experienced by organisms in temperate habitats; under these acclimation conditions, the phenotypic responses we observed were largely as expected. However, the combination of short day photoperiods and high temperature, representing conflicting seasonal cues, elicited very different responses between geographic regions for heat tolerance. This suggests that climate change may have differential effects on acclimation response and subsequent performance among populations that span a broad geographic range.

Our results are consistent with adaptive plasticity for chill coma recovery, suggesting that temperate North American populations are adapted to the colder and more variable thermal conditions they experience in natural habitats. The mechanism of plasticity in response to these environmental variables is unresolved, and may result from differential gene expression, activation of general physiological stress responses, or reflect differences in metabolic pools and allocation. The existence of adult stage plasticity over short-term exposure to temperature and photoperiod in D. melanogaster make it an excellent candidate to examine the mechanistic basis and principles underlying these responses.

Figure 2.

Phenotypic responses in the outdoor enclosure field assays, plotting chill coma revovery time (A) and survivorship following heat shock (B) as a function of Tmin (in degrees Centigrade) for populations from both the high and low latitude geographic regions. Regression lines are for illustration purposes only.