Abstract
It was previously shown that cGMP enhances cAMP-induced Ca2+-influx in Dictyostelium discoideum (Menz, S., Bumann, J., Jaworski, E. and Malchow, D., 1991, J. Cell Sci. 99, 187-191). This finding is based on experiments done with a strain with defective cGMP-hydrolysis, the streamer F cells. In this work we show, that these chemically mutagenized cells display different properties in their cAMP-induced light-scattering response and cAMP-cAMP-induced Ca2+-influx compared with a cGMP-phosphodiesterase knock out strain, pdeD KO, generated by homologous recombination. PdeD KO cells possess a reduced and delayed Ca2+-influx that is developmentally regulated. These findings partly contradict the results of streamer F cells, where cAMP-induced Ca2+-influx is prolonged and elevated. We show by the use of membrane-permeant cGMP-analogues, permeabilized cells and measurement on isolated vesicles that the cause for the reduced Ca2+-influx seems to be due to Ca2+-channel inhibition by cGMP.
Introduction
Asexual reproduction of the social amoebae Dictyostelium discoideum can be switched by starvation to a new life cycle phase whereby approximately 105 single cells begin to differentiate and aggregate in form of large streams. In the end these cells comprise a shared terminal structure called a fruitingbody consisting of a stalk and an encapsulated spore mass. To achieve multicellularity, cells are attracted by chemotaxis to a messenger, cyclic AMP, which is spatially restricted produced and likely secreted at the rear end of the cells (reviewed in (Kimmel et al., 2003)). Detection is mediated by the serpentine receptor cAR1 and its activation launches a signal cascade ultimately leading to the chemotactic movement of the cells. In addition, upon binding of cAMP to cells in suspension a Ca2+-influx occurs after 10 seconds lasting for 40 seconds and a rise in [Ca2+]i that peaks at 20 seconds. (Sonnemann et al., 1998). The elevation of [Ca2+]i requires
the release of Ca2+ from intracellular Ins(1,4,5)P3-sensitive and fatty acid sensitive Ca2+-stores (Schaloske et al., 1998).
Recent work has demonstrated that cGMP is a pivotal second messenger participating in the chemotactic response of the individual amoeba (reviewed by (Bosgraaf et al., 2002b)). Proteins that are involved in cGMP-signaling have been cloned and analyzed. Until now, three cGMP-hydrolyzing phosphodiesterases, PDE 3, PDE 5 (previously called PdeD / GbpA) and PDE 6 (formerly PdeE / GbpB), with distinct expression patterns establish the regulation of cGMP levels during development (Bosgraaf et al., 2002a; Goldberg et al., 2002; Meima et al., 2002; Meima et al., 2003).
Work in our laboratory aims to characterize the contribution of cGMP to maintenance of Ca2+ homeostasis. The cAMP-induced Ca2+-influx originates after a cGMP transient that reaches its maximum at 10 seconds, while the basal cGMP level is restored after 30 seconds. Streamer F mutants generated by chemical mutagenesis exhibited a dominant streaming phenotype during aggregation accompanied by prolonged and elevated cGMP transients (Ross et al., 1979;
Ross et al., 1981). Cyclic AMP-induced Ca2+-influx was likewise altered (Kuwayama et al., 1998; Menz et al., 1991). In addition, aequorin expressing streamer F cells starved for 4 hours displayed a delayed, elevated and prolonged rise in [Ca2+]i upon stimulation with cAMP (Nebl et al., 2002). These results suggested that cGMP enhances Ca2+-influx.
The altered cGMP transient in Streamer F cells is due to a disturbed cGMP-hydrolysis activity which is reduced by 80% in strain NP368 (Kuwayama et al., 2001). In wt cells PDE 5 is the major cGMP-hydrolyzing enzyme, while PDE 3 and PDE 6 have only a minor influence on regulating cGMP levels (Bosgraaf et al., 2002a). Recently, mutants deficient in PDE 5 have been generated by a single gene targeting strategy independently in two laboratories. The KO cells showed a prolonged and elevated cGMP-response similar to streamer F cells (Bosgraaf et al., 2002a; Meima et al., 2002). Due to the chemical mutagenesis generating streamer F cells it remains unclear whether Ca2+-influx entirely depends on cGMP in these cells or is a consequence of additional mutations.
Moreover, the regulation of Ca2+-influx by cGMP in wildtype cells needs to be elucidated. We have previously shown that Ca2+-influx can be increased in wt cells using the cGMP analog Sp-8-Br-cGMPS but its specificity was called into question
(Butt et al., 1990; Flaadt et al., 1993). Altogether this prompted us to study one of the PDE 5 KO mutants, pdeD KO, expressed in wt Ax2 background, with respect to its Ca2+-responses, a possible correlation with cGMP and to compare the results with those obtained from the streamer F mutants.
Material and Methods
8-pCPT-cGMP was provided by Biolog, Bremen (Germany); Filipin and arachidonic acid were obtained from Sigma, Munich (Germany); 3’, 5’-cyclic AMP was obtained from Roche, Mannheim (Germany). CaCl2 was bought from Fluka, Buchs (Switzerland), cGMP enzyme immuno assays were bought from Amersham, Freiburg (Germany)
Cultivation of Dictyostelium discoideum cells
Cells were grown at 23°C in axenic medium as described (Schaloske et al., 1997).
Medium for growing pdeD KO cells was supplemented with 10 µg/ml blasticidin.
Streamer F cells (strain NP368) and parental strain XP55 were cultured on 1/3 SM-Agar in association with Klebsiella planticola.
Differentiation was induced by washing the cells two times with ice-cold Sørensen phosphate (SP) buffer. Streamer F cells and XP 55 were separated from remaining bacteria by repeated washing with ice-cold SP buffer. Cells were then incubated at a cell density of 2*107 cells/ml.
Light-scattering recordings of cells in suspension
Light-scattering measurements of cells were done as described by Gerisch and Hess (Gerisch et al., 1974). The extinction of a cell suspension (2*107 cells/ml in SP buffer) aerated in a cuvette was monitored at 500 nm with a Zeiss PM6 spectrophotometer. Cyclic AMP pulses (0.1 µM) were applied to a cell suspension (2*107 cells/ml) every 6 minutes while monitoring light-scattering changes. 8-pCPT-cGMP was added after three cAMP control pulses 1-2 hours before the onset of spike-shaped oscillations. For every pulse, the magnitude of the first and second light scattering change was determined.
Measurement of cAMP-induced Ca2+ -influx
Net Ca2+-fluxes were measured after cAMP stimulation with a Ca2+-sensitive electrode in a cell suspension at 23°C as described previously (Bumann et al., 1984). In brief, 5*107 cells were incubated in nominally Ca2+ free buffer (Tricine 5
mM pH 7.0, supplemented with 5 mM KCl). Electrode potentials were recorded with a pH meter (Metrohm, Herisau, Switzerland). Extracellular Ca2+
concentrations were determined with a calibration curve from known standards.
Ca2+-flux was analysed 2.5-8 hours after induction of starvation. Cyclic AMP was applied in 5 minutes intervals unless specified otherwise. Permeabilization with the antibiotic filipin was done as described by Flaadt et al. (Flaadt et al., 1993).
Efficient permeabilization of the cells was achieved with 15-20 µg/ml filipin monitored by methylen blue staining.
Preparation of vesicle containing cell extracts
Cells differentiated for 2 hours were washed three times in ice-cold Hepes buffer (20 mM, pH 7.2) and resuspended at a density of 1,5*108 cells/ml. Cells were lyzed by passage through a nuclepore filter (5 µm) and collected in buffer containing 3% (w/v) sucrose, 50 mM KCl, 1 mM MgCl2, 20 µg/ml leupeptin, 1 µg/ml aprotinin and 2.5 mM DTT. Extracts were handled as described previously (Schaloske et al., 2000)
Determination of cGMP concentration
Cells stirred in buffer consisting of 5 mM Tricin-buffer and 5 mM KCl (pH 7.0) were stimulated with 0.1 µM cAMP 3 to 4.5 hours after starvation. The extracellular Ca2+-concentration was adjusted to 2 µM and cAMP-induced Ca2+-influx was monitored in parallel. Samples were taken before and after the addition of cAMP.
Subsequently, samples were treated as described (Wurster B., 1977). Total cGMP concentration was measured using an enzyme immuno assay specific for cGMP (Amersham, Freiburg).
Statistical analysis
Data were analyzed using SigmaStat®. Data are expressed as means ± S.D. if not indicated otherwise. Significance was tested with t-test and Mann Whitney Rank Sum test. An outcome of p < 0.05 was considered to be significant, when control or wt values were compared to mutant or test condition.
Results
Comparison of Streamer F mutants and pdeD KO
It was demonstrated by Menz et al. for the first time that streamer F mutants show a prolonged and elevated increase of cAMP-induced Ca2+-influx (Menz et al., 1991). Moreover, for these mutants, an altered light-scattering response of cells in suspension to a stimulus of cAMP was described by Newell et al. (Ross et al., 1981). Ax2 cells and parental strain XP55 respond with a decrease of optical density displaying a rapid first peak and a slower longer lasting second peak as depicted in Figure 8.
Figure 8: Light-scattering responses of Dictyostelium discoideum wt and mutants defective in cGMP-hydrolysis. Cells in suspension developed for 8 hours were stimulated with 0.1 µM cAMP as indicated by an arrow. First and second peaks are denoted by 1 and 2 respectively. One out of at least 3 independent stimulations for every strain is shown.
The second peak of a light-scattering response was correlated with cAMP-induced changes in cGMP concentration and is altered in streamer F cells. We confirmed the aberrant light-scattering response of streamer F mutant strain NP368 (Figure 8). Termination of the first phase took longer and subsequently extinction fell below baseline. The second peak was extremely prolonged as compared to the parental strain XP55. By contrast, a mutant defective in the gene, PdeD, coding for the cGMP-specific phosphodiesterase whose activity is assumed to be missing
in streamer F mutants (Meima et al., 2002), displayed a prolonged and elevated second peak of the cAMP-induced light-scattering response when compared to Ax2 and to streamer F mutants, respectively. The first peak, however, was similar compared to Ax2. NP368 and pdeD KO cells clearly displayed different light scattering behavior indicating that there must be differences between the cell lines beyond the lack of the PDE activity.
Figure 9 shows that cAMP-induced Ca2+-influx of pdeD KO was characterized by a delayed onset of influx and an reduced net amount of Ca2+ taken up with regard to Ax2.
Figure 9: Change of [Ca2+]e after cAMP-stimulation in wt and pdeD KO. Cells in suspension were developed for about 4 hours and stimulated with 0.1 µM cAMP as indicated by an arrow. [Ca2+]e
and Ca2+ control pulses are displayed. One out of 25 independent experiments is shown.
Strikingly, the delay of the onset of Ca2+-influx was very pronounced when compared to Ax2, but reminiscent of streamer F cells (Menz et al., 1991). The net influx was just the opposite of streamer F mutants: reduced instead of enhanced.
We conclude from these experiments that streamer F cells and pdeD KO mutants are different with respect to Ca2+-influx and the light-scattering response.
Secondly, pdeD KO cells provided further evidence that cGMP is linked to cAMP-induced Ca2+-influx.
Dependence of Ca2+-influx on extracellular Ca2+ and cAMP of pdeD KO cells
The cAMP-induced Ca2+-influx of wt cells depends on the amount of Ca2+ present in the buffer and on the concentration of the cAMP stimulus (Figure 10 and Figure 11). Ca2+-influx increases with elevation of the concentration of both parameters (Bumann et al., 1984). However, the Ca2+ taken up was significantly lower in pdeD KO. This difference in the amount of Ca2+ entry was more pronounced at higher [Ca2+]e (Figure 10).
Figure 10: Dose-response curve for the dependence of Ca2+-influx on [Ca2+]e of pdeD KO and wt.
Ca2+-influx was measured as described in material and methods. Cells were stimulated with 0.1 µM cAMP every 5 minutes and [Ca2+]e adjusted. Influx of wt and pdeD KO cells was recorded between 3 hours and 4.5 hours. Data represent one out of several routinely conducted experiments of wt cells confirming the results of Bumann et al.(Bumann et al., 1984). The data of three independent experiments for pdeD KO are shown.
Saturation of influx was reached at lower [Ca2+]e concentrations in pdeD KO.
When measured at the same high [Ca2+]e, maximal Ca2+-influx amounted to 45 ±
Extracellular Ca2+ (µM)
0 5 10 15 20 25 30
Ca2+ -influx (pmol/10
7 ce
lls)
0 50 100 150 200 250 300
pdeD KO pdeD KO pdeD KO Ax2
13 pmol/107 cells for pdeD KO and 259 ± 19 pmol/107 cells for Ax2 cells stimulated with 0.1 µM cAMP (p < 0,001; n ≥ 3).
Next we asked whether the cells possess the same sensitivity to cAMP. The mutant cells were still sensitive to low cAMP-concentrations as is shown in Figure 11.
Figure 11: Dose-response curve of Ca2+-influx for cAMP in pdeD KO and wt cells. Cyclic AMP stimuli were applied at [Ca2+]e concentrations between 2-3 µM. The asterisk denotes influx that was significantly different between wt and pdeD KO (p0.1 µM < 0.007, p1 µM < 0.013, p10 µM < 0.017, n ≥ 4 for pdeD KO). Wt measurements (n ≥ 3) confirm data of Bumann et al. (Bumann et al., 1984).
Cells responded to 1 nM cAMP even slightly better than wt cells. At a high cAMP concentration of 10 µM the maximum influx amounted to 54 ± 19 pmol/107 cells compared to 142 ± 43 pmol/107 cells of the wt at 4-5 hours of development. Since the [Ca2+]e was adjusted to 2-3 µM, the difference in net influx between wt and mutant was not as pronounced as in Figure 10. The mutant possess the same sensitivity to cAMP as the wt, but the regulation of Ca2+-influx by Ca2+ is strongly reduced.
cAMP (nM) Ca2+ -influx (pmol/107 cells)
0 50 100 150 200 250
pdeD KO Ax2
1 10 100 1000 10000
*
*
*
Developmental regulation of Ca2+-influx
Receptor-mediated Ca2+-influx is developmentally regulated (Bumann et al., 1984). A defective regulation could account for the observed differences in Ca2+ -influx. In wt cells the maximum net influx and the maximum rate of influx occurred at 6 hours of starvation. However, in pdeD deficient cells Ca2+-influx was maximal 4-5 hours of development and remained constant thereafter, as shown in Figure 5a.
Legend see next page
Time after induction of development (hours) Rate of Ca2+ -influx (pmol Ca2+ /107 cells/minute)
0 20 40 60 80 100
Time after induction of development (hours) Ca2+ -influx (pmol/107 cells)
0 10 20 30 40 50
6 7
3 4-5
3 4-5 6 7
*
*
*
*
*
* a
b
Figure 12 (previous page): Ca2+-influx is developmentally regulated in pdeD KO. In a) the net Ca2+-influx and in b) the rate of Ca2+-influx is shown. Stimulation of the cells was conducted with 0.1 µM cAMP at different times after starvation. [Ca2+]e amounted to 2-3 µM (n ≥ 5, mean ± S.E.M.).
The increase from 3 hours to 4-5 hours (p < 0.004), 3 hours to 6 and 7 hours (p < 0.001) of the net influx and from 3 hours to 4-5 hours (p < 0.017), 3 to 6 hours (p < 0.04) and 3 to 7 hours (p <
0.024) of the rate was significant as indicated by asterisk.
By contrast, the rate of Ca2+-influx rose up to 6 hours as depicted in Figure 5b, but did not decline substantially as reported for Ax2. Thus the extent of the divergence in Ca2+-influx of Ax2 from pdeD KO is less pronounced later in development.
These data demonstrate that the amount and rate of cAMP-induced Ca2+-influx in pdeD KO cells was less dependent on developmental time than in the wt.
The delay of Ca2+-influx depends on developmental time
As shown in Figure 9 the onset of pdeD KO Ca2+-influx was delayed as compared to Ax2. The onset was retarded to 44 ± 8 seconds after 3 hours after starvation, but it was accelerated significantly during development to 12 ± 7 seconds (n = 7), a value close to the one obtained with wt cells (Figure 13).
Figure 13: Developmental regulation of the time of onset of cAMP-induced Ca2+-influx. The delay was significantly different between 3 hours and 7 hours and between 3 hours and wt (n ≥ 6, p <
0.001).
In some experiments this shift appeared as early as 5 hours (data not shown).
Interestingly, the onset of the Ca2+-influx of streamer F mutants after cAMP-stimulation was also developmentally regulated. In XP55 the time until onset amounted to 12 ± 3 seconds (n = 14) and in streamer F cells to 34 ± 9 seconds (n
Time after induction of development (hours)
Time to onset of Ca2+-influx (seconds)
0 10 20 30 40 50 60
3 7
= 21) at 4 to 5 hours and returned to wt values at 8 hours of development (Menz, 1991).
However, a large cAMP stimulus overcame the delay. With 10 µM cAMP Ca2+ -influx started after 12 ± 3 seconds (n = 6) at 4 hours. This result suggests that the activation of unknown components of the influx pathway, possibly a second channel, is developmentally regulated and can be circumvented by a strong cAMP stimulus. Thus, the contribution of cGMP to regulation of Ca2+-influx depends on the stage of development.
Relationship of cAMP-induced Ca2+-influx and cGMP transient in pdeD KO
We measured the cGMP concentration and cAMP-induced Ca2+-influx in parallel to investigate whether the cGMP transient corresponds to the delayed onset of Ca2+ -influx. We confirmed the results of Meima et al. (Meima et al., 2002) that the cGMP concentration was elevated and prolonged after cAMP-stimulation as compared to Ax2 (Figure 14).
Figure 14: Time course of cAMP stimulated cGMP transient and Ca2+-influx in pdeD KO cells.
Experiments were conducted as described in material and methods. [Ca2+]e was adjusted to 2µM.
The addition of the stimulus is indicated. One out of 4 independent experiments is shown.
In 3 out of 4 experiments the maximum cGMP was reached after 30 - 70 seconds just before Ca2+-influx occurred. In the remaining experiment the cGMP peak occurred slightly after Ca2+-influx had terminated. The maximum cGMP concentration amounted to 72 ± 23 pmol/107 cells (n = 4) as compared to wt 21 ± 13 pmol/107 cells (n = 5). The decrease in cGMP-concentration was slow and accompanied Ca2+-influx. Basal levels of cGMP were restored and Ca2+-efflux terminated after about 200 seconds. These data show that cAMP-induced Ca2+ -influx is negatively regulated by cGMP causing a delay of the onset.
cGMP could act through the following mechanism: 1) by release of additional Ca2+
from intracellular stores 2) by inhibition of Ca2+-uptake into the stores 3 ) by inhibition of a Ca2+ channel at the plasma membrane and 4) by activation of a Ca2+
plasma membrane Ca2+-ATPase. However, considering the finding that there was no additional Ca2+ release after cAMP addition to intact pdeD KO cells until the onset of influx (Figure 9, 14, 17) hypothesis 4 seems unlikely to be the cause for the observed delay in Ca2+-influx.
Determination of Ca2+-release in permeabilized pdeD KO cells
To investigate the first and second hypothesis we permeabilized pdeD KO cells and stimulated subsequently with cAMP. Under these conditions we measure Ca2+-fluxes across intracellular membranes. Inhibition of Ca2+-uptake and/or additional Ca2+-release by cGMP should cause an increase of cytosolic Ca2+ that should be released into the medium (Schaloske et al., 2000). Figure 15 (next page) shows that the cells are still able to release Ca2+ from the stores in response to cAMP.
The amount of Ca2+-released by the mutant was not significantly different with 19 ± 6 pmol/107cells (n = 7) compared to 12 ± 5 pmol/107 cells (n = 15) by the wt (Flaadt et al., 1993). We conclude that cGMP does not cause an increased cytosolic Ca2+-concentration.
Figure 15: Ca2+-release from internal stores in the pdeD KO cells after cAMP stimulation. Cells developed for 3 to 4 hours were permeabilized as described in material and methods and stimulated with 0.1 µM cAMP as indicated. One out of 5 experiments is shown.
Measurement of Ca2+ uptake and release in vesicular preparations
Uptake into Ca2+-stores in vesicle preparations of pdeD KO cells remained unchanged. In addition, in vesicle preparations of wt cells no effects could be observed on uptake upon addition of cGMP or cGMP analogues. Cyclic GMP did not elicit Ca2+-release neither in wt nor pdeD KO vesicles under our experimental conditions (data not shown). We conclude that cGMP does not influence Ca2+ -fluxes in vitro and that the Ca2+-pump activity is not altered in the mutant.
Effect of 8-pCPT-cGMP on cAMP-induced Ca2+-influx of Ax2 and measurement of arachidonic acid-induced Ca2+-influx
As hypothesis 1, 2 and 4 are rendered unlikely it remains to be investigated whether cGMP inhibits a plasma membrane channel. 8-pCPT-cGMP is a hydrolysis resistant analogue with high membrane permeability and known to target plasma membrane Ca2+-channels (Wei et al., 1998). We found a time- and dose-dependent inhibition of Ca2+-influx in wt cells by 8-pCPT-cGMP (Figure 16a page 45) with a maximum inhibition at 25 µM by 37% (Fig. 16b). Interaction with the cAMP-receptor can be excluded since 75 µM of 8-pCPT-cGMP did not cause a reduction in the cAMP (0.1µM)-induced first light-scattering peak within 12 minutes (data not shown).
Arachidonic acid releases Ca2+ from Ca2+-stores and induces Ca2+-influx across the plasma membrane independently of the cAMP-receptor and without elevating cGMP levels (Schaloske et al. manuscript in preparation). We initiated Ca2+-influx by application of arachidonic acid alone (6 µM) and in combination with cAMP (1 µM). Since cAMP induces cGMP transients a block of Ca2+-influx by cGMP should
Arachidonic acid releases Ca2+ from Ca2+-stores and induces Ca2+-influx across the plasma membrane independently of the cAMP-receptor and without elevating cGMP levels (Schaloske et al. manuscript in preparation). We initiated Ca2+-influx by application of arachidonic acid alone (6 µM) and in combination with cAMP (1 µM). Since cAMP induces cGMP transients a block of Ca2+-influx by cGMP should