• G-proteins (guanine nucleotide-binding proteins that usually exhibit GTPase activities) and related signal transduction processes play important roles in mediating plant defense responses; here, a rice (Oryza sativa) cDNA clone, OsGAP1, encoding a GTPase-activating protein (GAP) that also contains a protein kinase C conserved region 2 (C2) domain is reported.
• An interacting G-protein partner for the OsGAP1 protein was identified by yeast two-hybrid library screening and confirmed by co-immunoprecipitation; the GTPase-activation activity of OsGAP1 on this interacting G-protein was demonstrated using in vitro assays.
• OsGAP1 was induced by wounding in rice and the presence of the R locus Xa14 enhances such induction.
• Gain-of-function tests in transgenic rice and Arabidopsis thaliana showed that constitutive expression of OsGAP1 led to increased resistance to bacterial pathogens in both monocots and dicots.
Crop yield can be severely reduced by pathogen attacks. For instance, the loss in rice (Oryza sativa) yield caused by diseases can reach 50% in some areas of Asia (Adhikari et al., 1995; Ronald, 1997). Successful examples of enhancing disease resistance by transformation of R genes have been reported in both dicots and monocots (Whitman et al., 1994; Song et al., 1995). However, R protein-mediated resistance exhibits strong race specificity (Song et al., 1995; Wang et al., 1998), and breakdown of resistance will occur as a consequence of the shift of pathogen populations in the field. The use of regulators of signal transduction pathways to achieve broad-spectrum, durable disease resistance against pathogens is an attractive alternative strategy.
G-proteins (guanine nucleotide-binding proteins that usually exhibit GTPase activities) are important components involved in various signal transduction pathways (de Vries et al., 2000; Assmann, 2002), including plant defense responses. G-proteins can be broadly classified into three categories (Assmann, 2002): the commonly known heterotrimeric G-proteins comprised of α, β and γ subunits, the single small G-proteins, and other ‘unconventional’ G-proteins. G-proteins cycle between the active and inactive forms by binding to GTP or GDP, respectively. In general, the activities of GTPase-activating proteins (GAPs) activate the intrinsic GTPase activities of G-proteins (the Gα subunit in the case of heterotrimeric G-proteins), and hence reduce and/or terminate signals generated from ligand-activated G-protein-coupled receptors (Jones & Assmann, 2004). Recycling of G-proteins is achieved through the activities of guanine-nucleotide exchange factors that catalyze the release of GDP and the subsequent uptake of GTP by G-proteins (Sprang, 2001).
The involvement of G-proteins in the plant defense response has been documented. For example, activation of G-proteins led to a defense-related oxidative burst in soybean (Glycine max) cells (Legendre et al., 1992) and induced synthesis of pathogenesis-related (PR) proteins in tobacco (Nicotiana tabacum) (Beffa et al., 1995). It has been proposed that the interaction between Avirulence (Avr) proteins and the corresponding R proteins can activate G-proteins to mediate expression of PR genes and signaling of reactive oxygen species (ROS) related to plant defense responses (Blumwald et al., 1998). Loss of function mutations in one of the two Arabidopsis thaliana Gγ subunit genes (to produce an Arabidopsis Gγ subunit gene 1 (AGG1)-deficient mutant) reduced resistance to necrotrophic pathogens, reduced induction of the plant defensin gene PDF1.2, and decreased sensitivity to methyl jasmonate (Trusov et al., 2007). In addition, the induction of a number of defense-related genes in response to Alternaria brassicicola infection was much reduced in Gβ subunit-deficient mutants of A. thaliana. Gβ-deficient mutants also exhibited decreased sensitivity to a number of methyl jasmonate-induced responses such as the induction of PDF1.2 (Trusov et al., 2006). In rice (Oryza sativa), the dwarf1 (d1) mutant, which is deficient in Gα, showed a delayed induction of the probenazole-inducible protein gene (PBZ1, a rice defense marker gene) when treated with probenazole (a plant defense chemical activator). The d1 mutant exhibited accelerated development of disease symptoms when infected with Xanthomonas oryzae pv. oryzae (Xoo) (Komatus et al., 2004).
In addition to heterotrimeric G-proteins, some small G-proteins such as OsRac1 in rice are also key signal transducers in the plant defense response (Ono et al., 2001; Suharsono et al., 2002). OsRac1 in rice is a key component mediating the defense response to both Xoo and the rice blast fungus Magnaporthe grisea, by regulating the production of ROS which induce the hypersensitive response (Ono et al., 2001; Suharsono et al., 2002). The heterotrimeric G-protein α subunit acts upstream of OsRac1 in such process (Suharsono et al., 2002). Furthermore, overexpression of a Ras-related G-protein, another type of small G-protein, in tobacco could also increase the salicyclic acid (SA) concentration and induce PR gene expression (Zhu et al., 1996).
It was also reported that the GAP of a Ras-related nuclear protein (Ran) family G-protein can physically interact with the nucleotide binding and leucine rich repeat (NB-LRR) resistance protein Rx to regulate viral resistance (Sacco et al., 2007; Tameling & Baulcombe, 2007), illustrating a vital role of GAPs in the plant defense response.
During a search for differentially expressed cDNA clones in near-isogenic lines (NILs) of rice containing Xa loci (R loci in the rice host against Xoo) using a suppression subtractive hybridization approach, one cDNA clone (OsGAP1) was obtained which encodes a putative GAP. The GTPase-activating activity of the OsGAP1 protein was demonstrated using an in vitro system. We also used gain-of-function tests in transgenic rice and A. thaliana to show the involvement of OsGAP1 in plant defense responses.
Materials and Methods
Plant material and chemicals
Rice (Oryza sativa L.) NILs (CBB14 and SN1033) used in this research were constructed previously (Zhang et al., 1996). The rice cultivar Aichi asahi and the Arabidopsis thaliana (L.) Heynh. wild type Columbia-0 (Col-0) used for transformation were laboratory stocks. The non-expresser of PR genes (npr1-3) mutant was obtained from Dr C. Deprés at Brock University, St. Catharines, Canada. Enzymes and reagents for molecular studies were obtained from Bio-Rad Laboratories (Hercules, CA, USA) and Roche Diagnostic Ltd (Basel, Switzerland). Chemicals for plant growth and tissue cultures were supplied by Sigma-Aldrich Co. (St Louis, MO, USA). MetroMix200 soil for the cultivation of A. thaliana was obtained from Hummert International Supplier (Earth City, MO, USA). DNA oligos were supplied by Integrated DNA Technologies, Inc. (Coraliville, IA, USA), Invitrogen Corp. (Carlsbad, CA, USA), and Tech Dragon Ltd (Hong Kong). Primers used in this research are listed in Supplementary Material Table S1.
Growth and pathogen inoculation of rice and A. thaliana
Pathogens used in this experiment included the Chinese Xanthomonas oryzae pv. oryzae (Xoo) strain LN44, the Japanese Xoo race T1 (T7147), and the Philippine Xoo race 6 (P6: PXO99). All the Xoo strains/races were cultured at 28°C on peptone sucrose agar plates. The bacterial inoculums were prepared by suspending bacterial culture in sterile distilled water at an optical density (OD) (600 nm) of 1.0.
Rice lines were grown on regular field soil in a glasshouse under natural sunlight. For each plant, tillers were randomly divided into three groups and the fully expanded leaves of each group were inoculated with one of the three Xoo strains/races. Rice plants at booting stage were inoculated with the Xoo strains/races using a leaf-clipping method (Kauffman et al., 1973; Zhang et al., 1996). Two weeks after inoculation, disease symptoms were scored by measuring the percentage of lesion area (of the whole leaf).
The same procedure was used for mock inoculation and wounding experiments except that the pathogen was replaced with water. For time-course experiments, a day 0 sample was collected before treatment. Other samples were collected at 2, 4, or 6 d after treatment at around the same time of day (between 08:00 and 10:00 h).
Arabidopsis thaliana was grown in a growth chamber (temperature 22–24°C; relative humidity (RH) 70–80%; light intensity 80–120 µmol m−2 s−1 on a 16 h light : 8 h dark cycle). Pseudomonas syringae pv. tomato DC3000 (Pst DC3000) was obtained from Dr C. Lo at the University of Hong Kong. Preparation of the Pst DC3000 culture, inoculation (using a dipping method at a concentration of 108 colony-forming units per ml in 10 mM MgSO4 supplemented with 0.02% (v/v) Silwet L-77), and subsequent titer determination were performed as previously described (Uknes et al., 1992; Falk et al., 1999; Kim & Delaney, 2002).
Constitutive expression of OsGAP1 in transgenic rice and A. thaliana
To construct transgenic rice lines overexpressing OsGAP1, the cDNA clone was first subcloned into the binary vector pSB130 (Yu et al., 2006) and placed under the control of the maize (Zea mays) ubiquitin promoter (Rooke et al., 2000). The recombinant construct was transformed into Agrobacterium tumefaciens strain EHA105 by electroporation. The recombinant construct was transformed into the japonica rice cultivar Aichi Asahi, using an A. tumefaciens-mediated transformation protocol described previously (Hiei et al., 1994; Zhu et al., 2007). Hygromycin-resistant lines were further tested for the presence of the transgene using PCR. For details of the recombinant construct, see Supplementary Material Fig. S1.
To express OsGAP1 in transgenic A. thaliana, the cDNA clone was inserted into a binary vector (Brears et al., 1993) and placed under the control of the cauliflower mosaic virus 35S promoter. The recombinant construct was transferred into A. tumefaciens strain GV3101 (pMP90) (Koncz & Schell, 1986) by electroporation. Transformation into the wild type Col-0 or the npr1-3 mutant was performed using a vacuum infiltration method (Bechtold & Pelletier, 1998). Tests for single insertion events were performed by statistical analysis (χ2 test) of the kanamycin resistance (encoded by the selection marker gene on the binary vector) phenotypes exhibited by the offspring. A 3 : 1 (resistant:sensitive) ratio in the T2 generation suggested a single insertion event. Homozygous lines obtained after the T3 generation were used in this study. The presence of the transgene was verified by PCR. For details of the recombinant construct, see Supplementary Material Fig. S2.
DNA sequencing, RNA extraction, reverse transcription and real-time PCR
DNA sequencing was performed using a thermal sequencing method according to the manufacturer's manual (4337455; Applied Biosystems, Foster City, CA, USA). Total RNA was extracted using a phenol extraction method (Ausubel et al., 1995). The cDNA samples were prepared by reverse transcription of DNase I (18068-015; Invitrogen)-treated RNA samples, using Superscript‘ II RNaseH (18064-071; Invitrogen) and a 18-mer oligo-dT as described in the manufacturer's manual.
Real-time PCR amplification of cDNA was performed using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems) in 96-well PCR plate with a dome cap. The reaction was carried out in a 20-µl reaction mixture containing 10 µl of 2X SYBR Green PCR Master Mix (4309155; Applied Biosystems) with 0.3 µM each of the forward and reverse primers. Primers for real-time PCR were designed using the program primer express (Applied Biosystems). At least one biological repeat was performed using independent plant samples to ensure that the gene expression pattern was consistently observed. All reactions were set independently at least four times and at least three sets of consistent data were used for analysis.
The expression levels of the rice actin gene (OsAc1D; accession number X15865; Wasaki et al., 2003) and the A. thalianaβ-tubulin gene (tub4; accession number M21415; Chu et al., 1993; van Wees et al., 2000; Ton et al., 2002) were used for signal normalization of real-time PCR results in rice and A. thaliana, respectively. The relative gene expression was calculated using the 2−ΔΔCT method (Livak & Schmittgen, 2001). To validate the application of the 2−ΔΔCT method, the comparability of amplification efficiencies between the target genes and the housekeeping genes was also examined. A single sharp peak was observed in the dissociation curve, confirming the accuracy of the real-time PCR.
Cloning of OsGAP1
Six- to eight-week-old rice lines were inoculated with the Xoo strain LN44. Leaf tissues were collected 4 d after inoculation and used to prepare total RNA. The Clontech PCR-Select cDNA Subtraction Kit (637401; Clontech, Mountain View, CA, USA) was used to perform suppression subtractive hybridization to obtain candidate clones. A candidate clone containing a partial sequence of OsGAP1 was subjected to 5′ rapid amplification of cDNA ends (5′RACE) using the SMART‘ RACE cDNA amplification kit (634914; Clontech). The primers used and PCR conditions for 5′RACE are described in the table footnote of Supplementary Material Table S1. All clones were stored in the plasmid vector pBlueScript KSII(+) and propagated in Escherichia coli strain DH5α.
Protein extraction, antibodies, and western blot analysis
Membrane-bound and soluble proteins were separated using a fractionation method (modified from Jiang & Rogers, 1998). Plant tissues were ground with homogenizing buffer (40 mM HEPES (pH 7.4), 250 mM sucrose, 3 mM MgCl2, 0.1 mM EDTA, 1 mM phenylmethylsulphonyl fluoride (PMSF) and 20 µg ml−1 leupeptin), followed by centrifugation at 300 g for 10 min. The supernatant was collected as the soluble fraction, and the pellet was further centrifuged at 10 000 g for 60 min. The resulting supernatant was pooled with the soluble fraction before 1% sodium dodecyl sulphate (SDS) was added. The membrane fraction containing proteins that were tightly membrane bound was obtained by resuspending the pellet in homogenizing buffer supplemented with 2% SDS. Both soluble and membrane fractions were boiled at 100°C for 10 min.
Primary antibody (polyclonal) targeting the OsGAP1 protein was raised using a commercial service (Invitrogen; Custom antibody) by injecting a synthetic peptide (‘N’-CRVIKKTTNPEWNDE-‘C’) into rabbits, and antibodies were purified with a synthetic peptide-conjugated affinity column before use. The antibodies against c-Myc (631206; Clontech) and HA (631207; Clontech) epitope tags were commercially available. Anti-rabbit or anti-mouse secondary antibody conjugated to an alkaline phosphatase (provided in the Western Breeze‘ Immunodetection Kit; WB7106; Invitrogen) was used for recognition of the primary antibody.
For western blot analysis, the proteins were electrophoretically separated on an SDS-polyacrylamide gel (4% stacking; 10% resolving) before being transferred to an activated polyvinylidene difluoride (PVDF) membrane pre-treated with absolute methanol for 20 min, and then to protein transfer buffer for 15 min, using the Trans-Blot® SD Semi-Dry Electrophoretic Transfer Cell (170-3940; Bio-Rad). The transfer, blocking (with Western Breeze‘ blocking solution), and detection (using the Western Breeze‘ Immunodetection Kit) steps were performed according to the manufacturer's manual.
Yeast two-hybrid, in vitro translation, and co-immunoprecipitation
For yeast two-hybrid experiments, the commercial kit BD Matchmaker‘ Library Construction & Screening Kit (K1615-1; Clontech) was used. Detailed procedures are described in the manufacturer's handbook (Yeast Protocols Handbook; PT3024-1; Clontech). OsGAP1 was amplified from the clone in the pBlueScript KSII(+) vector. Details of the primers used and PCR conditions are given in the table footnote of Supplementary Material Table S1. The amplified OsGAP1 fragment was restricted and inserted between the EcoRI and PstI sites of the plasmid vector pGBKT7 to form an in-frame fusion with the c-Myc epitope tag. The recombinant construct was subsequently transformed into the yeast strain Y187. Successful production of the OsGAP1 protein in the yeast cells was confirmed by western blot analysis using the anti-c-Myc antibody. To construct an AD domain fusion yeast library in the yeast strain AH109, a mixed sample of RNA from several rice lines (each containing one of the following R genes: Xa2, Xa12, Xa14, Pita2, Pib and Pik) inoculated with the corresponding incompatible pathogens (T2 for Xa2; P1 for Xa12; LN44 for Xa14; Ken54-04 for Pita2, Pib and Pik) after 4 d was used as the starting material; the procedure followed was that described in the manufacturer's manual.
Library screening was initiated by the mating between pGBKT7-OsGAP1 transformed Y187 and the AH109 yeast library constructed. Yeast diploid mating products were selected on SD minus Trp, Leu and His (SD/–3) agar plates and incubated at 30°C for 4 d. Only colonies 2–3 mm diameter in size were further streaked onto SD minus Trp, Leu, His and Ade (SD/–4) agar plates. Selected clones were further tested by colony-lift filter assay for β-galactosidase (lacZ) reporter gene activity.
The pGADT7-Rec plasmids containing putative OsGAP1 interacting partners were extracted from 5-ml yeast cultures in SD/–Trp, –Leu medium cultivated at 30°C overnight with shaking (OD at 600 nm > 1.0). Plasmid DNA was then transformed into E. coli DH5α competent cells using a CaCl2-mediated heat-shock method and selected on a Luria-Bertoni (LB) agar plate supplemented with ampicillin (100 mg l−1). Re-transformation of candidate clones into AH109 cells together with pGBKT7-OsGAP1 was performed to eliminate false positives resulting from mutations in yeast cells. Candidate clones were identified by DNA sequencing.
To obtain in vitro translated rice YchF domain-containing protein (OsYchF1), the full-length coding region of this clone was amplified and inserted into the pGADT7-Rec vector in-frame to the HA epitope tag. The primers used and PCR conditions are detailed in the table footnote of Supplementary Material Table S1. Plasmids linearized with HindIII were subjected to in vitro transcription using the RiboMAX Large Scale RNA Production Systems-T7 (P1300; Promega). The transcription reaction mixes were composed of approx. 3 µg of linearized plasmid DNA, 25 mM rNTPs, 1× transcription buffer and 1× enzyme mix (Promega). The reaction mixes were incubated at 37°C for 4 h. After transcription, RNA products were further incubated at 37°C for 15 min with DNaseI (Invitrogen; 1 U µl−1) to remove DNA templates before ethanol precipitation. Denaturing gel electrophoresis was performed to check the qualities and quantities of transcripts.
After verification of successful transcription, the mRNA was subjected to in vitro translation using wheat germ extract (L4330; Promega) and the Transcend‘ Biotin-Lysyl-tRNA System (L5061; Promega) in combination. About 10 µg of RNA sample was added to a 50-µl reaction solution which included 25 µl of wheat germ extract, 2 µl of 1 M potassium acetate, 2 µl of Transcend‘ biotin-lysyl-tRNA, 2 µl of 1 mM amino acid mixture (without leucine), 2 µl of 1 mM amino acid mixture (without methionine), and 40 units of RNasin® ribonuclease inhibitor. Successful production of in vitro translation products was confirmed using the Transcend‘ nonradioactive Translation Detection System (L5080; Promega).
For co-immunoprecipitation experiments, total protein was extracted from a rice line overexpressing OsGAP1 (modified from Boyes et al., 1998; Greve et al., 2003). About 100 µg of protein samples from rice were mixed with 40 µl of the in vitro translated fusion protein in a co-immunoprecipitation buffer containing 50 mM Tris/HCl (pH 7.5), 250 mM NaCl, 2 mM MgCl2, 0.5 mM CaCl2, 10% (v/v) glycerol, 1.5% (v/v) Triton X-100, 1 mM PMSF and 2 mg l−1 leupeptin (modified from Boyes et al., 1998; Greve et al., 2003). The BD Matchmaker‘ Co-IP Kit (630449; Clontech) was used, and the anti-HA epitope tag antibody was employed to pull down the protein complexes. The western signal was detected using the anti-OsGAP1 antibody. As a negative control for the co-immunoprecipitation experiment, a putative expressed protein from rice (accession number ABA98865) fused to HA was also produced using the same method.
Expression and purification of the GST-OsGAP1 and GST-OsYchF1 fusion proteins
The OsGAP1 cDNA containing the full-length coding region was subcloned into the EcoRI and SalI sites of the pGEX-4T-1 vector (GE Healthcare, Chalfont St Giles, UK) to form an in-frame fusion with the glutathione S-transferase (GST) coding sequence. Similarly, the OsYchF1 cDNA clone containing the full-length coding region was amplified from the cDNA pool of the CBB14 rice line (4 d post-inoculation with LN44), restricted with the EcoRI and SalI enzymes, and ligated into the pGEX-4T-1 expression vector to form an in-frame fusion with the GST coding sequence. The primer used and PCR conditions are detailed in the table footnote of Supplementary Material Table S1. After verification by DNA sequencing, the recombinant constructs were transformed into the BL21 E. coli strain. Expression of the GST-OsGAP1 and GST-OsYchF1 fusion proteins and the GST protein control was induced by adding 0.5 mM isopropyl-β-D-thiogalactopyranoside (IPTG) to the growth medium. Protein purification was performed with the GST SpinTrap‘ Purification Module (27-4570-03; GE Healthcare), following the procedures described in the user manual.
GTPase activity assay
GTPase activity was determined by monitoring the release of inorganic phosphate (Pi) during GTP hydrolysis using the EnzChek‘ Phosphate Assay Kit (E6646; Molecular Probes, Carlsbad, CA, USA) (Webb, 1992). Purified GST-OsGAP1, GST-OsYchF1, or GST protein (50 ng) was incubated with 200 µM GTP. The release of Pi was monitored by measuring the OD (360 nm) and quantification was performed by plotting a Pi standard curve, as detailed in the user manual. The GTPase activity of GST-OsYchF1 was reduced to about half or a quarter of the original value when the concentration of the protein extract was reduced 2- or 4-fold, respectively, indicating that the enzyme assay was within the linear range. To test GTPase-activating activities, 30 ng of GST-OsGAP1 or 30 ng of GST was mixed with 50 ng of GST-OsYchF1 before the GTPase activities were measured.
Data were analyzed using the Statistical Package for Social Sciences (ver. 12.0). The mean difference was analyzed using one-way analysis of variance followed by the Games–Howell or Tukey's post-hoc test.
Accession numbers of genes and clones
Accession numbers are used in the text. Corresponding At and Os locus tag/gene names (if available) are given in Supplementary Material Table S2.
Identification of OsGAP1, which encodes a C2 domain-containing protein in rice
Suppression subtractive hybridization experiments were performed using RNA samples isolated from the near-isogenic rice line CBB14 (tester; containing the R gene Xa14) and its susceptible recurrent parent SN1033 (driver; susceptible recurrent parent) after inoculation with the Xoo strain LN44 for 4 d (see the Materials and Methods). Several hundred clones were sequenced to obtain differentially expressed genes that encode putative new signal transduction components. One partial cDNA clone (spanning 681–960 bp of the clone NM_001053244) was identified as a putative GAP. The intact coding region of the cDNA clone (accession number EF584506) obtained via 5′RACE encodes 165 amino acid residues. The DNA sequence is 99% identical to that of a directly deposited rice cDNA clone (accession number NM_001053244). The BlastP program showed that the protein encoded by our clone is identical to an annotated protein kinase C conserved region 2 (C2) domain-containing protein-like clone from the rice genome sequence (accession number BAD15699). The genomic clone was located on chromosome 2. Further search using public databases did not reveal another copy of the gene in the rice genome.
The predicted amino acid sequence of the protein encoded by our clone also exhibited 59% identity to an A. thaliana clone (BAB02719) annotated as a GTPase-activating protein that also contains a C2 domain. We designated our clone Oryza sativa GTPase-activating protein 1 (OsGAP1) to reflect its putative activating activity toward GTPases. The amino acid residues encoded by OsGAP1 and BAB02719 were aligned (Fig. 1a). The position of the putative C2 domain was highlighted. Bioinformatics tools further suggested that the OsGAP1 protein does not possess a signal peptide, a targeting signal, or transmembrane domains (Supplementary Material Table S3). Western blot analysis showed that the OsGAP1 protein was found mainly in the soluble fraction (Supplementary Material Fig. S3), suggesting that the OsGAP1 protein is not tightly bound to membranes. Soluble protein fractions were used for subsequent experiments.
The OsGAP1 protein interacted with and activated the GTPase activities of the OsYchF1 protein
To show that the OsGAP1 protein indeed binds to the G-protein(s) and exhibits GTPase-activating activities, we performed yeast two-hybrid experiments to search for its interacting protein targets in rice, based on growth on selective media and positive blue color development in colony-lift assays. In the original yeast two-hybrid screen, a partial rice cDNA clone encoding a protein fragment (spanning 1175 to 1318 bp of the clone NM_001067741) was obtained. It is not an uncommon event in yeast two-hybrid experiments to obtain a partial cDNA clone in the initial screen (e.g. Moon et al., 1999). To determine the identity of the interacting protein partner obtained, we performed BlastX searches using public genome databases and found that our clone matched an annotated G-protein in rice (accession number BAD03576). The predicted amino acid sequence of BAD03576 exhibited 85% identity to an A. thaliana clone (accession number NP_174346) that is also annotated as a G-protein. The amino acid residues encoded by BAD03576 and NP_174346 were aligned (Fig. 1b). A YchF domain was found in both proteins, suggesting that the interacting partner of OsGAP1 is a YchF-type unconventional G-protein. We designated this new rice G-protein clone OsYchF1. The genomic clone of OsYchF1 was located on chromosome 8. Further search using public databases did not reveal another copy of the gene in the rice genome.
To verify the results of yeast two-hybrid experiments, co-immunoprecipitation assays were conducted. We inserted the coding region of OsYchF1 into the pGADT7-Rec vector to generate a fusion protein with an in-frame HA tag (HA-BAD03576). This recombinant construct was used for in vitro transcription followed by in vitro translation. Successful in vitro translation events were confirmed by detection of protein bands of the right size (Fig. 2a).
The in vitro translated product was incubated with protein extracts from transgenic rice lines overexpressing OsGAP1 (see below) and the anti-HA epitope tag antibody was used to pull down the protein complex (containing HA-OsYchF1 and its interacting proteins). Western blot analysis of the resulting protein complex using antibody against the OsGAP1 protein (Fig. 2b) confirmed co-immunoprecipitation of the OsGAP1 protein with the HA-OsYchF1 fusion protein. A negative control using an unrelated protein (accession number ABA98865) fused to HA did not co-precipitate OsGAP1. OsGAP1 transgenic rice (see later section) homogenate serving as a positive control for the antibodies was included. In addition, a control consisting of beads without the addition of anti-HA antibody gave no signal, confirming that there was no nonspecific binding of OsGAP1 to the beads. The interaction between the OsGAP1 and OsYchF1 proteins supported the notion that OsGAP1 may act on the G-protein(s) and OsYchF1 is one of the targets.
To demonstrate that OsYchF1 indeed carries GTPase activities that can be further activated in the presence of OsGAP1, we obtained purified OsYchF1 and OsGAP1 by constructing GST fusion proteins. The full-length coding region of OsYchF1 or OsGAP1 was fused in-frame and downstream to GST in the pGEX-4T-1 vector. Using GTP as the substrate, the amount of Pi released in the in vitro assay with the GST-OsYchF1 fusion protein was significantly higher than the background signals resulting from the GST or GST-OsGAP1 fusion protein (Fig. 2c). However, the mixing of GST-OsGAP1 and GST-OsYchF1 fusion proteins could further enhance the release of Pi (Fig. 2c). Such a phenomenon was not observed when the GST protein was used in place of GST-OsGAP1.
The expression of OsGAP1 in rice was induced by wounding
To verify that the OsGAP1 clone is differentially expressed in CBB14 as depicted from the suppression subtractive hybridization results described above, real-time PCR analyses were performed using reverse-transcribed RNA samples. An induction of the gene expression of OsGAP1 was observed when CBB14 plants were inoculated with an incompatible Xoo strain, LN44 (data not shown). However, similar induction was also found in the mock inoculation experiments in which the leaves were clipped without pathogen inoculation (data not shown). This suggests that the OsGAP1 clone may be wounding-inducible.
To further examine the effects of wounding on the expression of OsGAP1, both RNA and protein samples were collected 2, 4 or 6 d after wounding by clipping the leaves. Induction was observed in the steady-state levels of both the OsGAP1 transcript (by real-time PCR of reverse-transcribed RNA samples; Fig. 3a) and the OsGAP1 protein (by western blot analysis; Fig. 3b) in the CBB14 line, starting from day 2 after inoculation. Similar induction in SN1033 was observed from day 6 after inoculation. Whether the slight decrease in the amount of OsGAP1 transcript in SN1033 on day 4 was of significance awaited further analysis.
Overexpression of OsGAP1 in transgenic rice enhanced resistance to different strains/races of Xoo
To investigate whether OsGAP1 is involved in the rice defense response to Xoo, OsGAP1 was constitutively expressed under the maize ubiquitin promoter (Rooke et al., 2000). The construct was transformed into the rice cultivar Aichi asahi, which does not display resistance to Xoo (Liu et al., 2005). Individual plants of the T2 generation were screened for the presence of the transgene before being used for gene expression studies. Successful expression of the transgene was shown by an increased level of OsGAP1 transcripts (Fig. 4a). Expression of defense marker genes involved in different signaling pathways was also measured, including: PR1, which encodes pathogenesis-related 1 protein (Xiong et al., 2001); GRCWP, which encodes a glycine-rich cell wall protein (Xiong et al., 2001); and PBZ1, which is induced by probenazole and N-cyanomethyl-2-chloro-isonicotinamide (a group of compounds known to induce disease resistance) as well as the fungal blast pathogen Magnaporthe grisea (Nakashita et al., 2001). The expression of all three rice defense marker genes chosen was elevated in the transgenic rice lines without any pathogen inoculation (Fig. 4b). As a negative control, a transgenic rice line possessing the same vector but carrying an unrelated cDNA clone was also examined. No induction of the defense marker gene was observed (Supplementary Material Fig. S4a).
The protective effect resulting from overexpression of OsGAP1 was demonstrated by pathogen inoculation tests using three different strains/races of Xoo (LN44, T1, and P6). These strains/races exhibited different degrees of virulence on rice lines carrying different Xa genes. Broad-spectrum Xoo resistance in rice plants can be illustrated by their resistance to different Xoo strains/races (Zhang et al., 1996, 2001). Measurement of the average percentage of lesion area as a quantitative parameter indicated that transgenic lines (of the T4 generation) exhibited a significant protective effect (Fig. 4c). The phenotypes resulting from the inoculation test can be found in Supplementary Material Fig. S5. In a separate experiment, a transgenic rice line possessing the same vector but inserted with an unrelated clone was also subjected to an inoculation test. No similar protection effect was observed (Supplementary Material Fig. S4b).
Ectopic expression of OsGAP1 in transgenic A. thaliana enhances resistance to P. syringae pv. tomato DC3000 (Pst DC3000)
To determine whether the OsGAP1 cDNA originating from a monocot can also function in a dicot system, transgenic A. thaliana lines constitutively expressing OsGAP1 were constructed. The successful expression of the transgene was verified (Fig. 5a). The signal from the B-6-7 line was lowest, and was set to 1 for comparison of gene expression levels. No signal was obtained for Col-0 after prolonged PCR amplification.
The expression of four defense marker genes was tested, including PR1, PR2, PDF1.2 and Thionin 2.1 (Thi2.1). They are typical markers for the salicyclic acid (SA), jasmonate (JA) and/or ethylene (ET) pathways (Schweizer et al., 1997; Thomma et al., 1998; Spoel et al., 2003; Thibaud et al., 2004; Devoto & Turner, 2005). The expression levels of all four defense marker genes in transgenic seedlings were elevated when compared with the wild type Col-0 (Fig. 5b). The fold induction was particularly increased for PR1 and PDF1.2, which belong to two different signaling pathways (Thomma et al., 1998, 2001). In general, the degree of increase in defense marker gene expression was positively correlated with the level of OsGAP1 expression. For instance, the transgenic lines C-9-4 and J-7-5, which exhibited a higher level of OsGAP1 expression, also induced the expression of the four defense marker genes to a larger extent (compare Fig. 5a and 5b). In a separate experiment, we also demonstrated that A. thaliana transformed with an empty vector did not exhibit the same degree of gene expression induction (Supplementary Material Fig. S6).
Pst DC3000 is a pathogen commonly used to test the defense response in A. thaliana (Uknes et al., 1992; Falk et al., 1999; Kim & Delaney, 2002). When Pst DC3000 was inoculated into Col-0 or A. thaliana transformed with the empty vector cassette, disease symptoms (yellowing and necrosis) gradually appeared. Such disease symptoms were alleviated in all transgenic lines tested (Supplementary Material Fig. S7). All transgenic lines also exhibited a lower titer of pathogens in the rosette leaves, compared with the wild type Col-0 or A. thaliana transformed with the empty vector cassette (Fig. 5c).
The protective function of the OsGAP1 clone in transgenic A. thaliana requires the presence of non-expression of PR gene 1 (NPR1)
In A. thaliana, the NPR1 protein plays a key role in conducting the signals of the SA and JA pathways (Cao et al., 1994; Ryals et al., 1996; van Wees et al., 1999, 2000). An NPR1 homolog has also been reported in rice (NH1), although much less information on this is available (Chern et al., 2005). We transformed the OsGAP1 clone into the npr1-3 mutant, which is depleted in the NPR1 function (Yu et al., 2001). Successful expression of the transgene in the npr1-3 background was verified (Fig. 6a). No significant increase in the levels of any of the four defense marker genes was observed in the transgenic lines, when compared with the npr1-3 parent (Fig. 4b).
Subsequently, the transgenic lines were subjected to the challenge of Pst DC3000 together with the npr1-3 mutant and the wild type Col-0. No apparent protection phenotype was conferred by expressing the OsGAP1 clone in the npr1-3 mutant (data not shown). Consistent with the phenotype, the expression of OsGAP1 did not lower the titer of pathogens in rosette leaves of the npr1-3 mutant (Fig. 6c).
A cDNA clone (OsGAP1) encoding a C2 domain-containing protein was obtained in this study (Fig. 1a). C2 domain-containing proteins are frequently involved in signal transduction processes through interactions with other molecules (Jambunathan & McNellis, 2003). The C2 domain may be involved in phospholipid binding and some C2 domain-containing proteins will interact with membranes (Kopka et al., 1998). Phospholipids are important molecules regulating various signaling processes (Meijer & Munnik, 2003). C2 domain-containing proteins may therefore play a role in modulating or mediating phospholipid signals. The interaction between the C2 domain and phospholipids is also regulated by other factors, such as Ca2+ (Kopka et al., 1998). The OsGAP1 protein does not contain any transmembrane domains but has a C2 domain. Its interaction with membranes is weak, and the protein was found mainly in the soluble fraction when extracted using a simple procedure that separates membrane-bound and soluble proteins (Supplementary Material Fig. S3). Detailed subcellular localization of OsGAP1 should be addressed using a microscopy approach in future experiments.
Some C2 domain-containing proteins can bind to G-proteins and affect their activities (Wang et al., 1999). The gene product encoded by OsGAP1 exhibits strong homology to an annotated GAP in A. thaliana (Fig. 1a). However, no previous evidence was available to verify that this OsGAP1 homolog in A. thaliana can bind to G-proteins and activate GTPase activities.
To verify the biochemical nature of the OsGAP1 protein, we searched for its interacting protein partner. Only one candidate (OsYchF1) was identified, which is annotated as a YchF- type unconventional G-protein (Fig. 1b). Homologs of YchF proteins are highly conserved and can be found in all three domains of life (Mittenhuber, 2001). They belong to the TRAFAC class (related to translation factors) of the P-loop NTPases (Leipe et al., 2002). The crystalline structure of a bacterial YchF protein revealed binding sites for GTP and nucleic acid (Teplyakov et al., 2003). Some studies suggested that prokaryotic YchF proteins may be associated with virulence (Garbom et al., 2004) and iron utilization (Danese et al., 2004). However, the possible functions of eukaryotic YchF proteins remained largely unknown.
Using in vitro GTPase activity assays, the GTPase activity of OsYchF1 and the GTPase-activating protein activity of OsGAP1were successfully demonstrated (Fig. 2c). OsGAP1 may inhibit signals from OsYchF1 by converting it into a GDP-binding inactive form. Although it is less likely, it is also possible that GAPs can act as scaffolding proteins to form stable signaling complexes with G-protein couple receptors and G-proteins (Assmann, 2002). Whether OsGAP1 inhibits or enhances OsYchF1 activities through a physical interaction must be further investigated before a definitive conclusion can be drawn.
Although we showed that constitutive expression of OsGAP1 can enhance the defense response in plants (see below) and demonstrated GTPase-activating effects on OsYchF1, the possible roles of OsYchF1 in the disease response remain unclear. In addition to GTPase activities, other possible effects of OsGAP1 on OsYchF1 at the in planta level (such as the regulation of gene expression) will be addressed in future investigations to further delineate the interaction between OsGAP1 and OsYchF1 in relation to plant defense signaling.
One central question to be addressed is whether OsGAP1 is involved in the plant defense response. The first indication of such involvement lies in the wounding inducibility of OsGAP1. In the CBB14 rice line that contains the R locus Xa14, the expression of OsGAP1 was induced by wounding (Fig. 3). This induction of gene expression occurred much more rapidly in CBB14 than in the susceptible recurrent parent SN1033. As the invasion of the pathogen Xoo was via wounds (Ronald, 1997) and some Xa genes (e.g. Xa1) are wounding-induced (Song & Goodman, 2001), it is possible that the signaling pathways of rice bacterial blight defense responses and wounding responses may overlap (Song & Goodman, 2001). As the Xa14 gene is yet to be cloned, we cannot verify at this point whether the induction of Xa14 under pathogen inoculation or wounding will subsequently lead to the rapid induction of OsGAP1. OsGAP1 is unlikely to be the unidentified Xa14 as these genes are located on chromosome 2 (see the Results) and chromosome 11 (Zhang et al., 1996), respectively.
To demonstrate the involvement of OsGAP1 in the defense signal transduction pathway, we performed gain-of-function experiments in transgenic rice (the native plant) (Fig. 4) and transgenic A. thaliana (a dicot model; Fig. 5). We tested the defense responses using two parameters: the expression of defense marker genes and performance under pathogen attacks.
Overexpression of OsGAP1 in transgenic rice raised the basal transcriptional levels of all three defense marker genes (PR1, GRCWP, and PBZ1) tested (Fig. 4b). PR1 and PBZ1 are two PR genes that are also induced by overexpression of NH1 (the rice homolog of NPR1 in A. thaliana), a key signaling component in rice defense responses (Chern et al., 2005). The PBZ1 gene was originally identified as a gene induced by the chemical probenazole (PBZ) (Midoh & Iwata, 1996). PBZ is a plant defense chemical activator commonly used for broad-spectrum protection against rice fungal blast caused by M. grisea and rice bacterial blight caused by Xoo in Asia (Midoh & Iwata, 1996; Yoshioka et al., 2001). Its application to rice can enhance the activities of enzymes in the phenylpropanoid pathway (for SA, lignin and phytoalexin biosynthesis) and increase production of anti-conidial germination substances including intermediates and products in the octadecanoid pathway (for JA and oxylipin biosynthesis) (Midoh & Iwata, 1996; Yoshioka et al., 2001).
In plants, the defense response is a complicated regulatory process involving various plant hormones, including SA, JA, ethylene (ET), brassinosteroids, and abscisic acid (Píeterse & van Loon, 2004). Although the endogenous SA concentration in rice is very high and pathogen inoculation will not further increase the content of SA (Silverman et al., 1995), the presence of NH1 (its homolog NPR1 is a common regulatory component of SA and JA pathways in A. thaliana) (Cao et al., 1994; Ryals et al., 1996; van Wees et al., 1999, 2000) indicates a role of SA signaling in the rice defense response. The induction of expression of different defense marker genes by overexpression of OsGAP1 suggests that OsGAP1 may be involved in more than one signaling pathway to confer broad-spectrum disease resistance.
The protective effects of OsGAP1 against Xoo were demonstrated in an inoculation test (Fig. 4c). While the Xa14 locus can confer resistance only to LN44 (and not to T1 and P6) (Zhang et al., 1996), overexpression of OsGAP1 led to broad-spectrum resistance to all three Xoo races/strains tested. However, the resistance conferred by OsGAP1 was not a strong immune response, but rather a quantitative alleviation (with a significantly smaller lesion area on the leaves but not total resistance). It is not surprising that some signal component genes have been identified as major quantitative resistance loci (Wen et al., 2003).
To further illustrate the broad-spectrum resistance conferred by OsGAP1, the dicot model plant A. thaliana was employed. Transgenic A. thaliana lines ectopically expressing OsGAP1 exhibited an increase in the expression of the defense marker genes PR1, PR2, PDF1.2 and Thi2.1 (Fig. 5b). In A. thaliana, SA and JA/ET will initiate two interacting defense signaling pathways. Expression of PR1 and PR2 and expression of PDF1.2 and Thi2.1 signify the activation of the SA and JA/ET pathways, respectively (Thomma et al., 1998; Cheong et al., 2002). However, antagonistic effects on the expression of defense marker genes occur between the SA and JA/ET pathways (Turner et al., 2002; Beckers & Spoel, 2006; Balbi & Devoto, 2008). The expression of PR1 and PR2 can be further distinguished: while PR1 is strongly repressed by JA, the repression of PR2 is controlled by multiple factors (e.g. sucrose concentration) (Seo et al., 1995; Schweizer et al., 1997; Devoto & Turner, 2003; Thibaud et al., 2004). Moreover, while PDF1.2 is dependent on both the JA and ET signaling pathways, Thi2.1 is induced by JA and repressed by ET (Devoto & Turner, 2003). Similar to the results in transgenic rice described above, defense marker genes involved in different signaling pathway were induced by ectopic expression of OsGAP1, supporting the notion that the functions of OsGAP1 may be involved in more than one pathway.
The pathogen inoculation test using Pst DC3000 in transgenic A. thaliana lines showed a clear protection effect resulting from the ectopic expression of OsGAP1 (Fig. 5c), further confirming the ability of OsGAP1 to confer broad-spectrum resistance to bacterial pathogens and suggesting the presence of a similar defense pathway in A. thaliana.
The lower expression levels of defense marker genes in the transgenic lines B-6-7 and D-2-9 compared with those in the transgenic lines C-9-4 and J-7-5 (Fig. 5b) can be correlated with the lower levels of OsGAP1 transcripts in B-6-7 and D-2-9 (Fig. 5a). However, while the line C-9-4, which had the highest expression of OsGAP1, exhibited the lowest pathogen titer of all the transgenic lines (Fig. 5c), the differences among the transgenic lines were not statistically significant. This is partially a result of the high variance of the data. Moreover, the possibility cannot be excluded that OsGAP1 may act indirectly and its protective functions may require other components that are limiting.
Making use of the genetic resources in A. thaliana, we tried to position OsGAP1 relative to some known signaling components. Studies of the expression of defense marker genes in both rice and A. thaliana indicated that OsGAP1 may be related to both the SA and JA/ET pathways. In A. thaliana, SA signals alter the redox state of NPR1 and the transcription factor TGA1 (a basic leucine zipper transcription factor interacts with the TGACGT motif; Dong, 2004; Píeterse & van Loon, 2004; Fobert & Després, 2005). In their reduced states, both the NPR1 and TGA1 proteins localize in the nucleus and interact. The activated transcriptional processes lead to the expression of SA-induced genes. Nevertheless, NPR1 also mediates the signals from the JA pathway in induced systemic responses (ISRs) (Píeterse et al., 2002; Spoel et al., 2003; Píeterse & van Loon, 2004). Both NPR1 (named NH1 in rice) and TGA homologs have been identified in rice (Chern et al., 2001). NH1 in rice plays an important role in resistance to Xoo (Chern et al., 2005) (such resistance is also conferred by the Xa14 resistance locus in CBB14, whereas OsGAP1 is differentially expressed under wounding).
To determine the relative positions of OsGAP1 and NPR1 in A. thaliana defense response signaling, we made use of the npr1-3 mutant which is depleted in the NPR1 function (Yu et al., 2001). In a previous study, the presence of NPR1 was found to be essential for the functioning of a rice defense response component in A. thaliana (Cheung et al., 2007). In the npr1-3 background, OsGAP1 did not show the same induction of the four defense marker genes as in Col-0 (compare Fig. 6b with Fig. 5b). As in the transformation host npr1-3, the expression of PR1 was repressed in all transgenic lines with the npr1-3 mutation background. For a couple of lines, a 2–3-fold increase in PDF1.2 was observed. However, this became insignificant when compared with the fold induction of OsGAP1 lines with the Col-0 background. More importantly, the protective effects of OsGAP1 expression against Pst DC3000 diminished (Fig. 6c). Based on these results, we conclude that the function of the OsGAP1 protein in A. thaliana requires the presence of NPR1 and hence the OsGAP1 protein probably acts upstream from NPR1 in the defense mechanism, or functions to inhibit a negative regulator of NPR1. Whether OsGAP1 also requires NH1 (the rice NPR1 homolog) to function in rice remains to be confirmed. However, the observation that overexpression of either OsGAP1 or NH1 in rice can increase the expression of the defense marker genes PR1 and PBZ1 (Fig. 3b, and Chern et al., 2005) suggests some form of linkage between these two signaling components in the native system. For a working model, we propose that the protective function of OsGAP1 is mediated by inactivation of OsYchF1, while OsYchF1 negatively regulates a signaling step upstream from NPR1/NH1 or positively regulates a negative regulator downstream from NPR1/NH1.
In summary, we have identified a novel wounding-inducible cDNA clone (OsGAP1) in rice which encodes a C2 domain-containing protein and behaves as a GAP. OsGAP1 interacts with and activates the GTPase activities of the unconventional G-protein OsYchF1. Together with the findings that constitutive expression of OsGAP1 can enhance the defense response in both the native plant and a heterologous plant system, our results suggest the presence of a novel defense mechanism, probably acting via G-protein-related/mediated signal transduction pathways. Furthermore, the signaling pathways may also involve NPR1/NH1.
This work was supported by Hong Kong RGC Earmarked Grant CUHK4273/02M (to H-ML), the Hong Kong UGC AoE Plant & Agricultural Biotechnology Project AoE-B-07/09 and a SHARF Grant (to H-ML and SS-MS). Professors C. Deprés (Brock University) and C. Lo (Hong Kong University) kindly provided the seeds of the npr1-3 mutant and the pathogen strain Pseudomonas syringae pv. tomato DC3000, respectively. Technical support provided by T. Lau and S.-M. Chow (Chinese University of Hong Kong) is also greatly appreciated. We also thank Dr Q. Liu (Yangzhou University) for propagating the rice seed stocks.