Alvarez et al. BMC Microbiology (2019) 19:166 https://doi.org/10.1186/s12866-019-1534-3 RESEARCH ARTICLE Open Access Extracellular vesicles and soluble factors secreted by Escherichia coli Nissle 1917 and ECOR63 protect against enteropathogenic E. coli-induced intestinal epithelial barrier dysfunction Carina-Shianya Alvarez1,2,3, Rosa Giménez1,2, María-Alexandra Cañas1,2, Rodrigo Vera1,2, Natalia Díaz-Garrido1,2, Josefa Badia1,2*† and Laura Baldomà1,2*† Abstract Background: Enteric pathogens have developed mechanisms to disrupt tight junctions and increase gut permeability. Many studies have analysed the ability of live probiotics to protect intestinal epithelial cells against tight junction damage caused by bacterial pathogens. Escherichia coli Nissle 1917 (EcN) is among the probiotics that positively modulates the intestinal epithelial barrier by regulating expression and distribution of tight junction proteins. We previously reported that regulation of ZO-1, claudin-14 and claudin-2 is mediated by EcN secreted factors, either free-released or associated with outer membrane vesicles (OMVs). Factors secreted by commensal ECOR63 elicited comparable effects in intact epithelial T-84 and Caco-2 cell monolayers. Results: Here we analyse the ability of OMVs and soluble secreted factors to protect epithelial barrier function in polarized T-84 and Caco-2 cells infected with enteropathogenic Escherichia coli (EPEC). Transepithelial electrical resistance, paracellular permeability, mRNA levels and subcellular distribution of tight junction proteins were monitored in the absence or presence of EcN and ECOR63 extracellular fractions. EPEC downregulated expression of ZO-1 ZO-2, occludin and claudin-14 and altered the subcellular localization of ZO-1, occludin and F-actin cytoskeleton. OMVs and soluble factors secreted by EcN and ECOR63 counteracted EPEC-altered transepithelial resistance and paracellular permeability, preserved occludin and claudin-14 mRNA levels, retained ZO-1 and occludin at tight junctions in the cell boundaries and ameliorated F-actin disorganization. Redistribution of ZO-1 was not accompanied by changes at mRNA level. Conclusion: This study provides new insights on the role of microbiota secreted factors on the modulation of intestinal tight junctions, expanding their barrier-protective effects against pathogen-induced disruption. Keywords: Outer membrane vesicles, Escherichia coli Nissle 1917, ECOR63, EPEC, Tight junctions, Intestinal epithelial barrier * Correspondence: josefabadia@ub.edu; lbaldoma@ub.edu † Josefa Badia and Laura Baldomà contributed equally to this work. 1 Secció de Bioquímica i Biología Molecular, Departament de Bioquímica i Fisiologia, Facultat de Farmàcia i Ciències de l’Alimentació, Universitat de Barcelona, Barcelona, Spain Full list of author information is available at the end of the article © The Author(s). 2019 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Alvarez et al. BMC Microbiology (2019) 19:166 Page 2 of 12 Background the EPEC-mediated effects, which eventually leads to diar- The intestinal epithelium constitutes a physical and bio- rhea [9, 12, 13]. chemical barrier that limits translocation of luminal mi- Barrier disruption caused by enteric pathogens can be crobes, antigens, and toxins into host tissues. Several prevented or counteracted by certain probiotics and gut mechanisms are involved in this protective function, in- beneficial microbes [14–17]. One example is the pro- cluding production of an extracellular mucin layer, se- biotic Escherichia coli Nissle 1917 (EcN) with proven ef- cretion of antimicrobial peptides and the formation of fectiveness in the treatment of inflammatory intestinal intercellular tight junctions that seal the epithelial layer disorders [18–20] and acute diarrhea [21]. The EcN gen- against the luminal content, while preserving and con- ome encodes a great variety of interference and fitness trolling the passage of small ions and water-soluble sol- factors that enhance EcN survival in the gut (healthy or utes through the paracellular space. The intestinal inflamed) and play an important role in the beneficial ef- epithelium is also crucial to orchestrate host immune re- fects of this probiotic [22]. There is wide scientific evi- sponses to microbiota-derived signals [1, 2]. dence that EcN has immunomodulatory effects, mainly Adjacent epithelial cells are joined by several types of due to its ability to trigger activation/inactivation mech- intercellular connections that include tight junctions (TJs), anisms of the immune response that lead to a positive adherens junctions (AJs), and desmosomes. TJs are in the balance between pro- and anti-inflammatory local cyto- apical part of the polarized epithelium and regulate the kines [22, 23]. In addition, EcN positively modulates the paracellular permeability of the epithelial monolayer, intestinal epithelial barrier by triggering upregulation whereas AJs and desmosomes fulfil intercellular communi- and redistribution of the TJ proteins ZO-1, ZO-2 and cation functions [1]. TJs are formed by several types of claudin-14 [24–26]. transmembrane proteins such as occludin, claudins, junc- To identify the EcN factors that mediate regulation of tional adhesion molecules and tricellulin, as well as cyto- TJs, our group was focused on the study of bacterial se- solic scaffold proteins such as zonula occludens (ZO) creted factors, which can easily diffuse through the proteins (ZO-1, ZO-2 and ZO-3), which in turn anchor the mucin layer before reaching intestinal epithelial cells [27, integral membrane proteins to the actin cytoskeleton. This 28]. Nowadays, membrane vesicles secreted by gut bac- bridging is essential for maintaining TJ barrier integrity [3, teria are considered relevant players in microbiota-host 4]. TJs are highly dynamic structures controlled by a great interaction, as they allow long distance delivery of mi- variety of signaling pathways in response to multiple stimuli crobial effectors directly to the host [29]. We showed that regulate the expression and/or subcellular location of that outer membrane vesicles (OMVs) released by the TJ proteins. The regulatory mechanisms include transcrip- probiotic EcN and the commensal ECOR63 are taken up tional control, membrane trafficking, and post-translational by intestinal epithelial cells [30] and modulate the epi- modifications that determine the ability of TJ proteins to thelial barrier integrity through several mechanisms. In establish specific protein-protein interactions [5–8]. T-84 and Caco-2 cell monolayers, vesicles from these Disruption of gut epithelial TJs results in enhanced intes- strains reinforce TJs through activation of ZO-1 and tinal permeability, a condition that positively correlates with claudin-14 expression, downregulation of the leaky pro- a wide variety of diseases [9]. Enteric pathogens such as en- tein claudin-2 and modulation of ZO-1 subcellular dis- teropathogenic Escherichia coli (EPEC) have evolved strat- tribution. Free-secreted factors also contribute to this egies to disrupt TJ structures as a mechanism for regulation [31]. In addition, EcN OMVs induce IL-22 ex- increasing gut permeability and helping their dissemination pression in human colonic explants [32]. IL-22 acts on into the host tissues [8]. EPEC is a non-invasive pathogen epithelial cells and enhances the protective function of whose virulence depends on a T3SS secretion system that the intestinal barrier by inducing the expression of inject virulence factors and effector proteins directly into mucin and antimicrobial proteins. In a murine model of the cytoplasm of the infected cell [10]. Translocated bacter- experimental colitis, oral administration of EcN OMVs ial effectors alter functions of the target cells by different counteract the altered expression of pro-inflammatory mechanisms, which involve modulation of cell signalling cytokines and markers of intestinal barrier function such pathways and changes in the architecture of cellular struc- as Trefoil factor-3 (TFF3) [33]. This pleiotropic peptide tures such as the actin cytoskeleton, the microtubule net- mediates protection and repair of the epithelium by sev- work that directs vesicle trafficking, or connections eral mechanisms that include redistribution of ZO-1 to between TJ proteins. An intricate combination of bacterial intercellular junctions [34, 35]. factors leads to the formation of attaching and effacing le- Based on previous studies showing that the barrier sions on the gut mucosa, a characteristic feature of EPEC function of epithelial cell monolayers was strengthened infection. Effectors such as EspF, Map and NleA contribute by EcN and ECOR63 OMVs and free-secreted factors to TJ disruption [11, 12]. The subsequent alteration of para- [31], here we evaluate the barrier protective effects of cellular permeability of the intestinal epithelium is one of these extracellular fractions in a cellular model of Alvarez et al. BMC Microbiology (2019) 19:166 Page 3 of 12 epithelial barrier damaged by EPEC infection. The re- (COF-SN). As both soluble and vesicle-associated factors sponse was assessed through transepithelial electrical re- secreted by the probiotic EcN and the commensal sistance (TER) and paracellular permeability assays, ECOR63 can strengthen the epithelial barrier in cellular expression analysis of TJ proteins by RT-qPCR, and ana- models that mimic the normal intestinal epithelium [31], lysis of F-actin and TJ protein redistribution by confocal we sought to test whether these extracellular fractions fluorescence microscopy. could protect against EPEC-induced damage in polarized intestinal epithelial cells. Epithelial barrier function was Results analysed by measuring TER and FD-4 flux as markers of Membrane vesicles and soluble factors secreted by EcN epithelial resistance and permeability, respectively. To and ECOR63 prevent EPEC-mediated epithelial barrier this end, Caco-2 and T-84 cell monolayers grown in disruption in T-84 and Caco-2 cell monolayers Transwell membrane supports were infected with EPEC Previous studies of our group showed that cell-free su- E2348/69 (MOI 100) for 3 h. In parallel, cell monolayers pernatants collected from EcN cultures prevent disrup- were infected with EPEC in the presence of OMVs (0.1 tion of Caco-2 monolayers caused by EPEC infection mg/ml) or COF-SN (0.5 mg/ml) obtained from DMEM [36]. Therefore, the protective effect was attributed to cultures of EcN and ECOR63 strains. The concentration secreted factors. Cell-free supernatants contain all the of both extracellular samples was selected according to factors secreted by bacteria, either released through previous dose-response curves. The same amounts were membrane vesicles (OMVs) or in a free-soluble form used in previous studies on intact cell monolayers [31]. Fig. 1 OMVs and free-soluble factors (COF-SN) secreted by EcN and ECOR63 maintain barrier function in EPEC-infected cells. Confluent T-84 and Caco-2 cell monolayers grown in Transwell supports were infected with EPEC (MOI of 100) for 3 h in the absence or presence of EcN or ECOR63 OMVs (0.1 mg/ml) or COF-SN (0.5 mg/ml). Non-treated cells were processed in parallel as a control. Epithelial barrier function was analysed by measuring TER (a) and FD-4 flux (b) as markers of epithelial resistance and barrier permeability. a TER values were measured before and after 3-h infection. Data are presented as percentage of changes in TER (decrease) from the initial value. b After 3-h infection, cell monolayers were washed and treated apically with FD-4 (1 mg/ml). The fluorescence in the basolateral chamber was measured before and 1 h after the addition of FD-4. FD-4 flux values were calculated by subtracting the fluorescence intensity units (FI) measured at 0 h. Data were expressed as fold-change compared with non-infected control cells, whose values were normalized as 1. In all panels, data are from three independent biological experiments performed in triplicate. The TER baseline control values were 1290 ± 98 Ω.cm2 for T-84 monolayers and around 920 ± 80 Ω.cm2 for Caco-2 cells. a, Significance against untreated control cells (p ≤ 0.05); b, significance against control EPEC-infected cells (p ≤ 0.05) Alvarez et al. BMC Microbiology (2019) 19:166 Page 4 of 12 Untreated cell monolayers were used as a control. EPEC The implication of the ERK pathway in the modulation infection significantly reduced the TER values of polar- of barrier integrity mediated by these bacterial extracel- ized cell monolayers compared to untreated control lular fractions was evaluated in T-84 monolayers treated cells. However, simultaneous apical stimulation with with the specific ERK 1/2 inhibitor U0126 (25 μM). After COF-SN or OMVs from both microbiota strains neutral- 15 min pre-treatment, T-84 cells were incubated with ized the decrease in TER caused by EPEC, with TER EPEC (MOI 100) alone or simultaneously with OMVs or values that did not significantly differ from those of non- COF-SN, and TER was measured at 3 h. In the presence infected control cells (Fig. 1a). of the inhibitor U0126, EcN and ECOR63 secreted frac- To examine whether OMVs and COF-SN isolated tions were unable to compensate for the decrease in from EcN and ECOR63 also counteracted EPEC-induced TER caused by EPEC (Fig. 2a). The protective effect of increased paracellular permeability, the transport of the all secreted fractions was virtually lost by ERK 1/2 inhib- paracellular marker FD-4 was measured in T-84 and ition, except for ECOR63-derived COF-SN. In this case, Caco-2 monolayers challenged by EPEC infection. In a tendency to neutralize the drop in TER was observed, both cell lines, the unidirectional flux of FD-4 was sig- although values did not reach statistical significance nificantly higher in cell monolayers incubated with EPEC (Fig. 2a). These results indicate that the reinforcement of than in controls. When cells were infected in the pres- the epithelial barrier promoted by EcN and ECOR63 se- ence of EcN or ECOR63 extracellular fractions, there creted factors depends, at least in part, on ERK 1/2 sig- was a significant decrease in FD-4 flux with values close nalling. The regulatory effects were not restricted to to those of the non-treated control cells (Fig. 1b). conditions of damaged epithelial barrier. In fact, parallel Fig. 2 Analysis of host and bacterial factors involved in the strengthening activity of EcN and ECOR63 secreted fractions. The strengthening activity of COF-SN and OMVs depends on the ERK1/2 signalling pathway both in EPEC-infected cells (a) and intact cell monolayers (b). a-b Before infection/stimulation with OMVs or COF-SN, T-84 cell monolayers were pre-treated for 15 min with the ERK1/2 inhibitor U0126 (25 μM). TER values were measured before and after 3-h treatment. c-d Effect of heat treatment on the strengthening activity of EcN and ECOR63 COF-SN. EPEC- infected (c) or intact (d) T-84 cell monolayers were incubated with heated COF-SN (h-COF-SN), and TER values were measured before and after 3- h treatment. In all panels, data are presented as percentage of changes in TER from the initial value from three independent biological experiments performed in triplicate. TER initial values were between 1100 and 1300 Ω.cm2. a, Significance against untreated control cells (p ≤ 0.02); b, significance against control EPEC-infected cells (a, c) or cells treated with EcN or ECOR63 control extracellular fractions (b, d) (p ≤ 0.05) Alvarez et al. BMC Microbiology (2019) 19:166 Page 5 of 12 experiments performed in T-84 cells that were not chal- Interestingly, claudin-2 was also significantly downregu- lenged by EPEC infection showed that inhibition of ERK lated by EPEC. The mRNA levels of claudin-1 were not 1/2 activity also abolishes the ability of EcN and significantly affected by EPEC infection, although a trend ECOR63 extracellular fractions to increase the TER of towards downregulation was observed compared to un- the cell monolayers (Fig. 2b). Again, under conditions of treated T-84 control cells. Exposure to OMVs or COF- intact epithelial barrier ERK 1/2 inhibition only resulted SN from either EcN (white bars) or ECOR63 (grey bars) in partial loss of the strengthening activity of COF-SN during infection counteracted the EPEC-induced down- from ECOR63. regulation of occludin and claudin-14, whose expression To discover the nature of the free-soluble secreted fac- levels remained close to those of non-infected control tors that prevent EPEC-mediated reduction of TER, COF- SN were heat-inactivated before being added to the apical compartment of T-84 monolayers. As shown in Fig. 2c, heat treatment abolished the protective activity of EcN su- pernatants as the TER values did not differ significantly from those of single EPEC-infected cells. In contrast, the protective effect of ECOR63 COF-SN was preserved after heat inactivation. In the presence of heated ECOR63 COF-SN, EPEC-infected cells displayed TER values like those of control cells. Proteinase K treatment of EcN and ECOR63 COF-SN yielded results that were comparable with those obtained by heat inactivation (not shown). Overall, these results suggested that a protein factor medi- ates the protective activity of EcN COF-SN, whereas other factors of a different nature might contribute in the case of the ECOR63 supernatant. These effects were confirmed in experiments performed in non-infected T-84 mono- layers incubated with heat-inactivated supernatants. Again, heat treatment completely abolished the strength- ening activity of EcN COF-SN on intact cell monolayers but partially preserved the activity of the ECOR63 super- natant fraction (Fig. 2d). Effect of OMVs and soluble factors secreted by EcN and ECOR63 on the expression of TJ proteins in T-84 cell monolayers challenged by EPEC infection The prevention of TER reduction in EPEC-infected epi- thelial cells by OMVs and COF-SN from EcN and ECOR63 pointed to counteracting changes in TJ proteins. First, we undertook RT-qPCR analyses to assess whether the protection mediated by these bacterial fractions corre- lated with changes in the expression of relevant TJ- proteins known to be regulated by EcN and ECOR63 se- creted fractions under intact barrier conditions, such as ZO-1, claudin-14 and claudin-2 [31], or dysregulated by Fig. 3 Effect of the secreted fractions (COF-SN and OMVs) from EcN EPEC infection, such as occludin and claudin-1 [16]. and ECOR63 on the expression of TJ proteins in T-84 monolayers For this, T-84 monolayers (9 days post-confluence) infected with EPEC. Intestinal epithelial cells were incubated for 3 h were infected with EPEC (MOI 100) for 3 h in the ab- with EPEC at a MOI of 100 (black bars). Parallel infections were carried out in the presence of COF-SN (0.5 mg/ml) or OMVs (0.1 mg/ sence or presence of OMVs or COF-SN. Untreated T-84 ml) from EcN (white bars) or ECOR63 (gray bars). The relative mRNA monolayers were processed as a control. The relative levels of the indicated proteins were measured by RT-qPCR using mRNA levels of TJ proteins were measured by RT- β-actin as the reference gene. Data are presented as fold-change qPCR, using the β-actin gene as a reference (Fig. 3). compared to untreated control cells (dotted line) from three EPEC infection significantly decreases mRNA levels of independent biological experiments. a, Significance against untreated control cells (p ≤ 0.05); b, significance against EPEC- ZO-1, ZO-2, occludin, and claudin-14, all proteins with infected control cells (p ≤ 0.04) barrier strengthening activity (Fig. 3, black bars). Alvarez et al. BMC Microbiology (2019) 19:166 Page 6 of 12 cells (Fig. 3). Both kinds of samples (OMVs or COF-SN), infected with EPEC, altered distribution of both proteins regardless of their origin (probiotic or commensal was observed compared to untreated control cells. strains), had similar protective effects on occludin and The ZO-1 signal was mainly found in the cell boundar- claudin-14 expression, as no statistically significant dif- ies of untreated control Caco-2 monolayers, whereas in ferences were observed between them. In contrast, none EPEC-infected cells this protein was translocated from the of the microbiota secreted fractions could prevent the cell membrane to the cytoplasm. The microbiota secreted negative regulation of ZO-1 and ZO-2 elicited by EPEC. fractions evaluated in this study (COF-SN and OMVs As expected, the low mRNA levels of claudin-2 in from EcN and ECOR63) showed a clear potential to EPEC-infected cells remained unchanged by OMVs or neutralize the EPEC effect, thus retaining the peripheral COF-SN treatment. These results indicate that secreted location of ZO-1 anchored to the cell membrane. factors released by these gut microbiota strains compen- Regarding occludin, EPEC infection triggered its dis- sate for the EPEC-triggered negative regulation of occlu- sociation from the TJ- structures. The main observation din and claudin-14 expression, but not that of ZO-1 or was discontinuous labelling of occludin in the cell mem- ZO-2. Preservation of both transmembrane TJ proteins brane, which indicates disruption of TJs and loss of could explain, at least in part, the protection against intercellular membrane contacts. In addition, a slight in- EPEC injury observed in TER assays. crease in the occludin signal in the cytoplasmic com- partment was detected. When cell monolayers were EcN and ECOR63 secreted factors trigger redistribution of infected with the pathogen in the presence of EcN or ZO-1 and occludin to tight junctions in EPEC-infected ECOR63 secreted fractions, occludin was maintained in epithelial cells the cell membrane with a staining profile comparable to After EPEC infection, occludin and ZO-1 shift from the that of untreated control cells. TJ-structures located at the epithelial cell membrane to Fluorescence microscopy was also used to analyse F- intracellular compartments. The altered distribution of actin cytoskeleton alterations. Phalloidin staining re- these TJ-proteins greatly contributes to disruption of the vealed a continuous distribution of F-actin at the cell intestinal epithelial barrier induced by this pathogen [12]. boundaries of control cells. In contrast, EPEC infection To assess whether COF-SN and OMVs from EcN and caused disorganization of the F-actin architecture. The ECOR63 could prevent displacement of these proteins in red-phalloidin fluorescence was dispersed and lost from cells infected with EPEC we carried out an immuno- peri-junctional areas. As for ZO-1 and occludin, incuba- fluorescence microscopy analysis in Caco-2 monolayers tion with OMVs or COF-SN from EcN and ECOR63 (5–7 days after seeding). Cells were incubated with EPEC partially reversed the EPEC-induced alteration of the F- (MOI 100) for 3 h as a control for pathogen-mediated actin cytoskeleton. The compensatory effect was most effects, and parallel infections were performed in the relevant for the OMVs fractions. The F-actin labelling presence of COF-SN or OMVs from EcN and ECOR63. pattern clearly differed from that of EPEC-infected cells Then, Caco-2 cells were fixed and stained with specific and tended to resemble the distribution pattern of the antibodies against occludin and ZO-1 (Fig. 4). In cells control group at the intercellular borders (Fig. 5). Fig. 4 Immunofluorescence staining of ZO-1 and occludin in Caco-2 cell monolayers incubated for 3 h with EPEC in the absence or presence of COF-SN or OMVs from the indicated bacterial strains. Analysis was performed by laser scanning confocal spectral microscope with 63x oil immersion objective lens, and images were captured with a Nikon color camera (8 bit). Images shown are representative from three independent biological experiments and are coded with Fire look-up table. Calibration bar is shown on the left. Scale bar, 20 μm Alvarez et al. BMC Microbiology (2019) 19:166 Page 7 of 12 Taken together, these results indicate that OMVs and 38]. In the context of gastrointestinal infections, many free-secreted factors released by the probiotic EcN and studies have investigated the beneficial effects of certain the commensal ECOR63 can protect the intestinal epi- commensal and probiotic strains in ameliorating intes- thelium against injury caused by EPEC through mecha- tinal inflammation and injury caused by several enteric nisms that include modulation of gene expression and pathogens [39, 40]. Enhancement of gut barrier function subcellular redistribution of TJ proteins. is among the mechanisms used by probiotics to exert their beneficial effects in the prevention and treatment Discussion of infectious diarrhoea [41]. In fact, certain probiotics Gut microbiota is essential to preserve intestinal homeo- protect intestinal epithelial cells against injury on TJs stasis and human health. A fundamental function of this caused by various bacterial pathogens [14–17]. microbial community is development and maintenance It is widely known that EPEC alters intestinal epithelial of the intestinal barrier. Imbalances in microbiota com- TJs, thus leading to epithelial barrier dysfunction and in- position may result in increased intestinal permeability, creased permeability [13]. Consistently, this pathogen a condition that is linked to a wide range of illnesses [37, causes reduction of TER in monolayers of several epithe- lial cell lines [11, 16, 25, 42, 43]. EPEC-mediated disrup- tion of TJs has been extensively studied in several in vitro and in vivo models. These studies revealed that EPEC per- turbs interaction between TJ proteins by promoting ZO-1, occludin, and claudin-1 redistribution from the intercellu- lar junctions to a cytoplasmic location and disorganizing the peri-junctional F-actin cytoskeleton [44–46]. Several bacterial effectors injected into host cells by the T3SS act synergistically to elicit posttranslational modifications, in- volving phosphorylation/dephosphorylation mechanisms and the control of intracellular protein trafficking, that de- termine the subcellular location of the TJ proteins [12]. Moreover, EPEC also triggers downregulation of genes en- coding ZO-1, occludin, and claudin-1 in the epithelial cell line NCM460 [16]. In addition to the barrier strengthening activity of EcN and ECOR63 secreted fractions (COF-SN and OMVs) on intact epithelial cell monolayers [31], the results presented here prove that these extracellular fractions also attenuate alterations in epithelial barrier function caused by EPEC infection. The ability of OMVs and COF-SN isolated from both E. coli strains from human gut microbiota to prevent the decrease in TER and the increase in paracellular per- meability in EPEC-infected T-84 and Caco-2 cell mono- layers reflected their potential to counteract lesions in TJ integrity induced by this pathogen. The ERK 1/2 signal transduction pathway is involved in the barrier protective effect of these microbial factors, as the specific inhibitor U0126 abolished their ability to reinforce intact epithelial barrier and neutralized the damage caused by EPEC in TER assays. However, some differences were observed be- tween EcN and ECOR63, specifically in the extracellular fraction that contains free-soluble secreted factors. In the presence of the ERK inhibitor, the ECOR63 COF-SN frac- Fig. 5 Fluorescence microscopy analysis of F-actin in Caco-2 cell tion still displayed residual barrier protective/strengthen- monolayers incubated for 3 h with EPEC in the absence or presence ing activity. This suggests that an ECOR63 free-secreted of COF-SN or OMVs from the indicated bacterial strains. F-actin was factor (not released by EcN) can modulate the epithelial stained with TRITC-labelled phalloidin (red) and analysis was barrier function through other signalling pathway(s). This performed as described for Fig. 4. Images shown are representative observation is consistent with previous results, which of three independent biological experiments. Scale bar: 10 μm showed that, in addition to free-secreted factors common Alvarez et al. BMC Microbiology (2019) 19:166 Page 8 of 12 to both strains, ECOR63 releases other soluble factors that As stated above, the epithelial TJs are highly dynamic have a great impact on the epithelial barrier. The common structures whose strength is regulated by multiple stimuli. secreted factor that regulates ZO-1 and claudin-14 expres- Apart from transcriptional regulation [8], TJ proteins are sion was TcpC, a protein that is not exported through under the control of posttranslational mechanisms that extracellular vesicles protein [31]. Proteinase K and heat modulate their subcellular location and association with treatments of COF-SN fractions confirmed the protein na- F-actin and other proteins in TJ structures. Reversible ture of the common active secreted factor (TcpC) and the phosphorylation of TJ proteins is a crucial mechanism in presence of other non-proteinic factors in ECOR63 super- the regulation of the intestinal epithelial barrier, and sev- natants. Concerning the active effectors secreted through eral kinases and phosphatases have been described to act OMVs, it should be considered that the ability of EcN and on TJ proteins [5, 6]. The ability of EcN and ECOR63 se- ECOR63 OMVs to reinforce the epithelial barrier is not a creted fractions to counteract delocalization of ZO-1 and common feature to all E. coli microbiota strains, as OMVs occludin in EPEC-infected cells was confirmed by im- from the commensal ECOR12 do not display such activity munofluorescence staining followed by confocal laser [31]. Bacterial vesicles enclose biological components that scanning microscopy. For both E.coli microbiota strains, exist in the producer bacteria, including a high number of OMVs and COF-SN fractions protected epithelial mono- common microbe-associated molecular patterns, which layers against EPEC-induced redistribution of ZO-1 and are recognized by immune receptors in epithelial and im- occludin, maintaining their association with TJ structures mune cells, and also specific cargo. Therefore, some of the at the cell boundaries. This suggests their potential to effects are strain-specific. In this context, proteomic stud- counteract EPEC-mediated changes in the phosphoryl- ies revealed that EcN OMVs enclose specific proteins that ation state of these TJ proteins. This effect could be medi- may contribute to gut colonization and modulation of ated, at least in part, by ERK activation, since occludin is a host responses [47]. Some of these vesicular proteins may target substrate of activated ERK [6]. In Caco-2 cells, mediate positive modulation on the epithelial barrier al- ERK-mediated phosphorylation of occludin has posi- though metabolites, small RNAs or lipids cannot be ruled tive barrier effects in preventing TJ disruption by out. Thus, omics ‘approaches may help to identify the bio- hydrogen peroxide [49]. The actin cytoskeleton is also active factors(s) secreted through OMVs. relevant in maintaining TJ structures and modulation To decipher the molecular mechanisms underlying the of paracellular solute transport in intestinal epithelial protective effects against EPEC infection we undertook cells. Immunofluorescence analysis of F-actin revealed expression and subcellular localization analysis of TJ pro- that EPEC-induced alterations in F-actin cytoskeleton teins. Quantitative RT-PCR confirmed downregulation of can be partially counteracted by EcN and ECOR63 se- ZO-2, ZO-1 and occludin in EPEC-infected intestinal epi- creted fractions, specially by OMVs. As EPEC disrupts thelial cells [16, 25], and in addition revealed downregula- the peri-junctional actin-myosin ring by activating the tion of claudin-14. However, we did not observe myosin light chain kinase (MLCK) [50], it is likely significant changes in claudin-1 expression. Interestingly, that protection exerted by EcN and ECOR63 OMVs the mRNA levels of the leaky protein claudin-2 were di- occurs in an MLCK-dependent fashion. minished upon EPEC infection. Thus, unlike other enteric Overall, this study reveals that the protection exerted pathogens such as Salmonella [48], the increased perme- by secreted fractions of the probiotic EcN and the com- ability state induced by EPEC does not correlate with high mensal ECOR63 against EPEC-mediated disruption of claudin-2 levels. EcN and ECOR63 secreted fractions pre- the epithelial barrier involves compensatory regulation vented EPEC-induced downregulation of occludin and mechanisms at various levels (Fig. 6), targeted to: (i) claudin-14. In contrast, they could not counteract the mRNA expression of the TJ proteins occludin and EPEC-mediated decrease in ZO-1 or ZO-2 expression. Re- claudin-14, (ii) redistribution of ZO-1 and occludin from sults from ZO-1 clearly differ from those reported previ- the cytoplasm to TJ structures at the intercellular con- ously in cellular models of intact epithelial barrier, which tacts and (iii) maintenance of the F-actin cytoskeleton showed the ability of OMVs and COF-SN to upregulate structure. Considering the in vivo conditions in the hu- ZO-1 expression [31]. This fact suggests that the EPEC- man gut, it should be highlighted that crosstalk between activated mechanisms responsible for ZO-1 downregula- epithelial and immune cells is crucial for maintaining in- tion could not be counteracted by these microbiota extra- testinal homeostasis. Therefore, other mechanisms acti- cellular fractions. Although live EcN cells induced ZO-2 vated by EcN and ECOR63 secreted factors at the level expression in T-84 cells [25], the secreted EcN fractions of intestinal mucosa could synergistically protect and were unable to exert this positive effect, even under condi- reinforce the epithelial barrier. In this context, the ability tions of intact epithelial barrier [31]. These findings sug- of EcN OMVs to upregulate IL-22 and TFF3 in colonic gest that EcN-mediated upregulation of ZO-2 depends on tissue [32, 33] may magnify the strengthening effects of bacteria-associated factors. the probiotic vesicles on the intestinal epithelial TJs. Alvarez et al. BMC Microbiology (2019) 19:166 Page 9 of 12 Fig. 6 Schematic representation of mechanisms used by EPEC to disrupt the intestinal epithelial barrier and their prevention by EcN and ECOR63 secreted fractions (OMVs and soluble free-released factors) Conclusion Isolation of OMVs and cell-free supernatant fractions Microbiota-derived vesicles and secreted factors are key OMVs and cell-free supernatants were obtained as previ- players in inter-species communication in the gut. Now- ously described [31]. Briefly, DMEM cultures of EcN adays, the role of gut microbiota extracellular vesicles in and ECOR63 were centrifuged (10,000×g, 4 °C, 20 min) health and disease is an emerging topic as they act as ve- to remove bacterial cells. The supernatants were filtered hicles for the distribution and delivery of many bacterial through a 0.22 μm pore size filter, concentrated with a effectors to distal tissues. Therefore, their effects are not Centricon® Plus-70 filter device and fractionated by restricted to local intestinal environment. This study ultracentifugation (150,000×g,1 h at 4 °C) into OMVs broadens our current understanding of how probiotic and cell-OMV-free supernatants (COF-SN). OMVs were and beneficial gut microbes exert their positive effects washed and suspended in phosphate buffered saline on the intestinal epithelial barrier. We show that OMVs (PBS). Samples were quantified by protein concentration and free-secreted factors by the probiotic EcN and the [52] and sterility was confirmed by platting one aliquot commensal ECOR63 can protect the barrier integrity on LB plates. Reproducibility of each batch of OMVs against EPEC infection by counteracting EPEC-altered was assessed by negative stain electron microscopy [47]. mRNA levels of occludin and claudin-14, maintaining When indicated COF-SN were subjected to protein- subcellular localization of ZO-1 and occludin associated eliminating treatments. For this, aliquots of COF-SN with TJ structures at the cell boundaries, and preserving were heated at 95 °C for 15 min or incubated with pro- F-actin at the inter-cellular junctions. teinase K (100 μg/ml) for 1 h at 37 °C followed by 30 min incubation with 0.5 mM phenylmethylsulfonyl fluoride. Methods Bacterial strains and growth conditions Cell culture and infection conditions EcN (serotype O6:K5:H1) is a probiotic E. coli strain, ob- T-84 (CCL-248) and Caco-2 (ATCC HTB37) human in- tained from Ardeypharm (GmbH, Herdecke, Germany). testinal epithelial cell lines were from the American Strain ECOR63 was isolated from faecal samples col- Type Culture Collection. Caco-2 cells were grown in lected from a healthy individual [51]. Wild-type EPEC DMEM High Glucose supplemented with 10% (v/v) strain E2348/69 (O127:H6) was kindly provided by B.B. foetal bovine serum (FBS) and T-84 cells in DMEM/F12 Finlay. Cultures were grown aerobically at 37 °C in Glutamax medium with 5% (v/v) FBS. In addition, cul- Luria–Bertani broth (LB) or in Dulbecco’s Modified ture media contained 25 mM HEPES, 1% non-essential Eagle Medium (DMEM) supplemented with 1% non- amino acids, penicillin (100 U/ml) and streptomycin essential amino acids and 25 mM HEPES. Optical dens- (100 μg/ml). For propagation, cells were split once a ity at 600 nm (OD600) was used to monitor growth and week. Then, 2 × 105 Caco-2 or T-84 cells were seeded re- calculate the total number of bacterial cells. spectively in 55 cm2 dishes or in 75 cm2 flasks. Alvarez et al. BMC Microbiology (2019) 19:166 Page 10 of 12 For infection, EPEC cells grown overnight in LB were isolated by the TRIzol method. Concentration and purity diluted 1:50 with fresh medium and cultured to an of RNA samples were evaluated by the ratio A260/A280 OD600 of 0.5–1.0. Bacterial cells were collected by cen- measured by UV spectrometry. RNA integrity was con- trifugation and suspended in serum and antibiotic-free firmed by electrophoresis on a denaturing 1% agarose/ DMEM plus 0.5% mannose. One hour before infection, formaldehyde gel. epithelial cells, grown in adequate supports depending For gene expression analysis, cDNA synthesis and RT- on the experiment, were washed twice in PBS and the qPCR were performed as described previously [31]. Spe- medium was changed to serum and antibiotic-free cific oligonucleotides for ZO-1, ZO-2, occludin, claudin- medium. EPEC was added to the apical side of epithelial 1 and claudin-2 are listed in Table 1. Oligonucleotides monolayers at a multiplicity of infection (MOI) of 100 for claudin-14 were from BioRaD (PrimePCR Assay and incubated for 3 h in the absence or presence of CLDN14, qHsaCED0023020). The β-actin gene was used OMVs (0.1 mg/ml) or COF-SN (0.5 mg/ml). When indi- as a reference for normalization. Appropriate control re- cated, T-84 cell monolayers were pre-incubated for 15 actions were performed in parallel and melting curve min with the ERK 1/2 inhibitor U0126 at 25 μM before analysis was applied after amplification to exclude un- the addition of EPEC, OMVs or COF-SNs. specific products. All samples were amplified in tripli- cates to establish the mean cycle threshold value (ct). Transepithelial resistance and paracellular marker FD-4 Relative changes in gene expression versus untreated flux measurements control cells were calculated applying the ΔΔCt formula. T-84 cells were seeded on the apical compartment of 12 mm polycarbonate Transwell cell culture inserts with a pore size of 0.4 μm at a cell density of 1 × 105 cells/cm2 Immunofluorescence labelling and confocal microscopy and cultured for 9–10 days with medium changes on alter- analysis nate days. Caco-2 cells were cultured for 18–20 days fol- Caco-2 cells were grown in an 8-well chamber slide (Ibidi) lowing the same protocol. Assessment of monolayer for 6 days. After EPEC infection in the absence or presence integrity was followed by transepithelial electrical resist- of COF-SN or OMVs, cells were washed, fixed and perme- ance (TER) and microscopic evaluation. TER was mea- abilized as described before [31]. Nuclei were labelled with sured as previously reported [31]. EPEC infections were DAPI (0.125 μg/ml), and occludin and ZO-1 were stained performed at initial TER values greater than 1000 Ω.cm2. respectively with anti-occludin mouse IgG antibody After 3-h infection, cell monolayers were washed with (0.5 μg/ml, Invitrogen) and anti-ZO-1 rabbit IgG antibody PBS and kept in serum-free medium containing gentami- (5 μg/ml, Invitrogen) [31]. F-actin was stained with phal- cin 100 μg/ml for 1 h before TER measurement. loidin–tetramethylrhodamine B isothiocyanate (−TRITC) Paracellular permeability was evaluated by the flux of (dilution 1:50) as described previously [56]. Immunofluor- fluorescein isothiocyanate-labelled dextran FD-4 (4 KDa; escence was analysed in a Leica TCS SP2 confocal micro- Sigma) through differentiated cell monolayers. Prior to the scope with a 63 × 1.32NA oil immersion objective and Ar, experiment, monolayer integrity was checked by measuring Ar-UV and HeNe lasers. Fluorescence was recorded at TER. Following EPEC infection in the absence or presence 405 nm (blue; DAPI) and at 488 nm (green, Alexa Fluor of OMVs or COF-SN, cells were washed twice with PBS 488). Images (12-bit) were obtained at a resolution of and fresh serum-free medium containing gentamicin 0.232 × 0.232 × 0.488 μm/voxel (x, y, z respectively) and 100 μg/ml was added to both Transwell compartments. analysed using Fiji processing package [57]. Then, FD-4 was added to the apical side at a final concen- tration of 1 mg/ml. Aliquots of 0.1 ml were taken from the Table 1 Primer sequences used for RT-qPCR basolateral chamber at 0 and 1 h incubation and the fluor- Genes Primer sequences References escence intensity was measured using excitation/emission ZO-1 Fw: 5′-CGGGACTGTTGGTATTGGCTAGA-3′ [53] wavelengths of 490/510–570 nm in a Modulus™ Microplate Rv: 5′-GGCCAGGGCCATAGTAAGTTG-3′ fluorescence reader (Turner BioSystems). For each sample, ZO-2 Fw: 5′-CTAGCAGCGATCAACTTAGGGACAA-3′ [53] FD-4 flux values were calculated by subtracting the inten- Rv: 5′-CCCAGGAGTTTCATTACCAGCAA-3′ sity fluorescence units measured at 0 h. Data were occludin Fw: 5′- TCCTATAAATCCACGCCGGTTC − 3′ [53] expressed as fold-change compared with non-treated con- Rv: 5′- CTAAAGTTACCACCGCTGCTG − 3′ trol cells, whose values were normalized as 1. Claudin-1 Fw: 5′-GCCCCAGTGGAGGATTTACT-3′ [54] Rv: 5′-GTTTTGGATAGGGCCTTGGT-3′ RNA isolation and quantitative reverse transcription PCR Claudin-2 Fw: 5′-ACCTGCTACCGCCACTCTGT-3′ [48] Rv: 5′- CTCCCTGGCCTGCATTATCTC-3′ (RT-qPCR) T-84 cells were grown in 12 well plates for 9 days. After β-actin Fw: 5′-GCTCTGGCTCCTAGCACCAT-3′ [55] Rv: 5′-GCCACCGATCCACACAGAGT-3′ 3 h infection with EPEC (MOI 100), total RNA was Alvarez et al. BMC Microbiology (2019) 19:166 Page 11 of 12 Statistical analysis Received: 23 May 2018 Accepted: 27 June 2019 The SPSS software package (version 20.0, Chicago, IL, USA) was used for statistical analysis. Data were col- lected from at least three independent biological ex- References 1. Turner JR. Intestinal mucosal barrier function in health and disease. Nat Rev periments and values are presented as the mean ± Immunol. 2009;9:799–809. standard error (SE). The one-way ANOVA followed 2. Peterson LW, Artis D. Intestinal epithelial cells: regulators of barrier function by Tukey’s post-test was used to assess whether dif- and immune homeostasis. Nat Rev Immunol. 2014;14:141–53. 3. Chiba H, Osanai M, Murata M, Kojima T, Sawada N. Transmembrane proteins ferences between more than two groups were statisti- of tight junctions. Biochim Biophys Acta. 2008;1778:588–600. cally significant (p-value ≤0.05). 4. Lee SH. Intestinal permeability regulation by tight junctions: implication on inflammatory bowel diseases. Intest Res. 2015;13:11–8. 5. González-Mariscal L, Tapia R, Chamorro D. Crosstalk of tight junction components Abbreviations with signaling pathways. Biochim Biophys Acta. 2008;1778:729–56. COF-SN: Cell-OMV-free supernatant; DMEM: Dulbecco’s modified Eagle 6. Cummins PM. Occludin: one protein, many forms. Mol Cell Biol. 2012; medium; EcN: Escherichia coli Nissle 1917; ERK: Extracellular signal-regulated 32:242–50. protein kinase; FBS: Fetal bovine serum; FD-4: Fluorescein isothiocyanate- 7. Shen L. Tight junctions on the move: molecular mechanisms for epithelial labelled 4 kDa dextran; LB: Luria-Bertani broth; MLCK: Myosin light chain barrier regulation. Ann N Y Acad Sci. 2012;1258:9–18. kinase; OMVs: Outer membrane vesicles; PBS: Phosphate buffer saline; 8. Lu Z, Ding L, Lu Q, Chen Y-H. Claudins in intestines: distribution and PCR: Polymerase chain reaction; RT-qPCR: Quantitative reverse transcription functional significance in health and diseases. Tissue Barriers. 2013;1:e24978. PCR; TER: Transepithelial resistance; TFF3: Trefoil factor-3; TJ: Tight junction; 9. Suzuki T. Regulation of intestinal epithelial permeability by tight junctions. TRITC: Tetramethylrhodamine B isothiocyanate; ZO: Zonula occludens Cell Mol Life Sci. 2013;70(4):631–59. 10. Donnenberg MS, Whittam TS. Pathogenesis and evolution of virulence in Acknowledgments enteropathogenic and enterohemorrhagic Escherichia coli. J Clin Invest. We acknowledge laboratories Ardeypharm for providing strain EcN and 2001;107:539–48. Lucille Banham for English manuscript proofreading. We thank Dr. Manel 11. Thanabalasuriar A, Koutsouris A, Weflen A, Mimee M, Hecht G, Gruenheid S. Bosch from CCiT-UB for technical assistance with confocal fluorescence Bacterial virulence factor NleA is required for the disruption of intestinal microscopy. tight junctions by Enteropathogenic E. coli. Cell Microbiol. 2010;12:31–41. 12. Ugalde-Silva P, Gonzalez-Lugo O, Navarro-Garcia F. Tight junction disruption induced by type 3 secretion system effectors injected by enteropathogenic Authors’ contributions and enterohemorrhagic Escherichia coli. Front Cell Infect Microbiol. 2016;6:87. LB and JB conceived of the study with the participation of RG, C-SA, and M- 13. Weflen A, Alto N, Hecht G. Tight junctions and enteropathogenic E. coli. AC in experimental design. LB and JB wrote the manuscript and supervised Ann N Y Acad Sci. 2009;1165:169–74. the work. RG, M-AC and C-SA carried out data interpretation and statistical 14. Jonhnson-Henry KC, Donato KA, Shen-Tu G, Gordanpour M, Sherman PM. analysis. C-SA, RV and ND-G performed the experimental work. All authors re- Lactobacillus rhamnosus strain GC prevents Enterohemorrhagic Escherichia vised, read and approved the final manuscript. coli O157:H7-induced changes in epithelial barrier function. Infect Immun. 2008;66:1340–8. Funding 15. Qin H, Zhang Z, Hang X, Jiang Y. L. plantarum prevents enteroinvasive This study was supported by Grant AGL2016-79113-R, co-financed with Euro- Escherichia coli-induced tight junction proteins changes in intestinal pean Commission ERDF funds, from the Ministerio de Economía y Competiti- epithelial cells. BMC Microbiol. 2009;9:63. vidad, Spain (AEI/FEDER, UE) to LB. The funding agency had no role in the 16. Liu Z-H, Shen T-Y, Zhang P, Ma Y-L, Pat Moyer M, Qin H-L. Protective effects design of the study, data collection, analysis and interpretation, and writing of Lactobacillus plantarum against epithelial barrier dysfunction of human the manuscript. We acknowledge financial support from Universitat de Barce- colon cell line NCM460. World J Gastroenterol. 2010;16(45):5759–65. lona for publication in open access. C-SA acknowledges the fellowships 17. Yu Q, Yuan L, Deng J, Yang Q. Lactobacillus protects the integrity of granted by PRODEP and by Universidad Juárez Autónoma de Tabasco, intestinal epithelial barrier damaged by pathogenic bacteria. Front Cell Mexico. RV and ND-G acknowledge the post-doctoral and doctoral fellow- Infect Microbiol. 2015;5:26. ships, respectively, granted by CONICYT. 18. Kruis W, Frič P, Pokrotnieks J, Lukáš M, Fixa B, Kaščák M, Kamm MA, Weismueller J, Beglinger C, Stolte M, et al. Maintaining remission of ulcerative colitis with the probiotic Escherichia coli Nissle 1917 is as effective Availability of data and materials as with standard mesalazine. Gut. 2004;53:1617–23. All data generated or analysed during the study are included in the article. 19. Chibbar R, Dieleman IA. Probiotics in the management of ulcerative colitis. J Interested readers can contact the corresponding author for needed Clin Gastroeneterol. 2015;49:S50–5. information. 20. Kruis W, Chrubasik S, Boehm S, Stange C, Schulze J. A double-blind placebo- controlled trial to study therapeutic effects of probiotic Escherichia coli Ethics approval and consent to participate Nissle 1917 in subgroups of patients with irritable bowel syndrome. Int J Not applicable Color Dis. 2012;27(4):467–74. 21. Henker J, Laass M, Blokhin BM, Bolbot YK, Maydannik VG, Elze M, Wolff C, Schulze J. The probiotic Escherichia coli strain Nissle 1917 (EcN) stops acute Consent for publication diarrhoea in infants and toddlers. Eur J Pediatr. 2007;166:311–8. Not applicable 22. Behnsen J, Deriu E, Sassone-Corsi M, Raffatellu M. Probiotics: properties, examples, and specific applications. Cold Spring Harb Perspect Med Mar. Competing interests 2013;3(3):a010074. The authors declare that they have no competing interests. 23. Trebichavsky I, Splichal I, Rada V, Splichalova A. Modulation of natural immunity in the gut by Escherichia coli Nissle 1917. Nutr Rev. 2010;68:459–64. Author details 24. Ukena SN, Singh A, Dringenberg U, Engelhardt R, Seidler U, Hansen W. 1 Secció de Bioquímica i Biología Molecular, Departament de Bioquímica i Probiotic Escherichia coli Nissle 1917 inhibits leaky gut by enhancing Fisiologia, Facultat de Farmàcia i Ciències de l’Alimentació, Universitat de mucosal integrity. PLoS One. 2007;2:e1308. Barcelona, Barcelona, Spain. 2Institut de Biomedicina de la Universitat de 25. Zyrek AA, Cichon C, Helms S, Enders C, Sonnenborn U, Schmidt MA. Barcelona (IBUB), Institut de Recerca Sant Joan de Déu (IRSJD), Barcelona, Molecular mechanisms underlying the probiotic effects of Escherichia coli Spain. 3Present Address: División Académica de Ciencias Biológicas, Nissle 1917 involve ZO-2 and PKCZeta redistribution resulting in tight Universidad Juárez Autónoma de Tabasco, Tabasco, Mexico. junction and epithelial barrier repair. Cell Microbiol. 2007;9:804–16. Alvarez et al. BMC Microbiology (2019) 19:166 Page 12 of 12 26. Hering NA, Richter JF, Fromm A, Wieser A, Hartmann S, Günzel D, Bücker R, 50. Yhan R, Koutsouris A, Savkovic SD, Hecht G. Enteropathogenic Escherichia Fromm M, Schulzke JD, Troeger H. TcpC protein from E. coli Nissle improves coli-induced myosin light chain phosphorylation alters intestinal epithelial epithelial barrier function involving PKCζ and ERK1/2 signaling in HT-29/B6 permeability. Gastroenterology. 1997;113(6):1873–82. cells. Mucosal Immunol. 2014;7:369–78. 51. Ochman H, Selander RK. Standard reference strains of Escherichia coli from 27. Sánchez B, Urdaci MC, Margolles A. Extracellular proteins secreted by natural populations. J Bacteriol. 1984;157:690–3. probiotic bacteria as mediators of effects that promote mucosa-bacteria 52. Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurement with interactions. Microbiology. 2010;156:3232–42. the Folin phenol reagent. J Biol Chem. 1951;193(1):265–75. 28. Hevia A, Delgado S, Sánchez B, Margolles A. Molecular players involved in 53. Park H-Y, Kunitake Y, Hirasaki N, Tanaka M, Matsui T. Theaflavins enhance the interaction between beneficial bacteria and the immune system. Front intestinal epithelial barrier of Caco-2 monolayers through the expression of Microbiol. 2015;6:1285. AMP-activated protein kinase-mediated occludin, claudin-1, and ZO-1. Biosci 29. Kaparakis-Liaskos M, Ferrero RL. Immune modulation by bacterial outer Biotechnol Biochem. 2015;79:130–7. membrane vesicles. Nat Rev Immunol. 2015;15:375–87. 54. Satake S, Semba S, Matsuda Y, Usami Y, Chiba H, Sawada N, Kasuga M, 30. Cañas MA, Giménez R, Fábrega MJ, Toloza L, Baldomà L, Badia J. Outer Yokozaki H. Cdx2 transcription factor regulates claudin-3 and claudin-4 membrane vesicles from the probiotic Escherichia coli Nissle 1917 and expression during intestinal differentiation of gastric carcinoma. Pathol the commensal ECOR12 enter intestinal epithelial cells via clathrin- Int. 2008;58:156–63. dependent endocytosis and elicit differential effects on DNA damage. 55. Takahashi K, Hirose K, Saki K, Niwa Y, Wakashin H, Iwata A, Tokoyoda K, PLoS One. 2016;11:e0160374. Renauld JC, Iwamoto I, Nakayama T, et al. IL-22 attenuates IL-25 production 31. Alvarez CS, Badia J, Bosch M, Giménez R, Baldomà L. Outer membrane vesicles by lung epithelial cells and inhibits antigen-induced eosinophilic airway and soluble factors released by probiotic Escherichia coli Nissle 1917 and inflammation. J Allergy Clin Immunol. 2011;128:1067–76. commensal ECOR63 enhance barrier function by regulating expression of tight 56. Egea L, Giménez R, Lúcia D, Modolell I, Badía J, Baldomà L, Aguilar J. junction proteins in intestinal epithelial cells. Front Microbiol. 2016;7:1981. Increased production of the ether-lipid platelet-activating factor in 32. Fábrega MJ, Aguilera L, Giménez R, Varela E, Cañas MA, Antolín M, Badia J, intestinal epithelial cells infected by Salmonella enteritidis. Biochim Baldomà L. Activation of immune and defense responses in the intestinal Biophys Acta. 2008;1781:270–6. mucosa by outer membrane vesicles of commensal and probiotic 57. Schindelin J, Arganda-Carreras I, Frise E, Kaynig V, Longair M, Pietzsch T, Escherichia coli strains. Front Microbiol. 2016;7:705. Preibisch S, Rueden C, Saalfeld S, Schmid B, et al. Fiji: an open-source 33. Fábrega MJ, Rodriguez-Nogales A, Garrido-Mesa J, Algieri F, Badía J, platform for biological-image analysis. Nat Methods. 2012;9(7):676–82. Giménez R, Gálvez J, Baldomà L. Intestinal anti-inflammatory effects of outer membrane vesicles from Escherichia coli Nissle 1917 in DSS-experimental colitis in mice. Front Microbiol. 2017;8:1274. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in 34. Podolsky DK, Gerken G, Eyking A, Cario E. Colitis-associated variant of TLR2 published maps and institutional affiliations. causes impaired mucosal repair because of TFF3 deficiency. Gastroenterology. 2009;137:209–20. 35. Buda A, Jepson MA, Pignatelli M. Regulatory function of trefoil peptides (TFF) on intestinal cell junctional complexes. Cell Adhesion Comm. 2012;19:63–8. 36. Toloza L, Giménez R, Fábrega MJ, Alvarez CS, Aguilera L, Cañas MA, Martín- Venegas R, Badia J, Baldomà L. The secreted autotransporter toxin (Sat) does not act as a virulence factor in the probiotic Escherichia coli strain Nissle 1917. BMC Microbiol. 2015;15:250. 37. Hering NA, Fromm M, Schulzke JD. Determinants of colonic barrier function in inflammatory bowel disease and potential therapeutics. J Physiol. 2012;590:1035–44. 38. Oshima T, Miwa H. Gastrointestinal mucosal barrier function and diseases. J Gastroenterol. 2016;51:768–78. 39. Isolauri E, Kirjavainen PV, Salminen S. Probiotics: a role in the treatment of intestinal infection and inflammation? Gut. 2002;50(Suppl III):iii54–9. 40. Novak J, Katz JA. Probiotics and prebiotics for gastrointestinal infections. Curr Infect Dis Rep. 2006;8:103–9. 41. Britton RA, Versalovic J. Probiotics and gastrointestinal infections. Interdiscip Perspect Infect Dis. 2008;2008:290769. 42. Canil C, Rosenshine I, Ruschkowski S, Donnenberg MS, Kaper JB, Finlay BB. Enteropathogenic Escherichia coli decreases the transepithelial electrical resistance of polarized epithelial monolayers. Infect Immun. 1993;61(7):2755–62. 43. Spitz J, Yuhan R, Koutsouris A, Blatt C, Alverdy J, Hecht G. Enteropathogenic Escherichia coli adherence to intestinal epithelial monolayers diminishes barrier function. Am J Phys. 1995;268(2 Pt 1):G374–9. 44. Simonovic I, Rosenberg J, Koutsouris A, Hecht G. Enteropathogenic Escherichia coli dephosphorylates and dissociates occludin from intestinal epithelial tight junctions. Cell Microbiol. 2000;2:305–15. 45. Muza-Moons MM, Schneeberger EE, Hecht GA. Enteropathogenic Escherichia coli leads to appearance of aberrant tight junctions strands in the lateral membrane of intestinal epithelial cells. Cell Microbiol. 2004;6(8):783–93. 46. Zhang Q, Li Q, Wang C, Li N, Li J. Redistribution of tight junction proteins during EPEC infection in vivo. Inflammation. 2012;35:23–32. 47. Aguilera L, Toloza L, Giménez R, Odena A, Oliveira E, Aguilar J, Badia J, Baldomà L. Proteomic analysis of outer membrane vesicles from the probiotic strain Escherichia coli Nissle 1917. Proteomics. 2014;14(2–3):222–9. 48. Zhang YG, Wu S, Xia Y, Sun J. Salmonella infection upregulates the leaky protein claudin-2 in intestinal epithelial cells. PLoS One. 2013;8:e58606. 49. Basuroy S, Sheth P, Elias B, Naren AP, Rao R. MAPK interacts with occludin and mediates EGF-induced prevention of tight junction disruption by hydrogen peroxide. Biochem J. 2006;393:69–77.