Virulence of the Phytopathogen Pseudomonas syringae pv. Maculicola Is rpoN Dependent

Abstract

We cloned the rpoN (ntrA and glnF) gene encoding ς⁵⁴ from the phytopathogen Pseudomonas syringae pv. maculicola strain ES4326. The P. syringae ES4326 rpoN gene complemented Pseudomonas aeruginosa, Escherichia coli, and Klebsiella aerogenes rpoN mutants for a variety of rpoN mutant phenotypes, including the inability to utilize nitrate as sole nitrogen source. DNA sequence analysis of the P. syringae ES4326 rpoN gene revealed that the deduced amino acid sequence was most similar (86% identity; 95% similarity) to the ς⁵⁴ protein encoded by the Pseudomonas putida rpoN gene. A marker exchange protocol was used to construct an ES4326 rpoN insertional mutation, rpoN::Km^r. In contrast to wild-type ES4326, ES4326 rpoN::Km^r was nonmotile and could not utilize nitrate, urea, C₄-dicarboxylic acids, several amino acids, or concentrations of ammonia below 2 mM as nitrogen sources. rpoN was essential for production of the phytotoxin coronatine and for expression of the structural genes encoding coronamic acid. In addition, ES4326 rpoN::Km^r did not multiply or elicit disease symptoms when infiltrated into Arabidopsis thaliana leaves, did not elicit the accumulation of several Arabidopsis defense-related mRNAs, and did not elicit a hypersensitive response (HR) when infiltrated into tobacco (Nicotiana tabacum) leaves. Furthermore, whereas P. syringae ES4326 carrying the avirulence gene avrRpt2 elicited an HR when infiltrated into Arabidopsis ecotype Columbia leaves, ES4326 rpoN::Km^r carrying avrRpt2 elicited no response. Constitutive expression of ES4326 hrpL in ES4326 rpoN::Km^r partially restored defense-related mRNA accumulation, showing a direct role for the hrp cluster in host defense gene induction in a compatible host-pathogen interaction. However, constitutive expression of hrpL in ES4326 rpoN::Km^r did not restore coronatine production, showing that coronatine biosynthesis requires factors other than hrpL.

The rpoN gene encodes the alternate sigma factor ς⁵⁴, which works in conjunction with the NtrC class of transcriptional activators to control the expression of many genes in response to nutritional and environmental conditions (2, 54). For example, genes involved in nitrogen, hydrogen, and catabolite utilization are frequently regulated by ς⁵⁴ (6, 35, 48, 95). In the case of pathogenic bacteria, rpoN mediates expression of virulence-related factors such as pilin in Pseudomonas aeruginosa and flagellin in Vibrio anguillarum (24, 61, 86).

For some phytopathogenic bacteria, rpoN has been implicated indirectly as a regulator of pathogenicity-related genes known as the hrp gene cluster (17, 18). Pseudomonas syringae pv. syringae strain 61, for example, contains a 25-kb hrp cluster consisting of several complementation groups comprising at least 27 genes (8, 26). Several hrp genes encode proteins that have a high degree of homology to components of the type III secretory pathway of Yersinia species which are responsible for translocating Yersinia outer membrane proteins into mammalian cells (13, 15, 55, 83). By analogy, it is proposed that hrp-encoded proteins in phytopathogenic bacteria are involved in the transport of pathogenicity-related factors into plant cells.

The acronym hrp stands for hypersensitive response and pathogenicity. hrp genes are required not only for pathogenicity of a virulent pathogen but also for the elicitation of the hypersensitive response (HR) which occurs on some hosts (44, 45). The HR involves rapid, but localized, programmed plant cell death and is believed to restrict pathogen spread (1, 37). There is mounting evidence that the elicitation of an HR is mediated by the specific interaction between the products of a plant resistance gene (R gene) and a pathogen avirulence (avr) gene (43, 80, 85). It appears likely that at least some avr genes encode pathogenicity-related factors (34, 47, 71, 84) that are transported into plant cells via the hrp-encoded transport machinery (58, 59, 90). In the absence of a corresponding R gene product, the avr product enhances virulence; however, in hosts which have the corresponding R gene, recognition of the avr gene product enhances host resistance. Interestingly, most avr genes are also coordinately regulated with genes in the hrp cluster (27, 29, 46, 67, 71, 77, 97).

The HR is accompanied by the induction of defense-related genes (7, 91) that are differentially expressed depending on the particular pair of avr and R gene products eliciting the HR (70, 72). Defense gene induction also occurs in the absence of the HR during compatible pathogen-host interactions, although usually later and at lower levels than those occurring during an HR (11, 32, 66). Furthermore, hrp mutations that presumably block the export of avr gene products have been found to reduce, but not eliminate, defense gene induction (60). Collectively, these results suggest that there are a variety of signaling pathways that activate host responses.

In P. syringae the circuitry of hrp regulation appears to involve a transcriptional activation cascade. At the top of the cascade are two regulatory genes, hrpR and hrpS, which are required for expression of the remaining hrp genes in the cluster (12, 18). Both hrpR and hrpS encode proteins consist almost exclusively of the domain conserved among transcriptional activators such as NtrC, DctD, and NifA that work in concert with ς⁵⁴ (21, 22; reviewed in references 2 and 54). The hrpR product activates hrpS expression, and the hrpS product activates hrpL transcription (17, 18). HrpL is an alternate sigma factor homologous to AlgU of P. aeruginosa and is thought to activate transcription of the remaining genes in the hrp cluster (96, 97). The factor(s) involved in the regulation of hrpRS remains obscure. Nevertheless, the central role of hrpS in this cascade and the HrpS-NtrC homology predicts that rpoN would be required for activation of the hrp gene cluster in P. syringae.

Despite the highly conserved and clustered nature of hrp genes among phytopathogenic bacteria, transcriptional regulation of the hrp genes is achieved by different mechanisms in different species. In Ralstonia solanacearum, HrpB, an AraC-like transcriptional activator, controls hrp gene expression (14). Similarly, in Xanthomonas campestris pv. vesicatoria, hrp gene expression is regulated by an OmpR homolog, HrpG, which in turn activates HrpX, another AraC-like activator that activates the remaining hrp genes (79, 92). This latter regulatory cascade is consistent with the fact that rpoN is not required for hrp expression or pathogenicity in X. campestris pv. vesicatoria (25).

The experiments described here utilize a pathogenicity system that involves the infection of Arabidopsis thaliana with P. syringae pv. maculicola strain ES4326. Strain ES4326 belongs to the leaf spotting class of phytopathogenic pseudomonads (78), proliferates extensively in Arabidopsis ecotype Columbia leaves, and causes the development of water-soaked disease lesions (9, 11). In contrast, ES4326 carrying the avirulence gene avrRpt2 elicits a visible HR about 16 h after infiltration and proliferates 50- to 100-fold less than the wild-type strain ES4326 (11). Using this system, we describe experiments that examine the role of ς⁵⁴ in the pathogenicity of P. syringae ES4326. Our results indicate that ς⁵⁴ is an important virulence factor for P. syringae and is required for the elicitation of an HR by P. syringae in both host and nonhost plants.

MATERIALS AND METHODS

Bacterial strains and media.

Bacterial strains and plasmids used in this work are listed in Table 1. P. syringae pv. maculicola strain ES4326 and its derivatives were grown at 28°C in L broth (50), minimal M9 salts media (50), or King's B medium (36). Escherichia coli, Klebsiella aerogenes, and P. aeruginosa strains were grown at 37°C in L broth or M9 minimal salts medium. For clarity, ES4326 carrying plasmid pLH12 (which carries the avirulence gene avrRpt2) is referred to as ES4326 (avrRpt2). Carbon and nitrogen source utilization tests for ES4326 rpoN mutants were performed in M9 salts minimal medium by providing a carbon source at 10 mM and by replacing ammonium chloride with an alternative nitrogen source at 5 mM when required. Bacterial motility was tested on “swarm plates” consisting of 0.3% agar, 0.5% NaCl and 0.5% tryptone (38). Antibiotic concentrations for E. coli and P. syringae strains were as follows: streptomycin, 150 μg/ml; kanamycin, 25 μg/ml; tetracycline, 12 μg/ml; gentamicin, 20 μg/ml; and spectinomycin, 20 μg/ml. Interspecies complementation tests of the E. coli rpoN mutant by ES4326 rpoN were carried out on M9 minimal salts agar supplemented with 0.2% glutamine and 20 μg of 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) per ml (Sigma).

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Characteristics	Source or reference
Strains
E. coli
DH5α	F−lacZΔM15 endA1 recA1 hsdR17 supE44 thi-1 gyrA relA1 λ−	Bethesda Research Laboratories (20)
MM294		G. Walker (4)
Th1 λgln101	lacΔ rpoNΔ glnA-lacZ	D. Ow (62)
K. aerogenes HG63 rpoN	rpoN	D. Ow (62)
P. aeruginosa PAK N1	Gmr cassette in rpoN	S. Lory (31)
P. syringae pv. maculicola
ES4326	Wild-type	K. Davis (9)
ES4326 rpoN::Kmr	Kmr cassette in rpoN	This study
P. syringae pv. tomato
DC3000	Wild type	D. Cuppels (57)
DC3661	COR-defective mutant	D. Cuppels (57)
P. syringae pv. glycinea
PG4180.N9	COR-producing strain	C. Bender (88)
PG4180.P2	Gmr cassette in corR	C. Bender (65)
Plasmids
pRK2013	Kmr Tra+, mating helper	G. Ditta (10)
pPH1	Gmr Spr, chasing plasmid
pLAFR3	Tcr cosmid cloning vector	B. Staskawicz (81)
pBluescript SK(+)	Apr, cloning vector	Stratagene, Inc., La Jolla, Calif.
pGEM-7Zf(+)	Apr, cloning vector	Promega Inc., Madison, Wis.
pUC4K	Source of Kmr cassette from Tn903	Pharmacia, Inc., Bridgewater, N.J.
pJSR1	Apr, cosmid cloning vector	J. Shao (68)
pKI11	Source of P. aeruginosa rpoN gene	S. Lory (31)
pLH12	1.4-kb SalI fragment containing avrRpt2 in pLAFR3	R. Innes (93)
pPG101	17.2-kb insert of ES4326 DNA carrying rpoN in pLAFR3	This study
pPG102	4.5-kb BamHI-EcoRI fragment containing ES4326 rpoN in pLAFR3	This study
pLAFR-RK	pLAFR3 containing rpoN::Kmr	This study
pRN5; pNR9	pBluescript SK(+) carrying 4.5-kb HindIII fragment containing ES4326 rpoN gene in opposite orientations	This study
pRG960sd	Smr Spr, contains promoterless uidA with start codon and Shine-Dalgarno sequence	C. Bender (89)
pRGMU7	Smr Spr, contains 1.5-kb PstI-AatI fragment containing cmaABT promoter inserted into pRG960sd	C. Bender (87)
pHRPLC	lacZ-hrpL transcriptional fusion in pLAFR3	E. Hendrickson (23)

Bacterial genetics.

pLAFR3 derivatives were introduced into Pseudomonas strains via triparental matings with MM294(pRK2013) (10). Aspartate-utilizing pseudorevertants of strain ES4326 rpoN::Km^r were obtained by plating approximately 10⁹ CFU on M9 agar plates containing succinate and aspartate as carbon sources at 10 mM and the appropriate antibiotics (streptomycin and kanamycin).

Plant pathogenicity assays.

Arabidopsis ecotype Columbia was germinated, grown, and inoculated with ES4326 strains. Bacterial strains were grown overnight in King's B, subcultured and grown to mid-log phase, resuspended in 10 mM MgSO₄, and inoculated into the underside of the leaf at a titer of 10⁴ CFU/cm² using a disposable syringe. Growth of P. syringae strains in leaves was measured by individually grinding four to six 0.2-cm² leaf punches (excised with a no. 2 cork borer) in 10 mM MgSO₄, plating appropriate dilutions on King's B medium containing the appropriate antibiotics, and counting the CFU. For RNA blot analysis, entire Arabidopsis leaves infiltrated with ES4326 bacterial suspensions were harvested, frozen in liquid nitrogen at the indicated times, and stored at −80°C until needed (see below). Nicotiana tabacum (tobacco) cultivar Xanthi was grown under greenhouse conditions and inoculated with ES4326 strains and assayed for the HR as previously described (82).

Cosmid library constructions.

Total bacterial genomic DNA was prepared from strain ES4326 as described previously (3), partially digested with Sau3A, and size fractionated on a 14-ml sucrose gradient (50). DNA fragments of approximately 20 kb were purified and ligated with linearized pLAFR3 that had been prepared to promote the formation of concatemers (50). Packaging, infection, and plating of the cosmid clones were performed using the Giga Gold packaging kit according to the manufacturer's specifications (Stratagene, La Jolla, Calif.).

Nucleic acid manipulations.

Routine manipulations such as DNA blots and plasmid DNA isolation were performed as described earlier (3). Restriction enzymes, T4 DNA ligase, and calf intestine phosphatase were purchased from Boehringer Mannheim and New England BioLabs and used according to the manufacturer's specifications. Deletions in plasmids were created using the Erase-a-Base kit (Promega, Madison, Wis.). Isolation of Arabidopsis mRNA and RNA blot analysis was carried out as described previously (11). Hybridizations were performed at stringent conditions (2× SSC [1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate], 65°C) as described earlier (50). ³²P-labeled DNA probes for use in hybridizations were prepared as described previously for the Pal1, PR1, BGL2 (PR2), PR5, and GST1 pathogen-induced genes (16, 73).

Cloning and sequencing the ES4326 rpoN gene.

DNA blot analysis data indicated that ES4326 contained a single 4.5-kb HindIII fragment that hybridized to a 4.2-kb XhoI fragment in plasmid pKI11 containing the P. aeruginosa strain PAK rpoN gene (31). Approximately 1,400 clones from a cosmid library of strain ES4326 DNA were screened by colony hybridization using the rpoN probe derived from P. aeruginosa. A hybridizing clone, pPG101, carrying a 17.2-kb insert was identified and shown to contain the 4.5-kb HindIII fragment previously detected by Southern blot analysis (data not shown).

E. coli strain TH1 λgln101, which contains deletions in rpoN and lacZ and a glnA-lacZ reporter construct, was used as an assay for functional ES4326 rpoN clones by plating subclones of pPG101 onto M9 medium containing 0.2% glutamine and 20 μg of X-Gal per ml. The 4.5-kb HindIII fragment from pPG101 contained a functional rpoN gene that complemented the E. coli rpoN mutation in TH1 λgln101. The rpoN gene in this construct presumably contained its own promoter since this fragment activated the glnA-lacZ fusion in TH1 λgln101 when cloned in the HindIII site of pBluescript SK(+) in either orientation (plasmids pRN5 and pRN9). For subsequent use in Pseudomonas spp., the 4.5-kb fragment containing ES4326 rpoN was subcloned into cosmid pLAFR3. A 4.5-kb HindIII fragment from pRN5 was subcloned into pGem7Zf. Using the pGem7Zf polylinker sites, the rpoN gene was recloned as a 4.5-kb EcoRI-BamHI fragment in pLAFR3 to produce pPG102. Plasmids pRN5 and pRN9 were used to derive a series of nested deletions starting from either end of the 4.5 HindIII fragment. This analysis showed that the ES4326 rpoN gene was located near the left end of the 4.5-kb fragment (data not shown).

DNA sequence analysis was initiated at the middle XhoI site in Fig. 1 and continued in both directions using synthetic oligonucleotides for a total of approximately 1,900 bp. A single large open reading frame that encodes a protein that is highly homologous to ς⁵⁴ in other bacterial species was identified in the 1,900-bp region.

FIG. 1.

Restriction map of the ES4326 rpoN region. The stippled box indicates the region sequenced. The horizontal arrow indicates the direction of transcription. Restriction enzyme sites: HIII, HindIII; RI, EcoRI; RV, EcoRV; P, PstI; K, KpnI; X, XhoI.

Insertional mutagenesis of the ES4326 rpoN gene.

pGem7Zf containing the 4.5-kb rpoN fragment was digested with PstI, and a 1.24-kb fragment encoding the aminoglycoside 3′-phosphotransferase activity of Tn903 (Pharmacia) from pUC4K was ligated to the PstI-digested ends. A 5.8-kb EcoRI-BamHI fragment from this plasmid was then subcloned into pLAFR3. The resulting plasmid (pLAFR-RK) was used to recombine the mutated rpoN gene (referred to as rpoN::Km^r) into strain ES4326 by first mobilizing pLAFR-RK into ES4326 and then introducing plasmid pPH1, which confers gentamicin (Gm) resistance. Cultures were grown under selection for Km^r and Gm^r, and individual colonies were screened for tetracycline sensitivity (76). Southern blot analysis of chromosomal DNA prepared from a putative rpoN mutant confirmed the insertion of Km^r into the rpoN gene (data not shown).

COR preparation and assay.

Coronatine (COR) synthesis by strain ES4326 was assayed using two approaches. In the first procedure (74), a 5-ml culture of ES4326 grown overnight in King's B medium was used to inoculate 50 ml of Woolley's liquid medium (94), where potassium nitrate was replaced with 5 mM arginine to facilitate rapid growth of the ES4326 rpoN-Km^r mutant. Cultures (50 ml) were shaken at 20°C for 6 days, at which point the optical density at 600 nm (OD₆₀₀) was measured, and the cells were pelleted and weighed. The supernatants were acidified to pH 2.0 with HCl and extracted with 50 ml of ethyl acetate. The organic phase was lyophilized to dryness, and the residue was resuspended in 2.0 ml of H₂O/g of wet bacterial pellet. Then, 10-μl droplets containing dilutions of either purified COR or the COR preparation described above were inoculated into Arabidopsis and tomato leaves. Elicitation of red anthocyanin pigments on Arabidopsis leaves and chlorosis on tomato leaves was assayed 3 to 7 days later.

In the second method, P. syringae strains were grown at 18°C in Hoitink-Sinden medium optimized for COR production (HSC) (63), and the supernatants were analyzed for COR production by high-pressure liquid chromatography (HPLC) 5 days after inoculation (63). Each strain was inoculated into three replicate aliquots (10 ml) of HSC medium for evaluation of COR production, and each experiment was repeated.

Nucleotide sequence accession number.

The rpoN sequence from P. syringae pv. maculicola has been assigned accession number AF199600 in the GenBank database.

RESULTS

Cloning the strain ES4326 rpoN gene.

An interspecies hybridization strategy was used to isolate pPG101, a cosmid clone that carried a presumptive P. syringae ES4326 rpoN gene. Plasmid pPG101 complemented the inability of E. coli, K. aerogenes, and P. aeruginosa rpoN mutants (strains TH1 λgln101, HG63 rpoN, and PAKN1, respectively) to utilize 10 mM nitrate and 10 mM ammonia as sole nitrogen sources; the inability of the E. coli rpoN mutant to utilize histidine, arginine, or proline as sole nitrogen sources; and the lack of motility of the P. aeruginosa rpoN mutant. As described in Materials and Methods and as illustrated in Fig. 1, the ES4326 rpoN gene on pPG101 was mapped to a 4.5-kb-HindIII fragment, and a 1,900-bp region containing the presumptive rpoN gene was sequenced. DNA sequence analysis (Fig. 2) revealed that the presumptive ES4326 rpoN gene encodes a protein with 86% identity and 95% similarity to the rpoN gene of P. putida.

FIG. 2.

Amino acid alignment of the ES4326 protein with ς⁵⁴ from diverse bacterial species. The sequences are listed in decreasing order of conservation with ς⁵⁴ from P. syringae pv. maculicola. (P_Syr). Abbreviations and references: P_Put, Pseudomonas putida (30); P_Aer, Pseudomonas aeruginosa (33); A_Vin, Azotobacter vinelandii (53); S_Typ, Salmonella enterica serovar Typhimurium (66); E_Col, Escherichia coli (28). Shading: black, conserved amino acids; gray, conservative substitutions.

Construction and metabolic phenotypes of a strain ES4326 rpoN insertional mutant.

An ES4326 rpoN mutant was constructed by subcloning rpoN from pPG101, inserting a DNA cassette conferring kanamycin resistance into the PstI site (located at codon 162) of rpoN, transferring the mutated rpoN gene to pLAFR3, and marker exchanging the mutant gene into the ES4326 genome (see Materials and Methods). ES4326 rpoN::Km^r exhibited an array of phenotypes typical of rpoN mutants, including the inability to grow on nitrate and urea as sole nitrogen sources, lack of motility, and inability to grow on a variety of C₄-dicarboxylic acids as sole carbon sources, including aspartate, succinate, and fumarate, as well as the tricarboxylic acid intermediate α-ketoglutarate (data not shown). Unlike rpoN mutants of enteric bacteria (49, 86), however, ES4326 rpoN::Km^r was able to grow on glucose and ammonia as the sole carbon and nitrogen sources if ammonia was present at concentrations higher than 2 mM.

The ability of ES4326 rpoN::Km^r to utilize a variety of amino acids as nitrogen sources was tested on minimal M9 solid medium. The rpoN mutant failed to utilize aspartate, proline, histidine, and methionine as nitrogen sources, which supported growth of large colonies of wild-type ES4326. Both wild-type and the rpoN mutant formed large colonies on arginine, lysine, asparagine, glutamine, and glutamate, although the rpoN mutant grew somewhat slower than the wild type (5 days to form a large colony compared to 3 days for the wild type). The wild-type ES4326 formed small colonies, and the rpoN mutant formed even smaller colonies on serine, leucine, threonine, isoleucine, or alanine. Phenylalanine and cysteine were not utilized by the wild type or the rpoN mutant. As in the case of P. putida rpoN mutants, all amino acids that served as a sole nitrogen source for the ES4326 rpoN::Km^r mutant also served as sole carbon source with the exception of lysine, which served only as a nitrogen source. The best growth rate for ES4326 rpoN::Km^r was observed when the medium was supplemented with glutamate, where the growth was equivalent to the wild-type strain.

Strain ES4326 rpoN is nonpathogenic on Arabidopsis and cannot elicit an HR.

As described previously (11), infiltration of Arabidopsis leaves with ES4326 at a titer of 10³ to 10⁴ CFU/cm² of leaf area resulted in the development of characteristic disease symptoms, including spreading chlorosis, water-soaked lesions, and growth of the infiltrated bacteria to a titer of approximately 10⁷ CFU/cm² (Fig. 3A). In contrast, Arabidopsis leaves infiltrated with ES4326 rpoN::Km^r exhibited no symptoms, even when inoculated with 10⁸ CFU/cm². Furthermore, the titer of ES4326 rpoN::Km^r in Arabidopsis leaves remained consistently low for the duration of the experiment (Fig. 3A). ES4326 rpoN::Km^r carrying pPG102, which carries wild-type rpoN, exhibited the same pathogenic phenotype as ES4326, and its growth in Arabidopsis leaves was indistinguishable from that of the wild type (Fig. 3A). This result indicated that the nonpathogenic phenotype of ES4326 rpoN::Km^r was due to the disruption of the rpoN gene. However, because the complementing plasmid contains about 2,000 bp downstream of rpoN, these data do not rule out the possibility that the insertion in rpoN exerts polarity on a downstream gene and that this downstream gene is required for pathogenicity.

FIG. 3.

Growth of ES4326 rpoN::Km^r in Arabidopsis leaves. Six-week-old Arabidopsis (Columbia) seedlings were infiltrated with bacterial suspensions at a titer of 10⁴ CFU/cm² of leaf area, and bacterial populations were determined as described in Materials and Methods. Each value represents the average of at least four leaf discs. (A) Symbols: □, ES4326(pLAFR3); ■, ES4326 rpoN::Km^r (pLAFR3); ●, ES4326 rpoN::Km^r (pPG102) (rpoN⁺). The experiment was repeated with similar results. (B) Symbols: ○, ES4326(pLAFR3); ●, ES4326 (avrRpt2); ■, ES4326 rpoN::Km^r (pLAFR3); □, ES4326 rpoN::Km^r (avrRpt2). The experiment was repeated with similar results.

ES4326 carrying avrRpt2 on plasmid pLH12 [ES4326 (avrRpt2)] elicits an HR on Arabidopsis ecotype Columbia instead of disease symptoms, and the HR is accompanied by a 10- to 100-fold reduction in bacterial growth compared to ES4326 in infiltrated leaves (Fig. 3B). In contrast, ES4326 rpoN::Km^r (avrRpt2) failed to elicit an HR or any other visible symptom, even when the inoculum was 20 times higher than ES4326 (avrRpt2) (data not shown). As shown in Fig. 3B, ES4326 rpoN::Km^r (avrRpt2), like ES4326 (avrRpt2), failed to multiply in Arabidopsis leaves. Infiltration of ES4326 into tobacco leaves also results in the elicitation of an HR (not shown). In confirmation of the results obtained in Arabidopsis, infiltration of tobacco leaves with ES4326 rpoN::Km^r did not result in the appearance of any visible symptoms (not shown).

The pathogenicity defect of ES4326 rpoN::Km^r is not solely due to the inability to assimilate aspartate.

Because genes encoding dicarboxylic acid permease are regulated by DctD, an NtrC homolog, rpoN mutants in a variety of species cannot utilize dicarboxylic acids as carbon or nitrogen sources. R. meliloti mutants defective in aspartate aminotransferase cannot utilize aspartate and are defective in symbiosis, suggesting that aspartate may be a major carbon source for symbiotic bacterial cells (69). Therefore, we tested whether ES4326 rpoN::Km^r is nonpathogenic solely because it cannot assimilate aspartate. We selected a ES4326 rpoN::Km^r pseudorevertant, as described in Materials and Methods, that was able to utilize aspartate as a carbon source. This revertant still had an RpoN⁻ phenotype with respect to its inability to use nitrate and succinate and its lack of motility. The Asp⁺ revertant failed to grow in A. thaliana or to elicit an HR in tobacco or Arabidopsis (data not shown).

Strain ES4326 rpoN::Km^r fails to synthesize the phytotoxin coronatine.

One possible explanation for the reduced virulence of ES4326 rpoN::Km^r is the inability to produce virulence factors such as toxins. Many P. syringae pathovars, including ES4326, produce a chlorosis-inducing phytotoxin, coronatine, which is composed of an ethyl cyclopropyl amino acid linked to a polyketide moiety (19). COR production is regulated by a modified two-component regulatory system that controls the expression of essential COR biosynthetic genes. The regulators CorR and CorP are related to response regulators of the ROIII group, while CorS is similar to the corresponding histidine protein kinase sensors. To determine whether COR biosynthesis requires rpoN, crude COR was extracted from strains ES4326, ES4326 rpoN::Km^r, the COR-producing strain P. syringae pv. tomato DC3000, and DC3661, a COR⁻ mutant of DC3000. The COR extracted from DC3000 and ES4326 elicited typical COR-induced symptoms, chlorosis and anthocyanin accumulation, respectively, on tomato and Arabidopsis leaves. These symptoms were not detected with organic acids extracted from ES4326 rpoN::Km^r or DC3661. Further characterization was carried out by quantitatively analyzing COR production using HPLC. As shown in Table 2, ES4326 produced 51 mg of COR/g of protein, a level comparable to that produced by P. syringae pv. glycinea strain PG4180.N9, a high-yielding COR producer which has been used in many genetic investigations (65, 88). However, ES4326 rpoN::Km^r produced only 0.3 mg of COR, a level comparable to PG4180.P2, a corR mutant of PG4180.N9 which is considered completely defective in COR production (65).

TABLE 2.

COR production and cor gene transcriptional activity in selected strains of P. syringae^a

Strain	COR production (mg/g of protein)	Transcriptional activity (U of GUS/mg of protein)b
P. syringae pv. maculicola
ES4326	51.0	347.0
ES4326 rpoN	0.3	8.5
ES4326 rpoN + pHRPLC	0.0	ND
P. syringae pv. glycinea
PG4180.N9	43.4	390.0
PG4180.P2	0.3	12.6

^aValues for COR and glucuronidase (GUS) activity represent the average of two experiments with three replicates each. Means within each column were analyzed using Duncan's multiple-range test. The protein content in the cell lysates was determined with the Bio-Rad protein assay kit as recommended by the manufacturer. 

^bTranscriptional activity in the COR biosynthetic gene cluster was evaluated by introducing pRGMU7 into each strain and measuring the glucuronidase activity in the resulting transconjugants. ND, not determined. 

The structural genes encoding COR production belong primarily to two distinct transcriptional units: the cmaABT transcript which is essential for production of coronamic acid and the cfl-CFA transcript which encodes coronafacic acid (5). Coronamic acid and coronafacic acid function as the two key intermediates in the biosynthetic pathway to coronatine (5). The effect of the rpoN mutation on transcriptional activity in the COR biosynthetic gene cluster was investigated by measuring β-glucuronidase activity from pRGMU7, a construct containing the cmaABT promoter fused to uidA (87). As shown in Table 2, when pRGMU7 was introduced into the two COR-producing strains, ES4326 and PG4180.N9, transcriptional activity from the cmaABT promoter was comparable (347 and 390 U GUS, respectively). However, β-glucuronidase activity in the rpoN mutant containing pRGMU7 was extremely low and was comparable to the low level of expression in PG4180.P2(pRGMU7). It is important to note that PG4180.P2 is defective in corR, a gene which encodes a positive transcriptional activator of the cmaABT transcript (64). The present data suggest that a functional rpoN is required for expression of the cmaABT transcript in P. syringae pv. maculicola.

In the accompanying study (23), we demonstrate that rpoN in ES4326 is required for the expression of hrpL, which encodes an alternative sigma factor and is required for expression of the ES4326 hrp genes and avrRpt2 (12, 17, 29, 97). We also show that constitutive expression of hrpL on plasmid pHRPLC restores the ability of ES4326 rpoN::Km^r to elicit disease symptoms in A. thaliana and an HR in tobacco. However, pHRPLC did not restore COR production to ES4326 rpoN::Km^r (Table 2), indicating that COR biosynthesis is not dependent on hrpL in ES4326 but on a separate regulatory pathway that also requires rpoN.

Strain ES4326 rpoN::Km^r fails to activate high-level expression of Arabidopsis defense-related genes.

The infiltration of Arabidopsis leaves with ES4326 normally leads to the accumulation of mRNAs corresponding to a variety of Arabidopsis defense-related genes, including PR2 (BGL2), GST1, PR5 and PR1, which encode β-1,3-glucanase, glutathione S-transferase, a thaumatin-like protein, and a protein with unknown activity, respectively (42, 51, 52, 91). Each of these genes shows a different induction pattern. In general, GST1 and PR5 are induced within several hours after infection, whereas PR1 and BGL2 are induced later than 24 h postinfection. In contrast to ES4326, very little accumulation of the BGL2, GST1, PR1, and PR5 transcripts was seen following infiltration with ES4326 rpoN::Km^r (Fig. 4), which gave results similar to those for the MgSO₄ control. This demonstrated that a factor under rpoN control is necessary for defense gene induction.

FIG. 4.

Induction of Arabidopsis defense-related mRNAs following infiltration of ES4326 rpoN::Km^r. Leaves were infiltrated with bacterial suspensions normalized to 0.1 OD₆₀₀ as described in Materials and Methods. Filters containing 5 μg of total RNA per lane were probed with radiolabeled DNA fragments encoding the Arabidopsis GST1, PR5, BGL2, and PR1 genes. The filters were also probed with a radiolabeled probe corresponding to the Arabidopsis ubiquitin gene UBQ. Phosphorimager data are expressed as a ratio of defense gene to UBQ gene induction. Symbols: □, 10 mM MgSO₄; ◊, ES4326; ○, ES4326 rpoN::Km^r; ▵, ES4326 rpoN::Km^r (pHRPLC).

Even more rapid induction of defense-related genes is seen accompanying the HR in incompatible interactions (11, 32, 66). Two such genes, Pal1 and Gst1, which encode phenylalanine ammonia lysase and glutathione S-transferase, respectively, show early, high-level accumulation of mRNAs in response to the avirulence gene avrRpt2 (11, 16). ES4326 rpoN::Km^r carrying the avrRpt2 gene induces little accumulation of Pal1 or Gst1 mRNA compared to ES4326 (avrRpt2) (Fig. 5). Again, the rpoN mutant gave results similar to those for the MgSO₄ control.

FIG. 5.

Induction of Arabidopsis defense-related mRNAs following infiltration of ES4326 rpoN::Km^r carrying the avirulence gene avrRpt2. Leaves were infiltrated with bacterial suspensions normalized to 0.1 OD₆₀₀ as described in Materials and Methods. Filters containing 5 μg of total RNA per lane were probed with radiolabeled DNA fragments encoding the Arabidopsis Pal1 gene or the Arabidopsis GST1 gene. Lanes: 1, 10 mM MgSO₄; 2, ES4326 (avrRpt2); 3, ES4326 rpoN::Km^r (avrRpt2).

In the accompanying study we report that constitutive expression of the sigma factor HrpL restores disease and HR phenotypes to ES4326 rpoN::Km^r (23). Curiously, however, constitutive expression of hrpL did not restore in planta growth to ES4326 rpoN::Km^r. As seen in Fig. 4, constitutive expression of hrpL in ES4326 rpoN::Km^r restored partial activation of the PR1, PR2 (BGL2), and PR5 genes in infiltrated plants, suggesting that defense gene activation and the elicitation of disease symptoms are a consequence of HrpL-dependent bacterial factors rather than growth of the pathogen per se. One difficulty encountered in this experiment was the instability of pHRPLC in wild-type ES4326 in planta (23). Thus, it was not possible to examine PR gene expression in response to ES4326 (pHRPLC).

DISCUSSION

ES4326 rpoN gene.

The amino acid sequence of the ES4326 ς⁵⁴ protein is most closely related to the P. putida ς⁵⁴ protein, and the phenotype of the ES4326 rpoN mutant resembles the phenotype of P. putida rpoN mutants (38). Both P. syringae and P. putida rpoN mutants grow more slowly than their wild-type counterparts in all media tested and were unable to utilize several uncharged amino acids as substrates. Neither is a glutamine auxotroph, and both can utilize NH₄ as a sole nitrogen source. They can also utilize lysine as a nitrogen source though not as a carbon source. In P. putida, the inability of the rpoN mutant to grow on lysine is probably due to the fact that lysine decarboxylase is under ς⁵⁴ control (38). The major phenotypic difference that we observed between the P. putida and the P. syringae rpoN mutants was that the P. syringae rpoN mutant could not utilize proline and histidine as nitrogen sources, whereas the P. putida rpoN mutant could.

Lack of growth of strain ES4326 rpoN on low concentrations of ammonia and other nitrogen sources.

The inability of ES4326 rpoN::Km^r to grow on low concentrations of ammonia (<1 mM) and various amino acids is presumably caused by the lack of glnA expression (which encodes glutamine synthetase), which is under ς⁵⁴ control in several other bacterial species (54, 86). On the other hand, the ES4326 rpoN mutant is not a glutamine auxotroph, since it grew well when high concentrations of ammonia (>20 mM) or several amino acids other than glutamine were supplied as the sole nitrogen source. One explanation for these results is that P. syringae, like Rhizobium and Bradyrhizobium species, has more than one gene encoding glutamine synthetase, one that is expressed at high levels under ς⁵⁴ control and a second copy that is expressed at low levels and is ς⁵⁴ independent (41). Alternatively, P. syringae could have a single gene encoding glutamine synthetase, which requires ς⁵⁴ for the high-level expression needed for ammonia assimilation but which has sufficient basal expression to prevent auxotrophy. rpoN mutants of most soil bacteria, including Azotobacter vinelandii, P. putida, and Agrobacterium tumefaciens, are not glutamine auxotrophs (39, 40, 53, 74, 75).

Growth rate of strain ES4326 rpoN.

The ES4326 rpoN mutant displayed slower growth rates than the wild type in each medium examined, including M9 supplemented with 0.2 mM glutamine. Thus, it appears unlikely that the slower growth of the rpoN mutant can be explained solely on the basis of decreased levels of glnA expression. While it is possible that the growth deficit is due to a secondary mutation, plasmid pPG102, which carries the wild-type rpoN gene, fully complemented every rpoN-related phenotype tested, including growth and pathogenesis in Arabidopsis leaves.

Nonpathogenic phenotype of strain ES4326 rpoN.

Given the pleiotropic phenotype of rpoN mutants, it is not possible to state precisely why the ES4326 rpoN mutant failed to elicit disease symptoms and to grow in Arabidopsis leaves or to elicit an HR. In the accompanying study we show that the absence of a functional ς⁵⁴ in ES4326 blocks the transcription of hrp genes downstream of hrpRS (23), which would account for the nonpathogenic and HR-deficient phenotypes. However, we also report that the constitutive expression of hrpL in ES4326 rpoN::Km^r restored the elicitation of disease symptoms but failed to restore growth of ES4326 rpoN::Km^r in planta, implying that the absence of hrp functions is not the sole reason for the nonpathogenic phenotype of ES4326 rpoN::Km^r.

Our experiments using a pseudorevertant of the ES4326 rpoN mutant that was able to utilize aspartate eliminated the possibility that the rpoN mutant is nonpathogenic solely due to its inability to utilize this amino acid. However, it is possible that ES4326 rpoN::Km^r has another metabolic defect that contributes to the nonpathogenic phenotype. The rpoN mutant is also unable to utilize proline, histidine, and methionine, and if ammonia serves as the main nitrogen source during infection then leaf concentrations of less than 1 to 2 mM would likely stop the growth of ES4326 rpoN::Km^r. Finally, as discussed in the next section, rpoN is involved in the production of at least one known toxin, COR, which could help explain the reduced virulence.

rpoN-mediated regulation of COR synthesis.

The data in Table 2 demonstrate that ES4326 rpoN::Km^r does not produce COR, which contributes to lesion expansion, chlorosis, and bacterial multiplication in Arabidopsis (56). Although a COR⁻ mutant is not available for P. syringae pv. maculicola, pHRPLC, which expresses hrpL constitutively, restored some disease symptoms but not COR production to ES4326 rpoN::Km^r. However, because pHRPLC failed to restore in planta growth to the mutant (23), it remains possible that some of the growth defect in ES4326 rpoN::Km^r could be caused by loss of COR production.

The data in Table 2 show that rpoN is also required for the expression of the cmaABT transcript, which encodes proteins that produce coronamic acid, an intermediate in the COR pathway (5). This was surprising since a conserved −24(GG)/−12(GC) motif is lacking upstream of the cmaA transcriptional start site (87). Thus, ς⁵⁴ control of cmaABT expression is probably mediated indirectly through another regulatory gene whose expression is directly controlled by ς⁵⁴. Possible candidates for ς⁵⁴ control inside the COR gene cluster include corP and corR, which encode response regulators with uncharacterized upstream sequences. Alternatively, ς⁵⁴ might control the expression of regulatory genes unlinked to the COR biosynthetic gene cluster.

Reduced induction of the host defense response by ES4326 rpoN.

ES4326 rpoN::Km^r failed to induce defense gene induction in Arabidopsis during both compatible and incompatible interactions. These results contrast with those reported previously for hrp mutants. A nonpathogenic hrp deletion mutant of a compatible X. campestris pv. campestris strain elicited the expression of a variety of defense genes in the turnip to approximately 50% of their normal expression levels (60). Similarly, an incompatible X. campestris pv. armoraciae strain with an hrp deletion did not induce an HR in the turnip but still induced defense gene expression (60). In the P. syringae pv. phaseolicola-bean interaction, approximately the same level of defense gene induction occurred with incompatible wild-type and hrp deletion strains (32). One way to explain the discrepancy observed in defense gene induction by rpoN and hrp mutants is that important factors for defense gene induction may lie outside of the hrp pathway but under rpoN control. While there are reasons to believe that phytotoxins and avr genes may contribute to defense gene induction, they seem unlikely explanations for this phenomenon (5, 34, 47, 71). Although production of the phytotoxin COR is rpoN dependent, in ES4326, COR⁻ mutants of DC3000 elicited more defense gene induction than the wild-type strain (56). Similarly, avr gene products are thought to require a functioning hrp cluster for activity and are therefore unlikely candidates for hrp-independent defense induction factors (34, 47, 71, 84). Finally, the fact that X. campestris pv. vesicatoria rpoN mutants are fully pathogenic (25) indicates that, at least in the case of this species, rpoN does not regulate any essential pathogenicity factors.

Our experiments also indicate an important role for hrp-dependent factors in defense gene induction. When hrpL was constitutively expressed in ES4326 rpoN::Km^r, both hrp gene expression (23) and defense gene induction (Fig. 4) were restored. This result indicates that at least some inducing factors require genes downstream of hrpL for expression, function, or both. As mentioned above, avr genes are a likely source of hrp-dependent defense-inducing factors. We also report that the partial restoration of defense gene induction by hrpL is accompanied by restoration of disease symptoms and host cell death (23). This restored host cell death may also play a role in the activation of host defense responses. P. syringae products secreted by the Hrp system could result in necrotic or programmed cell death which in turn activates defense gene induction in neighboring cells.