1The relationship between heat shock protein Hsp70 expression level and the duration of heat-induced male sterility was investigated in four populations of Drosophila buzzatii Patterson and Wheeler. The effect of heat hardening on the duration of sterility was further examined after flies developed at either 25 or 31 °C. In addition, Hsp70 expression was measured in testes after development at three different thermal regimes.
2Four main hypotheses were tested: (i) Hsp70 is expressed in testes of D. buzzatii males even at non-stressful temperatures, and the level of expression increases with increasing rearing temperature. (ii) Hsp70 expression level differs between populations and is negatively correlated with the duration of poststress sterility. (iii) Experimentally induced Hsp70 expression at the pupal stage shortens the sterility period of flies reared above the temperature threshold of sterility. (iv) In contrast, a hardening treatment during the pupal stage of flies reared at 25 °C results in a longer time to fertility.
3The results matched the hypotheses, leading to the conclusion that higher Hsp70 expression reduces the duration of heat-induced male sterility.
Drosophila melanogaster males are sterile when reared at temperatures above 30 °C, but gain fertility within a few days if transferred to 25 °C (David, Arens & Cohet 1971). Very little intraspecific variation has been found within Drosophila species in this temperature threshold of sterility (Kuznetsova 1994; Vollmer et al. 2004). Concerning the duration of sterility after heat exposure and transfer to 25 °C, however, there are considerable differences among populations of Drosophila buzzatii Patterson and Wheeler (Vollmer et al. 2004). Taking into account the short average life span of Drosophila in the wild, estimated to be from a few days to a few weeks (Rosewell & Shorrocks 1987; Turelli & Hoffmann 1995), a delay of even a few days in sexual maturity should have profound effects on male fitness. Given that there is genetic variation in the duration of sterility, and male flies from high-temperature environments occasionally experience heat-induced sterility, flies in such environments should evolve mechanisms to shorten the duration of sterility compared with neighbouring populations from less harsh thermal environments.
Vollmer et al. (2004) examined the sterility period of four populations of D. buzzatii originating from different thermal environments. Two of the populations originated from Argentina, namely a population frequently exposed to high temperatures in the natural environment where it originated, Arg(H), and a population not exposed to daily high temperatures, Arg(L). The other two populations originated from Tenerife (Canary Islands) and were split into two populations in the laboratory, one of which was kept at a non-stressful temperature (25 °C) while the other was allowed to complete larval development at a high fluctuating temperature (38 °C (6 h)/25 °C (18 h)). The population from the high-temperature environment in nature was expected to have a shorter sterility period than the low-temperature population. The population experiencing high temperatures during larval and early pupal development in the laboratory, but kept at benign temperatures during maturation, was not expected to differ in sterility period from the corresponding population kept at 25 °C through the entire life cycle. Contrary to these expectations, both the low-temperature populations gained fertility faster than the corresponding high-temperature populations after development in a high-temperature fluctuating environment [38 °C (6 h)/25 °C (18 h)] in the laboratory. The fact that both sets of lines showed the same pattern suggested that this was not a coincidence. This unexpected finding could be due to an indirect selection response resulting from selection on a correlated trait, implying that the duration of sterility is negatively correlated with one or more traits being selected for in high-temperature environments.
The four populations have previously been examined for adult Hsp70 expression. These studies showed that the low-temperature populations had higher Hsp70 (an inducible 70 kDa heat shock protein) expression than the corresponding high-temperature populations (Sørensen et al. 1999; Sørensen, Dahlgaard & Loeschcke 2001), indicating that Hsp70 is down-regulated in populations that are frequently exposed to heat stress during development. Heat shock proteins (Hsps) are induced when organisms are exposed to a sub-lethal stress. Some Hsps are molecular chaperones, binding to denatured and unfolded proteins, preventing aggregation of non-native protein (Parsell & Lindquist 1993). Some of these proteins, as the constitutive Hsc70, are expressed continuously while others are exclusively expressed following stress. The latter disappear shortly after the stress ceases, like the inducible Hsp70. Still others, e.g. Hsp90, are both constitutive and inducible. Using a polyclonal antibody that recognizes both Hsc70 and Hsp70, Michaud et al. (1997) found that Hsp70 and/or Hsc70 is expressed in testes, spermatogonia and spermatids from unstressed males, and this expression increases following a heat shock. Additionally, Lakhotia & Prasanth (2001) found that the inducible form of Hsp70 is expressed in unstressed spermatogonia cells in D. melanogaster. As Hsp70 is present during spermatogenesis, it could partly protect the spermatogenesis from thermal damage, and the comparably long sterility period of the high-temperature populations could be due to down-regulation of Hsp70 expression in these populations.
In this study we examine if and how Hsp70 expression level is related to the sterility period in D. buzzatii. To examine whether Hsp70 is also expressed in unstressed D. buzzatii testes, and if expression level in testes increases with rearing temperature, Hsp70 expression was measured in testes from males reared at 25 °C, 31 °C or 38 °C (6 h)/25 °C (18 h).
To study the relationship between Hsp70 expression level and the duration of the sterility period, Hsp70 expression was measured in pupae in the same populations where sterility period after heat exposure had previously been studied (Vollmer et al. 2004). The four populations differ in developmental time, as do the sexes. Drosophila buzzatii males develop faster than females, so a greater proportion of early pupae will be males. Adult males have higher Hsp70 expression than females (Sørensen et al. 2001), and as it is impossible to determine the sex of the pupae it was necessary to synchronize the lines to obtain similar proportions of males in each sample.
If Hsp70 (or a factor correlated with Hsp70) to some degree protects spermatogenesis from heat damage, Hsp70 expression should be higher in the populations with a short sterility period than in the populations with a long duration of sterility. Based on the sterility data from Vollmer et al. (2004), we expected the level of Hsp70 expression to be highest in the low-temperature population from Argentina and lowest in the high-temperature population from Tenerife.
Hsp70 expression level is found to be stable over at least 50 generations in Drosophila cultures under laboratory conditions (Krebs et al. 2001), and although there has been no study of whether the duration of sterility is stable over many generations in the laboratory, the sterility period measured in the hardening experiment (see below) was comparable to the duration of sterility in Vollmer et al.'s experiment. Therefore, a correlation analysis was done between previously obtained sterility data (see Vollmer et al. 2004) and Hsp70 data from the same populations. This approach allowed us to further investigate the relationship between sterility period and Hsp70 expression in D. buzzatii. We hypothesize that the sterility period should be negatively correlated with Hsp70 expression.
The effect of a short pulse of heat hardening (30 min at 36, 37 or 38 °C) during pupal development on sterility period after rearing at 25 or 31 °C was examined in Arg(L), the population with the highest Hsp70 expression after 1 h at 38 °C. This procedure was used to directly test the effect of increasing Hsp70 level on the sterility period and to test for the possibility that the correlation between sterility period and Hsp70 expression was due to differences other than the Hsp70 expression levels between the four populations. We expected the hardening treatment to shorten the sterility period after development at 31 °C due to increased Hsp70 expression and an elongating effect of hardening on sterility period after development at 25 °C due to heat damage.
Materials and methods
origin of flies
We used two sets of populations. One set was collected in Tenerife, Canary Islands, in 1993 and split into two populations in the laboratory. One (Ten(H)) was kept at a high fluctuating temperature regime (38 °C (6 h)/25 °C (18 h) with 6 h light/18 h dark), where it completed larval and early pupal development in each generation. After pupation, bottles were transferred to 25 °C where adults emerged and were kept at 25 °C to allow males to reach fertility. The other population from Tenerife (Ten(L)) was kept at constant 25 °C at a 12 h light/dark cycle (Loeschcke & Krebs 1996). These populations from Tenerife have been kept in the laboratory for 117 generations. The other set of populations was collected in Argentina in March 1997. It consists of a population from Tilcara (Arg(L)) (23°29′ S, 65°39′ W; 2460 m above sea level), where monthly average maximum temperatures are in the range 23–24 °C and monthly average minimum temperatures are around 11–12 °C during summer and autumn (averaged from data obtained from two adjacent weather stations; http://www.meteofa.mil.ar/), and a population from Catamarca (Arg(H)) (28°29′ S, 65°24′ W; 590 m above sea level), where monthly average maximum temperatures are in the range 30–35 °C and monthly average minimum temperatures are around 20–21 °C during summer and autumn, i.e. experiencing higher overall temperatures than Tilcara. These populations have been kept in the laboratory for 47 generations at 25 °C at a 12 h light/dark cycle and differ in thermal adaptation, as described in Dahlgaard et al. (2001) and Sørensen et al. (2001).
maintenance of stocks
Each mass population was maintained in six bottles (200 ml) with 7 ml of instant Drosophila medium (Carolina Biological Supply, Burlington, NC, USA), with 20 pairs per bottle, which were allowed to oviposit for 24 h before transfer to new bottles. Mating was random in each generation, and adults were mixed among bottles.
hsp70 expression in pupae
The four populations were synchronized by collecting pupae from bottles where the first larvae pupated within 0–24 h before sampling. The pupae were placed in glass vials with medium and moistened stoppers to prevent desiccation. Control vials were kept at 25 °C. Vials for the heat-shock treatments were placed in racks and spaced evenly to ensure homogeneous heating. The racks were placed in preheated water baths with the water level above the bottom of the stoppers. Two temperatures were used: 31 and 38 °C. The pupae were 5–17 h old when Hsp was induced. Each treatment included four populations and seven replicates, giving a total of 28 vials per treatment. Each replicate contained 25 pupae collected randomly from at least five different bottles. After heat shock treatment the pupae were kept for 1 h at 25 °C and vials from all three treatments were then stored at −70 °C. Pupae were later homogenized and Hsp70 expression was measured using a monoclonal antibody (7.FB) specific for the inducible Hsp70 (Welte et al. 1993). Hsp70 expression levels were measured using enzyme-linked immunosorbent assay (ELISA). All populations and treatment combinations were represented on each ELISA plate with four subsamples per replicate, and the seven replicates were on different plates.
effect of hardening on sterility
In this experiment Arg(L) was used, as it had the highest Hsp70 expression in the pupal stage. Flies were reared at either 25 °C or 31 °C. At each rearing temperature pupae of an age of 0–12 h were collected and given one of four treatments. Three groups were given a heat-hardening treatment for 30 min at 36, 37 or 38 °C and were thereafter immediately returned to 25 or 31 °C. The fourth group was kept at the rearing temperature. When males eclosed they were collected under light CO2 anaesthesia when between 0 and 12 h old. The males were put in vials with medium, four in each vial together with four virgin females; 13 vials were set up from each treatment. The flies were transferred to new vials every 12 h until progeny were detected. Vials were kept at least 7 days before being checked for larvae/pupae. The time until fertility was reached by at least one of the four males in a vial was estimated from the mean of the period where the males eclosed to the mean of the period where the first viable eggs were laid.
hsp70 expression in testes
Males from the Arg(L) population were reared from first instar larvae to adulthood at 25 or 31 °C or at the fluctuating temperature (38 °C (6 h)/25 °C (18 h)). Adults were collected when 0–12 h old, kept at 25 °C for an hour, and frozen at −70 °C until dissection with five replicates per temperature and 25 males per replicate. The testes were separated from other tissues and frozen in homogenizing buffer at −70 °C. All samples were homogenized and kept on ice until administered to the microwell plate the same day. Hsp70 expression was assayed by ELISA, with four subsamples per replicate.
Unless stated otherwise all statistical analysis was performed using SPSS for Windows 10·0.
hsp70 expression in pupae
As no significant inducible Hsp70 expression was detected after exposure to 25 or 31 °C, these data were omitted from further analysis. The data from the 38 °C treatment were log-transformed after subtracting the plate means from each reading, adding 1 to eliminate negative values, and analysed by a mixed model nested anova (Sokal & Rohlf 1995). The mean absorbances are presented in Fig. 1. The effect of population was significant (F3,24 = 46·2, P < 0·001) and the effect of replicate within populations was not. Hsp70 expression in Arg(H) was significantly lower than in the low-temperature population Arg(L) (Tukey test α = 0·05). No significant difference was found between Arg(H) and Ten(L), but between the populations originating from Tenerife the difference in Hsp70 expression was significant (Tukey test α = 0·05).
effect of hsp70 level on the sterility period
A correlation coefficient r = −0·88 between the means of the sterility periods (from Vollmer et al. 2004) and Hsp70 expression level was estimated (P < 0·001, one-tailed Pearson correlation analysis).
effect of hardening on sterility
Data on the effect of hardening on sterility period after development at either 25 or 31 °C were analysed by a two-way anova. Results showed a significant effect of rearing temperature (F1,90 = 2607, P < 0·001) and hardening temperature (F3,90 = 5·5, P = 0·002), and a significant interaction between rearing temperature and hardening (df = 3, F = 7·3, P < 0·001).
The duration of sterility of flies reared at 31 °C was significantly different among treatments (one-way anova, F3,41 = 2·96, P = 0·043). The sterility period of non-hardened flies was significantly longer than if flies had received a hardening treatment. No significant difference in the duration of sterility was found between the hardening temperatures (Duncan test α = 0·05). The sterility period after hardening was on average 10 h shorter than the duration of sterility after development at 31 °C without hardening.
After development at 25 °C the time from eclosure until fertility was reached was significantly different among treatments (one-way anova, F3,41 = 3·9, P = 0·017). Mean duration of sterility after development at 25 °C without hardening was significantly shorter than the duration of sterility after hardening at 36 and 37 °C, but not at 38 °C (Duncan test α = 0·05). The sterility periods after the hardening treatments were not significantly different from each other, and the sterility period was on average 6·5 h longer in the hardened groups compared with the group without hardening (Fig. 2).
hsp70 expression in testes
Data were log-transformed and analysed by a mixed model nested anova (Sokal & Rohlf 1995). Both rearing temperature (F2,12 = 347, P < 0·001) and replicate (F12,45 = 2·5, P < 0·01) had significant effect on Hsp70 expression in testes (Fig. 3). Hsp70 expression was significantly different among all developmental temperatures, with a higher expression level in flies exposed to higher temperatures during development (Tukey test α = 0·05). In addition measures of Hsp70 expression after development at 25 °C were significantly different from zero (one tailed t-test; t = 18·43; P < 0·001).
Our results are in accordance with the hypothesis that Hsp70, or something correlated with Hsp70, reduces the damaging effect of heat on spermatogenesis, either by protecting spermatogenesis or by taking part in the repair of heat-damaged protein. The expression levels of the four populations of D. buzzatii were in the expected order, as predicted from sterility data when assuming that a high Hsp70 expression level shortens the sterility period. In addition, there was a strong correlation between previously obtained sterility data and Hsp70 expression. This correlation was subsequently confirmed by a hardening experiment in one population, to rule out the possibility that the association between Hsp70 expression and sterility period was due to genetic heterogeneity among the four populations. Accordingly, heat hardening significantly shortened the sterility period when flies were reared in a sterility-inducing environment. Although the duration of sterility is not significantly different among hardening temperatures, there is a clear trend towards a shorter sterility period if flies are hardened at higher temperatures, and thereby expressing higher levels of Hsp70. The experiment also revealed that the effect of hardening in shortening the sterility period after development at 31 °C was not an effect of heat hardening per se, independent of the rearing environment. This was demonstrated, as opposite effects of hardening in the two environments were found, resulting in a significant interaction term between rearing condition and hardening treatment (Fig. 2). In other studies of heat hardening's effect on heat-induced male sterility, Rinehart, Yocum & Denlinger (2000) found no effect of heat hardening in the Flesh Fly Sarcophaga crassipalpis, whereas Maisonhaute, Chihrane & Lauge (1999) found such an effect in Trichogramma brassicae. These opposite results may be due to the severity of the heat shock administered. If the heat shock is too severe, spermatogenesis may suffer heat damage beyond repair.
The correlation between Hsp70 expression level and the duration of sterility explains the unexpected result that the duration of sterility was longer in populations from high-temperature environments after development at a fluctuating high-temperature regime compared with the low-temperature populations (Vollmer et al. 2004). Hsp70 expression is in some cases down-regulated in populations from high-temperature environments in both nature and the laboratory (Sørensen et al. 1999, 2001; Bettencourt, Feder & Cavicchi 1999; Lansing, Justesen & Loeschcke 2000). The most plausible interpretation of down-regulation is that detrimental effects of Hsp70 expression (i.e. retarded development and reduced fecundity; Feder et al. 1992; Krebs & Loeschcke 1994; Silbermann & Tatar 2000) are larger than the corresponding benefits (lower mortality after heat shock; Loeschcke, Krebs & Barker 1994; Feder & Krebs 1997). The discovery of a possible functional role for Hsp70 in shortening the sterility period of males reared at sterility-inducing temperatures adds a new function to the ‘benefits list’ of Hsp70 induction. This new finding may help explain why a population from a high-temperature environment has been found to have higher Hsp70 expression level than a neighbour population from a lower temperature environment (Michalak et al. 2001).
The fact that a very short pulse of high temperature had a significant effect on the sterility period indicates that Hsp70 may be a powerful mechanism to reduce the effect of high temperature on fertility. As prolonged sterility has strong implications for male fitness, this could be part of the explanation for the contradictory results from selection for thermal resistance. Whether Hsp70 expression in males is lowered or increased by selection for increased thermal tolerance in nature may in part depend on the life expectancy of adult males. In environments with high mortality, the risk of males dying before becoming fertile might tip the balance towards higher expression of Hsp70.
One question remains to be investigated. Is the correlation between the duration of the sterility period and Hsp70 expression level due to a functional relationship between Hsp70 and the duration of sterility, and/or is Hsp70 expression correlated with the additional causative factor(s)? Another candidate for a causative factor is Hsp90, which is expressed continuously but is inducible as well. The induction of Hsp90, as of Hsp70, is controlled by the heat shock factor, HSF (Morimoto 1998). It is well established that Hsp90 plays a crucial role during spermatogenesis (Yue et al. 1999), and a universal phenotype of Hsp90 deficient mutants is sterility among males. In this process Hsp70 is also involved as it loads the receptor onto Hsp90 for further folding (Young, Moarefi & Hartl 2001). This interconnection between Hsp70, Hsp90 and steroid hormone receptors might explain the correlation between Hsp70 expression level and the sterility period, and give an indication of why the inducible Hsp70 was present in measurable amounts in D. buzzatii testes at 25 °C, supporting that both Hsp70 and Hsp90 are important for the trait.
Normally Hsp70 is not induced before a certain stress level is reached. In D. buzzatii we did not find Hsp70 expression in pupae at 31 °C and the threshold of thermal induction is 33 °C in whole adult flies (Kristensen, Dahlgaard & Loeschcke 2002). Here we found Hsp70 expression in testes from males reared at 25 °C, and expression was increased in testes developed at 31 °C. Expression of Hsp70 at 25 °C in testes has also been found in D. melanogaster (Krebs & Feder 1997); Lakhotia & Prasanth 2001), where Hsp70 is present in unstressed spermatogonia from the second instar larval stage and onwards. This developmental expression could indicate a function of the inducible form of Hsp70 in spermatogenesis also under normal temperatures.
We are grateful to Susan Lindquist for kindly providing the antibody 7.FB, to Stuart Barker, Jean David and three anonymous reviewers for helpful comments on the manuscript, and to the Danish Natural Sciences Research Council (frame and centre grant to VL) and the Faculty of Sciences, University of Aarhus (stipend to P.S.) for financial support.