Special Perspectives

G Protein Signaling in Yeast: New Components, New Connections, New Compartments

See allHide authors and affiliations

Science  01 Dec 2006:
Vol. 314, Issue 5804, pp. 1412-1413
DOI: 10.1126/science.1134041


Signaling by cell surface receptors and heterotrimeric guanine nucleotide–binding proteins (G proteins) is one of the most exhaustively studied processes in the cell but remains a major focus of molecular pharmacology research. The pheromone-response system in yeast (see the Connections Map at Science's Signal Transduction Knowledge Environment) has provided numerous major advances in our understanding of G protein signaling and regulation. However, the basic features of this prototypical pathway have remained largely unchanged since the mid-1990s. New tools available in yeast are beginning to uncover new pathway components and interactions and have revealed signaling in unexpected locations within the cell.

Many extracellular signals are detected by cell surface receptors and further transmitted inside the cell by G proteins, which serve as molecular switches (1). Activated receptors promote exchange of guanosine triphosphate (GTP) for guanosine diphosphate (GDP) bound to G protein α subunits, triggering the dissociation of the G protein α and βγ subunits and thus allowing them to signal. In yeast, mating pheromones activate a G protein and protein kinase cascade that includes Fus3 and Kss1, two members of the mitogen-activated protein (MAP) kinase family (2). G protein activation is eventually terminated by the intrinsic guanine triphosphatase (GTPase) activity of Gα subunits, a process that is accelerated by proteins known as regulators of G protein signaling (or RGS proteins) (1). Thus, the intensity of the signal depends on the opposing actions of receptors and RGS proteins. Proper functioning of G protein–mediated signal transduction mechanisms is extremely important, because defects in these pathways are implicated in a wide variety of diseases (3). Correspondingly, a substantial fraction of all pharmaceuticals act directly or indirectly on G protein–coupled receptors (4). Thus, it is possible that a new generation of drugs will act on RGS proteins and other newly discovered regulators of G protein activity (5).

The yeast Saccharomyces cerevisiae represents an attractive and appropriate model for investigating basic mechanisms of G protein and MAP kinase signaling (2) and in particular for finding new pathway regulators. Yeast is also emerging as a versatile model for systems-level analysis of complex signaling networks. The yeast genome was the first among eukaryotes to be sequenced. There exist comprehensive microarrays for transcription analysis, gene deletion arrays for genome-scale phenotypic analysis, and gene fusion arrays for systematic protein purification and in situ localization studies. The existence of these resources has profoundly transformed our approach to the identification and functional characterization of cell signaling regulators. Indeed, yeast is the only experimental system where it is possible to track the expression, localization, and activity of nearly every component of the G protein signaling pathway, from the cell surface to the nucleus (Fig. 1).

Fig. 1.

New features of RGS and G protein signaling in yeast. (Top) Activation of the receptor (Ste2) by mating pheromone leads to GTP binding and dissociation of the G protein α subunit (Gpa1) from the Gβγ subunits (Ste4 and Ste18). Gβγ activates multiple effectors at the plasma membrane, including components of a MAP kinase signaling cascade. The RGS protein Sst2 binds to the receptor and accelerates Gpa1 GTP hydrolysis. Casein kinase I (Yck1 and Yck2) phosphorylates the receptor, displacing Sst2 and promoting receptor endocytosis. (Bottom) Gpa1 is also present at the endosome. Activation by endocytosed receptor or another unknown factor may lead to GTP binding and dissociation of Gpa1 from the Gβ-like protein Vps15. Activated Gpa1-GTP then binds directly to the phosphoinositide 3-kinase (PI3K) Vps34. Elevated PI3P recruits proteins containing FYVE domains and PX domains to the endosome. Note that Vps15 and Vps34 are part of a multiprotein complex that also includes Vps30 and Vps38 or Atg14.

In the pheromone response pathway, the Gβγ subunits (Ste4 and Ste18) were long regarded as the sole signal-transmitting component of the G protein heterotrimer. The Gα subunit (Gpa1) was thought to regulate the amounts of free Gβγ: releasing Gβγ when in the activated GTP-bound form and sequestering Gβγ when in the inactivated GDP-bound form. Evidence of a positive signaling role for Gpa1 came from studies showing that GTPase-deficient (and thus constitutively active) Gpa1 mutants, including Gpa1Q323L, activate mating-specific gene transcription and morphological changes in the absence of added pheromone (6).

However, the question still remained: How does activated Gα transmit its signal and contribute to the mating response pathway? A study published this year brings us closer to answering this question (7). A systematic analysis of nearly 5000 gene deletion strains revealed seven genes required for Gpa1Q323L-mediated responses. Two of the genes, VPS34 and VPS15, encode the catalytic and regulatory subunits, respectively, of the sole phosphatidylinositol 3-kinase in yeast. Surprisingly, Vps34 and Vps15 are found at intracellular compartments, such as endosomes and Golgi, rather than at the plasma membrane where receptors and G proteins are normally thought to reside. Activated Gα localizes along with Vps34 at the endosome, binds to Vps34 directly, and triggers increased production of the Vps34 product, phosphatidylinositol 3-phosphate (PI3P). These observations suggest that Vps34 is a bona fide Gα effector capable of generating a second messenger. The regulatory subunit Vps15 has a seven WD40 domain repeat structure and binds directly to the inactive GDP-bound form of Gpa1, both hallmark features of known Gβ subunits. These findings imply that Vps15 functions as an alternate Gβ, but one acting at the endosome instead of at the plasma membrane. In further support of this model, it was shown that Vps15 is needed to target Gα to the endosome, whereas the canonical Gβ subunit Ste4 targets Gα to the plasma membrane (7).

Although it is widely accepted that G proteins signal at the plasma membrane, it has long been recognized that they are also present at intracellular compartments (8). Previous studies have implicated these intracellular G proteins in the regulation of membrane traffic (9). It has also become increasingly evident that membrane trafficking and signal transduction events are closely linked and that G protein signals can originate from intracellular compartments as well as from the plasma membrane (10, 11). Consequently, endosomes should be considered not only as a site of membrane sorting but also as a site of cell signaling (12).

How is the internal pool of G proteins regulated? One possibility is that Gα is initially activated by receptors at the plasma membrane and is then delivered to intracellular compartments. Another possibility is that the agonist-bound receptor transits to the endosome and activates Gα directly. Yet a third possibility is that Gα is activated by a distinct guanine nucleotide exchange factor located at the endosome. In fact, in other organisms a number of nonreceptor accessory proteins have been shown to activate Gα in vitro; of these a physiological role has been demonstrated for mammalian Ric-8A, which activates Gα proteins during cell division (1, 8, 13).

Why would having segregated pools of Gα be physiologically important? Signaling by Gpa1Q323L may provide some clues. Although it mimics most aspects of pheromone binding at the cell surface, Gpa1Q323L at the endosome preferentially activates just one of the two mating-specific MAP kinases and fails to trigger the cell division arrest that normally precedes mating. Thus, whereas Gα signaling at the plasma membrane is clearly important for the early steps of the pheromone response, Gα signaling at the endosome appears to convey a fundamentally different signal.

How does segregation of Gpa1 to two distinct membrane compartments relate to observed functional differences? As mentioned above, there could be differences in the activation step. Alternatively, there could be differences in how the two pools of G protein are inactivated. For example, it is possible that the internal pool of G protein is no longer attenuated by RGS proteins. Alternatively, both mechanisms might contribute. Recent studies in a variety of organisms suggest coordination between the receptors that activate the signal and the RGS proteins that inactivate the signal. RGS proteins show selectivity for particular Gα subunits, but some RGS proteins can physically interact with the receptors (14). One particularly intriguing example of receptor-RGS cooperation is provided by the plant protein AtRGS1, which has both RGS- and receptor-like domains, suggesting that it functions both as a G protein activator and inactivator (15). More recent findings reveal that the yeast RGS protein Sst2 can bind directly to the pheromone receptor Ste2 at theplasmamembrane (16). Presumably Sst2 interaction with its cognate receptor ensures that pathway regulation is both rapid and receptor-specific. Regulation is receptor-specific because of selective binding to Sst2. Regulation is rapid, because receptor binding positions the RGS domain of Sst2 in close proximity to its substrate, Gpa1. In this scenario, RGS positioning close to Gα sets the threshold for pheromone response, preventing signaling at low pheromone concentrations. As signaling at the plasma membrane wanes, endocytosed receptors might continue to signal from the endosome or other internal compartments, as reported for other receptors (10). Phosphorylation of the receptor promotes endocytosis and also results in dissociation of the RGS protein, suggesting that the RGS protein does not follow the receptor as it undergoes endocytosis. Thus, one way that endosomal signaling may differ from plasma membrane signaling is if the RGS protein is no longer in close proximity to the G protein.

Lastly, there may be as yet undiscovered pathway regulators acting at the endosome, perhaps in response to PI3P production. Unlike many other second messengers, however, PI3P cannot diffuse freely and remains bound to membranes at the site of synthesis. Therefore, PI3P might serve to recruit proteins with PI3P-binding domains to the endosome (17). Such protein recruitment may allow assembly of signaling complexes that are distinct from those known to exist at the plasma membrane. In support of this model, Gpa1Q323L promotes translocation of the PI3P-binding protein Bem1 to the endosome (7). Bem1 functions as an adaptor protein that promotes activation of other signaling components, including the small GTPase Cdc42 and the MAP kinase Fus3 (18). The adaptor protein Ste5 serves a similar role in transmitting the Gβγ signal at the plasma membrane (19).

If history is any guide, lessons learned in yeast will prove applicable to signaling events in more complex organisms. The identification of Sst2 and other RGS family members in the mid-1990s led to a dramatic rethinking of how G protein signaling pathways are organized and regulated. This view has now been further refined with the realization that receptors and RGS proteins cooperate to modulate G protein activity. Moreover, the old dogma of G protein–mediated signaling only at the plasma membrane must now be modified to include signaling from internal compartments. These discoveries benefited from the development of powerful genome-scale proteomic and genomic tools available in yeast. Our ability to fully exploit these tools is still evolving, but we can be confident that they will eventually provide us with a truly global or systems-level understanding of how cells respond to changes in their environment.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article