Two-dimensional gel analysis of the heat-shock response in marine snails(genus Tegula): interspecific variation in protein expression and acclimation ability Free

Acclimation to higher temperatures leads to wide-ranging physiological changes in ectothermic organisms (Cossins and Bowler, 1987; Prosser,1991; Withers,1992). In general, animals from relatively constant thermal environments are less capable of modifying physiological function during acclimation than animals from environments that undergo more frequent and greater temperature changes. Such comparisons in phenotypic plasticity have been documented for animals that differ either in their biogeographical distribution pattern (e.g. temperate vs tropical), habitat preference(shallow versus deep waters) or seasonal temperature (e.g. summer vs winter) (Prosser,1991). Expression patterns of heat-shock proteins (Hsps) have been shown to vary between species that differ in heat tolerance and within species due to short-term heat-hardening, longer term seasonal acclimatization or laboratory acclimation (Feder and Hofmann,1999; Hoffmann et al.,2003). However, our knowledge of how species that differ in heat tolerance, e.g. closely related eurytherms vs stenotherms, vary in expression patterns of Hsps with changing acclimation temperature is still limited (Hoffmann et al.,2003; Tomanek and Somero, 1999, 2002; Tomanek, 2002).

As molecular chaperones, Hsps stabilize and refold reversibly denatured proteins and help degrade irreversibly denatured proteins(Feige et al., 1996; Frydman, 2001; Hartl and Hayer-Hartl, 2002; Young et al., 2004). Their role in conferring increased heat tolerance has been established. The onset temperature (T_on) of Hsp70 synthesis, the major heat-induced protein in most organisms, correlates in general positively with a species' temperature range (Feder and Hofmann, 1999; Nakano and Iwama, 2002; Tomanek and Somero, 1999). T_on of Hsp expression, however,is not fixed and can shift to higher temperatures after laboratory acclimation(Tomanek and Somero, 1999) and seasonal acclimatization to increasing temperatures(Barua and Heckathorn, 2004; Buckley et al., 2001; Chapple et al., 1998; Dietz and Somero, 1992; Hamdoun et al., 2003; Hofmann and Somero, 1995; Roberts et al., 1997), an adjustment I refer to as the phenotypic plasticity of the heat-shock response(HSR). A shift in T_on in response to acclimation is equivalent to a decrease in Hsp synthesis, at least at the lower temperatures that normally elicit a stress response, and is in part due to Hsp70 itself,such that an increase in steady-state levels will shift the activation of its own transcription and synthesis towards higher temperatures(Morimoto, 1998; Tomanek and Somero, 2002; Voellmy, 2004). Interestingly,a heat-sensitive intertidal marine snail species of the genus Tegulashifts T_on of a Hsp70 band more after acclimation to higher temperatures than does a heat-tolerant congener(Tomanek and Somero, 1999). It is therefore possible that species that vary in heat tolerance activate the stress response at a common temperature following acclimation to warmer temperature. This suggests that species with the highest heat tolerance may have a relatively lower capacity to further adjust the stress response to warmer acclimation temperatures. If so, this limit to acclimation capacity of Hsp expression could explain the limited ability of more heat-tolerant organisms to increase tolerance to even greater heat stress(Cavicchi et al., 1995; Hoffmann et al., 2003; Stillman, 2003; Ushakov et al., 1977; Zatsepina et al., 2001).

However, it is not known if these results of a single Hsp band can be extrapolated to the entire complement of heat-induced proteins. Simply,previous acclimation studies have only used one-dimensional gel electrophoresis to analyze heat-induced protein bands that were close to a molecular mass class that is characteristic for a given Hsp family(Hofmann and Somero, 1996; Tomanek and Somero, 1999),despite the fact that the Hsp complement has been resolved since the early days of two-dimensional gel electrophoresis (2D-GE; Mirault et al., 1978).

In the present work, I address the effect of acclimation on the synthesis of the Hsp complement in three cool–temperate Tegula snail species that differ in their vertical distribution patterns along the subtidal to mid-intertidal axis in rocky shore habitats and differ in upper lethal temperatures by approximately 6.5°C(Tomanek and Somero, 1999). T. funebralis, which occupies the mid- to low-intertidal zone,experiences greater and more variable temperatures on a daily scale and is more heat-tolerant than the two low-intertidal to subtidal zone congeners T. brunnea and T. montereyi(Tomanek and Somero, 1999). Furthermore, T. funebralis differs from the two subtidal species in having (i) a higher T_on and a wider temperature range of Hsp synthesis, (ii) a faster rate to reach pre-stress levels of Hsp synthesis after a thermal exposure typical for the mid-intertidal zone, and (iii)smaller changes in inducible and steady-state levels of Hsp70 and generally higher levels of heat-shock transcription factor-1 (HSF-1) during acclimation to increasing temperatures (Tomanek,2002; Tomanek and Somero, 1999, 2000, 2002). These interspecific differences enable T. funebralis, but not T. brunnea, to better cope with the thermal variation in the mid-intertidal zone(Tomanek and Sanford, 2003)and further support the hypothesis that greater heat tolerance may correlate with a limited capacity to modify the stress response.

In this study, I tested for the first time how acclimation affects the incorporation of ³⁵S-labeled methionine and cysteine into over 30 heat-induced proteins or their variants in gill tissue, by separating the Hsp complement using 2D-GE. My results show that the more heat-tolerant species displays a far lower capacity to adjust Hsp synthesis following acclimation to warmer temperatures than the two heat-sensitive species. These results contradict the general observation that animals from thermally more variable environments have a greater capacity to acclimate or modify physiological function in response to changing (here increasing) temperatures than animals from environments with moderate temperature variation. Furthermore, it has been proposed that the benefit of acclimation to increasing temperatures is to prepare the animal for the possibility that the environment will become warmer(Huey et al., 1999; Leroi et al., 1994). I discuss how the differences in phenotypic plasticity between the Tegulacongeners support this interpretation in the context of the thermal characteristics of their respective vertical distribution ranges in the intertidal zone. Finally, the results also suggest that animals from thermally more-variable relative to less-variable environments may be more sensitive to future temperature increases.

Large adult animals of all three Tegula congeners were collected at Hopkins Marine Station of Stanford University in Pacific Grove, California,USA (36°36′N, 121°54′W) at the beginning of the months of June and July 2003 and acclimated to 13°C and 22°C in circulating seawater aquaria for 21 and 20 days, respectively. Animals were kept submersed and fed with freshly collected kelp (Macrocystis pyrifera).

After acclimation to 13°C and 22°C I dissected gill tissue under non-heat-shock-inducing conditions (13°C and 22°C, respectively). Tissues were placed into microcentrifuge tubes holding 300 μl of filtered(0.2 μm) seawater containing 10 mmol l^–1 glucose. Tubes were pre-incubated to either 13°C or 24°C for 13°C- and 22°C-acclimated animals, respectively. Gill tissue was aerated every 20 min. At the start of the experiment tissues were incubated at 13°C,24°C, 27°C and 30°C for 2.5 h (for further details, see Tomanek and Somero, 1999). Gill tissue was not directly transferred to 27°C and 30°C but instead incubated first to 24°C (13°C-acclimated animals) and 27°C (for the 30°C incubation) for 5 min to simulate a more gradual temperature increase. After incubation at the experimental temperature they were transferred to a common temperature of 13°C for 15 min before being incubated with ³⁵S-labeled methionine/cysteine (6.78 MBq ml^–1; Perkin Elmer) for 4 h. At the end of the experiment tissues were washed in filtered seawater, frozen on dry ice and stored at–70°C. Tissues were first homogenized using a sample grinding kit containing an abrasive grinding resin in 200 μl of lysis buffer containing 8 mol l^–1 urea, 4% CHAPS and 2% Pharmalyte 3-10 and subsequently precipitated using the 2-D clean-up kit (both kits were from Amersham Bioscience, Piscataway, NJ, USA). The precipitate was dissolved in 120 μl of rehydration buffer (8 mol l^–1 urea, 2% CHAPS,0.5% IPG buffer pH 4–7 (Amersham Bioscience), 0.002% Bromophenol Blue). Levels of ³⁵S-labelled amino acids incorporated into newly synthesized proteins were determined by precipitation with trichloroacetic acid (for details, see Tomanek and Somero,1999).

Samples were added (5×10⁵ cts min^–1) to the rehydration buffer (see above) during overnight rehydration of Immobiline DryStrip gels (7 cm long, pH 4–7; Amersham Bioscience). Rehydrated strips were run on a Multiphor II electrophoresis system (Amersham Bioscience)for 1 min at 200 V and for 2 h 50 min at 3500 V before being stored at–70°C.

For the second dimension, strips were equilibrated twice for 15 min in 10 ml of equilibration buffer (50 mmol l^–1 Tris-HCl, pH 8.8, 6 mol l^–1 urea, 30% glycerol, 2% SDS, 0.002% Bromophenol Blue);first with 100 mg dithiothreitol and second with 250 mg iodoacetamide. Strips were placed on ExcelGel 12–14% gradient gels (Amersham Bioscience) and run according to instructions. Afterwards gels were incubated in Amplify fluorographic reagent (Amersham Bioscience) for 30 min and subsequently dried overnight at 60°C (for the first 2 h) using a vacuum.

Gels were exposed to Hyperfilm (Amersham Bioscience) for 72, 48, 24, 12 and 6 h to obtain a range of exposure intensities, scanned on a densitometer(Sharp JX-330) and analyzed using Image Master 2D software (Amersham Bioscience). Specific spot volumes were normalized against total spot volume within a single gel and square root-transformed to normalize values for statistical analysis. Similar intensities were found for film that was exposed for 48 h after incubation at 13°C and 24°C, 24 h after 27°C incubation and 12 h after 30°C incubation. Expression values of a single Hsp of all species from all acclimation and incubation temperatures were fitted into a general linear model to estimate the average effect for each species, each acclimation and incubation temperature within a species, and an interaction effect between each combination of acclimation and incubation temperature (Software R, version 1.9.1). All treatments had an N-value of three (only exception: T. funebralis incubated to 24°C after acclimation to 22°C, N=1). The P-values(≤0.05 and ≤0.10) that are reported in Table 1 are for the average effect of acclimation on Hsp expression within an incubation temperature for a particular species.

Most heat-induced proteins in Tegula had a molecular mass between 68 and 90 kDa; additional ones were found around 40 and 24 kDa(Fig. 1A,B). Although the exact identity of these proteins has not been determined (see below), almost all of these heat-induced proteins were found at or close to a molecular mass that typically represents a family of functionally related Hsps, e.g. Hsp70, Hsp77 and Hsp90. Additionally, under thermal stress, cells preferentially synthesize Hsps while inhibiting the synthesis of the majority of other proteins(Lindquist, 1993; Storti et al., 1980). The downregulation of non-Hsp synthesis is illustrated in Fig. 1A. I am therefore assuming that the majority of heat-induced proteins, and especially the highly expressed ones, are Hsps or variants thereof.

The three Tegula species showed several clusters of Hsps between 68 and 90 kDa (Fig. 2A): one cluster with nine proteins of 71 kDa contained the most highly expressed proteins. There was a single 74 kDa protein. Two clusters contained four proteins of 76 kDa (T. funebralis and T. brunnea only) and up to six of 77 kDa. There were two (T. brunnea and T. montereyi) or three (T. funebralis) 88 kDa proteins. T. montereyi synthesized a cluster of four 72 kDa proteins, but lacked the 76 kDa cluster (Fig. 2A). T. brunnea did not express one of three heat-induced 40 kDa proteins(Fig. 2B). All species shared five proteins of about 24–25 kDa(Fig. 2C).

To examine the phenotypic plasticity of acute Hsp synthesis I acclimated snails of all three species to 13°C and 22°C, conditions within the seawater temperature range of T. funebralis, but at the high extreme of what T. brunnea and T. montereyi can tolerate. Of the>30 heat-induced proteins that I found in Tegula,Hsps71.1–71.4 were the most highly expressed (Figs 2A, 3). Acclimation decreased acute synthesis in three out of the four Hsp71s in the subtidal species, T. brunnea and T. montereyi, but not in the intertidal species T. funebralis (Table 1). The remaining four Hsp71s(Fig. 2A) were induced to one third or less of the level of Hsp71.1-4 (data not shown). Acclimation attenuated synthesis levels of these Hsps more in T. brunnea(Hsp71.5, 71.6 and 71.9) and T. montereyi (Hsp71.9) than in T. funebralis (Table 1). Four additional 71 kDa isoforms that were not shared by all species had more acidic isoelectric points (pI) than Hsp71.1 and were infrequently expressed (Figs 1B, 2A).

Hsp71.8 and 74.1 (Fig. 4)were heat-induced as well as constitutively expressed at 13°C(preliminarily confirmed with western analysis, data not shown). Acclimation to 22°C led to an increase in acute synthesis under heat-shock in both T.brunnea and T. montereyi (Hsp74.1 only; Table 1). The synthesis of both proteins changed with acclimation in T. funebralis(Table 1). Synthesis levels of Hsp71.8 decreased at an incubation of 24°C; Hsp74.1 levels increased overall (significant main effect for acclimation).

Acclimation had the greatest effect on synthesis of Hsp77s in T. brunnea (Hsp77.2; 77.4 and 77.5; Fig. 5), and a lesser effect in T. montereyi (Hsp77.1) and T. funebralis (Hsp77.4; Table 1). T. montereyi and T. funebralis expressed two additional proteins (Hsp77.1 and 77.6) that I did not detect in T. brunnea (Fig. 2A).

Hsp76.1–76.4 were only detected in T. funebralis and T. brunnea (Fig. 2A). Synthesis levels decreased after acclimation to 22°C in four proteins detected in T. brunnea but only in one (Hsp76.1) in T. funebralis (see Hsp76.3 in Fig. 5 and Table 1).

Gill tissue of T. montereyi expressed a unique cluster of four 72 kDa proteins. Acclimation led to the attenuation of Hsp72.4 synthesis only(Fig. 5 and Table 1).

In response to heat, all species synthesized at least two proteins of about 88 kDa (Fig. 2A, not seen in all gels). Acclimation attenuated the synthesis of Hsp88.1 in T. brunnea and T. montereyi, but not in T. funebralis(Fig. 6 and Table 1). Synthesis of the two additional proteins, Hsp88.2 and Hsp88.3, was downregulated in T. montereyi and T. brunnea, respectively(Table 1).

Three proteins of about 40 kDa were synthesized in response to heat in T. funebralis and T. montereyi(Fig. 2B). T. brunneadid not synthesize one of the proteins (Hsp40.3). Acclimation attenuated the acute synthesis of Hsp40.2 in T. brunnea and of Hsp40.1 in T. montereyi (Fig. 6 and Table 1).

Hsp24.1 was the most highly expressed small Hsp and reached levels similar to the major Hsp71s (Figs 2Cand 7). Acclimation affected its synthesis in all species except T. funebralis. However, whereas synthesis decreased overall after acclimation to 22°C in T. montereyi, in T. brunnea synthesis levels increased at an incubation temperature of 27°C following acclimation to 22°C. Acclimation changed the synthesis of Hsp24.2 and Hsp24.3 in T. funebralis and T. montereyi, respectively(Table 1). Acute synthesis of Hsp25.1 was significantly attenuated after acclimation to 22°C in T. brunnea (30°C incubation) and in T. montereyi (24°C and 27°C incubation); but overall upregulated in T. funebralis(Table 1). Hsp27.1 was only expressed in T. funebralis (Fig. 2C). Hsp27.2 was expressed in all three species and changed with increasing acclimation temperature in T. montereyi(Table 1).

Species differed not only in how acclimation affected acute Hsp synthesis but also in the T_on of Hsp synthesis within the fairly narrow range of incubation temperatures used in the present study, although the study was not designed to detect possible differences in T_on at lower temperatures (<24°C; Tomanek and Somero, 1999). For three out of the four shared Hsp77s T. brunnea and T. montereyi induced synthesis at 24°C, whereas T. funebralisdid not initiate synthesis until 27°C after acclimation to 13°C and 22°C (Fig. 5). Following acclimation to 13°C, the synthesis of all four Hsp76s was induced at 24°C in T. brunnea, but at 27°C in T. funebralis(Fig. 5). Hsp40.1 showed a significant effect with an increase in incubation temperature to 24°C in T. brunnea and T. montereyi, and with an increase to 27°C in T. funebralis (Fig. 6).

In Tegula congeners, phenotypic plasticity of the HSR is protein-specific and varies among species that differ in thermotolerance. Synthesis of almost all the major Hsps was attenuated after acclimation to 22°C (vs 13°C) in the heat-sensitive subtidal species, T. brunnea and T. montereyi (Figs 3, 4, 5, 6, 7 and Table 1). There were noticeable differences between the two species, however, specifically in the number of Hsp77s whose synthesis changed with acclimation. In contrast, in T. funebralis acclimation changed synthesis levels only for a small subset of Hsps (Hsp24.2, Hsp25.1, Hsp71.9, Hsp76.1 and Hsp77.4), including two constitutively expressed Hsps (Hsp71.8 and Hsp74.1), and not for the major Hsps (Hsp24.1, Hsp71.1-4, Hsp76.2-4 and Hsp77.1-6) (Figs 3, 4, 5, 6, 7 and Table 1).

The choice of acclimation temperatures (13°C and 22°C) was based on a previous study in which we used one-dimensional gel electrophoresis(Tomanek and Somero, 1999). In this study we found that an increase in acclimation temperature by 5°C can change T_on and the temperature of maximal Hsp synthesis(T_peak) for some Hsps. The upper acclimation temperature(22°C) was chosen because it is the highest temperature that the two subtidal species can tolerate over the 21-day acclimation period and is only a few degrees below temperatures at which T. funebralis shows the first signs of mortality when kept constantly submersed (personal observation). Although the results of the previous and the present study differ in detail,both show a greater shift in T_on (attenuation of acute Hsp synthesis) for the highly expressed Hsp70s in T. brunnea and T. montereyi than in T. funebralis. In the present study, however,I was able to resolve the bands of acutely expressed proteins into a complement of over 30 protein variants that are represented by spots that differ in molecular mass and isoelectric point (pI).

When compared to the Hsp complements of tropical fish of the genus Poeciliopsis, the eurythermal teleost Gillichthys mirabilis,Drosophila and intertidal limpets, Tegula snails show a surprisingly high number of Hsp variants (Garbuz et al., 2003, 2002; Kültz, 1996; Norris et al., 1997, 1995; Sanders et al., 1991; White et al., 1994). At this point I do not know if these variants are true isoforms (paralogous homologs)or represent post-translationally modified proteins. But a common molecular mass, proximate isoelectric points (pI), similar expression patterns and known post-translational modifications (PTM) of Hsps suggest that several of the proteins are more closely related to each other and are members of a common Hsp family. For example, Hsp71.1-4 are the most highly expressed Hsp70s, and the synthesis of three out of the four was significantly attenuated in T. brunnea and T. montereyi after acclimation to 22°C(Table 1). These proteins are likely paralogs of Hsp70.

Several highly induced Hsps are clustered around a similar molecular mass and pI (Hsp77s, 76s and 72s; Figs 1, 2). Since there is no report on the clustering of so many isoforms within an Hsp family, it is possible that some of these variants represent PTMs. For example, the mammalian BiP (or Grp78) binds to unfolded proteins in the endoplasmic reticulum and its function is in part regulated via phosphorylation(Gething, 1997); and a mammalian mitochondrial Hsp70 is characterized by its Ca²⁺-dependent autophosphorylation activity(Leusteck et al., 1989). Phosphorylation is commonly observed to regulate the activity and cause the variation in synthesis patterns of small Hsps(Arrigo and Landry, 1994; Norris et al., 1997). The extent and role of PTMs must be addressed further to assess their importance for creating the patterns of Hsp variants in Tegula and their contribution to the differences in phenotypic plasticity within a cluster of Hsps.

Interspecific differences in phenotypic plasticity of Hsp synthesis were not only common between T. funebralis on the one hand and the two subtidal species on the other, but the plasticity of the Hsp77 cluster varied even between T. brunnea and T. montereyi(Fig. 2A and Table 1). After acclimation,the synthesis of three out of four variants changed in T. brunnea,but only one did in T. montereyi. Furthermore, these two species either expressed the Hsp76 (T. brunnea) or Hsp72 (T. montereyi) cluster (Fig. 2A), but not both.

Two constitutively expressed Hsps, 71.8 and 74.1, were also expressed during heat-shock (Fig. 4). I confirmed their identity with an anti-Hsp70 antibody that I had used previously (Tomanek and Somero,2002). The synthesis of both proteins changed with acclimation in all three species; the direction of change differed among the species in case of Hsp71.8, but all species showed higher levels of Hsp74.1 synthesis after acclimation to 22°C (Fig. 4). Although acclimation affected the synthesis of several heat-induced proteins in T. funebralis, their expression levels were comparatively low. Thus, changes in the synthesis of Hsp71.8 and 74.1 suggest that the effect of acclimation on Hsps depends on differences in their expression pattern (constitutive and inducible vs inducible only).

Although T. funebralis tolerates higher temperatures(Tomanek and Somero, 1999),its ability to modify the stress response during thermal acclimation is limited in comparison to T. brunnea and T. montereyi (Figs 3, 4, 5, 6, 7 and Table 1). In addition to the evidence presented here, previous work showed that T. funebralis does not adjust steady-state levels of two Hsp70 bands, Hsp90 and the heat-shock transcription factor 1 (HSF1) after acclimation(Tomanek and Somero, 2002). On the other hand, T. funebralis recovers from a heat-shock at a faster rate (Tomanek and Somero,2000), has higher steady-state levels of HSF1(Tomanek and Somero, 2002),survives thermally stressful low-tide periods better than T. brunnea(Tomanek and Sanford, 2003),and synthesizes Hsps over a much wider temperature range than the two subtidal species (Tomanek and Somero,1999). All of these interspecific differences suggest that T. funebralis is adapted to cope with the greater thermal variation that it experiences in the mid-intertidal zone, but has a limited ability to modify the synthesis of the major stress proteins. It has been suggested that the function of acclimation is not `to maximize performance at warm temperatures, but rather to protect against the possibility that the environment will become even hotter'(Huey et al., 1999; Leroi et al., 1994). T. funebralis may have lost the capacity to modify the acute expression of many of its Hsps, in part because it is adapted to the frequent and extreme temperature changes that are typical for the mid-intertidal zone. By occupying a thermal zone that becomes warmer on a daily basis, for example during midday low-tide periods, T. funebralis may have opted for maximal possible biochemical protection from thermal insults, and thereby almost eliminated the need for any further modifications with increasing acclimation temperature. The interpretation of acclimation as a means by which to prepare for a warmer environment thus provides a conceptual framework to explain the limited acclimatory capacity in more eurythermal animals.

The limited plasticity of acute Hsp synthesis in T. funebralisalso suggests a possible explanation for the observation that more heat-tolerant organisms are less capable of increasing tolerance to even greater heat stress. Clones and siblings of five organismal groups, including Coelenterata, Arthropoda, Echinodermata and Amphibia, which showed greater heat tolerance initially, acquired less of an increase in tolerance after acclimation to higher temperatures(Ushakov et al., 1977);strains and populations of Drosophila that were found or raised in warmer thermal habitats or temperatures during laboratory selection showed not only higher heat tolerance but also a limited short-term capacity to increase tolerance to greater heat stress, e.g. heat-hardening(Cavicchi et al., 1995; Hoffmann et al., 2003; Zatsepina et al., 2001); and more heat-tolerant porcelain crabs of the genus Petrolisthes show a lower capacity to acclimate the upper thermal limits of cardiac function(Stillman, 2003). To date, no other cellular mechanism has been proposed to explain this relationship.

An important objective of this study and our previous studies has been to elucidate how the interspecific variation in Hsp expression patterns contributes to setting the vertical distribution limits of Tegulacongeners. Our previous comparisons of the T_on of Hsp synthesis and the higher body temperatures that snails experience in the mid-intertidal zone suggested that T. funebralis activates the HSR frequently under natural conditions(Tomanek and Somero, 1999). One of the hallmarks of the HSR is the strong and preferential synthesis of Hsps at elevated temperatures, while the synthesis of non-stress proteins is suppressed (Lindquist, 1993; Storti et al., 1980). Fig. 1A illustrates the attenuation of synthesis of most non-stress proteins at a temperature that T. funebralis frequently experiences in the mid-intertidal zone(30°C). Thus, T. funebralis not only has to bear the costs of increased chaperoning activity but also of the disruption of protein homeostasis during midday low-tide periods that are most likely to activate Hsp synthesis. We know from studies on Drosophila and yeast that there are substantial metabolic costs associated with the HSR(Krebs and Feder, 1997; Sanchez et al., 1992). This may explain the slower growth rates shown by T. funebralis in comparison to T. brunnea and T. montereyi(Frank, 1965; Paine, 1969; Watanabe, 1982). Two-dimensional gel analysis improves the identification and detection of Hsp synthesis and will allow us to clarify when and to what extent the de novo synthesis of Hsps occurs under natural conditions, and thus to assess the costs of being able to cope with a thermal niche as extreme as the mid-intertidal zone.

The present study raises several new questions that have to be answered to better understand the complex role of the interspecific variation in Hsp synthesis and acclimation ability in setting limits to the thermal environment an organism can occupy. What do the many heat-induced protein variants represent: Hsp-families, Hsp-isoforms or their PTMs? What limits phenotypic plasticity on the cellular level: higher levels of HSF1, the interaction of a multi-chaperone complex with HSF1, or regulatory steps downstream of HSF1 binding to the Hsp promoter (Buckley et al., 2001; Morimoto,1998; Tomanek and Somero,2002)? Why does acclimation affect Hsps differently (e.g. Hsp77s),even among species that are similar in heat tolerance: are the likely explanations transcriptional effects or organelle-specific functions? What are the relevant PTMs of heat-induced proteins in Tegula and how do they contribute to the interspecific differences in heat tolerance? More specifically, are the Hsp76 and Hsp72 clusters homologous, and if not, how do these proteins differ and what is their function (thus explaining the interspecific differences in T_on of Hsp76 shown in Fig. 5)?

Another urgent issue is the question of how global climate change will affect organisms that differ in heat tolerance and phenotypic plasticity. Although more heat-tolerant organisms seem poised to cope better with an increase in average temperature and the occurrence of thermal extremes, their physiological limits are close to their maximal body temperatures(Somero, 2002) and they show a limited capacity to further modify these limits, so they may turn out to be particularly vulnerable to even a small increase in temperature.

I want to thank Dr Duncan Temple Lang of the Department of Statistics,University of California, Davis, for advice on thestatistical analysis, and Drs George Somero and Dietmar Kültz for valuable suggestions that improved the manuscript. The study was supported by National Science Foundation grant IBN0133184 awarded to Dr George N. Somero and by the David and Lucile Packard Foundation through the PISCO program.