OPEN ACCESS ARTICLE

This Article Free via Open Access: OA OA Abstract Full Text (PDF) OA All Versions of this Article: 19/4/1235    most recent tpc.107.050096v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in Plant Cell Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (12) Citing Articles via Google Scholar Google Scholar Articles by Trusov, Y. Articles by Botella, J. R. Search for Related Content PubMed PubMed Citation Articles by Trusov, Y. Articles by Botella, J. R. Agricola Articles by Trusov, Y. Articles by Botella, J. R.

Heterotrimeric G Protein Subunits Provide Functional Selectivity in Gß Dimer Signaling in Arabidopsis^[OA]

^a Plant Genetic Engineering Laboratory, Department of Botany, School of Integrative Biology, University of Queensland, Brisbane, Queensland 4072, Australia
^b Departments of Biology and Pharmacology, University of North Carolina, Chapel Hill, North Carolina 27599-3280

² To whom correspondence should be addressed. E-mail j.botella{at}uq.edu.au; fax 61-7-33651699.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
The Arabidopsis thaliana heterotrimeric G protein complex is encoded by single canonical G and Gß subunit genes and two G subunit genes (AGG1 and AGG2), raising the possibility that the two potential G protein complexes mediate different cellular processes. Mutants with reduced expression of one or both G genes revealed specialized roles for each G subunit. AGG1-deficient mutants, but not AGG2-deficient mutants, showed impaired resistance against necrotrophic pathogens, reduced induction of the plant defensin gene PDF1.2, and decreased sensitivity to methyl jasmonate. By contrast, both AGG1- and AGG2-deficient mutants were hypersensitive to auxin-mediated induction of lateral roots, suggesting that Gß1 and Gß2 synergistically inhibit auxin-dependent lateral root initiation. However, the involvement of each G subunit in this root response differs, with Gß1 acting within the central cylinder, attenuating acropetally transported auxin signaling, while Gß2 affects the action of basipetal auxin and graviresponsiveness within the epidermis and/or cortex. This selectivity also operates in the hypocotyl. Selectivity in Gß signaling was also found in other known AGB1-mediated pathways. agg1 mutants were hypersensitive to glucose and the osmotic agent mannitol during seed germination, while agg2 mutants were only affected by glucose. We show that both G subunits form functional Gß dimers and that each provides functional selectivity to the plant heterotrimeric G proteins, revealing a mechanism underlying the complexity of G protein–mediated signaling in plants.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Heterotrimeric G proteins are an important element of transmembrane signal transduction, coupling stimuli as diverse as light, neurotransmitters, odorants, tastants, and hormones. They are found in a variety of eukaryotic organisms, including plants, fungi, and animals. The classical heterotrimer consist of three different subunits, , ß, and , which are organized in a highly conserved structure and typically bound to specific G protein–coupled receptors. Activation of the receptor by ligand binding induces a conformational change in G, catalyzing the exchange of GDP to GTP. GTP loading causes a protein conformational change that promotes dissociation of the heterotrimer into two functional signaling elements: the G subunit and the Gß dimer. These two elements (functional subunits) interact with specific effector molecules controlling downstream signaling. The inherent GTPase activity of the G subunit hydrolyzes its bound GTP, leading to the reassociation of G and the Gß dimer, returning the heterotrimer to its inactive GDP-bound state. While interaction between G and the Gß dimer is dependent on the conformational status of the G subunit, interaction between Gß and G is essentially nondissociable; therefore, the Gß dimer acts as a single functional unit in the cell (Gautam et al., 1998).

It was initially thought that signaling in animals only occurred via the activated G subunit, with the role of Gß being to inhibit the action of G by reforming the inactive heterotrimer and guiding G back to the receptor for reactivation. However, it is now established that the Gß dimer is an active signaling factor in at least as many processes as the G subunit (Clapham and Neer, 1997). Among others, the Gß dimer is able to interact with adenylyl cyclases, potassium channels, and phospholipases (Clapham and Neer, 1993; Scott et al., 2001). Aside from the activation of specific downstream effectors, the Gß dimer is involved in receptor recognition (Lim et al., 2001), membrane targeting, and activation of the G subunit (Evanko et al., 2000, 2001). Binding between G and Gß occurs at a molecular interface largely contained within the ß-propeller structure of Gß. With the exception of Gß5, there is little binding preference between G and Gß pairs. Therefore, it is assumed that G provides a major share of the structural requisite for the selective coupling of the heterotrimer to the receptor and the Gß dimer to its effectors (Gautam et al., 1990; Simon et al., 1991; Hou et al., 2000; Myung and Garrison, 2000; Azpiazu and Gautam, 2002; Chen et al., 2005; Myung et al., 2006). Recent evidence indicates that some animal Gß dimers can move from the plasma membrane to the Golgi upon receptor activation, providing an extra element of spatial segregation to the Gß dimer in G protein–mediated signaling. The G subunit type and the G subunit nucleotide exchange properties strongly influence the rate of translocation (Akgoz et al., 2004, 2006; Azpiazu et al., 2006).

A characteristic of mammalian systems is the existence of gene families for each of the G protein subunits. At least 23 G subunits, 6 Gß subunits (including an alternatively spliced variant), and 12 G subunits (Gautam et al., 1998; Balcueva et al., 2000) have been reported in humans, but not all possible combinations are present in the cell, with combinatorial multiplicity of Gß dimers being restricted by the specific expression patterns of the genes and selective interactions between different Gß and G subunits. Nevertheless, a wide range of Gß dimers, serving as distinct signal transduction elements involved in different processes, have been described (Camps et al., 1992; Katz et al., 1992; Chen et al., 1997; Clapham and Neer, 1997; Gautam et al., 1998; Bommakanti et al., 2000; Mirshahi et al., 2002; Krystofova and Borkovich, 2005).

In contrast with mammalian systems, only one canonical G subunit gene (GPA1) (Ma et al., 1990), one canonical Gß subunit gene (AGB1) (Weiss et al., 1994), and two G subunit genes (AGG1 and AGG2) (Mason and Botella, 2000, 2001) have been found in the Arabidopsis thaliana genome. The same number of G protein subunits were reported in the monocot species rice (Oryza sativa) (Ishikawa et al., 1995, 1996; Iwasaki et al., 1997; Kato et al., 2004); however, two G subunits were described for legume species (Kim et al., 1995; Gotor et al., 1996; Marsh and Kaufman, 1999). G proteins are implicated in a large variety of processes in plants (Jones and Assmann, 2004; Perfus-Barbeoch et al., 2004; Assmann, 2005; McCudden et al., 2005; Temple and Jones, 2007); nevertheless, specific signaling roles for the G subunit or Gß dimers remained elusive until recently. Analysis of T-DNA and ethyl methanesulfonate mutants lacking functional G or Gß subunits showed that both G and Gß could be involved in specific and independent pathways (Ullah et al., 2003; Joo et al., 2005; Chen et al., 2006a; Pandey et al., 2006; Trusov et al., 2006) as well as in the same processes (Ullah et al., 2003; Pandey et al., 2006). Studies using Arabidopsis demonstrated that the Gß-deficient agb1-1 and agb1-2 mutants have flowers with elongated peduncles, shortened flat-top siliques, rounded rosette leaves with crinkled surfaces, and increased root mass (Lease et al., 2001; Ullah et al., 2003). Detailed studies revealed that Gß modulates lateral root formation by interfering with auxin-dependent cell division (Ullah et al., 2003). It was shown that Gß-mediated signaling, but not G, plays a distinct part in plant resistance against necrotrophic pathogens (Llorente et al., 2005; Trusov et al., 2006). Specific changes in seed germination were also ascribed to Gß activity (Pandey et al., 2006; Trusov et al., 2006). Finally, analysis of transgenic tobacco (Nicotiana tabacum) plants with reduced Gß subunit levels due to antisense expression of the Gß subunit mRNA suggested that the Gß subunit is involved in regulation of the reproductive phase of the tobacco life cycle, particularly in stamen development and pollen maturation (Peskan-Berghofer et al., 2005).

Strong interaction between plant Gß and each G subunit was demonstrated in vitro (Mason and Botella, 2000, 2001) as well as in vivo (Kato et al., 2004; Adjobo-Hermans et al., 2006, Chakravorty and Botella, 2007). However, despite sequence similarity (48% amino acid identity), the interaction between each of the two Arabidopsis G subunits and Gß seems to be centered in different domains of the protein (Mason and Botella, 2000, 2001; Temple and Jones, 2007).

Nothing is known about the cellular and physiological roles of either of the two known G subunits, their possible functional redundancy, and whether the two potential dimers, Gß1 and Gß2, are involved in the same or different signaling pathways. We took advantage of the extensive phenotypic characterization of loss-of-function agb1 mutants, and using this inventory of phenotypes, we asked which of the G subunits acts with Gß to regulate a specific function. Fungal resistance, root development, and glucose sensing were the three well-characterized AGB1-signaling pathways examined in this study. By a genetic approach, we dissected the roles of the G subunits in G protein signaling in these pathways. Our results show that the different G subunits form independent signal-transducing Gß dimers and impart functional selectivity to the heterotrimeric G protein signaling network.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
The Expression Profiles of AGG1 and AGG2 Are Distinct but Together Overlap AGB1 Expression
Expression patterns for G and Gß subunit genes were previously reported in various plant species (Weiss et al., 1993; Huang et al., 1994; Kaydamov et al., 2000; Perroud et al., 2000; Chen et al., 2006c). In order to study the tissue-specific and developmental regulation of the AGB1, AGG1, and AGG2 genes, transgenic Arabidopsis (Col-0) plants were produced containing the promoter regions of each gene fused to the ß-glucuronidase (GUS) reporter gene. At least three independent lines were characterized for each of the promoter constructs. Transgenic plants did not show any obvious morphological alterations, suggesting that inserts did not disrupt functional genes. GUS histochemical assays revealed that all three genes are active during early seedling development, with GUS activity detected throughout the plant but highest at the hypocotyl–root junction in 2-d-old AGB1:GUS seedlings (Figure 1A ). AGG1:GUS staining was observed in the hypocotyl, while AGG2:GUS staining occurred in the upper part of the root, including root hairs, and gradually declined along the root (Figure 1A).

View larger version (88K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 1. In Situ AGB1, AGG1, and AGG2 Expression Patterns. Histochemical analysis of GUS expression in transgenic Arabidopsis plants carrying AGB1, AGG1, or AGG2 promoter:GUS fusions as indicated. (A) Shoot–root junction of 2-d-old, dark-grown seedlings. (B) Two-week-old light-grown seedlings. (C) Higher magnification of 2-week-old true leaves. (D) Four-week-old roots. (E) Cross section through 4-week-old roots.

View larger version (138K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 3. The G Subunit Is Involved in Defense against Necrotrophic Fungi. (A) Induction of GUS activity by A. brassicicola in leaves of transgenic plants expressing the designated promoter:GUS fusion constructs. Arrows indicate the region of infection. (B) Fluorometric assessment of GUS activity in leaves and roots of transgenic plants expressing the designated promoter:GUS fusion constructs after inoculation of roots with F. oxysporum. The bars represent expression ratios of pathogen-inoculated versus mock-inoculated plants. Error bars represent SE of three replicates. (C) Characteristic disease symptoms caused by F. oxysporum at 8 d after inoculation. (D) Lesion development at 3 d after inoculation with A. brassicicola.

Loss-of-Function Mutants for the G Subunits
In order to study the function of both G subunits in Arabidopsis, mutants carrying T-DNA insertions in AGG1 (agg1-1w, on the Wassilewskija [Ws] ecotype) and AGG2 (agg2-1, on the Columbia-0 [Col-0] background) genes were identified. An AGG1-deficient mutant in the Col-0 background was generated by genetic introgression over eight successive generations, resulting in a line designated agg1-1c (backcross to Col-0). In agg1-1w, the T-DNA insertion is positioned within the second intron, splitting the protein in approximately two equal halves, while in agg2-1, two tandem and opposing T-DNA insertions are located in the third intron, disrupting the C-terminal region of the hypothetical protein (Figure 2A ). RT-PCR analysis showed that neither allele (agg1-1c or agg2-1) produces a detectable functional transcript for its respective gene (Figure 2C). In addition, the absence of AGG1 expression in the agg1-1c mutants did not result in any observable changes in AGG2 expression, due to possible compensatory effects (Figure 2C; data not shown). The reverse applies to agg2-1 mutants. A double knockout of the AGG1 and AGG2 genes was obtained by hybridization of the agg1-1c and agg2-1 mutants (agg1 agg2). As expected, this line lacked detectable expression of each of the two G subunit genes (Figure 2C).

View larger version (39K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 2. Molecular Characterization of agg1-1, agg2-1, and RNAi Mutants. (A) T-DNA insertion sites in the agg1-1 and agg2-1 mutants. Gray boxes represent exons. Arrows show the positions of forward and reverse primers used for PCR and RT-PCR. The T-DNA insert is not in scale. ATG and TGA, start and stop codons, respectively; LB, T-DNA left border; RB, T-DNA right border. (B) RNAi construct used in the production of the agg1RNAi lines. A similar construct was generated with the AGG2 cDNA for the agg2RNAi lines. (C) RT-PCR analysis of the AGG1 and AGG2 transcripts. Total RNA extracted from 1-week-old seedlings was used for cDNA synthesis as a template for PCR. Forward and reverse primers depicted in (A) were used to perform PCR. Arabidopsis ACTIN2 was used as a control.

Gß-Mediated Defense against Necrotrophic Fungi Is Selectively Mediated by AGG1 but Not AGG2
It was shown previously that Gß-mediated signaling, but not G-mediated signaling, is involved in resistance against necrotrophic fungi (Llorente et al., 2005; Trusov et al., 2006). Therefore, we sought to determine whether there is a specific G subunit engaged with Gß in this process or whether both subunits play redundant or synergistic roles. In preliminary experiments, we analyzed the behavior of all three genes (AGB1, AGG1, and AGG2) in response to attack by necrotrophic pathogens using transgenic plants carrying the promoter:GUS fusion constructs. Alternaria brassicicola is an air-borne avirulent pathogenic fungus of Arabidopsis ecotype Columbia (Penninckx et al., 1996; Schenk et al., 2000, 2003; Thomma et al., 2000; van Wees et al., 2003), even though some isolates can reproduce at a very low rate under favorable conditions (van Wees et al., 2003). When plants were inoculated with a suspension of A. brassicicola spores, elevated GUS activity was detected 24 h after infection in AGB1:GUS and AGG1:GUS but not in AGG2:GUS transgenic plants (Figure 3A ). GUS staining was restricted to the inoculation site and did not spread throughout the entire leaf.

Fusarium oxysporum (f. sp conglutinans) is a soil-borne necrotrophic fungus that uses the root tip, secondary root formation foci, and wounds as entry points. It subsequently colonizes the plant by traveling through the vascular system (Mauchmani and Slusarenko, 1994; Agrios, 2005). In contrast with A. brassicicola, F. oxysporum is a virulent pathogen of Arabidopsis (Berrocal-Lobo and Molina, 2004). Surprisingly, inoculation of roots with F. oxysporum did not induce GUS activity in root tissue above background levels in any of the three reporter lines; however, significant induction was detected in leaves of AGB1:GUS and AGG1:GUS plants (Figure 3B). No induction was observed in AGG2:GUS plants; rather, a slight decrease in gene expression was observed in leaves and roots. Taken together, our findings indicate that leaf expression of AGB1 and AGG1 is systemically activated by F. oxysporum and locally by A. brassicicola.

To understand the roles of G1 and G2 in resistance against necrotrophic pathogens, we assayed the response of the T-DNA mutants agb1-2, agg1-1c, agg2-1, agg1 agg2, and the RNAi lines (agg1RNAi and agg2RNAi) to A. brassicicola and F. oxysporum inoculation. Roots of 2-week-old mutant and wild-type plants were infected with bud cell suspensions of F. oxysporum, and disease progression was monitored over time from the development of the first symptoms until plants died. Figure 3C illustrates the appearance of typical disease symptoms at an early stage of infection. The advanced chlorosis observed in veins and leaves of agb1-2, agg1-1c, agg1RNAi, and agg1 agg2 mutants gives a qualitative indication that there is increased susceptibility to F. oxysporum in these lines compared with the wild type as well as agg2-1 and agg2RNAi mutants. To quantify the levels of resistance, the number of decayed plants in all mutant lines and wild-type controls was determined (Figure 4A ). Plants lacking green leaves were considered decayed. agg1-1c, agg1RNAi, and agg1 agg2 lines showed similar dynamics to agb1-2, all of them exhibiting a faster rate of disease progression than wild-type plants, while the behavior of agg2-1 and agg2RNAi mutants resembled that of the wild type. To test whether the loss of AGG1 had a similar effect in the Ws background, we compared the agg1-1w mutant (in Ws) with wild-type Ws and the G subunit null mutant gpa1-1 (also in the Ws ecotype) (Ullah et al., 2001). Unfortunately, no agb1 mutants are yet available in the Ws background. We previously showed that the G subunit null mutants gpa1-3 and gpa1-4 (Col-0 ecotype) have slightly enhanced resistance to F. oxysporum (Trusov et al., 2006). After performing inoculation and disease evaluation as for Col-0 lines, it was evident that disease progressed faster in agg1-1w plants than in the Ws wild type, while gpa1-1, as expected, displayed slightly enhanced resistance (Figure 4C). The differences in disease progression observed between agg1, agb1, and agg1 agg2 mutants compared with wild-type Col-0 and agg2 mutants were statistically significant (P < 0.05). Similarly, the differences observed between agg1-1w and the wild type and between gpa1-1 and the wild type in the Ws ecotype were statistically significant (P < 0.01). All experiments were repeated at least twice with similar results.

View larger version (51K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 4. Differential Responses of G-Deficient Mutants to Pathogen Attack and MeJA Treatment. (A) Susceptibility of wild-type plants (Col-0) and Gß- and G-deficient mutants to F. oxysporum. For each genotype, 48 plants were inoculated and the average percentage of decayed plants per line was scored in three independent experiments. Error bars represent SE. (B) Inhibition of rosette growth after F. oxysporum inoculation expressed relative to the mean growth of the same genotype after mock inoculation. Mean values and corresponding SD were calculated from 48 inoculated and 24 mock-inoculated plants for each genotype. (C) Same as (A) for wild-type Ws ecotype and gpa1-1 and agg1-1w mutants. (D) Expression of the defense-related gene PDF1.2 in response to A. brassicicola infection. Two-week-old wild-type and mutant plants were sprayed with an A. brassicicola spore solution (106 spores/mL). Total RNA was extracted from infected leaf tissue at 20 h after inoculation. The blot was hybridized with a PDF1.2 probe, stripped, and reprobed with a ribosomal probe as a control. (E) Quantitative estimation of lesion development after A. brassicicola infection (106 spores/mL). The area covered by necrotic tissue was expressed as a percentage of the inoculated area. Data points represent averages with SD of at least 30 lesions for each genotype. Letters indicate statistically significant differences between genotypes (Student's t test, P < 0.05, n = 20). (F) Germination percentages of at least 100 seeds pretreated with 10 µM paclobutrazol and sown on 0.5x MS, 1% sucrose, and 0.8% agar plates with or without 50 µM MeJA. Germination was assessed at 2 d after transferring plates to 23°C in continuous light. Bars represent averages with SE of three independent experiments. (G) Root growth inhibition in response to MeJA treatment. Seedlings were grown for 14 d on 1x MS and 2% sucrose plates supplemented with or without 50 µM MeJA. At least 30 seedlings were measured for each genotype. Data are presented as percentages of the length of treated roots compared with their respective nontreated controls. Bars represent averages with SD. Letters indicate statistically significant differences between genotypes (Student's t test, P < 0.05, n = 30).

We previously showed that Gß is also involved in resistance to A. brassicicola (Trusov et al., 2006). Application of spores (10⁶ spores/mL) on the leaf surface of Arabidopsis plants causes necrotic lesions that are clearly different in the wild type and Gß-deficient mutants. agb1-2, agg1, agg2, and agg1 agg2 mutants along with wild-type Col-0 plants were inoculated with A. brassicicola (Figure 3D), and disease progression was quantified by measuring the necrotic lesion area (given as a percentage of the droplet-inoculated area) (Figure 4E). Statistical analysis showed two very distinct groups that are significantly different from each other (P < 0.05). Lesions on agb1-2, agg1, and agg1 agg2 mutant leaves occupied 50 to 60% of the inoculated area, in contrast with wild-type plants and agg2 mutants, in which an average of 30% of the inoculated area became necrotic. In agreement with these observations, RNA gel blot hybridization revealed that 20 h after infection with A. brassicicola, steady state levels of the plant defensin PDF1.2 transcript were reduced in agb1-2, agg1, and agg1 agg2 mutants compared with the wild type and agg2 mutants (Figure 4D).

It was previously established that the increased susceptibility to fungal necrotrophic pathogens that was observed in Gß-deficient mutants correlates with a decreased sensitivity to methyl jasmonate (MeJA). Therefore, we assayed MeJA sensitivity using a germination assay. All mutants showed reduced sensitivity to MeJA compared with wild-type plants (Figure 4F), although to different degrees: agb1-2 = agg1 agg2 < agg1 < agg2 < wild type. MeJA sensitivity was also assayed using root length inhibition assays (Figure 4G). Two statistically different groups (P < 0.05) were observed, the first one showing decreased sensitivity to MeJA in agb1-2, agg1 agg2, and agg1 mutants and the second one containing the wild type and agg2 mutants.

AGG1 and AGG2 Act Additively in Gß-Mediated Lateral Root Development
It has been established that Gß, but not G, attenuates auxin-induced cell division leading to lateral root proliferation, although it does not directly couple auxin signaling (Ullah et al., 2003; Chen et al., 2006a). Figure 5A shows the number of lateral roots in 2-week-old wild-type plants and mutants deficient in Gß, G1, G2, or both G subunits grown on vertical plates (0.5x MS, 1% sucrose, and 0.8% agar, 16:8 day:night cycle, 23°C). All mutants produced more lateral roots than wild-type plants, but three statistically distinct groups (P < 0.05) were observed within the mutants: agb1-2 and double agg1 agg2 mutants had the highest number of lateral roots, agg2-1 and agg2RNAi mutants produced fewer lateral roots, while agg1-1c and agg1RNAi had even fewer roots (Figure 5A). Alteration of the growth conditions, such as an increase in MS salt concentration (from 0.5x to 1x) and reduced temperature (from 23 to 21°C) substantially (more than three times) decreased the total number of lateral roots (Figure 5C, white bars) as well as the differences among the various mutants and between mutants and the wild type.

View larger version (23K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 5. Effect of the Loss of G Subunits on Lateral Root Formation. (A) Average number of lateral roots in 2-week-old seedlings grown on vertical plates (0.5x MS, 1% sucrose, and 0.8% agar, 23°C, 16:8 light:dark cycle). Error bars represent SE. Letters indicate statistically significant differences between genotypes (Student's t test, P < 0.05, n = 15). (B) Auxin-induced lateral root development. Seedlings were grown for 9 d on 5 µM NPA and transferred to plates with or without 0.1 µM NAA for an additional 5 d under continuous light on vertical plates. The SD is based on at least 15 seedlings. (C) High temperature–induced lateral root development. Seedlings were grown at 21 and 29°C for 10 d, and the number of lateral roots was scored. The SD is based on at least 15 seedlings.

Exposure of Arabidopsis plants to high temperature (29°C) results in an increase in endogenous auxin levels (Gray et al., 1998). Although that original work focused on the effect of endogenous auxin induction on hypocotyl elongation, an increased number of lateral roots was also observed (Gray et al., 1998). In addition, it has been established that shoot-derived auxin is required for the emergence of lateral root primordia (Reed et al., 1998). agb1-2, agg1, agg2, and agg1 agg2 mutants along with wild-type Col-0 plants were grown at either 21 or 29°C (1x MS), and the number of lateral roots was determined in 2-week-old plants. All genotypes showed a marked increase in the number of lateral roots when grown at high temperature, with the smallest effect (2.5-fold increase) observed in wild-type plants (Figure 5C). agg1-1c and agg1RNAi mutants displayed 5.5- and 4.6-fold increases, respectively, while agg2-1 and agg2RNAi showed 3.2- and 3.5-fold increases, respectively. Both agb1-2 and double agg1-1 agg2-1 mutants produced approximately seven times more lateral roots when grown at 29°C (Figure 5C). In addition, adventitious roots were frequently observed (80 to 90% of seedlings) on hypocotyls of agb1-2, agg1 agg2, and agg1 mutants but never in wild-type plants or agg2 mutants (data not shown).

AGG1 and AGG2 Are Involved in the Modulation of Acropetally and Basipetally Transported Auxin Activity, Respectively
AGG1 and AGG2 expression in roots is cell-specific (Figure 1D), correlating with acropetal and basipetal auxin streams, respectively (Mitchell and Davies, 1975; Jones, 1998). Therefore, we hypothesized that Gß1 represses lateral root development from the central cylinder by attenuating the activity of acropetally transported auxin, while Gß2 represses lateral root formation or growth through the cortex/epidermis by affecting basipetal auxin. It was established that shoot-derived auxin is the predominant source of auxin in young (5- to 7-d-old) Arabidopsis roots, controlling lateral root emergence during early development, while later in development, the root system gradually reduces the dependence on shoot-derived auxin by synthesizing a sufficient amount within the root tip at 10 d after germination (although shoot-derived auxin is still important for primordial outgrowth) (Bhalerao et al., 2002; Ljung et al., 2005). Therefore, seedlings were grown for 7 d (1x MS) to allow maximal root elongation before the root tip started to produce auxin, and then acropetal auxin transport was inhibited by the method described by Reed et al. (1998). Seedlings with the auxin transport inhibitor NPA block placed at the root tip had only acropetal auxin transport in the area of the root above the block, while seedlings with the NPA block placed at the shoot–root junction should develop lateral roots mainly under the control of basipetal transport, with the exception of the fraction of roots initiated by early acropetal auxin. The dynamics of lateral root emergence was recorded during the 2-week period after the application of the NPA block (Figures 6A and 6B ). As expected, the rate of lateral root production after both treatments was highest in the agb1-2 and agg1 agg2 mutants and lowest in wild-type plants. agg1-1c seedlings produced abundant lateral roots (statistically indistinguishable from agb1-2 and agg1 agg2), despite the arrest of basipetal transport (Figure 6A). Inhibition of acropetal transport resulted in an initially high number of lateral roots in agg1-1c seedlings (day 13 in Figure 6B), probably as a result of early acropetal auxin flux before the block was applied. After the initial peak, the rate of lateral root formation was similar to that in wild-type plants (Figure 6B). By contrast, suppression of basipetal transport reduced lateral root numbers in agg2-1 to wild-type levels (Figure 6A), while arrest of acropetal transport resulted in elevated levels of lateral roots, statistically indistinguishable from those of agb1-2 and agg1 agg2 mutants (Figure 6B). Similar behavior was exhibited by the RNAi lines (data not shown).

View larger version (49K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 6. Specific Roles of AGG1 and AGG2 in the Regulation of Auxin Response. (A) and (B) Dynamics of lateral root formation after the arrest of basipetal (A) and acropetal (B) auxin transport. Conditions are described in the text. At least 15 plants for each genotype were used in the assay. Error bars represent SD. (C) Adventitious root development on excised hypocotyl explants. Seedlings were grown for 4 d in the dark and then for 1 d under light. Hypocotyls were excised aseptically and transferred to plates containing 1 nM NAA. Excised hypocotyls were grown for an additional 10 d under continuous light and photographed. (D) Response to gravistimulation. Fifty to 60 seedlings of each genotype were grown on 1x MS plates for 5 d under continuous light and then moved into darkness for an additional 24 h, and the plates were rotated 90°. Bars represent average deviation of the angle (curvature) from the horizontal line. Asterisks indicate statistically significant differences relative to the wild type (***P < 0.001, *P < 0.05). Error bars indicate SD.

Rashotte and coworkers (2000) showed that inhibition of basipetal auxin transport in roots completely blocked its gravity response, while inhibition of acropetal transport only partially reduced it. Therefore, we assayed the gravitropic response of wild-type and G protein mutant roots by measuring the root angle (measured from the horizontal position) at 24 h after gravistimulation. Figure 6D shows that agb1-2, agg2-1, and agg1 agg2 mutants were less responsive to gravistimulation than wild-type plants and agg1-1c (P < 0.001). Interestingly, agg1-1c was slightly less responsive than the wild type (P < 0.05), probably due to a limited participation of the acropetal auxin in the gravity response (Rashotte et al., 2000).

AGG1 and AGG2 Are Involved in Different Responses during Germination
Two recent reports established that Gß signaling plays a role in germination (Pandey et al., 2006; Trusov et al., 2006). To determine the specific roles of each of the partner G subunits in this process, mutants lacking Gß, G1, G2, or both G subunits were subjected to germination tests. Since germination efficiency is extremely sensitive to the growth conditions experienced by the parental plant and postharvest storage, all seed lots were collected at the same time from plants grown simultaneously under the same conditions and were stored for 2 months at 4°C in the dark. Approximately 100 sterilized seeds of all tested lines were planted on the same Petri dish for a single treatment.

Germination and early development are regulated by many Gß-mediated signals, and glucose is arguably the best characterized of those signals to date (Ullah et al., 2002; Pandey et al., 2006; Wang et al., 2006). As shown in Figure 7A , there was a clear difference between wild-type and mutant plants when germinated in the presence of 6% glucose, while 4% glucose did not discriminate among the different genotypes and 2% glucose resulted in nearly 100% germination. Because light intensity also has an effect on germination, we used two different intensities of continuous light irradiation (63 and 150 µmol·m^–2·s^–1). The higher light intensity resulted in faster germination rates, reaching 90% by day 6 on glucose and by day 3 on mannitol (Figures 7C and 7E, respectively), obscuring any differences between genotypes. By contrast, the slower germination rates observed using a lower light intensity accentuated the differences among genotypes. When sown on glucose under low light intensity, agb1-2, agg1, and agg1 agg2 mutant seeds showed drastically reduced germination rates compared with wild-type seeds, with <50% germination after 2 weeks (Figure 7B). By contrast, at higher light intensities, the differences between wild-type and agb1 and agg1 mutant seeds were only observed at day 2 (Figure 7C). Interestingly, agg2 mutants also displayed significant inhibition of germination on glucose, albeit at notably lower levels than agg1 mutants. Again, the difference was statistically significant in lower light (Figure 7B), while at higher light this difference was insignificant (Figure 7C).

View larger version (30K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 7. Germination Assays in G-Deficient Mutants. (A) Germination rates of wild-type plants and the indicated mutants at 5 d after transfer to 23°C in the presence of different concentrations of glucose. (B) and (C) Germination dynamics of wild-type plants and mutants on medium (0.5x MS and 0.8% agar) containing 6% glucose under two different light intensities, 63 µmol·m–2·s–1 (B) and 150 µmol·m–2·s–1 (C). (D) and (E) Germination dynamics of wild-type plants and mutants on medium (0.5x MS and 0.8% agar) containing 6% mannitol under two different light intensities, 63 µmol·m–2·s–1 (D) and 150 µmol·m–2·s–1 (E). Error bars indicate SD.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION METHODS REFERENCES

 
Previously, the functional selectivity of G subunits was largely unrecognized, with the general view that G function is limited to anchoring the Gß dimer to the membrane. However, G recently emerged as an important element that provides effector specificity as well as receptor selectivity for the heterotrimer (Gautam et al., 1990; Hou et al., 2000; Akgoz et al., 2002; Azpiazu and Gautam, 2002; Myung et al., 2006).

The initial discovery of single G and Gß subunits in Arabidopsis challenged the concept that plants use combinatorial subunit composition to define G protein receptor/effector specificity (Arabidopsis Genome Initiative, 2000), as proven in mammalian systems (Robishaw and Berlot, 2004). With the recent discovery of two G subunits in Arabidopsis (Mason and Botella, 2000, 2001), we must now address this possibility. Since both plant G subunits share a number of similarities with animal G subunits, such as the strong interaction with Gß and the presence of isoprenylation domains, it is reasonable to expect that there are two operational Gß subunits in Arabidopsis. A number of logical questions follow, such as whether the two subunits mediate the same processes or whether they specialize in different developmental, biotic, or abiotic responses. In this respect, it is interesting that AGG1 and AGG2 in situ expression profiles show a high degree of tissue specificity and that, even though the sum of their individual expression patterns mimics the overall Gß expression, the two G gene expression patterns rarely overlap. This raises the possibility that G subunits impose selective functionality restricted by expression patterns.

The functions of the two G subunits are intrinsically linked to Gß, since, based on mammalian studies, the Gß dimer operates as a single signaling unit. The Gß subunit has been associated with a number of processes using loss-of-function mutants (Lease et al., 2001; Ullah et al., 2003; Llorente et al., 2005; Pandey et al., 2006; Trusov et al., 2006). However, according to the classical mechanism of heterotrimeric G protein action, the lack of a functional Gß subunit affects not only processes directly mediated by Gß but also those mediated by G; therefore, some of the processes affected in Gß mutants are actually regulated by G (Ullah et al., 2003). In general, those phenotypes shared by G- and Gß-deficient mutants are most likely due to disruption in processes mediated by G, while disruption of processes mediated by Gß results in different or even opposite phenotypes (Ullah et al., 2003). Therefore, to avoid complications in interpretation, we chose processes with predominant Gß signaling, namely, resistance against necrotrophic pathogens (Llorente, et al., 2005; Trusov, et al., 2006), auxin-regulated lateral root development (Ullah et al., 2003), and D-glucose inhibition of germination (Ullah et al., 2002; Chen et al., 2006b; Pandey et al., 2006; Wang et al., 2006).

Involvement of Gß1 in Resistance against Fungal Pathogens
Quantitative and in situ gene expression studies in transgenic Arabidopsis reporter lines using two different pathogens gave the first indication of the involvement of G1 along with Gß in the defense mechanisms against necrotrophic fungi. These observations were confirmed by the fact that the Gß-deficient mutant agb1-2 and all of the mutants lacking AGG1 (agg1-1c, agg1RNAi, and agg1 agg2) showed increased susceptibility to F. oxysporum, with no statistically significant differences observed between them. The increased susceptibility of G1-deficient mutants to F. oxysporum was shown for Col-0 and Ws. The slight increase in resistance observed for G-deficient mutants suggests that, in defense-related processes, G acts by sequestering the Gß1 dimer to the inactive heterotrimeric complex, thus effectively lowering the free available Gß1 pool (Llorente et al., 2005; Trusov et al., 2006). This is consistent with the finding that the expression of GPA1 is not altered by pathogen exposure (Y. Trusov and J.R. Botella, unpublished data). Even though A. brassicicola and F. oxysporum are both necrotrophic fungi, their infection mechanisms are different. As for F. oxysporum, the responses of all AGG1-deficient mutants and agb1-2 to A. brassicicola were statistically indistinguishable, being more severely affected than in the wild type. This finding suggests that the complete Gß1 dimer is required for defense. By contrast, mutants deficient in AGG2 but not AGG1 (agg2-1 and agg2RNAi) showed a wild-type phenotype in their behavior against both pathogens, thus precluding any significant role of the Gß2 dimer in pathogen resistance.

The susceptibility data are consistent with the molecular observations showing reduced induction of the plant defensin PDF1.2 by A. brassicicola in agb1-2 and all mutants lacking AGG1 (agg1-1c, agg1RNAi, and agg1 agg2) but wild-type induction in agg2 mutants. In addition, all AGG1-deficient mutants showed reduced responses to MeJA (statistically indistinguishable from Gß-deficient mutants), supporting the hypothesis that MeJA signaling could be the link between G proteins and the defense response (Trusov et al., 2006).

Regulation of Lateral Root Development by Gß1- and Gß2-Mediated Signaling
In the young Arabidopsis primary root, auxin transport occurs acropetally through the stele tissue from the first true leaves, where it is primarily synthesized (Bhalerao et al., 2002). This auxin stream initiates early lateral root primordia (Reed et al., 1998; Bhalerao et al., 2002) and augments root-mediated auxin synthesis (Ljung et al., 2005). At a later stage, the root meristem synthesizes auxin, which moves up from the root tip through the epidermis (Mitchell and Davies, 1975; Tsurumi and Ohwaki, 1978; Jones, 1990, 1998; Rashotte et al., 2001), influencing lateral root initiation (Bhalerao et al., 2002; Ljung et al., 2005). Thus, auxin in both streams initiates lateral root formation, but different signaling mechanisms had not been distinguished previously.

We showed that AGB1, AGG1, and AGG2 are each expressed in roots, with AGB1 expression being observed in the stele, cortex, and epidermis, whereas AGG1 expression is restricted to the stele and AGG2 is predominantly active in the cortex and epidermis. Interestingly, none of the genes was expressed in lateral root primordia or in pericycle cells, which become the initials to lateral root meristems. Gß attenuates auxin signaling during lateral root formation (Ullah et al., 2003), and we extended this finding by showing the G subunits provide specificity in this response. While both AGG1 and AGG2 are involved in the inhibition of auxin-dependent lateral root initiation and both possible dimers, Gß1 and Gß2, exert a synergistic effect in auxin signaling attenuation, neither Gß dimer type is able to compensate for loss of the other. A likely explanation is that each dimer acts on different branches of the auxin/lateral root pathway. This duality does not occur in hypocotyls, as Gß1, and not Gß2, attenuates auxin-induced adventitious roots in the hypocotyl.

Considering that AGG1 is expressed in the root stele, where acropetal auxin transport occurs, while AGG2 is expressed in the cortex and epidermis, which are known to accommodate basipetal auxin transport, we hypothesized that Gß1 and Gß2 could be specifically involved in signaling for each of the two auxin streams. Consistent with this, we found that inhibition of acropetal auxin transport at the shoot–root junction affected agg1 mutants, while agg2 mutants were more responsive to the inhibition of basipetal auxin transport arising from the root tip. Furthermore, support for our hypothesis was provided by studying the gravitropic response, a process that is dependent on basipetal auxin transport. The reduced responsiveness of agb1-2 and the agg2 mutants is consistent with a signaling role for basipetally moving auxin in the root. Taking into account the localization of the proteins, we speculate that Gß1 could mediate internal signals while Gß2 could be involved in external/environmental signaling. Brassinosteroids and ethylene are logical candidates to be such internal signals, since both brassinosteroids and ethylene signal transduction pathways are influenced by heterotrimeric G proteins at various stages of plant development (Ullah et al., 2002) and there is evidence that brassinosteroids and ethylene promote lateral root development by increasing acropetal auxin transport (Bao et al., 2004) and by increasing auxin content locally at pericycle founder cells (Aloni et al., 2006). On the other hand, it is well known that a wide range of soil characteristics, such as availability of water or nutrients, can dramatically affect lateral root development (Vanneste et al., 2005). Signaling from one or more of these factors could be coupled by Gß2.

Germination and G Protein Signaling
The role of G proteins in seed germination is intriguing and complicated, since these proteins affect gibberellic acid, abscisic acid, brassinosteroids, MeJA, ethylene, and auxin signaling (Ashikari et al., 1999; Ueguchi-Tanaka et al., 2000; Wang et al., 2001; Ullah et al., 2002; Lapik and Kaufman, 2003; Chen et al., 2004; Pandey et al., 2006) as well as D-glucose sensitivity (Ullah et al., 2002; Chen et al., 2006b; Pandey et al., 2006; Wang et al., 2006). The gpa1 and agb1 null mutants show a number of alterations in seed germination, suggesting that GPA1 and AGB1 are involved in this process, although their specific roles are not known (Ullah et al., 2002; Chen et al., 2006b; Pandey et al., 2006). Here, we focused on traits dependent on Gß-mediated signaling to establish the specificity of the G subunits. The D-glucose hypersensitive phenotype of the Gß null mutants is more severe than that for the G null mutants, implying that the predominant signaling element in D-glucose–regulated germination is the Gß dimer (Pandey et al., 2006; Wang et al., 2006). Our results indicate that both Gß1 and Gß2 dimers mediate this response, although their involvements are different. Gß1 is mostly involved in the osmotic component of the glucose response, although involvement in glucose signaling cannot be discounted, while Gß2 plays a role in glucose signaling but not in osmotic stress. The apparent contradiction of our results with the previously reported wild-type sensitivity of agb1-2 to a different osmotic agent, sorbitol (Pandey et al., 2006), can be explained by the masking effect that light intensity (used in that study) has on osmotic response (cf. Figure 6E with 6D). These data further illustrate the complexity of the germination process, implicating at least two independent signaling pathways involving Gß1 and Gß2 dimers and the additional effect of light intensity.

The fact that AGB1- and AGG1-deficient mutants are hypersensitive to osmotica raises the attractive possibility of the involvement of Gß1 signaling in osmoregulation (Zhu, 2002). The high expression levels observed for AGB1 and AGG1 in hydathods, highly specialized osmoregulatory organs, also suggests such a speculation.

Subunits Provide Functional Selectivity to the Gß Dimer
There are substantial similarities, but also important differences, between animal and plant heterotrimeric G proteins. They are structurally similar, suggesting a conserved mechanism of action (i.e., once a G protein–coupled receptor is activated, the associated G protein will dissociate and transduce the signal to downstream effectors through two functionally distinct subunits, G and Gß). However, plant G proteins lack the multiplicity of genes encoding each of the subunits, as in animals. It is this multiplicity that provides numerous combinatorial possibilities to the whole heterotrimer in order to mediate the action of hundreds of receptors in animal systems. Having single G and Gß subunits begs the question of how plant G proteins are involved in a large variety of plant processes (Jones, 2002; Assmann, 2004; Jones and Assmann, 2004). The existence of two different G subunits provides functional diversity to the entire heterotrimer for effector activation and receptor specificity. The similarities of the phenotypes displayed by Gß- and G-deficient mutants provide a functional association between the Gß subunit and each of the G subunits in plants, showing that both G subunits form functional Gß dimers. We also showed that the two G subunits serve independent, redundant, or complementary roles in planta, depending on the process and the tissue being studied. In some processes, such as defense against necrotrophic fungi, only one G subunit is involved (AGG1). In other processes, such as auxin signaling and the development of lateral roots, both subunits are involved but are mechanistically different in their operation. In other processes, such as germination, both G subunits are involved but with independent roles, with AGG2 implicated in glucose signaling and AGG1 mediating the response to osmotica (Figure 8 ).

View larger version (8K): [in this window] [in a new window] [as a PowerPoint slide]   Figure 8. Two Arabidopsis G Subunits Provide Functional Selectivity to the Gß Dimer. Summary of the involvement of each Gß dimer in pathogen resistance, germination, lateral root development, and gravitropism.