Introduction

Top Abstract Introduction Materials and Methods Results and Discussion References

Heterotrimeric G proteins are activated by seven transmembrane spanning receptors, which promote the dissociation of GDP and thereby allow intracellular GTP to bind. G proteins are deactivated by the hydrolysis of GTP to GDP and inorganic phosphate, and the rate of that reaction can be increased up to 2 orders of magnitude by GTPase activating proteins (GAPs) (Wilkie and Ross, 2000). Some effectors act as GAPs toward the G proteins that activate them; for example, phospholipase C (Chidiac and Ross, 1999) and p115Rho-GEF (Kozasa et al., 1998), respectively, increase the rates of GTP hydrolysis by G_q and G₁₂. However, most GAPs for heterotrimeric G proteins have no known effector function and belong to the regulators of G protein signaling (RGS) protein family. Most RGS proteins are GAPs for members of the G_i family (Hepler, 1999), and a subset of those also act as GAPs for G_q (Ingi et al., 1998; Scheschonka et al., 2000).

The most direct way to measure RGS GAP activities is via increases in the rates at which their targeted G proteins hydrolyze GTP. With isolated G subunits, GTP turnover is limited by the dissociation of GDP, and therefore the rate of GTP hydrolysis needs to be measured under pre-steady-state conditions using G protein prebound to GTP (Wang et al., 1998a). Alternatively, GDP dissociation is not rate-limiting in the presence of activated receptors, and RGS GAP effects therefore can also be observed under steady-state conditions with receptor-coupled G proteins in the presence of agonist (Wang et al., 1998a). Although the pre-steady-state method is advantageous in that it allows changes in G protein GTP hydrolysis rate constants to be quantified unambiguously and also excludes the possible confounding effects of auxiliary proteins, the steady-state method may be more biologically relevant. As an alternative to assaying GTP hydrolysis, evidence of RGS protein GAP activity can also be observed via downstream events such as second messenger generation or mitogen-activated protein kinase activity; however, changes in these endpoints can also arise from other processes, such as steric effects or competition between RGS and effector proteins for activated G proteins (Hepler et al., 1997).

RGS2 is unique among the RGS proteins in its apparent selectivity toward G_q. Indeed, initial findings showed that the pre-steady-state GTPase activities of isolated G_i family members (G_i1 and G_o) were not stimulated by RGS2 (Ingi et al., 1998), even at protein concentrations up to 3000-fold higher than necessary to detect the effects of RGS4 (Heximer et al., 1997). This lack of effect can be attributed to three amino acid substitutions in the RGS domain of RGS2, because an RGS2 triple mutant (C106S,N184D,E191K) containing the corresponding residues from RGS4 increased the rate of hydrolysis by G_i to the same extent that RGS4 did (Heximer et al., 1999). Notwithstanding its failure to act as a GAP on isolated G_i, however, wild-type RGS2 has been found to inhibit cellular signaling events downstream of receptor-activated G_i (Herlitze et al., 1999; Heximer et al., 1999; Ingi et al., 1998; Potenza et al., 1999) and also to increase steady-state G_i GTPase activity in a reconstituted system containing agonist-activated m2 muscarinic receptor and phospholipids (Ingi et al., 1998). A possible explanation for the discrepancy between the pre-steady-state and other assays is that the additional proteins present in the latter experiments, namely receptor and G, may promote interactions between RGS2 and G_i. Because free G actually inhibits RGS GAP effects on isolated G (Chidiac and Ross, 1999; Wang et al., 1998b), it follows that receptors may support or promote RGS-G protein interactions. This view is supported by the identification of putative receptor binding domains in the C- and N-terminal regions of RGS4 (Zeng et al., 1998) and by the finding that the potencies of RGS proteins in attenuating signaling via G_q can vary depending on the identity of the activating receptor (Xu et al., 1999). Thus RGS proteins may be targeted in vivo to receptor-G protein complexes in addition to free G proteins.

Some studies have found RGS2 to be without effect on events subsequent to G_i activation by receptors, whereas other RGS proteins assayed in parallel have been inhibitory (Doupnik et al., 1997; Bowman et al., 1998; Reif and Cyster, 2000). The lack of inhibition by RGS2 in these cases suggests that its affinity for G_i is low compared with that of other RGS proteins. Taken from extrapolations from yeast pheromone response assays, RGS4 has been estimated to be approximately 8-fold more potent than RGS2 in blocking or deactivating G_i (Heximer et al., 1999). However, the potency of RGS2 as a GAP for G_i has never been established. For the present study, we developed an assay to measure and compare RGS2 and RGS4 GAP activities in a cellular membrane environment using postnuclear membranes from Sf9 cells infected with baculoviruses encoding the m2 muscarinic receptor plus G_i2, G₁, and G₂.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results and Discussion References

Plasmids encoding histidine-tagged RGS proteins were a generous gift from Dr. John Hepler (Emory University, Atlanta, GA). Baculoviruses encoding G_i2, G₁, G_2, and the m2 muscarinic receptor (tagged at the N terminus with a c-myc epitope) were kindly provided by Dr. Terry Hebert (Montreal Heart Institute, Montreal, Quebec). [³²P]GTP was purchased from ICN Pharmaceuticals (Costa Mesa, CA), and [³H]quinuclidinyl benzilate was purchased from PerkinElmer Life Sciences (Boston, MA). GTP, ATP, carbachol, atropine, and phenylmethylsulfonyl fluoride were purchased from Sigma Chemical (St. Louis, MO). Leupeptin and aprotinin were purchased from Roche Diagnostics (Indianapolis, IN). Cell culture reagents were purchased from Invitrogen (Carlsbad, CA).

Receptor and G Protein Expression in Sf9 Cells and Membrane Preparation. Sf9 insect cells at a density of 2 × 10⁶ cells/ml were infected with baculoviruses encoding the following: N-terminal c-myc-tagged m2 muscarinic receptor, G_i, G₁, and G_2. For some experiments, one or more viruses were omitted. After 48 h of infection, the cells were centrifuged at 228g for 5 min, resuspended in phosphate-buffered saline, and recentrifuged. The resulting pellets were resuspended in one third of the original volume of lysis buffer (20 mM Tris, pH 8.0, 0.1 mM PMSF, 10 µg/ml leupeptin, 1 µg/ml aprotinin). After a 15-min incubation on ice, the cells were lysed with the use of a Polytron homogenizer (Brinkmann Instruments, Rexdale, ON, Canada), followed by a 10-min centrifugation at 500g. The supernatant was retained and centrifuged for 30 min at 48,000g. The supernatant was discarded and the pellets resuspended in 0.01 volume of lysis buffer, then placed in aliquots and stored at 80°C. This procedure yielded 10.4 ± 3.0 fmol of m2 muscarinic receptor per milligram of membrane protein, as determined by the specific binding of [³H]quinuclidinyl benzilate.

RGS Protein Purification. For RGS4, 4 L of LB broth with ampicillin (0.05 mg/ml) were inoculated with Escherichia coli NH6 RGS4 pQE60 and incubated with vigorous aeration at 37°C until mid-log phase (optical density at 600 nm of 0.55). A 3-h induction was commenced by adding 1 mM isopropylthio--D-galactoside. The bacteria were harvested and resuspended in 60 ml of buffer A (0.05 M HEPES, pH 8.0, 0.1 M NaCl, 0.02 M -mercaptoethanol, 1% Triton X-100, 0.1 mM PMSF, 10 µg/ml leupeptin, 1 µg/ml aprotinin) and then snap frozen to 80°C. After thawing, 0.2 mg/ml lysozyme was added, and the suspension was mixed and incubated on ice for 30 min followed by the addition of 25 µg/ml DNase in the presence of 0.5 mM MgCl₂ for 20 min. The mixture was centrifuged at 140,000g for 30 min at 4°C, and the volume of the supernatant increased to 100 ml with buffer A supplemented with glycerol and imidazole (final concentrations of 20% and 0.02 M, respectively). Equilibrated Ni-nitrilotriacetic acid affinity resin (1.5 ml) was added and gently rotated in the cold for 90 min and then loaded onto a 5-ml column, washed with 30 ml of buffer A with 0.5 M NaCl, and then washed with 30 ml of buffer A with no Triton X-100. RGS4 was eluted with 0.2 M imidazole. A Superdex 75 HR 10/30 column (Amersham Biosciences Inc., Piscataway, NJ) was equilibrated with buffer B (0.05 M HEPES, pH 8.0, 0.15 M NaCl, 0.001 M DTT, 0.1 mM PMSF). The RGS4 fraction was loaded and eluted in a clean peak. The fractions were pooled, placed in aliquots, and stored at 80°C. For RGS2, a similar method was used in the purification process. LB broth (4 L) with ampicillin (0.05 mg/ml) were inoculated with E. coli BL 21 (DE3) RGS2.H10 pET 19b. The same procedure for the purification of RGS4 was followed except that the concentration of NaCl was 0.5 M throughout the imidazole-elution step and then reduced to 0.25 M for the final gel-filtration step and subsequent pooling of peak fractions. For both RGS2 and RGS4, the pooled Superdex fractions were estimated to be >95% pure as determined by Coomassie staining.

Assay of GTP Hydrolysis. The steady-state hydrolysis of [³²P]GTP by Sf9 membranes was measured in the absence and presence of purified RGS proteins. Unless indicated otherwise, 50-µl reaction mixtures containing 20 mM HEPES, pH 7.5, 1 mM EDTA, 1 mM DTT, 0.1mM PMSF, 10 µg/ml leupeptin, 1 µg/ml aprotinin, plus 10 to 50 mM NaCl and 10 mM MgCl₂ [7.5 mM free Mg²⁺, as calculated using the program "Bound and Determined" (Brooks and Storey, 1992)], were incubated at 30°C for 15 min with 1 µM GTP, 500 µM ATP, [³²P]GTP (1 × 10⁶ cpm/assay), either 100 µM carbachol or 10 µM atropine, and membranes (2 µg/assay). The assay was stopped by adding 950 µl of ice-cold 5% (w/v) Norit in 0.05 M NaH₂PO₄; the mix was then centrifuged, and the level of ³²P_i in the resulting supernatant was determined by liquid-scintillation counting. The nonspecific membrane GTPase signal was estimated by adding 1 mM of unlabeled GTP to the above assay mix, and this value was subtracted from the total counts per minute. In each experiment, separate controls were carried out to determine the GTPase activity attributable to trace contaminants in the purified RGS protein preparations. This was taken as the difference in ³²P_i between samples with and without maximal levels of RGS2 or RGS4, but lacking membranes, and then scaled according to RGS protein concentration at each assay point; the scaled value was then subtracted from experimental measurements of membrane-dependent GTPase activity as appropriate. Typically, the GTPase activity of the purified RGS proteins contributed less than 2% to the total signal. Agonist-dependent GTPase activity was taken as that observed in the presence of the muscarinic receptor agonist carbachol minus that in the presence of the antagonist atropine.

Analysis of Data. In the determination of K_m for GTP, the receptor-specific signal at multiple GTP concentrations was estimated by fitting data acquired in the presence of atropine to a straight line and then subtracting that from data acquired in the presence of carbachol. The value of K_m for GTP for receptor-specific GTPase activity was determined by Lineweaver-Burk analysis. Analyses of RGS2 and RGS4 dose-response data were carried out by nonlinear regression using the fitting program SigmaPlot 4.0 (SPSS Science, Chicago, IL). RGS4 data acquired in the presence of 7.5 mM free magnesium and RGS2 data were analyzed according to the four-parameter Hill equation included with the program. RGS4 in the presence of 0.5 mM magnesium yielded a bell-shaped dose-response pattern, and these data were analyzed as the sum of two rectangular hyperbolae plus a nonspecific component (eq. 1):

(1)

where ns is the nonspecific signal, K_up and K_down correspond to the midpoints, whereas max₁ and max₂ correspond to the y values of the rising and descending rectangular hyperbolae, respectively. In some experiments, max₂ was poorly defined, and its value was constrained during the fitting procedure such that its absolute value would not exceed that of max₁. Throughout the text, mean values are reported ± S.D.

	Results and Discussion

Top Abstract Introduction Materials and Methods Results and Discussion References

RGS proteins limit receptor signaling by accelerating GTP hydrolysis by G proteins, thereby shortening the lifespan of the activated GTP-bound state. This is likely to be their primary function in vivo, yet relatively few studies have measured RGS effects on GTP hydrolysis by receptor-activated G proteins. In general, the detection of receptor-stimulated G protein GTPase activity can be hampered by high background activities arising from endogenous nucleotidases in subcellular preparations. To avoid this shortcoming, GTPase assays can be carried out using purified receptors and G proteins co-reconstituted into phospholipid vesicles (Ingi et al., 1998; Mukhopadhyay and Ross, 1999), although such an approach is technically demanding. As a simpler alternative, we functionally reconstituted m2 muscarinic receptors and heterotrimeric G_i2 in Sf9 cells by coinfecting with baculoviruses encoding the receptor and the three G protein subunits.

Sf9 insect cells are useful for the study of G protein-linked receptors because of the high levels of functional protein that can be expressed and post-translationally modified. Although the endogenous G protein complement in these cells can be sufficient for the measurement of receptor-stimulated second messenger production (Chidiac et al., 1994), measuring changes in the nucleotide binding activity of G proteins is facilitated by the coexpression of the latter with a given receptor (Barr et al., 1997). We could not detect carbachol-stimulated GTPase activity in membranes prepared from cells infected only with baculoviruses encoding the m2 muscarinic receptor, but a clear agonist signal did emerge with cells additionally infected with baculoviruses encoding the three G protein subunits G_i2, G₁, and G₂ (Fig. 1A). Using this paradigm, we also were able to observe an effect of purified RGS4 on the steady-state GTPase activity of receptor-coupled G_i (Fig. 1B). If any G protein subunit was omitted, GAP activity was decreased or eliminated (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Stimulation of GTP hydrolysis by m2 muscarinic receptor and RGS proteins. GTPase activity was measured in the absence of drug (striped bars), in the presence of the agonist carbachol (), or in the presence of the antagonist atropine (). A, membranes were prepared from Sf9 cells infected with wild-type baculovirus (WT), baculovirus encoding the m2 muscarinic receptor (M2), or baculoviruses encoding the receptor plus Gi2, G1, and G2 (M2 + Gi) were prepared as described and were assessed for GTPase activity as indicated. B. Membranes from Sf9 cells infected with baculoviruses encoding the receptor and three G protein subunits were assayed for GTPase activity alone and in the presence of 5 µM of either RGS2 or RGS4, as indicated.

Similar to previous findings with GAP effects on receptor-activated G proteins (Biddlecome et al., 1996; Cavalli et al., 2000), the K_m of m2 muscarinic receptor-activated G_i2 for GTP was increased in the presence of RGS4 (Fig. 2). Such observations are consistent with an increase in the rate of GTP hydrolysis with no change in nucleotide binding properties, as discussed previously (Cavalli et al., 2000). Alternatively, the increase in K_m could conceivably reflect an RGS4-related decrease in the GDP dissociation rate constant or GTP association rate constant, but this seems unlikely because a previous study showed RGS4 to have no effect on nucleotide exchange with isolated G_i1 (Berman et al., 1996).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Effect of RGS4 on GTP concentration dependence. GTPase activity was measured at multiple GTP concentrations in the presence of atropine (), carbachol (), atropine plus 400 nM RGS4 (×), or carbachol plus 400 nM RGS4 (). The value of Km for GTP was determined by Lineweaver-Burk analysis as described under Materials and Methods, and the means of this experiment plus two others were calculated to yield values of 130 ± 80 nM in the presence of carbachol alone and 500 ± 80 nM in the presence of carbachol plus RGS4.

Effects of Salts on RGS GAP Activities. NaCl and KCl both were found to have inhibitory effects on RGS4 GAP activity (Fig. 3, A and B), with 250 mM of either salt decreasing RGS4 GAP activity by approximately 75%. Similarly, both salts also decreased the GAP activity of RGS2 (Fig. 3, C and D). These inhibitory effects seem to be related to RGS GAP activity rather than to either the intrinsic GTPase activity of the G protein or the agonist-stimulated GTPase activity, because both of those rates were essentially constant over the entire range of NaCl and KCl concentrations tested. To minimize the inhibitory effects of NaCl, which was present in the purified protein stocks, concentrations of the salt were kept at or below 50 mM in all further experiments.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Effects of NaCl and KCl on RGS2 and RGS4 GAP activity. GTPase activity was measured at the indicated concentrations of NaCl (A and C) or KCl (B and D) in the presence of atropine (), carbachol (), or carbachol plus RGS protein () (A and B: 400 nM RGS4; C and D: 3 µM RGS2).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Effect of MgCl2 on GTPase activity. GTPase activity was measured at increasing concentrations of MgCl2, and the calculated concentrations of free Mg2+ are shown on the abscissa. Assays were carried out in the presence of atropine (), carbachol (), or carbachol plus 2 µM of either RGS2 () or RGS4 ().

Concentration dependence of RGS2 and RGS4 GAP Activities. The results shown in Fig. 5 indicate by direct observation that (1) RGS2 acts as a G_i GAP in a membrane environment containing an activated G_i-coupled receptor and (2) the potency of RGS2 is lower than that of RGS4. Assays of RGS concentration dependence were carried out at two different levels of MgCl₂. At 7.5 mM free magnesium, RGS4 was approximately 5 times as potent as RGS2 in stimulating agonist-dependent GTPase activity (Table 1). RGS4 increased the agonist-dependent signal approximately 10-fold, whereas the highest concentrations of RGS2 tested increased it by a factor of 5 (Fig. 5A). At 0.5 mM free magnesium, the maximal effects of RGS2 and RGS4 were similar, approximately 10-fold greater than the agonist-dependent GTPase signal (Fig. 5B). It is difficult to compare concentration dependence under these conditions because the shapes of the curves are different; however, the midpoint of the ascending phase (K_up, Table 1) suggests a 9-fold greater potency of RGS4. The increased maximal effect of RGS2 relative to RGS4 at the lower magnesium concentration is consistent with the patterns of magnesium dependence observed in Fig. 4.

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Concentration dependence of RGS2 and RGS4 GAP activities at high and low magnesium levels. Agonist-dependent GTPase activities were measured in the presence of 7.5 mM (A) and 0.5 mM (B) calculated free Mg2+ (total added MgCl2 was 10 mM and 2 mM, respectively) in the absence of RGS protein () or at the RGS2 () and RGS4 () concentrations indicated on the abscissae. Data were fitted to either eq. 1 (RGS4 in B) or the Hill equation (others), and average parameters from these and replicate data sets are presented in Table 1.

i