Dual Regulation Role of GH3.5 in Salicylic Acid and Auxin Signaling during Arabidopsis- Pseudomonas syringae Interaction 1

Salicylic acid (SA) plays a central role in plant disease resistance, and emerging evidence indicates that auxin, an essential plant hormone in regulating plant growth and development, is involved in plant disease susceptibility. GH3.5 , a member of the GH3 family of early auxin-responsive genes in Arabidopsis, encodes a protein possessing an in vitro adenylation activity on both indole-3-acetic acid (IAA) and SA. Here we show that GH3.5 acts as a bifunctional modulator in both SA and auxin signaling during pathogen infection. Overexpression of the GH3.5 gene in an activation-tagged mutant gh3.5-1D led to elevated accumulation of SA and increased expression of PR-1 in local and systemic tissues in response to avirulent pathogens. In contrast, two T-DNA insertional mutations of GH3.5 partially compromised the systemic acquired resistance (SAR) associated with diminished PR-1 expression in systemic tissues. The gh3.5-1D mutant also accumulated high levels of free IAA after pathogen infection and had impaired different R gene-mediated resistance, which was also observed in the GH3.6 activation-tagged mutant dfl1-D that impacted the auxin pathway, indicating an important role of GH3.5/GH3.6 in disease susceptibility. Furthermore, microarray analysis showed that the SA and auxin pathways were simultaneously augmented in gh3.5-1D after infection with an avirulent pathogen. The SA pathway was amplified by GH3.5 through inducing SA-responsive genes, basal defense components; while the auxin pathway was derepressed through up-regulating IAA biosynthesis and down-regulating auxin repressor genes. Taken together, our data reveals novel regulatory functions of GH3.5 in the plant-pathogen interaction.

Previous studies have shown that the levels of IAA, the primary plant auxin, increased in plant tissues infected by P. syringae pv. tomato (Pst) DC3000 (O'Donnell et al., 2003). Moreover, the type III effector AvrRpt2 modulates host IAA levels to promote pathogen virulence and disease development in Arabidopsis (Chen et al., 2004;Kunkel et al., 2005). Recent microarray analysis has shown that infection with Pst DC3000 activates genes related to IAA biosynthesis, and represses Aux/IAA family and auxin transporter genes, suggesting that Pst DC3000 impacts auxin signaling probably through activating IAA production, altering IAA movement, and derepressing the auxin pathway (Thilmony et al., 2006). Conversely, the Arabidopsis microRNA miR393a is induced by the flagellin-derived peptide flg22 and represses auxin signaling through down-regulating the auxin receptor genes, resulting in increased resistance to P. syringae (Navarro et al., 2006). Together, these results suggest that auxin plays an important role in the disease susceptibility pathway(s).
However, little is known about genes that regulate auxin signaling in response to pathogen infection and the downstream targets of an altered auxin signaling responsible for enhanced disease susceptibility.
Auxin rapidly induces numerous genes called early auxin response genes. The best-characterized early auxin response genes include three major classes: Aux/IAAs, SAURs and GH3s (Hagen and Guilfoyle, 2002). The Arabidopsis GH3 family consists of 19 members, six of which are known to adenylate IAA in vitro (Staswick et al., 2002). Further studies have shown that GH3 genes encode IAA-amido synthetases that are involved in auxin homeostasis through conjugating amino acids to IAA (Staswick et al., 2005). However, the physiological function and regulatory target genes of these GH3 members is still largely unknown. It was also shown that GH3.5 (At4g27260) adenylated both IAA and SA in vitro (Staswick et al., 2002), suggesting that GH3.5 could function in modulating and integrating both auxin and SA signaling in the plant-pathogen interaction.
In this work, we investigate the functions of GH3.5 in response to P. syringae infection. We show that GH3.5 positively regulates the SA signaling pathway in plant defense, and modulates the auxin pathway to enhance host susceptibility. Our findings demonstrate that GH3.5 is a key modulator that both positively and negatively affects various aspects of plant defense, revealing another dimension to the complex and dynamic plant-pathogen interaction. Figure 3B, gh3.5-1D plants accumulated much more both free and total SA in local and systemic tissues than wild-type plants at 48 hpi, indicating that GH3.5 overexpression positively regulates SA accumulation in both local and systemic tissues. A similar result was also observed in the plants infected with Pst DC3000(avrRpt2) (Supplemental Fig. S3). Taken together, our results demonstrate that the PR-1 transcript and SA levels were increased rather than decreased in gh3.5-1D mutant, suggesting that GH3.5 overexpression might also enhance a susceptibility-related pathway that contribute to the compromised R-mediated local resistance.

gh3.5-1D plants after infection with Psm(avrRpm1). As shown in
Previous in vitro analysis showed that GH3.5 exhibited adenylation function on SA (Staswick et al., 2002). We indeed observed that salicyloylaspartate (SA-Asp), the only identified endogenous SA-amido conjugate in planta (Bourne et al., 1991), accumulated to higher levels in the gh3.5-1D leaves infected with Psm(avrRpm1), as compared to Col-0 (Supplemental Fig. S4). Since the only identified active SA form in vivo is free SA (Sticher et al., 1997), there may be a possibility that accumulated SA-Asp might serve as a pool for the increased SA generation observed in gh3.5-1D.

SAR Response Is Impaired in Loss-of-Function Mutants of GH3.5
To further investigate the role of GH3.5 in response to pathogen, we obtained two T-DNA insertion lines of GH3.5 from the Salk T-DNA collection (Alonso et al., 2003). The SALK_014376 T-DNA line contains the T-DNA insertion 260-bp upstream of the GH3.5 start codon and was previously named gh3. 5-1 (Staswick et al., 2005), while SALK_151766 contains the T-DNA insertion 1809-bp downstream of the GH3.5 start codon and was designated gh3.5-2 (Fig. 4A). RT-PCR analysis showed that the GH3.5 transcript was induced in wild-type plants upon infection with Psm(avrRpm1) (Fig. 4B and see below). Although no difference in the GH3.5 transcript level was detected between wild-type and gh3.5-1 plant at 0 hpi, Psm(avrRpm1) infection led to no induction of the GH3.5 gene, indicating that gh3.5-1 is a partial loss-of-function mutant. In contrast, no GH3.5 transcript was detected in gh3.5-2 regardless of pathogen infection, indicating that gh3.5-2 is a loss-of-function mutant. Consistent with the predicted functional redundancy among the members of the GH3 family (Staswick et al., 2005), both mutants display no obvious morphological phenotypes compared with the wild-type, although they showed slight increase in sensitivity to exogenous auxins (Supplemental Fig. S1).

Overexpression of DFL1 Impairs R-Mediated and Basal Resistance
Despite the accumulation of SA and the PR-1 transcript in the gh3.5-1D mutant, it still exhibited compromised R-mediated local resistance, and had a normal rather than increased SAR and basal resistance. The data suggest that GH3.5 overexpression might enhance a susceptibility-related pathway that counteracts SA-mediated disease resistance. Given the facts that the protein product of the GH3.5 gene can conjugate IAA to amino acids in vitro and a recent revelation of the auxin involvement in disease susceptibility (Staswick et al., 2002;Navarro et al., 2006), we hypothesized that an altered auxin pathway is likely responsible for the enhanced local susceptibility phenotype of the gh3.5-1D mutant. To test our hypothesis, we analyzed the disease resistance of a similar activation-tagged dwarf mutant dfl1-D (Nakazawa et al., 2001). DFL1 (GH3.6) is the closest family member of GH3.5, and a previous study showed that DFL1 exhibited IAA-amido synthetase activity similar to GH3.5 in vitro (Staswick et al., 2005), suggesting that both proteins regulate IAA homeostasis/signaling through a similar mechanism. Unlike GH3.5, GH3.6 did not show activity on SA in vitro (Staswick et al., 2002), so dfl1-D should be suitable to study the effect of the auxin pathway on disease susceptibility. Consistent with the in vitro result, we did not observe changes in SA levels in dfl1-D compared to the corresponding wild-type (Fig. 5A).
Bacterial growth assay showed that dfl1-D was indeed more susceptible to Pst DC3000(avrRpt2), Psm(avrRpm1) and Pst DC3000(avrRps4), resulting in more bacterial growth than in the corresponding wild-type controls at 3 dpi (Fig. 5B). This indicates that dfl1-D also impairs the same R gene-mediated resistance. In contrast to gh3.5-1D, basal resistance to Pst DC3000 (mock) was decreased in dfl1-D plants (Fig.   5C). Although dfl1-D was capable of mounting SAR, enhanced growth of Pst DC3000 was still apparent in the secondary infected leaves of dfl1-D (Fig. 5C), showing a phenotype similar to that of the previously reported eds (enhanced disease susceptibility) mutants (Rogers and Ausubel, 1997).

IAA Levels Are Increased in gh3.5-1D during Pathogen Infection
Considering GH3.5 has IAA-amido synthase activity (Staswick et al., 2002(Staswick et al., , 2005, we measured levels of free IAA and IAA-amido conjugates in gh3.5-1D in response to virulent and avirulent pathogen attack. As shown in Table 1, the IAA levels were not different between gh3.5-1D and Col-0 with mock treatment, similar to the observation with dfl-1D (Staswick et al., 2005). Infection with Pst DC3000 but not Pst DC3000(avrRpt2) greatly increased the IAA levels in wild-type plants, suggesting an important role for IAA in susceptibility during virulent pathogen infection.
Surprisingly, the IAA levels were significantly increased in gh3.5-1D plants infected with Pst DC3000(avrRpt2) as compared to that of the wild-type plants, and were further increased when the mutant was infected with Pst DC3000, indicating that GH3.5 overexpression positively regulates IAA accumulation during pathogen infection. The increased IAA level might contribute to the enhanced susceptibility to avirulent pathogen. We did not observe a significant elevation of the three known IAA conjugates in gh3.5-1D compared with the wild-type controls, indicating GH3.5 does not regulate the accumulation of those amido-conjugated IAAs under the conditions of our study or they were immediately hydrolyzed after formed. We further measured IAA levels in the systemic tissues of the gh3.5-1D and wild-type plants that were either mock-treated or pre-inoculated with Pst DC3000(avrRpt2). There was no difference in IAA levels between the gh3.5-1D and wild-type (data not shown), indicating that gh3.5-1D did not affect the systemic IAA levels during SAR establishment.

GH3.5 Is Induced by Pathogen and SA
Because GH3.5 plays diverse roles in the plant-pathogen interactions, its expression pattern could be indicative of its function. Northern hybridization indicated that the induction of GH3.5 was slightly different in responses to the avirulent and virulent strains in wild-type plants. Induction was detectable as early as 6 h with a major peak at 48 h after inoculation with both avirulent Pst DC3000(avrRpt2) and Psm(avrRpm1) (Fig. 6, A and B). In contrast, induction was delayed until 24 h with a peak occurring around 2 dpi by Pst DC3000 and 3 dpi by Psm (Fig. 6B). It was previously shown that GH3.5, also known as AtGH3a, was induced by IAA (Tanaka et al., 2002). We next tested if GH3.5 could be induced by SA. As shown in Figure 6C, SA induced GH3.5 as early as 3 h with the peak level at 12 h after treatment. Consistent with these results, GH3.5 expression was up-regulated in cpr1 and cpr6, and slightly increased in cpr5 ( Fig. 6D), which are known to contain high levels of SA (Clarke et al., 2000). It should be noted that the degree of GH3.5 induction is much lower by SA than by pathogen (Fig. 6, A and B), suggesting that pathogen-induced expression of the GH3.5 gene might be also mediated by an SA-independent pathway.

Transcriptional Profiling of gh3.5-1D Plants in Response to Pathogen
To reveal the underlying molecular mechanism of GH3.5 action in modulating the SA and auxin pathways, we performed transcriptional profiling of wild-type Col-0 and gh3.5-1D plants with or without infection by Pst DC3000(avrRpt2) using the Affymetrix Arabidopsis ATH1 GeneChip. With 3 biological replicates of pairwise experiments (Supplemental Table S1) and high stringent selection criteria (Thilmony et al., 2006), we identified 273 and 831 differentially regulated genes in uninoculated gh3.5-1D and gh3.5-1D infected for 48 h, in comparison with Col-0 plants respectively, and a total of 1013 non-redundant genes that exhibited statistically significant (≥ 2-fold) changes of expression levels (Supplemental Table S2 and S3).
For more detailed analyses, we first performed hierarchical clustering to study the genes related to auxin biosynthesis and response ( Fig. 7  genes functioning as repressors in auxin signaling (Tiwari et al., 2001) were down-regulated, consistent with GH3.5 role in the auxin pathway. A great number of auxin-related genes were either up-or down-regulated in gh3.5-1D infected with pathogen. Among those up-regulated genes were three IAA biosynthesis-related genes (ASA1, CYP79B2 and AAO1), two GH3 genes (DFL1/GH3.6 and GH3.3), and two IAA-amido hydrolase genes (ILR1and ILL5). The up-regulation of the IAA biosynthesis-related and IAA-amido hydrolase genes could contribute to the higher IAA levels observed in gh3.5-1D in response to pathogen (Table 1). In contrast, down-regulated genes included fifteen SAURs and five Aux/IAAs. Taken together, these results suggest that GH3.5 might regulate auxin signaling by increasing IAA biosynthesis and derepressing the auxin pathway during pathogenesis.
We next profiled metabolic pathways using MAPMAN software (Thimm et al., 2004). The genes involved in the breakdown of starch, sucrose, lipids, and nucleotides, and the metabolism of amino acids were remarkably up-regulated in gh3.5-1D at 48 hpi ( Fig. 8). This suggests high levels of nutrient accumulation during pathogen infection in gh3.5-1D plants. Moreover, the MAPMAN analysis revealed that many transporter genes were up-regulated in gh3.5-1D as compared to the wild-type at 48 hpi (Supplemental Fig. S6), including several ABC-type transporter genes and genes encoding transporters for oligopeptides, amino acids and sugars. The strong activation of these genes could result in nutrient efflux to benefit invading pathogen in gh3.5-1D.
Consistent with a positive role of gh3.5-1D in the SA-mediated defense pathway, many defense-related genes were more strongly activated by pathogen in gh3.5-1D as compared to wild-type plants (Table 2). These genes could be grouped into three main categories: (1) SA-induced genes such as α -DOX1 (de León, 2002), PAD3, PAL1 and WRKY transcription factor genes; (2) genes involved in cell wall modification known to be a critical aspect of the plant basal defense, and (3) other defense-related genes such as FRK1 (Asai et al., 2002). WRKY induction was particularly interesting, since some WRKYs such as WRKY18 were recently identified to play important roles in SAR (Wang et al., 2006).
To confirm the expression patterns of the genes identified by microarray analysis, we next conducted RT-PCR experiments to detect the expression levels of four defense-related and three auxin-related genes in gh3.5-1D and wild-type plants infected with Pst DC3000(avrRpt2). As expected, these genes were differentially induced in gh3.5-1D compared to wild-type plants (Supplemental Fig. S7).

GH3.5 is a Novel Positive Modulator of SA Signaling
SA-mediated SAR has been extensively studied, revealing that SA is necessary and sufficient for induction SAR. It is well established that the transcription cofactor protein NPR1-regulated expression of PR genes is required for the induction of SAR (Durrant and Dong, 2004). Our current study reveals that GH3.5 is a novel positive modulator in regulating SA signaling during SAR establishment, with several lines of evidence. First, the systemic induction of PR-1 by avirulent strains was increased and associated with higher levels of SA in gh3.5-1D compared with Col-0. Second, systemic PR-1 induction was repressed and SAR was partially compromised in the two GH3.5 insertional mutants. Third, expression of WRK18, a positive regulator of SAR (Wang et al., 2006), was induced in gh3.5-1D compared with the wild-type after infection with Pst DC3000(avrRpt2). Therefore, GH3.5 positively regulates the SA-mediated SAR most likely by enhancing SA biosynthesis, up-regulating expression of WRKY18, and consequently activating downstream genes.
SA also plays a central role in R-mediated local resistance (Nimchuk et al., 2003).
We found that SA levels were significantly elevated in gh3.5-1D than the wild-type.
Consistent with the increased SA levels in gh3.5-1D after pathogen infection, phenylalanine ammonia-lyase gene PAL1 was up-regulated. Furthermore, the global expression experiment revealed that many defense-responsive genes were up-regulated in gh3.5-1D as compared to Col-0 after infection with Pst DC3000(avrRpt2), which included four known SA-responsive WRKY genes (Dong et al., 2003;Wang et al., 2006), consistent with the notion that GH3.5 plays a positive role in the SA signaling pathway. Moreover, we also observed up-regulation of the genes involved in cell wall modification and basal defense response. Cell wall modification is recognized as a critical aspect of the plant basal defense, suggesting that GH3.5 is also involved in basal defense. Consistent with this deduction, the basal defense marker gene, FRK1 (Asai et al., 2002), was greatly induced in gh3.5-1D. In contrast to disease resistance, little is known about the mechanism of plant susceptibility (Vogel et al., 2002;Nomura et al., 2005). It has long been recognized that many pathogenic microbes can produce IAA, which is proposed to alter host cellular processes to favor pathogen infection (Yamada, 1993;Jameson, 2000;Spaepen et al., 2007). Recent molecular genetic studies and global gene expression experiments have implicated auxin as an important disease susceptibility factor (Navarro et al., 2006;Siemens et al., 2006;Thilmony et al., 2006). Consistent with those studies, our result further reveals an important role for GH3.5, an early auxin-responsive gene, in the auxin-elicited susceptibility.

GH3.5 Plays an Important Role in the Auxin-Elicited Susceptibility
It is intriguing that GH3.5 positively regulates SA signaling but weakens different R gene-mediated local resistance to diverse avirulent strains in gh3.5-1D. We hypothesized that R-mediated resistance was counteracted by auxin-mediated susceptibility due to GH3.5 overexpression. Indeed, dfl1-D, a similar activation-tagged dwarf mutant that overexpresses GH3.6 showed impaired R-mediated resistance, which was also more susceptible to virulent strains than the wild-type. The predicted decrease of basal disease resistance is likely not apparent in gh3.5-1D plants due to counteraction by the simultaneously enhanced SA pathway.
This evidence also accounts for the normal R-induced SAR observed in gh3.5-1D plants in which a predicted enhanced SAR response, indicated by increased SA levels and PR-1 induction, was counteracted by susceptibility events.
Regulation of auxin biosynthesis is a key step in the auxin-mediated response.
Interestingly, GH3.5 positively regulates IAA accumulation during pathogen infection revealed by the IAA levels, and our microarray assays showed that several genes involved in IAA biosynthesis were more strongly up-regulated in gh3.5-1D than in wild-type plants after pathogen infection. This finding suggests that GH3.5 regulates IAA biosynthesis-related genes to increase IAA accumulation in planta during infection, in addition to IAA generation by P. syringae that harbors the iaaM and iaaH genes encoding enzymes catalyzing the conversion of tryptophan to IAA via indole-3-acetamide (Glickmann et al., 1998;Buell et al., 2003 microarray assays also showed that GH3.5 overexpression suppresses expression of Aux/IAA genes, thus likely derepressing the auxin signaling pathway, consistent with the previous observation in wild-type Col-0 plants infected with Pst DC3000 and Plasmodiophora brassicae (Siemens et al., 2006;Thilmony et al., 2006). An important outcome of regulation of auxin signaling by GH3.5 in infected gh3.5-1D plants is the up-regulation of genes involved in metabolism and transport of nutrition across the plasma membrane (Fig. 8, Supplemental Fig. S6). How P. syringae, which colonizes the apoplast, modifies the host metabolism to promote nutrient synthesis or release to sustain its growth is a fundamental question that remains to be answered (Alfano and Collmer, 1996). Our current study probably provides new clues regarding this aspect.

Could GH3.5 Regulate SA and Auxin Homeostasis during Pathogen Infection?
In plant, free SA and IAA accumulate at very low levels and most of SA and IAA are found in conjugated forms (Sticher et al., 1997;Ljung et al 2002). SA can form SA-glucoside (SAG) and methyl salicylate in various plants (Sticher et al., 1997). The roles of these conjugated forms of salicylates are diverse: SAG is not active in disease resistance and considered a main storage form of SA (Ryals et al., 1996); methyl salicylate acts as an airborne signal in SAR (Shulaev et al., 1997). SA-Asp, the only known amido derivative of SA, has been reported in grape and bean (Steffan et al., 1988;Bourne et al., 1991). As GH3.5 exhibited adenylation activity on SA in vitro, it is possible GH3.5 can form SA-amino acid conjugates in vivo. Consistent with this prediction, we have shown that SA-Asp is also present in Arabidopsis and GH3.5 positively regulates the SA-Asp levels in gh3.5-1D. However, we did not observe a lower SA-Asp level in the GH3.5 loss function mutant (Supplemental Fig. S4), suggesting that GH3.5 might be not the only amido synthase for SA-Asp. Since there is not any available information on SA-Asp function in planta, it is currently unclear whether SA-Asp is involved in defense responses in gh3.5-1D. Recently, another member of GH3 family, GH3.12/PBS3, was shown to regulate SA and SAG levels, suggesting its important role in SA metabolism (Nobuta et al., 2007;Jagadeeswaran et al., 2007). Further analysis of SA-Asp might provide insight into the biochemical basis of SA metabolism and signaling.
IAA can form conjugates with sugars, amino acids, and small peptides. It has shown that IAA conjugation is involved in IAA transport, storage and metabolism (Ljung et al., 2002). Earlier studies reported that certain members of the GH3 family, including GH3.5, encode IAA-amido synthases that might maintain IAA homeostasis by converting excess auxin into amino acid conjugates that are either inactive or degraded (Staswick et al., 2005). This finding has contributed to our understanding of auxin homeostasis. However, overexpression of GH3.6 in dfl1-D does not alter the IAA level despite the increased IAA-Asp accumulation (Staswick et al., 2005).
Similar to dfl1-D, gh3.5-1D accumulated the same level of IAA as the wild-type.
Moreover, the IAA levels were significantly elevated after induced by P. syringae.
One possible interpretation is that GH3.5 may act differently in regulating IAA homeostasis in different plant physiological processes. In support of this hypothesis, the transgenic tobacco overexpressing the iaaL gene, which encodes an IAA-lysine synthase converting IAA to IAA-lysine (Roberto et al., 1990), could decrease or unaffect or increase IAA levels depending on the organs and developmental stages of plant (Spena et al., 1991). Unlike in dfl1-D, the amount of IAA-Asp in gh3.5-1D was similar to the wild-type even after infection by pathogen. The induction of two IAA-amido hydrolase ILR1 and ILL5 genes by Pst DC3000(avrRpt2) in our microarray assays, which can cleave IAA-amino acid conjugates to free IAA (Bartel and Fink, 1995), may reflect the possibility that IAA-amino acid conjugates transiently form and rapidly hydrolyze in gh3.5-1D, explaining our failure to detect the elevation of three typical IAA-amino acid conjugates in these tissues. On the other hand, gh3.5-1D was less responsive than wild-type while the T-DNA knockout mutants became more sensitive to exogenous IAA and NAA (Supplemental Fig. S1).
This suggests that the GH3.5 protein could remove excessive auxin through its activity of amido synthetase.

Dual Roles of GH3.5 in the Arabidopsis-P. syringae Interaction
The results presented here reveals that GH3.5 acts as a bifunctional modulator in two distinct signaling pathways during infection by P. syringae: the SA-mediated pathway for disease resistance and the IAA-mediated pathway for disease  and Col-0, because IAA levels are greatly different with growth times and tissues as previously described (Spena et al., 1991).
Based on our current data, we propose a functional model for GH3.5 in the Arabidopsis-P. syringae interactions (Fig. 9). During the virulent pathogen infection in the wild-type, bacterial growth requires an increase in nutrition. Pathogen induces GH3.5 expression, while GH3.5 modulates the auxin pathway to enhance disease susceptibility by activating IAA production and derepressing the auxin pathway. At the same time, the increased IAA level feedback regulates GH3.5 expression, resulting in an amplifying effect on auxin response (Fig. 9A). To maintain the endogenous IAA at an appropriate level, GH3.5 may act as an IAA-amido synthetase to form IAA-amido conjugates which are either degraded or hydrolyzed to cycle IAA.
However, this infection process must pay a penalty: GH3.5 simultaneously modulates the SA pathway to activate defense-related genes, resulting in increased defense

Overexpression of GH3.5
The full-length GH3.5 cDNA was generated by RT-PCR with the primers,  Methods for bacterial growth assays were performed as described (Katagiri et al., 2002). Heterozygous gh3.5-1D (+/-) plants were used for all of the experiments in the bacterial growth assay unless otherwise indicated.

SAR Assays
Two lower leaves of five-week-old plants were inoculated with Psm(avrRpm1) bacterial suspension in 10 mM MgCl 2 at 10 7 cfu mL -1 . Three upper leaves were infiltrated with Pst DC3000 at 10 5 cfu mL -1 three days later. Bacterial growth was assayed in the secondary infected leaves. Bacterial growth titers were counted at 0 and 3 days after inoculation.

RNA and Protein Analysis
Total RNA was isolated from leaf tissues using TRIzol reagent according to the manufacturer's instructions (GIBCO BRL). Ten µg RNA samples were separated on a 1 % formaldehyde-agarose gel, and then blotted onto Hybond-N + membranes (Amersham). A 353-bp fragment of the GH3.5 transcript was amplified from genomic DNA using the primers 5′-TAATCAGTATAAGACGCCGAGATGC-3′ and 5′-TCGAGAAAGAGTGATGAGAGTTGGTT-3′, and was labeled with [α-32 P] dCTP using a random primer labeling kit (Takara) for hybridization and autoradiography.
Northern blot analysis was also performed to determine the transcript levels of the pathogenesis-related genes PR-1 (At2g14610) and PDF1.2 (At5g44420). The filters were reprobed with a 2.5-kb fragment of 18S Arabidopsis rDNA for loading normalization. The same GH3.5 primers were also used for RT-PCR to detect the (612 amino acids) was ligated into pET-32a to produce the fusion protein in the Escherichia coli strain DE3. The GH3.5 fusion protein was used to immunize rabbit to produce antiserum. Western blotting was performed using the SuperSignal West chemiluminescence kit according to the manufacturer's protocol (Pierce).

SA and SA-Asp Measurement
Leaves of Col-0, gh3.5-1D, gh3.5-2, dfl1-D and Ler (the wild-type for dfl1-D) plants infected with Psm(avrRpm1) or PstDC3000(avrRpt2) at 10 7 cfu ml -1 or with 10 mM MgCl 2 (mock) were harvested at 0 (control) and 48 hpi. Free and total SA were extracted from leaf tissues and analyzed by HPLC according to the method described by Dewdney et al. (2000). For SA-Asp measurement, standard SA-Asp was synthesized according to the method described by Bourne et al. (1991), and quantification of SA-Asp was also performed by HPLC accordingly. DC3000 and Pst DC3000(avrRpt2) at 10 5 cfu mL -1 or with 10 mM MgCl 2 (mock).

Extraction and Quantification of Auxins
Leaves were harvested three days after inoculation and lyophilized. Quantitative analysis of IAA and IAA-amino acid conjugates were performed as previously described (Staswick et al., 2005).

Microarray Analysis
Leaves of plants of 5-week-old wild-type and gh3.5-1D were inoculated with Pst Raw data were analyzed with Affymetrix GeneChip® Operating Software (GCOS, Version 1.4) using Affymetrix default analysis settings and global scaling as normalization method. The trimmed mean target intensity of each array was arbitrarily set to 100. Data was then compared between sample chips from the same biological replicate producing a signal log 2 ratio (SLR), a change call and change P-value which calculated from the GeneChip fluorescence signal intensity data using the Affymetrix GCOS software. SLRs, change calls and change P-values were determined for each inoculation sample compared with its corresponding mock control.
Reproducibly differentially expressed probe sets were selected from the total normalized data (Table S1) method selected, and TreeView version 1.60 (Eisen et al., 1998; http://rana.lbl.gov/EisenSoftware.htm). MapMan version 1.6.0 (Thimm et al., 2004) was used for analysis of the functional classes and metabolic pathways following pathogen inoculation. The complete set of microarray data has been deposited to the National Center for Biotechnology Information GEO database (http://www.ncbi.nlm.nih.gov/geo/) in a MIAME-compliant format (GSE6556 and GSM151694 to GSM151705 ).

Supplemental Data
The following materials are available in the online version of this article.         Table S4).  A, In the compatible interaction, GH3.5 is activated to modulate the auxin pathway resulting in enhanced disease susceptibility, through activating IAA biosynthesis and derepressing auxin signaling. IAA also induces GH3.5 to enlarge those processes. In addition, GH3.5 might also function as an IAA-amido synthetase to regulate IAA homeostasis.

Supplemental
B, In the compatible/ incompatible interactions, GH3.5 positively modulates the SA pathway to enhance plant defense response, through elevating SA biosynthesis, activating SA-induced genes, WRKYs and basal defenses-related genes. The feedback regulation of GH3.5 by SA amplifies those effects. In this case, GH3.5 might also synthesize SA-Asp with unknown function during the interactions.