Rev1 wbdR tagged vaccines against Brucella ovis

Sheep brucellosis is a worldwide extended disease caused by B. melitensis and B. ovis, two species respectively carrying smooth or rough lipopolysaccharide. Vaccine B. melitensis Rev1 is used against B. melitensis and B. ovis but induces an anti-smooth-lipopolysaccharide response interfering with B. melitensis serodiagnosis, which precludes its use against B. ovis where B. melitensis is absent. In mice, Rev1 deleted in wbkC (Brucella lipopolysaccharide formyl-transferase) and carrying wbdR (E. coli acetyl-transferase) triggered antibodies that could be differentiated from those evoked by wild-type strains, was comparatively attenuated and protected against B. ovis, suggesting its potential as a B. ovis vaccine.


Introduction
Brucellosis is one of the most common zoonoses worldwide, causing significant loss to animal production and severely affecting human health [1]. The etiological agents of brucellosis are Gram-negative bacteria of the genus Brucella. This genus includes several zoonotic species among which B. abortus preferentially infects cattle, B. suis swine and wild-life and B. melitensis goats and sheep [2]. Sheep can also be infected by B. ovis, a nonzoonotic species [2]. These species are smooth (S) or rough (R), depending on the presence or absence respectively of O-polysaccharide (O-PS) chains in the outer membrane lipopolysaccharide (LPS). Whereas B. abortus, B. melitensis and B. suis carry a S-LPS, B. ovis is a naturally a R species [2].
B. ovis infection causes genital lesions and reduced fertility constituting one of the most important causes of reproductive failure in sheep [3]. Animal vaccination is the most suitable method for controlling brucellosis in areas with moderate to high prevalence of the disease.
Currently, no specific vaccines against B. ovis infection are available, but the S live attenuated B. melitensis Rev1 vaccine, widely used for vaccination against ovine and caprine brucellosis caused by B. melitensis, is also effective against B. ovis [4]. However, Rev1 is virulent in humans, induces abortions when used in pregnant animals [4] and is resistant to streptomycin, an antibiotic of choice for brucellosis treatment [5]. While these problems can be solved by using appropriate vaccination strategies and biosafety precautions [4,6], Rev1 also induces a strong antibody response to the O-PS section of S-LPS [7] thus hampering differentiation between true infected and vaccinated animals (DIVA problem) in the routine diagnosis of B. melitensis. Because of this, Rev1 is banned in the countries where B. melitensis has been eradicated, resulting in a subsequent increase in B. ovis infections in sheep. Since B. ovis is naturally R, some attempts to circumvent the problems associated with Rev1 vaccination in B. melitensis-free areas have been based on the use of B. ovis. Accordingly, some investigations have examined subcellular vaccines carrying B. ovis fractions rich in envelope components in lipid-muramyl dipeptide or nanoparticle adjuvants [8,9]. However, these formulations either provide less protection than Rev1 or become too costly. Also, B. ovis mutants in LPS core genes [10] [11,12]. Yet, industrial production of these vaccines would require solving the problem posed by B. ovis CO 2 -dependence [13] with the subsequent reassessment of their properties. Moreover the ABC transporter mutant requires encapsulation [12]. R mutants of S Brucella species (i.e., the so-called R vaccines) are more easily produced and, as they lack the O-PS, are often assumed to solve the Rev1 DIVA problem. However, R vaccines still interfere in S-LPS ELISA [14][15][16] because of the cross-reactivity with the core epitopes shared by the S and R-LPS or, in the wzm/wzt and related B. melitensis 115 spontaneous R mutants, presence of a cytoplasm O-PS precursor [17][18][19]. Another approach was to delete protein BP26. However, whereas the BP26-deleted Rev1 provides protection against B. ovis, the ancillary BP26 ELISA lacks adequate diagnostic sensitivity [20,21]. More recently, a Rev1 construct expressing the green fluorescent protein as a tagging antigen has been proposed. However, this method requires the simultaneous injection of this protein and a booster injection to trigger antibody persisting in the ancillary ELISA-green fluorescent protein DIVA test [22].
The aim of this research is to investigate an alternative to circumvent the serological diagnosis problems caused by the Rev1 vaccine while keeping its good attenuation and protection characteristics. Our approach differs from those summarized above in that, instead of deleting S Brucella antigens or epitopes or introducing a foreign antigen, we have modified a Brucella immunodominant antigen. For this purpose, we applied to Rev1 the strategy proposed by Martínez-Gómez et al. [23] to modify the epitopic structure of Brucella S-LPS by substituting the N-formyl-perosamine of the O-PS by N-acetyl-perosamine. We present here the experiments carried out in the mouse model as a first step to investigate the validity of this approach.

DNA manipulations
Plasmid and chromosomal DNA were extracted with Q1Aprep ® spin Miniprep Kit (Qiagen GmbH, Hilden, Germany) and Ultraclean Microbial DNA Isolation Kit (Mo Bio Laboratories), respectively. When needed, DNA was purified from agarose gels using a QIAquick Gel extraction kit (Qiagen). DNA sequencing was performed by "Servicio de Secuenciación del Centro de Investigación Médica Aplicada" (Pamplona, Spain). Primers were synthesized by Sigma-Genosys Ltd. (Haverhill, United Kingdom).

Construction of mutants
For the construction of Rev1::Tn7wbdRKm R mutants, we used the plasmid pYRI-27 (pUC18R6KT-miniTn7T-Km-PwbdR) described in Martínez-Gómez et al. [23]. The acquisition of this vector by Brucella after tetra-parental mating with conjugative E. coli S17.1 λpir and E. coli HB101 (pRK2013) and E. coli SM10 λpir (pTNS2) was selected by Km and Pmx resistance. The correct insertion and orientation of the miniTn7 carrying wbdR was examined by PCR as previously described [23].
To obtain a wbdR construct with no Km resistance (Rev1::Tn7wbdR), we used the plasmid pRCI-65 (pNPTS138Cm R ΔKm) described in Martínez-Gómez et al. [23]. This suicide plasmid containing the KmR deletion allele was transformed into E. coli β2150, a diaminopimelic acid (DAP) auxotrophic donor strain, to avoid the use of antibiotic during the conjugation process [24] and transferred into Rev1::Tn7wbdRKm R by conjugation. The integration of the suicide vector and disruption of the target gene were selected by Km sensitivity and confirmed by PCR using oligonucleotides KmR-F1 and KmR-R4 (Additional file 2).
Rev1::Tn7wbdRΔwbkC was constructed in a similar way. The suicide plasmid pYRI-31 (pJQKmΔwbkC) [23] was used to delete the wbkC gene of Rev1::Tn7wbdR by allelic change. The resulting mutator plasmid was introduced in Rev1::Tn7wbdR by conjugation using the E. coli β2150. The loss of the plasmid concomitant with the deletion of wbkC gene was checked by PCR using oligonucleotides wbkC-F1 and wbkC-R4 (Additional file 2).
The strain B. ovis PA-KmR, used as challenge in mouse experiments (see below), was obtained using the modified miniTn7 site-specific integration vector technology [25,26].

Characterization of the mutants
Mutants were characterized by the standard Brucella typing procedures described in Alton et al. [27]: colonial morphology, urease, susceptibility to thionine blue, fuchsine and safranine, acriflavine agglutination, crystal violet dye exclusion test (S/R colony morphology), agglutination with anti-A and anti-M monospecific sera, and sensitivity to phages. For co-agglutination, bacteria resuspended in 25 µL of saline on a glass slide were mixed with an equal amount staphylococci sensitized with anti-Formyl-Acetyl, anti-Acetyl and anti-Formyl sera, as previously described [23].

LPS extraction
LPS from Rev1 and Rev1 wbdR tagged was obtained by the proteinase-K sodium dodecyl sulfate (SDS) protocol [28,29] with some modifications. Briefly, cells inactivated with phenol were suspended in 2% SDS-60 mM Tris-HCl buffer (pH 6.8), heated at 100 °C for 10 min and treated with proteinase K (60 µL of a 2.5 mg/mL stock per each mL of suspension, 3 h at 55 °C). This was followed by two consecutive precipitations with 3 volumes of methanol with 1% sodium acetate-saturated methanol at −20 °C. After centrifugation, pellets were resuspended by sonication in 3 mL of 60 mM HCl-Tris (pH 6.8), digested with nucleases and treated again with proteinase K (3 h at 55 °C). After a third precipitation, the pellet containing LPS was recovered in 1 mL of distilled water and frozen.

SDS-PAGE and Western blots
Samples were mixed 1:1 with Sample buffer 2x (Bio-Rad), heated at 100 °C for 10 min, and analyzed in 15% polyacrylamide gels in Tris-HCl-glycine and stained by the periodate-alkaline silver method [30]. For Western blots, LPS were analyzed in SDS-PAGE gel (12% polyacrylamide) (see above) and electrotransferred onto nitrocellulose blotting sheets (Amersham-GE Healthcare Life Scientific, Germany; 0.45 µm pore size). The polyclonal sera used were the following: Anti-S Brucella serum (polyclonal serum from a rabbit infected with B. melitensis 16 M and bled at day 45), Formyl-Acetyl serum and Acetyl serum. These last two sera were obtained by the method described in Martínez-Gómez et al. [23].

Multiplex PCR
To discriminate between Rev1 wbdR tagged strains and all known Brucella species and vaccine strains, we designed a pair of primers based on the wbdR gene and added them to the 8 pairs of primers used in the Bruce-ladder assay [31]. These new primers consisted of 5´-TGA TGT TTT GGC AGG AAA GA-3′ (wbdR forward) and 5′-TAG CCC CAG GAG CAA ATG TA-3′ (wbdR reverse). We designed them with a Tm similar to that of the other primers used in the multiplex PCR and amplifying a band of 347 bp that did not prevent the visualization of the other specific bands of species or vaccine strains.

Studies in mice: virulence and protection
Seven-week-old female BALB/C mice (ENVIGO, Harlan) were kept in cages in BSL-3 facilities (ES/31-2010-000132) with water and food al libitum. Procedures were in accordance with the current European (directive 86/609/EEC) and Spanish (RD 53/2013) legislations, supervised by the corresponding Ethical Committee for Animal Experimentation and authorised by Aragón (reports No. 2014-20 and 2014-21) and Navarra (CEEA 045/12) Governments. For virulence assessment, 10-mice groups were inoculated intraperitoneally (IP) with 1 × 10 5 or 1 × 10 8 colony forming units (CFU) of the corresponding strain and mean CFU values per spleen were determined at 1 (n = 5) and 5 (n = 5) weeks after inoculation as described elsewhere [32]. To evaluate the protective efficacy of Rev1::Tn7wbdRΔwbkC vaccine, 5-mice groups were vaccinated subcutaneously (SC) with two different doses (1 × 10 5 and 1 × 10 8 CFU/ mouse). Mice (n = 5) inoculated SC with the Rev1 reference vaccine (1 × 10 5 CFU/mouse) or sterile Buffered Saline (BSS; 0.85% NaCl, 0.1% KH 2 PO 4 , 0.2% K 2 HPO 4 ; pH 6.85) were used as effective-vaccine and unvaccinated controls, respectively. Four weeks after vaccination, mice were challenged IP with 5 × 10 6 CFU/mouse of B. ovis PA-KmR and the mean CFU/spleen values of this strain were determined 2 weeks after. Inocula were prepared by harvesting the BAB or BAB-S grown bacteria in sterile BSS, adjusting spectrophotometrically (600 nm) the bacterial suspension and making proper serial tenfold dilutions. Mice were inoculated with 0.1 mL and doses were retrospectively assessed by plating inocula countable dilutions. Differentiation between challenge and vaccine strains was achieved by plating on BAB supplemented with streptomycin and BAB-S supplemented with Km (incubated in CO 2 atmosphere). Results were expressed as the mean log 10 CFU/spleen ± SD (n = 5) of the corresponding mutant or challenge strain and units of protection were calculated by subtracting the mean log 10 CFU values of the vaccine group from those of unvaccinated controls. Statistical comparisons were made by one-way ANOVA and Fisher's Protected Least Significant Differences (PLSD) post hoc tests. The adequate virulence of B. ovis PA-KmR challenge strain was proved in a previous experiment in BALB/c mice by IP infection (5 × 10 5 CFU/mouse) and bacterial spleen counting (in BAB-S and BAB-S supplemented with Km) 3 and 8 weeks later. This strain showed identical multiplication to that of B. ovis PA (not shown).

Insertion of wbdR into Rev1 genome modifies the epitopic structure of the vaccine
To obtain a tagged Rev1 vaccine, we followed two strategies. First, we obtained Rev1::Tn7wbdR inserting wbdR in the chromosome II of Rev1 vaccine using the previously described Tn7 methodology [8]. Second, we constructed a wbdR tagged Rev1 lacking N-formyl-perosamine in its O-PS (Rev1::Tn7wbdRΔwbkC) deleting wbkC (the formyl transferase gene) from Rev1::Tn7wbdR. As expected, these constructs could be differentiated from Rev1, vaccine B. abortus S19, or representative strains of wild-type B. abortus, B. melitensis and B. ovis by PCR using the Multiplex-PCR combining Bruce-ladder and wbdR specific primers (Figure 1).
Rev1::Tn7wbdR and Rev1::Tn7wbdRΔwbkC were identical to the parental strain in colony morphology, growth, oxidase and urease tests and dye sensitivity The LPS extracts of Rev1::Tn7wbdR and Rev1::Tn7wbdRΔwbkC contained the typical R-LPS (i.e. low molecular weight) and S-LPS (higher molecular weight) fractions present in the parental strain (Figure 2). However, as observed before for their wbdR tagged B. melitensis and B. abortus counterparts [23], the S-LPS fraction of the wbdR constructs displayed an apparent average molecular weight lower than those of the parental strains (Figure 2A). To analyze possible epitopic changes associated with wbdR tagging, whole cells and LPS extracts were tested by co-agglutination and Western blot, respectively, using immune sera specifically recognizing O-PS carrying only N-formylperosamine (anti-Formyl), both N-formyl-perosamine and N-acetyl-perosamine (anti-Formyl-Acetyl) or only N-acetyl-perosamine (anti-Acetyl) [23]. Rev1::Tn7wbdR co-agglutinated with staphylococci sensitized with anti-Formyl-Acetyl, anti-Acetyl and anti-Formyl sera showing that its O-PS contained N-acetyl and N-formyl-perosamine. In contrast, Rev1::Tn7wbdRΔwbkC did not co-agglutinate with anti-Formyl serum, while keeping the anti-Formyl-Acetyl and anti-Acetyl reactivity (Additional file 4). Western blots with proteinase-K extracted S-LPS ( Figure 2B) confirmed the epitopic changes and demonstrated that the differences in apparent average molecular weight observed by SDS-PAGE (Figure 2A) corresponded in fact to O-PS heterogeneity.

wbdR tagging decreases the residual virulence of Rev1
We vaccinated BALB/c mice with 10 5 CFU/mouse (standard dose for Brucella S vaccines) of Rev1::Tn7wbdR, Rev1::Tn7wbdRΔwbkC and Rev1 and determined CFU numbers in spleen. One week post-infection, CFU/spleen values were significantly lower in mice infected with Rev1::Tn7wbdR or Rev1::Tn7wbdRΔwbkC than in those infected with Rev1 ( Figure 3A). At week 5, though both Rev1::Tn7wbdR and Rev1::Tn7wbdRΔwbkC also showed lower CFU, only the values of Rev1::Tn7wbdRΔwbkC were statistically different ( Figure 3A). However, attenuation with respect to Rev1 was clearly observed for both tagged vaccines at post-infection weeks 1 and 5 when the experiment was carried out with 10 8 CFU/mouse, the dose recommended for R vaccines ( Figure 3B).

Mice vaccinated with Rev1 wbdR constructs can be discriminated by ELISA
We then analyzed by iELISA the sera of mice vaccinated with 10 8 CFU using first the S-LPS of Ba::Tn7wbdRΔwbkC as antigen. As expected, and no matter the vaccine, no serum from these mice showed ELISA reactivity one week post-infection (not shown). At week 5, whereas most sera of mice inoculated with the wbdR constructs showed some reactivity, sera of mice infected with Rev1 or B. melitensis did not react (Figure 4A). Interestingly, the reverse picture was obtained in the iELISA with wild-type S-LPS as antigen: whereas mice infected with Rev1 or B. melitensis developed reacting antibodies, sera of mice inoculated with wbdR constructs displayed almost no reaction ( Figure 4B) with small but consistently repeated (not shown) differences between Rev1::Tn7wbdR and Rev1::Tn7wbdRΔwbkC.

Rev1::Tn7wbdRΔwbkC protects against B. ovis in the mouse model
Based on the ELISA results and on its lack of reactivity with the anti-Formyl serum (Additional file 4), we chose Rev1::Tn7wbdRΔwbkC to study the efficacy of wbdR tagged vaccines against B. ovis. As shown in Table 1 doses tested. However, only the 10 8 CFU/mouse dose (used for rough vaccines in the mouse model) conferred a protection similar to Rev1 at the standard dose (Table 1). Interestingly, even at this higher dose, the vaccine CFU remaining in the spleens was markedly lower than those of Rev1 (p < 0.01).

Discussion
The results presented in this work demonstrate that the O-PS of vaccine Rev1 can be antigenically tagged by genomic insertion of wbdR, the acetyltransferase gene involved in the synthesis of the O-PS repeating unit of E. coli O157:H7. Using the mouse model, we studied the protective capacity of wbdR tagged Rev1 against B. ovis and tested an ancillary ELISA DIVA test. In previous and preliminary experiments, we examined whether wbdR tagged brucellae (wbdRΔwbkC) could be used as antigen in a simple test similar to the Rose Bengal test for brucellosis. However, we observed that the modified brucellae had a marked tendency to autoagglutinate, a feature that is reminiscent of the autoagglutination typical of R brucellae [16] and shows profound cell surface modifications. For this reason, we ruled out the agglutination test and applied an iELISA using the LPS purified from a wbdRΔwbkC construct [23] adjusted to minimize any cross-reactivity between acety-tagged and native O-PS. In this assay, we observed differences between the antibodies triggered by wbdR tagged vaccines when compared with those triggered by Rev1; whereas sera of mice vaccinated with Rev1::Tn7wbdR or Rev1::Tn7wbdRΔwbkC showed moderate reaction, sera corresponding to the Rev1 vaccinated mice did not react. Moreover, when the iELISA with wild-type S-LPS was used, contrary to the diagnostically interfering reaction showed by Rev1, the sera of mice inoculated with wbdR constructs displayed almost no reaction. In a previous work, we observed that the LPS of Brucella crossreacting E. hermanii serotypes reacts strongly with sera from infected cattle [21,33] which are known to contain an overwhelming majority of antibodies of overlapping C specificities [34]. The O-PS of these E. hermanii strains contains N-acetylated perosamine in an α (1-3) and α (1-2) linkage frequency and arrangement different from those in Rev1 (i.e. B. melitensis biovar 1) and cross-react with Brucella M monoclonal antibodies in gel immunoprecipitation [33]. The O-PS in the wbdRΔwbkC constructs does not react in ELISA with C/Y-A = M or al C/Y-A > M in ELISA [23]. These data suggest epitopic differences that, in addition to differences in the intensity of the stimulus in an experimental infection in mice and a natural infection in cattle, in the subsequent different avidity of the antibodies and in the standardization of the ELISA, could account for these apparently contradictory results. Our results suggest that the use of wbdR tagged Rev1 and its ancillary ELISA might be a suitable strategy to solve the DIVA problem associated with the unmodified vaccine, but it is clear that experiments in sheep are necessary for a definite assessment.
In keeping with previous studies in virulent brucellae [23], the most profound modification was obtained in Rev1::Tn7wbdRΔwbkC and, accordingly, we tested its protective capacity in the mouse model. We observed that a 10 8 CFU/mouse dose was necessary to reach the protection levels obtained with Rev1 at 10 5 CFU/mouse. The recommended dose for brucellosis R vaccines is also 10 8 CFU/mouse [35] and, like the autoagglutination commented above this suggests surface similarities between the Rev1::Tn7wbdRΔwbkC construct and R brucellae. The lack of O-PS in the latter increases the exposure of lipid A-core and Omp charged groups and it has been postulated that the OP-S acts as a negative modulator of unspecific interactions of these bacteria with host cells, thus favoring invasion for specific routes [18]. On these bases, it can be hypothesized that either the N-acetylperosamine O-PS cannot replace the N-formyl-perosamine with regards to surface physicochemical properties and/or that the shorter O-PS of Rev1::Tn7wbdRΔwbkC constructs does not provide enough steric hindrance to the Omps and the lipid A-core section, as suggested by its sensitivity to the R/C phage. These surface changes may account for our observation that, when tested for the residual virulence (Figure 3), we found that the CFU counts/spleen of Rev1::Tn7wbdRΔwbkC were lower than those of Rev1 even at a dose one thousand times higher, a result in agreement with the low numbers of CFU remaining in the spleens in the vaccination experiments (Table 1). Taken together, these results in mice suggest that Rev1::Tn7wbdRΔwbkC could be not only a DIVA vaccine but also a vaccine with reduced residual virulence in the host. Research is in progress to evaluate Statistical comparison (n = 5) of mean log 10 B. ovis CFU/spleen: a p < 0.001 vs. PBS (unvaccinated), b p > 0.05 vs. Rev1 and log 10 vaccine CFU/spleen c p < 0.01 vs. Rev1.

Vaccine (dose)
Mean ± SD of log 10 CFU in spleen of: