INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Pathogens have developed remarkable and diverse strategies of host cell infection and tissue dispersion (7). Many invasive bacteria have common approaches of host cell interaction, but each species has evolved a subset of unique tactics that exploit normal host cell function, promoting survival and enhancing virulence (7). Invasive bacteria elicit their own uptake into typically nonphagocytic host cells. Entry by induced phagocytosis usually involves host cell receptors, such as integrin or epidermal growth factor receptor (13, 19), and the complicated manipulation of the host cell cytoskeleton by the bacterium. Some bacteria, such as the enteric Salmonella, Yersinia, Shigella, and Listeria spp., usurp only microfilaments for entry (7, 8, 12, 14, 32, 44). Internalization by most strains of the oral periodontopathogen Actinobacillus actinomycetemcomitans, including strain SUNY 465, the invasion prototype, is also actin dependent (2, 9, 10, 29, 30). Polymerized actin accumulates beneath the host cell cytoplasmic membrane at the site of entry of these organisms (8, 9). Other bacteria, such as Citrobacter freundii and Camplobacter jejuni, initiate distinctive microtubule entry processes (16, 33). Certain others, such as Porphyromonas gingivalis, Edwardsiella spp., Neiserria gonorrhoeae, enteropathogenic Escherichia coli, Haemophilus influenzae, and Vibrio hollisae, require both microtubules and microfilaments for entry (6, 20, 24, 31, 36, 43). Some organisms persist and/or multiply within the phagocytic vacuole in which they are internalized (25, 41). Others escape from the vaculole, multiply within the host cell cytoplasm, and spread to adjacent epithelial cells (7, 11, 27, 29, 34, 39, 44). Both Shigella flexneri and Listeria monocytogenes multiply in the cytoplasm, move intracellularly, and spread to adjacent cells (12, 35, 39, 40, 44). The intra- and intercellular movement is based on continuous actin assembly at one pole of the bacterium which is generated by the asymmetrical expression of specific proteins on the replicating bacterial surface (1, 34, 35, 44). At the host cell plasma membrane, moving bacteria generate elongated cell surface protrusions that are engulfed by neighboring cells (22, 44).

Previously we reported that strain SUNY 465, the A. actinomycetemcomitans invasion prototype, egresses from host epithelial cells soon after internalization (27). These A. actinomycetemcomitans organisms can be recovered in the assay milieu. The spread of A. actinomycetemcomitans to adjacent monolayers was also demonstrated (27). Like that of S. flexneri and L. monocytogenes, the spread involves protrusions that appear to be generated by bacteria pushing out the cell plasma membranes. However, the spread does not seem to be mediated by the engulfment of protrusions by adjacent cells as with Shigella and Listeria. Instead, it appears that A. actinomycetemcomitans bacteria spread to neighboring cells by travelling through protrusions that connect adjacent cells. Whereas the host cell protrusions that mediate the spread of A. actinomycetemcomitans contain microfilaments, there is no indication that they are directly involved in its cell-to-cell spread (27).

Prior studies in our laboratory demonstrated that microtubule-disrupting agents did not inhibit SUNY 465 internalization into epithelial cells, indicating that microtubules are not required for its entry (42). However, colchicine, which causes microtubule depolymerization and reduces microtubule mass (4), was shown to increase markedly the number of intracellular SUNY 465 organisms (26, 42). Immunofluorescence microscopy revealed that intracellular A. actinomycetemcomitans localized almost exclusively with microtubule-organizing centers of taxol-induced asters in KB cells (29). These observations suggested a role for microtubules in the egression and/or intra- and intercellular spread of this organism. In this study we used both quantitative gentamicin protection assays and immunofluorescence microscopy to further examine the effects of microtubule modulators (colchicine, nocodazole, and taxol) and cytochalasin D, a microfilament inhibitor, on SUNY 465 invasion and egression. Immunofluorescence microscopy was also used to demonstrate the presence of microtubules in host cell protrusions and the binding of A. actinomycetemcomitans SUNY 465 specifically to the plus ends of filaments of microtubule asters in a KB cell extract (KBE).

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacteria. A. actinomycetemcomitans SUNY 465, SUNY 523, and 652, Haemophilus aphrophilus ATCC 19415, H. influenzae KW20, E. coli HB101, S. flexneri M90T, and L. monocytogenes ATCC 19111 were used (2, 28, 30). Cells were maintained frozen at 70°C in 10% dimethyl sulfoxide (DMSO) (Sigma Chemical Co., St. Louis, Mo.). All bacteria except H. influenzae were cultured in Trypticase soy broth (Difco Laboratories, Detroit, Mich.) supplemented with 0.6% yeast extract (Difco). Solid medium was prepared by adding agar (Difco) to liquid medium to a final concentration of 1.5% (wt/vol). H. influenzae was cultured in Todd-Hewitt broth (Becton Dickinson, Cockeysville, Md.) or agar. Bacteria were cultivated at 37°C; A. actinomycetemcomitans strains and H. aphrophilus were in a humidified atmosphere of 10% CO₂ in air.

Cell culture. The KB cell line (derived from a human oral epidermoid carcinoma) was maintained in RPMI 1640 medium (Sigma) supplemented with 5% fetal bovine serum (GIBCO, Grand Island, N.Y.) and 50 µg of gentamicin (Sigma) per ml. KB cells were cultured in 75-cm² flasks at 37°C in a humidified atmosphere of 5% CO₂ in air. Cultures were split by treatment with 0.02% EDTA (Sigma) followed by trypsin (GIBCO) to detach cells.

Standard quantitative invasion assay. Approximately 10⁵ KB cells in antibiotic-free medium were seeded onto glass coverslips in wells of 24-well tissue culture plates and incubated for 16 to 18 h. Overnight cultures of the A. actinomycetemcomitans invasion prototype strain, SUNY 465, were diluted in fresh broth and harvested during early exponential growth. Bacteria were pelleted by centrifugation and suspended in antibiotic-free medium, and the semiconfluent cell monolayers were inoculated with bacterial suspensions adjusted to obtain a multiplicity of infection of 1,000 bacteria to 1 KB cell. Bacteria were centrifuged onto the monolayers at 900 × g for 10 min at room temperature and incubated at 37°C for 2 h. Extracellular, unattached bacteria were removed by washing monolayers two times with phosphate-buffered saline that contained 1.0 mM CaCl₂ and 0.5 mM MgCl₂ (PBS). Monolayers were incubated for 1 h in the presence of cell culture medium that contained 100 µg of gentamicin per ml to kill extracellular A. actinomycetemcomitans. The medium was removed, and monolayers were washed twice with PBS. One milliliter of 0.5% Triton X-100 (Sigma) solution in PBS was added to lyse the KB cells and release internalized bacteria. Two milliliters of PBS was added to each well to dilute the detergent, appropriate dilutions were spread onto Trypticase soy broth-yeast extract plates, and CFU were enumerated. To assess A. actinomycetemcomitans egression from KB cells into the medium, the standard quantitative invasion assay was modified as follows. After the gentamicin treatment step, the medium was removed, monolayers were washed with PBS, fresh antibiotic-free medium was added, and the monolayers were incubated further. At various times, the medium was removed for analysis and fresh medium was added to the monolayers. Assay medium was analyzed for the presence of A. actinomycetemcomitans by plating aliquots and enumerating CFU. All quantitative determinations were carried out in quadruplicate.

Modulating biochemicals. Taxol, nocodazole, and colchicine (all microtubule inhibitors), cytochalasin D (a microfilament inhibitor), and brefeldin A (a drug that interferes with the Golgi network) were used. All were obtained from Sigma. Stock solutions were prepared in DMSO as follows: taxol (pacitaxel), 10 mM; nocodazole [methyl-5-(2-thienylcarbonyl)-1H-benzimidazol-2-yl] carbonate, 5 mg/ml; colchicine, 1 mg/ml; cytochalasin D, 5 mg/ml; and brefeldin A, 1 mg/ml. Taxol was stored at 4°C, and colchicine, cytochalasin D, nocodazole, and brefeldin A were stored at 20°C. Stock solutions were diluted in antibiotic-free medium to obtain the following final assay concentrations: taxol 10 µM; nocodazole, 10 µg/ml; colchicine, 5 µg/ml; cytochalasin D, 5 µg/ml; and brefeldin A, 1 µg/ml. Taxol, colchicine, and brefeldin A were added to monolayers 30 min prior to the addition of bacteria. The monolayers were incubated in nocodazole for 1 h on ice and then warmed to 37°C for 30 min prior to the addition of bacteria (38). Microtubule inhibitors and brefeldin A were present during all incubations but not during PBS washes. However, since pretreatment of KB cells with cytochalasin D markedly inhibits A. actinomycetemcomitans SUNY 465 invasion (30), the treatment regimen for it had to be modified to enable entry. Cytochalasin D was added after 90 min of infection and was also present during the gentamicin treatment step. Monolayers treated with appropriate amounts of DMSO diluted in antibiotic-free medium served as controls.

Immunofluorescence microscopy of invasion. Immunofluorescence microscopy (15) was carried out to monitor SUNY 465 internalization and to examine host cell protrusions for the presence of microtubules. Standard invasion assays were carried out as described above, except that the gentamicin step was omitted. At various times after infection, the assay medium was removed and cells were washed and fixed for 20 min in 3.7% formaldehyde (in PBS). If microtubules were stained, monolayers were incubated first in 0.2 mg of dithio-bis(succinimidyl proprionate) (Sigma) per ml for 5 min and then in 100 mM PIPES [piperazine-N,N'-bis(2-ethanesulfonic acid)] (pH 6.9) (Sigma)-1 mM EGTA (Sigma)-4% polyethylene glycol 6000 (J. T. Baker Co, Phillipsburg, N.J.) for 5 min prior to fixation. After being washed, monolayers were incubated for 15 min in SUNY 465 antiserum (42) diluted 1:2,500. Monolayers were washed and incubated in tetramethylrhodamine isothiocyanate (TRITC)-conjugated anti-rabbit immunoglobulin G (IgG) (Sigma, Catalog no. T-5268) (1:100 dilution) for 15 min in the dark. (All subsequent steps, except washes, were also carried out in the dark.) After being washed, monolayers were permeabilized by incubation in 0.1% Triton X-100 for 15 min. Monolayers were washed and incubated again in SUNY 465 immune serum and in monoclonal anti--tubulin (mouse) (Sigma, Catalog no. T-9026) (1:500 dilution) for 15 min to label microtubules. Monolayers were washed and incubated in fluorescein isothiocyanate (FITC)-labeled anti-rabbit IgG (Sigma, Catalog no. T-5268) and in TRITC-conjugated anti-mouse IgG (Sigma, Catalog no. T-5393) (1:100 dilution). Monolayers were washed, and coverslips were mounted in Vectashield (Vector Laboratories, Burlingame, Calif.) and sealed with nail polish. Indirect immunofluorescence was examined under oil immersion with a Nikon EC 400 (fluorescence/phase) microscope and photographed with a Nikon N6006 camera with Kodak Ektachrome professional slide film with an ASA of 400.

KBE microtubule binding assay. Binding of bacteria to microtubules was studied by using a cell-free method recently developed in our laboratory (37). The bacterial strains used and their relevant characteristics are listed in Table 1. KB cells were grown for 2 to 4 days in 1.0 liter of RPMI 1640 cell culture medium with 10% fetal bovine serum at 37°C in a humidified atmosphere of 5% CO₂ in air. The culture was centrifuged at 20,000 × g for 20 min, and the supernatant was decanted. Cells were suspended in 5 volumes of 30 mM Tris-HCl (pH 7.5)-120 mM KCl-5 mM Mg acetate [Mg(OAc)₂] and centrifuged as before. The supernatant was decanted, and the cell pellet was suspended in 2 volumes of 10 mM Tris-HCl (pH 7.5)-10 mM KCl-1.5 mM Mg(OAc)₂ for 5 min to enable swelling and then homogenized for 15 s. The salt concentration in the mixture was adjusted by the addition of 0.1 volume of 230 mM Tris-HCl (pH 7.5)-1.27 M KCl-40 mM Mg(OAc)₂, and it was centrifuged again as before. The supernatant was removed by aspiration and passed through a 0.8-µm/0.2-µm-pore-size Acrodisc PF filter (Gelman Sciences, Ann Arbor, Mich.) to yield the KBE. Bacteria (2 µl, ~10⁷ cells) and KBE (7 µl) were incubated for 20 min in a 37°C H₂O bath, followed by the addition of 1 µl of 10 mM taxol and further incubation to mediate the formation of small asters of microtubules. Unattached and weakly binding bacteria were removed by centrifugation of the reaction milieu through a two-layer sucrose cushion as follows. The bacterium-aster mixture was underlaid with 500 µl of 20% sucrose (in PBS); the 20% layer was then underlaid with 500 µl of 40% sucrose-0.1% glutaraldehyde (in PBS) and centrifuged at 100 × g, producing a soft pellet from which the supernatant was carefully removed. The pellet (bacterium-associated microtubules or microtubules) was suspended in 50 µl of 0.1% glutaraldehyde in PBS. For microscopy, 10 µl of the suspension was applied to a 12-mm-diameter polylysine-treated coverslip in a 24-well tissue culture plate, the plate was centrifuged at 100 × g for 15 min or until dry, and the coverslip was washed twice with PBS. To visualize bacterium-aster interaction, antibodies specific for microtubules and bacteria or dyes (propidium iodide and SYTO 9) that stain microtubules or bacteria nonspecifically (LIVE/DEAD BacLight Viability Kit; Molecular Probes, Eugene, Oreg.) were used. Propidium iodide stains DNA in dead bacteria; SYTO 9 stains microtubules. The dye strategy can be used with any bacterium, so it is cost-effective for screening different species. In addition, it does not require the washing steps, as does the antibody method, so it reduces stress on the asters. The antibody strategy has the advantage of specificity; it can be used to identify specific bacteria in a mixed population. When the antibody strategy was used, coverslips were overlaid with appropriate dilutions of antibacterial antibody and monoclonal anti--tubulin (1:1,000 dilution) and incubated at room temperature for 20 min. Primary antibodies were removed, and the coverslips were washed twice with PBS, overlaid with secondary antibodies (anti-rabbit FITC conjugate and anti-mouse Texas Red or TRITC conjugate at 1:500 dilutions in PBS), and incubated at room temperature in the dark for 20 min. Antibodies were removed, and coverslips were washed twice with PBS, mounted on microscope slides with VectaShield, and sealed with nail polish. When the dye strategy was used, after the initial incubation, each dye mixture supplied with the BacLight Kit was added to microtubule-bacterium complexes at 0.5 ml per well. The complexes were incubated for 5 min at 37°C and applied to the two-layer sucrose cushion. After centrifugation as described above, the supernatant was removed and the complexes were washed as follows. Glutaraldehyde in PBS (1.0 ml) was added gently, and the plate was centrifuged at less than 100 × g for 5 min. The supernatant was removed, and the procedure was repeated. The supernatant was removed, and 5 µl was applied to a coverslip and mounted on a slide as described above. Both strategies elicit the same results. The antibody strategy was used for A. actinomycetemcomitans SUNY 465, L. monocytogenes, H. aphrophilus, and E. coli HB101; the dye strategy was used for all other bacteria.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Based on prior studies, we hypothesized that the microtubule network provided a potential mechanistic track by which A. actinomycetemcomitans could move within the host cell and spread to adjacent cells. In the present study we used in vitro techniques to examine the localization of microtubules in infected KB cells, the effects of microtubule inhibitors on A. actinomycetemcomitans invasion of KB cells and its egression from the cells into the assay medium, and the interaction of A. actinomycetemcomitans and other bacterial species with KB cell microtubules in a cell extract.

Microtubule inhibitors increased the number of intracellular A. actinomycetemcomitans organisms recovered from infected KB cells. The number of A. actinomycetemcomitans organisms recovered from KB cells treated with nocodazole or colchicine, drugs that bind to and inhibit microtubule polymerization, or with taxol, a drug that binds to microtubules and prevents their depolymerization, was three- to fourfold greater than control values (Fig. 1). None of the inhibitors affected the viability or growth rate of A. actinomycetemcomitans (data not shown). Immunofluorescence microscopy confirmed the quantitative results. The number of intracellular A. actinomycetemcomitans organisms was increased markedly, and most organisms occurred in clusters in KB cells treated with taxol (Fig. 2a) compared with controls (Fig. 2b). Colchicine treatment produced similar results (data not shown).

View larger version (45K): [in this window] [in a new window]   FIG. 1.   Taxol, nocodazole, and colchicine increased the number of intracellular organisms recovered from A. actinomycetemcomitans-infected KB cells. Taxol (10 µM), nocodazole (10 µg/ml), and colchicine (5 µg/ml) were added to monolayers 15 min prior to the addition of A. actinomycetemcomitans and were present in the assay medium at these concentrations throughout the assay. The bars represent ratios of internalized A. actinomycetemcomitans in the presence of the microtubule modulator to internalized A. actinomycetemcomitans in the absence of the modulator. Values are the means for quadruplicate samples from a typical experiment. The standard deviation was less than 25% of the mean in all cases.

View larger version (93K): [in this window] [in a new window]   FIG. 2.   Immunofluorescence microscopy of taxol-treated KB cells infected with A. actinomycetemcomitans SUNY 465. (a) Untreated KB cells; (b) taxol (10 µM)-treated KB cells. Taxol was added to KB cells 15 min prior to infection to stabilize the microtubules. The staining and filter set used allows visualization of only internalized bacteria (arrows) and not KB cells; thus, the KB cells (grey) are not well defined. Note that the number of intracellular A. actinomycetemcomitans organisms is increased greatly and bacteria occur in clusters in KB cells treated with taxol.

Microtubule inhibitors, but not microfilament inhibitors, prevented the egression of A. actinomycetemcomitans from KB cells into the assay medium. Egression assays were carried out to determine the effects of microtubule and microfilament inhibitors and brefeldin A on the exit of SUNY 465 from KB cells. In this modification of the standard invasion assay, KB cells are not lysed after the gentamicin treatment step. Instead, fresh antibiotic-free medium is added and the KB cells and medium are monitored subsequently for bacteria. Time zero is defined as the time at which the fresh medium is added. Values for time zero are determined in standard invasion assays carried out concurrently, whereas 180-min values reflect the actual CFU recovered from cells at that time. The kinetics of accumulation of SUNY 465 in the culture medium in egression assays carried out in the presence of colchicine and taxol are shown in Fig. 3a. Between 0 and 180 min, A. actinomycetemcomitans SUNY 465 was recovered from the assay culture medium of control monolayers. By contrast, few bacteria were recovered in the assay medium of monolayers that had been treated with either colchicine or taxol. Roughly 45% of internalized (0 min) A. actinomycetemcomitans organisms were recovered at 180 min from cells treated with the microtubule inhibitors, but only 1 to 2% were recovered at this time from control cells (Fig. 3a). Neither cytochalasin D nor brefeldin A, a drug which interferes with normal organelle trafficking and enhances lysosome movement to the cell periphery, inhibited the egression of A. actinomycetemcomitans from KB cells; i.e., the number of A. actinomycetemcomitans organisms recovered from KB cells at 180 min was the same as that for the controls. As stated previously, the addition of cytochalasin D prior to infection results in inhibition of invasion of A. actinomycetemcomitans SUNY 465 (30). Thus, it was surprising to find that the addition of cytochalasin D after infection resulted in a transient increase in intracellular organisms, i.e., the zero time point (Fig. 3b).

View larger version (43K): [in this window] [in a new window]   FIG. 3.   Effects of inhibitors on the egression of A. actinomycetemcomitans SUNY 465 from KB cells. (a) Kinetics of A. actinomycetemcomitans accumulation into the assay medium. A. actinomycetemcomitans egresses from control cells and accumulates in the medium (). Bacteria do not accumulate in the assay medium if either taxol () or colchicine () is present. Bars represent additive values (i.e., the 60-min bar represents the 40 × 103 CFU recovered at 30 min plus 20 × 103 CFU recovered during the next 30 min, and so on) from a typical experiment carried out in quadruplicate. The standard deviation was less than 20% of the mean in all cases. The inset shows SUNY 465 in untreated and taxol-treated cells at time zero () and 180 min (). (b) SUNY 465 recovered from KB cells after treatment with cytochalasin D or brefeldin A. , control cells; , treated cells. KB cells were treated with brefeldin A prior to infection, whereas cytochalasin D was added 90 min after infection.

Host cell protrusions contained microtubules. A. actinomycetemcomitans organisms internalized within KB cells are frequently found at the periphery of the cell close to the cell membrane. Organisms are also often found in long protrusions that appear to be extensions of the host cell membrane that have been forced out by bacteria pushing against it (27). Protrusions either are associated with a single host cell (rudimentary protrusions) or interconnect adjacent cells (intercellular protrusions) and mediate the spread of A. actinomycetemcomitans from one epithelial cell to another (27). Immunofluorescence labeling of infected KB cells was carried out to analyze the cells for the distribution of microtubules and internalized bacteria. Typical interphase microtubule arrays with microtubules radiating out from the nucleus into the periphery of the cell in delicate lacelike threads were observed (Fig. 4a). Internalized SUNY 465 localized to the same region of the cytoplasm. Furthermore, microtubules also occurred in both cell-to-cell and rudimentary protrusions in which bacteria were also evident (Figs. 4b and c).

View larger version (32K): [in this window] [in a new window]   FIG. 4.   Immunofluorescence micrographs of KB cells 60 min after infection with A. actinomycetemcomitans SUNY 465. KB cell microtubules were labeled with mouse monoclonal -tubulin and TRITC-conjugated anti-mouse IgG. Bacteria were labeled with rabbit immune serum specific for SUNY 465 and secondary IgG conjugated to either FITC or TRITC. Bacteria were differentially stained by a double-labeling technique, and internal and external organisms were distinguished by comparing fields with red (TRITC) and green (FITC) fluores cence filters. Internalized organisms are visualized only with the FITC filter, whereas external organisms are visualized with both FITC and TRITC filters; i.e., if an organism is visualized with both filters, it is external. Microtubules are visualized only with the TRITC filter. The micrographs shown here were taken with a multiuse filter which allows the simultaneous visualization of both bacteria and microtubules. (a) Microtubules (white arrowheads) were dispersed throughout the cytoplasm of KB cells. Internalized A. actinomycetemcomitans (black arrows) localized to the same region of the cytoplasm. (b and c) Microtubules also occurred in both cell-to-cell (b) and rudimentary (c) protrusions. Bacteria in protrusions are indicated by the arrow in panel b and by arrowheads in panel c.

Only bacteria that are both invasive and egressive interacted with microtubules. A number of bacteria were tested for their ability to bind to taxol-induced microtubule asters in the KBE microtubule binding assay (Fig. 5). The A. actinomycetemcomitans invasion prototype strain, SUNY 465, bound specifically to the plus ends of the asters (Fig. 5a). Other organisms which can also invade and spread intercellularly also bound to the asters; A. actinomycetemcomitans 652 (Fig. 5b) and L. monocytogenes (Fig. 5c) both bound to the plus ends of microtubules in the asters, while S. flexneri (Fig. 5d) appeared to bind primarily to the minus ends. Neither H. aphrophilus (Fig. 5e), a noninvasive, nonpathogenic oral species closely related to A. actinomycetemcomitans, nor the noninvasive E. coli HB101 (Fig. 5f) bound to the asters. H. influenzae (Fig. 5g) and A. actinomycetemcomitans SUNY 523 (Fig. 5h), two organisms that invade but do not egress from host cells, did not interact with asters either. These results suggest that only bacteria which can invade and egress from host cells can interact with host cell microtubules. Table 1 contains a summary of these results. Untreated asters (Fig. 5i) and those to which bacteria bound strongly and specifically were well defined (e.g., A. actinomycetemcomitans SUNY 465). By contrast, the treatment of asters with bacteria that did not bind (e.g., A. actinomycetemcomitans SUNY 523, Haemophilus species, and E. coli) consistently resulted in fewer asters, and those that were present appeared to be disintegrating (Fig. 5).

View larger version (55K): [in this window] [in a new window]   FIG. 5.   Immunofluorescence micrographs of the interaction of bacteria with taxol-induced microtubule asters. A. actinomycetemcomitans SUNY 465 (a) and 652 (b) and L. monocytogenes (c), organisms that can invade and egress from host cells, all bound primarily to the plus ends of microtubules in asters. S. flexneri, another organism that can invade and egress, bound to the minus ends (d). The noninvasive organisms H. aphrophilus (e) and E. coli HB101 (f) did not bind. H. influenzae (g) and A. actinomycetemcomitans SUNY 523 (h), organisms that invade but do not egress from the host cell, did not bind to the asters either. (i) Asters not subjected to bacteria.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial entry into mammalian cells may involve microfilament-dependent mechanisms, microtubule-dependent mechanisms, or both. Once internalized, some bacteria usurp host cell actin to move within cells and to spread to adjacent cells (1, 34, 40, 44). However, intracellular movement that involves host cell microtubules has not been reported.

The present study provides evidence that A. actinomycetemcomitans interacts in a specific manner with host cell microtubules and that microtubules may be involved in its intracellular spread and egression from host cells. An increase in intracellular A. actinomycetemcomitans was reproducibly observed with colchicine, nocodazole, and taxol. Both taxol and colchicine were also effective in reducing the number of A. actinomycetemcomitans organisms egressing from KB cells into the medium. Why should taxol and colchicine, inhibitors which produce such opposite effects on microtubules, both prevent the egression of A. actinomycetemcomitans from host cells? Microtubules are long, stiff polymers of tubulin molecules that continually polymerize and depolymerize, a process tightly linked to their function, e.g., the transport of organelles and vesicles (18, 23). Most likely, the microtubule-mediated transport of A. actinomycetemcomitans would also require continual microtubule polymerization and depolymerization; thus, inhibitors of both processes would elicit the same result.

It is clear that microfilaments are involved in A. actinomycetemcomitans SUNY 465 entry into epithelial cells (2, 9, 10, 27, 28). Thus, the finding in this study that cytochalasin D treatment postinfection resulted in a transient increase in the number of intracellular bacteria was somewhat surprising. A possible explanation for this involves our recent demonstration that A. actinomycetemcomitans strains may utilize either actin-dependent or actin-independent mechanisms of invasion. Whereas the majority of strains appear to utilize the actin-dependent mode, a few utilize an actin-independent mode (2). It is possible that A. actinomycetemcomitans SUNY 465 utilizes primarily the actin-dependent pathway but that certain conditions can lead to utilization of the actin-independent mechanism. In support of this is the fact that there is always residual invasion of A. actinomycetemcomitans SUNY 465 after pretreatment of KB cells with cytochalasin D. An alternative explanation for increased numbers of intracellular bacteria following cytochalasin D treatment postinfection is that microfilaments may play some role in the egression process. Microfilaments may provide a structural framework within protrusions but may not actually be involved mechanistically in the movement and exit process. In this regard, infected host cells treated with cytochalasin D appear to have increased numbers of projections.

The finding that only bacteria that can both invade and egress from host cells bound to microtubules suggests that microtubules may be directly involved in the movement process and supports our hypothesis that microtubules play a major role in the intra- and intercellular movement of A. actinomycetemcomitans SUNY 465. Since the intracellular movement and spread of both Listeria and Shigella have been attributed to host cell microfilaments (1, 34, 35, 44), the interaction of these organisms with microtubules was somewhat surprising. However, it has been reported that the comet tails involved in Listeria movement have a tubulin-like component (3). Furthermore, there is increasing evidence that microtubules and microfilaments are strongly linked both structurally and functionally (17), suggesting highly mutualistic entities.

Microtubules extend throughout the cytoplasm of animal cells, where they provide tracks for the movement and intracellular positioning of membrane-bound organelles and vesicles (5, 18, 23). During mitosis they organize into spindle-shaped arrays responsible for the correct segregation of duplicated chromosomes (18, 21). The two ends (plus and minus) of a microtubule are different, and they polymerize at different rates; the end that elongates fastest is the plus end. Plus ends of microtubules are frequently located near the plasma membrane, whereas minus ends are usually embedded in a microtubule-organizing center. Microtubule-associated proteins bind to microtubules and act both to stabilize them against depolymerization and to mediate their interaction with other cell components. An important class of microtubule-associated proteins is the motor proteins (e.g., kinesins and dyneins) that move along microtubules transporting specific cargo, e.g., organelles and vesicles (18, 45). Dyneins move toward the minus ends of microtubules, whereas kinesins move toward the plus ends.

The specificity of the binding of the invasive and egressive strains, i.e., A. actinomycetemcomitans and L. monocytogenes binding to the plus ends of microtubules and S. flexneri binding to the minus ends, may indicate that the surfaces of these organisms have a kinesin- or dynein-like protein that mediates the bacterium-microtubule interaction. Our observation that bacteria that do not bind to asters appear to have a detrimental effect on aster stability may have some relevance in this regard. Bacterium-aster interaction could also be mediated by the binding of the bacteria to a specific kinesin or dynein or by the binding of the bacteria to cargo that is being transported. Both kinesins and dyneins exist in many forms, each of which carries a distinct cargo (45). Thus, rather than interacting with a specific motor protein, these organisms might interact with cargo (e.g., a vesicle) that interacts with a specific type of motor protein. The fact that A. actinomycetemcomitans SUNY 465 bound primarily to the plus ends of microtubule asters in the in vitro KBE model in this study, whereas we previously observed binding to microtubule-organizing centers in taxol-induced asters in intact cells (29), is not surprising. The association of A. actinomycetemcomitans with the microtubule-organizing center could represent the difference between the in vivo and in vitro situations or an earlier or different step in the movement process.

This study indicates that microtubules play a role in the intracellular movement and cell-to-cell spread of A. actinomycetemcomitans. It provides the first evidence that host cell dispersion of an intracellular pathogen may involve the usurption of microtubules and furthers our understanding of the interaction of this pathogen with oral epithelial cells. The study also indicates that microfilaments may play some role in the process. Given the intricate structural and functional relationships of these cytoskeletal entities (e.g., it was recently reported that microtubule- and actin-based transport motors can interact directly [17]), some role for microfilaments in the movement process would not be surprising. Future studies should seek to identify A. actinomycetemcomitans proteins and microtubule- and microfilament-associated factors (e.g., motor proteins) involved and to determine mechanisms by which this novel host cell spreading process is effected.