HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 277, Issue 37, 34434-34442, September 13, 2002

This Article Abstract Full Text (PDF) All Versions of this Article: 277/37/34434    most recent M204259200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gille, A. Articles by Seifert, R. Search for Related Content PubMed PubMed Citation Articles by Gille, A. Articles by Seifert, R. Social Bookmarking                      What's this?

Distinct Interactions of GTP, UTP, and CTP with G_s Proteins^*

Andreas Gille, Hui-Yu Liu^§, Stephen R. Sprang^¶, and Roland Seifert

From the  Department of Pharmacology and Toxicology, the University of Kansas, Lawrence, Kansas 66045-7582, and the ^¶ Howard Hughes Medical Institute and Department of Biochemistry, the University of Texas Southwestern Medical Center, Dallas, Texas 75390-9050

Received for publication, May 1, 2002, and in revised form, June 4, 2002

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Early studies showed that in addition to GTP, the pyrimidine nucleotides UTP and CTP support activation of the adenylyl cyclase (AC)-stimulating G_s protein. The aim of this study was to elucidate the mechanism by which UTP and CTP support G_s activation. As models, we used S49 wild-type lymphoma cells, representing a physiologically relevant system in which the ₂-adrenoreceptor (₂AR) couples to G_s, and Sf9 insect cell membranes expressing ₂AR-G_s fusion proteins. Fusion proteins provide a higher sensitivity for the analysis of ₂AR-G_s coupling than native systems. Nucleoside 5'-triphosphates (NTPs) supported agonist-stimulated AC activity in the two systems and basal AC activity in membranes from cholera toxin-treated S49 cells in the order of efficacy GTP  UTP > CTP > ATP (ineffective). NTPs disrupted high affinity agonist binding in ₂AR-G_s in the order of efficacy GTP > UTP > CTP > ATP (ineffective). In contrast, the order of efficacy of NTPs as substrates for nucleoside diphosphokinase, catalyzing the formation of GTP from GDP and NTP was ATP  UTP  CTP  GTP. NTPs inhibited ₂AR-G_s-catalyzed [-³²P]GTP hydrolysis in the order of potency GTP > UTP > CTP. Molecular dynamics simulations revealed that UTP is accommodated more easily within the binding pocket of G_s than CTP. Collectively, our data indicate that GTP, UTP, and CTP interact differentially with G_s proteins and that transphosphorylation of GDP to GTP is not involved in this G protein activation. In certain cell systems, intracellular UTP and CTP concentrations reach ~10 nmol/mg of protein and are higher than intracellular GTP concentrations, indicating that G protein activation by UTP and CTP can occur physiologically. G protein activation by UTP and CTP could be of particular importance in pathological conditions such as cholera and Lesch-Nyhan syndrome.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

G proteins consist of an  subunit and a complex and serve as signal transducers between agonist-activated GPCRs¹ and effector systems (1-4). Upon binding of an agonist, GPCRs undergo a conformational change causing GDP dissociation from G. GDP dissociation is the rate-limiting step of the G protein cycle. Agonist-occupied GPCRs then form a ternary complex with the nucleotide-free G protein. The ternary complex possesses high agonist affinity. Subsequently, GPCRs promote binding of GTP to G. The binding of GTP to G induces the active conformation of the G protein, leading to the dissociation of the heterotrimer into G-_GTP and the complex. Both G-_GTP and can regulate the activity of effector systems. G possesses GTPase activity. The GTPase cleaves GTP into GDP and P_i and thereby deactivates the G protein. G-_GDP and reassociate, completing the G protein cycle.

Intriguingly, not only the purine nucleotide GTP but also pyrimidine nucleotides exhibit effects on G proteins. Particularly, various natural and synthetic uracil nucleotides disrupt the complex between the photoexcited light receptor rhodopsin and the retinal G protein transducin, but the uracil nucleotides are less effective in this regard than the corresponding guanine nucleotides (5). [-³²P]GTP hydrolysis and [³⁵S]GTPS binding competition studies showed that pyrimidine nucleotides bind to G proteins with low affinity (5-8). Moreover, early studies revealed that UTP and CTP support GPCR-mediated AC activation in membranes (9-11). However, it remained unclear whether the effects of UTP and CTP on AC were mediated via NDPK, catalyzing the formation of GTP from GDP and NTP (12, 13), or via direct interaction of UTP and CTP with G_s (6).

The aim of the present study was to elucidate the mechanism by which UTP and CTP support G_s activation. To achieve our aim we have studied AC regulation in S49 membranes. S49 cells are a widely used and physiologically relevant model system for the analysis of ₂AR/G_s/AC interactions (14-17). Additionally, we have studied fusion proteins of the ₂AR with individual G_s isoforms, i.e. ₂AR-G_sS, ₂AR-G_sL, and ₂AR-G_olf, expressed in Sf9 insect cells. Fusion proteins provide close proximity of the coupling partners and ensure efficient GPCR-G protein-effector coupling (18, 19). In addition, fusion proteins allow for the analysis of the coupling of a given GPCR to various G isoforms under defined experimental conditions (20-23). Here, we report on distinct interactions of GTP, UTP, and CTP with G_s proteins.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- The generation of baculoviruses encoding for ₂AR-G_sS, ₂AR-G_sL, and ₂AR-G_olf was described elsewhere (20, 23, 24). [-³²P]GTP (6,000 Ci/mmol), [-³²P]ATP (3,000 Ci/mmol), [³²P]P_i (8,500-9,000 Ci/mmol), [³H]DHA (85-90 Ci/mmol), and [³H]GDP (30-39 Ci/mmol) were from PerkinElmer Life Sciences. Unlabeled ATP (special quality, < 0.01% GTP as assessed by high performance liquid chromatography), GTP, UTP, and CTP were of the highest quality available and were obtained from Roche Molecular Biochemicals. [-³²P]UTP and [-³²P]CTP (~6,000 Ci/mmol each) were synthesized as described (5, 25). S49 cells were obtained from the Cell Culture Facility of the University of California at San Francisco. ISO, SAL, (±)-alprenolol and CTX were obtained from Sigma.

Cell Culture and Membrane Preparation-- Sf9 cells were cultured and infected with recombinant baculoviruses as described (20, 24, 26). S49 cells were grown at 37 °C in suspension in Dulbecco's modified Eagle's medium supplemented with 4.5 g/liter D-glucose, 2 mM L-glutamine, 1,000 units/ml penicillin, 100 µg/ml streptomycin, and 10% (v/v) heat-inactivated horse serum in a humidified atmosphere containing 7% (v/v) CO₂. S49 cells were maintained at a density of 0.2-2.0 × 10⁶ cells/ml. To inactivate the GTPase of G_s, S49 cells were treated with 1 µg/ml CTX for 24 h before membrane preparation (27). Dulbecco's modified Eagle's medium was from Cellgro Mediatech (Herndon, VA). All other constituents for the culture of S49 cells were obtained from BioWhittaker (Walkersville, MD). S49 and Sf9 membranes were prepared according to the protocol described previously (24). Membranes were suspended in binding buffer (12.5 mM MgCl₂, 1 mM EDTA, and 75 mM Tris-HCl, pH 7.4) at a concentration of ~1-2 mg of protein/ml and stored at -80 °C until use. Immediately prior to [³H]DHA binding, AC, NTPase, and NDPK experiments, membrane aliquots were thawed, suspended in binding buffer, and centrifuged for 15 min at 4 °C and 15,000 × g to remove, as far as possible, any remaining endogenous nucleotides (23).

[³H]DHA Binding-- The expression levels of ₂AR-G_sS, ₂AR-G_sL, and ₂AR-G_olf were determined by saturation binding using the ₂AR antagonist [³H]DHA (20, 24). 500-µl tubes contained Sf9 membranes (10-20 µg of protein/tube) expressing fusion proteins, 10 nM [³H]DHA, and binding buffer. Nonspecific binding was determined in the presence of 10 µM (±)-alprenolol. Incubations were performed for 90 min at 25 °C and shaking at 250 rpm. Bound [³H]DHA was separated from free [³H]DHA by filtration through GF/C filters (Schleicher & Schuell). For generation of concentration/response curves for the inhibitory effects of NTPs on high affinity agonist binding, reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing ₂AR-G_s fusion proteins, 1 µM SAL, 1 nM [³H]DHA as radioligand, and NTPs at increasing concentrations (26). For determination of the extent of ternary complex formation and its sensitivity to disruption by NTPs, reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing ₂AR-G_sL, 1 nM [³H]DHA as radioligand, and ISO at increasing concentrations in the absence and presence of NTPs at a fixed concentration (1 mM each) (23, 24).

AC Activity-- The determination of AC activity in S49 membranes and Sf9 membranes expressing ₂AR-G_s fusion proteins was performed under identical experimental conditions and followed the protocol published previously (23). Briefly, 30-µl tubes contained membranes (20-50 µg of protein/tube), 5 mM MgCl₂, 0.4 mM EDTA, 30 mM Tris-HCl, pH 7.4, and NTPs at various concentrations in the absence or presence of ISO. Tubes were incubated for 3 min at 37 °C before the addition of 20 µl of reaction mixture containing (final) [-³²P]ATP (1.0-1.5 µCi/tube) plus 40 µM ATP, 2.7 mM mono(cyclohexyl)ammonium phosphoenolpyruvate, 0.125 IU of pyruvate kinase, 1 IU of myokinase, and 0.1 mM cAMP. Reactions were conducted for 20 min at 37 °C. Stopping of reactions and separation of [-³²P]ATP from [³²P]cAMP were performed as described previously (23).

NTPase Activity-- High affinity GTPase activity in Sf9 membranes expressing ₂AR-G_s fusion proteins was determined as described (23). Briefly, tubes (80 µl) contained membranes (10 µg of protein/tube), 1.0 mM MgCl₂, 0.1 mM EDTA, 0.1 mM ATP, 1 mM adenylyl imidodiphosphate, 5 mM creatine phosphate, 40 µg of creatine kinase, 30 nM unlabeled GTP, 10 µM ISO, and 0.05% (mass/volume) bovine serum albumin in 50 mM Tris-HCl, pH 7.4. Reaction mixtures were incubated for 3 min at 25 °C before the addition of 20 µl of [-³²P]GTP (0.2 µCi/tube). Nonenzymatic [-³²P]GTP hydrolysis was determined in the presence of a large excess of unlabeled GTP (1 mM). Reactions were conducted for 20 min at 25 °C. Stopping of reactions and recovery of ³²P_i were performed as described (23).

UTPase and CTPase activities in Sf9 membranes were determined essentially as GTPase activity except that reaction mixtures contained [-³²P]UTP or [-³²P]CTP (up to 2.5 µCi/tube) instead of [-³²P]GTP. In addition, reaction mixtures contained unlabeled UTP or CTP (0.1-100 µM) instead of unlabeled GTP. Nonenzymatic UTP and CTP hydrolysis was determined in the presence of 1 mM unlabeled UTP and CTP, respectively.

NDPK Activity-- NDPK activity in S49 wild-type lymphoma membranes and Sf9 membranes was determined as described for soluble transducin preparations with modifications (28). Briefly, reaction mixtures contained S49 membranes (5.0-10.0 µg of protein/tube) or Sf9 membranes (0.5-1.0 µg of protein/tube), NTPs at various concentrations, 0.1% (mass/volume) bovine serum albumin, 1 mM MgCl₂, and 0.1 mM EDTA in 50 mM Tris-HCl, pH 7.4. Reaction mixtures were preincubated for 3 min at 37 °C before the addition of 0.5 µM carrier-free [³H]GDP. The total reaction volume was 50 µl. Reactions were conducted for 10 min. To obtain blank values, tubes containing all components described above except for membranes were processed in parallel. Stopping of reactions, separation of nucleotides on poly(ethyleneimine)-cellulose TLC plates (Schleicher & Schuell), elution of nucleotides, and counting of radioactivity were performed exactly as described (28).

Molecular Modeling-- Potential energy minimization and molecular dynamics simulations were carried out using the AMBER 6.0 program package (29), using the force field of Cornell et al. (30). Initial models for G_s·Mg²⁺·NTP complexes were based on the coordinates of G_s·Mg²⁺·GTPS in the complex with the catalytic domains of AC (PDB 1AZS) (31). We replaced the S substituent of GTPS by an sp2-hybridized oxygen atom. Models of complexes with UTP and CTP were generated by replacing the guanine base with uracil and cytosine, respectively, maintaining the glycosyl torsion angle (O4'-C1'-N9-C8) of GTPS in the starting model. Partial charges assigned to phosphate groups were obtained from Dr. N. Duclert-Savatier (Institut Pasteur, Paris, France).² Protein-bound water molecules observed in the crystal structure were included in the model. The nonbonded cutoff distance was set at 10 Å. The stereochemistry of the Mg²⁺ ligand field was restrained as an octahedral complex with the six coordinating oxygen atoms of the - and -phosphate oxygens, the hydroxyl groups of Ser-54 (G_sL and G_sS) and Thr-204 (G_sL) and two water molecules, with oxygen-Mg²⁺ distances of 2.1 Å. This geometry was maintained by pseudo-van der Waals potentials among all pairs of atoms within the octahedral complex. In constructing models of the G_s·Mg²⁺·UTP and G_s·Mg²⁺·CTP complexes, we placed a water molecule at the site occupied by the guanine exocyclic C (2) amine in the GTP complex. The included water molecule bridges the uracil exocyclic C (2) keto with the O1 carboxylate oxygen of Asp-295 (G_sL) and Asp-280 (G_sS). The water molecule placed at this site did not move after energy minimization of the model complexes.

Energy relaxations were carried out using the SANDER module of AMBER, using steepest descent minimization for the first 250 cycles, followed by 250 cycles of conjugate gradient minimization. In all cases the total computed potential energy typically declined smoothly from values of 14,500 to 16,750 kcal·mol¹, attaining a constant value after 400-450 cycles of minimization. For molecular dynamics simulations, a box of TIP3P water molecules was used to solvate the protein, leaving a 10 Å border between the edge of the box and the closest atoms of the protein. The system was heated to 300 K using the temperature scaling scheme of Berendsen et al. (32) and periodic boundary conditions. Simulations were carried out for 10 ps in steps of 1 fs.

Miscellaneous-- Protein was determined using the Bio-Rad DC protein assay kit. Data shown in Figs. 1-5 and Table I were analyzed by nonlinear regression, using the Prism III program (GraphPad, Prism, San Diego).

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

UTP, CTP, and GTP Differentially Support AC Activation in S49 Membranes-- S49 cells express the ₂AR, the G_s splice variants G_sS and G_sL, and AC (17, 33, 34). In the absence of added UTP, CTP, or GTP, the full AR agonist ISO at a maximally stimulatory concentration (10 µM) had no stimulatory effect on AC activity in S49 membranes (Fig. 1). However, these experimental conditions do not imply the complete absence of NTP because the AC assay contained 40 µM ATP as AC substrate (see "Experimental Procedures"). UTP, CTP, and GTP had little effects on basal AC activity in the absence of ISO. In agreement with the literature (1-4), GTP was potent (EC₅₀, 230 nM; 95% c.i., 150-340 nM) and effective at supporting AC activation by ISO (Fig. 1C). UTP was much less potent (EC₅₀, 82 µM, 95% c.i., 43-160 µM) in this regard than GTP, but it was only moderately less efficient (~80% efficacy) than GTP (Fig. 1A). CTP only poorly supported AC activation by ISO, both in terms of potency and efficacy (Fig. 1B).

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Effects of UTP, CTP, and GTP on AC activity in S49 wild-type lymphoma membranes. AC activity in S49 wild-type (wt) lymphoma membranes was determined as described under "Experimental Procedures." Reaction mixtures contained S49 membranes (20-50 µg of protein/tube) and NTPs at the concentrations indicated on the abscissa with solvent (basal) () or with 10 µM ISO (). 109 designates the absence of added UTP, CTP, or GTP. In additional experiments, reaction mixtures contained membranes from S49 cells that had been treated with 1 µg/ml CTX for 24 h before membrane preparation. The experiments with membranes from CTX-treated cells were only conducted with solvent (basal) (). The basal AC activity in membranes from untreated S49 cells in the absence of NTP and ISO amounted to 0.48 ± 0.51 pmol/mg/min and was set as 0%. The AC activity in untreated S49 membranes in the presence of 10 µM GTP and 10 µM ISO was set as 100% (control) and amounted to 2.36 ± 0.75 pmol/mg/min. All other AC activities were referred to those calibration points. Data were analyzed by nonlinear regression analysis and were best fitted to sigmoidal concentration/response curves. Data shown are the means ± S.D. of six to eight experiments performed in duplicate.

CTX-catalyzed ADP-ribosylation of Arg-201 (G_sL) and Arg-186 (G_sS) blocks the GTPase of G_s (1, 27, 35). As a result, GTP, like the GTPase-resistant GTP analog GTPS, becomes an efficient AC activator even in the absence of GPCR agonist (1, 27, 35). In fact, in membranes from CTX-treated S49 cells, GTP was far more efficient at activating AC than were GTP plus ISO in control membranes (Fig. 1C). However, CTX did not increase the potency of GTP (EC₅₀, 560 nM; 95% c.i., 310-1,000 nM). As was the case with GTP, CTX greatly increased the efficacy of UTP at activating AC (Fig. 1A). The efficacy of UTP at activating AC in membranes from CTX-treated S49 cells amounted to ~70% of the efficacy of GTP. In contrast to the data obtained for GTP, CTX increased the potency of UTP ~4-fold (EC₅₀, 21 µM; 95% c.i., 11-38 µM). Although CTX also substantially enhanced AC activation by CTP, this NTP was, nonetheless, much less efficient than UTP and GTP (Fig. 1B). Thus, CTX greatly amplifies the maximum effects of NTPs on basal AC activity without altering their relative efficacies.

Rationale for Conducting Further Studies with Sf9 Membranes Expressing ₂AR-G_s Fusion Proteins-- We wished to answer the questions of whether UTP and CTP, like GTP, disrupt the ternary complex consisting of agonist-occupied ₂AR and nucleotide-free G_s, whether UTP and CTP inhibit [-³²P]GTP hydrolysis by G_s, and whether G_s hydrolyzes [-³²P]UTP and [-³²P]CTP. However, the extent of ternary complex formation in S49 membranes is rather limited (17). Thus, we were concerned that this system would not be sensitive enough to dissect potentially small differences in efficacies of NTPs on high affinity agonist binding. In addition, S49 membranes are not sensitive models for the analysis of the GTPase activity of G_s (36). We were also interested in answering the question of whether the different G_s isoforms, i.e. G_sS, G_sL, and G_olf, respond similarly to UTP and CTP. However, S49 cells express a mixture of G_sS and G_sL, and the sensitivity of the G_s-deficient S49 cyc cells as reconstitution system for the planned studies is limited too (33, 34).

Sf9 membranes expressing ₂AR-G_s fusion proteins possess a sufficiently high sensitivity for dissecting differential effects of NTPs on ternary complex formation, surpassing the sensitivity of native and recombinant nonfused systems (17, 24, 26). Additionally, ₂AR-G_s fusion proteins are sensitive models for NTPase studies (24, 26). Moreover, ₂AR-G_s fusion proteins are suitable systems for dissecting biochemical differences between G_s isoforms (20, 22, 23). Based on these considerations, we decided to conduct all further studies with ₂AR-G_s fusion proteins.

UTP, CTP, and GTP Differentially Disrupt the Ternary Complex in Sf9 Membranes Expressing ₂AR-G_sS, ₂AR-G_sL, and ₂AR-G_olf-- We examined binding of a fixed concentration of the antagonist [³H]DHA (1 nM) to fusion proteins in the presence of an agonist (SAL) at a fixed subsaturating concentration (1 µM) and NTPs at increasing concentrations. NTPs decrease the affinity of the ₂AR for SAL and, as a result, increase [³H]DHA binding (20, 23, 26). GTP potently and efficiently disrupted the ternary complex in membranes expressing ₂AR-G_sS (EC₅₀, 180 nM; 95% c.i., 110-310 nM), ₂AR-G_sL (EC₅₀, 90 nM; 95% c.i., 50-150 nM), and ₂AR-G_olf (EC₅₀, 590 nM; 95% c.i., 310-1,100 nM) (Fig. 2). The efficacy of UTP at disrupting the ternary complex in ₂AR-G_s fusion proteins amounted to 53-61% of the efficacy of GTP. UTP was far less potent than GTP in membranes expressing ₂AR-G_sS (EC₅₀, 30 µM; 95% c.i., 9-101 µM), ₂AR-G_sL (EC₅₀, 12 µM; 95% c.i., 2-69 µM), and ₂AR-G_olf (EC₅₀, 19 µM; 95% c.i., 7-62 µM). The efficacy of CTP at disrupting the ternary complex amounted to less than 25% of the efficacy of GTP, preventing us from calculating meaningful EC₅₀ values for CTP. Of particular importance was the finding that ATP did not disrupt the ternary complex in membranes expressing ₂AR-G_sS or ₂AR-G_olf.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Effects of UTP, CTP, GTP, and ATP on ternary complex formation in Sf9 membranes expressing 2AR-GsS, 2AR-GsL, or 2AR-Golf. Concentration/response curves for NTPs are shown. [3H]DHA binding in Sf9 membranes was performed as described under "Experimental Procedures." Reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing 2AR-GsS, 2AR-GsL, or 2AR-Golf, 1 nM [3H]DHA, 1 µM SAL, and NTPs (, GTP; , UTP; , CTP; , ATP) at the concentration indicated on the abscissa. 1010 designates the absence of added NTP. A, membranes expressing 2AR-GsS at 2.6-4.4 pmol/mg; B, membranes expressing 2AR-GsL at 3.0-4.8 pmol/mg; C, membranes expressing 2AR-Golf at 3.3-4.2 pmol/mg. Data were analyzed by nonlinear regression and were best fitted to sigmoidal concentration/response curves. Data shown are the means ± S.D. of three to five independent experiments performed in triplicate.

To corroborate the conclusion that the order of efficacy of NTPs at disrupting the ternary complex is GTP > UTP > CTP, we determined the extent of ternary complex formation in ₂AR-G_sL by competing [³H]DHA binding by the full agonist ISO in the absence and presence of NTPs at a saturating concentration (1 mM). Fig. 3 shows the agonist competition curves, and Table I summarizes the nonlinear regression analysis of these binding experiments. As reported before (20, 24), ISO inhibited [³H]DHA binding according to a biphasic function, with ~35% of the ₂ARs being in a state of high agonist affinity (Fig. 3A). GTP (1 mM) substantially shifted the ISO competition curve to the right and converted the competition curve into a steep monophasic function, reflecting complete disruption of the ternary complex. In contrast, in the presence of 1 mM UTP, the agonist competition curve was still biphasic (~10% high affinity agonist binding remaining), and the low affinity component of the competition curve was not shifted as far to the right as with GTP (Fig. 3B). These data confirm the notion that UTP disrupted the ternary complex only incompletely. CTP had only minimal inhibitory effects on ternary complex formation in membranes expressing ₂AR-G_sL, i.e. the fraction of high affinity binding sites was not decreased compared with control conditions (Fig. 3C and Table I). In addition, CTP increased the K_i values for ISO only slightly relative to control. Taken together, the analysis of the effects of NTPs on ternary complex formation in membranes expressing ₂AR-G_s fusion proteins clearly showed that the order of efficacy is GTP > UTP > CTP > ATP (ineffective).

View larger version (12K): [in this window] [in a new window]   Fig. 3.   Effects of UTP, CTP, and GTP on ternary complex formation in Sf9 membranes expressing 2AR-GsL. Concentration/response curves for ISO are shown. [3H]DHA binding in Sf9 membranes was performed as described under "Experimental Procedures." Reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing 2AR-GsL at 4.5-7.4 pmol/mg. Reaction mixtures additionally contained ISO at the concentrations indicated on the abscissa in the presence of solvent (control) () or various NTPs () at a concentration of 1 mM each. 1010 designates the absence of ISO. Data were analyzed for best fit to monophasic or biphasic competition isotherms (F test). Data shown are the means ± S.D. of five independent experiments performed in triplicate. The dashed lines without data points shown in B and C represent the competition isotherms in the presence of solvent (control) and 1 mM GTP depicted in A to illustrate the relative position of the isotherms in the presence of UTP and CTP.

View this table: [in this window] [in a new window]   Table I Nonlinear regression analysis of the effects of GTP, UTP, and CTP on ternary complex formation in Sf9 membranes expressing 2AR-GsL [3H]DHA binding in Sf9 membranes was performed as described under "Experimental Procedures." Reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing 2AR-GsL at 4.5-7.4 pmol/mg. Reaction mixtures additionally contained ISO at the concentrations indicated on the abscissa of Fig. 3 in the presence of solvent (control) or various NTPs at a concentration of 1 mM each. The data shown in Fig. 3 were analyzed for best fit to monophasic or biphasic competition isotherms (F test). Data shown are the means of five independent experiments performed in triplicate. Numbers in parentheses represent the 95% confidence intervals. Kh and Kl designate the dissociation constants for the high and low affinity state of 2AR-GsL, respectively. Rh indicates the percentage of high affinity binding sites.

Competition of ₂AR-G_s-catalyzed [-³²P]GTP Hydrolysis by GTP, UTP, and CTP-- To address the question of whether UTP and CTP bind to the nucleotide binding pocket of G_s, we stimulated G_s-catalyzed [-³²P]GTP hydrolysis in ₂AR-G_s fusion proteins by ISO and competed [-³²P]GTP hydrolysis with unlabeled NTPs. For all three ₂AR-G_s fusion proteins we obtained monophasic competition curves, indicating that [-³²P]GTP and NTPs competed for binding to a single site (Fig. 4). The K_i values for GTP at the three fusion proteins ranged between 210 and 490 nM. The K_i values for UTP ranged between 170 and 910 µM, and those for CTP ranged between 3.4 and 4.4 mM. Thus, the order of affinity of G_sS, G_sL, and G_olf for NTPs is GTP > UTP > CTP.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Competition of [-32P]GTP hydrolysis in Sf9 membranes expressing 2AR-Golf, 2AR-GsS, or 2AR-GsL by GTP, UTP, and CTP. GTPase activity in Sf9 membranes was determined as described under "Experimental Procedures." Reaction mixtures contained Sf9 membranes (10 µg of protein/tube) expressing various fusion proteins, 30 nM [-32P]GTP, 10 µM ISO, and unlabeled GTP (), UTP (), or CTP () at increasing concentrations. 108 designates the absence of unlabeled UTP, CTP, or GTP. GTPase activities in the absence of competitor (control) were as follows. 2AR-Golf (expressed at 13.7 pmol/mg), 8.5 ± 0.4 pmol/mg/min; 2AR-GsS (expressed at 4.0 pmol/mg/min), 3.3 ± 0.3 pmol/mg/min; 2AR-GsL (expressed at 6.0 pmol/mg), 5.2 ± 0.6 pmol/mg/min. These GTPase activities were defined as 100% (control). Competition curves were extended until complete inhibition of GTP hydrolysis in Sf9 membranes. This point was defined as 0%. All other data points were referred to those calibration points. Data were analyzed by nonlinear regression and were best fitted to monophasic competition curves. Data shown are the means ± S.D. of three independent experiments performed in duplicate.

We also addressed the question of whether ₂AR-G_s fusion proteins hydrolyze [-³²P]UTP and [-³²P]CTP. However, despite using high amounts of [-³²P]NTPs/tube (up to 2.5 µCi/tube) and low basal UTPase and CTPase activities in Sf9 membranes, we could not detect ISO-stimulated UTP and CTP hydrolysis in Sf9 membranes expressing any of the three fusion proteins using UTP and CTP concentrations between 0.1 and 100 µM (data not shown).

UTP, CTP, and GTP Differentially Support AC Activation by ₂AR-G_sS, ₂AR-G_sL, and ₂AR-G_olf Expressed in Sf9 Membranes-- In the absence of GTP, UTP, or CTP, i.e. in the presence of ATP alone, membranes expressing ₂AR-G_sS exhibit a higher basal AC activity than membranes expressing ₂AR-G_sL and ₂AR-G_olf, and ISO efficiently reduced AC activity in membranes expressing ₂AR-G_sS but not in membranes expressing ₂AR-G_sL or ₂AR-G_olf (compare Fig. 5, A, D, and G). These differences are explained by the fact that ₂AR-G_sS possesses a higher GDP affinity than ₂AR-G_sL and ₂AR-G_olf (20, 23). Accordingly, ISO efficiently promotes GDP dissociation from G_sS and thereby reduces the concentration of G_sS-GDP. Because G_s-GDP is more efficient at activating AC than nucleotide-free G_s, ISO reduces AC activity in the absence of GTP, UTP, or CTP. In membranes expressing any of the three fusion proteins, NTPs supported ISO stimulation of AC in the order of efficacy GTP ~ UTP > CTP, and UTP and CTP were considerably less potent than GTP (Fig. 5, A-I). Although there are certain differences in the regulation of AC by NTPs between S49 and Sf9 membranes with respect to the effect of ISO in the presence of ATP alone and the stimulatory effects of NTPs on basal AC activity, the overall pattern of AC regulation by NTPs is similar. Specifically, the order of efficacy of NTPs in all systems is GTP  UTP > CTP, and pyrimidine nucleotides are less potent than GTP. Fusion proteins facilitate detection of UTP and CTP effects on G_s because the relative efficacies of CTP and particularly UTP with respect to ISO stimulation of AC are larger with ₂AR-G_s fusion proteins than with S49 membranes.

View larger version (28K): [in this window] [in a new window]   Fig. 5.   Effects of UTP, CTP, and GTP on AC activity in Sf9 membranes expressing 2AR-GsS, 2AR-GsL, or 2AR-Golf. AC activity in Sf9 membranes was determined as described under "Experimental Procedures." Reaction mixtures contained Sf9 membranes (20 µg of protein/tube) expressing 2AR-GsS, 2AR-GsL, or 2AR-Golf, and NTPs at the concentrations indicated on the abscissa with solvent (basal) () or with 10 µM ISO (). 109 designates the absence of added UTP, CTP, or GTP. A-C, membranes expressing 2AR-GsS at 2.3-2.6 pmol/mg; D-F, membranes expressing 2AR-GsL at 4.8-5.4 pmol/mg; G-I, membranes expressing 2AR-Golf at 3.6-4.1 pmol/mg. Data were analyzed by nonlinear regression and were best fitted to sigmoidal concentration/response curves. Data shown are the means ± S.D. of three to five independent experiments performed in duplicate. Note that the scale of the ordinate in A-F is different from the scale in G-I. The different scales were chosen to facilitate comparison of the relative effects of NTPs at the various fusion proteins. Differences in the absolute efficacies of 2AR-Gs fusion proteins at activating AC were reported before (20, 23).

We observed certain differences in the effects of UTP and CTP at the various fusion proteins in the AC assay. Specifically, the maximum agonist-stimulated AC activities achieved with CTP in membranes expressing ₂AR-G_sS and ₂AR-G_sL amounted to ~65% of that obtained with UTP, whereas in membranes expressing ₂AR-G_olf the maximum AC activity achieved with CTP amounted to only ~40% of the activity obtained with UTP (compare Fig. 5, A with B, D with E, and G with H, respectively). There were also differences in the relative stimulatory effects of ISO in the individual fusion proteins in the presence of different NTPs. In membranes expressing ₂AR-G_sS, the maximum stimulatory effects of ISO amounted to 74% (UTP), 55% (CTP), and 57% (GTP). The corresponding values for ₂AR-G_sL were 74% (UTP), 115% (CTP), and 60% (GTP), respectively. For ₂AR-G_olf the stimulations were 195% (UTP), 77% (CTP), and 206% (GTP).

NDPK Activity in S49 and Sf9 Membranes-- In S49 membranes, GTP was the most potent (EC₅₀ ~ 1 µM) but least efficient NDPK substrate (Fig. 6A). UTP and ATP were less potent (EC₅₀ ~ 5 µM) but more efficient as phosphoryl group donors than GTP. CTP was less potent (EC₅₀ ~ 10 µM) than UTP and ATP at serving as NDPK substrate but similarly efficient. Although GTP exhibited similar potency (EC₅₀ ~1 µM) and relative efficacy as the NDPK substrate in Sf9 and S49 membranes (compare Fig. 6, A and B), the properties of ATP, UTP, and CTP were different in the two systems. The order of efficacy of NTPs in Sf9 membranes was ATP > UTP > CTP, and the order of potency was ATP (EC₅₀ ~ 5 µM) > UTP (EC₅₀ ~10 µM) > CTP (EC₅₀ ~ 20 µM).

View larger version (16K): [in this window] [in a new window]   Fig. 6.   [3H]GTP formation from [3H]GDP and NTP in S49 wild-type (wt) lymphoma membranes and Sf9 membranes. [3H]GTP formation from [3H]GDP and NTP was determined as described under "Experimental Procedures." Reaction mixtures contained S49 wild-type lymphoma membranes (5.0-10.0 µg of protein/tube) (A) or membranes from Sf9 cells infected with 2AR-Golf baculovirus (0.5-1.0 µg of protein/tube) (Sf9 inf., B), 0.5 µM carrier-free [3H]GDP and NTPs at the concentrations indicated on the abscissa. Data shown are the means ± S.D. of three experiments performed in duplicate. Note that the scale of the ordinate in A and B is different. The different scales were chosen to facilitate comparison of the relative effects of NTPs in the two membrane systems.

Molecular Modeling of the Interactions of GTP, UTP, and CTP with G_s-- To provide an explanation for the different affinities and efficacies of NTPs at G_s proteins we constructed models for the GTP, CTP, and UTP complexes of G_s based on the structure of the G_s·Mg²⁺·GTPS complex (31) and subjected the complexes to potential energy minimization. However, none of the energy-minimized G_s·Mg²⁺·NTP models differed substantially from the corresponding G_s·Mg²⁺·GTPS complex (0.15-0.18Å root mean square deviation between pairs of corresponding C atoms) (data not shown). Furthermore, the position, orientation, and conformation of NTP·Mg²⁺ were unaltered after energy minimization. Hence, energy minimization experiments did not provide an explanation for the differences in affinity and efficacy of NTPs at G_s.

We then decided to investigate the molecular models of the GTP, UTP, and CTP complexes of G_s by molecular dynamics simulation. After a 10-ps simulation at 300 K, the G_s·Mg²⁺·GTP, G_s·Mg²⁺·UTP, and G_s·Mg²⁺·CTP models diverged from the corresponding G_s·Mg²⁺·GTPS complex by 0.94, 0.95, and 1.15 Å, root mean square, respectively, over all C pairs. The deviations were smaller (0.6-0.8 Å) for the set limited to C atoms within 6 Å of the NTPs. The models did not change further after 500 cycles of energy minimization. All of the guanine ring-G_s hydrogen bonds and van der Waals interactions observed in the structure of the G_s·Mg²⁺·GTPS complex (31, 37) were retained in the model of the G_s·Mg²⁺·GTP complex after molecular dynamics simulation (Fig. 7A), although the ribosyl group shifted ~1 Å. Likewise, the molecular geometry of the triphosphate moiety and its interaction with Mg²⁺ were unperturbed, indicating that the simulation with GTP preserves experimentally observed properties of the interaction of G_s with GTPS.

View larger version (16K): [in this window] [in a new window]   Fig. 7.   Three-dimensional model of the interaction of GTP, UTP, and CTP with the nucleotide binding pocket of Gs. Molecular dynamics simulations were performed as described under "Experimental Procedures." The environments of the NTP bases of Gs·Mg2+·NTP complexes after molecular dynamics simulation are shown. A, Gs·Mg2+·GTP complex. For clarity, Asn-292 (GsL), which forms a hydrogen bond with the N(7) imine(31), is not shown. B, Gs·Mg2+·UTP complex. For clarity, the side chain of Val-367 (GsL) is not shown. C, Gs·Mg2+·CTP complex. Carbon atoms of Gs are colored orange, carbon atoms of the NTPs are colored green. Nitrogen and oxygen atoms in the protein and NTPs are colored blue and red, respectively. Selected hydrogen bonds are depicted as green dotted lines. The isolated red sphere represents the included water molecule. In B and C, the Gs·Mg2+·UTP and Gs·Mg2+·CTP complexes, respectively, are superimposed on the Gs·Mg2+·GTP complex that is shown as charcoal stick model.

In contrast, the molecular dynamics simulations of the G_s·Mg²⁺·UTP and G_s·Mg²⁺·CTP complexes had different outcomes. The position and orientation of the uracil ring of UTP remained near the starting position after 10 ps of dynamics simulation (Fig. 7B). The N3 imine stayed in position to donate a hydrogen bond to the O2-carboxylate oxygen of Asp-295, and the hydrogen bond network involving the 2-ketooxygen, the imposed water molecule, and O1 of Asp-295 were intact, although with suboptimal geometry. In contrast, the cytosine ring of CTP rotated ~10° from its initial orientation (Fig. 7C). The rotation was accommodated, without substantial change in ribose ring pucker or in the position of the triphosphate moiety, by a 30° rotation around the ribosyl C(5')-O(5') bond. By so rotating, the cytosine ring avoided steric conflict with the side chain of Asn-292 (G_sL) and Asn-277 (G_sS) (not shown). Val-367 (G_sL) and Val-352 (G_sS) also moved away from the cytosine ring (Fig. 7C). As another consequence of the rotation, the cytosine C(4) exocyclic amine was in position to donate a hydrogen bond to Asp-295; this interaction is not possible in the starting structure. However, neither the N(3) imine nor the C(2) exocyclic ketooxygen atom of the cytosine ring was within hydrogen bonding distance of the imposed water molecule, which remained bonded to Asp-295.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Transphosphorylation of GDP to GTP Cannot Explain the Effects of UTP and CTP on G_s Proteins-- UTP and CTP disrupt the ternary complex between the ₂AR and G_s proteins (Figs. 2 and 3 and Table I) and support AC activation by unmodified and ADP-ribosylated G_s (Figs. 1 and 5). An explanation for those observations could be transphosphorylation of endogenous GDP to GTP by UTP or CTP via NDPK (12, 13, 38), but several findings argue against this notion. First, the concentration of G_sS-GDP in washed Sf9 membranes is higher than the concentration of G_sL-GDP and G_olf-GDP (20, 23). Thus, for transphosphorylation, membranes expressing ₂AR-G_sS should have been a considerably more efficient system than membranes expressing ₂AR-G_sL and ₂AR-G_olf. However, UTP and CTP were not more efficient at disrupting the ternary complex and supporting AC activation in membranes expressing ₂AR-G_sS than in membranes expressing ₂AR-G_sL and ₂AR-G_olf (Figs. 2 and 5). Second, a given NTP should have had the same relative efficacies in the AC and high affinity agonist binding assays. However, UTP was less efficient than GTP at disrupting the ternary complex in ₂AR-G_s fusion proteins (Figs. 2 and 3 and Table I). In contrast, UTP and GTP were equally efficient at supporting ISO stimulation of AC in Sf9 membranes expressing ₂AR-G_s fusion proteins (Fig. 5). Third, in S49 and Sf9 membranes, ATP was a potent (EC₅₀ ~5 µM) and efficient phosphoryl group donor for GTP formation (Fig. 6), but we failed to detect a stimulatory effect of ISO on AC activity in S49 and Sf9 membranes in the presence of ATP (40 µM) (Figs. 1 and 5). Fourth, CTP and UTP were similarly efficient NDPK substrates in S49 membranes (Fig. 6A), but CTP was much less efficient than UTP at supporting AC activation in this system (Fig. 1, A and B). Finally, in S49 and Sf9 membranes UTP was a more efficient phosphoryl group donor for [³H]GTP formation than GTP (Fig. 6), but UTP was not more efficient than GTP with respect to disruption of the ternary complex and AC activation.

Evidence That GTP, UTP, and CTP Stabilize Distinct Conformations of G_s-- We then considered the hypothesis that UTP and CTP exert their effects on AC and ternary complex formation directly by binding to the nucleotide binding pocket of G_s. Indeed, UTP and CTP have already been shown to bind with low affinity to various G proteins including G_s (5-8). In agreement with the results of the earlier studies, our GTPase competition studies with ₂AR-G_s fusion proteins showed that NTPs bind to G_sS, G_sL, and G_olf in the order of affinity GTP > UTP > CTP (Fig. 4). We noted that the apparent affinities of UTP and CTP in the agonist binding and AC studies with fusion proteins (Figs. 2 and 5) were considerably higher than in the GTPase competition studies (Fig. 4). An important difference between these experiments is that the GTPase studies were conducted in the presence of GTP, whereas the agonist binding and AC studies were conducted in the absence of GTP (see "Experimental Procedures"). These data raise the intriguing hypothesis that in the presence of GTP, access of UTP and CTP to G_s is restricted. Evidence for restricted access of nucleotides to G proteins was already obtained in a previous study (39).

If the NTPs studied had stabilized the same conformation in G_s we would have expected NTPs to exhibit the same maximum effects on ternary complex formation and AC activation at saturating concentrations. However, this was clearly not the case. The overall order of efficacy of NTPs with respect to these parameters was GTP  UTP > CTP (Figs. 1-3 and 5). The enhancing effects of CTX on maximum AC activation by NTPs (Fig. 1) are particularly intriguing. CTX unmasks a strong stimulatory effect of GTP on AC by blocking the GTPase activity of G_s (1, 27, 35). Our present data are in agreement with this concept (Fig. 1C). CTX also greatly enhanced the stimulatory effects of UTP and CTP on AC activity (Fig. 1, A and B), suggesting that G_s hydrolyzes UTP and CTP as well. However, we could not obtain evidence for the presence of UTPase and CTPase activity in ₂AR-G_s proteins, although Sf9 membranes expressing these proteins are highly sensitive systems in detecting NTPase activity (24, 26). These data imply that the mechanism of deactivation of G_s-UTP and G_s-CTP is NTP dissociation rather than NTP hydrolysis. Our data indicate that CTX-catalyzed ADP-ribosylation induces a conformational change in G_s which is independent of the well established GTPase inhibition and allows UTP and CTP to interact more productively with the G protein. In support of this hypothesis is the fact that CTX-catalyzed ADP-ribosylation of G_s increased the apparent affinity of the G protein for UTP (Fig. 1A). In contrast to UTP, we did not observe an increase of affinity of G_s for GTP after CTX-catalyzed ADP-ribosylation (Fig. 1C). Taken together, our data indicate that GTP, UTP, and CTP interact with G_s proteins in nonidentical fashions.

The molecular dynamics simulations are consistent our experimental data. Specifically, the simulation experiments indicate that CTP binds with lower affinity to G_s than UTP because optimal interactions require a conformational change in the protein with no net gain in stabilizing interactions relative to UTP (Fig. 7). In fact, CTP binds to G_s proteins with lower affinity than UTP (Fig. 4). In addition, UTP and CTP are expected to bind to G_s with lower affinity than GTP because they form fewer stabilizing hydrogen bonds to the G protein. In particular, hydrogen bonds analogous to that between the guanine N(7) and Asn-292 (31, 37, 40) cannot be formed with uracil and cytosine. The experimental data are in agreement with the modeling studies (Figs. 1, 2, 4, and 5).

Compared with CTP, UTP is more easily accommodated within the binding site of G_s and forms more hydrogen bonds with the protein, assuming the participation of a water molecule captured from the solvent. The differences in interactions of GTP, UTP, and CTP with G_s could be interpreted in such a way that each of the NTPs stabilizes a distinct conformation of G_s or that certain NTPs stabilize less productive states of the G protein than GTP. Specifically, G_s-GTP possesses a conformation that is highly efficient at disrupting the high affinity interaction with the agonist-occupied ₂AR and at activating AC, whereas G_s-CTP possesses a conformation that is inefficient in these regards. Of particular interest, G_s-UTP possesses a conformation that is as efficient as G_s-GTP at activating AC (Fig. 5) but less efficient than G_s-GTP at disrupting the high affinity interaction with the agonist-occupied ₂AR (Figs. 2 and 3 and Table I). In agreement with our present data on G_s, guanine nucleotides are more efficient at disrupting the complex between photoexcited rhodopsin and transducin than the corresponding uracil nucleotides (5). These data indicate that G proteins do not simply act as on/off switches but rather exist in states with different functional capacities that are differentially stabilized by NTPs.

Although the overall pattern of the effects of GTP, UTP, and CTP on the various ₂AR-G_s fusion proteins was similar (Figs. 2 and 5), we noted that CTP was particularly effective at supporting ISO stimulation of AC in Sf9 membranes expressing ₂AR-G_sL (Fig. 5E). These data indicate that there are subtle differences in the conformations of G_sS-CTP, G_sL-CTP, and G_olf-CTP, exhibiting different efficacies at activating AC relative to the corresponding conformations of G_s-GTP and G_s-UTP. Studies with hydrolysis-resistant phosphorothioate analogs of UTP and CTP will be important to substantiate the concept of distinct active G_s conformations.

Physiological and Pathological Relevance of G Protein Activation by Pyrimidine Nucleotides-- Differential G_s activation by NTPs was observed in Sf9 membranes expressing ₂AR-G_s fusion proteins (Figs. 2, 3, and 5) and in a physiological system, the S49 membranes. The effects of UTP and CTP on fused and nonfused G_s proteins were observed at concentrations  10 µM (Figs. 1-5). The bulk intracellular UTP and CTP concentrations in various neuronal and astroglial cells may be as high as ~10 nmol/mg of protein and may e