Advertisement

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Murinization of Internalin Extends Its Receptor Repertoire, Altering Listeria monocytogenes Cell Tropism and Host Responses

  • Yu-Huan Tsai,

    Affiliations Institut Pasteur, Biology of Infection Unit, Paris, France, Inserm U1117, Paris, France, Université Paris Diderot, Sorbonne Paris Cité, Paris, France

    ⨯
  • Olivier Disson,

    Affiliations Institut Pasteur, Biology of Infection Unit, Paris, France, Inserm U1117, Paris, France

    ⨯
  • Hélène Bierne,

    Affiliations Institut Pasteur, Unité des Interactions Bactéries Cellules, Paris, France, Inserm, U604, Paris, France, INRA, USC2020, Paris, France

    ⨯
  • Marc Lecuit

    marc.lecuit@pasteur.fr

    Affiliations Institut Pasteur, Biology of Infection Unit, Paris, France, Inserm U1117, Paris, France, Institut Pasteur, French National Reference Center and World Health Organization Collaborating Center on Listeria, Paris, France, Université Paris Descartes, Sorbonne Paris Cité, Institut Imagine, Paris, France, Necker-Enfants Malades University Hospital, APHP, Division of Infectious Diseases and Tropical Medicine, Paris, France

    ⨯

Abstract

Listeria monocytogenes (Lm) is an invasive foodborne pathogen that leads to severe central nervous system and maternal-fetal infections. Lm ability to actively cross the intestinal barrier is one of its key pathogenic properties. Lm crosses the intestinal epithelium upon the interaction of its surface protein internalin (InlA) with its host receptor E-cadherin (Ecad). InlA-Ecad interaction is species-specific, does not occur in wild-type mice, but does in transgenic mice expressing human Ecad and knock-in mice expressing humanized mouse Ecad. To study listeriosis in wild-type mice, InlA has been “murinized” to interact with mouse Ecad. Here, we demonstrate that, unexpectedly, murinized InlA (InlAm) mediates not only Ecad-dependent internalization, but also N-cadherin-dependent internalization. Consequently, InlAm-expressing Lm targets not only goblet cells expressing luminally-accessible Ecad, as does Lm in humanized mice, but also targets villous M cells, which express luminally-accessible N-cadherin. This aberrant Lm portal of entry results in enhanced innate immune responses and intestinal barrier damage, both of which are not observed in wild-type Lm-infected humanized mice. Murinization of InlA therefore not only extends the host range of Lm, but also broadens its receptor repertoire, providing Lm with artifactual pathogenic properties. These results challenge the relevance of using InlAm-expressing Lm to study human listeriosis and in vivo host responses to this human pathogen.

Author Summary

Co-evolution of microbes with their hosts can select stringently specific host-microbe interactions at the cell, tissue and species levels. Listeria monocytogenes (Lm) is a foodborne pathogen that causes a deadly systemic infection in humans. Lm crosses the intestinal epithelium upon the interaction of its surface protein InlA with E-cadherin (Ecad). InlA-Ecad interaction is species-specific, does not occur in wild-type mice, but does in transgenic mice expressing human Ecad and knock-in mice expressing humanized mouse Ecad. To study listeriosis in wild-type mice, InlA has been “murinized” to interact with mouse Ecad. Here, we demonstrate that in addition to interacting with mouse Ecad, InlAm also uses N-cadherin as a receptor, whereas InlA does not. This artifactual InlAm-N-cadherin interaction promotes bacterial translocation across villous M cells, a cell type which is not targeted by InlA-expressing bacteria. This leads to intestinal inflammation and intestinal barrier damage, both of which are not seen in humans and humanized mouse models permissive to InlA-Ecad interaction. These results challenge the relevance of using InlAm-expressing Lm as a model to study human listeriosis and host responses to this pathogen. They also illustrate that caution must be exercised before using “murinized” pathogens to study human infectious diseases.

Citation: Tsai Y-H, Disson O, Bierne H, Lecuit M (2013) Murinization of Internalin Extends Its Receptor Repertoire, Altering Listeria monocytogenes Cell Tropism and Host Responses. PLoS Pathog 9(5): e1003381. https://doi.org/10.1371/journal.ppat.1003381

Editor: Dana J. Philpott, University of Toronto, Canada

Received: November 28, 2012; Accepted: April 9, 2013; Published: May 30, 2013

Copyright: © 2013 Tsai et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Funding: The Biology of Infection Unit is supported by Institut Pasteur, Institut National de la Santé et de la Recherche Médicale, Fondation pour la Recherche Médicale, Ville de Paris, Fondation BNP Paribas, the Listress EU program, LabEx IBEID and the European Research Council. YHT is the recipient of a fellowship from Institut Pasteur (Pasteur-Paris University International PhD Program) funded by Institut Carnot. HB is supported by Institut Pasteur, INRA, the French National Research Agency (ANR blanc) and Ligue Contre le Cancer. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Co-evolution of microbes with their hosts can select stringently specific host-microbe interactions at the cell, tissue and species levels [1]. Species-specific host-microbe interactions, which are the rule rather than the exception, pose a challenge for the use of laboratory animal models to study human pathogens, including Listeria monocytogenes (Lm), the etiological agent of listeriosis, a deadly foodborne infection. Lm is able to actively cross the intestinal barrier, reach the systemic circulation and cross the blood-brain and placental barriers, leading to its dissemination to the central nervous system and the fetus [2].

The mouse is a genetically amenable model that is widely used to investigate human diseases [3], [4]. To obtain a mouse model in which the pathogenic properties of a given pathogen are similar to what is observed in human, species specificity can be circumvented by humanizing the mouse by transgenesis [5], [6], [7], [8], knock-in [9], knock-out [10] or xenograft techniques [11]. One can also adapt the pathogen to the mouse by multiple passages on cell lines [12], [13] or in vivo [14], or specifically “murinize” a pathogen ligand so that it interacts with the mouse ortholog of a species-specific human receptor [15], [16].

The Lm surface protein InlA interacts with E-cadherin (Ecad) and mediates Lm entry into epithelial cells, which express this adherens junction protein [17], [18]. Cadherins constitute a family of calcium-dependent cell adhesion receptors. Ecad is expressed mainly in epithelia, whereas N-cadherin (Ncad) is found primarily in neuronal cells and endothelial cells together with VE-cadherin [19], [20]. Ncad can also be coexpressed with Ecad in epithelial cells [21]. Importantly, Ncad has been reported to not act as a receptor for InlA, and so far Ecad is the only known classical cadherin acting as a receptor for InlA [18]. In contrast to Ecad from human, guinea pig, rabbit and gerbil, mouse Ecad (mEcad) and rat Ecad are not recognized by InlA and do not promote bacterial entry [9], [22]. The interaction of InlB, another Lm invasion protein, with its host receptor is also species-specific [23]. InlB recognizes the hepatocyte growth factor receptor Met of human, mouse, rat and gerbil but not that of guinea pig and rabbit [9], [23], [24].

Two mouse lines have been established to study InlA-Ecad interaction in vivo: a transgenic mouse line expressing human Ecad (hEcad) in enterocytes (hEcad Tg) [6], and a humanized mEcad knock-in mouse line (E16P KI) with an E16P amino acid substitution which enables mEcad to interact with InlA without affecting Ecad homophilic interactions and allows Lm internalization [9], [22]. Using these two humanized mouse models, we have demonstrated that InlA mediates Lm crossing of the intestinal epithelium upon targeting of luminally-accessible Ecad around goblet cells [6], [9], [25], and that InlA and InlB act interdependently to mediate the crossing of the placental barrier [9]. Epidemiological investigations have confirmed the relevance of these experimental findings, and shown that InlA is implicated in Lm crossing of human intestinal and placental barriers [9], [26].

In 2007, Wollert et al. engineered a genetically modified InlA with the purpose of increasing its binding affinity to hEcad [16]. Two amino acid substitutions in InlA, S192N and Y369S, were shown to enhance InlA binding affinity to hEcad [16]. Neither S192N nor Y369S substitution has been observed in the more than 500 Lm isolates InlA sequences we have checked (our unpublished results). Wollert et al. published that this increased affinity for hEcad translates into an increased bacterial entry into human epithelial cells (Caco-2) [16]. Importantly, Wollert et al. also showed that this modified InlA binds the extracellular cadherin domain 1 (EC1) of mEcad in solution with a comparable affinity to that of the wild-type (wt) InlA for hEcad EC1 [16]. They hypothesized that this interaction would allow Lm expressing this “murinized” InlA (InlAm) to cross intestinal barrier and would render wt mice orally permissive to Lm infection, a phenotype which is mediated by InlA in permissive models [6]. In support of this hypothesis, Wollert et al. found an increased intestinal, spleen and liver bacterial loads of wt mice orally inoculated with Lm expressing InlAm, yet only after 3 to 4 days post infection, which is later than in models permissive to InlA-Ecad interaction [6], [9], [16]. Moreover, the ability of InlAm to mediate mEcad-dependent Lm internalization into host cells has never been tested. In addition, InlAm unexpectedly promoted pronounced inflammation and intestinal epithelial cell damages in wt mice [16], whereas wt InlA mediates the crossing of the intestinal barrier without inducing significant intestinal response and tissue damage in hEcad transgenic mice [6], [27].

This prompted us to investigate the detailed properties of InlAm in cultured cells, as well as the in vivo cell and tissue tropisms of bacteria expressing InlAm, as compared to that of its isogenic parental Lm strain that expresses wt InlA. Here, we demonstrate that InlAm promotes bacterial entry not only into mEcad-positive but also into mEcad-negative mouse cells. We show that InlAm-mediated entry into mEcad-negative cells is mouse Ncad (mNcad)-dependent. Importantly, InlAm-mNcad interaction allows bacteria to specifically target Ncad-positive villous M cells in vivo, a cell type which is not targeted by Lm in humanized mouse models permissive to InlA-Ecad interaction. This leads to enhanced intestinal inflammatory responses and disruption of the intestinal barrier integrity, both of which are not observed in Lm-infected humanized mice and human listeriosis. Together, these results demonstrate that the murinization of InlA not only extends Lm host range, but also broadens its receptor repertoire, consequently changing Lm cell tropism and enhancing host immune responses to Lm. These results challenge the relevance of using InlAm-expressing Lm to study human listeriosis and in vivo host responses to this human pathogen.

Results

Murinization of InlA promotes bacterial entry into mEcad-expressing cells but has no impact on bacterial entry into hEcad-expressing cells

We first investigated whether the increased affinity of InlAm to hEcad translates into an enhanced invasion of hEcad-expressing cells, as proposed by Wollert et al. [16]. To this end, we assessed InlAm-dependent entry into LoVo cell, a human epithelial cell line expressing hEcad [22]. Lm wt strain and Lm expressing InlAm (Lm-inlAm) invaded LoVo cells at similar levels (Figure 1A). Because Lm can be internalized by InlA-independent pathways such as InlB-Met, we transferred either inlA or inlAm onto the chromosome of Listeria innocua (Li), a naturally non-invasive and non-pathogenic Listeria species, in which heterologous expression of inlA has been shown to confer invasiveness [17], [18], [28]. Li expressing either InlA (Li-inlA) or InlAm (Li-inlAm) were equally invasive in LoVo cells (Figure 1B). These results indicate that contrary to what is reported by Wollert et al. [16], the increased affinity of InlAm to hEcad does not translate into an increased level of bacterial entry. Both Li-inlA and Li-inlAm recruited hEcad when incubated with LoVo cells, suggesting that hEcad is involved in both InlA- and InlAm-mediated entries (Figure 1E, upper panel). Because purifed InlAm interacts with the purified EC1 domain of mEcad, Wollert et al. have proposed, although not tested, that InlAm would mediate bacterial entry into mEcad-expressing cells [16]. We therefore tested the ability of InlAm to promote bacterial entry into the mouse epithelial cell line Nme, which expresses mEcad [29]. InlAm promoted bacterial entry into mEcad-expressing Nme cells, although to a lower level than InlA in hEcad-expressing LoVo cells (Figure 1C and D). Li-inlAm also recruited mEcad during cell invasion, whereas as expected, Li-inlA does not (Figure 1E, lower panel). Together, these results show that (i) the increased affinity of InlAm to hEcad does not enhance bacterial entry into hEcad-expressing cells, and (ii) the murinization of InlA confers to Lm an enhanced ability to be internalized into mEcad-expressing cells [16].

thumbnail
Download:
Figure 1. Respective ability of InlA and InlAm to promote bacterial entry into hEcad- and mEcad-expressing cells.

Bacterial entry into hEcad-expressing human epithelial cells (LoVo) (A for Lm and B for Li) and mEcad-expressing mouse epithelial cells (Nme) (C for Lm and D for Li) was performed by counting intracellular gentamicin resistant bacteria following 1 hr of infection and 1 hr of gentamicin incubation. No significant difference was observed between InlA- and InlAm-expressing bacteria when invading LoVo cells, whereas InlAm expression promoted bacterial entry into Nme cells. Values are expressed as a mean + SD (n = 3). Statistical analysis was performed with the unpaired Student's t test. (E) Recruitment of Ecad was performed by incubating Li-inlA or Li-inlAm with the cells cultured on coverslips for 1 hr, followed by PBS wash and fixation before staining. The cells were stained with the anti-Li antibody and anti-hEcad or anti-mEcad antibody. Right panels show separated channels and merge of boxed regions, showing the recruitment of cadherin by bacteria. Both Li-inlA and Li-inlAm recruit hEcad when incubated with LoVo cells. Li-inlAm but not Li-inlA is able to recruit endogenous mEcad in Nme cells. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.g001

InlAm promotes mEcad-independent entry into mouse cells

Monk et al. have reported that Lm-inlAm invades mouse CT26 cells more efficiently than Lm [13]. Strikingly, CT26 cells do not express mEcad (Figure 2A) [30], yet we confirmed that InlAm mediates bacterial entry into these cells (Figure 2B). Because classical cadherins exhibit a high level of conservation in their EC1 domains (Figure S1A), we tested whether Li-inlAm would recruit another classical cadherin than mEcad in CT26 cells. We labeled CT26 cells with a pan-cadherin antibody, which recognizes the cytoplasmic domain of classical cadherins [31]. CT26 cells were strongly stained with the pan-cadherin antibody (Figure S1B), indicating that they likely express classical cadherin proteins. Furthermore, this pan-cadherin-immunoreactive protein was recruited in CT26 cells by Li-inlAm but not Li-inlA (Figure S1B). Immunoblotting and immunostaining revealed that CT26 cells express Ncad (Figures 2C and D), a classical cadherin known to be expressed in endothelial cells, neurons and some transformed epithelial cells [20]. Importantly, Li-inlAm, but not Li-inlA, recruited Ncad in CT26 cells (Figure 2D). We next tested other cell lines for Ncad expression. We found that Nme cells (which also express mEcad and are permissive to InlAm-mediated entry), human HeLa cells, and guinea pig 104C1 cells all express Ncad (Figure 2C). As in CT26 cells, InlAm promoted bacterial entry into HeLa and 104C1 cells, although these two cell lines do not express Ecad and are therefore not permissive to InlA-dependent entry (Figure S2) [23]. These results suggest that the murinization of InlA confers to this protein the ability to interact with Ncad from different species, and to enter into host cells expressing Ncad.

thumbnail
Download:
Figure 2. InlAm promotes bacterial entry into mEcad-negative mouse cells and recruits endogenous Ncad.

(A) Nme cells express mEcad, whereas no mEcad expression can be detected in CT26 cells. Actin expression was used as a loading control. (B) Bacterial entry was performed as described in Figure 1. InlAm- but not InlA-expression promotes bacterial entry into mouse CT26 cells. Values are expressed as a mean + SD (n = 3). Statistical analysis was performed with the unpaired Student's t test. (C) Nme cells, CT26 cells, HeLa cells and 104C1 cells express Ncad. The lysate of BHK21 cells which do not express any detectable classical cadherins was used as a negative control. (D) Recruitment of Ncad was performed as described in Figure 1. The coverslips were stained with the anti-Li antibody and anti-Ncad antibody. Right panels show separated channels and merge of boxed regions, showing the recruitment of Ncad by Li-inlAm. Scale bar, 20 µm. See also Figures S1 and S2.

https://doi.org/10.1371/journal.ppat.1003381.g002

mNcad is a receptor for InlAm but not InlA

To investigate if mNcad serves as a receptor for InlAm-mediated entry into CT26 cells, CT26 cells were treated with mNcad-specific siRNAs or scrambled control siRNAs. Treatment of CT26 cells with mNcad siRNAs led to a reduced expression of mNcad which correlated with a significantly decreased InlAm-dependent entry (Figures 3A and B). To directly assess the ability of mEcad and mNcad to act as receptors for InlAm, we used the BHK21 cell line, which is of hamster origin and does not express any known classical cadherin [32], and transfected this cell line with plasmids encoding either hEcad, mEcad or mNcad. As expected, both InlA and InlAm mediated bacterial entry into hEcad-expressing cells (Figure 3C). Moreover, InlAm mediated entry into mEcad-expressing cells, whereas as previously shown, InlA did not (Figure 3C) [22]. Most importantly, we also demonstrated that InlAm mediated bacterial entry into Ncad-expressing cells, whereas, as previously shown, InlA did not (Figure 3C) [18].

thumbnail
Download:
Figure 3. InlAm mediates mouse Ncad-dependent internalization.

(A) Mouse CT26 cells were transfected with scrambled siRNAs or mNcad-specific siRNAs. Bacteria internalization was evaluated by counting intracellular gentamicin resistant bacteria. Values are expressed as a mean + SD (n = 3). Statistical analysis was performed with the unpaired Student's t test. (B) The expression of mNcad was decreased in the cells transfected with mNcad–specific siRNAs compared to that transfected with scrambled siRNAs. Decrease of mNcad expression in CT26 cells reduced the entry of InlAm-expressing bacteria but not that of InlA-expressing bacteria into CT26 cells. (C) To evaluate the function of different cadherins as receptors, BHK21 cells were transiently transfected with pcDNA3 expression vector harboring the cDNAs of each cadherin. Bacterial invasion were evaluated by counting intracellular gentamicin resistant bacteria. Expression of hEcad provides gain-of-function for both InlA- and InlAm-expressing Li to invade. InlAm but not InlA expression promotes bacterial entry to mEcad- and mNcad-expressing cells. Neither mouse P-cadherin (mPcad) nor mouse VE-cadherin (mVEcad) expression promotes InlA- and InlAm-dependent entry. Values are expressed as a mean + SD (n = 3). Statistical analysis was performed with the unpaired Student's t test. See also Figure S1.

https://doi.org/10.1371/journal.ppat.1003381.g003

To investigate whether the InlAm receptor repertoire extends to other members of classical cadherins, we tested the ability of mouse P-cadherin (mPcad) and VE-cadherin (mVEcad) to serve as receptors for InlAm (Figure S1A). Neither mPcad nor mVEcad acted as a receptor for InlAm or InlA (Figure 3C). Taken together, these data confirm that InlA exhibits a species-specific and narrow repertoire for Ecad and mediates entry into hEcad- but not mEcad-expressing cells, and demonstrate that by widening InlA species spectrum from human to mouse Ecad, murinization of InlA extends its receptor repertoire to Ncad.

Murinization of InlA extends the cell tropism of Lm at the intestinal level

In order to investigate if these in vitro results translate into an in vivo phenotype, and study in particular the cell tropism of InlAm-expressing bacteria, we investigated Ncad luminal accessibility at the intestinal epithelium level, which is the portal of InlA-mediated entry of Lm. In contrast to luminally-accessible Ecad which is mostly observed as rings surrounding goblet cells [25], mNcad was accessible on the apical pole of villous M cells (Figure 4, Movie S1), but not M cells of Peyer's patches (Movie S2) in wt mice. The expression of luminally-accessible Ncad was also detected on the apical pole of villous M cells in E16P KI mice (Figure S3, Movie S3). These results suggest that InlAm may allow bacteria to target villous M cells upon mouse oral inoculation.

thumbnail
Download:
Figure 4. Ncad expression is detected on the apical pole of villous M cells.

Immunolabeling of the small intestine tissue section (A) and whole mount tissue of a part of ileum (B and C) of wt mice was performed. (A) The small intestine tissue section was stained by anti-mouse IgG followed by anti-Ncad mouse IgG and anti-mouse IgG conjugated with another fluorophore to distinguish mouse IgG-positive cells and Ncad-positive cells. Scale bar, 50 µm. (B and C) Intestinal tissue of wt mice was fixed and stained for luminally accessible (acc) Ncad with antibody against extracellular domain of Ncad (clone GC4) before tissue permeabilization, M cells with NKM 16-2-4 antibody, wheat germ agglutinin (WGA) and nuclei after tissue permeabilization. Projection of a 25 µm thick reconstructed intestinal villus (B) and one of the xy plane (C) are shown. Right panels show separated channels and merge of boxed regions in (C), showing Ncad on the apical side of NKM 16-2-4-positive cells. NKM 16-2-4 antibody is a monoclonal antibody raised against α(1,2) fucose moiety in absence of neighboring sialic acids, a specific marker on M cells surface. WGA was used to stain the mucus of goblet cells and cell membrane. Scale bar, 20 µm. See also Figure S3, Movies S1, S2 and S3.

https://doi.org/10.1371/journal.ppat.1003381.g004

To specifically investigate whether InlAm-expressing bacteria target cells that express luminally-accessible Ncad, we inoculated orally wt mice with Li-inlA or Li-inlAm, and for comparison we inoculated humanized E16P KI mice orally with Li-inlA. As expected from our recent results [25], Li-inlA were found in goblet cells 5 hrs post oral inoculation of E16P KI mice (Figures 5C and D). In contrast, Li-inlAm targeted both goblet cells (Figures 5A and D) and villous M cells (Figures 5B and D, Movie S4) in wt mice. We next performed a detailed quantification of the location of bacteria in the intestinal epithelium (i.e. goblet cells, villous M cells, other epithelial cells). This demonstrated that, contrary to InlA, which targets almost exclusively goblet cells in E16P KI mice (82%), InlAm preferentially targets villous M cells (56%) in wt mice, and to a lower degree goblet cells (34%) (p<0.001, χ2 test analysis) (Figure 5D). In agreement with these results obtained with Li-inlAm, Lm-inlAm also targeted both goblet cells (Figures S4A and D, S5A, Movie S5) and villous M cells (Figures S4B and D, S5B, Movie S6) in both wt and E16P KI mice, in contrast to Lm which exclusively targeted goblet cells, only in E16P KI mice (Figures S4C and D, S5C, Movie S7). Together, these results demonstrate that while InlA- and InlAm-Ecad interactions both contribute to the targeting of goblet cells, InlAm-mNcad interaction allows bacteria to target villous M cells, a cell type which is not targeted when InlA interacts only with its native receptor Ecad.

thumbnail
Download:
Figure 5. InlAm mediates targeting of villous M cells.

The intestinal ileum was taken from E16P KI mice and wt mice orally inoculated by 1010 Li-inlA and Li-inlAm at 5 hr post infection, respectively. The intestinal tissues were fixed and stained with WGA for goblet cells, NKM 16-2-4 monoclonal antibody for M cells, and for bacteria and nuclei. (A and B) The confocal Z-plane of an ileal villus from Li-inlAm infected wt mice demonstrates that Li-inlAm was able to target goblet cells (A) and villous M cells (B). Right panels show separated channels and merge of boxed regions, showing the location of bacteria in villous epithelia. (C) The confocal Z-plane of an ileal villus from Li-inlA infected E16P KI mice shows that Li-inlA targeted goblet cells. (D) Relative location of bacteria in mice intestinal epithelia of villi is shown. The total number of Li-inlAm in wt mice intestinal villi epithelia was set to 100. 20 villi from two mice ileal loops were counted in each set. Scale bar, 20 µm. See also Figures S4, S5, S6, Movies S4, S5, S6, S7.

https://doi.org/10.1371/journal.ppat.1003381.g005

InlAm-mNcad interaction has an impact on Lm systemic dissemination in orally inoculated mice

To investigate the impact of InlAm-mNcad interaction on the infection process, we inoculated orally wt and E16P KI mice with Lm-inlAm or Lm. In Lm-infected E16P KI mice in which InlA-Ecad interaction is functional, InlA promoted Lm invasion of the small intestinal tissue and bacterial dissemination to spleen and liver as early as 2 days post infection (dpi) (Figure 6). In contrast, in Lm-inlAm infected wt mice, in which both InlAm-Ecad and InlAm-Ncad interactions are functional, Lm bacterial loads in the small intestinal tissue, spleen and liver were not significantly increased at 2 dpi compared to Lm-infected wt mice, but were at 4 dpi (Figure 6). This delayed systemic dissemination was also observed when comparing Lm-inlAm to LmΔinlA in E16P KI mice (Figure S7). These results demonstrate that, although promoting Lm crossing of the wt mouse intestinal barrier, InlAm delays bacterial systemic dissemination relative to InlA in E16P KI mice, and therefore alters the kinetics of Lm infection in vivo.

thumbnail
Download:
Figure 6. Respective invasive potential of Lm and Lm-inlAm in orally inoculated E16P KI and wt mice.

Mice were orally inoculated with 1010 bacteria for 2 (n = 8, upper panel) or 4 (n = 6, lower panel) days. Bacterial loads in the ileum loops of small intestine, in spleens and in livers were shown. Statistical analysis was performed with the Mann-Whitney u test. See also Figure S7.

https://doi.org/10.1371/journal.ppat.1003381.g006

InlAm-mNcad interaction leads to enhanced intestinal response and compromised intestinal barrier function

Given the changes in infection kinetics induced by InlAm, and the artifactual route of translocation taken by InlAm-expressing bacteria at the intestinal epithelium level, we investigated whether InlAm-Ncad-mediated targeting of villous M cells would have an impact on host responses. Strikingly, oral inoculation of Lm-inlAm led to a significant neutrophil recruitment in wt (Figures 7A and B), E16P KI (Figures S8A and B) and hEcad Tg mice (Figures S8A and B), which was not observed with Lm in E16P KI (Figures 7A and B) and in hEcad Tg mice (Figures S8A and B). Importantly, neutrophil infiltration correlated only with InlAm-mediated invasion, and did not reflect bacterial load in the villi, which was actually the highest in Lm-infected humanized mice, in which no neutrophil infiltration was observed (Figures 7A–C, S8A–C). Moreover, a significant increase in IFN-γ and IL-1β expression was observed in the intestinal tissue of wt mice infected with Lm-inlAm, whereas no significant increase was observed in Lm-infected wt and humanized mice (Figures 7 D and E). Together, these results indicate that InlAm-Ncad-mediated intestinal invasion per se leads to exacerbated host responses compared to InlA-Ecad-mediated intestinal invasion, and are not a reflect of enhanced bacterial tissue invasion.

thumbnail
Download:
Figure 7. Lm expressing InlAm stimulate inflammatory response in the small intestine and compromise intestinal epithelial barrier integrity.

The intestinal ileum was taken from E16P KI and wt mice orally inoculated by 1010 Lm and Lm-inlAm at 24 (A–C) or 48 (D–F) hr post infection, respectively. (A) Anti-Ly6G antibody staining indicates neutrophils (red, highlighted by the open arrowheads). Tissues were stained for Lm (green, highlighted by the arrows) and counterstained with WGA (grey) for goblet cells, respectively. Scale bar, 20 µm. (B) No obvious difference on neutrophil numbers was observed between wt and E16P KI mice infected by Lm, whereas orally Lm-inlAm infection in wt mice induced neutrophil recruitment at 24 hr post infection. Statistical analysis was performed with the Mann-Whitney u test (n = 20 from 2 mice). (C) No obvious difference on bacterial numbers was observed between Lm and Lm-inlAm in wt mice intestinal villi, whereas a significantly increased invasion of Lm in the villi was obsevred in E16P KI mice at 24 hr post infection. Statistical analysis was performed with the Mann-Whitney u test (n = 20 from 2 mice). (D and E) RNA was extracted from the ileum loops of infected or PBS-treated mice 48 hr post infection (n = 4). Following reverse transcription reaction, gene expression was quantified by qPCR with normalization to the GAPDH transcript. Values are expressed as a mean + SD of the fold change relative to that in PBS-treated mice. No significant difference on IFN-γ (D) and IL-1β (E) expression was observed among PBS-treated, Lm and LmΔinlA-infected E16P KI mice. In contrast, Lm-inlAm oral infection induced 5 to 15 fold increase of IFN-γ and IL-1β gene expression in intestinal tissue compared to Lm-infected and PBS-treated wt mice. Statistical analysis was performed with the unpaired Student's t test. (F) Biotin (red) penetration into intestinal lamina propria was done to address intestinal barrier integrity during infection. Mice were sacrificed 2 days post infection. Biotin was injected into ileum loop followed by PBS wash and fixation. Tissues were stained for Lm (green, highlighted by the arrows) and counterstained with WGA (grey) for goblet cells, respectively. Biotin is located within lamina propria of the villi from Lm-inlAm infected mice but not Lm infected wt and E16P KI mice. Scale bar, 20 µm. See also Figure S8.

https://doi.org/10.1371/journal.ppat.1003381.g007

We next assessed intestinal barrier integrity upon infection by testing the intratissular diffusion of biotin administered intraluminally (see Material and Methods) [33]. In wt and humanized mice infected by Lm for two days, biotin localized exclusively to the luminal side of the small intestine (Figures 7F and S8D). In contrast, although the intestinal villi of Lm-inlAm infected wt and humanized mice were not heavily infected, biotin accessed the lamina propria (Figures 7F and S8D). These findings indicate that InlAm-Ncad-mediated intestinal invasion leads to a disruption of intestinal barrier integrity. Together, these results demonstrate that the murinization of InlA profoundly modifies the pathogenic properties of Lm by altering its intestinal portal of entry, host intestinal responses and intestinal barrier integrity.

Discussion

InlA interaction with Ecad allows Lm translocation across the intestinal epithelium and is therefore a critical event in the development of systemic listeriosis, one of the deadliest foodborne infections in human. Because InlA does not interact with mEcad, the discovery and characterization of this key step were made in species permissive to InlA-Ecad interaction (guinea pig, gerbil) and humanized mouse models (hEcad Tg and E16P KI mouse lines) [6], [9]. A genetically engineered Lm strain expressing a murinized InlA (InlAm) enabling interaction with mEcad in vitro has been proposed to constitute an attractive alternative model to study human listeriosis in wt mice [16]. A practical advantage of this latter system is that it can be readily used to infect several different mouse lines. However, a systematic study comparing the properties of Lm expressing InlAm to that of its isogenic parental strain has not been performed, neither in vitro nor in vivo.

Here we show that InlAm is able to recruit mEcad and mediate mEcad-dependent entry into cultured cells. We also show that InlAm mediates entry into goblet cells of wt mice, which express luminally-accessible mEcad. These results confirm that the S192N and Y369S substitutions confer to InlA a phenotype in wt mice which is observed in humanized mice permissive to InlA-Ecad interaction [25].

Importantly, we also uncover that InlAm is able to recruit Ncad and mediate Ncad-dependent internalization. This artifactual interaction translates in vivo into InlAm-dependent targeting of villous M cells, intestinal inflammatory responses, disruption of intestinal barrier integrity and delayed bacterial systemic dissemination in wt mice, as well as in humanized mice. Such stricking phenotypes are not observed in humanized mice orally-inoculated with wt Lm, suggesting that they depend on InlAm-Ncad interaction and invasion of villous M cells, but not on InlAm-Ecad interaction and invasion of goblet cells (Figure 8). It is important to note that these phenotypes are also present in E16P KI and hEcad Tg mice infected with Lm-inlAm, indicating that intestinal inflammation is a direct consequence of InlAm-mediated intestinal invasion, and proving that the absence of inflammation in Lm-infected humanized mice is not a side effect of mouse humanization, but is a genuine property of InlA-dependent intestinal invasion. These results are in agreement with the observation by Wollert et al. that infection with Lm-inlAm leads to severe intestinal inflammation and tissue damage in wt mice [16], and with our earlier observation that InlA has little impact on Lm intestinal responses in mice permissive to InlA-Ecad interaction [6], [27]. This indicates that the murinization of InlA, in addition to broadening the host range of Lm, also extends its receptor repertoire to another member of the classical cadherin family, Ncad, therefore modifying its cell tropism, host responses and the dynamics of infection.

thumbnail
Download:
Figure 8. Revisited model of InlAm-expressing bacteria at the intestinal epithelium.

In humanized mEcad-expressing E16P mouse, as well as in human, Lm targets accessible Ecad around intestinal goblet cells to cross intestinal epithelium, without inducing significant intestinal response. The wt InlA of Lm does not interact with mEcad, thus limits the ability of Lm to cross wt mouse intestinal epithelium. Murinization of InlA enables Lm-inlAm to interact with mEcad and also Ncad of mouse. Lm-inlAm therefore invades not only the goblet cells but also the villous M cells expressing accessible Ncad in the intestinal epithelia of mice. Targeting of villous M cells, which is not observed in Lm-infected E16P KI mice and is not predicted to occur in humans, results in severe intestinal inflammation which induces polymorphonuclear neutrophil (PMN) infiltration, intestinal barrier damage and a delayed systemic dissemination in wt mice, all of which are not observed in human listeriosis and Lm-infected humanized mice.

https://doi.org/10.1371/journal.ppat.1003381.g008

The engineering of InlAm was based on the rational protein design of a modified InlA that would increase InlA-hEcad binding affinity [16]. Indeed, S192N and Y369S substitutions in InlA lead to a 6,700-fold increase in the binding affinity of InlA to hEcad [16]. Here we have shown that this does not translate into increased invasion of hEcad-expressing cells. Before drawing this conclusion, we ensured that the BHK21 cell line we used does not express other cadherins than the one we intended to study. A possible reason for the observed increased level of invasion of Lm-inlAm in Caco-2 cells observed by Wollert et al. is the coexpression of Ecad and Ncad in these cells [21]. These results suggest that InlA-hEcad interaction, although it is of relatively low affinity (KD = 8±4 µM) [16], has been naturally selected to mediate an optimal level of infection.

We have shown that InlB, another major invasion protein of Lm, does not play a significant role for the crossing of the intestinal barrier [23]. In contrast, InlB has been reported to promote Lm expressing InlAm to invade intestinal villi [34]. Our results shed light onto these apparent contradictory results and raise the possibilty that InlAm-Ncad mediated invasion of villous M cells may involve the InlB pathway.

Shigella flexneri, the etiological agent of bacillary dysentery is associated with strong polymorphonuclear infiltration, severe local inflammation, disruption of intestinal barrier integrity, yet no systemic dissemination [35], [36]. In contrast, listeriosis in human and humanized mice is characterized by the paucity of intestinal symptoms, the absence of polymorphonuclear intestinal infiltration, little local inflammation, the absence of intestinal barrier disruption, but systemic dissemination [6], [27], [36], [37]. We have demonstrated that Lm-inlAm triggers pro-inflammatory response and disrupts epithelial integrity in intestinal tissue of wt and humanized mice, and exhibits a delayed systemic dissemination, compared to Lm-infected humanized mice. These observations strongly suggest that the targeting of villous M cells by InlAm-expressing bacteria triggers pro-inflammatory host responses which contain bacterial invasion but lead to intestinal epithelium damages. This fits with the observation that antigen delivery via villous M cells stimulates immune reponses [38]. Like InlAm, Als3 is a Candida albicans invasin that binds both Ecad and Ncad to invade host cells [39]. Candida albicans has been shown to favor gut inflammation and promotes food allergy accompanied by gut epithelial barrier hyperpermeability, the underlying mechanisms of which are so far unclear [40], [41]. Our study indicates that Candida albicans may use Als3 to target Ncad-positive villous M cells, and thereby trigger intestinal inflammation. The specific functions of villous M cells remain poorly understood, yet villous M cells are a particularly abundant constituent of the intestinal epithelium. Our results show that InlAm- and Als3-expressing microorganisms would be particularly instrumental to study villous M cell functions.

Repeated infection of mice in vivo or mouse cells in vitro allows the obtention of “murinized” pathogens adapted to the mouse. Despite the great adaptability of microbes, evolutionary constraints limit pathogen variability [42]. A mutation beneficial under certain environmental conditions may end up as disadvantageous in another, highlighting the fine-tuning of host-microbe interactions. The structure-based rational design of InlAm was proposed as a subtle and elegant way to electively “murinize” a microbial ligand with least impact on the pathogen. However, we provide here evidence that the rationally designed InlAm has gained the unfortunate ability to interact with another surface protein than its cognate receptor Ecad. Even though InlAm mediates Lm crossing of the intestinal barrier, a phenotype which is strictly dependent on InlA-Ecad interaction, the way by which Lm crosses the intestinal barrier in an InlAm-dependent manner differs from what observed with wt Lm in humanized mice and humans, as does the resulting infection process. This illustrates that murinization of human-specific pathogens, although an elegant and rational approach, may unfortunately mislead rather than ease the understanding of human infectious diseases' pathophysiology. Caution must therefore be exercised before engineering and using “murinized” pathogens to study human infectious diseases.

Materials and Methods

Bacterial and cell culture

Bacterial strains, plasmids and primers are listed in Table S1. Note that the sequences of inlA, inlAm in Lm and in Li were confirmed by sequencing, as well as the integration sites of inlA and inlAm in Li and the deletion site of inlA in Lm. Listeria and Escherichia coli strains were respectively cultivated in BHI and LB at 37°C with shaking at 180 rpm. To deliver plasmids into Li, E. coli S17-1 (colistin and nalidixic acid sensitive) cells were transformed with the plasmids followed by conjugation with Li (colistin and nalidixic acid resistant). Mammalian cell lines used in this study were routinely cultured at 37°C in 5% CO2. Except for the culture medium for BHK21 which was supplemented with 5% fetal bovine serum, all the cell culture media were supplemented with 10% fetal bovine serum. Human epithelium LoVo cells were cultured in F12K nutrient GlutaMax medium. Mouse epithelium Nme cells were cultured in DMEM GlutaMax medium supplemented with 10 µg/ml insulin. Mouse CT26 and guinea pig 104C1 cells were cultured in RPMI1640 GlutaMax medium supplemented with HEPES buffer and sodium pyruvate. Human HeLa cells were cultured in MEM GlutaMax medium. Hamster BHK21 cells were cultured in GMEM GlutaMax medium supplemented with tryptose phosphate buffer and HEPES buffer. All the culture medium and related chemicals were purchased from Gibco (Invitrogen). Transient transfection of mammalian cells was performed with jetPRIME transfection kit (Polyplus transfection). The scrambled (sc-37007) and mouse Ncad specific siRNAs (sc-35999) were purchased from Santa Cruz. For the transfection of siRNAs, mouse CT26 cells were seeded into the 24-well plates for 1 day and then transfected with scrambled siRNAs (25 nM) or mNcad-specific siRNAs (25 nM) followed by 1 day incubation and replacement of transfection medium with growth medium another 1 day of incubation before infection. For the transfection of plasmid DNAs, BHK21 cells were transiently transfected with pcDNA3 expression vector harboring the cDNAs of each cadherin (1 µg DNA for each well in a 24-well plate) followed by 2 days incubation before infection.

Construction of plasmids

The strategy to express inlA or inlAm in Li is as described based on integrative plasmid pAD containing a constitutive promoter [43]. The primers EagI_UTRhly-F and UTRhly-R were used to amplify the hly 5′ UTR of Lm EGDe. Full length of inlA and inlAm were amplified from the genomic DNA of Lm EGDe and Lm-inlAm, respectively, with the primers UTRhly_inlA-F and SalI_inlA-R2. The resulting PCR products were ligated to hly 5′ UTR by splicing-by-overlap-extension (SOE) PCR. The final SOE PCR products, containing the entire hly 5′ UTR sequence fused to the start codon of the inlA (hly 5′ UTR-inlA) or inlAm, (hly 5′ UTR-inlAm), were then cloned in pCR-Blunt (Invitrogen) and verified by sequencing. Plasmids containing correct sequence and pAD-cGFP were digested by EagI and SalI. The backbone of pAD-cGFP was ligated with hly 5′ UTR-inlA and hly 5′ UTR-inlAm to form pAD-inlA and pAD-inlAm.

The mouse N-cadherin (mNcad) cDNA was bought from Open Biosystems (Thermo Scientific) and the cDNAs of mouse P-cadherin (mPcad) and mouse VE-cadherin (mVEcad) were from Riken Fantom Clones (Dnaform). To form pcDNA3-mNcad, mNcad cDNA was cloned into EcoRI-NotI site of pcDNA3. The plasmids pcDNA3-mPcad and pcDNA3-mVEcad were constructed by inserting mPcad and mVEcad cDNAs into NotI-KpnI site of pcDNA3, respectively.

Invasion assay

Cell suspensions from confluent monolayers were seeded at a concentration of 5×104 cells per well in 24-well tissue culture plates and grown for 40–48 hr in an antibiotics-free medium at 37°C. Lm and Li strains were grown to OD600 at 0.8 and 0.6 in BHI, respectively. Bacterial culture were then washed with PBS and diluted in cell culture medium without serum. Bacterial suspensions were added to the cells at a multiplicity of infection (MOI) of approximately 50 and incubated for 1 hr. Following wash with complete medium, 10 µg/ml of gentamicin was added to kill the extracellular bacteria for 1 hr. The cells were then washed by complete medium and PBS, and homogenized in PBS supplemented with 0.4% Triton X-100, followed by serial dilution and colony forming units (CFUs) counting. For cadherin recruitment assay, the procedure was the same as the invasion assay except that the cell attachment buffer (HEPES 20 mM, NaCl 150 mM, glucose 50 mM, MgCl2 1 mM, CaCl2 2 mM, MnCl2 1 mM, 0.1% BSA) was used for infection and PBS (Ca2+/Mg2+) (Gibco) was applied to wash the non-attached bacteria stringently followed by fixation.

Animals

Eight to 10-week old C57BL/6 female mice (JANVIER) and isogenic mEcad E16P KI female mice were food restricted overnight but allowed free access to water. Lm culture was prepared as described [6], and inoculated with a feeding needle intragastrically [44]. Mice were then immediately allowed free access to food and water. All the procedures were in agreement with the guidelines of the European Commission for the handling of laboratory animals, directive 86/609/EEC (http://ec.europa.eu/environment/chemicals/lab_animals/home_en.htm) and were approved by the Animal Care and Use Committee of the Institut Pasteur, as well as by the ethical committee of “Paris Centre et Sud” under the number 2010-0020.

Immunofluorescence labeling and immunoblotting

Preparation of tissue sections and whole mount tissues were as described [9], [25]. The following antibodies and fluorescent probes were used for immunostaining and Western blot: anti-hEcad clone HECD-1 mouse monoclonal antibody (Invitrogen), anti-mEcad clone ECCD-2 rat monoclonal antibody (Invitrogen), anti-β-actin clone AC-15 mouse monoclonal antibody (Sigma), anti-Ncad clone 32/N-cadherin mouse monoclonal antibody (BD), anti-Ncad clone GC-4 mouse monoclonal antibody (Sigma), anti-pan cadherin clone CH-19 monoclonal antibody (Sigma), anti-M cell clone NKM 16-2-4 rat monoclonal antibody (Miltenyl Biotec), R6 anti-Li rabbit polyclonal antibody and R11 anti-Lm rabbit polyclonal antibody [45], Rat anti-mouse Ly-6G (BD), wheat germ agglutinin (WGA) conjugated with Alexa Fluor 647 (Jackson ImmunoResearch), Alexa Fuor 488 goat anti-rabbit (Invitrogen), Alexa Fluor 488 or Alexa Fluor 546 goat anti-mouse (Invitrogen), Alexa Fluor 647 donkey anti-rat (Jackson ImmunoResearch), Alexa Fluor 546 goat anti-rat (Invitrogen), Cy3-conjugated streptavidin (Jackson ImmunoResearch) and Hoechst 33342 (Invitrogen).

Biotin penetration experiment

Biotin was used as a molecule to address the integrity of intestinal epithelium as described previously [33]. Briefly, 2 mg/ml of EZ-link Sulfo-NHS-Biotin (Pierce) in PBS was slowly injected into the lumen of ileum loop via the open end adjacent to cecum immediatedly after removal of the entire ileum. After 3 min, the loop was opened followed by PBS wash and 4% paraformaldehye fixation.

Intestinal tissue genes expression quantification

Four mice for each condition were sacrificed 2 days post infection. 1 cm-long of ileal loop of each animal was applied for RNA extraction. The RNA isolation, reverse transcription and quantitative real time PCR (qRT-PCR) were performed as described [46]. Primers used for qRT-PCR were pre-designed, validated RT2 qPCR primer pairs (SABioSciences, Qiagen) as follows: IFNG (IFN-γ, PPM03121A), IL1B (IL-1β, PPM03109F) and GAPDH (PPM02946E).

Statistical analysis

Values are expressed as mean + SD. Statistical comparisons were made using the unpaired Student's t test, Mann-Whitney u test or the χ2 test as indicated. p values<0.05 were considered significant. Significant differences are marked with an asterisk for p<0.05, two asterisks for p<0.01, three asterisks for p<0.001 and four asterisks for p<0.0001.

Supporting Information

Figure S1.

Murinization of InlA allows bacteria to recruit cadherin in Ecad-negative mouse cells, related to Figure 2. (A) Amino acids sequence alignment of first extracellular domains (EC1) of type I classical cadherins and mouse VE-cadherin, a type II classical cadherin. (B) Recruitment of mNcad in CT26 cells was performed as described in Figure 1. The coverslips were stained with the anti-Li antibody and anti-Pan-cadherin (Pan-cad) antibody. Right panels show the boxed regions of separated channels and merge, demonstrating the recruitment of cadherin proteins specifically by Li-inlAm. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.s001

(PDF)

Figure S2.

Murinization of InlA promotes bacterial entry into Ecad-negative, Ncad-positive human and guinea pig cells, related to Figure 2. Human HeLa cells and guinea pig 104C1 cells are Ecad-negative and Ncad-positive cells. Cell invasion ability was evaluated by counting intracellular gentamicin resistant bacteria following 1 hr of infection (MOI 50) and 1 hr of gentamicin (10 µg/ml) incubation. No difference in bacterial entry is seen between Lm and its isogenic inlA null mutant (LmΔinlA), whereas the Lm harboring inlAm of which chromosomal inlA is replaced by inlAm (Lm-inlAm) promoted bacterial entry into both HeLa and 104C1 cells. Values are expressed as a mean + SD (n = 3). Statistical analysis was performed with the unpaired Student's t test.

https://doi.org/10.1371/journal.ppat.1003381.s002

(PDF)

Figure S3.

Ncad expression is detected on the apical pole of villous M cells in E16P KI mice, related to Figure 4. Intestinal tissue of wt mice was fixed and stained for luminally accessible (acc) Ncad with antibody against extracellular domain of Ncad (clone GC4) before tissue permeabilization, M cells with NKM 16-2-4 antibody, WGA and nuclei after tissue permeabilization. Projection of a 30 µm thick reconstructed intestinal villus (A) and one xy plane (B) are shown. Right panels show separated channels and merge of boxed regions in (B), showing Ncad on the apical side of NKM 16-2-4-positive cells. See also S7. NKM 16-2-4 antibody is a monoclonal antibody raised against α(1,2) fucose moiety in absence of neighboring sialic acids, a specific marker on M cells surface. WGA was used to stain the mucus of goblet cells and cell membrane. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.s003

(PDF)

Figure S4.

Lm-inlAm target goblet cells and villous M cells in wt mice, related to Figure 5. The intestinal ileum was taken from wt or E16P KI mice orally inoculated by 1010 Lm or Lm-inlAm at 5 hr post infection. The intestinal tissues were fixed. Vibratome sections were stained with WGA for goblet cells, NKM 16-2-4 monoclonal antibody for M cells, and for bacteria and nuclei. (A and B) The confocal Z-plane of an ileal villus from Lm-inlAm infected wt mice demonstrates that Lm-inlAm was able to target goblet cells (A, see also Figure S5A and Movie S5) and villous M cells (B, see also Figure S5B, and Movie S6). Right panels show separated channels and merge of boxed regions, showing the location of bacteria in villous epithelia. (C) The confocal Z-plane of an ileal villus from Lm infected E16P KI mouse shows that Lm targeted goblet cells (see also Figure S5C and Movie S7). (D) Relative location of bacteria in mice intestinal epithelia of villi is shown. The total number of Lm-inlAm in wt mice intestinal villi epithelia was set to 100. 20 villi from two mice ileal loops were counted in each set. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.s004

(PDF)

Figure S5.

Intracellular location of bacteria targeting goblet and villous M cells, related to Figure 5. Orthogonal views of the infected cells in wt mice infected with Lm-inlAm (A and B, related to Figures S4A and B, respectively) and in E16P KI mice infected by Lm (C, related to Figure S4C) presented in Figure S5 were shown. These images demonstrate that the bacteria highlighted in the Figure S4 were intracelullar. See also Movies S5, S6 and S7.

https://doi.org/10.1371/journal.ppat.1003381.s005

(PDF)

Figure S6.

Lm-inlAm attached to villous M cells and Lm-inlAm having reached the lamina propria underlying villous M cells, related to Figure 5. The intestinal ileum was taken from the wt mice orally inoculated by 1010 Lm or Lm-inlAm at 5 hr post infection. The intestinal tissues were fixed. Vibratome sections were stained with WGA for goblet cells, NKM 16-2-4 monoclonal antibody for M cells, and for bacteria and nuclei. Results shown are two different confocal Z-planes of an ileal villus from Lm-inlAm infected wt mice. Lm-inlAm was found to attach to the apical pole of villous M cell in the upper panel and reach the lamina propria in the lower panel. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.s006

(PDF)

Figure S7.

Respective invasive potential of Lm and Lm-inlAm in orally inoculated E16P KI mice, related to Figure 6. Mice were orally inoculated by 1010 bacteria for 2 (n = 6, upper panel) or 4 (n = 6, lower panel) days. Bacterial loads in the ileum loops of small intestine, the spleens and livers were shown. Statistical analysis was performed with the Mann-Whitney u test.

https://doi.org/10.1371/journal.ppat.1003381.s007

(PDF)

Figure S8.

Lm-inlAm induced neutrophil infiltration and compromised intestinal epithelial barrier integrity in E16P KI and hEcad Tg mice, related to Figure 7. The intestinal ileum was taken from E16P KI and hEcad Tg mice orally inoculated by 1010 LmΔinlA, Lm and Lm-inlAm 24 hr (A to C) or 48 hr (D) post infection. (A) Anti-Ly6G antibody staining indicates neutrophils (red, highlighted by the open arrowheads). Tissues were stained for Lm (green, highlighted by the arrows) and counterstained with WGA (grey) for goblet cells and epithelia. Scale bar, 20 µm. (B) No obvious difference on neutrophil numbers was observed between LmΔinlA and Lm infection in hEcad Tg mice, whereas Lm-inlAm infection induced neutrophil infiltration in the intestinal villi compared to Lm in both E16P KI and hEcad Tg mice. (C) The number of bacteria in each infected villus was also quantified. Bacteria load of Lm in the intestinal villi was higher than that of Lm-inlAm in both E16P KI and hEcad Tg mice upon oral infection 24 hpi. In order to compare the result of Lm-inlAm with Lm in E16P KI mice, the data of Lm-infected E16P KI mice shown here in B and C were from figure 7B and C, respectively. Statistical analysis was done with Mann-Whitney u test (n = 20 villi from 2 mice). (D) Biotin was injected into ileum loop followed by PBS wash and fixation. Tissues were stained for Lm (green, highlighted by the arrows) and counterstained with WGA (grey) for goblet cells and epithelia. Biotin is located within lamina propria of the villi from Lm-inlAm infected mice but not Lm infected mice. Scale bar, 20 µm.

https://doi.org/10.1371/journal.ppat.1003381.s008

(PDF)

Movie S1.

Luminally accessible Ncad is expressed on the apical poles of villous M cells in wt mice, related to Figure 4. Whole mount intestinal tissue of a wt mouse was stained before permeabilization for accessible mNcad (green) and NKM 16-2-4 for M cells (red), and after permeabilization for nuclei (blue) and WGA for goblet cells (grey). Intestinal villus is oriented with the villus tip facing the viewer. The luminally accessible apical surface of villous M cells is labeled with the anti-Ncad antibody. Images were acquired as a z stack by confocal microscopy and assembled as a three-dimensional reconstruction with Imaris software.

https://doi.org/10.1371/journal.ppat.1003381.s009

(MOV)

Movie S2.

Peyer's patch M cells do not express luminally accessible Ncad in wt mice, related to Figure 4. Whole mount intestinal tissue of a wt mouse was stained before permeabilization for accessible mNcad (green) and NKM 16-2-4 for M cells (red), and after permeabilization for nuclei (blue) and WGA for goblet cells (grey). The luminally accessible apical surface of Peyer's patch M cells is not labeled with the anti-Ncad antibody. Intestinal Peyer's patch is oriented with the tip facing the viewer. Images were acquired as a z stack by confocal microscopy and assembled as a three-dimensional reconstruction with Imaris software.

https://doi.org/10.1371/journal.ppat.1003381.s010

(MOV)

Movie S3.

Luminally accessible Ncad is expressed on the apical poles of villous M cells in E16P KI mice, related to Figure 4. Whole mount intestinal tissue of an E16P KI mouse was stained before permeabilization for accessible mNcad (green) and NKM 16-2-4 for M cells (red), and after permeabilization for nuclei (blue) and WGA for goblet cells (grey). Intestinal villus is oriented with the villus tip facing the viewer. The luminally accessible apical surface of villous M cells is labeled with the anti-Ncad antibody. Images were acquired as a z stack by confocal microscopy and assembled as a three-dimensional reconstruction with Imaris software.

https://doi.org/10.1371/journal.ppat.1003381.s011

(MOV)

Movie S4.

Li-inlAm targets both villous M cells and goblet cells in the intestinal villi upon oral inoculation of wt mice, related to Figure 5. Ileal loop of a wt mouse orally infected by Li-inlAm was taken 5 hr post infection, followed by fixation and staining for Li (green), M cells (red), goblet cells (grey) and nuclei (blue) after permeabilization. Images were acquired and assembled as described for Movie S1.

https://doi.org/10.1371/journal.ppat.1003381.s012

(MOV)

Movie S5.

Lm-inlAm targets goblet cells in the intestinal villi upon oral inoculation of wt mice, related to Figure 5. Ileal loop of a wt mouse orally infected by Lm-inlAm was taken 5 hr post infection, followed by fixation. Vibratome section was stained for Lm-inlAm (green), M cells (red), goblet cells (grey) and nuclei (blue) after permeabilization. Images were acquired and assembled as described for Movie S1.

https://doi.org/10.1371/journal.ppat.1003381.s013

(MOV)

Movie S6.

Lm-inlAm targets villous M cells in the intestinal villi upon oral inoculation of wt mice, related to Figure 5. Ileal loop of a wt mouse orally infected by Lm-inlAm was taken 5 hr post infection, followed by fixation. Vibratome section was stained for Lm-inlAm (green), M cells (red), goblet cells (grey) and nuclei (blue) after permeabilization. Images were acquired and assembled as described for Movie S1.

https://doi.org/10.1371/journal.ppat.1003381.s014

(MOV)

Movie S7.

Lm targets goblet cells in the intestinal villi upon oral inoculation of E16P KI mice, related to Figure 5. Ileal loop of a wt mouse orally infected by Lm was taken 5 hr post infection, followed by fixation. Vibratome section was stained for Lm (green), M cells (red), goblet cells (grey) and nuclei (blue) after permeabilization. Images were acquired and assembled as described for Movie S1.

https://doi.org/10.1371/journal.ppat.1003381.s015

(MOV)

Table S1.

Bacterial strains, plasmids and primers used in this study.

https://doi.org/10.1371/journal.ppat.1003381.s016

(DOC)

Acknowledgments

We thank Colin Hill for providing Lm strain EGDe expressing murinized inlA and its isogenic parental strain EGDe. We thank Pascale Cossart for providing pcDNA3-hEcad, pcDNA3-mEcad and pAD-cGFP plasmids, LoVo, Nme and 104C1 cell lines, and helpful discussion. We thank the members of the Biology of Infection Unit and the Imagopole Facility of the Institut Pasteur for their help and support, and Thomas Cantinelli from the National Reference Center Listeria for inlA sequences analysis. We thank Serge Mostowy for critical reading of the manuscript.

Author Contributions

Conceived and designed the experiments: YHT OD HB ML. Performed the experiments: YHT OD HB. Analyzed the data: YHT OD HB ML. Wrote the paper: YHT OD HB ML.

References

  1. 1. Brown NF, Wickham ME, Coombes BK, Finlay BB (2006) Crossing the line: selection and evolution of virulence traits. PLoS Pathog 2: e42.
  2. 2. Lecuit M (2007) Human listeriosis and animal models. Microbes Infect 9: 1216–1225.
  3. 3. Buer J, Balling R (2003) Mice, microbes and models of infection. Nat Rev Genet 4: 195–205.
  4. 4. Lecuit M, Cossart P (2002) Genetically-modified-animal models for human infections: the Listeria paradigm. Trends Mol Med 8: 537–542.
  5. 5. Dorner M, Horwitz JA, Robbins JB, Barry WT, Feng Q, et al. (2011) A genetically humanized mouse model for hepatitis C virus infection. Nature 474: 208–211.
  6. 6. Lecuit M, Vandormael-Pournin S, Lefort J, Huerre M, Gounon P, et al. (2001) A transgenic model for listeriosis: role of internalin in crossing the intestinal barrier. Science 292: 1722–1725.
  7. 7. Ren RB, Costantini F, Gorgacz EJ, Lee JJ, Racaniello VR (1990) Transgenic mice expressing a human poliovirus receptor: a new model for poliomyelitis. Cell 63: 353–362.
  8. 8. Sun H, Ringdahl U, Homeister JW, Fay WP, Engleberg NC, et al. (2004) Plasminogen is a critical host pathogenicity factor for group A streptococcal infection. Science 305: 1283–1286.
  9. 9. Disson O, Grayo S, Huillet E, Nikitas G, Langa-Vives F, et al. (2008) Conjugated action of two species-specific invasion proteins for fetoplacental listeriosis. Nature 455: 1114–1118.
  10. 10. Mathur R, Oh H, Zhang D, Park SG, Seo J, et al. (2012) A mouse model of Salmonella Typhi infection. Cell 151: 590–602.
  11. 11. Song J, Willinger T, Rongvaux A, Eynon EE, Stevens S, et al. (2010) A mouse model for the human pathogen Salmonella Typhi. Cell Host Microbe 8: 369–376.
  12. 12. Chua BH, Phuektes P, Sanders SA, Nicholls PK, McMinn PC (2008) The molecular basis of mouse adaptation by human enterovirus 71. J Gen Virol 89: 1622–1632.
  13. 13. Monk IR, Casey PG, Hill C, Gahan CG (2010) Directed evolution and targeted mutagenesis to murinize listeria monocytogenes internalin A for enhanced infectivity in the murine oral infection model. BMC Microbiol 10: 318.
  14. 14. Marchetti M, Arico B, Burroni D, Figura N, Rappuoli R, et al. (1995) Development of a mouse model of Helicobacter pylori infection that mimics human disease. Science 267: 1655–1658.
  15. 15. Bitzegeio J, Bankwitz D, Hueging K, Haid S, Brohm C, et al. (2010) Adaptation of hepatitis C virus to mouse CD81 permits infection of mouse cells in the absence of human entry factors. PLoS Pathog 6: e1000978.
  16. 16. Wollert T, Pasche B, Rochon M, Deppenmeier S, van den Heuvel J, et al. (2007) Extending the host range of Listeria monocytogenes by rational protein design. Cell 129: 891–902.
  17. 17. Gaillard JL, Berche P, Frehel C, Gouin E, Cossart P (1991) Entry of L. monocytogenes into cells is mediated by internalin, a repeat protein reminiscent of surface antigens from gram-positive cocci. Cell 65: 1127–1141.
  18. 18. Mengaud J, Ohayon H, Gounon P, Mège R-M, Cossart P (1996) E-Cadherin Is the Receptor for Internalin, a Surface Protein Required for Entry of L. monocytogenes into Epithelial Cells. Cell 84: 923–932.
  19. 19. Dejana E (2004) Endothelial cell-cell junctions: happy together. Nat Rev Mol Cell Biol 5: 261–270.
  20. 20. Derycke LD, Bracke ME (2004) N-cadherin in the spotlight of cell-cell adhesion, differentiation, embryogenesis, invasion and signalling. Int J Dev Biol 48: 463–476.
  21. 21. Makagiansar IT, Yusuf-Makagiansar H, Ikesue A, Calcagno AM, Murray JS, et al. (2002) N-cadherin involvement in the heterotypic adherence of malignant T-cells to epithelia. Mol Cell Biochem 233: 1–8.
  22. 22. Lecuit M, Dramsi S, Gottardi C, Fedor-Chaiken M, Gumbiner B, et al. (1999) A single amino acid in E-cadherin responsible for host specificity towards the human pathogen Listeria monocytogenes. Embo J 18: 3956–3963.
  23. 23. Khelef N, Lecuit M, Bierne H, Cossart P (2006) Species specificity of the Listeria monocytogenes InlB protein. Cell Microbiol 8: 457–470.
  24. 24. Shen Y, Naujokas M, Park M, Ireton K (2000) InIB-dependent internalization of Listeria is mediated by the Met receptor tyrosine kinase. Cell 103: 501–510.
  25. 25. Nikitas G, Deschamps C, Disson O, Niault T, Cossart P, et al. (2011) Transcytosis of Listeria monocytogenes across the intestinal barrier upon specific targeting of goblet cell accessible E-cadherin. J Exp Med 208: 2263–2277.
  26. 26. Jacquet C, Doumith M, Gordon JI, Martin PM, Cossart P, et al. (2004) A molecular marker for evaluating the pathogenic potential of foodborne Listeria monocytogenes. J Infect Dis 189: 2094–2100.
  27. 27. Lecuit M, Sonnenburg JL, Cossart P, Gordon JI (2007) Functional genomic studies of the intestinal response to a foodborne enteropathogen in a humanized gnotobiotic mouse model. J Biol Chem 282: 15065–15072.
  28. 28. Lecuit M, Ohayon H, Braun L, Mengaud J, Cossart P (1997) Internalin of Listeria monocytogenes with an intact leucine-rich repeat region is sufficient to promote internalization. Infect Immun 65: 5309–5319.
  29. 29. Vleminckx K, Vakaet L Jr, Mareel M, Fiers W, van Roy F (1991) Genetic manipulation of E-cadherin expression by epithelial tumor cells reveals an invasion suppressor role. Cell 66: 107–119.
  30. 30. Leung N, Turbide C, Balachandra B, Marcus V, Beauchemin N (2008) Intestinal tumor progression is promoted by decreased apoptosis and dysregulated Wnt signaling in Ceacam1-/- mice. Oncogene 27: 4943–4953.
  31. 31. Geiger B, Volberg T, Ginsberg D, Bitzur S, Sabanay I, et al. (1990) Broad spectrum pan-cadherin antibodies, reactive with the C-terminal 24 amino acid residues of N-cadherin. J Cell Sci 97 (Pt 4) 607–614.
  32. 32. Redfield A, Nieman MT, Knudsen KA (1997) Cadherins promote skeletal muscle differentiation in three-dimensional cultures. J Cell Biol 138: 1323–1331.
  33. 33. Guttman JA, Samji FN, Li Y, Vogl AW, Finlay BB (2006) Evidence that tight junctions are disrupted due to intimate bacterial contact and not inflammation during attaching and effacing pathogen infection in vivo. Infect Immun 74: 6075–6084.
  34. 34. Pentecost M, Kumaran J, Ghosh P, Amieva MR (2010) Listeria monocytogenes internalin B activates junctional endocytosis to accelerate intestinal invasion. PLoS Pathog 6: e1000900.
  35. 35. Phalipon A, Sansonetti PJ (2007) Shigella's ways of manipulating the host intestinal innate and adaptive immune system: a tool box for survival? Immunol Cell Biol 85: 119–129.
  36. 36. Sansonetti PJ, Di Santo JP (2007) Debugging how bacteria manipulate the immune response. Immunity 26: 149–161.
  37. 37. Munoz P, Rojas L, Bunsow E, Saez E, Sanchez-Cambronero L, et al. (2012) Listeriosis: An emerging public health problem especially among the elderly. J Infect 64: 19–33.
  38. 38. Nochi T, Yuki Y, Matsumura A, Mejima M, Terahara K, et al. (2007) A novel M cell-specific carbohydrate-targeted mucosal vaccine effectively induces antigen-specific immune responses. J Exp Med 204: 2789–2796.
  39. 39. Phan QT, Myers CL, Fu Y, Sheppard DC, Yeaman MR, et al. (2007) Als3 is a Candida albicans invasin that binds to cadherins and induces endocytosis by host cells. PLoS Biol 5: e64.
  40. 40. Kumamoto CA (2011) Inflammation and gastrointestinal Candida colonization. Curr Opin Microbiol 14: 386–391.
  41. 41. Yamaguchi N, Sugita R, Miki A, Takemura N, Kawabata J, et al. (2006) Gastrointestinal Candida colonisation promotes sensitisation against food antigens by affecting the mucosal barrier in mice. Gut 55: 954–960.
  42. 42. Arnold SJ (1992) Constraints on phenotypic evolution. Am Nat 140 Suppl 1: S85–107.
  43. 43. Balestrino D, Hamon MA, Dortet L, Nahori MA, Pizarro-Cerda J, et al. (2010) Single-cell techniques using chromosomally tagged fluorescent bacteria to study Listeria monocytogenes infection processes. Appl Environ Microbiol 76: 3625–3636.
  44. 44. Disson O, Nikitas G, Grayo S, Dussurget O, Cossart P, et al. (2009) Modeling human listeriosis in natural and genetically engineered animals. Nat Protoc 4: 799–810.
  45. 45. Dramsi S, Levi S, Triller A, Cossart P (1998) Entry of Listeria monocytogenes into neurons occurs by cell-to-cell spread: an in vitro study. Infect Immun 66: 4461–4468.
  46. 46. Bierne H, Travier L, Mahlakoiv T, Tailleux L, Subtil A, et al. (2012) Activation of type III interferon genes by pathogenic bacteria in infected epithelial cells and mouse placenta. PLoS One 7: e39080.