Evolution of thermotolerance and the heat-shock response: evidence from inter/intraspecific comparison and interspecific hybridization in the virilis species group of Drosophila. I. Thermal phenotype

Investigations in evolutionary physiology normally fall into two separate realms. The first concerns the large differences in phenomes among species or higher taxa, whereas the second concerns the often lesser variation within species (Feder et al., 2000). These realms are ordinarily distinct because of reproductive isolation among species. Thus, because species and higher taxa usually cannot be interbred experimentally, investigation of their physiological differences often must proceed in the absence of experimental manipulative genetics. Genetic studies within species, however, often must forego the larger physiological variation that interspecific studies offer. Here, we report on variation in thermotolerance and the heat-shock response in a remarkable instance in which these two normally separate realms of investigation coalesce. The virilis group of Drosophila comprises 12 species living in greatly differing habitats but amenable to interbreeding in the laboratory(Patterson and Stone, 1952; Throckmorton, 1982). We have exploited these attributes to investigate the genetic basis of interspecific differences in physiology in combination with comparative inference.

The virilis group is divisible into two phylads: virilisand montana (Patterson and Stone,1952). In the virilis phylad, two species, D. virilis and D. lummei, populate opposite ends of a climatic/geographic spectrum but overlap in its center. According to many investigators (Evgenev et al.,1982; Nurminsky et al.,1996; Patterson and Stone,1952; Spicer, 1991, 1992; Throckmorton, 1982), D. virilis is the most primitive species of the virilis phylad and is probably ancestral to it if not to the entire virilis group. Its distribution, which extends across Eurasia, is primarily below 40° N latitude. D. lummei, considered the closest relative of D. virilis, occurs from just above 40° to just above 65° N latitude and from Sweden east to the Pacific coast of Asia. In the New World, D. novamexicana and D. texana are a similar but less widely distributed species pair from the virilis phylad. D. novamexicana occurs in hot, arid environments in New Mexico, Arizona,Colorado, Utah (Hsu, 1952) and elsewhere, whereas D. texana inhabits more mesic environments in central Texas.

If the microclimates that these species experience differ in the same ways as their climatic ranges, then data for many other species suggest that their thermal phenotypes (Feder,1996) should differ correspondingly. Specifically, for organisms in general (Cossins and Bowler,1987), and Drosophila in particular(David et al., 1983; Feder and Krebs, 1998; Hoffmann et al., 2003; Stratman and Markow, 1998),tolerance of high temperatures and its underlying mechanisms are correlated with the typical thermal environments of species (i.e. countergradient variation). From this prior work, we focus on several discrete aspects: basal thermotolerance (tolerance of acute exposure to hyperthermia in naive organisms or cells), inducible thermotolerance [the change in thermotolerance when mild hyperthermia (= pre-treatment; PT) precedes exposure to more-severe hyperthermia] and the heat-shock response. The heat-shock response comprises the PT-inducible expression of heat-shock proteins (Hsps), molecular chaperones and other proteins that contribute to inducible thermotolerance(Feder and Hofmann, 1999; Ulmasov et al., 1992; Zatsepina et al., 2001). The most important of these in other Drosophila species is Hsp70, a member of the DnaK-Hsp70 superfamily(Feder and Krebs, 1998; Zatsepina et al., 2001). Prior work on these aspects in the virilis group, however, is limited. D. virilis exceeds D. lummei in basal thermotolerance(Garbuz et al., 2002; Mitrofanov and Blanter, 1975),as expected. In the cold, D. lummei is far more tolerant than D. virilis, as expected, and also undergoes a diapause absent in all other D. virilis group species (Lumme,1982). Heat-inducible loci have been localized cytogenetically in D. virilis (Evgenev et al., 1970; Peters et al., 1980). Although Sinibaldi and Storti (1982)reported that the two species do not differ in Hsps induced by heat shock,Garbuz et al. (2002) have recently shown that both the D. virilis–D. lummei and D. novamexicana–D. texana species pairs differ in Hsp induction. In both cases, the former member of each pair synthesizes greater amounts of Hsps after heat shock. The present study extends the work of Garbuz et al. (2002) by comparing additional strains of D. virilis and D. lummei, focusing on more-intense heat shock and examining hsp mRNA expression and its regulation. Heat-inducible loci have been localized cytogenetically in D. virilis (Evgenev et al.,1978; Peters et al.,1980).

Ordinarily, understanding the evolution of such interspecific variation would be amenable only to comparative inference and inaccessible to genetic experimentation. In the 1970s, however, Evgenev and colleagues were able to cross D. lummei with a D. virilis strain (160) bearing recessive markers on all autosomes and subsequently developed strains in which a single lummei chromosome or portion thereof was integrated into a uniform D. virilis background(Evgenev and Sidorova, 1976). In the present study, we combine comparative inference and interspecific genetics to elucidate the molecular basis for interspecific variation in thermotolerance. We demonstrate (1) replicated countergradient variation of basal thermotolerance in two geographically distinct sets of virilisphylad species; (2) countergradient variation in inducible thermotolerance and the heat-shock response in D. virilis and D. lummei; and (3)the applicability of interspecific genetics to inducible thermotolerance (and prospectively other aspects of the thermal phenotype) in the virilisspecies group of Drosophila.

We compared laboratory strains of four species from the virilisphylad and one reference strain of Drosophila melanogaster(Table 1). All strains were maintained at 25°C on a yeast, molasses and agar medium for many generations before analysis. Hybrids of D. virilis strain 9 and D. lummei strain 200 were created by mass mating of heterospecific parents of the opposite sex.

These procedures were identical to those of Garbuz et al.(2002). Eclosing individuals were sequestered daily and, when 4 days old, were transferred by aspiration to a preheated polypropylene vial (Nalge Nunc International, Rochester, NY, USA),which was immersed in a thermostatted water bath for 30 min. Each vial usually contained 35–50 animals but occasionally as few as 20. To determine basal thermotolerance, vials were placed at one of a series of heat-shock temperatures ranging from 38.5°C to 43°C. To determine inducible thermotolerance, similar determinations ensued after flies first underwent pre-treatment at one of a series of temperatures from 35°C to 38°C for 30 min and 25°C for 1 h or 3 h. Tolerance was assessed as the proportion of flies in a vial that could walk 48 h after heat shock.

These procedures were identical to those of Garbuz et al.(2002). 10 salivary glands from third-instar larvae were labeled in 20 μl of Schneider's insect medium without methionine (Sigma, St Louis, MO, USA) after the addition of 1 μl(1.85 MBq) of [³⁵S]_L-methionine (Amersham Biosciences Corp., Piscataway, NJ, USA) for 1 h at 25°C after various treatments. Two-dimensional gel electrophoresis and other procedures applied were as described (O'Farrell et al.,1977; Ulmasov et al.,1992). The position of major Hsps and actin was determined by both autoradiography and subsequent staining of gels with silver(Creighton, 1990).

For immunoblotting, after sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE) of larval lysate prepared as above, the proteins were transferred to nitrocellulose membrane (Hybond ECL; Amersham)according to the manufacturer's protocol and reacted with monoclonal antibodies specific to the entire Drosophila Hsp70 family (7.10.3)and only Hsp70 (7FB) as previously described(Zatsepina et al., 2001). Immune complexes were detected via chemiluminescence (ECL kit;Amersham) and diaminobenzidine (DAB) (Sigma) with appropriate peroxidase-conjugated anti-rat secondary antibodies.

RNA was prepared by the standard method with 4 mol l^-1 guanidine isothiocyanate (Chomczynski and Sacchi,1987), separated by agarose gel electrophoresis and transferred to a membrane for hybridization (Sambrook and Fritsch, 1989) with a ClaI–BamHI fragment containing the Drosophila melanogaster hsp70 gene cloned into the BamHI site of pUC13 (McGarry and Lindquist, 1985). Hybridization was overnight at 42°C in 50%formamide, followed by two 20 min washes in 2×SSC, 0.2% SDS at 42°C,two 20 min washes in 1×SSC, 0.2% SDS at 42°C, and one 20 min wash in 0.2×SSC, 0.2% SDS at 68°C.

Flies were frozen and pulverized in liquid nitrogen, and the powder was suspended (1:5) in a buffer containing 20 mmol l^-1 Hepes, pH 7.9,25% (v/v) glycerol, 0.42 mol l^-1 NaCl, 1.5 mmol l^-1MgCl₂, 0.2 mmol l^-1 EDTA, 0.5 mmol l^-1phenylmethylsulfonyl fluoride (PMSF) and 0.5 mmol l^-1dithiothreitol, which was centrifuged at 100 000 g for 20 min. The supernatants were frozen in liquid nitrogen and stored at -70°C. The protein concentration of the extracts was estimated with a modified Lowry method (Ulmasov et al.,1992).

Consensus HSE probe (Wu et al.,1988) was prepared by annealing partially complementary oligonucleotides (ATCCGAGCGCGCCTCGAATGTTCTAGAA and CTCGCGCGGAGCTTACAAGATCTTTTCCA) in 10 mmol l^-1 potassium phosphate buffer, pH 8.2, in the presence of 0.1 mmol l^-1 NaCl. Single-stranded termini were filled with Klenow polymerase and[³²P]ATP (Sambrook and Fritsch,1989). For the gel mobility-shift assay, extracts containing 50μg of protein were mixed with 0.5 ng of [³²P] heat shock element(HSE) in binding buffer as described(Mosser et al., 1993). Binding-reaction mixture was incubated at room temperature (20°C) for 20 min. Free probe was separated from HSE–HSF complexes by electrophoresis in 5% polyacrylamide gels (Mosser et al.,1993). The gels were dried and exposed to X-ray film (Kodak X-Omat) at -70°C.

The basal thermotolerance of adult flies of all D. virilis strains used in the study exceeded that for its sister species D. lummei(Fig. 1; Table 1). The LT₅₀s for a 30-min heat shock are generally 1.3°C greater for D. virilis strains than for D. lummei and are inversely related to their latitude of origin. Similarly, basal thermotolerance for D. novamexicana exceeded that for its related species D. texana. Although both are from similar latitudes, D. novamexicana is from a more arid and warmer climate than D. texana(Table 1).

Within D. virilis and D. lummei, we failed to detect differences as large as those among species. This minor intraspecific variation in basal thermotolerance, moreover, was not inversely correlated with latitude. With the exception of strain 160, the LT₅₀ for D. virilis strains ranging from 32° to 53° N were within 0.2°C of one another and, with the exception of strain 1102, the LT₅₀ for D. lummei strains ranging from 45° to 60°N were within 0.1°C of one another. Strain 160 is a marker strain with at least one known recessive mutation on each autosome (broken bk in ch. II; gapped gp in ch. III; cardinal cd in ch. IV; peach pe in ch.V; and glossy gl in ch. VI); all other strains are wild-type.

Reciprocal hybrids of D. virilis and D. lummei had basal thermotolerances intermediate to those of the two parental species(Fig. 2A). The direction of the cross did not affect basal thermotolerance appreciably.

In D. virilis strain 9, pre-treatment improved thermotolerance(Fig. 2B), and the change in thermotolerance (i.e. LT₅₀ after PT minus LT₅₀ without PT) was correlated with the PT temperature. The impact of PT on thermotolerance in D. virilis strain 160 was nearly identical to that in strain 9, although the absolute values of LT₅₀s differed in these two strains (Fig. 1; data after PT not shown for strain 9). By contrast, pre-treatment reduced thermotolerance in D. lummei, and the change in thermotolerance was inversely correlated with the PT temperature(Fig. 2B). PT at 36°C and 37°C reduced survival of 40–40.5°C heat shock to 0%.

A D. virilis parent, regardless of sex, was sufficient for positive inducible thermotolerance in D. virilis–D. lummei hybrids (Fig. 2B). Although results differed slightly according to the direction of the cross, PT temperature and the interval between PT and heat shock, PT increased the LT₅₀ of D. virilis–D. lummei hybrids by 0.25–0.5°C, to where it was slightly less than the basal thermotolerance of pure D. virilis. Although this PT was modest, it clearly differs from the decreased thermotolerance after PT in D. lummei. In the hybrids, the effect of increasing PT temperature was intermediate to that in the parental species; it neither increased nor decreased inducible thermotolerance.

As Garbuz et al. (2002;figs 5, 6) have previously reported,under normal physiological conditions (25°C), D. virilis strain 9 and D. lummei strains 200 and 1102 did not differ in total protein synthesis, as indicated by [³⁵S]_L-methionine incorporation. At 1 h after mild heat shock (37.5°C), these D. lummei strains synthesized no less protein than did D. virilisstrain 9. By contrast, at 1 h after more-intense heat shock (40°C), these D. lummei strains clearly synthesized less protein than did D. virilis strain 9. These differences are manifest in all major classes of heat-shock proteins that are typically distinguishable in one-dimensional electrophoresis of synthesized proteins. Additional determinations of protein synthesis via [³⁵S]_L-methionine incorporation(Figs 3, 4) corroborate and extend these conclusions. First, as inclusion of two additional D. virilis strains(160 and 1433) demonstrates, these are true interspecific differences rather than a feature unique to D. virilis strain 9. Second, these interspecific differences are even greater after more-intense heat shock(40.5°C and 41°C), although both species are capable of some protein synthesis at all temperatures studied. After these intense heat shocks, Hsp40 and small Hsps are especially reduced in D. lummei relative to D. virilis. Third, interspecific hybrids (D. lummei strain 200× D. virilis strain 9) exhibit the D. virilis pattern of protein synthesis. The same is true of the reciprocal cross (data not shown). Finally, the species pair D. novamexicana and D. texana, which correspond in thermotolerance to D. virilis and D. lummei, respectively, exhibit parallel differences in protein synthesis (Fig. 4). Interestingly, D. virilis strain 160, which both has numerous mutations and is the most thermosensitive of the D. virilis strains(see above), is not obviously defective in terms of protein synthesis after heat shock. Also, both D. lummei and D. virilis exhibit expression of a high-molecular-mass heat-shock protein under some conditions(Fig. 3).

According to both Garbuz et al.(2002) and the present study(Figs 3, 4), Hsp70 is quantitatively the major heat-shock protein in the species and strains examined. To examine the conclusions of the previous paragraph in detail for this specific Hsp, we have replotted the data from Fig. 2Aof Garbuz et al. (2002) and included data for additional strains and conditions(Fig. 5). Indeed, Hsp70 levels for D. lummei strains are within the range for the various D. virilis strains 3 h after a 37.5°C heat shock. After more-severe heat shock, Hsp70 levels in the D. lummei strains are below the range for D. virilis strains, corresponding to their differing thermotolerances(Figs 1, 2). Data for the more-thermotolerant D. novamexicana and less-thermotolerant D. texana recapitulate this pattern (Fig. 5B). Immunoblots of Hsp70 levels(Fig. 5C) clearly emphasize the differing Hsp70 levels in these species after intense heat shock. Detailed examination of the Hsp70 family reveals that the differences between Hsp70s of D. virilis and D. lummei are qualitative as well as quantitative (Fig. 6; see also Garbuz et al., 2002). D. virilis exhibits three isoforms recognizable by antibody 7FB, which in D. melanogaster reacts only with Hsp70; D. lummei exhibits only two of these three isoforms. D. virilis also exhibits two inducible isoforms not recognized by 7FB but recognizable by antibody 7.10.3,which reacts with all Hsp70 family members in most species examined; D. lummei exhibits only one of these two isoforms. Interspecific hybrids(D. lummei strain 200 × D. virilis strain 9) exhibit the D. virilis pattern (Fig. 6), as does the reciprocal cross (data not shown).

The thermal sensitivity of hsp70 mRNA transcription in general resembled that for Hsp70 protein. Thus, D. lummei synthesized no detectable hsp70 mRNA during heat shock at 40–41°C, whereas D. virilis strain 9 did so in abundance(Fig. 7). These results for D. virilis strain 9 exemplify those obtained for other D. virilis strains. Despite its low thermotolerance(Fig. 1) and unexceptional Hsp70 protein levels (Fig. 3), D. virilis strain 160 typically synthesized more hsp70 mRNA than did other D. virilis strains(Fig. 7; and other data not shown).

As indicated by electrophoretic mobility-shift assays(Fig. 8), D. virilisunderwent HSF trimerization at temperatures of >32°C, whereas D. lummei exhibited activation at temperatures of >31°C. This difference, although small, is again consistent with the environmental regimes of these species.

Many aspects of the thermal phenotypes of the species under study exhibit countergradient variation similar to that in other taxa(Feder and Hofmann, 1999; Ulmasov et al., 1992; Zatsepina et al., 2001); i.e. the magnitude and threshold for traits correspond to the thermal environments in which the species occur. Thus, the lower-latitude species D. virilis exceeds the higher-latitude species D. lummei in basal thermotolerance (see also Garbuz et al.,2002), inducible thermotolerance and temperature threshold for HSF activation. Similarly, the more xeric D. novamexicana exceeds the more mesic D. texana in the former two traits. As also shown previously for protein (Garbuz et al.,2002), D. virilis and D. lummei express similar amounts of heat-shock mRNA and protein at low to moderate heat-shock temperatures. Intense heat shock (e.g. 40–41°C) almost abolishes heat-shock mRNA and protein expression in D. lummei, whereas considerable expression persists after such heat shock in D. virilis. For example, Hsp70 concentration is similar in D. virilis and D. lummei after 37.5°C heat shock but is markedly greater in D. virilis than in D. lummei after more-intense heat shocks(Fig. 5A); a similar pattern is evident for D. novamexicana and D. texana(Fig. 5B). In D. lummei, moreover, the reduction in protein level after severe heat shock is even greater for the small Hsps and Hsp40 than for Hsp70 (Figs 3, 4).

Within D. virilis and D. lummei, by contrast, little such countergradient variation is evident. We speculate that extensive gene flow among populations swamps incipient adaptation to local conditions, a situation that may not be universal in Drosophila species(Michalak et al., 2001). As in D. melanogaster (Krebs et al.,2001), adaptation to laboratory conditions does not seem to contribute to this intraspecific similarity, as the two strains from Tashkent,Uzbekistan (T53 and T61) exhibit virtually identical patterns of thermotolerance and Hsp induction despite the fact that T61 was captured recently and the T53 more than 30 years ago. The only conspicuous departure from this pattern is for D. virilis strain 160, a marker strain with at least one known recessive mutation on each autosome. The basal thermotolerance for this strain is considerably lower than for all other D. virilis strains studied, suggesting that the mutations it bears interfere with this form of thermotolerance. Its expression of heat-shock proteins, which underlies inducible rather than basal thermotolerance, is as in other D. virilis strains, however.

The absence of inducible thermotolerance in the high-latitude species D. lummei is indeed distinctive. We know of no other Drosophila species that shares this absence. Remarkably, this species expresses a full complement of heat-shock proteins after heat pre-treatment(although in lesser magnitude than in D. virilis), suggesting that these proteins are not sufficient for inducible thermotolerance and that some unknown component is deficient.

An unusual feature of the virilis species group is its capacity for interspecific introgression of genetic material in the laboratory. Here,we demonstrate corresponding patterns in basal and inducible thermotolerance(Fig. 2), overall protein synthesis after heat shock (Figs 3, 4) and Hsp70 electromorphs(Fig. 6). For thermotolerance,the pattern for hybrids is intermediate to that for the two parental species and independent of the direction of the cross. For quantitative and qualitative variation in Hsp expression, the D. virilis pattern behaves as a dominant trait. For example, D. virilis exhibits three Hsp70-family protein isoforms recognizable by both antibody 7.10.3 and 7FB, D. lummei exhibits only two isoforms, and hybrids exhibit three isoforms (Fig. 6). Similarly, D. virilis exhibits two inducible Hsp70-family protein isoforms recognizable by antibody 7.10.3, D. lummei exhibits only one isoform,and hybrids exhibit two isoforms (Fig. 6). The Drosophila virilis group has great potential to elucidate the genetic basis for interspecific differences in thermal phenotype for two reasons. First, the group is far more diverse than the four species examined here; in essence, an independent but related phylad (the montana phylad) replicates the patterns of latitudinal and geographic replacement seen in the virilis phylad. Species within each phylad can be crossed with one another to produce partially fertile progeny;moreover, some species belonging to different phylads can also be crossed(Evgenev et al., 1982; Patterson and Stone, 1952). Second, markers present on each chromosome make possible the introgression of a single chromosome or even a portion of a chromosome bearing, for example,the hsp70 gene cluster, whose location is known(Evgenev et al., 1978). This potential is an unexploited and exciting opportunity for evolutionary physiology.

Finally, Garbuz et al.(2002) interpreted restriction fragment length polymorphism for hsp70 genes to suggest that the D. virilis–D. lummei and D. novamexicana–D. texana species pairs exhibit corresponding differences in hsp70 gene family copy number. Although less likely than differences in copy number, these data are also consistent with nucleotide polymorphisms in restriction sites and a constant number of gene copies. Part II of this series of research papers will show that the D. virilis–D. lummei species pair indeed differ in gene copy number,with the high-latitude D. lummei having lost some hsp70genes present in the low-latitude D. virilis. Thus, the mRNA, protein and thermotolerance differences reported in the present manuscript have at least part of their basis in the copy number of their encoding genes.

In Moscow, the research was supported by Russian Grants for Basic Science 0004-48-285 and 0004-32-243. In Chicago, research was supported by an International Supplement to NSF grant IBN99-86158. We thank A. Helin and D. Lerman for technical assistance.