Skip to main content
Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

The Differential Interaction of Brucella and Ochrobactrum with Innate Immunity Reveals Traits Related to the Evolution of Stealthy Pathogens

  • Elías Barquero-Calvo ,

    Contributed equally to this work with: Elías Barquero-Calvo, Raquel Conde-Alvarez

    Affiliation Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica

  • Raquel Conde-Alvarez ,

    Contributed equally to this work with: Elías Barquero-Calvo, Raquel Conde-Alvarez

    Affiliation Department of Microbiology, University of Navarra, Navarra, Spain

  • Carlos Chacón-Díaz,

    Affiliations Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica, Centro de Investigación en Enfermedades Tropicales, Facultad de Microbiología, Universidad de Costa Rica, San José, Costa Rica

  • Lucía Quesada-Lobo,

    Affiliation Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica

  • Anna Martirosyan,

    Affiliation Centre d'Immunologie de Marseille-Luminy, Aix Marseille Université, Faculté de Sciences de Luminy, INSERM U631, CNRS UMR6102, Marseille, France

  • Caterina Guzmán-Verri,

    Affiliation Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica

  • Maite Iriarte,

    Affiliation Department of Microbiology, University of Navarra, Navarra, Spain

  • Mateja Mancek-Keber,

    Affiliation Department of Biotechnology, National Institute of Chemistry, Hajdrihova, Ljubljana, Slovenia

  • Roman Jerala,

    Affiliation Department of Biotechnology, National Institute of Chemistry, Hajdrihova, Ljubljana, Slovenia

  • Jean Pierre Gorvel,

    Affiliation Centre d'Immunologie de Marseille-Luminy, Aix Marseille Université, Faculté de Sciences de Luminy, INSERM U631, CNRS UMR6102, Marseille, France

  • Ignacio Moriyón,

    Affiliation Department of Microbiology, University of Navarra, Navarra, Spain

  • Edgardo Moreno , (EM); (ECO)

    Affiliation Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica

  • Esteban Chaves-Olarte (EM); (ECO)

    Affiliations Programa de Investigación en Enfermedades Tropicales, Escuela de Medicina Veterinaria, Universidad Nacional, Heredia, Costa Rica, Centro de Investigación en Enfermedades Tropicales, Facultad de Microbiología, Universidad de Costa Rica, San José, Costa Rica



During evolution, innate immunity has been tuned to recognize pathogen-associated molecular patterns. However, some α-Proteobacteria are stealthy intracellular pathogens not readily detected by this system. Brucella members follow this strategy and are highly virulent, but other Brucellaceae like Ochrobactrum are rhizosphere inhabitants and only opportunistic pathogens. To gain insight into the emergence of the stealthy strategy, we compared these two phylogenetically close but biologically divergent bacteria.

Methodology/Principal Findings

In contrast to Brucella abortus, Ochrobactrum anthropi did not replicate within professional and non-professional phagocytes and, whereas neutrophils had a limited action on B. abortus, they were essential to control O. anthropi infections. O. anthropi triggered proinflammatory responses markedly lower than Salmonella enterica but higher than B. abortus. In macrophages and dendritic cells, the corresponding lipopolysaccharides reproduced these grades of activation, and binding of O. anthropi lipopolysaccharide to the TLR4 co-receptor MD-2 and NF-κB induction laid between those of B. abortus and enteric bacteria lipopolysaccharides. These differences correlate with reported variations in lipopolysaccharide core sugars, sensitivity to bactericidal peptides and outer membrane permeability.


The results suggest that Brucellaceae ancestors carried molecules not readily recognized by innate immunity, so that non-drastic variations led to the emergence of stealthy intracellular parasites. They also suggest that some critical envelope properties, like selective permeability, are profoundly altered upon modification of pathogen-associated molecular patterns, and that this represents a further adaptation to the host. It is proposed that this adaptive trend is relevant in other intracellular α-Proteobacteria like Bartonella, Rickettsia, Anaplasma, Ehrlichia and Wolbachia.


The class α-Proteobacteria includes microorganisms capable of establishing close associations with eukaryotic cells in a wide range of lifestyles. Members of the genus Agrobacterium are pericellular to plant cells and induce tumors, whereas Brucella, Bartonella, Phyllobacterium and Sinorhizobium are facultative extracellular-intracellular bacteria that behave as pathogens or endosymbionts, and Rickettsia, Anaplasma, Ehrlichia and Wolbachia are obligate intracellular pathogens of mammals and arthropods [1], [2]. Remarkably, animal pathogens of this group have the ability to avoid immediate recognition by innate immunity, thus following a stealthy strategy of which Brucella can be considered as a model [3]. In contrast, some bacteria close to Brucella are free living environmental microorganisms like Ochrobactrum, Daeguia, Crabtreella and Mycoplana. Although these four genera are included in the Brucellaceae (, only Ochrobactrum has been reported to display some degree of pathogenicity. Ochrobactrum anthropi, primarily a rhizosphere inhabitant, has been isolated from immunocompromised individuals or patients subjected to dialysis, catheterization, surgical interventions or transplantation [4], [5], [6], [7], [8], [9], [10], [11] and often shows a broad antibiotic resistance [12], [13]. Ochrobactrum intermedium is another opportunistic member of the genus. Interestingly, Ochrobactrum is the Brucellaceae member closest to Brucella according to several molecular markers and genome comparisons [14].

Owing to its close phylogenetic relatedness with the highly virulent Brucella [15], Ochrobactrum has received some attention. Phenotypic analysis reveals that O. anthropi displays envelope molecules known to be critical in Brucella virulence. They include phosphatidylcholine and a lipopolysaccharide (LPS) with a lipid A carrying very long chain fatty acids (VLCFA). Also, like other Brucellaceae, O. anthropi free lipids contain acyl chains with a higher number of carbons than those found in typical Gram negative bacteria. Along with the VLCFAs of lipid A, these structural features are thought to be relevant in the construction of a firm envelope which, in the case of Brucella, has been proposed to be important in pathogenicity [16], [17]. Despite these similarities, there are differences in important outer membrane properties. Most notably, while Brucella is highly permeable to hydrophobic substances and resistant to bactericidal cationic peptides, Ochrobactrum is impermeable and sensitive, albeit not to the same extent as typical Gram negative bacteria [18]. Interestingly, these differences have been correlated with some structural variations in the LPSs of these bacteria [18]. LPS typically bears a marked pathogen-associated molecular pattern (PAMP) and is thus a key target of innate immunity [19]. Indeed, the altered PAMP of Brucella LPS (BaLPS) greatly contributes to the stealthy strategy and virulence of this pathogen [3], [20]. All this evidence leads to the hypothesis that the Brucellaceae ancestors carried such molecules that without extensive remodeling resulted in features adequate to evade innate immunity [3].

To examine this possibility, we compared O. anthropi with Brucella abortus. We found that these bacteria display a significant divergence in the interaction with host cells and the innate immune system, and traced some of the differences to changes in orthologs of key Brucella virulence determinants. The results not only help to understand the widely different degrees of pathogenicity of these bacteria but also how intracellular pathogens following a stealthy strategy may have emerged in the α-Proteobacteria.


O. anthropi does not multiply intracellularly

We first compared the replication of O. anthropi and B. abortus in mice after intraperitoneal injection, a route commonly used in virulence studies [21]. As expected, B. abortus reached high numbers one week after infection and maintained those numbers throughout the experiment, reflecting the chronic nature of brucellosis (Fig. 1A). O. anthropi, on the other hand, was recovered in four orders of magnitude and CFU numbers diminished throughout the experiment (Fig. 1A). However, under the conditions used, clearance was not complete, indicating some ability to persist. Although at 109 O. anthropi bacterial doses not deaths were recorded, piloerection and diarrhea were observed at 48 h after infection, suggesting moderate endotoxicity.

Figure 1. O. anthropi does not replicate in mice or in cells but expresses Brucella-associated virulence factors.

(A) Groups of six mice were inoculated intraperitoneally with 107 (white circles) or 109 (red circles) of O. anthropi or 106 of B. abortus (black circles) CFU/mouse. At the indicated times the CFUs per spleen were calculated. (B) HeLa cells treated with CNF for 2 h or non-treated were infected with O. anthropi or B. abortus at a MOI of 500. Extracellular bacteria were killed with gentamicin (100 µg/ml). At the indicated times intracellular CFU were determined. (C) Raw 264.7 macrophages were infected with the indicated bacteria at a MOI of 500 and processed as in “B”. (D) Lysates of O. anthropi (Oa) and B. abortus (Ba) were probed with anti-VirB8 and anti-BvrR antibodies. Genomic DNA (1) or cDNA (2) from O. anthropi was used as template for PCR amplification of the indicated genes. No amplification from total O. anthropi RNA preparation (minus RT) was observed (not shown). All values of O. anthrophi in A, B and C were significantly different at p<0.001 with respect to B. abortus.

We then studied the ability of O. anthropi to invade and multiply in cells. For this, we first infected HeLa cells with O. anthropi or B. abortus and killed extracellular bacteria with gentamicin. Even at early times, O. anthropi was recovered in three orders of magnitude less than B. abortus, thus showing its inability to induce internalization (Fig. 1B). This was confirmed by intracellular-extracellular differential fluorescence, which showed that less that 0.1% of HeLa cells interacted with O. anthropi and that most of the visualized bacteria were located extracellularly. Moreover, whereas the CFU numbers of intracellular B. abortus increased throughout the experiment, those of O. anthropi declined steadily (Fig. 1B). In order to bypass the internalization deficiency and to determine if O. anthropi was able to multiply intracellularly, we used the Escherichia coli Cytotoxic Necrotizing Factor, a toxin that confers a phagocytic phenotype to HeLa cells [22]. This treatment increased the intracellular O. anthropi in one order of magnitude (Fig. 1B) and most visualized bacteria were intracellular. Nevertheless, O. anthropi was not able to achieve sustained replication, even though some survival was observed (Fig. 1B). Then, we repeated the experiments in Raw 264.7 murine macrophages. Although at early times the intracellular numbers of O. anthropi paralleled those of B. abortus, the replication curves coursed with opposite tendencies because, whereas B. abortus replicated, O. anthropi did not (Fig. 1C). Noteworthy, the later bacteria remained cultivable 48 h after infection (Fig. 1C). Altogether, these results indicate that O. anthropi possesses a very limited ability to invade and multiply intracellularly in cells.

O. anthropi expresses orthologs of Brucella virulence determinants

We have shown that the cell envelopes of Brucella play a critical role in virulence [23], [24]. In part, this is due to some unusual physicochemical properties linked to peculiar lipids (i.e. phospholipids, ornithine lipids, LPS lipid A and lipoproteins) and to the properties of the periplasmic cyclic β-glucans. Moreover, the overall outer membrane protein profile [25] and some lipids [26] (Conde-Álvarez et al, unpublished) are under the control of the BvrR/BvrS system, a critical Brucella virulence regulator. Also, the envelope is the place where the type IV secretion system VirB is expressed. To contrast these Brucella features with those of O. anthropi, we first looked for the corresponding orthologs in genomic databases. O. anthropi carried orthologs of Brucella genes known to be involved in the synthesis of the periplasmic cyclic glucans, ornithine lipids, and phosphatidylcholine. Flagellin genes (coding for a PAMP-bearing protein) were present in O. anthropi and B. abortus, even though both predicted proteins are modified in the consensus sequence recognized by TLR5 [27]. In addition, B. abortus lipoprotein genes omp19, omp16 and omp10 had orthologs in O. anthropi. Other B. abortus putative lipoprotein ORFs were also present in O. anthropi, although the one encoded by open reading frame (ORF) BAB1_1548 carried a significant deletion encompassing 29 amino acids as compared to its ortholog. However, the converse was not true since orthologs of the putative lipoproteins corresponding to ORFs Oant_3560, Oant_1600, and Oant_2497 were not found in B. abortus (Table 1), suggesting a more simple lipoprotein profile in the cell envelope of the latter bacteria. Putative orthologous of genes coding for specialized acyltransferases and acylcarrier proteins (LpxXL and AcpXL) necessary to synthesize and incorporate VLFACs to lipid A were also present.

Table 1. Genomic comparison of Brucella 2308 and O. anthropi LMG3331 ORFs involved in the synthesis of surface exposed or OM PAMPs.

We then examined the presence and expression of the two component sensory-regulatory system BvrR/BvrS in O. anthropi. The predicted proteins of the consecutive ORFs Oant_0827 and Oant_0828 show 90.5% and 97.5% identity with BAB1_2092 (B. abortus BvrR) and BAB1_2093 (B. abortus BvrS), respectively. Primers designed on these sequences gave positive amplification reactions from both O. anthropi genomic DNA and cDNA, thus demonstrating transcription (Fig. 1D). In addition, an antibody against B. abortus BvrR readily detected a single band in O. anthropi lysates with a molecular weight consistent with that of the predicted protein (Fig. 1D).

Finally, we searched for orthologs of the type IV secretion system VirB. O. anthropi genome contains two DNA regions with an organization and identity compatible with the presence of a virB operon. The first one contains twelve consecutive ORFs (from Oant_674 to Oant_685) with an organization similar to that of the virB operon in B. abortus. With the exception of Oant_0679, which encodes a hypothetical protein of unknown function, all other ORFs have orthologs in the B. abortus virB operon, with percentage identities between the homologous predicted proteins ranging from 22.9% to 45.5%. In contrast, the second DNA region contains only seven ORFs (Oant_4564 to Oant_4562, Oant_4559 to Oant_4557 and Oant_4555) and the predicted proteins show a lower identity (from 18.1% to 35.1%) with the corresponding B. abortus VirB products. Primers designed on the sequences of the first DNA region for the virB4 (Oant_0677), virB5 (Oant_0678) and virB8 (Oant_0682) orthologs gave a positive amplification from O. anthropi genomic DNA and cDNA indicating their expression (Fig. 1D). Moreover, while the BvrR orthologous protein strongly reacted in Western blots, antibodies specific for B. abortus VirB8 showed only a weak band of the appropriate molecular weight in O. anthropi lysates (Fig. 1D). These results demonstrate the closer proximity between the BvrR/BvrS orthologs and indicate that in the Brucellaceae there is higher degree of divergence in the type IV secretion VirB systems committed to association with eukaryotic cells than in the BvrR/BvrS systems devoted to cell envelope homeostasis.

O. anthropi triggers leukocyte recruitment that lay between those induced by B. abortus and S. enterica

The initial phases of B. abortus infection are characterized by a negligible activation of proinflammatory mechanisms which contrasts with those triggered by most Gram negative pathogens [3]. We first tested to what extent this property was shared with O. anthropi. As shown in Fig. 2A, O. anthropi induced recruitment of leukocytes into the peritoneal cavity of mice, with more than a twofold increase in 24 hours. The analysis of these leukocytes indicated a rapid switch, with neutrophils increasing from 3% to more than 40% at two hours after intraperitoneal inoculation (Fig. 2B).

Figure 2. O. anthropi recruits leukocytes in the peritoneum.

Groups of 6 mice were intraperitoneally inoculated with 106 CFU/ml of O. anthropi. Thereafter, peritoneal exudates were obtained and leukocytes counted. (A) Total number of peritoneum leukocytes (B) Relative number of mononuclear (monocytes and lymphocytes) and polymorphonuclear neutrophil leukocytes.

We also monitored the changes in blood after peritoneal inoculation in comparison with B. abortus and S. enterica. Although there was a quick initial decrease in the circulating leukocytes in the three cases, numbers remained low in mice inoculated with S. enterica. In contrast, there was a rapid return to basal levels in B. abortus inoculated mice and an intermediate phenotype, slightly closer to that of B. abortus, obtained with O. anthropi (Fig. 3A). The evaluation of the blood leukocyte populations showed a moderate change in the proportion of neutrophils in mice inoculated with B. abortus (Fig. 3B) and a rapid increase of the same cells in mice inoculated with S. enterica (Fig. 3C). Again, O. anthropi triggered an intermediate response (Fig. 3D).

Figure 3. O. anthropi infection induces an intermediate blood leukocyte response.

Groups of 6 mice were inoculated intraperitoneally with106 CFU/ml of the indicated bacteria, bleed at different times and leukocytes counted in blood. (A) Total leukocytes number of mice infected with the indicated bacteria. (B) Relative blood leukocyte number of mice infected with B. abortus. (C) Relative blood leukocyte number of mice infected with S. enterica. (D) Relative blood leukocyte number of mice infected with O. anthropi. Values of O. anthropi in “A” were significantly different at p<0.01 (*) and p<0.001 (**) with respect to B. abortus. In “B, C and D” the standard error was less that 5% at all points.

Neutrophils control O. anthropi but not B. abortus infections

The above results indicated that O. anthropi induces a cellular response compatible with acute bacterial infections and suggested that neutrophils could play a protective role. In addition, we previously observed that neutrophils do not play a significant role in B. abortus infections in mice [3]. Thus, we evaluated the relevance of these leukocytes in controlling O. anthropi using chronic neutropenic mice. Forty-eight hours after infection, all the neutropenic but none of the non-neutropenic mice were dead (Fig. 4A). The role of neutrophils in controlling O. anthropi was confirmed by the results of ex-vivo experiments with rat neutrophils (Fig. 4B).

Figure 4. Neutrophils are required for the control of O. anthropi infections.

(A) Neutrophils were depleted from twenty mice by injection of the anti-neutrophil RB6 antibody and ten additional mice were instead injected with PBS alone. One group of ten mice anti-neutrophil treated and the group of ten mice injected with PBS alone were then infected intraperitoneally with 109 CFU/mouse of O. anthropi. The last group of ten mice treated with antibody anti-neutrophil was inoculated with PBS alone. The lethality of the bacteria was recorded at the indicated times. (B) O. anthropi was mixed with purified rat neutrophils at a ratio of 5±4 bacteria/cell. Control bacteria were not incubated with neutrophils. At the indicated times the viable CFU were determined and the percentage of bacterial replication was calculated. In “B” Values of p<0.01 (*) and p<0.001 (**) are indicated.

O. anthropi activates complement through non-LPS determinants

Both Brucella cells and LPS are very poor complement activators [3]. To assess whether O. anthropi shares this property, we incubated rabbit serum with B. abortus and O. anthropi and then measured the lytic activity remaining in the supernatants using antibody-sensitized erythrocytes. Whereas B. abortus practically did not consume complement, O. anthropi completely abrogated the lytic activity, demonstrating an effective surface binding and the subsequent activation (Fig. 5A). Next, we tested the bactericidal activity of bovine serum against these bacteria. In agreement with the high avidity of the complement for the O. anthropi surface, this bacterium was considerable more susceptible to the bactericidal action of serum than B. abortus (Fig. 5B). Since LPS is one of the main molecules responsible for complement activation by Gram negative cell surfaces, we tested whether the O. anthropi LPS (OaLPS) consumed complement. Surprisingly, OaLPS was less effective than BaLPS in depleting complement activity, even both were less effective than the LPS from S. enterica (SeLPS) (Fig. 5C). Therefore, other O. anthropi surface molecules, such as capsule polysaccharides, may be responsible for binding complement.

Figure 5. O. anthropi cells but not its LPS fix complement.

(A) Rabbit serum was incubated with the indicated amounts of O. anthropi or B. abortus cells. After incubation, the ability of the serum to lyze pre-sensitized erythrocytes was determined and expressed as percentage of hemolysis. (B) The indicated bacteria were incubated in the presence of bovine serum for 90 min and their relative resistance was calculated.(C) Rabbit serum was incubated with the indicated concentrations of OaLPS, BaLPS or SeLPS. After incubation the ability of the serum to lyse pre sensitized erythrocytes was determined. Values of O. anthropi in were significant different at p<0.001 (**) with respect to B. abortus.

O. anthropi and its LPS trigger cytokine responses that lay between those of B. abortus and S. enterica

Proinflammatory cytokines, particularly TNF-α, increase during acute Gram negative infections but not at the onset of brucellosis. When we studied this in comparative terms, we found that murine macrophages infected with O. anthropi released significantly higher amounts of TNF-α than cells infected with B. abortus (Fig. 6A), and this observation held true in vivo since mice displayed higher TNF-α blood levels when inoculated with O. anthropi than with B. abortus (Fig. 6B). O. anthropi, however, was not as efficient as S. enterica in the induction of TNF-α release in vivo (Fig. 6B).

Figure 6. O. anthropi induces TNF-α that lays between B. abortus and S. enterica.

(A) Murine Raw 264.7 macrophages were infected for 30 min with O. anthropi or B. abortus at a MOI of 500. Extracellular bacteria were killed by addition of gentamicin. At the indicated times samples from the culture media were taken and processed for TNF-α quantification. (B) Groups of 6 mice were inoculated intraperitoneally with O. anthropi. At the indicated times mice were bled and TNF-α levels were determined. The levels of TNF-α induction after 3 h of an equivalent inoculation of B. abortus or S. enterica. Value of p<0.001 (**) is indicated.

Since LPS is the foremost PAMP-bearing molecule of Gram negative bacteria, we compared the ability of OaLPS, BaLPS, and E. coli LPS (EcLPS) to trigger cytokine responses in mice. OaLPS consistently induced higher blood levels of TNF-α, IL-1 and IL-10 than BaLPS (Fig. 7A–C) but lower than those induced by EcLPS, even if the latter was used at ten times less concentration (Fig. 7A–C). Consistent with these results, OaLPS induced a higher cytokine release in cultured macrophages than BaLPS (Fig. 7D insert). The induction of TNF-α release by OaLPS was TLR4- but not TLR2-dependent, because only macrophages derived from TLR4-knockout mice failed to release this cytokine (Fig. 7D). This pattern was similar to the induction of TNF-α by EcLPS, which also depended on TLR4 and not on TLR2 (Fig. 7D). Under the conditions used, the TNF-α levels produced by macrophages stimulated with BaLPS were significantly lower in both TLR2- and TLR4-deficient cells (Fig. 7D).

Figure 7. OaLPS induces cytokine responses that lay between BaLPS. and SeLPS.

(A, B and C) Groups of 6 mice were inoculated intraperitoneally with BaLPS, EcLPS or OaLPS at the indicated concentrations. Mice were bled at 2 or 5 h and interleukins quantitated by ELISA. (D) Bone marrow macrophages from wild type TLR2−/− and TLR4−/− knockout mice were incubated with the indicated LPSs. After incubation TNF-α was quantified in the supernatant of cells at 24 h. The insert shows Raw 264.7 macrophages incubated with 25 µg of OaLPS or BaLPS and at the indicated times TNF-α quantified from cell culture supernatants. Values of p<0.05 (*) and p<0.005 (**) with respect to BaLPS are indicated.

OaLPS partially induces dendritic cell maturation

Dendritic cells are key elements linking innate and adaptive immunity. Since the evidence presented above situates the induction of innate immunity by O. anthropi between those elicited by enteric bacteria and B. abortus, we wanted to know whether this was also true in these cells. To this purpose, we measured the ability of the respective LPSs to trigger the formation of the dendritic cell aggregosome-like induced structures (DALIS) characteristic of the maturation of these cells. As expected, EcLPS induced more DALIS than BaLPS. Moreover, in keeping with all the above-described results, OaLPS displayed an activity that lay between BaLPS and EcLPS (Fig. 8A). To confirm this, we measured the expression of maturation markers in LPS-treated dendritic cells. As compared to BaLPS, OaLPS induced a significantly higher expression of surface CD40, CD80, CD86 and MHC-II which was, however, lower than that triggered by EcLPS (Fig. 8B).

Figure 8. OaLPS induces dendritic cell maturation that lay between BaLPS. and SeLPS.

(A) BMDC were incubated with 0.1 µg of EcLPS, 10 µg of BaLPS or 10 µg of OaLPS in 1 ml culture and labeled with the anti-FK2 antibody for detection of DALIS at 24 h. The percentage of BMDC positive for DALIS is indicated. (B) BMDC incubated as in “A” were labeled with fluorescent anti-CD40, CD-80, CD-86 and MHC II. Values of median fluorescence correspond to four independent experiments. Values of p< 0.05 (*) and p<0.005 (**) with respect to BaLPS are indicated.

OaLPS partially activates NF-κB and its affinity for MD-2 lies between that of BaLPS and SeLPS

Most biological effects of LPS depend on its interaction with the TLR4 co-receptor MD-2, an event that triggers a cascade of signals leading to the NF-κB-dependent activation of immune response genes. On the basis of the above-described results, we hypothesized that OaLPS would have a higher binding to MD-2 than BaLPS but lower than SeLPS, and that this would lead to an NF-κB activation between both types of LPSs. To test this possibility, we compared first the ability of OaLPS, BaLPS, and SeLPS to displace the hydrophobic probe bis-ANS from the binding site of human MD-2 (hMD2) [28]. We found that SeLPS was able to displace approximately 30% of bis-ANS from the hMD-2-binding site at concentrations from 1.25–5 µg/ml (Fig. 9A). Displacement by BaLPS, on the other hand, remained close to the background values obtained with an equal amount of water (Fig. 9A). At the highest concentration tested, OaLPS caused a displacement similar to that induced by SeLPS. However, at lower concentrations it approached the negligible values of BaLPS (Fig. 9A). The interaction with hMD-2 was also measured as the ability of the LPSs to block the recognition of hMD2 by anti-hMD-2 monoclonal antibody (Fig. 9B). SeLPS inhibited this antibody in a dose-dependent manner, with a 50% reduction in the signal at a concentration of 2 µg/ml. BaLPS did not reduce significantly the binding of the anti-hMD-2 antibody at any of the concentrations tested. Again, OaLPS showed an activity that laid between BaLPS and SeLPS since it blocked the antibody at concentrations higher than those of SeLPS (Fig. 9B). Finally, we assessed the activation of the terminal end of the signaling pathway by measuring the relative activity of a luciferase reporter under the control of the NF-κB promoter in HEK 293 cells transfected with expression vectors containing hMD-2, human TLR4 (hTLR4) and human CD14 (hCD14). As can be seen in Fig. 9C, 10 to 25 times higher concentrations of OaLPS were necessary to generate a relative luciferase activity similar to that induced by 1 µg/ml of SeLPS. Consistent with other observations, BaLPS induced a weaker luciferase activity (Fig. 9C).

Figure 9. OaLPS shows an intermediate affinity for MD-2 and partially activates NF-kB.

(A) Displacement of bis-ANS by LPS from the bis-ANS/hMD2 complex. The bis-ANS/hMD2 complex (50 nM/50 nM) was incubated for 5 minutes to reach stable fluorescence (F0). Then, increasing amounts of the indicated LPS or the same volume of water were added and fluorescence displacement measured. (B) Detection of free hMD2 by ELISA after incubation (0.75 mM of hMD2) with increasing concentrations of the indicated LPS. (C) NF-κB expression measured by relative luciferase activity of HEK293 transfected cells with hTLR4, hMD2, human CD14, NFκB-luc and Renilla containing plasmids upon respective LPS stimulation. The responsiveness to LPS was determined by the NF-κB-responsive luciferase reporter plasmid, which was normalized using constitutively active Renilla luciferase. Value of p<0.005 (**) with respect to BaLPS is indicated. In SeLPS all values beyond 0.1 µg were significant at p<0.005, with respect to BaLPS.


We proposed that B. abortus follows a stealthy strategy which relies on the absence, modification and inaccessibility of PAMP-bearing molecules behind a firm outer membrane [3]. A related hypothesis is that this pathogen has evolved from a free living α-Proteobacteria bacterium with envelope structural features already representing a first step towards innate immunity evasion [3], [18], [29]. According to this, the Brucella ancestor underwent further changes in such a structural scaffold upon contact with innate immunity effectors and, in parallel, reduced its genome, removed plasmids and trimmed-down its metabolic alternatives [30]. The results presented here reinforce these hypotheses in three ways: i), they show that the proinflammatory response to Brucella is significantly lower than that to the closest known neighbor bacteria which, significantly, is only an opportunistic pathogen; ii) they demonstrate the existence of envelope molecules prone to low recognition in Brucellaceae members that are primarily soil inhabitants and; iii) they support the notion that relatively few and not drastic structural changes in a key PAMP-bearing molecule result in very different lifestyles.

The early clinical evolution of human brucellosis is characterized by a negligible pro-inflammatory response [31], a phenomenon reproduced in experimental models [3]. In contrast, O. anthropi infection triggered leukocytosis, neutrophilia, platelet aggregation and acute fever [32], [33], all indicative of sepsis. The relevance of this response in O. anthropi infections is best exemplified by the prominent role played by neutrophils. These cells were bactericidal for O. anthropi and their absence converted an otherwise harmless microbe into a lethal pathogen. This agrees with the opportunistic character of O. anthropi and sets a clear cut limit with brucellosis, where neutrophils do not seem to play a significant role in controlling the infection [3]. It is tempting to speculate that conserved molecules involved in Brucella virulence like phosphatidylcholine, the VLCFA in lipoproteins and lipid A, the periplasmic cyclic glucans, and the orthologous BvrR/BvrS and VirB systems [34], [35], [36] participate in this survival. In the natural environment of Ochrobactrum, these structures and systems may be related to resistance to antibiotic peptides, osmotic stress, the transference of plasmids, survival against protozoan predation, or other modes of interaction with eukaryotic cells of the rhizosphere [4], [37].

Similarly to the proinflammatory response, It is striking that the adaptive immune response provoked by O. anthropi also lies between that induced by typical extracellular and intracellular pathogens [38]. Indeed, while Brucella infection is characterized by a protective Th1 immune response [39], O. anthropi infection stimulates a mix of Th1 and Th2 type response which does not protect against Brucella [38]. However, when switched to a Th1-biased response by addition of the TLR9 agonist unmethylated CpG, O. anthropi induces some protection against Brucella. Since protection against Brucella is dependent on TLR9 and Myd88 [40], [41] and, since unmethylated CpG is not accessible unless envelopes are breached, then the toughness of the Brucella envelope should play an important role in the stealthy behavior. The results presented here and previously [18] evidences this and demonstrates that O. anthropi differs from Brucella in crucial envelope properties.

It is feasible to propose that Brucella ancestors had already envelope molecules displaying determinants that departure from the PAMPs targeted by innate immunity receptors. Since PAMPs are conserved because they are functionally essential, one may ask why this departure from the canonical structure was possible in the Brucellaceae, and what kind of selective forces were behind it. In most Proteobacteria, cell envelopes are impermeable to a large variety of noxious substances through the combined action of outer leaflet characteristics that do not allow the partition of hydrophobic substances, a property linked to the high density of negatively charged groups in the outer membrane [42], [43]. This leads to the conservation of those groups and, consequently, to their ready recognition as PAMPs by innate immunity, as illustrated by the wide array of bactericidal peptides that exert their action upon binding to the negative charged groups in core and lipid A.

Whereas O. anthropi keeps the permeability barrier, Brucella does not and is concomitantly less sensitive to bactericidal peptides [18]. Significantly, this property also lays between the high resistance of Brucella envelopes and the sensitivity of enterobacterial envelopes to bactericidal cationic peptides [18]. This shows that Brucella and O. anthropi represent a lineage endowed with outer membranes and LPSs with some degree of resistance to bactericidal peptides and, since the primary habitat of most Brucellaceae is the soil, it is likely that this property resulted from the selective pressure of antibiotic peptides produced by soil microorganisms. However, in Brucella the permeability barrier to which sensitivity to bactericidal peptides is linked, is no longer necessary to confront mucosal surfaces or intracellular environments. Therefore, this property could have been relinquished by the Brucella ancestor, keeping and increasing its resistance to host bactericidal peptides.

An analysis of the respective LPSs, supports above developed hypotheses (Fig. 10). The similarity of the biological activity of the BaLPS and OaLPS makes unlikely that the respective O-chains play a significant role since they are very different structures [44], [45] and, at least for B. abortus there is evidence for horizontal acquisition [46]. On the other hand, the VLCFAs of both bacteria represent a first departure of the canonical lipid A structure (Fig. 10) that seems ancestral in α-2 Proteobacteria [47]. Since VLCFAs carry an hydroxyl group at the iso position and are postulated to span the outer membrane [47], they could provide a firm anchorage that contributes to explain the integrity of the outer membranes observed in the polymyxin B binding experiments [18]. Detail core structures have not been elucidated for any Brucella species or O. anthropi, but the available information on the Brucella closer relative O. intermedium and on the sugar and lipid compositions in B. abortus and O. anthropi also shows a departure from enteric bacteria, with a comparatively lower number of negatively charged groups (Fig. 10). Moreover, quantitative analysis suggests a further charge reduction in Brucella LPS core that is consistent with the loss of the permeability barrier and its greater resistance to bactericidal peptides (Fig. 10).

Figure 10. Model of Brucella, Ochrobactrum and Escherichia lipid As and core oligosaccharide moieties. [18], [19], [45].

The presence of three KDO and phosphorylated heptoses confers five negative charges to the Escherichia coli core oligosaccharide. The core oligosaccharide of Ochrobactrum LPS possesses galacturonic acid, in contrast to Brucella LPS core which does not possess this negatively charged acidic sugar. The differential sensitivity of Escherichia, Ochrobactrum and Brucella to cationic peptides expressed as sensitivity to polymyxin B (PlxB) units (U) is linked to the different LPSs properties such as uptake hydrophobic probes (NPN), PlxB binding, and LPS zeta potential (ζ) dependence after saturation with polymyxin B [18]. Notice that upon saturation with PlxB, zeta potential became positive (1 mV) for Ochrobactrum LPS while remaining negative (−5 mV) for Brucella smooth LPS, suggesting hindered access to inner targets. Ochrobactrum and Brucella LPSs does not show significant differences in lipid A structure and acyl chain fluidity and display very similar structure [18], [19], [45]. It has been proposed that the quinovosamine sugar in Ochrobactrum and Brucella LPS links the O core oligosaccharide with the O-chain polysaccharide [21].

The departure from the canonical PAMP in both the VLCFAs of lipid A and in the negatively charged groups of the lipid A-core of Ochrobactrum, enhanced further in B. abortus, is in agreement with the results of the LPS-hMD-2 binding experiments. MD-2 comprises a hydrophobic pocket and some cationic domains that correspond to two functional types of LPS recognition sites, one for the lipid A portion and the other for the anionic groups [28], [48], [49]. In the case of the lipid A, it would be impossible to accommodate all acyl chains, particularly VLCFA, into the hydrophobic pocket of MD-2. A model of TLR4 activation, that suggests that the secondary acyl chain of the lipid A protrudes out of the MD-2 pocket and directly interacts with hydrophobic side chains of TLR4 (F440 and F463), diving the receptor dimerization and activation, has been suggested [50]. In case of OaLPS we propose that the VLCFA interact directly with TLR4, and in this case the terminal hydroxyl group increases its propensity to be exposed to the solvent. The key requirement for activation of TLR4.MD-2 complex is the delivery of monomeric LPS, by LBP and CD14, which may also be determined by the LPS core, which differs between the Ochrobactrum and Brucella LPSs (Fig. 10). Since these LPS structures are rooted in the phylogeny of the group, this illustrates how relatively few structural changes on an orthologous structure may contribute to adaptation to widely different lifestyles.

Not all properties of O. anthropi were related to LPS. The complete bacterium readily consumed complement and was killed by serum, but its LPS did not trigger the lytic cascade. Consequently, differences in other surface structures, such as O. anthropi lipoproteins or extracellular polysaccharides [51], should be responsible for complement activation and the ensuing proinflammatory activities. Bacterial capsular polysaccharides have been shown to consume complement and stimulate cytokine production through TLR activation [52], [53], [54].

Animal pathogenic α-Proteobacteria have molecules not readily recognized by innate immunity, and these bacteria also tend to establish chronic infections [55], [56], [57], [58], [59]. For instance, Bartonella possesses a low agonist non-canonical lipid A with VLCFA, a low number of lipoproteins and flagella not recognized by TLR5 [27], [59], [60]. In Rickettsia, there is an absence of flagella a reduced number of lipoproteins and a non-canonical LPS [61], [62]. The insect Wolbachia parasites do not posses LPS, flagellum, fimbria or pili, and contain a low number of lipoproteins and an unusual peptidoglycan [63]. Intracellular Ehrlichia and Anaplasma have low number of lipoproteins and do not have LPS, peptidoglycan or flagellin [64], [65]. Therefore it seems plausible that the α-Proteobacteria intracellular pathogen ancestor may have been endowed with structures prone to low recognition. In addition, animals could have developed an innate immune repertoire that preferably recognizes β and γ-Proteobacteria rather than α-Proteobacteria molecular patterns. In the case of Brucella, it is possible that evolutionary pressures linked to innate immunity forced for the selection of organisms displaying low PAMPs recognition [66]. The selective forces working in the second scenario could be linked to the early endosymbiotic relationship between eukaryotic cells and α-Proteobacteria ancestors [67] which would have led to the identification of some bacterial molecules as “self” in relation to those of β and γ-Proteobacteria, which appeared later in evolution [68]. It is worth noting, however, that all these α-Proteobacteria have orthologous type IV secretion systems, which seem to be essential for controlling the intracellular survival.

We propose that, Brucella evolved from an opportunistic pathogen, non markedly endotoxic ancestor within a mammal environment [30], [69] by: i), removing microbial molecular determinants such as capsules and fimbriae; ii), modifying surface molecules such as flagella, lipoproteins and LPS to provide high resistance to bactericidal cationic substances and complement and, concomitantly, avoid recognition by TLRs; iii), keeping structures such as peptidoglycan and unmethylated DNA under a tough envelope; and iv), maintaining the regulatory systems (e.g. BvrR-BvrS) that control the homeostasis of an outer membrane whose overall structure is essential in the interaction with the host. This furtive strategy, which also required a remodeling of ancestral type IV secretion systems to sneak and reproduce in cells, does not preclude the development of determinants actively hampering the activation of immunity at later stages [70], [71], [72].

Finally, the recorded phenotypic, biological and pathogenic differences between Brucella and Ochrobactrum, strengthens arguments for maintaining these two bacterial clusters as separate genera in the Brucellaceae and reinforce the concept that non-drastic molecular variations generate marked divergences that result in different biological groups.

Materials and Methods

Ethics Statement

All animals were handled and sacrificed according to the approval and guidelines established by the “Comité Institucional para el Cuido y Uso de los Animales of the Universidad de Costa Rica”, and in agreement with the corresponding law “Ley de Bienestar de los Animales No 7451” of Costa Rica (http://www.protecnet. The recommendations stated in the Weatherall report “The use of non-human primates in research” (, were followed accordingly.

Bacterial strains, LPSs and biological reagents

Salmonella enterica sv. typhimurium strain SL1344, virulent B. abortus 2308, and O. anthropi LMG 3331T have been previously described [1], [45], [73]. BaLPS, OaLPS, SeLPS and EcLPS were purified and characterized as previously described [18], [24]. S. enterica (strainHL83) LPS was provided by K. Brandenburg (Forschungszentrum Borstel, Germany). E. coli CNF was purified as previously described [22]. BvrR was cloned as glutathione-S-transferase (GST) fusion protein in pGEX-2TK vector, expressed in E. coli BL21(DE3), and purified following standard protocols. Rabbit anti-BvrR antibodies were produced by four intramuscularly applied boosts of GST-BvrR (250 µg) in complete (first boost) or incomplete (second to fourth boost) Freund adjuvant (Sigma-Aldrich Co.). Antibodies from 2 ml of serum were adsorbed to GST-BvrR-Sepharose beads, eluted with 0.2 M glycine, pH 2.5, and collected in 1 M Tris, pH 9,0. The antibodies were concentrated by ultrafiltration and stored at -20°C in 50% glycerol. The anti-Ochrobactrum antibody was produced in a rabbit by 4 boosts of intramuscular injections of formaldehyde fixed bacteria. The specificity of the antibody was confirmed by immunofluorescence against a panel of related α-Proteobacteria. Enzyme linked immunosorbent assay (ELISA) kits for cytokine determination were from BD Biosciences (San Diego, CA.). Monoclonal antibody RB6-8C5 against murine granulocytes was a gift from Bärbel Raupach, MPIIB, Berlin, Germany. Allophycocyanin conjugated-anti-CD11c antibody (HL3) and phycoerythrin-conjugated anti-CD40, -CD80, -CD86, and -MHC class II were all from Pharmingen. Mouse antibody FK2 (Biomol) was provided by D. Williams (University of Toronto). The properties and reactivity of anti-VirB8 antibody have been described before [74].

Infections and bacterial counts in experimental animals

The characteristics, source and maintenance of Balb/c, C57BL/6 and the KO counterparts TLR4−/− and TLR2−/−, mice have been described previously [3], [72]. Generation of neutropenic mice was performed as described before [3]. For calculating the bacterial multiplication rates in spleens, groups of 20 eight-week-old female BALB/c were intraperitoneally inoculated with 0.1 ml of a suspension of the indicated bacteria species. Animals for each group were killed by cervical dislocation at appropriate times. Spleens were aseptically removed and individually homogenized in 1 ml PBS. Dilutions from this homogenate were seeded on TSA and CFU numbers determined. Alternatively mice were inoculated with 0.1 ml (pyrogen-free sterile PBS) of the indicated LPS preparation by the intraperitoneal route. Wistar rats were maintained in the animal facility of the Veterinary School of the National University, Costa Rica. Cytokine or leukocyte counts were performed in the peritoneum and blood of infected or LPS inoculated mice as described previously [3].

Infections and bacterial counts in macrophages and epithelial cells

Human cervix carcinoma cells (HeLa; American Type Culture Collection No. CCL-2) were grown at 37°C under 5% CO2 in Eagle's minimal essential medium supplemented with 5% fetal bovine serum, 2.5% sodium bicarbonate and 1% glutamine (all from Gibco, Inc.). Murine macrophages (Raw 264.7; American Type Culture Collection No. TIB-71) were grown at 37°C under 5% CO2 in Dulbecco's medium supplemented with 10% fetal bovine serum, 2.5% sodium bicarbonate and 1% glutamine (all from Gibco). Penicillin (100 units/ml) and streptomycin (100 µg/ml) routinely added, were excluded from cell cultures during Ochrobactrum and Brucella infections. Before infection, 24-well-plates monolayers were washed in PBS and kept at 4°C. Infections were carried out using cultures of O. anthropi or B. abortus in logarithmic phase diluted in Eagle's minimal essential medium to reach the desired multiplicities of infection (MOI). Plates were centrifuged at 300×g at 4°C, incubated for 30 min at 37°C under 5% CO2 and washed 3 times with PBS. Extracellular bacteria were killed by adding Eagle's minimal essential medium supplemented with 100 µg/ml gentamicin for 1 h for B. abortus or 4 h for O. anthropi. Cells were further incubated for the indicated times in the presence of 5 µg/ml gentamicin for B. abortus or 100 µg/ml gentamicin for O. anthropi. Plates were then washed with PBS. Cells were lyzed with 0.1% Triton X-100 for 10 min. Aliquots were diluted and plated in tryptic soy agar and incubated at 37°C for determination of CFU. For immunofluorescence experiments HeLa cells or Raw 264.7 macrophages grown on 12 mm glass slides were inoculated with B. abortus or O. anthropi as described above and kept in the presence of gentamicin for the indicated times. After incubation cells were fixed with 3.7% paraformaldehyde-PBS and free aldehyde groups were quenched with 50 mM NH4Cl-PBS. After permeabilization with 0.5% Triton-PBS, total bacteria were labeled with rabbit anti-Ochrobactrum antibodies followed by TRITC conjugated anti-rabbit antibodies.

Confocal microscopy and flow cytometry of LPS treated dendritic cells

Immunofluorescence confocal microscopy and flow cytometry of LPS treated murine bone marrow derived dendritic cells (BMDC) were performed as described before [72]. Briefly, paraformaldehyde fixed cells processed by immunofluorescence were examined on a Zeiss LSM 510 laser scanning confocal microscope for image acquisition. Antibody against a conserved cytoplasmic epitope found on MHC-II I-Aβ which strongly labels BMDC but not macrophages was used throughout the experiments. For confirmation of the dendritic cell phenotype, double labeling with anti- MHC-II and anti-CD11c antibody was performed. Quantification was always done by counting at least 100 cells in 4 independent experiments, for a total of at least 400 host cells analyzed. DALIS BMDC were collected and stained immediately before fixation. Isotype controls were included as well as BMDC non-LPS treated as control for autofluorescence. Cells were always gated on CD11c for analysis and at least 100000 events were collected to obtain a minimum of 10000 CD11c positive events for analysis. A FACS calibur cytometer (Becton Dickinson) was used and data was analyzed using FlowJo software (Tree Star).

Electrophoretic and immunochemical analysis

The different bacterial strains were grown to exponential phase in TSB at 37°C. The bacteria were concentrated by centrifugation (10000×g, 10 min), resuspended in Laemmli sample buffer [75] and heated at 100°C for 20 min. Protein concentration was determined by BioRad DC method according to the manufacturer's instructions and equal amounts of protein (20 µg) were loaded onto a 12.5% gel for SDS-PAGE. Separated proteins were transferred to PVDF membrane and probed with the indicated antibodies. Membranes were further incubated with peroxidase conjugated anti-mouse or anti-rabbit antibodies and the detected bands were visualized by chemiluminescence reaction exposed to X-ray films.


Genomic DNA was isolated by CTAB method according to standard procedures [76]. Specific amplification from gDNA or cDNA was achieved with the following primers designed after analysis of O. anthropi genome: for locus Oant_0677: oantvirB4f GGATTACACCGTGACCTCAAC and oantvirB4r GCCTGATAACATGCGTCCATAA; for locus Oant_0678: oantvirB5.3 CGATGCTCTTCATCAGCAGATTGAG, oantvirB5.5 CTTTGCGACATCCGCCATATC; for locus Oant_0682: oantvirB8.3 GATCAAGACCGCATGATTCAGG, oantvirB8.5 GTCACGGCTTCGTCGTAAGTG. For amplification of gDNA or cDNA encoding putative bvrR and bvrS, the following primers were used: for Oant_0827 (bvrS): oantbvrS.3 GCATTCTCTACATGAATCAGTTCC and oantbvrS.5 GTGATGGCGTCAGGCTTTG. And for Oant_0828 (bvrR): oantbvrR.3 GACGACGACCGCAACATCCTG and oantbvrR.5 CGAAGATCGCAAGATTGGGC. Positive controls of amplification of genomic DNA or cDNA were achieved using a pair of primers oantl12.3 CTGTCGGCACTGACCGTTCTCG and oantl12.5 CGCCACCAGCAGCAGCAACTG interrogating locus Oant_1946, encoding the ribosomal protein L7/L12. Cycling conditions were: 94°C for 4 min, 30 cycles at 94°C for 30 s, 50 or 65°C for 1 minute and 68°C for 30 seconds and continued with a 68°C incubation for 10 minutes. The PCR products were analyzed on agarose gels using standard procedures. For RNA extraction bacterial cultures were adjusted to 109 CFU/ml and disrupted with 0.5% Zwittergent 3–16 at 37°C for 1 h. Total RNA was extracted using RNeasy Midi Kit (Qiagen, Co.) according to the manufacturer's instructions. Eluted RNA was treated with Ambion Turbo DNAse following the robust protocol. Synthesis of cDNA was carried out using Revert Aid M-MuLV Reverse Transcriptase with random hexamers and following manufacturers' Instructions. (Fermentas). The obtained cDNA was used as template for detection of the genes mentioned above by PCR using the same primers described. Minus RT controls were performed using RNA samples during PCR.

Complement depletion and bactericidal action of serum

Complement consumption was estimated as previously described [77], [78]. Briefly, rabbit serum was incubated with live packed bacteria or LPS for 30 min at 37°C. After incubation, the complement activity on the serum was detected as the ability to lyze sheep erythrocytes pre-sensitized with guinea pig anti-sheep antibodies. For the estimation of complement bactericidal activity, exponentially growing bacteria were adjusted to 104 CFU/ml in saline and dispensed in triplicate in microtiter plates (45 µl per well) containing fresh normal bovine serum (90 µl/well). After 90 min of incubation at 37°C, brain heart infusion broth was dispensed (200 µl/well), mixed with the bacterial suspension and 100 µl was plated on TSA. Results were expressed as the percentage of CFU with respect to the inoculums in three assays.

Binding of LPS to hTLR4 co-receptor hMD-2 protein

Binding of LPS to MD-2 was assayed by two different methods: displacement of bis-ANS from hMD-2 by LPS, and binding to hMD-2 by competitive enzyme linked-immunosorbent assay (ELISA). Recombinant hMD-2 was produced in E. coli and isolated as described before [79]. In both assays, LPS was subjected to three cycles of heating at 56°C followed by cooling to 4°C, left over night and sonicated before using. Binding of bis-ANS to hMD-2 was measured at 20°C using excitation at 385 nm and measuring the emission fluorescence spectra between 420 and 550 nm. Then, from 5 to 80 µl of a LPS stock at 0.125 mg/ml was added (controls were prepared with the same volume of water) to preincubated bis-ANS/hMD-2 complex (50 nM/50 nM). The F0 value was the fluorescence of bis-ANS/hMD-2 complexes after 5 min of incubation (to reach stable fluorescence).The FLPS value was the fluorescence after LPS addition to the complex. Fluorescence was measured on Perkin Elmer fluorimeter LS 55. Quartz glass cuvettes (5×5 and 10×5 mm optical path, Hellma Suprasil) were used and bis-ANS was obtained from Sigma (St. Louis, Missouri, U.S.A.). The ELISA for determination of LPS binding to hMD-2 was performed in 96-well plates (NUNC immunoplate F96 cert. Maxi-sorp, Roskilde, Denmark). Chicken polyclonal anti-hMD-2 (GenTel, Madison, WI, U.S.A) (5 µg/ml) in 50 mM Na2CO3 (pH 9.6) was used to coat the microtiter plate at 4°C overnight. Excess binding sites were blocked with 1% BSA in PBS buffer (pH 7.2) for 1 h at room temperature, and rinsed three times with the same buffer. During the blocking step, hMD-2 (0.75 µM) was preincubated with 0 µg/ml to 20 µg/ml LPS at 37°C and, as a negative control, LPS was also preincubated in absence of hMD-2. These solutions were added to the plate, which was then incubated for 1 h at 37°C. After rinsing, only hMD-2 without LPS was detected by incubation with 0.1 µg/ml of mouse monoclonal anti-hMD-2 (clone 9B4 e-Bioscience San Diego, CA., U.S.A.) in PBS buffer at 37°C for 1 h, followed by incubation with 0.1 µg/ml peroxidase-conjugated goat anti-mouse IgG (Santa Cruz, CA., U.S.A.), also in PBS buffer at 37°C for 1 h. After plate washing, ABTS (Sigma) was added, the reaction was stopped with 1% SDS after 15 min, and the absorbance at 420 nm measured using a Mithras LB940 apparatus. (Berthold Technologies). S. enterica (strainHL83) LPS, used as a control.

Stimulation of NF-κB transcription by purified LPS. HEK293 cells (ATCC CRL-1573) were transiently transfected with 3 ng hTLR4, 6 ng hMD-2, 5 ng hCD14 expression vectors, 50 ng of NF-κB or IP-10 promoter dependent luciferase reporter, and 5 ng constitutive Renilla reporter plasmids, using lipofectamine 2000 (Invitrogen). After 6 h, media were changed to Dulbeccos's modified Eagle's medium supplemented with 10% fetal bovine serum and different amounts of LPS were added to the cells. As a negative control, we used cells transfected with the same plasmids but without addition of LPS. After 20 h, cells were lyzed and analyzed for reporter gene activities using a dual luciferase reporter assay system on a Mithras LB940 apparatus (Berthold Technologies). The data of luciferase activity were normalized using Renilla luciferase readings.

Statistical methods

Student's t test for was used for determining the statistical significance in the different assays. For bacterial colonization experiments, the Mann-Whitney test was performed accordingly (


Bärbel Raupach (MPIIB, Berlin, Germany), David O'Callaghan (INSERM U431, Université de Montpellier 2, 34095 Montpellier Cedex 05, France) and D. Williams (University of Toronto) for supplying the anti-RB6, anti VirB8 and anti FK2 antibodies, respectively.

Author Contributions

Conceived and designed the experiments: EBC RC CCD JPG IM EM ECO. Performed the experiments: EBC RC CCD LQL AM CGV MMK ECO. Analyzed the data: EBC RC CCD MI RJ JPG ECO. Contributed reagents/materials/analysis tools: MMK RJ. Wrote the paper: IM EM ECO.


  1. 1. Moreno E, Moriyón I (2006) The genus Brucella. In: Dworkin M, Falkow S, Rosenberg E, Schleifer KH, Stackebrandt E, editors. The Prokaryotes. 3 ed. New York: Springer. pp. 315–456.
  2. 2. Batut J, Andersson SG, O'Callaghan D (2004) The evolution of chronic infection strategies in the alpha-proteobacteria. Nat Rev Microbiol 2: 933–945.
  3. 3. Barquero-Calvo E, Chaves-Olarte E, Weiss DS, Guzmán-Verri C, Chacon-Díaz C, et al. (2007) Brucella abortus uses a stealthy strategy to avoid activation of the innate immune system during the onset of infection. PLoS ONE 2: e631.
  4. 4. Berg G, Eberl L, Hartmann A (2005) The rhizosphere as a reservoir for opportunistic human pathogenic bacteria. Environ Microbiol 7: 1673–1685.
  5. 5. Gill MV, Ly H, Mueenuddin M, Schoch PE, Cunha BA (1997) Intravenous line infection due to Ochrobactrum anthropi (CDC Group Vd) in a normal host. Heart Lung 26: 335–336.
  6. 6. Mahmood MS, Sarwari AR, Khan MA, Sophie Z, Khan E, et al. (2000) Infective endocarditis and septic embolization with Ochrobactrum anthropi: case report and review of literature. J Infect 40: 287–290.
  7. 7. Stiakaki E, Galanakis E, Samonis G, Christidou A, Maraka S, et al. (2002) Ochrobactrum anthropi bacteremia in pediatric oncology patients. Pediatr Infect Dis J 21: 72–74.
  8. 8. Berman AJ, Del Priore LV, Fischer CK (1997) Endogenous Ochrobactrum anthropi endophthalmitis. Am J Ophthalmol 123: 560–562.
  9. 9. Greven CM, Nelson KC (2001) Chronic postoperative endophthalmitis secondary to Ochrobactrum anthropi. Retina 21: 279–280.
  10. 10. Inoue K, Numaga J, Nagata Y, Sakurai M, Aso N, et al. (1999) Ochrobactrum anthropi endophthalmitis after vitreous surgery. Br J Ophthalmol 83: 502.
  11. 11. Ozdemir D, Soypacaci Z, Sahin I, Bicik Z, Sencan I (2006) Ochrobactrum anthropi endocarditis and septic shock in a patient with no prosthetic valve or rheumatic heart disease: case report and review of the literature. Jpn J Infect Dis 59: 264–265.
  12. 12. Nadjar D, Labia R, Cerceau C, Bizet C, Philippon A, et al. (2001) Molecular characterization of chromosomal class C beta-lactamase and its regulatory gene in Ochrobactrum anthropi. Antimicrob Agents Chemother 45: 2324–2330.
  13. 13. Higgins CS, Avison MB, Jamieson L, Simm AM, Bennett PM, et al. (2001) Characterization, cloning and sequence analysis of the inducible Ochrobactrum anthropi AmpC beta-lactamase. J Antimicrob Chemother 47: 745–754.
  14. 14. Foster JT, Beckstrom-Sternberg SM, Pearson T, Beckstrom-Sternberg JS, Chain PS, et al. (2009) Whole Genome-Based Phylogeny and Divergence of the Genus Brucella. J Bacteriol.
  15. 15. Velasco J, Romero C, López-Goñi I, Leiva J, Díaz R, et al. (1998) Evaluation of the relatedness of Brucella spp. and Ochrobactrum anthropi and description of Ochrobactrum intermedium sp. nov., a new species with a closer relationship to Brucella spp. Int J Syst Bacteriol 48 Pt 3: 759–768.
  16. 16. Moriyón I, Berman DT (1982) Effects of nonionic, ionic, and dipolar ionic detergents and EDTA on the Brucella cell envelope. J Bacteriol 152: 822–828.
  17. 17. Moriyón I, Lopez-Goñi I (1998) Structure and properties of the outer membranes of Brucella abortus and Brucella melitensis. Int Microbiol 1: 19–26.
  18. 18. Velasco J, Bengoechea JA, Brandenburg K, Lindner B, Seydel U, et al. (2000) Brucella abortus and its closest phylogenetic relative, Ochrobactrum spp., differ in outer membrane permeability and cationic peptide resistance. Infect Immun 68: 3210–3218.
  19. 19. Moriyón I (2004) Against gram-negative bacteria: the lipopolysaccharide case. In: Gorvel JP, editor. Intracellular pathogens in membrane interactions and vacuole biogenesis. New York: Kluwer Academic/Plenum Publishers. pp. 204–230.
  20. 20. Lapaque N, Moriyón I, Moreno E, Gorvel JP (2005) Brucella lipopolysaccharide acts as a virulence factor. Curr Opin Microbiol 8: 60–66.
  21. 21. González D, Grillo MJ, De Miguel MJ, Ali T, Arce-Gorvel V, et al. (2008) Brucellosis vaccines: assessment of Brucella melitensis lipopolysaccharide rough mutants defective in core and O-polysaccharide synthesis and export. PLoS ONE 3: e2760.
  22. 22. Chaves-Olarte E, Guzmán-Verri C, Meresse S, Desjardins M, Pizarro-Cerda J, et al. (2002) Activation of Rho and Rab GTPases dissociates Brucella abortus internalization from intracellular trafficking. Cell Microbiol 4: 663–676.
  23. 23. Sola-Landa A, Pizarro-Cerda J, Grillo MJ, Moreno E, Moriyón I, et al. (1998) A two-component regulatory system playing a critical role in plant pathogens and endosymbionts is present in Brucella abortus and controls cell invasion and virulence. Mol Microbiol 29: 125–138.
  24. 24. Freer E, Moreno E, Moriyón I, Pizarro-Cerda J, Weintraub A, et al. (1996) Brucella-Salmonella lipopolysaccharide chimeras are less permeable to hydrophobic probes and more sensitive to cationic peptides and EDTA than are their native Brucella sp. counterparts. J Bacteriol 178: 5867–5876.
  25. 25. Lamontagne J, Butler H, Chaves-Olarte E, Hunter J, Schirm M, et al. (2007) Extensive cell envelope modulation is associated with virulence in Brucella abortus. J Proteome Res 6: 1519–1529.
  26. 26. Manterola L, Moriyón I, Moreno E, Sola-Landa A, Weiss DS, et al. (2005) The lipopolysaccharide of Brucella abortus BvrS/BvrR mutants contains lipid A modifications and has higher affinity for bactericidal cationic peptides. J Bacteriol 187: 5631–5639.
  27. 27. Andersen-Nissen E, Smith KD, Strobe KL, Barrett SL, Cookson BT, et al. (2005) Evasion of Toll-like receptor 5 by flagellated bacteria. Proc Natl Acad Sci U S A 102: 9247–9252.
  28. 28. Mancek-Keber M, Jerala R (2006) Structural similarity between the hydrophobic fluorescent probe and lipid A as a ligand of MD-2. Faseb J 20: 1836–1842.
  29. 29. Lapaque N, Forquet F, de Chastellier C, Mishal Z, Jolly G, et al. (2006) Characterization of Brucella abortus lipopolysaccharide macrodomains as mega rafts. Cell Microbiol 8: 197–206.
  30. 30. Moreno E (1998) Genome evolution within the alpha Proteobacteria: why do some bacteria not possess plasmids and others exhibit more than one different chromosome? FEMS Microbiol Rev 22: 255–275.
  31. 31. Ariza J (1999) Brucellosis: an update. The perspective from the Mediterranean basin. Reviews in Medical Microbiology 10: 125–135.
  32. 32. Yu WL, Lin CW, Wang DY (1998) Clinical and microbiologic characteristics of Ochrobactrum anthropi bacteremia. J Formos Med Assoc 97: 106–112.
  33. 33. Arora U, Kaur S, Devi P (2008) Ochrobactrum anthropi septicaemia. Indian J Med Microbiol 26: 81–83.
  34. 34. Arellano-Reynoso B, Lapaque N, Salcedo S, Briones G, Ciocchini AE, et al. (2005) Cyclic beta-1,2-glucan is a Brucella virulence factor required for intracellular survival. Nat Immunol 6: 618–625.
  35. 35. Conde-Alvarez R, Grillo MJ, Salcedo SP, de Miguel MJ, Fugier E, et al. (2006) Synthesis of phosphatidylcholine, a typical eukaryotic phospholipid, is necessary for full virulence of the intracellular bacterial parasite Brucella abortus. Cell Microbiol 8: 1322–1335.
  36. 36. Moreno E, Moriyón I (2002) Brucella melitensis: a nasty bug with hidden credentials for virulence. Proc Natl Acad Sci U S A 99: 1–3.
  37. 37. Weller SA, Stead DE, Young JP (2004) Acquisition of an Agrobacterium Ri plasmid and pathogenicity by other alpha-Proteobacteria in cucumber and tomato crops affected by root mat. Appl Environ Microbiol 70: 2779–2785.
  38. 38. He Y, Vemulapalli R, Schurig GG (2002) Recombinant Ochrobactrum anthropi expressing Brucella abortus Cu,Zn superoxide dismutase protects mice against B. abortus infection only after switching of immune responses to Th1 type. Infect Immun 70: 2535–2543.
  39. 39. Baldwin CL, Goenka R (2006) Host immune responses to the intracellular bacteria Brucella: does the bacteria instruct the host to facilitate chronic infection? Crit Rev Immunol 26: 407–442.
  40. 40. Copin R, De Baetselier P, Carlier Y, Letesson JJ, Muraille E (2007) MyD88-dependent activation of B220-CD11b+LY-6C+ dendritic cells during Brucella melitensis infection. J Immunol 178: 5182–5191.
  41. 41. Weiss DS, Takeda K, Akira S, Zychlinsky A, Moreno E (2005) MyD88, but not toll-like receptors 4 and 2, is required for efficient clearance of Brucella abortus. Infect Immun 73: 5137–5143.
  42. 42. Nikaido H (2003) Molecular basis of bacterial outer membrane permeability revisited. Microbiol Mol Biol Rev 67: 593–656.
  43. 43. Vaara M (1999) Lipopolysaccharide and the permeability of the bacterial outer membrane. In: Brade H, Opal SM, Vogel SN, Morrison DC, editors. Endotoxin in health and disease. New York: Marcel Dekker. pp. 31–38.
  44. 44. Caroff M, Bundle DR, Perry MB, Cherwonogrodzky JW, Duncan JR (1984) Antigenic S-type lipopolysaccharide of Brucella abortus 1119–3. Infect Immun 46: 384–388.
  45. 45. Velasco J, Moll H, Knirel YA, Sinnwell V, Moriyón I, et al. (1998) Structural studies on the lipopolysaccharide from a rough strain of Ochrobactrum anthropi containing a 2,3-diamino-2,3-dideoxy-D-glucose disaccharide lipid A backbone. Carbohydr Res 306: 283–290.
  46. 46. Godfroid F, Cloeckaert A, Taminiau B, Danese I, Tibor A, et al. (2000) Genetic organisation of the lipopolysaccharide O-antigen biosynthesis region of Brucella melitensis 16M (wbk). Res Microbiol 151: 655–668.
  47. 47. Bhat UR, Mayer H, Yokota A, Hollingsworth RI, Carlson RW (1991) Occurrence of lipid A variants with 27-hydroxyoctacosanoic acid in lipopolysaccharides from members of the family Rhizobiaceae. J Bacteriol 173: 2155–2159.
  48. 48. Kennedy MN, Mullen GE, Leifer CA, Lee C, Mazzoni A, et al. (2004) A complex of soluble MD-2 and lipopolysaccharide serves as an activating ligand for Toll-like receptor 4. J Biol Chem 279: 34698–34704.
  49. 49. Jerala R (2007) Structural biology of the LPS recognition. Int J Med Microbiol 297: 353–363.
  50. 50. Nu AR, Va LJ, Oblak A, Pristov-Ek P, Gioannini TL, et al. (2009) Essential roles of hydrophobic residues in both md-2 and toll-like receptor 4 in activation by endotoxin. J Biol Chem.
  51. 51. Calvo C, Silva-Castro GA, Uad I, García Fandino C, Laguna J, et al. (2008) Efficiency of the EPS emulsifier produced by Ochrobactrum anthropi in different hydrocarbon bioremediation assays. J Ind Microbiol Biotechnol.
  52. 52. Acosta F, Ellis AE, Vivas J, Padilla D, Acosta B, et al. (2006) Complement consumption by Photobacterium damselae subsp. piscicida in seabream, red porgy and seabass normal and immune serum. Effect of the capsule on the bactericidal effect. Fish Shellfish Immunol 20: 709–717.
  53. 53. Regueiro V, Campos MA, Pons J, Alberti S, Bengoechea JA (2006) The uptake of a Klebsiella pneumoniae capsule polysaccharide mutant triggers an inflammatory response by human airway epithelial cells. Microbiology 152: 555–566.
  54. 54. Sen G, Khan AQ, Chen Q, Snapper CM (2005) In vivo humoral immune responses to isolated pneumococcal polysaccharides are dependent on the presence of associated TLR ligands. J Immunol 175: 3084–3091.
  55. 55. Choi KS, Scorpio DG, Dumler JS (2004) Anaplasma phagocytophilum ligation to toll-like receptor (TLR) 2, but not to TLR4, activates macrophages for nuclear factor-kappa B nuclear translocation. J Infect Dis 189: 1921–1925.
  56. 56. Fenn K, Blaxter M (2006) Wolbachia genomes: revealing the biology of parasitism and mutualism. Trends Parasitol 22: 60–65.
  57. 57. Lee EH, Rikihisa Y (1996) Absence of tumor necrosis factor alpha, interleukin-6 (IL-6), and granulocyte-macrophage colony-stimulating factor expression but presence of IL-1beta, IL-8, and IL-10 expression in human monocytes exposed to viable or killed Ehrlichia chaffeensis. Infect Immun 64: 4211–4219.
  58. 58. von Loewenich FD, Scorpio DG, Reischl U, Dumler JS, Bogdan C (2004) Frontline: control of Anaplasma phagocytophilum, an obligate intracellular pathogen, in the absence of inducible nitric oxide synthase, phagocyte NADPH oxidase, tumor necrosis factor, Toll-like receptor (TLR)2 and TLR4, or the TLR adaptor molecule MyD88. Eur J Immunol 34: 1789–1797.
  59. 59. Zahringer U, Lindner B, Knirel YA, van den Akker WM, Hiestand R, et al. (2004) Structure and biological activity of the short-chain lipopolysaccharide from Bartonella henselae ATCC 49882T. J Biol Chem 279: 21046–21054.
  60. 60. Babu MM, Priya ML, Selvan AT, Madera M, Gough J, et al. (2006) A database of bacterial lipoproteins (DOLOP) with functional assignments to predicted lipoproteins. J Bacteriol 188: 2761–2773.
  61. 61. Blanc G, Ogata H, Robert C, Audic S, Suhre K, et al. (2007) Reductive genome evolution from the mother of Rickettsia. PLoS Genet 3: e14.
  62. 62. Amano K, Fujita M, Suto T (1993) Chemical properties of lipopolysaccharides from spotted fever group rickettsiae and their common antigenicity with lipopolysaccharides from Proteus species. Infect Immun 61: 4350–4355.
  63. 63. Wu M, Sun LV, Vamathevan J, Riegler M, Deboy R, et al. (2004) Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: a streamlined genome overrun by mobile genetic elements. PLoS Biol 2: E69.
  64. 64. Lin M, Rikihisa Y (2003) Ehrlichia chaffeensis and Anaplasma phagocytophilum lack genes for lipid A biosynthesis and incorporate cholesterol for their survival. Infect Immun 71: 5324–5331.
  65. 65. Rikihisa Y (2006) Ehrlichia subversion of host innate responses. Curr Opin Microbiol 9: 95–101.
  66. 66. Sansonetti PJ (2006) The innate signaling of dangers and the dangers of innate signaling. Nat Immunol 7: 1237–1242.
  67. 67. Esser C, Ahmadinejad N, Wiegand C, Rotte C, Sebastiani F, et al. (2004) A genome phylogeny for mitochondria among alpha-proteobacteria and a predominantly eubacterial ancestry of yeast nuclear genes. Mol Biol Evol 21: 1643–1660.
  68. 68. Gupta RS (2005) Protein signatures distinctive of alpha proteobacteria and its subgroups and a model for alpha-proteobacterial evolution. Crit Rev Microbiol 31: 101–135.
  69. 69. Moreno E, Stackebrandt E, Dorsch M, Wolters J, Busch M, et al. (1990) Brucella abortus 16S rRNA and lipid A reveal a phylogenetic relationship with members of the alpha-2 subdivision of the class Proteobacteria. J Bacteriol 172: 3569–3576.
  70. 70. Cirl C, Wieser A, Yadav M, Duerr S, Schubert S, et al. (2008) Subversion of Toll-like receptor signaling by a unique family of bacterial Toll/interleukin-1 receptor domain-containing proteins. Nat Med 14: 399–406.
  71. 71. Forestier C, Deleuil F, Lapaque N, Moreno E, Gorvel JP (2000) Brucella abortus lipopolysaccharide in murine peritoneal macrophages acts as a down-regulator of T cell activation. J Immunol 165: 5202–5210.
  72. 72. Salcedo SP, Marchesini MI, Lelouard H, Fugier E, Jolly G, et al. (2008) Brucella control of dendritic cell maturation is dependent on the TIR-containing protein Btp1. PLoS Pathog 4: e21.
  73. 73. Francis CL, Starnbach MN, Falkow S (1992) Morphological and cytoskeletal changes in epithelial cells occur immediately upon interaction with Salmonella typhimurium grown under low-oxygen conditions. Mol Microbiol 6: 3077–3087.
  74. 74. Rouot B, Alvarez-Martinez MT, Marius C, Menanteau P, Guilloteau L, et al. (2003) Production of the type IV secretion system differs among Brucella species as revealed with VirB5- and VirB8-specific antisera. Infect Immun 71: 1075–1082.
  75. 75. Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680–685.
  76. 76. Romero C, Gamazo C, Pardo M, Lopez-Goñi I (1995) Specific detection of Brucella DNA by PCR. J Clin Microbiol 33: 615–617.
  77. 77. Moreno E, Berman DT, Boettcher LA (1981) Biological activities of Brucella abortus lipopolysaccharides. Infect Immun 31: 362–370.
  78. 78. Hudson L, Hay FC (1976) Practical Immunology. Oxford: Blackwell Scientific.
  79. 79. Gruber A, Mancek M, Wagner H, Kirschning CJ, Jerala R (2004) Structural model of MD-2 and functional role of its basic amino acid clusters involved in cellular lipopolysaccharide recognition. J Biol Chem 279: 28475–28482.
  80. 80. Comerci DJ, Altabe S, de Mendoza D, Ugalde RA (2006) Brucella abortus synthesizes phosphatidylcholine from choline provided by the host. J Bacteriol 188: 1929–1934.
  81. 81. Fretin D, Fauconnier A, Kohler S, Halling S, Leonard S, et al. (2005) The sheathed flagellum of Brucella melitensis is involved in persistence in a murine model of infection. Cell Microbiol 7: 687–698.
  82. 82. Tibor A, Wansard V, Bielartz V, Delrue RM, Danese I, et al. (2002) Effect of omp10 or omp19 deletion on Brucella abortus outer membrane properties and virulence in mice. Infect Immun 70: 5540–5546.
  83. 83. Raetz CR, Reynolds CM, Trent MS, Bishop RE (2007) Lipid A modification systems in gram-negative bacteria. Annu Rev Biochem 76: 295–329.