Abstract
The gram-positive bacillus Listeria monocytogenes gains entry into host cells through a phagosome membrane that forms around entering bacteria. During the early stages of internalization the invading bacteria appear to modify the protein composition of the forming phagosome membrane in J774 cells. MHC class II molecules on the cell surface and exposed surface molecules available for biotinylation are excluded from the bacteria-host cell membrane interface and from the forming phagosome. This exclusion of MHC class II molecules from the early phagosome may partially help to explain previous reports suggesting that L. monocytogenes is able to interfere with antigen presentation. Inside the host cell, MHC class II molecules are delivered to the phagosome membrane. This is followed by delivery of LAMP 1, a marker of late endocytic compartments, and fusion with low-pH compartments. The bacteria then escape into the cell cytoplasm, possibly assisted by rapid delivery of this low-pH environment.
I
The gram-positive bacterium Listeria monocytogenes enters and is able to survive in mammalian cells. This organism is a contaminant of the food we eat (Cossart and Mengaud 1989), and although it is a common pathogen in farm animals L. monocytogenes does not normally infect humans (Kaufman 1988). The bacteria, taken into membrane-bound phagosomes, use a thiol-activated cytolysin listeriolysin O (LLO) to disrupt the membrane (Cossart et al. 1989; Bielecki et al. 1990) and gain access to the cell cytoplasm (Mackaness 1962; Tilney and Portnoy 1989; Kaufman 1993). Subsequent actin polymerizations on the bacterial surface induce motility in the cell and form long cytoplasmic protrusions at the cell surface that spread bacteria to neighboring cells (Tilney and Portnoy 1989).
Experimental listeriosis in mice is a widely studied model for intracellular bacterial infections. With this model, it has been shown that L. monocytogenes is able to generate responses from both the MHC class I and class II pathways (Ziegler and Unanue 1981; Pamer et al. 1991; Harding and Geuze 1992). Entry into host cell cytoplasm brings the bacteria into contact with the MHC class I pathway, indicating that only viable bacteria are involved. To effect MHC class II presentation, the bacteria must enter phagolysosomal compartments involved in protein degradation (Harding and Geuze 1992). In experimental studies, this involves internalization of heat-killed or LLO mutant bacteria by host cells (Harding and Geuze 1992). Presentation of processed bacterial antigens by class II MHC molecules on the cell surface occurs within 20 min after uptake by peritoneal macrophages (Harding and Geuze 1992). The rapid processing that occurs within these cells is partially a result of endosome and lysosome fusion with phagosomes (Lang et al. 1988; Hart and Young 1991; Mayorga et al. 1991; Harding and Geuze 1992; Pitt et al. 1992b; Rabinowitz et al. 1992). The contents of these compartments, which include cathepsin B and MHC class II (Harding and Geuze 1992), are delivered to phagosomes.
In contrast, viable L. monocytogenes exhibit a fine control over phagosome maturation. Newly formed phagosomes accumulate rab 5 and N-ethylmaleimide (NEM)-sensitive factor (NSF), which causes upregulation of fusion with early endosomes (Alvarez-Dominguez et al. 1996). Subsequent phagosome maturation is inhibited, possibly by the modification of the phagosome membrane either by removal or addition of fusogenic agents (Alvarez-Dominguez et al. 1997; Collins et al. 1997) or by an rab5a-dependent mechanism (Alvarez-Dominguez and Stahl 1999). This control over phagosome formation may help to explain how the bacterium can inhibit antigen processing and presentation by macrophages (Cluff et al. 1990).
Although events occurring during the early stages of L. monocytogenes internalization are better understood than for most pathogenic bacteria, many questions remain unanswered. The interactions between viable bacteria and host cell plasma membranes have not yet been fully investigated. Similarly, intracellular events during the initial internalization of bacteria are only currently being elucidated. For this reason, a preliminary immunocytochemical study of the early phagosome formation process is reported here. In this study the processes of phagosome maturation are visualized during the critical early stages of infection when bacteria first associate with and enter host cells.
Materials and Methods
Cells
All tissue culture reagents were obtained from Gibco Laboratories (Grand Island, NY) and were sterilized either by autoclave or filtration. The macrophage-like J774A.1 cultured cell line (ATCC TIB-67) was grown in Dulbecco's minimal essential medium (DMEM) containing 5% fetal calf serum (FCS) at 37C in a 5% CO2 atmosphere. The cells were maintained as monolayers in 75-cm2 flasks and passaged twice a week by washing once in PBS, scraping them into serum-free DMEM, and then diluting the resuspended cells in fresh DMEM with 5% FCS.
Bacteria
The wild-type L. monocytogenes strain (10403S), which belongs to serotype 1, was a gift from Dr. D. A. Portnoy and was previously described (Jones and Portnoy 1994). Bacteria, taken from overnight standing cultures were incubated aerobically in BHI broth at 37C until they reached mid-log phase of growth (2 hr). Spectrophotometer readings at 660 nm, initially calibrated against hemocytometer counts, were used to estimate bacterial numbers. The bacteria were pelleted, washed in DMEM, resuspended in fresh DMEM containing 10% FCS, and added to J774 cell cultures.
Western Blotting
SDS-PAGE separation of J774 cell lysates was performed as described previously (Laemmli 1970). Separated proteins were blotted onto Immobilon-P, blocked overnight at 4C with 10 mM Tris buffer (pH 7.4) containing 5% skim milk and 15 mM sodium chloride. The blots were washed and then incubated sequentially with specific antibodies to MHC class II antigens and secondary goat antibodies conjugated to alkaline phosphatase (Promega; Madison, WI). The antibodies were visualized by a colorimetric alkaline phosphatase reaction.
Invasion Assays
J774 cells were seeded into sterile 9-mm culture dishes at a dilution of 1 × 104 per ml and incubated for 2 days at 37C before addition of L. monocytogenes suspensions. Electron microscopy was performed using these cultures. For light microscopy, J774 cells were grown on sterile glass coverslips at the bottom of culture dishes.
After 2 days of incubation at 37C, the J774 cells were counted (approximately 1 × 105 per dish) and infected with 20 mid-log phase bacteria (OD 0.6 at 660 nm) bacteria per cell (2 × 106). The cells and bacteria were left at 37C for 5 min, washed with DMEM, and then either fixed immediately or incubated in DMEM with 10% FCS for increasing time intervals. They were then fixed in buffered aldehyde and prepared for examination by either light or electron microscopy. Infection times were recorded from the time the bacteria were first added to the cells as time zero.
For some control experiments, cells were incubated for 10 min with 1-μm diameter latex beads coated with avidin (Jahraus et al. 1998). Phagocytosis of these beads was stopped by immersing the cells in buffered aldehyde. Cryosections through them were examined immunocytochemically.
Surface Biotinylation and Incubation with DAMP
Cells to be biotinylated were cooled on ice, washed with PBS, and then exposed to 2 mg/ml sulfo-NHS-biotin (Pierce; Rockford, IL) in PBS for 5 min on ice. They were then washed with cold PBS and then with cold DMEM containing 10% serum. Cold DMEM containing L. monocytogenes was then added to the cells, left for 5 min, and either fixed immediately on ice or warmed to 37C and incubated for increasing times before being fixed.
For some experiments, the J774 cells were incubated in DMEM containing 10% FCS and 20 μM 3-(2,4-dinitroanilino)-3′-amino-N-methyldipropylamine (DAMP) at 37C (Anderson et al. 1984). After 30 min the cells were washed to remove free DAMP and then infected with L. monocytogenes as described above.
Antibodies
Intracellular organelles of infected J774 cells were probed on thawed thin cryosections or on permeabilized whole-cell preparations with specific antibodies to LAMP 1 (Developmental Studies Hybridoma Bank; Iowa City, IA), biotin (Jackson Immunoresearch; West Grove, PA), and mouse Ia MHC class II (Amigorena et al. 1994). The accumulation of DAMP inside cells was detected using specific polyclonal antibodies to dinitrophenol (DNP), a gift from I. Mellman. Bacteria binding to the cell surface were labeled with L. monocytogenes-specific polyclonal antiserum (Difco Laboratories; Detroit, MI). Antibodies were visualized by TEM using protein A-gold (PAG: produced by and purchased from J. W. Slot, Utrecht University, The Netherlands). The surfaces of J774 cells were probed by FACS analysis using FITC-conjugated monoclonal antibodies (PharMingen; San Diego, CA) to examine the distribution of I-Ed (antibody number 14-4-45) and I-Ad (number AMS-32.1) within the total cell population. Unconjugated secondary antibodies, used as a bridge between monoclonals and PAG, and secondary antibodies conjugated with fluorescent dyes were purchased from Cappel (Cochranville, PA).
BSA-Gold Incubation
Colloidal gold of 15-nm particle size, prepared as previously described (Slot and Geuze 1981,1985), was coupled to bovine serum albumin (BSA) by adding 2.4 mg of BSA dissolved in water per 1 ml of gold sol at pH 6.0. The gold probe was concentrated by centrifugation, dialyzed against DMEM, then suspended in DMEM, containing 5% FCS to a final concentration with an OD of 4 at 525 nm. This was then added to J774 cultures, incubated for 30 min at 37C, removed, and internalized gold chased into lysosomes by incubating at 37C for 3 hr or 24 hr.
Light and Electron Microscopy
For light microscopy, cells grown on glass coverslips were fixed in 4% phosphate-buffered formaldehyde, permeabilized with 0.01% Triton X-100 in PBS, and labeled with specific and fluorescent secondary antibodies. After labeling, the coverslips were examined by epifluorescent illumination with an Axiophot light microscope (Zeiss).
For immunocytochemistry, some cell monolayers were fixed for 1 hr in 0.5% glutaraldehyde buffered in 100 mM sodium phosphate (pH 7.4). Other cells were fixed in 4% phosphate buffered-formaldehyde for 1 hr and then left overnight in 8% formaldehyde in the same buffer. After fixation, the cells were prepared for cryosectioning and immunocytochemistry using previously described methods (Webster 1999). Briefly, the sections were retrieved using sucrose droplets, mounted on coated specimen grids, labeled with primary antibodies, unconjugated bridging antibodies (when appropriate), PAG, and finally contrasted by incubating and drying in the presence of aqueous methyl cellulose and uranyl acetate (final concentration of 0.3% uranyl acetate). When anti-biotin antibodies were used, PBS containing 1% (v/v) cold water fish skin gelatin was used as a blocking agent. Control experiments consisted of treating sections with PAG alone, with non-relevant antibodies and PAG, and with rabbit anti-mouse antibodies and PAG.
For routine morphological examination, cell monolayers were fixed in 2.5% glutaraldehyde buffered in 100 mM sodium cacodylate (pH 7.4), scraped from the culture dish as described, pelleted, postfixed in osmium tetroxide, en blocstained in 1% uranyl acetate in 50 mM sodium maleate (pH 5.2), dehydrated in ethanol, and embedded in epoxy resin. Sections were examined and photographed using a transmission electron microscope (FEI-Philips; Hillsboro, OR) operating at 80 kV.
Quantification
Co-localizations of L. monocytogenes with intracellular markers was examined after labeling with specific antibodies and fluorescent secondary antibodies. Samples were examined by epifluorescent illumination. Approximately 200–300 cells were sequentially sampled for each experiment.
Stereological estimates of anti-biotin labeling density over J774 cell membranes were performed on sequentially sampled electron micrographs using cross-lattice overlays. Gold particle numbers were combined with numbers of intersects between the membrane under investigation and a test line of known length (Griffiths 1993).
Results
Viability of L. monocytogenes
To assess the viability of the L. monocytogenes used in this work, growth and infection rates were examined. Although all incubation times used in this study were never longer than 20 min post infection (p.i.), longer incubation times were used to examine the normal internalization process of the bacteria being used. After 5-min exposure to bacteria, approximately 15% of J774 cells were observed with associated bacteria, with an average of 1 bacterium per cell (Table 1). The numbers of cells with associated bacteria increased with time (Table 1), with 35% of cells being affected after 2 hr and 74% of cells after 3 hr. The number of bacteria per cell also increased during this time, with average numbers increasing slowly for the first 2 hr but increasing to 10 per cell after 3 hr. Labeling of non-permeabilized cells showed only small numbers of extracellular bacteria associated with the J774 cells (data not shown).
The numbers of actin tails associated with internalized L. monocytogenes, visualized with fluorescent phalloidin, were first detected 2 hr p.i. The mean number of actin tails was slightly less than the number of bacteria per cell but increased over time (Table 1).
Infection and internalization rates of Listeria monocytogenes in J774 macrophage-like cells, estimated by counting fluorescently labeled cells and bacteria
Adding L. monocytogenes to J774 cell cultures had no obvious effect on the subcellular morphology of the cells. Intracellular bacteria were easily detected by light and electron microscopy and sufficient bacteria were found by electron microscopy to give unbiased overviews of early intracellular events.
Nonspecific Antibody Binding
Although monospecific antibody probes were used for labeling sections, controls to test for nonspecific binding were performed. Sections of L. monocytogenes-infected J774 cells were probed with non-relevant antibodies and PAG, PAG alone, or rabbit anti-mouse antibodies followed by PAG. No signal was detected with either the non-relevant antibodies or PAG (data not shown), and only small amounts of gold were detected on sections treated with anti-mouse antibodies (Figure 1A and 1B). Small numbers of gold particles were seen over the nuclear envelope, throughout the cytoplasm, over the nucleus, and occasionally inside bacteria (data not shown). There was no signal over host cell membrane or over phagosome membranes (data not shown).
Interaction of L. monocytogenes with Host Cell Surface
In the first few minutes of infection, L. monocytogenes bound to the host cell surface with a tight apposition of bacterial and host cell membranes (Figure 1A). This close interaction of host and bacterial membranes continued into early phagosomes (Figure 1B). No specific accumulation of DAMP-containing structures (Figure 1C) or BSA-gold-filled lysosomes (Figure 1D) was detected in the intracellular region close to extracellular surface-bound bacteria.
In cryosections, the MHC class II labeling of J774 cells could be easily detected on the cell surface and in intracellular structures. At the site where L. monocytogenes were attached to the plasma membrane of cells expressing surface MHC class II, an exclusion of labeling from the J774 cell surface was observed (Figure 2A). This suggested that L. monocytogenes was able to exclude some surface proteins from the plasma membrane during the binding phase before internalization. To confirm this, J774 cells were surface-biotinylated and then infected with L. monocytogenes. Biotinylation on the cell membrane can be detected by applying specific anti-biotin antibodies and PAG. In sections, this label was present on the whole length of the surface membrane. However, at the site where L. monocytogenes attached to the plasma membrane an exclusion of the specific biotin labeling from the interface between bacterium and cell surface was observed (Figure 2B and 2C). The biotin signal was excluded when there was a tight apposition of bacterial and host cell membranes (Figure 2B) and when the association appeared to be a loose one (Figure 2C). Quantification of the labeling density of anti-biotin antibody at this site showed a significant reduction in the signal compared to plasma membrane where L. monocytogenes was not attached (Table 2).
It is possible that the absence of labeling between the L. monocytogenes and the host cell was because the antigens at this site were inaccessible for antibody binding. For this reason, sections were probed with antibodies that recognized bacterial surface antigens. These antibodies did label the interface between bacteria and host cell membrane (Figure 2D). To confirm that antigens at this site were accessible for antibody binding, bacteria were biotinylated, bound to cells, sectioned, and probed with anti-biotin antibodies. The anti-biotin antibodies bound to the zone of contact between the biotinylated L. monocytogenes and the J774 cell membrane (Figure 2E). Both experiments clearly demonstrate that the site of bacterial and host cell membrane apposition was easily accessible to antibodies.
To further investigate the plasma membrane site where particles bind as a first step to internalization, J774 cells were incubated for 10 min with avidin-coated latex beads. Sections through aldehyde-fixed cells with attached latex beads were probed with anti-MHC class II antibody. Antibody binding was detected at the interface where the plasma membrane and latex bead surface were interacting (Figure 3). The method of internalization appeared to be very different from that occurring during L. monocytogenes internalization. For each bead examined associating with a cell membrane, there was a close apposition of membrane to the bead surface, with the cell membrane wrapping completely around the latex bead (Figure 3). In contrast, L. monocytogenes did not appear to stimulate the long phagocytic processes on the J774 cell membrane that engulfed latex beads.
Expression of MHC Class II by J774 Cells
The J774 cell line is not usually considered to be an MHC class II-expressing cell line. However, class II-specific antibodies produced specific labeling patterns on cryosections of J774 cells. Light and electron microscopic quantification of the cells showed that approximately 70% of the cells in the population expressed MHC class II (data not shown).
To confirm the expression of MHC class II by the J774 cells, Western blots of homogenized cells were probed with antibodies to class II MHC antigens. A polypeptide pattern consistent with MHC class II expression in J774 cells was recorded (Figure 4). This pattern was similar to that seen in Western blots of A20 β-lymphocytes, a known MHC class II-expressing cell line, probed with the same antibodies (Figure 4).
Sections through J774 cells exposed to viable L. monocytogenes (LM) for 5 min. (
Micrographs of cryosections through immunolabeled J774 cells exposed to L. monocytogenes (LM) for 5 min. (
Labeling density of anti-biotin antibodies on the membranes of biotinylated J774 cells infected with L. monocytogenes
Cells probed with antibodies to I-Ed epitopes specific to MHC class II molecules from Balb/c mice (the original source of the J774 cell line) and subjected to a fluorescently activated cell sorter (FACS) analysis demonstrated a positive signal on approximately 65% of the cells (Figure 5). In contrast, a less specific epitope, I-Ad, was seen on only 5% of the cells (Figure 5).
The Early Phagosome and Intracellular Organelles
After incubation of cells and bacteria for 5 min at 37C, bacteria were found inside the cells in membrane-bound vacuoles. The membrane of early phagosomes showed reduced labeling with antibodies to class II MHC antigens (Figure 6A) and with anti-biotin antibodies in cells that had been surface-biotinylated (Figure 6B). With the biotin labeling, the mean number of gold particles associated with phagosome membranes was similar to the depleted signal estimated for the plasma membrane at the region of contact with L. monocytogenes (Table 2). Some phagosomes did not label with the anti-MHC class II antibodies (data not shown).
After 10-min exposure to bacteria, some L. monocytogenes-containing phagosomes still did not label with the anti-MHC class II antibodies. Structures in close proximity to the early L. monocytogenes-containing phagosomes did show MHC class II labeling (Figure 7A). In other cells, it appeared that MHC class II-positive structures were fusing with the phagosome membrane (Figure 7B). Although many of the internalized bacteria had escaped into the cytoplasm at the 25-min time point (data not shown), phagosomes still containing bacteria were present and showed increased labeling with the MHC class II antibodies (Figure 7C). In multiple-label experiments the phagosome membranes showed variable labeling with antibodies to LAMP 1 and MHC class II. Although many phagosomes that labeled with anti-class II antibodies had little or no LAMP 1 label (Figure 7D and 7E), some were present that did label with this marker (Figure 7E). In some cells, LAMP 1-positive structures were seen close to L. monocytogenes-containing phagosomes (Figure 8).
The variable association of LAMP 1 and MHC class II labeling with L. monocytogenes-containing phagosomes observed by TEM was quantified by light microscopy. Co-localization of bacteria with either LAMP 1 or MHC class II antibodies was quantified on cells fixed at regular intervals after exposure to bacteria. At the first time point, when bacteria were first added to the cells, no co-localization of bacteria with MHC class II or LAMP 1 was observed (data not shown). However, after the 5-min p.i. time point, approximately 50% of the cell associated bacteria co-localized with MHC class II (Figure 8). After 10 min the number of bacteria co-localized with MHC class II had risen to 70%. This number dropped to just over 20% at 15 min p.i. and steadily dropped, until by 2 hr p.i. there was only approximately 10% co-localization (Figure 9).
In comparison, co-localization with LAMP 1 was first observed 15 min p.i. with ∼10% of the bacteria being associated with LAMP 1. This rose to 20% by 25 min and remained at this value until it rose to 50% at 3 hr p.i. (Figure 9).
Anti-MHC Class II Antibodies Do Not Bind to Surface Fc Receptors
To confirm that MHC class II labeling on cryosections was not an artifact due to crossreactivity with the Fc receptor, antibodies to mouse Fc receptor (24G2) were used as a blocking step before labeling with anti-class II antibodies. Sequential sections through cells infected for 5 min either were incubated with anti-MHC class II alone (Figure 10A) or were treated with 24G2 before labeling with anti-MHC class II (Figure 10B and 10C). As expected, early phagosomes in untreated sections labeled with MHC class II (Figure 10A). Pretreatment of sections with 24G2 did not inhibit MHC class II labeling on the cell surface (Figure 10B) or on the L. monocytogenes-containing phagosome membrane (Figure 10C), suggesting that the MHC class II antibodies do not bind to Fc receptors in J774 cells. Similarly, sections labeled with the anti-Fc receptor antibody showed only a low labeling signal on the surface and over intracellular sites in J774 cells, which were not associated with intracellular L. monocytogenes (data not shown), also suggesting that the MHC class II antibody labeling was specific.
Micrograph of a cryosection through J774 cells incubated for 10 min with avidin-coated latex beads. This section, labeled with anti-MHC class II antibody and PAG, shows specific binding of collodial gold on the plasma membrane of the J774 cell. This label is not excluded from the region where the latex bead has bound to the cell membrane (arrowheads). The membrane attaching to the bead can be seen wrapping around the entire circumference of the bead profile (arrow). Bar = 0.5 μm.
Western blotting analysis of A20 β-lymphocytes, a known MHC class II-expressing cell line, and J774 cells. The blot, probed with rabbit anti-mouse Ia MHC class II antibodies, shows a typical labeling pattern for MHC class II present over the A20 cells with an almost identical pattern over the J774 cells. Both α and β side chains are present in both cell types.
Interaction of Phagosomes with Low-pH Compartments
To examine the process of acidification of L. monocytogenes-containing compartments, cells were incubated with DAMP before infection. Sections through aldehyde-fixed cells were probed with anti-DNP antibodies and PAG to identify intracellular sites of DAMP accumulation. Many membrane-bound organelles labeled with the anti-DNP antibodies, indicating sites of low pH. The cell nucleus (data not shown) and mitochondria labeled with anti-DNP antibodies, indicating an accumulation of DAMP (Figure 11A). Many other organelles showed accumulation of DAMP, some of which were observed fusing with L. monocytogenes-containing phagosomes (Figure 11A). The anti-DNP signal was found accumulating around the bacteria within the lumen of phagosomes (Figure 11B).
Discussion
Exposure of J774 cells to a large inoculum of bacteria for short incubation times has made it possible to examine early synchronous events of L. monocytogenes internalization using immunocytochemical methods. In this way it has been possible to demonstrate that, during the early stages of L. monocytogenes association with J774 cells, the initial binding to host plasma membrane is accompanied by a rearrangement of some surface proteins on the host cell surface. This results in the formation of an early phagosome membrane inside the cell with a peptide profile different from that of the plasma membrane. Surface molecules available for biotinylation, and MHC class II molecules at the site of attachment, are excluded from the phagosome membrane. The exclusion of biotin signal and MHC class II molecules from between the bacterium and the host cell membrane suggests that manipulation of the cell surface by the bacteria is occurring either before or during the binding event. This hypothesis is supported by the observation that MHC class II molecules are not excluded from the interface between cell membrane and phagocytosed latex beads and by an exclusion of surface molecules from the membranes of early phagosomes containing L. monocytogenes.
FACS analysis of J774 cells labeled with monoclonal antibodies to I-Ed, a Balb/c mouse-specific epitope of MHC class II molecules, show it to be present on the surface of approximately 65% of the cells. In contrast, only small amounts of I-Ad are present.
Once inside the cells the L. monocytogenes are again exposed to MHC class II molecules when organelles containing MHC class II are delivered to and fuse with the phagosome membrane just before the bacteria escape into the cell cytoplasm. The phagosome membrane also acquires lysosomal glycoproteins from organelles that fuse at this early internalization stage. Organelles with low luminal pH also fuse with the L. monocytogenes-containing phagosome during the first 25 min after exposure to bacteria.
Micrographs of cryosections through immunolabeled J774 cells exposed to L. monocytogenes (LM) for 5 min, showing labeling of early phagosomes. (
Cryosections through J774 cells exposed to L. monocytogenes (LM) for 10 min, showing labeling patterns over phagosome membranes. (
Micrograph of a section through J774 cells exposed to L. monocytogenes for 25 min, showing a L. monocytogenes-containing phagosome labeled with anti-LAMP 1. The antibody labels the phagosome membrane as well as a small spherical profile (arrow) close to the phagosome. Bar = 0.5 μm.
The technical limitations of the labeling methods used in this study could lead to conclusions that some of the observations are the result of technical artifacts. For this reason it was important to perform extensive control experiments to demonstrate specificity. It is possible that the exclusion of label at sites where L. monocytogenes is bound to the cell membrane was a result of antigen inaccessibility for antibody binding. To exclude this possibility, sites with loosely bound bacteria on the cell membrane were examined. These sites showed exclusion of MHC class II and biotin label even though antibody accessibility appeared possible. In addition, antibodies that recognized antigens on the bacterial surface were used to label the site of binding between plasma membrane and attached bacteria. These antibodies were able to label this site. Biotin on the bacterial surface was also accessible to antibody binding. Finally, anti-MHC class II labeling was detected at the interface where latex beads bound to J774 cells before internalization. Together, these observations suggest that the loss of labeling on the host plasma membrane was probably a result of a bacterial effect and not the result of physical exclusion of labeling. This is supported by the modified peptide profiles observed on early phagosomes that appear similar to those observed on the cell membrane.
Graph showing the percent co-localization of L. monocytogenes and membrane markers of J774 cells. During the early stages of internalization the bacteria associate with anti-MHC class II signal. Only a small number co-localize with LAMP 1. Over time, the co-localization with MHC class II declines but association with LAMP 11 increases.
Micrographs of sequential cryosections through J774 cells exposed to L. monocytogenes (LM) for 5 min. (
Micrographs of thawed frozen sections through J774 cells that have accumulated DAMP in low-pH compartments. The cells were incubated in the presence of L. monocytogenes (LM) for 25 min before fixation. Antibodies to DNP reveal the location of DAMP within the cells. (
The use of antibodies that recognize mouse Igs could also be a source of incorrectly observed immunolabeling on mouse macrophages. However, the control experiments that show only a limited amount of signal on these cells effectively rule out this possibility. Similarly, control experiments rule out the possibility that nonspecific binding of antibody to Fc receptors is playing a role in the labeling experiments.
Although J774 cells have previously been considered to be a non-MHC class II-expressing cell, it appears that approximately 60–70% of the cells in culture do express MHC class II molecules on their cell surface and at intracellular sites. The peptide profile produced on blots and the presence of a Balb/c-specific MHC class II epitope associated with the J774 cells support the observation that these cells express MHC class II molecules. It remains to be determined whether infected J774 cells are good candidates for stimulation and presentation of L. monocytogenes antigens. However, presentation of antigen using the class II MHC pathway has been reported to occur in these cells (Muno et al. 2000). Peptide association with MHC class II appeared to occur within 15 min after internalization.
When living L. monocytogenes are used to infect bone marrow macrophages, most of the intracellular bacteria are destroyed before escape from the phagosome has been accomplished (de Chastellier and Berche 1994). By using J774 cells, which have low bactericidal activity, the present study has shown that MHC class II molecules can interact with internalized viable L. monocytogenes only after internalization. Contact with MHC class II molecules is first restricted by excluding this molecule from parts of the host cell surface and then by excluding it from the early forming phagosome.
Internalized L. monocytogenes appear to encounter MHC class II molecules only after they have entered the cells and before they have escaped into the cell cytoplasm. Once in the cytoplasm the L. monocytogenes pass into adjacent cells via LAMP 1-positive compartments (this study and (Marquis et al. 1997) and appear to avoid re-entry into MHC class II containing-compartments. This reduced exposure to MHC class II molecules may help to explain the previously reported inhibition of antigen processing and presentation by macrophages infected with viable L. monocytogenes (Cluff et al. 1990). A similar study using heat-killed L. monocytogenes also reported intracellular interference with antigen processing in peritoneal macrophages. (Leyva-Cobian and Unanue 1988). Further study of MHC class II distribution on the membranes of the cell surface and early phagosomes during uptake of heat-killed L. monocytogenes is needed to compare the reported findings with the current results using viable bacteria.
In antigen-presenting cells, the exact site where peptide-MHC class II complexes are formed is still undetermined. Newly synthesized MHC class II molecules pass from the Golgi complex to specialized MHC class II-containing endosomal compartments (CIIVs) (Lotteau et al. 1990; Amigorena et al. 1994; West et al. 1994). In peritoneal macrophages, it has been suggested that antigen presentation may occur in tubulovesicular lysosomal structures (Harding and Geuze 1992). However, such structures were not observed in the L. monocytogenes-infected J774 cells. Because we have no information on the MHC class II presentation of bacterial antigens on the surface of J774 cells infected with living L. monocytogenes, it would be unwise to speculate on possible sites of peptide-MHC class II complex formation. However, the observation that the majority of L. monocytogenes co-localized with MHC class II antigens during the early stages of infection, but only few were seen associated with LAMP 1 antigens, suggests that interaction with intracellular MHC class II molecules occurs in compartments that do not contain LAMP 1.
The ability of pathogens to modify phagosome membranes appears to be a common event (Garcia-del Portillo and Finley 1995). In some cases, e. g., with Mycobacterium tuberculosis and Leishmania mexicana (Russell et al. 1992,1996; Clemens and Horwitz 1995), maturation of the phagosome membrane is impaired, resulting in a retention of endosomal characteristics. Other pathogens, such as the protozoan Toxoplasma gondii, exist within a membrane devoid of plasma membrane proteins (Sinai and Joiner 1997; Sinai et al. 1997). Exclusion of surface molecules from the phagosome membrane, as exhibited by L. monocytogenes, is shared by other organisms, such as Legionella pneumophilia (Clemens and Horwitz 1993,1995), which can exclude MHC class I and class II molecules from the vacuole membrane. Conversely, Streptococcus pneumoniae, when entering surface-biotinylated host cells, enters phagosomes with heavily biotinylated membranes, presumably originating from internalized plasma membrane (Tuomanen and Masure 1997).
It is commonly accepted that, during phagocytosis, newly formed phagosomes have a polypeptide profile similar to that of the plasma membrane (Muller et al. 1980a,b). This profile is changed, within minutes, by interactions with organelles of the endocytic pathway (Keilian and Cohn 1980; Muller et al. 1980a; Pitt et al. 1992a,b) and by recycling events (Mayorga et al. 1991; Pitt et al. 1992a). The current study shows that L. monocytogenes, in addition to controlling early fusion events and phagosome maturation (Alvarez-Dominguez et al. 1996,1997; Collins et al. 1997), is also able to exclude cell surface molecules, including MHC class II, from the phagosome membrane. This supports previous observations that L. monocytogenes is not entering cells by a classical or constitutive phagocytic pathway.
In conclusion, intracellular pathogens must utilize host cell processes for the regulation of pathogenesis. The results of this work suggest that L. monocytogenes enters cells through a route that is different from the phagocytic pathway. In this way the organism can partially limit exposure to MHC class II by excluding this molecule from the phagocytic vacuole as it is formed. The subsequent proteolysis of the organism resulting from exposure to lysosomes is negated by a rapid escape into the cell cytoplasm, possibly assisted by rapid delivery of low pH organelles to the phagosome. Peritoneal macrophages are able to destroy most of the intracellular bacteria before they can escape from the phagosome (de Chastellier and Berche 1994). This observation, taken together with the rapid delivery of MHC class II molecules to the phagosome seen in this study, suggests that host cells are not entirely defenseless against attack.
Footnotes
Acknowledgements
I thank Linda Chicoine for excellent technical assistance, Dan Portnoy for the L. monocytogenes cultures, Ira Mellman for immunoreagents, and Gareth Griffiths for advice on incubating latex beads with J774 cells.