An ent-kaurene-derived diterpenoid virulence factor from Xanthomonas oryzae pv. oryzicola
Summary
- Both plants and fungi produce ent-kaurene as a precursor to the gibberellin plant hormones. A number of rhizobia contain functionally conserved, sequentially acting ent-copalyl diphosphate and ent-kaurene synthases (CPS and KS, respectively), which are found within a well-conserved operon that may lead to the production of gibberellins.
- Intriguingly, the rice bacterial leaf streak pathogen Xanthomonas oryzae pv. oryzicola (Xoc) contains a homologous operon. Here, we report biochemical characterization of the encoded CPS and KS, and the impact of insertional mutagenesis on virulence and the plant defense response for these genes, as well as that for one of the cytochromes P450 (CYP112) found in the operon.
- Activity of the CPS and KS found in this phytopathogen was verified – that is, Xoc is capable of producing ent-kaurene. Moreover, knocking out CPS, KS or CYP112 led to mutant Xoc that exhibited reduced virulence. Investigation of the effect on marker gene transcript levels suggests that the Xoc diterpenoid affects the plant defense response, most directly that mediated by jasmonic acid (JA).
- Xoc produces an ent-kaurene-derived diterpenoid as a virulence factor, potentially a gibberellin phytohormone, which is antagonistic to JA, consistent with the recent recognition of opposing effects for these phytohormones on the microbial defense response.
Introduction
It has long been evident that microbes produce phytohormones to manipulate their interaction with plant hosts (Bari & Jones, 2009). Indeed, the gibberellins (GAs) were first isolated from the fungal pathogen Gibberella fujikuroi (Fusarium moniliforme), which causes the bakanae or foolish seedling disease in rice (Oryza sativa), with its characteristic excessive growth phenotype (Yabuta & Sumiki, 1938). Although best known for their role in promoting plant growth, GAs have more recently been recognized as antagonists of the plant defense response mediated by jasmonic acid (JA) and ethylene (Bari & Jones, 2009). For example, in rice, the GAs have been shown to negatively regulate resistance to both the bacterial blight pathogen Xanthomonas oryzae pv. oryzae (Xoo) and fungal blast pathogen Magnaporthe oryzae (Yang et al., 2008; Qin et al., 2013). Moreover, recently, the production of GA by G. fujikuroi has also been reported to play a role in the virulence of this rice plant pathogen (Wiemann et al., 2013).
In addition to G. fujikuroi, it has been found that certain (rhizo)bacteria also produce GAs (Bottini et al., 2004). This includes Bradyrhizobium japonicum USDA 110 (Mendez et al., 2014), whose genome contains an operon consisting of three cytochromes P450 (CYPs), a ferredoxin (Fd), a short-chain alcohol dehydrogenase/reductase (SDR), a putative prenyl transferase and two putative prenyl cyclases, which has been hypothesized to be involved in the production of GA (Tully et al., 1998). Both plants and fungi produce ent-kaurene as a key precursor in GA biosynthesis (Hedden et al., 2001), and it has been shown that the putative prenyl cyclases from B. japonicum act sequentially to produce ent-kaurene via ent-copalyl diphosphate (CPP) from (E,E,E)-geranylgeranyl diphosphate (GGPP), and these are therefore an ent-copalyl diphosphate synthase (CPS) and ent-kaurene synthase (KS), respectively (Morrone et al., 2009). Moreover, homologs of this operon are found in a number of other rhizobia, several of which are known to produce GAs, and representative examples from the three other major genera of rhizobia (Mesorhizobium loti, Sinorhizobium fredii and Rhizobium etli) have been shown to at least retain the capacity for the production of ent-kaurene (Hershey et al., 2014). In particular, the CPS and KS homologs also produce ent-kaurene from GGPP which, furthermore, has been shown to be the product of the associated prenyl transferase, which is then a GGPP synthase (GGPS). These results strengthen the association of this operon with GA biosynthesis (Fig. 1).
Intriguingly, there is a homolog of this putative GA biosynthesis operon in the genome sequence of the rice bacterial leaf streak pathogen Xanthomonas oryzae pv. oryzicola (Xoc) strain BLS256 (Bogdanove et al., 2011). Here, it is demonstrated that the putative CPS and KS from Xoc similarly produce ent-kaurene from GGPP. In addition, insertional disruption of the genes for not only the diterpene synthases, but also one of the CYPs, leads to Xoc strains with impaired pathogenicity, indicating that an ent-kaurene-derived diterpenoid, potentially a GA, acts as a virulence factor in the Xoc–rice interaction. Expression profiling of phytohormone marker genes indicates that the ent-kaurene-derived diterpenoid produced from Xoc downregulates the rice plant defense response, most directly that mediated by JA.
Materials and Methods
Unless otherwise noted, all molecular biology reagents were purchased from Invitrogen (Carlsbad, CA, USA) and all other chemicals from Fischer Scientific (Waltham, MA, USA). Escherichia coli was grown at 37 or 16°C using either NZY (for cloning) or TB medium (for expression). Xoc was cultured on GYE (2% glucose, 1% yeast extract) medium at 28°C. When necessary, 1.8% (v/v) agar was added to the relevant medium to pour plates. Where applicable, antibiotics were used at the following concentrations: chloramphenicol, 30 μg ml−1; carbenicillin, 50 μg ml−1; spectinomycin, 50 μg ml−1; and/or kanamycin, 50 μg ml−1. Liquid cultures were grown with vigorous shaking (200 rpm), generally in 250-ml Erlenmeyer flasks containing 50 ml of medium.
The presence of the diterpenoid biosynthetic operon in Xoc was first detected by BLAST searches of the nonredundant nucleotide database at GenBank in 2010 using the distinctive rhizobial KS gene as the query sequence (specifically that from B. japonicum). Identification of the KS homolog was followed by analysis of the surrounding sequence, which proved to have the entire core operon defined from our studies in rhizobia (Hershey et al., 2014). The presence of homologous sequence in the various pathovars/strains of X. translucens was only more recently discovered by similar analysis carried out in 2013. Sequence analysis/alignments were carried out using the CLC Main Workbench program version 6.8.4 (CLC bio, Aarhus, Denmark).
Biochemical characterization of the CPS and KS from the Xoc operon was carried out as described previously (Morrone et al., 2009). Briefly, the XocCPS and XocKS genes were amplified via PCR from Xoc BLS256 genomic DNA, prepared as described previously (Bogdanove et al., 2011), using gene-specific primers (Supporting Information Table S1), cloned by topoisomerization into pENTR/SD/D-TOPO (Invitrogen) and verified by complete gene sequencing. For co-expression, XocCPS and XocKS were subcloned by directional recombination into pDest14 (Invitrogen) and a previously described GGPS containing pGG-DEST vector (Cyr et al., 2007), respectively. The resulting pDest14::XocCPS and pGG-DEST::XocKS constructs were co-transformed into E. coli strain C41 (Lucigen, Middleton, WI, USA), together with pIRS, which overexpresses genes from the endogenous methylerythritol phosphate-dependent isoprenoid precursor supply pathway and has been shown to increase flux towards terpenoid metabolism (Morrone et al., 2010). The resulting recombinant strains were cultured in 1 l of liquid medium at 37°C until they reached an optical density at 600 nm (OD600 nm) of c. 0.6, the pH was adjusted to 7.1, the cultures were shifted to 16°C for 1 h before induction with IPTG (isopropyl-β-d-thiogalactopyranoside) added to a final concentration of 0.5 mM, followed by further fermentation at 16°C for c. another 72 h. The cultures were then extracted with an equal volume of hexane, which was separated out, dried in a rotary evaporator, and the residue was resuspended in 100 μl of hexane. This concentrated extract was analyzed by gas chromatography (GC) carried out on a Varian (Palo Alto, CA, USA) 3900 GC with Saturn 2100 ion trap mass spectrometer (MS) in electron ionization (70 eV) mode. Samples (1 μl) were injected in splitless mode at 50°C and, after holding for 3 min at 50°C, the oven temperature was raised at a rate of 14°C min−1 to 300°C, where it was held for an additional 3 min. MS data from 90 to 600 m/z were collected starting 12 min after injection until the end of the run. The production of ent-kaurene was verified by comparison of mass spectra and retention time with an authentic standard (enzymatically produced by the characterized CPS and KS from Arabidopsis thaliana).
For the analysis of GA formation, Xoc was cultured in 500-ml flasks containing 300 ml of liquid GYE medium. The cultures were incubated for 3, 6, 9 and 12 d, respectively, on a rotary shaker (200 rpm) at 28°C. Cells were separated from the spent medium by centrifugation (15 min at 4000 g). For the analysis of GA content, the supernatant was acidified to pH 2.5 with acetic acid, and then extracted with an equal volume of ethyl acetate saturated with acetic acid (1% v/v). This organic extract was separated and passed over a 1-ml HP-20 resin column, which was then eluted with 3 ml each of 1% acetic acid in distilled H2O, and then 40% and 80% methanol (v/v in distilled H2O with 1% acetic acid). Each of these fractions was dried in a rotary evaporator and the residue was resuspended in 100 μl of acetic acid-saturated ethyl acetate for analysis by GC-MS as described above. For the analysis of the ent-kaurene intermediate, the total culture was extracted directly with an equal volume of hexane, which was separated out and passed over a 1-ml silica gel column to remove contaminating polar compounds, with the resulting organic extract dried under a gentle stream of N2, and the residue resuspended in 100 μl of hexane for analysis by GC-MS, again as described above.
To generate insertional mutant strains of Xoc, internal fragments from the CPS, KS and CYP112 genes (700, 400 and 500 bp in length, respectively) were amplified from genomic DNA with gene-specific primers (Table S1). These PCR fragments were cloned into the pZeroblunt plasmid (carries kanamycin resistance) to create suicide vectors for insertional inactivation of CPS, KS and CYP112. The resulting plasmids were electroporated into Xoc for intragenic integration. Kanamycin-resistant colonies appearing at 28°C were taken to be integrating mutants, in which a single crossover homologous recombination event had taken place. This was confirmed by PCR from genomic DNA (prepared using the Wizard Genomic DNA Purification Kit; Promega, Madison, WI, USA) from each mutant, using the primer pairs (Table S1) CPS200-F and M13-R for cps, KS50-F and M13-R for ks and CYP112.50-F and M13-R for cyp112, generating PCR fragments of 1000, 700 and 900 bp in length, respectively, if the relevant suicide vector had been inserted. In each case, the lack of such a PCR product with wild-type Xoc was also confirmed.
To investigate the polarity of the various mutant strains, reverse transcription-polymerase chain reaction (RT-PCR) analysis was carried out on these strains and compared with wild-type Xoc. Single colonies of Xoc strains were inoculated into 5 ml GYE and grown at 28°C to OD600 nm of c. 0.5. Bacterial cells were collected by centrifugation (15 min at 4000 g), and total RNA was extracted using the RNeasy Kit (Qiagen, Venlo, the Netherlands). The extracted RNA was treated with DNase I (Invitrogen) and purified. cDNAs were obtained from RT reactions containing 500 ng total RNA using the iScript Reverse Transcription kit (BioRad, Hercules, CA, USA), following the manufacturer's instructions. This was used as the template for PCR with gene-specific primers (CPS430-R and CPS1103-R for CPS; KS249-F and KS649-R for KS; CYP112.351-F and CYP112.850-R for CYP112; CYP114.106-F and CYP114.459-R for CYP114; IDI1-F and IDI301-R for IDI), and 16S rRNA primers as a positive internal control (Table S1).
To complement the cps and ks mutant strains, XocCPS and XocKS were transferred by directional recombination from the pENTR/SD/D-TOPO-derived constructs described above to the Xanthomonas expression vector pKEB31 which contains the necessary DEST cassette (Cermak et al., 2011). The resulting pKEB31/DEST::XocCPS and pKEB31/DEST::XocKS vectors were transformed by electroporation into the Xoc cps and ks mutant strains, respectively. These two complementation strains (cps + CPS and ks + KS, respectively) were confirmed by sequencing PCR products of the introduced CPS and KS, respectively.
Quantitative real-time PCR (qPCR) was carried out with total RNA, extracted using the RNeasy Plant Mini Kit, from infected wild-type Nipponbare rice leaves collected either quickly (c. 30 min) or 1, 2, 3, 4 or 5 d post-inoculation. Template cDNA was generated from 500 ng of DNAse-treated RNA as described above. This was diluted eight-fold for use in ABsolute qPCR SYBR Green Mix (Thermo, Waltham, MA, USA) reactions (25 μl), with the relevant gene-specific pairs of primers (Table S1), carried out using a Stratagene Mx4000 instrument located in the Iowa State University Genomics Technologies Facility. Transcription FactorIIAγ5 (LOC_Os05g01710) was used as internal control for rice genes, and the Xoc reference gene was the ribosomal 16S subunit. Each data point represents the average of two independent biological replicates, each measured with two technical replicates. For Xoc gene analysis, transcript levels in bacterial culture were also measured using DNAse-treated total RNA isolated from wild-type Xoc as described above.
Relative virulence was assessed using paired-infection assays carried out with Xoc cells suspended in 10 mM magnesium chloride to an OD600 nm of 0.5, which were used to spot infiltrate leaves of 8-wk-old rice plants with a needleless syringe. Leaves of rice plants were inoculated with one strain on each side of the midrib (e.g. mutant on one side and the wild-type on the other). After 10 d, lesion lengths were measured for each paired inoculation, and a paired, two-tailed Student's t-test was performed across all three replicates. Each experiment was carried out three separate times, each time with 10 leaves.
For bacterial colony counting, three leaves, each infected with only one strain of Xoc, were cut and ground in a sterile mortar and pestle with 2 ml of sterile water. This suspension was then serially diluted (10-fold) and 5-μl drops from the 10−1–10−6 dilutions were placed on peptone sucrose plates with cephalexin antibiotics, with each measurement (i.e. plate) carried out in triplicate. These plates were incubated at 28°C until individual colonies were apparent for at least some of the spots, which were then counted, allowing the calculation of the original titer of viable bacteria from each leaf.
Results
It has been noted previously that a KS from B. japonicum exhibits a distinct sequence relative to other characterized bacterial diterpene synthases (Morrone et al., 2009). This distinctive sequence has been used to identify other rhizobia that contain homologs to KS, and an associated diterpenoid biosynthetic operon, with functional conservation of at least the capacity for the production of ent-kaurene (Hershey et al., 2014). With the recent discovery of a KS from Streptomyces platensis, albeit for the production of the antibiotic platensimycin rather than GA phytohormones (Smanski et al., 2011), we became interested in the wider distribution of such biosynthesis. Although the KS from S. platensis is not closely related to the KSs found in rhizobia, sharing < 26% amino acid (aa) sequence identity with these (which share > 67% aa sequence identity with each other), broader BLAST searches (i.e. outside of the previously investigated Rhizobiales order) revealed the presence of homologs in the Xanthomonas genus. The initial example found was from the genome sequence of Xoc strain BLS256 (Bogdanove et al., 2011). Moreover, examination of the genomic context of this XocKS (Xoc_0077) revealed that it was situated in an operon exhibiting strong similarity to the putative GA biosynthetic gene cluster found in the rhizobia. Specifically, it contains all eight genes found in B. japonicum which have been defined as the core operon in the rhizobia more generally (i.e. CYP112, CYP114, Fd, SDR, CYP117, GGPS, CPS and KS), with those from Xoc sharing 70–78% nucleotide (nt) sequence identity with their reported rhizobial homologs, close to the 80–92% nt sequence identity which they share with each other. In addition, there are two other genes that appear to be part of the Xoc operon (Fig. 1c). As also found in R. etli and M. loti, on the 3′ end of the Xoc operon is an isopentenyl diphosphate isomerase (IDI), which shares c. 77% nt sequence identity with the IDIs from these two rhizobia (again close to their 83% nt sequence identity to each other). There is also an additional CYP, CYP115, on the 5′ end of the operon in Xoc. This shares some similarity to a partial CYP pseudogene located in a similar 5′ position in the B. japonicum operon (Tully et al., 1998), but appears to be full length.
Despite the strong similarity, the greater divergence of the Xoc operon from the previously characterized rhizobial homologs prompted us to investigate whether the encoded enzymes were functionally conserved. Accordingly, we biochemically characterized the CPS and KS from Xoc. These were cloned from Xoc genomic DNA and investigated using a previously described modular metabolic engineering system (Cyr et al., 2007), much as described for the rhizobial enzymes. Briefly, co-expression of the CPS and KS from Xoc, in recombinant E. coli also expressing a GGPS, led to the production of kaurene (Fig. 2). Production of the ent-kaurene stereoisomer, consistent with the production of the GA phytohormones, was demonstrated by co-expressing the Xoc diterpene synthases with plant counterparts whose stereospecificity had already been established (again, together with GGPS co-expression in each case). In particular, the KS from Xoc was co-expressed with the ent-CPP-producing CPS from A. thaliana (Sun & Kamiya, 1994), whereas the CPS from Xoc was co-expressed with the ent-CPP-specific KS from Cucurbita maxima (Yamaguchi et al., 1996), leading to the production of ent-kaurene in both cases. Moreover, analogous co-expression studies with plant diterpene synthases exhibiting alternative stereochemical specificity further demonstrated that XocCPS and XocKS were specific for the enantiomeric (9R,10R) isomer of CPP (data not shown). Accordingly, the CPS and KS from Xoc cooperatively produce ent-kaurene, exhibiting functional conservation with their homologs from rhizobia, which is consistent with broader functional conservation of the entire operon.
Given that ent-kaurene is the olefin intermediate in GA biosynthesis in both plants and fungi (Hedden et al., 2001), we hypothesized that Xoc might produce GAs. Unfortunately, we could not detect any of the common bioactive GAs (e.g. GA1, GA3, GA4 or GA7), or even ent-kaurene, from Xoc cultures. It is similarly difficult to detect GAs (and also ent-kaurene) from rhizobial cultures (Hershey et al., 2014), although GA metabolism can be observed with bacteroids isolated from root nodules (Mendez et al., 2014). Indeed, in the rhizobia, it is clear that the operon is strongly induced on interaction with the host plant during nodulation (Pessi et al., 2007). Accordingly, we hypothesized that the Xoc operon is similarly induced on interaction with its host plant during infection. This was examined by comparing the expression level of several genes from the Xoc operon in infected leaves relative to that observed in bacterial culture over a 5-d time course. Increases of up to four-fold in transcript levels were observed, with the highest levels consistently observed after the first day (Fig. 3). Thus, similar to the rhizobial expression pattern, the Xoc operon is induced during plant–microbe interactions.
To investigate the role of ent-kaurene production in Xoc, we generated knockout mutants of its CPS and KS. In particular, these were disrupted via single-crossover homologous recombination using pZeroBlunt-based constructs (Supporting Information Figs S1, S2). Somewhat surprisingly, these were nonpolar insertions, as expression of the downstream KS could be detected in the resulting cps (CPS::pZeroBlunt) strain, and expression of the downstream IDI could be detected in the resulting ks (KS::pZeroBlunt) strain, although the targeted gene was clearly no longer expressed in both strains (Fig. S2). The effect of these mutations was investigated using a previously described paired infection assay that enables comparison of relative virulence (Wang et al., 2007). Briefly, the mutant and parental (wild-type) strains were inoculated, via infiltration with a needleless syringe, in a pairwise fashion on opposite sides of the midvein of the same leaf. Lesion length was then measured 10 d after inoculation. With both cps and ks, the resulting lesions were significantly shorter than those produced by the paired wild-type Xoc (Fig. 4a). To verify that the observed reduction in lesion length reflects decreased bacterial growth (i.e. virulence), the numbers of viable bacteria following inoculation with equivalent titers of each of the Xoc strains (i.e. wild-type, cps and ks) were compared. With this assay, there is a clear decrease in virulence, with significantly reduced numbers of viable Xoc from inoculation with both cps and ks relative to the wild-type, observed within 1 d (Fig. 4b).
To confirm that this reduction in virulence was caused by the loss of CPS and KS, we generated complementation strains. These strains (cps + CPS and ks + KS) were then tested in paired infection assays against either the parental mutant strains or the original wild-type strain. The results demonstrated that complementation effectively restored the virulence of both the cps and ks strains to levels equivalent to that of the wild-type (Fig. 5), confirming that these diterpene synthases contribute to the virulence of Xoc in rice, presumably via the production of ent-kaurene as an intermediate in the biosynthesis of a more elaborate diterpenoid virulence factor.
To investigate whether it is the ent-kaurene product of CPS and KS or a further oxidized diterpenoid that acts as the virulence factor, the cytochrome P450 CYP112, presumably involved in such further elaboration, was targeted by insertional mutagenesis, creating the knock-out strain cyp112 (CYP112::pZeroBlunt), which was also found to be nonpolar, with the continued expression of CPS and KS suggesting continued production of ent-kaurene (Fig. S2). In paired infection tests, cyp112 was also found to exhibit significantly reduced lesion length relative to wild-type Xoc (Fig. 6). Thus, it appears to be an ent-kaurene-derived diterpenoid, potentially a GA, rather than ent-kaurene itself, which acts as a virulence factor for Xoc.
Particularly given the reported antagonism between GAs and the JA-mediated plant defense response (Bari & Jones, 2009), we hypothesized that the Xoc diterpenoid virulence factor might act through effects on phytohormone signaling. This was investigated by analysis of the expression profile of plant defense-related marker genes in rice following infection with either cps or wild-type strains of Xoc. Transcript levels of lipoxygenase (OsLOX) and allene oxide synthase (OsAOS2) are increased by JA-mediated signaling, whereas pathogenesis-related genes, such as OsPR1a and OsPR1b, are increased by salicylic acid (SA)-mediated signaling (Navarro et al., 2008; Yoshii et al., 2010; Qin et al., 2013). Hence, normalized transcript levels for these four rice genes were measured by qPCR over a 5-d time course following infection with either wild-type or cps Xoc, with significant differences evident for each. The most immediate effect was a greater increase in transcript levels of the JA signaling marker genes OsLOX and OsAOS2, which was observed on the first day post-inoculation for leaves infected with cps relative to those infected with wild-type Xoc (Fig. 7a,b). By contrast, no significant difference was observed over this time period between leaves infected with either wild-type or cps Xoc for the SA signaling marker genes OsPR1a and OsPR1b (Fig. 7c,d). Thus, the Xoc ent-kaurene-derived diterpenoid virulence factor appears to act, at least in part, through the suppression of the JA-mediated plant defense response.
Notably, this diterpenoid biosynthetic operon does not appear to be present in any of the sequenced strains of the other pathovar Xoo, or other isolates that group separately from Xoc and Xoo (Lee et al., 2005; Ochiai et al., 2005; Salzberg et al., 2008; Triplett et al., 2011). However, BLAST searches revealed the presence of a homologous operon in the recently reported genome sequence of X. translucens pv. graminis ART-Xtg29 (Wichmann et al., 2013). Moreover, homologous sequence can also be found in the genome sequences for the two strains of X. translucens pv. translucens, DSM 18974 and DAR61454, that are currently available at the National Center for Biotechnology Information (NCBI) (Genome accession 14 066). The operons from these two strains are essentially identical to each other, sharing c. 99% nt sequence identity. In addition to the eight core operon genes found in the rhizobia (CYP112, CYP114, Fd, SDR, CYP117, GGPS, CPS and KS), these examples from X. translucens also include the accessory IDI on the 3′ end and additional full-length CYP115 on the 5′ end. The shared presence of the unique full-length CYP115 suggests a homologous origin for the operons found in the X. translucens genomes and that from Xoc. However, between X. translucens and Xoc, these operons exhibit quite distinct sequences, sharing only c. 70% identity at the nt level, indicating that any such common origin is quite distant and these may represent separate horizontal gene transfer events. By contrast, the operons from the two distinct pathovars of X. translucens are more closely related to each other (c. 89% nt sequence identity), and presumably share a much more recent common origin, presumably via vertical descent rather than horizontal gene transfer (Fig. S3).
Discussion
Horizontal gene transfer of this diterpenoid biosynthetic operon among the rhizobia has been demonstrated previously (Hershey et al., 2014). However, its presence in the Xanthomonas genus, which falls into the separate Gammaproteobacteria class, represents a surprisingly wide distribution. Nevertheless, these bacteria share an obvious common association with plants, albeit in strikingly different contexts, as symbionts and pathogens, respectively. Intriguingly, it appears that the operon may have been separately acquired by Xoc and X. translucens via horizontal gene transfer, presumably from the rhizobia.
The operon has been associated with GA production in the rhizobia, and such production in the Xanthomonas phytopathogens would be consistent with the recent discovery that GAs are antagonistic to JA-mediated plant defense (Robert-Seilaniantz et al., 2007). However, it should be noted that the presence of the additional, unique cytochrome P450 CYP115 in the Xanthomonas-associated operons indicates that these may produce a further elaborated diterpenoid relative to the rhizobia. For example, by analogy with the production of the JA-isoleucine (JA-Ile) mimic coronatine by the phytopathogen Pseudomonas syringae (Fonseca et al., 2009), the Xanthomonas operon might lead to the production of a GA mimic resistant to degradation by plant catabolism. Alternatively, the resulting diterpenoid might inhibit the catabolism of endogenously produced GAs.
The observed rice host marker gene expression analysis is consistent with the potential production of GA. Specifically, the lack of such production by cps Xoc most immediately leads to a more robust JA-mediated defense response, consistent with antagonism between the ent-kaurene-derived diterpenoid and JA signaling, as has been reported for GA as a result of interactions between the DELLA and JAZ proteins involved in the GA and JA response mechanisms, respectively (Hou et al., 2010; Wild et al., 2012; Yang et al., 2012). Regardless of the exact identity of the ent-kaurene-derived diterpenoid product of this operon, the results presented here demonstrate that this compound acts as a virulence factor promoting Xoc phytopathogen growth, ostensibly by suppressing the defense response of its host plant rice via effects on JA signaling. This mechanism may also underlie the use of GAs as virulence factors by the rice fungal pathogen G. fujikuroi (Wiemann et al., 2013), and is consistent with the emerging paradigm that a key pathogen virulence strategy is such modulation of plant hormone signaling (Robert-Seilaniantz et al., 2011).
Much as found in the rhizobia, this diterpenoid biosynthesis operon exhibits a scattered distribution within the Xanthomonas genus. The operon characterized here from the X. oryzae pathovar Xoc is not found in the Xoo pathovar. As Xoo and Xoc are otherwise very closely related (Nino-Liu et al., 2006; Lu et al., 2008; Bogdanove et al., 2011), this operon may contribute to the differentiation between the rice leaf streak and blight diseases caused by Xoc and Xoo, respectively. Alternatively, production of the resulting ent-kaurene-derived diterpenoid virulence factor may only provide a selective advantage under the conditions encountered by X. oryzae in the context of rice leaf streak rather than blight (e.g. infection of the mesophyll vs the xylem, respectively). However, the presence of homologous sequence in various isolates of X. translucens suggests a broader role for the resulting ent-kaurene-derived diterpenoid virulence factor. In particular, together the sequenced isolates of X. translucens collectively cause leaf diseases on other small-grain cereal crops, such as wheat and barley, as well as forage grasses, such as ryegrass, with some infecting the vasculature and others the mesophyll. Thus, this ent-kaurene-derived diterpenoid may serve as a virulence factor for Xanthomonas phytopathogens of the Poaceae/grass plant family more generally. Building on the role in pathogenicity shown here for the Xoc–rice interaction, future investigation of the exact role of this ent-kaurene-derived diterpenoid in the other Xanthomonas species in which it is present will help to elucidate its role in tissue specificity vs Xanthomonas–Poaceae interactions more broadly.
Acknowledgements
The authors thank Shan Qi (Cornell University) for her help with the inoculation assays and helpful discussions. We also thank Aaron Hummel for productive discussions. We are thankful to Brice Floyd for technical assistance with the quantitative RT-PCR analyses. This work was supported by grants from the National Science Foundation (MCB0919735) and US Department of Agriculture (USDA-AFRI-NIFA 2014-67013-21720) to R.J.P.
