OPEN ACCESS ARTICLE

This Article Free via Open Access: OA OA Abstract Full Text (PDF) Supplemental Data OA All Versions of this Article: 147/2/636    most recent pp.108.117655v2 pp.108.117655v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Google Scholar Articles by Trusov, Y. Articles by Botella, J. R. PubMed PubMed Citation Articles by Trusov, Y. Articles by Botella, J. R. Agricola Articles by Trusov, Y. Articles by Botella, J. R.

CELL BIOLOGY AND SIGNAL TRANSDUCTION

G1 + G2 Gβ: Heterotrimeric G Protein G-Deficient Mutants Do Not Recapitulate All Phenotypes of Gβ-Deficient Mutants^{1,[C],[W],[OA]}

Plant Genetic Engineering Laboratory, Department of Botany, School of Integrative Biology, University of Queensland, Brisbane, Queensland 4072, Australia (Y.T., J.R.B.); and Biology Department, Pennsylvania State University, University Park, Pennsylvania 16802–5301 (W.Z., S.M.A.)

	ABSTRACT

TOP ABSTRACT RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

 
Heterotrimeric G proteins are signaling molecules ubiquitous among all eukaryotes. The Arabidopsis (Arabidopsis thaliana) genome contains one G (GPA1), one Gβ (AGB1), and two G subunit (AGG1 and AGG2) genes. The Gβ requirement of a functional G subunit for active signaling predicts that a mutant lacking both AGG1 and AGG2 proteins should phenotypically resemble mutants lacking AGB1 in all respects. We previously reported that Gβ- and G-deficient mutants coincide during plant pathogen interaction, lateral root development, gravitropic response, and some aspects of seed germination. Here, we report a number of phenotypic discrepancies between Gβ- and G-deficient mutants, including the double mutant lacking both G subunits. While Gβ-deficient mutants are hypersensitive to abscisic acid inhibition of seed germination and are hyposensitive to abscisic acid inhibition of stomatal opening and guard cell inward K⁺ currents, none of the available G-deficient mutants shows any deviation from the wild type in these responses, nor do they show the hypocotyl elongation and hook development defects that are characteristic of Gβ-deficient mutants. In addition, striking discrepancies were observed in the aerial organs of Gβ- versus G-deficient mutants. In fact, none of the distinctive traits observed in Gβ-deficient mutants (such as reduced size of cotyledons, leaves, flowers, and siliques) is present in any of the G single and double mutants. Despite the considerable amount of phenotypic overlap between Gβ- and G-deficient mutants, confirming the tight relationship between Gβ and G subunits in plants, considering the significant differences reported here, we hypothesize the existence of new and as yet unknown elements in the heterotrimeric G protein signaling complex.

β

β

In animal systems, G proteins mediate the signaling of over 800 receptors (G protein-coupled receptors; Pierce et al., 2002; Fredriksson et al., 2003; Zhang et al., 2006). Multiple family members exist for each of the three subunits, and different combinatorial possibilities provide the required specificity for multitudinous G protein-based signaling pathways (Gautam et al., 1998; Balcueva et al., 2000; Wettschureck and Offermanns, 2005; Marrari et al., 2007). In open contrast, plants only contain one or two genes for each of the subunits (Ma et al., 1990; Poulsen et al., 1994; Weiss et al., 1994; Gotor et al., 1996; Iwasaki et al., 1997; Seo et al., 1997; Marsh and Kaufmann, 1999; Ando et al., 2000; Mason and Botella, 2000, 2001; Perroud et al., 2000; Kang et al., 2002; Hossain et al., 2003a, 2003b; Kato et al., 2004; Misra et al., 2007). In Arabidopsis (Arabidopsis thaliana), a single G (Ma et al., 1990; Ma, 1994), a single Gβ (Weiss et al., 1994), and two G (Mason and Botella, 2000, 2001) subunit genes have been identified.

Despite the fact that only two combinatorial possibilities are feasible for the heterotrimers in Arabidopsis, G proteins are involved in multiple processes (Assmann, 2004; Jones and Assmann, 2004; Perfus-Barbeoch et al., 2004). Recent genetic studies using G- and Gβ-deficient or -overproducing mutants have demonstrated the involvement of G proteins in abscisic acid (ABA) and brassinosteroid sensitivity during seed germination and early plant development (Ullah et al., 2002; Lapik and Kaufman, 2003; Chen et al., 2004, 2006; Pandey et al., 2006; Warpeha et al., 2006), stomatal regulation (Wang et al., 2001), D-Glc signaling (Huang et al., 2006; Wang et al., 2006), light perception (Okamoto et al., 2001; Warpeha et al., 2006, 2007), rosette leaf, flower, and silique development (Lease et al., 2001; Ullah et al., 2003), plant defense against necrotrophic fungi (Llorente et al., 2005; Trusov et al., 2006), and auxin signaling in roots (Ullah et al., 2003; Trusov et al., 2007). Similar functional multiplicity has been observed in rice (Oryza sativa; Ashikari et al., 1999; Fujisawa et al., 1999; Ueguchi-Tanaka et al., 2000; Suharsono et al., 2002; Komatsu et al., 2004; Oki et al., 2005).

In animals and fungi, the existence of a functional G subunit is a compulsory prerequisite for the functioning of the entire heterotrimer, and lack of G subunits results in the obliteration of both Gβ- and G-mediated pathways (Gilman, 1987; Kisselev et al., 1994; Manahan et al., 2000; Krystofova and Borkovich, 2005; Myung et al., 2006). There is one notable exception to this rule: the human neurospecific Gβ5 subunit forms functional dimers with some members of the regulator of G protein signaling (RGS) subfamily C proteins instead of with conventional G subunits (Snow et al., 1998; Sondek and Siderovski, 2001). Plant G proteins appear to behave similarly to their animal counterparts, and tight physical interaction between Gβ and each of the G subunits has been demonstrated in vitro (Mason and Botella, 2000, 2001) and in vivo (Kato et al., 2004; Adjobo-Hermans et al., 2006). Moreover, it has been shown that Arabidopsis Gβ subunit localization on the plasma membrane requires a G subunit (Obrdlik et al., 2000; Adjobo-Hermans et al., 2006). Evidence for functional interaction of the subunits was recently provided using overexpression of a truncated G1 subunit lacking the isoprenylation motif (which is responsible for anchoring the β dimer to the membrane) and mutants lacking each of or both G subunits (Chakravorty and Botella, 2007; Trusov et al., 2007). It has also been shown that different G subunits confer specificity to the Gβ dimer, with the Gβ1 dimer mediating signal transduction events during plant defense against necrotrophic fungi, acropetal auxin signaling in roots, and osmotic stress regulation of seed germination, while Gβ2 is involved in basipetal auxin signaling in roots and D-Glc sensing during germination (Trusov et al., 2007).

However, despite the above-mentioned similarities between plant and animal G proteins, a number of unusual properties observed in the plant subunits have led to suggestions that, in some cases, the animal paradigm may not necessarily hold true in plants. For instance, it was established that unlike animal subunits, Arabidopsis G and Gβ are capable of tethering to the plasma membrane independently and do not rely on each other (Adjobo-Hermans et al., 2006; Zeng et al., 2007; Wang et al., 2008). It was also shown that the GTPase activity of the Arabidopsis G subunit is very low, leading to the hypothesis by some authors that G might be in the activated state by default (Willard and Siderovski, 2004; Johnston et al., 2007a; Temple and Jones, 2007). In addition, the interaction between G and the Gβ dimer in rice has been reported to be relatively weak compared with that in animal systems (Kato et al., 2004), although formation of the heterotrimer in vivo has been demonstrated (Kato et al., 2004; Adjobo-Hermans et al., 2006). Recently, it was shown that in Arabidopsis both G and Gβ are associated with large macromolecular complexes of approximately 700 kD (Wang et al., 2008). Finally, Arabidopsis G subunits are capable of being targeted to the plasma membrane in mutants lacking functional G and Gβ subunits (Zeng et al., 2007).

Here, we present data conflicting with the established heterotrimeric G protein model. We found that a number of phenotypic alterations observed in G- or Gβ-deficient mutants cannot be detected in mutants lacking G1, G2, or both G subunits. Our results raise the possibility that G subunits are not required for some G- and Gβ-mediated processes in Arabidopsis or that additional nonconventional G subunits exist in this species.

	RESULTS

TOP ABSTRACT RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

 

The Expression Profiles of AGG1 and AGG2 in Reproductive Organs Do Not Match the AGB1 Expression Pattern

We previously reported that GUS staining patterns in transgenic Arabidopsis ecotype Columbia (Col-0) plants carrying promoter fusions of each of the AGG1 and AGG2 genes with the GUS reporter gene were tissue specific and that together they overlap AGB1 expression patterns in most plant tissues and developmental stages (Anderson and Botella, 2007; Trusov et al., 2007). Nevertheless, a number of small but important differences can also be observed, especially in reproductive tissues. Analysis of GUS expression in flowers and siliques revealed that AGB1 is moderately expressed in sepals and stamen filaments (Fig. 1A ), with high expression found in stigma and pollen (Fig. 1, B and C). In siliques, GUS staining was observed at both ends, gradually disappearing toward the center (Fig. 1D). In AGG1:GUS plants, GUS expression was only detected in the stigma of mature flowers (Fig. 1, B and C), possibly correlating with pollination, while in siliques only the abscission zone showed staining (Fig. 1D). In AGG2:GUS transformants, GUS expression was evident in the apex of stamen filaments at a very early developmental stage and disappeared before the flower opened (Fig. 1C). No GUS staining was detected in siliques of AGG2:GUS plants. Discrepancies were also observed in germinating seeds, where distinct GUS staining in AGB1:GUS plants was observed much earlier (24 and 48 h after imbibition) than in AGG1:GUS or AGG2:GUS plants. However, expression in AGG1:GUS or AGG2:GUS increased to detectable levels and overlapped with AGB1 expression in 4-d-old seedlings (Fig. 1E), in agreement with a previous report (Trusov et al., 2007).

View larger version (97K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. AGB1, AGG1, and AGG2 expression patterns in flowers. Histochemical analysis of GUS expression in transgenic Arabidopsis plants carrying AGB1:GUS, AGG1:GUS, or AGG2:GUS fusions. A, Fully opened flowers. B, Stigma of young (unopened bud) and mature flowers. C, Anther, stamen filaments, and pollen. D, Green-stage siliques. E, Germinating seedlings at 24, 48, and 96 h after imbibition.

ABA Sensitivity of G1- and G2-Deficient Mutants in Germination

Heterotrimeric G protein involvement in different aspects of seed germination has been established by a number of studies (Ullah et al., 2002; Lapik and Kaufman, 2003; Chen et al., 2004, 2006; Pandey et al., 2006; Liu et al., 2007; Warpeha et al., 2007). Involvement of the heterotrimeric G protein signaling components GPA1, AGB1, GCR1, and RGS1 in ABA inhibition of seed germination is well documented (Ullah et al., 2002; Chen et al., 2004, 2006; Pandey et al., 2006). It was recently proposed that a second putative G protein-coupled receptor (GCR2) is a plasma membrane receptor for ABA (Liu et al., 2007); however, these results have been challenged (Gao et al., 2007; Johnston et al., 2007b; Illingworth et al., 2008).

Roles for the Arabidopsis G subunits, AGG1 and AGG2, in D-Glc and osmotic sensing during germination were recently reported (Trusov et al., 2007), but no data are available at present on their involvement in ABA sensing. Therefore, we compared the responses of G-, Gβ-, and G-deficient mutants to ABA-mediated inhibition of germination (Fig. 2 ). We analyzed germination rates for seven different seed lots for each genotype (stored in an identical environment for approximately 1 month after harvest) under a number of experimental conditions. Each lot was tested at least twice. In all tests, all genotypes showed 100% germination on control medium (no ABA added) by day 3 after light exposure (data not shown). In accordance with previous reports showing hypersensitivity of G- and Gβ-deficient mutants to ABA during germination (Pandey et al., 2006), gpa1-4 and agb1-2 mutants showed enhanced ABA-mediated inhibition of germination compared with the wild type (Fig. 2, A and B). In contrast, four independent single G-deficient mutants, agg1-1c, agg1-2, agg2-1, and agg2-2, as well as the double agg1 agg2 mutant showed levels of ABA sensitivity similar to wild-type plants and in one case (agg1-2 in 5 µM ABA) even showed hyposensitivity (P < 0.05; Fig. 2, A and B). In some isolated experiments, G1- or G2-deficient mutants showed either decreased or slightly increased ABA sensitivity compared with the wild type, but the responses never reached the hypersensitivity levels displayed by Gβ-deficient mutants. Figure 2, A and B, shows germination rates for the wild type and all mutants in a representative experiment. The differences in ABA sensitivity between gpa1-4 and agb1-2 mutants and all of the other genotypes analyzed (the wild type and G-deficient mutants) were statistically significant (P < 0.05).

View larger version (50K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 2. Sensitivity to ABA-induced inhibition of seed germination in G protein complex mutants. A and B, Seeds from matched seed lots were surface sterilized and plated on 0.5x Murashige and Skoog medium plates in the presence of 2 µM (A) or 5 µM (B) ABA. Plates were transferred to 100 µmol m–2 s–1 light and 23°C. Germination was recorded at 3, 4, 5, and 6 d after transfer of the plates under light and expressed as a percentage of total number of seeds. The experiment was repeated three times, and data were averaged (n > 100 for each experiment). The error bars represent SE. C, Rescue effect of Suc on germination inhibited by 5 µM ABA. Error bars indicate SE values obtained from four measurements. Asterisks indicate values statistically significantly different from the wild type (P < 0.05).

G-Deficient Mutants Do Not Display the agb1-2 Deetiolated Phenotype in Darkness

Partially deetiolated seedlings of G- and Gβ-deficient mutants have short hypocotyls with visibly increased girth and a characteristic open hook (Ullah et al., 2001, 2003; Wang et al., 2006). We analyzed hypocotyl elongation rate and hook development in darkness as well as light inhibition of hypocotyl elongation in gpa1-4, agb1-2, agg1-1c, agg1-2, agg2-1, agg2-2, and agg1 agg2 mutants as well as in the wild type. To ensure synchronized germination, all seeds were stratified for 5 d and induced to germinate under 150 µmol m^–2 s^–1 continuous light during 24 h. Afterward, plates with seeds were placed vertically either in a dark cabinet or under continuous light (90 µmol m^–2 s^–1) for 24, 48, and 72 h.

Figure 3, A and B , shows hypocotyl elongation dynamics of all mutant genotypes and the wild type grown in darkness and light, respectively. In agreement with previous reports (Ullah et al., 2001, 2003), hypocotyl elongation rates for G- and Gβ-deficient mutants (gpa1-4 and agb1-2) in darkness were lower compared with those for the wild type, with statistically significant differences after 24 and 48 h (P < 0.01 and P < 0.05, respectively). Under light, gpa1-4 and agb1-2 hypocotyls were significantly shorter than wild-type hypocotyls at all three time points (at least P < 0.05). In open contrast to the behavior of gpa1-4 and agb1-2 mutants, hypocotyl elongation in either dark or light conditions in each of the individual G1- or G2-deficient mutants or the double agg1 agg2 mutant was not statistically different from that in the wild type.

View larger version (39K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 3. Seedling development of G protein complex mutants grown in darkness or under light. A and B, Hypocotyl elongation rates of the wild type and the mutants at 24, 48, and 72 h after germination in darkness (A) or under 90 µmol m–2 s–1 light (B). Error bars indicate SE. At least 20 seedlings were measured. C, Degree of hook opening in dark-grown wild-type and mutant seedlings measured approximately 24 h after germination. SE values indicated by error bars are based on a minimum of 20 seedlings. Closed hooks were treated as having zero degree of opening. Triple asterisks indicate values statistically significantly different from the wild type (P < 0.001). The inset shows representative seedlings in order from left to right: Col-0, gpa1-4, agb1-2, agg1-1c, agg1-2, agg2-1, agg2-2, and agg1 agg2.

None of the Morphological Alterations of Aerial Organs Observed in G- and Gβ-Deficient Mutants Is Present in G-Deficient Mutants

Phenotypic analyses of G- and Gβ-deficient mutants have revealed a number of developmental and morphological abnormalities (Ullah et al., 2003). In order to determine whether G-deficient mutants showed similar traits, we compared the wild type, gpa1-4, agb1-2, agg1-1c, agg1-2, agg2-1, agg-2-2, and agg1 agg2 grown under "standard" conditions as specified in "Materials and Methods." Quantitative analysis of a number of morphological characteristics was carried out at defined growth stages (Boyes et al., 2001) during development (Table I ; Fig. 4 ). Mean values of these traits in the mutant lines were analyzed for significant deviation from the corresponding values in the wild type by pair-wise, two-sample Student's t test. Unless stated otherwise, the mean values were derived from analysis of at least 30 plants grown in a checkerboard pattern.

View larger version (89K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 4. Comparison of characteristic morphological traits in the wild type and gpa1-4, agb1-2, and G mutants. A, Seven-day-old soil-grown seedlings. B, Average rosette leaves from a 30-d-old plant. C, Fully open mature flowers. D, Fully expanded siliques. [See online article for color version of this figure.]

Inflorescence emergence, defined by the appearance of flower buds, occurs approximately 2 d earlier in agg1-1c, agg1-2, and double agg1 agg2 mutants than in the wild type and gpa1-4, agg2-1, and agg2-2 mutants. The agb1-2 mutants showed delayed flowering, with inflorescence emergence occurring 2 d later than in the wild type (Table I). Despite the delay in inflorescence emergence (P < 0.05), the first agb1-2 flower opened at the same time as for the wild type, gpa1-4, and G2-deficient mutants (data not shown). The final inflorescence height was noticeably lower in agb1-2 (P < 0.01), while all other mutants were similar in height to the wild type. Apical dominance is known to be regulated by basipetal auxin flow from the apical meristem (Jones, 1998; Dun et al., 2006; Leyser, 2006). Attenuation of auxin signaling by Gβ and both G subunits has been established previously in roots (Ullah et al., 2003; Trusov et al., 2007). In the floral stem, our data showed increased apical dominance in agb1 mutants, as evidenced by a decreased number of branches, which is consistent with previous reports (Ullah et al., 2003). In contrast, all of the G-deficient mutants showed a wild-type floral stem branching pattern. Contrary to a previous report (Ullah et al., 2003), we found that the number of open flowers at the midflowering stage was higher in agb1-2 plants (P < 0.01) compared with all other genotypes analyzed (Table I). Flowers were significantly smaller in agb1-2 and gpa1-4 mutants, while all G-deficient mutants had flowers of similar size compared with the wild type (Table I; Fig. 4C). Even though we measured only the diameter of the fully open flower, sepal size was also affected in agb1-2 and gpa1-4 mutants, as can be seen in Figure 4C.

Alteration of the silique shape was first described for the agb1-1 mutant (Lease et al., 2001), and similar observations were made for agb1-2 (Ullah et al., 2003). Our measurements of agb1-2 concur with previous reports showing shorter and wider siliques (P < 0.001), with the characteristic blunt (flat) tips (Table I; Fig. 4D). Similarly to agb1-2, the gpa1-4 mutants produced shorter (although to a lesser extent) and wider siliques, with the characteristic blunt tip. Curiously, these results differ from those reported by Ullah et al. (2003), who analyzed two different G-deficient mutants, gpa1-1 and gpa1-2, and found their siliques to be slightly longer than wild-type siliques and having a wild-type tip shape. It is noteworthy that gpa1-1 and gpa1-2 mutants were obtained in the Wassilewskija (Ws) ecotype, while gpa1-4 (as well as another G-deficient mutant, gpa1-3) is in the Col-0 background. To analyze the effects of growth conditions or ecotype in silique development, we simultaneously grew gpa1-1, gpa1-2, gpa1-3, and gpa1-4 mutants and corresponding wild-type plants. Interestingly, siliques of gpa1-1 and gpa1-2 were indeed slightly longer than those of the Ws wild type, with "normal" acute tips, as described by Ullah et al. (2003), while both gpa1-3 and gpa1-4 siliques were shorter than those in the Col-0 wild type, and both had blunt tips (Table II ). This observation is quite important, as it demonstrates that some effects of GPA1 deficiency are dependent on the genetic background and are not necessarily universal, even within the same species. Siliques in all G-deficient mutants showed wild-type characteristics.

Guard Cells of G-Deficient Mutants Show Wild-Type Responses to ABA

ABA is a well-studied phytohormone in guard cell signaling and is an important component of many stress responses in plants. In Arabidopsis, G-deficient mutants show alterations in a number of guard cell responses, such as hyposensitivity to ABA inhibition of stomatal opening and reduced ABA responsiveness of guard cell inward K⁺ channels (Wang et al., 2001; Coursol et al., 2003; Mishra et al., 2006). Our recent studies (L.M. Fan, unpublished data) show that Gβ-deficient mutants show the same alterations in guard cell ABA responses as observed for G-deficient mutants. Therefore, we assessed both ABA inhibition of light-induced stomatal opening and the ABA promotion of stomatal closure in G-deficient plants. In agg1-1c, agg2-1, and agg2-2 single mutants, stomatal responses to 50 µM ABA were not statistically different from those of the wild type. A wild-type ABA response was also observed in the double agg1 agg2 mutant (Fig. 5, A and B ). To ensure that a subtle alteration in ABA sensitivity was not overlooked in these experiments, the assays were repeated with a lower concentration of ABA (20 µM). As shown in Figure 5, C and D, stomatal aperture responses of the G-deficient single and double mutants at this lower ABA concentration still remained indistinguishable from those of wild-type plants.

View larger version (36K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 5. ABA regulates stomatal movements similarly in Col and agg1-1c, agg2-1, agg2-2, and agg1 agg2. A, ABA (50 µM) inhibition of light-induced stomatal opening in Col, agg1-1c, agg2-1, agg2-2, and agg1 agg2. Data shown are means ± SE from three replicates with n > 150 stomata for each experiment. B, ABA (50 µM) induction of stomatal closure in Col, agg1-1c, agg2-1, agg2-2, and agg1 agg2. Data shown are means ± SE from three replicates with n > 150 stomata for each experiment. C, ABA (20 µM) inhibition of light-induced stomatal opening in Col, agg1-1c, agg2-1, agg2-2, and agg1 agg2. Data shown are means ± SE from three replicates with n > 150 stomata for each experiment. D, ABA (20 µM) induction of stomatal closure in Col, agg1-1c, agg2-1, agg2-2, and agg1 agg2. Data shown are means ± SE from three replicates with n > 150 stomata for each experiment.

View larger version (29K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 6. ABA regulates K+ currents similarly in guard cells of Col, agg1-1c, agg2-1, agg2-2, and agg1 agg2. A, Typical whole cell recordings of guard cell K+ currents with or without 50 µM ABA. Time and voltage scales shown in the top right panel apply to all panels. B, Current/voltage relationship (mean ± SE) of time-activated whole cell K+ currents as illustrated in A. Number of guard cells was as follows: Col (11), Col + ABA (13), agg1-1c (10), agg1-1c + ABA (10), agg2-1 (eight), agg2-1 + ABA (16), agg2-2 (six), agg2-2 + ABA (seven), agg1 agg2 double mutant (16), and agg1 agg2 double mutant + ABA (17).

	DISCUSSION

TOP ABSTRACT RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

β

β

Here, we show that there are considerable discrepancies between the behavior of Gβ- and G-deficient mutants, including the double agg1 agg2 mutant. AGB1 gene expression patterns in reproductive organs do not match AGG1, AGG2, or their combination. Hypocotyl elongation, both in darkness and under light, as well as hook development are altered in the Gβ-deficient mutant but not in any of the single G-deficient mutants or in the double agg1 agg2 mutant. Furthermore, agb1-2 was hypersensitive to ABA during germination, while all G-deficient mutants displayed wild-type sensitivity. Striking discrepancies were observed in the morphological and developmental phenotypic characterization of the mutants. It is remarkable that none of the aerial morphological traits for which the Gβ-deficient mutants show statistically significant differences from the wild type can be observed in the different G-deficient mutants. The only instance in which G1-deficient mutants, and the double agg1 agg2 mutant, show statistically significant differences from the wild-type controls is in flowering time; however, this effect was opposite to that observed in the Gβ-deficient mutant (Table I). Finally, in gpa1 (Wang et al., 2001; Coursol et al., 2003; Mishra et al., 2006) and agb1 (L.M. Fan, unpublished data), stomatal opening (but not stomatal closure) exhibits hyposensitivity to inhibition by ABA. In contrast, no ABA hyposensitivity of either ABA inhibition of stomatal opening or ABA promotion of stomatal closure was observed in any of the single G-deficient mutants or the double agg1 agg2 mutant. Stomatal aperture responses result from a complex web of cellular signaling events that regulate multiple effectors (Li et al., 2006). Therefore, we also assessed one particularly well-defined subcellular event: ABA inhibition of inward K⁺ channels in guard cells (Blatt, 1990; Schwartz et al., 1994). In G- and Gβ-deficient lines, the response of inward K⁺ currents to ABA is abrogated (Wang et al., 2001; Coursol et al., 2003; L.M. Fan, unpublished data). However, in all of the G-deficient mutants, guard cell K⁺ channel regulation by ABA was present at the same magnitude as in the wild type. Taken as a whole, our results demonstrate that the absence of the G subunits does not completely phenocopy the lack of the Gβ subunit.

To explain the observed discrepancies, two not necessarily exclusive hypotheses could be proposed. The first is that additional G subunits exist in Arabidopsis, which could possibly be expressed in reproductive organs to explain the expression discrepancies observed between AGB1 and the two known G subunit genes, AGG1 and AGG2. Although exhaustive search of the fully sequenced Arabidopsis genome did not reveal any additional canonical G subunits, these subunits display poor sequence conservation; therefore, the presence of atypical subunits cannot be discarded. This hypothesis allows retention of the heterotrimeric G protein dogma, developed mainly for mammalian systems, which states that all subunits are interdependent and required for proper signaling of the heterotrimer (Gilman, 1987).

A second hypothesis predicts some functional autonomy of the G protein subunits in plants, consistent with several recent observations. In animals, interdependence of G subunits and the corresponding Gβ dimers for correct subcellular localization has been established (Takida and Wedegaertner, 2003), while in plants, the subunits do not depend on each other for plasma membrane targeting to the same extent (Adjobo-Hermans et al., 2006; Zeng et al., 2007; Wang et al., 2008). Moreover, both Arabidopsis G subunits localized to the plasma membrane in G- and Gβ-deficient mutants (Zeng et al., 2007). Other unusual properties of the plant G protein complex recently led Temple and Jones (2007) to suggest that the plant heterotrimer is lagging behind in evolutionary terms from its animal counterparts. This antiquity could allow some functional autonomy for the individual subunits, since it is logical to propose that the heterotrimer most probably originated from three initially independent proteins. It is possible, therefore, that in some processes, plant G and Gβ subunits might act independently from each other and from G subunits. Notably, some G-dependent traits, such as susceptibility to necrotrophic fungi, methyl jasmonate sensitivity during root elongation and seed germination, and lateral root number, were altered in an opposite way in gpa1-4 compared with agb1-2 (Trusov et al., 2006, 2007). At the same time, many traits reported here, including ABA sensitivity in seed germination and stomatal regulation, floral organ shape, and hypocotyl elongation in darkness, which were similarly altered in gpa1-4 and agb1-2 mutants, were not altered in G mutants. This could imply that G and Gβ, but not always G, are part of an as yet uncharacterized complex, which governs those traits. Interestingly, in rice, all individual heterotrimeric G protein subunits were found as a part of 400-kD multiprotein complexes, but a fraction of the Gβ dimers were also found not to be associated with those complexes (Kato et al., 2004). In Arabidopsis, membrane-associated 400-kD protein complexes were reported for the ERECTA receptor-like kinase (Shpak et al., 2003), and knockouts of ERECTA and ERECTA-like genes display similar phenotypes to G- and/or Gβ-deficient mutants in a number of traits (Lease et al., 2001; Shpak et al., 2004) as well as increased susceptibility to necrotrophic pathogens (Llorente et al., 2005), suggesting that G proteins could be an integral part of those complexes. Moreover, during the preparation of this article, it was reported that in Arabidopsis roughly 30% of the native G and all of the overexpressed cyan fluorescent protein (CFP)-Gβ are associated with large (approximately 700 kD) multiprotein complexes found in the plasma membrane fraction (Wang et al., 2008). Unfortunately, no information on G-containing complexes was provided for either of the G subunits (Wang et al., 2008).

The question arises whether plant Gβ can act independently of G, as has already been described for Gβ5 in animal systems. The mammalian Gβ5 subunit can form functional dimers with a number of RGS proteins that contain a "G-like motif" known as the GGL domain, precluding formation of the canonical dimer Gβ5 (Snow et al., 1998; Witherow et al., 2000; Sondek and Siderovski, 2001; Witherow and Slepak, 2003; Willars, 2006). There are no known GGL domain-containing plant proteins; reports about plant GGL domain proteins, however, are currently available. In addition, the WD40 propeller structure of the Gβ subunit allows interaction with multiple proteins, hence providing a scaffold for large protein assemblies. In Arabidopsis and tobacco (Nicotiana tabacum), the Gβ subunit was identified in plasma membrane Triton X-100-insoluble microdomains along with other signaling components, including kinases and small GTP-binding proteins (Peskan and Oelmuller, 2000; Peskan et al., 2000; Shahollari et al., 2004). Thus, it is not unlikely that Gβ could function independently from G as an integral part of multiprotein complexes to mediate some processes while functioning as a Gβ dimer in others. In this scenario, knockout of AGB1 would destroy both the Gβ dimers and the hypothetical Gβ-containing complexes, while the absence of G subunits would only obliterate signaling that was directly dependent on Gβ dimers.

In both hypotheses, it is necessary to explain how Gβ can be targeted to the plasma membrane in the absence of both G subunits. A hybrid CFP-AGB1 fusion protein showed diffuse localization in the cytoplasm unless coexpressed with either AGG1 or AGG2 subunits (Adjobo-Hermans et al., 2006). This study demonstrated sufficiency of the G subunits for the plasma membrane localization of Gβ. However, since this study was based on high ectopic expression of the AGB1, AGG1, and AGG2 genes, it is possible that other proteins, or complexes, can also anchor Gβ to the plasma membrane under normal circumstances. In this respect, it will be interesting to determine whether AGB1 is membrane localized in the agg1 agg2 double mutant. Just as for all other known G subunits, the ability of AGG1 and AGG2 to be targeted to the plasma membrane crucially relies on prenylation of the C-terminal CAAX motif (Zeng et al., 2007). In Arabidopsis, two prenylation enzymes, geranylgeranyltransferase I (PGGT-I) and farnesyltransferase (PFT), have been identified (Caldelari et al., 2001; Running et al., 2004). Interestingly, knockout of PFT results in ABA hypersensitivity and phenotype alterations similar to those observed in G- and Gβ-deficient mutants, while knockout of PGGT-I, which has been shown to prenylate both AGG1 and AGG2 subunits (Zeng et al., 2007), leads to wild-type phenotype and wild-type ABA sensitivity in seed germination (Johnson et al., 2005; Zeng et al., 2007). Taken together with the fact that knockout of both G subunits does not completely phenocopy Gβ-deficient mutant phenotypes, these observations suggest that there are additional, probably farnesylated, proteins interacting with AGB1 in Arabidopsis. It is possible, therefore, that currently unknown G subunits, GGL-containing RGSs, or some components of the multiprotein complexes described by Wang et al. (2008) can be prenylated by PFT and are able to target AGB1 to the plasma membrane. It is interesting that in rice, the G subunit RGG2 lacks the C-terminal prenylation motif but nevertheless is detected in the plasma membrane fraction associated with Gβ (Kato et al., 2004).

	CONCLUSION

TOP ABSTRACT RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

 
Additional studies will prove or disprove the feasibility of the hypotheses described above. Nevertheless, whatever the result of those studies, our data unambiguously reveal that the variety of heterotrimeric G proteins in plants is not limited to the two canonical heterotrimers Gβ1 and Gβ2 found so far. This adds an additional level of complexity to the molecular mechanisms used by G proteins in plants and provides a new degree of functional selectivity to that reported previously for the two known heterotrimers (Trusov et al., 2007).

	MATERIALS AND METHODS

TOP ABSTRACT RESULTS DISCUSSION CONCLUSION MATERIALS AND METHODS LITERATURE CITED

 

Plant Material

The Arabidopsis (Arabidopsis thaliana) agg1-1c mutant allele of AGG1 (At3g63420), the agg2-1 mutant allele of AGG2 (At3g22942), the double mutant agg1 agg2, and the agb1-2 mutant were described previously (Ullah et al., 2003; Trusov et al., 2007). New alleles agg1-2 and agg2-2 (both in the Col-0 background) were produced as T-DNA mutants by GBI-Kat (Rosso et al., 2003). agg1-2 seeds were obtained from GBI-Kat (accession no. 736A08), while agg2-2 lines were obtained from the Nottingham Arabidopsis Science Centre (accession no. N375172). For each new line, homozygous plants were selected using a three-primer PCR approach. The exact position of the T-DNA insertion was determined by amplifying and sequencing a genomic DNA fragment between the T-DNA end and the 3' end of the gene in the chromosome. Absence of full-length AGG2 mRNA for agg2-2 was confirmed by reverse transcription (RT)-PCR (data not shown). In agg1-2, the insert is located in the promoter region, and full-size AGG1 mRNA was detected in the mutant. However, northern analysis revealed significant down-regulation in AGG1 mRNA levels in mutant plants (roughly 30% of the wild-type level; Supplemental Fig. S1).

Mutant Characterization

Mature plants were grown under a long-day (16 h of light/8 h of dark) and 23°C regimen for 6 weeks, and an additional 2 weeks were allowed for seed maturation. Morphological characteristics were measured at defined developmental stages as described elsewhere (Boyes et al., 2001). For destructive analysis (flower, silique, and leaf shape traits), plants were removed from the rest of the population randomly, dissected, and photographed if necessary.

Seed and Seedling Assays

All plates contained 0.5x Murashige and Skoog basal salts (PhytoTechnology Laboratories) and 0.8% agar. Stock solutions of ABA at the designated concentrations were added to autoclaved medium cooled to approximately 55°C. Since germination is extremely sensitive to the growth conditions experienced by the parental plant and to postharvest storage, all seed lots for seed and seedling assays were collected at the same time from plants grown simultaneously under the same conditions. The seeds were stored at 4°C in the dark. Seeds were dry sterilized by 3 h of incubation in a chamber filled with chlorine gas. Approximately 150 sterilized seeds of all tested lines were planted on the same petri dish with a designated treatment. After sowing, all seeds were stratified for at least 48 h at 4°C in darkness. Germination was defined as an obvious protrusion of the radicle.

For hypocotyl elongation assays, seeds were induced under continuous light (150 µmol m^–2 s^–1) for 24 h, then seedlings were grown on vertical plates for 1 to 3 d in a dark cabinet. The plates were photographed and hypocotyl length was measured.

Isolation of RNA and Transcript Analysis

Total RNA for northern analysis and RT-PCR was extracted as described previously (Trusov and Botella, 2006; Purnell and Botella, 2007). Probes for northern blots were labeled using a Rediprime II P32 radiolabeling kit (Amersham). Membranes were hybridized overnight in Church buffer (Church and Gilbert, 1984) at 65°C, washed twice in 0.1% SSC and 0.1% SDS solution as described (Petsch et al., 2005), and exposed to PhosphorImager plates for analysis (Molecular Dynamics). For RT-PCR, reverse transcriptions were carried out using the SuperScript III RT kit according to the manufacturer's instructions (Invitrogen) as described previously (Moyle et al., 2005). PCR amplifications were performed using GoTaq Green Master mix (Promega) in 35 cycles with the following parameters: 94°C for 30 s, 54°C for 30 s, and 72°C for 1 min. The primers used for the AGG1 and AGG2 genes were described previously (Trusov et al., 2007).

Stomatal Aperture Experiments and Guard Cell Electrophysiology

Arabidopsis plants were grown in soil mix (Potting Mix; Miracle-Gro) in growth chambers with 8-/16-h light/dark and 22°C/20°C cycles. Fully expanded young leaves from 4-week-old plants were used for both stomatal aperture assays and guard cell protoplast isolation. All of the protocols for stomatal aperture assays and whole cell K⁺ current recordings from guard cells with and without ABA treatment were as described for previous analyses of G-deficient mutant plants (Wang et al., 2001; Coursol et al., 2003). For whole cell K⁺ current analysis, recordings obtained at 10 min after formation of the whole cell configuration were used and were analyzed as described (Coursol et al., 2003). Whole cell capacitances were used to normalize current amplitude (pA/pF) to avoid the influence of cell size. Data were compared with Student's t test, and results with P < 0.01 were considered significantly different.

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure S1. Down-regulation of AGG1 gene expression in agg1-2 T-DNA mutants.

	ACKNOWLEDGMENTS

 
We thank Dr. Mike Mason and Dr. David Chakravorty for critical reading of the manuscript.

Received February 13, 2008; accepted April 22, 2008; published April 25, 2008.

	FOOTNOTES

 
¹ This work was supported by the Australian Research Council (Discovery Grant nos. DP0344924 and DP0772145), the U.S. Department of Agriculture (grant no. 2006–35100–17254), and the National Science Foundation (grant no. MCB–0209694).

The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy described in the Instructions for Authors (www.plantphysiol.org) is: José Ramón Botella (j.botella{at}uq.edu.au).

^[C] Some figures in this article are displayed in color online but in black and white in the print edition.