An elaborate heterotrimeric G-protein family from soybean expands the diversity of plant G-protein networks


Author for correspondence:
Sona Pandey
Tel.: +1 314 587 1471


  • The repertoire of heterotrimeric G-proteins in plant species analyzed thus far is simple, with the presence of only two possible canonical heterotrimers in Arabidopsis and rice vs hundreds in animal systems. We assessed whether genome duplication events have resulted in the multiplicity of G-protein in plant species like soybean that would increase the complexity of G-protein networks.
  • We identified and amplified four Gα, four Gβ and two Gγ proteins, analyzed their expression profile by quantitative PCR during different developmental stages. We purified the four Gα proteins and analyzed their guanosine-5′-triphosphate (GTP)-binding and GTPase activity. We performed yeast-based interaction analysis to assess the interaction specificity of different G-protein subunits.
  • Our results show that all 10 G-protein genes are retained in the soybean genome and ubiquitously expressed. The four Gα proteins seem to be plasma membrane-localized. The G-protein genes have interesting expression profiles during seed development and germination. The four Gα proteins form two distinct groups based on their GTPase activity. Yeast-based interaction analyses predict that the proteins interact in most of the possible combinations, with some degree of interaction specificity between duplicated gene pairs.
  • This research identifies the most elaborate heterotrimeric G-protein network known to date in the plant kingdom.


Heterotrimeric G-proteins (G-proteins hereinafter) consisting of α, β and γ subunits are important signal transducers in all eukaryotes (Jones & Assmann, 2004). The Gα subunit of the heterotrimer undergoes a signal-dependent transition between guanosine diphosphate (GDP)- and guanosine-5′-triphosphate (GTP)-bound forms to modulate signal transduction. In the inactive GDP-bound state, Gα interacts with Gβγ and is associated with a cell surface G-protein-coupled receptor (GPCR). Upon signal perception, a change in GPCR conformation facilitates exchange of GTP for GDP on Gα, which results in dissociation of the heterotrimer. The active Gα˙GTP and the released Gβγ interact with different effectors to propagate the signal. Intrinsic GTPase-activity returns Gα to its inactive form (Cabrera-Vera et al., 2003; Offermanns, 2003). These fundamental aspects of G-protein signaling are conserved across phyla (Jones & Assmann, 2004).

In metazoans, G-protein-coupled signaling mediates vision, olfaction, taste, neurotransmission, and hormone responses (Cabrera-Vera et al., 2003; Offermanns, 2003). In these organisms, the multiplicity of G-proteins and GPCRs, in addition to downstream signaling components, provides for complex signal transduction architecture. For example, humans have 23 Gα (including splice variants), five Gβ, 12 Gγ, > 800 GPCRs, and scores of accessory proteins that form an elaborate network of specific GPCR–Gαβγ combinations to transduce a specific signal. Studies in Arabidopsis and rice provide knowledge of plant G-protein signaling and suggest much simpler architecture. The fully sequenced genomes of both these species contain only one conventional Gα, one Gβ and two Gγ subunits and a few GPCRs each (Perfus-Barbeoch et al., 2004), although additional splice variants, or nonconventional genes especially for Gγ subunits may also exist. These limited number of components, however, regulate diverse signaling pathways, including hormone signaling and cross-talk, environment sensing, cell division, ion channel regulation, disease resistance and cell death (Temple & Jones, 2007; Chen, 2008).

Homologs of G-proteins are found in other plant species (Supporting Information, Table S1), but the unavailability of genetic and genomic resources limits insight into their biological function. Moreover, in some cases, the phenotypes observed in Arabidopsis G-protein mutants are not recapitulated in other plant species. For example, in both Arabidopsis and rice, mutation in Gα results in hyposensitivity to gibberellic acid (Ueguchi-Tanaka et al., 2000; Ullah et al., 2001, 2002), but only in rice does this cause plant dwarfism (Ashikari et al., 1999; Fujisawa et al., 1999). Similarly, tobacco plants with suppressed levels of Gβ show defects in anther shape and pollen development, which results in nongerminating pollen grains (Peskan-Berghöfer et al., 2005). These phenotypes are not shared with the corresponding Arabidopsis agb1 mutants (Ullah et al., 2003). Importantly, such examples question whether the relatively simple architecture of the Arabidopsis G-protein signaling pathway is the norm in the plant kingdom, considering that an estimated 30–80% of plant species, including major crops like soybean, have undergone one or more rounds of genome duplication (Wendel, 2000), raising the possibility of expansion of G-protein gene families in such species.

Expanding on studies of model monocots and dicots, growing evidence suggests that G-protein signaling plays a critical role in a range of economically important legume crops during normal growth and development as well as during symbiosis. Some examples include the G-protein-regulated blue and red light signaling pathways in pea (Warpeha et al., 1991), induction of root hair deformation in Vigna, and expression of nodulation-specific genes in Medicago by mastoparan, a known G-protein agonist (Kelly & Irving, 2003; Sun et al., 2007). Additionally, some well-established effectors of G-protein signaling, for example phospholipase C (PLC) and D (PLD) are involved in mastoparan-induced root hair deformation and development of root nodules (Pingret et al., 1998; den Hartog et al., 2001; Charron et al., 2004; Santos-Briones et al., 2009). G-proteins and PLC are also involved in an elicitor-induced oxidative burst in soybean cells (Legendre et al., 1992). Lastly, there is one report of molecular characterization of G-proteins from legumes, where one Gα, one Gβ, and two Gγ cDNAs were isolated from pea and studied for their role in abiotic stress signaling (Misra et al., 2007).

We have designed the present study to identify and characterize soybean heterotrimeric G-proteins. Sequence homology-based screenings and Southern hybridizations identified multiple homologs of G-protein subunits in some plant species, including soybean (Kim et al., 1995; Gotor et al., 1996), but a systematic study to identify and comparatively analyze multiple G-protein subunits with respect to the plant polyploidy has not been performed to date. Soybean was chosen because of the presence of an allopolyploid genome along with the recent availability of suitable genetic and molecular resources (Schmutz et al., 2010) and a lack of any comprehensive data from a legume on the important roles played by G-proteins in regulating fundamental growth and development processes. This study describes multiple G-protein subunits from soybean with distinct biochemical activity, details their expression profiles, and depicts the specificity of interaction between different subunits. This work reveals the most elaborate network of G-proteins in a plant known to date.

Materials and Methods

Plant material and growth conditions

Soybean (Glycine max L.) cv Jack seeds were grown in growth chamber (26 : 20°C, day : night temperature, photoperiod of 14 : 10 h, 800 μmol m−2 s−1 light intensity, and 60% humidity). Different developmental stages of soybean plants were collected, immediately frozen in liquid nitrogen and stored at −80°C. For germination assays, dry seeds were surface-sterilized with 1% bleach followed by extensive washing with distilled water. The seeds were then placed on pre-wet filter papers in Petri dishes under 16 : 8 h light : dark regime and collected every 6 h for 30 h. By this time, an obvious radicle had emerged from the seeds. For stress treatments, 7-d-old soybean seedlings were treated with ABA (100 μM, 6 h), mannitol (300 mM, 6 h), salt (200 mM NaCl, 6 h), cold (4°C, 6 h), and heat (42°C, 6 h). Roots were harvested after 6 h and used for gene expression analysis using real-time quantitative PCR (qRT-PCR).

Cloning of soybean G-protein genes

Soybean G-protein genes were identified by BLAST analysis of the latest soybean genome assembly ( with Arabidopsis and rice G-protein sequences as queries. Full-length G-protein genes were amplified from soybean seedling cDNA using gene-specific primers (Table S2). All 10 G-protein genes were cloned into the pENTR/D-TOPO vector and confirmed by sequencing.

RNA isolation and qRT-PCR

Total RNA was isolated from different tissues of soybean plants using Trizol reagent (Invitrogen). One microgram of total RNA was used for cDNA synthesis using Superscript III cDNA synthesis kit (Invitrogen). All cDNA samples were diluted 1 : 50 in sterile water. qRT-PCR assays were performed as described before (Pandey & Assmann, 2004) with gene-specific primers (Table S2) and SYBR green master mix (Quantas BioSciences, Gaithersburg, MD, USA). Soybean Actin gene was used as an internal control for gene expression studies (Subramanian et al., 2004). To assess the effect of different stresses, a soybean homolog of the Arabidopsis RAB18 gene was used as a positive control (Welin et al., 1994). The real-time PCR amplification was repeated three times and data were averaged. Sequencing and melt curve analysis of amplicons confirmed their specificity.

Extraction of soybean proteins and western blot analysis

Soybean tissues (leaves, roots or seeds) were ground to a fine powder in liquid nitrogen and extracted with three volumes of extraction buffer (50 mM Tris, pH 7.5, 10 mM EDTA, 10% glycerol, 1 mM PMSF (phenylmethylsulfonyl fluoride) and 1 × plant protease inhibitor cocktail; Sigma). Lysate was cleared by centrifugation at 12 000 g for 20 min at 4°C. Cleared lysate was centrifuged at 100 000 g at 4°C for 90 min to separate the microsomal (pellet) and soluble fractions. Protein concentration was measured using the Bradford assay. Western blotting of proteins was performed according to Pandey & Assmann (2004).

Localization of GmGα proteins

GmGα1-4 and AtGPA1 were cloned into the pEarleyGate 101 (Earley et al., 2006) destination vector using LR clonase mix (Invitrogen). Sequence-confirmed recombinant plasmid containing the GmGα1-4::YFP and AtGPA1::YFP were transformed into Agrobacterium tumefaciens strain GV3101 for subsequent plant transformation. AtGPA1::CFP was used as a positive control (Chen et al., 2003).

Tobacco leaves were infiltrated with a log-phase culture of A. tumefaciens containing either the gene of interest or an empty vector control according to Voinnet et al. (2003). Infiltrated plants were incubated in darkness for 36 h followed by 24 h in light. The leaves were observed under a Nikon Eclipse E800 (Nikon Corp. Tokyo, Japan) microscope with epi-fluoroscent modules for cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) fluorescence detection. At least two independent infiltrations were performed for each construct.

Recombinant protein purification and G-protein activity assay

GmGα1-4 and AtGPA1 were cloned into the pET-28a vector (Novagen, Gibbstown, NJ, USA) and recombinant proteins were purified using Ni2+-affinity chromatography (Jez & Cahoon, 2004). Protein aliquots were snap-frozen in liquid nitrogen and stored at −80°C.

Real-time fluorescence-based GTP-binding and GTP-hydrolysis assays were performed using 4,4-difluoro-4-bora-3α,4α-diaza-s-indacene-GTP Fluorophore (BODIPY-GTP FL, Invitrogen) as described previously (Pandey et al., 2009). Assays were performed at 25°C in a 200 μl reaction volume in assay buffer (20 mM Tris, pH 8.0 and 10 mM MgCl2). For competition with nonlabeled nucleotides, 5 μM of GTP, GDP, ATP or ADP was added to the assay buffer before starting the reaction. The reaction was started by addition of labeled nucleotide. Fluorescence (λex485, λem530 nm) was recorded every 3 s for up to 10 min using a fluorescence microplate reader (FLUOstar Optima, BMG Lab Technologies, Cary, NC, USA).

Protein–protein interaction assays

The interaction assay between GmGα and GmGβ was performed using the mating-based yeast split ubiqutin system (Obrdlik et al., 2004; Pandey & Assmann, 2004). Briefly, full-length GmGβ1-4 were cloned into vector containing the C-terminal half of ubiquitin (Cub; Leu+). Next, GmGα1-4 were cloned into vectors containing the N-terminal half of ubiquitin (both wild-type (Nubwt) and a low-affinity NubG version; Trp+). Combinations of GmGβ and GmGα vectors were transformed into yeast haploid strains THY.AP4 (MATa) and THY.AP5 (MATα), respectively. Interaction was determined by growth of diploid yeast colonies on minimal media lacking leucine, tryptophan, histidine and adenine, but containing 1 mM methionine.

To determine the interaction between GmGβ and GmGγ, GATEWAY-based yeast-two-hybrid assay was performed (ProQuest Two Hybrid System, Invitrogen). Briefly, GmGβ1-4 were cloned into pDEST32 bait vector (containing DNA-binding domain) and GmGγ1-2 were cloned into pDEST22 prey vector (containing DNA-activating domain). Both vectors were co-transformed into yeast host strain MaV203 (Invitrogen) as per the manufacturer’s instructions. The interaction of proteins was quantified by β-galatosidase (β-gal) expression assay using o-nitrophenyl-β-d-galactopyranoside (ONPG) as a substrate.


Identification of genes encoding heterotrimeric G-protein subunits from soybean

To identify the complete repertoire of canonical G-protein subunits in soybean, we queried the database using Arabidopsis Gα, Gβ and Gγ sequences. BLAST (Altschul et al., 1990) analysis with AtGPA1 identified four highly homologous protein sequences in soybean database. Two loci (Glyma04g05960.1 and Glyma17g34450.1) corresponded to the previously reported soybean Gα cDNA sequences SGA1 and SGA2 (Kim et al., 1995; Gotor et al., 1996; here called GmGα1 and GmGα2, respectively). The other two loci (Glyma14g11140.1 and Glyma06g05960.1) were called GmGα3 and GmGα4, respectively. The GmGα proteins are c. 90% identical between themselves and share c. 85 and 75% identity with AtGPA1 and OsRGA1, respectively (Fig. 1, Table 1). Signature motifs for GTP binding, GTP hydrolysis, ADP ribosylation and an invariant glutamine residue previously used to generate constitutively active versions of AtGPA1 and OsRGA1 (Okamoto et al., 2001; Oki et al., 2005) are conserved in each GmGα (Fig. 1). Predicted signature sequences for palmitoylation (MGXXXS) and myristoylation (MGXXCS) are present at the N-terminus of each GmGα (Fig. 1). The C-termini of the GmGα are similar to AtGPA1, but distinct from OsRGA1. The N-termini of GmGα2 and GmGα3 contain an eight-amino-acid extension, which predicts two possible translational start sites similar to what is suggested for the genes of rice, tomato, tobacco, and wheat (Kaydamov et al., 2000).

Figure 1.

 Amino acid sequence alignment of GmGα proteins. The sequence alignment of Arabidopsis GPA1 (AtGPA1), rice RGA1 (OsRGA1), and the four GmGα proteins was performed using Clustal W ( Consensus sequences for guanosine-5′-triphosphate (GTP)-binding and hydrolysis are labeled with G1–G5 (Bourne et al., 1991). P/M, the predicted site for N-terminal palmitoylation/myristoylation (MGXXCS); closed triangle, the conserved ADP ribosylation site in each of the Gα proteins; *, the conserved glutamine important for the GTPase activity of Gα proteins. The amino acid positions are numbered in accordance with the AtGPA1.

Table 1.   Amino acid sequence identity (%) of soybean Gα, Gβ and Gγ proteins with corresponding proteins from Arabidopsis and rice
OsRGA1 ***76.676.675.976.6
GmGα1  ***89.989.494.8
GmGα2   ***97.789.9
GmGα3    ***89.4
GmGα4     ***
OsRGB1 ***77.577.577.578.3
GmGβ1  ***98.789.990.5
GmGβ2   ***89.289.7
GmGβ3    ***95.5
GmGβ4     ***
AtAGG2 ***57.436.651.550.5
OsRGG1  ***
OsRGG2   ***31.830.3
GmGγ1    ***92.7
GmGγ2     ***

The sequence for the GmGα4 gene is mis-annotated in the current version of the soybean genome (Schmutz et al., 2010; The reported sequence predicts an additional 71 amino acids at the N-terminus and a C-terminal deletion that removes the entire G5 domain (Fig. S1). This predicted cDNA could never be amplified in our experiments, even when using multiple possible translation start sites, different primer combinations, and varied tissue types. Our experimental data predict that GmGα4 has similar translation start and stop sites to GmGα1. The translation start codon of the experimentally amplified GmGα4 is 213 bp downstream of the predicted sequence. Three different exon–intron junctions are also mis-annotated in the reported GmGα4 gene sequence (Fig. S1). The reported GmGα4 sequence has insertions of 20 and 28 bp at exons 3 and 4, respectively, and a 70 bp deletion of exon 12, thereby translating a very divergent Gα protein. It is possible that the reported GmGα4 sequence is a splice variant of the sequence amplified in this study and is only expressed under certain conditions or specific tissues.

Four homologs of the Arabidopsis Gβ protein (AtAGB1) were identified at loci Glyma11g12600.1, Glyma12g04810.1, Glyma06g01510.1 and Glyma04g01460.1 and called GmGβ1, GmGβ2, GmGβ3, and GmGβ4, respectively. The GmGβ proteins are highly homologous (89–98% identity) and share c. 80% identity with AtAGB1 and OsRGB1 (Table 1). The GmGβ proteins contain all the conserved features found in canonical Gβ, including the seven WD(Trp-Asp)-repeat motifs (Fig. S2) and sequences important for the interaction with Gα and Gγ proteins (Temple & Jones, 2007).

To complete the inventory of G-protein subunits from soybean, we performed BLAST analysis with the Arabidopsis Gγ proteins (AtAGG1 and AtAGG2) and identified two loci, Glyma10g03610.1 and Glyma02g16190.1 (called GmGγ1 and GmGγ2, respectively). The GmGγ proteins share c. 50% sequence similarity with AGG1 and AGG2 and c. 90% with each other. The GmGγ proteins contain the predicted N-terminal coil-coiled domain and a CAAX motif (A, aliphatic, X, any amino acid) at the C-termini (Fig. S3). The ‘DPLL’ motif and other amino acids involved in forming the hydrophobic contact surface with Gβ are also conserved (Temple & Jones, 2007). The GmGα, and genes form duplicated gene pairs as is shown by the analysis of their phylogeny and chromosome location (Figs S4–S6).

Full-length cDNAs were isolated for each of the 10 G-protein subunit genes using soybean seedling cDNA. Given the multiplicity of G-protein subunits, 32 Gαβγ combinations are possible in soybean, making it the most diverse plant G-protein network known to date. Only two such combinations exist in Arabidopsis and rice (Assmann, 2005).

Tissue expression of soybean heterotrimeric G-protein genes

To confirm that all 10 G-protein subunits genes are expressed in soybean, we quantified their absolute expression by measuring the copy number of each transcript using 7-d-old primary root (V1R) as reference tissue. qRT-PCR was used to analyze expression patterns with gene-specific primers (Table S2). Standard curves were generated for each primer pair using a 100-fold serial dilution of the corresponding plasmid DNA. A linear correlation coefficient (R2) of 0.97–0.99 was observed over 10 000-fold dilution range and reflects strong PCR efficiency of each primer pair. All 10 G-protein subunit genes were expressed at comparable levels in soybean roots (Table S3).

The multiplicity of subunits and their comparable expression levels led us to perform a detailed expression analysis during different developmental stages and in various tissue types. All 10 soybean genes were ubiquitously expressed (Fig. 2a,b), similar to the expression pattern reported for Arabidopsis and rice (Weiss et al., 1993, 1994; Huang et al., 1994; Chen et al., 2006; Anderson & Botella, 2007; Izawa et al., 2010). The G-protein genes were expressed in both developing and mature tissues. In general, the expression of G-protein genes was lower in roots than in aerial tissues. Except for GmGα4, which is highly expressed in the primary stem and the first trifoliate leaves at V1 stage, the soybean G-protein genes were expressed at comparable levels in all different developmental stages and tissue types with moderate differences (two- to four-fold) in some cases (Fig. 2a,b).

Figure 2.

 Expression of soybean G-protein genes in different tissue types. The stages are defined as: VC, cotyledon; V1-R, primary root at stage V1 (appearance of the first set of unfolded trifoliolate leaves); V1-S, primary stem at stage V1; V1-TF, first trifoliate leaf; Vn-R, mature root; Vn-S, mature stem; Vn-L, mature leaves; Apex, shoot apex; Vn-F, flower; S4, seed stage S4. Real-time quantitative PCR (qRT-PCR) amplifications were performed three times independently for each target, and the data were averaged. The expression values across different tissue types were normalized against soybean Actin gene expression (Subramanian et al., 2004). Error bars represent the standard error of the mean. (a) Relative expression of the GmGα genes. Expression in V1-R roots was set at 1. (b) Relative expression of the GmGβ and GmGγ genes. Expression in V1-R roots was set at 1. (c) Relative expression in nodules and hairy roots. Expression in mature roots (Vn-R) is set at 1.

We also examined the expression of G-protein genes in nodules, a legume-specific tissue, and in hairy roots. Interestingly, GmGα1-3 showed significantly increased expression in nodules (expression in root set at 1), whereas GmGα4 showed high expression in hairy roots. All GmGβ and GmGγ genes were also expressed in these tissues with a c. twofold increase in RNA levels in nodules (Fig. 2c). Expression of GmGβ and GmGγ was generally low in hairy roots (Fig. 2c).

Expression of soybean G-protein genes during seed development and germination

Molecular genetic studies of Arabidopsis and rice indicate that G-proteins play important biological roles during seed germination and development (Perfus-Barbeoch et al., 2004). Seeds are the most important part of soybean plants because of their importance as food and feed. Since all G-protein genes are expressed in developing soybean seeds (stage S4, Fig. 2a), we focused on further characterizing the expression of individual soybean G-protein genes at different stages of seed development and during seed germination.

The development of soybean seeds is classified into well-defined stages (S1–S8) based on their FW (Miernyk et al., 2010). The G-protein genes were expressed during S1–S8 and in mature dry seeds and showed interesting expression patterns.

GmGα1 and GmGα4 RNA levels were relatively constant from S1 to S8 and in dry seeds, whereas GmGα2 and GmGα3 showed a significant increase following S6 (Fig. 3a). Specifically, the expression level of GmGα2 was sevenfold higher from S7 to S8, when the seeds begin to desiccate. GmGα protein could be detected throughout seed development and maturation as shown by western blotting with AtGPA1 antibodies (Fig. 3a, inset). The apparent decrease in the level of GmGα proteins during stages S7 to dry seeds seems to contradict the gene expression data; however it is notable that during this phase, the majority of seed proteome consists of storage proteins, and we are using total seed proteins for the detection of GmGα proteins. Two of the GmGβ genes, GmGβ2 and GmGβ3, also showed an 80-fold and 180-fold increase, respectively, in RNA level during S7 and S8. In comparison, GmGβ1 and GmGβ4 were expressed at a relatively constant level (Fig. 3b). The two GmGγs showed a modest, but statistically significant decrease (P < 0.05) in RNA level, especially during S1–S4 (Fig. 3c).

Figure 3.

 Expression of soybean G-proteins during different stages of seed development. The seed development stages (S1–S8) are as per Miernyk et al. (2010). Real-time quantitative PCR (qRT-PCR) amplifications were performed three times independently for each target, and the data were averaged. The expression values across different seed stages were normalized against soybean Actin gene expression and expression in seeds at stage S1 was set at 1. Dry seed (DS) stage was also used for the analysis. Error bars represent the standard error of the mean. (a) Relative expression of the GmGα genes in during different stages of seed development; inset shows the expression of GmGα proteins during different stages of seed development; (b) relative expression of the GmGβ genes; (c) relative expression of the GmGγ genes.

RNA levels of GmGα, GmGβ and GmGγ were also analyzed during seed germination, starting with dry seeds until 30 h post-imbibition when germination is complete. The expression of all four GmGα genes increased post-imbibition (Fig. 4a), although each gene followed a distinct pattern of expression. The relative expression of GmGα1 and GmGα4 peaked at 12 h post-imbibition followed by a decrease. GmGα2 and GmGα3 had the highest expression 6 h post-imbibition, after which the level of GmGα2 was maintained for 24 h. At 30 h post-imbibition, all genes had RNA levels either similar to or lower than their expression in dry seeds. All GmGβ showed increased expression at the onset of germination (Fig. 4b). The GmGβ1 and GmGβ2 gene pair exhibited maximum expression at 6 h post-imbibition and the GmGβ3 and GmGβ4 gene pair showed maximum expression at 12 h post-imbibition. The RNA level of all four GmGβs decreased below their RNA levels in dry seeds at 18 h post-imbibition. The two GmGγ genes showed a twofold increase in expression during germination (Fig. 4c). These data strongly suggest the possible involvement of G-proteins during soybean seed development and germination, with some specificity of regulation in the duplicated gene pairs.

Figure 4.

 Expression of soybean G-proteins during seed germination. Seed germination was followed starting from dry seeds (0 h) up to 30 h when an obvious radicle had protruded. Seed samples were collected at every 6 h following imbibition of dry seeds. Real-time quantitative PCR (qRT-PCR) amplification experiments were performed three times independently for each target, and the data were averaged. The expression values across different seed germination stages were normalized against soybean Actin gene expression and expression in dry seeds at 0 h was set at 1. Error bars represent the standard error of the mean. (a) Relative expression of the GmGα genes; (b) relative expression of the GmGβ genes; (c) relative expression of the GmGγ genes during seed germination.

Expression of soybean G-proteins in response to ABA and abiotic stresses

Transcriptional regulation of G-proteins in response to abiotic stresses and/or ABA may occur. The pea Gα and Gβ transcripts showed up-regulation in response to salt, hydrogen peroxide, and heat stress (Misra et al., 2007). A Brassica napus Gα transcript is also up-regulated in response to prolonged (24 h) ABA and salt treatments (Gao et al., 2010). We tested the effects of different abiotic stresses and ABA treatment on expression of the 10 G-protein genes using young soybean seedlings. We used ABA, mannitol, cold, heat, and salt to treat soybean seedlings (stage V1). Roots and shoots were harvested separately and used for the source of cDNA for qRT-PCR. A soybean homolog of Arabidopsis RAB18 (an ABA- and stress-responsive gene; Welin et al., 1994), referred to here as GmRab18, was used as a positive control. No significant differences in the expression of GmGα, GmGβ or GmGγ RNA levels were observed under the conditions tested (Fig. S7), whereas the GmRAB18 RNA level was significantly higher in stress-treated roots than in the control. We analyzed 2 kb upstream promoter regions of G-protein subunit genes reported in using the PLAnt Cis-acting regulatory DNA Elements database (PLACE; Many cis-regulatory elements related to light and hormone signaling and tissue type-dependent expression, were over-represented, suggesting regulation of soybean G-protein genes during multiple growth and development processes (Table S4). By contrast, very few regulatory elements related to stress responsiveness were present in the promoters of the G-protein genes. Moreover, the duplicated gene pairs contain similar cis-elements in the upstream regions, corroborating their somewhat similar expression patterns.

Localization of soybean Gα proteins

Similar to mammalian Gα proteins, the Arabidopsis and rice Gα localize to the plasma membrane (Iwasaki et al., 1997; Weiss et al., 1997). Targeting of Gα to the plasma membrane is determined by N-terminal sequences for palmitoylation and myristoylation (Marrari et al., 2007). The consensus sequence for myristoylation (MGXXXS) is conserved in all four GmGα proteins (based on the second translational start site for GmGα2 and GmGα3). Although the consensus for palmitoylation is not well defined, the conserved cysteine at position 5 in GmGα1 and GmGα4 and at position 13 in GmGα2 and GmGα3 is predicted to be an S-acylation target by the CSS Palm 2.0 palmitoylation prediction program (Ren et al., 2008). These sequence features suggest plasma membrane localization of GmGα.

To determine the localization of the soybean Gα proteins, we assessed their relative expression in soluble and microsomal fractions of soybean roots using western blot analysis with AtGPA1 antibodies. Gα protein was primarily detected in the microsomal fractions (Fig. 5a), similar to what is reported for AtGPA1 (Pandey et al., 2006). We further confirmed the subcellular localization of GmGα-YFP fusion proteins using transient Agrobacterium infiltration-based expression in tobacco leaf. The AtGPA1-YFP fusion protein constructed in same vector as GmGα-YFP was used as a positive control. As shown for the positive control, all four GmGα-YFP proteins were also localized at the periphery, consistent with their plasma membrane locale (Fig. 5b).

Figure 5.

 Localization of GmGα proteins. (a) Western blot of GmGα proteins in microsomal (MF) and soluble fractions (SF) of the total seedling proteins with anti-AtGPA1 antibodies. Arabidopsis seedling proteins are used as controls. (b) Localization of AtGPA1:YFP and GmGα1-4:YFP in transiently transformed tobacco leaves. Inset: brightfield images of the represented leaves. A. thaliana, Arabidopsis thaliana; G. max, Glycine max

GTP-binding and GTP-hydrolysis activity of recombinant GmGα proteins

To establish the GmGα as authentic G-proteins, we assayed their GTP-binding and GTPase activities. Full-length cDNAs corresponding to AtGPA1 and GmGα1-4 were expressed as recombinant proteins in Escherichia coli and purified by Ni2+-affinity chromatography. Each protein was purified to > 95% purity and was immunoreactive with anti-AtGPA1 antibodies (Fig. 6a,b).

Figure 6.

 Purification and G-protein activity assay of GmGα proteins. (a) Purified recombinant GmGα proteins along with GPA1. (b) Western blot of GmGα proteins and AtGPA1 with AtGPA1 antibodies. (c) Guanosine-5′-triphosphate (GTP) binding and hydrolysis of recombinant GmGα proteins and GPA1 measured using GTP-4,4-difluoro-4-bora-3α,4α-diaza-s-indacene (BODIPY) FL in real-time fluorescence assays. Data are one of two independent experiments, each with three replicates (mean ± SD). (d) Specificity of GTP binding and hydrolysis of GmGα2 protein. Nonfluorescent GTP and guanosine diphosphate (GDP) compete efficiently for fluorescent GTP binding but not nonfluorescent ATP or ADP. Data are mean of three replicates, error bars ± SD.

We used a fluorescence-based real-time assay to evaluate the G-protein activity of each GmGα using BODIPY-GTP FL, where an increase in fluorescence over time corresponds to GTP binding. The inherent GTPase activity of Gα causes GTP hydrolysis, which is represented as a decrease in fluorescence over time in this assay (Willard & Siderovski, 2004; Pandey et al., 2009). All four GmGα proteins exhibit the signature GTP-binding and GTP-hydrolysis activities of known G-proteins (Fig. 6c), such as the AtGPA1 control. The rate of GTP binding of all four proteins is comparable; however, clear differences exist in their GTPase activity. GmGα1 is similar to GmGα4 in its slow rate of GTP hydrolysis, as shown by the relatively flat slope of the downward curve (Fig. 6c). These proteins show GTPase activity comparable to AtGPA1 (Johnston et al., 2007). By contrast, GmGα2 and GmGα3 display a faster rate of GTP hydrolysis, as shown by the steeper slope of the curves (Fig. 6c). This functional analysis based on the activity assay in a heterologous system indicates that the genome duplication event in soybean leading to two different groups of G-proteins has introduced significant differences in their biochemical properties.

We also tested the specificity of GTP binding and hydrolysis using a competition assay with nonfluorescent nucleotides (GTP, GDP, ATP and ADP). GTP and GDP efficiently compete with fluorescent GTP (Figs 6d, S8). No effect was observed with either ADP or ATP (Figs 6d, S8) and confirms the specificity of each of GmGα towards GTP/GDP binding and GTP hydrolysis.

Protein–protein interaction analysis between soybean G-protein subunits

To verify the presence of multiple heterotrimeric combinations of soybean G-proteins, we performed targeted yeast-based interaction assays. Interactions between GmGα and GmGβ in all 16 possible combinations were assessed using the membrane-based split ubiquitin system (Obrdlik et al., 2004; Pandey & Assmann, 2004) with GmGβ as bait proteins and GmGα as prey proteins, respectively. Growth assays on selective media supplemented with 1 mM methionine confirmed interactions between GmGα and GmGβ (Fig. 7a). As expected, GmGα and GmGβ interact in most of the possible combinations (13 out of 16), but display a degree of interaction specificity between the duplicated gene pairs. GmGβ1 and GmGβ2 (a duplicated GmGβ pair) interact strongly with GmGα2 and GmGα3 (a duplicated GmGα pair), but with less efficiency for GmGα1 and GmGα4. GmGβ3 and GmGβ4 showed a strong and exclusive interaction with GmGα1 and GmGα4 (Fig. 7a).

Figure 7.

 Interaction between soybean G-protein subunits. (a) Interaction between GmGα and GmGβ proteins using split-ubiquitin-based interaction assay. The picture shows yeast growth on selective media with 1 mM methionine. In all cases, Gα proteins were used as Nub (N-terminal half of ubiquitin) fusions in both orientations and Gβ proteins were used as Cub (C-terminal half of ubiquitin) fusions. Nubwt (N-terminal half of wild-type ubiquitin) fusion constructs were used as positive controls for interaction and Nub vector fusions were used as negative controls. Two biological replicates of the experiment were performed with identical results. (b) Interaction between GmGβ (in pDEST32) and GmGγ (in pDEST22) proteins using yeast-two-hybrid-based colorimetric assay. The assays were performed in triplicate and data were averaged. Error bars represent the standard error of the mean. Two biological replicates of the experiment were performed with similar results.

To determine the interaction between GmGβ and GmGγ, we used a conventional yeast two-hybrid assay, wherein GmGβ and GmGγ were used as bait and prey proteins, respectively. We performed a β-galactosidase activity assay to determine the quantitative differences between their interactions. GmGγ1 and GmGγ2 exhibit strong interaction with GmGβ2 and GmGβ4 but, surprisingly, no interaction with GmGβ1 and GmGβ3 (Fig. 7b). In these assays, GmGγ1 is a relatively stronger interactor compared with GmGγ2. The interaction data suggest the existence of subunit-specific heterotrimers of G-proteins in soybean.


Multiplicity and phylogeny of heterotrimeric G-proteins in the plant kingdom

Comprehensive analysis of plant G-proteins that integrate molecular, genetic, and biochemical characterization and their roles in regulating specific signal transduction pathways is mostly limited to two model species – Arabidopsis and rice (Perfus-Barbeoch et al., 2004). To date, a single canonical gene has been reported from three monocots and 10 dicots (except for pea, wheat and tobacco with two  genes each; Assmann, 2002). Similarly, most plant and fungal species (except tobacco, wild oat, and moss) have a single conventional Gβ subunit based on genome sequencing (Table S1). The presence of a limited repertoire of conventional G-proteins with two possible heterotrimeric combinations (αβγ1 and αβγ2) in both Arabidopsis and rice, along with the slow GTP-hydrolysis activity of Arabidopsis Gα, has led to the hypothesis that plants have a simple and mechanistically different G-protein signaling network compared with mammals (Temple & Jones, 2007). The identification of four Gα, four Gβ and two Gγ sequences, leading to 32 possible heterotrimeric combinations in soybean and the identification of two distinct types of GTP-hydrolysis activities in GmGα, indicate that G-protein signaling networks are more diverse and complex in the plant kingdom than was previously thought.

Amino acid alignment and comparison of soybean Gα, β and γ subunits with other known plant G-proteins confirmed the presence of all signature sequences required for their function (Figs 1, S2, S3). A distinct feature of GmGα2 and GmGα3 is the presence of an N-terminal eight-amino-acid extension that results in two possible translational start sites. Although similar in-frame multiple translational start sites have been reported in the tobacco, rice, tomato and wheat Gα sequences, its significance is unclear (Kaydamov et al., 2000). Both these groups of proteins show predominant plasma membrane localization (Fig. 5); however, detailed analysis by organelle-specific biochemical fractionation or stable expression in nonvacuolated cell type, such as dividing root tip, would further strengthen this observation. The extreme C-termini of Gα proteins determine the specificity of their interaction with receptors (Lambright et al., 1994). Based on the available plant Gα sequences, the C-termini of monocot and dicot Gα seem to be quite divergent (Ashikari et al., 1999; Hossain et al., 2003) and the C-termini of the GmGα are similar to those of other dicot Gα. The availability of more Gα sequences from other plant species will substantiate this observation and determine the role of the dicot- and monocot-specific Gα C-termini, if any, in eliciting specific cellular responses.

It is notable that we have identified only two Gγ sequences from soybean; a number similar to that reported from Arabidopsis and rice. If the Gγ proteins of soybean are to follow the two genome duplication events of Gα and Gβ, then eight proteins are predicted which would result in 128 theoretically possible heterotrimeric combinations. However, the identification of plant Gγ subunits based on sequence alone is difficult because of their small size (7–10 kDa) and lack of significant similarity with mammalian homologs (Mason & Botella, 2000, 2001). This was also evident when homology-based approaches could not reveal Gγ proteins from the Arabidopsis genome, requiring their identification by interaction with Gβ (Mason & Botella, 2000, 2001). Overall, plant Gγ proteins are more closely related to each other, as compared with the nonplant proteins, and all known plant Gγ proteins have been identified as Arabidopsis Gγ homologs (Kato et al., 2004; Misra et al., 2007). This does not rule out the presence of additional unidentified Gγ proteins. Improved annotation of the soybean genome and screening for GmGβ interacting proteins may divulge additional GmGγ, which would add greatly to the diversity of possible heterotrimeric combinations.

With the exception of Gγ proteins, the multiplicity of soybean G-proteins corresponds well with the two genome duplication events (Schmutz et al., 2010), which is also corroborated by their phylogeny (Figs S4, S5). Interestingly, the moss Gβ sequences are in the same clade as the higher plant Gβ sequences, but this is not observed with the moss Gα-like sequence. The reported moss Gα-like protein (XP_001772174, 611 amino acids) is significantly larger than other known plant Gα proteins (c. 380 amino acids). Analysis of the moss Gα proteins identified all signature motifs required for G-protein activity, as well as motifs for βγ-binding and receptor interaction (Fig. S9). Biochemical characterization of this protein with respect to its signature GTP-binding and hydrolytic activity will ascertain if it is a true Gα ortholog. Alternatively, the presence of classic Gβ-like proteins in moss is also intriguing if it does not possess a classic Gα-like protein.

Genome duplication events serve as important sources for introducing functional diversity. Duplicated genes usually evolve to exhibit altered expression levels and tissue-specific expression, or they become nonfunctional (Adams, 2007; Jackson & Chen, 2009). Even though soybean genome duplication is fairly recent (dating back to 59 and 13 million yr ago, Schmutz et al., 2010), our analysis confirms that all four and genes are retained and expressed at comparative levels and that the Gα proteins are active. Additionally, transcript profiling, biochemical activity and protein–protein interaction analysis suggest that the first round of duplication event has conferred some specificity in their expression level and function. The GTPase activities of GmGα1 and GmGα4 are distinct from those of GmGα2 and GmGα3 (Fig. 6). In addition, GmGβ3 and GmGβ4 interact with GmGα1 and GmGα4, but not with GmGα2 and GmGα3 (Fig. 7). Since the syntenic gene pairs are still very similar at the sequence level, future research focused on determining the roles of individual amino acids conferring such specificity promises to reveal important information about the regulation of G-protein signaling in plant species.

Developmental regulation of soybean heterotrimeric G-protein expression

In mammalian systems, the expression of Gαβγ is usually cell- and tissue-specific (Melien, 2007; Oldham & Hamm, 2008). Similarly, the two Arabidopsis Gγ proteins exhibit some degree of tissue specificity with its overall expression overlapping expression of Gβ (Anderson & Botella, 2007; Trusov et al., 2007, 2008). The presence of multiple copies of each of the G-protein subunits in soybean provides the opportunity to assess their specific expression patterns. qRT-PCR analysis of G-protein expression confirmed nearly ubiquitous expression of all 10 genes in all tissues examined with some degree of quantitative differences (Fig. 2). For example, nodules show higher expression of GmGα1–3, but not GmGα4. This is interesting because multiple pharmacological experiments suggest the involvement of G-proteins in signaling processes related to nodulation (Pingret et al., 1998; den Hartog et al., 2001; Kelly & Irving, 2003; Charron et al., 2004; Sun et al., 2007; Santos-Briones et al., 2009).

We focused on transcriptional regulation of G-protein subunits during seed development and germination given their involvement during these processes in Arabidopsis and rice (Fujisawa et al., 1999; Ullah et al., 2002; Chen et al., 2004; Pandey et al., 2006). The soybean G-protein genes are expressed throughout seed development and in dry seeds (Fig. 3). During the late maturation phase (S6 onwards) when seeds start to dry, the RNA levels of GmGα2, GmGα3, GmGβ2 and GmGβ3 increase significantly. This developmental stage, when the seed is preparing for quiescence, is highly regulated by ABA and these specific G-protein subunits might be involved in ABA signal transduction. During early phases of seed germination, the expression of the soybean G-protein genes was significantly up-regulated (Fig. 4). A similar up-regulation of AtGPA1 has been reported previously (Pandey et al., 2006) Interestingly, the duplicated gene pairs for both the GmGα and GmGβ genes showed a similar overall expression profile during seed imbibition and germination (Fig. 4).

GTP-binding and GTPase activity of GmGα proteins

A key find of our experiments is the presence of two distinct types of GTPase activity in the syntenic pairs of recombinant GmGα proteins. Signaling through G-proteins is determined by the biochemical activity of Gα subunits in the heterotrimer, which balances the rate of GTP binding, GTP hydrolysis, and GDP release (Cabrera-Vera et al., 2003). Our results indicate that the GmGα proteins are grouped into two categories with distinct biochemical properties.

Although the overall rates of GTP binding for the GmGα are comparable with that of AtGPA1 (Fig. 6), the hydrolysis rates can differ. GmGα1 and GmGα4 indeed exhibit relatively slow rates of GTP hydrolysis, similar to AtGPA1 (Fig. 6). In comparison, GmGα2 and GmGα3 display more rapid rates of GTP hydrolysis. Earlier studies showing the slow GTPase activity of AtGPA1 led to the suggestion that plant G-proteins are always active and do not undergo a GPCR-induced nucleotide exchange upon perception of a signal (Johnston et al., 2007). The presence of a fast GTPase activity in these two GmGα proteins questions the hypothesis that plant G-protein signaling is fundamentally different from nonplant systems. The difference in GTPase activity of GmGα2 and GmGα3 vs GmGα1, GmGα4, and AtGPA1 suggests the presence of an integrated GPCR-based signal transduction mechanism for G-protein signaling in soybean. Further detailed kinetic characterization of the GmGα proteins will help to address some of these issues and, importantly, may suggest specific mechanisms for controlling signaling events associated with various physiological and developmental processes.

Interaction specificity of soybean heterotrimeric G-proteins

The presence of multiple G-protein subunits and their ubiquitous expression led us to analyze the specificity of subunit interactions. Although the four GmGα proteins are quite similar at the sequence level and the proteins interacted in most of the possible combinations, some specificity in their interaction was observed. GmGβ1 and GmGβ2 interact strongly with GmGα2 and GmGα3, but less efficiently with GmGα1 and GmGα4, whereas GmGβ3 and GmGβ4 showed strong and exclusive interactions with GmGα1 and GmGα4 (Fig. 7a). The interaction data agree with the phylogeny and genome duplication events associated with the GmGα and GmGβ gene pairs.

Both Arabidopsis Gγ proteins (48% sequence identity) interact strongly with the sole Gβ protein, but with some difference in their interactions. Deletion of the C-terminal domain of AtAGG1 reduces the strength of its interaction with AtAGB1, whereas deletion of the C-terminal domain of AtAGG2 completely abolishes binding (Mason & Botella, 2000, 2001). Given the 90% sequence identity between the GmGγ, it is surprising to observe a strong specificity of interaction for GmGβ2 and GmGβ4 (Fig. 7b). It may be possible that additional factors are required for interaction of Gγ with other Gβ proteins or there may be unidentified Gγ proteins that interact specifically with GmGβ1 and GmGβ3. Overall, our data suggest the presence of subunit specific heterotrimeric combinations in soybean. Additional specificity of interaction arising as a response to a particular signal or localization in different tissue types may also occur, but it has not been assessed in any plant species. Further analysis of specific interactions between different G-protein subunits using in vivo approaches will help evaluate the biological significance of specific heterotrimer or dimer formation.


This work represents a necessary first step towards characterization of the most elaborate heterotrimeric G-protein network in a plant to date. Since an estimated 30–80% of living plant species are polyploid, this research provides a glimpse of the possible complexity of plant G-protein signaling networks. Development of RNAi knockout and overexpression lines of respective G-protein subunit genes will help us to elucidate the molecular mechanisms by which heterotrimeric G-proteins signaling controls various developmental processes in soybean.


We thank Dr Jan A. Meirnyk, University of Missouri, Columbia, for providing different stages of developing soybean seeds and Dr Alan M. Jones University of North Carolina, Chapel Hill, for anti-GPA1 antibodies. We also thank Drs Christopher Taylor, Senthil Subramanian and Manjula Govindarajulu for their help with various soybean-related resources. N.C.B. was supported by a short-term visiting scholar fellowship from NIPGR, India. This work was supported by USDA/AFRI (2010-65116-20454) grant to S.P.