Ca2+/Calmodulin Stimulates GTP Binding to the Ras-related Protein Ral-A*

Ral-A is a Ras-related GTP-binding protein that has been suggested to be the downstream target of Ras proteins and is involved in the tyrosine kinase-mediated, Ras-dependent activation of phospholipase D. We reported recently that Ral-A purified from human erythrocyte membrane binds to calmodulin in a Ca2+-dependent manner at a calmodulin binding domain identified near its C-terminal region (Wang, K. L., Khan, M. T., and Roufogalis, B. D. (1997) J. Biol. Chem. 272, 16002–16009). In this study we show the enhancement of GTP binding to Ral-A by Ca2+/calmodulin. The stimulation up to 3-fold by calmodulin was Ca2+-dependent, with half-maximum activation occurring at 180 nm calmodulin and 80 nm free Ca2+ concentration. The present work supports a regulatory role of Ca2+/calmodulin for the activation of Ral-A and suggests a possible direct link between signal transduction pathways of Ca2+/calmodulin and Ral-A proteins.

Calmodulin is a ubiquitous, highly conserved Ca 2ϩ sensor protein that translates the Ca 2ϩ signal into a wide variety of cellular processes. In response to an increase in the intracellular concentration of Ca 2ϩ , calmodulin undergoes a conformational change that results in binding to its target proteins and acts as a trigger to modify their functions. Calmodulin-binding proteins therefore play an important role in intracellular Ca 2ϩ signaling and in various physiological functions, including glycogen metabolism, secretion, muscle contraction, actin/cytoskeletal organization, and cell division (1,2). Ras proteins are integral to signal transduction of extracellular signals to the nucleus, thus regulating a diverse spectrum of intracellular processes (3,4). Accumulating evidence indicates the existence of a direct link between signaling pathways of Ca 2ϩ /calmodulin and Ras proteins. RasGRF, a neuronal Ras-guanine nucleotide releasing factor, has been shown to be activated in response to Ca 2ϩ via direct binding of calmodulin (5). In rat pheochromocytoma PC12 cells, Ca 2ϩ influx through voltage-sensitive Ca 2ϩ channels induces neurite growth via activation of Ras, which in turn activates the serine/threonine protein kinase, Raf (6, 7). Peppelenbosch et al. (8) have reported that epidermal growth factor-induced Ca 2ϩ influx is mediated by Rac proteins. Rapid activation of Rap1 in human platelets has been shown to be mediated by an increased intracellular Ca 2ϩ concentration (9). In addition, an increasing number of Ras-related proteins and Ras protein effectors have been identified as calmodulin-binding proteins, supporting the involvement of Ca 2ϩ /calmodulin in regulation of Ras-related GTPase function. IQGAP1, a novel RasGAP-related and Ras-binding protein, has been found to bind to calmodulin and the interaction between IQGAP1 and Cdc42 is modulated by Ca 2ϩ /calmodulin (10).
Ral proteins represent a distinct family of Ras-related GTPbinding protein. RalGDS was found to interact with ras p21 specifically and to function as an effector target in Ras signaling pathways, inducing cellular transformation in parallel with activation of the Raf/mitogen-activated protein kinase cascade (11)(12)(13). Ral-A has been shown to be involved in the tyrosine kinase-mediated activation of phospholipase D, suggesting that the signaling pathway from Ras to Ral leads to the regulation of phospholipid metabolism (14). Ral may also have a functional role in regulating the cytoskeleton through its interaction with the effector protein RLIP (Ral interacting protein) and Cdc42 (15). In addition, the diverse subcellular localization of Ral-A, not only in plasma membrane, but also in endocytotic vesicles, synaptic vesicles (16,17), and specialized secretory organelles (18), suggests that Ral proteins may be involved in exo-or endocytosis and membrane traffic. However, the physiological activation of Ral by extracellular stimuli is largely unknown.
We have reported that Ral-A, purified from human erythrocyte plasma membrane, can bind to calmodulin in a Ca 2ϩ -dependent manner, and its calmodulin binding domain has been identified and characterized near its C-terminal region (19). The Ca 2ϩ /calmodulin regulation of Ral-A was strongly supported by the evidence that calmodulin could block Ral-A phosphorylation by PKA, 1 PKG, and PKC and that Ral-A was phosphorylated by CaM kinase II in vitro. 2 In this study we report the effect of Ca 2ϩ /calmodulin on GTP binding activity of Ral-A. Up to 3-fold Ca 2ϩ /calmodulin-dependent stimulation of GTP binding to Ral-A was observed. The results raise the possibility that calmodulin may be a potential effector for Ral-A GTPase activation and serve as a molecular switch in response to changes of intracellular Ca 2ϩ concentration over the physiological range.

EXPERIMENTAL PROCEDURES
Materials-Packed red blood cells were obtained from the Red Cross blood bank, Sydney (New South Wales, Australia). [␣-32 P]GTP (3000 Ci/mmol) was purchased from NEN Life Science Products. Calmodulin was obtained from Calbiochem. Calmodulin-agarose was from Sigma. Mag-Fura-2 tetrapotassium salt was obtained from Molecular Probes (Eugene, OR). Sephadex G-25 was purchased from Pharmacia Biotech (Uppsala, Sweden). Spin columns were obtained from Promega (Madison, WI). Other materials and chemicals were the highest grade available from commercial sources.
Protein Purification-Ral-A protein was purified from human red blood cell plasma membrane as described previously (19). Briefly, calmodulin-depleted human plasma membranes were prepared according to the method of Wang et al. (20). The membrane solute in 200 mM KCl, 1 mM MgCl 2 , 200 M CaCl 2 , 20 mM Tris-HCl, pH 7.4, 0.55% (w/v) Triton X-100, and 20% (v/v) glycerol was applied to a calmodulin agarose column and the column was washed extensively with washing buffer (200 mM KCl, 1 mM MgCl 2 , 200 M CaCl 2 , 20 mM Tris-HCl, pH 7.4, 0.1% (w/v) Triton X-100, and 20% (v/v) glycerol) until no protein was detected in the wash. The protein was eluted from the column by a gradient of increasing concentrations of EDTA from 2 to 5 mM in the elution buffer (200 mM KCl, 20 mM Tris-HCl, pH 7.4, 0.1% (w/v) Triton X-100, and 20% (v/v) glycerol). The fractions were analyzed by 6 -14% SDS-polyacrylamide gel electrophoresis and stored at Ϫ80°C.
Free Ca 2ϩ Concentration Measurement-The fractions containing purified Ral-A were pooled. To determine the precise concentration of EDTA in the protein sample, excess CaCl 2 was added to the sample and Mag-Fura-2, a Ca 2ϩ -sensitive fluorescence probe, was used to measure the free Ca 2ϩ available. This method is based on the fact that Mag-Fura-2 displays shifts in its excitation spectra upon calcium binding and the ratio of the fluorescence intensities at the characteristic maxima for completely bound and unbound forms is indicative of the concentration of Ca 2ϩ present (21). The fluorescence ratio measurements of Mag-Fura-2 (I 331 /I 383 ) in the sample buffer were conducted using a Luminescence Spectrometer (Perkin-Elmer model LS 50B). Excitation was measured at 331 nm when Mag-Fura-2 was completely bound to calcium, and 383 nm when Mag-Fura-2 was unbound to calcium, with emission at 510 nm. Ca 2ϩ concentration was calculated from the ratio (I 331 /I 383 ) and calibration data using the program in the "Intracellular Biochemistry Application" (Perkin-Elmer model LS 50B). EDTA concentration in the protein sample was then obtained by subtracting the measured free Ca 2ϩ concentration from the total added CaCl 2 concentration. When the Ca 2ϩ dose-response curve for calmodulin stimulation of Ral-A GTP binding activity was performed, various external CaCl 2 concentrations were added to yield the free Ca 2ϩ concentrations required, as calculated by a computer program from Fabiato et al. (22). This program calculates the free ionic concentrations resulting from specified total concentrations of cations (Ca 2ϩ and Mg 2ϩ ) and ligands (EDTA) that have been used in the assay.
Guanine Nucleotide Binding-GTP binding activity of Ral-A was determined using a spin column filtration assay. Approximately 1.2 ml of Sephadex G-25 filtration matrix pre-equilibrated in 200 mM KCl and 20 mM Tris-HCl, pH 7.4, was added to individual spin columns. The assembly was then centrifuged at 800 ϫ g in a Jouan CR 4 -11 centrifuge (St-Nazaire Cedex, France) with a E4 swinging bucket rotor for 5 min. A second centrifugation is necessary to remove remaining buffer. 200 l of Ral-A preparation (about 20 pmol) was incubated for the required time at 37°C with 1 Ci of [␣-32 P]GTP (3000 Ci/mmol) in the assay buffer containing 200 mM KCl, 20 mM Tris-HCl, pH 7.4, 5 mM MgCl 2, and required CaCl 2 in the absence and in the presence of calmodulin. Controls were obtained by omission of Ral-A in the assay mixture. Aliquots of 60 l were withdrawn in triplicate and directly applied to the center of the spin column matrix. The assembly was subjected to centrifugation at 800 ϫ g for 5 min at 4°C. Unincorporated free [␣-32 P]GTP remains in the matrix and is effectively removed from [␣-32 P]GTP-bound Ral-A. 5 ml of scintillation fluid was added to the flow-through collected from the spin columns, and the radioactivity of [␣-32 P]GTP-bound protein was measured by scintillation counting in a liquid scintillation analyzer (1900A, TRI-CARB, Packard Pty. Ltd.). The [␣-32 P]GTP binding activity of Ral-A was determined as the radioactivity of the sample minus the radioactivity of the relevant control and expressed as nanomoles of [␣-32 P]GTP/milligrams of protein.

RESULTS AND DISCUSSION
Since Ral-A, purified from human erythrocyte membrane, has been described as a Ca 2ϩ /calmodulin-binding protein (19), we investigated a possible role of Ca 2ϩ /calmodulin to modulate the GTP binding of Ral-A. To test this hypothesis [␣-32 P]GTP binding activity of Ral-A was determined using a spin column filtration assay in the absence and presence of calmodulin. Although a nitrocellulose filtration assay is commonly used to measure [␣-32 P]GTP binding activity of Ras-related GTPases, spin column chromatography eliminates the possibility that calmodulin at high concentration might alter Ral-A binding to nitrocellulose membranes and generate false results. As shown in Fig. 1A, the presence of Ca 2ϩ /calmodulin enhances GTP binding to Ral-A significantly. The increase in GTP binding to Ral-A in the presence of 240 nM calmodulin required Ca 2ϩ , as the omission of Ca 2ϩ completely abolished the stimulatory effect of calmodulin. Ca 2ϩ alone does not affect GTP binding to Ral-A, indicating that GTP binding of Ral-A is not directly activated by Ca 2ϩ , but rather requires calmodulin in its Ca 2ϩbound form. Fig. 1B shows that the binding of GTP to Ral-A in the presence and absence of Ca 2ϩ /calmodulin is time-dependent and that Ca 2ϩ /calmodulin stimulates the initial rate of GTP binding.
To assess how varying the concentration of GTP affects [␣-32 P]GTP binding to Ral-A in response to Ca 2ϩ /CaM, an experiment was performed in which unlabeled GTP at high concentration (0.5 M) was displaced by addition of increasing amounts of [␣-32 P]GTP. Fig. 2 shows that Ca 2ϩ /CaM enhances the [␣- 32  Ral-A has not reached saturation over the [␣-32 P]GTP concentration range examined. Nevertheless, accumulating evidence shows that only a small amount of Ral, in the order of 10%, is in the active GTP-bound form in intact cells (23,24). Based on these observations we suggest that most of the Ral-A purified from human erythrocyte may be in its inactive GDP-bound form and Ca 2ϩ -dependent calmodulin binding to Ral-A may accelerate the displacement of the bound GDP, thereby promoting GTP binding. The displacement of GDP may be incomplete under the conditions used.
To further assess the influence of Ca 2ϩ on calmodulin stimulation of GTP binding, a Ca 2ϩ dose-response curve was carried out. Since Ral-A was purified by elution from a calmodulin agarose column by a gradient EDTA elution buffer, the EDTA concentration in Ral-A was required for the accurate determination of free Ca 2ϩ and was determined as described under "Experimental Procedures." A computer program (22) was then used to calculate the free Ca 2ϩ concentration in the final assay buffer when EDTA, Ca 2ϩ , and Mg 2ϩ were present. Fig. 3A shows the -fold calmodulin stimulation of Ral-A GTP binding activity as a function of increasing concentration of free Ca 2ϩ . In the presence of 240 nM calmodulin, half-maximum stimulation was achieved when the free Ca 2ϩ concentration was about 80 nM. With increasing concentration of calmodulin in the presence of saturating free Ca 2ϩ (1.1 M), a stimulation of GTP binding activity up to 3-fold was observed, as shown in Fig. 3B. Approximately 180 nM calmodulin was required to induce halfmaximum stimulation. The data in Fig. 3, A and B, was fit to the Hill equation , where the Hill coefficient ␥ for free Ca 2ϩ is 0.244, and ␥ for calmodulin is 0.352. The negative cooperativity allows sensitivity of response over a wide concentration range (Fig. 3, A and B, insets). These results are consistent with the known requirement of Ca 2ϩ and calmodulin for the interaction between calmodulin and most of its target proteins (25,26). For example, in vitro study shows the concentration of calmodulin producing half-maximal activation of immobilized CaM kinase II in response to Ca 2ϩ spikes was 271 nM for ␣ subunit and 70 nM for ␤ subunit, respectively (27).
It is known that guanine nucleotide exchange factors, such as RalGAP, RalGDS, and RalGDI, are responsible for direct interaction with and regulation of the "on" and "off" status of Ral protein through the conformational transitions induced by the cycle of GTP/GDP exchange and GTP hydrolysis. Although Hinoi et al. (28) have demonstrated that post-translational modification of Ral enhances the action of RalGDS to stimulate GDP/GTP exchange on Ral, the modes of activation and action of RalGDS or RalGAP have not yet been clarified. The Ca 2ϩ / calmodulin effect on GTP binding activity of Ral-A differs from that of Ral-A regulatory factors. The maximum stimulation of GTP binding by Ca 2ϩ /calmodulin of about 3-fold is lower than the enhanced proportion of Ral-A GTP-bound form induced by RalGDS in COS-7 cells (11). RalGAP partially purified from human platelet cytosol also stimulates the intrinsic GTPase activity of Ral-A by at least 6-fold (29). These results indicate that activation by Ca 2ϩ /calmodulin may be less effective than those of Ral-A regulatory factors and thus it might act in conjunction with regulatory factors in vivo. The sensitivity of GTP binding of Ca 2ϩ (half-maximal activation at 80 nM) is within the range of the increase in Ca 2ϩ from a resting level of about 20 nM to levels of 1 M or more in cells responding to incoming signals. Calmodulin, acting as a Ca 2ϩ -sensor protein, is present in large excess in cells, although the free concentration of Ca 2ϩ -bound calmodulin is limiting and essentially all the Ca 2ϩ -bound calmodulin present in the cell must be bound to targets (30). The very low physiological levels of free Ca 2ϩ -bound calmodulin indicates that small changes in the affinity of a typical target could significantly affect the level of its activity at a submaximal intracellular Ca 2ϩ , as demonstrated recently for smooth muscle myosin light chain kinase activity (31). Recent evidence supports the notion that the sensitivity of Ral-A to Ca 2ϩ has potential physiological significance. Wolthuis et al. (24) reported that an elevation of intracellular Ca 2ϩ through Ca 2ϩ influx stimulated by ionomycin, or release from intracellular Ca 2ϩ stores induced by thapsigargin, induced a rapid activation of Ral in platelets. In addition, Hofer et al. (32) found that endogenous levels of activated GTP-bound Ral (Ral-GTP) was increased by treatment with the Ca 2ϩ ionophore ionomycin in Rat fibroblasts. However, the Ca 2ϩ sensitivity of Ral activation has not yet been reported in these papers. Although the precise mechanism of Ral activation by Ca 2ϩ is unknown, it is speculated that Ca 2ϩ -dependent calmodulin binding to Ral may have a direct effect on Ral activation (24).
Recently, a growing number of calmodulin binding proteins have been identified among Ras-related GTPases, including members of the Rad family of GTPases (including Rad, Gem, and Kir) and Rin family of GTPases (including Rin, Rit, and Ric) (33)(34)(35)(36). However, [ 3 H]GTP or [ 3 H]GDP binding activity of GST/Rad was not affected by calmodulin, neither was its GTP hydrolysis activity (34), and calmodulin inhibits GTP binding to GST/Kir and GST/Gem (33). Furthermore, the effects of calmodulin on Rad, Kir, and Gem were seen in the absence of Ca 2ϩ (33,34). Park et al. (37) have reported that the interaction of Ca 2ϩ /calmodulin with Rab3A induces its dissociation from synaptic membranes. The same effect can be achieved by Rab-GDI (37), but compared with Ca 2ϩ /calmodulin with less effect and a less stringent requirement for GDP than that of RabGDI (37). Similarly to this finding, our work has shown that the stimulatory effect of calmodulin on Ral-A GTP binding is Ca 2ϩdependent and appears to be less effective than that of Ral-A regulatory effectors, suggesting that the regulation by Ca 2ϩ / calmodulin and GTP/GDP exchange factors of this Ras protein may occur by a different mechanism and possibly in conjunction with other Ral effectors. While Ca 2ϩ /CaM enhances the initial rate and apparent sensitivity of GTP binding to Ral-A, it needs to be determined whether this occurs via enhancement of GDP/GTP exchange, by increases in the number and/or affinity of GTP binding sites, or by displacement by calmodulin of an autoinhibitory domain from the active site (38). This work shows a direct functional regulation of Ral-A by Ca 2ϩ /calmod-ulin in the physiological range and suggests a potential role of Ral-A in Ca 2ϩ -regulated pathways.