|Home | About | Journals | Submit | Contact Us | Français|
Porphyromonas gingivalis, a periodontal pathogen, secretes outer membrane vesicles (MVs) that contain major virulence factors, including major fimbriae and proteases termed gingipains, although it is not confirmed whether MVs enter host cells. In this study, we analyzed the mechanisms involved in the interactions of P. gingivalis MVs with human epithelial cells. Our results showed that MVs swiftly adhered to HeLa and immortalized human gingival epithelial cells in a fimbria-dependent manner and then entered via a lipid raft-dependent endocytic pathway. The intracellular MVs were subsequently routed to early endosome antigen 1-associated compartments and then were sorted to lysosomal compartments within 90 min, suggesting that intracellular MVs were ultimately degraded by the cellular digestive machinery. However, P. gingivalis MVs remained there for over 24 h and significantly induced acidified compartment formation after being taken up by the cellular digestive machinery. In addition, MV entry was shown to be mediated by a novel pathway for transmission of bacterial products into host cells, a Rac1-regulated pinocytic pathway that is independent of caveolin, dynamin, and clathrin. Our findings indicate that P. gingivalis MVs efficiently enter host cells via an endocytic pathway and survive within the endocyte organelles for an extended period, which provides better understanding of the role of MVs in the etiology of periodontitis.
Outer membrane vesicles (MVs), ubiquitously shed from gram-negative bacteria by a mechanism involving cell wall turnover, consist of outer membrane lipids, as well as a subset of outer membrane proteins and soluble periplasmic components (56). Gram-negative pathogens have developed type I to VII secretion systems to transport active virulence factors in an extracellular manner (16, 49). Delivery by MVs is thought to be a distinct typeable secretion system that may play a part in the strategy utilized by bacterial pathogens to establish a colonization niche, modulate host defense and response, and impair host cell function. Thus, MVs have been proposed to be natural vehicles, or bacterial “bombs” (26). MVs from pathogenic strains such as Pseudomonas aeruginosa (2), Helicobacter pylori (18), and Actinobacillus actinomycetemcomitans (8), as well as pathogenic and nonpathogenic Escherichia coli (7, 53), have been shown to contain various active virulence factors such as toxins, proteases, adhesins, and lipopolysaccharide. However, the molecular mechanism of the entry of MVs into host cells for virulence factor delivery is largely unknown. Recently, cellular entry by MVs derived from enterotoxigenic E. coli was closely analyzed (24). Those results showed that the MVs were taken up by a human epithelial cell line via lipid rafts, which was dependent on monosialoganglioside and intracellular MVs accumulated in nonacidified compartments inaccessible to the extracellular milieu. Thus, it was concluded that enterotoxigenic E. coli MVs function as specifically targeted transport vehicles that mediate entry of heat-labile enterotoxin and other bacterial envelope components into host cells. These findings suggest a novel mechanism for bacterial virulence factor transmission into host cells, although no other related results have been reported.
Periodontitis, one of the most common infectious diseases seen in humans (54), is characterized by gingival inflammation, as well as loss of connective tissue and bone from around the roots of the teeth, which leads to eventual tooth exfoliation. This infectious disorder is caused by a small subset of periodontal gram-negative anaerobic bacteria (17). Among those, Porphyromonas gingivalis is considered to be a bona fide pathogen that causes several forms of severe periodontal disease, and has also been reported to contribute to systemic conditions such as cardiovascular diseases (34) and preterm low birth weight (38). P. gingivalis releases MVs in an extracellular manner, which retain the full components of outer membrane constituents, including lipopolysaccharide, muramic acid, a capsule, fimbriae, and proteases termed “gingipains” (13, 33). Long fimbriae (referred to herein as simply fimbriae) reportedly mediate bacterial adherence to and entry into periodontal cells (1), while gingipains contribute to the destruction of periodontal tissues (20). Therefore, fimbriae and gingipains provide adhesive and proteolytic abilities to MVs, which, together with their small size (20 to 500 nm), are suspected to enable them to penetrate intact mucosa and enter underlying host tissues and cells (32). However, the mechanisms involved in the interactions of P. gingivalis MVs with the host cell remain unknown, and no study has actually demonstrated that MVs adhere to or enter host cells. Thus, we considered it important to determine how P. gingivalis MVs are taken up by human epithelial cells, as well as whether intracellular MVs are sorted to lytic compartments or accumulate in nonacidified compartments as enterotoxigenic E. coli MVs.
In the present study, we performed molecular dissection to elucidate the mechanisms underlying entry into and intracellular trafficking of P. gingivalis MVs in human cervical HeLa and immortalized human gingival epithelial (IHGE) cells.
P. gingivalis strain ATCC 33277 and its derived fimbria-null mutant strain KDP150 (ΔfimA) (48), which was kindly provided by K. Nakayama (Nagasaki University, Japan), were used in this study. The organisms were anaerobically maintained in blood agar plates, and grown in trypticase soy broth (Nissui, Tokyo, Japan) supplemented with hemin (5 μg/ml; Wako Pure Chemical Industries) and menadione (1 μg/ml; Sigma-Aldrich), as described previously (35). P. gingivalis MVs were prepared as described previously (52). Briefly, P. gingivalis cells were removed from the culture by centrifugation at 10,000 × g for 30 min at 4°C. Following filtration of the supernatant with Millipore filters (pore size, 0.22 μm), MVs were collected as pellet by ultracentrifugation at 100,000 × g for 3 h at 4°C.
Cytochalasin D, nocodazole, and methyl-β-cyclodextrin (MβCD) were purchased from Sigma-Aldrich (St. Louis, MO), and wortmannin from Wako Pure Chemical Industries (Osaka, Japan). These reagents were dissolved in an aliquot of dimethyl sulfoxide (DMSO; 1.1 μg/ml) and added to the culture medium. Alexa Fluor 568-conjugated phalloidin and transferrin were purchased from Invitrogen (Carlsbad, CA), as was LysotrackerRed. The dominant-negative (DN) Eps15 construct GFP (green fluorescent protein)-Eps15Δ95/295, in which the second and third Eps15 homology (EH) domains are deleted (4), was kindly provided by A. Dautry-Varsat (Institut Pasteur, France). Using this plasmid, we constructed dominant mutant GFP-Eps15ΔEH, lacking the EH domains, which are required for clathrin-dependent endocytosis. GFP-dynamin 1-K44A and GFP-dynamin 2-K44A (22) were generous gifts from K. Nakayama (Kyoto University, Japan). The mutants Dyn1-K44A and Dyn2-K44A are defective in GTP binding and show a DN effect. Caveolin-1-GFP (40) was a kind gift from A. Helenius (Swiss Federal Institute of Technology, Switzerland). GFP-glycosylphosphatidylinositol (GPI), a fusion construct containing the GPI-anchoring sequence to GFP (25), was provided by G. Kondoh (Kyoto University, Japan). The DN Rac1-S17N and Cdc42-S17N vectors were generous gifts from K. Kaibuchi (Nagoya University, Japan) and H. Miki (University of Tokyo, Japan), respectively.
Human cervical epithelial HeLa cells (CCL-2) were obtained from the American Type Culture Collection (Manassas, VA). The cells were grown in Dulbecco's modified Eagle's medium (DMEM; Sigma) supplemented with 10% fetal bovine serum (FBS; JRH Biosciences, Inc., Lenexa, KS), penicillin (100 U/ml), and streptomycin (50 μg/ml) at 37°C in 5% CO2. IHGE cells (27) were kindly provided by S. Murakami (Osaka University, Japan) and maintained in Humedia KB-2 without FBS (Kurabo, Osaka, Japan), as described previously (23). An hour before each infection assay, the culture medium for IHGE cells was changed to fresh DMEM with 10% fetal calf serum (23). To construct DN mutants, HeLa cells were seeded on glass coverslips in 24-well plates (6 × 104 cells per well) and then transiently transfected with 1.0 μg of the plasmid using LipofectAMINE 2000 reagent (Invitrogen), according to the manufacturer's recommendations. Transfected cells were further incubated for 24 h and used for the assays.
Drug treatment experiments were performed as described previously (50), with minor modifications. Briefly, HeLa and IHGE cells were washed twice with serum-free DMEM and incubated with the following drugs in DMEM at 37°C for 30 min prior to the assay. To disorganize the cytoskeletal architecture, cytochalasin D (1 μg/ml) and nocodazole (25 μM) were used, while MβCD (10 mM) was used to analyze the role of cholesterol in MV entry and wortmannin (100 nM) was used to inhibit phosphatidylinositol 3′ kinase (PI3K) activity. All reagents were included in the medium throughout the entry experiments.
HeLa and IHGE cells (6 × 104 cells per well) were seeded onto glass coverslips in 24-well plates the day before the assay. Two hours prior to each assay, culture media for both HeLa and IHGE cells were changed to fresh DMEM with 10% FBS. The cells were washed twice and then incubated in serum-free DMEM for 15 min with MVs (30 μg/ml). After washing with DMEM, the cells were incubated in DMEM containing 10% FBS for various times following the addition of MVs. For immunostaining, the cells were washed with phosphate-buffered saline (PBS), fixed with 3% paraformaldehyde in PBS for 15 min, and permeabilized with 0.1% Triton X-100 in PBS for 5 min. After being washed twice with PBS, the cells were incubated in blocking solution (0.1% gelatin in PBS) for 10 min and subsequently with primary antibodies diluted with blocking solution at room temperature for 1 h. MVs were probed with rabbit polyclonal antibodies against native fimbriae of P. gingivalis ATCC 33277, which were prepared as described previously (55). To identify acidified compartments, HeLa cells were incubated with MVs for 150 min and further incubated with LysotrackerRed (50 nM) in DMEM at 37°C for 60 min. The cells were analyzed using a laser-scanning confocal microscope (model LSM510; Carl Zeiss, Thornwood, NY) and Zeiss LSM Image Browser software (Carl Zeiss), at a magnification of 630×. The laser power and pinhole size were set based on the optimal trade-off between the z resolution and signal/noise ratio of the images in each experiment by LSM Image Browser software to avoid artifacts at different depths within the cells. To analyze the distribution of MVs that entered the cells, those located on the inside 1 μm distant from the outermost actin filament were counted manually. The perimeter of the cells was determined after incubation with sulfo-NHS-LC-biotin (100 μg/ml) at 4°C for 15 min following drug treatment with cytochalasin D. To confirm our colocalization observations, a z-series with 0.3-μm intervals was scanned and images of the x-z and y-z planes were reconstructed with the orthogonal section tool of the LSM510 software. More than 120 cells were analyzed with three independent experiments performed.
Mouse anti-clathrin monoclonal antibodies were purchased from BD Biosciences (San Jose, CA), while the mouse anti-early endosome antigen 1 (EEA1) monoclonal antibodies were from Abcam (Cambridge, MA), and mouse anti-lysosome-associated membrane protein 1 (LAMP1) monoclonal antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA). Alexa Fluor-conjugated secondary antibodies (goat anti-mouse immunoglobulin G [IgG] and goat anti-rabbit IgG), Alexa 488-conjugated streptavidin, and Alexa 568-conjugated phalloidin, used for fluorescence microscopy, were from Invitrogen. EZ-link sulfo-NHSLC-biotin was from Pierce (Rockford, IL).
Conventional electron microscopy was performed as described previously (11). Briefly, HeLa cells (6 × 104 cells per 24-well culture dish) were seeded onto collagen-coated plastic coverslips the day before the assay. The cells were washed twice with serum-free DMEM and incubated for 15 min with MVs (30 μg/ml). Then, the cells were washed and further incubated in 10% FBS-DMEM for 15 min, after which they were fixed in 2% glutaraldehyde in 0.1 M Na-phosphate buffer (pH 7.4) for 1 h. Finally, the cells were washed in the same buffer three times. Ultrathin sections were doubly stained with uranyl acetate and lead citrate, and observed under an H7000 electron microscope (Hitachi, Tokyo, Japan). Immunoelectron microscopy using antifimbrial antibodies was also performed using a gold enhancement method, as described previously (29).
All data are presented as the mean ± standard deviation. Statistical analyses were performed using an unpaired Student's t test. Multiple comparisons were performed by one-way analysis of variance and Sheffe's test using STAT View software (SAS Institute, Inc., Cary, NC).
First, we examined whether P. gingivalis MVs adhered to and entered HeLa and IHGE epithelial cells. As shown in Fig. Fig.1A,1A, MVs apparently interacted with the epithelial cells. At 15 min after addition of the MVs, most were found to be associated with the cellular plasma membrane, although they scarcely existed within the cells. The number of intracellular MVs increased with extended incubation time, with nearly all MVs found to have entered the cells at 90 min, where they remained after 24 h (Fig. 1A and B). There was no difference observed between the process of entry of MVs into HeLa and IHGE cells. In contrast, MVs prepared from a fimbria-null mutant were found to negligibly interact with HeLa and IHGE epithelial cells (data not shown).
Next, we determined whether MVs were internalized within the HeLa and IHGE cells by endocytic compartments associated with EEA1 and LAMP1. We found that MVs became colocalized with EEA1 in an incubation-time-dependent manner, as a number of MVs were clearly colocalized with EEA1 at 30 min within both cell types, while colocalization was decreased after 90 min (Fig. (Fig.2A).2A). In contrast, MVs began to be colocalized with LAMP1 at 30 min and the number increased with incubation time, until LAMP1 surrounded most after 24 h and no intracellular MVs were observed at 36 h (data not shown), suggesting that they were degraded within the lysosomes. These intracellular transient characteristics of MVs were quantified using projection image and vertical (z) optical sections (x-z and y-z planes) (Fig. 2B and C). The ratios (%) of MV colocalization with EEA1 and LAMP1 in single-scan images (Fig. (Fig.2A)2A) were found to be similar to those in z-series images (Fig. (Fig.2B).2B). Most of the MVs were sorted to endocytic compartments associated with EEA1 and LAMP-1 within 90 min and remained there for more than 24 h in both cell types (Fig. 2B and C). In addition, the MVs apparently induced formation of acidified compartments and merged with them (Fig. (Fig.2D2D).
Immunoelectron microscopic analysis was also performed to confirm that anti-native fimbria antibodies accurately trace intracellular MVs (Fig. (Fig.3).3). At 90 min after the addition of MVs, individual endocytic compartments were found to contain multiple MVs (Fig. 3B and C) and the antibodies had traced the MVs enwrapped with those compartments (Fig. 3F and G). The outlines of the enwrapped MVs changed to become indistinct after 24 h (Fig. 3D and E), although they were detected by antifimbria antibodies. The results presented in Fig. Fig.1,1, ,2,2, and and33 indicate that P. gingivalis MVs enter cells via an endocytic pathway and are sorted to lysosomal organelles, suggesting that intracellular MVs are ultimately degraded by the cellular digestive machinery. In addition, fimbriae were shown to be associated with MVs, indicating that antifimbria antibodies are an appropriate tracer for MVs. It is notable that the MVs were able to survive within the lysosomal compartments for more than 24 h, which is significantly longer than the normal range (2 to ~3 h) for protein degradation (46).
To identify the entry route for MVs into epithelial cells, their colocalization with specific markers for caveolin, GPI-AP1, and clathrin was examined using HeLa cells. The MVs apparently merged with GPI-AP1 (Fig. 4A and B), whereas there was negligible colocalization with caveolin and clathrin at 15 min after addition (Fig. (Fig.4A).4A). The involvement of Rac1, Cdc42, and dynamin in MV entry was also examined using the overexpression of appropriate dominant mutants. Entry of MVs was prevented in cells that expressed the DN form of Rac1, while inhibition also occurred slightly in Cdc42-DN cells, though it did not reach the level of significance (Fig. 4C and D). Inhibition of the clathrin- and dynamin-dependent pathways showed no effect. These results indicate that Rac1 is involved in MV entry, whereas Cdc42, clathrin, and dynamin are not. During our experiments, we found that IHGE cells were too fragile to be transfected with plasmid vectors for construction of DN mutants. Therefore, we did not employ IHGE cells for the experiments shown in Fig. Fig.44.
To examine whether P. gingivalis MVs use lipid rafts, we disturbed raft formation by treatment with MβCD, which removes cholesterol from the membrane (50). To evaluate the role of PI3K, which is required to activate Rac1 and Cdc42 (3), wortmannin (PI3K inhibitor) was utilized, as were cytochalasin D (actin polymerization inhibitor) and nocodazole (microtubule assembly inhibitor). As shown in Fig. Fig.5A,5A, MV entry was significantly inhibited by MβCD, wortmannin, and cytochalasin D, whereas nocodazole had no effect. There were no significant differences between HeLa and IHGE cells with regard to these inhibitory effects (Fig. (Fig.5B).5B). These results indicate that lipid rafts, PI3K activities, and actin polymerization are required for MV entry.
It has been reported that a significant portion of the lifestyle of P. gingivalis is harbored within microbial biofilm (dental plaque) that develops on gingival margins (28). Therefore, MVs, bacterial bombs released from biofilm, may be involved in a wide range of pathological processes to attack host cells. Entry and nuclear targeting by gingipains in HeLa cells were previously demonstrated (44). However, despite a great amount of published data regarding the virulence factors of P. gingivalis, nothing has been reported regarding the interaction of P. gingivalis MVs with host cells. The present experiments showed that MVs swiftly adhered to HeLa and IHGE cells in a fimbria-dependent manner and then entered via a lipid raft-dependent endocytic pathway, after which they were transported to lysosomes and degraded. It was initially reported that enterotoxigenic E. coli MVs underwent pinocytosis via cholesterol-rich lipid rafts (24). The present study is the second known report regarding the molecular mechanism of the cellular entry/trafficking of MVs.
Cellular endocytosis encompasses several diverse mechanisms by which cells internalize macromolecules and particles into transport vesicles derived from the plasma membrane. Cells often internalize macromolecules and particles into transport vesicles via clathrin- and/or caveolin-independent endocytosis pathways, which require GTPase dynamin activity to generate endocytic carriers (43, 45), as well as Rho family GTPases, such as Rac1 and Cdc42, for endocytic actin rearrangement (3, 14). In contrast, the present study demonstrated that P. gingivalis MVs were internalized via a Rac1-regulated pinocytic pathway that is independent of caveolin, dynamin, and clathrin. Although the mechanisms that govern caveola- and clathrin-independent endocytosis remain poorly understood (6), a growing number of dynamin-independent endocytic routes for various cell surface proteins including lipid rafts are being reported (6, 31, 36, 37, 43). Lipid rafts are membrane subdomains enriched in cholesterol and sphingolipids, as well as specific membrane proteins such as glycosyl-phosphatidylinositol (GPI)-anchored proteins (GPI-APs) and caveolae. These rafts have been suggested to be entry sites for several different types of bacteria (15, 30, 42). In addition, molecules clustered on lipid rafts appear to provide sorting signals for selective endocytosis via a Cdc42-regulated, and dynamin- and clathrin-independent pinocytic pathway, which has been termed the GPI-AP-enriched early endosomal compartments (GEECs) pathway (5). A recent report showed that H. pylori VacA cytotoxin underwent pinocytosis to late endosomes via the GEECs pathway (12). However, the involvement of the present pinocytic pathway, which is Cdc42 regulated and dynamin and clathrin independent, in the internalization of bacterial products has not been reported. To the best of our knowledge, the entry route of P. gingivalis MVs presented here is a novel pathway mediated by varying involvement of regulatory GTPases.
In our previous study using fluorescent beads coated with MVs as homogenous artificial intruders, we suggested that P. gingivalis cells entered cells through an actin-mediated pathway that is controlled by PI3 and Rac1 in a caveolin- and dynamin-dependent manner (52). Those results are different from the present findings. Endocytic pathways differ based on the size of the endocytic vesicle: for example, macropinocytosis (>1 μm), clathrin-mediated endocytosis (~120 nm), caveolin-mediated endocytosis (~60 nm), and clathrin- and/or caveolin-independent endocytosis (~90 nm) (6). MVs are much smaller (20 to 500 nm) than bacterial cells (≥1 μm); thus, the discrepancies between our former and present findings indicate that the endocytic pathways differ in regard to the sizes of the cargos taken up into the cells.
P. gingivalis MVs were degraded in lysosomal compartments of HeLa and IHGE cells, which indicates characteristics in contrast to those of enterotoxigenic E. coli MVs, which were able to escape from the endocytic degradation pathway and accumulate in nonacidified compartments (24). P. gingivalis MVs may not be efficient transport vehicles that mediate entry of bacterial cytotoxic products into cellular cytoplasm and organelles, except for endosomes/lysosomes. However, in the present study, MVs survived within endocytic organelles for more than 24 h, with a significant induction of acidified compartment formation seen. Such cellular stress was previously suggested to cause lysosome-specific initiation of cellular impairment (10). Based on the pathological features of chronic periodontitis caused by P. gingivalis, daily and continuous attacks by these bacterial bombs from biofilm seem able to destroy periodontal tissue. Therefore, P. gingivalis MVs may be a powerful offensive weapon in the longstanding battle between bacterial biofilm and the host. On the other hand, our previous study showed that the entry of fluorescent beads coated with MVs did not significantly induce cellular death (19). Additional analysis of the entry-associated effects of P. gingivalis MVs on cellular function is needed.
Collectively, our results show a novel molecular basis for the entry of P. gingivalis MVs into host epithelial cells. In addition, MVs have been reported to possess various virulent abilities (9, 13, 21, 39, 41, 47, 51). We intend to conduct additional studies to elucidate the cellular functions impaired by invading MVs.
This research was supported in part by a grant from the 21st Century Center of Excellence program entitled “Origination of Frontier BioDentistry,” held at Osaka University Graduate School of Dentistry, as well as Grants-in-Aid for Scientific Research (B) from the Ministry of Education, Culture, Sports, Science, and Technology.
Editor: V. J. DiRita
Published ahead of print on 3 August 2009.