In our previous studies of heat-shock protein (hsp) expression in congeneric marine gastropods of the genus Tegula, we observed interspecific and acclimation-induced variation in the temperatures at which heat-shock gene expression is induced (Ton). To investigate the factors responsible for these inter- and intraspecific differences in Ton, we tested the predictions of the ‘cellular thermometer’ model for the transcriptional regulation of hsp expression. According to this model, hsps not active in chaperoning unfolded proteins bind to a transcription factor, heat-shock factor-1 (HSF1), thereby reducing the levels of free HSF1 that are available to bind to the heat-shock element, a regulatory element upstream of hsp genes. Under stress, hsps bind to denatured proteins, releasing HSF1, which can now activate hsp gene transcription. Thus, elevated levels of heat-shock proteins of the 40, 70 and 90 kDa families (hsp 40, hsp70 and hsp90, respectively) would be predicted to elevate Ton. Conversely, elevated levels of HSF1 would be predicted to decrease Ton. Following laboratory acclimation to 13, 18 and 23°C, we used solid-phase immunochemistry (western analysis) to quantify endogenous levels of two hsp70 isoforms (hsp74 and hsp72), hsp90 and HSF1 in the low- to mid-intertidal species Tegula funebralis and in two subtidal to low-intertidal congeners, T. brunnea and T. montereyi. We found higher endogenous levels of hsp72 (a strongly heat-induced isoform) at 13 and 18°C in T. funebralis in comparison with T. brunnea and T. montereyi. However, T. funebralis also had higher levels of HSF1 than its congeners. The higher levels of HSF1 in T. funebralis cannot, within the framework of the cellular thermometer model, account for the higher Ton observed for this species, although they may explain why T. funebralis is able to induce the heat-shock response more rapidly than T. brunnea. However, the cellular thermometer model does appear to explain the cause of the increases in Ton that occurred during warm acclimation of the two subtidal species, in which warm acclimation was accompanied by increased levels of hsp72, hsp74 and hsp90, whereas levels of HSF1 remained stable. T. funebralis, which experiences greater heat stress than its subtidal congeners, consistently had higher ratios of hsp72 to hsp74 than its congeners, although the sum of levels of the two isoforms was similar for all three species except at the highest acclimation temperature (23°C). The ratio of hsp72 to hsp74 may provide a more accurate estimate of environmental heat stress than the total concentrations of both hsp70 isoforms.
In the face of thermal, chemical or physiological stresses that cause unfolding of proteins, cells preferentially express heat-shock (stress) proteins (hsps), a process called the heat-shock (or stress) response. As molecular chaperones, hsps protect proteins from denaturation and facilitate either their refolding into native conformations or, in the case of irreversibly damaged proteins, their removal from the cell through proteolysis (Bukau and Horwich, 1998; Frydman, 2001; Hartl, 1996; Morimoto, 1998; Parsell and Lindquist, 1993, 1994; Sherman and Goldberg, 1996). The heat-shock response is regulated in part by a transcription factor, heat-shock factor-1 (HSF1), which specifically stimulates the transcription of hsp-encoding genes (Fig. 1) (Voellmy, 1996; Wu, 1995).
Regulation of HSF1 activity is complex and is determined in part by intrinsic properties of the molecule, notably by heat-induced changes in its conformation (Farkas et al., 1998; Goodson and Sarge, 1995; Larson et al., 1995; Zhong et al., 1998), which enable inactive monomers to assemble into trimers that can bind to the heat-shock element (HSE), a gene regulatory element found upstream of most hsp-encoding genes (Bienz and Pelham, 1987; Fernandes et al., 1994). These studies suggested that HSF1 itself ‘senses’ thermal stress and may be an important type of ‘cellular thermometer’. However, several studies have shown that the temperature at which HSF1 is activated is not fixed (Abravaya et al., 1991; Clos et al., 1993), which implies that extrinsic factors also play important roles in regulating the activity of HSF1. Among these factors are hsps of the 40, 70 and 90 kDa size classes (hsp40, hsp70 and hsp90, respectively), which interact directly with HSF1 in a multi-protein complex under in vivo conditions (Fig. 1) (Abravaya et al., 1992; Ali et al., 1998; Baler et al., 1992, 1996; Bharadwaj et al., 1999; Duina et al., 1998; Marchler and Wu, 2001; Mosser et al., 1993; Rabindran et al., 1994; Shi et al., 1998; Zou et al., 1998).
It has been proposed that these hsps keep HSF1 ‘locked’ in the multi-protein complex under non-stressful conditions and thereby prevent it from binding to the heat-shock element and stimulating transcription of hsp genes. Upon environmental or physiological stresses that cause proteins to unfold, this complex dissociates because hsps begin to bind preferentially to unfolded proteins rather than to HSF1 (Parsell and Lindquist, 1993, 1994). Free HSF1 monomers are then able to move to the nucleus (Mercier et al., 1999; Westwood et al., 1991), trimerize and bind to the HSE. HSF1 trimers that are transcriptionally competent when bound to the HSE are hyperphosphorylated (Holmberg et al., 2001) (and see references therein). To render HSF1 transcriptionally active, an additional unknown structural transition is required (Shi et al., 1998; Zuo et al., 1995). This cascade of events is, in general, reversed during the attenuation of the heat-shock response, but additional proteins play a regulatory role (Satyal et al., 1998).
In this regulatory scheme, which is termed the ‘cellular thermometer’ model of hsp expression, hsps play an autoregulatory role in governing their own synthesis by determining the levels of free HSF1 that exist in the cell (Craig and Gross, 1991; DiDomenico et al., 1982; Lindquist, 1993; Morimoto, 1998). Thus, under conditions at which endogenous levels of hsps increase, for example, in response to acclimation to increasing temperature, the formation of the HSF1–hsp70–hsp90–hsp40 complex is favored, leading to a higher Ton (the temperature at which heat-shock gene expression is induced) for synthesis of hsps. Although HSF1 transcription is not induced by heat stress and HSF1’s activity is regulated post-translationally (Rabindran et al., 1991; Sarge et al., 1993), the cellular thermometer model predicts that elevated concentrations of HSF1 would, other things such as hsp70 and hsp90 levels being equal, favor lower Ton values for hsp gene transcription because of the presence of large amounts of free HSF1 for binding to the HSE.
The experiments described below were conducted to test these and other predictions of the cellular thermometer model in the context of the observed interspecific variation in Ton and in the kinetics of induction of hsps in gastropods of the genus Tegula (Tomanek and Somero, 1999, 2000). Congeners of the turban snail Tegula from temperate regions inhabit widely differing thermal environments, including cool subtidal habitats (e.g. T. brunnea and T. montereyi) and warmer and thermally variable intertidal habitats (e.g. T. funebralis). These three congeners have been shown to differ in (i) heat tolerance, (ii) Ton for the synthesis of hsp38, hsp70 and hsp90, (iii) temperatures at which synthesis of hsps is maximal (Tpeak), (iv) upper thermal limits for hsp synthesis (Toff), (v) rates of hsp induction and (vi) the duration of hsp synthesis following heat shock (Tomanek and Somero, 1999, 2000). Thus, these closely related species (Hellberg, 1998) provide an excellent study system for elucidating interspecific variation in the regulatory mechanisms that govern the expression of hsps. In addition, acclimatory changes in the heat-shock responses of these species occur: Ton and Tpeak shift upwards with warmer acclimation temperatures (Tomanek and Somero, 1999). Thus, the applicability of the cellular thermometer model can be tested in the context of acclimation-induced variation within a species as well as in the analysis of interspecific variation.
To test the predictions of the cellular thermometer model, we applied solid-phase immunochemical methods (western analysis) to quantify endogenous levels of hsp70, hsp90 and HSF1 in field-acclimatized and laboratory-acclimated (13, 18 and 23°C) Tegula. Because T. funebralis encounters greater heat stress in its habitat and grows significantly more slowly than T. brunnea and T. montereyi (Frank, 1965; Paine, 1969), we also analyzed the relative levels of hsp70 isoforms that appear to be associated either with normal protein synthesis or with repair of thermal damage to proteins. Our results indicate that the cellular thermometer model can account for intraspecific but not interspecific variation in Ton and that the relative levels of expression of different isoforms of hsp70 may provide a better measure of heat stress than the total expression of all hsp70 isoforms.
Materials and methods
Organisms, distribution patterns and collection sites
The three Tegula congeners used in this study differ in their biogeographic patterns and vertical distributions as follows: Tegula brunnea and T. montereyi inhabit the subtidal to low-intertidal zones of the eastern Pacific Ocean from Cape Arago, Oregon, USA (43°25′N), to the Channel Islands, California, USA (34°0′N), and from Sonoma County, California (38°17′N), to the Channel Islands, respectively (Abbott and Haderlie, 1980; Riedman et al., 1981; Watanabe, 1984). Tegula funebralis is found in the low- to mid-intertidal zone and has a wider latitudinal distribution range, from Vancouver Island, British Columbia, Canada (48°25′N), to central Baja California, Mexico (28°0′N) (Abbott and Haderlie, 1980; Riedman et al., 1981).
Specimens of all species were collected at Hopkins Marine Station of Stanford University in Pacific Grove, California, USA (36°36′N, 121°54′W). Large adults were used exclusively in all experiments, and the sizes of specimens were similar among all three species.
Thermal acclimation and experimental design
Specimens of all species were collected for acclimation in mid-July 1997 and either dissected immediately (field-acclimatized control group) or kept in temperature-controlled (13, 18 or 23°C) circulating seawater aquaria for 30–34 days. Acclimation temperatures were chosen to be within the range of sea surface temperatures found within the latitudinal distribution range of the species, except for 23°C, which is above the commonly experienced sea surface temperatures of T. brunnea and T. montereyi.
Specimens were kept constantly immersed and fed regularly with freshly collected giant kelp (Macrocystis pyrifera).
Gill tissue was dissected under conditions that do not induce heat shock (13°C) and immediately placed in 200 μl (T. funebralis, 15.0–25.0 mg wet mass) or 300 μl (T. brunnea and T. montereyi, 30.0–45.0 mg wet mass) of homogenization buffer [32 mmol l–1 Tris-HCl, pH 7.5 at 4°C, 2 % (w/v) SDS, 1 mmol l–1 EDTA, 1 mmol l–1 Pefabloc (Boehringer Mannheim), 10 μg ml–1 pepstatin and 10 μg ml–1 leupeptin]. Samples were stored at –70°C. To prepare homogenates for immunoblotting, the frozen samples were thawed in a dry bath for 5 min at 100°C and then homogenized with a silicone pestle. Homogenates were incubated at 100°C for 5 min, homogenized a second time, and then centrifuged at 15 800 g for 15 min. The supernatant was removed and stored at –70°C. Protein concentrations were determined using the Micro-BCA assay (Pierce) according to the manufacturer’s instructions.
Gel electrophoresis and immunodetection protocol
Proteins were separated electrophoretically according to size on a mini-gel apparatus (BioRad) for 45–50 min at 200 V (5 μg of protein per lane for hsp70 and 25 μg of protein per lane for hsp90 and HSF1). Subsequently, proteins were transferred for either 75 min (hsp70 and hsp90) or 90 min (HSF1) at 80 V (BioRad Protean apparatus) onto nitrocellulose membranes (Nitrobind, Schleicher and Schuell) soaked for at least 2 h in transfer buffer [25 mmol l–1 Tris-base, 0.193 mol l–1 glycine, 20 % (v/v) methanol, pH 8.3 at 20°C]. Membranes were dried overnight on tissue paper.
Hsp70/hsp90 western protocol
Nitrocellulose membranes were blocked with blocking buffer [25 mmol l–1 Tris-Cl, pH 7.5 at 20°C, 150 mmol l–1 NaCl, 0.1 % (v/v) Tween, 0.02 % (w/v) Thimerosol, 5 % (w/v) non-fat dried milk] for 1 h, subsequently washed twice for 5 min with Tris-buffered saline (TBS; 25 mmol l–1 Tris-Cl, pH 7.5 at 20°C, 150 mmol l–1 NaCl) and then incubated with a solution of a monoclonal rat antibody (IgG) against hsp70 (clone 7.10; Affinity BioReagent, MA3-001; 1:2500 dilution of hsp70 antibody in buffer A) [BA: TBS, 2.5 % (w/v) bovine serum albumin, 0.02 % (w/v) Thimerosol] for 1 h or, in the case of hsp90, with a solution of a monoclonal rat antibody (IgG2a) against hsp90 (StressGen, SPA-835; 1:500 dilution of hsp90 antibody in BA) for 90 min. This was followed by a 5 min wash with TBS, two 5 min washes with TBS containing 0.1 % (v/v) Tween and a final 5 min wash with TBS. Subsequently, the membrane was incubated for 30 min with a rabbit anti-rat bridging antibody (IgG) solution (1:2000 dilution in BA; Vector, AI-4000) followed by four 5 min washes with TBS, TBS containing 0.1 % Tween (twice) and TBS again. Finally, we incubated membranes with a horseradish-peroxidase Protein A solution (1:5000 dilution in BA; BioRad) for 30 min. The incubation was followed by a 5 min wash with TBS, three 10 min washes with TBS containing 0.1 % Tween and a final 5 min wash with TBS. Membranes were overlaid with a solution of enhanced chemiluminescent (ECL) reagent (Amersham) according to the manufacturer’s instructions for 1 min. Under dark-room conditions, we exposed membranes onto pre-flashed Hyperfilm (Amersham) for 5, 10 (hsp70) and 45 s (hsp90 and HSF1) 10, 30 and 50 min after ECL treatment to obtain various exposures within the linear range of detection. All samples were run at least twice. Results for hsp70 were also confirmed with an iodinated tertiary antibody (125I-protein A, data not shown).
HSF1 western protocol
Nitrocellulose membranes were blocked with blocking buffer (see above) for 1 h and subsequently incubated with a polyclonal rabbit HSF1 antibody prepared against recombinant human HSF1 (Affinity BioReagents, PA3-017; 1:500 dilution in BA) for 1 h. Membranes were then washed for 3 min with TBS, TBS containing 0.1 % Tween and TBS and subsequently incubated with a goat anti-rabbit IgG antibody conjugated to horseradish peroxidase (StressGen, SAB-300) for 30 min. After repeating the three washing steps, we detected bands by treating the blot with ECL according to the manufacturer’s instructions. Results obtained with this human HSF1 antibody were confirmed using another antibody also prepared against human HSF1 in rabbit (StressGen, SPA-901).
Image analysis, quantification of expression of heat-shock proteins and statistical analyses
Film images were scanned on a densitometer (Sharp JX-330), and the digitized images were analyzed with image-analysis software (ImageMaster 1D, Version 2.01, Pharmacia) to quantify the band intensities of two hsp70 isoforms, one with a molecular mass of approximately 72 kDa (hsp72), the other of approximately 74 kDa (hsp74), hsp90 and HSF1. We express band intensities relative to a known amount of a bovine heat-shock cognate 70 (80 ng; StressGen, SPP-750), a bovine hsp90 (120 ng; StressGen, SPP 780) and a human HSF1 standard of 80 kDa (20 ng; StressGen, SPA-900WB) to account for variation among western blots (Fig. 2).
Differences in endogenous levels of hsp70 isoforms (see Figs 3, 4), their levels relative to each other (see Fig. 5) and their total sum (see Fig. 6), hsp90 (see Fig. 7) and HSF1 (see Fig. 8) were tested among species within treatments (four comparisons of T. funebralis relative to T. brunnea and T. montereyi) and among treatments within species (three comparisons of the 23°C treatment relative to all other treatments). The resulting seven comparisons were of particular interest to us because of the differences in induction temperatures observed previously (Tomanek and Somero, 1999). We used the least significant difference (LSD)-test for a post-hoc pairwise comparison (SYSTAT software, Systat, Inc.) after a two-way analysis of variance (ANOVA). For the two-way ANOVA, species and treatment were used as independent categorical variables, and endogenous levels of hsps and HSF1 were used as the dependent variable. LSD was used because we were specifically interested in seven planned comparisons. We accounted for the fact that these comparisons are not independent by multiplying the P-value we obtained by the number of comparisons made (in this case seven). Differences are described as statistically significant when the resulting P-level was ⩽0.05.
Endogenous levels of hsp70 isoforms and hsp90
The hsp70 antibody we used for western analysis detected one low-molecular-mass (hsp72) and one high-molecular-mass (hsp74) isoform (Fig. 2). Endogenous levels of hsp72 were higher in T. funebralis than in T. brunnea and T. montereyi in field-acclimatized specimens and following acclimation to 13 and 18°C (Fig. 3). No interspecific differences in levels of hsp72 were found at 23°C. In T. funebralis, no significant differences in levels of hsp72 were observed among the four treatments. In contrast, hsp72 levels increased with increasing acclimation temperature (23°C in comparison with 18 and 13°C) in both subtidal Tegula congeners.
Endogenous levels of hsp74 were invariably higher in T. brunnea and T. montereyi than in T. funebralis, although these differences were not always significant (Fig. 4). Levels of hsp74 increased following acclimation to 23°C, in comparison with 18°C, in both subtidal congeners but, as for hsp72, T. funebralis showed no significant differences among the four treatments.
To determine whether the three congeners differed in the relative levels of the weakly (hsp74) and strongly (hsp72) heat-induced hsp70 isoforms, we calculated the ratio of hsp72 to hsp74 levels in the four treatment groups (Fig. 5). T. funebralis had, under all four treatments, an approximately twofold higher ratio of hsp72 to hsp74 relative to the two subtidal congeners. Within a species, the ratio of hsp72 to hsp74 did not vary among treatment groups.
Hsp70 isoforms may act together in repressing the induction of hsp synthesis, i.e. the inhibition of HSF1 may be the summed effect of all hsp70 isoforms in the cell. We therefore calculated the total levels of hsp72 plus hsp74 in the experimental groups (Fig. 6). The total levels of the two isoforms were similar in all three congeners for any treatment group. Acclimation had no effect on the total amount of hsp70 isoforms in T. funebralis, but acclimation to 23°C led to significantly higher total hsp70 levels in the two subtidal species.
Endogenous levels of HSF1
To determine whether the inter- and intraspecific variation in Ton observed in our previous study was related to different endogenous levels of HSF1, we used an antibody against human HSF1 to quantify HSF1 levels in differently acclimated congeners of Tegula (Figs 2, 8). Note that human HSF1 (80 kDa) has a lower molecular mass than Tegula HSF1 (approximately 100 kDa) (Fig. 2). This is not unusual considering how widely the molecular mass of HSF1 varies among species (yeast, 150 kDa; Drosophila melanogaster, 110 kDa; human, 80 kDa) (Wu, 1995).
Endogenous levels of HSF1 were significantly higher in T. funebralis than in T. brunnea and T. montereyi in field-acclimatized and laboratory-acclimated specimens (Fig. 8). Within a species, HSF1 levels changed little with increasing acclimation temperature in most cases. However, levels of HSF1 were always lowest in the field-acclimatized group (P<0.05 for all post-hoc comparisons between the field-acclimatized and all other acclimation groups, except for 23°C acclimation in the two subtidal congeners; statistical results not shown). Thus, laboratory acclimation had an effect on endogenous levels of HSF1 even at a temperature (13°C) that represents a typical sea surface temperature at the collection site.
Regulation of interspecific variation in the heat-shock responses of Tegula congeners
Endogenous levels of hsps and HSF1 were measured to elucidate the mechanisms underlying the inter- and intraspecific variation in hsp gene expression discovered in our earlier studies of hsp synthesis in congeners of Tegula (Tomanek and Somero, 1999, 2000). Specifically, we tested predictions based on the cellular thermometer hypothesis, which emphasizes the central regulatory role played by a multi-protein complex comprising HSF1 and hsps in regulating hsp gene transcription (see Introduction). Following acclimation to 13 and 18°C under constant submersion, we found that the heat-tolerant, low- to mid-intertidal species T. funebralis induced de novo synthesis of hsp70, the major hsp in Tegula, at 27°C, whereas the heat-sensitive subtidal to low-intertidal T. brunnea and T. montereyi induced hsp synthesis at 24°C (Tomanek and Somero, 1999). All three species induced hsp synthesis at 27°C after acclimation to 23°C, a temperature close to the upper end of the sea surface temperatures found within the distribution range of the two subtidal congeners (Tomanek and Somero, 1999). Our prediction, based on the cellular thermometer model, was that higher endogenous levels of two repressors of the heat-shock response, hsp70 and hsp90, would be found in T. funebralis following acclimation to 13 and 18°C, but not to 23°C. This initial hypothesis assumed that the levels of HSF1 would be the same for all three species.
Levels of hsp72, but not of hsp74 or hsp90, agreed with this prediction (Figs 3, 4, 7). Thus, endogenous levels of hsp72 appear to provide a potential basis for the interspecific differences in Ton found in specimens acclimated to 13 and 18°C. However, two other observations call into question the role of hsp70 isoforms in accounting for these interspecific differences in Ton. First, consistently higher levels of HSF1 were present in T. funebralis than in T. brunnea and T. montereyi (Fig. 8). Second, total levels of hsp72 plus hsp74 did not differ among species in the specimens acclimated to 13 and 18°C. According to the cellular thermometer model, the higher levels of HSF1 in T. funebralis, vis à vis the equal levels of total hsp70 proteins in the three species, would favor a lower rather than a higher Ton in this species. Thus, interspecific differences in the Ton of hsp synthesis cannot be explained by a cellular thermometer model, which has as its central mechanism the inhibition of HSF1 activity by hsp70 and hsp90.
The higher levels of HSF1 present in T. funebralis may, however, account for another difference observed between this species and T. brunnea: a more rapid induction of hsp synthesis by T. funebralis following severe heat stress (Tomanek and Somero, 2000). Higher endogenous levels of HSF1 could facilitate a faster conversion of inactive HSF1 monomers to active trimers and increased binding of HSF1 to the HSE following heat-shock. When HSF1 was overexpressed in Escherichia coli, the increased levels of HSF1 led to constitutive binding of HSF1 trimers to the HSE (Clos et al., 1990; Rabindran et al., 1991; Sarge et al., 1993). Whether warm-adapted species such as T. funebralis generally contain higher levels of HSF1 than more cold-adapted species remains to be determined. In only a few cases have comparisons of this type been made. In fact, Zatsepina et al. (2000) reported that two heat-tolerant desert lizard species had lower levels of HSF1 than a distantly related more heat-sensitive species. On the basis of current understanding of the mechanisms that regulate hsp gene expression, it appears likely that the total concentration of HSF1 in the cell is an important, but only partial, determinant of Ton and the rate of hsp gene expression. As suggested by the model in Fig. 1, conformational changes and the phosphorylation state of HSF1 also contribute to regulation of hsp genes (Shi et al., 1998; Zuo et al., 1995). In a recent study of hsp gene expression in Mytilus spp., Buckley et al. (2001) showed that binding of HSF1 to the HSE occurred at temperatures several degrees below Ton. This observation suggests that additional steps that occur subsequent to the binding of HSF1 to the HSE may be of importance in transcriptional regulation of hsp genes. Nonetheless, higher levels of HSF1 present in T. funebralis are likely to poise the hsp70 system towards rapid expression.
Regulation of acclimation-induced intraspecific variation in the heat-shock responses of Tegula congeners
Acclimation to a temperature (23°C) close to the habitat upper thermal extreme for the subtidal species induced a shift in Ton in T. brunnea and T. montereyi (from 24 to 27°C), but not in T. funebralis (27°C) (Tomanek and Somero, 1999). Corresponding increases in hsp72, hsp74, the sum of both hsp70 isoforms and hsp90 (Figs 3, 4, 6, 7) levels in both subtidal species, but not in T. funebralis, following acclimation to 23°C suggest that the 3°C increase in Ton could be due to elevated levels of HSF1-binding hsps. This conclusion is supported by the observation that HSF1 levels were the same under all acclimation conditions within a species (Fig. 8). Thus, it seems that the cellular thermometer model can explain acclimatory changes in Ton. Similar intraspecific changes in induction temperatures and endogenous levels of hsp90 or hsp70 have been observed for summer- and winter-acclimatized gobies of the genus Gillichthys (Dietz and Somero, 1992) and for intertidal mussels of the genus Mytilus (Buckley et al., 2001; Roberts et al., 1997), respectively.
Heat stress and growth in Tegula congeners: using hsp70 isoforms as indicators of physiological state
Multiple isoforms of hsp70 are present in most species, and these isoforms generally differ in the extent to which they are induced by heat stress (Parsell and Lindquist, 1993, 1994). It is common for a constitutively expressed hsp70 cognate (hsc70) to be present in cells under all conditions, where it functions as a chaperone during normal protein synthesis. In contrast, some hsp70 isoforms may be absent or present only at low levels until the cell encounters heat stress. This type of variation in hsp70 expression has been observed in marine mussels of the genus Mytilus.
In the bay mussel Mytilus trossulus, levels of a high-molecular-mass isoform of hsp70 (hsp76) exhibited little variation with exposure temperature, whereas levels of a small hsp70 (hsp68) varied several-fold as a function of habitat temperature conditions (Hofmann and Somero, 1995). In some specimens from a cold-acclimatized population, only the larger isoform could be detected, indicating that the low-molecular-mass isoform is not constitutively expressed. Similar observations were made in the ribbed mussel Mytilus californianus: the low-molecular-mass hsp70 isoform was more strongly induced by high temperatures than the larger isoform (Roberts et al., 1997), and subtidal specimens acclimatized to low temperatures sometimes lacked the smaller isoform altogether (M. Lopuch and G. N. Somero, unpublished observations).
In the three species of Tegula, two hsp70 isoforms, hsp72 and hsp74, were detected in all specimens examined, but hsp72 exhibited stronger induction during heat stress (Figs 2–4). As in the case of the congeners of Mytilus, the larger of the two isoforms may be a constitutively expressed chaperone that is present at all times to assist in normal protein synthesis. Hsp72 may be a stress-induced isoform whose synthesis is strongly upregulated during heat stress. These differences in heat-inducibility between hsp70 isoforms in Mytilus and Tegula suggest that attempts to use endogenous levels of hsp70 as an index of environmental thermal stress may require separate analyses to be made of each isoform. Although hsp74 levels may be elevated under extreme heat stress (e.g. at 23°C in the subtidal species, Fig. 4), its endogenous levels under non-stressful conditions may reflect the capacity of the tissue for normal protein synthesis. In contrast, levels of the strongly heat-inducible low-molecular-mass isoforms may be a valid index of heat stress (Hofmann and Somero, 1995; Roberts et al., 1997) (present study). Total levels of low- and high-molecular-mass hsp70 isoforms may obscure important relationships related to physiological state, including propensity for growth.
The interspecific differences observed in the field-acclimatized snails, as well as in the 13 and 18°C acclimation groups support the latter possibility (Figs 3–5). In T. funebralis, levels of hsp72 are generally higher, and those of hsp74 lower, than in the subtidal species. However, the total amounts of hsp72 plus hsp74 do not differ among species. On the basis of the proposed different roles of the two hsp70 isoforms, it appears that a different balance between normal protein synthetic activity (indexed by levels of hsp74) and stress-induced repair of unfolded proteins (indexed by levels of hsp72) is present in the three species. In T. funebralis, which experiences greater heat stress than the two subtidal species (Tomanek and Somero, 1999), there may be a greater expenditure of energy for repair of heat damage to proteins, resulting in a lesser channeling of energy into normal protein synthesis.
This hypothesized interspecific difference in allocation of energy between repair and protein synthesis, based on the ratio of levels of hsp72 to hsp74, is consistent with observed differences among the congeners in growth rates: T. brunnea and T. montereyi grow approximately 3–6 times faster than T. funebralis under field conditions (Frank, 1965; Paine, 1969; Watanabe, 1982). A recent study on the intertidal predatory whelk Nucella ostrina also illustrated how heat stress and concomitant higher levels of hsp70 can affect metabolic activity and possibly growth rates (Dahlhoff et al., 2001).
In conclusion, using levels of molecular chaperones to gain insights into the physiological state of specimens requires that the distinct roles of these proteins in normal protein synthesis and in repair of heat-induced damage to proteins be taken into consideration. Total levels of all hsp70 isoforms may fail to provide a valid estimate of either protein synthetic capacity (growth potential) or level of heat stress.
These studies were supported by National Science Foundation grant IBN97-27721 and by the David and Lucile Packard Foundation through the PISCO program.