Trp fluorescence reveals an activation-dependent cation-π interaction in the Switch II region of Gα_i proteins

Abstract

Crystal structures of Gα_i (and closely related family member Gα_t) reveal much of what we currently know about G protein structure, including changes which occur in Switch regions. Gα_t exhibits a low rate of basal (uncatalyzed) nucleotide exchange and an ordered Switch II region in the GDP-bound state, unlike Gα_i, which exhibits higher basal exchange and a disordered Switch II region in Gα_iGDP structures. Using purified Gα_i and Gα_t, we examined the intrinsic tryptophan fluorescence of these proteins, which reports conformational changes associated with activation and deactivation of Gα proteins. In addition to the expected enhancement in tryptophan fluorescence intensity, activation of GαGDP proteins was accompanied by a modest but notable red shift in tryptophan emission maxima. We identified a cation-π interaction between tryptophan and arginine residues in the Switch II of Gα_i family proteins that mediates the observed red shift in emission maxima. Furthermore, amino-terminal myristoylation of Gα_i resulted in a less polar environment for tryptophan residues in the GTPase domain, consistent with an interaction between the myristoylated amino terminus and the GTPase domain of Gα proteins. These results reveal unique insights into conformational changes which occur upon activation and deactivation of G proteins in solution.

Introduction

Activation of heterotrimeric G proteins coupled to 7-transmembrane receptors results in activated GαGTP and Gβγ subunits which activate downstream effectors. Hydrolysis of bound GTP to GDP facilitates re-association of GαGDP and Gβγ, a prerequisite for productive coupling to activated receptors. The conformational changes in Gα subunits that accompany nucleotide exchange and hydrolysis are sensed by three Switch regions: Switch I, II, and III. These regions adopt distinctive conformations in crystal structures of Gα proteins bound to GDP, GDP.Gβγ, GTPγS, and GDP-AlF₄,^1–9 facilitating interactions of Gα with Gβγ^8,9 and a variety of effectors.^10–14 Although much of what we know about Gα structure and function has been gleaned from crystal structures, the Switch II of Gα_i which binds Gβγ^8,9 is disordered in Gα_iGDP crystal structures in the absence of Gβγ, as is the N-terminus of isolated Gα proteins. Crystallographically disordered regions in proteins are often important interfaces for protein–protein interaction. The conformation of Switch II may be important for binding to not only Gβγ, but also other Gα effectors such as GIV, a recently identified GEF for Gα_i proteins.¹⁵ Similarly, a peptide derived from the GoLoco region of RGS14 binds Switch II of Gα_iGDP¹⁶ to competitively inhibit Gβγ binding. Other Gα effectors which bind Switch II of Gα proteins in a conformationally sensitive manner include adenylyl cyclase (which binds activated Gα_s)¹¹ and the RGS domain of RhoGEF which binds activated Gα₁₃.¹⁷

The Trp in Switch II (W211 in Gα_i) reports conformational changes upon activation.^20,21 The increase in emission intensity which accompanies the activation-dependent movement of the Switch II region into a hydrophobic pocket^1–7 provides a convenient and reliable method to assess the functional integrity of purified proteins.^22,23 Gα_i contains two other Trp residues, W258 in the GTPase domain and W131 in the helical domain. Crystal structures show that the Trp in the helical domain of Gα_i and Gα_t proteins remains buried in both the inactive and active states,^3,4,6 and NMR studies have shown it to be refractory to activation.²⁴ A conformational change in a protein which relocates a Trp residue into a more solvent-excluded environment is typically accompanied by an increase in emission intensity (such as upon Gα activation) and a shift in emission maxima (em_max) to lower wavelengths (called a blue shift). However, in Gα_i proteins, this activation-dependent enhancement in intensity is accompanied by a small but notable red shift in em_max (toward a higher wavelength). Investigation of this unexpected red shift revealed an activation-dependent cation-π interaction between Arg and Trp residues in Switch II as determined by steady-state fluorescence, and this finding was confirmed using site-directed mutagenesis. The relative differences in the Trp fluorescence of Gα_iGDP proteins before and after activation indicate an overall conformational similarity in solution between the Switch II region of Gα_i and Gα_t proteins.

Both Switch II and the N-terminus of Gα proteins participate in binding to Gβγ. In nature, the N-terminus of Gα_i family proteins are permanently, cotranslationally myristoylated,¹⁸ which enhances Gα-Gβγ association,¹⁹ thus its conformation (along with that of Switch II) may influence the duration of Gβγ signaling. We also looked at the environment of Trp residues in myristoylated proteins, and found a myristoylation-dependent reduction in solvent exposure for Trp residues in the GTPase domain of Gα_i proteins. Together these results shed light on conformational changes that occur in solution upon activation and deactivation of Gα subunits.

Results

There are three Trp residues in Gα_i [Fig. 1(A)]: W211 and W258 in the GTPase domain, and a Trp in the helical domain, W131, which does not report activation-dependent changes.^4,24 Because of the well-known ability of the Trp in Switch II of Gα proteins to report such changes, functional Gα proteins typically demonstrate a minimum 40% increase in Trp emission intensity upon activation with AlF₄.²⁵ AlF₄ activates Gα proteins by formation of a GαGDP-AlF₄ complex which mimics the pentavalent transition state for GTP hydrolysis. In this work, we investigated the activation-dependent red shift in Trp em_max in Gα_i [Fig. 1(B)], which was also present after a prolonged incubation with GTPγS (not shown), as well as upon activation of Gα_i HI [Fig. 1(C)], a Gα_i protein lacking several solvent exposed cysteines which does not perturb function.²⁶

Figure 1.

Intrinsic Trp fluorescence of Gα_i. (A) Location of 3 native Trp residues (red) in Gα_iGDP-AlF₄ (PDB 1GFI) with nucleotide rendered in space fill. NT, last resolved residue (33) in N-terminus. (B, C) Normalized emission of 400 nM Gα_i (B, wt Gα_i; C, Gα_i HI), ex/em 280/310–400 nm, before (dotted) and after (solid) activation with AlF₄. Control, inset: Gα proteins demonstrate ≥40% increase in intrinsic Trp fluorescence upon AlF₄ activation (raw data).

To gain insight into the cause of the red shift, we sought to compare structures of Gα_i before and after activation, however, Gα_iGDP is disordered in the Switch II region.^7,27 Therefore, structures of the closely related Gα_i family member, Gα_t were examined, since Gα_t shares a high sequence and structural homology with Gα_i. Comparison of the Switch II conformation in Gα_tGDP and Gα_tGDP-AlF₄ structures revealed an activation-dependent stacking of positively charged R204 (R208 in Gα_i) over the π electrons of W207 (W211 in Gα_i), reducing the distance between them by more than 2 Å [Fig. 2(A,B)]. A computational study of energetically favorable cation-π interactions determined that the CD and CZ atoms of Arg (δ-carbon and guanidinium carbon, respectively) are typically within 6 Å of the aromatic ring of Trp,²⁸ as is the case here [Fig. 2(A)]. In Gα proteins, this interaction is aided by a network of interactions which contribute to the positioning of the Arg-Trp pair upon activation, including a salt bridge between this Arg and a Glu adjacent to Switch III [E241 in Gα_t, Fig. 2(B), bottom panel], as well as arginine's interactions with glycine residues that contact bound nucleotide (Supporting Information Fig. 1). Structural studies show that activation tucks Switch II into a hydrophobic pocket, resulting in a closer proximity between the Arg and Trp, which may polarize Trp emission in the Switch II region. If so, this would explain the shift of the Trp em_max to higher, rather than lower wavelength.

Figure 2.

Activation-dependent increase in proximity between Arg and neighboring Trp in Switch II of Gα_t. (A): Switch II Trp in Gα_i family member Gα_t is 2.2 Å closer to positively charged Arg upon activation, Gα_tGDP-AlF₄ (PDB file 1TAD) (teal) overlay with Gα_tGDP (PDB file 1TAG) (purple). Note residues R204 and W207 in Gα_t correspond to residues R208 and W211 in Gα_i. (B, top panel): Switch II of Gα_tGDP shown with surface rendered. (B, bottom panel): Switch II tucks into protein upon activation, shown with surface of Gα_tGDP-AlF₄ rendered.

Before ascertaining if the Switch II Arg was polarizing Trp emission in Gα_i proteins, we first examined this proposal in Gα_t, which is ordered in the Switch II region in both GDP- and GDP-AlF₄-bound states. Unmyristoylated Gα_t used in the structural determinations was prepared by treating Gα_t obtained from rod outer segments with EndoLysC² to cleave the myristoylated N-terminus of Gα_t at residue 33, followed by gel filtration. After purification, the molecular weight of the truncated Gα protein was confirmed by SDS-PAGE [Fig. 3(A), ΔNT]; this protein retained the ability to report activation-dependent changes [Fig. 3(B), inset]. We scanned the emission of the unmyristoylated Gα_t before and after activation, and found the em_max underwent a red shift [Fig. 3(B)] similar to the one observed for Gα_i, consistent with a polarization of the Switch II Trp upon activation, linking structural changes in Gα_t to shifts in em_max. In addition, comparison of unmyristoylated Gα_i and Gα_t proteins reveal a lower em_max for Gα_t [Fig. 3(C)] regardless of activation, most likely due to the fact that Gα_t has a Tyr instead of the more solvent exposed Trp at the position homologous to W258 in Gα_i. Trp emission dominates protein spectra between 330 and 350 nm, while Tyr has an em_max at 303 nm. The em_max of Tyr is unaffected by changes in its environment, unlike Trp which demonstrates shifts in em_max in response to the macroenvironment and microenvironment of tryptophan's indole ring.

Figure 3.

Trp fluorescence of unmyristoylated proteins. (A) Myristoylated Gα_t (native) is enzymatically cleaved at the N-terminus to yield the unmyristoylated ΔNT Gα_t protein used in structure determination.⁶ (B) Emission of 500 nM ΔNT-Gα_t protein was scanned as above before (dotted line) and after (solid) activation with AlF₄ as in Figure 1. Control, inset: ΔNT-Gα_t demonstrated ≥40% increase in Trp fluorescence upon AlF₄ activation. (C) Summary of em_max before and after activation for indicated Gα proteins (n ≥ 3, ±SEM; statistically significant changes noted by asterisk(s), ***P < 0.001, **P < 0.01). Dark gray bars, far left, Gα_i; light gray, center, Gα_i HI; black, far right, ΔNTGα_t.

Given that the activation-dependent red shift observed in Gα_i (Fig. 1) was recapitulated in Gα_t (Fig. 3), and in light of the associated increase in proximity between Arg and Trp in Switch II of Gα_t (Fig. 2), the red shift we observed in Gα_i may indicate a similar geometry. If so, R208 and W211 of Gα_i would be predicted to play critical roles. To investigate this, we examined the em_max of 4 Gα_i HI proteins: W211C, R208C, R208L, and W258F, residues all located in the GTPase domain, where activation-dependent changes are known to occur. Gα_i HI proteins were used in this study because they have been shown to tolerate cysteine mutations in the Switch II region²⁹ and elsewhere throughout the protein.^26,29,30 Residues W211 and R208 are part of the Switch II region of Gα_i, and W258 is located on a more solvent exposed loop adjacent to Switch II in the GTPase domain [Fig. 1(A)].

Although the Gα proteins used in this study generally demonstrated a minimum of 40% increase in Trp intensity upon activation, mutation of residues implicated in the cation-π interaction precluded assessment of functionality by this method. Therefore, we used an extrinsic probe linked to GTPγS to confirm the ability of these proteins to exchange nucleotide. Uncatalyzed exchange of GDP for BD-GTPγS is reported by an increase in emission from BD-GTPγS that occurs upon its binding to Gα_i subunits.³¹ Because Gα_i subunits are known to bind to BD-GTPγS at a slower rate and with a lower affinity than unlabeled GTPγS,³¹ overall BD-GTPγS binding was measured after full exchange was allowed to occur. The indicated Gα_i proteins showed the ability to exchange GDP for BD-GTPγS [Fig. 4(A), solid bars], which was eliminated by pre-boiling these proteins [Fig. 4(A), far right bars]. The mutant proteins were also monodisperse on gel filtration, exhibited retention time on gel filtration chromatography according to their predicted molecular weight, and migrated essentially the same as wild-type proteins on SDS PAGE, consistent with proper expression and folding of these proteins.

Figure 4.

A: Extrinsic nucleotide exchange. Solid bars (controls): Gα proteins exchange GDP for BD-GTPγS, in contrast to heat-denatured Gα (boiled, last three bars). Striped bars: nucleotide exchange in Gα proteins that were preactivated by AlF₄ before assay. B,C: Red shift is markedly reduced by mutation of either W211 or R208, unlike mutation of Trp outside of Switch II, W258 (A, B: n ≥ 3, ±SEM; **P < 0.01; ***P < 0.001).

Since AlF₄ binding to wild-type Gα_iGDP is known to reduce nucleotide exchange with GTP analogs due to additional stabilizing contacts introduced by binding AlF₄,^4,32 we next measured nucleotide exchange using Gα proteins pre-activated by AlF₄ [Fig. 4(A), striped bars]_. The stabilizing contacts introduced by AlF₄ reduced binding of BD-GTPγS by Gα_i HI as expected, and did so to a somewhat lesser extent in the R208 mutants, however mutation of W211 resulted in the complete loss of AlF₄-mediated stabilization of bound GDP. This protein exchanged GDP for BD-GTPγS regardless of activation by AlF₄ [Fig. 4(A), striped bars].

Mutation of either R208 or W211 attenuated the red shift upon activation [Fig. 4(B,C)], despite the ability of these proteins to exchange nucleotide [Fig. 4(A)]. Mutation of the same Switch II Trp in Gα_o also prevents detection of activation by Trp fluorescence, without preventing activation-dependent conformational changes or the binding to labeled GTPγS.³³ Both R208 and W211 are direct participants in the cation-π interaction and play a role in stabilization of the Switch II region upon activation through a combination of electrostatic, hydrophobic and van der Waals interactions. Mutation of W258 did not impair the activation-dependent red shift in em_max [Fig. 4(B)], and it resulted in a lower em_max than the parent Gα_iGDP HI protein [Fig. 4(B) vs. Fig. 3(C)], indicating a high degree of solvent exposure for this Trp. These studies were performed in the absence of Gβγ and receptor which contain numerous Trp residues that obscure detection of changes in Gα emission by increasing background signal.

Using Gα_i proteins coexpressed with N-myristoyl transferase, we found that N-terminal myristoylation reduced the activation-dependent red shift in wild-type Gα_i, and eliminated it in native Gα_t [Fig. 5(A) vs. Fig. 3(C)], without impairing activation-dependent increases in intensity [Fig. 5(B), shown for Gα_t]. Interestingly, myristoylation also reduced the em_max of GDP-bound Gα_i proteins with a native Trp at position 258 [Fig. 5(C)], but did not have the same effect on Gα proteins lacking this or an equivalent Trp, Gα_i HI W258F, and Gα_t. Thus, W258 in the GTPase domain senses myristoylation in GDP-bound Gα_i subunits. Myristoylation also decreases the em_max of activated, wild-type Gα_i and Gα_t [Fig. 5(D)], consistent with a myristoylation-dependent modulation of the cation-π interaction in these proteins, either directly or indirectly, however, the exact nature of the interaction has yet to be determined. Taken together, myristoylation has effects on both W258 and W211, which are located near the last resolved residues in the N-terminus of Gα_i [Fig. 1(A)]. These data are in agreement with a previous model suggesting the myristoylated N-terminus of Gα_i proteins interact with residues on the surface of Gα_i,^18,27 with myristoylation-dependent effects detected for Trp residues W211 and W258, located in and near the Switch II region in the GTPase domain of Gα_i proteins.

Figure 5.

Effect of myristoylation on em_max. (A) Comparison of em_max of myristoylated proteins before and after activation. (B) Normalized spectra of Gα_t (native, myristoylated) before and after activation with AlF₄. Inset: raw data before and after activation. (C,D) Effect of myristoylation on em_max of GDP (C) and GDP-AlF₄ bound proteins (D). (A, C, D: n ≥ 3, ± SEM; *P < 0.05; **P < 0.01; ***P < 0.001).

These results suggest a model whereby upon GTP hydrolysis, Gα_iGDP adopts a conformation in the Switch II region [Fig. 6(A,B)] which resembles the ordered Switch II region seen in structures of Gα_tGDP [Fig. 6(B), purple]. The red shift upon activation provides information about the position of Gα_i Switch II residues in the GDP-bound state relative to that in the GTP-bound state, indicating a qualitatively similar distancing of the Trp and Arg in the Switch II of GDP-bound Gα_i and Gα_t proteins after GTP hydrolysis has occurred. The activation-dependent cation-π interaction between R208 and W211 in Gα_i (separated by one turn of the Switch II α-helix) is stabilized by a network of interactions, including a conserved pair of flexible glycines, G202 and G203 (Supporting Information Fig. 1), which allosterically link the Switch II region to activation-dependent changes in the nucleotide binding site. Computational analysis using the program CaPTURE²⁸ estimates the strength of this interaction to be approximately −2.0 kcal/mol, which reflects both electrostatic and van der Waals contributions, resulting in an energetically relevant interaction between these residues.

Figure 6.

Model. (A) Crystal structures of activated Gα_i (PDB file 1GFI, top panel, overview; bottom panel, Switch II region enlarged, nucleotide shown in spacefill) demonstrate that the Trp π electrons are within 6 Å from the positively charged Arg in the Switch II helix. (B) Ordered regions of Gα_iGDP (PDB file 1BOF) shown in blue, Switch II region of Gα_tGDP in purple (from structural alignment of PDB file 1TAG and 1BOF), and nucleotide shown in spacefill. Top, overview; bottom panel, enlarged, depicting spatial separation of 8 Å or more between the Switch II Arg and Trp after GTP hydrolysis has occurred.

Discussion

This study provides insights into the Switch II conformation of Gα_iGDP in solution. Crystal structures of Gα proteins demonstrate that upon activation, the Switch II tryptophan moves into a more solvent-protected environment, resulting in enhanced Trp emission intensity, and a small but unexpected red shift in em_max. This was originally noted without elaboration in one of the first reports of Gα protein isolation.²³ We used a combination of mutational and biochemical approaches to identify the cause of this shift, an activation-dependent cation-π interaction, mediated by Arg and Trp in the Switch II helix of Gα_i and Gα_t proteins. The homology between Gα_i and Gα_t was used as a link between this shift and Switch II conformation, as the Switch II of Gα_tGDP is ordered in crystal structures, unlike Gα_iGDP. These solution studies indicate a relatively similar orientation of Arg and Trp in Switch II of Gα_i and Gα_t proteins both before and after activation.

A recent study indicated that almost 40% of energetically relevant cation-π interactions involve an Arg-Trp pair,²⁸ and many of these help stabilize α-helices, as in the case of Gα proteins. The majority position Trp's aromatic ring 6 Å or less from the CD and CZ atoms of Arg, over which the charge is distributed.²⁸ In Gα_i, the CZ of R208 moves to within 6 Å of the aromatic ring of W211 upon activation, with the CD of R208 within 5 Å of the ring, polarizing Trp emission. In Gα_t, this results in a 2.2 Å reduction in the distance between the Arg-Trp pair. Similarly, in the protein ricin, ligand binding results in a 1.7 Å reduction between its Arg-Trp pair and a 3–4 nm red shift.³⁴ Common cation-π interactions^35,36 position the cation parallel to the plane of tryptophan's ring of p-orbitals.³⁷ Energetically relevant Arg-Trp interactions are estimated at −2.9 ± 1.4 kcal/mol,²⁸ which includes the −2.0 kcal/mol calculated for the Arg-Trp pair in Gα_i, in agreement with a proposed cation-π interaction suggested from comparison of Gαβγ_i and Gα_iGTPγS structures.³⁶ Van der Waals and hydrophobic contacts along the cation side chain contribute to the favorable energetics of the Arg-Trp pair, along with charge delocalization which maximizes Arg's interaction with Trp's indole ring. In addition, Arg often forms conventional hydrogen bonds with other residues while simultaneously interacting with Trp,^38–40 as seen in the activation-dependent interaction of R208 with E245 just next to Switch III, which allosterically links changes in nucleotide binding to Switch III conformation in Gα_i proteins.

Residues R208 and W211, as well as G202 and G203, are highly conserved among G proteins. Mutation of W211 abrogates the stabilizing influence normally imparted by formation of a GDP-AlF₄ complex. The structure of Gα_iGDP-AlF₄ shows that W211 is hydrogen bonded to G202, with G203 making hydrogen bonds to AlF₄ and R208, thus allosterically linking Switch II conformation to nucleotide binding. The flexibility of these conserved glycines may also facilitate Q204's interactions with the catalytic water molecule during GTP hydrolysis, which together with R178, Q171 and T181 help orient this water toward the developing positive charge on the γ-phosphate.⁴ In our study, GTPγS binding also resulted in red-shifted Trp emission, as interactions that the backbone amide hydrogens of G202 and G203 also make with the γ phosphate help bring R208 and W211 within range to polarize W211 emission. Altogether, a concerted network of allosteric interactions between these conserved residues and others in the GTPase domain may drive Switch II conformation during GTP binding and hydrolysis.

Despite structural similarities between Gα_i and Gα_t, the magnitude of the red shift differs somewhat between these proteins. In unmyristoylated Gα proteins, the relatively larger red shift for Gα_t is likely due to an additional more solvent exposed Trp present in Gα_i proteins, W258, that is absent in Gα_t. Mutation of W258 in Gα_i HI reduces its em_max and results in a Gα_t-like emission profile, due to the location of W258 on a solvent exposed loop of Gα_i, and indicating a qualitatively similar environment for Switch II tryptophans in Gα_t and Gα_i proteins in solution. The conformation of Switch II after GTP hydrolysis (before Gβγ binding) may be important in regulation of basal signaling, evidenced by a peptide that binds Switch II and peels it away from the core of Gα_i-GDP, increasing nucleotide exchange.⁴¹ We recently determined the structure of a Gα_i protein with mutations near the receptor-binding carboxy terminus which displayed elevated nucleotide exchange and altered Switch II conformation,³² consistent with an allosteric linkage between receptor binding, Switch II conformation and nucleotide exchange.

Not reflected in any Gα structure is N-terminal myristoylation, which has been shown in some systems to act as a myristoyl switch,^42,43 with effects on spatial and temporal regulation of signaling⁴⁴ in the visual system. Our previous work suggested an immobile and solvent excluded environment for myristoylated N-terminal residues even in the absence of Gβγ, using a series of extrinsic probes linked to specific N-terminal residues in Gα_i HI proteins.¹⁸ The current study has uncovered an effect of N-terminal myristoylation on Trp residues far removed in sequence from the N-terminus, suggesting an interaction of the myristoylated N-terminus with or near Trp residues in the GTPase domain. Mutation of W258 to Phe eliminates its contribution to overall Trp emission, but the preservation of a hydrophobic residue at this position likely does not impair any potential interactions that residue 258 may make with the myristoylated N-terminus. Although myristoylated proteins exhibited the requisite activation-dependent increases in Trp intensity upon activation, indicative of the movement of the Switch II Trp into a hydrophobic pocket, the associated red shifts were reduced by myristoylation of Gα_i and eliminated in myristoylated Gα_t, indicating a myristoylation-dependent modulation of the cation-π interaction. Together these results shed light on the environment of the Switch II region of Gα_i proteins before and after activation, and suggest a conformationally sensitive effect of N-terminal myristoylation on residues in the GTPase domain of Gα proteins, consistent with an intramolecular binding site for the myristoylated N-terminus of Gα_i family proteins.

Materials and Methods

Materials

GDP and GTPγS were purchased from Sigma-Aldrich (Milwaukee, WI). BODIPY FL-GTPγS, thioester, was purchased from Invitrogen (Madison, WI). The endoproteinase EndoLysC was purchased from Calbiochem (San Diego, CA). All other reagents and chemicals were of the highest available purity.

Preparation of Gα proteins

Gα_t was obtained and prepared as described previously.² Gα_il and Gα_il HI were used as is or as a basis for site-specific mutagenesis in selected residues. Construction, expression, and purification of Gα_il and Gα_il HI proteins (unmyristoylated and myristoylated) were performed as described previously.^18,26 Briefly, a parent Gα_il HI lacking solvent-exposed cysteines which was previously shown to have properties similar to wild type²⁶ was generated with an expression vector encoding rat Gα_i1, with the following amino acids substituted: (C66A-C214S-C305S-C325A-C351I), a hexahistidine tag inserted between amino acid residues M119 and T120, and the Switch II mutants described previously have an alanine at residue 3, which does not perturb function.^26,29 Mutations were introduced using the QuickChange system (Stratagene, La Jolla, CA); DNA sequencing confirmed all mutations. Proteins were expressed with N-myristoyl transferase to obtain myristoylated protein, and the same protein expressed in the absence of the N-myristoyl transferase vector to produce unmyristoylated proteins. Proteins were expressed and purified as detailed previously¹⁸ in E. coli BL21-Gold (DE3) with or without a N-myristoyl transferase (NMT) vector pbb131, which also encodes kanamycin resistance for selection purposes (NMT vector generously provided by M. Linder, Washington University). Coomassie staining of urea SDS PAGE gels⁴⁵ demonstrate Gα_i proteins are fully myristoylated¹⁸ when coexpressed with the NMT vector. Purified proteins were stored at −80°C in buffer A containing 50 mM Tris, 100 mM NaCl, 2 mM MgCl₂, 10 μM GDP, pH 7.5, and 10% (v/v) glycerol; wild-type Gα containing native, solvent-exposed cysteines was additionally supplemented with 5 mM β-mercaptoethanol or 1 mM DTT before storage at −80°C. After purification, all proteins used in this study were greater than 85% pure, as estimated by Coomassie staining of SDS-PAGE gels.

Intrinsic Trp fluorescence

Proteins without mutations in Switch II demonstrated ≥40% increase in intrinsic Trp emission upon AlF₄ activation, relative to emission in the GDP-bound state, as previously described.²⁰ Briefly, Gα proteins (200 nM) in 50 mM Tris, 100 mM NaCl, 1 mM MgCl₂, pH 7.5 containing 10 μM GDP were monitored by excitation/emission 280/340 nm before and after the addition of AlF₄ (10 mM NaF and 50 μM AlCl₃) using a Varian Cary Eclipse Fluorometer. For proteins with mutations in Switch II region which impair the ability to report activation-dependent changes via intrinsic fluorescence, an extrinsic fluorescence using BD-GTPγS was used to assess ability to undergo activation-dependent changes (see below).

Nucleotide exchange of GDP and GDP-AlF₄ as measured by BODIPY-GTPγS fluorescence

Basal nucleotide exchange was measured as fold-increase in emission of 1 μM BODIPY-GTPγS (BD-GTPγS, λ_ex = 490 nm, λ_em = 515 nm) in buffer containing 50 mM Tris, 100 mM NaCl, 1 mM MgCl₂, 10 μM GDP, pH 7.5, at 21°C, in the presence and absence of 75 μM AlF₄ before and 60 min after the addition of Gα_i subunits (200 nM) to the cuvette containing BD-GTPγS. Boiled proteins (b) were heated at 95°C for 5 min before analysis.

Structural analysis/depictions

Computational determination of the energetics of the cation-π interaction in 1GFI was determined using the CaPTURE program.²⁸ Molecular graphics images were produced using the UCSF Chimera package.^46,47 Graphs and statistical analysis were performed using GraphPad Prism 4.0, GraphPad Software, San Diego, California.

Glossary

Abbreviations:

AU

arbitrary units

BD-GTPγS

BODIPY-GTPγS

emmax

maximal emission wavelength

ex/em

excitation/emission

GDP

guanosine diphosphate

GEF

guanine nucleotide exchange factor

Gi

family of G proteins coupled to inhibition of adenylyl cyclase

Gs

family of G proteins coupled to activation of adenylyl cyclase

Gt

G protein of the rod outer segment, transducin

GTP

guanosine triphosphate

GTPγS

guanosine-5′-O-(3-thiotriphosphate)

Gα

α subunits of G proteins

Gβγ

βγ dimer of heterotrimeric G proteins which binds Gα

HI

Gαil isoform lacking several solvent-exposed cysteines

N-terminus

amino terminus

SEM

standard error of mean

wt

wild-type.