Influence of Hormones, Second Messengers and pH on the Expression of Metabolic Responses to Anoxia in a Marine Whelk

Many marine molluscs are excellent facultative anaerobes and can live for days or weeks in the absence of environmental oxygen (Hochachka, 1980). The biochemical adaptations supporting anoxic tolerance have been extensively studied in several species of marine molluscs (de Zwaan, 1983; Livingstone & de Zwann, 1983; Kreutzer et al. 1985; Storey, 1985, 1988b). Two adaptations are fundamental to anoxic survival: (1) metabolic rate depression, i.e. the ability to lower anoxic metabolic rate to a level only 5–10% of normoxic rate, and (2) an altered energy metabolism involving the coupled catabolism of glycogen and aspartate, the production of alternative end-products (e.g. alanine, succinate, propionate), and additional sites for substrate-level phosphorylation of ADP linked to organic acid synthesis.

Since glycolysis is the central pathway of anaerobic ATP synthesis, facultative anaerobes have elaborated strict regulatory controls over glycolytic rate. In these animals, no Pasteur effect is observed during anoxia, and glycolytic rate is actually depressed compared to that in the aerobic state, as part of the anoxia-induced metabolic rate depression (de Zwaan & Wijsman, 1976; Ebberink & de Zwaan, 1980; Storey, 1985, 1988a; K. B. Storey, J. D. Duncan & D. A. Kelly, in preparation). The molecular mechanisms controlling glycolytic rate depression, first investigated in studies of the marine whelk Busycon canaliculatum (Storey, 1984, 1985; Plaxton & Storey, 1984a,b, 1985, 1986), are proving to be universal among anoxia-tolerant vertebrates and invertebrates (Storey, 1985, 1987, 1988b; Rahman & Storey, 1988; Brooks & Storey, 1988). These controls include: (1) covalent modification of regulatory enzymes (via reversible protein phosphorylation), (2) reversible associations of enzymes into complexes bound to the subcellular particulate fraction, and (3) allosteric regulation, in particular fruc-tose-2,6-bisphosphate (fructose-2,6-P₂) control over phosphofructokinase. The regulation of pyruvate kinase (PK) by enzyme phosphorylation and déphosphorylation has additional importance for marine molluscs. Control of PK regulates carbon flux through the ‘phosphoenolpyruvate branchpoint’, directing the flow of glycogen carbon into alanine versus succinate as end-products (Holwerda et al. 1983; Plaxton & Storey, 1984a).

Although the molecular mechanisms involved in regulating glycolytic rate (and overall metabolic rate) during anoxia are now fairly well understood, the method of signal transmission remains unknown. Clearly, the actions of protein kinases and phosphatases are central to the anaerobic response; these enzymes are typically regulated in the cell by low molecular weight effectors (e.g. cyclic AMP, Ca²⁺), often in response to hormone stimulation (Krebs & Beavo, 1979; Mieskes et al. 1987; Pallen et al. 1987; Connelly et al. 1987; Lincoln, 1983). Furthermore, anaerobiosis is always accompanied in vivo by intracellular acidification (Ellington, 1983a,b;,Walsh et al. 1984; Portner et al. 1984). The present study uses an in vitro system of isolated muscle tissues to investigate the roles of hormones, second messengers and tissue pH in triggering or potentiating the metabolic responses to anoxia in the whelk.

All chemicals were purchased from Sigma Chemical Co. (St Louis, MO) or Boehringer-Mannheim (Montréal, PQ), and were of the highest quality available. Whelks (Busycon canaliculatum) were obtained from the Marine Biological Laboratory at Woods Hole (MA) and were held in aerated artificial sea water (1000mosmoll^-1) at 8–10°C until use.

Kinetic properties of PK were determined in a cuvette containing 1 ml of buffer A (50mmoll^-1 imidazole, pH7·0, containing 50mmoll^-1 KC1 and 5mmoll^-1 MgCl₂), 2 mmol l^-1 Mg-ADP, 0·2 mmol l^-1 NADH and excess dialysed lactate dehydrogenase. For the determination of values (the inhibitor concentration that reduces enzyme velocity by 50%), phosphoenolpyruvate (PEP) concentrations were 0·5 mmol I^-1 for radular retractor muscle and 0·3 mmol l^-1 for foot muscle. Initial velocities were measured at 340nm and 22°C, and the reaction was started by the addition of sample homogenate (see below). Fructose-2,6-P₂ was extracted and assayed by the method of van Schaftingen (1984).

Tissues from aerobic whelks were rapidly dissected out and separated into small pieces; ventricles were divided into six pieces (approximately 0·08 g each), radular retractor muscles were separated into bundles of strands (approximately 0·3 g each), and foot muscle was cut into thin (1–2mm) shoes (approximately 0·5g). Muscle samples were then placed in vials containing 10 ml of sea water previously bubbled for 1 h with either air (O₂, aerobic), a mixture of 95 % N₂ and 5 % CO₂ (anoxic) or 100% N₂ (anoxic). The sea water contained the appropriate concentrations of effector(s), or was adjusted to the appropriate pH value at 22°C. After 4 h of incubation at 22°C (with continuous bubbling), each tissue sample was divided in half. One half was immediately frozen in liquid nitrogen and then transferred to —80°C for long-term storage; this sample was used for the assay of fructose-2,6-P₂ and PK activity. The other half was quickly rinsed in distilled water and homogenized immediately in distilled water (1:2 w/v). The pH of this homogenate was quickly measured to obtain an estimate of the tissue pH value. Although this method does not provide an absolute measurement of the cytosolic pH value, the results obtained using this method are similar to values previously reported using n.m.r. (Ellington, 1983a, 1985; Wiseman & Ellington, 1987) and can be used to monitor relative changes in cellular pH. For the analysis of pyruvate kinase kinetic constants, frozen muscle samples were homogenized 1:4 (v/w) in buffer B (50mmoll^-1 imidazole-HC1, pH7·5, containing 0·1mmoll^-1 phenyl-methylsulphonyl fluoride, 30 mmol l^-1 β-mercaptoethanol, 100 mmol l^-1 NaF, 5 mmol l^-1 EDTA and 5 mmol l^-1 EGTA) for 20 s using a Polytron PT 10 homogenizer. Homogenates were centrifuged at 15 000 g for 10 min at 10 °C, and the supernatant was desalted by centrifugation through a Sephadex G-25 ‘spun column’ (Helmerhorst & Stokes, 1980) previously equifibrated with buffer C [40 mmol l^-1 imidazole, pH 7·0, containing 20% (v/v) glycerol, 10 mmol l^-1 potassium phosphate, 0-lmmolP¹ EDTA and 25mmoll^-1 β-mercaptoethanol].

This procedure rapidly removes all low molecular weight species (M_r < 5000) from the homogenate with greater than 95 % recovery of PK activity; the spun column procedure effectively acts to dialyse PK into buffer C. Fructose-2,6-P₂ samples were prepared by grinding frozen tissue in 50 mmol l^-1 NaOH (at 80°C) at a 1:20 dilution. Fructose-2,6-P2 was then measured according to van Schaftingen (1984).

Tissues were excised from freshly killed aerobic whelks and homogenized 1:2 (w/v) in buffer B plus 25 % glycerol. Low molecular weight species were removed by centrifugation through a Sephadex G-25 spun column pre-equilibrated with buffer C. The eluant was then mixed 1:2 (v/v) with 50 mmol l^-1 imidazole buffer, pH7·0, containing 30 mmol l^-1 NaF, 20% glycerol, 0·1 mmol l^-1 EDTA, 10 mmol l^-1 potassium phosphate and 10 mmol l^-1β-mercaptoethanol. Samples were divided in half and incubated for 16 h at 22 °C in the presence of 10 mmol l^-1 ATP plus 20 mmol I^-1 EDTA (control) or 10 mmol l^-1 ATP plus 20 mmol l^-1 MgCl₂ (experimental). Changes in PK kinetic parameters depended on the presence of exogenously added MgCl₂. Immediately before assaying PK activity, these samples were centrifuged through a spun column to remove NaF, ATP, EDTA and magnesium (Helmerhorst & Stokes, 1980).

Table 1 demonstrates the dependence of PK kinetic parameters on the phosphorylation state of the enzyme. Both the K_m value for PEP (expressed as the ratio V₅/V_0·5) and the I₅₀ value for alanine (expressed as the ratio V_0·5/V_+ala) increase only when the tissue is incubated in the presence of Mg²⁺ and ATP. In these experiments, the cell-free extracts were passed through a Sephadex G-25 spun column to remove low molecular weight metabolites and salts, prior to incubation in the presence of ATP with (+MgCl₂) or without (+EDTA) divalent ion. This procedure ensures that the reaction is dependent on externally added magnesium ions and ATP. Note that the kinetic parameters changed only when magnesium ions were present (protein kinase is active only in the presence of Mg^24-: ATP substrate) and not when EDTA was added to the mixture, despite the inclusion of ATP in both experimental cases. Table 1 shows that phosphorylation of PK resulted in an enzyme with a lower affinity for PEP and a higher sensitivity to alanine (measured at 0·5 mmol l^-1 PEP). This agrees with previous studies on radular retractor muscle PK; anoxia stimulated the incorporation of ³²P into the enzyme and produced similar changes in enzyme kinetic parameters (Plaxton & Storey, 1984a,b). Table 1 also shows that the extent of PK modification was not altered by the addition of either 4 mmol l^-1 cyclic AMP or 2 mmol I^-1 CaCl₂ plus 50μgml^-1 phorbol 12-myristate 13-acetate (PMA). The PK kinetic parameters were, however, measured after a long (16 h) incubation which corresponds to the maximal change observed under these conditions. These experiments demonstrate the importance of phosphorylation in determining the functional PK kinetic parameters, and imply that effectors influencing protein kinases and phosphatases may play a pivotal role in the cellular anoxic response.

The effects of anoxia in vitro on metabolic parameters in whelk foot muscle, radular retractor and ventricle are shown in Table 2 for fructose-2,6-P₂ and Table 3 for the kinetic properties of PK. Anaerobic incubation had no effect on the content of the bisphosphate in foot or ventricle, but fructose-2,6-P₂ in anoxic radular retractor muscle was significantly reduced to 63% of the aerobic value. The effects of anoxia on PK kinetics were consistent in all three tissues. PK from anoxic tissues showed a greatly reduced affinity for PEP as a substrate (K_m increased 6·6-to 17·2-fold over control values) and an increased inhibition by L-alanine (I₅₀ dropped to 17–60 % of control). These results are virtually identical to the metabolic responses of these tissues during anoxia in vivo (Storey, 1988a; Plaxton & Storey, 1984a,b, 1985) and we feel that this validates the use of an isolated muscle preparation as the model system for studying hormone, second messenger and pH control over anaerobic metabolism. The only exception noted was the fructose-2,6-P₂ content of the ventricle after aerobic incubation; this was 10-fold lower than the level in freshly excised tissue from an aerobic whelk (Storey, 1988b).

Experiments were undertaken to determine whether inhibitors of prostaglandin synthesis, hormones or second messengers could mimic or modulate the metabolic responses to anoxia by isolated whelk tissues. Muscle samples were incubated under aerobic and anoxic conditions in the presence of added hormones: 10⁻⁷moll^-1 bovine glucagon, 10⁻⁷moll^-1 bovine insulin, 10⁻⁶moll^-1 epinephrine, 10⁻⁶moll^-1 norepinephrine, 10⁻⁶moll^-1 octopamine and 10⁻⁵moll^-1 serotonin. No effects of hormones were found; neither levels of fructose-2,6-P2 nor kinetic parameters of PK (K_m for PEP, I₅₀ for L-alanine) were altered by the presence of added hormones in either aerobic or anoxic incubations of any of the muscle tissues (data not shown).

The response to anoxia observed in the individual tissues (Table 3) could also result from alterations in the prostaglandin synthesis pathway present in the plasma membranes. This could arise through one of two different mechanisms that combine to alter plasmalemma nucleotide cyclase activity by changes in membrane fluidity. These changes could result from differences in either phospholipase A₂ activity or plasma membrane cyclo-oxygenase activity (through limited O₂ availability), since both enzymes are involved in the breakdown of phospholipids in the membrane, and changes in their activity are often associated with cyclic nucleotide-mediated responses (Saintsing & Dietz, 1983). Addition of either TOjixmoll^-1 dexamethasone (phospholipase A₂ inhibitor) or 0-4mmoll^-1 aspirin (cyclo-oxygenase inhibitor) to tissues incubated under aerobic or anoxic conditions did not result in aerobic enzyme when added to anoxic incubations, nor did it result in anoxic enzyme when added to aerobic incubations.

Equivalent experiments were subsequently carried out to test the effects of second messenger compounds: dibutyryl 3β,5β-cyclic AMP (db-cAMP, to stimulate protein kinase A) or Ca²⁺ + ionophore A23187 + phorbol myristate acetate (C+A+P, to stimulate Ca²⁺/calmodulin protein kinase and/or Ca²⁺/phospho-lipid-dependent protein kinase C). Table 2 (fructose-2,6-P₂) and Table 3 (PK kinetic parameters) show the effects of these additions. Although no consistent pattern of second-messenger effects was observed, some responses were obtained. db-cAMP reduced fructose-2,6-P₂ content in aerobic radular retractor and in ventricle to a level equivalent to that in anoxic tissues. Calcium plus PMA also reduced the fructose-2,6-P₂ level in anoxic ventricle. Addition of C+A+P significantly altered the PK K_m value for PEP in all three tissues under aerobic incubation conditions. C+A+P also altered the I₅₀ value for alanine in foot. In foot and ventricle, the direction of change in the K_m was the same as that found in anoxia, but the magnitude of the change was much smaller (a twofold increase only). Addition of db-cAMP affected only the I₅₀ value for alanine of radular retractor PK; the value was significantly higher than in the anoxic control.

To analyse the role of pH in the development of biochemical responses to anoxia, tissues were incubated under aerobic or anoxic conditions in sea water adjusted to altered pH values. Although the technique used for measurement of tissue pH was extremely simple, the values generated were highly reproducible within treatment groups and, for ventricle and radular retractor, the data agree closely with pH determinations made by other methods (Ellington, 1983a, 1985; Wiseman & Ellington, 1987) and thus serve as a good guide to changes in cellular pH values. Anoxia alone led to a significant drop (P < 0·03) in pH of the isolated muscles, the decreases being 0·49,1·0 and 0·74 pH units for foot, radular retractor and ventricle, respectively, over the 4 h incubation period (Table 4). However, the separate influences of pH and anoxia on the kinetic properties of PK could be evaluated by altering the pH of the sea water (with added buffers) so that aerobic tissues had a pH value equivalent to that of anoxic tissues and vice versa. Table 4 clearly shows that phosphorylation-mediated changes in the kinetic properties of PK (increased K_m for PEP, decreased I₅₀ for alanine) were correlated with anoxia and not with measured tissue pH. In all cases, when the pH of an anoxic tissue was similar to that of the corresponding aerobic tissue, anoxia-induced changes in either the K_m or the I₅₀ value for PK were still apparent; the kinetic parameters for the N₂/9·5 condition were not different from the values for N₂/SW. Furthermore, acidification under aerobic conditions did not produce the large changes in enzyme kinetic parameters characteristic of anoxia. Thus, the kinetic properties of PK in aerobic radular retractor, foot and ventricle samples, incubated at low pH, were not different from aerobic control values.

To determine whether CO₂ levels were critical in controlling enzyme phosphorylation in isolated whelk tissues, the effect of incubating tissues in the absence of CO₂ (bubbling tissues with 100% N₂) was compared with incubations performed at high CO₂ levels (95% N₂, 5% CO₂, see Table 4). The results of Table 5 show that PK kinetic parameters change from aerobic values to anoxic values regardless of the CO₂ content of the incubating atmosphere. Note that the tissue pH in anoxic heart and radular retractor dropped significantly both in the absence (100% N₂) and in the presence (95% N₂, 5% CO₂) of CO₂. This demonstrates that atmospheric CO₂ is not important in altering tissue pH in these tissues under the conditions of the incubation. The measured pH of anoxic foot muscle, however, did depend on the CO₂ content of the incubation atmosphere: no tissue acidification was observed in the absence of CO₂, whereas a pronounced acidification occurred when foot was incubated in 95% N₂, 5% CO₂. These results show that, although tissue pH was apparently altered by changing the CO₂ concentration of the incubation atmosphere, the tissue response to anoxia (shown by an increase in PK covalent modification) was not affected by the changes in CO₂ concentration.

The metabolic responses that characterize anaerobiosis may result from: (1) a direct response of enzyme(s) to changing oxygen tension, (2) the detection of changing O₂ levels by specific receptor(s) in each cell followed by triggering of second messenger release to coordinate intracellular responses, or (3) the selective detection of changing oxygen levels by a specialized receptor body followed by nervous or hormonal transmission to spread the signal to all tissues. The third possibility can undoubtedly be discounted, at least for marine invertebrates. The present study demonstrates that neither added hormones nor prostaglandin pathway inhibitors affect the biochemical responses to anoxia and, moreover, isolated tissues respond to anoxia in a manner identical (changes in fructose-2,6-P₂, phosphorylation of PK) to that seen in vivo (Plaxton & Storey, 1984a, 1985; Storey, 1988b). The same conclusion can be drawn from earlier studies; the patterns of substrate utilization and end-product accumulation were essentially equivalent when measured after anoxia in vivo or in vitro (Collicutt & Hochachka, 1977).

It is clear that the regulation of protein kinases and protein phosphatases is central to glycolytic control in anoxia. First, anoxia-induced covalent modifications of PK, phosphofructokinase and glycogen phosphorylase have been documented in marine molluscs (Holwerda et al. 1983; Plaxton & Storey, 1984a,b, 1985; Storey, 1984, 1988b, see also Table 1). Second, changes in fructose-2,6-P₂ levels predict that the phosphorylation state of 6-phosphofructo-2-kinase, an enzyme regulated by reversible phosphorylation in other animals, should be altered during anoxia (Hue & Rider, 1987). Third, an analysis of ³²P labelling patterns showed widespread phosphorylation of cellular proteins in whelk radular retractor muscle during anoxia (Plaxton & Storey, 1984b). It is probable, therefore, that protein phosphorylation is a key to glycolytic regulation during anaerobiosis and also has a major role in coordinating the depression of metabolic rate overall. The anoxia signal may be mediated through the actions of one of the known second messenger compounds in the cell, to stimulate the action of the appropriate protein kinase.

Four types of protein kinases are known which are responsive to: (a) cyclic AMP (protein kinase A), (b) cytoplasmic free Ca²⁺ (Ca²⁺/calmodulin-dependent protein kinase), (c) free Ca²⁺, diglycerides and phospholipids (protein kinase C), or (d) cyclic GMP (cyclic GMP-activated protein kinase) (see Krebs & Beavo, 1979; Lincoln, 1983; Mieskes et al. 1987; Pallen et al. 1987; Connelly et al. 1987; Ochs, 1988). The present experiments show selected effects of cyclic AMP and Ca²⁺/phospholipid stimulation on the kinetic properties of PK in whelk muscles. In particular, PK in aerobic muscles showed Ca²⁺-mediated modifications, indicating a role for Ca²⁺/calmodulin protein kinase or protein kinase C in the regulation of the muscle enzyme. However, none of the second-messenger-mediated effects on PK in aerobic muscles were of a magnitude equivalent to those stimulated by anoxia and second messengers did not enhance or alter the anoxia-induced modifications of PK. It is apparent, therefore, that the link between the anoxia signal and the coordinated covalent modification of cellular proteins is not, a simple function of one of the common second messengers used in this study.

Numerous studies have shown a correlation between decreased tissue pH and anaerobiosis in marine invertebrates (Ellington, 1983a,b; Portner et al. 1984; Walsh et al. 1984). These and other studies (Busa & Nuccitelli, 1984) devoted considerable attention to the potential role(s) of changing intracellular pH in metabolic regulation. The influence of pH in transitions to, and from, dormancy is now well documented for several systems (Busa & Nuccitelli, 1984), but no previous studies have attempted to dissociate the influence of pH from that of anoxia in the expression of the metabolic responses of anaerobiosis. In the present paper, the effects of anoxia and pH are separated by manipulating tissue pH. The results demonstrate (Table 4) that the kinetic changes in PK, which characterize anaerobiosis and anoxia-induced metabolic rate depression, correlate with anoxia and not with altered pH values. High tissue pH (equal to or greater than the aerobic, control value) could not prevent the changes in PK kinetic properties induced by anoxia. Acidosis under aerobic conditions produced only small alterations in PK properties of a magnitude much less than that seen during anoxia, even when tissue pH was lowered well below the value found in anoxic tissues. Since pH changes develop slowly over the course of anoxia (Ellington 1983a,b-, Portner et al. 1984), and since anoxia-tolerant animals take steps to minimize pH changes during anoxia (metabolic rate depression, alternative products to lactic acid), it is not surprising that pH does not trigger the biochemical responses of metabolic depression. However, tissue acidosis is a real consequence of anoxia in all animal systems and a low-pH environment may create a ‘metabolic context’ (Busa & Nuccitelli, 1984) that favours the induction of metabolic rate depression. Changing cellular pH may similarly have other facilitating effects, such as altering relative activities of protein kinases versus protein phosphatases.

Experiments to determine the effects of anoxia on pyruvate kinase phosphorylation were also performed in the presence and absence of CO₂ in the incubating atmosphere (Table 5). These experiments clearly show that, although using CO₂-rich mixtures had some effect in acidifying foot tissue, phosphorylation of PK occurred regardless of the CO₂ content of the incubating atmosphere. These results show that the regulation of enzyme phosphorylation by CO₂ levels is unlikely in this animal, and agree with the results showing that enzyme phosphorylation is largely independent of the tissue incubation pH. These results also indicate that increased CO₂ concentrations are not the cellular signal mediating the changes in PK phosphorylation but, rather, the absence of O₂ is the universal factor which triggers anoxia responses in whelk tissues.

Anoxia-induced metabolic rate depression is widespread throughout the animal kingdom, and the biochemical mechanisms that form the basis of this phenomenon are expressed in numerous invertebrate and vertebrate facultative anaerobes (Storey, 1985, 1988a). Enzyme/protein control via reversible phosphorylatipn is a particularly well-documented mechanism of metabolic depression. The present study demonstrates, however, that common routes of protein kinase control (via hormones and/or cyclic AMP, Ca²⁺ or Ca²⁺/phospholipid second messengers) are not the means by which the anoxic signal is transmitted. Alternative mechanisms must be sought. These could include: (1) a requirement for a different type of second messenger or for synergistic action by cyclic AMP and Ca²⁺, (2) direct triggering by oxygen itself of cellular protein kinases or phosphatases (perhaps via an O₂-binding protein), or (3) an ‘anoxic shock’ response that modifies transcription/translation in whelk tissues to induce rapid synthesis of specific protein kinase(s).

The authors thank J. M. Storey for reading the manuscript critically. Supported by an NSERC operating grant to KBS and an NSERC postdoctoral fellowship to SPJB.