Listeriolysin O-dependent host surfaceome remodeling modulates Listeria monocytogenes invasion.

Listeria monocytogenes is a pathogenic bacterium that invades epithelial cells by activating host signaling cascades, which promote bacterial engulfment within a phagosome. The pore-forming toxin listeriolysin O (LLO), which is required for bacteria phagosomal escape, has also been associated with the activation of several signaling pathways when secreted by extracellular bacteria, including Ca2+ influx and promotion of L. monocytogenes entry. Quantitative host surfaceome analysis revealed significant quantitative remodeling of a defined set of cell surface glycoproteins upon LLO treatment, including a subset previously identified to play a role in the L. monocytogenes infection process. Our data further shows that the lysosomal-associated membrane proteins LAMP-1 and LAMP-2 are translocated to the cellular surface and those LLO-induced Ca2+ fluxes are required to trigger the surface relocalization of LAMP-1. Finally, we identify late endosomes/lysosomes as the major donor compartments of LAMP-1 upon LLO treatment and by perturbing their function, we suggest that these organelles participate in L. monocytogenes invasion.

INTRODUCTION

Listeria monocytogenes is a Gram-positive bacterium responsible for listeriosis, a food-borne disease characterized by gastroenteritis and bacteremia in immuno-compromised individuals, meningitis in the new-born and abortion in pregnant women (Lecuit 2007; Swaminathan and Gerner-Smidt 2007). Listeria monocytogenes pathogenicity relies on its ability to cross several host tissue barriers including the intestine, the placenta and the blood-brain barrier, and to survive within macrophages as well as to invade and replicate in epithelial cells (Pizarro-Cerdá, Kühbacher and Cossart 2012; Pizarro-Cerdá et al.2016). Cellular invasion requires the presence of the bacterial surface proteins internalin (InlA) and/or InlB and their interaction with the cellular receptors E-cadherin and/or Met, respectively (Gaillard et al.1991; Mengaud et al.1996; Shen et al.2000; Gessain et al.2015; Quereda et al.2018). Activation of host cell receptors leads to recruitment of a clathrin-based internalization machinery that promotes actin polymerization and plasma membrane rearrangements orchestrating bacterial engulfment (Pizarro-Cerdá and Cossart 2009; Pizarro-Cerdá, Bonazzi and Cossart 2010; Bonazzi et al.2012). Following cellular invasion, L. monocytogenes disrupts its internalization compartment through the action of the pore-forming toxin listeriolysin O (LLO), leading to bacterial escape to the host cell cytoplasm (Quereda, Cossart and Pizarro-Cerdá 2016). Cytoplasmic bacteria, through the activity of the surface protein ActA, promote their own movement via an actin-based motility system that leads to cell-to-cell spread (Gouin, Welch and Cossart 2005; Van Troys et al.2008).

Bacterial invasion of non-polarized epithelial cells by means of the InlB-dependent internalization pathway is a complex process that requires interaction of many molecular partners. Met is a receptor tyrosine kinase that upon interaction with InlB promotes its own phosphorylation and that of the molecular adaptors Cbl, Gab1 and Shc, which are involved in the recruitment of a type I PI 3-kinase to bacterial entry sites (Ireton et al.1996; Ireton, Payrastre and Cossart 1999). Phosphatidylinositol(3,4,5)P3 produced by the PI 3-kinase is redistributed in lipid rafts and is required for Rac activation (Seveau 2004; Seveau et al.2007; Pizarro-Cerdá, Kühbacher and Cossart 2014; Gessain et al.2015). Downstream of Rac, several WASP-related proteins promote actin polymerization by the Arp2/3 complex while ENA/Vasp favor actin filament elongation (Bierne et al.2005). The actin depolymerization process is also highly regulated at bacterial entry sites (Kühbacher et al.2012) and disassembly of an actin-rich phagocytic cup is required for cell invasion (Bierne et al.2001). In parallel, Cbl ubiquitinates Met through its ubiquitin-ligase activity and therefore favors the recruitment to L. monocytogenes invasion foci of a clathrin-associated machinery that promotes additional actin polymerization (Veiga and Cossart 2005; Veiga et al.2007; Bonazzi et al.2011).

Besides InlB, ActA and LLO have been reported to foster cellular invasion by L. monocytogenes in epithelial cells (Suárez et al.2001; Dramsi and Cossart 2003; Vadia et al.2011). In particular, LLO is a secreted pleiotropic bacterial effector that acts extracellularly on host plasma membranes by promoting Ca2+ influx in target cells (Dramsi and Cossart 2003), modulating mitochondrial dynamics (Stavru and Cossart 2011; Stavru et al.2013), perturbing the telomerase function (Samba-Louaka, Stavru and Cossart 2012), inducing histone post-translational modifications (Hamon et al.2007; Hamon and Cossart 2011) and impairing host cell protein sumoylation (Ribet et al.2010; Malet et al.2018). Using a quantitative chemoproteomic strategy (Wollscheid et al.2009; Hofmann, Bausch-Fluck and Wollscheid 2013), we assessed the impact of LLO exposure on the host cell surface exposed N-glycoproteins, here referred as surfaceome. Cross-correlation of these results with functional data from a previous genome-wide siRNA screen for L. monocytogenes infection (Kühbacher et al.2015) revealed key molecular players that positively or negatively affect bacterial infection. We show in particular that LLO promotes fusion of late endosomal compartments with the plasma membrane at bacterial entry foci and participates in efficient L. monocytogenes invasion of host cells.

MATERIALS AND METHODS

Cell culture and bacteria

HeLa CCL-2 cells from the American Type Culture Collection were cultured at 37°C in DMEM supplemented with 10% fetal calf serum (FCS) in a humidified 10% CO2 atmosphere. Listeria monocytogenes strain EGDe.PrfA* was cultured in brain heart infusion at 37°C. For infection experiments, bacteria were grown for 3 to 5 h in order to obtain cultures in logarithmic phase or overnight to reach the stationary phase. One ml of bacteria culture was washed three times with 1 ml PBS, the OD at 600 nm was measured to calculate the bacteria number and cells were infected in DMEM at a multiplicity of infection of 25 to 50 bacteria per cell (Pizarro-Cerdá, Lecuit and Cossart 2002; Kühbacher, Cossart and Pizarro-Cerdá 2014). To synchronize infection, bacteria were centrifuged on cells for 5 min at 1000 rpm at room temperature. For treatment of cells with purified LLO, cells were starved for at least 3 h with DMEM. Purified LLO was diluted in DMEM at indicated concentrations.

Probes, recombinant proteins and drugs

EGTA, horseradish peroxidase (HRP), H2O2, triton X-100, saponin and 3,3΄-Diaminobenzidine tetrahydrochloride (DAB) were purchased from Sigma. Recombinant LLO and LLOw492A were produced as previously described (Pizarro-Cerdá, Lecuit and Cossart 2002; Glomski, Decatur and Portnoy 2003).

Plasmid transfection

Transfection with a plasmid coding for canine EGFP-LAMP-1 (described in Lozach et al.2011) was performed 24 h prior to assays using the FuGENE HD transfection procedure (Roche Applied Science) according to the manufacturers protocol. The final plasmid concentration was 0.5 μg/ml.

Flow cytometry

Cells were infected or treated with recombinant proteins at 37°C for indicated times. Subsequently cells were detached for 5 min at 37°C using calcium and magnesium free PBS supplemented with 0.5mM EDTA. Cells were then fixed for 20 min with 4% paraformaldehyde (PFA) in PBS on ice and stained for surface LAMP-1 or LAMP-2 or treated with isotype antibody for 1h in PBS supplemented with 1% bovine serum albumin (BSA). Fluorescently labeled Antibodies (anti-LAMP-1-Phycoerythrin 555801, anti-LAMP-2-FITC 560946 and mouse isotype control IgG-Phycoerythrin 555749 (BD Bioscience)) were diluted according to the manufacturers recommendations. For staining of total LAMP-1 0.05% saponin was added to the staining solution. Fluorescent signals were measured using a FACSCalibur flow cytometer (BD Biosciences). Dead cells were excluded by gating.

Hexosaminidase assay

To measure the release of lysosomal hexosaminidase, cells in 96 well plates were stimulated by infection or treatment with recombinant LLO for 20 min in DMEM without phenol red. 87.5 μl of supernatants were then transferred to a new 96 well plate. The rest of the supernatant was aspirated and the cells were lysed for 15 min at 37°C by addition of 100 μl PBS supplemented with 0.1% Triton X-100. Lysates were then diluted 1:10 and 87.5 μl were transferred to a new 96 well plate. 87.5 μl PBS was added to blank wells. 12.5 μl of 6 mM 4-methyl-umbellyferyl-N-acetylb-d-glucosaminide (69 585, Sigma) in citrate-phosphate buffer, pH 4.5 was added to each well and after incubation for 15 min at 37°C the reaction was stopped by addition of 25 μl 2 M Na2CO3 and 1.1 M glycine. Fluorescence at 450 nm was measured after excitation at 365 nm using a spectrofluorometer (Flx-Xenius, Safas).

Immunofluorescence staining and live-cell imaging

For immunofluorescence staining 50 000 cells were seeded on 12 mm round glass coverslips in 24 well plates the day before the experiment. After fixation with 4% PFA, cells were stained with a Phycoerythrin coupled antibody against LAMP-1 for 40 min in staining solution (PBS, 1% BSA, 0.05% saponin). For labeling of extracellular LAMP-1, staining solution without saponin was used. Differential bacteria staining was carried out by first staining extracellular bacteria with a rabbit-derived antibody (R11) and a secondary Alexa Fluor 488-coupled antibody (Molecular Probes) both for 30 min in staining solution without saponin followed by thorough washes and staining of total bacteria with DAPI or with R11 and a secondary Alexa Fluor 350- or Cy3-coupled antibody. Coverslips were mounted on microscopy glass slides with Fluoromount (Interchim) and images were acquired with a 63x or 100x objective installed on an inverted wide-field fluorescence microscope (Axiovert 200M, Carl Zeiss Microscopy), equipped with an EMCCD camera (Neo, Andor) and the imaging software MetaMorph (Molecular Devices). For live-cell imaging, 200 000 cells were seeded and transfected with a plasmid coding for canine LAMP-1-GFP the day before in 35mm glass bottom dishes (Mattek). For infections and image acquisition medium was replaced by DMEM without phenol red. Infections were synchronized by centrifugation for 5min at 1000 rpm before imaging. Images were acquired with a 37°C pre-warmed microscope (Axiobserver Z1, Zeiss) equipped with an EMCCD camera (Evolve, Photometrics), a spinning disk head (CSU X-1, Yokogawa) and a laser launch system (405, 491, 546 and 642 nm, Photometrics) in a 5% CO2-containing atmosphere. Images were taken every 5 to 20 sec. For imaging of LLO-treated cells, 1 nm LLO was added during image acquisition as indicated in the corresponding figure.

Exosome purification

For each condition, three 125 cm2 cell culture flasks with 70% to 90% confluent HeLa cells were prepared. Before treatment with mock DMEM or 1 nM LLO in DMEM for 20 min, cells were starved in DMEM for 3 h. Supernatants of treated cells were collected and centrifuged first at 2000 rcf for 20 min, supernatants were again transferred to new tubes and centrifuged at 10 000 rcf for another 30 min to remove cell debris. Cleared supernatants were centrifuged at 100 000 rcf for 70 min to concentrate the exosome fraction which was then washed by removal of the supernatant, re-suspension in PBS and repetition of the previous centrifugation step. The washed exosome fraction was kept at –80°C as aliquotes of 25 μl. For fluorescent detection of exosomes, purified fractions were seeded on glass coverslips and fixed with 4% PFA. Exosomes were then stained for CD63 with a mouse-anti-CD63 antibody TS63 and a secondary Alexa Fluor 488-coupled antibody.

Lysosome/late endosome quenching

To inactivate late endocytic compartments, the endocytic pathway was loaded with HRP by treatment with DMEM, supplemented with 1% BSA and 5 mg/ml HRP for 1.5 h at 37°C before washout of HRP three times with Ca2+ and Mg2+ containing PBS and a chase for 30 min at 37°C in DMEM, supplemented with 1% BSA to allow the HRP to accumulate in late endocytic compartments and be cleared from the early endocytic pathway. Cells were then transferred to ice and treated with 100 μg/ml DAB and 0.0025 % H2O2 in Ca2+ and Mg2+ containing PBS for another 30 min. After washing three times with PBS and once with DMEM cells were transferred again to 37°C for 5 min and then immediately infected or treated with LLO.

Cell Surface Capture technology

HeLa cells were expanded on 14 cm dishes in DMEM supplemented with 10% FCS and 4 mM L-Glutamine. Cells were starved overnight in DMEM at 37°C and then treated for 20 min with purified LLO at a concentration 1 nM. Control cells were incubated in DMEM at 37°C for 20 min. Cell surface proteins were captured using the cell surface capture (CSC) technology as described elsewhere (Wollscheid et al.2009; Hofmann, Bausch-Fluck and Wollscheid 2013). Briefly, cells were washed on plates twice in ice-cold PBS then in ice-cold PBS pH 6.5 and oxidized in PBS pH 6.5 supplemented with 2 mM sodium meta-periodate for 15 min at 4°C. The oxidized cells were washed with PBS pH 6.5, biocytin hydrazide was added to a final concentration of 77 μM in PBS pH 6.5 and incubated at 4°C for 1h. After washing in PBS, cells were scraped and resuspended in 50 mM ammonium bicarbonate and lysed by sonication. Nuclei were pelleted and the supernatant collected. Rapigest was added to a final concentration of 0.1% and the solution was sonicated for 3 min. The proteins were reduced with 5 mM Tris(2-carboxyethyl)phosphine and alkylated with 10 mM iodoacetamide. Proteins were enzymatically digested overnight with trypsin at a substrate:enzyme ratio of 50:1. After trypsin heat inactivation, streptavidin beads were added to the peptides and incubated for 2 h at room temperature. The beads were washed with 5 M sodium chloride, triton X-100 based buffer, 50 mM ammonium bicarbonate, 100 mM sodium bicarbonate adjusted to pH 11 and finally with 50 mM ammonium bicarbonate. Peptide-N-glycosidase F was added to the beads and incubated overnight at 37°C. Released peptides were desalted on a C18-column and separated by reversed-phase chromatography on a high-performance liquid chromatography (HPLC) column (75-μm inner diameter, New Objective) that was packed in-house with a 10-cm stationary phase (Magic C18AQ, 200 Å, 3 micron Michrom Bioresources) and connected to a nano-flow HPLC combined with an autosampler (EASY-nLC II, Proxeon). The HPLC was coupled to a LTQ-Orbitrap XL mass spectrometer (Thermo Scientific) equipped with a nanoelectrospray ion source (Thermo Scientific). Peptides were loaded onto the column with 99.9% buffer A (99.9% H2O, 0.1% FA) and eluted with 300 nl/min over a 60-min linear gradient from 7%–35% buffer B (99.9% ACN, 0.1% FA). After the gradient, the column was washed with 80% buffer B and re-equilibrated with buffer A. Mass spectra were acquired in a data-dependent manner, with an automatic switch between MS and MS/MS scans. High- resolution MS scans were acquired in the Orbitrap (60 000 FWHM, target value 106) to monitor peptide ions in the mass range of 300–1600 m/z, followed by collision-induced dissociation MS/MS scans in the ion trap (minimum signal threshold 250, target value 104, isolation width 2 m/z) of the five most intense precursor ions. To avoid multiple scans of dominant ions, the precursor ion masses of scanned ions were dynamically excluded from MS/MS analysis for 30 s. Singly charged ions and ions with unassigned charge states were excluded from MS/MS fragmentation. RAW files were converted into the open mzXML format and searched by the SEQUEST (Eng, McCormack and Yates 1994) search engine against the UniProt/Swiss-Prot database (v57.15 of H. sapiens and L. monocytogenes) concatenated with common contaminants. Search parameters for the peptide identification included a precursor mass tolerance of 0.05 Da, a minimum of one tryptic terminus and a maximum of two internal tryptic missed cleavage sites. Cysteine carbamidomethylation (+ 57.021 Da) was set as a static amino acid modification and methionine oxidation (+15.995 Da) as well as asparagine deamidation (+ 0.9840 Da) were set as variable modifications. The PeptideProphet and ProteinProphet tools of the Trans-Proteomic Pipeline (Keller et al.2005) (TPP version 4.5) were used for the probability scoring of peptides spectrum matches.and protein inference. Protein identifications were filtered to reach an estimated false-discovery rate of ≤1%. Peptide feature intensities were extracted using the Progenesis LC-MS software (Nonlinear Dynamics). Matched peptides were filtered for the presence of the N[115]-X-S/T motif (wherein N[115] stands for a deamidated asparagine residue, X stands for any amino acid except proline and S/T for serine or threonine, respectively). Constant normalization was performed on the log-transformed intensities of peptide features across samples. Relative protein abundance and statistical significance between stimulated and control samples were assessed using the MSstats library (Clough et al.2009) in the statistical software package R.

RESULTS

LLO alters the host surfaceome

LLO pores in the plasma membrane cause ion fluxes between the cytoplasm and the extracellular space that have been associated with increased L. monocytogenes cell invasion (Dramsi and Cossart 2003; Stavru and Cossart 2011). In particular, influx of extracellular calcium may induce secretory processes, leading to changes in the surfaceome protein composition. To understand how LLO impacts L. monocytogenes cell invasion, we analyzed LLO-dependent changes of the surfaceome composition of HeLa cells using the chemoproteomic CSC enrichment strategy, targeting specifically the cell surface-exposed N-glycoproteins combined with quantitative mass spectro-metry-based proteomics (Wollscheid et al.2009; Hofmann, Bausch-Fluck and Wollscheid 2013).

The cell surface abundance of 10 glycoproteins was significantly increased and 20 glycoproteins were significantly decreased at the cell surface after treatment for 20 min with 1 nM LLO (Fig. 1A, labeled proteins in upper quadrants). Out of these 30 glycoproteins (Fig. 1B, black bars in all panels), 24 have been previously studied in a genome-wide siRNA screen for L monocytogenes infection (Kühbacher et al.2015) (Fig. 1B, left and center panels). These 24 glycoproteins can be divided in four classes: (i) proteins that inhibit infection (positive infection Z-score) that were increased on the cell surface upon LLO treatment (SLC6A15, EDEM3) (Fig. 1B, left panel), (ii) proteins that are required for infection (negative infection Z-score) that were decreased on the cell surface upon LLO treatment (IL6ST, SLC3A2, FRAS1, AXL, SC29A1, LRRC4C) (Fig. 1B, left panel), (iii) proteins that are required for infection and are increased at the cell surface (NUP210, THY1, LAMP2, HYOU1) (Fig. 1B, center panel) and (iv) proteins that inhibit infection and are decreased at the cell surface (PTGFRN, BST2, CD70, PTPRF, PTK7, BTN2A1, DCBLD1, VASN, IGSF3, ITGA5, SLC1A5, GAPDH) (Fig. 1B, center panel).

Figure 1.

Changes of the cell surface proteome after LLO treatment A, HeLa cells were treated with 1nM LLO or medium only for 20 min. Subsequently cells were detached and processed for cell surface capturing of glycoproteins followed by mass spectrometry-based proteomics analysis. The log2 fold change and adjusted P-value of differentially changed protein cell surface expression are shown. Tests for differential expression were performed using a linear mixed model and P-values were adjusted using Benjamini and Hochberg method. B, Log2 fold change of significantly enriched and depleted proteins at the cell surface (black bars, all panels) and the corresponding Z-score of infection from previously published genome-wide siRNA screens (white bars, left and center panels). Proteins are divided in four categories: (i) increased cell surface enrichment and positive Z-score; (ii) decreased cell surface exposure and negative Z-score; (iii) increased cell surface enrichment and negative Z-score and (iv) decreased cell surface exposure and positive Z-score. The right panel shows molecules which display significant changes at the cell surface distribution but which were not present in our previous siRNA screen. Dotted line: + 1 and -1 Z-score limit.

The first two classes are comprised of proteins that when combining the translocation phenotype and the infection phenotype most likely result in inhibition of infection (Fig. 1B, left panel) while the latter two classes are expected to promote L. monocytogenes infection (Fig. 1B, center panel). Therefore, LLO changes the host surfaceome in two ways: membrane associated proteins that inhibit infection like the amino acid transporter SLC1A5 are depleted in abundance from the cell surface location, while the heat shock protein HYOU1, which was previously shown to be required for L. monocytogenes infection (Kühbacher et al.2015) is enriched at the cell surface and the expected outcome in both cases is to promote L. monocytogenes invasion (Fig. 1B, center panel). Interestingly, we also observe a strong decrease in the abundance of proteins in the plasma membrane location that, according to our previous siRNA screen results, are highly important for L. monocytogenes infection (Fig. 1B, left panel). This is the case, for example for the membrane transporters SLC3A2, SLC29A1 as well as LRRC4C which has been associated with signaling by the receptor tyrosine kinase AXL. Taken together, these results indicate that LLO-dependent cell surface changes can affect positively or negatively L. monocytogenes invasion in HeLa cells.

LLO induces translocation of late endosomal compartments to the cell surface

To better understand the cellular functions leading to LLO-dependent changes of the cell surface and modulation of L. monocytogenes invasion, we performed a cellular component enrichment analysis using the gene ontology enrichment tool of the STRING database (www.string-db.org) (Szklarczyk et al.2015). This analysis showed that the 10 significantly enriched proteins at the cell surface upon LLO treatment are related to the endoplasmic reticulum (ER) (FKBP9, GLT25D1, NUP210, HYOU1, THY1), to extracellular exosomes (LGAL3BP, LAMP1, LAMP2, HYOU1, THY1) and to late endosomes/lysosomes (LAMP1, LAMP2) (Fig. 2A and 2B).

Figure 2.

Assignment of translocated proteins to cellular components. A, String network (stringency: 0.4) of significantly enriched cell surface proteins upon LLO treatment. Proteins belonging to the indicated cellular components are highlighted in red. B, Schematic representation of a human cell indicating the proteins signifcantly enriched at the cell surface upon LLO treatment, and the cellular components to which these proteins are associated.

LLO dependent translocation of the late endosomal/lysosomal membrane protein LAMP-1 was further characterized by fluorescence microscopy, and surface translocation of LAMP-1 could be readily observed as a smooth extracellular staining (Fig. 3A). LLO dependent translocation of the late endosomal/lysosomal membrane proteins LAMP-1 and LAMP-2 was also characterized by flow cytometry: a marked increase of both proteins at the cell surface was detected already as early as 5 min after LLO treatment and afterwards increased only slightly (Fig. 3B, right panel). This process was concentration dependent and up to a 6- and 2-fold increase in the surface levels of LAMP-1 and LAMP-2, respectively, could be observed on cells exposed to 1 nM LLO for 20 min (Fig. 3B, left panel). Cellular treatment with LLOw492A that binds to cells but does not form pores (Michel et al.1990) did not change the localization of LAMP-1 or LAMP-2, confirming that the LLO pore-forming activity is required for the translocation of the lysosomal proteins. In order to follow the dynamics of this translocation, EGFP-LAMP-1-expressing HeLa cells were recorded by live-cell microscopy during LLO-treatment.

Figure 3.

Translocation of late endosomal/lysosomal compartments to the cell surface in response to LLO. A, Cells grown on glass coverslips in a 24 well plate were treated for 20 min with 1 nM LLO. Subsequently, cells were fixed and immunostained for extracellular LAMP-1 and for DNA using Dapi after permeabilization Bar: 10 μm (magnification bar: 1 μm). B, HeLa cells were treated with 1 nM LLO or LLOW492A (binds to cells but is not able to form pores) for different time intervals (right), or were treated independently with different concentrations of LLO or LLOW492A for 20 min (left). Subsequently, cells were detached, fixed and stained for LAMP-1 or LAMP-2 in the absence of cell permeabilization. Fluorescence was measured by flow cytometry and values are given as fold change of the geometric mean of the fluorescent signal, compared to non-treated cells. Results represent the average of at least three independent experiments + /- standard error of the mean. C, HeLa cells transfected with EGFP-LAMP-1 were grown for 24 h in 35 mm imaging dishes. Dishes were then transferred to a 37°C pre-warmed microscope in a 5% CO2-containing atmosphere and image stacks were taken before and after addition of LLO to a final concentration of 1 nM. The boxed region in the upper panels is enlarged in the lower panels. Bar: 1 μm. D, HeLa cells were pretreated for 20 min with 20mM HEPES or 10 mM EGTA and 20mM HEPES and subsequently left untreated or treated for 20 min with 1 nM LLO in the presence of the different pretreatments. Following this, cells were detached, fixed and fluorescently stained for surface accessible LAMP-1 in the absence of permeabilization. The fluorescent signal was measured by flow cytometry. Geometric mean of fluorescence intensity was normalized to the average of all samples. Bars represent the average and geometric mean of three independent experiments (Student's t-test: *, p < 0.05). E, HeLa cells were treated with 1 nM LLO for 20 min and the β-hexosaminidase activity in the cellular supernatant (and cell lysates) was measured fluorometrically. Values represent the fold change of the ratio of β-hexosaminidase activity of supernatants/cell lysates. Results show the average of three independent experiments relative to the mock control + /- SEM. F, Exosomes were purified from the supernatant of control and LLO-treated cells by ultracentrifugation, seeded on glass coverslips, fixed, stained for CD63 and quantified by microscopy. Bars represent the mean and standard deviation of six visual fields of a representative experiment.

LLO promoted a major change not only in the dynamic movement of intracellular LAMP-1 positive vesicles but also in the distribution of EGFP-LAMP-1: before LLO treatment, EGFP-LAMP-1 is detectable as bright mobile spots (Fig. 3C); upon treatment, some of these structures disappear (inset, Fig. 3C) and remaining spots become immobile. Together these results suggest that LAMP-1 positive intracellular compartments such as late endosomes or lysosomes are very rapidly translocated to the cell surface upon LLO treatment and that this process is dependent on the pore-forming activity of the toxin. Treatment of cells with EGTA reduced the appearance of EGFP-LAMP1 at the cell surface upon LLO stimulation (Fig. 3D), confirming that Ca2+ influx is required for this process. Interestingly, we observed that the L. monocytogenes invasion protein InlB synergizes with LLO and also promotes translocation of LAMP-1 and LAMP-2 to the cell surface, although to a much lower extent in comparison to LLO (Supplementary Fig. 1).

If late endocytic compartments fuse with the plasma membrane, luminal lysosomal enzymes are expected to be secreted to the extracellular space. To determine whether L. monocytogenes infection promotes the fusion of late endosomes or lysosomes with the plasma membrane, we measured the presence of the luminal lysosomal enzyme β-hexosaminidase in the extracellular space upon cell treatment with LLO. As shown in Fig. 3E, hexosaminidase activity was detected in cell culture medium of cells treated with LLO. Multivesicular late endosomes can fuse with the plasma membrane and release their intraluminal vesicles as structures named exosomes (Stoorvogel et al.2002). We applied an ultracentrifugation-based protocol to purify exosomes and examined preparations from culture supernatants of control cells versus cells treated with LLO: the exosome preparation from control cells already contains some vesicles with a diameter between 80–150 nm but LAMP-1 positive vesicles/exosomes were only occasionally detected (Supplementary Fig. 2A, I); by contrast, the preparation from LLO-treated cells presented more frequently exosomes displaying a LAMP-1 labeling (Supplementary Fig. 2A, II inset). Isolation and quantification of these compartments by fluorescence microscopy using antibodies against the exosomal marker CD63 confirmed that exosome secretion is highly increased upon cellular exposure to LLO (Fig. 3F). By western blot we also detected higher amounts of CD63 in exosomal fractions isolated from LLO-treated cells (Supplementary Fig. 2B). In conclusion, we determined that LLO treatment leads to the secretion of intracellular compartments of late endosomal nature.

Listeria monocytogenes elicits secretion of late endosomal compartments at invasion foci

During L. monocytogenes invasion, the LLO concentration is presumably highest close to the bacterium where the toxin is released. We therefore tested if late endosomes fuse with the host cell plasma membrane at bacterial invasion foci. As shown in Fig. 4A, extracellular LAMP-1 is enriched in the vicinity of bacteria during invasion. Specific fluorescent labeling of extracellular bacteria revealed moreover that endosomal exocytosis at the entry site occurs early during invasion before vacuole closure (Fig. 4B). By spinning disk confocal microscopy, we could follow in living cells the association of EGFP-LAMP-1-positive vesicles with L. monocytogenes during the initial steps of cellular invasion (Fig. 4C). As for purified LLO, L. monocytogenes induced LAMP-1 translocation to the cell surface was blocked by Ca2+ depletion (Fig. 4D).

Figure 4.

Late endosomes/lysosomes fuse with the plasma membrane at the L. monocytogenes entry site during invasion. A, HeLa cells grown on coverslips were infected with L. monocytogenes EGDe.PrfA* for 20 min. To synchronize infection cells were centrifuged for 5 min at room temperature after inoculation. Cells were then fixed, stained for extracellular LAMP-1 and then permeabilized and stained for DNA allowing to visualize total bacteria using Dapi. The arrow points towards a group of LAMP-1 positive bacteria Bar: 10 μm (magnification bar: 1 μm). B, Cells were infected as in A and stained for extracellular LAMP-1 and extracellular bacteria without permeabilization. Subsequently cells were permeabilized and stained for total DNA. The arrow points towards a LAMP-1-positive extracellular L. monocytogenes in the process of cell invasion. Bar: 10 μm (magnification bar: 1 μm). C, HeLa cells grown on 35 mm imaging dishes and transfected with EGFP-LAMP-1 were infected with L. monocytogenes EGDe.PrfA* and centrifuged for 5 min at room temperature to synchronize infection before image acquisition in a spinning disk microscope at 37°C. Upper panel: phase contrast channel; lower panel: GFP channel. Arrows point to LAMP-1-positive vesicles that associate with- and seemingly fuse with the L. monocytogenes-containing vacuole during the course of bacterial invasion. Bar: 1 μm. D, HeLa cells were pretreated for 20 min with 20mM HEPES or 10 mM EGTA and 20mM HEPES and left uninfected or infected for 20 min with L. monocytogenes EGDe.PrfA* in the presence of the different treatments. Following this, cells were detached, fixed and fluorescently stained for surface accessible LAMP-1 in the absence of permeabilization. The fluorescent signal was measured by flow cytometry. Geometric mean of fluorescence intensity was normalized to the average of all samples. Bars represent the mean and SEM of three independent experiments.

Applying transmission electron microscopy on L. monocytogenes infected cells to investigate the morphological features of the late endosomal compartments that fuse with the plasma membrane upon infection, we detected bacteria within membrane-bound vacuoles that contained small vesicles of 50 to 100 nm diameter (Supplementary Fig. 3A), which correlates with the size of the intraluminal vesicles found in late endosomes (Van Niel et al.2001), indicating an interaction or fusion with these compartments at bacterial entry sites. These results prompted us to investigate whether the unconventional lipid lysobisphosphatidic acid (LBPA), which is particularly detected in intraluminal membranes of late endosomal compartments, could also be found at L. monocytogenes entry sites, and by immunofluorescence microscopy LBPA was detected in a subset of bacterial entry foci together with LAMP-1 (Supplementary Fig. 3B). Altogether, these observations suggest that L. monocytogenes promotes local exocytosis of late endosomal compartments during cellular invasion before bacterial engulfment is completed.

LLO-dependent secretion of late endosomal compartments at L. monocytogenes invasion foci participates in bacterial invasion

To investigate the functional relevance of late endosomal secretion during L. monocytogenes cell entry, we first internalized soluble horseradish peroxidase (HRP) in cells for 1.5 h and then performed a chase for 30 min to accumulate HRP in late endosomal compartments (Futter et al.1996). We then perturbed the organization of late endosomes by crosslinking the HRP with 3,3΄-diaminobenzidine (DAB): in these conditions, we detected a clear reduction in the surface LAMP-1 levels after LLO treatment (Fig. 5A) or L. monocytogenes infection of HeLa cells, (Fig. 5B). We also applied this late endosomal cross-linking procedure and performed invasion assays quantified by differential staining of total and extracellular bacteria: as shown in Fig. 5C, HRP/DAB treatment inhibits cellular invasion by L. monocytogenes, suggesting that the delivery of these organelles to the plasma membrane, and in particular to bacterial entry sites, facilitates bacterial internalization.

Figure 5.

Perturbation of late endosomal/lysosomal compartments blocks LAMP-1 translocation and reduces L. monocytogenes invasion. HeLa cells were let untreated or were incubated with 5 mg/ml HRP for 1.5 h followed by 30 min of chase in the absence of the enzyme and 30 min of lysosomal/late endosomal cross-linking by DAB/H2O2 exposure on ice. Cells were then treated with mock DMEM or treated with 1 nM LLO in DMEM for 20 min (A) or infected for 20 min with L. monocytogenes EGDe.PrfA* (B). After fixation cells were stained for LAMP-1 and analyzed by flow cytometry. Geometric mean of fluorescence intensity was normalized to the average of all samples. Mean and SEM of three independent experiments are presented (Student's t-test: *, p < 0.05). C, Cells treated with HRP and DAB/H2O2 (as in A and B) were infected with L. monocytogenes EGDe.PrfA* for 40 min and extracellular bacteria were stained differentially (A). Results represent the average intracellular bacteria (% of total) + /- standard error of the mean (Student's t-test: ***, p < 0.001).

DISCUSSION

In this work, we assessed host surfaceome protein abundance changes taking place upon cellular exposure to extracellular LLO, and we identify cellular components which can differentially modulate L. monocytogenes internalization. In particular, we show that LLO induces the translocation of late endosomal/lysosomal compartments to the bacterial entry foci and that this late endosomal mobilization promotes efficient L. monocytogenes invasion of host cells.

We observed a significant downregulation in the abundance of a subset of proteins in the cell surface location, as well as the translocation of proteins from intracellular compartments to the cell surface. Cross-correlation of these results with those of a previous genome-wide siRNA screen to investigate L. monocytogenes infection led to the discovery of proteins like the amino acid transporters SLC3A2 and SLC29A1, the receptor tyrosine kinase AXL as well as the leucine rich repeat protein LRRC4C, which in our previous siRNA screen were shown to be required for bacterial infection and in our present study are efficiently internalized upon LLO treatment. It is tempting to speculate that during host colonization, internalization of infection-promoting factors might give early invading bacteria an advantage to secure their niche, a potentially important concept that should be evaluated in the context of competing pathogens under physiological conditions. Other molecules like the Na+-dependent amino acid transporter SLC6A15 or the endoplasmic reticulum degradation enhancer EDEM3, which in our previous siRNA screen were shown to inhibit infection, are translocated to the cell surface location and increased in abundance upon LLO treatment in our present study. These molecules are therefore associated to novel signaling cascades that block bacterial entry into host cells.

We were particularly interested in proteins which were translocated to the surface upon LLO cell exposure and which contribute positively to the L. monocytogenes infection process. Among these, we identified a cluster of ten proteins which are functionally associated with the endoplasmic reticulum, late endosomes/lysosomes and exosomes. We validated our CSC results by monitoring the translocation of the protein LAMP-1 to the surfaceome using flow cytometry, fixed-cell fluorescence microscopy and live-cell spinning disk confocal microscopy. Our results confirmed that LLO induces the cell surface translocation of LAMP-1 in a concentration-, time- and Ca2+-dependent manner, which correlates with the concomitant export of exosomes and the release of lysosomal contents in the extracellular space. Intact L. monocytogenes also produce this LAMP-1 translocation to bacterial entry foci in a Ca2+-dependent manner, and by perturbing the function of late endosomes/lysosomes we do not only show that these compartments are the donor pools of LAMP-1, we also demonstrate that perturbation of this translocation inhibits L. monocytogenes entry.

Targeting of lysosomal compartments to the plasma membrane of fibroblasts and epithelial cells has been reported as a Ca2+-dependent repair mechanism triggered upon mechanical wounding aggressions, including membrane disruption by the pore-forming toxin streptolysin O (Rodríguez et al.1997; Reddy, Caler and ANDREWS 2001), a process dependent on Synaptotagmin VII (Rao 2004; Czibener et al.2006). Membrane injury by Salmonella enterica, a bacterial pathogen that harbors a type III secretion system that induces Ca2+-influx, activates this repair mechanism which limits bacterial intracellular growth (Roy et al.2004). Interaction of lysosomes with invading pathogens has also been described in the case of the parasites Trypanosoma cruzi (Tardieux et al.1992; Kima, Burleigh and Andrews 2000; Fernandes et al.2011) and Leishmania donovani (Forestier et al.2011); in the particular case of T. cruzi, host cell membrane wounding by the parasite flagellum induces a Ca2+ influx which mobilizes LAMP-1 positive lysosomes to the plasma membrane: secretion of an acid sphingomyelinase from lysosomes to the outer leaflet of the plasma membrane induces the production of ceramide to promote membrane repair, and T. cruzi takes advantage of ceramide-rich microdomains to favor the formation of its parasitophorous vacuole (Fernandes et al.2011). The results we report in this manuscript are highly reminiscent of the lysosome-based membrane repair mechanisms described in previous works and represent, to our knowledge, the first description of late endosomal/lysosomal compartments exocytosis favoring the entry of a bacterial pathogen.

As mentioned above, LLO has been previously reported to promote Ca2+ entry within epithelial cells and this influx was shown to favor bacterial entry, but the molecular mechanism(s) regulated by this ion during L. monocytogenes invasion were not characterized at the time (Dramsi and Cossart 2003). LLO-induced Ca2+ influx has also been shown to induce fragmentation of mitochondria and it was proposed to indirectly affect bacterial infection by modulating mitochondrial function (Stavru et al.2011). More recently, it has been demonstrated that LLO promotes entry by a dynamin- and F-actin-dependent but InlA/InlB- and clathrin-independent pathway that induces membrane extensions and particle engulfment (Vadia et al.2011; Vadia and Seveau 2014). Our results are in agreement with these previous observations and indicate therefore that late endosomes/lysosomes participate to this LLO-dependent entry process, and suggest that LLO synergizes with the InlB-dependent pathway to promote efficient plasma membrane remodeling during the early L. monocytogenes invasion steps. Along this line, it is interesting to note that we also identified an increase in the plasma membrane levels of LAMP-1 upon cell exposure to purified InlB, which correlate with a recently described role for host cell exocytosis in InlB-mediated internalization (Van Ngo et al.2017), confirming therefore a synergy between the InlB- and the LLO-dependent internalization pathway.

In this context, it is important to mention that LAMP-1 surface translocation can also occur through an unconventional endo-secretory pathway that implicates a population of early endocytic compartments and requires in part the function of the clathrin adaptor AP-1 (Laulagnier et al.2011). Remarkably, AP-1 fosters optimal phagocytosis in macrophages (Braun et al.2004; 2007), participates in the polarized exocytosis of LAMP-2-positive lysosomes in epithelial cells during membrane repair (Xu et al.2012), and has been also implicated in L. monocytogenes internalization through an InlB-dependent invasion pathway (Pizarro-Cerdá et al.2007). The investigation of the molecular intersection between these signaling cascades during L. monocytogenes infection represents a promising research avenue to explore in the future.

AUTHOR CONTRIBUTIONS

A. Kühbacher, K. Novel, B. Wollscheid and J. Pizarro-Cerdá designed research; A. Kühbacher, K. Novel, J.J. Quereda, M. Sachse, M. Moya-Nilges and J. Pizarro-Cerdá performed experiments; A. Kühbacher, K. Novel, B. Wollscheid and J. Pizarro-Cerdá analyzed the data; and A. Kühbacher, P. Cossart and J. Pizarro-Cerdá wrote the paper.

Click here for additional data file.