G Protein-coupled receptors: Multi-turnover GDP/GTP exchange catalysis on heterotrimeric G proteins.

G protein-coupled receptors and heterotrimeric G proteins can diffuse laterally in the plasma membrane such that one receptor can catalyze the activation (GDP/GTP exchange) of multiple G proteins. In some cases, these processes are fast enough to support molecular signal amplification, where a single receptor maintains the activation of multiple G proteins at steady-state. Amplification in cells is probably highly regulated. It depends upon the identities of the G receptor and G protein - some do and some don't - and upon the activities of GTPase-activating proteins, membrane scaffolds, and other regulatory partners.

The idea that a single G protein-coupled receptor (GPCR) can sequentially activate multiple heterotrimeric G proteins on the surface of the plasma membranes derived from 2 separate experimental arguments made prior to 1990, and is now generally accepted. GPCRs catalyze GDP/GTP exchange to promote G protein activation, and the definition of a catalyst demands turnover of multiple substrate molecules. Massively catalytic G protein activation is best demonstrated in the mammalian photoreceptor membrane, where the majority of the membrane protein is the GPCR rhodopsin and the vast majority of GTP binding sites is the Gt (see essays by Arshavsky, Liebman, ...). In other cells, where both GPCR and its G protein targets may be well below 0.1% of plasma membrane protein, catalytic turnover by GPCRs remains hard to quantitate. In what membranes it occurs, which receptors and G proteins do it, if and how it is regulated, and which protein(s) is (are) the diffusing species are all variables.

The first suggestion that GPCRs and/or G proteins diffuse laterally and that a receptor can activate multiple G proteins came from the work of Levitzki and coworkers,- who used adenylyl cyclase activity as a surrogate measure of Gs activation; Gs was itself discovered at about the same time. They showed that covalent inactivation of β-adrenergic receptors in erythrocyte membranes progressively decreased the apparent first-order rate of activation of adenylyl cyclase by non-hydrolyzable GTP analogs at a fixed agonist concentration without decreasing maximal stimulation. Such inactivation also shifted the EC50 for agonist to higher concentrations. The number of catalytic turnovers for each receptor before cyclase was fully activated could be estimated from these data to be about 10. With the work of Selinger and coworkers,- who showed that β-adrenergic agonists promote the GDP/GTP exchange step in the receptor-stimulated GTPase cycle, the idea of exchange catalysis by receptors entered the review literature by 1980, if not earlier. Even at this point, however, it was also clear that not all receptors recycle freely. Braun and Levitzki showed that an adenosine receptor appeared by the above criteria to be tightly coupled to Gs, with no sign of multi-molecular turnover.

Accurate quantitation of catalytic turnover of G proteins by GPCRs is difficult in intact cells or isolated plasma membranes, but became possible when the purified proteins were reconstituted into phospholipid vesicles. In an early effort, Pedersen and Ross found that a β-adrenergic receptor could catalyze nucleotide exchange on about 6 Gs molecules, and Asano et al., using more accurate quantitation, demonstrated at least 8 turnovers. A kinetic analysis of Asano’s data similar to that performed by Levitzki and coworkers argued that a single receptor could turn over about 50 Gs molecules. Using a reconstituted system similar to that of Asano et al., Cerione et al. also found that the α2-adrenergic receptor could catalyze GDP/GTP exchange on about 7 molecules of Gi. Mirroring Braun, Senogles et al. then found that the reconstituted D2 dopamine receptor did not turn over multiple Gi molecules. Some receptors do; some don’t. It is uncertain whether receptors that do or do not regulate multiple G protein molecules in phospholipid vesicles might behave differently in cells, although some co-fractionation studies have argued that these behaviors are consistent.

Given that a receptor can sequentially catalyze GDP/GTP exchange on multiple G proteins, at least in some cases, the more relevant physiologic question is whether it does so within the lifetime of the G protein’s GTP-activated state. If it can, then a single agonist-liganded receptor can maintain the activation of multiple G proteins at steady-state, leading to molecular amplification of the signal. Catalytic amplification can also increase agonist potency (decrease EC50) because fewer agonist-bound receptors can activate the same number of G proteins.

Even with catalytic turnover, though, amplification will only occur if the rate of receptor-catalyzed GDP/GTP exchange is faster than the rate of hydrolysis of G protein-bound GTP. Rate constants for G protein deactivation, with or without stimulation by GTPase-activating proteins (GAPs), have not been measured directly in living cells or in intact membranes other than the photoreceptor disc. The deactivation rate constant for Gs and Gi in cells is estimated to be ~0.05–0.1 s−1, based on the rates of signal termination,, and the rates of hydrolysis of GTP bound to purified Gαi and Gαs, are similar to this value. Other G proteins are slower: ~0.02 s−1 for Gαq and ~0.002 s−1 for Gαz, Gα12 and Gα13. GAPs can increase these rates as much as 2000-fold., Somewhat faster rates of G protein activation and deactivation in cells have been estimated according to effector activation/deactivation rates (above) or with fluorescent G protein activation biosensors., While they provide a beautiful insight into the time course of signaling, on/off kinetics has not clarified the amplification question. To do this would require stoichiometric information that is not available from ensemble fluorescence data. In general, stoichiometric relationships among the protein components of plasma membrane signaling networks are poorly understood. They are central to issues of molecular amplification, and they also dictate whether these networks act as first-order (linear) or non-first-order (nonlinear) transducers.

Beyond the not-so-simple activation/deactivation kinetic issues are the questions of cellular constraints on receptor and G protein diffusion. Clearly some receptors and G proteins are highly constrained in their motion. Constraints include stable binding to scaffolding proteins on the molecular scale (1–2 nm) and membrane domain restrictions on the 100–1000 nm scale. Other mechanistic constraints may also arise. For example, we have proposed that one way in which G protein modules use GAPs to accelerate response rates is to promote long-term (10 s vs 0.1 s) receptor-G protein binding., This mechanism overcomes inhibition by the GAP, but at the cost of molecular amplification.

If catalytic amplification occurs at the initial receptor-G protein step in the signaling pathway, we have to know the absolute rate constants for actual nucleotide exchange and GTP hydrolysis, and we don’t for any system I know of. One can envision double-label, single-molecule approaches to the problem even though the technology is not now available. I predict that the question of molecular amplification in cells will depend on such optical techniques perfected to give us both kinetic and stoichiometric data, as well as molecular-scale location. Imaging technology is pushing toward this goal.