AMP-activated protein kinase (AMPK) in the rock crab, Cancer irroratus: an early indicator of temperature stress

Marine crustaceans are exposed to frequent seasonal and diel temperature changes. These temperature changes have a profound impact on ectothermal animals' energy metabolism and scope for exercise. Additionally, the potential impacts of global climate change on marine ecosystems emphasize the need to understand the effects of rising sea temperatures on marine invertebrates (for a review, see Osovitz and Hofmann,2007). To predict how species will be affected by these temperature changes, we need both an indicator of heat stress and an understanding of the underlying mechanisms. Many physiological studies investigating temperature stress identify heat stress through the expression of heat shock proteins (HSPs) (e.g. Feder and Hofmann, 1999; Hoffmann et al., 2003; Tomanek,2005). HSPs are ubiquitously expressed proteins that act as chaperones and aid in the conformational stabilization of other proteins(Feige et al., 1996; Frydman, 2001). Though HSPs were identified and named for their activation during temperature stress, many other stressors, including osmotic shock, hypoxia and exercise, induce HSP expression (for a review, see Hochachka and Somero, 2002). Named after their molecular weight, the HSP70 family includes constitutive (or cognate) HSP73, and stress-inducible HSP72(Reading et al., 1989; Brown et al., 1993). Tomanek and Somero have shown that a heat shock response in snails is already detectable 1–2 h after heat stress and that the temperature at which HSP70 expression occurs can change after acclimation(Tomanek and Somero, 1999). Iwama acknowledges that HSPs are good indicators of general cellular stress,but points out that due to the variety of possible triggers of HSP expression,specific conclusions about HSP up- or down-regulation may not be valid(Iwama et al., 2004). The time lag between the actual heat stress and the detectable accumulation of HSPs raises the question of whether an additional, faster mechanism exists that would aid in withstanding thermal challenges.

In mammals, AMP-activated protein kinase (AMPK) has been described as a`low fuel sensor' (Hardie and Carling,1997) or a `metabolic master switch'(Winder and Hardie, 1999),which reflects AMPK's central role in cellular energy metabolism. AMPK is thought to constantly monitor the energy status of the cell and, if needed,regulate the anabolic and catabolic processes to ensure a constant ATP concentration (Hardie et al.,2006). This regulation is achieved by AMPK through phosphorylation and subsequent activation or inhibition of rate-limiting enzymes of all the major energy metabolism pathways. By decelerating ATP-consuming pathways such as glycogen synthesis, fatty acid synthesis, protein synthesis and cholesterol synthesis, and in parallel accelerating ATP-producing pathways such as glucose uptake, glycolysis and fatty acid β-oxidation, the activation of AMPK prevents ATP depletion and promotes replenishment of the ATP pool(Fig. 1) (for a review, see Hardie et al., 2006; Hardie and Sakamoto,2006).

The AMPK protein is a heterotrimer with a catalytic α-subunit and regulatory β- and γ-subunits. AMPK derives its main activation by phosphorylation of the α-subunit at threonine 172 (T172). AMP plays a major role in activating AMPK by four effects in parallel: (1) allosteric activation of AMPK kinase (AMPKK); (2) binding of AMP to AMPK, rendering it a poorer substrate for protein phosphatases; (3) binding of AMP to AMPK, making it a better substrate for the upstream kinase, AMPKK; and (4) allosteric activation of AMPK (Hardie et al.,1999). AMP is a good indicator of cellular stress because an increased ATP hydrolysis rate leads to a rapid accumulation of AMP in the cell. ATP hydrolysis first increases the cellular ADP concentration. The ADP is then converted by the adenylate kinase reaction (2 ADP→ATP+AMP) to ATP and AMP (Hardie et al., 2003). Therefore, during increased ATP use, AMP accumulates well before any changes in cellular ATP or ADP concentration occur. This is especially true for muscle tissues with the creatine phosphate or arginine phosphate system. These high energy phosphagens rapidly provide more ATP, so that the cellular ATP concentration remains constant despite a high ATP hydrolysis rate(Bessman, 1985). In a previous study on mice, we described AMPK activation under conditions of a raised AMP but constant ATP concentration (Frederich and Balschi, 2002).

Hypoxia, exercise and osmotic shock are known to activate AMPK in mammals through AMPKK activation and AMP accumulation. Cold stress has been shown to affect AMPK activity in frogs (Bartrons et al., 2004) and in the brown adipose tissue of mice(Mulligan et al., 2007); for a recent review on AMPK-activating factors see Hardie et al.(Hardie et al., 2006). However, AMPK activity and its regulation during heat stress, especially in invertebrates, have not yet been investigated thoroughly. Furthermore, most AMPK studies focus on vertebrates, especially mammals. Only a few studies have investigated AMPK in invertebrates in species such as the brine shrimp, Artemia franciscana (Zhu et al.,2007), and the fruit fly, Drosophila melanogaster(Lee et al., 2007).

AMPK is remarkably highly conserved during evolution with high sequence similarity between humans and other mammals [rat, mouse, rabbit, pig(Hardie et al., 1998)]. AMPK has also been described in the fruit fly, D. melanogaster(Pan and Hardie, 2002), and the nematode worm, C. elegans(Gao et al., 1995), as well as in plants such as cauliflower and tobacco(Kelner et al., 2004). We recently identified AMPK in the rock crab, Cancer irroratus, and demonstrated tissue-specific AMPK activation during hypoxia(Pinz et al., 2005). It is therefore likely that the AMPK cascade is a central mechanism for regulating energy metabolism found in most, if not all eukaryotes.

Marine invertebrates switch to anaerobiosis during heat stress (see Discussion). This anaerobic metabolism is characterized by a limited ATP yield, and an accumulation of AMP is expected due to the concomitant increase in metabolic rate. Because of the ubiquity of the AMPK cascade, we predicted that AMPK is activated during temperature stress. We tested this hypothesis in a marine decapod crustacean and compared AMPK activation with a more established marker for heat stress, the heat shock protein 70 (HSP70).

Male rock crabs, Cancer irroratus (Say 1817), with an average carapace width of 101.9±9.7 mm (mean±s.d.) were obtained from a local lobster fisherman in Saco, ME, USA, kept in a flow-through seawater system in the Marine Science Center of the University of New England at 12–15°C, and fed squid and fish ad libitum.

Animals (five per experiment) were placed in a darkened 100 l tank at 12°C overnight. The next day animals were exposed to a fast progressive temperature increase (6°C h^–1) and killed at 12, 16, 18,20, 22, 24, 26 or 28°C (30°C was also used in some trials). Animals were killed at the respective temperature with a cut through the cerebral ganglion and the heart was removed and either stored in RNAlater® solution(Ambion, Austin, TX, USA) or flash-frozen with Wollenberger tongs pre-cooled in liquid nitrogen. The flash frozen samples were stored at –80°C until analysis of AMPK activity and HSP70 protein levels.

In a second set of experiments, animals were exposed to the same progressive temperature increase up to the sublethal temperature of 26°C and temperature was then kept constant at 26°C. Animals (five per time point) were killed and tissue harvested as described above at 0, 1, 2, 4 and 6 h after reaching 26°C.

To investigate the ability of the animals to respond to experimental stimulation at different temperatures, animals were subjected to the fast progressive temperature increase described above. At 12, 16, 18, 20, 22, 24,26, 28 and 30°C the animals were turned upside down and placed on a flat surface underwater. The time (reaction time) to return to the upright position was monitored. Animals were counted as `not responding' if they did not turn within 15 min.

To monitor the animals' heart rate during the temperature incubations,photoplethysmographs (iSiTEC, Bremerhaven, Germany) connected to a digital recording device (PowerLab, Mountain View, CA, USA) were glued to the carapace above the heart as described in detail by Depledge(Depledge, 1984) or Frederich and Pörtner (Frederich and Pörtner, 2000).

The lactate concentration in the heart tissue was measured using a photometric test according to Bergmeyer(Bergmeyer, 1985) to characterize the onset of anaerobiosis. Briefly, tissue was ground under liquid nitrogen and the frozen tissue powder transferred to 1.2 mol l^–1 perchloric acid to precipitate the protein. After neutralization with 1 mol l^–1 K₂HPO₄,the sample was centrifuged, and the lactate concentration was measured in the supernatant at 340 nm as NADH accumulation by the lactate dehydrogenase reaction. The lactate concentrations were normalized to the protein concentration in the extract and are presented as nanomoles per gram of protein (Bradford, 1976).

Total HSP70 protein, inducible HSP70 protein, total AMPK protein and AMPK phosphorylation (i.e. AMPK activity) were quantified by western blots. Tissue samples were ground under liquid nitrogen and homogenized in a buffer containing phosphatase inhibitors to prevent dephosphorylation of AMPK (in mmol l^–1: Tris-HCl 20, NaCl 50, NaF 50, NaPPi 30, sucrose 250, ZnCl₂ 10, sodium vanadate 100, DTT 2, PMSF 50, pepstatin A 5 and leupeptin 5, with 0.4 mg ml^–1 digitonin; pH 7.4 at 4°C). Proteins from the homogenate (50 μg) were separated on a 10%polyacrylamide/SDS gel at 180 V for 30 min. The proteins were then transferred to a nitrocellulose membrane at 70 V for 2 h and the membrane blocked with 3%non-fat dry milk and 0.1% gelatin. Primary rabbit anti-AMPK (total AMPK),anti-pT172 antibodies (Upstate, Lake Placid, NY, USA), mouse anti-HSP70(Sigma, St Louis, MO, USA), anti-inducible HSP70 (Biovision, Mountain View,CA, USA) and secondary goat anti-rabbit or goat anti-mouse antibodies (Biorad,Hercules, CA, USA), respectively, were used to visualize the proteins. Mouse anti-actin antibodies (Sigma) were used as a loading control. The membranes were developed with a colorimetric kit (Biorad, Opti-4CN system). Scanned bands were quantified using Image-J software (NIH, Bethesda, MD, USA).

The AMPK activity, quantified as T172 phosphorylation, could also be interpreted as AMPKK activity rather than AMPK activity itself. Because AMPK has a multitude of downstream targets, the activity of each target (e.g. acetyl-CoA carboxylase, HMG CoA reductase, phosphofructokinase 2, protein synthase, and many others) could be used to measure AMPK activity. Each of these downstream targets, however, is regulated in its respective activity by other regulators as well and, therefore, does not reflect AMPK activity alone. Many studies, including one of ours(Frederich and Balschi, 2002),have shown that an increase in phosphorylated T172 is representative of the respective cellular effects of activated AMPK. These days, it is common to quantify AMPK activity as T172 phosphorylation.

We measured the heat shock response as both the total HSP70 and the inducible HSP70 to ensure that we did not miss any small upregulation of inducible HSP that might otherwise be masked by a constitutively expressed HSP in the cell.

To design degenerate primers, we searched GenBank for AMPKα and AMPKγ protein sequences from various invertebrate and vertebrate species. Obtained sequences were aligned using the MultAlin tool(http://bioinfo.genotoul.fr/multalin/multalin.html). Degenerate forward and reverse primers for PCR (see Table 1) were designed based on highly conserved areas in the aligned sequences. For HSP70, we used a primer pair that Voznesensky and colleagues(Voznesensky et al., 2004)prepared for the copepod, Calanus finmarchicus.

Total RNA from C. irroratus hepatopancreas was purified using the Total RNA Isolation System (Promega, Madison, WI, USA) and reverse transcribed(Super-Script First Strand Synthesis System, Invitrogen, Carlsbad, CA, USA). cDNA was amplified with the respective primer pairs via PCR with an annealing temperature of 45°C. DNA was sequenced at the Mount Desert Island Biological Laboratory (MDIBL, Salisbury Cove, ME, USA) sequencing core facility on an ABI 3100 sequencer. The DNA sequences obtained were converted to a predicted amino acid sequence using the NCBI open reading frame finder(http://www.ncbi.nlm.nih.gov),and a BLAST search confirmed the cDNA as AMPK or HSP70, respectively. These methods are described in detail by Towle and colleagues(Towle et al., 2001).

Specific primers for AMPK and HSP70 (see Table 1) were designed with the idtDNA primer design tool(www.idtDNA.com)based on the sequences obtained with the respective degenerate primers. Specific primers for 18S were designed based on Spears et al.(Spears et al., 1992) (GenBank accession no. M91050). Expression of AMPK and HSP70 mRNA was quantified in duplicate by quantitative real-time PCR using the Stratagene Brilliant SYBR Green qPCR Kit (Stratagene, La Jolla, CA, USA) on a Stratagene MX3005s instrument. After 40 cycles with an annealing temperature of 55°C, a melting curve analysis confirmed that only one DNA product was amplified. The 18S gene was used as a reference gene and one sample (undiluted and diluted 1:10, 1:100, 1:1000) with high HSP70 mRNA expression served as an internal standard. AMPK and HSP70 mRNA levels are shown as relative increase above the AMPK or HSP70 mRNA level, respectively, for heart tissues at the control temperature of 12°C.

Data were tested for significant difference by ANOVA and a Tukey post-hoc analysis or repeated measures ANOVA, dependent on the data set (GraphPad InStat; www.graphpad.com). P<0.05 was considered significant. Data are shown as means± s.e.m.

The partial amino acid sequence for the AMPKα subunit for C. irroratus (GenBank submission no. FJ496868) confirmed the high conservation that has been reported for several other species(Hardie et al., 1998). More than 60% of the obtained C. irroratus sequence (453 amino acids) is conserved in comparison to other arthropods and vertebrates(Fig. 2). We could not determine whether the C. irroratus sequence more closely resembles the mammalian AMPKα1 or α2 subunit based on our partial sequence. However, more importantly, the region flanking the regulatory threonin 172(T172) site where the AMPKα subunit is phosphorylated by an upstream AMPKK is highly conserved. The region flanking the T172 position for C. irroratus contains the amino acid sequence VDGEFL - RpTSCGSPNY,compared with rat SDGEFLRpTSCGSPNY. The antibodies used to quantify AMPK phosphorylation in C. irroratus (see below) were raised against the rat specific antigen KDGEFLRpTSCGSPNY. Except for the very first amino acid, rat and crab sequences are identical. The antigen used also varies in this first position. Exactly the same Cancer irroratus sequence of 15 amino acids flanking the T172 position was identified for Carcinus maenas, Homarus americanus and Calanus finmarchicus (data not shown). With this high sequence conservation of the peptide, the use of heterologous antibodies is not problematic. Similarly high conservation was observed in the 180 amino acid sequence of the AMPKγ subunit (GenBank submission no. FJ496867, 54% similar to mouse, NM_153745). As expected, the obtained HSP70 sequence of 178 amino acids (GenBank submission no. FJ496866)is highly conserved as well and shows more than 80% similarity with the mouse HSP70 sequence (AAC84170).

Between 12 and 18°C, the crabs needed 4.1±0.9 s to right themselves after being turned upside down(Fig. 3A). The crabs became slower at 20, 22 and 24°C (28.5, 58.9 and 46.2 s, respectively). With a high variability (between 1 and 280 s) this increase was not statistically significant. At 26°C, the average reaction time decreased to 20.3±9.8 s. However, 20% of the crabs did not return to the upright position (Fig. 3B) and are not included in this time average. At 28°C, 80% were not responding at all,and the remaining animals needed 157.7±20.6 s to react(P<0.05, repeated measures ANOVA). None of the animals were able to right themselves at 30°C.

During the progressive temperature increase, the heart rate increased,following a Q₁₀ of 2.2 between 12 and 26°C, reaching a maximum of 153±27 beats min^–1 at 26°C. Heart rate remained constant between 26 and 30°C, and then dropped at temperatures above 30°C (Fig. 3C).

Lactate concentration in the heart remained constant at 7.3±2.6 nmol g^–1 protein between 12 and 26°C, but increased significantly (P<0.05, ANOVA) 2.2-fold to 16.1±5.6 nmol g^–1 protein above 26°C(Fig. 3D). The concordance between the lack of scope for exercise, the maximum heart rate (see above) and an increase in lactate accumulation indicates that the animals reached their critical temperature (T_c, see Discussion) between 26 and 28°C, as indicated by the dashed line in Fig. 3. However, a reduction in the scope for exercise is already evident between 18 and 20°C, indicated by the dotted line in Fig. 3(T_p).

AMPK activity (western blot for T172 phosphorylation) did not differ between 12 and 18°C. AMPK activity started to increase above 18°C and reached a maximum at 28°C [9.9(±2.3)-fold, P<0.05,ANOVA; Fig. 4A]. The high variability of the data is consistent with the high variability of the scope for exercise of the individual animals (as shown by the reaction time data in Fig. 3A,B). To test for a discontinuity in the data, a Q-BASIC program to identify critical points(Yeager and Ultsch, 1989) was used. We identified two significantly different linear regressions(y=1.21–5.26E–03x, R²=0.8659, P<0.05; y=–12.76+0.75x,R²=0.8067, P<0.05). The two regressions intersect at 18.5°C. The increase in AMPK activity above 18°C coincides with the decrease in reaction time (Fig. 3A). AMPK protein (Fig. 4B) and AMPKα mRNA levels(Fig. 4C) showed the same small decreases and increases, which remained statistically insignificant(P>0.05, ANOVA) during the progressive temperature increase.

The increase of AMPK activity above 18°C occurred before any significant changes in HSP70 protein or mRNA levels were detected(Fig. 4D–F). A non-significant upward trend of inducible HSP70 protein and HSP70 mRNA can be seen at 28°C. This might indicate the onset of the heat shock response,which would coincide with the inability of most animals to respond after being turned on their backs (Fig. 3B)and the onset of anaerobic metabolism (Fig. 3D).

The fast, progressive temperature increase described above elicits a quick and immediate cellular response to thermal stress, as shown by the rapid phosphorylation of AMPK. To test whether AMPK and HSP70 are affected differentially during prolonged thermal stress, we exposed the animals for various periods of time to the sublethal temperature of 26°C. Exposure to 26°C for up to 6 h led to a constantly high heart rate above 150 beats min^–1 (Fig. 5A). The lactate concentration in the heart peaked after 4 h at 23.4±8.5 nmol g^–1 protein and remained above control levels (15.3±0.8 nmol g^–1 protein) after 6 h(Fig. 5B). AMPK activity remained constant at the high level that was reached at 26°C(Fig. 5C). AMPKα mRNA levels increased continuously throughout exposure to 26°C, reaching 5.6±2.2 times more after 6 h at 26°C than in the 12°C controls,but reached statistical significance after 4 and 6 h, only at the P<0.1 level, ANOVA (Fig. 5D). HSP70 protein levels (total and inducible) remained constantly low with no significant changes(Fig. 5E,F). HSP70 mRNA levels rose constantly throughout the 6 h, up to 6.8(±1.7)-fold above control(Fig. 5G). Thus, after this prolonged exposure to high temperatures, both AMPKα mRNA and HSP70 mRNA increased. The HSP70 protein did not follow the same trend as HSP70 mRNA. Our experimental protocol most likely did not account for the time lag of protein synthesis from mRNA. However, this was not the intention of the experiment and, with the AMPK activity rising before either HSP70 protein or mRNA increased, was not followed further.

AMPK is the subject of active research in the medical community, which focuses on AMPK functions in homoeothermic mammals in response to cellular ischemia and metabolic stress. AMPK plays a role in obesity(Kola et al., 2008), heart failure (Hardie, 2008) and diabetes (Koh et al., 2008). In all of these diseases, the cellular ATP homeostasis is disturbed. Because AMPK is an important regulator of ATP homeostasis, it plays a major role in developing or actually treating the respective disease. However, little is known about AMPK in invertebrates and the effects of temperature on AMPK activity. As expected, we found a high sequence conservation for the AMPK protein in crustaceans. The sequence conservation across diverse phyla indicates an important function in cellular metabolism, and maintenance of cellular ATP concentration is probably the most important form of homeostasis. We have shown that AMPK is expressed in C. irroratus, that temperature stress leads to AMPK activation, and that this activation occurs well before the well-characterized heat shock response with HSP70.

It is likely that a deeper understanding of the physiological processes involved in withstanding temperature stress will allow for predictions of the potential impacts of temperature change on animals. Once stress markers have been identified, they can be used to evaluate the stress level of the respective animal. Because extreme temperatures lead to anaerobiosis in the tissues, temperature tolerance and hypoxia tolerance are related. Therefore,elucidating the cellular mechanisms involved enhances our understanding of how common changes in the environment, such as temperature and oxygen concentration, will affect marine crustaceans' survival and their geographical distribution range. This is especially important in the context of global climate change or increasing hypoxic benthic areas (e.g. Diaz and Rosenberg, 2008).

We chose the rock crab, C. irroratus, as a model species because it can easily be obtained and maintained. It is one of the three major decapod species of economic importance in the Gulf of Maine(Palma et al., 1999), and the crab fishery is currently expanding, especially in Canada(Gendron et al., 2001). A decline in the rock crab population could also have a negative impact on the lobster fisheries in the Gulf of Maine because rock crabs play an important role in the diet of the American lobster, Homarus americanus(Gendron et al., 2001). A multitude of earlier studies have characterized the physiological response of decapod crustaceans to temperature stress. The current study builds on the wealth of existing knowledge and contributes to an enhanced understanding of the cellular and molecular processes affected by temperature stress. Preliminary data from an earlier study showed that hypoxia affects AMPK activity (Pinz et al., 2005),and long-term temperature stress affects AMPKγ mRNA expression differentially in different tissues of C. irroratus(Frederich et al., 2006). For the current, more comprehensive study, we chose to focus on the heart only,because the temperature-induced effects on heart rate are well described and can easily be monitored by a heart rate sensor. We are aware that heart rate is a sub-optimal measure of cardiac workload. However, for the purpose of this study, the temperature-induced increase in heart rate indicates increased performance and, therefore, increased energy demand of this organ.

As outlined in the Introduction, AMPK is phosphorylated and activated through AMPKK by a change in the cellular free AMP concentration. The AMP concentration does not change only in the context of temperature stress. Potentially every stress that affects cellular energy metabolism and ATP hydrolysis rates, such as exercise, hypoxia, salinity stress and many others,will lead to changes in cellular AMP and, consequently, affect AMPK activity. Preliminary data show that salinity stress affects AMPK mRNA expression and AMPK activity in salmon, Salmo salar, as well as AMPK activity in the green crab, Carcinus maenas (M.F. and J.A.J., unpublished observations). Jibb and Richards (Jibb and Richards, 2008) show in their recent study that 0.5 h of hypoxia in goldfish leads to a 5.5-fold increase in AMPK activity in the liver. Consequently, we expect further studies to show that AMPK, as a central regulator of cellular energy metabolism, is involved in many kinds of stress response. Furthermore, we expect a dual mechanism to achieve this regulation. First, a fast and immediate response through AMPK activation by phosphorylation of AMPKα at the T172 site. This provides increased AMPK activity within seconds. Second, a slower but longer lasting response through increased AMPK mRNA and consequently AMPK protein expression. This provides a long-term adjustment to varying energy demand. Both mechanisms are supported by our study. A third possibility is a differential expression of AMPK subunit isoforms. Whether invertebrates express the same set of isoforms as mammals(see Introduction) is currently not clear. We are aware of only one invertebrate study that claims to demonstrate two different AMPKαisoforms, in the brine shrimp, Artemia franciscana(Zhu et al., 2007).

The activation of AMPK through heat stress can be viewed in the broader theoretical framework of Shelford's law of tolerance(Shelford, 1931) and critical temperatures. Shelford describes an optimum and a pessimum range for animals(and plants) for each respective environmental parameter. The ability to survive is obviously at a maximum in the optimum range and gradually decreases towards the upper and lower pessimum range. The following describes how physiological parameters can identify thresholds between those ranges and how AMPK activity can be used in a new way to characterize an additional range in this model.

Exposure of marine invertebrates to extremes of high and low temperature causes a mismatch of oxygen (O₂) demand and O₂ supply in the tissues of the animals, despite sufficient O₂ availability in the environment (Frederich and Pörtner, 2000). The thresholds for these effects have been defined in crustaceans and other invertebrates as the critical temperature(T_c) (Zielinski and Pörtner, 1996; Frederich and Pörtner, 2000; Pörtner, 2002). Critical temperatures are characterized by the onset of anaerobic metabolism and the subsequent accumulation of anaerobic end-products such as lactate, as well as the failing of ventilatory and circulatory activity (e.g. Zielinski and Pörtner,1996; Frederich and Pörtner, 2000; Pörtner, 2002; Peck et al., 2002; Braby and Somero, 2006). Prolonged exposure to temperatures above the upper T_c or below the lower T_c results in anaerobic metabolism that finally leads to death due to energy depletion(Frederich and Pörtner,2000; Pörtner,2002).

We detected an accumulation of lactate between 26 and 28°C in C. irroratus, concomitant with attaining the maximum heart rate. Inducible HSP70 protein, measured by western blot, as well as HSP70 mRNA, measured by quantitative real-time PCR, showed a small but statistically insignificant upward trend at this threshold. AMPK activity, however, was already well elevated at T_c, with the increase commencing between 18 and 20°C. The response time after experimental stimulation increased significantly above 26°C and coincided with lactate accumulation, maximum heart rate, AMPK activity and an upward trend in HSP. The critical temperature is therefore likely to occur between 26 and 28°C.

The threshold for the onset of increased AMPK activity coincides with the temperature at which the response time first slowed down (18°C). This initial slowing of response time was not statistically significant. However,it might be biologically significant for a crab in its environment because the ability to escape from a predator is crucial for survival. With a high variability among individuals, some animals will be affected by heat stress earlier and will be more vulnerable, as reflected in the higher standard error in response time as well as AMPK activity. The onset of increased AMPK activity coincides with the average maximum summer temperature in the area where the animals were caught as well. Records of the Gulf of Maine Ocean Observing System(www.gomoos.org)from 2002 to 2007 for Casco Bay (buoy CO2) at depths of 2 and 20 m show daily average temperatures peaking in July at approximately 19°C and hourly maximum temperatures of 20°C (Fig. 6). Animals used in the experiments were caught between late June and early September and experiments were performed in the fall. The animals therefore had a thermal history of maximum temperatures between 19 and 20°C. This is very close to the observed onset of AMPK activity between 18 and 20°C.

In an earlier study (Frederich and Pörtner, 2000), we identified critical temperatures in the spider crab, Maja squinado, and also an earlier threshold that we called `pejus temperature, T<sub>p</sub>' (pejus; latin for `getting worse'). The upper and lower T_p encompass the range of maximum performance and coincide with the normal habitat temperature. Animals are exposed to temperatures in the pejus range, between T_pand T_c, only occasionally. A recent study by Pörtner and Knust (Pörtner and Knust,2007) shows that eelpout in the North Sea are exposed briefly every summer to temperatures above their pejus temperature, but rarely to temperatures above their critical temperature. However, a reduction in growth rate and relative abundance was clearly correlated to temperatures in the pejus range, between T_p and T_c. Therefore, the pejus temperature represents an important threshold that describes the upper limit of regular function for an animal. Fig. 6 summarizes the optimum,pejus and pessimum range, as well as T_p, T_c and the respective cellular processes described in the present study. Habitat temperature data indicate why the cellular processes at the T_p are more often relevant for the animals' survival than the processes at T_c. Specifically, processes within the pejus range, such as AMPK activity and the subsequent increased ATP synthesis and reduced ATP use in anabolic pathways, help animals to save ATP.

The actual value for T_p in C. irroratus, here shown to be between 18 and 20°C, is likely to change with temperature adaptation. During the winter, when ambient water temperatures are around 3°C, C. irroratus cannot survive much longer than 24 h at 23°C (M.F., personal observation). While the actual temperatures for T_c and T_p shift with seasons and probably vary among populations, depending on specific thermal conditions, the mechanism is likely to remain the same. Animals are fairly tolerant to temperature variations within their range of optimum performance between the upper and the lower T_p. At the upper pejus temperature,coinciding with the average seasonal maximum temperature, AMPK activity increases to maintain an adequately high cellular ATP concentration. When animals are exposed to much higher temperatures, anaerobic metabolism sets in and survival is limited to a very short period of time. When exposed to temperatures between T_p and T_c for longer periods of time (in this study 6 h) the traditional heat shock response starts, as shown by the increase in HSP70 mRNA after several hours at 26°C(Fig. 5G).

In conclusion, to our knowledge this is the first study that specifically investigates the effects of heat stress on AMPK activity in any animal. We have shown that a fast, progressive temperature increase activates AMPK well before the heat shock response can be observed via HSP70. The temperature-related increase in AMPK activity coincides with a decrease in responsiveness and therefore might be part of a mechanism that has significant implications for the survival of the animals in their environment. We suggest that the AMPK cascade represents a cellular mechanism to conserve ATP, which allows the animals to endure short exposures to temperatures above the average maximum temperature of their habitat. AMPK therefore, may be a potential early cellular marker for heat stress in an animal.

LIST OF ABBREVIATIONS

 
• AMPK

  AMP-activated protein kinase

 
• AMPKK

  AMP-activated protein kinase kinase

 
• HSP70

  heat shock protein 70

 
• Tc

  critical temperature

 
• Tp

  pejus temperature