cAMP controls cytosolic Ca2+ levels in Dictyostelium discoideum

Background Differentiating Dictyostelium discoideum amoebae respond upon cAMP-stimulation with an increase in the cytosolic free Ca2+ concentration ([Ca2+]i) that is composed of liberation of stored Ca2+ and extracellular Ca2+-influx. In this study we investigated whether intracellular cAMP is involved in the control of [Ca2+]i. Results We analyzed Ca2+-fluxes in a mutant that is devoid of the main cAMP-phosphodiesterase (PDE) RegA and displays an altered cAMP metabolism. In suspensions of developing cells cAMP-activated influx of extracellular Ca2+ was reduced as compared to wild type. Yet, single cell [Ca2+]i-imaging of regA- amoebae revealed a cAMP-induced [Ca2+]i increase even in the absence of extracellular Ca2+. The cytosolic presence of the cAMP PDE inhibitor 3-isobutyl-1-methylxanthine (IBMX) induced elevated basal [Ca2+]i in both, mutant and wild type cells. Under this condition wild type cells displayed cAMP-activated [Ca2+]i-transients also in nominally Ca2+-free medium. In the mutant strain the amplitude of light scattering oscillations and of accompanying cAMP oscillations were strongly reduced to almost basal levels. In addition, chemotactic performance during challenge with a cAMP-filled glass capillary was altered by EGTA-incubation. Cells were more sensitive to EGTA treatment than wild type: already at 2 mM EGTA only small pseudopods were extended and chemotactic speed was reduced. Conclusion We conclude that there is a link between the second messengers cAMP and Ca2+. cAMP-dependent protein kinase (PKA) could provide for this link as a membrane-permeable PKA-activator also increased basal [Ca2+]i of regA- cells. Intracellular cAMP levels control [Ca2+]i by regulating Ca2+-fluxes of stores which in turn affect Ca2+-influx, light scattering oscillations and chemotactic performance.


Background
Starving Dictyostelium discoideum amoebae form a multicellular organism by chemotactic aggregation. The signaling molecule that mediates aggregation and development is cAMP. Aggregation proceeds in a rhythmic fashion; cAMP is secreted periodically by cells in the center of the aggregate. Cells in the neighbourhood respond by an oriented inward movement and secrete cAMP themselves to relay the signal. In cell suspensions periodic synthesis and release of cAMP leads to rhythmic shape changes that cause alterations in light transmittance and spike-shaped and sinusoidal light scattering oscillations [1]. The marked rhythmic behaviour of the cell population is also apparent by oscillations of other parameters, e.g. extracellular concentrations of Ca 2+ , K + or H + (for review see [2]). Recently, changes in [Ca 2+ ] i were postulated to comprise the (or at least a part of the) master oscillator controlling oscillation patterns [3,4]. A short [Ca 2+ ] i -transient induced by addition of CaCl 2 or calmodulin antagonists alters light scattering oscillations and can even reset the oscillation phase [3]. The height of the [Ca 2+ ] i -increase determines whether light scattering and the accompanying cAMP oscillations are abolished or augmented: large [Ca 2+ ] i -transients inhibit cAMP and light scattering oscillations [3] whereas small [Ca 2+ ] i -elevations enhance oscillations of both parameters [4]. From these experiments it was concluded that Ca 2+ exerts a dual control over the production of the first messenger cAMP (for a detailed model see [4]). cAMP controls its own synthesis as binding of the agonist to cell surface receptors induces a transient [Ca 2+ ] ielevation [5][6][7]. However, until now the question as to whether there is an interaction between cAMP acting intracellularly as second messenger and [Ca 2+ ] i in D. discoideum has not been resolved. In other cell systems such as nerve cells crosstalk between the cAMP and the Ca 2+ signaling pathway exists that is necessary to generate oscillations of both parameters [8].
In order to gain insight into a possible connection between intracellular cAMP and [Ca 2+ ] i we used a mutant defective in the phosphodiesterase RegA. RegA is one out of two cAMP-specific phosphodiesterases (for an overview of classes of PDEs in Dictyostelium see [9]) that is inhibited by IBMX and comprises part of an eukaryotic phosphorelay system [10,11]. RegAmutants are rapid developers; their differentiation is shifted towards the stalk pathway [12,13]. Chemotactic migration is characterized by an increased frequency of lateral pseudopod extension as compared to wild type amoebae [14]. We found that the mutant displayed an altered [Ca 2+ ] i -response pattern upon stimulation with cAMP with an augmentation of Ca 2+ -release from stores and a concomitant decrease of extracellular Ca 2+ -entry. Light scattering oscillations and the underlying cAMP oscillations were drastically reduced in regAcells. Chemotaxis was influenced by the extracellular presence of EGTA. We conclude that indeed, intracellular cAMP signaling and the regulation of [Ca 2+ ] i are linked at the level of Ca 2+ -storage compartments.

Extracellular and intracellular [Ca 2+ ]-recordings
To test whether the absence of the main cAMP-specific phosphodiesterase affects regulation of [Ca 2+ ] i we analyzed extracellular Ca 2+ -fluxes in cell suspensions and studied [Ca 2+ ] i in single amoebae. cAMP-induced Ca 2+influx in suspensions of regAcells occurred with a similar time course as in wild type. Yet, influx was reduced by approximately 40% (Fig. 1). The loss of RegA should lead to an altered cAMP metabolism. Indeed, the basal total amount of cAMP was increased fourfold (13 ± 3 pmol/10 7 regAcells; mean ± s.e.m. of 16  The effect of the increased basal cAMP concentration on the [Ca 2+ ] i -regulation in regAamoebae might be caused by a change in the characteristics of Ca 2+ -fluxes of internal stores. A positive influence of cAMP via PKA-mediated phosphorylation of both, IP 3 -receptors and ryanodine receptors on release of stored Ca 2+ has been reported (for review see [19]). We therefore tested the response of regAamoebae upon stimulation with cAMP in the presence of the chelator BAPTA. We found that even after the addition of 1 mM BAPTA cAMP activated a transient increase in [Ca 2+ ] i (Fig. 4). The elevation was smaller than that observed in nominally Ca 2+ -free medium and amounted to an average of 44 ± 3 nM above basal [Ca 2+ ] i (mean ± s.e.m. of 18 determinations in 2 independent experiments). In wild type amoebae a cAMP-stimulated [Ca 2+ ] iincrease is not detectable in the presence of BAPTA; the occurrence of a transient [Ca 2+ ] i -elevation in regAcells indicates an augmented release of Ca 2+ from stores in the mutant. Support for an effect of cAMP via PKA came from experiments where we incubated cells with the membrane permeant activator of PKA, Sp-5,6-DCl-cBIMPS [20,21]. Basal [Ca 2+ ] i was increased in regAcells upon treatment with 30 µM Sp-5,6-DCl-cBIMPS for 60 min (139 ± 2 nM; mean ± s.e.m. of 15 determinations in 3 independent experiments); agonist-induced [Ca 2+ ] i -transients in nominally free Ca 2+ -buffer were unaltered in height (87 ± 8 nM; mean ± s.e.m.) as compared to control cells. In addition, we found that preincubation of wild type amoebae with 30 µM Sp-5,6-DCl-cBIMPS reduced cAMP-activated Ca 2+influx in cell suspensions by 26 ± 8% (mean ± s.e.m. of 3 independent experiments).

Light scattering and extracellular Ca 2+ oscillations depend on internal cAMP levels
We had shown previously that artificial changes of [Ca 2+ ] i , either by affecting Ca 2+ -stores or by activating Ca 2+ -influx alter light scattering oscillations [3,4]. When light scattering was analyzed in regAsuspensions two types of responses were observed. On one hand, regular oscilla-tions with a phase length of 4.3 ± 1 min (mean ± s.d. of 61 determinations in 6 independent experiments) occurred (Fig. 5A). The amplitude of these oscillations was reduced as compared to wild type (Fig. 5B), i.e. by 78%. On the other hand, irregular light scattering changes were detected (Fig. 5C). Determination of cAMP levels revealed that cAMP scarcely oscillated in regA - (Fig. 5D) and increased on average by a factor of 2.9 ± 0.6 (mean ± s.e.m. of 5 independent experiments). The response upon addition of cAMP was also different: after an increased first light scattering peak and the occurrence of a second peak light scattering did not return to the baseline as in wild type suspensions but fell well below (Fig. 6). The alteration in light scattering responses in the mutant might be due to a shift in sensitivity to cAMP. As a control we tested the reaction upon stimulation with cAMP and found that regAcells reacted when 3 nM cAMP was added (not shown) which indicates that the mutant strain is practically as sensitive as wild type. Measurement of [Ca 2+ ] e in regAcell suspensions revealed irregular [Ca 2+ ] e oscillations, similar to the results obtained for light scattering (Fig. 7).

Chemotaxis of regAamoebae
It had been reported previously that regAcells have a reduced capacity to suppress lateral pseudopod formation [14]. In accordance with the data presented by Wessels et al. [14] we also observed augmented lateral pseudopod Ca 2+ -influx after cAMP stimulation is reduced in regAcells Figure 1 Ca 2+ -influx after cAMP stimulation is reduced in regAcells. The amount of influx (pmol/10 7 cells) after addition of 1 µM cAMP is plotted versus extracellular [Ca 2+ ]. Average influx was higher in wild type than in regAamoebae (mean ± s.d. from at least 6 determinations in 3 independent experiments each).
extension upon challenge of aggregation competent amoebae with a cAMP filled glass capillary (not shown). The reduction in chemotactic polarization was reflected by a decrease in the average chemotactic speed as com-pared to wild type amoebae (Fig. 8). Pretreatment with EGTA to empty Ca 2+ -storage compartments dose-dependently inhibited chemotaxis of regAand wild type. The EGTA-incubated cells were rounded and extended only tested (P < 0.001) as compared to control cells; Mann Whitney rank sum test). Wild type cells were unaffected by preincubation with 5 mM EGTA for up to 1 hour whereas at 10 mM EGTA chemotaxis was reduced.

Discussion
The cytosolic concentration of Ca 2+ was demonstrated to control light scattering oscillations by affecting the synthesis of cAMP; depending on the height of an artificial [Ca 2+ ] i -transient the production of cAMP which in this case serves as first messenger was either augmented or blocked [3,4]. The results presented in this study provide evidence for a reciprocal influence of the second messengers cAMP and Ca 2+ in Dictyostelium cells. We observed altered agonist-induced Ca 2+ -fluxes and [Ca 2+ ] i -transients in the regAmutant cell line where the absence of the main cAMP-hydrolyzing PDE led to a fourfold increased basal cAMP level. One could argue that the effect on [Ca 2+ ] i was not a consequence of the increased basal concentration of cAMP but rather due to a potentially altered pattern of gene expression in the mutant strain. Indeed, this is possible and could result in a different signal perception and/ or processing. However, we consider an alteration in gene expression unlikely to be responsible for the augmented [Ca 2+ ] i -transients upon cAMP-stimulation since the same effect could be evoked in wild type amoebae by loading of the PDE inhibitor IBMX into the cytosol. In addition, IBMX evoked an increase in basal [Ca 2+ ] i in both, wild type and mutant cells. In regAthe inhibitor should act on the additional cAMP-PDE (PDE-E) [16,17] and therefore increase cAMP levels even further. In wild type amoebae hydrolysis of cAMP should be retarded as well. Yet, the threshold of the cAMP concentration required to increase basal [Ca 2+ ] i might not be achieved as consistently as in the mutant since IBMX must act on both PDEs.
The sensitizing effect of the increased amount of cAMP on [Ca 2+ ] i could be caused by several factors. Ca 2+ -flux characteristics can be changed by influencing Ca 2+ -channels and/or Ca 2+ -ATPases located on both, the plasma membrane and membranes of internal stores. When we analyzed Ca 2+ -fluxes with a Ca 2+ -sensitive electrode influx

Strain Condition
Basal  Ca 2+ -channel. In other cell systems activation of the PMCA and of Ca 2+ -channels by an increase in cAMP levels was shown [22][23][24] but our data argue against a stimulating effect on plasma membrane Ca 2+ -channel or PMCA activity in Dictyostelium amoebae.
The second target of action of cAMP are intracellular stores. Indeed, we showed for the first time that in Dicty-ostelium a cAMP-activated [Ca 2+ ] i -elevation occurred in the extracellular presence of the Ca 2+ -chelator BAPTA. This argues for an alteration of Ca 2+ -uptake into and/or Ca 2+ -release from stores. An as yet unknown negative regulation of Ca 2+ -sequestration could cause accumulation of Ca 2+ in the cytosol; until now, however, activation of SERCA-type Ca 2+ -ATPases was found only (for review see [19]). On the other hand, release of Ca 2+ could have been augmented by the high basal cAMP level in the mutant. cAMP-dependent phosphorylation of the IP 3 -receptor by PKA results in increased sensitivity for IP 3 in pancreatic acinar cells [25]; the same holds true for the ryanodine receptor [19]. Stimulation of PKA activity is plausible since pretreatment with the PKA-activator Sp-5,6-DCl-cBIMPS elevated basal [Ca 2+ ] i and reduced agonist-evoked Ca 2+ -entry. Membrane permeable Sp-5,6-DCl-cBIMPS was shown to be virtually ineffective in inducing gene expression and to be highly selective for PKA vs cAMP receptor activation at the concentration employed [21]. In summary, we propose the following model: in the mutant sensitivity of the Ca 2+ -release system is enhanced by an augmented PKA-mediated phosphorylation which is due to increased basal cAMP levels. This results in larger amounts of Ca 2+ being liberated upon stimulation. In Dictyostelium release of Ca 2+ from stores was also found after addition of calmidazolium [26] which was shown to inhibit calmodulin-dependent and independent activity of calcineurin [27]. Calcineurin in turn was proposed to be responsible for termination of Ca 2+ -release by dephosphorylating the IP 3 -receptor [28]. In regAaugmented release of Ca 2+ leads to a reduction of Ca 2+ -entry across the plasma membrane as a negative feedback.
We suggest the alteration in [Ca 2+ ] i to be responsible for the irregular light scattering and extracellular [Ca 2+ ]-oscil-lations of regAcells. Previously, Wessels et al. [14] have shown that the mutant cannot propagate a cAMP wave since wild type amoebae no longer aggregated correctly when mixed with mutant cells. Indeed, we found that peak cAMP levels during light scattering oscillations were very low in regAas compared to wild type. This effect is plausible, as the increased sensitivity of the Ca 2+ second messenger system exerts a negative feedback on cAMP synthesis: large [Ca 2+ ] i -transients inhibit production of cAMP [3]. An interplay of cAMP and [Ca 2+ ] i -oscillations and their mutual dependence has also been shown in neurons: absence of either, cAMP or [Ca 2+ ] i -oscillations resulted in failure of the other component to oscillate [8].
In Dictyostelium the strong decrease in peak cAMP oscillation levels affected [Ca 2+ ] e -oscillations which were irregular. The basis is probably an influence on [Ca 2+ ] ioscillations. Such oscillations were suggested to occur but have not been demonstrated in single cells until now, presumably due to the small size of the amoebae and the characteristics of the wave [29].
With respect to chemotaxis, reduced suppression of lateral pseudopod formation was shown in regAcells and an essential role of RegA for a correct response in a natural cAMP wave and chemotactic migration was assigned [14]; subsequently, a similar result was found in a mutant expressing a constitutively active PKA [30]. When we analyzed chemotaxis towards a cAMP-filled glass capillary we observed the same behaviour as described by Wessels et al. [14]. In principle, it is possible that the reduced capacity of regAcells to polarize was due to a difference in the developmental stage as compared to wild type cells. However, regAdevelops much faster than wild type which suggests an even more efficient chemotaxis as this response increases during differentiation to aggregation competence. Alternatively, an altered or dampened signaling response caused by a lower number of cAMP receptors present on the cell surface could have caused the reduced chemotactic response. We consider this to be unlikely for the following reason. Aggregation-competent Dictyostelium amoebae possess roughly 50.000 cAMP receptors at the cell surface [31]. Yet, for chemotactic orientation and polarization in a cAMP gradient the difference in receptor occupancy between the front and the rear end of the amoebae is important rather than the absolute number of stimulated receptors [31]. So even if regAexpressed less receptors than wild type this should not influence the accuracy of the response. We propose the reduced polarization capacity of regAamoebae to be caused by their altered [Ca 2+ ] i -regulation. In the mutant strain the threshold for generation of an agonist-induced [Ca 2+ ] i -increase is lower than in wild type. The [Ca 2+ ] i -elevation is not as tightly controlled and occurs even in the presence of BAPTA. The characteristics of a [Ca 2+ ] iincrease are important for the resulting cytoskeletal Chemotactic speed of wild type and regAamoebae Figure 8 Chemotactic speed of wild type and regAamoebae. The effect of preincubation with EGTA for 30 min was assayed. Chemotactic velocity of amoebae was affected dose dependently by EGTA treatment; when compared to the wild type the speed of the regAstrain was significantly reduced at lower concentrations of EGTA. Velocity of wild type and regAcells is shown (median of at least 2 independent experiments).
rearrangements and whether pseudopods are formed correctly. Indeed, artificial induction of a small global [Ca 2+ ] i -transient by incubation with calmidazolium caused overall pseudopod protrusion [26]. In migrating cells the establishment of a [Ca 2+ ] i -gradient at the rear end was shown [5,32] [33].
When we analyzed the influence of pretreatment with EGTA on chemotactic behaviour of wild type and regAcells we found that the mutant was more sensitive. When compared to wild type, lower doses of EGTA were sufficient to reduce chemotactic speed. The effect of EGTA treatment is most probably due to emptying of the storage compartments [34]; the presence or absence of extracellular Ca 2+ affects the Ca 2+ -content of stores [35,36]. RegAcells are more sensitive than wild type amoebae because of the lower threshold for Ca 2+ release and thus a more rapid depletion of Ca 2+ in the cells.
Cell culture D. discoideum axenic wild type Ax2 was grown as described [4]; the mutant regA -(kindly provided by Dr. P. Thomason) was grown in the presence of blasticidinS. Cells were washed by repeated centrifugation and resuspension of the cell pellet in cold Sørensen phosphate buffer (17 mM Na + /K + -phosphate, pH 6.0; SP-buffer). Amoebae were shaken at 2 × 10 7 cells/ml, 150 rpm and 23°C until use. The time, in hours, after induction of development is designated t x .

Recording of light scattering
At t 2.5 -t 4 2 ml of cell suspension was pipetted into cuvettes and aerated. Light scattering oscillations were recorded at 500 nm with a photometer as described [4].

Determination of cAMP
The total amount of cAMP was determined using the cAMP enzyme immuno assay (Biotrak, Amersham Pharmacia Biotech, Freiburg, FRG) according to the manufacturer's instructions. Samples were prepared as outlined previously [4].
Washed cells (2-5 µl) were placed on glass coverslips and incubated in a humid chamber until use. When experiments were done in nominally Ca 2+ -free medium, 85-88 µl of H5-buffer was added 1 min before the [Ca 2+ ]-imaging experiment. To test the response of amoebae in the presence of BAPTA, 75-78 µl of H5-buffer was pipetted to the cells; 10 µl of 10 mM BAPTA was added during the [Ca 2+ ]-imaging experiment and 10-12 sec later cAMP was given. When the response of cells was to be analyzed in the presence of extracellular CaCl 2 , H5-buffer (85-88 µl) with 1 mM CaCl 2 was added to the cells 15 min before the [Ca 2+ ]-imaging experiment to load stores (see also [18]). cAMP-stimulation was done by adding 10 µl of 10 µM cAMP (± 1 mM CaCl 2 ) to the cells. To load cells with IBMX, they were electroporated with fura2-dextran in the presence of 250 µM of the inhibitor. The cytosolic concentration of IBMX is in the range of maximally 2-5% of the concentration present during electroporation [6]. Measurement of regAwas done at t 4 and wild type [Ca 2+ ]imaging was done at t 7-8 . In another series of experiments we treated regAcells with Sp-5,6-DCl-cBIMPS, a membrane permeant activator of PKA [20]. Incubation was done with 37 µM of the activator for 60 min prior to the [Ca 2+ ]-imaging experiment.

Chemotaxis of regAcells
Chemotactic performance of the amoebae depends on the degree of differentiation, so their shape was checked prior to the chemotaxis assay. 200 µl of cells at 2 × 10 7 cells/ml were placed on a coverslip and allowed to settle for at least 30 min. The morphology of the cells was controlled microscopically: when elongated and thus aggregation competent cells were present, an aliquot of cells from the suspension was diluted for the chemotaxis assay. RegAwas tested at t 4 -t 5 , wild type was measured at t 7 -t 10 . 250 µl of cells in 5 mM Hepes, pH 7.0 (1 × 10 5 cells/ml) were placed on glass coverslips. After 30 min cells were challenged with a cAMP (100 µM) filled glass capillary and chemotaxis was recorded for 40-45 min either on vidoe tape or images were stored directly on a hard disk. In addition, experiments were done with cells incubated with 2-10 mM EGTA for 30 min to empty Ca 2+ -storage compartments. Analysis of chemotaxis was done as outlined previously [34].