Ca2+/Calmodulin-independent Activation of Calcineurin from Dictyostelium by Unsaturated Long Chain Fatty Acids*

This study describes a novel mode of activation for the Ca2+/calmodulin-dependent protein phosphatase calcineurin. Using purified calcineurin fromDictyostelium discoideum we found a reversible, Ca2+/calmodulin-independent activation by the long chain unsaturated fatty acids arachidonic acid, linoleic acid, and oleic acid, which was of the same magnitude as activation by Ca2+/calmodulin. Half-maximal stimulation of calcineurin occurred at fatty acid concentrations of approximately 10 μm with either p-nitrophenyl phosphate or RII phosphopeptide as substrates. The methyl ester of arachidonic acid and the saturated fatty acids palmitic acid and arachidic acid did not activate calcineurin. The activation was shown to be independent of the regulatory subunit, calcineurin B. Activation by Ca2+/calmodulin and fatty acids was not additive. In binding assays with immobilized calmodulin, arachidonic acid inhibited binding of calcineurin to calmodulin. Therefore fatty acids appear to mimic Ca2+/calmodulin action by binding to the calmodulin-binding site.

The enzyme has some unusual features as compared with its counterparts from other species, e.g. N-terminal and C-terminal extensions (9) resulting in its high molecular mass of 73 kDa. Expression of the single calcineurin A gene is developmentally regulated both at the mRNA and the protein level (9). Biochemical characterization of the protein purified from recombinant bacteria and an overexpressing Dictyostelium cell line revealed enzymatic properties that are fairly comparable with calcineurins from other sources (10). Pharmacological evidence suggests at least three distinct roles for Dictyostelium calcineurin. Using the calcineurin inhibitors FK 506 and cyclosporin A, Horn and Gross (11) have provided evidence that calcineurin is involved in both pathways of Dictyostelium differentiation. Moniakis et al. (12) have shown by using cyclosporin A that the Ca 2ϩ regulated expression of the Ca 2ϩ ATPase, PAT1, is under the control of calcineurin. The calcineurin inhibitor deltamethrin, but not the inactive analog perimethrin, was shown to inhibit chemotaxis toward the attractant cAMP, suggesting a role in chemotaxis for calcineurin (13).
Ca 2ϩ signaling in Dictyostelium is essential for chemotaxis toward cAMP (for review see Ref. 14). The binding of cAMP to its cell surface receptors leads to Ca 2ϩ influx across the plasma membrane and to an increase in the cytosolic Ca 2ϩ concentration (15)(16)(17). It was shown that several inhibitors for phospholipase A 2 blocked cAMP induced Ca 2ϩ influx to a great extent, suggesting a prominent role for phospholipase A 2 in the signal transduction pathway leading to Ca 2ϩ uptake (18). Phospholipase A 2 releases fatty acids from phospholipids. It was demonstrated that long chain fatty acids mimic the action of cAMP by bypassing both cAMP-receptor stimulation and activation of phospholipase A 2 , directly inducing Ca 2ϩ influx (18). Moreover long chain fatty acids liberate Ca 2ϩ from internal stores and thus induce capacitative Ca 2ϩ entry (19).
Dictyostelium cells, which live as amoebae in the soil, are subject to large environmental changes, e.g. fluctuations in ion supply. Thus robust mechanisms supporting Ca 2ϩ signaling are required to maintain chemotactic capability. One group of candidates for signaling molecules are long chain fatty acids (18,19). In this study we have investigated direct effects of long chain fatty acids on Dictyostelium calcineurin activity and found a novel, Ca 2ϩ /calmodulin-independent mode of activation for calcineurin.
from Dictyostelium was purified from an overexpressing cell line as described previously (10). Purification and characterization of recombinant Dictyostelium calcineurin B will be described elsewhere. 1 Phosphatase Activity Assays-Phosphatase activity assays with pnitrophenyl phosphate (pNPP) 2 as substrate were performed routinely in duplicate or triplicate according to Hellstern et al. (10). Briefly, calcineurin A (final concentration, 140 nM) was preincubated in the absence or presence of calcineurin B (140 nM) and/or calmodulin (1 M) in buffer A (50 mM Tris-HCl, 0.5 mM MnCl 2 , 0.5 mM dithiothreitol, 0.2 mM CaCl 2 , 20 mM MgCl 2 , pH 7.8) for 1 h. Subsequently fatty acid solutions or 1% ethanol (vehicle) as a control were added 1 min prior to addition of pNPP (20 mM). The reaction (total volume, 150 l) was stopped after 1 h by adding an equal volume of 1 M Na 2 CO 3 , 20 mM EDTA, 2 mM EGTA, pH 8.0. All incubations were performed at 30°C. The absorbance of the dye was measured at 405 nm and corrected for blank absorbance.
For determination of phosphatase activity with RII phosphopeptide as a substrate, the reaction was initiated by adding 35 M of the phosphopeptide (total volume after addition, 50 l). To terminate the reaction, 2 volumes of Biomol Green reagent were added after 40 min. The test was linear from 10 to 60 min. The dye was allowed to develop for 25 min before measuring the absorbance at 620 nM. In the experiments with calcineurin A alone, the final concentration of the phosphatase was 420 nM. In the presence of both calcineurin A and B, the concentration of each was 252 nM. The concentration of calmodulin was 1 M throughout. 1 unit is defined as the release of 1 nmol phosphate min Ϫ1 mg protein Ϫ1 .
Binding of Calcineurin A to Calmodulin-Sepharose 4B-Binding of calcineurin to calmodulin-Sepharose 4B in the absence or presence of arachidonic acid or arachidonic acid methyl ester was determined at 4°C in a total volume of 250 l containing 170 nM calcineurin A, arachidonic acid, or arachidonic acid methyl ester (20 M each) or 1% ethanol (vehicle), and 100 l of a calmodulin-Sepharose 4B slurry equilibrated in buffer B (50 mM Tris-HCl, 150 mM NaCl, 2 mM CaCl 2 , 1 mM EDTA, 0.5 mM dithiothreitol, pH 7.8). The mixture was incubated under constant agitation for 30 min. The Sepharose was pelleted by centrifugation in an Eppendorf centrifuge for 2 s. 150-l aliquots of the supernatant were withdrawn, and 100 l was loaded onto SDS-polyacrylamide gels and subsequently subjected to Western blotting. For washing the resin the volume was adjusted to 250 l with buffer B, resuspended, and recentrifuged. This procedure was repeated twice. After the second wash the supernatant was removed thoroughly. The protein was eluted by adjusting the volume to 250 l with elution buffer (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 20 mM EGTA, 0.5 mM dithiothreitol, pH 7.8), resuspending the resin, and incubation for 15 min. Elution was repeated once. Samples for SDS-polyacrylamide gel electrophoresis were withdrawn after each step as above.
To elute calcineurin A with arachidonic acid, the phosphatase was bound to calmodulin-Sepharose 4B in buffer B as above. The resin was washed twice with the same buffer. Subsequently buffer B containing 20 M arachidonic acid or 20 M arachidonic acid methyl ester or 1% ethanol was added, and the samples were incubated for 15 min. The gel was washed once with buffer B, and the final elution was carried out with elution buffer as above. Aliquots for SDS-polyacrylamide gel electrophoresis were taken as described above.
Other Methods-The critical micellar concentration of fatty acids was determined as described by Chattopadhyay and London (22). Ethanolic stock solutions of arachidonic acid, linoleic acid, and arachidonic acid methyl ester (100 mM each) were stored under nitrogen at Ϫ80°C. Diluted fatty acid solutions were prepared from stock aliquots immediately prior to the experiment and kept under nitrogen until use. All other fatty acid solutions were freshly prepared from solids or oils. Fig. 1A shows a dose-response curve for activation of purified Dictyostelium calcineurin A by arachidonic acid (20:4) in the absence of calmodulin and calcineurin B. Half-maximal stimulation of the phosphatase activity with pNPP as substrate occurred between 5 and 10 M of the fatty acid. Maximal stimulation was observed at 30 M arachidonic acid and resulted in an approximately 4-fold activation. Arachidonic acid and calmodulin activated the phosphatase to the same extent. Calcineurin activities in the presence of 30 M arachidonic acid and a saturating dose of calmodulin (1 M) are not significantly different. Fig. 1B shows the activation of the holoenzyme by arachidonic acid. The dose-response relation is very similar to the one above, with half-maximal stimulation at 10 M arachidonic acid and maximal stimulation at 30 M resulting in a 3-3.5-fold stimulation. Again, there was no significant difference between the maximal stimulation by arachidonic acid and calmodulin. Half-maximal activation was observed well below the critical micellar concentration of 20 M for arachidonic acid in the buffer system used for the phosphatase assays. Because there was no influence of calcineurin B on the activation process, subsequent experiments with pNPP as substrate were performed in the absence of the regulatory subunit.

Effect of Arachidonic Acid on Calcineurin Activity-
Substrate Specificity of the Activation of Calcineurin by Arachidonic Acid-The synthetic nondecapeptide corresponding to the regulatory subunit of the cAMP-dependent protein kinase (type II) is a model peptide substrate for calcineurin (23). As shown previously (10), the basal activity of Dictyostelium calcineurin A toward the RII peptide in the absence of the regulatory subunit calcineurin B (Fig. 2) is lower than toward the nonpeptide substrate pNPP (0.4 Ϯ 0.3 unit versus 152 Ϯ 32 units for RII peptide (n ϭ 4) and pNPP (n ϭ 25), respectively). Similar observations have been reported for recombinant Neurospora crassa calcineurin A (24).
Calcineurin B activates the catalytic subunit by lowering the K m (5). This is reflected by the rise in calcineurin activity upon addition of calcineurin B. In the absence of calcineurin B, Ca 2ϩ /calmodulin activated calcineurin to approximately the same extent. Full activation was obtained in the presence of both calcineurin B and calmodulin. The activation by calmodulin in the absence of calcineurin B could be mimicked by the addition of 50 M arachidonic acid. Addition of both calcineurin B and arachidonic acid again yielded full activation of the phosphatase. 50 M arachidonic acid is a saturating dose for stimulation of the phosphatase activity. Half-maximal activity was obtained at 10 M of the fatty acid (data not shown). Activation of calcineurin by arachidonic acid is therefore also valid for a peptide substrate.
Effects of Other Unsaturated Fatty Acids on Calcineurin Activity-The unsaturated long chain fatty acids linoleic acid (18:2) and oleic acid (18:1) exerted effects on calcineurin similar to those of arachidonic acid with respect to half-maximal and maximal stimulating concentrations (Fig. 3). The maximal stimulation by linoleic acid shown in Fig. 3 seems to be lower than that of oleic acid or arachidonic acid, but there is no significant difference between the maximal stimulation by linoleic acid and by Ca 2ϩ /calmodulin (t test at the 0.05 level).
Specificity of Stimulation by Unsaturated Fatty Acids-To investigate the specificity of the effect of long chain unsaturated fatty acids on calcineurin and to obtain insight into the mechanism of stimulation, the methyl ester of arachidonic acid and the saturated fatty acids arachidic acid (20:0) and palmitic acid (16:0) were used. The methyl ester had no effect on the phosphatase activity (Table I). Moreover, both of the saturated fatty acids neither stimulated basal activity nor influenced the Ca 2ϩ /calmodulin-stimulated activity at concentrations that are saturating for the unsaturated fatty acids (50 M). The data in Table I also show that the stimulation of the phosphatase activity by unsaturated fatty acids is not additive to Ca 2ϩ / calmodulin stimulation because the stimulation in the presence of both unsaturated fatty acids and Ca 2ϩ /calmodulin did not exceed the values in the presence of either compound. Taken together, these results suggest that both the carboxy group and the unsaturated character of the fatty acids are required for stimulation of calcineurin. Calmodulin and unsaturated fatty acids may exert their effects through a similar mechanism, possibly by binding to the same site.
Reversibility of Calcineurin Activation by Arachidonic Acid-Next we analyzed whether the activation of calcineurin by arachidonic acid is reversible. Bovine serum albumin binds to arachidonic acid with high affinity (K d , 62 nM) (25) and should therefore effectively abolish activation by the fatty acid. Activation of calcineurin by 30 M arachidonic acid resulted in an approximately 3-fold stimulation over basal activity (Fig. 4).  Addition of 15 M fatty acid-free bovine serum albumin reduced the phosphatase activity to basal levels, indicating that complexing the fatty acid stops the activation process. The enzyme was thereafter susceptible to stimulation by Ca 2ϩ /calmodulin as shown in Fig. 4. Activation by Ca 2ϩ /calmodulin after capturing the fatty acid was of the same magnitude as the response to arachidonic acid, demonstrating that the structure of calcineurin had not been disrupted irreversibly by arachidonic acid. Inhibition of Calcineurin Binding to Immobilized Calmodulin by Arachidonic Acid-Several lines of evidence suggested that unsaturated fatty acids and Ca 2ϩ /calmodulin exert their effects on calcineurin activity through similar mechanisms by binding to the same site. Therefore we tested whether arachidonic acid interferes with binding of purified Dictyostelium calcineurin to immobilized calmodulin. In a first set of experiments we examined binding of calcineurin to calmodulin in the presence of arachidonic acid (Fig. 5, A-C). Arachidonic acid inhibited binding of the phosphatase to calmodulin-Sepharose (Fig. 5B), whereas with equal concentrations of arachidonic acid methyl ester (Fig. 5C) most of the protein bound to the gel and could subsequently be eluted by buffer containing EGTA. In a second set of experiments we investigated whether a buffer containing arachidonic acid was able to elute prebound calcineurin from calmodulin-Sepharose (Fig. 5, D-F). This is indeed the case, as shown in Fig. 5E. About half of the prebound protein could be displaced in a single elution step from calmodulin (Fig. 5E, lane E*), whereas all the protein remained bound to calmodulin in the presence of arachidonic acid methyl ester (Fig. 5F) and could only be eluted from calmodulin-Sepharose by incubation with EGTA. DISCUSSION We show here that the long chain unsaturated fatty acids arachidonic acid, linoleic acid, and oleic acid activate calcineurin from Dictyostelium. The activation is independent of Ca 2ϩ /calmodulin. Half-maximal activation occurred at a concentration (5-10 M) well below the critical micellar concentration (20 M), indicating that it is not merely an effect of calcineurin binding to micelles. Politino and King (26) have described activation of calcineurin from bovine brain by binding to acidic phospholipids. This effect is clearly distinct from the one reported here because phospholipids bind to the regulatory subunit calcineurin B (27). Activation of Dictyostelium calcineurin by unsaturated fatty acids is independent of the presence of calcineurin B. Stimulation of enzymatic activity is specific for unsaturated fatty acids because saturated fatty acids were not active. Moreover, the negatively charged carboxy group is necessary for the activation process because the methyl ester derivative of arachidonic acid was inactive. Stimulation of calcineurin by fatty acids was demonstrated with chemically distinct substrates, the chromogenic substrate pNPP, and the RII phosphopeptide substrate corresponding to a phosphorylation site within the RII subunit of the cAMP-dependent protein kinase. The effect is readily reversible as shown by relieving the activation of calcineurin through capturing the fatty acids by albumin. Calcineurin activation by fatty acids could be a means for in vivo regulation. We suggest that stimulation by fatty acids or Ca 2ϩ /calmodulin are alternative mechanisms.
What are the structural determinants for fatty acid binding to calcineurin A? The most likely candidates are the hydrophobic binding sites for calcineurin B and Ca 2ϩ /calmodulin. The following data exclude the calcineurin B-binding site as a candidate: (i) arachidonic acid activates the phosphatase both in the absence and in the presence of calcineurin B (Fig. 1). Even in the presence of low Ca 2ϩ concentrations the binding of calcineurin B to the catalytic subunit is nearly irreversible (5), FIG. 4. Reversibility of the activation of calcineurin by arachidonic acid. Phosphatase assays were carried out in the presence of 140 nM calcineurin A. Left bar, calcineurin A was preincubated in buffer A for 50 min. Subsequently, 30 M arachidonic acid (AA) was added and incubated for 10 min, and then 20 mM pNPP was applied to start the reaction. The reaction was stopped after 1 h as outlined under "Experimental Procedures." Middle bar, same as above but 1 min before starting the reaction 15 M bovine serum albumin (BSA) was added. Right bar, calcineurin A was preincubated for 1 or 10 min in the presence of 30 M arachidonic acid. After this period bovine serum albumin was added, incubated for 1 min before the addition of calmodulin (CaM, 1 M), and incubated for a further 60 or 50 min. Thereupon the reaction was started and finally stopped as above. Data are the means Ϯ S.E. of duplicates from three independent experiments. Data are given as the activity relative to basal activity measured in the absence of arachidonic acid or calmodulin. 150-l samples were prepared for SDS-polyacrylamide gel electrophoresis, and 100 l were run on a 9% SDS-polyacrylamide gel and subsequently subjected to Western blotting. Immunoreactivity was detected with an antiserum (dilution 1:5000) raised against recombinant Dictyostelium calcineurin. The position of molecular mass standards are shown on the left. Dictyostelium calcineurin A migrates at a higher apparent molecular mass than that derived from the cDNA sequence. The experiment was performed twice with similar results. and the calcineurin A/B complex is only dissociated in the presence of urea, SDS, or guanidine hydrochloride (Ref. 6 and references therein) and (ii) if arachidonic acid disturbed calcineurin B binding to calcineurin A, then the phosphatase activity would not increase upon addition of calcineurin B with RII phosphopeptide as a substrate (Fig. 2). Our results suggest that Ca 2ϩ /calmodulin and unsaturated fatty acids share the same binding domain and exert their effects through similar mechanisms: (i) arachidonic acid inhibits binding of calcineurin A to calmodulin and displaces calcineurin from immobilized calmodulin (Fig. 5); (ii) activation by fatty acids and by Ca 2ϩ / calmodulin is not additive (Table I); (iii) not only the full stimulation by Ca 2ϩ /calmodulin of the RII phosphopeptide phosphatase activity in the presence of calcineurin B could be mimicked by arachidonic acid but also the partial stimulation in the absence of calcineurin B (Fig. 2); and (iv) maximal activation by unsaturated fatty acids and Ca 2ϩ /calmodulin is in the same range ( Figs. 1 and 2).
Although these data argue strongly in favor of direct competition of Ca 2ϩ /calmodulin and arachidonic acid for binding to the same site, we cannot distinguish between fatty acid binding to the enzyme and to Ca 2ϩ /calmodulin in our competition experiments. Binding of Ca 2ϩ to calmodulin results in exposure of hydrophobic and acidic amino acids on the protein surface that interact with a basic amphipathic ␣-helix within the target protein (28). Therefore hydrophobic and ionic interactions might be required for binding of fatty acids to the basic amphipathic ␣-helix, which forms the Ca 2ϩ /calmodulin-binding site (8). Accordingly, not only the unsaturated hydrophobic character of the lipids is essential for proper activation but also the additional negatively charged carboxy residue. Therefore, the acidity of Ca 2ϩ /calmodulin renders an interaction with acidic lipids unlikely.
Several calmodulin-binding proteins were found to be activated by long chain fatty acids. Human erythrocyte Ca 2ϩ ATPase was activated by oleic and linoleic acid (29 -31), bovine brain cyclic nucleotide phosphodiesterase was stimulated by oleic acid (31,32), and porcine cyclic nucleotide phosphodiesterase was activated by both unsaturated and saturated fatty acids (33). Myosin light chain kinase from chicken gizzard was stimulated by arachidonic acid but not by arachidic acid (34). Activation of Ca 2ϩ /calmodulin-dependent enzymes by binding of fatty acids to their Ca 2ϩ /calmodulin-binding sites could therefore represent a more common, alternative mode of regulation of calmodulin-binding proteins.
Two other protein Ser/Thr phosphatases were recently shown to be activated by unsaturated fatty acids, the phosphatases type 5 (35,36) and type 2C (37). In both cases activation occurred at fatty acid concentrations greater than 100 M, which was above the critical micellar concentration. In case of protein phosphatase 5 it was demonstrated that lipid vesicles consisting of fatty acids stimulate the enzymatic activity by binding to a motif consisting of 34 amino acids called tetratricopeptide repeat (TPR) domain. The TPR domain in protein phosphatase P5 is supposed to form a shield in front of the active site, and binding of lipid vesicles was suggested to result in displacement of the TPR domain allowing access of substrates to the active site (36). We could not identify a TPR domain in Dictyostelium calcineurin. Activation of calcineurin by binding of unsaturated fatty acids to the calmodulin-binding domain might, however, represent a similar mechanism by relieving inhibition of the C-terminal autoinhibitory domain.
The physiological significance of calcineurin activation by unsaturated fatty acids is so far uncertain. No data on the release of fatty acids into the cytosol of Dictyostelium are available. However, in glucose-stimulated pancreatic islets, a rise in the mass of free arachidonic acid sufficient to yield an overall cellular concentration of 38 -75 M was measured (38), making regulation of soluble proteins such as calcineurin possible.
Unterweger and Schlatterer (39) have shown that an increase in the cytosolic Ca 2ϩ concentration is necessary for orientation and locomotion of Dictyostelium cells since clamping the cytosolic Ca 2ϩ concentration by introducing the Ca 2ϩ buffer [1,2-bis(o-aminophenoxy)ethane-N,N,NЈ,NЈ-tetraacetic acid] inhibited chemotaxis. In the presence of extracellular EGTA, Dictyostelium cells were able to protrude pseudopods and to migrate, albeit at lower velocity, indicating that liberation of Ca 2ϩ from stores is sufficient for the chemotactic response (39). The calcineurin inhibitor deltamethrin was reported to inhibit chemotaxis in Dictyostelium, suggesting a role for calcineurin (13). Recently Persechini and Cronk (40) have shown that the free cytosolic Ca 2ϩ concentration necessary to produce sufficient Ca 2ϩ /calmodulin for in vivo activation of high affinity Ca 2ϩ /calmodulin-binding proteins is in the range of several hundred nM, presuming that the K d of Ca 2ϩ /calmodulin binding to the calmodulin-binding protein is 2 nM. The K d for Ca 2ϩ /calmodulin binding to Dictyostelium calcineurin is in the low nanomolar range (41). At the low extracellular Ca 2ϩ concentration in the range of 200 nM the free cytosolic Ca 2ϩ concentration cannot increase above this value merely by opening Ca 2ϩ channels in the plasma membrane. Ca 2ϩ influx from the extracellular medium would thus not allow activation of calcineurin under these conditions. Nevertheless Dictyostelium cells are able to perform chemotaxis at the extracellular Ca 2ϩ concentration of 200 nM (42). 3 Recently it was demonstrated that arachidonic acid releases Ca 2ϩ from intracellular stores and thus induces capacitative Ca 2ϩ entry (18,19). A direct effect of arachidonic acid on Ca 2ϩ -or H ϩ -ATPases could be excluded (19). The mechanism of arachidonic acid-induced Ca 2ϩ release remained unclear. The stimulatory effect of unsaturated fatty acids described here could mean that calcineurin is involved in Ca 2ϩ release.
A testable model for the signal transduction pathway resulting in Ca 2ϩ signaling would imply receptor occupation by cAMP, activation of phospholipase A 2 , and release of unsaturated fatty acids that activate calcineurin. Calcineurin might trigger liberation of Ca 2ϩ , which would be sufficient for the chemotactic response (39). In addition, Ca 2ϩ release leads to capacitative Ca 2ϩ entry even at low extracellular Ca 2ϩ concentrations (18). Taken together, this leads to an increase in the Ca 2ϩ /calmodulin concentration resulting in the alternative activation of calcineurin. As soon as the cytosolic Ca 2ϩ concentration reaches resting levels through Ca 2ϩ ATPase activity, calcineurin activation is relieved.