Expression of the Gene for Resistance to Phaseolotoxin (argK) Depends on the Activity of Genes phtABC in Pseudomonas syringae pv. phaseolicola

The bacterium Pseudomonas syringae pv. phaseolicola produces phaseolotoxin in a temperature dependent manner, being optimally produced between 18°C and 20°C, while no detectable amounts are present above 28°C. Phaseolotoxin is an effective inhibitor of ornithine carbamoyltransferase (OCTase) activity from plant, mammalian and bacterial sources and causes a phenotypic requirement for arginine. To protect the cell from its own toxin, P. syringae pv. phaseolicola synthesizes a phaseolotoxin-resistant OCTase (ROCT). The ROCT is the product of the argK gene and is synthesized only under conditions leading to phaseolotoxin synthesis. The argK gene is included in a chromosomal fragment named Pht cluster, which contains genes involved in the synthesis of phaseolotoxin. The aim of the present work was to investigate the possible involvement of other genes included in the Pht cluster in the regulation of gene argK. We conducted transcriptional analyses of argK in several mutants unable to produce phaseolotoxin, transcriptional fusions and electrophoretic mobility shift assays, which allowed us to determine that genes phtABC, located within the Pht cluster, participate in the transcriptional repression of gene argK at temperatures not permissive for phaseolotoxin biosynthesis. This repression is mediated by a protein present in both toxigenic and nontoxigenic strains of P. syringae and in E. coli, and requires the coordinated participation of phtA, phtB and phtC products in order to carry out an efficient argK repression.


Introduction
Production of phaseolotoxin, a non host specific toxin, has been described in Pseudomonas syringae pv. phaseolicola, which infects bean (Phaseolus vulgaris L), P. syringae pv. actinidiae, which infects kiwi (Actinidia chinensis), and in strain CFBP3388 of P. syringae pv. syringae, isolated from vetch (Vicia sativa) [1,2,3]. There are natural strains of these pathovars that do not produce the toxin and that do not contain the DNA responsible for its synthesis [3,4,5,6] indicating that the ability to produce phaseolotoxin has been acquired after pathovar delineation as a recent event [7]. Analysis of the conservation of the phaseolotoxin biosynthesis genes in a broad collection of P. syringae pv. phaseolicola and pv. actinidiae strains, as well as in pv. syringae strain CFBP3388, suggests that genes for the biosynthesis of phaseolotoxin have a complex evolutionary history and have been acquired by pathovars of P. syringae at least twice during evolution [6].
The production of phaseolotoxin is temperature dependent, being optimally produced between 18uC and 20uC, while no detectable amounts are present above 28uC [8,9,10]. Phaseolotoxin is composed of two moieties: the inorganic moiety, N d N9-sulfodiaminophosphinyl, and the L-ornithyl-alanyl-homoarginine tripeptide [1,11]. Targets of this toxin are the enzymes ornithine carbamoyltransferase (OCTase; EC 2.1.3.3) [12], which catalyzes the formation of citrulline from ornithine and carbamoylphosphate in the arginine biosynthetic pathway, and ornithine decarboxylase, which participates in the biosynthesis of polyamines [13]. Phaseolotoxin is an effective inhibitor of OCTase activity from plant, mammalian and bacterial sources and causes a phenotypic requirement for arginine. This property led to the development of a rapid bioassay that evaluates growth inhibition of a bacterial culture exposed to this toxin [14].
To protect itself from its own toxin, P. syringae pv. phaseolicola synthesizes a phaseolotoxin-resistant OCTase (ROCT) [12,15,16,17,18]. The ROCT, which is the product of the argK gene and is expressed under conditions leading to phaseolotoxin synthesis, is a polypeptide composed of 327 amino acid residues with a molecular mass of 36.52 kDa, [18,19,20]. ROCT is necessary for P. syringae pv. phaseolicola under conditions of phaseolotoxin synthesis, because it ensures an optimal supply of the arginine required for growth and synthesis of phaseolotoxin [21].
Several efforts have been made to determine how argK is regulated. It has been postulated that argK might be regulated at 28uC under negative control by a repressor protein [20]. This repressor protein could bind to specific DNA motifs found in the argK promoter that have been postulated to be involved in thermoregulation of phaseolotoxin synthesis [22,23]. Additionally, carbamoylphosphate is able to induce argK expression, bypassing the temperature control [21]. Since carbamoylphosphate is a compound resembling the inorganic moiety of phaseolotoxin, N d N9-sulfodiaminophosphinyl, these results suggested that argK is being directly regulated by a molecule that may be a precursor of phaseolotoxin and only indirectly by temperature [21]. On the other hand, the OCTase present under conditions nonpermissive for phaseolotoxin synthesis in P. syringae pv. phaseolicola, encoded by gene argF, is negatively regulated by ArgR [24]. However, production of ROCT and synthesis of phaseolotoxin occur independently of ArgR. Therefore there is not any apparent metabolic link between the genes for phaseolotoxin synthesis/ resistance and the genes involved in the primary metabolism [24].
Genes required for the biosynthesis of and resistance to phaseolotoxin are grouped in the so-called Pht cluster and are included into a genomic region that has the characteristics of a pathogenicity island [15,25,26,27,28]. The Pht cluster is composed of 23 genes and it is flanked by insertion sequences and transposases. The 23 genes are organized in five transcriptional units ( Figure 1A), two monocistronic and three polycistronic, with one of them overlapping a larger operon. Mutagenesis of 14 genes within the Pht cluster resulted in three mutants showing low levels of toxin production, while a Tox-phenotype was shown for the eleven remaining mutants [25].
The two-component system and global regulators GacS/GacA also participate in the regulation of phaseolotoxin biosynthesis [29], and in a gacA 2 background it was evident the downregulation of genes within operons phtA, phtD, phtL and phtM, whose expression was negligible at 18uC. Interestingly, gene argK losses its temperature-dependent expression in a gacA 2 background and, in contrast with the rest of the genes within the Pht cluster, it becomes constitutive at both 18uC and 28uC [29].
We investigated the possible involvement of other genes included in the Pht cluster in the regulation of the argK gene. To this end, we analyzed the transcription of argK in several mutants unable to produce phaseolotoxin, and we also conducted electrophoretic mobility shift assays, which allowed us to determine that genes phtABC, included in the Pht cluster, are required to control argK transcription in response to temperature in P. syringae pv. phaseolicola NPS3121. We also report that to carry out an efficient argK repression, it is necessary the coordinated participation of the products of phtA, phtB and phtC.

argK transcriptional analysis in a group of mutants
We analyzed the effect on argK expression of mutations on different genes of the Pht cluster, including polar mutants YNorf1P, SAorf5P, SAorf10P and AT3, altered in genes phtA, phtE, phtL and amtA, respectively (Table 1) [25,30]. To assess the expression pattern of gene argK at 18uC and 28uC in these mutants with respect to the wild type strain NPS3121, we conducted Reverse Transcription-PCR analysis (RT-PCR) aimed to amplify specific fragments derived from cDNA.
In mutant YNorf1P, the argK gene showed an increased expression at 28uC, unlike what happens in strain NPS3121 at the same temperature ( Figure 1B), indicating that a mutation on the phtA operon resulted in alleviation of the repression of argK at a nonpermissive temperature for phaseolotoxin synthesis. These results are compatible with previous reports postulating that in P. syringae pv. phaseolicola, the argK gene could be regulated under negative control by a repressor protein at 28uC [20]. Conversely, Figure 1. Participation of genes from the phaseolotoxin biosynthesis cluster in the expression of gene argK from P. syringae pv. phaseolicola NPS3121. A. Graphic representation of the phaseolotoxin biosynthesis cluster (Pht cluster) of P. syringae pv. phaseolicola NPS3121. The Pht cluster contains five transcriptional units, including two monocistronic (argK and phtL) and three polycistronic operons (phtA, phtD and phtM) [25]. There is a secondary promoter (P D ) capable of driving the expression of genes from phtD to phtK. Genes are represented by arrows, with the direction of the arrow indicating the direction of transcription. B. Analysis of the argK transcriptional pattern in P. syringae pv. phaseolicola NPS3121 and polar mutants by reverse transcription-PCR. RT-PCR amplicons were separated by electrophoresis, and reversed images of the gels are shown; the strains analyzed and the corresponding mutated genes are described on top of their corresponding lanes. The small numbers under the lanes represent the temperature at which expression was assayed: 1 indicates 18uC and 2 indicates 28uC. doi:10.1371/journal.pone.0046815.g001 argK Regulation in P. syringae pv. phaseolicola PLOS ONE | www.plosone.org transcription of gene argK in mutants SAorf5P, SAorf10P and AT3, affected in genes phtE, phtL and amtA, respectively, was similar to that in wild type strain NPS3121 at both temperatures. Since a mutation into gene phtE, which belongs to the phtA operon, did not modify argK expression at 28uC, it is likely that only genes located upstream to phtE could be participating in argK regulation ( Figures 1A;1B).

Mutant YNorf1P showed an increase in OCTase activity at 28uC
Based in the previous RT-PCR analyses showing an increase in argK expression at 28uC in mutant YNorf1P ( Figure 1B) and considering that the argK gene codes for the phaseolotoxin-resistant OCTase, we analyzed the OCTase specific activity at 28uC in YNorf1P in comparison to that of the wild type strain NPS3121. In order to discard any activity corresponding to the phaseolo-toxin-sensitive OCTase, we preincubated the reaction mixture with a phaseolotoxin containing supernatant; later, OCTase activity was determined as previously reported [24]. The results obtained are shown in Figure 2. In agreement with the results obtained by RT-PCR, we observed a significant increase in the OCTase activity in YNorf1P compared with NPS3121, indicating that the argK expression level observed at 28uC was directly related to an increase in the OCTase activity.
Effect of phtABC genes on argK expression pattern in P. syringae pv. phaseolicola strain CYL233 Since argK transcription was unambiguously de-repressed at 28uC in mutant YNorf1P, we decided to determine which genes from the phtA operon could participate in argK regulation. The phtA operon contains 11 genes, from phtA to phtK, which are transcribed divergently to the argK gene, although it also possesses an internal promoter immediately downstream of phtC, capable of driving expression of phtD to phtK ( Figure 1A). Additionally, RT-PCR analyses indicated that a polar mutation in gene phtE did not modify argK expression at 18uC and 28uC ( Figure 1B). Therefore, we focused our analysis in the participation of genes phtA, phtB and phtC in the argK regulation considering that phtE gene belongs to the phtD operon. To carry out our experiments we decided to use the P. syringae pv. phaseolicola wild type strain CYL233 with the aim to discard the participation of other genes from the Pht cluster. It was reported that strain CYL233 is naturally unable to synthesize phaseolotoxin because it lacks the entire Pht cluster for phaseolotoxin biosynthesis [5,6]. PCR analyses using primers directed to all genes from the Pht cluster did not yield any amplification product using DNA from CYL233 as template (data no shown), supporting the idea that this strain really lacks the entire Pht cluster. Plasmids pSAK; pSAK-A; pSAK-B; pSAK-C; pSAK-AB; pSAK-BC; pSAK-AC and pSAK-ABC were constructed containing genes argK; argK-phtA; argK-phtB; argK-phtC; argK-phtAB; argK-phtBC; argK-phtAC and argK-phtABC cloned into pUCP20, respectively (Table 1; Figure 3A). These constructions were electroporated into P. syringae pv. phaseolicola strain CYL233 and the argK expression pattern was evaluated by Northern blot analysis. a) argK expression in P. syringae pv. phaseolicola CYL233 wild type strain. Strains were grown in M9 medium at 18uC and 28uC. In strain CYL233(pSAK) we observed argK transcripts at both temperatures, as it occurs in mutant YNorf1P and in contrast with the expression observed in the wild type strain NPS3121, where argK expression is only observed at 18uC ( Figure 4A). These results indicate that the molecule that repress argK gene expression at 28uC is either not present or inactive in the wild type strain CYL233, suggesting that this repressor molecule could be coded within the Pht cluster. On the other hand, the argK expression observed at 18uC in strain CYL233 was similar to that in strain NPS3121 ( Figure 4A), showing that the Pht cluster is dispensable to carry out an efficient argK transcription. b) Effect of phtA, phtB and/or phtC on argK transcription in strain P. syringae pv. phaseolicola CYL233. We evaluated the effect of phtA, phtB and phtC on argK transcription. To achieve this we constructed different plasmids containing argK and a combination of genes phtA, phtB and phtC transcribed divergently to argK ( Figure 3A) that were transferred to strain P. syringae pv. phaseolicola CYL233. Northern blot analyses revealed a clear argK transcriptional repression effect when this gene was transcribed together with phtABC in strain CYL233 at both temperatures ( Figure 4B). This repression effect was not so efficient with other combination of genes phtA, phtB and phtC ( Figure 4B).
Effect of phtABC genes on argK transcription in trans in strain P. syringae pv. phaseolicola NPS3121 To evaluate this, we introduced plasmids pSAK and pSAK-ABC in the wild type strain NPS3121 and evaluated the expression pattern of argK in each derivative ( Figure 4C). In strain NPS3121(pSAK), we could still observe argK expression at 28uC ( Figure 4C). This result was not unexpected, since pSAK is a plasmid that occurs in multiple copies [31] and the cloned argK gene would probably titrate the putative repressor. However, in a NPS3121 derivative containing pSAK plus genes phtABC (plasmid pSAK-ABC, Figure 3), the transcription of argK was regulated by temperature.
Induction of argK expression with carbamoylphosphate in strain P. syringae pv. phaseolicola CYL233 A model of phaseolotoxin regulation has been proposed in which, at permissive temperature for phaseolotoxin production, 18uC, an inducer molecule could bind the postulated repressor molecule of argK to release it from the argK operator allowing its expression [21]. It has also been proposed that such inducer could be a precursor of phaseolotoxin. Carbamoylphosphate, which presents a similar structure to that of the inorganic moiety of phaseolotoxin, induces argK expression at 28uC [21]. We decided to determine whether this molecule was able to eliminate the argK transcriptional repression caused by phtABC. Northern blot analyses of RNA extracted from cells grown in M9 medium at 28uC until the end of the logarithmic phase, showed that there was not a de-repression effect caused by carbamoylphosphate in strain CYL233(pSAK-ABC) ( Figure 4D).
Transcriptional fusions to the uidA reporter gene in strain P. syringae pv. phaseolicola CYL233 In order to investigate the regulatory mechanism exerted by phtA, phtB and phtC products on argK transcription, we decided to construct transcriptional fusions to uidA (GUS) reporter gene ( Figure 3B). Transcription of the GUS reporter gene was determined in the wild type strain CYL233 by a qualitative colorimetric assay using X-Gluc as substrate. The strains were grown in M9 medium at 18uC and 28uC, and the presence or absence of blue color in the culture medium was evaluated. Reporter activation was detected for cultures of CYL233(pSAF-P KA ), indicating that the argK promoter (P KA ) was able to drive uidA expression ( Figure S1). Likewise, we observed argK promoter activity for all the constructions containing genes phtA, phtB or phtC, either by themselves or in combination ( Figure S1). The only exception was CYL233(pSAF-ABC), containing genes phtABC, which resulted in cultures that did not develop any color; moreover, with this strain we could not detect even the background blue color produced with the promoterless construct, pSAF. These results complement our observations obtained by Northern blot hybridization, and support the hypothesis that the phtABC products participate in the transcriptional regulation of gene argK, interacting either directly or indirectly with the argK promoter.

DNA electrophoretic mobility shift assay
Results obtained using transcriptional fusions suggest that the phtA, phtB and phtC products participate in the regulation of gene argK at the transcriptional level. To analyze the possibility that the phtABC products could bind to the promoter region of argK we performed an electrophoretic mobility shift assay. The argK promoter has been previously defined and shown to be a Pribnow (s70) type promoter with appropriate 210 and 235 regions [19,20].
Experiments were carried out by using a 288-bp DNA probe from the argK promoter region containing the 210 and 235 regions. A clear retardation signal of the P K probe was observed when crude extracts from NPS3121, CYL233 and CY-L233(pSAK-ABC) strains, grown at 18uC or 28uC in M9 medium, were added to the retardation mixture ( Figure 5A). These results indicate that a molecule coded within the chromosome of P. syringae pv. phaseolicola was bound to this probe. Also, a second retardation signal was observed when crude extracts from CYL233(pSAK-ABC) strain grown at 28uC were added to the retardation mixture ( Figure 5A). Specific binding to the argK promoter probe was demonstrated when the nonlabeled P K probe efficiently replaced the labeled probe, causing the almost complete disappearance of the retardation signal. A nonlabeled argF promoter probe failed to compete the binding of the labeled P K probe ( Figure 5A); indicating that unspecific binding to the probe was not occurring.
We observed a second retardation signal when the P K probe was incubated with crude extracts from strain CYL233 containing genes phtABC ( Figure 5A); however, it is unlikely that the products of these genes will directly bind to the argK promoter. The deduced product of phtA belongs to the ''P-loop containing nucleoside triphosphate hydrolases'' superfamily (InterPro SSF52540), whereas there were no hits for the products of phtB and phtC in an InterProScan comparison. Therefore, these products probably do not have DNA binding domains that would suggest a possible function as repressor proteins. Additionally, database comparisons using available BLAST [32] did not find any similarity with proteins related to transcription factors or DNA binding proteins.
We decided to investigate whether the proteins binding the argK promoter were present in other P. syringae pathovars and in E. coli. Results are shown in Figure 5B, where we observed only one retardation signal in all strains, indicating that all analyzed strains produce the unknown protein that binds to the argK promoter.

Discussion
Regulation of the phaseolotoxin biosynthesis cluster is very complex and there are at least three regulatory circuits that differentially affect its five transcriptional units. The global regulators GacA/GacS are required for the expression of all the transcriptional units, except for argK, which becomes constitutive in a gacA 2 mutant background [29]. Additionally, the expression of the phtM transcriptional unit depends on the activity of gene argK Regulation in P. syringae pv. phaseolicola PLOS ONE | www.plosone.org phtL, also included in the Pht cluster [25]. Finally, the global regulator IHF (Integration Host Factor) participates in the repression control of the phtD operon [33].
Expression of gene argK, conferring resistance to phaseolotoxin, appears to depend on different regulatory circuits than the rest of the Pht cluster. It is well known that fragments containing genes involved in toxin synthesis commonly carry regulatory elements that directly control the expression of biosynthetic genes [34]. An example are corP, corS, and corR genes, a modified two-component regulatory systems involved in the regulation of coronatine synthesis by P. syringae pv. glycinea PG4180. These regulatory genes have been located within the coronatine gene cluster [34,35,36]. In support of early predictions about argK regulation [20,21,22,23], our results indicate that regulation of argK in P. syringae pv. phaseolicola NPS3121 is carried out by a repressor protein that prevents its expression at 28uC; additionally, we show that genes phtABC, from the Pht cluster, are essential for this repression.
The unknown protein participating in the repression of gene argK appears to be normally present in a variety of bacteria, because crude cell extracts from different P. syringae and E. coli strains contained a protein able to bind the argK promoter. This is not surprising, because it is well known that several regulatory genes involved in phytotoxin synthesis could also be coded outside the toxin biosynthesis cluster [34,37,38]. An example are the syringomycin and syringopeptin biosynthesis clusters from P. syringae pv. syringae, which are regulated by a two components regulatory system, GacS/GacA, and the regulator SalA, all coded outside the toxins gene clusters [37,39]. Additionally, several genes located outside the Pht cluster are involved in phaseolotoxin synthesis and regulation in P. syringae pv. phaseolicola NPS3121. These genes include locus PSPPH_4550, coding for a putative nonribosomal peptide synthetase involved in phaseolotoxin production [29]; the GacS/GacA system, which is involved in the global regulation of phaseolotoxin biosynthesis genes [29] and the integration host factor, which participates in the repression of the phtD operon [33]. These facts, along with the results from this work support the idea that, following its acquisition by horizontal transfer, the Pht cluster has integrated into the regulatory circuits of P. syringae pv. phaseolicola.
The putative argK repressor protein was also present in the nontoxigenic strain P. syringae pv. phaseolicola CYL233, as shown by a specific retardation of the Pk promoter in gel-shift assays. Unlike what happened with other bacteria, we observed a second retardation signal when using extracts from strain CYL233 containing phtABC and grown at 28uC, which necessitates of further experiments for clarification. Nevertheless, it is unlikely that the products of genes phtABC bind directly to the argK promoter to repress transcription, because they did not contain any conserved regulatory or DNA-binding domain in comparison with the InterPro databases. Gene phtA belongs to the P-loop containing nucleoside triphosphate (NTP) hydrolases superfamily (InterPro SSF52540), whose members can function as kinases with very different specificities as different kinds of motor proteins and as batteries to drive reactions through conformational change [40], suggesting an enzymatic role for the corresponding phtA product. We can therefore speculate that the products of the phtABC genes participate in the biosynthesis of a phaseolotoxin precursor that Figure 4. Effect of cloned argK and phtABC genes on the argK expression in P. syringae pv. phaseolicola. The expression of argK was evaluated by Northern blot in derivatives of strain CYL233 (nontoxigenic; panels A and B) and strain NPS3121 (toxigenic, panel C) harboring plasmids with diverse combinations of genes argK, phtA, phtB and phtC, as indicated above each blot (see Figure 3 for a description of each plasmid). Panel D shows the expression of argK in the absence or presence of carbamoylphosphate (CP) in strains NPS3121 and CYL233 at 28uC. Blots were hybridized with an internal probe specific for argK, and the signal corresponding to the monocistronic argK RNA is marked. The asterisks indicate the position of a band of approximately 2.3-kb, corresponding to a previously described possible alternative argK transcript. Strain CYL233(pUCP20) was used as negative control of argK expression, whereas the wild type strain NPS3121 was used as a positive control. The numbers on top of the Northern blots represent the temperatures at which expression was assayed: 1 indicates 18uC and 2 indicates 28uC. doi:10.1371/journal.pone.0046815.g004 can also bind the argK repressor molecule, leading to active repression. Genes phtABC lead to a very efficient repression of argK transcription in the nontoxigenic strain CYL233, when all were cloned in a high-copy number vector; remarkably, this repression was not alleviated at 18uC, a temperature at which argK is normally expressed. However, argK displayed the wild type expression pattern (i.e., expressed at 18uC but not at 28uC) when using the same construction in the toxigenic strain NPS3121. Carbamoylphosphate is one of the substrates for OCTase in the arginine biosynthesis pathway, and exogenously supplied carbamoylphosphate was shown to induce argK expression at 28uC in strain NPS3121 [21]. However, carbamoylphosphate did not relieve the repression of argK in strain CYL233. We do not have yet a satisfactory explanation for this result, although a likely possibility is that other genes of the Pht cluster are involved in the transformation of carbamoylphosphate into an inducer of argK transcription. This is the first study that identify genes involved in argK regulation in P. syringae pv. phaseolicola. Further work to understand this intriguing and interesting regulatory mechanism is currently under way in our laboratory.

Molecular biology techniques
Routine techniques were performed as described [41]. Purification of DNA from agarose gels and plasmid minipreps were performed using Qiagen columns and kits (Valencia, Ca, USA). Chromosomal DNA from P. syringae pv. phaseolicola was obtained as described [45]. DNA fragments used as probes for Northern blots were labelled with [a-32 P]dCTP using the Rediprime random primer labeling kit (Amersham). Protein concentration in bacterial lysates was estimated by the procedure of Bradford [46].

Analysis of argK transcriptional pattern by Reverse
Transcription-PCR analysis DNA-free RNA was obtained from cultures grown in M9 medium at 18uC and 28uC when they reached an O.D. 600 of 0.8. The RNA integrity was verified in a denaturing agarose gel and used for reverse transcription (RT) and PCR using the SuperScript One-Step kit (Invitrogen). Samples of RNA (20 ng) were used in each RT-PCR reaction. A 608-bp DNA amplicon from gene argK was obtained by RT-PCR using specific primers (Table S1). Controls used for the set of primers were: 1) PCR without the reverse transcription step to verify the absence of DNA; 2) RT-PCR performed without RNA templates to detect any contaminating DNA/RNA; 3) PCR performed using chromosomal DNA as template to ensure primer fidelity; 4) Amplification of a portion of the 23S ribosomal RNA operon using suitable primers as an internal control of the reaction. The RT reaction was performed at 50uC for 40 min, followed by PCR amplification at 94uC for 2 min for 1 cycle; 94uC for 35 s, 58uC for 30 s, 72uC for 1 min for 25 cycles; 72uC for 15 min for 1 cycle. The quantity of amplicon between samples was compared.

OCTase activity assays
The OCTase activity was determined by measuring OCTase specific activity as previously described [24] by using 5 ml cultures of strains NPS3121 (wild type) and Ynorf1P (phtA 2 ) grown in M9 medium at 28uC during 24 h. Cells were disrupted with a VirTis sonicator (model VirSonic 60), and 3 ml of this crude extract were used for the assay. The reaction mixtures were incubated at 37uC for 20 min and OCTase activity was determined. To eliminate OCTase activity from samples due to the phaseolotoxin-sensitive OCTase, coded by gene argF, phaseolotoxin-containing supernatant was added to the reaction mixture. To determine the amount of supernatant enough to achieve an efficient OCTase inhibition, we constructed a standard curve (OCTase activity vs. supernatant) (data no shown).The addition of a phaseolotoxin-containing supernatant to the reaction mixture and a preincubation for 25 min was made before OCTase activity was determined. OCTase specific activity was reported as nmol of citrulline produced per mg of protein per min at 37uC.
Construction of plasmids containing argK, phtA, phtB and/or phtC genes The argK, phtA, phtB and phtC genes were obtained by PCR using primers designed to include suitable restriction sites (Table S1). The argK, argK-phtA, argK-phtAB and argK-phtABC amplicons were cloned into the PstI-SmaI sites of the pUCP20 vector to create plasmids pSAK, pSAK-A, pSAK-AB, and pSAK-ABC, respectively. The phtB, phtC and phtBC amplicons were cloned into the SmaI site of pSAK to create pSAK-B, pSAK-C and pSAK-BC, respectively. The phtC amplicon was cloned into the SmaI site of pSAK-A to create pSAK-AC (Table 1, Figure 3A). All constructions were confirmed by restriction pattern and electroporated into P. syringae pv. phaseolicola.

RNA extraction and Northern blot analysis
Total RNA was isolated from cultures of P. syringae pv. phaseolicola grown in M9 medium at 18uC or 28uC until an O.D. 600 of 0.8 using the TRIzol Reagent as recommended by the manufacturer (Invitrogen). Genomic DNA was removed by digestion with Rnase-free DNase (Roche). Samples of total RNA (25 mg) were denatured by treatment with formamide and separated by electrophoresis using 1.3% denaturing agarose gels. The RNA was transferred to Hybond N + nylon membranes (Amersham) and cross-linked by exposure to UV radiation. Hybridization was performed using NorthernMax Prehybridization/Hybridization buffer (Ambion) and a DNA probe for the argK gene. The argK probe was labeled with [a-32 P]dCTP by using the RediPrime Random Primer labeling kit. Hybridization was carried out overnight at 60uC. The membranes were washed twice with 26 SSC-0.1% sodium dodecyl sulfate (16 SSC is 0.15 M NaCl and 0.015 M sodium citrate) for 5 min at room temperature, followed by a wash with 16 SSC-0.1% sodium dodecyl sulfate for 15 min at 60uC. The membranes were exposed to a phosphorus screen and the signal detected in a Storm 860 apparatus. The analysis was made using ImageQuant version 1.1 (BioRad).

Induction assays with carbamoylphosphate
Induction assays were conducted as previously reported [21] using cultures of P. syringae pv. phaseolicola grown in M9 medium at 28uC to an O.D. 600 of 0.8. Carbamoylphosphate disodium salt (Sigma) was added to the cultures at 1 mM and, after a 30 min incubation, RNA was purified from cultures.
Colorimetric GUS assays of transcriptional uidA fusions The 1.8-kb EcoRI-SalI fragment containing a promoterless uidA gene was obtained from pWM6 [47]. This fragment was cloned into pUCP20 to construct pSAF, containing the promoterless uidA gene preceded by a SD sequence. Fragments containing the argK promoter and phtA promoter (P KA ), P KA -phtA, P KA -phtAB, P KA -phtABC were obtained by PCR using primers designed to include suitable restriction sites (Table S1) and were cloned into the BamHI-SmaI sites of pSAF to create pSAF-P KA , pSAF-A, pSAF-AB and pSAF-ABC respectively. The phtB, phtC and phtBC amplicons were cloned into the SmaI site of pSAF-P KA to made pSAF-B; pSAF-C and pSAF-BC respectively. The phtC amplicon was cloned into the SmaI site of pSAF-A to create pSAF-AC (Table 1, Figure 3B). The orientations were determined by restriction pattern and constructions were electroporated into strain P. syringae pv. phaseolicola CYL233. The activation of the GUS reporter was determined by a colorimetric assay [48]. To analyze GUS activity, P. syringae pv. phaseolicola strains carrying uidA fusions were grown in M9 medium at 18uC and 28uC and when they reached an O.D. 600 of 0.8 were used. Culture dilutions (1:10) were made and X-Gluc (20 mg/ml) was added. They were incubated at 18uC and 28uC during 6 h and 3 h, respectively, and the presence or absence of blue color in the culture medium was evaluated.
DNA Electrophoretic mobility shift assay DNA electrophoretic mobility shift was determined as previously reported [24,49]. A 288-bp DNA probe from the argK promoter (P K ) containing 210 and 235 regions, was obtained by PCR using specific primers (Table S1). The probe was labeled with [c-32 P]dATP by using the T4 polynucleotide kinase as recommended by the manufacturer (Invitrogen). Crude extract from cultures were obtained by sonication. The cell pellet was resuspended in extraction buffer [25 mM Tris (pH 8.0), 0.1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, 0.02 mM phenylmethyl-sulfonyl-fluoride] and sonicated using a VirTis sonicator (model VirSonic 60) with three pulses for 10 s with pauses for 10 s. Protein concentration in bacterial lysates was estimated by the procedure of Bradford [46]. The retardation reaction mixture [20 mM Tris (pH 8.0), 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 5% glycerol, 10 ng/ml of poly(dI-dC), 5 mg of crude extract], was preincubated for 15 min at room temperature, the radioactively labeled probe was added, and the mixture was incubated for 25 min at room temperature. Samples were loaded onto 5% polyacrylamide gels and separated at 7 mA for 90 min using Tris-acetate-EDTA buffer. In the competition assays, nonlabelled argF promoter probe and nonlabeled pK probe were added. Gels were dried and exposed to phosphorus screen and signal detected in a Storm 860 apparatus and analyzed using ImageQuant version 1.1 (BioRad). Figure S1 Qualitative assays of argK expression using transcriptional fusions with the uidA gene. The wild type strain CYL233 was transformed with plasmids containing a transcriptional fusion between the promoter of argK and the uidA gene, and different combinations of genes phtA, phtB and/or phtC (see Figure 3 for a description of each plasmid). Cultures were then incubated in the presence of X-Gluc at 18uC and 28uC during 6 h and 3 h, respectively. Each strain is indicated above the corresponding tube. The activation of the GUS reporter was evaluated as presence or absence of blue color in the culture medium. (TIF)