A Sinorhizobium meliloti and Agrobacterium tumefaciens ExoR ortholog is not crucial for Brucella abortus virulence

Brucella is a facultative extracellular-intracellular pathogen that belongs to the Alphaproteobacteria class. Precise sensing of environmental changes and a proper response mediated by a gene expression regulatory network are essential for this pathogen to survive. The plant-related Alphaproteobacteria Sinorhizobium meliloti and Agrobacterium tumefaciens also alternate from a free to a host-associated life, where a regulatory invasion switch is needed for this transition. This switch is composed of a two-component regulatory system (TCS) and a global inhibitor, ExoR. In B. abortus, the BvrR/BvrS TCS is essential for intracellular survival. However, the presence of a TCS inhibitor, such as ExoR, in Brucella is still unknown. In this work, we identified a genomic sequence similar to S. meliloti exoR in the B. abortus 2308W genome, constructed an exoR mutant strain, and performed its characterization through ex vivo and in vivo assays. Our findings indicate that ExoR is related to the BvrR phosphorylation state, and is related to the expression of known BvrR/BrvS gene targets, such as virB8, vjbR, and omp25 when grown in rich medium or starving conditions. Despite this, the exoR mutant strain showed no significant differences as compared to the wild-type strain, related to resistance to polymyxin B or human non-immune serum, intracellular replication, or infectivity in a mice model. ExoR in B. abortus is related to BvrR/BvrS as observed in other Rhizobiales; however, its function seems different from that observed for its orthologs described in A. tumefaciens and S. meliloti.


Introduction
Members of the Brucella genus are worldwide distributed zoonotic pathogens that belong to the Alphaproteobacteria class. Infection in humans is associated with diverse and unspecific symptoms, including fever, lassitude, general malaise, weight loss, headache, low back pain, and arthralgias [1]. Brucella abortus infects bovines, where it preferentially replicates in reproductive organs, causing abortion, infertility, decreased milk production, reproductive failure, transcriptional activators related to quorum sensing [31][32][33]. Furthermore, direct regulation of VjbR over the virB promoter has been demonstrated [22,29]. To date, ExoR has not been described for B. abortus or any other Brucella spp. Due to the homology between ExoS and BvrS, it is hypothesized that the Brucella ExoR ortholog could be functionally related to BvrR/BvrS [28]. To elucidate this hypothesis, an exoR mutant derived from the reference strain B. abortus 2308 Wisconsin was constructed and studied [34]. The results show that ExoR affects VirB, Omp25, and VjbR expression dynamics and BvrR phosphorylation, but its absence does not affect B. abortus virulence ex vivo or in vivo.

B. abortus 2308W genome encodes an exoR ortholog
Using Basic Local Alignment Search Tool (BLASTp) and the ExoR protein sequence from S. meliloti (GenBank Accession WP_003534542.1) [35,36] a similar sequence (51% identity), possibly encoding a 267 amino acid residues long protein was found in the B. abortus 2308W genome (GenBank Accession ERS568782) [34]. This sequence was found under locus tag BAW1_0856, between open reading frames annotated as a predicted protein and an exodeoxyribonuclease III. Based on this information, we constructed a B. abortus 2308W mutant with an in-frame deletion of 171 internal amino acid residues. Deletion of the 515 bp was confirmed by PCR, Southern blot assays [37], and Sanger sequencing, as described in the Methods section and S1 Fig. No significant differences in growth kinetics in TSB ( Fig 1A) and sugar assimilation, according to API 50 CH assay, were detected between the mutant and the wild-type strain.
The exoR mutation does not affect B. abortus resistance to polymyxin B and nonimmune human serum. The B. abortus LPS smooth phenotype was not altered by the exoR mutation, as judged by the acriflavine agglutination test. Exposure to non-immune human serum and increasing polymyxin B concentrations at pH 7 and pH 6 did not reveal differences between the wild-type and the exoR mutant strain (Fig 1B, 1C and 1D). As expected, the bvrR and bvrS mutants, included as controls, were sensitive to polymyxin B and non-immune human serum (Fig 1C and 1D). Both strains showed significant differences when compared to the wild-type and the exoR mutant.
ExoR impacts the expression dynamics of proteins related to BvrR/BvrS in nutrient rich conditions. We assessed the expression of BvrR and BvrS by western blot of cell lysates obtained at different moments of growth in TSB. We observed similar BvrR and BvrS expression patterns compared to the wild-type strain in the mutant strain (Fig 2). The BvrR phosphorylation state was analyzed by Phos-tag SDS-PAGE and western blot [22,38]. As expected, two bands were recognized, corresponding to BvrR-P (upper band) and non-phosphorylated BvrR (Fig 2). BvrR-P remained constant in the wild-type strain over time, ranging from 11 to 20% of total BvrR, until 28 hours of growth where there was no phosphorylation signal. In the exoR mutant, BvrR-P constantly decreased, and its signal disappeared more radidly, going from 15 to 6%, and being undetectable at 24 hours.
We also evaluated three proteins whose expression is positively regulated by BvrR/BvrS: VirB8, VjbR, and Omp25 (Fig 3). VirB8 and VjbR reached a maximum expression around late-log and mid-log in the wild-type strain, respectively, and then decreased until reaching the stationary phase, as previously reported [29]. In the exoR mutant, VirB8 and VjbR expression showed a similar trend; nevertheless, the expression declined before it occurred in the wildtype, around mid-log. Omp25 showed slight differences in expression according to the wildtype strain time of growth as reported [39]. In the exoR mutant, expression levels gradually decreased from the early log until the stationary phase of growth.
Low nutrients and pH do not significantly alter B. abortus exoR phenotype as compared to nutrient rich conditions. BvrR phosphorylation increases when exponential growing bacteria are exposed to minimal medium and pH 5.0 [22]. These conditions mimic the intracellular environment, where BvrR/BvrS is critical for survival [23,24,40]. A possible role for ExoR related to the activation of BvrR/BvrS was assessed using similar assays. Hence, the polymyxin B and non-immune human serum assays were carried out using bacteria after 4-hours incubation in minimal medium and pH 5.0 (Fig 4A and 4B). No significant difference was found between the wild-type and the mutant strain. As expected, the bvrR mutant, included as a control, was susceptible to polymyxin B and non-immune human serum. This strain showed significant differences when compared to the wild-type strain and the exoR mutant.
The expression of VirB8 and VjbR in the wild-type strain under these conditions was similar to the previously reported and related to BvrR transient phosphorylation [29]. The exoR mutant showed a similar expression pattern as compared to the wild-type for both proteins , exoR, bvrS, and bvrR mutants, were exposed to non-immune human serum at 37˚C for 90 min. These experiments are representative of at least three performed. (C) and (D) The indicated strains, B. abortus 2308W (WT), exoR, bvrS, and bvrR mutants, were exposed to increasing concentrations of polymyxin B for 48 hours at 37˚C, at pH 7 and pH 6, respectively. The minimum inhibitory concentrations shown are representative of at least three experiments performed. The bvrS and bvrR mutant strains were used as susceptibility controls. � , P < 0,05 (Anova and Kruskal-Wallis test for multiple comparisons).
The exoR mutant successfully infects mice and cell cultures. The ability to escape the host immune response and to survive intracellularly is critical for B. abortus pathogenicity [41,42]. Cellular and murine models were used to evaluate the exoR mutant strain infective abilities. Fig 5A and 5B show that the exoR mutant infected HeLa and Raw macrophage cells similar to the wild-type strain. Fig 5C shows that the dynamics of infection of the wild-type strain in mice were consistent with previous results [26,43]. We found no significant differences between the wild-type and the exoR mutant, indicating that ExoR is not required for B. abortus infection in mice or cell culture.

Discussion
BvrR/BvrS homologs in S. meliloti and A. tumefaciens are required to associate with plant cells [14,36]. Similarly, B. abortus needs BvrR/BvrS to infect the host animal cell successfully [44]. However, the RSI invasion switch functioning in both plant cell microbes, where ExoR is a protagonist, seems dissimilar in B. abortus.

Fig 2. BvrR/BvrS expression and BvrR phosphorylation dynamics in the exoR mutant.
Western blot analysis of BvrS, BvrR, and BvrR-P expression according to growth phase in the B. abortus exoR mutant (exoR) and B. abortus 2308W (WT). Both strains were grown in TSB for 48 hours. Representative hours of different growth phases were analyzed (mid-, late-log, and stationary) indicated on top. For western blot, equal amounts (20 μg) of whole-bacterium lysates were separated by 12,5% SDS-PAGE. For phosphorylation analyses, samples were separated by 10% SDS-PAGE containing Phos-tag. PVDF membranes were incubated with anti-BvrS, anti-BvrR, and and reprobed with anti-omp19 antibodies indicated on the right side. After incubation, the immune complexes were detected by chemiluminescence reaction. Omp19 was used as a loading control. The blots for BvrS, BvrR and Phos-tag are from independent gels. https://doi.org/10.1371/journal.pone.0254568.g002 In the RSI invasion switch, ExoR levels are critical for host interaction and subsequent invasion in S. meliloti. exoR95:: Tn5 loss-of-function mutant overproduces succinoglycan and has a dramatic decrease of flagellar gene expression and other genes related to motility and chemotaxis [18,36]. A similar effect was observed in an A. tumefaciens exoR null mutant [9]. In both cases, host interactions are compromised. According to our mice model result, this does not seem to be the case for B. abortus.
However, ExoR seems to be related to BvrR/BvrS. Proteins whose gene expression is under the control of BvrR/BvrS [29,45] showed a different expression pattern in the exoR mutant as compared to the wild-type strain. The transcriptional regulator VjbR, the T4SS VirB component VirB8 (essential for intracellular survival) [23,29,46,47] and Omp25 (with a potential role in virulence) [48][49][50], were expressed in the ΔexoR mutant at lower levels, particularly starting at late log growth phase, as compared to the wild strain. Nevertheless, the exoR mutant was able to infect and survive in cell and mice models. These results were unexpected considering the high degree of identity between the BvrR, BvrS, and ExoR homologs. If the B. abortus ExoR homolog is functioning as its counterparts, i.e., as a sensor protein inhibitor, the TCS's upregulated genes would be overexpressed. For example, in A. tumefaciens, ChvG/ChvI upregulates the expression of an Omp25 orthologue, known as AopB [20,38]. This protein influences the stability, permeability, and topology of the membrane, as does Omp25 and other OMPs in B. abortus [51,52]. Since ExoR represses ChvG by direct interaction, inhibiting the signal transferred through the TCS, aopB was highly upregulated in the A. tumefaciens exoR null mutant [20]. The same has been observed for succinoglycan genes that are upregulated by ExoS/ChvI from S. meliloti [13]. Surprisingly, in B. abortus exoR mutant Omp25 was down expressed. Moreover, BvrR and BvrS did not show differences in their expression compared to the wild-type. This suggests that their expression is independent of ExoR, but possible posttranslational modifications (including phosphorylation and dephosphorylation) might not, since the BvrR phosphorylation dynamics was different as compared to the wild-type strain. The exoR mutant showed a gradual fade of the BvrR-P signal during bacterial growth, which occurs faster than in the wild-type strain. There is no direct evidence linking the phosphorylation of the response regulator and ExoR in other bacterial RSI switches; but, in A. After incubation, bacteria were exposed to non-immune human serum at 37˚C for 45 min or (B) exposed to increasing concentrations of polymyxin B for 48 hours at 37˚C at pH 7. The Anova and Kruskal-Wallis test for multiple comparisons was used. � , P < 0,05. (C) The wild-type and exoR mutant were grown as indicated in (A) for 6 hours, and 1 mL culture of each strain was resuspended in minimal medium, pH 5 for 0.5, 2, 4, and 6 hours. After incubation, lysates were prepared, separated by 10% SDS-PAGE and assessed using anti-VjbR and anti-VirB8 antibodies by western blotting. Omp19 detection was used as a loading control. For comparison, a bacterial sample from each strain, with no exposure to minimal medium, pH 5 was included and is labeled as "C".
https://doi.org/10.1371/journal.pone.0254568.g004 tumefaciens and S. meliloti ExoR absence induces a phenotype similar to that of bacteria with a constitutive TCS. This does not seem to be the case in B. abortus 2308W.
In A. tumefaciens ChvG/ChvI is activated under low-pH conditions and ExoR represses it under neutral conditions. A null exoR mutant shows a phenotype at neutral pH similar to that of acid-exposed wild-type bacteria, demonstrating that ExoR is required to inactivate the system [20]. In S. meliloti, ExoR inhibits flagella production and stimulates succinoglycan production. Mutants with defective ExoR mimic non-motile succinoglycan-producing bacteria, demonstrating the need for ExoR to inhibit ExoS [18]. This inhibitory ExoR function on BvrR/ BvrS was not clearly evident in our mutant under neutral pH conditions. Therefore, we also evaluated the possibility that the exoR mutation has an evident effect under starving and acidexposed conditions, which are also relevant for BvrR phosphorylation, and induction of VjbR and VirB expression in the wild-type strain [22]. Our results show that, as in rich medium, pH 7, the exoR mutation has no effect after exposure to minimal medium and pH 5, on the B. abortus ability to resist polymyxin B or non-immune human serum treatment, and that there are subtle differences in expression of VjbR and VirB8 as compared to the wild-type strain.
Nevertheless, there seems to be a partial functional similarity between ExoR from B. abortus and its orthologs. ExoR is not required for B. abortus virulence in the analyzed models. However, there is a functional link with BvR/BvrS, clearly shown by altering BvrR phosphorylation patterns and expression levels of some of the downstream BvrR/BvrS targets. One possible interpretation is that ExoR is inhibiting, directly or indirectly, the dephosphorylation of BvrR and consequently, the expression of the analyzed genes. Unlike the studies published for S. meliloti and A. tumefaciens, we conclude B. abortus exoR mutant maintains the ability to invade and replicate inside cells and survive inside the host. Since the introduced deletion in exoR did not remove the gene entirely, we cannot exclude that the remaining protein segment kept some of the functions of the native ExoR. More studies are needed to understand the role of the conserved ExoR, not only in B. abortus but also in other members of the genus, as well as other members of the Alphaproteobacteria.

Strain construction
B. abortus exoR mutant was constructed as reported elsewhere [54,55]. Briefly, in-frame deletion was generated by PCR overlap using genomic DNA of B. abortus as a template. Primers were designed using the available genome sequence ERS568782, corresponding to reference strain B. abortus 2308W to perform a 515 bp internal deletion (position +172 to +687 from the first codon ATG), leaving only 35% of the coding sequence (S1 Fig). Final validation of successful deletion was done using Sanger DNA sequencing of selected fragments and Southern blot (see below and S1 Fig).
Plasmid and chromosomal DNA were extracted with QIAprep Spin Miniprep and DNeasy Blood and Tissue Kit (Qiagen).

Carbohydrate assimilation pattern
The carbohydrate assimilation pattern was performed using the API 50 CH kit (Biomerieux) according to the manufacturer's instructions. Briefly, a bacterial suspension of both strains, mutant and wild-type, grown in TSB at pH 7, with agitation of 200 rpm to exponential phase, were used to rehydrate each of the wells. The strips were incubated at 37˚C for 18 hours. During incubation, metabolism produces color changes. Positive and negative test results were annotated to obtain a profile. The profile from three different assays was compared to the wild-type strain profile.

Southern blot
Whole genomic DNA (0.5 μg) from Brucella strains was digested with 5U of the restriction endonucleases BamHI and PstI (Fermentas) in a reaction volume of 20 μl and according to the manufacturer's instructions. The resulting fragments were separated in a 0.7% agarose gel. After electrophoresis, hybridization was carried out with minor variations as described previously [37]. Primers bruabI0884.3 and bruabI0884.5 were used to generate a 295 bp amplicon from B. abortus 2308W DNA (S1 Table). These primers amplify a region from position 68 bp to -363 from exoR first codon (S1 Fig), used as a probe after digoxigenin labeling (DIG DNA Labeling and detection kit, Roche). The labeled probe was denatured and added to the hybridization buffer. Hybridization was performed overnight. After exposure to ultraviolet light, the membrane was then washed, blocked, incubated with anti-DIG alkaline phosphatase-conjugated FAB-antibody, and washed according to the manufacturer's instructions. Luminescence was recorded on X-Ray films (Kodak) for 4 hours.

Polymyxin B sensitivity assay
Polymyxin B (Sigma, USA) sensitivity assays were performed in triplicate as described with some modifications [56]. Both strains, mutant, and wild-type, were grown until exponential phase in TSB at pH 7, with the agitation of 200 rpm. Bacteria of each strain were adjusted to 5x10 6 CFU/mL in TSB. 100 μL sample of each strain was mixed with 100 μL of different polymyxin B concentrations in a microplate (the final concentration of the first well was 250 ug/ mL). After 48 h of incubation at 37˚C in two different pH (6 and 7), the minimum inhibitory concentration for polymyxin was calculated.

Sensitivity to the bactericidal action of non-immune serum
Exponentially growing bacteria grown in TSB at pH 7, with agitation of 200 rpm, were adjusted to 10 4 CFU/mL in PBS and dispensed in duplicate in microtiter plates (200 μl per well) containing fresh human serum (400 μl/well). After 90 min of incubation at 37˚C, 100 μl was plated on tryptic soy agar. Results were expressed as the percentage CFU as compared to controls performed with decomplemented serum at room temperature for 30 min. To assess complement consumption, serum was homogenized with dehydrated yeast Saccharomyces cerevisiae for 1 h at 37˚C. After incubation, the serum was centrifugated at 14000 rpm for 5 min [25].

Western blot and BvrR phosphorylation
Detection of BvrS, BvrR, VirB8, VjbR, Omp25, and determination of BvrR phosphorylation was performed as previously described [22]. Bacteria grown to exponential phase in TSB at pH 7, with the agitation of 200 rpm, were concentrated by centrifugation, resuspended in Laemmli sample buffer, and heated at 100˚C for 20 min. The protein concentration was determined by the Bio-Rad DC method according to the manufacturer's instructions. Equal amounts of protein (20 μg) were loaded onto a 10% gel for SDS-PAGE. Separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane and probed with the indicated antibodies. To analyze the phosphorylated status of BvrR, samples were solubilized in Laemmli sample buffer without heating, and equal amounts of protein (20 μg) were loaded onto a 10% gel for SDS-PAGE containing Phos-tag (100 mM) and MnCl2 (0.2 mM). Recombinant BvrR phosphorylated with the phosphate universal donor carbamoyl phosphate was used as a positive control. The percentage of BvrR-P from total BvrR was calculated for each indicated condition by densitometry from at least three independent experiments.
For loading control each membrane was reprobed. First, the membranes were washed with PBS Tween 0,1% for 30 minutes and incubated in glycine 0.1M pH 2.5 for 1 hour. After incubation, membranes were probed with Omp19 or Bscp31 antibody.

Exposure of Brucella strains to minimal and low pH medium
Bacterial strains were grown in vitro in TSB to exponential growth phase [22]. A volume of each culture corresponding to 10 4 CFU/mL or 5x10 6 CFU/mL was calculated for the sensitivity to the bactericidal action of non-immune serum or the polymyxin sensitivity assays, respectively. Bacteria were centrifuged at 10 000 x g for 3 min and resuspended in minimal medium at pH 5.0 or in a nutrient-rich medium (TSB) at pH 7.0 for 4 h at 37˚C at 200 rpm. After incubation, bacteria were concentrated by centrifugation at 10,000 x g for 3 min and tested as described above. For the VirB8 and VjbR protein expression assays, aliquots of 1mL of bacteria grown in TSB to exponential growth phase were centrifuged, resuspended in minimal medium, pH 5 for 0.5, 2, 4, and 6 hours. After incubation, each sample was centrifuged and tested as described above. For comparison, a 1mL aliquot with no exposure to minimal medium, pH 5 was included.

Gentamicin-protection assay and intracellular replication quantification
Murine RAW 264.7 macrophages (ATCC TIB-71) or HeLa epithelial cells (ATCC clone CCl-2) were cultivated and infected with B. abortus strains at exponential and stationary growth phase as previously described with some modifications [22]. Cells were seeded in 24-well tissue culture plates and multiplicities of infection (MOI) of 100 for macrophages and 500 for HeLa epithelial cells. The number of intracellular viable B. abortus CFU was determined at different hours post-infection. Cells were washed twice with phosphate-buffered saline (PBS) and treated with Triton X-100 (0.01%). Lysates were serially diluted and plated on tryptic soy agar dishes for the quantification of CFU.

Virulence assays in mice
BALB/c female (18-24 grams) mice were intraperitoneally (i.p.) inoculated with 10 6 UFC of either B. abortus 2308W or B. abortus 2308W exoR mutant at exponential phase grown in TSB at pH 7, with agitation of 200 rpm. Mice were sacrificed at 3-and 9-weeks post-infection. Spleen counts were determined as described elsewhere [57]. Protocols of experimentation were revised and approved by the Welfare Commission of the Veterinary School at Universidad Nacional, Costa Rica, under protocol number FCSA-EMV_CBAB-009-2015, and agreed with the corresponding law, Ley de Bienestar de los Animales, of Costa Rica (law 7451 on animal welfare). Only certified veterinarians carried out these experiments and all efforts were made to minimize suffering.

Statistical analysis
Statistical analyses were performed using GraphPad Prism software. All results are presented as means ± SD from at least three independent experiments unless otherwise stated. Nonparametric statistics were used. Anova and Kruskal-Wallis test were used for multiple comparisons and Mann-Whitney test for the comparison of two strains.

S1 Fig. Mutant construction and Southern blot. (A)
Schematic representation of the inframe deletion strategy used to construct the exoR mutant. Primers exoR-F1 and exoR-R2 were used to generate fragment 1, and fragment 2 was generated using exoR-F3, and exoR-R4 (S1 Table). Both fragments were ligated by PCR overlapping using nucleotides exoR-F1 and exoR-R4. The resulting deletion allele was cloned in the pCR™ 2.1 vector (Invitrogen™) and subcloned into the BamHI-XBaI site of the suicide plasmid pJQKm. Plasmid pJQKm containing the deleted allele was introduced in B. abortus 2308W by conjugation. Colonies corresponding to integrating the suicide vector in the chromosome were selected using Nalidixic acid (NaI, 25 μg/mL) and Kanamycin (Km, 50 μg/mL) resistance. Excision of the suicide plasmid leading to the mutant's construction by allelic exchange was selected by 5% sucrose resistance and Km sensitivity. The resulting colonies were screened using primers exoR-F1 and exoR-R4. Mutant colonies generated a 780 bp fragment and the parental strain a 1200 bp fragment. (B) Southern blot analysis. BamH1 and PstI restriction sites were chosen according to the B. abortus 2308W genome sequence, at positions -3009 and +1552, respectively, from exoR first codon. The resulting fragments were separated in a 0.7% agarose gel. After electrophoresis, the protocol was carried out with minor variations as described in Methods. Briefly, the gel was rinsed in 0.25 M HCL, denatured (1.5 M NaCl, 0.5 M NaOH), and neutralized twice (1.5 M NaCl, 0.5M Tris HCl pH8). The DNA was then transferred to a nylon membrane (Roche) overnight in 10x SSC pH 7 (0.15 M sodium citrate, 1.5 M NaCl). Primers bruabI0884.3 and bruabI0884.5 (S1 Table) were used to generate a 295 bp amplicon from B. abortus 2308W DNA. These primers amplify a region from position 68 bp to -363 from exoR first codon. The amplicon was purified from agarose gels using QIAquick Gel Extraction Kit (Qiagen) and labeled with digoxigenin for use as a probe according to the manufacturer's instructions (DIG DNA Labelling and detection kit, Roche). The labeled probe was denatured and added to the hybridization buffer. Hybridization was performed overnight. After exposure to ultraviolet light, the membrane was then washed, blocked, incubated with anti-DIG alkaline phosphatase-conjugated FAB-antibody, and washed according to the manufacturer's instructions. Luminescence was recorded on X-Ray films (Kodak) for 4 hours. (TIF) S1