Special Viewpoints

G Proteins Go Green: A Plant G Protein Signaling FAQ Sheet

See allHide authors and affiliations

Science  07 Oct 2005:
Vol. 310, Issue 5745, pp. 71-73
DOI: 10.1126/science.1118580


Plants, like animals, use signal transduction pathways based on heterotrimeric guanine nucleotide–binding proteins (G proteins) to regulate many aspects of development and cell signaling. Some components of G protein signaling are highly conserved between plants and animals and some are not. This Viewpoint compares key aspects of G protein signal transduction in plants and animals and describes the current knowledge of this system in plants, the questions that still await exploration, and the value of research on plant G proteins to scientists who do not study plants. Pathways in Science's Signal Transduction Knowledge Environment Connections Maps database provide details about the emerging roles of G proteins in several cellular processes of plants.

The turn of the 21st century, accompanied by completion of the sequencing of the genome of the model plant species, Arabidopsis thaliana, and the identification of G protein mutants in Arabidopsis and rice (Oryza sativa), has marked the start of a minirevolution in our understanding of G protein signaling in plants. This Viewpoint highlights the components, mechanisms, and functions of G protein signaling in plants as compared with metazoans (14).

Heterotrimeric G proteins are secondary messengers composed of three dissimilar subunits: Gα, Gβ, and Gγ. Gα and the nondissociable Gβγ dimer link signals emanating from altered activation states of G protein–coupled receptors (GPCRs) that contain seven transmembrane (7TM) domains to multitudinous intracellular effectors (Fig. 1). In humans, there are 20-some Gα subunits, five Gβ subunits, and about a dozen Gγ subunits. G proteins mediate vision, olfaction, and some aspects of gustation, and they are intimately involved in numerous neuroendocrine signaling pathways (5, 6). Thus, it is not surprising that G protein–centered signaling cascades are targets of an estimated one-third to one-half of the pharmaceuticals currently marketed (7, 8).

Fig. 1.

The G protein cycle. Components and processes found in both plants and animals are indicated in green, with darker green indicating more definitive evidence (in plants) than lighter green. Those components and processes to date reported only for plants are in purple and those reported only for animals are indicated in red. GEF (guanine nucleotide exchange factor) and GDI (guanine nucleotide dissociation inhibitor) are two Gα binding proteins so far only identified in animals. In addition to desensitization and internalization, arrestins also act as adapter proteins in the regulation of intracellular signaling. (inset) The round-leaf phenotype exhibited by Gα (gpa1) knockout plants of Arabidopsis (right), in this case, a plant with the mutant allele gpa1-4, in comparison with a wild-type plant of the Columbia-0 ecotype (Col-0) (left).

1) Do plants have heterotrimeric G proteins and are they essential? The genomes of diploid plant species encode single canonical G protein α subunit and G protein β subunit proteins: GPA1 and AGB1 in Arabidopsis and RGA1 and RGB1 in rice (1, 2). Two Arabidopsis proteins, AGG1 and AGG2, are likely Gγ subunits on the basis of their ability to interact with plant Gβ subunits, stretches of sequence similarity with known Gγ proteins including a predicted C-terminal prenylation site, and structural modeling (911). The corresponding rice homologs are RGG1 and RGG2 (12).

Knockout mutants of GPA1, AGB1, and the likely plant GPCR, GCR1, as well as their double- and triple-mutant combinations, are viable and healthy, although not phenotypically identical to wild-type plants, under standard laboratory conditions (Fig. 1, inset) (4). However, the extent to which such mutations would confer reduced fitness in the natural environment awaits assessment (13).

2) Do plants have any noncanonical G protein subunits? This viability of the G protein knockouts leads one to ask whether there are noncanonical G protein subunits in plants. Indeed, identified in plant genomes are the “extra large G protein” genes (XLGs) (2). The three XLG proteins of Arabidopsis and the four rice XLGs are twice as large as the typical metazoan Gα protein and are typified by the presence of a carboxy half, with ∼50% similarity to the plant Gα subunit GPA1, and a unique amino half, with a cysteine-rich region and a nuclear localization site. Arabidopsis XLG1 has been verified biochemically as a guanosine 5′-triphosphate (GTP)–binding protein. The plant XLG proteins have no sequence similarity outside of the Gα domain with mammalian XLα proteins, which are Gαs splice variants (14) and, thus, are unlikely to be evolutionarily or functionally related.

A group of plant proteins, the RACK1s or Arcs, with three family members in Arabidopsis, has a predicted three-dimensional structure similar to that of Gβs (15). However, to date, there is no evidence that the XLGs or the RACKs interact with the canonical G protein subunits (that is XLGs with Gβγ and RACKs with Gα or Gγ), so their role as nonconventional subunits remains hypothetical.

3) Do plant genomes encode G protein regulatory proteins commonly found in metazoans? In known plant genomes, there are no obvious homologs of genes encoding arrestins or GRKs (G protein–coupled receptor kinase) (3). Plant genomes are rich in kinase-encoding genes (16), and another class of kinases may act as GRKs in plants. The Arabidopsis genome contains one RGS (regulator of G protein signaling) gene, RGS1, which encodes a protein with a predicted 7TM domain structure, followed by a domain containing an RGS box. On the basis of its unique structure, it is intriguing to speculate that RGS1 may have dual function as a receptor and a guanosine triphosphatase (GTPase)–activating (GAP) protein (17).

4) Do plants have GPCRs, and do they share ligands with mammalian GPCRs? GCR1 and RGS1 are good candidates for plant GPCRs because they physically interact with GPA1 in planta, and because their genetic elimination affects plant processes, such as hormone sensitivity and cell division, known to be regulated by GPA1 (17, 18). The Arabidopsis protein GCR1 has limited similarity within its predicted 7TM region to the slime mold cyclic AMP receptor, CAR1 (18, 19). Because the 800 to 1000 metazoan GPCRs (20) are typified not by sequence conservation but rather by their 7TM structure, the absence in plant genomes of additional genes with sequence similarity to metazoan GPCRs does not necessarily lead to the conclusion that plants are deficient in these receptors. Indeed, plants have many predicted 7TM proteins (21). For one plant 7TM family, the MLO proteins, the 7TM structure of a family member from barley has been experimentally confirmed (22); whether any of the MLOs couple with Gα awaits evaluation (23). Another question is whether dimerization, which is recognized for maturation and signaling of some metazoan GPCRs (24), is also important for the activity of plant 7TM proteins.

No ligands of GCR1 or RGS1 have yet been identified by ligand-binding experiments. However, gpa1 knockout mutants have reduced sensitivity to sphingosine-1-phosphate (S1P), the ligand of the five mammalian S1P [previously referred to as the protein product of the endothelial differentiation gene (EDG)] GPCRs (25), as well as to the related metabolite, phytosphingosine-1-phosphate (26, 27); rice mutants in the Gα subunit gene, RGA1, also show decreased responsiveness to sphingolipid compounds (28, 29). In addition, rgs1 mutants exhibit altered sensitivity to glucose (17), and sugars are ligands for both yeast and mammalian GPCRs (30). Thus, there is evidence to suggest that some molecules act as GPCR ligands in both animals and plants.

5) Do plants exhibit signaling by GPCRs through G protein–independent pathways? In Arabidopsis, knockout mutation of GCR1 results in reduced sensitivity of seed germination to the germination-promotive hormones gibberellin (GA) and brassinolide. However, the double mutants gcr1 gpa1 and gcr1 agb1 have even less sensitivity to these hormones than the single gcr1 mutants. Epistasis analysis predicts that, if GCR1 signals only through the G protein, then these single and double mutants should exhibit identical phenotypes, but they do not. Thus, GCR1 may have a signaling role independent of the G protein (19), as has also been observed for some metazoan GPCR-mediated responses and has been termed “signaling at zero G” (31).

6) Do plant G proteins exhibit a G protein cycle of guanine nucleotide exchange and GTP hydrolysis? In the classical paradigm, agonist activation of G protein signaling is initiated when GTP binds with Gα and signaling is terminated by the intrinsic GTPase activity of Gα. Recombinant plant Gα subunits also exhibit GTPase activity, although rates are lower than for typical mammalian Gα subunits (2, 32, 33). Moreover, GTPase activity of recombinant GPA1 is accelerated when assayed in the presence of the C4 domain of Arabidopsis RGS1, which includes the RGS box (17). Localization of the rice heterotrimer in the plasma membrane has been verified, and the α, β, and γ subunits cofractionate in gel filtration experiments (12). This cofractionation is disrupted when membranes are treated with the G protein activator, guanosine 5′-O-(3′-thiotriphosphate) (GTP-γ-S), or when a constitutively active version of the Gα protein is expressed (12). In addition, gpa1 and agb1 knockout plants share a subset of phenotypes (4). Thus, the classical mode of G protein signaling is likely to operate in plants, although the possibility also exists for unique modes of action, as suggested by the novel bipartite structure of the Arabidopsis RGS1 protein.

We are learning that “inactive,” guanosine diphosphate (GDP)–bound Gα subunits also can play an active role, e.g., in the regulation of mammalian cell division through interaction of GDP-bound Gα with proteins that bind the GDP-bound but not the GTP-bound form of the subunit (34). There is no experimental evidence addressing this possibility in plants, but it is tantalizing that Gα and Gβ knockout mutants both exhibit alterations in cell division (11, 35).

7) Do plants and animals share similar G protein effector proteins? Several plant effector proteins have been uncovered, and more are likely to be identified, especially because, to date, only two direct interaction partners have been identified for GPA1 and none for RGA1 or the Gβγ dimers. In in vitro pull-down and yeast two-hybrid assays, Arabidopsis Pirin1, which is a postulated transcriptional cofactor, and phospholipase D α1 directly bind to GPA1 (36, 37). Increased phospholipase C (PLC) activity in tobacco cell lines overexpressing GCR1 or GPA1 also implicates this class of lipases as plant G protein targets (38). In addition, the TUBBY protein (39), which in mammalian cells exhibits PLC-β and Gαq/11-dependent translocation from the plasma membrane to the nucleus, has plant homologs (40). Through phenotypic analysis of gpa1 knockout plants, inwardly rectifying K+ channels and anion channels have been identified as targets of G protein signaling in Arabidopsis (26, 41), and application of recombinant tomato Gα protein to excised plasma membrane patches increases the open probability of Ca2+-permeable channels (32). Thus, there is some commonality between plant and animal G protein effectors.

8) Do G proteins regulate similar cellular and developmental processes in plants and animals? Cell division, ion channel regulation, and disease response are processes regulated by G proteins in both plants and animals. Arabidopsis gpa1 mutants exhibit decreased cell division in the hypocotyl (the seedling stem) and in leaves, and exhibit a rounded leaf shape (35). agb1 mutants exhibit proliferation of lateral roots, a phenomenon linked to differential sensitivity to the phytohormone auxin (11).

An example of ion channel regulation by G proteins in plants can be seen in guard cells, in which the responses to another plant hormone, abscisic acid (ABA), are altered in gpa1 mutants. Guard cells are specialized pairs of cells in the plant epidermis in which osmotically driven cell swelling or shrinkage regulates the width of the stomatal pore found between each guard cell pair, thereby controlling both carbon dioxide uptake and water vapor loss through the pore. ABA inhibits inwardly rectifying K+ channels and activates anion channels in wild-type guard cells, but channels of gpa1 guard cells show decreased ABA responsiveness (26, 41).

G proteins have been implicated in many aspects of human disease (68). In plants, rice Gα mutants exhibit reduced disease resistance responses to the highly destructive rice blast fungus and to fungal sphingolipid elicitors (28, 29), as described in the Connections Map (42).

9) What is the evidence for species-specificity in plant G protein function? Arabidopsis gpa1 knockout plants are within the normal size range, and the most obvious phenotype of the mature vegetative plant is that the rosette leaves are rounded, rather than lanceolate (Fig. 1, inset). By contrast, rice plants harboring a nonfunctional RGA1 gene are dwarfs (3, 4). The dwarfing in rice has been traced, at least in part, to reduced sensitivity of the Gα mutant stems to the plant hormone GA, which promotes cell elongation (43). With regard to seed germination, pharmacological and biochemical studies have shown that the Gα subunit of barley is involved in transducing the effects of ABA, whereas Arabidopsis gpa1 mutant seeds exhibit hypersensitivity to this hormone, rather than the insensitivity predicted by the results from barley (36, 44, 45).

However, not enough is known yet about the complete signaling pathways to conclude that these phenotypic differences necessarily result from fundamentally different G protein signaling mechanisms in the different species. It is also important to note that the Gα mutants of rice and Arabidopsis also show some phenotypic similarities, such as reduced sensitivity of seed metabolism to GA (19, 43, 44). The Connections Maps for rice and Arabidopsis G protein signaling in seed germination should prove useful tools for exploring these comparisons (46, 47).

10) Do plants have some pathways that signal through Gα and other pathways that signal through Gβγ? Some phenotypes, such as rounded rosette leaves, are exhibited similarly by both gpa1 and agb1 knockout mutants (35, 48). Plausibly, these phenotypes are mediated by Gα, and loss of Gβ confers the same phenotype because, in the absence of Gβ, Gα cannot couple appropriately with its GPCRs, effectors, or both.

For pathways signaling through Gβγ, one might predict that knockout of Gα, which may increase the availability of free Gβγ to interact with its effectors, would lead to the opposite phenotype than knockout of Gβ, which would eliminate Gβγ-effector coupling. Lateral root production appears to be one process that is regulated in this manner: Compared with wild type, agb1 mutants exhibit increased numbers of lateral roots while gpa1 mutants show decreased lateral root production. These results suggest that AGB1 acts as a negative regulator of lateral root production (11). agb1 mutants also exhibit more damage following acute O3 treatment, whereas gpa1 mutants exhibit more resistance than wild type (49). Thus, genetic evidence indicates that plants, like animals, have some pathways that signal through Gα and other pathways that signal through Gβγ.

11) What are the advantages to researchers studying nonplant systems of research on plant G protein signaling? Plants afford several advantages to the study of heterotrimeric G proteins over those offered by other systems. Because of the plethora of G protein subunits in animals, it is not possible to create knockout mammals in which the effects of complete abrogation of G protein signaling can be assessed. By contrast, because plants have few G protein subunits, it is feasible to create knockout plants in which G protein signaling has been genetically eliminated. Plants and mammals share multicellularity, and some human disease-related genes are actually more similar to Arabidopsis genes than to counterpart genes in other model systems, such as yeast and C. elegans (16). Because plants and mammals also share some G protein effectors, changes in cellular processes detected in G protein knockout plants may also inform our knowledge of mammalian G protein signaling. A second advantage of such knockout plants is that they can be used as heterologous expression systems for the study of mammalian G protein signaling, in the absence of the “native” G protein subunits that complicate data interpretation when mammalian subunits are heterologously expressed in mammalian cell lines. Moreover, the cost and regulatory burden of maintaining transgenic plants as seeds is minimal.

Finally, it is worth pointing out that some exogenous ligands of human GPCRs, including key analgesics, are plant metabolites. One can anticipate that with the growth of the field of plant metabolomics, more plant-based GPCR ligands of pharmaceutical interest will be identified.

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article