Effect of Halothane on the Guanosine 5′ Triphosphate Binding Activity of G-Protein αⁱSubunits

The rate of [³⁵S]GTPγS binding to recombinant Gαⁱ¹was measured, following a modified version of the method of Sternweis and Robishaw. 5For each time point in the [³⁵S]GTPγS binding assay, the reaction was carried out at 30°C in a 250-μl tube, which was sealed with a Teflon stopper. Final assay concentrations in 150 μl assay were 50 mm HEPES (pH 8), 1 mm EDTA, 1 mm dithiothreitol, 10 mm MgCl², 0.1% polyoxyethylene 10 lauryl ether (C¹²E¹⁰), 0.7 μm GTPγS, 0.4 μCi [³⁵S]GTPγS, 30 nm GDP, 8 nm Gαⁱ¹, and 0.25 μm bovine serum albumin. Anesthetic was incorporated when appropriate by bubbling halothane for 15 min with a vaporizer into buffer containing 25 mm HEPES (pH 8), 1 mm EDTA, 1 mm dithiothreitol, and 10 mm MgCl², which was then added to the assay tubes without changing the final volume. The reactions were quenched with 1 ml ice-cold stop buffer (25 mm Tris [pH 8], 20 mm MgCl², and 100 mm NaCl). This was immediately filtered under vacuum through 0.45-μm high-affinity nitrocellulose filters from Millipore (Bedford, MA), washed five times with 3 ml stop buffer, dried, and counted in 4 ml liquid scintillation cocktail. Protein-free blanks were run with each assay and subtracted from the total counts before further processing. The Gαⁱ¹concentration was determined from the binding at 3 h (assuming that binding was complete at this time), and all data were normalized to this value.

The increase in intrinsic tryptophan fluorescence accompanying GTPγS binding, described by Higashijima et al. , 6was used as a complementary technique to monitor the binding of GTPγS to recombinant subunits. In this technique, 1 μm Gαⁱ¹in 1.15-ml solution of 25 mm HEPES (pH 8), 1 mm EDTA, 1 mm dithiothreitol, 10 mm MgCl², 0.1% C¹²E¹⁰, 10 μm GTPγS, and 300 nm GDP was sealed in a 1-mm quartz cell with a Teflon stopper. Note that the GTPγS concentration was in excess of the G protein but by an order of magnitude less than that used in the [³⁵S]GTPγS binding assay to minimize the inner filter effect at the excitation wavelength. In both the fluorescence and [³⁵S]GTPγS assays, there was an approximately 25-fold excess of GTPγS to GDP. The intrinsic protein tryptophan fluorescence (λ^ex= 290 nm, λ^em= 340 nm) was recorded at 20-s intervals for 60 min with a Spex Fluorolog fluorimeter (Jobin Yvon Inc., Edison, NJ). During the assay, the cell holder was maintained at 30°C, and the contents were stirred continuously with a magnetic stir bar. The baseline fluorescence of the GDP-bound subunit, measured just prior to nucleotide addition, was subtracted, and the data were analyzed as net increase in fluorescence. As with the [³⁵S]GTPγS assay, halothane was added from a stock solution prepared by bubbling. Halothane quenches the fluorescence of tryptophan (isolated amino acid) by approximately 1.8%/mm, 7which was accounted for when analyzing the fluorescence intensity curves.

GTPase activity was measured by the release of [³²P]-labeled inorganic phosphate, based on the procedure described by Higashijima. 8Gαⁱ¹, 0.2 μm, was incubated at 30°C in 25 mm HEPES (pH 7.8), 100 mm NaCl, 10 mm MgSO⁴, 1 mm EDTA, 1 mm dithiothreitol, 0.1 μm GDP, 5 μm GTP, 0.025 mCi [γ³²P]GTP, and 0.5 mg/ml bovine serum albumin with and without halothane (control), which was diluted from an aqueous stock solution. At 10 and 20 min after initiation, aliquots from each assay tube were quenched with a solution of 7% (w/v) activated charcoal in 2 N HCL and 0.35 m NaH²PO⁴and then vortexed for 10 s. The tubes were centrifuged at maximum speed in a microfuge E centrifuge for 10 min to pellet the charcoal. Aliquots of the supernatant from each assay sample time point were transferred to a scintillation vial with 3 ml water, and the amount of [³²P]-labeled inorganic phosphate was determined by Cerenkov counting.

GTPγS binding to Gαⁱβγ heterotrimeric complex associated in crude membrane prepared from ASM tissue was measured using a modification of previously described methods. 9,10Porcine tracheal smooth muscle strips (260–280 mg) were frozen and pulverized in liquid N²to a fine powder with a mortar and pestle, which was kept cold with dry ice. The powdered tissue was then suspended for 15 min by vigorous mixing in 1.5 ml ice-cold extraction buffer and then centrifuged at 4°C for 10 min at 2,500 g . The extraction buffer contained 20 mm HEPES (pH 8.0), 1 mm EDTA, and the protease inhibitors 0.1 mm phenylmethanesulfonyl fluoride, 2 μg/ml aprotinin, and 10 μg/ml leupeptin. The crude membrane pellet was washed three times with 1 ml extraction buffer, resuspended in the assay buffer, and homogenized with 20 strokes in a Dounce tissue grinder. The assay buffer contained 50 mm Tris-HCl (pH 7.4), 4.8 mm MgCl², 2 mm EDTA, 100 mm NaCl, and 1 μm GDP. The crude homogenate was incubated on ice for 1 h and then centrifuged at 4°C for 10 min at 2,500 g .

The resulting supernatant contained intact Gαⁱβγ heterotrimeric complex (fig. 1, A ). To immunoprecipitate, an immunoaffinity purified rabbit antibody directed against the C-terminal nine amino acids of Gα^i-3(Calbiochem, San Diego, CA) was used at a 1:200 dilution. This antibody has cross-reactivity with Gα^i-1, Gα^i-2, and Gα^i-3but not with Gα^q/11(fig. 1, B ). Further centrifuging the supernatant at 100,000 g  for 1 h pelleted the membrane completely, which removed any trace of Gαⁱβγ heterotrimeric complex from the supernatant, as determined by Western blot (data not shown). Thus, the supernatant can be characterized as a colloidal suspension of small membrane particles containing the Gαⁱβγ heterotrimeric complex.

To measure the protein concentration in the supernatant, 50 μl was first solubilized by the addition of 12.5 μl NaOH, 0.1 N, and boiled for 30 min. The protein concentration was measured by the Lowry method, using the detergent compatible assay from Bio-Rad (Hercules, CA). The supernatant was diluted with cold assay buffer to 2.5 mg/ml.

The [³⁵S]-GTPγS binding to Gαⁱsubunits in the supernatant was performed by modification of previously described methods. 9To minimize variability between tissue preparations, assays were run in duplicate on supernatant from the same tissue preparations. Supernatant (55 μl) was incubated at 30°C without (control) or with 1 mm halothane for 10 min. The reactions were then initiated by the addition of 6 nm (final concentration) of [³⁵S]GTPγS (≈ 2.7 mCi/fmol) to each assay tube. The reactions were quenched after 1, 5, 10, and 30 min with 0.6 ml ice-cold immunoprecipitation buffer containing 50 mm Tris-HCl (pH 7.5), 20 mm MgCl², 150 mm NaCl, 0.5% of the nonionic detergent IGEPAL CA630, 100 μm GDP, 100 μm GTP, and 20 μg/ml aprotinin. Then, the suspensions were incubated with 70 μl protein A-Agarose beads that had been precoated with rabbit anti-Gαⁱ³antirabbit immunoglobulin directed against an epitope at the C terminus of Gα^i-3. These mixtures were incubated for 2 h at 4°C on a rocker to ensure continuous mixing. After this incubation, the mixture was centrifuged at 2,500 g , and the supernatant was removed by aspiration. The pelleted beads were then washed five times with 1 ml immunoprecipitation buffer. The beads were then transferred onto very low protein binding filters (Millipore, Billerica, MA) and washed a final time with 20 ml immunoprecipitation buffer. The dried filter paper retaining the beads was placed in a scintillation vial containing 4 ml liquid scintillant (Ultima Gold; Packard Bioscience, Boston, MA) and counted. The nonspecific binding of [³⁵S]-GTPγS to the beads in the absence of protein was used as the blank value and was determined simultaneously for each study.

Halothane concentrations in assay solutions were estimated from assay tubes handled under identical conditions and run simultaneously with the assay. Halothane concentrations in solution were measured by gas chromatography after hexane extraction by the method of Van Dyke and Wood. 11 

Results are presented as mean ± SD; n refers to the number of pigs for the crude membrane preparations, or the number of assay replicates for recombinant protein. The GTPase-specific activities of recombinant Gαⁱ¹, with and without halothane, were compared by paired t  test.

The GTPγS binding curves of the recombinant protein, generated with the intrinsic fluorescence and [³⁵S]GTPγS binding assays, and the endogenous protein, generated with [³⁵S]GTPγS immunoprecipitation assay, were fit with the equation y = a (1 − e^−kt) from reference 12using nonlinear least squares fitting. In the equation, the independent variable is t, time, the dependent variable is a, the amount of Gαⁱ¹bound to GTPγS, the parameter k is the on rate of the ligand for the bimolecular reaction, and the parameter a vertically scales the curve. For recombinant Gαⁱ¹, the k parameters were compared for significance by t  test (n = 3). For the effect of suramin on the endogenous subunit, the k parameter for each curve fit was used to calculate the amount of Gαⁱbound to GTPγS at 15 min. The amounts of Gαⁱ¹bound to GTPγS, calculated from the fits to the control and suramin curves, were compared for significance on an experiment by experiment basis by Student paired t  test.

The effect of halothane or suramin on the amount of recombinant Gαⁱ¹bound to [³⁵S]GTPγS at specified times was examined using a paired t  test. The effect of halothane, atropine, or acetylcholine on the amount of endogenous Gαⁱbound to [³⁵S]GTPγS binding at 10 min was assessed by one-way repeated-measures analysis of variance.

Significance was assigned at P < 0.05. Sigma Plot and Sigma Stat (Jandel Scientific, San Rafael, CA), respectively, were used to perform the curve fitting and statistics.

Hexahistidine-tagged recombinant Gαⁱ¹subunit, subcloned from rat and expressed in Escherichia coli , was generously provided by Alfred G. Gilman, 13M.D. Ph.D. (Professor of Pharmacology, University of Texas Southwestern Medical Center, Dallas, Texas). [³⁵S]GTPγS was purchased from Amersham Biosciences (San Francisco, CA). Unless specifically mentioned in the text, other materials were purchased from Sigma (St. Louis, MO).

The steady state GTPase-specific activity of the recombinant Gαⁱ¹was 0.033 ± 0.018 mole Pⁱmole Gαⁱ¹⁻¹min⁻¹under control conditions and 0.035 ± 0.015 mole Pⁱmole Gαⁱ¹⁻¹min⁻¹in the presence of 1.1 ± 0.2 mm halothane, a difference that is not significant (P = 0.628, n = 15).

Gαⁱ¹bound [³⁵S]GTPγS in a time-dependent fashion (fig. 2, A ), with a k of 0.095 ± 0.002 min⁻¹(n = 3).

Suramin, which is known to inhibit GTPγS binding to Gα subunits by inhibiting GDP dissociation, 14significantly (P < 0.02, n = 3) decreased the mole fraction of GTPγS-bound Gαⁱ¹at 15 min from 0.62 ± 0.10 to 0.49 ± 0.03 (fig. 2, A ). However, because GTPγS is a nonhydrolyzable analog of GTP, which is more tightly bound, near-stoichiometric binding was achieved within 3 h even in the presence of suramin (n = 3).

The mole fractions of Gαⁱ¹bound to [³⁵S]GTPγS were 0.49 ± 0.02 and 0.60 ± 0.02 at 10 and 20 min, respectively. The addition of halothane (1.26 ± 0.07 mm) did not significantly change these values (0.48 ± 0.005 and 0.63 ± 0.016, respectively, P = 0.14, n = 3;fig. 2, A ).

The a parameter from the curve fit to the [³⁵S]GTPγS-binding time course was substituted into the control fluorescence curve fit equation, which scales the fluorescence curve fit to the same ordinate units for direct comparison. An assumption that the observed fluorescence intensity change corresponds exclusively to GTPγS binding at the Gαⁱ¹active site was made and is justified on the basis of previous observations. 6,8Furthermore, on the basis of our [³⁵S]GTPγS binding experiments, which demonstrate that inhibition of binding with suramin changes the rate of GTPγS binding (k) but not the final stoichiometry (a), it was possible to scale all subsequent fluorescence curve fits with the a parameter from the control [³⁵S]GTPγS-binding time course. This allowed the comparison of the radioactive and fluorescence data on the same ordinate scale. Under control conditions, the intrinsic fluorescence intensity curve fit had a k of 0.091 ± 0.015 min⁻¹(n = 3, fig. 2, B ), which is nearly identical to that obtained in the control [³⁵S]GTPγS-binding time course. Thus, the two methods provide comparable results, which are consistent with previously reported values for this subunit. 15,16 

Suramin (10 μm) significantly decreased the value of k (to 0.046 ± 0.004 min⁻¹) compared with control conditions (n = 3, P < 0.01;fig. 2, B ). The increase in fluorescence observed at 15 min in the presence of suramin was on average 33% smaller than the increase in the control fluorescence at that time. This is similar to the 22% decrease in the mean mole fraction of [³⁵S]GTPγS bound to Gαⁱin the presence of 10 μm suramin at 15 min. In contrast to the effect of suramin, halothane (0.59 ± 0.03 mm) did not significantly (n = 3, P = 0.1) affect the rate of fluorescence increase (k = 0.083 ± 0.002 min⁻¹;fig. 2, B ).

[³⁵S]GTPγS binding increased in a time-dependent manner with a mean rate that was approximately twice that measured in the recombinant Gαⁱ¹(k = 0.17 ± 0.08 min⁻¹, n = 7).

Suramin (10 μm) did not significantly affect the mean of the fitted k values (0.18 ± 0.13 min⁻¹, P = 0.92, n = 7;fig. 3). However, when compared on a per-experiment basis with Student t  test, suramin significantly reduced the amount of [³⁵S]GTPγS bound at 10 min, as calculated from the fit parameters (5.7 ± 1.7 and 4.2 ± 1.4 fmol/mg for control vs.  suramin, P < 0.02, n = 7).

To further characterize this preparation, muscarinic receptor function was examined in a separate set of studies. Neither 100 μm acetylcholine nor 10 μm atropine affected [³⁵S]GTPγS binding at 10 min of assay (n = 4;fig. 4), suggesting that muscarinic receptors are not functionally coupled to heterotrimeric G proteins in this preparation.

Halothane (0.99 ± 0.17 mm) did not significantly affect [³⁵S]GTPγS binding measured at 10 min (n = 6;fig. 4).

The major finding of this study is that halothane, at concentrations sufficient to produce both anesthesia in vivo  and relaxation of intact ASM in vitro , does not inhibit the binding of GTPγS to recombinant Gαⁱ¹subunits in aqueous solution or its subsequent hydrolysis. Likewise, similar concentrations of halothane do not inhibit the binding of GTPγS by endogenous Gαⁱβγ heterotrimer in crude membrane preparations from ASM tissue.

In the current study, we examined the effects of halothane on the Gαⁱsubunit to explore the mechanism responsible for the observed inhibition of muscarinic receptor–induced increases in calcium sensitivity by halothane in ASM. 3Our prior work localized the site of halothane's action to the muscarinic G protein–coupled receptor complex. 4We and others have also shown that the Gαⁱsubunit is responsible at least in part for mediating these effects. 4,17–19 

The anesthetic target is not currently known, but is assumed to be a component of the G protein–coupled receptor complex. Studies of halothane interaction with rhodopsin-transducin have failed to demonstrate direct binding of the anesthetic to the G protein. 2,20However, Pentyala et al.  1found that volatile anesthetics (including halothane, isoflurane, and sevoflurane) at low concentrations (approximately 0.25 mm, equivalent to approximately 1 minimum alveolar concentration) modulated the binding of guanine nucleotides to purified Gα subunits in aqueous solution, inhibiting the exchange of GTPγS for bound GDP. This effect was observed for several subunits studied (including Gⁱand G^s) but not all (including G^oand monomeric G proteins). This finding suggested potential anesthetic effects directly on the guanine-nucleotide binding (catalytic) site that may be specific to Gα subunit isotype. However, these isoform-specific results are not totally consistent with the behavior of the corresponding isoforms in cardiac muscle membranes, where Gαⁱand Gα^Scouple M²muscarinic and β-adrenergic receptors, respectively, to adenylate cyclase. In preparations of these membranes, halothane interferes with muscarinic receptor inhibition of adenylate cyclase through Gαⁱbut does not affect β-adrenergic receptor stimulation via  Gα^S. 17,21,22 

In the current study, the GTPase-specific activity and the [³⁵S]GTPγS binding kinetic profile of recombinant Gαⁱ¹under control conditions (fig. 2) were similar to those previously reported for Gαⁱisoforms. 15,16The results of the [³⁵S]GTPγS binding assays were verified by comparison to the results of the increase in intrinsic fluorescence, run under similar conditions. At times less than 60 min, Gαⁱ¹in the presence of suramin bound less GTPγS than the control, presumably because suramin inhibited the off rate of GDP, 14which is the rate-limiting step to GTP binding and hydrolysis. 8These findings suggest the validity of our GTPγS binding assay.

On the other hand, halothane did not significantly affect the amount of GTPγS bound to Gαⁱ¹compared with control conditions. Our results are not consistent with the conclusion of Pentyala et al.  1that clinically relevant concentrations of halothane reduce the binding of GTP. We observed neither an inhibition of GTPγS binding nor an inhibition of GTP hydrolysis.

What could explain the apparent contradiction between our results and those of Pentyala et al.  1? The [³⁵S]GTPγS binding results were obtained with similar assay methods. One possible point of departure is different sources of Gα subunits. However, the binding kinetics and hydrolysis rate that we measured for the Gαⁱ¹are in accordance with published literature and are invariate in the presence of up to 1 mm halothane. Thus, we cannot explain the reason for the differences in our results.

To confirm our results, we also studied the ability of endogenous membrane–associated G-protein heterotrimer from crude membrane preparations of ASM tissue to bind GTPγS. This preparation also provided a milieu in which to test whether halothane might affect unstimulated Gαⁱβγ GTPγS binding through direct interaction on the Gαⁱsubunit or indirect effects through the heterotrimer. The lack of effect of acetylcholine, a receptor agonist, or atropine, a receptor antagonist, on Gαⁱβγ GTPγS binding demonstrated that the heterotrimer was functionally uncoupled from the muscarinic receptor. Thus, the measured Gαⁱβγ GTPγS binding was not regulated by membrane receptors. However, this membrane preparation probably contains other regulators of G-protein activity that may affect the GTP binding activity of Gαⁱ. For example, the basal GTPγS binding activity of recombinant Gα^qin aqueous solution is undetectable but is measurable in membrane preparations. 23,24This factor may explain the more rapid kinetics of binding observed in the current immunoprecipitation study compared with the kinetics measured for recombinant Gαⁱsubunits in aqueous solution.

Despite these differences, the pattern of results was similar between the two preparations. Suramin had an inhibitory effect on GTPγS binding by Gαⁱβγ in both the membrane preparation and the recombinant Gαⁱ, confirming the ability of both assays to detect a decrease in GTPγS binding. However, unlike the recombinant protein, suramin did not significantly alter the mean rate (k) of binding to endogenous protein in the membrane preparation compared to the control. This may arise partly because of the presence of intracellular effectors, which enhance the apparent rate of GTPγS binding by Gαⁱin the membrane preparation compared to the isolated recombinant protein. If these effectors act to enhance the off rate of GDP from Gαⁱ, they might partially counteract the effect of suramin, which is to inhibit the GDP off rate. Halothane did not inhibit the basal GTPγS binding activity of endogenous Gαⁱin the membrane preparation (fig. 4). Taken together with a similar lack of effect on recombinant Gαⁱ, this finding suggests that factors such as the presence of Gβγ subunits do not modulate any effects of halothane on G-protein function in situ . This finding does not appear to lend support to the proposal that anesthetics stabilize the binding of the Gαⁱsubunit to the Gβγ dimer, although a more detailed study of such an effect is required to definitively address this. 25It is possible that features necessary for an effect of halothane on basal GTPγS binding activity (such as functional coupling to receptors) are lost in the membrane preparation assayed, so this conclusion cannot be definitive. It is also possible that halothane could affect the activity of other Gα subunits that regulate contractility in ASM, such as Gα^q. 26,27However, the current results suggest that the target of anesthetic effects to inhibit G protein–coupled receptor responses is the receptor itself, or the interface between receptor and the heterotrimeric G protein, rather than a site on the Gα subunit that regulates guanine nucleotide exchange.

Halothane did not inhibit the basal GTPγS binding or hydrolysis activity of isolated, recombinant Gαⁱ¹subunits or the unregulated GTPγS binding of endogenous Gαⁱsubunits in a crude membrane preparation of ASM. These results suggest that halothane, which inhibits receptor-activated Gαⁱ-coupled pathways in ASM, must functionally target a component of the G protein–coupled receptor complex other than Gαⁱ.

The authors thank Alfred G. Gilman, M.D. Ph.D. (Professor of Pharmacology, University of Texas Southwestern Medical Center, Dallas, Texas), for generously supplying the Gαⁱ¹; Dr. Franklyn Prendergast, M.D., Ph.D. (Professor of Biochemistry and Molecular Biology, Mayo Clinic, Rochester, Minnesota) for the use of spectroscopic instrumentation; and Darrell Loeffler (Technician, Department of Anesthesiology, Mayo Clinic, Rochester, Minnesota) for expert technical assistance.