Local adaptation of stress related traits in ... - Wiley Online Library

doi:10.1111/j.1420-9101.2009.01725.x

Local adaptation of stress related traits in Drosophila buzzatii and Drosophila simulans in spite of high gene flow P. SARUP, J. FRYDENBERG & V. LOESCHCKE Aarhus Centre for Environmental Stress Research (ACES), Ecology and Genetics Department of Biology, University of Aarhus, Aarhus C, Denmark

Keywords:

Abstract

clinal variation; cold shock; FST; heat shock; Hsp70; knock-down resistance; male sterility; marginal populations; starvation; thermal stress.

We addressed the question if local adaptation to a thermal gradient is possible in spite of a high gene flow among closely spaced populations of two species of Drosophila from the island of La Gomera (Canary Islands). Variation in multiple traits related to stress resistance in different life stages was measured in both species in flies collected from five localities at different altitudes and thereby with different climatic conditions. Based on microsatellite loci, the populations were not genetically differentiated. However, 18 of the 24 independent traits measured showed significant differentiation among populations of Drosophila buzzatii, but only nine of 25 for Drosophila simulans. This difference in the number of traits might reflect higher habitat specificity and thus higher potential for local adaptation of D. buzzatii than D. simulans. We found clinal variation, as some traits showed significant linear regressions on altitude, but more on altitude cubed.

Introduction When investigating adaptive evolution in plants and animals, studies of geographic variation in fitness related traits have proved profitable. In particular, studies of intraspecies variation along an environmental gradient can reveal possible directional selection (Chown & Klok, 2003; Palo et al., 2003; Sørensen et al., 2005; Collinge et al., 2006; Sarup et al., 2006). For thermal adaptation, studies of populations collected along latitudinal or altitudinal gradients are often used, as climate varies in a fairly predictable manner along such clines (for a review see Hoffmann et al., 2003). In many cases, however, the populations studied were separated by large distances as compared with the dispersal abilities of Drosophila. This may lead to misinterpretation of the results, as isolation by distance alone can create clinal patterns in allele frequencies (Vasemagi, 2006). In this paper, we investigate geographical variation in populations collected at closely spaced sites along an altitudinal gradient on the island of La Gomera (Canary Islands). This spatial separation is well within the dispersal range Correspondence: Pernille Sarup, Aarhus Centre for Environmental Stress Research (ACES), Ecology and Genetics Department of Biology, University of Aarhus, Ny Munkegade, Buildg. 1540, DK-8000 Aarhus C, Denmark. Tel.: +45 8942 3219; fax: +45 8942 2722; e-mail: [email protected]

of Drosophila, at least that of Drosophila melanogaster, which as adults are able to disperse 10–15 km overnight (Coyne & Milstead, 1987; Coyne et al., 1987). Even though La Gomera is a small island (37 800 ha), the climatic conditions are very diverse. Due to its volcanic origin, it consists of a single central massif reaching 1487 meters above sea level (m.a.s.l.), which drops steeply to the sea. It is situated 100 km off the North African coast in a comparably cool ocean current. At sea level, the climate is dry and warm with palm trees but at higher elevations it is cooler and humid with laurel forest, because of the cooling of the moist air in the trade winds as they are forced up the mountain. This diversity in climate and habitats within a few kilometres gives the potential for adaptive differentiation of closely spaced populations, if natural selection is strong enough to override the potentially high gene flow. Evidence of such local adaptation in spite of very high rates of gene flow has been reported in evolution canyon, Israel (Korol et al., 2006), where environmental parameters change dramatically over just 400 m. We studied several traits related to thermal resistance in two Drosophila species, Drosophila simulans Alfred Sturtevant a cosmopolitan, and close relative to D. melanogaster, and Drosophila buzzatii Patterson and Wheeler, which has a more restricted distribution. The two species differ greatly in ecology: D. simulans has spread around the globe with humans (Lachaise et al., 1988), but it is

ª 2009 THE AUTHORS. J. EVOL. BIOL. 22 (2009) 1111–1122 JOURNAL COMPILATION ª 2009 EUROPEAN SOCIETY FOR EVOLUTIONARY BIOLOGY

1111

1112

P. SARUP ET AL.

less associated with humans than D. melanogaster, and does not normally enter human constructions (David et al., 2004). Drosophila simulans breed in necrotic fruit and vegetables. Drosophila buzzatii has a narrower niche, as it is cactophilic, living and breeding in rotting cladodes of Opuntia spp. During the last 200 years, it has spread from South America to Europe, Africa and Australia following human introduction of the Opuntia cacti (Barker, 1982; Fontdevila, 1989). In addition, the two species seem to differ in mobility. Drosophila buzzatii is larger and moves more slowly and less frequently than D. simulans in both nature and in the laboratory (P. Sarup, personal observation). Thus, it might be easier for the latter species to migrate between the collection sites as it breeds in necrotic fruit from orchards and gardens which are more continuously dispersed along the gradient. Many different stress resistance traits could potentially vary along an altitudinal gradient. Survival after exposure to extreme but ecological relevant temperatures could be an important trait for the fitness of individuals and populations. Latitudinal gradients in larval survival after heat shock have been found (Sarup et al., 2006), but for heat shock survival after hardening (prior exposure to an Hsp70 inducing temperature) the picture is less clear. Survival after hardening and heat shock is strongly related to Hsp70 expression level, which in turn is negatively correlated with high temperature adaptation both in adult D. melanogaster and D. buzzatii (Bettencourt et al., 1999; Sørensen et al., 2001). Hsp70 expression level also seems to influence the duration of heat-induced sterility in D. buzzatii, where heat adapted populations had longer sterility periods (Sarup et al., 2004). A much clearer pattern appears when knock-down resistance is considered: populations adapted to high temperature environments remain active longer at high temperature (Sørensen et al., 2001; Hoffmann et al., 2002). Generally, high temperature goes together with desiccation in nature and evidence for adaptive patterns exist (Karan et al., 1998; Hoffmann et al., 2003; Sørensen et al., 2005), but see also Hoffmann et al. (2001, 2002, 2005). Increased starvation resistance has been linked to selection for many kinds of stress resistance (Bubliy & Loeschcke, 2005b) but can also arise because of direct selection for this trait (Harshman et al., 1999). For D. buzzatii, food resources could be scarcer at the highest collection sites due to slower development of the rots, but also at the lowest site food was limited because of a low number of cacti and the low humidity at that locality. There was no obvious pattern in food availability for D. simulans, as the distribution of small gardens and orchards did not show a clear relation with altitude, but a negative association between starvation resistance and latitude has previously been found in D. simulans (Arthur et al., 2008). Increased survival after cold shock has been found in adults from low temperature populations of D. melanogaster (Hoffmann et al., 2002) and D. simulans

after acclimation (Hoffmann & Watson, 1993). Little is known about which life stage of D. buzzatii and D. simulans survives the winter. If they overwinter as juveniles, high altitude populations could be expected to have higher larval and ⁄ or pupal survival after a cold shock than low altitude populations. Drosophila simulans is more cold tolerant than D. buzzatii and might also overwinter as adults (J. David, personal communication). In this study, we expected to find more genetic population differentiation in thermal resistance related traits in D. buzzatii than in D. simulans, because of the presumed higher gene flow between populations of the latter species, and ⁄ or due to higher habitat specificity of D. buzzatii. We expected to find similar patterns of local adaptation in the two species in traits that were influenced by sufficiently strong natural selection. We expected a positive relation between altitude and survival after cold shock, Hsp70 expression level, and starvation resistance (at least in D. buzzatii). Further, a negative relation between altitude and knock-down time, the duration of sterility, survival after heat shock and desiccation resistance could be expected, if thermal selection was strong enough to drive local adaptation at the collection sites.

Materials and methods Origin and maintenance of flies Flies were collected in late March 2003 at five sites of increasing altitude on the western part of the island of La Gomera (see Fig. 1 in Bubliy & Loeschcke, 2005a). Geographic co-ordinates were determined with a global positioning system: (1) 2805.90¢N, 1720.73¢W, 20 m. a.s.l.; (2) 2806.53¢N, 1719.39¢W, 224 m.a.s.l.; (3) 2806.91¢N, 1718.87¢W, 374 m.a.s.l.; (4) 2806.88¢N, 1719.26¢W, 620 m.a.s.l.; (5) 2808.33¢N, 1718.84¢W, 886 m.a.s.l. Distances between localities ranged from 650 (3 and 4) to 5450 m (1 and 5). Over this short distance, a pronounced temperature and humidity gradient was observed during our expeditions to La Gomera. The temperature difference between the lowest and the highest site ranged from 6 C in the night and up to 11 C at noon. At collection site 1, the climate was hot and dry as compared with the relatively cool and humid climate at locality 5. Localities 2–4 were intermediate for temperature and humidity. In the laboratory, as explained in Bubliy & Loeschcke (2005a), the flies sampled at the five localities were used to start five mass populations of D. buzzatii and five sets of D. simulans ⁄ D. melanogaster isofemale lines, corresponding to the different localities. The isofemale lines allowed separating females of D. simulans from those of its sibling species D. melanogaster, which was not found to be numerous in our collections. The second laboratory generation of D. simulans was set up as mass populations obtained by mixing the lines using between 25 and 32


Local adaptation and high gene flow

isofemale lines per population. For D. buzzatii, 15–24 wild females and 9–48 males were used to establish the first laboratory generation of the mass populations, which were then expanded in the second generation. The mass populations of D. buzzatii were maintained with 50 mating pairs in each of six bottles with 21 mL of instant Drosophila medium with live yeast added (Carolina Biological Supply, Burlington, NC, USA). Throughout the experiments, instant Drosophila medium was used for this species unless stated otherwise. The mass populations of D. simulans were maintained with 50 mating pairs in each of six bottles with 35 mL of a standard Drosophila oatmeal-sugar-yeast-agar medium with live yeast added. To avoid crowding effects in experimental flies, only 20 mating pairs per bottle were allowed to lay eggs for 24 h. In both D. buzzatii and D. simulans, in every generation, the offspring from bottles were mixed within populations. All experiments were conducted within 3–6 generations after collection. This prevented maternal and grandmaternal effects (Kjærsgaard et al., 2007), arising from the different environmental conditions at collection sites from affecting the data, while also limiting the influence of adaptation to laboratory conditions (Krebs et al., 2001). Hsp70 expression Hsp70 expression was assayed at two temperatures, as clinal variation in this trait can differ depending on the temperature used for heat hardening (Sarup et al., 2006), in both D. buzzatii (37 or 38 C) and the less heat resistant D. simulans (35 or 36 C) in larvae, pupae and adult males and females. Hsp induction took place in preheated waterbaths for 1 h with the water above the bottom of the stoppers, followed by a 1 h recovery at 25 C before freezing at )80 C. The vials were placed in racks in the waterbaths and spaced evenly to ensure homogeneous heating. Third instar larvae for heat treatment were collected by inserting paper into the bottles and leaving for 6 h. The paper with wandering 3rd instar larvae was then transferred to empty plastic vials with moistened stoppers to prevent desiccation and then heated. On another set of papers, the larvae were allowed to pupate and then transferred to empty plastic vials on the moist paper. The pupae were heat exposed 2–3 days after pupation. Adult flies < 24 h old were collected, sexed under light CO2 anaesthesia and transferred to food vials. On day 3, they were transferred to fresh vials. When flies were 5–6 days old, they were placed in empty plastic vials and heat exposed. To prevent desiccation, the stoppers were moistened with water. Each treatment included six replicates containing 15 individuals. Later, larvae, pupae and flies were homogenized and the level of Hsp70 expression was assayed using a monoclonal inducible Hsp70 antibody (7.FB, Velazquez et al., 1980, 1983). Enzyme-linked immunosorbent assay (ELISA) was conducted in five

replicate microwell plates following described in Sørensen et al. (1999).

the

1113

protocol

Knock down resistance Flies < 24 h old were collected and transferred to food vials at a density of 50 (equal number of both sexes). Every second day, they were transferred to fresh vials. When flies were between 5 and 7 days old, the knockdown test (Huey et al., 1992) was conducted with five replicates per population as described in Sørensen et al. (2001). In this test, the flies are subjected to a high constant temperature (41.4 C ± 0.1 for D. buzzatii and 37.7 C ± 0.1 for D. simulans) and the time until heat coma is scored. Sterility period Sterility period was scored after males developed at one of three experimental temperature regimes: constant 25 C, constant 31 C and a fluctuating temperature regime 25 C (18 h) ⁄ 38 C (6 h) for D. buzzatii and constant 25 C, constant 29 C and fluctuating 20 C (18 h) ⁄ 34 C (6 h) for D. simulans. Flies were allowed to lay eggs in bottles for 24 h at 25 C. Bottles were placed at the appropriate temperature 2 days after the beginning of oviposition, where they remained until emergence of adults. To prevent desiccation of the vials at the high temperature, a tray with water was placed in the bottom of the incubators and stoppers were wetted every second day. After hatching, males (0–12 h old) were placed on oatmeal-sugar-agar-yeast medium with 6-day-old virgin females (raised at 25 C) from their respective population. For each combination of population and treatment, 15 vials with one male and four females were placed at 25 C. Every 24 h, the flies were transferred to new vials. Vials were kept at 25 C and later evaluated for the presence of larvae. Time until sexual maturity was calculated from the mean of the period where the males eclosed to the mean of the period where the first viable eggs were laid. Very few males that were permanently sterile were not included in the analysis. This measure of heat-induced sterility includes the period where males naturally are sexually immature. Heat shock survival Heat shock survival was assayed in larvae and adult flies. First instar larvae were collected in food vials at a density of 25 larvae per vial with 10 replicates per population and treatment. The larvae were kept at 25 C until heat shocked. Two experimental groups were used and a control was kept at 25 C throughout development. At day 4 after collection, one group was hardened for 1 h at 37 C (D. buzzatii) or for 1 h at 36 C (D. simulans),


1114

P. SARUP ET AL.

followed by 1 h recovery at 25 C before being heat shocked for 1 h at 41.5 C (D. buzzatii) or 39.5 C (D. simulans). The other group was directly heat shocked at 40.5 C for 1 h (D. buzzatii) or 38.5 C for 55 min (D. simulans). After heat shock, the larvae were returned to 25 C for development and survivors were scored as number of eclosing flies. Adult flies < 24 h old were collected from each population, set up (sexes separate) at a density of 25 per vial with 10 replicates per sex and population, and heat shocked according to the protocol described for Hsp70 expression in adults. One group was hardened at 37 C (D. buzzatii) or 35 C (D. simulans) for 1 h, followed by 1 h at 25 C to allow the flies to recover before being heat shocked at 41.0 C for 1 h (D. buzzatii) or 38.1 C for 75 min (D. simulans). The other group was directly heat shocked at 40.5 C for 1 h (D. buzzatii) or 37.5 C for 55 min (D. simulans). After the heat shock, flies were transferred to fresh food vials and allowed 24 h of recovery at 25 C before being scored as either alive or dead. Flies were considered as surviving if they were able to walk after a light touch with a brush. As survival was calculated as proportions, arcsine-square-root transformation was applied to improve normality and homogeneity of variances. Cold shock Cold shock survival was assayed in larvae and pupae. First instar larvae were collected from each population

Locus name

Chromosomal location (cytological location)

Drosophila buzzatii M013 2 M034 2 M52 2 M87 4 M090 5 M109 2 M122 4 M142 4 M223 5 M225 5 M290 2 M411 2 M493 2 M681 4 Drosophila simulans ac004640 2 droyanetsb 2 aC004118 2 drogPad 2 dmrhob 3 drogpdha 2 dmu14395 3 DMU43090 3

(54F1–55A1) (22D1-2) (35B2-B3) (47A) (62A) (25F5) (65D1-D3) (99D6-D9)

with 10 replicates per population, and transferred to food vials at a density of 25 larvae per vial. The larvae were kept at 25 C until cold shocked. At day 4 after collection, the larvae were put at 0 C for 15 h (D. buzzatii) or 20 h (D. simulans). After treatment, the larvae were returned to 25 C for development and survivors were scored as number of eclosing flies. To evaluate the survival rate after cold shock in pupae, larvae were allowed to pupate on filter paper, and then transferred to empty plastic vials on the moist paper with 25 pupae per vial and 10 vials per population. The pupae were exposed to 0 C for 2–3 days after pupation for 20 h (D. buzzatii) or 30 h (D. simulans). Desiccation resistance Adult flies < 24 h old were collected from each population, and set up (sexes separate) at a density of 25 flies per vial with five vials per population and sex. The 4–5 days old flies were transferred to empty food vials and placed in a desiccator for 45 h (D. buzzatii) or 16.5 h (D. simulans) after which the flies were scored for survival. Starvation resistance Freshly emerged flies (< 12 h old) were sexed under light CO2 anaesthesia and transferred to agar vials (to prevent desiccation), with 25 flies per vial and five vials per sex and population. The flies were starved for 5 days Table 1 Properties and references for the loci used in this study.

Size range

Number of alleles

References

74–88 252–258 141–155 163–203 185–195 151–167 217–221 194–206 227–229 166–172 167–179 234–248 165–169 255–275

4 2 2 5 3 6 2 5 2 4 4 4 4 6

Barker et al., 2009 Barker et al., 2009 Frydenberg et al., 2002 Frydenberg et al., 2002 Barker et al., 2009 Barker et al., 2009 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002 Frydenberg et al., 2002

244–258 84–86 266–268 160–168 198–208 130–132 266–288 180–188

2 2 2 4 5 2 8 4

Colson Colson Colson Colson Colson Colson Colson Colson

et et et et et et et et

al., al., al., al., al., al., al., al.,

1999 1999 1999 1999 1999 1999 1999 1999



0.00032 (25) 0.012 (24) 0.024(45)

0.0011 (4)* 0.084 (4)** 0.16(4)*** 0.25 (1)***

1184 (85) 853 (85) 670 (87)

Only tests with significant effects of population are shown. Degrees of freedom in parentheses. *P < 0.05; **P < 0.01; ***P < 0.001.

0.045 (44)

0.076(35)

0.088 (4)** 0.015 (40)

0.34 (4)*** 0.90 (1)*** 0.034 (4) 0.017 (40) 0.82 (4)*** 1.1 (4)*** 1.3 (4)*** 16 229 (4)***

17 226 6195 2026 3097 Population (1) Sex (2) (1) · (2) Error

4678 (4)***

5054 (4)*

197 (4)** 0.0017 (1) 0.024 (4) 0.028 (94)

0.13 (4)* 0.65 (1)*** 0.069 (4) 0.040 (90)

Hardened Hardened 25 ⁄ 38 C 31 C 25 C Knock-down

Duration of male sterility

Source

In D. simulans none of the results from the thermal resistance assays were significantly correlated within measurement method or in the total data set after significance levels were corrected using the Dunn-Sida´k method. In D. buzzatii, only larval Hsp70 expression at

ANOVA

Results

Table 2 Results of

Correlations between population means for the thermal resistance assays both within each measurement method (i.e. within Hsp70 expression among temperatures and life stages, etc.) and the total data set were analysed to evaluate the number of independent traits measured. A N O V A was used when appropriate to establish general effects of treatment and sex within experiments. In all traits with a significant population effect, the relation with altitude, altitude squared and altitude cubed was tested using linear regression, as nonlinear associations with clinal variables have been reported previously (Sarup et al., 2004). The explanatory variable with the highest R2 was chosen as the best descriptor. All statistical tests on quantitative data were performed with J M P v. 7 (SAS Institute Inc., 2007). The microsatellite loci were checked for null alleles using F R E E N A (Chapuis & Estoup, 2007). We analysed genetic differentiation among populations (FST; Wright, 1951) using F R E E N A and F S T A T 2.9.3.2 (Goudet, 2002) obtaining 95% confidence intervals (95% CI) by both bootstrapping and jackknifing. Number of immigrants in the first generation was estimated using G E N E C L A S S 2 (Piry et al., 2004).

(mean squares) on resistance traits in Drosophila buzzatii.

Statistical analysis

Heat shock adults

Heat shock larvae

Starvation

Desiccation

Cold shock larvae

DNA was extracted from between 19 and 36 wild caught flies for each combination of population and species with equal numbers of males and females using the hexadecyltrimethylammonium bromide (CTAB) method modified from Doyle and Doyle (1987). Drosophila buzzatii was genotyped using 14 microsatellite loci (Db013, Db034, Db052, Db087, Db090, Db109, Db122, Db142, Db223, Db225, Db290, Db411, Db493 and Db681) [see Frydenberg et al., 2002 (loci in this paper named M instead of Db) and Barker et al., 2009 for primer sequences, chromosomal location and genetic variation]. The number of alleles in D. buzzatii ranged from 2 to 6 (Table 1). Drosophila simulans was genotyped using 8 microsatellite loci (locus ac004640, droyanetsb, ac004118, drogPad, dmrhob, drogpdha, dmu14395 and DMU43090) (see Colson et al., 1999 and references therein for primer sequences and genetic variation). The number of alleles in this species ranged from 2 to 8 (Table 1).

(4)** (1) (4) (38)

Genetic differentiation in neutral markers

Larvae 38 C

Hsp70 expression

(D. buzzatii) or 43 h (D. simulans) before being scored for survival.

Pupae 37 C


1115

1116

P. SARUP ET AL.

37 C and larvae heat shock resistance were significantly correlated (r = 0.998; P < 0.0001) after significance levels were corrected using the Dunn-Sida´k method either within measurement method or in the total data set. A comparison of results for D. buzzatii and D. simulans showed a high correlation between heat shock resistance between the two species after hardening in males (Pearson correlation; r = 0.904; P = 0.03, nonsignificant after Dunn-Sida´k correction) and females (Pearson correlation; r = 0.995; P < 0.001) but no correlation between the sexes within species. No other traits showed significant correlations between species before or after correcting for the number of tests. For D. buzzatii, no effect of population was detected in pupal cold shock resistance (F4,40 = 1.18, P = 0.33), or in Hsp70 expression level in males at 37 C (F4,25 = 0.64, P = 0.64) and 38 C (F4,25 = 0.56, P = 0.70), in females at 37 C (F4,25 = 0.62, P = 0.65) and 38 C (F4,25 = 1.92, P = 0.14), pupal expression at 38 C (F4,25 = 0.18, P = 0.95) and larval at 37 C (F4,25 = 0.99, P = 0.43). There were significant effects of population (Table 2) in duration of male sterility at all temperatures tested (Fig. 1a), desiccation resistance, both sexes (Fig. 1b), starvation resistance, both sexes (see below, Fig. 1c), Hsp70 expression in larvae at 38 C and pupae at 37 C (Fig. 1d), male and female heat shock (Fig. 1e) and larval cold shock and heat shock survival (Fig. 1f) with or without hardening, and knock-down resistance, both sexes (Fig. 1g). In starvation resistance, there was a significant interaction between the effects of sex and population (Table 2), but still the effect of population was significant in both sexes (males: F4,20 = 24.13, P < 0.0001; females: F4,20 = 41.98, P < 0.0001). In D. buzzatii altitude cubed gave the highest R2 in knock-down resistance (R2 = 0.09; ß = 7.59 · 10)8; P = 0.03), cold shock resistance in larvae (R2 = 0.10; ß = 2.07 · 10)10; P = 0.025), duration of male sterility at 25 C (R2 = 0.05; ß = )2.55 · 10)8; P = 0.021), starvation resistance in males (R2 = 0.56; ß = 8.34 · 10)10; P < 0.0001) and Hsp70 expression in pupae at 37 C (R2 = 0.25; ß = 3.88 · 10)11; P < 0.0001). Altitude squared gave the highest R2 in female starvation resistance (R2 = 0.69; ß = 5.17 · 10)7; P < 0.0001). Considering only linear effects of altitude there were significant regressions in three traits. The variation in altitude could explain 36.2% of the variation among populations in

male starvation resistance (ß = 0.602; P = 0.001), 68.7% in females starvation resistance (ß = 0.829; P < 0.001) and 19% of the variation in pupal Hsp70 expression (ß = 0.436; P = 0.016). For D. simulans, no effects of population were detected in pupal cold shock resistance (F4,40 = 1.17, P = 0.50), male desiccation resistance (F4,20 = 2.68, P = 0.06), adult heat shock survival without hardening (F4,92 = 0.664, P = 0.62), larval heat shock survival with (F4,45 = 1.37, P = 0.26) or without hardening (F4,45 = 0.29, P = 0.89), duration of male sterility after development at 25 C (F4,62 = 1.02, P = 0.40) or 29 C (F4,62 = 0.792, P = 0.54), knock-down resistance (F4,40 = 0.96, P = 0.44) and Hsp70 expression in males at 36 C (F4,25 = 0.51, P = 0.72), in females at 35 C (F4,25 = 1.73, P = 0.18) and 36 C (F4,25 = 1.19, P = 0.34), in pupae at 35 C (F4,25 = 2.08, P = 0.11) and in larvae at 35 C (F4,25 = 1.73, P = 0.18) and 36 C (F4,25 = 1.73, P = 0.18). The effect of population was significant (Table 3) for duration of male sterility after development at 25 ⁄ 34 C (Fig. 2a), larval cold shock survival (Fig. 2a), female desiccation resistance (see below) (Fig. 2b), starvation resistance in both sexes (Fig. 2c), Hsp70 expression in pupae at 36 C and males at 35 C (Fig. 2d) and adult heat shock survival with hardening in both sexes (Fig. 2e). With respect to desiccation resistance, there was a significant interaction between the effects of sex and population (Table 3), and the effect of population was significant in females (F4,19 = 5.30, P < 0.01). In D. simulans, only two traits showed significant regressions on altitude. Altitude squared gave the highest R2 in female starvation resistance (R2 = 0.36; ß = )3.26 · 10)7; P = 0.0017). Altitude gave the highest R2 in female desiccation resistance (ß = )0.514; P = 0.012). Considering only linear effects of altitude the variation in altitude could explain 33.2% of the variation among populations in female starvation resistance (ß = )0.576; P = 0.003). For the traits where both males and females were tested, except adult heat shock survival without hardening, there was a significant effect of sex in D. simulans. In D. buzzatii for these traits, there was a significant effect of sex except for knock-down resistance and adult heat shock survival without hardening. No null alleles were detected in any of the loci. There was no significant genetic differentiation among populations in either species. In D. buzzatii, we found a global

Fig. 1 Traits measured in Drosophila buzzatii that were significantly different among populations. (a) Time to fertility in males ± SE after development at 25 C (18 h) ⁄ 38 C (6 h) (triangles), 25 C (closed circles) and 31 C (open circles) plotted against altitude. (b) Desiccation resistance ± SE in males (closed symbols) and females (open symbols) plotted against altitude. (c) Starvation resistance ± SE in males (closed symbols) and females (open symbols) plotted against altitude. The upper regression line is for females (P < 0.001), the lower for males (P = 0.001). (d) Hsp70 expression level ± SE level in pupae at 37 C (closed symbols) with regression line (P = 0.016), and in larvae at 38 C (open symbols) plotted against altitude. (e) Adult survival after heat shock ± SE without hardening in both sexes (triangles), and survival after heat shock ± SE with hardening in males (closed circles) and females (open circles) plotted against altitude. (f) Larval survival after cold shock ± SE (triangles), heat shock without hardening ± SE (closed circles), and heat shock with hardening ± SE (open circles) plotted against altitude. (g) Knock-down resistance ± SE plotted against altitude.



(a)

(b)

(c)

(d)

(e)

(f)

(g)


1117

1118

P. SARUP ET AL.

Table 3 Results of

ANOVA

(mean squares) on resistance traits in Drosophila simulans.

Duration of male sterility

Heat shock adults

Source

25 ⁄ 34 C

With hardening

Starvation

Desiccation

Cold shock larvae

Pupae 36 C

Male 35 C

Population (1) Sex (2) (1) · (2) Error

1544 (4)**

0.23 (4)* 0.64 (1)* 0.18 (4) 0.093 (90)

0.14 (4)* 1.5 (1)*** 0.057 (4) 0.027 (40)

0.096 (4)* 1.7 (1)*** 0.12 (4)* 0.029 (37)

0.16 (4)*

0.10 (4)*

0.0049 (4)**

0.024 (45)

0.022 (26)

0.00069 (24)

2.2 (64)

Hsp70 expression level

Only tests with significant effects of population are shown. Degrees of freedom in parentheses. *P < 0.05; **P < 0.01; ***P < 0.001.

FST = 0.00 (SE = 0.003) and a 95% CI of ()0.006; 0.005) obtained by bootstrapping. None of the pair wise FST values (results not shown) or FIS values (Table 4) were significant. In D. simulans, we found a global FST = 0.015 (SE = 0.015) and a 95% CI of ()0.009; 0.044) obtained by bootstrapping. The FIS values (Table 4) were not significant. No deviations from Hardy-Weinberg expectations were found in either species (D. buzzatii P = 0.07, D. simulans P = 0.053; test based on 1300 randomizations). In both species, one of the genotyped individuals was classified as a first generation immigrant by G E N E C L A S S 2.

Discussion Of 24 independent tests in D. buzzatii and 25 in D. simulans, we found significant genetic variation in resistance among populations for 18 traits in D. buzzatii and nine in D. simulans, confirming our hypothesis that D. buzzatii had greater potential for local adaptation than D. simulans. We found no evidence of this being caused by lower gene flow between D. buzzatii populations as no differentiation between populations in neutral markers could be detected in either species. In addition, we see a comparable number of first generation immigrants in both species. The higher differentiation for resistance traits in D. buzzatii is possibly an outcome of a higher selection pressure because of the narrower ecological niche of this species. An alternative explanation could be that the quantitative genetic differentiation among populations is a result of genetic drift during recurrent bottlenecks (i.e. seasonal fluctuations). However, this is not likely as this differentiation should easily be detected by the neutral markers, and this also cannot explain the significant regressions between some of the measured traits and altitude. Drosophila buzzatii also showed more clinal variation than D. simulans, especially when we tested for a nonlinear response. Here, in six of the 18 traits with significant population effects we found nonlinear effects of altitude compared with two of nine traits in D. simulans. At higher altitude, the environmental conditions become more hostile for the heat-adapted D. buzzatii. We were not able to find this species or its host (Opuntia sp.)

at above 886 m.a.s.l., and at localities situated at the extreme altitudes fewer flies of this species were caught in bait buckets per day (V. Loeschcke, personal observation). However, this was not the case for D. simulans. Therefore, the nonlinearity of altitudinal effects in D. buzzatii may reflect that marginal populations fail to adapt to the local environment (Bridle & Vines, 2007). The failure to adapt can be related to low Ne in marginal populations, and might have two contrasting explanations. Either, the populations lack the genetic variation needed for local adaptation or adaptation is prevented by too high migration rate from more abundant midrange populations, well adapted to the environment of their origin but maladapted to conditions at range margins (Haldane, 1956). Local adaptation in these Drosophila populations on La Gomera also depends on a balance between natural selection and gene flow, with the outcome depending on the effective population size (Ne). However, in this case it is not likely that the high altitude populations suffered from low genetic variation due to restricted gene flow. Asymmetric gene flow towards the high altitude populations might limit local adaptation here, especially if Ne is small (Eckert et al., 2008), and this could result in the nonlinear clinal effects observed. We expected to find similar patterns of local adaptation in the two species at least in some of the traits, but the resistance pattern between species was only significantly correlated for heat shock resistance after hardening. Notably, the correlation coefficients were very high (r > 0.9), and both males and females were correlated across species even though there was no correlation between male and female survival within species. This indicates that the correlations across species in this trait are not random. However, the regressions of this trait on altitude were nonsignificant in all cases, as has been seen in several studies of climatic adaptation (Hoffmann et al., 1997). There seems to be strong selection for local adaptation in this or a related trait that could not be explained by altitude, and this selection pressure is strong enough to override the large gene flow in both species. Some of the collection sites were more exposed to the sun than others, and when operating on a small spatial scale sun radiation might be a more important determi-



(a)

1119

(b)

(c)

(d)

(e)

Fig. 2 Traits measured in Drosophila simulans that were significantly different among populations. (a) Time to fertility in males ± SE after development at 25 C (18 h) ⁄ 34 C (6 h) (closed symbols) and larval survival after cold shock ± SE (open symbols) plotted against altitude. (b) Desiccation resistance ± SE in females (open symbols), with a significant regression (P = 0.012) plotted against altitude. (c) Starvation resistance ± SE in males (closed symbols) and females (open symbols), with a significant regression (P = 0.003) plotted against altitude. (d) Hsp70 expression level ± SE in pupae at 36 C (open symbols) and in males at 35 C (closed symbols) plotted against altitude. (e) Adult survival after heat shock ± SE with hardening in males (closed circles) and females (open circles) plotted against altitude.

nator of local climate than elevation. Heat shock resistance is often portrayed as an ecologically nonrelevant trait, as very few flies in nature experience such a steep increase (and drop) in temperatures as normally used in this assay. However, the correlation between species in this study implies that heat shock resistance after hardening might be relevant for the flies’ fitness in nature or that it is linked to or reflects performance with respect to a trait that is fitness related.

The regression of knock-down time on altitude squared only explained 9% of the variation among D. buzzatii populations, and in D. simulans the effect of population was nonsignificant. This is puzzling as there is growing evidence suggesting that knock-down resistance is an ecologically relevant heat resistance trait (for review, see Hoffmann et al., 2003), although some studies do not find altitudinal variation in this trait (Collinge et al., 2006; Arthur et al., 2008). However, if we


1120

P. SARUP ET AL.

Table 4 Number of individuals genotyped and FIS and corresponding P values within each population. Population Drosophila buzzatii 1 2 3 4 5 Drosophila simulans 1 5

n

FIS (P-value)

19 36 30 23 36

0.105 0.081 )0.011 )0.067 0.031

27 27

(0.34) (0.17) (0.97) (0.23) (0.62)

0.120 (0.12) 0.101 (0.19)

disregard the highest D. buzzatii population, which had longer knock-down time than the other populations, there is a significantly lower knock-down resistance at higher altitude (linear regression, R2 = 0.18, ß = )0.102, P < 0.01). Starvation resistance was the only trait that showed clinal variation in both D. buzzatii and D. simulans, in the latter only in females. The regression with altitude was positive in D. buzzatii as expected, and as previously found for this species in Argentina (Sørensen et al., 2005). The increased starvation resistance at higher altitude was not coupled with size in this case, as a study of morphometric traits in the same populations revealed no relation between altitude and size measured as thorax length (Bubliy & Loeschcke, 2005a). In D. simulans females, the regression with altitude was negative, comparable to the negative association between altitude and starvation resistance Arthur et al. (2008) found in D. simulans females; but we did not observe a gradient in the food availability that could indicate stronger selection for starvation resistance at low altitude. An alternative interpretation of the opposite trends in starvation resistance in the two species could be that the cline in starvation resistance reflected that D. buzzatii perceived the higher altitudes as most stressful (cold stress), especially since the flies were collected in the spring, whereas D. simulans was comparably more stressed at low altitudes (heat stress). Increased starvation resistance has been linked to selection for many forms of stress resistance including both cold and heat resistance (Bubliy & Loeschcke, 2005b). Thus, the opposite trends in these species might be due to D. buzzatii being adapted to higher temperatures than D. simulans. Although no clear trends were observed in most resistance traits the combined effect of weak selection for different cold resistance traits during the winter in D. buzzatii and heat resistance traits in D. simulans might result in clear regressions in starvation resistance. Compared with the other traits measured, little variation was found in Hsp70 expression, and only two of the treatments in each species revealed significant differences among populations. However, there were significant population effects for pupal Hsp70 expression in both

species, and expression level at 37 C in D. buzzatii was among the traits that showed clinal variation. Low altitude populations had lower Hsp70 expression than high altitude populations. This is the first record in pupae of a pattern that has been found in larvae (Sarup et al., 2006) and repeatedly in adult D. melanogaster and D. buzzatii in both nature and simulated natural conditions in the laboratory (Bettencourt et al., 1999; Sørensen et al., 2001). We found genetic variation between populations in the duration of sterility at all temperatures investigated in D. buzzatii and after rearing at high fluctuating temperatures in D. simulans. Like the survival after heat shock and hardening, most of the variance in sterility period could not be explained by altitude. Only the time to sexual maturation at 25 C in D. buzzatii produced a significant regression on altitude cubed, but in this trait there was no common pattern in the two species. However, this might be expected given the different clinal patterns previously found in D. buzzatii and the sibling species of D. simulans: D. melanogaster. In D. buzzatii, Vollmer et al. (2004) found prolonged duration of heat induced male sterility in populations originating from hot environments while Rohmer et al. (2004) found the opposite pattern in D. melanogaster. Both species showed signs of local adaptation in desiccation resistance, but the regression with altitude was only significant in D. simulans females. Here, the lowest population had the highest desiccation resistance as could be expected (Sørensen et al., 2005). In this study, heat shock resistance after hardening, desiccation and starvation resistance and Hsp70 expression in pupae seem to be the traits that are under the strongest selection for local adaptation, as they show clinal variation and differences between populations in both species. We conclude that there is a large potential for local adaptation in stress related traits in Drosophila, even in situations were gene flow between populations is so high that it prevents genetic differentiation between populations at neutral markers. However, this might depend on a large Ne in order to avoid local adaptations being swamped by immigrants.

Acknowledgments The authors are grateful to Heidi Arvidsson, Doth Andersen and Mia F. Nielsen for technical assistance, to Stuart Barker, Cino Pertoldi and two anonymous reviewers for helpful comments on the manuscript, to Corneel Vermeulen for providing some of the microsatellite primers, to S. Lindquist and M. Evgenev for kindly providing the 7.FB antibody, to the Sr. Consejero de Medio Ambiente, Excmo. Cabildo Insular de La Gomera, for permitting us to collect flies, and to the Danish Natural Sciences Research Council (frame and centre grant to V.L.) and the Faculty of Sciences, University of Aarhus, Carlsbergfondet and the Lundbeck Foundation (stipend to P.S.) for financial support.



References Arthur, A.L., Weeks, A.R. & Sgro, C.M. 2008. Investigating latitudinal clines for life history and stress resistance traits in Drosophila simulans from eastern Australia. J. Evol. Biol. 21: 1470–1479. Barker, J.S.F. 1982. Population genetics of Opuntia breeding Drosophila in Australia. In: Ecological Genetics and Evolution: the Cactus-Yeast-Drosophila Mode (J.S.F.B.a.W.T. Barker Starmer & W.T. Starmer, eds), pp. 209–224. Academic Press, NY. Barker, J.S.F., Frydenberg, J., Gonza´lez, J., Davies, H.I., Ruiz, A., Sørensen, J.G. & Loeschcke, V. 2009. Bottlenecks, population differentiation and apparent selection at microsatellite loci in Australian Drosophila buzzatii. Heredity, in press (doi: 10.1038 ⁄ hdy.2008.127). Bettencourt, B.R., Feder, M.E. & Cavicchi, S. 1999. Experimental evolution of Hsp70 expression and thermotolerance in Drosophila melanogaster. Evolution 53: 484–492. Bridle, J.R. & Vines, T.H. 2007. Limits to evolution at range margins: when and why does adaptation fail? Trends Ecol. Evol. 22: 140–147. Bubliy, O.A. & Loeschcke, V. 2005a. Variation of life-history and morphometrical traits in Drosophila buzzatii and Drosophila simulans collected along an altitudinal gradient from a Canary island. Biol. J. Linn. Soc. 84: 119–136. Bubliy, O.A. & Loeschcke, V. 2005b. Correlated responses to selection for stress resistance and longevity in a laboratory population of Drosophila melanogaster. J. Evol. Biol. 18: 789– 803. Chapuis, M.P. & Estoup, A. 2007. Microsatellite null alleles and estimation of population differentiation. Mol. Biol. Evol. 24: 621–631. Chown, S.L. & Klok, C.J. 2003. Altitudinal body size clines: latitudinal effects associated with changing seasonality. Ecography 26: 445–455. Collinge, J.E., Hoffmann, A.A. & McKechnie, S.W. 2006. Altitudinal patterns for latitudinally varying traits and polymorphic markers in Drosophila melanogaster from eastern Australia. J. Evol. Biol. 19: 473–482. Colson, I., Macdonald, S.J. & Goldstein, D.B. 1999. Microsatellite markers for interspecific mapping of Drosophila simulans and D. sechellia. Mol. Ecol. 8: 1951–1955. Coyne, J.A. & Milstead, B. 1987. Long-distance migration of Drosophila. 3. Dispersal of Drosophila melanogaster alleles from a maryland orchard. Am. Nat. 130: 70–82. Coyne, J.A., Bryant, S.H. & Turelli, M. 1987. Long-distance migration of Drosophila. 2. Presence in desolate sites and dispersal near a desert oasis. Am. Nat. 129: 847–861. David, J.R., Allemand, R., Capy, P., Chakir, M., Gibert, P., Petavy, G. & Moreteau, B. 2004. Comparative life histories and ecophysiology of Drosophila melanogaster and D. simulans. Genetica 120: 151–163. Doyle, J.J. & Doyle, J.L. 1987. A rapid DNA isolation procedure for small amounts of fresh leaf tissue. Phytochem. Bull. 19: 11– 15. Eckert, C.G., Samis, K.E. & Lougheed, S.C. 2008. Genetic variation across species’ geographical ranges: the centralmarginal hypothesis and beyond. Mol. Ecol. 17: 1170–1188. Fontdevila, A. 1989. Founder effects in colonising populations: the case of Drosophila buzzatii. In: Evolutionary Biology of Transient Unstable Populations (A. Fontdevila, ed.), pp. 74–95. Springer Verlag, Berlin.

1121

Frydenberg, J., Pertoldi, C., Dahlgaard, J. & Loeschcke, V. 2002. Genetic variation in original and colonizing Drosophila buzzatii populations analysed by microsatellite loci isolated with a new PCR screening method. Mol. Ecol. 11: 181–190. Goudet, J. 2002. Fstat. A Program to Estimate and Test Gene Diversities and Fixation Indices (version 2.9.3.2). URL: http:// www2.unil.ch/popgen/softwares/fstat.htm Haldane, J.B.S. 1956. The relation between density regulation and natural selection. Proc. R. Soc. Lond. B Biol. Sci. 145: 306– 308. Harshman, L.G., Hoffmann, A.A. & Clark, A.G. 1999. Selection for starvation resistance in Drosophila melanogaster: physiological correlates, enzyme activities and multiple stress responses. J. Evol. Biol. 12: 370–379. Hoffmann, A.A. & Watson, M. 1993. Geographical variation in the acclimation responses of Drosophila to temperature extremes. Am. Nat. 142: S93–S113. Hoffmann, A.A., Dagher, H., Hercus, M. & Berrigan, D. 1997. Comparing different measures of heat resistance in selected lines of Drosophila melanogaster. J. Insect Physiol. 43: 393–405. Hoffmann, A.A., Hallas, R., Sinclair, C. & Mitrovski, P. 2001. Levels of variation in stress resistance in Drosophila among strains, local populations, and geographic regions: patterns for desiccation, starvation, cold resistance, and associated traits. Evolution 55: 1621–1630. Hoffmann, A.A., Anderson, A. & Hallas, R. 2002. Opposing clines for high and low temperature resistance in Drosophila melanogaster. Ecol. Lett. 5: 614–618. Hoffmann, A.A., Sorensen, J.G. & Loeschcke, V. 2003. Adaptation of Drosophila to temperature extremes: bringing together quantitative and molecular approaches. J. Therm. Biol. 28: 175–216. Hoffmann, A.A., Shirriffs, J. & Scott, M. 2005. Relative importance of plastic vs genetic factors in adaptive differentiation: geographical variation for stress resistance in Drosophila melanogaster from eastern Australia. Funct. Ecol. 19: 222– 227. Huey, R.B., Crill, W.D., Kingsolver, J.G. & Weber, K.E. 1992. A method for rapid measurement of heat or cold resistance of small insects. Funct. Ecol. 6: 489–494. Karan, D., Dahiya, N., Munjal, A.K., Gibert, P., Moreteau, B., Parkash, R. & David, J.R. 1998. Desiccation and starvation tolerance of adult Drosophila: opposite latitudinal clines in natural populations of three different species. Evolution 52: 825–831. Kjærsgaard, A., Faurby, S., Andersen, D.H., Pertoldi, C., David, J.R. & Loeschcke, V. 2007. Effects of temperature and maternal and grandmaternal age on wing shape in parthenogenetic Drosophila mercatorum. J. Therm. Biol. 32: 59–65. Korol, A., Rashkovetsky, E. & Nevo, E. 2006. Enigmatic flies: is Drosophila in the ‘‘evolution canyon’’ a model for incipient sympatric speciation? Isr. J. Zool. 52: 507–525. Krebs, R.A., Roberts, S.P., Bettencourt, B.R. & Feder, M.E. 2001. Changes in thermotolerance and Hsp70 expression with domestication in Drosophila melanogaster. J. Evol. Biol. 14: 75– 82. Lachaise, D., Cariou, M.L., David, J.R., Lemeunier, F., Tsacas, L. & Ashburner, M. 1988. Historical biogeography of the Drosophila melanogaster species subgroup. Evol. Biol. 22: 159–225. Palo, J.U., O’Hara, R.B., Laugen, A.T., Laurila, A., Primmer, C.R. & Merila, J. 2003. Latitudinal divergence of common frog (Rana temporaria) life history traits by natural selection:


1122

P. SARUP ET AL.

evidence from a comparison of molecular and quantitative genetic data. Mol. Ecol. 12: 1963–1978. Piry, S., Alapetite, A., Cornuet, J.M., Paetkau, D., Baudouin, L. & Estoup, A. 2004. GENECLASS2: a software for genetic assignment and first-generation migrant detection. J. Hered. 95: 536–539. Rohmer, C., David, J.R., Moreteau, B. & Joly, D. 2004. Heat induced male sterility in Drosophila melanogaster: adaptive genetic variations among geographic populations and role of the Y chromosome. J. Exp. Biol. 207: 2735–2743. Sarup, P., Dahlgaard, J., Norup, A.M., Jorgensen, K.T., Hebsgaard, M.B. & Loeschcke, V. 2004. Down regulation of Hsp70 expression level prolongs the duration of heat-induced male sterility in Drosophila buzzatii. Funct. Ecol. 18: 365–370. Sarup, P., Sorensen, J.G., Dimitrov, K., Barker, J.S.F. & Loeschcke, V. 2006. Climatic adaptation of Drosophila buzzatii populations in southeast Australia. Heredity 96: 479–486. SAS Institute Inc. 2007. JMP (version 7). Sørensen, J.G., Michalak, P., Justesen, J. & Loeschcke, V. 1999. Expression of the heat-shock protein HSP70 in Drosophila buzzatii lines selected for thermal resistance. Hereditas 131: 155–164. Sørensen, J.G., Dahlgaard, J. & Loeschcke, V. 2001. Genetic variation in thermal tolerance among natural populations of Drosophila buzzatii: down regulation of Hsp70 expression and variation in heat stress resistance traits. Funct. Ecol. 15: 289– 296. Sørensen, J.G., Norry, F.M., Scannapieco, A.C. & Loeschcke, V. 2005. Altitudinal variation for stress resistance traits and thermal adaptation in adult Drosophila buzzatii from the New World. J. Evol. Biol. 18: 829–837.

Vasemagi, A. 2006. The adaptive hypothesis of clinal variation revisited: single-locus clines as a result of spatially restricted gene flow. Genetics 173: 2411–2414. Velazquez, J.M., Didomenico, B.J. & Linquist, S. 1980. Intracellular localization of heat shock proteins in Drosophila. Cell 20: 679–690. Velazquez, J.M., Sonoda, S., Bugaisky, G. & Lindquist, S. 1983. Is the major Drosophila heat shock protein present in cells that have not been heat shocked? J. Cell Biol. 96: 286–290. Vollmer, J.H., Sarup, P., Kaersgaard, C.W., Dahlgaard, J. & Loeschcke, V. 2004. Heat and cold-induced male sterility in Drosophila buzzatii: genetic variation among populations for the duration of sterility. Heredity 92: 257–262. Wright, S. 1951. The genetical structure of populations. Ann. Eugen. 15: 323–354.

Supporting information Additional Supporting Information may be found in the online version of this article: Appendix S1 Expected and observed heterozygosities of genotyped alleles. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article. Received 8 November 2008; revised 5 February 2009; accepted 6 February 2009
