Dictyostelium discoideum mutant Synag 7 with altered g-protein–adenylate cyclase interaction

The receptor-linked adenylate cyclase system in Dictyostelium discoideum provides a useful model for comparison with the hormone and neurotransmitter-regulated adenylate cyclase systems in vertebrates. In the cellular slime mould D. discoideum, cyclic AMP functions as a hormone-like signal during chemotaxis (Konijn, 1970), morphogenesis (Schaap et al. 1984), and cell differentiation (Kay, 1982). Extracellular cyclic AMP is detected by cell surface cyclic AMP receptors (Devreotes, 1983), and results in several intracellular responses, including the activation of guanylate and adenylate cyclase (Gerisch, 1987; Janssens & Van Haastert, 1987), phospholipase C (Europe-Finner & Newell, 1987), and the phosphorylation of cyclic AMP receptors (Klein et al. 1985, 1987). Intracellular cyclic GMP reaches a peak at 10 s after stimulation and is possibly involved in cyclic AMP-induced chemotactic movement. Adenylate cyclase activity increases at about 10–30 s after cyclic AMP addition, and reaches a maximal level at 60–120 s. The cyclic AMP produced is secreted, and triggers more distal cells, thus relaying the signal (Van Haastert, 1984”). Surrounding cells respond to the cyclic AMP signal by chemotaxis and by secreting a pulse of cyclic AMP themselves (see reviews, Devreotes, 1983; Gerisch, 1987; Janssens & Van Haastert, 1987).

Several lines of evidence in D. discoideum have previously suggested that surface cyclic AMP receptors are coupled to intracellular effectors via G-proteins. Binding of cyclic AMP is complex, showing interconversions of binding states in vivo (Van Haastert & De Wit, 1984; Van Haastert, 1985; Van Haastert et al. 1986), which are promoted by guanine nucleotides in vitro (Van Haastert, 19846; Van Haastert et al. 1986). Alternatively, cyclic AMP increases [³H]GTP binding to isolated membranes, at the same time accelerating the dissociation rate of GTP (De Wit & Snaar-Jagalska, 1985). In addition, receptor agonists stimulate high-affinity GTPase activity (Snaar-Jagalska et al. 19886). Recently, it has been shown that GTP stimulates the formation of myo-inositol 1,4,5-trisphosphate (Europe-Finner & Newell, 1987). Also adenylate cyclase can be stimulated (Theibert & Devreotes, 1986), and inhibited (Van Haastert et al. 1987) by GTP. However, stimulation by GTP and its analogues can only be observed in cell homogenates shortly after lysis, or in membranes at temperatures between 0 and 10°C and in the presence of a cytosolic factor (Theibert & Devreotes, 1986). Preincubation of membranes under phosphorylation conditions converted stimulation of adenylate cyclase by guanine nucleotides to inhibition (Van Haastert et al. 1987). Inhibition of adenylate cyclase by GTP is blocked by pretreatment of cells with pertussis toxin (Van Haastert et al. 1987). Moreover, desensitization to persistent cyclic AMP stimuli is blocked after treatment with pertussis toxin in vivo (Snaar-Jagalska & Van Haastert, unpublished results). These observations suggest that adenylate cyclase regulation by G-protein, while present in D. discoideum, might operate somewhat differently from the way it does in vertebrates.

In vertebrate cells, surface receptor-linked adenylate cyclase systems were successfully investigated in S49 lymphoma cells, deficient or defective in stimulated G-protein (Gs). An aggregation-defective mutant of D. discoideum was isolated (Frantz, 1980). In this mutant, designated synag 7 or N7, cyclic AMP and/or guanine nucleotides do not activate adenylate cyclase (Theibert & Devreotes, 1986; Van Haastert et al. 1987). The wild-type pattern of guanosine nucleotide regulation was restored by the addition of a high-speed supernatant from wild-type cells. In the present study the interaction of surface cyclic AMP receptors with G-proteins was investigated in membranes from the mutant and in the presence of wild-type supernatant. Addition of this supernatant restored the low GTP binding and GTPase activity in synag 7 membranes. These results suggest that the supernatant factor absent in the mutant may represent a new molecule required for receptor- and G-protein-mediated activation of adenylate cyclase.

[2,8-³H]cyclic AMP (1·5 TBq mmol⁻¹), [8-³H]GTP (sodium salt; 10·6 Ci mmol⁻¹) were obtained from Amersham. [γ-³²P]GTP (37·94 Ci mmol⁻¹) was purchased from New England Nuclear, cyclic AMP, ATP, ATPγS, AppNHp, GTP, creatine phosphate and creatine kinase were obtained from Boehringer-Mannheim. Dithiothreitol (DTT) and bovine serum albumin (BSA) were from Sigma.

The synag 7 mutant of wild-type D. discoideum NC-4 has been characterized by Dr Frantz (1980) and was kindly provided by Dr P. N. Devreotes. Wild-type and mutant cells were grown as described (Van Haastert & Van der Heijden, 1983), harvested in 10mM-Na/K phosphate buffer (PB), pH 6·5, washed and starved in PB by shaking at a density of 10⁷ cells ml⁻¹. During starvation of mutant cells pulses of cyclic AMP were given at 6-min intervals and at a concentration of 10⁻⁷M. After 5·6 h, cells were collected by centrifugation, washed twice with PB, and the pellet was resuspended in 40 mM-Hepes/NaOH, 0·5mM-EDTA, 250mM-sucrose, pH 7·7, to a density of 2×10⁸ cells ml⁻¹. Crude membranes were prepared by pressing the cell suspension through a Nuclepore filter (pore size 3 gM) at 0°C; if it is not otherwise indicated, the membranes were washed once and finally resuspended in PB to the equivalent of 1×10⁸ cells ml⁻¹.

The association of [³H]cyclic AMP with D. discoideum cells at 20°C was detected in a volume of 100 μl containing PB, 10 mM-DTT, 30nM-[³H]cyclic AMP, and 8×10⁶ cells. At the times indicated the cells were separated from the extracellular medium by centrifugation through silicon oil as described (Van Haastert & De Wit, 1984c). The dissociation of bound [³H]cyclic AMP was measured after equilibration of cells with 2 nM-[³H]cyclic AMP for 45 s at 20°C. Dissociation at t = 0 was induced by the addition of 0·lmM-cyclic AMP. Inhibition of cyclic AMP binding to membranes was measured at 20°C in an incubation volume of 100 μl containing 80 μl membranes, 2nM-[³H]cyclic AMP and different concentrations of guanine nucleotides. The incubation period was 75 s, followed by centrifugation through silicon oil. Non-specific binding was determined by including 0·1 mM-cyclic AMP in the incubation mixture and was subtracted from all data shown.

Binding of [³H]GTP to membranes was measured at 0°C in an incubation volume of 100gl containing 10mM-Tris-HCl, pH8·0, 100n_M-[³H]GTP, 1 mM-ATP, 5mM-MgCl₂, and 80gl membranes. Bound [³H]GTP was separated from free [³H]GTP by centrifugation through silicon oil at 10000g for 30 s. Non-specific binding was determined in the presence of 0·1 mM unlabelled GTP.

GTPase was determined as described (Snaar-Jagalska et al. 1988b). The reaction mixture was preincubated at 25 °C for 5 min and contained [γ-^3ZP]GTP (0·1 μCi/assay), 2 mM-MgCl₂, 0·lmM-EGTA, 0·2mM-AppNHp, 0·lmM-ATPyS, 10mM-DTT, 5 mM-creatine phosphate (Tris salt), 0·4mgml⁻¹ creatine kinase and 2 mg ml⁻¹ BSA in 50 mM-triethanolamine · HC1, pH 7·4, in a total volume of 100 μl. The reaction was initiated by the addition of 30 μl of membranes to 70 μl of the reaction mixture and conducted for 3 min. The reaction was terminated by the addition of 0·5 ml sodium phosphate buffer (50 mM), pH 2·0, containing 5% (w/v) activated charcoal. The reaction tubes were centrifuged for 5 min at 10000g at 4°C and the radioactivity of the supernatant was determined using Čerenkov radiation.

D. discoideum and synag 7 cells were grown as described above. Cells were resuspended to 1×10⁸ cells ml⁻¹ at 0°C in 10mM-Tris-HCl, pH 8·0, 1 mM-MgSC₄, 0·2MM-EDTA, 200MM-SU-crose, and lysed by pressing through a Nucle pore filter. The lysate was centrifuged at 10000g for 10 min. The supernatant was removed, centrifuged at 10000g for 10min and used for reconstitution assays.

NC-4 and synag 7 membranes were prepared as described above (‘Culture conditions and membrane isolation’), washed and resuspended in the lysis buffer: 10 mM-Tris · HC1, pH 8·0, 1 mM-MgSO₄ to a final density equivalent to 10⁸ cells ml⁻¹.

For reconstitution of GTP-binding, a mixture of supernatant and membranes (1:1) was incubated at 0°C for 10 min, then added (80μl) to GTP binding mixture. After 2 min, incubation samples were centrifuged for 3 min at 10 000gat 2·4°C, and the supernatant was aspirated. The pellet was dissolved in 80 μl 1 M-acetic acid, 1-2ml Emulsifier was added and radioactivity was determined.

For reconstitution of GTPase, the supernatant and membrane mixture was washed after 10 min incubation, resuspended in 10 mM-triethanolamine · HC1, pH 7·4, to the same density and used in the GTPase assay.

The chemotactic response in mutant synag 7 was slightly less sensitive to cyclic AMP than in wild-type cells, but in both cases the concentration of cyclic AMP that induced a response in 50% of the populations was between 10⁻⁹ and 10~⁸M (Table 1). Because synag 7 cells belong to different batches of cells, differences in their growing conditions may affect the chemotactic sensitivity to cyclic AMP. Therefore, we conclude that chemotaxis in synag 7 is not significantly altered. Cells of synag 7 had normal basal cyclic AMP levels (Schaap et al. 1986), but activation of adenylate cyclase by cyclic AMP in vivo was about 30-fold lower than in the parent strain NC-4 and was not improved when the cells were exposed to cyclic AMP pulses during starvation (Schaap et al. 1986). In mutant cells cyclic AMP-induced activation of guanylate cyclase was about 50% lower than in wild-type (Schaap et al. 1986). This response was restored when synag 7 cells were pulsed with 10⁻⁷ M-cyclic AMP during differentiation (Table 1). Furthermore, treatment of mutant cells with 1 μM-cyclic AMP for 15 min induced loss of cyclic AMP binding to 10% of control values. This treatment of cells with cyclic AMP also induced the covalent modification of the receptor, which caused an increase in its apparent molecular weight from 40×10³ to 43×10³M_r (Table 1).

Adenylate cyclase in the mutant could not be activated by cyclic AMP in vivo or by GTPγS in vitro. Inhibition of adenylate cyclase by GTPγS in the mutant membranes was not affected (Van Haastert et al. 1987). The wildtype pattern of adenylate cyclase regulation was restored in synag 7 lysates by the addition of a high-speed supernatant from wild-type cells (Theibert & Devreotes, 1986; Van Haastert et al. 1987). This indicates that the defect of the mutant is probably located in a cytosolic cofactor that is required for guanine-nucleotide-mediated activation of adenylate cyclase in D. discoideum.

In this paper the characteristics of cell surface cyclic AMP receptors, G-protein, and the interaction between cyclic AMP receptors and G-protein are described.

The association of 30 nM-[³H]cyclic AMP to aggregative wild-type and synag 7 cells is shown in Fig. 1. Binding of cyclic AMP to both types of cells rapidly increases, reaching a maximum after about 6 s, and subsequently declines to an apparent equilibrium value approached at about 45–60 s. It has been demonstrated that this decrease in cyclic AMP binding appears to be due to an interconversion of high-affinity to low-affinity binding forms of the receptor (Van Haastert & De Wit, 1984c; Van Haastert, 1985; Van Haastert et al. 1986).

The dissociation of the [³H]cyclic AMP-receptor complex was measured to obtain more information on these binding types (Fig. 2). Wild-type and mutant cells were incubated with [³H]cyclic AMP until equilibrium was reached, and then excess unlabelled cyclic AMP was added. The release of bound cyclic AMP from the mutant and wild-type receptor was identical. These types of experiment demonstrate that D. discoideum cells contain multiple forms of the receptor that have different affinities and rate constants of dissociation (Van Haastert & De Wit, 1984; Van Haastert, 1985; Van Haastert et al. 1986). These forms of receptor most probably reflect the interaction with G-protein (Van Haastert, 19846; Van Haastert et al. 1986).

The inhibition of cyclic AMP binding by guanine nucleotides is shown in Fig. 3. Wild-type and synag 7 membranes were incubated with 2 nM-[³H] cyclic AMP to reach binding equilibrium in the absence and presence of GDPβS, GTPγS (Fig. 3), and GTP, GDP, GppNHp (data not shown). Binding of cyclic AMP was reduced to 35% in the presence of GTPγS and to about 55% in the presence of GDPβS (Fig. 3). The inhibition by GTPγS, GDPβS, GTP, GDP and GppNHp was essentially identical in wild-type and synag 7 membranes. These results indicate that the synag 7 mutant expresses normal levels of surface cyclic AMP receptor, which interact with G-protein as in the wild-type strain.

The kinetics of association of 100nM-[³H]GTP with wild-type and synag 7 membranes are presented in Fig. 4. Binding equilibrium in wild-type membranes was reached within 60s. Analysis of the association rate of [³H]GTP binding (inset) to wild-type membranes indicated fast- and slow-binding types with half-times of association of about 4 s and 23 s, respectively. The binding at equilibrium was enhanced about 29% by 10 μM-cyclic AMP without an obvious change in the association kinetics.

The total [³H]GTP binding to synag 7 membranes was about 47% lower than in control membranes and reached equilibrium after 90s incubation (Fig. 4B). The fast-binding type was absent in the mutant membranes (inset, 4A). However, stimulation of GTP-binding to synag 7 membranes by cyclic AMP had the same relative level as in wild-type membranes (33%).

The kinetics of dissociation of bound [³H]GTP are shown in Fig. 5. Wild-type and synag 7 membranes were incubated with 100nM-[³H]GTP until equilibrium was reached and then excess unlabelled GTP was added. The release of bound [³H]GTP was multiphasic, since a semi-logarithmic plot was non-linear. The rates of dissociation were essentially identical and accelerated by 10 μM-cyclic AMP in wild-type and mutant membranes.

For Scatchard analysis, wild-type and mutant membranes were incubated with different concentrations of [³H]GTP and binding was measured at equilibrium (Fig. 6). The curve for [³H]GTP binding to wild-type membranes was slightly convex, suggesting a complex binding process. These data confirm the complex kinetics of association and dissociation of [³H]GTP (Fig. 5) and the recently described [³⁵S] GTPγS binding by two binding sites with, respectively, high (K_d = 0·2 μM) and low (K_d = 6·3 μM) affinity (Snaar-Jagalska et al. 1988a). However, the high- and low-affinity components for [³H]GTP-binding do not differ much (apparent K_d of about 2qM), making accurate calculations difficult; the total concentration of binding sites in wild-type membranes is 85 nM. Scatchard analysis in synag 7 suggests that the [³H]GTP binding to synag 7 membranes is also heterogeneous. High- and low-affinity components can be easily deduced with apparent K_d values of 0-8 UM and 21 UM, respectively. The total number of binding sites in synag 7 membranes was similar to that in wild-type membranes.

These observations suggest that the synag 7 mutant has altered GTP binding. Reduction of equilibrium binding to mutant membranes is probably caused by loss of the fast-binding type. This defect does not disturb the signal-transduction pathway from cell surface cyclic AMP receptor to a putative G-protein, since cyclic AMP binding to mutant membranes is altered by guanine nucleotides and, alternatively, cyclic AMP increases [³H]GTP binding to isolated membranes, at the same time accelerating the dissociation rate of GTP.

It has been shown that stimulation of adenylate cyclase by guanine nucleotides in D. discoideum membranes required a cytosolic factor. In mutant synag 7 adenylate cyclase was activated by guanine nucleotide in the presence of a cytosolic factor from wild-type only, not from mutant cells (Theibert & Devreotes, 1986; Van Haastert et al. 1987).

Membranes of wild-type and mutant synag 7 were incubated for 10 min at 0°C with PB buffer or supernatants isolated from wild-type or mutant lysates. Subsequently [³H]GTP binding was determined at equilibrium (Table 2). In the presence of PB buffer [³H]GTP binding to mutant membranes was about 40% lower than to wild-type membranes. Supernatant of the mutant slightly (10%) increased binding to both types of membranes. Addition of wild-type supernatant increased [³H]GTP binding to wild-type membranes from 100% to about 157%. Furthermore, wild-type supernatant restored [³H]GTP binding to synag 7 membranes (from 61% to 155%) to the same level. The component of the supernatant that restored GTP binding to synag 7 was inactivated by heating to 60°C for 5 min.

These results suggest that the supernatant factor absent in synag 7 mutant is essential for normal GTP binding.

It has been shown that GTP hydrolysis in D. discoideum membranes is caused by at least two enzymes with high and low affinity and that the high-affinity GTPase is stimulated by cyclic AMP (Snaar-Jagalska et al. 1988b). The hydrolysis of different concentrations of [γ-³²P]GTP by high-affinity GTPase in the absence and presence of 10μM-cyclic AMP is shown in Fig. 7 as an Eadie-Hofstee plot. Basal activity in membranes of the synag 7 mutant measured at 0·01 μM-GTP was about 55% lower than in wild-type membranes. However, the relative stimulation by 10μM-cyclic AMP was only slightly lower in synag 7 (44%) than in wild-type (50%). The stimulatory effect of receptor agonist on GTP hydrolysis by high-affinity GTPase occurred without a change in the V_max value and was apparently caused by an increase in enzyme affinity for GTP from 6·1 |UM to 4·2μM in wild-type and from 7·0 to 3·8,μM in mutant membranes.

Subsequently, the GTPase activity of the synag 7 mutant was investigated under reconstitution conditions with added wild-type supernatant (Table 3). Because the supernatant contains high levels of non-specific nucleoside triphosphatase activity, membranes had to be washed after incubation with supernatant before they could be used in the GTPase assay. The supernatant of either wild-type or synag 7 did not alter the GTPase activity in wild-type membranes. In contrast, the GTPase activity of synag 7 was significantly (P<0·01) increased by preincubation with wild-type supernatant, but not with the synag 7 supernatant. It should be noted that, in contrast to GTP binding, GTPase activity in synag 7 membranes is not completely restored by the wild-type supernatant; washing of the membranes and reversibility of the reconstitution could account for the partial recovery of GTPase activity in synag 7 membranes.

Mutants are valuable tools for investigating transmembrane signal transduction; therefore, the biochemical properties of the D. discoideum mutant synag 7 were characterized. In this aggregation-defective mutant cyclic AMP and/or guanine nucleotides are not sufficient to activate adenylate cyclase (Theibert & Devreotes, 1986; Van Haastert et al. 1987). The wild-type pattern of guanine nucleotide regulation of adenylate cyclase is restored to the synag 7 mutant by the addition of a highspeed supernatant from wild-type cells. Except for the activation of adenylate cyclase, cyclic AMP induces all responses in synag 7 investigated to date (Table 1). synag 7 cells express normal surface cyclic AMP receptors, which are regulated by guanine nucleotides as in wildtype cells. Cyclic AMP induces the down-regulation of the receptors, as well as their covalent modification. Cyclic AMP increases [³H]GTP binding to mutant membranes, and at the same time accelerates the dissociation rate of bound GTP.

However, the association kinetics and equilibrium binding of [³H]GTP are different if compared with those of wild-type membranes. In mutant membranes equilibrium [³H]GTP binding is reduced and a fast-associating component is absent. Also high-affinity GTPase activity is reduced but still stimulated by the receptor agonist. These results suggest that in mutant synag 7 surface cyclic AMP receptors interact normally with G-protein. Considering that the coupling between receptors and adenylate cyclase involves the transduction of the signal via G-protein, the defect in the synag 7 mutant must be localized at the point of interaction between G-protein and adenylate cyclase. Abnormal GTP binding and basal GTPase activity suggest that the defect is caused by an altered activation of G-protein or that the defect is localized at the point of interaction of an active GaGTP with basal adenylate cyclase. Addition of wild-type supernatant to synag 7 membranes completely restored GTP binding and partly restored GTPase activities. These results suggest that this supernatant factor is essential for normal GTP binding, GTPase activity and activation of adenylate cyclase. The reconstitution occurred in the absence of detergent, which indicates that a soluble component is involved. This factor is heat-labile, and has a M_r greater than 20000. It is unlikely that this factor is the stimulatory G-protein or Gsrr, since stimulation of adenylate cyclase by GTPγS is associated with the membrane fraction and not with the cytosol fraction of a cell homogenate (Van Haastert et al. 1987). It is also unlikely that this factor represents the β subunit, because the β subunit is thought to serve as a membrane anchor (Spiegel, 1987). Preincubation of wild-type and mutant membranes under phosphorylation conditions converted stimulation of adenylate cyclase by GTPγS to inhibition. This inhibition was not affected by the cytosolic factor involved in stimulation of adenylate cyclase, but was abolished when membranes were obtained from cells treated with pertussis toxin (Van Haastert et al. 1987). These observations suggest that in wild-type and mutant cells an inhibitory G-protein in addition to a stimulatory Gs is present.

Recently a GTPase-activating protein (GAP) has been described in mammalian cells, which is required for optimal GTPase activity of ras proteins (Trahey & McCormick, 1987). The novel soluble heat-labile factor absent in the synag 7 mutant shows functional homology with GAP, but the identity of these proteins has not been shown. It is tempting to suggest that the supernatant factor absent in mutant synag 7 may belong to a family of proteins that regulate the function of G-proteins.

We thank Peter N. Devreotes for stimulating discussions. This work was supported by the Organisation for Fundamental Research (Medigon) and the C. and C. Huygens Fund, which are subsidized by the Netherlands Organisation for Scientific Research (NWO).