The mechanism of Gα_q regulation of PLCβ3-catalyzed PIP2 hydrolysis

Significance

For certain cellular signaling processes, the background activity of signaling enzymes must be minimal and stimulus-dependent activation robust. Nowhere is this truer than in signaling by PLCβ3 (Phospholipase Cβ), whose activity regulates intracellular Ca²⁺, phosphorylation by Protein Kinase C, and the activity of numerous ion channels and membrane receptors. In this study we show how PLCβ3 enzymes are regulated by two kinds of G proteins, Gβγ and Gα_q. Enzyme activity studies and structures on membranes show how these G proteins act by separate, independent mechanisms, leading to a product rule of costimulation when they act together. The findings explain how cells achieve robust stimulation of PLCβ3 in the setting of low background activity, properties essential to cell health and survival.

Abstract

PLCβ (Phospholipase Cβ) enzymes cleave phosphatidylinositol 4,5-bisphosphate (PIP2) producing IP3 and DAG (diacylglycerol). PIP2 modulates the function of many ion channels, while IP3 and DAG regulate intracellular Ca²⁺ levels and protein phosphorylation by protein kinase C, respectively. PLCβ enzymes are under the control of G protein coupled receptor signaling through direct interactions with G proteins Gβγ and Gα_q and have been shown to be coincidence detectors for dual stimulation of Gα_q and Gα_i-coupled receptors. PLCβs are aqueous-soluble cytoplasmic enzymes but partition onto the membrane surface to access their lipid substrate, complicating their functional and structural characterization. Using newly developed methods, we recently showed that Gβγ activates PLCβ3 by recruiting it to the membrane. Using these same methods, here we show that Gα_q increases the catalytic rate constant, k_cat, of PLCβ3. Since stimulation of PLCβ3 by Gα_q depends on an autoinhibitory element (the X-Y linker), we propose that Gα_q produces partial relief of the X-Y linker autoinhibition through an allosteric mechanism. We also determined membrane-bound structures of the PLCβ3·Gα_q and PLCβ3·Gβγ(2)·Gα_q complexes, which show that these G proteins can bind simultaneously and independently of each other to regulate PLCβ3 activity. The structures rationalize a finding in the enzyme assay, that costimulation by both G proteins follows a product rule of each independent stimulus. We conclude that baseline activity of PLCβ3 is strongly suppressed, but the effect of G proteins, especially acting together, provides a robust stimulus upon G protein stimulation.

Phospholipase Cβ (PLCβ) enzymes cleave phosphatidylinositol 4,5-bisphosphate (PIP2) in the plasma membrane to produce inositol triphosphate (IP3) and diacylglycerol (DAG), (1, 2). PIP2 regulates the function of many membrane proteins including ion channels, IP3 increases cytoplasmic Ca²⁺ via the IP3 receptor, and DAG activates protein kinase C, which itself regulates numerous target proteins (3–5). Because PIP2, IP3, and DAG are critical to so many cellular processes, their tight regulation by PLCβs is essential to normal cellular function. PLCβ enzymes are under the control of G protein coupled receptor (GPCR) signaling through direct interaction with G proteins, Gα_q and Gβγ (6–8). Basal activity of PLCβs is maintained at very low levels in cells via two autoinhibitory elements, the X-Y linker, which occupies the active site, and the Hα2’ in the proximal c-terminal domain (CTD), whose mechanism of autoinhibition is not well understood (9–13).

PLCβs are aqueous-soluble enzymes that must partition onto the membrane to carry out PIP2 hydrolysis, which has posed a challenge to obtaining a quantitative description of their catalysis and regulation by G proteins. To overcome this challenge, we recently developed methods to measure both the partitioning of PLCβ enzymes between aqueous solution and membrane surfaces and the hydrolysis of PIP2 by membrane-bound enzyme (14). With these methods, we showed that PLCβ3 is a Michaelis-Menten enzyme and that Gβγ-dependent activation is mediated by increasing its local concentration at the membrane surface. Gβγ does not significantly change the catalytic rate constant (k_cat) of PLCβ3 nor does it alter its autoinhibitory elements in structures of the PLCβ3·Gβγ(2) complex (14).

The mechanism of activation by Gα_q is not understood, particularly the potential role of the membrane in activation. Specifically, it is not clear whether Gα_q activates by membrane recruitment like Gβγ or whether it increases k_cat through an allosteric mechanism. The lipid anchor on Gα_q is not required for activation of PLCβs, in contrast to Gβγ, suggesting that membrane recruitment might not underlie Gα_q-dependent activation (10, 11, 15, 16). However, nonlipidated Gα_q has been shown to maintain its association with membranes in cells and in vitro, raising the possibility that membrane recruitment could still play a role even in the absence of a covalent lipid group (15). It has also been established that Gα_q does not activate PLCβs in the absence of a membrane environment, suggesting that the membrane does play a role in activation (13).

Structural studies of the PLCβ3·Gα_q complex in the absence of membranes revealed that Gα_q binds to the proximal and distal CTD of PLCβ3 and Gα_q binding displaces the autoinhibitory Hα2’ away from its binding site on the catalytic core by ~50 Å (10, 16). These observations led to the proposal that Gα_q actives PLCβs by relieving Hα2’ autoinhibition. However, the mechanism of autoinhibition by Hα2’ is unknown: it is only known that removing the Hα2’ or disrupting its contacts with the catalytic core increases the basal activity of PLCβs (9, 11).

Some PLCβs, including PLCβ3, can also be activated by Gβγ and Gα_q simultaneously. This dual activation, which underlies many physiological functions, was proposed to play a role in coincidence detection under costimulation of Gα_i and Gα_q-coupled receptors in cells (8, 17, 18). Dual activation was proposed to be synergistic, or greater than the sum of the activation of each G protein on its own (19).

The goal of the present study is to understand the mechanism of activation of PLCβ3 by Gα_q and of dual activation by Gα_q and Gβγ. Using functional experiments, membrane partitioning studies, and structural studies on membrane surfaces, we show that nonlipidated Gα_q activates PLCβ3 by increasing its catalytic rate constant, k_cat, without affecting membrane recruitment. We also show that Gα_q-stimulated enhancement of k_cat is mediated by the X-Y linker autoinhibitory element. Thus, the X-Y linker is a suppressor of k_cat that is partially relieved by Gα_q. Finally, we show that nonlipidated Gα_q and Gβγ regulate PLCβ function independently, the former through k_cat and the latter through membrane recruitment. Consequently, dual stimulation yields activity enhancement equal to the product of the two independent stimuli.

Results

Activation of PLCβ3 by Nonlipidated Gα_q.

PLCβ3 is an aqueous-soluble enzyme that partitions onto the membrane surface to catalyze PIP2 hydrolysis. As we will describe below, we use the partition coefficient of PLCβ3 to calculate its membrane surface concentration from its aqueous concentration set by experimental design (14). For reasons that will become apparent, we use two different concentration units in this study. In some circumstances, we specify molar concentration using the notation [quantity]_molar. In other circumstances, we specify mole fraction (mf  ) in mole% (100 × mf   ) using the notation [quantity]. Because moles of solvent (water for aqueous solutions and lipids for membranes) are in vast excess of moles of solute (PLCβ3 for aqueous solutions and PLCβ3 and PIP2 for membranes), we approximate mf as moles solute over moles solvent (14).

To measure PIP2 hydrolysis by PLCβ3 on membrane surfaces, we employed an enzyme assay described in a recent publication (14). Briefly, the reaction takes place on a planar lipid bilayer formed over a hole in a partition separating two aqueous chambers (Fig. 1A). The lipid bilayer is composed of 2 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (DOPE):1 1-palmitoyl-2-oleoyl-glycero-3-phosphocholine (POPC):1 1-palmitoyl-2-oleoyl-sn-glycero-3-phospho-L-serine (POPS) (wt:wt:wt), plus a predetermined concentration of PIP2 ([PIP2] = 1.0 mol%), and a PIP2-gated ion channel is incorporated into the bilayer via vesicle fusion. The ion channel, a modified PIP2-dependent, G protein-dependent inward rectifier K⁺ channel, called GIRK-ALFA, is calibrated so that the normalized K⁺ current level can be converted to membrane PIP2 concentration (Fig. 1 B–D) (14). Upon addition of PLCβ3 under continuous mixing, after an approximately 2 s delay, the GIRK-ALFA current decreases due to hydrolysis of PIP2 by the added PLCβ3 (Fig. 1B). Using the predetermined calibration curve (Fig. 1C), normalized current as a function of time (Fig. 1B) can be converted to PIP2 concentration as a function of time, as shown (Fig. 1D). The latter curve corresponds very well (typical R² > 0.9) to a Lambert W Function (aka Product Log Function) (20), which describes a linear decay initially (when PIP2 concentration is high) and an exponential decay at later times (when PIP2 concentration is low) (Fig. 1D). The Lambert W Function derives from integration of the well-known Michaelis-Menten enzyme equation,

$$v=\frac{d\left[PIP2\right]}{d\tau }=-\frac{V_{\mathit{max}}\left[PIP2\right]}{K_M+\left[PIP2\right]},$$ [1]

Fig. 1.

Summary of PLCβ functional assay and analysis. (A) Cartoon schematic of planar lipid bilayer setup used to measure PLCβ3 function. (B–D) Summary of analysis of PLCβ3-dependent current decays. The PIP2 calibration curve for GIRK-ALFA (C) is used to convert the normalized current decay (using 29 nM PLCβ3 and lipidated Gβγ) (B) to PIP2 decay (D) (14). Points on the normalized current decay are matched to [PIP2] and time. The resulting PIP2 decay as a function of time is fit (R² = 0.97) to the Lambert W Function (SI Appendix, Eq. S3) derived through integration of the Michaelis-Menten equation with free parameters V_max and K_M shown. (E) Corresponding Michaelis-Menten curve with the K_M and V_max values determined in (D).

which, when integrated from τ = 0 to τ = t, yields

$$\left[PIP2\left(t\right)\right]=K_{M}ProductLog\frac{e^{\frac{\left(\left[PIP2\left(0\right)\right]-t{V}_{\mathit{max}}\right)}{K_M}}\left[PIP2\left(0\right)\right]}{K_M}.$$ [2]

An instructive description of the relationship between the Michaelis-Menten equation (Eq. 1) and the Lambert W Function (Eq. 2) and the suitability of the latter to our studies is given in SI Appendix, Appendix 1. In practice, we fit the normalized current data, i.e., Figs. 2A and 3 and SI Appendix, Fig. S2, directly with a function that is the Lambert W Function transformed by the calibration curve that converts normalized current to PIP2 concentration (Fig. 1C and SI Appendix, Eq. S3). This function has three free parameters, V_max, K_M, and C, the latter a dimensionless (I/I_max) current accounting for the small (almost inconsequential) nonspecific “leak” current observed at the longest recorded times (Figs. 1B and 2A and SI Appendix, Fig. S2A). This enzyme assay permits reproducible estimates of V_max and K_M for PLCβ3 over a wide range of experimental conditions (14).

Fig. 2.

Activation of PLCβ3 by nonlipidated Gα_q. (A) Representative PLCβ3-dependent normalized current decay with 29 nM enzyme in the absence of Gα_q (gray) or in the presence of 1.0 μM Gα_q (black) fit to SI Appendix, Eq. S3 (red curve). In the absence of Gα_q (gray), V_max = 0.0031 mol%/s, K_M = 0.52 mol% (R² = 0.97). In the presence of Gα_q, V_max = 0.091 mol%/s, K_M = 0.42 mol%, C = 0.00092, R² = 0.99. (B) V_max as a function of Gα_q concentration for 29 nM PLCβ3 fit to $V_{max}\left[G{\alpha }_q\right]=\left(V_{max}\left[G{\alpha }_q\to \infty \right]-V_{max}\left[G{\alpha }_q=0\right]\right)*\left(\frac{G{\alpha }_q}{G{\alpha }_q+E{C}_{50}}\right)+V_{max}\left[G{\alpha }_q=0\right]$  , for V_max[Gα_q→∞] and EC₅₀ where V_max[Gα_q = 0] is the V_max in the absence of Gα_q, 0.0026 mol%/s (14), EC₅₀ = 120 nM and V_max[Gα_q→∞] = 0.095 mol%/s, R² = 0.91. Individual points are from at least 3 repeats and the error bars are SEM. (C) K_M for 29 nM PLCβ3 as a function of Gα_q concentration. Dashed line highlights the mean of all values. Individual points are from at least 3 repeats and the error bars are SEM. (D) Membrane partitioning curve for wild-type PLCβ3 (100 nM and 300 nM) in the presence of 200 nM Gα_q Q209L (blue) for 2DOPE:1POPC:1POPS (wt:wt:wt) LUVs with Fraction Partitioned (F_p) on the Y axis. Points are the average from 2 repeats for each lipid concentration and error bars are range of mean. Data were fit to Eq. 5 to determine K_x (dashed blue curve). K_x = 4.2 × 10⁴, R² = 0.68. Data points and the fit to Eq. 5 for PLCβ3 (100 nM and 300 nM) alone (black) and in the presence of lipidated Gβγ (pink) are shown for comparison, (14).

Fig. 3.

Involvement of the X-Y linker in Gα_q-dependent activation and demonstration of dual activation of PLCβ3 by Gα_q and Gβγ. (A and B) Representative normalized current decay for PLCβ3 lacking the entire X-Y linker (ΔX-Y all) using 290 pM of enzyme in the absence (A) or presence (B) of 200 nM Gα_q fit to SI Appendix, Eq. S3 to determine V_max and K_M (red curves), R² = 0.96, R² = 0.98 for A and B. (C and D) Representative wildtype PLCβ3-dependent normalized current decay using 29 nM enzyme in the presence of lipidated Gβγ and 1.0 nM Gα_q (C) or 10 nM Gα_q (D) fit to SI Appendix, Eq. S3 (red curve) to determine V_max and K_M, R² = 1.0 for C and D.

To ensure that PLCβ3 was not affected by product inhibition under our assay conditions, we tested its function in the presence of Gβγ and 1.0 mol% DAG or 1.0 μM IP3 (SI Appendix, Fig. S3). Current decays and determined values for V_max and K_M are indistinguishable from experiments without DAG and IP3 (SI Appendix, Fig. S3), indicating that PLCβ3 is not inhibited by the products of its catalyzed reaction in our experimental setup.

To study the activation of PLCβ3 by Gα_q we used nonlipidated Gα_q (10, 11, 15). Because the GTP bound form of Gα_q is required to activate PLCβ3, we used a hydrolysis-deficient mutant (Q209L) that remains constitutively bound to GTP (SI Appendix, Fig. S1C). Activation by this mutant is similar to wild-type Gα_q (21–24), and migration on a size exclusion column as a complex with PLCβ3 is indistinguishable from wild-type Gα_q (SI Appendix, Fig. S1 A and B). When Gα_q is added to the enzyme assay in the absence of PLCβ3, it does not affect GIRK-ALFA current (SI Appendix, Fig. S1D).

Fig. 2A shows the influence of nonlipidated Gα_q on PIP2 hydrolysis. In the presence of 29 nM PLCβ3 in aqueous solution, the decay of GIRK-ALFA current is slow (reduction of ~15% over 30 s), reflecting slow hydrolysis of PIP2. In the presence of 1.0 μM Gα_q, by contrast, the current reduction is faster, reflecting more rapid hydrolysis. The red curves correspond to fits to SI Appendix, Eq. S3 and yield V_max = 0.0031 mol%/s, K_M = 0.52 mol% (R² = 0.97) in the absence of Gα_q and V_max = 0.091 mol%/s, K_M = 0.42 mol% (R² = 0.99) in the presence of 1.0 μM Gα_q. By performing these experiments in multiples, with different concentrations of Gα_q (0 to 1,000 nM) in the aqueous solution that interfaces the lipid bilayer, we observe that Gα_q increases V_max without affecting K_M (Fig. 2 B and C and SI Appendix, Fig. S2 A–F). The red dashed curve in Fig. 2B is a rectangular hyperbola with a half activation concentration for Gα_q about 120 nM. The maximum increase in V_max elicited by Gα_q is about 35-fold above V_max in the absence of Gα_q. Previous work showed a 20 to 50-fold enhancement of PLCβ3 catalysis, but any relationship to V_max or K_M was unknown in earlier studies of Gα_q (9–11). We note that the effect of Gα_q to increase V_max without affecting K_M is exactly what we observed for Gβγ stimulation of PLCβ3 (14). However, as we will show below, the origins of these apparently similar effects are mechanistically distinct.

Gα_q Modifies k_cat of PLCβ3 Catalysis.

Because PLCβ3 is an aqueous-soluble enzyme that must partition onto the membrane surface to catalyze PIP2 hydrolysis, the kinetic parameter V_max is the product of two separate quantities, expressible as

$$V_{\mathit{max}}={[PLC\beta 3}_m]\ast k_{\mathit{cat}},$$ [3]

where [PLCβ3_m] is the local mole fraction concentration of PLCβ3 on the membrane surface (subscript m) and k_cat is the first-order rate constant for the hydrolysis of PIP2 by a PLCβ3·PIP2 complex on the membrane surface. In principle, to increase V_max, Gα_q could affect either or both quantities. It is not clear whether the nonlipidated Gα_q used in our experiments retains membrane binding (10, 11, 15, 16). To examine whether nonlipidated Gα_q affects the membrane concentration of PLCβ3, we measured whether Gα_q changes the degree to which PLCβ3 partitions onto the membrane, i.e., whether Gα_q recruits PLCβ3 to the membrane surface. The concentration of PLCβ3 at the membrane is determined by its partition coefficient, K_x, which is the ratio of the mole fraction PLCβ3 on the membrane [PLCβ_m] to the mole fraction of PLCβ3 in aqueous solution [PLCβ_w]:

$$K_x=\frac{\left[{PLC\beta }_m\right]}{\left[{PLC\beta }_w\right]}.$$ [4]

We used a vesicle spin-down assay to measure the fraction of Gα_q or PLCβ3 in the absence and presence of Gα_q that binds to the membrane (F_p). This was done by incubating large unilamellar vesicles (LUVs) consisting of 2DOPE:1POPC:1POPS (wt:wt:wt) with Gα_q or PLCβ3 (±Gα_q), centrifuging the mixture, and then measuring the fraction of Gα_q or PLCβ3 associated with the membranes (SI Appendix, Fig. S4 A–C). These experiments were carried out at several lipid concentrations and the measured F_p for Gα_q or PLCβ3 (±Gα_q) was graphed as a function of lipid concentration (Fig. 2D and SI Appendix, Fig. S4 E–G). The dashed curves correspond to the function

$$\begin{array}{c}{F}_p=\frac{\left[{PLC\beta 3}_m\right]{\left[L\right]}_{\mathit{molar}}}{\left[{PLC\beta 3}_m\right]{\left[L\right]}_{\mathit{molar}}+\left[{PLC\beta 3}_w\right]{\left[W\right]}_{\mathit{molar}}}\hfill \\ =\frac{K_x{\left[L\right]}_{\mathit{molar}}}{K_x{\left[L\right]}_{\mathit{molar}}+{\left[W\right]}_{\mathit{molar}}},\hfill \end{array}$$ [5]

where [W]_molar is the molar concentration of water, ~55 M, and [L]_molar is half the total lipid concentration, recognizing that proteins can only access the outer leaflet of the LUVs. Therefore, K_x is the only free parameter (Fig. 2D and SI Appendix, Fig. S4 F and G) (14, 25). The results show that neither wildtype nor the Q209L mutant Gα_q affect partitioning of PLCβ3 onto these membrane surfaces (Fig. 2D and SI Appendix, Fig. S4 B, C, and F). Moreover, Gα_q alone, at least the nonlipidated version used in these experiments, does not partition onto membranes in our experiments, in contrast to previously reported results (15) (SI Appendix, Fig. S4 A and E).

Having established that nonlipidated Gα_q used in these studies does not increase the membrane-bound concentration of PLCβ3, from Eq. 3 we conclude that the increase in V_max in the presence of Gα_q must come from an increased k_cat. In the experiments shown in Fig. 2B, the aqueous concentration of PLCβ3 was 29 nM = 5.3 × 10⁻⁸ mol%, which, using the partition coefficient K_x = 2.9 × 10⁴ (14) and Eq. 4, yields a membrane concentration for PLCβ3, [PLCβ_m] = 1.5 × 10⁻³ mol%. Therefore, from V_max = 0.091 mol%/s (Fig. 2A) and Eq. 3, we calculate k_cat ~60 s⁻¹ in the presence of a maximally activating concentration of Gα_q, which is about 35-fold higher than k_catin the absence of Gα_q (Fig. 2A and Table 1) (14). This finding contrasts the influence of Gβγ on PLCβ3 function, which increases V_max almost entirely through membrane recruitment with little effect on k_cat (Table 1) (14). We note that in the cellular environment where Gα_q is lipidated and membrane-associated, it is likely to also increase [PLCβ_m] in addition to the established increase in k_cat.

Table 1.

Kinetic parameters for PLCβ3

Condition	KM (mol%)	kcat (s−1)	Fold-increase*
PLCβ3 alone†	0.43 ± 0.05	1.7 ± 0.5	-
PLCβ3 + lipidated Gβγ†	0.42 ± 0.05	3.2 ± 0.5	1.9
PLCβ3 + Gαq‡	0.51 ± 0.04	56.9 ± 8	34
PLCβ3 ΔX-Y all	0.33 ± 0.04	1,977.5 ± 150	1,160
PLCβ3 ΔX-Y all + Gαq§	0.31 ± 0.02	2,485 ± 250	1,460
PLCβ3 ΔX-Y contact	0.36 ± 0.02	588.5 ± 90	346
PLCβ3 ΔX-Y contact + Gαq§	0.34 ± 0.04	1,161.3 ± 140	679

^*Over wild-type basal activity.

^†Previously determined (14).

^‡For saturating Gα_q, 300 nM.

^§For 200 nM Gα_q.

Our observation that Gα_q increases k_cat without discernably affecting K_M places constraints on the rate constants of a Michaelis-Menten kinetic reaction scheme. Specifically, for the reaction $(PLC\beta 3 +$ $PIP2\stackrel{k_1}{\underset{k_{-1}}{\leftrightarrows }}PLC\beta 3·PIP2\stackrel{k_{cat}}{\to }PLC\beta + IP3+DAG)$ , where k₁ and k₋₁ are the forward and reverse rate constants for PLCβ3·PIP2 complex formation and k_cat the catalytic step, K_M is

$$K_M=\frac{k_{-1}+k_{\mathit{cat}}}{k_1}.$$ [6]

The most likely explanation for a 35-fold change in k_cat with little effect on K_M is that k₋₁ >> k_cat so that the value of the numerator is little affected by changes in the smaller quantity, k_cat. In the framework of the above reaction scheme, this would imply that most encounters between PLCβ3 and PIP2 dissociate prior to hydrolysis.

Gα_q-Dependent Activation Is Dependent on the X-Y Linker.

A natural explanation for how Gα_q increases k_cat is that it somehow destabilizes the interaction between the autoinhibitory X-Y linker and the active site, allowing it to be displaced with a higher probability. To test this possibility, we expressed and purified PLCβ3 lacking the entire X-Y linker (R471-V584, PLCβ3 ΔX-Y all) or the segment of the linker in direct contact with the active site (T575-V584, ΔX-Y contact) and tested their basal and Gα_q-dependent catalytic activity. If Gα_q-dependent activation is mediated through the X-Y linker, then the maximum fold-activation by Gα_q should be significantly reduced, which has been previously reported (12, 13, 26).

Both X-Y linker mutants exhibited significantly increased basal (i.e., unstimulated by G proteins) V_max: ~2,300-fold for PLCβ3 ΔX-Y all and ~700-fold for PLCβ3 ΔX-Y contact, consistent with the autoinhibitory function of the linker (Fig. 3A and SI Appendix, Fig. S2G). Membrane partitioning experiments showed that membrane association is enhanced only ~two-fold in the ΔX-Y all construct (SI Appendix, Fig. S4 D and G). Therefore, the increase in basal activity is primarily due to an increase in k_cat, ~1,100-fold for PLCβ3 ΔX-Y all and ~350-fold for PLCβ3 ΔX-Y contact (Table 1). This observation also indicates that partitioning is not significantly influenced by the X-Y linker. In addition, the K_M values for the deletion mutants were not significantly different than wildtype (Table 1), suggesting that the linker does not simply act as a competitive inhibitor, blocking PIP2 from binding to the active site. The small difference in basal activity between the two constructs, ~three-fold, suggests that most of the autoinhibitory impact is mediated by the residues in direct contact with the active site.

Addition of 200 nM Gα_q, which produces a ~20-fold increased V_max in wild-type PLCβ3 (Fig. 2B), had less than a two-fold effect on V_max for PLCβ3 ΔX-Y and ~two-fold for PLCβ3 ΔX-Y contact (Fig. 3B, SI Appendix, Fig. S2H, and Table 1). Thus, an intact autoinhibitory X-Y linker is required for Gα_q-dependent activation. Because the lack of Gα_q-dependent activation is comparable in the two mutants, stimulation by Gα_q is likely mediated through the residues that directly contact the active site. Taken together, these results suggest that the presence of the X-Y linker in the active site is a major suppressor of k_cat and that Gα_q-dependent activation is mediated through partial relief of this suppression.

One might wonder whether the relative insensitivity of the catalytic rate to Gα_q in the ΔX-Y mutants could reflect PIP2 depletion near the active site owing to the relatively high catalytic rates in these mutants, i.e., substrate access becomes diffusion-limited. Based on a calculation presented in SI Appendix, Appendix 2, we think this is unlikely to be the case. More likely, allosteric regulation of the active site of PLCβ3 is mediated at least in part through the inhibitory X-Y linker and manifests kinetically through the altered k_cat that we observe.

Simultaneous Activation of PLCβ3 by Gα_q and Gβγ.

We have demonstrated that nonlipidated Gα_q and lipidated Gβγ activate PLCβ3 through different mechanisms, Gβγ through membrane recruitment to increase the membrane concentration of enzyme and Gα_q by increasing the catalytic rate constant. Given these observations, we suspected that dual activation of PLCβ3 by both G proteins would combine both mechanisms, which would lead to a product, rather than a sum, of the two effects (Eq. 3). To test this idea, we measured PLCβ3 activity in the presence of a high concentration of lipidated Gβγ and 1.0 nM or 10 nM Gα_q (Fig. 3 C and D). The current decays in the bilayer enzyme assay were very rapid, but still well fit by the transformed Lambert W Function, thus permitting determination of V_max and K_M (Fig. 3 C and D). Fig. 3 C and D show that Gα_q induces a concentration-dependent increase in V_max in the presence of Gβγ, as was observed in the absence of Gβγ (Table 2). Moreover, the fold-increase of V_max in the presence of Gα_q compared to that in the absence of Gα_q is approximately the same whether Gβγ is present or not (Table 2). This supports the independent action of Gα_q and Gβγ and the conclusion that together both G proteins increase V_max by a produce rule.

Table 2.

Effect of dual activation with Gα_q and Gβγ on V_max

Condition	Vmax (mol%/s)	Total fold-increase	Fold-increase over 0 Gαq
PLCβ3 alone*	0.0026 ± 0.0007	-	-
PLCβ3 + 1.0 nM Gαq	0.0068 ± 0.0007	2.6 ± 0.3	2.6 ± 0.3
PLCβ3 + 10 nM Gαq	0.0076 ± 0.001	2.9 ± 0.4	2.9 ± 0.4
PLCβ3 + Gβγ*	0.17 ± 0.02	65	-
PLCβ3 + Gβγ + 1.0 nM Gαq	0.34 ± 0.1	129 ± 38	1.8 ± 0.5
PLCβ3 + Gβγ + 10 nM Gαq	0.55 ± 0.1	213 ± 44	2.9 ± 0.6

^*Previously reported (14).

Structure of the PLCβ3·Gα_q Complex on Lipid Vesicles.

We determined the structure of the PLCβ3·Gα_q complex associated with lipid vesicles at 3.4 Å (Fig. 4 and SI Appendix, Fig. S5 and Table S1). The sample was prepared by combining PLCβ3 and wildtype Gα_q bound to GDP-AlF₄, purifying the complex using size exclusion chromatography (SI Appendix, Fig. S1A), and then mixing the purified complex with lipid vesicles composed of 2DOPE:1POPC:1POPS (wt:wt:wt). The structure of the complex contains density for the PLCβ3 catalytic core and the proximal CTD, but the CTD linker and the distal CTD are disordered (Fig. 4 A and B), suggesting conformational heterogeneity of the domains with respect to each other. The overall complex is very similar to the previously determined crystal structure, including the X-Y linker engaged in the active site, with a Cα rmsd of 0.84 Å (10) (Fig. 4 B and C). Despite the disordered distal CTD, which was previously shown to be part of the PLCβ3·Gα_q interface (10), the interface between the PLCβ3 catalytic core and Gα_q is extensive, burying ~1,500 Ų and involving 56 residues, 27 from Gα_q and 29 from PLCβ3 (Fig. 4E and SI Appendix, Fig. S7A and Table S2 and S3). Compared to the structure of the catalytic core in the absence of Gα_q, the only conformational difference is the displacement of the Hα2’ away from the catalytic core (Fig. 4D). Despite its proximity to the catalytic site, the Hα2’ displacement does not induce additional changes in that region (Fig. 4D). Membrane association of the complex also does not produce conformational differences other than the additional heterogeneity between the catalytic core and the distal CTD (Fig. 4C).

Fig. 4.

Structure of the PLCβ3·Gα_q complex on lipid vesicles. (A) Primary structure arrangement of PLCβ3 enzymes. Sections are colored by domain. The PH domain is yellow, the EF hand repeats are pink, the C2 domain is light teal, the Y domain is green, the X domain is light blue, the X-Y linker and the pCTD are red. Domains in gray (CTD linker and Distal CTD) are not observed in our structures. pCTD is proximal CTD. (B) Sharpened, masked map of the PLCβ3·Gα_q complex colored by protein. PLCβ3 is yellow and Gα_q is pink. The autoinhibitory elements in PLCβ3, the X-Y linker and the pCTD, are colored red. (C) Structural alignment of the PLCβ3·Gα_q complex on membranes, colored by domain as in A, with the previously determined crystal structure of the complex [PDBID: 4GNK, (10)], in gray. Cα rmsd is 0.84 Å. Calcium ion from the cryo-EM structure is shown as a yellow sphere and the active site is denoted with an asterisk. (D) Structural alignment of PLCβ3·Gα_q complex on membranes, colored by protein-PLCβ3 is yellow and Gα_q is pink, with the previously determined cryo-EM structure of the apo catalytic core [PDBID: 8EMV, (14)] colored in gray. The X-Y linker and the pCTD from the Gα_q complex are colored red, the calcium ion from the cryo-EM structure is shown as a yellow sphere and the active site is denoted with an asterisk. The Hα2’ from the apo structure is colored in blue to highlight its position on the catalytic core and an arrow denotes the Gα_q-dependent movement of the Hα2’. (E) Surface representation of the PLCβ3·Gα_q interface peeled apart to show extensive interactions. Residues on PLCβ3 that interact with Gα_q are colored in pink and residues on Gα_q that interact with PLCβ3 are colored in yellow. Interface residues were determined using the ChimeraX interface feature using a buried surface area cutoff of 15 Ų.

Despite our finding that the X-Y linker is involved functionally in Gα_q-dependent activation, we do not observe structural differences at the active site or its interface with the X-Y linker, even with extensive classification targeting that region. This observation is not too surprising, however, given the magnitude of activation of Gα_q compared to the activity in the absence of linker. The basal k_cat and maximal Gα_q-stimulated k_cat are only ~0.09% and ~3%, respectively, of the activity in the absence of the linker, suggesting that Gα_q does not alter the probability of its occupancy in the active site enough to be observable in structural experiments. In other words, if we take the activity in the absence of the linker as zero occupancy of the linker in the active site, then even in the presence of saturating Gα_q, the linker would only be displaced 3% of the time, which is not easily detectable using cryo-EM.

Structure of the PLCβ3·Gβγ(2)·Gα_q Complex on Lipid Vesicles.

We also determined the structure of the PLCβ3·Gβγ(2)·Gα_q complex bound to lipid vesicles to 3.4 Å resolution (Fig. 5 and SI Appendix, Fig. S6 and Table S1). We reconstituted lipidated Gβγ into vesicles comprised of 2DOPE:1POPC:1POPS (wt:wt:wt) as previously described (14), mixed PLCβ3, and wildtype Gα_q bound to GDP-AlF₄ and added the complex to the Gβγ-containing lipid vesicles for grid preparation. The structure contains the PLCβ3 catalytic core and proximal CTD, two Gβγ molecules, and one Gα_q molecule (Fig. 5 and SI Appendix, Fig. S6). The CTD linker and distal CTD were disordered, as in the other structures of PLCβ3·G protein complexes on membranes (14). The structure is very similar to the structures of PLCβ3 in complex with each G protein on its own, with no additional conformational changes observed (Fig. 5 B and C). The X-Y linker is present in the active site and the Hα2’ is in the Gα_q-bound conformation (Fig. 5 A and B). Each of the PLCβ3·G protein interfaces is unaltered by the presence of the additional G protein (Fig. 5 and SI Appendix, Fig. S7 B–D and Tables S2 and S3). These observations are consistent with the functional experiments, which show that binding of one G protein does not influence the other, and that they act independently to give a product rule for catalytic enhancement when both G proteins are present.

Fig. 5.

Structure of the PLCβ3·Gβγ(2)·Gα_q complex on lipid vesicles. (A) Sharpened, masked map of the PLCβ3·Gβγ(2)·Gα_q complex colored by protein. PLCβ3 is yellow, Gα_q is pink, Gβ 1 is dark teal, Gγ 1 is purple, Gβ 2 is light blue, Gγ 2 is gray. The X-Y linker and pCTD are colored red. (B and C) Structural alignment of PLCβ3·Gβγ(2)·Gα_q complex on lipid vesicles, colored as in A, with the PLCβ3·Gα_q complex on lipid vesicles in gray (B, Cα rmsd = 0.62 Å) or with the PLCβ3·Gβγ(2) complex on lipid vesicles in gray (C, Cα rmsd = 0.63 Å), [PDBID: 8EMW, (14)].

Membrane Association of PLCβ3·G Protein Complexes.

Unmasked classification on the aligned particle subsets for each complex yielded reconstructions with density for the lipid bilayer, allowing us to study the orientation of each complex on the membrane (Fig. 6). Two different membrane-associated reconstructions of the PLCβ3·Gα_q complex were observed, in which the catalytic core associates with the membrane and orients the active site toward the membrane (Fig. 6 B and C). There were no differences in the protein components of each reconstruction, suggesting that the complex tilts on the membrane as a rigid body (Fig. 6 B and C). This orientation differs from PLCβ3 in the absence of G proteins, where the catalytic core extends away from the membrane (Fig. 6A) (14). This orientation also differs from the PLCβ3·Gβγ(2) complex where the two Gβγs anchor the catalytic core to the membrane on the opposing side, resulting in the catalytic site tilting away from the membrane (14).

Fig. 6.

Membrane association of PLCβ3 in the presence of G proteins. (A) Unsharpened reconstruction of PLCβ3 bound to lipid vesicles in the absence of G proteins shown for comparison (14). PLCβ3 is colored in yellow and the membrane is colored in gray. (B and C) 3D reconstructions of two different orientations of the PLCβ3·Gα_q complex on the membrane surface. The reconstructions are colored by protein, PLCβ3 is yellow, Gα_q is pink, and the membrane is gray. The PLCβ3 X-Y linker is colored red to highlight the active site within the catalytic core. (D) 3D reconstructions of six 3D classes of the PLCβ3·Gβγ(2)·Gα_q complex on membranes showing different positions of the complex with respect to the membrane arranged by degree of tilting. The reconstructions are colored by protein as in B and C, and Gβ 1 is dark teal, Gγ 1 is purple, Gβ 2 is light blue, Gγ 2 is gray. The N terminus of Gα_q is colored blue for reference.

The PLCβ3·Gα_q orientation seems more poised for catalysis as the active site is oriented directly toward the membrane (Fig. 6 B and C). It appears likely that this orientation is driven by the Gα_q-induced conformational change of the Hα2’ because, when Gα_q binds, the Hα2’ is displaced from the catalytic core and an underlying hydrophobic patch is exposed on the surface of the catalytic core (SI Appendix, Fig. S8 A and B). The point of membrane association in the complex is very close to this hydrophobic patch, suggesting that it plays a role in positioning the complex on the membrane (SI Appendix, Fig. S8 C and D). If we fit the catalytic core in the absence of G proteins into the density of the PLCβ3·Gα_q complex on the membrane, the Hα2’ protrudes near the membrane density, suggesting that it could hinder membrane association in this configuration. This suggests that Gα_q binding to PLCβ3, even without the lipid anchor, could indirectly play a role in orienting the PLCβ3 catalytic core on the membrane. Such an orientation effect would apply to PLCβ3s that have partitioned onto the membrane, rather than on the partitioning step, consistent with the observation that Gα_q does not alter membrane association. It is possible that this orientation of the PLCβ3·Gα_q complex on the membrane contributes to the Gα_q-mediated displacement of the X-Y linker from the active site.

For the complex with both G proteins, the orientation resembles that of the PLCβ3·Gβγ(2) complex, where the two Gβγs firmly anchor the PH domain and EF hands to the membrane and the other side of the catalytic core tilts away (Fig. 6D) (14). We also observed variation in orientation of the complex with PLCβ3 and both G proteins on the membrane, as in the PLCβ3·Gβγ(2) complex. Six reconstructions with at least 4 Å resolution were observed with differing tilt angles of the catalytic core with respect to the membrane, ranging from 34° in the most tilted to 13° in the least tilted (Fig. 6D). There are no changes to the protein components in these reconstructions, suggesting again that the complex tilts on the membrane as a rigid body. The membrane orientation seems to be driven by the Gβγs under these conditions, which we speculate is due to the lipid anchor on Gβγ and the lack thereof on Gα_q.

However, the observed orientations are not incompatible with a lipid anchor on Gα_q, which would likely be present in a cell. In our structures, the N terminus of Gα_q is disordered until position 38 (Fig. 6D, blue region) and the lipid modifications are placed on cysteines at positions 9 and 10. Even in our most tilted reconstruction, where the Gα_q N terminus is ~85 Å from the membrane (Fig. 6D), the disordered portion is long enough for the lipid anchors to reside in the membrane. This observation is consistent with other structures of Gα subunits in complex with their effectors, including adenylyl cyclase and TRPC5 (27–30), where the Gα is positioned ~50 Å from the membrane and the N terminus is disordered. These observations are consistent with our functional experiments showing that Gα_q activates PLCβ3 by increasing k_cat rather than through membrane recruitment. However, it is possible that a lipidated Gα_q might also recruit PLCβ3 to the membrane in addition to increasing its k_cat.

Discussion

In a recent study, we analyzed the structural and enzymatic properties of PLCβ3 in the absence and presence of Gβγ on lipid vesicles (14). We found that PLCβ3 catalyzes PIP2 hydrolysis in accordance with Michaelis-Menten enzyme kinetics with a very small k_cat (~1.7 s⁻¹) but that Gβγ can increase net catalysis by binding to PLCβ3 and thus recruiting it to the membrane. It is known that Gα_q also increases net catalysis (8, 10, 11, 16). In this study, we investigate the influence of Gα_q on PLCβ3 activity. We used Gα_q that does not contain a lipid anchor. Our essential findings are as follows: 1) Nonlipidated Gα_q increases V_max in a concentration-dependent manner, following a rectangular hyperbola, consistent with 1:1 binding of Gα_q to PLCβ3. The apparent equilibrium constant for binding is ~120 nM, and maximal activation is ~35-fold greater than the basal (i.e., in the absence of Gα_q) catalytic rate. 2) Gα_q without a lipid anchor does not partition onto the membrane surface nor does it influence the degree to which PLCβ3 partitions onto the membrane surface. Thus, Gα_q without a covalent lipid anchor increases V_max by increasing k_cat. 3) The ability of Gα_q to increase k_cat depends on the presence of the X-Y linker autoinhibitory element on PLCβ3. 4) Gα_q and Gβγ act independently to increase V_max. Consequently, when both G proteins are applied simultaneously, the net increase in PLCβ3 catalytic activity is given by the product of the two individual effects. Under the conditions in which we have studied PLCβ3 enzyme activity, maximal dual stimulation can increase PIP2 hydrolysis greater than 2,000-fold. 5) Structures of PLCβ3 on lipid membrane vesicles alone, with Gα_q, with Gβγ, and with both G proteins together, show that two Gβγ and one Gα_q bind to PLCβ3 simultaneously and independently, consistent with their influence on PLCβ3 catalysis. In summary, two Gβγ localize (i.e., recruit) PLCβ3 to the membrane. Independently, Gα_q increases k_cat. Mutational studies support the hypothesis that Gα_q regulates k_cat allosterically through the autoinhibitory X-Y linker (Fig. 7).

Fig. 7.

Hypothesized mechanism of activation of PLCβ enzymes by Gβγ and Gα_q. When PLCβ3 binds the membrane, the active site is positioned away from the membrane and the enzyme is autoinhibited by both the X-Y linker and the Hα2’, resulting in low activity in the absence of G proteins. (A) Free Gβγ binds to membrane-associated PLCβ3, increases its concentration at the membrane and orients the active site for catalysis, leading to an increase in PIP2 degradation. However, the k_cat is limited by both the X-Y linker and the Hα2’ (shown in red and dark red, respectively). (B) Free Gα_q binds to membrane associated PLCβ3, displaces the autoinhibitory Hα2’ (shown in dark red) and the X-Y linker is more frequently absent from the active site, resulting in an increase in k_cat and PIP2 turnover. (C) Free Gα_q and Gβγ both bind to membrane-associated PLCβ3, leading to a combination of the activation effects of each G protein. The final result is increased PLCβ3 on the membrane surface with reduced autoinhibition (both the Hα2’ and the X-Y linker) at the membrane, leading to robust PIP2 hydrolysis. The distal CTD of PLCβ3 was omitted for clarity.

There is one difference in the conditions of our partitioning experiments and the kinetic experiments for PLCβ3 function: The partitioning experiments are carried out in the absence of PIP2. We could not include PIP2 in the partitioning experiments because it would be hydrolyzed throughout the measurement. However, if PIP2 did influence the partition coefficient for PLCβ3, it would not affect our conclusion that Gα_q (without a lipid anchor) does not alter the local concentration PLCβ3 in the membrane and thus increases V_max by increasing k_cat. As shown in Fig. 2D, Gα_q does not alter the fraction of PLCβ3 partitioned, whereas Gβγ does. Enhanced partitioning caused by Gβγ accounts for most of its effect on catalysis (14). That Gα_q does not enhance partitioning is independent of the precise value of the PLCβ3 partition coefficient. Thus, we can attribute the ability of Gα_q to increase the V_max of PLCβ3 by ~35-fold as an increase in k_cat, not its local concentration.

In the enzyme assay, k_cat for PLCβ3 without Gα_q stimulation is ~1.7 s⁻¹ (14), with maximal Gα_q stimulation ~60 s⁻¹, and with the X-Y linker removed by mutation ~2,000 s⁻¹. If we take 2,000 s⁻¹ as the magnitude of k_cat without autoinhibition, then wild-type PLCβ3 in the absence of Gα_q is inhibited by the X-Y linker more than 99.9% of the time and in the presence of a maximally activating concentration of Gα_q it is still inhibited about 97% of the time. On top of this, the partition coefficient of PLCβ3 is such that nearly all of it in a cell is in the aqueous solution, not on the membrane, in the absence of G protein stimulation (14). Why has nature so severely suppressed the catalytic activity of this enzyme? The answer, we propose, is that excessive background activity of PLCβ3 activity will have severe consequences for the stability of cells. In fact, naturally occurring mutations show this to be the case (31–33). Not only does PIP2 regulate the activity of many membrane channels, transporters and receptors, but of equal importance, the products of PLCβ3-mediated PIP2 hydrolysis, DAG and IP3, regulate protein kinase C and the IP3 receptor, which control phosphorylation of many proteins and intracellular Ca²⁺ concentration, respectively. Therefore, we propose that there is strong evolutionary “pressure” to minimize baseline PLCβ3 activity. Combining the results in our previous study (14) and in the present study, we can understand how, in the setting of intense catalytic suppression, catalysis still occurs in abundance when it is called for (Fig. 7). Gβγ, by binding to PLCβ3, recruits it to the membrane (Fig. 7A). Simultaneously, Gα_q can increase k_cat ~35 fold through partial relief of X-Y linker inhibition (Fig. 7B). We show that under the conditions of our experiments, together these two regulatory mechanisms can enact a greater than 2,000-fold increase in PLCβ3 activity (Fig. 7C). Lipidated Gα_q may further concentrate PLCβ3 on the membrane, leading to an even greater increase in activity upon receptor stimulation.

While our experiments leave little question about the involvement of the X-Y linker in Gα_q-dependent activation, it remains unclear exactly how Gα_q binding alters the association of the linker in the active site. The only observed conformational change in the protein upon Gα_q binding is the displacement of the Hα2’ away from the catalytic core (Fig. 4D). Perhaps the displacement of this helix increases the dynamics in the catalytic core, allowing the X-Y linker to be displaced more frequently as previously proposed (34, 35). We also observed a Gα_q-dependent change in orientation of the catalytic core on the membrane, which could be related to the Hα2’ displacement (Fig. 6 B and C and SI Appendix, Fig. S8). This change in membrane orientation is consistent with previous results showing that the membrane plays a role in Hα2’ autoinhibition and that Gα_q only activates PLCβ3 in the presence of membranes (13). In the Gα_q-dependent orientation, the PLCβ3 active site is oriented toward the membrane, which could potentially displace the X-Y linker through repulsion of its adjacent acidic stretch by the negatively charged lipids. Such a mechanism has been previously proposed (8, 13), but our observations offer a new subtlety in that the linker could be transiently displaced based on the orientation of the catalytic core on the membrane rather than a stable displacement following membrane partitioning. Involvement of the Hα2’, as in either of these potential mechanisms, leads to the proposal that the autoinhibitory function of the Hα2’ is related to its coupling to the X-Y linker. However, previous studies have proposed that the autoinhibition by the Hα2’ and the X-Y linker are independent (9). Further experiments are necessary to fully understand the mechanism of X-Y linker displacement by Gα_q and Hα2’ autoinhibition.

As described above, the results from our reconstitution experiments have many implications for signaling in the cellular environment. For example, the observed affinity of PLCβ3 for Gα_q is relatively high, suggesting that a low level of receptor stimulation can lead to robust PLCβ3 signaling. This effect would be further amplified in the cellular context with lipidated Gα_q, which might also increase the local concentration of PLCβ3 on the membrane. Furthermore, because Gβγ and Gα_q activate PLCβ3 by different mechanisms and coactivate as the product of the two influences of each G protein, PLCβ3 is well poised to serve as a coincidence detector of costimulation by Gα_i and Gα_q coupled receptors, even under low levels of costimulation, which would be important for many physiological processes (8, 17, 18).

Materials and Methods

Protein Expression, Purification, and Reconstitution.

All proteins were purified according to previously established protocols using affinity chromatography and size exclusion chromatography. Detailed methods are described in SI Appendix, Materials and Methods: Protein Expression and Purification and Protein Reconstitution.

PLCβ3 Functional Assay.

PLCβ3 activity was measured using a planar lipid bilayer setup and a PIP2-dependent ion channel to report PIP2 concentration in the membrane over time. Detailed methods are described in SI Appendix, Materials and Methods: Bilayer Experiments and Analysis.

Membrane Partitioning Experiments.

Gα_q or fluorescently labeled PLCβ3 was mixed with LUVs and pelleted. Protein in the pellet and supernatant was quantified using fluorescence. Detailed methods are described in SI Appendix, Materials and Methods: PLCβ3 and Gα_q Vesicle Partition Experiments.

PLCβ3·G Protein Complex Structure Determination.

PLCβ3·Gα_q complex was mixed with liposomes with or without Gβγ prior to sample vitrification. Cryo-EM data were collected using a Titan Krios with a Gatan K3 direct detector according to the parameter values in SI Appendix, Table S1 and analyzed according to the procedures outlined in SI Appendix, Figs. S5 and S6. Atomic models from previously determined structures were fit into our density maps, refined using PHENIX real-space refine (36), and manually adjusted. Detailed methods are described in SI Appendix, Materials and Methods: Cryo-EM Sample Preparation and Data Collection, Cryo-EM Data Processing, and Model Building and Validation.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Cryo-EM maps and atomic models for all structures described in this work have been deposited to the Electron Microscopy Data Bank (EMDB) (PLCβ·Gα_q: EMDB-42475 (37) and PLCβ·Gβγ·Gα_q: EMD-42476 (38)) and the Protein Data Bank (PDB) (PLCβ·Gα_q: 8UQN (39) and PLCβ·Gβγ·Gα_q: 8UQO (40)), respectively.