Defining the Metabolic Functions and Roles in Virulence of the rpoN1 and rpoN2 Genes in Ralstonia solanacearum GMI1000

The alternative sigma factor RpoN is a unique regulator found among bacteria. It controls numerous processes that range from basic metabolism to more complex functions such as motility and nitrogen fixation. Our current understanding of RpoN function is largely derived from studies on prototypical bacteria such as Escherichia coli. Bacillus subtilis and Pseudomonas putida. Although the extent and necessity of RpoN-dependent functions differ radically between these model organisms, each bacterium depends on a single chromosomal rpoN gene to meet the cellular demands of RpoN regulation. The bacterium Ralstonia solanacearum is often recognized for being the causative agent of wilt disease in crops, including banana, peanut and potato. However, this plant pathogen is also one of the few bacterial species whose genome possesses dual rpoN genes. To determine if the rpoN genes in this bacterium are genetically redundant and interchangeable, we constructed and characterized ΔrpoN1, ΔrpoN2 and ΔrpoN1 ΔrpoN2 mutants of R. solanacearum GMI1000. It was found that growth on a small range of metabolites, including dicarboxylates, ethanol, nitrate, ornithine, proline and xanthine, were dependent on only the rpoN1 gene. Furthermore, the rpoN1 gene was required for wilt disease on tomato whereas rpoN2 had no observable role in virulence or metabolism in R. solanacearum GMI1000. Interestingly, plasmid-based expression of rpoN2 did not fully rescue the metabolic deficiencies of the ΔrpoN1 mutants; full recovery was specific to rpoN1. In comparison, only rpoN2 was able to genetically complement a ΔrpoN E. coli mutant. These results demonstrate that the RpoN1 and RpoN2 proteins are not functionally equivalent or interchangeable in R. solanacearum GMI1000.


Introduction
The alternative sigma factor σ 54 or RpoN regulates an assortment of diverse biological processes and is an essential protein for some bacteria [1]. RpoN proteins are structurally and mechanistically distinct from sigma factors belonging to the σ 70 family [2]. For example, RpoN proteins have a modular structure consisting of three domains (regions I, II & III). Region I is at the N-terminus and is involved in interactions with transcriptional regulator proteins known as enhancer-binding proteins or EBPs [3][4][5][6]. Region II serves as a linker between regions II and III; region II is absent in some RpoN proteins [7]. Region III (C-terminus) is involved in interacting with RNA polymerase [8,9] and contains the DNA-binding motifs responsible for promoter recognition [9][10][11]. RpoN recognizes a unique −24/−12 promoter that has a consensus sequence of TGGCACG-N 4 -TTGC [12]. This sequence is remarkably conserved with some nucleotides in the −24 and −12 sites displaying >90% conservation across a number of bacterial species [12].
Our knowledge of RpoN as a regulator of gene expression is largely derived from studies on bacteria harboring a single rpoN gene, and thus, a single RpoN protein. Under these circumstances, the RpoN protein has evolved to interact efficiently with its arsenal of partner EBPs to direct transcription from −24/−12 promoters. At the center of these interactions is region I (residues 1-50) of RpoN and the conserved amino acid motif 'GAFTGA' localized in a loop structure in the RpoN-interaction domain of EBPs [13]. These two domains make direct contacts with one another, and therefore, play a significant role in allowing RpoN to communicate with EBPs that differ in overall structure and function.
The situation can be more complex when a bacterium possesses multiple RpoN proteins. For example, the photosynthetic bacterium Rhodobacter sphaeroides has four genes encoding for RpoN proteins (RpoN1-4), which share 50-60% homology to one another [14]. The RpoN3 and RpoN4 proteins had no observable function in this bacterium [14]. In contrast, RpoN1 was specifically required for transcription of genes associated with nitrogen fixation while RpoN2 was necessary for motility [14]. The specializations of the RpoN1 and RpoN2 proteins were a result of differences in their promoter recognition (at position −11) and EBP interactions [15].
A number of soil-dwelling and plant-associated bacteria do have genes encoding for dual RpoN proteins. For example, the two RpoN proteins in Bradyrhizobium japonicum are nearly identical (>80% homology) and were previously found to be interchangeable [16]. The plant pathogen Ralstonia solanacearum is widely known for its role as the causative agent of bacterial wilt disease in economically important plant species such as banana, tomato, potato and peanut [17]. The genome of this plant pathogen consists of a 3.7 Mb chromosome and a 2.1 Mb megaplasmid [18]. Interestingly, each of these replicons harbors an rpoN gene. The rpoN1 gene is chromosomal while rpoN2 is carried on the megaplasmid. Unlike B. japonicum, the RpoN1 and RpoN2 proteins in R. solanacearum GMI1000 are not identical. They share *60% homology towards one another with the majority of differences being found in regions I and II. Indeed, a recent study showed that the rpoN genes in R. solanacearum GMI1000 are not genetically redundant [19]. Specifically, phenotypic traits such as nitrate utilization, natural competence, twitching motility and virulence were dependent only on the rpoN1 gene. Inactivation of the rpoN2 gene had no affect on these phenotypes.
In an effort to expand on these previous findings, and therefore, more fully understand metabolic functions dependent on RpoN in R. solanacearum GMI1000, we constructed a series of rpoN-deletion mutants (ΔrpoN1, ΔrpoN2 and ΔrpoN1 ΔrpoN2) and subsequently tested their growth on an array of compounds. In addition to being essential for nitrate utilization, the rpoN1 gene was required for the assimilation of dicarboxylates, ethanol, ornithine, proline, propionate and xanthine. We did not observe any metabolic role for the rpoN2 gene. However, despite the lack of an observable role for metabolic function in R. solanacearum GMI1000, heterologous expression of the rpoN2 gene did genetically complement a ΔrpoN mutant of Escherichia coli. In contrast, heterologous expression of rpoN1 had no effect on restoring RpoN function in ΔrpoN E. coli. Plasmid-derived expression of the rpoN1 gene but not rpoN2 completely rescued the growth defects of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 R. solanacearum GMI1000 mutants. These findings not only confirm that the rpoN genes are not genetically redundant, but importantly, demonstrate that RpoN1 and RpoN2 are not interchangeable (functionally equivalent proteins) in R. solanacearum GMI1000.

Electroporation of R. solanacearum
A single colony of R. solanacearum was inoculated into 10 mL of CPG medium, and the culture was grown for 24 h at 30°C, 200 rpm. The entire 10 mL culture was centrifuged, and the cells were washed two times (1 st wash 5.0 mL, 2 nd wash 1.0 mL) with 10% (v/v) glycerol. After the second wash, cells were suspended in a final volume of 0.1 mL of 10% (v/v) glycerol. The cell suspension was given 0.05-1.0 μg of plasmid DNA and the plasmid-cell mixture was transferred to a 2-mm gap electroporation cuvette. Electroporation was performed at 2500 V in an ECM 399 (Harvard Apparatus). Cells were recovered in 1.0 mL of CPG at 30°C, 200 rpm for 1.0 h. Transformants were selected on CPG supplemented with the appropriate antibiotics.
Standard DNA procedures DNA was purified using Promega nucleic acid purification kits. Restriction enzymes, ligases and polymerases were products of New England BioLabs. Phusion polymerase (New England BioLabs) was used for all PCR applications. PCR was done using cycling times and parameters as recommended for the Phusion polymerase. Oligonucleotides used for PCR applications were purchased from Integrated DNA Technologies and are listed in Table 1. Cloned DNA was verified by sequencing (Genewiz).

Cloning of the rpoN genes
The rpoN1 and rpoN2 genes were PCR amplified with the primers BL568.f/BL568.r and BL569. f/BL569.r, respectively. The desired rpoN1 and rpoN2 PCR products (~1.5 kb) were gel-purified and separately cloned into pCR-Blunt (Invitrogen) according to the manufacturer's instructions. The rpoN1 and rpoN2 genes were then individually subcloned into the KpnI/XbaI sites of pBBR1MCS-2 [22] to give pBRL587 and pBRL584, respectively.

Deletion of rpoN genes in R. solanacearum GMI1000
A gene deletion methodology that was originally developed for Pseudomonas aeruginosa [23] was adopted and used to inactivate the rpoN1 and rpoN2 genes from R. solanacearum GMI1000. Construction of the rpoN deletion plasmids was done according to previously  published procedures [23]. The 5' and 3' ends (~250 bp) of the rpoN1 ORF were PCR amplified with the primers BL462.f/BL462.r and BL463.f/BL463.r, respectively. Similarly, primers BL464. f/BL464.r and BL465.f/BL465.r were used to PCR amplify the 5' and 3' ends, respectively, of the rpoN2 ORF. The Gm r marker was PCR amplified from the pPS856 plasmid [24] with the primers Gm-F/Gm-R. The desired PCR products were gel-purified. The purified 5' and 3' end fragments of each rpoN ORF were fused to Gm r marker via PCR with the primers GW-attB1/GW-attB2. The resulting rpoN1::Gm r and rpoN2::Gm r cassettes were gel-purified and individually cloned into pDONR221 (Invitrogen) using BP clonase II (Invitrogen). Lastly, LR clonase II (Invitrogen) was used to shuttle the rpoN1::Gm r or rpoN2::Gm r cassette from pDONR221 into the gene replacement vector pEX18ApGW [23]. The final pBRL535 and pBRL536 plasmids carried the rpoN1::Gm r and rpoN2::Gm r cassette, respectively, in pEX18ApGW. For the deletion plasmid pBRL560 (rpoN2::Tc r pEX18ApGW), primers BL464.f/BL464.r2 and BL465.f2/ BL465.r were used to PCR amplify the 5' and 3' ends of the rpoN2 ORF. The Tc r marker was PCR-amplified from pBR322 (NEB) with the primers Tc-F/Tc-R. Fusion PCR and subcloning of the rpoN2::Tc r cassette was done using identical procedures as described above.
To generate the single ΔrpoN1 and ΔrpoN2 mutants, 0.5-1.0 μg of pBRL535 or pBRL536 was electroporated into (0.1 mL) R. solanacearum GMI1000. Following a 1.0 h recovery period (30°C, 200 rpm), the entire culture was plated onto CPG supplemented with Gm. The plates were incubated at 30°C for 48 h. Gm r colonies were patched onto CPG supplemented with either Gm or Cb in order to differentiate from double and single crossovers, respectively. After 48 h at 30°C, patched clones displaying Gm r and Cb s were identified as putative ΔrpoN mutants. The rpoN1 and rpoN2 loci were PCR amplified from the ΔrpoN1 and ΔrpoN2 mutants. PCR products were gel-purified, cloned into pCR-Blunt (Invitrogen) and the inserts present in the recombinant plasmids were sequenced to verify the desired ΔrpoN mutation.
A similar procedure was used to create the double ΔrpoN1 ΔrpoN2 mutant. Briefly, the pBRL560 plasmid was electroporated into ΔrpoN1 R. solanacearum GMI1000, and colonies were selected for on CPG supplemented with Gm and Tc. Gm r Tc r colonies were patched onto CPG supplemented with Cb, and colonies displaying Cb s were identified as putative ΔrpoN1 ΔrpoN2 mutants. The rpoN loci were PCR-amplified, cloned and sequenced to verify the mutations in the double ΔrpoN1 ΔrpoN2 mutant.
Growth of ΔrpoN mutants of R. solanacearum GMI1000 on various carbon, nitrogen and sulfur sources Growth analyses were done in triplicate. Strains were first grown on solid CPG at 30°C for 48 h. Single colonies were inoculated into 1.0 mL of liquid CPG, and the inoculums were grown at 30°C, 200 rpm for 24 h. Next, 10 μL of a CPG-grown seed culture was used to inoculate 1.0 mL of a low-salt minimal media (15 mM K phosphate, 0.7 mM Na citrate, 0.5 mM   Table 2.

Plant assays
Tomato (Solanum lycopersicum) cultivar (cv.) Bonny Best and cv. Hawaii 7996 were grown in growth chambers (Conviron) with 12 h light-and-dark cycles at 25°C with 70% relative humidity. R. solanacearum strains were grown on solid CPG supplemented with 0.005% (w/v) of 2,3,5-tripheny tetrazolium chloride (TTC) at 30°C for 48 h. Single colonies that displayed a virulent morphology (mucoid and pinkish-white in color) [20] were grown in liquid CPG at 30°C for 24 h. Cells from each of these cultures were harvested, washed and then diluted in sterile water to a cell density of 1 x 10 6 CFU mL -1 . These inoculums were used for the virulence assays as described below.
Virulence of R. solanacearum strains for tomato cv. Bonny Best and cv. Hawaii 7996 was determined using standard procedures with some modifications [20,25]. Briefly, stems of healthy plants were wounded by puncturing into pith at the leaf axial using a disposable pipette tip. A drop of bacterial inoculum (1 x 10 6 CFU mL -1 ) was applied using a pipet tip, and the tip containing 10 μL of inoculum was left at the puncture site. Infected plants were returned to the growth chamber, and the day temperature was increased to 30°C. After the inoculum was taken up by the plant (10-24 h), the pipet tip was discarded.
Disease symptoms were evaluated daily and scored on a 0-5 disease index scale as follows: 0, healthy or no leaf wilting; 1, 10% of the leaf area wilted; 2, 11 to 40% of the leaf area wilted; 3, 41 to 60% of the leaf area wilted; 4, 61 to 85% of the leaf area wilted, 5, 85% of the leaf area wilted or dead leaf. Disease wilting index (dwi) was calculated using the following formula [20]: dwi (%) = [(sum of rating 0-5)] x 0.2 x (number of leaves evaluated) x 100. Plant assays were performed at least in duplicates using 8-9 plants per treatment.
The plants were sampled at 4 d post inoculation (dpi) for cv. Bonny Best and 6-7 dpi for cv. Hawaii7996. One-centimeter stem pieces just above the infection site were cut using a sterile razor blade. The stem pieces were weighed, sterilized with 70% (v/v) ethanol and then rinsed twice with sterile water. The stem pieces were ground using pestle in 800 μL of sterile water and serial dilutions were plated on CPG supplemented with 0.005% (w/v) TTC. Plates were incubated at 30°C for 18-24 h, and the number of bacterial colonies present were recorded. All experiments were performed at least in duplicates for each bacterial stain for both tomato cultivars. Data was analyzed using student t-tests, and values reported represent a mean of the colonies (± SD) per gram of fresh tissue.

Results and Discussion
Assimilation of a limited but diverse group of compounds is dependent only on the rpoN1 gene in R. solanacearum GMI1000 RpoN is well known for being involved in the assimilation of a wide variety of compounds among bacteria. Therefore, the ΔrpoN1, ΔrpoN2 and ΔrpoN1 ΔrpoN2 mutants of R. solanacearum GMI1000 were challenged with an array of carbon, nitrogen and sulfur sources ( Table 2). Preliminary experiments showed that all three rpoN mutants grew to cell densities (OD 600nm values) equivalent to that of wild-type in minimal media supplemented with 20 mM glucose and 20 mM NH 4 Cl (S1 Table). Based on this result, glucose was chosen as the carbon source for nitrogen utilization experiments while NH 4 Cl served as the nitrogen source for carbon utilization experiments.
Surprisingly, only the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants were observed to have either zero or reduced growth on a limited number of compounds (Fig 1). The ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants could not grow on ethanol, propionate, fumarate, malate, succinate and α-KG as carbon sources. For nitrogen sources, nitrate, ornithine, proline and xanthine did not support the growth of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants. The ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants had decreased growth on alanine and serine when provided as nitrogen sources. The ΔrpoN2 mutant exhibited wild-type growth on all compounds. Furthermore, all rpoN mutants showed wild-type growth on aliphatic sulfur sources even though RpoN is known to regulate such assimilation in other bacteria (S1 Table) [26].
The metabolic deficiencies of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants are in agreement with the putative EBPs and RpoN-controlled genes of R. solanacearum GMI1000 (Table 3). For example, the rpoN1 gene was necessary for the growth of R. solanacearum GMI1000 on C 4 -dicarboxylates and α-KG. In a number of bacteria, dicarboxylate utilization is dependent on RpoN and an EBP known as DctD, which together activate transcription of genes encoding for dicarboxylate transport proteins [27][28][29]. In R. solanacearum GMI1000, the Rsp0009 (dctD1) and Rsp0332 (dctD2) genes encode for homologs of DctD, and RpoN promoters are found upstream of genes encoding for C 4 -dicarboxylates and α-KG transport proteins. It is expected that R. solanacearum GMI1000 also uses an RpoN-DctD mechanism to regulate uptake of dicarboxylates.
RpoN and the EBP PrpR regulate the degradation of propionate in some bacteria [33,34]. Indeed, the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants of R. solanacearum GMI1000 did not grow on propionate as a carbon source. Rsp0123 is a predicted homolog of PrpR, and there is an RpoN promoter upstream of an operon encoding for enzymes of a propionate catabolic pathway in R. solanacearum GMI1000. Propionate catabolism in R. solanacearum GMI1000 is iron-sulfur cluster repair protein Because the RpoN-interaction domain is conserved in EBPs, the RpoN-interaction domain of E. coli NtrC was used in a BlastP search against the protein database of R. solanacearum GMI1000. This searched returned a total of thirteen putative EBPs, which is in agreement with a previously published assessment [30]. Nine EBPs possess the signature GAFTGA motif while the remaining four have GSFTGA or GAYTGA. A partial listing of potential gene targets regulated by RpoN  likely regulated by RpoN1 and PrpR. Rsc3129 has homology to the EBP AcoR [35,36] and is predicted to regulate transcription of an acetaldehyde dehydrogenase (ExaC), a key enzyme in the breakdown of acetoin, ethanol and ethanolamine [37,38]. The rpoN1 gene was required for growth on ethanol, but we observed no growth for the wild-type strain or rpoN mutants on acetoin or ethanolamine. Pectin, sucrose, galactose and hexoses in general are common substrates for R. solanacearum in the environment. Deletion of the rpoN genes had no impact on the consumption of these compounds. The utilization of nitrate, however, was completely dependent on the rpoN1 gene. This finding is in agreement with a previous study [19]. Nitrate assimilation was observed to be a crucial component for R. solanacearum in the host plant environment [39]. The NtrCresponse was involved in nitrate utilization [39]. Nitrate might serve as a valuable nitrogen source for R. solanacearum during host plant interactions.
The rpoN genes were not required for the use of amino acids as carbon sources. However, ornithine and proline were the only amino acids that could not be used as nitrogen sources by the ΔrpoN1 mutants. The ΔrpoN1 mutants grew poorly on serine and alanine as nitrogen sources but nonetheless still produced wild-type cell densities after 72 h of incubation. The strict rpoN1-dependency for ornithine could be attributed to the Rsc3410 locus, which encodes for a protein sharing some homology to the glutamine transporter GlnH [40]. Ornithine transport might be mediated through Rsc3410. There are no putative RpoN-controlled genes encoding for catabolic enzymes of proline degradation in R. solanacearum GMI1000. Proline degradation is regulated by Nac (nitrogen assimilation regulator) in some bacteria [41,42], and the nac gene in R. solanacearum GMI1000 is preceded by an RpoN promoter. This might account for the RpoN1-dependency of proline utilization.

RpoN1 and RpoN2 are not interchangeable
The rpoN1 and rpoN2 genes were individually cloned under the lac-promoter on the broadhost range plasmid pBBR1MCS-2 [22]. The plasmid-encoded rpoN1 and rpoN2 genes were electroporated into the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants, and the recombinant strains were challenged with compounds whose assimilation was previously found to be rpoN1-dependent. As shown in Fig 2, plasmid-derived expression of either rpoN1 or rpoN2 recovered the growth of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants on C 4 -dicarboxylates. Additionally, expression of rpoN1 rescued the growth of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants on ethanol, α-KG, nitrate, ornithine, proline and propionate. In sharp contrast, expression of rpoN2 did not allow for the growth of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants on these compounds. Complementation via RpoN2 was specific for C 4 -dicarboxylates, indicating that RpoN1 and RpoN2 are not equivalent proteins. Although this observation was unexpected at first, subsequent studies with the rpoN1/rpoN2 genes in ΔrpoN E. coli confirmed that the RpoN1 and RpoN2 proteins are not functionally identical.

RpoN1 and RpoN2 have different functionality in ΔrpoN E. coli
The RpoN1 and RpoN2 proteins share 63% and 55% homology to E. coli RpoN, respectively. Therefore, we decided to determine if the rpoN1 and rpoN2 genes were capable of genetically complementing a ΔrpoN E. coli strain [21]. The plasmid-encoded rpoN1 and rpoN2 genes were introduced into ΔrpoN E. coli strain, and the recombinant ΔrpoN E. coli strains were assayed under growth conditions known to be RpoN-dependent, including nitrogen assimilation (NtrC-regulated response) [43], propionate catabolism (PrpR-regulated response) [33] and acetoacetate metabolism (AtoC-regulated response) [44].
Heterologous expression of either the rpoN1 or rpoN2 gene did recover the growth of the ΔrpoN E. coli on nitrogen sources such as asparagine and xanthine (Fig 3). However, only the expression of rpoN2 completely rescued the growth of ΔrpoN E. coli under nitrogen limitation (NH 4 < 0.5mM) and when arginine was the sole nitrogen source. Additionally, rpoN2 (but not rpoN1) restored the growth of ΔrpoN E. coli on propionate and acetoacetate to *25% and 100% of wild-type levels, respectively. These results indicate that the RpoN1 and RpoN2 proteins are not functionally equivalent.
Alignment of RpoN1, RpoN2 and E. coli RpoN revealed that all three proteins have *80% homology in region I (residues 1-50) and 60% homology in region III (C-terminal 350 residues). For RpoN2 and E. coli RpoN, they also share weak homology in region II (residues 51-100) (Fig  4). This homology does not exist between RpoN1 and E. coli RpoN, and thus, would explain why RpoN2 was more sufficient in complementing ΔrpoN E. coli than that of RpoN1 (Fig 3).
Region II of RpoN has been implicated in EBP interactions [7], and NtrC of R. solanacearum GMI1000 has 70% homology to E. coli NtrC. Based on our results, it would appear that RpoN2 could function with NtrC of E. coli to facilitate nitrogen assimilation. RpoN1 can also interact with E. coli NtrC but to a much lesser extent than that of RpoN2 [as judged by RpoN1 being able to only partially compensate for nitrogen assimilation in ΔrpoN E. coli]. R. solanacearum NtrC exhibits 60% homology to E. coli AtoC. Expression of rpoN2 restored the growth of ΔrpoN E. coli on acetoacetate, suggesting that RpoN2 can partner with AtoC to regulate transcription from AtoC-controlled genes. R. solanacearum PrpR and E. coli PrpR have *50% homology. The differences in sequence between the PrpR EBPs are reflected in the inability of RpoN2 to fully restore the growth of ΔrpoN E. coli on propionate.
The rpoN1 gene is required for virulence of R. solanacearum GMI1000 on tomato Dicarboxylates and nitrate were previously implicated as key nutrients for R. solanacearum in the host plant environment [25,39]. Because the rpoN1 gene was needed for the growth of R. solanacearum GMI1000 on these compounds, it was suspected that rpoN1 might also be a necessary factor in the pathogenesis of this pathogen. Indeed, this suspicion was recently Defining RpoN functions in Ralstonia solanacearum GMI1000 confirmed in a study, which found that the rpoN1 gene was necessary for virulence of R. solanacearum GMI1000 on tomato plants (Lycopersicum esculentum cv. Marmande) [19]. Therefore, we decided to determine if the ΔrpoN1 mutants generated in our study were also avirulent. To this end, tomato plants consisting of cv. Bonny Best (susceptible) and cv. Hawaii 7996 (resistant) were inoculated with a bacterial suspension comprising of wild-type, ΔrpoN1, ΔrpoN2 or ΔrpoN1 ΔrpoN2 R. solanacearum GMI1000. The inoculated plants were then examined on a daily basis for signs of wilt disease, which was quantified and used to calculate a disease-wilting index (dwi). Wilting was first observed on the older leaves of cv. Bonny Best plants inoculated with wild-type and ΔrpoN2 R. solanacearum GMI1000. The overall pattern of disease was similar for both tomato cultivars but the extent of wilting was severe in cv. Bonny Best compared to cv. Hawaii 7996 (Fig 5). Entire cv. Bonny Best plants wilted earlier (4 dpi) than cv. Hawaii 7996 (7 dpi). However, ΔrpoN1and ΔrpoN1 ΔrpoN2-inoculated cv. Bonny Best and cv. Hawaii 7996 plants did not develop any significant wilting symptoms after 4 and 7 days, respectively (Fig 5). Mock plants treated with sterile water had no wilting symptoms (Fig 5). At 2 dpi, wild-type and ΔrpoN2-inoculated cv. Bonny Best plants had 35-50% dwi, which increased to 60-67% at 3 dpi. In contrast, progression of disease was reduced in ΔrpoN1and ΔrpoN1 ΔrpoN2-inoculated cv. Bonny Best plants with dwi reaching only up to 0-7% at 3 dpi (Fig 6A). The wild-type and ΔrpoN2 inoculated cv. Hawaii 7996 population had 23-45% dwi at 6 dpi, which increased up to 40-45% at 7 dpi. ΔrpoN1and ΔrpoN1 ΔrpoN2-inoculated cv. Hawaii-7996 population did not show any visible disease symptoms. At 6 dpi, the dwi for these plants was 0-4%, which did not significantly change at 7 dpi (0-8%) (Fig 6B). We next evaluated the in planta growth of the rpoN mutants. Plants were inoculated with each strain at a titer of 1 x 10 6 CFU mL -1 and bacterial growth (CFUs per gram of fresh tissue) was determined at 4 dpi for cv. Bonny Best and 7 dpi for cv. Hawaii 7996. The bacterial growth could not be determined at 7 dpi for the cv. Bonny Best plants, because all plants were dead by this particular dpi. At 4 dpi in cv. Bonny Best, wild-type cells and the ΔrpoN2 mutant produced 1000-fold more CFUs than that of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants. Similarly, at 7 dpi in cv. Hawaii 7996, wild-type cells and the ΔrpoN2 mutant produced~100-fold more CFUs than that of the ΔrpoN1 and ΔrpoN1 ΔrpoN2 mutants (Fig 7). These results are consistent with the patterns of wilting disease caused by wild-type and ΔrpoN2 strains on both tomato cultivars. Collectively, these findings demonstrate that rpoN1 but not rpoN2 is critical for virulence of R. solanacearum on tomato and they are not functionally redundant for virulence.

Conclusions
Although two rpoN genes are present in R. solanacearum GMI1000, our results, in combination with previous findings clearly indicate that the rpoN1 and rpoN2 genes are not genetically redundant. Functions commonly associated with RpoN regulation were dependent only on the rpoN1 gene. Additionally, this study demonstrates that the RpoN1 and RpoN2 proteins are not equivalent or interchangeable. Why R. solanacearum GMI1000 possesses the genetic machinery to encode dual RpoN proteins that are not interchangeable is an intriguing question. It is possible that the EBPs of R. solanacearum GMI1000 have evolved to preferentially interact with RpoN1 to activate transcription from RpoN promoters while RpoN2 may have evolved to play more of an antagonistic role in R. solanacearum GMI1000, i.e., repressing transcription from genes possessing RpoN promoters. Future studies aimed at identifying the RpoN-EBP partnerships in R. solanacearum GMI1000 are expected to clarify the regulatory roles of the RpoN1 and RpoN2 proteins. Supporting Information S1 Table. Growth data of wild-type, ΔrpoN1, ΔrpoN2, ΔrpoN1 ΔrpoN2 R. solanacearum GMI1000 on various carbon, nitrogen and sulfur sources. Each strain (n = 3) was grown in minimal media in which the indicated compound served as either the sole carbon, nitrogen or sulfur source. The mean absorbance value at 600 nm (± standard deviation) is reported for each strain. (DOCX)