Report

A Seven-Transmembrane RGS Protein That Modulates Plant Cell Proliferation

See allHide authors and affiliations

Science  19 Sep 2003:
Vol. 301, Issue 5640, pp. 1728-1731
DOI: 10.1126/science.1087790

Abstract

G protein–coupled receptors (GPCRs) at the cell surface activate heterotrimeric G proteins by inducing the G protein alpha (Gα) subunit to exchange guanosine diphosphate for guanosine triphosphate. Regulators of G protein signaling (RGS) proteins accelerate the deactivation of Gα subunits to reduce GPCR signaling. Here we identified an RGS protein (AtRGS1) in Arabidopsis that has a predicted structure similar to a GPCR as well as an RGS box with GTPase accelerating activity. Expression of AtRGS1 complemented the pheromone supersensitivity phenotype of a yeast RGS mutant, sst2Δ. Loss of AtRGS1 increased the activity of the Arabidopsis Gα subunit, resulting in increased cell elongation in hypocotyls in darkness and increased cell production in roots grown in light. These findings suggest that AtRGS1 is a critical modulator of plant cell proliferation.

Heterotrimeric G proteins couple multiple signal transduction pathways from seven-transmembrane (7TM) GPCRs to down-stream effectors in mammalian cells (1). RGS proteins accelerate the intrinsic guanosine triphosphatase (GTPase) activity of the Gα subunit, thus returning the heterotrimer to its basal GDP-bound state (2). In contrast to metazoans, the Arabidopsis genome contains only one canonical Gα subunit (AtGPA1), one G protein β (Gβ) subunit (AtAGB1), and two G protein γ (Gγ) subunits (AtAGG1 and AtAGG2), and neither a 7TM receptor nor a cognate ligand has been identified in plants. Nonetheless, plants use heterotrimeric G protein signaling to regulate growth and development (3, 4). The phenotypes of null mutants of the Gα and Gβ subunits indicate that a heterotrimeric G protein controls cell proliferation processes in Arabidopsis (5, 6).

A search of the Arabidopsis genome database for other potential components of heterotrimeric G protein signaling (7) revealed a single open-reading frame (ORF) of 459 amino acids (aa), hereafter called AtRGS1, that encodes an extended N-terminal region and a C-terminal RGS box (Fig. 1). The first 250 amino acids are predicted to form a 7TM domain with a topology reminiscent of GPCRs: an extracellular N-terminus and an intracellular C-terminus (Fig. 1A). All other known RGS proteins lack 7TM domains. The 7TM region of AtRGS1 has weak overall similarity to the metabotropic glutamate GPCR subfamily (“Family C”) that also includes calcium-sensing, odorant, pheromone, and γ-aminobutyric acid type B (GABAB) receptors (8). However, the large, ligand-binding, N-terminal ectodomain that typifies most members of this GPCR subfamily is lacking in the AtRGS1 ORF (Fig. 1A). The predicted topology of AtRGS1 places cysteine residues (Cys84 and Cys153) at the entry to transmembrane domain 3 (TM3) and the second extracellular loop (fig. S1), respectively, in similar positions to a common disulfide linkage found frequently in GPCRs of all subfamilies (9). A database search of Arabidopsis ORFs using the N- and C-terminal domains of AtRGS1 failed to yield homologs; thus, it appears that AtRGS1 represents the single member of this family.

Fig. 1.

Prediction of transmembrane regions, overall topology, and domain architecture of the AtRGS1 protein. (A) The probability of AtRGS1 amino acids being extracellular, transmembrane, or intracellular, as predicted using a transmembrane domain hidden Markov model (15), is plotted below the schematic representation of the AtRGS1 ORF (GenPept accession number NP_189238). (B) Multiple sequence alignment of the RGS box regions of human hRGS19, rat rRGS4, bovine bRGS9, and Arabidopsis AtRGS1 proteins. Conserved amino acids identified by ClustalW (16) are boxed in black. The nine α helices observed within the nuclear magnetic resonance (NMR) solution structure (17) of hRGS19 are numbered with roman numerals and overlined in blue. Closed circles denote conserved residues forming the RGS box hydrophobic core; open circles highlight conserved residues making direct contacts with Gα in the RGS4/Gαi1 crystal structure (10). Predicted α helical (α) and β strand (β) secondary structure within the AtRGS1 RGS box, based on the PSI-Pred algorithm (18), is denoted underneath the AtRGS1 sequence. Primary sequences in the alignment are human hRGS19/GAIP (SwissProt accession number P49795), rat rRGS4 (P49799), bovine bRGS9 (O46469), and AtRGS1 (GenPept NP_189238).

The C-terminal 211 amino acids (aa 249–459) of AtRGS1, which contains the RGS box was expressed and purified from E. coli as a glutathione S-transferase (GST) fusion protein (GST-RGS1box) and was tested for its ability to bind to and accelerate the GTPase activity of AtGPA1. GST-RGS1box associated with recombinant AtGPA1 in vitro (Fig. 2A). This interaction was dependent on the addition of aluminium tetrafluoride (AlF4), a planar ion that stabilizes Gα subunits in a transition state for GTP hydrolysis (10). RGS box proteins that act as GTPase-accelerating proteins (GAPs) for Gα subunits bind most avidly to the GDP-AlF4 form of their Gα subunit targets (10). GST-RGS1box also accelerated the GTPase activity of AtGPA1 in a dose-dependent manner (Fig. 2, B and C). A 14-fold increase in inorganic phosphate production was observed at a threefold excess of GTP-loaded AtGPA1 over GST-RGS1box protein.

Fig. 2.

AtRGS1 is a Gα GTPase-accelerating protein. (A) Guanine nucleotide dependence of the AtRGS1-AtGPA1 interaction. Purified His6-AtGPA1 and GST-RGS1box (aa 249–459) proteins were incubated with GDP, GTPγS, or GDP-AlF4 and then precipitated with glutathione agarose. Complex formation was detected by immunoblotting (IB) with specific antibodies as indicated. (B) GST-RGS1box accelerates the intrinsic GTPase activity of AtGPA1 in vitro. Single-turnover GTPase assays, using GTP-loaded His6-AtGPA1 (575 nM) in the absence or presence of 182 nM GST-RGS1box protein, were initiated with the addition of 5 mM MgCl2 at 0 s. Inorganic phosphate release was measured in real-time using fluorescent phosphate binding protein [λex = 425 nm, λem = 465 nm (excitation and emission wavelengths, respectively)] and is expressed as change in fluorescence units over time (19). Basal level of protein-independent inorganic phosphate production is also denoted (CONTROL). (C) Dose-dependent AtRGS1 GAP activity. Single turnover GTPase assays were performed as in (B) using 500 nM AtGPA1 and varying concentrations of GST-RGS1box protein (0 to 4.6 μM). The apparent initial rate of phosphate production (kobs in s1), calculated from changes in fluorescence, is plotted against GST-RGS1box protein concentration. (D) AtRGS1 expression reverts the pheromone supersensitivity phenotype of Sst2-deficient S. cerevisiae. Wild-type and Sst2-null (sst2Δ) yeast were independently transformed with either a vector containing the C-terminal domain of AtRGS1 (AtRGS1box; aa 249–459), or empty vector (pYES), along with a pheromone-responsive FUS1 promoter-lacZ reporter plasmid. Cells were then treated with indicated concentrations of α factor pheromone, and resultant β-galactosidase activity was measured. Median effective concentration (EC50) values (and 95% confidence intervals) for pheromone response are as follows: wild-type yeast (squares) = 2.3 μM (1.8 to 2.8 μM); sst2Δ plus empty vector (circles) = 0.029 μM (0.024 to 0.035 μM); sst2Δ plus AtRGS1 box (inverted triangles) = 0.37 μM (0.29 to 0.49 μM).

Loss-of-function mutations to Sst2, the archetypal RGS protein of the budding yeast Saccharomyces cerevisiae, render haploid yeast supersensitive to pheromone signaling that is mediated by a canonical GPCR-linked signal transduction pathway (11). The C-terminal domain of AtRGS1 (AtRGS1box; aa 249–459) was tested for its ability to complement the supersensitivity phenotype of the yeast Sst2 deletion mutant, sst2Δ. As observed with some mammalian RGS proteins, expression of AtRGS1 box partially restored the normal dose-response curve for α factor pheromone induction of β-galactosidase reporter gene expression placed under the control of a pheromone pathway-specific promoter (pFUS1-lacZ) (11), suggesting that the C-terminal domain of AtRGS1 exerts GAP activity on the yeast Gα subunit Gpa1. In addition, halo assays of pheromone-induced lethal arrest also revealed that expression of this domain can attenuate the pheromone supersensitivity of haploid sst2Δ yeast (7) (fig. S2).

Interaction between full-length AtRGS1 and AtGPA1 was shown by complementation of split ubiquitin domain fusions in yeast (12). The isolated AtRGS1 C-terminus [AtRGS1(Δ7TM); aa 249–459], as well as the full-length AtRGS1, interacted with both a constitutively active, GTPase-deficient form [Gln222 → Leu222 (Q222L)] of AtGPA1 (AtGPA1QL) and wild-type AtGPA1 (Fig. 3A). However, the N-terminal 7TM domain of AtRGS1 [AtRGS1(ΔRGS); aa 1–248] did not interact with AtGPA1 in this assay. Full-length AtRGS1 fused to a c-myc epitope tag (AtRGS1-myc) was immunoprecipitated from Arabidopsis cell lysates with antiserum to AtGPA1, preferentially in the presence of AlF4 (Fig. 3B). In the reciprocal experiment, AtGPA1 immunoprecipitated with antibody to myc, also preferentially in the presence of AlF4. These results indicate that the full-length AtRGS1 protein interacts with endogenous AtGPA1 preferentially in the transition-state mimetic form and, thus, has the property of an RGS protein in vivo.

Fig. 3.

AtRGS1 interacts with AtGPA1 in vitro and in vivo. (A) The C-terminus of AtRGS1 interacts with wild-type AtGPA1 and AtGPA1QL, a constitutively active mutant form of AtGPA1, in the yeast split-ubiquitin system. Interactions between AtGPA1 (or AtGPA1QL) and full-length AtRGS1 (aa 1–459), N-terminal 7TM (AtRGS1ΔRGS; aa 1–248), and C-terminus (AtRGS1Δ7TM; aa 249–459) were analyzed on the basis of yeast growth in selective media. The MLO1-calmodulin interaction was used as a positive control (20). (B) AtRGS1 immunoprecipitates with AtGPA1. Arabidopsis suspension cells were transformed with 35S::c-Myc epitope-tagged AtRGS1 binary vector. Total protein extracts were immunoprecipitated (IP) using antibody to c-myc (αMyc) or AtGPA1 (αGPA1) and then immunoblotted (IB) with the indicated antibody. (C) Both AtRGS1 and AtGPA1 localize at the plasma membrane. 35S::AtRGS1-GFP and 35S::AtGPA1-GFP binary constructs were transformed separately into Arabidopsis suspension cells. GFP was visualized by fluorescence microscopy in intact cells and protoplasts. (D) AtRGS1 is localized to the plasma membrane of cortical cells in the differentiated zone of Arabidopsis roots. Asterisks indicate the positions of apical and basal membranes. The picture was taken from the differentiation zone of a root of a 5-day-old Arabidopsis seedling transformed with 35S::AtRGS1-GFP. (E) AtRGS1-GFP and (F) AtGPA1-CFP accumulate at the nascent cell plate in dividing Arabidopsis cells. Arrows indicate the nascent cell plate. Cells were taken from a population of suspension cells transformed with 35S::AtRGS1-GFP or 35S::AtGPA1-CFP binary vector 4 days after subculture.

Expression of AtRGS1 and AtGPA1 as fusions with green fluorescent protein (AtRGS1-GFP and AtGPA1-GFP) showed localization of both proteins to the plasma membranes of postmitotic cells in culture (Fig. 3C). Plasma membrane localization of AtRGS1-GFP was maintained in fully differentiated, postmitotic cells in the intact root (Fig. 3D). In dividing Arabidopsis cells, both proteins accumulated at the nascent cell plate (Fig. 3, E and F), suggesting a role in cytokinesis.

Null alleles of AtGPA1 have reduced proliferation of some cell types throughout development (5, 6). Therefore, we expected that if AtRGS1 is a negative regulator of AtGPA1, null mutants of AtRGS1 (Atrgs1-1, Atrgs1-2) would exhibit increased proliferation of some cell types, and AtRGS1 should be expressed in some or all sites of stem cell proliferation. An AtRGS1::GUS transcriptional fusion transgene was predominantly expressed in shoot and root apical meristems where AtGPA1 is also expressed (13) (Fig. 4E). Atrgs1 null mutant seedlings (Fig. 4, A and B) had longer hypocotyls in the dark as a result of increased cell elongation (Fig. 4C; fig. S3). In contrast, null alleles of AtGPA1 (Atgpa1-1-4) (5, 14) caused shorter hypocotyls, although this phenotype is due to fewer cells (5). Expression of the constitutively active AtGPA1QL increased etiolated hypocotyl length due to increased cell elongation, similar to the Atrgs1-null mutants (Fig. 4, C and G).

Fig. 4.

AtRGS1 modulates cell proliferation in Arabidopsis. (A) Transferred DNA (T-DNA) insertion sites in AtRGS1. LB, T-DNA left border; RB, T-DNA right border. Gray boxes represent exons. The T-DNA insert is not drawn to scale. The gray arrows at LB indicate the T-DNA left border primer, and the black arrows indicate the AtRGS1 specific primers used for mutant isolation. (B) Reverse transcription polymerase chain reaction (RT-PCR) analysis for AtRGS1 transcript. The AtRGS1 transcript was present in total RNA from wild-type Arabidopsis but absent in the Atrgs1-1 and Atrgs1-2 mutants. As a control, actin primers that amplify a 901 base pair product were added together with AtRGS1 primers in each PCR reaction. (C) AtRGS1 null allele phenotypes of 2-day-old, dark-grown seedlings. Atrgs1-1, Atrgs1-2, and null alleles of AtGPA1 (gpa1-3 and gpa1-4) are in Columbia (Col) ecotype background. Null alleles of AtGPA1, gpa1-1, and gpa1-2, and the transgenic lines overexpressing a constitutively active form of AtGPA1 [AtGPA1QL (D) and (E)] are in the Wassilewskija (WS) ecotype background. The null alleles of AtRGS1 had the same number of epidermal cells in the hypocotyl as wild-type seedlings (fig. S3). (D) AtRGS1 null allele phenotypes of 3-day-old, light-grown seedlings. (E) AtRGS1::GUS expression in light-grown Arabidopsis seedlings. The regions of shoot and root meristems are indicated with arrows. (F) Independent transgenic lines (designated ROX lines S2, S3, and S9) overexpressing AtRGS1 driven by a dexamethasone (Dex)-inducible promoter produce a similar phenotype as the loss of AtGPA1. Shown here are 2-day-old, dark-grown seedlings. Plants transformed with empty pTA7002 vector were controls. Dex was applied at 0.5 μM. (G) Null mutants of AtRGS1 have increased cell elongation in hypocotyls grown in darkness and increased cell production in roots grown in light. The hypocotyl lengths were taken from 2-day-old, dark-grown seedlings. The cell production in roots was measured using 5-day-old, light-grown seedlings and calculated as the rate of root growth divided by the average cortex cell length.

In light-grown seedlings, both null mutants of AtRGS1 and lines overexpressing AtGPA1QL produced longer primary roots compared with wild-type Arabidopsis or null mutants of AtGPA1 (Fig. 4D). This increased root growth phenotype resulted from increased cell production in root meristems (Fig. 4G). These data suggest that increased activity of AtGPA1, either by expression of constitutively active AtGPA1QL or through loss of AtRGS1 expression, results in increased cell proliferation in the apical root meristem. Inducible overexpression of AtRGS1 (7) (fig. S4) produced a similar phenotype as loss of AtGPA1 (Fig. 4F), further suggesting that AtRGS1 antagonizes the activation of AtGPA1. In addition, some loss-of-function gpa1 phenotypes, such as paclobutrazol and sugar sensitivity, are the opposite in the Atrgs1 mutants, indicating a role for activated GPA1 in other signaling pathways throughout development (figs. S5 and S6).

Heterologous expression of full-length AtRGS1 protein in Sf9 insect cells has not yet provided adequate expression levels for a biochemical test of GAP activity in vitro. Nevertheless, the evidence that null mutants of AtRGS1 phenocopy the constitutively active mutant form of AtGPA1 (AtGPA1QL), that overexpression of AtRGS1 antagonizes the activation of AtGPA1 (Fig. 4), and that full-length AtRGS1 interacts with AtGPA1 in an AlF4 -dependent manner (Fig. 3) suggests that AtRGS1 exerts GAP activity on AtGPA1 in vivo. Our results here support earlier findings that cell proliferation in plants is regulated by heterotrimeric G protein subunits and further extend those findings by showing that this regulation is cell type specific. It also reveals that cell proliferation control by the Arabidopsis G protein mechanistically involves either the unsequestered Gβγ subunit or the activated Gα subunit as the predominant regulatory element, depending on cell type.

Supporting Online Material

www.sciencemag.org/cgi/content/full/301/5640/1728/DC1

Materials and Methods

Figs. S1 to S6

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article