Purification of OMVs was performed according to the protocol described by Avila-Calderón et al. (2012). In short, the B. melitensis 16M and VTRM1 strains were cultured on tryptic soy agar plates supplemented with 0.5% yeast extract (TSA-YE) (Becton Dickinson-BD^TM) and incubated 24 h at 37°C. The bacteria were harvested with a rubber policeman and suspended in 250 mL sterile PBS 0.1M (Gibco^®). The bacterial suspension was centrifuged at 10,000 × g for 30 min (Thermo Scientific^TM Sorvall^TM Legend^TM XT/XF.). The supernatant was decanted through a 0.22 μm filter (Millipore Corp.), and a sterility test was performed by culturing an aliquot on a TSA-YE plate followed by incubation for 1 week at 37°C. The sterile supernatant was centrifuged at 100,000 × g for 2 h at 4°C (Beckman Coulter Optima L-90K). The pellet containing the isolated OMVs was washed twice using 25 mL sterile PBS and OMVs were suspended in 1 mL sterile PBS (OMVs). Total protein concentration was determined using the PIERCE-BCA (Thermo Fisher Scientific Inc.) reagents, following the manufacturer’s recommendations. OMVs were purified through a density-gradient with the aid of the OptiPrep protocol (Sigma-Aldrich, Inc.), following the Fernandez-Moreira et al. (2006). Concisely, sterile PBS-OptiPrep solutions were prepared at final concentrations of 10, 15, 20, 25, and 30%. A volume of 2.6 mL from these OptiPrep solutions were subsequentially layered in an ultracentrifuge tube from the highest to the lowest density. OMVs were placed at the bottom of the tube. Tubes were centrifuged at 100,000 × g for 16 h at 4°C. Finally, OMVs could be observed as an opalescent band within the density gradient. Bands containing OMVs were collected, washed twice using sterile PBS at 100,000 × g for 2 h at 4°C and they were ultimately resuspended in 500 μL PBS. OMVs were divided into 0.5 mL aliquots and stored at −20°C until usage. These purified OMVs were used in all the experiments.

Twenty-five microliters from the OMVs suspension (25 μg of protein) from both strains were placed onto copper grids coated with formvar and they were subsequently dried using filter paper. Then, 1% phosphotungstic acid was added onto the samples. The grids were allowed to dry for 10 h and they were observed with the aid of a transmission electron microscope (JEOL model JEM 10-10). In order to observe OMVs released from whole bacteria, cells from the B. melitensis strain 16M and B. melitensis VTRM1 were grown on TSA plates for 36 h at 37°C. Subsequently, molten soft agar was poured onto the plates to cover the growth. Once the agar had solidified, small agar cubes (2 mm) were cut. Blocks were fixed using 2.5% glutaraldehyde in PBS, rinsed with Sorensen’s PBS, dehydrated with ethanol, and prepared for transmission electron microscopy (TEM). Thin-section preparations were stained with OsO4, and they were observed using the JEOL model JEM 10–10, transmission electron microscope. Images were obtained using the aforementioned transmission electron microscope at ENCB’s (IPN, Mexico City, Mexico) Microcopy Facility.

The SDS-PAGE protocol was performed using 15% acrylamide gel slabs, in accordance with Laemmli’s method (Laemmli, 1970). The gel slabs were stained with Coomassie blue and Silver Stain Plus Kit (Biorad). The apparent molecular masses of proteins contained in purified OMVs were determined by comparing their electrophoretic mobility against that of wide range molecular mass markers, using the SigmaGel^® V1.0 computer program.

After separation of proteins contained in OMVs through SDS-PAGE, the acrylamide gel was stained with Coomassie blue. The gel was then cut into four sections. Each gel section was analyzed as follows: Adjustment of the instrument parameters was performed with a Calmix solution containing N-butylamine, caffeine, Met-Arg-Phe-Ala (MRFA) and Ultramark 1621 (Pierce LTQ Velos ESI Positive Ion Calibration Solution). These reagents are used to calibrate the LTQ Veils module with ion trap and the Orbitrap module with FT (Fourier Transform) mass detector in the ESI positive ionization mode. The N-butylamine is used to extend the mass calibration at lower m/z (73.14 Da) values. This type of calibration allows for the determination of molecular masses with accuracy variations no higher than 5 ppm. Each gel section was reduced with dithiothreitol (Sigma-Aldrich; St Louis, MO, United States), which was alkylated with iodoacetamide (Sigma-Aldrich) and digested “in gel” with Trypsin (Promega Sequencing Grade Modified Trypsin; Madison, WI, United States). The digestion reaction was performed in a solution containing 50 mM ammonium bicarbonate (pH 8.2) and incubated for 18 h at 37°C. Peptides produced by enzymatic cleavage were desalinated using Zip Tip C18 (Millipore; Billerica, MA, United States) and applied into an LC-MS (Liquid Chromatography-Mass Spectrometry) system composed of an ACCELA pump (Thermo-Fisher, San Jose, CA, United States) coupled to an LTQ-Orbitrap Velos mass spectrometer (Thermo-Fisher Co., San Jose, CA, United States) with a nano-electrospray (ESI) ionization source. For the nano-flow liquid chromatography in line, a gradient system containing 5–80% solvent B (acetonitrile with 0.1% formic acid) was used with a set run-time of 120 min using an RP-C18 capillary column (0.75 μm internal diameter and 20 cm long). The system flow was set to 300 nL/min. The total ion scan (Full Scan) was performed on the Orbitrap analyzer with a resolution mass of 60,000. Peptide fragmentation was performed using the CID (Collision-Induced Dissociation) and the ion trap (IT) methods. These fragmentation methods are used because they provide more information about the peptide. All spectra were acquired using the positive detection mode. Execution and capture of fragmentation data were performed depending on the total ion scan according to pre-determined charges (ions with z2⁺, z³⁺, and z⁴⁺) charges were fragmented with an isolation width of 2.0 (m/z), normalized collision energy of 35 arbitrary units, Q activation of 0.250, activation time of 10 ms and maximum injection time of 10 ms per micro-scan. During automatic data capture, the dynamic ion exclusion was used as following: (i) 500 ion exclusion list, (ii) 30 s pre-exclusion time and (iii) 90 s exclusion time. Protein identification was performed by inputting spectrometric raw-data into the Proteome Discoverer v1.4 program. (Thermo-Fisher Co., San Jose, CA, United States) by means of the Sequest HT search engine. As for protein identity, the B. melitensis 16M genome was downloaded from the Uniprot database. An FDR-False Discovery Rate (Minimum) of 0.01 and FDR 0.05 (Maximum) was used in addition to the inverted database (Decoy database) as a tool for the “Percolator” validation program. The maximum tolerance of molecular mass difference for the precursor ion when compared to the theoretical versus experimental values (precursor mass tolerance) was 20 ppm and the tolerance for fragments obtained by dissociation of the precursor ion (fragment mass tolerance) was 0.6 Da. Constant modifications [carbamidomethylation of cysteines (C) and variables such as oxidation of methionines (M) and deamination of asparagine (N) and glutamine (Q)] were established for the automatic search. The samples were processed in duplicate.

Hits obtained from proteomic identification were analyzed with the aid of the by BlastP database, using the B. melitensis 16M genome (obtained from NCBI¹). OrthoVenn2 analysis for enrichment of gene ontology terms was also used.² Each protein along with its putative function was looked for in the PSORTb v3.0. ExPASy Bioinformatics Resource Portal³ and ProtCompB⁴ were used to determine the subcellular location of each protein. The protein Homology/analogy Recognition Engine V 2.0 (PHYRE2) database was used to find homologous proteins.⁵ The pathway enrichment was made using the KEEG database and the GhostKOALA tool.⁶

Outer membrane vesicles protect proteins that could be delivered to the host cells via fusion of the vesicles with the host cell or through endocytosis (McCaig et al., 2013). The purpose of this experiment was to evaluate resistance of OMVs against the effect of detergents and proteinase K. Purified OMVs (100 μg) from the B. melitensis 16M and VTRM1 strains were treated separately with 10 μL of proteinase K (Thermo Fisher Scientific) (10 mg/mL), SDS (0.02%), proteinase K 10 μL (10 mg/mL) plus SDS (0.5%) (Sigma-Aldrich, Inc), and sodium deoxycholate (0.5%) (Sigma-Aldrich, Inc). OMVs subjected to the different treatments were incubated at 37°C for 2 and 24 h. After incubation, phenylmethylsulfonyl fluoride (PMSF; Merck Millipore) (0.1 mM) was added to inhibit proteinase K, then all the samples were processed for SDS-PAGE analysis and subsequently stained with Silver Stain Plus Kit (Biorad) following the manufacturer’s recommendations.

Peripheral blood mononuclear cells (PBMCs) were separated from buffy coats provided by healthy donors. In short, blood units were diluted with sterile PBS and layered into a centrifuge tube with an equal volume of Ficoll-Paque Premium medium (GE Healthcare). Blood was centrifuged at 2, 000 × g for 25 min at room temperature, followed by two consecutive whases. The cells were then counted using a Neubauer chamber and adding Trypan blue. The cell suspension was adjusted a 1 × 10^6/mL concentration in RPMI (Gibco^®) supplemented with 10% fetal calf serum (FCS), 50 μg/mL gentamicin (Gibco^®) and 2.5 μL/mL Fungizone (Gibco^®). PBMCs platting was adjusted to 1 × 10^6/mL/well, using a 24-well plate, and incubated under a CO₂ atmosphere. After that, PBMCs were treated with different concentration of OMVs. For each experiment, PBMCs were obtained fresh and they were incubated overnight under same conditions. Healthy donors signed a written consent to participate in this study.

The apoptosis, DNA damage and cell proliferation kit was used as per manufacturer’s instructions (BD Pharmingen^TM). PBMCs were stimulated with either a 1, 10, and 25 μg/mL suspension of OMVs or 10 μL a of 1 mM Carbonyl cyanide 3-chlorophenylhydrazone solution (CCCP; Abcam) (de Graaf et al., 2004), the latter was used as a positive control for apoptosis. Furthermore, a 10 μg/mL suspension prepared using Escherichia coli-LPS (Sigma-Aldrich, Inc.) was used as a positive control for cell proliferation for a period of 24 h. Subsequently, bromodeoxyuridine (BrdU) was added to the cell suspension, followed by incubation for 1 h at a final 10 μM concentration. The cells were collected, washed using FACS buffer (PBS, 1% FCS, sodium azide 0.01%), and stained with monoclonal antibodies (mAbs) according to the manufacturer’s instructions: anti-human H2AX-Alexa Flour 647, and anti-cleaved PARP-PE. PARP (Poly [ADP-Ribose] Polymerase) is a nuclear chromatin-associated enzyme that is involved in DNA repair. During apoptosis, Caspase-3 cleaves PARP, resulting in its inactivation and a general inability of cells to repair DNA damage. For this reason, the 89-kDa-cleaved PARP fragment serves as a marker for cellular apoptosis. A total number of 50,000 events were acquired using the LSRFortessa cytometer (BD Biosciences), and the resulting data was analyzed with the aid of the FlowJo^® software v.10 (FlowJo, LLC, Ashland, OR, United States).

Peripheral blood mononuclear cells were stimulated with either a 1, 10, or 25 μg/mL suspension of OMVs from both strains or a 5 μg phytohemagglutinin (PHA) solution for 12, 24, and 48 h. Plates were incubated at 37°C under 5% CO₂. Monoclonal antibodies (mAbs) coupled to the following fluorochromes were used: anti-human CD91-eFluor 660 (eBioscience), anti-human CD3-APC (BD Pharmingen^TM), anti-human CD19-PE-Cy7 (BD Biosciences), anti-human PD-L1-PE (BioLegend), anti-human PD-1-FITC (BioLegend), anti-human CD86-PE (BioLegend), anti-human CD69-TRI-COLOR (Molecular Probes). After stimulation, 100 μL of supernatant were collected from wells and stored at −70°C until usage. Then, PBMCs were collected and washed with FACS buffer, followed by incubation with diluted mAbs in FACS buffer for 1 h at 4°C (protecting the reaction from light). Finally, cells were washed with FACS buffer, fixed with PBS-PFA (Paraformaldehyde) 1%, and resuspended in 400 μL of FACS buffer. Samples were analyzed in LSRFortessa^TM cytometer (BD Biosciences), providing a total number of 30,000 events, and the resulting data was analyzed with the aid of the FlowJo^® software.

In order to assess the potential regulatory effects of OMVs from B. melitensis on the expression of activation and inhibition surface markers, PBMCs were pre-incubated together with either OMVs or PHA. Briefly, 1 × 10⁶ cells^/mL/well were incubated together with either 25 μg of OMVs from both strains or 5 μg of PHA with the supplemented medium at 37°C under 5% CO₂ for a period of during 12 h. Then, the medium was replaced with freshly prepared supplemented medium. The cells that had been pre-incubated with OMVs were re-stimulated with 5 μg of PHA to evaluate the potential inhibitory effects of OMVs on the expression of surface markers. To determine possible regulatory effects of OMVs on the expression of the surface markers, cells pre-incubated with PHA were re-stimulated with 25 μg of OMVs from both strains. The expression of the activation (CD69, CD86) and inhibition (PD-1, PD-L1) surface markers were measured by flow cytometry as mentioned above.

Cytokine quantification of Th1/Th2/Th17 cytokines; IL-2, IL-4, IL-6, IL-10, TNF, IFNγ, and IL-17A was performed with the aid of the CBA Human Th1/Th2/Th17 cytokine kit (BD Biosciences, San Diego, CA United States) and using the supernatants of cells that had been stimulated with either OMVs or PHA for 24 h, following the manufacturer’s instructions. Samples were read by using the LSRFortessa^TM cytometer (BD Biosciences) with CBA template, a total of 2000 events were recorded, and the resulting data was analyzed with the FCAP Array^TM software v3.0 (BD Biosciences).

Peripheral blood mononuclear cells were adjusted to 1 × 10⁶ and they were plated on sterile cover slips, followed by stimulation with 25 μg/mL of either OMVs or PHA for 2 and 6 h. Then, the cells were fixed using increasing concentrations of PFA (0.5, 1, and 2%) in PBS for 30 min. PBMCs were fixed onto cover slips by duplicated and permeabilized with Triton X-100 0.5% in PBS. Subsequently, one set of cells were treated with PBS/BSA 1% plus human serum 1% during 1 h at room temperature and it was later washed with PBS/BSA 0.2%. After that, 50 μL of mAb anti-tubulin (Biolegend) diluted 1:100 in PBS/BSA 1%, were added to the cells, to be incubated overnight at 4°C in a humid chamber. Cover slips were washed three times and incubated with mAb goat anti-mouse IgG Alexa fluor-488 (Invitrogen^TM, Thermo Fisher Scientific Inc.) during 1 h at room temperature, under darkness. On the other hand, the other remained set of cover slips of fixed cells were incubated with rhodamine-phalloidin (Molecular Probes^TM) diluted 1:100 in PBS during 30 min under darkness. Both sets of cover slips were washed and stained with DAPI 1 mg/mL diluted 1:10 in PBS for 1 min. Cover slips were washed and mounted using DABCO solution. Cells were analyzed with the aid of a confocal microscopy Leica TCS SP8 AOBS (Acousto-Optical Beam Splitter) DMI6000 (Leica Microsystems, Germany). A minimum of three fields were captured for analysis and nuclei were measured with ImageJ software.

The present work was approved by the ethic committee of Escuela Nacional de Ciencias Biológicas, Instituto Politécnico Nacional with the approval number 4265. Written informed consent was signed by healthy donors.

Two-way ANOVA analysis with Bonferroni post-tests and One-way ANOVA analysis with Tukey/Dunnett post-test were used. GraphPad Prism V5.01 was used for statistical calculations.

Outer membrane vesicles released from the whole cell of B. melitensis 16M and VTRM1 and purified vesicles were observed as showed in the micrographs (Figure 1). Protein profiles of OMVs purified from B. melitensis 16M and VTRM1 displayed bands ranging from 10 to 88 kDa, along with two notably major bands of 20 and 23 kDa (Figure 2A). OMVs from B. melitensis 16M treated with SDS, sodium deoxycholate and a mixture of proteinase K and SDS did not show changes in protein profile. Only a prominent band of 45 kDa was consistently observed in OMVs after both treatments (Figures 2B,D). A 45 kDa band was also observed in OMVs from B. melitensis VTRM1, treated with SDS and proteinase K (Figures 2C,E). The protein profile of OMVs of B. melitensis VTRM1, treated with proteinase K and SDS for 2 h, yielded two 16–18 kDa bands (Figure 2C). OMVs belonging to B. melitensis 16M, treated with deoxycholate, and a combination of proteinase K and SDS for 24 h showed a decreased in intensity of the bands corresponding to 18–20 kDa (Figure 2D). OMVs from the VTRM1 strain, treated with proteinase K for 24 h, displayed bands ranging from 14 to 65 kDa, while the resulting bands obtained from treatment with proteinase K together with SDS, showed several low-molecular-weight bands which are indicate the presence of partially degraded proteins. On the other hand, OMVs from the VTRM1 strain, treated with SDS or deoxycholate for 24 h, also affected proteins. However, two major bands in the vicinity of 18 and 20 kDa were still evident after SDS treatment (Figure 2E). In general, OMVs from B. melitensis 16M were more resistant to proteinase K as well as detergent treatment, whereas OMVs from B. melitensis VTRM1 were more sensitive.

The resulting peptide sequences obtained from the mass spectrometry analysis were used to scan through the databases. This analysis revealed 51 hits (proteins) for B. melitensis VTRM1 and 254 hits for B. melitensis 16M. A scan result was only accepted if the score or the coverage was higher than 20 and at least two tryptic peptides as well as their fragment ions matched the protein (Avila-Calderón et al., 2012, 2018). These hits were analyzed with the BlastP tool using the genomic sequence of B. melitensis 16M obtained from the NCBI. The proteins unambiguously identified were 131 for the B. melitensis 16M and 43 for the B. melitensis VTRM1 (Supplementary Table S1). The proteins contained in OMVs from both B. melitensis 16M and VTRM1 were identified according to aminoacid length, molecular weight (Mw), isoelectric point (pI), locus tag, subcellular location and COG (cluster of orthologous groups) functional classification (Supplementary Table S1). Supplementary Table S2, shows the proteins identified in two prominent bands obtained from the electrophoretic profile of OMVs from both strains, corresponding to 20 and 23 kDa. Any uncharacterized or hypothetical proteins lacking COG classification, were annotated along with the following information: string relation, protein domains (motifs) or homologous proteins found using the PHYRE2 analysis to characterize or assign a putative function. According to the subcellular location of proteins identified in OMVs, some notable differences regarding proportions were observed. Identified proteins in OMVs from the 16M strain were 29% cytoplasmic, 19% periplasmic, 37% outer membrane, 11% inner membrane, and 4% extracellular. In the case of OMVs from VTRM1, the identified proteins were 67% cytoplasmic, 5% periplasmic, 14% outer membrane, 14% inner membrane and no proteins from the extracellular location were found (Figure 3A). In order to compare the identified proteins found in OMVs from the smooth and rough strains, an analysis of orthologous proteins was performed. Proteins found in OMVs from B. melitensis16M were grouped into 25 clusters (22 proteins), while proteins contained in OMVs from B. melitensis VTRM1 were classified into 23 clusters. OMVs from both strains shared a total of 22 clusters, and only 3 clusters were exclusive for OMVs from B. melitensis16M (Figure 3B). The proteins shared in OMVs from both species are: Lon protease, Omp25, Omp31, and Omp16 (Supplementary Table S1). The 22 orthologous clusters shared in OMVs from both strains were grouped based on their biological-process ontology into five groups considering translation as the most important ontology group (GO: 0006412). Based on the molecular function process, orthologous proteins were grouped into nine groups and the most important ontologies were nucleic acid-binding (GO: 0003676) and structural molecule activity (GO: 0005198) (Supplementary Figure S1). Moreover, an uncharacterized orthologous protein was identified by this research team, BMEI0542, which showed a high degree of homology compared to OprG from Pseudomonas aeruginosa (98.8% confidence by PHYRE2). The latter protein has been associated to iron-uptake and cytotoxicity on the human bronchial epithelial cell line (McPhee et al., 2009) (Supplementary Table S1). A total of 103 proteins were found in OMVs from B. melitensis 16M and 21 proteins in OMVs from VTRM1 which were not classified into the orthologous clusters (singletons). The enrichment pathway showed a high number of proteins in OMVs from B. melitensis 16M that are related to genetic-information processing, environmental-processing information, signaling, cellular processes and carbohydrate metabolism. In the case of proteins identified in OMVs from the VTRM1 strain, these were mainly related to genetic-information processing, signaling and cellular processes (Figure 3C).

It has previously been demonstrated that purified, lipidated-Omp19, from B. abortus 2308 induces apoptosis in T lymphocytes (Velásquez et al., 2012). On the other hand, B. abortus LPS induces DNA damage through the production of reactive oxygen species in polymorphonuclear cells (Barquero-Calvo et al., 2015). LPS and Omp19 are present in OMVs produced by B. melitensis, thus, both the analysis of apoptosis induction and DNA damage was performed. To determine whether OMVs display a cytotoxic-effect upon host cells, DNA damage and apoptosis were analyzed using PBMCs stimulated with OMVs from both B. melitensis strains. DNA damage was analyzed through the expression of the phosphorylated histone H2AX, while the induction of apoptosis was evaluated by measuring expression of the cleaved PARP [Poly (ADP-ribose) polymerase-1]. OMVs from both smooth and rough B. melitensis strains did not induce the expression of either cleaved PARP (cPARP) or phosphorylated H2AX on gated cells from the lymphocyte and monocyte region. Although gated cells from the monocyte region showed slight increase in cPARP expression, it was not statistically significant (one-way ANOVA with Dunnett post-test, 95% confidence interval) compared with the unstimulated control cells (Supplementary Figures S2A,B). These results demonstrated that OMVs show no cytotoxic effect in PBMCs. Moreover, OMVs did not induce cell proliferation in gated cells from the monocyte and lymphocytic regions (Supplementary Figure S2C).

Sonicated B. melitensis Rev1 induced lymphocyte activation through CD69 expression (Almajid, 2011). On the other hand, recombinant Omp25, identified in OMVs from B. melitensis, decreased IL-12 production through PD-1 signaling in monocyte/macrophages (Cui et al., 2017). Thus, it was expected that OMVs produced by B. melitensis induce the expression of surface activation or inhibition markers. Results showed no differences in the expression of either the activation (CD69 and CD86) or inhibition (PD-1 and PD-L1) markers on PBMCs stimulated with OMVs at the different time intervals tested (two-way ANOVA with Bonferroni post-test, 95% confidence interval) (Supplementary Figures S3, S4). The percentage of CD3⁺CD69⁺, CD19⁺CD69⁺, CD3⁺PD-1⁺, and CD19⁺PD-1⁺ cells showed an increase after 24 h and a decrease 48 h after treatment with different concentrations of OMVs; however these changes were not statistically significant. To determine how the manner in which OMVs from B. melitensis regulate the expression of activation and inhibition cell surface markers on PBMCs, the cells were pre-incubated with OMVs before a subsequent PHA-treatment, which showed that the number of T cells (CD3⁺) expressing CD69 decreased. It should be noted that significant differences was observed only when cells were treated with OMVs from the VTRM1 strain compared with cells stimulated with PHA in absence of OMVs (p < 0.005) (one-way ANOVA with Tukey post-test, 95% confidence interval). Additionally, the number of T cells expressing the PD-L1 marker decreased significantly when PBMCs were pre-incubated with OMVs from both strains, in comparison with PHA stimulated cells (p < 0.05) (Figures 4A–D). Pre-incubation with OMVs not only decreased the number of CD3⁺ PD-L1⁺ cells, but also the expression level of the PD-L1 molecule expressed by mean fluoresce intensity (MFI) in gated T-cells (p < 0.05) (Figures 4B,C). Conversely, incubation of PBMCs with OMVs from B. melitensis 16M prior to or after the stimulation with PHA, increased the number of B cells (CD19⁺) expressing PD-1, in comparison with the positive control (p < 0.005) (Figures 4E–G). Therefore, OMVs seem to exert an inhibitory effect on the expression of the surface markers in both T and B cells, which in turn modulates the expression of the CD69 and PD-1/PD-L1 molecules. OMVs decreased the number of CD3⁺ CD86⁺, CD3⁺ PD-1⁺, and CD91⁺ PD-1⁺ cells but not to a significant degree (Supplementary Figures S5, S6).

To determine whether the vesicles drive the immune response through a Th1, Th2 or Th17 profile, the production of each specific cytokine was measured. Figure 5 shows that OMVs produced by both B. melitensis strains did not induce the production of either the Th2- (IL-4 and IL-10) or Th1-type cytokines (IL-2, IFNγ) (one-way ANOVA with Tukey post-test, 95% confidence interval). IL-17A decreased in the presence of different concentrations of OMVs, while production of TNFα and IL-6 increased using 10 and 25 μg/mL of OMVs from B. melitensis 16M (p < 0.005, p < 0.05 respectively), and 25 μg/mL of OMVs from the rough VTRM1 strain induced cytokine production (p < 0.005) (Figure 5).

It has previously been reported that Brucella is able to induce cytoskeleton rearrangements (Czibener et al., 2016). In order to evaluate the effect of OMVs release by Brucella on the cytoskeleton ultrastructure, PBMCs were treated with OMVs and noticeable morphological changes were observed by confocal microscopy. After incubation with OMVs for 2 h, PBMCs showed small actin foci within the cells and the nucleus showed a different affinity to the DAPI stain (Figures 6A,B). After 6 h stimulation with OMVs, the PBMCs showed evident actin and tubulin depolymerization, as well as morphological changes in the nucleus. The observed cells displayed an enlarged nucleus and different affinity for the DAPI stain. Moreover, the microtubule-organizing center was not well-defined in some cells that had been stimulated with OMVs (Supplementary Figures S7A,B). Even though these changes were observed in cells treated with OMVs from both strains, the effect was more evident in cells treated with OMVs from the rough strain. The nucleus areas of the treated cells were measured and after incubation with OMVs for about 2 h, the nuclei exhibited a significant increase in area size, though only when using OMVs from the smooth strain (2 h, p < 0.001; 6 h, p < 0.005) (Figures 6C,D) (one-way ANOVA with Tukey post-test, 95% confidence interval).