Arachidonate and related unsaturated fatty acids selectively inactivate the guanine nucleotide-binding regulatory protein, Gz.

G is a member of the family of trimeric guanine nucleotide-binding regulatory proteins (G proteins), which plays a crucial role in signaling across cell membranes. The expression of G is predominately confined to neuronal cells and platelets, suggesting an involvement in a neuroendocrine process. Although the signaling pathway in which G participates is not yet known, it has been linked to inhibition of adenylyl cyclase. We have found that arachidonate and related unsaturated fatty acids suppress guanine nucleotide binding to the α subunit of G. This inhibition of nucleotide binding by cis-unsaturated fatty acids is specific for G; other G protein α subunits are relatively insensitive to these lipids. The IC for inhibition by the lipids closely corresponds to their critical micellar concentrations, suggesting that the interaction of the lipid micelle with G is the primary event leading to inhibition. The presence of the acidic group of the fatty acid is critical for inhibition, as no effect is observed with the corresponding fatty alcohol. While arachidonic acid produces near-complete inhibition of both GDP and guanosine 5′-(3-O-thio)triphosphate binding by G, release of GDP from the protein was unaffected. Furthermore, the rate of inactivation of G by arachidonate is essentially identical to the rate of GDP release from the protein, indicating that GDP release is required for inactivation. These observations indicate that the mechanism of inactivation of G by unsaturated fatty acids is through an interaction of an acidic lipid micelle with the nucleotide-free form of the protein. Although the physiologic significance of this finding is unclear, similar effects of unsaturated fatty acids on other proteins involved in cell signaling indicate potential roles for these lipids in signal modulation. Additionally, the ability of arachidonate to inactivate this adenylyl cyclase-inhibitory G protein provides a molecular mechanism for previous findings that treatment of platelets with arachidonate results in elevated cAMP levels.

G z is a member of the family of trimeric guanine nucleotide-binding regulatory proteins (G proteins), which plays a crucial role in signaling across cell membranes. The expression of G z is predominately confined to neuronal cells and platelets, suggesting an involvement in a neuroendocrine process. Although the signaling pathway in which G z participates is not yet known, it has been linked to inhibition of adenylyl cyclase. We have found that arachidonate and related unsaturated fatty acids suppress guanine nucleotide binding to the ␣ subunit of G z . This inhibition of nucleotide binding by cisunsaturated fatty acids is specific for G z␣ ; other G protein ␣ subunits are relatively insensitive to these lipids. The IC 50 for inhibition by the lipids closely corresponds to their critical micellar concentrations, suggesting that the interaction of the lipid micelle with G z␣ is the primary event leading to inhibition. The presence of the acidic group of the fatty acid is critical for inhibition, as no effect is observed with the corresponding fatty alcohol. While arachidonic acid produces near-complete inhibition of both GDP and guanosine 5-(3-O-thio)triphosphate binding by G z␣ , release of GDP from the protein was unaffected. Furthermore, the rate of inactivation of G z␣ by arachidonate is essentially identical to the rate of GDP release from the protein, indicating that GDP release is required for inactivation. These observations indicate that the mechanism of inactivation of G z␣ by unsaturated fatty acids is through an interaction of an acidic lipid micelle with the nucleotide-free form of the protein. Although the physiologic significance of this finding is unclear, similar effects of unsaturated fatty acids on other proteins involved in cell signaling indicate potential roles for these lipids in signal modulation. Additionally, the ability of arachidonate to inactivate this adenylyl cyclase-inhibitory G protein provides a molecular mechanism for previous findings that treatment of platelets with arachidonate results in elevated cAMP levels.
Trimeric guanine nucleotide binding regulatory proteins (G proteins) 1 comprise a class of membrane-associated proteins that participate in a wide variety of signal transduction path-ways by communicating the external signal from cell surface receptors to intracellular effector molecules (1,2). These G proteins are ␣␤␥ heterotrimers, consisting of two functional subunits, an ␣ subunit containing bound guanine nucleotide, and a ␤␥ complex. In the resting (GDP-bound) state, a G protein can interact with a liganded receptor in a fashion that drives the exchange of GDP for GTP on the ␣ subunit. The ␣-GTP and ␤␥ subunits then dissociate, and both subunit complexes can interact with, and modulate the activity of, downstream effectors. The signal is terminated by an intrinsic GTPase activity of the ␣ subunit; subsequent reassociation with the ␤␥ complex returns the system to its resting state. Effector molecules for G proteins include adenylyl cyclase, certain subtypes of phospholipase C, and various ion channels (3,4).
G proteins are classified through the identity of their ␣ subunit. The high sequence homology among these polypeptides has led to the cloning of several forms for which precise physiologic roles have not yet been ascribed. One such isotype is G z␣ (5,6). The distribution of G z␣ is limited primarily to platelets and neurons, implicating this G protein in some specific role in these tissues (5)(6)(7)(8). The protein has been purified from bovine brain as well as a bacterial expression system and shown to possess biochemical properties distinct from other G protein ␣ subunits (9). For example, nucleotide exchange by G z␣ is highly dependent on free magnesium concentrations. At free magnesium concentrations greater than 10 Ϫ5 M, GTP binding by G z␣ is nearly completely suppressed. This effect is not seen with other G proteins; in fact, the presence of high magnesium concentrations generally stimulates their rates of nucleotide exchange (2). Magnesium-dependent suppression of nucleotide exchange is observed, however, with members of the monomeric family of GTP-binding proteins, e.g. Ras (10). G z␣ also has a very slow intrinsic rate of GTP hydrolysis, more similar to that of Ras and Ras-related proteins than ␣ subunits (9). Although G z is formally a member of the G i family, it is insensitive to ADP-ribosylation catalyzed by pertussis toxin (9), a modification that inactivates the other members of the G i family (11). A property that G z␣ does share with most members of the G i family is an ability to mediate inhibition of adenylyl cyclase (12,13). In addition, G z␣ serves as an excellent substrate for activated protein kinase C both in vitro and in intact platelets (14), and evidence has been obtained that this phosphorylation blocks subunit interactions of this G protein (15).
Several reports have appeared recently, indicating that particular biogenically active lipids can interact in vitro with signaling proteins and modulate their activities. For example, arachidonate and related unsaturated fatty acids physically associate with, and inhibit the activity of, the Ras GTPase activating protein known as GAP (16,17). Such lipids can also regulate the association of the Ras-related protein, Rac, with a specific GDP dissociation inhibitor (18). Similarly, cis-unsatur-ated fatty acids such as oleate and arachidonate have been shown to activate protein kinase C (19). While the mechanism by which lipids modulate the activities of these proteins is not completely defined, their interaction raises interesting possibilities for the role of lipids in cellular regulation. In this study, we demonstrate that cis-unsaturated fatty acids block GTP␥S binding by G z␣ . The mechanism of inactivation involves a specific effect of lipid micelles on the nucleotide-free form of the protein. These observations are of particular interest since the tissues in which G z␣ is found are known to accumulate significant levels of arachidonic acid in response to certain activating stimuli (20,21), and thus the potential exists for cross-talk between arachidonate-producing pathways and those controlled by G z .

EXPERIMENTAL PROCEDURES
Production and Purification of Recombinant G Protein ␣ Subunits-Recombinant G z␣ was expressed in Escherichia coli and purified as described previously (9). The protein was stored at Ϫ80°C in 50 mM HEPES, 1 mM EDTA, 1 mM DTT, and 5 mM MgCl 2 , supplemented with 2 mg/ml bovine serum albumin. Recombinant G s␣ was purified from a bacterial expression system as described (22). Recombinant G i␣1 , G i␣3 , and G o␣ were generous gifts of Maurine Linder (Washington University School of Medicine, St. Louis MO) (23).
Lipid Storage and Micelle Preparation-All lipids were purchased from Sigma, dissolved in ethanol at final concentrations of 50 mM, and stored under N 2 at Ϫ80°C. For the preparation of pure micelles, the required amount of ethanolic lipid solution was dried under vacuum and suspended in 50 mM HEPES, pH 8.0, 1 mM EDTA, and 1 mM DTT. This suspension was subjected to bath sonication until homogenous. Mixed micelles were prepared by dissolving the dried lipids in the same buffer containing 0.1% Lubrol (ICN) at 30°C with vortexing or brief sonication.
Guanine Nucleotide Binding Assays-Guanine nucleotide binding by G protein ␣ subunits was quantitated as described previously (24). Briefly, 1-5 pmol of the protein to be analyzed was diluted to 30 l with 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, and, where indicated, 0.1% Lubrol. 30 l of GTP␥S binding mix consisting of 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, and 2 M [ 35 S]GTP␥S (specific activity, ϳ10,000 cpm/pmol) was then added. In experiments investigating competition between arachidonate and GTP␥S, the concentration of the nucleotide was varied as indicated in the appropriate figure legend. Reactions were initiated by addition of protein, and, unless otherwise indicated, incubation conditions were set as a function of the intrinsic rates of exchange of the various ␣ subunits. These were G z␣ , 30 min at 30°C; G i␣1 and G i␣3 , 20 min at 30°C; G o␣ , 2 min at 20°C; G s␣ , 6 min at 20°C. Free Mg 2ϩ concentration during incubation was 700 nM unless otherwise indicated. Reactions were terminated by the addition of 2 ml of ice-cold 20 mM Tris-Cl, pH 8.0, 25 mM MgCl 2 , and 100 mM NaCl. Samples were kept on ice until filtration through BA85 nitrocellulose filters. Filters were dried, and radioactivity was determined by liquid scintillation spectrophotometry.
For experiments assessing the time course of GDP dissociation from G z␣ , 11 pmol of the protein were incubated at 30°C for 60 min in the presence of 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, 100 mM NaCl, 0.05% Lubrol, and 0.5 M [ 3 H]GDP (specific activity, ϳ26,000 cpm/pmol). Arachidonic acid (300 M) or palmitic acid (300 M) in 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, and 0.1% Lubrol was added, and samples were incubated for an additional 2 min. Samples were then spiked with unlabeled GDP such that the final concentration of GDP in the "chase" was 50 M. The addition of arachidonate and GDP were of small enough volume as not to significantly perturb the relative concentration of protein or detergent. At the time points indicated in the appropriate figure, aliquots of G z␣ were removed to ice-cold buffer (20 mM Tris-Cl, pH 7.7, 100 mM NaCl, 25 mM MgCl 2 ) and stored on ice until filtration through BA85 nitrocellulose filters. Filters were dried, and radioactivity was determined by liquid scintillation spectrophotometry.
For experiments demonstrating the recovery of binding activity with time, G z␣ (7.6 pmol) was incubated under the standard reaction conditions plus 300 M arachidonte. After 5 min, the reaction was diluted 10-fold with 50 mM HEPES, 1 mM EDTA, 1 mM DTT, 0.05% Lubrol, and 2 M GTP␥S. At the times indicated in the appropriate figure, aliquots were removed from the incubation into ice-cold buffer (20 mM Tris-Cl, pH 7.7, 100 mM NaCl, 25 mM MgCl 2 ), and bound nucleotide was determined.
For experiments assessing the time course of inactivation of G z␣ by arachidonate, ϳ7 pmol of protein were incubated in the presence or absence of 300 M arachidonic acid for up to 90 min. At the time points indicated, aliquots of the incubation mixture containing 0.4 pmol of G z␣ were removed and immediately subjected to a 60-min GTP␥S binding assay. Transferring the protein from the pre-incubation to the GTP␥S binding reaction effectively diluted the arachidonate to 30 M in the samples in which the pre-incubation was performed in the presence of 300 M lipid. Additional changes from the standard reaction mixture included the presence of 5 M GDP during the pre-incubation to stabilize the G protein and inclusion of 10 M [ 35 S]GTP␥S (specific, activity ϳ10,000 cpm/pmol) in the GTP␥S binding mix.
Fluorimetric Determination of Critical Micellar Concentrations of Lipids-CMC values for lipids were determined by fluorescence spectroscopy as described by Chattopadhyay and London (25) using the fluorescent probe 1,6-diphenyl-1,3,5-hexatriene. A stock 500 M suspension of lipid (see above) was diluted to the appropriate concentration in 500 l of 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, 5 mM MgCl 2 , and 1.5 mg/ml bovine serum albumin, and this mixture was then added to 500 l of 50 mM HEPES, pH 7.6, 1 mM EDTA, 1 mM DTT, and 1 M GTP. Following equilibration to room temperature, 1,6-diphenyl-1,3,5hexatriene (2 l of a 1 mM solution in tetrahydrofuran) was added, and the tubes were incubated in the dark for at least 30 min. Fluorescence measurements were performed on a Perkin-Elmer 650 -40 fluorescence spectrophotometer set at excitation and emission wavelengths of 358 and 430 nm, respectively.

Cis-unsaturated Fatty Acids Inhibit GTP␥S Binding by
G z␣ -Reports of the effects of arachidonic acid on proteins involved in cell signaling prompted us to evaluate arachidonate and related lipids for potential effects on G protein activities.
The initial experiments focused on G z , since as noted in the Introduction, this G protein is predominantly expressed in tissues with highly active phospholipase A 2 pathways. We first tested various fatty acids for their effect on the ability of G z␣ to bind GTP␥S, a non-hydrolyzable analog of GTP. As shown in Fig. 1, the 20-carbon unsaturated fatty acid, arachidonic acid, dramatically inhibited the ability of G z␣ to bind guanine nucleotide. GTP␥S binding by G z␣ was suppressed in a dose-dependent fashion, and suppression was nearly complete at a concentration of 120 M arachidonic acid. Essentially identical results were obtained when the binding of [ 3 H]GDP, rather then GTP␥S, was examined (results not shown). The carboxyl group on the fatty acid was essential for inhibition of nucleotide binding, as the equivalent fatty alcohol, arachidonyl alcohol, was not inhibitory. Interestingly, arachidic acid, the saturated 20-carbon fatty acid, also had no effect on GTP␥S binding by G z␣ , indicating the requirement for the double bonds for inhibition.
The ability of arachidonate to suppress GTP␥S binding by G z␣ prompted us to examine whether other unsaturated fatty acids exert the same effect. This was indeed found to be the case. Oleic acid, linoleic acid, and linolenic acid all suppressed nucleotide binding by G z␣ in the same dose-dependent fashion as arachidonic acid (Fig. 1). Oleic acid and linoleic acid both completely suppressed GTP␥S binding by G z␣ at a concentration of 175 M, and linolenic acid was completely inhibitory at 250 M. However, the trans-unsaturated fatty acid, elaidic acid, was only slightly inhibitory at these concentrations (data not shown).
The steepness of the inhibition curves for the unsaturated fatty acids indicated that the inhibition was not due to a simple binding event but rather to some sort of cooperative process. One such process, which is quite obvious when working with lipids, is the formation of micelles, which is a highly cooperative aggregation event. Accordingly, we determined the CMC for the lipids under the same conditions (e.g. ionic strength, Mg 2ϩ concentration) as for the GTP␥S binding experiments. CMC values for the various lipids were determined using a fluorescence technique (24), and it was found that the CMC values corresponded nearly identically to the observed IC 50 for the inhibition of GTP␥S binding (Table I). For example, the observed IC 50 and the CMC for arachidonic acid were 60 and 73 M, respectively. These observations provide strong evidence that the abilities of the unsaturated fatty acids to suppress GTP␥S binding by G z␣ is micelle dependent; i.e. it is an interaction of the protein with an anionic lipid micelle, which is responsible for the inhibition.
To facilitate manipulation of the lipid in subsequent studies, we assessed whether the inhibition by arachidonate of the ability of G z␣ to bind GTP␥S occurred when the fatty acid was present in a mixed micelle. The data in Fig. 2 show that this is the case, as the same type of inhibition is observed in response to increasing arachidonic acid when the fatty acid is present in a mixed micelle with the non-ionic detergent, Lubrol. The doseresponse curve is shifted substantially to the right as would be expected for a process that depends on the mole fraction of lipid in the micelle (26). In fact, the IC 50 for inhibition of GTP␥S binding by G z␣ shifts in proportion to the mole fraction of the lipid (results not shown).
Inhibition of Nucleotide Binding by Arachidonate Is Specific for G z␣ -We next determined the specificity of G proteinarachidonate interactions by determining the effect of the fatty acid on GTP␥S binding by other G protein ␣ subunits. As in the previous experiments, GTP␥S binding by G z␣ was inhibited 50% at 150 M arachidonate, while at 300 M, nucleotide binding was nearly completely suppressed (Fig. 3). GTP␥S binding by G i␣3 , a member of the G i subfamily to which G z␣ belongs, was not significantly inhibited at 150 M arachidonate but was inhibited ϳ50% at 300 M. The other G protein ␣ subunits tested, including that of the closely related G i1 as well as G o and G s , were not inhibited by arachidonate at any concentra-tion tested. This specificity for G z␣ over other G protein ␣ subunits suggests a potential role for arachidonate in G z signaling and prompted us to investigate the mechanism of this lipid effect on GTP␥S binding by G z␣ .
Mechanism of Action-The binding of GTP␥S by G proteins is a two-step process involving GDP release (the rate-limiting step) and subsequent diffusion-controlled GTP␥S binding by the nucleotide-free protein (2). To explore the mechanism by which arachidonic acid inhibits GTP␥S binding, we first assessed the effect of arachidonic acid on the rate of GDP release from the protein. As release of bound GDP from ␣ subunits is the rate-limiting step in nucleotide exchange, we expected that this process would be inhibited by arachidonic acid in a fashion similar to that of GTP␥S binding. Quite surprisingly, however, release of [ 3 H]GDP from G z␣ occurred with the same rate constant (Fig. 4) in the presence of 300 M palmitic acid (a non-inhibitory fatty acid) or in the presence of 300 M arachi-   2. Effect of a non-ionic detergent on arachidonic acid-dependent inhibition of GTP␥S binding by G z␣ . G z␣ was subjected to the GTP␥S binding reaction as described under "Experimental Procedures" for 20 min at 30°C in the presence of the indicated concentrations of arachidonic acid and 0.05% Lubrol (E). For comparison, the data obtained in the absence of added Lubrol are shown (q, see Fig. 1). Data shown are from a single experiment that has been repeated at least three times. Maximal GTP␥S binding is defined as the amount of GTP␥S binding observed in the absence of arachidonate under each condition; the presence of Lubrol had a negligible effect on the binding. donic acid, a concentration that essentially completely suppressed GTP␥S binding to the protein (see Fig. 2). These data indicate that the effect of arachidonic acid is exerted at the step of GTP␥S binding. This would be highly unusual since GTP binding by G proteins is normally diffusion controlled, and thus its rate would have to be reduced by many orders of magnitude before an effect on the overall binding reaction would be observed.
One possibility for the selective effect of arachidonate on the GTP␥S binding step is that the lipid micelle could interact specifically with the unoccupied nucleotide binding site on G z␣ and effectively compete for GTP␥S binding. If this were the case, inhibition of nucleotide binding by arachidonic acid should be reduced by increasing the concentration of competing nucleotide. To explore this possibility, we measured the effect of arachidonic acid on GTP␥S binding in the presence of increasing GTP␥S concentrations. However, assessment of the arachidonate-mediated inhibition over a 50-fold range of GTP␥S revealed that binding was nearly completely suppressed at all concentrations of competing nucleotide (Fig. 5). Since inhibition of nucleotide binding by arachidonic acid was unaffected at GTP␥S concentrations as high as 25 M, which is Ͼ1000-fold above the K d of G protein ␣ subunits for GTP␥S (2), it is considered highly unlikely that the lipid micelle is competing for the nucleotide binding site of the protein.
An alternative explanation for the effect of arachidonate on GTP␥S binding, but not on GDP release, by G z␣ is that the fatty acid could somehow interact with and inactivate the nucleotide-free form of the G protein that is a transient intermediate in the exchange process. To examine this possibility, we assessed the time dependence of the inactivation of G z␣ by arachidonate. If arachidonate could exert its effect only on the nucleotide-free form of G z␣ , a recovery of binding activity should be observed if the protein is exposed to high arachidonate and then is diluted to an ineffective concentration. This recovery of binding activity would then reflect the fraction of the protein that had not yet released its GDP. Furthermore, if the arachi-donic acid is selectively inactivating the nucleotide-free form of G z␣ , then the rate of inactivation of GTP␥S binding should correspond to the rate of GDP release. Indeed, the evidence indicates that this is the case (Fig. 6). In the first experiment (Fig. 6A), binding activity was measured after G z␣ was first incubated with 300 M arachidonate for 5 min and then diluted to 30 M arachidonate. While GTP␥S binding activity was detected, the level of nucleotide binding recovered was significantly less than the control levels in which only 30 M fatty acid had been present throughout.
This same type of experiment was performed over a range of pre-incubation times with 300 M arachidonate from 5 to 90 min; in each case, the quantity of G z␣ capable of binding nucleotide was assessed after a 10-fold dilution to the ineffective concentration of the lipid (i.e. 30 M). The results of this analysis, shown in Fig. 6B, revealed in each case a loss of GTP␥S binding activity that was not recovered by subsequent dilution. This was not due simply to protein lability, as pre-incubation of G z␣ in the absence of arachidonate did not result in a loss of GTP␥S binding activity. An equally important finding from this experiment is that the time dependence in the loss of the binding activity of G z␣ could be fit to an exponential with a decay constant of 0.028 min Ϫ1 , which is nearly identical to the rate constant for GDP release from G z␣ under the same conditions (9). Taken together, these data indicate that, in the presence of arachidonate, G z␣ is able to release GDP normally but is then rapidly inactivated when the lipid micelle interacts with the nucleotide-free form of the protein. The incubation conditions were adjusted for each ␣ subunit as described under "Experimental Procedures." Data shown represent the mean of three separate determinations with the 100% control value being the binding observed in the absence of added arachidonic acid. AA, arachidonic acid. DISCUSSION The role of lipids in cellular signaling has received increasing attention in recent years (27). It is now clear that lipids such as arachidonic acid and diacylglycerol actively participate as second messengers in signaling pathways (28,29). Examples are also beginning to emerge of arachidonate and other cis-unsaturated fatty acids directly modulating activities of signaling proteins. For example, these fatty acids can associate with and alter the activity of Ras-GAP (17,30). Cis-unsaturated fatty acids have also been shown to regulate association between the monomeric G protein, Rac, and its GDP dissociation inhibitor (18). Arachidonate and other unsaturated fatty acids have also been shown to activate certain isozymes of the protein kinase, protein kinase C (19).
In this report, we have identified an additional effect of cis-unsaturated fatty acids on a signaling protein, that being the inactivation of a G protein ␣ subunit, specifically G z␣ . Arachidonate-dependent inactivation of GTP␥S binding by G z␣ was quite specific for this ␣ subunit, as treatment of a number of other ␣ subunits had only minimal effects on their abilities to bind nucleotide. The inactivation was dependent upon the presence of an acidic group on the lipid and correlated with the formation of a lipid micelle. Several cis-unsaturated fatty acids were potent inhibitors of GTP␥S binding by G z␣ with a dose dependence that matched the lipid's respective CMC. This suggests that it is an interaction between the charged surface of a micelle with G z␣ that is required for its inhibition. These results are similar to the results of Serth et al. (17), who observed the inhibition of Ras-GAP in the presence of fatty acids and acidic phospholipids but not in the presence of neutral lipids, and only under conditions in which the active lipids formed micellar structures.
The inhibition of GTP␥S binding by G z␣ seen upon the addition of arachidonic acid could have been exerted at either of two distinct steps in the process, these being dissociation of bound GDP or association of the GTP␥S. The former step was initially considered the most likely, as GDP dissociation from G proteins is ϳ10 7 -fold slower than association of guanine nucleotides FIG. 6. Arachidonic acid selectively inactivates the nucleotidefree form of G z␣ . A, G z␣ was incubated in a batch reaction as described under "Experimental Procedures" at 30°C in the presence of either 30 M (f) or 300 (q, E) M arachidonate. In one of the reaction mixtures containing 300 M arachidonate (q), the mixture was diluted 10-fold with 50 mM HEPES, 1 mM EDTA, 1 mM DTT, 2 M GTP␥S (specific activity, 12,000 cpm/pmol) after 5 min of incubation. At the times indicated, aliquots containing ϳ0.5 pmol of G z␣ was removed to ice-cold 20 mM Tris-Cl, 25 mM MgCl 2 , 100 mM NaCl, and bound nucleotide was determined. B, G z␣ was incubated in a batch reaction as described under "Experimental Procedures" at 30°C in 50 mM HEPES, 10 mM EDTA, 1 mM DTT, 2.65 mM MgCl 2 , 5 M GDP, 0.05% Lubrol, and either 0 (E) or 300 M (q) arachidonate. At the indicated times, aliquots containing ϳ0.4 pmol of G z␣ were removed and added to a GTP␥S binding reaction mixture containing 10 M GTP␥S and a free Mg 2ϩ concentration of 700 nM. For the experiment conducted in the absence of arachidonic acid (E), the GTP␥S reaction mixtures contained an added 30 M arachidonic acid so that the lipid concentration in all binding assays was held constant. GTP␥S binding was performed at 30°C for 60 min for all data points. Data points represent means of six separate determinations. AA, arachidonic acid. (31). To identify the step in G z␣ nucleotide exchange affected by arachidonate, we directly determined the effect of arachidonic acid on the rate of GDP release. Quite surprisingly, GDP release was virtually unaffected by concentrations of arachidonate, which essentially completely suppress GTP␥S binding, indicating that GTP␥S binding was the step being affected. An assessment of the time dependence of arachidonate inhibition of GTP␥S binding revealed that (a) inhibition of GTP␥S binding was not reversible even after 60 min of incubation and (b) the rate of this inactivation corresponded precisely with that of GDP release by G z␣ . Taken together, these experiments indicate that the effect of arachidonate on the ability of G z to bind guanine nucleotides is dependent on an association of the lipid micelle with the nucleotide-free form of the protein, resulting in an alteration of the protein that renders it inactive.
While lipid-mediated modulation of G protein activity by irreversible inactivation seems an unlikely mode of regulation in the cell, the selectivity of the process for G z over other G proteins, as well as the unique distribution of G z , provides strong reasons to suspect that the process is physiologically relevant. As noted above, the cell types in which G z is found, such as platelets and chromaffin cells, are known to possess high levels of phospholipase A 2 activity (20,21). These cells are also known to produce substantial levels of arachidonate in response to external stimuli (20,32,33). Also of note in this regard are previous studies showing that treatment of platelets with high levels of exogenous arachidonate results in increased intracellular cAMP accompanied by reduced aggregation (34,35). In one of these studies, treatment of platelets with an adenylyl cyclase inhibitor restored aggregation in the presence of arachidonate, indicating that the fatty acid was exerting its effect at or upstream of adenylyl cyclase (35). The finding that arachidonate can inactivate a G protein that is both present in platelets and implicated in the inhibition of adenylyl cyclase thus provides a potential molecular mechanism for these effects.
Finally, it is certainly possible that in the context of an intact cell an increase in the concentration of arachidonic acid might be only transiently inhibitory, i.e. the cellular environment could provide protection of the apoprotein form of G z␣ from permanent inactivation by the lipid. Possibilities here include a protective factor in these cells that associates with G z␣ or one that reverses the association between G z␣ and inhibitory lipids. Identification of the pathway in which G z␣ participates will likely shed some light on these results and on the possibilities for the novel means of G protein regulation they may represent.