Roles of Agrobacterium tumefaciens RirA in Iron Regulation, Oxidative Stress Response, and Virulence

Abstract

The analysis of genetics and physiological functions of Agrobacterium tumefaciens RirA (rhizobial iron regulator) has shown that it is a transcription regulator and a repressor of iron uptake systems. The rirA mutant strain (NTLrirA) overproduced siderophores and exhibited a highly constitutive expression of genes involved in iron uptake (fhuA, irp6A, and fbpA) compared to that of the wild-type strain (NTL4). The deregulation in the iron control of iron uptake in NTLrirA led to iron overload in the cell, which was supported by the observation that the NTLrirA mutant was more sensitive than wild-type NTL4 to an iron-activated antibiotic, streptonigrin. The NTLrirA mutant was more sensitive than the parental strain to oxidants, including hydrogen peroxide, organic hydroperoxide, and a superoxide generator, menadione. However, the addition of an iron chelator, 2,2′-dipyridyl, reversed the mutant hypersensitivity to H₂O₂ and organic hydroperoxide, indicating the role of iron in peroxide toxicity. Meanwhile, the reduced level of superoxide dismutase (SodBIII) was partly responsible for the menadione-sensitive phenotype of the NTLrirA mutant. The NTLrirA mutant showed a defect in tumorigenesis on tobacco leaves, which likely resulted from the increased sensitivity of NTLrirA to oxidants and the decreased ability of NTLrirA to induce virulence genes (virB and virE). These data demonstrated that RirA is important for A. tumefaciens during plant-pathogen interactions.

The regulation of intracellular iron concentrations is a critical task for bacteria. Acquiring enough iron to grow is as crucial as preventing iron overload and its consequential toxicity. The Fur (ferric uptake regulator) proteins from many bacteria have been shown to play a major role in controlling genes involved in iron transport, storage, consumption, and the overall maintenance of intracellular iron homeostasis (2). Under high-iron conditions, Fur binds to its corepressor ferrous ion (Fe²⁺). The Fe²⁺-Fur complex binds to the promoter region of iron uptake genes, thus shutting down iron uptake and preventing iron overload toxicity. When iron is scarce, Fur exists in the apo form and no longer can repress iron uptake genes, and iron uptake resumes (2). However, Fur-mediated iron regulation does not occur in all bacteria.

Rhizobia are symbiotic soil bacteria that form N₂-fixing nodules on the roots of leguminous plants. These bacteria have a high demand for iron during symbiosis, since nitrogenase and other iron-containing proteins are required for N₂ fixation. Rhizobia have atypical regulation of iron homeostasis. The Fur-like protein has been shown to play only a minor role in the regulation of iron uptake. Instead, Fur physiologically functions in response to manganese by repressing the transcription of the sitABCD operon, which encodes a Mn²⁺ uptake system, under manganese-replete conditions (7, 9, 10, 18, 19). The role of Fur-like proteins seems to be restricted to the regulation of the manganese uptake gene, and thus Fur was renamed Mur (7, 9). The RirA (rhizobial iron regulator) protein evolved to carry out typical Fur functions in the regulation of iron-responsive genes for maintaining iron homeostasis in rhizobia (8, 29). The RirA protein belongs to the Rrf2 family of transcription regulators. The RirA homologues are found exclusively in members of alphaproteobacteria, including the rhizobia of Rhizobium, Sinorhizobium, and Mesorhizobium, the human pathogen Bartonella, the animal pathogen Brucella, and the phytopathogen Agrobacterium.

RirA is an Fe-S protein. RirA acts as a repressor of many iron-responsive genes under iron-replete conditions. The RirA regulon includes genes for the synthesis (vbs) and uptake (fhu) of the siderophore vicibactin, genes involved in heme uptake (hmu and tonB), genes encoding hemin-binding proteins (hbp), genes that probably participate in the transport of Fe³⁺ (sfu), a putative ferri-siderophore ABC transporter (rrp1), a gene that specifies an extracytoplasmic-function RNA polymerase δ factor (rpoI), genes for the synthesis of Fe-S clusters (suf), an iron response regulator (irrA), and rirA itself (4, 28). The iron-responsive operator (IRO) motif (TGA-N9-TCA) has been identified as a DNA-binding site for RirA. Computational analysis has been used to search for IRO motifs in the genomes of alphaproteobacteria, leading to the identification of target genes in the RirA regulon (20).

Agrobacterium tumefaciens, a gram-negative member of the alphaproteobacteria, is a soil-borne plant pathogen that causes crown gall tumor disease in dicotyledonous plants. The pathogenesis involves the attachment of A. tumefaciens to the wounded plant cells and the subsequent transfer of a segment of its tumor-inducing (Ti) plasmid into plant cells (36). The virulence (vir) genes located on the Ti plasmid are important primarily for tumorigenesis. The vir genes are induced in response to acidic pHs (approximately 4.8 to 5.5) and phenolic compounds, such as acetosyringone (AS), that are released by wounded plant cells. Plants generate high levels of reactive oxygen species, such as H₂O₂ and superoxide radicals, which are used as an important initial defense mechanism to inhibit bacterial invasion and proliferation. Pathogenic bacteria need to carry out antioxidant responses in order to survive in host plants and ultimately cause disease. Catalase (Kat) and superoxide dismutase (Sod) are key antioxidant enzymes for degrading H₂O₂ and superoxides, respectively. These enzymes have been shown to be virulence factors that are involved in the tumorigenesis of A. tumefaciens (21, 35). Iron is an essential metal for bacterial growth and also is required as a cofactor for Kat and Sod. However, excessive amounts of iron can be toxic due to its ability to induce the production of highly deleterious hydroxyl radicals via the Fenton reaction (13). Controlling intracellular iron levels and modulating antioxidant responses therefore are crucial tasks for pathogenic bacteria during infection. The iron-dependent transcriptional regulator has been reported to play a key role in both adjusting intracellular iron levels and controlling oxidative stress responses during host plant-pathogen interactions (11, 26). The role of RirA in iron regulation and oxidative stress response has not been studied in A. tumefaciens. Here, the physiological functions of A. tumefaciens RirA in iron regulation and oxidative stress response are shown, and its regulatory roles in bacterial stress survival and virulence are demonstrated.

MATERIALS AND METHODS

Bacterial growth conditions.

Bacterial strains and plasmids are listed in Table 1. A. tumefaciens strains were grown aerobically in Luria-Bertani (LB) medium at 28°C with shaking at 150 rpm, supplemented with 100 μg ml⁻¹ carbenicillin (Cb), 25 μg ml⁻¹ chloramphenicol (Cm), 90 μg ml⁻¹ gentamicin (Gm), 30 μg ml⁻¹ kanamycin (Km), or 10 μg ml⁻¹ tetracycline (Tc) as required. Escherichia coli strains were used for routine DNA cloning experiments and grown aerobically in LB medium at 37°C, supplemented with 100 μg ml⁻¹ Ap, 30 μg ml⁻¹ Gm, 30 μg ml⁻¹ kanamycin (Km), or 15 μg ml⁻¹ Tc as required.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)a	Source or reference
A. tumefaciens
NTL4	Wild-type strain, a Ti plasmid-cured derivative of strain C58	S. K. Farrand
NTLrirA	rirA mutant, derivative of NTL4 in which rirA was disrupted by pKNOCKrirA, Kmr	This study
E. coli
DH5α	supE Δlac(φ80dlacZΔM15) hsdR recA endA gyrA thi relA	12
BW20767	leu-63::IS10 recA1 creC510 hsdR17 endA1 zbf-5 uidA(ΔMlu1)::pir+thi RP4-2-tet::Mu-1kan::Tn7	16
Plasmids
pGEM-T-easy	Cloning vector, Apr	Promega
pKNOCK-Km	Suicide vector, Kmr	1
pKNOCKrirA	pKNOCK-Km containing a 188-bp fragment of the rirA coding region, Kmr	This study
pBBR1MCS-4	Expression vector, Apr	15
pRirA	Full-length rirA coding region cloned into pBBR1MCS-4, Apr	This study
pSodBIII	Full-length sodBIII coding region cloned into pBBR1MCS-4, Apr	21
pCMA1	pTiC58traM::nptII, Cbr, Kmr	S. K. Farrand
pSM243cd	virB::lacZ fusion, Cbr, Kmr	S. K. Farrand
pSM358cd	virE::lacZ fusion, Cbr, Kmr	S. K. Farrand

^aKm^r, kanamycin resistance; Ap^r, ampicillin resistance; Cb^r, carbenicillin resistance.

Molecular techniques.

General molecular techniques were performed using standard protocols (22). Plasmid DNA was isolated using the QIAprep kit (Qiagen). DNA restriction and modifying enzymes were purchased from Promega, Fermentas, or New England Biolabs and used according to the suppliers’ recommendations. PCR products and restriction fragments were purified using PCR clean-up and gel extraction kits (Qiagen). Sequencing was carried out on an ABI 310 automated DNA sequencer (Applied Biosystems) using a BigDye terminator cycle sequencing kit (PE Biosystems). Plasmids (50 to 100 ng) were transferred into A. tumefaciens strains by electroporation (6). The primers used are listed in Table 2.

TABLE 2.

Primers used in this study

Gene-primer name and purpose	Sequence (5′→3′)	PCR product size (bp)
Gene inactivation
rirA-BT1201	GGCTTACGGCGTATCTGAGC	188
rirA-BT1202	CGTCCTCAAAGCACTCAGCC
Complementation
rirA-BT1701	GCGGTCAACAAGGCTGTTCG	593
rirA-BT1205	CGATCAGGCTGCCGGAAGTG
RT-PCR
16S rRNA-BT1421	GAATCTACCCATCTCTGCGG	280
16S rRNA-BT1422	AAGGCCTTCATCACTCACGC
fhuA-BT1346	GGTGACGAAGGGTATTGGCG	202
fhuA-BT1096	CGCGACATAACCTTTCACCG
fbpA-BT2728	CTGCGGATGTGCTGCTGACG	190
fbpA-BT2729	GTCACCCGCTCCTTCGAAGC
irp6A-BT2734	CACCGTCAAGGATGTGACCG	232
irp6A-BT2735	CTTCATGCCGCCGAAAGTCG
sodBI-BT534	CAGCACTACAACCACGTC	198
sodBI-BT535	TTCAAGCTTGCCGTTCTT
sodBII-BT641	GCGCAGTCCGGCATATTCA	200
sodBII-BT640	ACCAGCCACGCCCAGCCCG

Construction of A. tumefaciens rirA mutant.

The disruption of the rirA gene (Atu0201) (34) was performed by the method described previously (14). The internal DNA fragment of the rirA coding region was amplified by PCR with primers BT1201 and BT1202, using Pfu DNA polymerase and genomic DNA isolated from wild-type NTL4 as the template. The 188-bp PCR product was cloned into the unique SmaI site of the pKNOCK-Km suicide plasmid (2 kb in size) (1), which is unable to replicate in A. tumefaciens, generating the recombinant plasmid pKNOCKrirA. The cloned DNA region was confirmed by automated DNA sequencing. pKNOCKrirA was transferred into wild-type NTL4 by conjugation (6). The single homologous recombinants were selected on LB agar plates containing 25 μg ml⁻¹ chloramphenicol and 30 μg ml⁻¹ kanamycin. The correct integration of pKNOCKrirA into the rirA locus was confirmed by Southern blot analysis.

Cloning of full-length rirA.

The full length of the wild-type rirA gene was amplified from wild-type NTL4 genomic DNA by PCR using primers BT1701 and BT1205 and Pfu DNA polymerase. The PCR products were cloned into the expression vector pBBR1MCS-4 (15), which had been digested with SmaI. The cloned DNA region was confirmed by automated DNA sequencing. The resulting plasmid, named pRirA, was used for rirA mutant complementation experiments.

Siderophore detection.

Siderophore production was analyzed using a chrome azural S (CAS) agar plate (24). Solid CAS medium was made by adding 10 ml of CAS stock (24) to 100 ml of YEM (yeast extract-mannitol) medium (31) containing 1.5% agar. Overnight cultures grown in LB medium (5 μl at an optical density at 600 nm [OD₆₀₀] of 0.1) were spotted onto a YEM-CAS plate containing 200 μM 2,2′-dipyridyl (Dy) and incubated at 28°C for 2 days. Siderophore production is indicated by the presence of a halo zone around the bacteria. This occurs because siderophores produced by bacteria remove iron from the original green CAS-Fe³⁺ complex contained in the plate, resulting in a change in the color of the dye.

RNA extraction and reverse transcriptase PCR (RT-PCR) analysis.

Bacteria grown overnight in LB medium were subcultured into 20 ml of fresh LB medium to give an OD₆₀₀ of 0.1. Exponential-phase cells (OD₆₀₀ of 0.5 after incubation for 4 h) were treated with 50 μM FeCl₃, 250 μM Dy, 50 μM MnCl₂, or 200 μM menadione (MD) for 15 min. Cells were harvested by centrifugation at 6,000 rpm for 5 min. Total RNA was extracted from untreated and treated cells using the modified hot phenol method (25). Briefly, the cell pellet was suspended in 300 μl of 0.3 M sucrose and 10 mM sodium acetate (NaOAc). Lysis buffer (300 μl) containing 2% sodium dodecyl sulfate and 10 mM NaOAc was added, and the mixture was incubated at 65°C for 5 min with gentle mixing. The hot phenol (300 μl), maintained at 65°C, was added. After incubation at 65°C for 5 min with occasional mixing, the phases were separated by centrifugation at 12,000 rpm for 10 min. The aqueous phase was reextracted twice with an equal volume of hot phenol as described above, followed by extraction with an equal volume of chloroform. RNA was precipitated by adding 10% of the volume of 3 M NaOAc and two volumes of absolute ethanol. After overnight incubation at −20°C, the RNA was pelleted by centrifugation at 12,000 rpm for 15 min and washed once with 1 ml of 70% ethanol. The RNA pellet was dried and suspended in diethylpyrocarbonate (DEPC)-treated sterile distilled water.

The RNA samples were treated with DNase I using the DNA-free kit (Ambion) according to the manufacturer's protocols. Reverse transcription (converting mRNA to cDNA before PCR) was accomplished using SuperScript II RT (Invitrogen) with random hexamer primers (BioDesign, Thailand). Reverse-transcribed RNA samples (0.1 μg) from each condition were used in the PCR. Control reactions, where RT was omitted, were run in parallel to ensure that there was no DNA contamination. Positive controls were performed using genomic DNA isolated from wild-type NTL4. Gene-specific primers for fhuA, irp6A, fbpA, sodBI, sodBII, and 16S rRNA (Table 2) were used for separate PCRs using the Taq PCR master mix kit (Qiagen). PCRs were carried out with an initial denaturation step at 95°C for 5 min, followed by 25 cycles of denaturation at 95°C for 30 s, annealing at 55°C (sodBI and sodBII) or 58°C (fhuA, irp6A and fbpA) for 30 s, and extension at 72°C for 30 s, with a final extension step at 72°C for 5 min. RT-PCR products were visualized through gel electrophoresis on a 2% agarose gel and ethidium bromide staining.

Sensitivity to SNG.

Cells grown in LB medium for 24 h were washed once with fresh LB medium, and their concentration was adjusted to give an OD₆₀₀ of 0.1. Cells were treated with streptonigrin (SNG) at a concentration of 40 μg ml⁻¹ in the absence or presence of 100 μM Dy. SNG was prepared as a stock solution at 10 mg ml⁻¹ in dimethyl sulfoxide. Control cells (untreated) received equivalent amounts of dimethyl sulfoxide. The cells were incubated at 28°C with shaking for 3 h and were diluted (10-fold serial dilutions). An aliquot (10 μl) of each dilution was spotted onto an LB agar plate and incubated at 28°C for 2 days. Each strain was tested in duplicate, and the experiment was repeated at least twice to ensure the reproducibility of the results.

Sensitivity to oxidants.

Cells grown in LB medium for 24 h were washed once and then were 10-fold serially diluted in fresh LB medium. An aliquot (10 μl) of each dilution was spotted onto LB agar plates containing 450 μM H₂O₂, 600 μM tert-butyl hydroperoxide (tBOOH), or 450 μM MD in the absence or presence of 100 μM Dy. Plates then were incubated at 28°C for 2 days. Each strain was tested in duplicate, and the experiment was repeated at least twice to ensure the reproducibility of the results.

Sod activity gel assay.

Crude bacterial lysates were prepared using bacterial suspensions in 50 mM sodium phosphate buffer, pH 7.0, containing 1 mM phenylmethylsulfonyl fluoride, a protease inhibitor. Cell suspensions were lysed by brief sonication followed by centrifugation at 12,000 × g for 10 min. Cleared lysates were used for total protein determination and Sod activity gel assays. Protein concentrations were determined using the Bradford Bio-Rad protein assay. To visualize Sod activity on the gel, 25 μg protein from lysate samples was separated on a 10% nondenaturing gel, followed by staining with nitroblue tetrazolium-riboflavin photochemical stain (5).

Tumor formation assay.

A. tumefaciens strains containing pCMA1 plasmid were used to infect tobacco (Nicotiana tabacum) leaves according to the method described previously (14, 17).

vir gene induction assay.

Overnight cultures of wild-type NTL4 and mutant NTLrirA strains containing pSM243cd (virB::lacZ) or pSM358cd (virE::lacZ) in LB medium were washed and adjusted to an OD₆₀₀ of 0.5 in 10 ml of induction broth, pH 5.5 (6), in the absence or presence of 50 μM acetosyringone (AS). In some experiments, cells were grown in media with 50 μM AS in the presence of either 50 μM FeCl₃ or an iron chelator, 50 μM Dy. Cells were further incubated at 28°C with shaking for 24 h. Cells were harvested, and β-galactosidase activity was measured as previously described (14). β-Galactosidase activity is presented in units per milligram of protein.

RESULTS AND DISCUSSION

A. tumefaciens RirA is the repressor of iron uptake systems, and loss of RirA leads to intracellular iron overload.

The A. tumefaciens genome contains an rirA homologue (Atu0201), rirA_At (34). The rirA_At gene is 471 bp long and is flanked upstream by an fbpA gene (Atu0202) encoding a putative iron ABC transporter periplasmic binding protein. The rirA and fbpA genes are transcribed in the opposite direction. Downstream of the rirA_At gene is a gene encoding a conserved hypothetical protein (Atu0200) similar to ammonia monooxygenase and a gene cluster (Atu0199, Atu0198, Atu0197, and Atu0196) encoding a putative proline/glycine/betaine ABC transport system. The protein encoded by rirA_At has a predicted molecular mass of 17.1 kDa and shows high sequence similarity to RirA proteins from Rhizobium leguminosarum (88% identity) and Sinorhizobium meliloti (85% identity). While R. leguminosarum rirA is preceded by an iolA gene encoding a semialdehyde decarboxylase (29), there is an fbpA gene located immediately upstream of S. meliloti rirA that is similar to that observed in A. tumefaciens. The S. meliloti fbpA gene was found to be regulated by rirA (8). Unlike in A. tumefaciens, a cluster of genes (dppA-dppF) involved in the uptake of a heme precursor is found immediately downstream of rirA in those two rhizobia.

Several studies have clearly shown that RirA evolved to carry out typical Fur functions in controlling iron uptake systems (29, 30). The inactivation of rirA led to the deregulation of iron uptake, which typically was indicated by the overexpression of siderophore synthesis and transport genes. We have shown that the inactivation of A. tumefaciens fur has no effect on the siderophore synthesis and iron transport genes, suggesting that fur is not the major regulator of iron uptake genes (14). An A. tumefaciens rirA mutant (NTLrirA) was constructed as described in Materials and Methods. First, the effect of rirA inactivation on siderophore production was investigated. As shown in Fig. 1A, the NTLrirA mutant strain harboring a plasmid vector (NTLrirA/pBBR) produced more siderophores than the wild-type strain (NTL4/pBBR), as indicated by a larger halo zone surrounding NTLrirA/pBBR than that surrounding NTL4/pBBR. Furthermore, the complementation of the mutant strain by the expression of the functional rirA gene on the plasmid pRirA (NTLrirA/pRirA) could restore the siderophore production to levels similar to that of wild-type NTL4/pBBR (Fig. 1A). This confirmed that the overproduction of siderophores in NTLrirA was due to the loss of the rirA gene.

FIG. 1.

(A) Analysis of siderophore production was performed using siderophore indicator CAS agar plates. NTL4/pBBR and NTLrirA/pBBR strains are the wild type and the rirA mutant, respectively, expressing the plasmid vector pBBR1MCS-4. NTLrirA/pRirA is the rirA mutant expressing functional RirA from the plasmid pRirA. Cells were spotted onto YEM-CAS agar plates containing 200 μM Dy and incubated at 28°C for 2 days. The halo zone surrounding the bacteria indicates the production of siderophores. (B) RT-PCR was used to monitor the mRNA levels for fhuA, irp6A, fbpA, and 16S rRNA. RNA samples were isolated from wild-type NTL4 and mutant NTLrirA cultures treated with 250 μM Dy or 50 μM FeCl₃ (Fe) for 15 min.

The role of A. tumefaciens rirA in controlling iron uptake was further investigated by monitoring the expression of fhuA (Atu4022), irp6A (Atu3391), and fbpA (Atu0202) genes encoding iron ABC transporter periplasmic binding proteins by using RT-PCR analysis. RNA samples were isolated from NTL4 and NTLrirA cells grown under iron-replete (LB plus 50 μM FeCl₃) and iron-depleted (LB plus 250 μM Dy) conditions. Under iron-replete conditions, the expression of fhuA, irp6A, and fbpA was repressed compared to that of wild-type NTL4 under iron-depleted conditions (Fig. 1B). The iron-mediated repression of these genes was lost in the NTLrirA mutant. The amounts of RT-PCR products from iron-replete NTLrirA samples were as large as those from iron-depleted NTLrirA samples. These data demonstrated that A. tumefaciens rirA was the repressor of iron uptake systems and that the inactivation of rirA led to a constitutive high-level expression of genes involved in the process. In addition, the data provided experimental verification of the predicted members of the RirA regulon (20). The iron-dependent regulation of fbpA may be more complex than those of fhuA and irp6A. In the wild-type NTL4 under iron-replete conditions, the expression of fbpA was not fully repressed compared to those of fhuA and irp6A (Fig. 1B). There might be another mechanism to activate fbpA expression under iron-replete conditions.

The derepression of iron uptake in the NTLrirA mutant led us to ask whether the NTLrirA mutant experienced intracellular iron overload. SNG is an iron-activated antibiotic. Increased intracellular iron concentrations have been shown to correlate with an increased sensitivity to SNG (14, 26, 32). The relative intracellular iron levels in NTL4 and the NTLrirA mutant were assessed by SNG sensitivity assays. The results in Fig. 2 show that the NTLrirA/pBBR mutant was 10-fold more sensitive to 40 μg ml⁻¹ SNG treatment than wild-type NTL4/pBBR. This suggests that the NTLrirA mutant had intracellular free iron levels that were higher than that of wild-type NTL4. The expression of the functional rirA gene in the NTLrirA mutant (NTLrirA/pRirA) could restore resistance to SNG to wild-type (NTL4/pBBR) levels. Moreover, the addition of Dy, a cell membrane-permeable iron chelator that chelates intracellular iron, could protect the NTLrirA mutant from SNG killing. This further supported the conclusion that the loss of rirA led to iron overload conditions in the NTLrirA mutant.

FIG. 2.

Sensitivity to SNG. Cells were left untreated or were treated with 40 μg ml⁻¹ SNG in the absence (SNG) or presence of 100 μM Dy (SNG+Dy) at 28°C for 3 h. Cells then were diluted and spotted onto LB agar plates (LA). Tenfold serial dilutions are indicated above each column. NTL4/pBBR and NTLrirA/pBBR strains are the wild type and the rirA mutant, respectively, expressing the plasmid vector pBBR1MCS-4. NTLrirA/pRirA is the rirA mutant expressing functional RirA from the plasmid pRirA.

Role of RirA in oxidant resistance.

An oxidative burst is the first-line plant defense response against microbial infection (33). Therefore, the ability of pathogens to respond to oxidants is important for survival and successful infection. We tested whether the inactivation of rirA affected the sensitivity of A. tumefaciens to oxidants including H₂O₂, tBOOH, and a superoxide generator, MD. The results in Fig. 3 showed that the NTLrirA mutant was more sensitive to oxidants than wild-type NTL4. NTLrirA/pBBR was 10-fold more sensitive than NTL4/pBBR to 450 μM H₂O₂, 600 μM tBOOH, and 450 μM MD. The hypersensitive phenotype of NTLrirA to oxidants could be complemented by expressing a functional rirA gene on the plasmid pRirA (NTLrirA/pRirA). Interestingly, the addition of the iron chelator Dy to the medium also could reverse the hypersensitive phenotype of NTLrirA to H₂O₂ and tBOOH but not to MD. The growth of NTLrirA/pBBR was similar to that of NTL4/pBBR on LB agar plates containing 450 μM H₂O₂ and 100 μM Dy or 600 μM tBOOH and 100 μM Dy (Fig. 3). The data implied that increased sensitivities to H₂O₂ and tBOOH toxicity in the NTLrirA mutant most likely were due to the iron overload condition. This condition likely resulted from the derepression of the expression of iron uptake genes (Fig. 1). This is consistent with exposure to peroxides, leading to the Fenton reaction and the generation of highly reactive oxygen species, which are responsible for the phenotype of the NTLrirA mutant with an increased sensitivity to peroxides. In contrast, the hypersensitivity of the NTLrirA mutant to MD could not be reversed by the addition of an iron chelator (Fig. 3), suggesting that mechanisms other than the iron overload condition mediated MD toxicity in the NTLrirA mutant. A possible mechanism is a defect in the NTLrirA mutant in enzymes involved in the superoxide detoxification.

FIG. 3.

Sensitivity to oxidants. Cells were diluted and spotted on LB agar plates (LA) containing 450 μM H₂O₂, 600 μM tBOOH, or 450 μM MD in the absence or presence of 100 μM Dy. Plates were incubated at 28°C for 48 h. Tenfold serial dilutions are marked above each column. Cells spotted on an LB agar plate were used as a control. NTL4/pBBR and NTLrirA/pBBR strains are the wild type and the rirA mutant, respectively, expressing the plasmid vector pBBR1MCS-4. NTL4/pRirA and NTLrirA/pRirA strains are the wild type and the rirA mutant, respectively, expressing functional RirA from the plasmid pRirA. NTLrirA/pSodBIII is the rirA mutant expressing SodBIII from the plasmid pSodBIII.

To test this hypothesis, an Sod activity gel assay was performed using cell lysates from wild-type NTL4 and the NTLrirA mutant (Fig. 4A). A. tumefaciens contains three Sods, namely SodBI, SodBII, and SodBIII (21). All Sod enzymes have iron cofactors. These sod genes exhibit differential expression patterns, and their gene products are found in different cellular locations. SodBI is the major cytoplasmic Sod enzyme, and its gene is constitutively expressed throughout all growth phases, whereas sodBII is scarcely expressed under normal conditions. sodBII has an MD-inducible expression pattern and is regulated by a superoxide sensor and transcription regulator, SoxR (21). SodBII is a cytoplasmic enzyme. sodBIII is highly expressed during the stationary phase of growth, and the enzyme is located in the periplasmic space. At present, the regulators of sodBI and sodBIII genes have not been identified. The results in Fig. 4A clearly showed that SodBIII could not be detected in the NTLrirA mutant. SodBI and SodBII migrated to the same position on an Sod activity gel, which contributed to the major visible band (21). The RT-PCR analysis was performed in order to measure the expression levels of sodBI and sodBII in the NTLrirA mutant. The results showed that sodBI RT-PCR products obtained from the NTLrirA mutant were not different from those from wild-type NTL4 (Fig. 4B). We also tested whether the MD-induced expression of sodBII was affected by rirA inactivation. The sodBII RT-PCR products were detected only in MD-treated samples; moreover, the amounts of sodBII RT-PCR products were similar in wild-type NTL4 and mutant NTLrirA. These data indicated that the levels of SodBIII and not SodBI or SodBII are affected by the inactivation of rirA. Next, we determined whether the reduction in SodBIII levels was responsible for the MD-sensitive phenotype of the NTLrirA mutant. The MD-sensitive phenotype of the NTLrirA mutant could not be fully complemented by the overexpression of sodBIII alone, as observed for NTLrirA/pSodBIII (Fig. 3). This observation could be explained by a previous observation that SodBIII has a minor role in protecting A. tumefaciens from superoxide stress (21). The inactivation of sodBIII alone has no effect on the MD resistance levels. However, the contribution of sodBIII to MD resistance is revealed only in the double sod mutant strains (21). It is likely that in addition to reductions in SodBIII levels, other not-yet-identified RirA-regulated genes also participated in the hypersensitivity of the NTLrirA mutant to MD.

FIG. 4.

(A) Sod activity gel. Equal amounts of protein (25 μg) from cell lysates of NTL4 and NTLrirA were loaded and separated on a 10% nondenaturing gel. Sod isozymes were visualized by activity staining as described in Materials and Methods. Arrowheads indicate the positions of SodBI, SodBII, and SodBIII activity. (B) RT-PCR analysis of sodBI and sodBII transcripts. RNA samples were isolated from wild-type NTL4 and mutant NTLrirA cultures left untreated (Un) or treated with 200 μM MD for 15 min.

RirA is required for induction of virulence genes and tumorigenesis in A. tumefaciens.

A. tumefaciens induces the formation of crown gall tumors by transferring T-DNA from the bacterium's Ti plasmid into host plant cells (36). The effect of rirA inactivation on the virulence of A. tumefaciens was evaluated by the analysis of tumor formation on tobacco leaf pieces infected with wild-type and rirA mutant strains containing Ti plasmid pCMA1. The mutant NTLrirA/pCMA1 showed significantly less virulence than the wild-type strain NTL4/pCMA1 (Fig. 5A). The tumors that formed on tobacco leaf pieces infected with NTLrirA/pCMA1 were much fewer and smaller than those caused by NTL4/pCMA1. Furthermore, the attenuated virulence of the NTLrirA/pCMA1 mutant could be complemented by pRirA, as shown by the fact that the tumor-inducing ability of NTLrirA/pCMA1/pRirA was completely restored to NTL4/pCMA1 levels (Fig. 5A). In contrast, NTLrirA/pCMA1/pBBR1MCS-4 could not complement the reduced-virulence phenotype of the mutant (data not shown). These data confirmed that the loss of rirA led to the defect in virulence.

FIG. 5.

Virulence assay. (A) Tumor formation was determined on tobacco leaf squares infected with A. tumefaciens wild-type NTL4 or mutant NTLrirA containing the pCMA1 plasmid (NTL4/pCMA1 and NTLrirA/pCMA1, respectively). The NTLrirA mutant was complemented by the expression of functional RirA from the plasmid pRirA (NTLrirA/pCMA1/pRirA). The controls are tobacco leaf squares without infection. Representative leaf pieces (n = 30) are shown. (B) virB gene induction. (C) virE gene induction. Wild-type (NTL4/pBBR), rirA mutant (NTLrirA/pBBR), and complementation (NTLrirA/pRirA) strains containing pSM243cd (virB::lacZ) or pSM358cd (virE::lacZ) were grown in induction broth (pH 5.5) in the absence (Un) or presence of 50 μM AS at 28°C for 24 h. β-Galactosidase activity was measured. Also shown are the induction of virB (D) and virE (E) in induction broth (pH 5.5) containing 50 μM AS and in the presence of 50 μM FeCl₃ (Fe) or 50 μM Dy.

There are reports that the attenuated virulence of A. tumefaciens mutants results from defects in vir gene induction (3, 27). Thus, we sought to determine whether a mutation in rirA affected vir gene induction. The β-galactosidase activity was measured in wild-type and mutant strains containing plasmid pSM243cd (virB::lacZ) or pSM358cd (virE::lacZ) grown in induction broth (pH 5.5) and in the absence or presence of 50 μM AS. There was no β-galactosidase activity from cells grown in the absence of AS, whereas β-galactosidase activity was largely increased when cells were grown in the presence of 50 μM AS. This was consistent with the fact that vir genes are induced in response to phenolic compounds (36). The inactivation of rirA resulted in a reduction in the magnitude of virB and virE gene induction by AS (Fig. 5B and C), as shown by the β-galactosidase levels from NTLrirA/pBBR containing pSM243 or pSM358, which were lower than those from NTL4/pBBR containing pSM243 or pSM358. The reduced induction of virB and virE by AS could be restored in the complemented NTLrirA/pRirA strains containing pSM243 or pSM358 (Fig. 5B and C, respectively). These data indicated that RirA played a role in the induction of vir gene expression. The inactivation of rirA caused a reduction in virB and virE induction but did not abolish the induction. This implied that the RirA protein could have an indirect role in modulating the level of vir gene induction. One possible explanation for the impairment of vir gene induction by AS in the NTLrirA mutant is related to the levels of intracellular iron. The NTLrirA mutant had intracellular free iron levels that were higher than that of wild-type NTL4 (Fig. 2). We observed that the induction levels of vir genes by AS from cells grown in iron-replete medium were lower than those from cells grown in iron-depleted medium. The induction of virB (Fig. 5D) and virE (Fig. 5E) by AS in both wild-type NTL4 and the NTLrirA mutant could be enhanced when cells were grown in the presence of the iron chelator Dy. The impairment of virB and virE induction in the NTLrirA mutant could be partially restored by the addition of an iron chelator, as shown by the induction levels from NTLrirA/pBBR cells grown under iron-depleted conditions (Dy) and from NTL4/pBBR cells grown under iron-replete conditions (Fe). However, under iron-depleted conditions (Dy), the β-galactosidase levels from NTLrirA/pBBR were lower than those from NTL4/pBBR. These results suggested that mechanisms other than the iron overload condition also contributed to the impairment of vir gene induction by AS in the NTLrirA mutant.

The host plant environment is low in iron (23). A. tumefaciens might sense a low-iron environment as a signal for the induction of virulence genes upon entry to the host plant. During plant-pathogen interactions, an oxidative burst is the first line of plant defense against microbial infection. The NTLrirA mutant showed increased sensitivity to oxidants compared to that of wild-type NTL4 (Fig. 3). Hence, both the increased sensitivity of NTLrirA to oxidants and the decreased ability of NTLrirA to induce virulence genes likely contributed to the attenuated virulence of the NTLrirA mutant on tobacco leaves.